CO2/oxalate cathodes as safe and efficient alternatives in high energy density metal–air type rechargeable batteries

Károly Németh *ab and George Srajer a
aAdvanced Photon Source, Argonne National Laboratory, Argonne, Illinois 60439, USA. E-mail: nemeth@ANL.Gov
bPhysics Department, Illinois Institute of Technology, Chicago, Illinois 60616, USA

Received 2nd October 2013 , Accepted 15th November 2013

First published on 15th November 2013


Abstract

We present theoretical analysis on why and how rechargeable metal–air type batteries can be made significantly safer and more practical by utilizing CO2/oxalate conversions instead of O2/peroxide or O2/hydroxide ones, in the positive electrode. Metal–air batteries, such as the Li–air one, may have very large energy densities, comparable to that of gasoline, theoretically allowing for long range all-electric vehicles. There are, however, still significant challenges, especially related to the safety of their underlying chemistries, the robustness of their recharging and the need of supplying high purity O2 from air to the battery. We point out that the CO2/oxalate reversible electrochemical conversion is a viable alternative of the O2-based ones, allowing for similarly high energy density and almost identical voltage, while being much safer through the elimination of aggressive oxidant peroxides and the use of thermally stable, non-oxidative and environmentally benign oxalates instead.


1 Introduction

High energy density batteries are expected to revolutionize transportation and allow for long-range all-electric vehicles.1–5 The basis of exceptionally high gravimetric energy density metal–air batteries lies on one hand in the great free energy change of metal–O2 reactions, on the other hand in the fact that O2 is available from air and does not have to be carried on the vehicle. It has been pointed out in recent years that high energy density rechargeable Li–O2 batteries can be built when O2 is converted to peroxide ions, O2−2 during the discharge of the battery.6 In practice, such Li–O2/peroxide batteries also produce Li-superoxide, LiO2. Both of these discharge products are aggressive oxidants that are difficult to control and may lead to unwanted side reactions. They may oxidize the electrolyte and the porous carbon matrix often used as part of a gas-diffusion electrode in which the discharge product is deposited. A high concentration of peroxides deposited in a carbon or general combustible matrix may lead to thermal runaway reactions, even to explosive combustions, making such batteries a safety hazard. Other sources of hazards are posed e.g. by flammable electrolyte components and very reactive dendrites on the surface of bulk lithium electrodes. Many of these hazards can be avoided by appropriate materials and techniques, discussed e.g. in ref. 2. For example, all solid state Li–O2 batteries7 eliminate hazards related to dendrite formation and flammable electrolytes, however they continue to deposit peroxide in a mix of carbon and ceramic material still leaving potential for explosive combustion of carbon. Aqueous Li–O2 batteries8 produce lithium-hydroxide, LiOH, instead of Li2O2 in the cathode reaction O2 + 2H2O + 4e → 4OH. In these batteries the anode bulk lithium is immersed in an aprotic organic electrolyte separated by a solid Li-ion conducting ceramic membrane9 from the aqueous solution of LiOH. The aqueous Li–O2 battery has demonstrated only limited or energy inefficient rechargeability so far.2,10 Aqueous electrolytes are also disadvantageous as they allow for the development of explosive hydrogen gas when operated outside their safe voltages windows.11 For mechanical stability (e.g. avoiding punctuation) the Li-ion conducting membranes tend to be thick and heavy, their ionic conductivity is low.2

Rechargeable metal–air batteries also have to use high purity O2 to avoid the formation of large amounts of carbonates due to the reaction of peroxides and hydroxides with CO2 and that of nitrides in the case of non-aqueous metal–air batteries. Such reactions would lead to the elimination of practical rechargeability,12 though for a primary, non-rechargeable battery, the addition of large amounts of CO2 (≈20–80%) in the gas feed, as an assist to O2, leads to 2–3 fold increase of the realizable energy density of such primary batteries.13,14 The carbonates produced from a mix of O2 and CO2 feed require very large overpotentials during the charging, at least as long as no catalyst will be found to reduce this overpotential to a practical value.10 The generation of pure O2 from air, free of H2O, N2 and CO2, needed for metal–O2 batteries without an on-board O2-tank, represents a great problem in itself. Unfortunately, large amounts of O2 can only be stored in heavy tanks at high pressure, and the compression of O2 requires such a large amount of energy that it renders batteries with on-board O2-tanks impractical.2

Additional problems arise from the storage of the discharge products. Solid discharge products deposited in the porous positive gas diffusion electrodes clog the pores of the electrode and increase its internal electrical resistivity, thereby decreasing its energy storage capacity.2 Ideally, discharge products would be removed from the space between the electrodes and would be stored in a separate container, following the operating principle of flow batteries. The optimal solvents for the dissolution of peroxides should allow for sufficiently fast ion transport for fast discharge and charge, i.e. for high current densities. These solvents should also be safe, ideally non-flammable and non-reactive with peroxides and whiskers on anodes. The best solvents developed for metal–air battery electrolytes appear to be blends of ionic liquids (organic salts that are liquid at room temperature) and polar aprotic organic solvents.15 Even with such solvents, explosive combustion of the electrolytes cannot be ruled out when the concentration of peroxides becomes large.

2 Results and discussion

2.1 Thermodynamics and safety considerations: achieving high energy density and improved safety

The conversion of O2 to peroxides, such as Li2O2, instead of oxides or hydroxides, is necessary to achieve practical rechargeability in metal–O2 batteries.2,6

The above mentioned issues with metal–O2/peroxide batteries motivated us to seek alternative electrochemistries utilizing air-available species other than O2 allowing for safer and more practical high energy density batteries for long range all-electric vehicles. Besides O2, only CO2 is an air-available molecule with rich electrochemistry that has been studied extensively. The conversion of CO2 to oxalate ions, according to the reaction 2CO2 + 2e → C2O2−4, can be carried out with 100% selectivity and great efficiency using existing catalysts16 or properly chosen electrolytes/electrodes.14,24 All other major studied reduction products of CO2 would either involve hydrogen and would thus be less robust, or would lead to the evolution of poisonous carbon monoxide. While the deposition of aggressive oxidant peroxides in combustible battery materials, such as carbon-based electrodes and electrolytes, may lead to explosive combustions when peroxides are in high concentrations, there are no such problems with oxalates, as they are non-oxidative, thermally stable and environmentally benign species. For example, Li2C2O4 decomposes only at about 500 °C.25 Due to these features, oxalates are significantly safer discharge products than peroxides. Oxalates are also known to be easily and quantitatively oxidizable to CO2, in both aqueous and organic electrolytes on various types of electrodes, including graphite.14,16,24,26 Thus, reversible electrochemistries can be based on CO2/oxalate conversions.

Remarkably, with the application of the copper-complex catalyst from ref. 16, the CO2 reduction standard electrode potential becomes as high as U0(CO2(g)/C2O2−4) = −0.03 V and the back-oxidation happens at U0(C2O2−4/CO2(g)) = +0.81 V,16 leading to greatly reduced overpotentials and energy efficiency of the corresponding conversions. Coupled with a Li-anode (U0(Li(s)/Li+) = −3.04 V,17), this Li–CO2/oxalate battery would have an open circuit voltage of ≈3.0 V, which is practically identical with that of the Li–O2/peroxide battery.2 The identical voltages also imply close reaction Gibbs free energies (≈−575 kJ mol−1), as both processes transfer two electrons per molecule of product.

Tables 1 and 2 present the respective formation and reaction energies, while Table 3 lists energy densities and capacities of several Li–air type batteries in comparison with gasoline (n-octane) and a typical Li-ion battery. These data indicate that Li–CO2/oxalate batteries may compete with Li–O2/peroxide ones for applicability in long-range all-electric vehicles. Indeed, the gravimetric energy density of the Li2C2O4 formation is 11.5 kW h kg−1, slightly greater than that of Li2O2 (11.3 kW h kg−1), within 11% to n-octane (12.8 kW h kg−1), in reference to the weight of Li, assuming air based O2 or CO2 intake. When O2 or CO2 is carried on the vehicle, or in reference to the weight of the discharge products, the gravimetric energy density of Li2C2O4 is 1.6 kW h kg−1, about 2.2 times smaller than that of Li2O2, due to the larger weight of Li2C2O4, still about 3 times larger than that of a LiCoO2-based Li-ion battery. The practical energy densities of the oxalate and peroxide based batteries would differ significantly less, as they would also involve the mass of other battery components, such as electrolytes, membranes, current collectors, cases, etc. Furthermore, in case of the corresponding Na batteries (instead of Li ones), the weight-factor would be reduced to 1.7 from 2.2. The gravimetric and volumetric capacities (Li-densities) of Li2C2O4 are lower than those of Li2O2 by a factor of about 2.2. While the gravimetric capacity of Li2C2O4 is about twice as much as that of LiCoO2, the volumetric one is only about 80% of it.

Table 1 Standard enthalpies (ΔfH) and Gibbs free energies (ΔfG) of formation of molecules and crystals relevant for energy storage reactions discussed in the present study. Gas and solid phases are referred to as (g) and (s), respectively. The ΔfG value of Li2C2O4 (s) has been calculated from the ΔfG values of Li(s) and CO2(g) and the 3.0 V open circuit voltage of a Li–CO2/oxalate cell based on the standard electrode potential of U0(CO2(g)/C2O2−4) = −0.03 V from ref. 16 and the standard electrode potential of U0(Li(s)/Li+) = −3.04 V from ref. 17
Molecule/crystal ΔfH (kJ mol−1) ΔfH (kJ mol−1) Ref.
C8H18(g) −208.70 −347.95 18
O2(g) 0.00 −61.17 19
H2O(g) −241.83 −298.13 19
CO2(g) −393.52 −457.26 19
Li(s) 0.00 −8.67 19
Li2O(s) −598.73 −610.01 19
Li2O2(s) −632.62 −649.44 19
Li2CO3(s) −1216.04 −1242.96 19
Li2C2O4(s) −1377.21 −1510.61 16 and 20


Table 2 Standard reaction enthalpies (ΔrH) and Gibbs free energies (ΔrG) of energy storage reactions discussed in the present study, as calculated from ΔfH and ΔfH data of Table 1. Data of the LiCoO2-based Li-ion battery are from ref. 21–23
System Reaction ΔrH ΔrG
(kJ mol−1) (kJ mol−1)
n-Octane C8H18(g) + 12.5O2(g) → 8CO2(g) + 9H2O(g) −5116 −5259
Li-ion (LiCoO2) LixC6(s) + Li1−xCoO2(s) → C6(s) + LiCoO2(s) −213
Li–O2/oxide 2Li(s) + 1/2O2(g) → Li2O(s) −599 −562
Li–O2/peroxide 2Li(s) + O2(g) → Li2O2(s) −633 −571
Li–[O2 + CO2]/carbonate 2Li(s) + CO2(g) + 1/2O2(g) → Li2CO3(s) −823 −738
Li–O2/oxalate 2Li(s) + 2CO2(g) → Li2C2O4(s) −590 −579


Table 3 Theoretical gravimetric and volumetric energy (ΔrH) densities and capacities of energy storage reactions, as well as open circuit voltages (OCV), densities of products and rechargeabilities. ΔrH values have been taken from Table 2. Note that the OCV for the Li–CO2/oxalate cell is based on the standard electrode potential of U0(CO2(g)/C2O2−4) = −0.03 V from ref. 16 and on U0(Li(s)/Li+) = −3.04 V from ref. 17. OCV-s of the other cells are based on ref. 6 and 21–23. The OCV of the Li–(O2 + CO2)/carbonate cell is identical to that of the Li–O2/peroxide one, as of ref. 13, i.e. the addition of CO2 to a Li–O2/peroxide cell produces extra heat instead of electrical energy, while its rechargeability is debated.10,12,13 O2 or CO2 may be supplied from air or from a gas tank carried on the vehicle leading to different energy densities. Discharge capacities are referenced to bulk lithium and are identical for all Li–air type systems (3830 mA h kg−1 and 2045 mA h cm−3), while charge capacities are referenced to solid discharge products. Densities of solids are based on crystal structures at standard state
System OCV (V) Density of product (kg L−1) Energy density Charge capacity Rechargeability
Gravimetric (W h kg−1) Volumetric (W h L−1) Gravimetric (mA h g−1) Volumetric (mA h cm−3)
(Air) (Tank) (Air) (Tank)
n-Octane 12[thin space (1/6-em)]814 9008 N
Li-ion (LiCoO2) 3.6 5.05 568 2868 273 1379 Y
Li–O2/oxide 2.9 2.02 11[thin space (1/6-em)]151 5204 5955 10[thin space (1/6-em)]512 1787 3610 N
Li–O2/peroxide 3.0 2.25 11[thin space (1/6-em)]329 3448 6050 7758 1165 2621 Y
Li–[O2 + CO2]/carbonate 3.0 2.10 11[thin space (1/6-em)]329 + 3313 2143 + 627 6049 +1769 4500 + 1316 724 1520 Y/N
Li–CO2/oxalate 3.0 2.14 11[thin space (1/6-em)]488 1577 6134 3375 525 1125 Y


The energy and capacity density values of Li2C2O4 are quite practical to allow for long range electric vehicles. A car that takes up 13 US (liquid) gallons (about 49 L) of n-octane, stores about 88.6 kW h energy for useful work, assuming 20% engine efficiency. The equivalent of this energy would be released by the production of 7.7 gallons of Li2C2O4 at a rather poor electric motor efficiency of 90%. The amount of oxalates needed may further be reduced with greater motor efficiency (96% as of ref. 27), electrically storing energy from braking or due to reduced idle time. Solvents for (partial) dissolution of the oxalates (or peroxides) may require additional space though. The weight of 7.7 gallons of Li2C2O4 is 62 kg, only about 27 kg more than that of 13 gallons of gasoline (35 kg). The increase in the weight of the vehicle by the battery may be reduced by lighter, carbon-composite based cars, and by electric motors lighter than internal combustion engines, to maintain the same driving distance and approximately the same volume for energy storage as known for present day gasoline driven cars.

The energy storage efficiency of the Li–CO2/oxalate battery can be estimated from the voltage ratio of the discharge and charge processes at constant current. Using the voltages of 3.0 and 3.8 V for discharge and charge, respectively, associated with the copper-complex catalyst of ref. 16, the energy storage efficiency of the Li–CO2/oxalate battery is estimated to be 79%. In practice, this value would be significantly lower but likely higher than that of Li–O2/peroxide batteries, 65%, due to considerable overpotentials on charge in the latter ones.4

In principle, a 1[thin space (1/6-em)]:[thin space (1/6-em)]1 molar mix of O2 and CO2 could lead to a further increased energy density in Li–[O2 + CO2]/carbonate batteries, however, in practice either the O2 or the CO2 gets reduced, leaving electrically utilizable energy densities (per mol of product) at the level of Li–O2/peroxide or Li–CO2/oxalate batteries. When O2 gets reduced, the presence of CO2 leads to carbonate formation, producing a lot of additional heat, formally through the reaction of Li2O with CO2.13 The voltage of such a Li–[O2 + CO2]/carbonate battery was found identical to that of the Li–O2/peroxide one.13 The extra heat is disadvantageous as it will require additional cooling of the battery during discharge. It also indicates that a large amount of energy, one-third of that of the electrical energy stored, is needed to free up CO2 from carbonates on charge. These facts may render all metal–O2/peroxide batteries inefficient for rechargeable use, unless CO2-free O2-intake is applied.12 The current best rechargeable Li/ambient-air batteries produce mostly Li2CO3 during discharge10 making the recharging process highly energy inefficient.

Interestingly, in some electrolytes based on ionic liquids, it is the CO2 that gets reduced to oxalate and O2 does not formally participate in reactions.14 Selective reduction of CO2 to oxalate from ambient air through a catalyst16 and simultaneous reduction of O2 to peroxide is also possible, potentially allowing for a rechargeable Li–air battery by eliminating the presence of carbonates.

Pure O2 or CO2 could also be supplied from a tank stored on board of the vehicle. At a relatively safe and practical 120 bar pressure and 30 °C temperature, the density of CO2 is 0.802 kg L−1, while that of O2 is only 0.160 kg L−1. With good thermal insulation, potentially even solid CO2, ‘dry ice’, could be stored on board, with a high density of 1.6 kg L−1. For the above mentioned 88.6 kW h useful energy, the space required for CO2 storage would be between 8.8 and 18 gallons, while that of O2 would be at 32 gallons. The compression of O2 requires far more work though than that of CO2. This is obvious also from the economical availability of ‘dry ice’, or the use of CO2 as working fluid in air conditioning in both vehicles and buildings.28 Even though the work invested in the compression of these gases may be returned to some extent when a higher pressure gas is applied on the respective electrodes,2 CO2 appears far more advantageous for on-board storage than O2, for its far better compressibility. CO2 can be collected efficiently from air as well, via CO2-sponge materials, based on economically available ion-exchange resins29,30 allowing for CO2 intake from air at rates of 0.25–0.83 g CO2 m−2 s−1 (ref. 29) (up to 3.0 kg CO2 m−2 h−1) despite the low, ≈0.04 mol% concentration of CO2 in air.

On charge, oxalate salts from metal–CO2/oxalate batteries will be converted to CO2 that will not react with other battery components, such as the electrolyte or the porous carbon electrode, at the charging electrode potential. Peroxide salts from metal–O2/peroxide batteries will, however, be converted to highly reactive oxygen species such as singlet oxygen31 that may react with the electrolyte and the porous carbon or other combustible battery components. While materials and techniques are being explored to avoid such unwanted side reactions during the charging of a metal–O2/peroxide battery,2 metal–CO2/oxalate batteries completely eliminate this problem through the robustness of quantitative oxidation of oxalates to CO2.16,26

Fig. 1 shows one possible implementation of a simple Li–CO2/oxalate battery.32 The porous positive gas-diffusion electrode contains catalysts, such as the copper-complex16 mentioned above, for the reversible conversion of CO2 to oxalate ions. During discharge, CO2 is converted to oxalate ions that migrate through an anion-exchange membrane33 into the central compartment of the cell where they mix with Li+ ions released from a bulk Li-anode, while electrons move from the anode to the cathode. In this implementation, the aprotic electrolyte in the central compartment may be composed of a blend of an ionic liquid and a polar organic solvent, similarly to that in ref. 15. The negative ions of the ionic liquid must exclusively be oxalate ions, so that no other negative ions get oxidized in the positive electrode on charge. The positive ions of the ionic liquid may be quaternary ammonium ions, such as pyrrolidinium, imidazonium or tetrabutylammonium ones. Such ionic liquids have already been explored to some extent.34,35 The polar organic solvent may be a mix of propylene carbonate and dimethoxyethane (monoglyme), similarly to electrolytes in Li–O2/peroxide batteries, its main role is to dissolve the organic oxalate ionic liquid and decrease its viscosity. This electrolyte may also be used in the porous positive electrode to assist the transport of oxalate ions from/to the catalyst. Anion exchange membranes are used instead of cation ones to avoid the deposition of discharge products in the positive electrode and thus minimize the amount of catalysts needed. The rate of oxalate ion transport in the cathode, through the membrane and in the central compartment determines the current density of the battery (rate of charge/discharge capacity) and should be subject of optimization. On charge, oxalate ions migrate to the positive electrode where they get oxidized back to CO2, while Li+ ions get reduced and deposited on the negative electrode and electrons move from the positive electrode to the negative one through the outer circuit. Many variants of the above described simple implementation of a Li–CO2/oxalate battery are possible and some are discussed to a larger extent in ref. 32.


image file: c3ra45528a-f1.tif
Fig. 1 Schematic view of one possible implementation of a Li–CO2/oxalate battery. The positive electrode is made of a porous electrically conducting material (brown color), which contains catalyst (green color) for reversible conversion of CO2 to oxalate. During discharge (panel A), oxalate ions (C2O2−4) are produced from CO2 diffusing in the cathode from outside the cell. The oxalate ions migrate through the anion exchange membrane (yellow color) and mix with Li+ ions released from the Li-anode in the aprotic electrolyte (blue color) that may be composed of a suitable ionic liquid. During the charging process (panel B) Li is deposited in the negative electrode, while oxalate ions are converted back to CO2 in the positive electrode and leave the cell.

2.2 Kinetics considerations: optimization of the power density

In order to maximize the power density, i.e. the rate of charging/discharging the battery, the kinetic processes involved in the electron transfers and ion transports have to be made quick. The application of CO2 instead of O2 offers several specific advantages in this respect as well.

In case CO2 is taken from a tank and is supplied as a moderately high pressure (10–140 bar) gas or supercritical fluid into the battery, it can serve both as a solvent and a source of electroactive species in the same time. Fundamental aspects of electrochemistry in supercritical CO2 have been investigated by Abbott et al.36,37 The use of high pressure or supercritical CO2 appears attractive for achieving very fast ion transport and thus high charge/discharge rates, as diffusion may be very fast in such high-pressure or supercritical medium when the concentration of voids is large enough.38 As liquid or fluid CO2 has low solubility for salts, a polar modifier needs to be added, such as 1,1,1,2-tetrafluoroethane (HFC 134a),37 propylene carbonate39,40 or glymes.41 Both propylene carbonate and glymes are excellent solvents (or solutes) for (or in) pressurized CO2 and mix well with supercritical CO2 fluid. Propylene carbonate at room temperature and 10 to 55 bar pressure can dissolve 10 to 60 mol% CO2, respectively,39 while diglyme at 40 °C and 10 to 71 bar pressure dissolves 20 to 85 mol% CO2, respectively.41 Note that the critical point of pure CO2 is at Tc = 31 °C and pc = 74 bar. The compression of the pressurized/supercritical electrolyte consumes a fraction of the energy stored in the battery, however, this fraction can be kept small (<3%) for isothermal compression and even smaller when electrochemical compression is used, as discussed by Christensen et al. in ref. 2 for Li–O2 batteries.

In addition to the polar modifier, supporting electrolyte is also needed to increase the ionic conductivity of the electrolyte, at least until it is saturated with discharge products. The supporting electrolyte is typically composed of organic salts (ionic liquids), such as tetrabutylammonium tetrafluoroborate,37 or imidazolium salts40 and may be composed of the oxalate ion containing organic salts mentioned above, as well. Additionally, many organic salts show anomalous melting point depression in pressurized CO2 (ref. 42 and 43) which in principle allows such pressurized room temperature ionic liquids to be used as electrolytes in themselves, i.e. without the glymes or organic carbonates or other polar modifiers while also maintaining increased concentration of voids for faster ion transport.

A flow battery based on pressurized/supercritical CO2 is depicted in Fig. 2. It is composed of a negative electrode with protected metal (Li, Na, Mg, Al, etc.) source, a positive electrode current collector with a catalyst for CO2 reduction and oxalate oxidation covering its surface, and a flow of CO2-rich solution or supercritical fluid between the electrodes. As the flow passes between the electrodes it becomes enriched in oxalate salts during discharge. The oxalate salts will be deposited in a container where the pressure may decrease due to increased volume and the CO2 and the polar modifier or solvents will be recompressed and circulated back to between the electrodes. In case a non-compressible solvent is used, it may be pumped through a filter leaving the excess discharge products in the product container. On charge, the flow of the supercritical fluid or that of the pressurized solution will be reversed, oxalate ions will be oxidized back to CO2, metal ions reduced and the regenerated CO2 will be stored back in the CO2 tank. The charge/discharge rates of such a battery are expected to primarily depend on the catalyst density on the surface of the positive electrode, as the concentration of the CO2 is very high and transport of electroactive species to and from the catalyst is expected to be very fast due to the directed flow of the voids-rich fluid or liquid. The rates also depend on the ionic conductivity of the protective skin (or cation selective membrane) on the negative electrode. Highest discharge rates are expected for supercritical fluid electrolytes for the greatest concentration of voids supporting ion transport.


image file: c3ra45528a-f2.tif
Fig. 2 Schematic view of a rechargeable Li–CO2/oxalate flow battery. The purple stripe refers to a Li-ion selective membrane, the orange one to the catalytic CO2/oxalate electrode. The arrows indicate the flow of the electrolyte during charge and discharge. For maximum charge/discharge rates a pressurized solution or supercritical fluid electrolyte is applied consisting of high pressure (10–140 bar) mixture of CO2, polar modifier (propylene carbonate, glyme, etc.) and supporting electrolyte (organic salt, ionic liquid). As the electrolyte passes between the protected Li anode and the catalytic CO2/oxalate cathode, it dissolves the discharge product Li2C2O4 and deposits it in the product container either through locally expanding volume and reduced pressure, or through a filter. The flow of the electrolytic fluid is recompressed after the deposition of Li2C2O4 and circulated back toward the electrodes. On charge, the flow of the electrolyte is reversed, Li2C2O4 is dissolved in the electrolyte in the product container and is converted back to Li and CO2 on the electrodes. The increased pressure causes CO2 to migrate back to the CO2 container through a CO2 selective membrane and valve. A similar flow battery may be realized at normal pressure as well, utilizing deep eutectic solvents as electrolytes based on organic and inorganic oxalates, as discussed in the text. The CO2 source unit may be a CO2 tank (for high pressure applications) or an atmospheric CO2 absorbing unit (for normal pressure applications).

In case CO2 is not supplied from a tank but taken up from the air through a CO2 absorbing material mentioned above, a similar flow battery scheme can be utilized, with a circulated electrolyte at atmospheric pressure. In order to maximize the speed of transport of oxalate ions, the use of deep eutectic solutions is suggested as primary component of the electrolyte, as it opens the way to hopping based fast oxalate ion transport as opposed to slow simple diffusion. It is expected that the mix of oxalate ion containing solid organic salts (such as those mentioned above with pyrrolidinium, imidazonium or tetrabutylammonium cations) and the discharge product metal-oxalates forms room temperature liquids based on principles of deep eutectic mixtures. Examples of such deep eutectic liquids with oxalate salts exist, and provide alternatives to ionic liquids as solvents. For example, the mixture of choline chloride and oxalic acid forms a room temperature liquid electrolyte, even though the individual components are solids themselves. Electrochemistry in such deep eutectic solvents with carboxylic acids has been studied by Abbott et al. and by LeSuer et al.38,44 The ionic conductivity of such deep eutectic liquids was found to be similar to that of ionic liquids.38 In the proposed mix of organic oxalates and metallic oxalates it is expected that the oxalate ions would chelate the metal cations and oxalate transport would happen through hopping of oxalate ions between neighboring chelating sites, instead of diffusion. Since the concentration of the chelating sites is very high in the proposed deep eutectic liquids, a fast oxalate ion transport is expected. Note, that only the transport of oxalate ions is expected to be fast in such liquids, other anions would be transported much slower. Also note that the previously studied choline chloride and oxalic acid mixture does not allow for hopping based oxalate transport as it does not contain metal cations as chelating sites, thus the proposed hopping based oxalate ion transport is yet to be investigated experimentally. The deep eutectic solvent may be diluted by the addition of polar aprotic solvents, such as propylene carbonate and dimethoxyethane, to decrease its viscosity, especially for the fully charged state where the metal oxalate component may be missing as is yet to be produced during the discharge. When the deep eutectic liquid becomes saturated with the dissolved metal oxalates, the precipitating excess oxalates would be deposited in the discharge product container. On charge, the oxalates in the product container would be dissolved in the deep eutectic electrolyte and brought to the electrodes where metal and CO2 would be generated as the battery is being charged. Also note that the cell in Fig. 1 can also be used in a flow battery design, in this case the CO2 is supplied through a gas diffusion electrode separated by an oxalate ion conducting membrane from the electrolyte. The above described flow battery designs also have the advantage of minimizing the amount of the electrolyte and other additional components relative to the reactants and the products. Many variants of the CO2/oxalate battery are possible utilizing the principles described here.

A primary battery with Na anode and CO2/oxalate cathode has recently been realized by Das et al.14 as an unexpected by-product of Na–(O2 + CO2) battery experiments, observing that oxygen admixture to CO2 can act as a catalyst for the production of oxalate salts during discharge in appropriately chosen ionic liquid electrolytes while no oxide or carbonate discharge products are formed. The voltage of this battery (≈2.25 V, OCV) is about 18% smaller than that of an analogous Na–(O2 + CO2)/(peroxide + carbonate + oxalate) battery presented in the same work (≈2.75 V, OCV, see Fig. 1 of ref. 14) or that of a Na–O2/(peroxide) battery (2.6 V, theoretical OCV). This experimental result indicates the feasibility of metal–air type batteries with CO2/oxalate cathodes, and also indicates that oxalate formation may be similarly energetic as peroxide formation when proper catalysts and electrolytes are used. This observation is in agreement with what we described above about the thermodynamics of CO2/oxalate conversion using experimental data with the copper complex catalyst by Angamuthu et al.16 and the measured enthalpies and Gibbs free energies of formation of Li2C2O4.

3 Conclusions

We have proposed to use CO2/oxalate electrodes instead of O2/peroxide or O2/hydroxide ones in high energy density rechargeable metal–air type batteries. The advantages of CO2/oxalate electrodes may be realized in terms of significantly improved safety and environmental friendliness, high energy and power density, robust and more efficient rechargeability and efficient on-board storage or air-based intake of CO2.

Acknowledgements

K. N. gratefully acknowledges helpful discussions with Drs M. Balasubramanian, G. Crabtree, N. Markovic, L. Trahey, and M. van Veenendaal at Argonne National Laboratory and A. K. Unni, C. U. Segre and L. Shaw at IIT. This research was supported by the U.S. DOE Office of Science, under contract no. DE-AC02-06CH11357.

References

  1. M. Armand and J.-M. Tarascon, Nature, 2008, 451, 652–657 CrossRef CAS PubMed.
  2. J. Christensen, P. Albertus, R. S. Sanchez-Carrera, T. Lohmann, B. Kozinsky, R. Liedtke, J. Ahmed and A. Kojic, J. Electrochem. Soc., 2012, 159, R1–R30 CrossRef CAS PubMed.
  3. G. Girishkumar, B. McCloskey, A. C. Luntz, S. Swanson and W. Wilcke, J. Phys. Chem. Lett., 2010, 1, 2193–2203 CrossRef CAS.
  4. A. Kraytsberg and Y. Ein-Eli, J. Power Sources, 2011, 196, 886–893 CrossRef CAS PubMed.
  5. F. Li, T. Zhang and H. Zhou, Energy Environ. Sci., 2013, 6, 1125 CAS.
  6. T. Ogasawara, A. Débart, M. Holzapfel, P. Novák and P. G. Bruce, J. Am. Chem. Soc., 2006, 128, 1390 CrossRef CAS PubMed.
  7. B. Kumar, J. Kumar, R. Leese, J. P. Fellner, S. J. Rodriques and K. M. Abraham, J. Electrochem. Soc., 2010, 157, A50–A54 CrossRef CAS PubMed.
  8. S. J. Visco and Y. S. Nimon, Active Metal/Aqueous Electrochemical Cells and Systems, Patent, US7666233, 2010.
  9. J. Fu, Solid State Ionics, 1997, 96, 195–200 CrossRef CAS.
  10. T. Zhang and H. Zhou, Nat. Commun., 2013, 4, 1817 CrossRef PubMed.
  11. C. Wessells, R. Ruffo, R. A. Huggins and Y. Cui, Electrochem. Solid-State Lett., 2010, 13, A59–A61 CrossRef CAS PubMed.
  12. S. R. Gowda, A. Brunet, G. M. Wallraff and B. D. McCloskey, J. Phys. Chem. Lett., 2013, 4, 276–279 CrossRef CAS.
  13. K. Takechi, T. Shiga and T. Asaoka, Chem. Commun., 2011, 47, 3463 RSC.
  14. S. K. Das, S. Xu and L. A. Archer, Electrochem. Commun., 2013, 27, 59–62 CrossRef CAS PubMed.
  15. B. G. Kim, J.-N. Lee, D. J. Lee, J.-K. Park and J. W. Choi, ChemSusChem, 2013, 6, 443–448 CrossRef CAS PubMed.
  16. R. Angamuthu, P. Byers, M. Lutz, A. L. Spek and E. Bouwman, Science, 2010, 327, 313–315 CrossRef CAS PubMed.
  17. W. M. Haynes, CRC Handbook of Chemistry and Physics 2011–2012, Taylor & Francis, 2011 Search PubMed.
  18. W. D. Good, J. Chem. Thermodyn., 1972, 4, 709–714 CrossRef CAS.
  19. NIST-JANAF Themochemical Tables, ed. M. W. Chase Jr, American Chemical Society and the American Institute of Physics, Washington DC, 4th edn, 1998, pp. 1–1951 Search PubMed.
  20. B. B. Letson and A. R. Taylor, Heats of formation of lithium chloride and lithium oxalate, including details on the construction and operation of a solution calorimeter, U.S. Dept. of the Interior, Bureau of Mines, 1965, vol. 6583 Search PubMed.
  21. K. Mizushima, P. Jones, P. Wiseman and J. Goodenough, Mater. Res. Bull., 1980, 15, 783 CrossRef CAS.
  22. R. Yazami and P. Touzain, A reversible graphite-lithium negative electrode for electrochemical generators, J. Power Sources, 1983, 9, 365 Search PubMed.
  23. M. M. Thackeray, et al., Lithium-Oxygen (Air) Electrochemical Cells and Batteries, Patent, US8313721, 2012.
  24. L. Skarlos, Preparation of Oxalic Acid, Patent, US3720591, 1973.
  25. D. Dollimore and D. Tinsley, J. Chem. Soc. A, 1971, 3043–3047 RSC.
  26. B. Šlukić, R. Baron and R. G. Compton, Electroanalysis, 2007, 19, 918–922 CrossRef.
  27. T. Oikawa, T. Tajima, K. Masumoto, H. Akita and H. Kawaguchi, International Compressor Engineering Conference, 2002, pp. C12–4.
  28. B. Dienhart, et al., Optimized CO2 Operated Air Conditioning System, Patent, US6588223, 2003.
  29. K. S. Lackner, Eur. Phys. J.: Spec. Top., 2009, 176, 93–106 CrossRef.
  30. K. S. Lackner and A. Wright, Laminar Scrubber Apparatus For Capturing Carbon Dioxide From Air And Methods Of Use, Patent, US7833328, 2010.
  31. J. Hassoun, F. Croce, M. Armand and B. Scrosati, Angew. Chem., 2011, 50, 2999–3002 CAS.
  32. K. Nemeth, M. van Veenendaal and G. Srajer, Electrochemical energy storage device based on carbon dioxide as electroactive species, Patent, US8389178, 2013.
  33. J. M. Bernatowicz, et al., Methods and apparatus for the formation of heterogenious ion exchange membranes, Patent, US6716888, 2004.
  34. A. Rachocki, K. Pogorzelec-Glaser, A. Pietraszko and J. Tritt-Goc, J. Phys.: Condens. Matter, 2008, 20, 505101 CrossRef.
  35. A. G. Pinkus and R. Hariharan, Poly(methylene-oxalate) and precursors thereto, Patent, US3720591, 1993.
  36. A. P. Abbott and J. C. Harper, J. Chem. Soc., Faraday Trans., 1996, 92, 3895–3898 RSC.
  37. A. P. Abbott and C. A. Eardley, J. Phys. Chem. B, 2000, 104, 775–779 CrossRef CAS.
  38. A. P. Abbott, D. Boothby, G. Capper, D. L. Davies and R. K. Rasheed, J. Am. Chem. Soc., 2004, 126, 9142–9147 CrossRef CAS PubMed.
  39. P. D. Mantor, O. Abib, Jr, K. Y. Song and R. Kobayashi, J. Chem. Eng. Data, 1982, 27, 243–245 CrossRef CAS.
  40. H. Kawanami, A. Sasaki, K. Matsuia and Y. Ikushima, Chem. Commun., 2003, 896–897 RSC.
  41. D. Kodama, M. Kanakubob, M. Kokuboa, S. Hashimotoa, H. Nanjob and M. Katoa, Fluid Phase Equilib., 2011, 302, 103–108 CrossRef CAS PubMed.
  42. A. M. Scurto and W. Leitner, Chem. Commun., 2006, 3681–3683 RSC.
  43. A. M. Scurto, E. Newton, R. R. Weikel, L. Draucker, J. Hallett, C. L. Liotta, W. Leitner and C. A. Eckert, Ind. Eng. Chem. Res., 2008, 47, 493–501 CrossRef CAS.
  44. C. A. Nkuku and R. J. LeSuer, J. Phys. Chem. B, 2007, 111, 13271–13277 CrossRef CAS PubMed.

This journal is © The Royal Society of Chemistry 2014
Click here to see how this site uses Cookies. View our privacy policy here.