Increased Stability Toward Oxygen Reduction Products for Lithium-Air

Nov 14, 2011 - Lithium-Air Batteries with Oligoether-Functionalized Silane. Electrolytes ... Chemical Sciences and Engineering Division,. ‡. Materia...
0 downloads 0 Views 915KB Size
Download PDF
ARTICLE pubs.acs.org/JPCC

Increased Stability Toward Oxygen Reduction Products for Lithium-Air Batteries with Oligoether-Functionalized Silane Electrolytes Zhengcheng Zhang,† Jun Lu,† Rajeev S. Assary,‡,^ Peng Du,† Hsien-Hau Wang,‡ Yang-Kook Sun,|| Yan Qin,† Kah Chun Lau,‡ Jeffrey Greeley,§ Paul C. Redfern,† Hakim Iddir,‡ Larry A. Curtiss,*,‡,§ and Khalil Amine*,† Chemical Sciences and Engineering Division, ‡Materials Science Division, §Center for Nanoscale Materials, Argonne National Laboratory, Argonne, Illinois 60439, United States ^ Chemical and Biological Engineering, Northwestern University, Evanston, Illinois 60208, United States Center for Information and Communication Material, Department of Chemical Engineering, Hanyang University, Seoul 133-791, South Korea

)

†

bS Supporting Information ABSTRACT: The successful development of Li-air batteries would significantly increase the possibility of extending the range of electric vehicles. There is much evidence that typical organic carbonate based electrolytes used in lithium ion batteries form lithium carbonates from reaction with oxygen reduction products during discharge in lithium-air cells so more stable electrolytes need to be found. This combined experimental and computational study of an electrolyte based on a tri(ethylene glycol)-substituted trimethylsilane (1NM3) provides evidence that the ethers are more stable toward oxygen reduction discharge species. X-ray photoelectron spectroscopy (XPS) and FTIR experiments show that only lithium oxides and no carbonates are formed when 1NM3 electrolyte is used. In contrast XPS shows that propylene carbonate (PC) in the same cell configuration decomposes to form lithium carbonates during discharge. Density functional calculations of probable decomposition reaction pathways involving solvated oxygen reduction species confirm that oligoether substituted silanes, as well as other ethers, are more stable to the oxygen reduction products than propylene carbonate. These results indicate that the choice of electrolyte plays a key role in the performance of Li-air batteries.

1. INTRODUCTION While today’s lithium ion batteries may provide acceptable power for hybrid electric vehicles, they do not provide enough energy for electric vehicles for long-distance driving. This range limitation and the absence of a battery charging infrastructure has limited public interest in electric vehicles. Lithium-air batteries offer, in principle, a much greater theoretical gravimetric energy density compared to conventional lithium-ion batteries.1,2 The inherent energy potential of lithium-air can approach that of gasoline, but still significant challenges remain to be overcome. A breakthrough in Li-air battery technology would significantly increase the possibility of extending the electric range of these vehicles with the additional advantage of reduced battery cost. Li-air batteries are a class of electrochemical cells comprising a lithium anode in combination with an oxygen cathode that typically operates in an ambient air environment.3 6 Nonaqueous Li-air cells provide a potential of ∼3 V and therefore a significantly higher specific energy than other metal-air cells such as Zn and Al. Because oxygen is supplied as a fuel to the cell r 2011 American Chemical Society

during discharge, Li-air cells differ from conventional battery systems; they might be construed, rather, as being hybrid batteryfuel cell systems. During electrochemical discharge, the lithium anode is oxidized by releasing an electron to the external circuit to produce lithium ions in the electrolyte, whereas the oxygen is reduced at a cathode surface to form lithium peroxide or lithium oxide which then can be recharged internally.7 There are many challenges facing the successful development of Li-air batteries, including new materials for the anode, cathode, and electrolyte, as well as in the engineering of the cells. Most of the work has been focused on developing new catalyst and cathode structures with limited work carried out on understanding the role of the electrolyte on cell performance.7 9 There has been recent experimental and theoretical evidence that the organic carbonates (e.g., propylene carbonate, ethylene Received: September 9, 2011 Revised: November 7, 2011 Published: November 14, 2011 25535

dx.doi.org/10.1021/jp2087412 | J. Phys. Chem. C 2011, 115, 25535–25542

The Journal of Physical Chemistry C carbonate, and dimethyl carbonate) commonly used in Li-ion batteries10,11 are not stable against oxygen reduction products formed during discharge.2,12 18 It has been reported that, during discharge, in addition to formation of lithium peroxide, propylene carbonate is decomposed, resulting in the formation of lithium carbonate and other species.15,16 As a result of the problems with propylene carbonate as an electrolyte, there have been recent studies of other electrolytes for Li-air batteries, including ether-based ones.7,19 22 McCloskey et al19 have reported an investigation of dimethoxyethane (DME), which contains two ether groups, and found no evidence of decomposition on discharge, indicating that it is stable to oxygen reduction products. They did find, however, that it was oxidatively unstable on charge around 4.7 V and not reversible. In contrast, Fruenberger et al22 reported an investigation into the stability of n-glymes (n = 2 4) in Li-air cells and found evidence that they were unstable during discharge. They also proposed a reaction mechanism based on hydrogen radical abstraction from a CH2 group by superoxide radical anion. In this work, we have carried out a combined experimental/computational study of the stability during discharge of electrolytes based on propylene carbonate and a tri(ethylene glycol)-substituted trimethylsilane (1NM3) in lithium air batteries. The experimental studies are based on X-ray photoelectron spectroscopy (XPS) and Fourier transform infrared (FTIR) studies of charge and discharge products. Density functional studies have been carried out on several possible decomposition mechanisms for the two solvents. The results confirm the instability of propylene carbonate, while indicating that the oligoether substituted silane is stable during discharge. In section II the experimental and theoretical methods are described. In section III the results are presented and discussed.

2. EXPERIMENTAL SECTION Electrochemical measurements were carried out in plastic Swagelok cells composed of a Li metal anode, electrolyte 1.0 M LiPF6 in 1NM3 or 1.0 M LiPF6 in propylene carbonate (Mitsubishi Chemical) impregnated into a glass fiber separator (Fisher Scientific) and a porous (11 mm diameter) air cathode fabricated by casting a mixture of Super P Li carbon black, the appropriate amount of electrolytic manganese dioxide (EMD) as catalyst, and a polymer based on poly(vinylidene fluoride) as binder in weight ratio 70:20:10. The Swagelok cells were sealed in a glass chamber filled with 1 atm of high purity oxygen and located within a chamber which was thermostatted at 298 K. The first discharge test was carried out by using 0.1 mA current with cutoff voltage of 2.0 V; the charge test was performed by using 0.1 mA current and was terminated when the capacity reached the same capacity obtained from the previous discharge process or 4.5 V, whichever comes earlier. All the electrochemical measurements were carried out using a Maccor cycler. All XPS measurements were made using a KratosTM Axis Ultra DLD surface analysis instrument. The base pressure of the analysis chamber during these experiments was 3  10 10 Torr, with operating pressures around 1  10 9 Torr. Spectra were collected using a monochromatic Al Kα source (1486.7 eV) and a 300  700 μm spot size. The Al source was operated at 13 mA of emission current with the target anode set to 15 kV, for a resulting power of 195 W. For survey spectra the data were collected using a pass energy of 160 eV (fixed analyzer transmission mode), a step size of 1 eV, and a dwell time of 200 ms. Highresolution regional spectra were collected using a pass energy of

ARTICLE

Figure 1. First charge and discharge cycles of a Li-air cell with propylene carbonate (PC) and 1NM3.

20 eV (fixed analyzer transmission mode), a step size of 0.1 eV, and a dwell time of 300 ms. For low signal-to-noise regions, multiple passes were made and the results averaged together as noted. Prior to introduction into the load lock vacuum chamber of the instrument, all air-sensitive samples were loaded into an inert transfer module, which interfaces with the instrument. Samples were prepared for analysis in an Ar-filled glovebox, with no more than 1 ppm O2 and 1 ppm H2O. Nonconductive samples showed evidence of differential charging, resulting in peak shifts and broadening. Photoelectron peak positions were shifted back toward their true values, and their peak widths were minimized by flooding the samples with low-energy electrons and ions from the charge neutralizer system on the instrument. Further peak position correction was made by referencing the C 1s peak position of adventitious carbon for a respective sample (284.8 eV, PHI Handbook) and shifting all other peaks in the spectrum accordingly. Fitting was done using the program XPS Peak. Each relevant spectrum was fit to a Shirley-type background to correct for the rising edge of backscattered electrons that shifts the baseline higher at high binding energies. The end points of the background region were averaged over the 7 9 nearest points to increase the accuracy. Peaks were fit as asymmetric Guassian/Lorentzians, with 0 30% Lorentzian character. The fwhm of all subpeaks were constrained to 0.7 2 eV, as dictated by instrumental parameters, lifetime broadening factors, and broadening due to sample charging. With this native resolution set, peaks were added, and the best fit, using a least-squares type fitting routine, was obtained while adhering to the constraints mentioned above.

3. RESULTS AND DISCUSSION 3.1. Theoretical. We have employed B3LYP23 density functional

theory (DFT) for the calculations of the stability of the electrolytes toward solvated oxygen reduction species. The stability of two solvents was considered in this study: (1) propylene carbonate (PC) and (2) mono(ethylene glycol) trimethyl silane (1NM1). The latter has a single ether group attached to a trimethylsilane and is used as a model for the longer chain analogue 1NM3, used in the experimental studies. The 6-31+G(d) basis set was used to obtain transition state structures for breaking C O bonds. This basis set 25536

dx.doi.org/10.1021/jp2087412 |J. Phys. Chem. C 2011, 115, 25535–25542

The Journal of Physical Chemistry C

ARTICLE

Figure 2. XPS spectra of Li 1s and C 1s core peaks of the cathode carbon electrodes after discharge and charge when using PC or 1NM3 as solvent in the electrolyte. Standard compounds Li2O2, Li2O, and Li2CO3 are listed at the bottom of each spectrum for reference. PC-DSCG refers to the sample after the first discharge in PC-based electrolyte; PC-CG refers to the sample after the first charge in PC-based electrolyte; 1NM3-DSCG refers to the sample after the first discharge in 1NM3 electrolyte; and 1NM3-CG refers to the sample after the first charge in 1NM3 electrolyte. We note that the intensities do not necessarily correlate with amounts of products present and that the XPS samples only to depths of about 5 nm.

was also used to calculate vibration energies and solvation energies for calculating the corresponding reaction barrier enthalpies. Optimized structures for calculating hydrogen and proton abstraction reaction energies were obtained with the 6-31G(2df,p) basis sets. This basis set was also used for calculating the solvation energies. Single-point energies were then calculated with the 6311+G(2df,p) basis set to obtain Gibbs free energies (at 298.15 K, 1 atm). The SMD solvation model was used to compute the solvation free energies at the B3LYP level of theory.24 Dielectric constants (ε) for 1NM1 of 5.67 (average of the experimental dielectric values 3.6 to 7.5 for 1NM1, 1NM2, and 1NM3) using the chloro-benzene solvent model and 64.4 for propylene carbonate using water as the solvent model were employed in the calculations. In addition, the high-level G4MP2 quantum chemical method27,28 was used to assess selected energies from the B3LYP method. The Gaussian 09 software was used for these calculations.25 3.2. Propylene Carbonate. A detailed investigation was carried out of the charge and discharge products in propylene carbonate-based electrolyte, which has been commonly used in many Li-air electrochemical reactions.8 Figure 1 shows the first discharge and charge cycle of a lithium-air cell using 1.0 M LiPF6 in propylene carbonate with an electrolytic manganese dioxide (EMD) catalyst. The average potential of the discharge curve is 2.5 V, while the charge potential is initiated at around 4.2 V,

which is similar to what has been reported previously for this type of system.21 XPS was used to investigate the charge and discharge products during the first cell cycling and the data is shown in Figure 2. The Li 1s XPS spectra of the discharge products from cells based on propylene carbonate is shown in Figure 2A. The Li 1s spectra for some standard compounds (Li2O, Li2O2, and Li2CO3), which are potential discharge products, are included at the bottom of Figure 2A. The results in Figure 2A indicate that the discharge products in PC (PC-DSCG) are Li2CO3 (or Li-alky-carbonate) and lithium oxides. On the basis of this XPS data, it is not possible to distinguish between Li2O2 and Li2O in the discharge. Upon charging the cell (PC-CG), both C 1s and Li 1s spectra (parts A and B of Figure 2) show the disappearance of the lithium carbonate species, indicating that it is decomposed under charge. The XPS data shows that lithium oxides remain after charging, but based on this data we cannot determine how much, if any, was decomposed during charge. The XPS data also indicates the presence of LiF upon charging, perhaps due to a decomposition of LiPF6. The XPS data provide strong evidence that lithium carbonate or other carbonates are formed during the discharge and that they decompose during the charging process. Density functional theory (DFT) calculations were carried out to investigate reaction mechanisms for decomposition of PC as a means of further 25537

dx.doi.org/10.1021/jp2087412 |J. Phys. Chem. C 2011, 115, 25535–25542

The Journal of Physical Chemistry C

ARTICLE

Figure 3. Comparison of the computed barriers (enthalpies) for activation of (A) PC decomposition by O2 anion radical (O2 ), LiO2 radical, LiO2 anion (LiO2 ), and Li2O2 and (B) 1NM1 decomposition by O2 anion radical, LiO2 radical, LiO2 anion, and Li2O2. The reaction of LiO2 anion with PC is exothermic by 40.5 kcal/mol and is a barrierless process. The O2 anion radical barriers at G4MP2 are 12.1 kcal/mol (for PC) and 21.4 kcal/mol (for 1NM3).

understanding the experimental results. Oxygen reduction products can take a variety of forms that can lead to electrolyte decomposition and could be found either in solution or at the cathode interface. In this study, we have considered only reactions occurring in solution, including four possible solvated species, to assess the stability of the electrolytes. These include molecular, neutral, and anionic species of Li2O2, LiO2 , LiO2, and O2 . The likelihood of any of these species being present is dependent on the electrolyte conditions, and assessing this is beyond the scope of this study. For example, the superoxide anion, O2 , has been postulated by many to be present in some recent studies,12,16 although its existence does depend on the electrolyte conditions.26 Reactions at the lithium oxide surface could also be responsible for the degradation of the solvent. In DFT studies using a Li2O2 trimer as a model for the Li2O2 surface we found the decomposition is also favorable.27 The barriers for the first step in the decomposition of PC by the four solvated species described above were calculated by density functional theory. The first step in the decomposition of PC to Li2CO3 or other lithium alkyl carbonates is ring-opening (C O bond breaking) on the basis of previous theoretical studies of the PC reduction mechanisms.17,28 The energy barriers for this first decomposition step in reactions of PC with Li2O2, LiO2 , LiO2, and O2 and the transition state structures are shown in Figure 3. The calculated energy barriers for all four species are quite small, ranging from no barrier to 23 kcal/mol, with LiO2 being the most reactive. These conclusions are also supported by high level quantum chemical calculations at the G4MP2 level of theory29,30 for the O2 species, which are also

given in Figure 3. The results for O2 attack on PC are also consistent with a recent DFT study reported by Blanco and Brantsev31 for this reaction. After ring-opening of PC, the reaction is thermodynamically downhill, assuming a second electron transfer, to form Li2CO3 and other products such as formaldehyde and acetaldehyde. The calculated energies, including barriers, for such a decomposition pathway for PC by LiO2 are given in Figure 4. The apparent barrier for the second step in this reaction, i.e., C O bond breaking is below the previous reactants, indicating that the decomposition will be favorable. Li alkyl carbonates or other products could also be formed by attack of the oxygen reduction products on PC as suggested by Freunberger et al.15 Thus, the choice of solvent for electrolyte appears to be critical factor in improving the performance of lithium-air batteries. We also note that the decomposition of the lithium carbonate observed by XPS will take place at a higher potential than lithium oxide based on formation energies (see Supporting Information) and could be responsible for the large overpotential observed during cell cycling when using propylene carbonate. 3.3. Tri(ethylene glycol)-Substituted Methyltrimethyl Silane (1NM3). To identify more stable electrolytes that do not generate side reactions during the discharge process, we have investigated the reactivity of a new oligoether silane solvent 1NM3 using both density functional computations and experiments. Recently, we have reported that a series of oligoether silane compounds have very good properties as electrolytes, such as low glass transition temperatures, effective ionic transport, low viscosity, and good conductivity in the presence of lithium 25538

dx.doi.org/10.1021/jp2087412 |J. Phys. Chem. C 2011, 115, 25535–25542

The Journal of Physical Chemistry C

Figure 4. Calculated decomposition pathway for propylene carbonate molecule by LiO2 from density functional theory. The first step A f B is barrierless. The enthalpy of activation is 23.6 kcal/mol for Cf D but is much further below the starting reactants. Note that the addition of lithium ion at position (a) or (b) in structure B results in the formation of structure B.

ions.32 36 The oligoether silane compounds are referred to as 1NMx, where x is the number of ethylene oxide (EO) units such as 1NM1 (one EO unit) and 1NM3 (three EO units).

The 1NMx series is based on linear ether chains, so it is in the same class of solvents as dimethoxyethane and the n-glymes investigated by others for Li-air batteries.7,19,22 However, the addition of the silicon group gives it improved properties as an electrolyte.32 35,37 In this work, we have used 1NM3 for the investigation of the performance of these silicon-based ethers in the Li-air cells. In most of the density functional calculations we used 1NM1 as a model for investigating the stability of 1NM3

ARTICLE

during discharge. The shorter ether chain of 1NM1 saves computational time but provides a reasonable assessment of the stability of this type of solvent toward different discharge products. Computationally, we have investigated two types of possible reactions during discharge: (1) cleavage of C O ether bonds and (2) hydrogen or proton abstraction reactions. In the first case we have investigated C O bond breaking adjacent to the silicon in 1NM1, which is a possible first step in the decomposition during discharge since the Si O bond is stronger than the C O bonds.38 The calculated energy barriers for the first step in the decomposition of 1NM1 by the four solvated species (Li2O2, LiO2 , LiO2, and O2 ) considered in section 3a for PC decomposition are summarized in Figure 3. The results in the Figure clearly show that the barriers for C O bond breaking in 1NM1 are much larger than the barriers for ring-opening in PC and in most cases nearly double in magnitude. For the oxygen radical anion, which many people speculate is a likely oxygen reduction species, the barrier for C O bond breaking is 29 kcal/mol, compared to16 kcal/mol for PC. We note that the solvated LiO2 species has a relatively small barrier (12 kcal) for breaking a C O bond in 1NM1 (Figure 3) so that under some conditions it is possible decomposition of an ether group could occur based on these calculations just as PC. In the longer chain species, 1NM3, nucleophilic attack by the solvated oxygen reduction species can occur at various sites. Three C O bonds of 1NM3 were considered for attack by O2 , and the activation barriers were evaluated. The calculated barriers of breaking the three different C O bonds are in the range of 30 to 36 kcal/mol, which is similar to that found for the C O bond breaking in 1NM1. Thus, bond breaking in ether type groups by oxygen reduction species is much less likely in the oligoether silanes than in propylene carbonates. In addition, the higher-level G4MP2 calculations confirm the DFT energies for the solvated O2 species (see Figure 3). In DFT studies using a Li2O2 trimer as a model for the Li2O2 surface we found the decomposition of the ether is also much less favorable than for PC.27 In addition to the barriers for C O bond breaking in 1NM1, we have investigated the barriers for hydrogen and proton abstraction by two solvated species, the O2 radical anion and the Li2O2 molecule, which are illustrated in Scheme 1. This type of mechanism has been proposed by Fruenberger et al22 to account for the decomposition of n-glyme ethers observed in their Li-air cells. They suggested that the hydrogen abstraction mechanism occurs during the first discharge and causes decomposition due to attack of the O2 radical anion on a methylene group. They postulated that the resulting products would decompose to form carbonates. The results of our calculations of the hydrogen abstraction (reaction 1) as well as proton abstraction (reaction 2) mechanisms (Scheme 1) for 1NM1 are given in Table 1. In the case of reaction 1 the product is a 1NM1 anion, while in the case of reaction 2 it is a 1NM1 radical. Four possible C H sites were considered (A D), as shown in Scheme 1. The reaction energies are all endothermic (12 64 kcal/mol). No reaction barriers were calculated, but both hydrogen and proton abstraction will likely have barriers and make these mechanisms even more unfavorable. Table 1 also summarizes the hydrogen and proton abstraction reaction energies by a solvated Li2O2 monomeric species (reactions 3 and 4). The results indicate that hydrogen or proton abstraction by a solvated Li2O2 species is also very endothermic (26 83 kcal/mol). Thus, hydrogen and proton abstraction is not a favorable first step in 25539

dx.doi.org/10.1021/jp2087412 |J. Phys. Chem. C 2011, 115, 25535–25542

The Journal of Physical Chemistry C

ARTICLE

Scheme 1. Schematic Representation of Possible Proton Abstraction and Hydrogen Abstraction Reactions of 1NM1 with O2 Anion Radical

Table 1. Computed Free Energies of Various Hydrogen and Proton Abstraction Reactions of 1NM1 at 298 K Using the B3LYP/6-311+G(2df,p)//B3LYP/6-31G(2df,p) Level of Theory (see Scheme 1 for O2 reactions; Li2O2 reactions are similar).a (All energies in kcal/mol) H position

H abstraction by O2 (reaction 1)

proton abstraction by O2 (reaction 2)

H abstraction by Li2O2 (reaction 3)

proton abstraction by Li2O2 (reaction 4) 77.5

A

64.2, 56.1b

32.4

78.2 (71.2)

B

11.9, 11.0 b

29.0

25.9 (26.8)

74.1

C

31.7, 31.7 b, 30.1c

30.3, 30.1c

45.6 (46.9)

75.4

D

44.5 (39.2) b

37.8

58.5 (53.3)

83.0

a

The solvation energies are computed at the B3LYP/6-31G(2df,p) level of theory using the SMD solvation model. b Result from G4MP2 calculations. c Result from B3LYP calculation on monoglyme (CH3OCH2CH2CH3).

the decomposition mechanism of oligoether silanes by potential oxygen discharge products based on these calculations using a continuum model for the surrounding electrolyte. Results are also given in Table 1 for hydrogen and proton abstraction from a diglyme ether molecule for which the reaction energies are also endothermic. Motivated by our DFT calculations, we carried out a detailed experimental investigation of the charge and discharge products in a 1NM3-based electrolyte. Figure 1 shows the first discharge and charge cycle of a lithium-air cell using 1.0 M LiPF6 in 1NM3 with an electrolytic manganese dioxide (EMD) catalyst. The average potential of the discharge curve is about 2.5 V, while the charge potential is initiated at around 3.4 V and rises to about 4.2 V before reaching the charge cutoff. The full cycling performance data indicates a degradation of discharge capacity that occurs in the subsequent cycles (see Supporting Information). XPS measurements were used to investigate the charge and discharge products during the first cell cycling, and the data are shown in Figure 2. The results clearly show distinct differences of the 1NM3 discharge products compared to those of PC (Figure 2A). Specifically, the only discharge products for 1NM3 (1NM3DSCG) are lithium oxides, and no lithium carbonates are observed. The lithium oxides are partially decomposed upon charging according to the results in Figure 2C (1NM3-CG). Thus, the XPS data shown in Figure 2C are consistent with the density functional predictions that 1NM3 is more stable than PC during discharge. FTIR data also provide clear evidence that the 1NM3-based electrolyte is stable to decomposition during discharge. Figure 5

shows the IR spectra of the discharged product (without solvent rinse to avoid the potential reaction of the discharged product and the solvent). No Li2CO3 signature (strong peak at ∼1416 cm 1 and a shoulder at ∼1500 cm 1)39 can be identified on the air cathode, lithium anode, or the separator, indicating the lithium air cell with 1NM3 electrolyte does not form Li2CO3, which is present in the cathode at discharged state with PC electrolyte, as shown by our XPS data. The FTIR spectra also provide no evidence for the formation of alkycarbonates on discharge using 1NM3 as a solvent. These results are consistent with the above theoretical calculations on ether stability. We note that in the first charging cycle for 1NM3 (Figure 1) the initial potential is around 3.4 eV and then rises to over 4 V, at which time the charge is terminated by the criteria for charge time being equal to the discharge time. The voltage plateau at 3.4 V is probably associated with the decomposition of the lithium oxide observed by XPS. The plateau at 3.4 eV found for 1NM3 suggests that the lithium oxide does not necessarily have large charge overpotential as has been generally assumed for Li-air cells. 3.4. Discussions. The following picture emerges of the role of electrolyte stability in Li-air batteries from this experimental/ theoretical study of 1NM3 and PC electrolytes under similar conditions in a Li-air cell. The measured XPS and FTIR data for the 1NM3 electrolytes provides evidence that only lithium oxides are formed during discharge and are partially decomposed during charge. In contrast, when PC is used in the same cell configuration, oxygen reduction products cause decomposition of carbonatebased electrolytes as evidenced by XPS data, and the decomposition of Li2CO3 results in a large overpotential of the cell. 25540

dx.doi.org/10.1021/jp2087412 |J. Phys. Chem. C 2011, 115, 25535–25542

The Journal of Physical Chemistry C

ARTICLE

confirm that the 1NM3 species should be more stable during discharge than PC. Results for other ethers give similar results in the calculations. (3) There is a large initial reduction in the cell charge overpotential when only lithium oxides form on discharge with use of the 1NM3 electrolyte compared to PC electrolyte, which produces carbonates on discharge. This suggests that large overpotentials in Li-air batteries can be significantly lowered with sufficient control of the discharge products. In summary, these results provide evidence that electrolyte solvent stability plays a key role in the performance of Li-air batteries and making advances in new electrolytes will be a key factor in reducing the large overpotential and improving reversibility.

’ ASSOCIATED CONTENT

bS Figure 5. FTIR spectra of cathode, Li anode, separator, and electrolyte at discharge stage used in the 1NM3/LiPF6 Li-air cell with a MnO2 catalyst. The peak at 1458 cm 1 that appears in the anode, cathode, and separator is from the 1NM3 C O stretching frequency. The reference spectrum for Li2CO3 is given in the Supporting Information.

These results are consistent with DFT calculations indicating that the 1NM3 is more stable to the highly active oxygen reduction species formed during discharge than PC based on calculations for breaking C O bonds and hydrogen or proton abstraction reactions. Although the 1NM3-based solvent is stable to oxygen reduction species during discharge, the cycling performance of the Li-air battery using this solvent was poor because of a lack of reversibility (see Supporting Information). The lack of reversibility is probably related to the failure to decompose all of the lithium oxide on charge (see XPS data, Figure 2). The behavior of the 1NM3-based cell during the first charge, i.e., rising from 3.4 to 4.2 V before the cutoff, is also consistent with some of the lithium oxide not being decomposed, suggesting that some of the lithium oxide requires a higher potential for decomposition. The possibility of several forms of Li2O2 being formed during discharge has also been suggested by McCloskey et al19 in their investigations based on dimethoxyethane as a solvent. It is clear that further exploration of how both solvents and salts in electrolytes control the discharge and charge processes is needed to improve the cycling performance in Li-air cells. We are carrying out further investigations for these effects.

4. CONCLUSIONS The following conclusions can be drawn from this combined experimental and computational study of an electrolyte based on an oligoether silane 1NM3 in a Li-air battery: (1) XPS and FTIR experiments show that only lithium oxides and no carbonates are formed when this oligoether silanebased electrolyte is used. In contrast XPS shows that PC in the same cell configuration decomposes to form lithium carbonates during discharge. (2) Density functional calculations of probable decomposition reaction pathways involving C O cleavage and proton/ hydrogen abstraction to potential oxygen reduction species

Supporting Information. Decomposition potential for Li2CO3, 1NM3 C O bond activation transition state structures, XPS spectra of C 1s core peaks of the standard compounds Li2O2 and Li2O, reference FTIR spectrum of Li2CO3, and cycling performance of the Li-air battery using the 1NM3 solvent. This information is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: ([email protected]) (L.C.); ([email protected]) (K.A.). Phone: 1-630-252-7380 (L.C.); 1-630-252-3838 (K.A.). Author Contributions #

These authors contributed equally.

’ ACKNOWLEDGMENT We gratefully acknowledge grants of computer time from EMSL, a national scientific user facility located at Pacific Northwest National Laboratory, ANL Laboratory Computing Resource Center (LCRC), and Center of Nanoscale Materials (CNM). This work was also supported by the Human Resources Development of the Korea Institute of Energy Technology Evaluation and Planning (KETEP ) grant funded by the Korea government Ministry of Knowledge Economy (No. 20114010203150 ). The submitted manuscript has been created by UChicago Argonne, LLC, Operator of Argonne National Laboratory (“Argonne”). Argonne, a U.S. Department of Energy Office of Science laboratory, is operated under Contract No. DE-AC02-06CH11357. The U.S. Government retains for itself, and others acting on its behalf, a paid-up nonexclusive, irrevocable worldwide license in said article to reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly, by or on behalf of the Government. ’ REFERENCES (1) Ogasawara, T.; Debart, A.; Holzapfel, M.; Novak, P.; Bruce, P. G. J. Am. Chem. Soc. 2006, 128, 1390–1393. (2) Girishkumar, G.; McCloskey, B.; Luntz, A. C.; Swanson, S.; Wilcke, W. J. Phys. Chem. Lett. 2010, 1, 2193–2203. (3) Abraham, K. M.; Jiang, Z. J. Electrochem. Soc. 1996, 143, 1–5. (4) Abraham, K. M.; Jiang, Z.; Carroll, B. Chem. Mater. 1997, 9, 1978–1988. 25541

dx.doi.org/10.1021/jp2087412 |J. Phys. Chem. C 2011, 115, 25535–25542

The Journal of Physical Chemistry C (5) Read, J.; Mutolo, K.; Ervin, M.; Behl, W.; Wolfenstine, J.; Driedger, A.; Foster, D. J. Electrochem. Soc. 2003, 150, A1351–A1356. (6) Kowaluk, I.; Read, J.; Salomon, M. Pure Appl. Chem. 2007, 79, 851–860. (7) Read, J. J. Electrochem. Soc. 2006, 153, A96–A100. (8) Debart, A.; Paterson, A. J.; Bao, J.; Bruce, P. G. Angew. Chem., Int. Ed. 2008, 47, 4521–4524. (9) Lu, Y.-C.; Gasteiger, H. A.; Parent, M. C.; Chiloyan, V.; Shao-Horn, Y. Electrochem. Solid-State Lett. 2010, 13, A69–A72. (10) Amine, K.; Chen, C. H.; Liu, J.; Hammond, M.; Jansen, A.; Dees, D.; Bloom, I.; Vissers, D.; Henriksen, G. J. Power Sources 2001, 97 8, 684–687. (11) Abraham, D. P.; Twesten, R. D.; Balasubramanian, M.; Petrov, I.; McBreen, J.; Amine, K. Electrochem. Commun. 2002, 4, 620–625. (12) Albertus, P.; Girishkumar, G.; McCloskey, B.; Sanchez-Carrera, R. S.; Kozinsky, B.; Christensen, J.; Luntz, A. C. J. Electrochem. Soc. 2011, 158, A343–A351. (13) Xiao, J.; Hu, J.; Wang, D.; Hu, D.; Xu, W.; Graff, G. L.; Nie, Z.; Liu, J.; Zhang, J.-G. J. Power Sources 2011, 196, 5674–5678. (14) Bryantsev, V. S.; Blanco, M.; Faglioni, F. J. Phys. Chem. A 2010, 114, 8165–8169. (15) Freunberger, S. A.; Chen, Y.; Peng, Z.; Griffin, J. M.; Hardwick, L. J.; Barde, F.; Novak, P.; Bruce, P. G. J. Am. Chem. Soc. 2011, 133, 8040–8047. (16) Mizuno, F.; Nakanishi, S.; Kotani, Y.; Yokoishi, S.; Iba, H. Electrochemistry 2010, 78, 403–405. (17) Vollmer, J. M.; Curtiss, L. A.; Vissers, D. R.; Amine, K. J. Electrochem. Soc. 2004, 151, A178–A183. (18) Veith, G. M.; Dudney, N. J.; Howe, J.; Nanda, J. J. Phys. Chem. C 2011, 115, 14325–14333. (19) McCloskey, B. D.; Bethune, D. S.; Shelby, R. M.; Girishkumar, G.; Luntz, A. C. J. Phys. Chem. Lett. 2011, 2, 1161–1166. (20) Laoire, C. O.; Mukerjee, S.; Plichta, E. J.; Hendrickson, M. A.; Abraham, K. M. J. Electrochem. Soc. 2011, 158, A302–A308. (21) Hassoun, J.; Croce, F.; Armand, M.; Scrosati, B. Angew. Chem., Int. Ed. 2011, 50, 2999–3002. (22) Freunberger, S. A.; Chen, Y.; Drewett, N. E.; Hardwick, L. J.; Barde, F.; Bruce, P. G. Angew. Chem., Int. Ed. 2011, 50, 8609–8613. (23) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652. (24) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 2009, 113, 6378–6396. (25) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, N. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, € Foresman, J. B.; Ortiz, J. V.; J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O; Cioslowski, J.; Fox, D. J. Gaussian 09; Gaussian, Inc.: Wallingford, CT, 2009. (26) Laoire, C. O.; Mukerjee, S.; Abraham, K. M.; Plichta, E. J.; Hendrickson, M. A. J. Phys. Chem. C 2010, 114, 9178–9186. (27) Assary, R. S.; Lau, K. C.; Redfern, P.; Greeley, J.; Curtiss, L. To be published. (28) Wang, Y.; Balbuena, P. B. J. Phys. Chem. B 2002, 106, 4486–4495. (29) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. J. Chem. Phys. 2007, 126, 084108. (30) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. J. Chem. Phys. 2007, 127, 124105. (31) Bryantsev, V. S.; Blanco, M. J. Phys. Chem. Lett. 2011, 2, 379–383.

ARTICLE

(32) Zhang, Z. C.; Lyons, L. J.; Jin, J. J.; Amine, K.; West, R. Chem. Mater. 2005, 17, 5646–5650. (33) Zhang, L. Z.; Zhang, Z. C.; Harring, S.; Straughan, M.; Butorac, R.; Chen, Z. H.; Lyons, L.; Amine, K.; West, R. J. Mater. Chem. 2008, 18, 3713–3717. (34) Chen, Z. H.; Wang, H. H.; Vissers, D. R.; Zhang, L. Z.; West, R.; Lyons, L. J.; Amine, K. J. Phys. Chem. C 2008, 112, 2210–2214. (35) Oh, B.; Vissers, D.; Zhang, Z.; West, R.; Tsukamoto, H.; Amine, K. J. Power Sources 2003, 119, 442–447. (36) Rossi, N. A.; West, R. Polym. Int. 2009, 58, 267–272. (37) Zhang, L. Z.; Lyons, L.; Newhouse, J.; Zhang, Z. C.; Straughan, M.; Chen, Z. H.; Amine, K.; Hamers, R. J.; West, R. J. Mater. Chem. 2010, 20, 8224–8226. (38) Assary, R. S.; Curtiss, L. A.; Redfern, P. C.; Zhang, Z.; Amine, K. J. Phys. Chem. C 2011, 115, 12216–12223. (39) Zhuang, G. V.; Yang, H.; Blizanac, B.; Philip N. Ross, J. Electrochem. Solid-State Lett. 2005, 8, A441–A445.

25542

dx.doi.org/10.1021/jp2087412 |J. Phys. Chem. C 2011, 115, 25535–25542