Enhanced electrochemical performance of Li–O₂ battery based on modifying the solid-state air cathode with Pd catalyst

Abstract

Pd nanoparticles were deposited onto carbon nanotubes (CNTs) via a simple wet chemical method and were employed to fabricate a highly efficient cathode for a Li–O₂ battery using a solid-state air cathode. The results demonstrate that the introduced Pd catalyst showed promising catalytic activity, making the battery exhibit improved rate ability and excellent cycling performance. The fabricated CNTs@Pd electrode tended to induce the poorly crystalline Li₂O₂ to form preferentially in the solid-state air cathode, which could be decomposed at a lower potential in comparison to the Pd-free electrode. The Li–O₂ battery using a solid-state CNTs@Pd electrode exhibited a long cycle life up to 50 cycles without capacity fading (capacity limit of 500 mA h g⁻¹). The unique architecture of the CNTs@Pd electrode minimised the side reactions relating to carbon and the discharge product. Our results provide a snapshot toward developing a high performance solid-state Li–O₂ battery.

---

1. Introduction

Li–O₂ batteries are considered to be a promising next-generation electrochemical energy storage technology; the energy density of the Li–O₂ battery is almost one order of magnitude higher than that of conventional Li-ion batteries,^1,2 which is comparable to gasoline.³ Much research currently focuses on using organic electrolytes, however, their practical application is constrained by several serious challenges, including low rate capability, poor cycling performance and the decomposition of the components in the Li–O₂ battery.^4–6 In addition, the oxygen solubility in the organic liquid electrolytes is very low, and the evaporation, flammable and flowing natures of the organic liquid electrolytes inevitably cause safety problems. Therefore, introducing a stable electrolyte is one of the challenges to improve the safety and practicality of the Li–O₂ battery.

Replacing the organic electrolytes with non-volatile, non-flammable electrolytes is an option. Possible electrolytes such as polymer electrolytes and ionic liquids have been investigated. However, their use still faces great challenges: (1) the ionic conductivity of polymer electrolytes such as poly(ethylene oxide) (PEO) reaches acceptable values only at high temperatures above 70 °C, and PEO used at the cathode side is unstable and has been detected to decompose during cycling; (2) the cost of ionic liquids is high and their stability is under intense scrutiny. Promising alternatives are inorganic solid-state electrolytes such as Li–Al–Ge–PO₄ (LAGP) or Li–Al–Ti–PO₄ (LATP), which has acceptable Li-ion conductivity at room temperature and possesses relatively high chemical and electrochemical stability. Many efforts have been made to fabricate semi-solid or all solid-state Li–O₂ batteries by introducing a solid-state electrolyte.^7–9 Recently, we fabricated a Li–O₂ battery using a solid-state air cathode with a simple method and obtained a high specific capacity.¹⁰ However, many challenges such as low rate capability and poor cycling performance need to be overcome before practical application is realized. It is undoubted that the cathode plays a critical role of the performance-determining step. Therefore, for practical application of the solid-state Li–O₂ battery, there is an urgent need to optimize the cathode.

Among the components of the air cathode, carbon materials have a high surface area and superior electrochemical activity for the oxygen reduction reaction (ORR); the only drawback is the formation of interfacial Li₂CO₃ via the reaction of Li₂O₂ with carbon, which results in the decomposition of carbon and deteriorates the resulting electrochemical performance. Although non-carbonaceous materials such as Au,¹¹ TiC,¹² TiO₂,¹³ Ru/ITO,¹⁴ Ru/STO¹⁵ are expected to be attractive candidates, if solely used as carbon-free cathodes, the specific capacity per unit mass or area is very low due to their high molecular weight or low catalytic activity. In addition, the synthesis of a porous structure with a high surface area to deposit the discharge product is not an easy task. Therefore, carbon materials have been widely used until now. Introducing a catalyst on the surface of carbon is an efficient approach to prevent or decrease the direct contact between carbon and Li₂O₂ and avoid or reduce the decomposition of carbon.¹⁶ Various catalysts have been investigated for use in organic electrolyte Li–O₂ batteries.^17,18 However, following research pointed out that some catalysts actually play a role in catalyzing the decomposition of electrolytes.^19,20 It is speculated that the catalyst in organic electrolyte Li–O₂ batteries provides limited insight into its real catalytic activity. In contrast to organic electrolytes, inorganic solid-state electrolytes are considered to be a reliable alternative, and nearly no catalyst catalyzes their decomposition under normal circumstances. We believe that the performance of a solid-state Li–O₂ battery could be further improved by modifying the air cathode, and investigation of the effect of catalysts on the electrocatalytic activity of solid-state air cathodes is very important. However, research about this topic is rare.

Herein, Pd nanoparticles were deposited onto carbon nanotubes (CNTs) via a chemical reduction method. Unlike previous studies that investigated the catalytic activity in organic electrolytes, the obtained CNTs@Pd composite was firstly introduced into a solid-state air cathode and the performance of a Li–O₂ battery using the solid-state air cathode was investigated and evaluated in detail. Our results demonstrate that the electrochemical performance (rate ability and cycle performance) of the Li–O₂ battery using a solid-state air cathode modified with a Pd catalyst was significantly enhanced, which is attributed to the high electrocatalytic activity of Pd and the unique architecture that minimises the possible side reactions. These results exhibited the real role of the Pd catalyst in a Li–O₂ battery.

2. Experimental

2.1 Synthesis of the CNTs/Pd composite

CNTs were prepared by a floating catalyst chemical vapor deposition method, using ethylene as the carbon source and ferrocene as the catalyst; detailed procedures for the preparation can be seen elsewhere.²¹ Prior to the synthesis of the CNTs/Pd composite, the CNTs were treated in a solution of 75% HNO₃ and 25% H₂SO₄ for 30 min. After that the CNTs were washed several times in DI water until neutral, and finally, the resulting CNTs were dried at 80 °C overnight. This process imported oxygen-containing functional groups onto the surface of the CNTs and enhanced the interfacial adhesion between the CNTs and the metal, which facilitates the chemisorption of the Pd precursor.²² The obtained CNTs were immersed into a 3 mM PdCl₂ EG solution; the weight ratio of CNTs : Pd²⁺ was controlled at around 8 : 2. Then, ethylenediamine tetraacetic acid was added as a chelating agent and NaOH was added to adjust the pH to about 9.0. After that, a freshly prepared NaBH₄ solution was added dropwise into the above solution and stirred for 60 min. Finally, the mixture was filtered, washed and dried in a vacuum oven at 80 °C overnight to obtain the final CNTs/Pd composite.

2.2 Electrochemical measurements

The electrochemical performance was carried out using a Swagelok-type Li–O₂ battery composed of a clean Li metal disk, a piece of LATP sheet and a porous solid-state air cathode. 50 µl of organic electrolyte (0.9 M lithium bis(trifluoromethanesulfonyl)imide in tetraethylene glycol dimethyl ether (TEGDME)) impregnated into a Celgard 2500 separator was introduced to avoid direct contact between the Li anode and LATP sheet because Li metal reacts easily with LATP when contacted directly. More details have been reported in our earlier work.⁹ The solid-state air cathode contained 15 wt% carbon material (pure CNTs or CNTs/Pd composite), 80 wt% LATP powder and 5 wt% binder (polyvinylidene fluoride). All of the components were coated on a nickel foam current collector (1.5 cm²) and dried at 120 °C under vacuum for 12 h. The typical loading of both the CNTs and CNTs@Pd for the CNT electrode and Pd-loaded electrode were the same (1 mg). The battery was assembled in an argon-filled glove box and was tested in 1 atm of O₂. The galvanostatic tests were carried out using a LAND tester (Wuhan LAND Electronic Co., Ltd.) at 25 °C. The current density and specific capacity were calculated based on the weight of the pure CNTs (or CNTs@Pd) for the CNT electrode (or CNTs@Pd electrode).

2.3 Characterizations

After electrochemical measurement, the battery was disassembled in the glove box for further analysis. X-ray diffraction (XRD) data was collected using a Dandong DX-2600 diffractometer with Cu Kα radiation. Fourier transform infrared spectroscopy (FTIR) was carried out on a Nicolet 6700 spectrometer (Thermo Fisher Scientific). Raman spectra were measured using the Renishaw inVia Raman spectrometer system. The micrograph was investigated by scanning electron microscopy/energy dispersive X-ray spectroscopy (SEM/EDS, JSM-7500F) and transmission electron microscopy (TEM, Tecnai G2 20 S-TWIN). The specific surface areas and the pore size distributions of the CNTs and the CNTs@Pd composite were determined by nitrogen adsorption–desorption isotherms, which were performed on an SSA-4200 analyzer.

3. Results and discussion

The Raman spectra of the CNTs before and after acid treatment are shown in Fig. 1a; it was found that two principal bands existed between 1000 and 2000 cm⁻¹. The band at 1350 cm⁻¹ is common for disordered sp² carbon, and is called the D-band; the band at 1580 cm⁻¹ is close to that in well-ordered graphite, and it is often called the G-band. The I_D/I_G ratio is often applied to measure the defect nature of carbon.²³ The I_D/I_G ratio of the raw CNTs was calculated as 0.808, indicating a relatively intact sp² hybridization structure. After the acid treatment, the value increased to 0.901, and the increased value of I_D/I_G indicated the introduction of structural defects to the CNTs. Refined analysis was obtained from the FTIR spectra, as shown in Fig. 1b. The peaks at 3427 and 1076 cm⁻¹ correspond to the –OH stretching modes of the absorbed water vapor in ambient air and the C–O stretching modes of ethers, esters, alcohols and phenol compounds,^24,25 respectively. After the acid treatment, another two peaks appeared; the peaks at 1700 cm⁻¹ and 1384 cm⁻¹ correspond to the C=O and O–H stretching modes of the –COOH groups respectively, meanwhile, it could be found that the intensity of the peak at 3427 cm⁻¹ increased after the acid treatment, corresponding to the –OH stretching mode of the –COOH group.²⁶ All these points mentioned above indicated the introduction of oxygen-containing functional groups by the acid treatment, which is favorable to the following Pd-loading process.

Fig. 1 Raman (a) and FTIR (b) spectra of the CNTs before and after acid treatment.

From the TEM image in Fig. 2a, it could be seen that the raw CNTs exhibited a tubular structure and had an average diameter of 20–30 nm, with a tube length of several micrometers. After Pd-loading, the outer surface of the CNTs became rough and some nanoparticles appeared on the CNTs, as shown in Fig. 2b. The weight ratio of the deposited Pd to C was about 20 : 80, as confirmed by EDS in Fig. 2c. The high resolution transmission electron microscopy of the CNTs@Pd is given in Fig. S1a;† the lattice fringes of the Pd nanoparticles are about 0.225 nm apart, which matches well with the Pd (111) plane. From the TEM image it is clear to find that the particle size of the deposited Pd ranges from 5.3 nm to 9.5 nm and they have an average size of about 7.1 nm, as shown in the histogram of the distribution of the Pd particle sizes (Fig. S1b†). The deposition of Pd nanoparticles can be further confirmed by XRD, as shown in Fig. 2d. The peaks at 25.9° and 42.8° are attributable to the CNTs, and the peaks at 2θ values of 40.1°, 46.4°, 68.1° are in agreement with the (111), (200) and (222) crystal planes of Pd, respectively (JCPDS card 46-1043). The Brunauer–Emmett–Teller analysis of the nitrogen absorption/desorption isotherm in Fig. 2e indicated that the specific surface areas of the raw CNTs and CNTs@Pd were 113.90 m² g⁻¹ and 102.49 m² g⁻¹, respectively. The lower surface area of CNTs@Pd can be attributed to the deposition of Pd nanoparticles. However, as the Pd content in CNTs@Pd was about 20 wt%, the specific surface area of CNTs@Pd was still high, suggesting that the Pd nanoparticles deposited on the surface of the CNTs made great contribution towards the specific surface area. In addition, it is well known that CNTs have pores at the end of the tubes, and after the deposition of Pd nanoparticles, the CNTs@Pd retained the pore structure, as shown in the pore size distribution analysis in Fig. 2e, further suggesting the well maintained tubular structure.

Fig. 2 TEM images of the pristine CNTs (a) and CNTs@Pd (b). EDS of the CNTs@Pd (c). XRD pattern of the pristine CNTs and CNTs@Pd (d). Nitrogen adsorption–desorption isotherms of the pristine CNTs and CNTs@Pd (e). Inset: pore size distributions.

To evaluate the effect of a Pd catalyst on the electrochemical performance of the Li–O₂ battery using a solid-state air cathode, the first discharge–charge curves of the battery with the CNTs@Pd electrode were compared with those of the battery using the CNT electrode at a current density of 100 mA g⁻¹, as shown in Fig. 3a. It could be found that the two kinds of air cathode showed similar discharge capacities. At the discharge terminal, the Pd-free electrode delivered a capacity of 7455 mA h g⁻¹ and the Pd-loaded electrode delivered a capacity of 7800 mA h g⁻¹ (or 9750 mA h g⁻¹ based on the weight of CNTs). However, the discharge voltage plateau was 2.80 V for the CNTs@Pd electrode, which was 260 mV higher than that of the CNT electrode, reflecting an improved ORR catalytic activity. During the following recharge process, we employed an equal-capacity charging mode in which the charging capacity matches the previous discharging capacity, while the charge potential is limited to 4.5 V; in other words, the recharge process will stop at whichever condition is reached first, as reported in the previous literature.²⁷ It was found that both electrodes showed a similar charge process and the charge terminal voltage did not exceed 4.5 V. Although there was no significant difference in the charge voltage plateau for both electrodes, a steeper increase in voltage could be found at the end of the charge process for the CNT electrode compared with the CNTs@Pd electrode; at the charge terminal, the voltage of the CNTs@Pd electrode was 4.3 V, which was about 100 mV lower than that of the CNTs electrode. The result suggests that Pd nanoparticles can efficiently catalyze the decomposition of the discharge product and thus improve the reversibility of the CNTs@Pd electrode.

Fig. 3 Discharge–charge profiles of the Li–O₂ battery with CNT and CNTs@Pd electrodes (a). XRD patterns of the pristine, discharged and recharged CNT (b) and CNTs@Pd electrodes (c). FTIR spectra of the CNT and CNTs@Pd electrodes (d).

The reaction products of the Li–O₂ batteries with two kinds of electrodes were then analysed by XRD, as shown in Fig. 3b and c. Compared with the CNT electrode, the peaks related to Li₂O₂ were clearly evident after discharge, as highlighted in Fig. 3b. After recharge, all the discharge products were removed from the electrode, indicating that the recharge process made Li₂O₂ decompose. For the CNTs@Pd electrode, the intensity of Pd was weak, due to the limited amount of Pd and the strong background of Ni and LATP, as shown in Fig. 3c. Compared with the XRD patterns of the pristine electrodes, the characteristic diffraction peaks of Li₂O₂ were also observed for the discharged electrodes, and no other crystalline phase could be detected, indicating that Li₂O₂ is the dominant crystalline product. However, it was apparent that the peaks of Li₂O₂ in the CNTs@Pd electrode were broader and their intensity was much less than those in the CNT electrode. This result indicated that the amount of crystalline Li₂O₂ in the discharged CNTs@Pd electrode was much less than that in the discharged CNT electrode. Considering that the discharge capacity of the CNTs@Pd electrode was slightly higher than that of the CNT electrode, the results indicated that some Li₂O₂ was not detectable in the XRD pattern. Further refined analysis of the CNT and CNTs@Pd electrodes was obtained by FTIR, as shown in Fig. 3d. No significant Li₂CO₃ could be identified at the discharge terminal, excluding the possibility of a side reaction in which crystalline Li₂O₂ is consumed by a reaction with carbon that was reported in previous work. Therefore, a possible reason is that the formed Li₂O₂ has a non-crystalline structure. A different crystalline discharge product was found in an organic-based electrolyte Li–O₂ battery when using a noble metal or some other metal oxide as catalyst, which was probably attributed to the strong binding of superoxide to the catalyst surface.^28–31 For the Li–O₂ battery using a solid-state air cathode, the mechanism of the formation of poorly crystalline Li₂O₂ is unclear. It is speculated that the existence of Pd nanoparticles on the CNTs influenced the binding strength between Li and O^*₂ and resulted in the growth of poorly crystalline Li₂O₂. Further detailed investigations on this are needed in the future. The subsequent decomposition of Li₂O₂ in the recharge process was also evident from the XRD pattern, indicating that the discharge product was reversibly decomposed during the following charging process. From the observed high discharge voltage and the poorly crystalline discharge product of the CNTs@Pd electrode, it is currently clear that the Pd nanoparticles not only contribute to a higher discharge voltage but also lead to the formation of poorly or non-crystalline Li₂O₂.

It was reported that the nucleation of Li₂O₂ took place on the surface of the Pd catalyst during the discharge process of an organic electrolyte Li–O₂ battery; in our experiment, it has been detected that the crystallinity of Li₂O₂ is different. Refined analysis was obtained by SEM and the morphologies of the discharged air cathodes for the CNT and CNTs@Pd electrodes are shown in Fig. 4a and b, respectively. It could be found that the morphologies of both electrodes exhibit a very pronounced difference. The discharged CNT electrode contained dendritic-shaped particles, and the discharged CNTs@Pd electrode contained numerous continuous uniform nanorods. The particle size of the discharged CNTs@Pd electrode was much smaller than that of the CNT electrode, which mainly attributed to the different nucleation and growth process. For the CNT electrode, high crystalline Li₂O₂ can freely nucleate and grow on the CNTs, facilitating the formation of large particles. Combining the high discharge voltage with the poorly crystalline discharge product of the CNT@Pd electrode, we speculated that the free migration of Li₂O₂ was suppressed and the growth of poorly crystalline Li₂O₂ preferentially started near the Pd surface and numerous continuous nanorods were formed.

Fig. 4 SEM images of the discharge product of Li–O₂ batteries with CNT (a) and CNTs@Pd electrodes (b). EIS spectra of the two kinds of electrodes before (c) and after discharge (d).

It is presumed that poorly crystalline Li₂O₂ has many defects, which may transform the insulating Li₂O₂ into an electronic conductor.^32–34 EIS was performed to investigate the resistance changes of the Li–O₂ batteries with CNT and CNTs@Pd electrodes before and after discharge, the spectra are shown in Fig. 4c and d, respectively. It was interesting to find that the battery with the CNTs@Pd electrode showed relatively smaller battery impedance in comparison to the battery with the CNT electrode. Although the addition of Pd has no contribution to the interfacial electron transfer resistance in the high-frequency region, the charge-transfer resistance in the medium frequency region in the pristine CNTs@Pd electrode was obviously lower than that in the pristine CNT electrode, indicating that the ORR kinetics were enhanced in the CNTs@Pd electrode. The charge-transfer resistance of the discharged CNTs@Pd electrode was largely lower than that of the CNT electrode, this result indicated less passivation of the CNTs@Pd electrode which is probably attributed to the high electron conductivity of Pd and the enhanced electronic conductivity of poorly crystalline Li₂O₂.

The enhanced ORR kinetics in the CNTs@Pd electrode can promote the formation of Li₂O₂ and facilitate the achievement of improved electrochemical performance. Fig. 5 shows the discharge curves and capacity retention of the two kinds of electrodes at different current densities. The discharge capacity and voltage deteriorated when the current density was increased for both electrodes, which was mainly attributed to electrochemical polarization. However, the Li–O₂ battery with the CNTs@Pd electrode displayed a higher discharge voltage and capacity retention than the battery with the CNT electrode at the investigated current densities. The CNT electrode delivered a discharge capacity of 3363 mA h g⁻¹ at 200 mA g⁻¹ and 1529 mA h g⁻¹ at 300 mA g⁻¹; the corresponding capacity retentions were 45.1 and 20.5%, respectively. The CNTs@Pd electrode delivered an improved discharge capacity of 6285 mA h g⁻¹ at 200 mA g⁻¹ and 3536.7 mA h g⁻¹ at 300 mA g⁻¹; the corresponding capacity retentions were 80.6 and 45.3%, respectively. Also of note is the fact that this discharge capacity was even higher than that of the organic-based electrolyte Li–O₂ battery at 100 mA g⁻¹, even though the discharge capacity of the CNTs@Pd electrode demonstrated an effective improvement compared to that of the CNT electrode, as shown in Fig. S3.† This comparison is in agreement with our previous report and further indicates the superiority of the Li–O₂ battery using a solid-state air cathode.¹⁰ The high rate performance can be attributed to the unique structure of the CNTs@Pd electrode: (1) numerous Pd nanoparticles offered abundant reaction sites and accelerated the nucleation and growth of the discharge products; (2) the unique architecture of the CNTs@Pd electrode helped to enhance the electron and oxygen transportation paths. The improved rate performance confirms the positive catalytic activity of Pd nanoparticles in the ORR process for the Li–O₂ battery using a solid-state air cathode.

Fig. 5 Discharge curves (a) and capacity retention (b) of the two kinds of electrodes at different current densities.

Another considerable improvement of the electrochemical performance of the Li–O₂ battery with the CNTs@Pd electrode is the cycling stability. The cycling performance was measured following the widely used capacity-limiting method,^35–37 at a current density of 100 mA g⁻¹ with a capacity restriction of 500 mA h g⁻¹. Fig. 6a displays the selected discharge–charge curves of the solid-state Li–O₂ battery with CNTs air cathode, and the corresponding discharge and charge capacities against cycle number is exhibited in Fig. 6b. The battery with the CNT air cathode could run 20 cycles and the discharge capacity reached the fixed value. However, a gradual decrease of the cut-off voltage was observed during the cycling process, the 15th discharge voltage of the CNT air cathode dropped to 2.34 V. Meanwhile, from the cycling performance in Fig. 6b, it was noted that the partial recharge capacity did not reached a fixed value when the recharge voltage reached 4.5 V, indicating that the recharge was incomplete. On the contrary, the cycling performance of the CNTs@Pd electrode was significantly improved; it could be found that the battery with the CNTs@Pd electrode maintained 50 cycles and no obvious fading of the discharge and charge capacities could be found, as shown in Fig. 6c and d. Although a decrease in the cut-off voltage was also observed during the cycling process, this situation was better than that of the CNT electrode, as the discharge voltage was higher than that of the CNT electrode at the same number of cycles. The 15th discharge voltage of the CNTs@Pd air cathode was about 2.71 V, which was about 370 mV higher than that of the CNT air cathode, and the discharge voltage was as high as 2.48 V after 50 cycles. In addition to the enhanced ORR catalytic activity, the CNTs@Pd electrode also displayed better OER performance. From Fig. 6a and c it can be found that the 1st charge process of the CNTs@Pd electrode stopped at 4.20 V, about 300 mV lower than that of the CNT air cathode. After 50 cycles, the CNTs@Pd electrode still exhibited excellent rechargeable behavior and the charge voltage stopped at 4.42 V. The lower charge overpotential under the capacity-limiting test is a common phenomenon, because a slight amount of the discharge product can easily be decomposed. The better cycling performance of the CNTs@Pd electrode compared to the CNT electrode might be attributed to the following: firstly, experimental work and calculations have indicated that the functional group or metal catalyst can act as nucleation sites for the discharge product, leading to the small and uniform deposition of Li₂O₂ particles.^38,39 The introduced Pd nanoparticles in the CNTs@Pd electrode induced Li₂O₂ to preferentially form near the active site of Pd, which led to the observed smaller sized Li₂O₂. As confirmed in Fig. 4, the formed smaller sized Li₂O₂ is easier to decompose during the following recharge process.⁴⁰ Secondly, it was found from Fig. 3c that the crystallinity of the discharge product Li₂O₂ was poor for the CNTs@Pd electrode, and the low crystalline Li₂O₂ enhanced the charge-transfer property of the cathode, as confirmed in Fig. 4, leading to the discharged CNTs@Pd electrode suffering lower surface passivation than the discharged CNT electrode, which facilitated the decomposition of Li₂O₂. The combination of the above two reasons resulted in the more thorough decomposition of the discharged CNTs@Pd electrode, and the complete decomposition of the discharge product tends to result in excellent cycling performance, since accumulation of the isolated discharge product during cycling leads to serious passivation of the air cathode, and will ultimately lead to a low capacity, high overpotential and limited cycling performance in a Li–O₂ battery.

Fig. 6 The discharge–charge curves and cycling performances of the Li–O₂ battery with CNT (a and b) and CNTs@Pd (c and d) electrodes.

In addition, the improved cycling performance of the CNTs@Pd electrode also attributed to the suppression of side reactions. As mentioned in previous reports, the cycling performance of a non-aqueous Li–O₂ battery is greatly affected by unwanted side reactions such as the decomposition of carbon and electrolyte, which results in the formation of Li₂CO₃ and deterioration of the electrochemical performance. In our experiment, although we replaced liquid electrolytes with solid-state electrolyte, avoiding the decomposition problem of the liquid electrolyte, the unwanted side reaction between carbon materials and Li₂O₂ could not be totally avoided, and this situation was confirmed by FTIR, as shown in Fig. 7. For the CNT electrode, weak Li₂CO₃ peaks could be found after 6 cycles and evident Li₂CO₃ peaks could be detected after 12 cycles, indicating that isolated Li₂CO₃ was generated and this gradually accumulated with an increase in the cycling number. For the CNTs@Pd electrode, although no significant Li₂CO₃ peaks could be observed after 6 cycles, weak transmission peaks ascribed to Li₂CO₃ could be detected after 12 cycles and more obvious Li₂CO₃ peaks could be detected after 50 cycles. It seems that the decomposition of carbon also existed in the CNTs@Pd electrode, however, the deposition of Pd truly effectively reduced the formation of Li₂CO₃, since the peak intensity of Li₂CO₃ after 50 cycles was significantly weaker than that of the CNT electrode after 12 cycles. The main reason probably behind this is that for the CNT electrode, the discharge product Li₂O₂ covered the surface of CNTs, as shown in the schematic in Fig. S2a.† Li₂O₂ is in direct contact with the CNTs and easily undergoes oxidative decomposition to form Li₂CO₃ during the recharge process. Unlike the CNT electrode, although Li₂O₂ also covered the surface of the CNTs, the loaded Pd nanoparticles separate the direct contact of Li₂O₂ with the CNTs to some extent, as shown in the schematic in Fig. S2b;† in other words, the loaded Pd nanoparticles play a role in limiting the degree of the unwanted side reaction.

Fig. 7 FTIR spectra of the CNT and CNTs@Pd electrodes at different cycles.

It should be noted that the generation of Li₂CO₃ was not thoroughly avoided after long-term cycling for the CNTs@Pd electrode, which was due to the fact that Pd nanoparticles were not fully coated on the surface of the CNTs, and the exposed region with no Pd loaded could induce the possible side reaction. It is speculated that the cycling performance could be further improved if the side reaction could be thoroughly avoided, and a possible solution is to design a carbon-free cathode. However, for the practical application of Li–O₂ batteries, the energy density is of great importance, and the density of carbon-free material is much larger than most of the carbon materials, so carbon materials are still widely employed, as discussed in the Introduction. Finally, for different carbon materials, it should be noted that properties such as density, specific surface area, porosity, and even defects, are various, and if the cathode is assembled with different carbon materials, the specific capacity is different even though the cathode has the same loading weight of carbon, since they cannot provide the same space for the deposition of Li₂O₂. In our case, we choose CNTs as the carbon material to investigate the catalytic activity of Pd on the Li–O₂ battery using a solid-state air cathode and improve the electrochemical performance. In the future, other suitable carbon materials and catalysts can also be attempted to further improve the battery performance.

4. Conclusion

In summary, Pd nanoparticles were deposited on CNTs via a wet chemical method and the as-prepared composite was firstly used in a solid-state air cathode. The obtained composites maintained the tubular structure, and the CNTs@Pd electrode endowed the Li–O₂ battery with excellent electrochemical performance. The battery with the CNTs@Pd electrode delivered a capacity of 7800 mA h g⁻¹ based on the weight of CNTs + Pd (9750 mA h g⁻¹ based on the weight of CNTs), and the discharge voltage was 260 mV higher than that of the CNT electrode, which delivered a capacity of 7455 mA h g⁻¹ based on the weight of the CNTs. Meanwhile, the CNTs@Pd electrode accelerated the nucleation and growth of the solid-state discharge products and therefore improved the rate performance. In comparison to the CNT electrode, the CNTs@Pd electrode tended to induce the formation of poorly crystalline Li₂O₂. Moreover, the employment of Pd into the solid-state air cathode facilitated the achievement of a long cycle life up to 50 cycles without capacity fading, which could be attributed to the positive catalytic activity of Pd and the suppression of side reactions related to carbon and Li₂O₂. These findings are potentially important to design solid-state Li–O₂ batteries with high performance for practical application.

Acknowledgements

This research was supported by the Synergistic Innovative Joint Foundation of AEP-SCU (No. XTCX2014001).

Notes and references

1. K. M. Abraham and Z. Jiang, J. Electrochem. Soc., 1996, 143, 1.
2. P. G. Bruce, S. A. Freunberger, L. J. Hardwick and J. M. Tarascon, Nat. Mater., 2012, 11, 19.
3. G. Girishkumar, B. McCloskey, A. C. Luntz, S. Swanson and W. Wilcke, J. Phys. Chem. Lett., 2010, 1, 2193.
4. B. D. McCloskey, D. S. Bethune, R. M. Shelby, G. Girishkumar and A. C. Luntz, J. Phys. Chem. Lett., 2011, 2, 1161.
5. B. D. McCloskey, A. Speidel, R. Scheffler, D. C. Miller, V. Viswanathan, J. S. Hummelshøj, J. K. Nørskov and A. C. Luntz, J. Phys. Chem. Lett., 2012, 3, 997.
6. R. Padbury and X. Zhang, J. Power Sources, 2011, 196, 4436.
7. H. Kitaura and H. Zhou, Adv. Energy Mater., 2012, 2, 889.
8. H. Kitaura and H. Zhou, Energy Environ. Sci., 2012, 5, 9077.
9. P. Kichambare, S. Rodrigues and J. Kumar, ACS Appl. Mater. Interfaces, 2012, 4, 49.
10. X. F. Wang, D. Zhu, M. Song, S. R. Cai, L. Zhang and Y. G. Chen, ACS Appl. Mater. Interfaces, 2014, 6, 11204.
11. Z. Peng, S. A. Freunberger, Y. Chen and P. G. Bruce, Science, 2012, 337, 563.
12. M. M. O. Thotiyl, S. A. Freunberger, Z. Peng, Y. Chen, Z. Liu and P. G. Bruce, Nat. Mater., 2013, 12, 1050.
13. G. Zhao, R. Mo, B. Wang, L. Zhang and K. Sun, Chem. Mater., 2014, 26, 2551.
14. F. Li, D. M. Tang, Y. Chen, D. Golberg, H. Kitaura, T. Zhang, A. Yamada and H. Zhou, Nano Lett., 2013, 13, 4702.
15. F. Li, D. M. Tang, Z. Jian, D. Liu, D. Golberg, A. Yamada and H. Zhou, Adv. Mater., 2014, 26, 4659.
16. Z. Jian, P. Liu, F. Li, P. He, X. Guo, M. Chen and H. Zhou, Angew. Chem., Int. Ed., 2014, 53, 442.
17. Y. C. Lu, Z. Xu, H. A. Gasteiger, S. Chen, K. Hamad-Schifferli and Y. Shao-Horn, J. Am. Chem. Soc., 2010, 132, 12170.
18. Y. C. Lu, H. A. Gasteiger, M. C. Parent, V. Chiloyan and Y. Shao-Horn, Electrochem. Solid-State Lett., 2010, 13, A69.
19. B. D. McCloskey, R. Scheffler, A. Speidel, D. S. Bethune, R. M. Shelby and A. C. Luntz, J. Am. Chem. Soc., 2011, 133, 18038.
20. C. J. Barile and A. A. Gewirth, J. Electrochem. Soc., 2013, 160, A549.
21. H. Liu, Y. Zhang, D. Arato, R. Li, P. Mérel and X. Sun, Surf. Coat. Technol., 2008, 202, 4114.
22. T. C. Lin and B. R. Huang, Sens. Actuators, B, 2012, 162, 108.
23. Y. G. Li, W. Zhou, H. L. Wang, L. M. Xie, Y. Y. Liang, F. Wei, J. C. Idrobo, S. J. Pennycook and H. J. Dai, Nat. Nanotechnol., 2012, 7, 394.
24. C. W. Yang, X. G. Hu, D. L. Wang, C. S. Dai, L. Zhang, H. B. Jin and S. Agathopoulos, J. Power Sources, 2006, 160, 187.
25. U. J. Kim, C. A. Furtado, X. M. Liu, G. G. Chen and P. C. Eklund, J. Am. Chem. Soc., 2005, 127, 15437.
26. Y. Wang, Z. Iqbal and S. Mitra, J. Am. Chem. Soc., 2006, 128, 95.
27. J. Lu, Y. Qin, P. Du, X. Y. Luo, T. P. Wu, Y. Ren, J. G. Wen, D. J. Miller, J. T. Miller and K. Amine, RSC Adv., 2013, 3, 8276.
28. S. H. Oh, R. Black, E. Pomerantseva, J. H. Lee and L. F. Nazar, Nat. Chem., 2012, 4, 1004.
29. Y. C. Lu, D. G. Kwabi, K. P. C. Yao, J. R. Harding, J. Q. Zhou, L. Zuin and Y. Shao-Horn, Energy Environ. Sci., 2011, 4, 2999.
30. E. Yilmaz, C. Yogi, K. Yamanaka, T. Ohta and H. R. Byon, Nano Lett., 2013, 13, 4679.
31. Y. Lei, J. Lu, X. Luo, T. Wu, P. Du, X. Zhang, Y. Ren, J. Wen, D. J. Miller, J. T. Miller, Y. K. Sun, J. W. Elam and K. Amine, Nano Lett., 2013, 13, 4182.
32. J. S. Hummelshoj, J. Blomqvist, S. Datta, T. Vegge, J. Rossmeisl, K. S. Thygesen, A. C. Luntz, K. W. Jacobsen and J. K. Nørskov, J. Chem. Phys., 2010, 132, 071101.
33. M. D. Radin, J. F. Rodriguez, F. Tian and D. J. Siegel, J. Am. Chem. Soc., 2012, 134, 1093.
34. M. D. Radin and D. J. Siegel, Energy Environ. Sci., 2013, 6, 2370.
35. J. Zhang, Y. Zhao, X. Zhao, Z. Liu and W. Chen, Sci. Rep., 2014, 4, 1.
36. D. Xu, Z. L. Wang, J. J. Xu, L. L. Zhang and X. B. Zhang, Chem. Commun., 2012, 48, 6948.
37. Y. H. Chen, S. A. Freunberger, Z. Q. Peng, O. Fontaine and P. G. Bruce, Nat. Chem., 2013, 5, 489.
38. L. J. Wang, J. Zhang, X. Zhao, L. L. Xu, Z. Y. Lyu, M. Lai and W. Chen, RSC Adv., 2015, 5, 73451.
39. X. Jie, D. Mei, X. Li, X. Wu, D. Wang, G. L. Graff, W. D. Bennett, Z. Nie, L. V. Saraf and I. A. Aksay, Nano Lett., 2011, 11, 5071.
40. Y. Hu, X. Han, F. Cheng, Q. Zhao, Z. Hu and J. Chen, Nanoscale, 2013, 6, 177.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra15206b

---