Ant-Cave-Structured Nanopore-Embedded CoMn₂O₄ Microspheres with Stable Electrochemical Reaction for Li-Air Battery

The degrees of phase formation and crystallography for the as-prepared Co-Mn compounds were characterized by powder XRD. As shown in Fig. 1a, the sample without post annealing exhibited phase-pure carbonate compounds: (Co,Mn)CO₃ JCPDS 085-1109. The relatively broad diffraction peaks suggested that the nanoparticles were in an amorphous state in the as-grown samples. The sample was dispersed in a dilute citric acid solution and then heat-treated at 600 °C, as shown in Fig. 1b, in which all of the diffraction peaks were well matched to those of spinel structured CoMn₂O₄ (JCPDS 055-0685) compound. Thus, the Co/Mn intermixed spinel compounds with different molar ratios were successfully synthesized by the specified solvothermal reactions.

Zoom In Zoom Out Reset image size

Rietveld refinement of the obtained diffraction profiles was performed using GSAS.²³ The background was successfully fitted using a Chebyshev function with 12 coefficients, and a pseudo-Voigt function was used to fit the obtained diffraction profile. In the initial cycles, the lattice constants and profile coefficients of the selected function were refined. Subsequently, the atomic coordinates and thermal parameters were refined. The refinement parameters were converged such that reasonable goodness-of-fit parameters could be obtained. The structure crystallized in the tetragonal I4₁/amd space group with traces of impurity was identified as Co₃O₄. The weight percentage of the impurity phase was estimated to be less than 5 wt%. The unit cell parameters of CoMn₂O₄ were determined as a = b = 5.7197(0) Å, and c = 9.2570(1) Å. Figure 2 shows the results of the Rietveld refinement and the converged refinement parameters are summarized in Table SI (available online at stacks.iop.org/JES/167/080537/mmedia). Fractional coordinates and displacement parameters of the atoms in the unit cell are presented in Table SII.

Zoom In Zoom Out Reset image size

We further investigated the particle morphology and chemical composition of the as-synthesized CoMn₂O₄, which controlled the overall electrochemical behavior. Figure 3a show that the as-synthesized spinel compound has rather rough-surfaced spherical particles, preserving the morphology of the samples without heat treatment, as shown in Figs. S1a–S1c. The particles are in micron-size consisting of crystalline nanoparticles arranged as bunch of grapes and irregularly piled up to form a particulate electrode (Fig. 3b). The surface of the porous particles was quite rough, and we could clearly observe the primary nanoparticles interconnected with pores in high-resolution images. These unique inner-spacious morphologies of the resultant microspheres were induced by the uniformly emitting CO₂ gas during post annealing at 600 °C. The surface area was measured by N₂ adsorption/desorption analysis (Fig. S1d). From the results, we can realize that the as-prepared CoMn₂O₄ particles have a proper specific surface area with pores, which is expected to facilitate lithium-ion insertion/extraction during the overall electrochemical reaction with short pathways for rapid lithium-ion and electron conduction.

Zoom In Zoom Out Reset image size

High-resolution TEM (HR-TEM) images clearly showed that the polygonal shaped CoMn₂O₄ nanoparticles (with size less than 50 nm) were loosely packed within the microspheres (Figs. 3c and 3d). Moreover, we could clearly observe the spacious voids that were formed within the interconnected nanoparticles entirely positioned within all spherical particles. We have already reported electrochemically reactive porous CoO_x which is one of the promising spinel compounds.²⁴ In that work, via both SEM and HR-TEM, we have found that the controlled particle morphology was stable enough even after cycling at a higher current density (above 100 cycles at 500 mA g⁻¹). Thus, we can expect that the exceptional particle morphology of the ant-cave structured CoMn₂O₄ will improve the cycling properties by mitigating the stress induced by detrimental volume change. This has been discussed in a subsequent section with respect to the charge/discharge results. By dispersing the sample in the concentrated citric-acid solution followed by annealing (as shown in Fig. 3b), the overall primary particles were completely separated, while some particles were slightly agglomerated.

Finally, the chemical composition and metal oxidation state of the synthesized CoMn₂O₄ were verified by XPS. Figure 4 shows the core level spectra of Co2p and Mn2p. The Co2p spectra of CoMn₂O₄ are fitted by six peaks. The main peaks at 778.7 eV and 794.2 eV were assigned to Co(III), whereas the peak located at 779.7 eV and 795.4 eV were attributed to Co(II). The two additional satellite peaks at about 784.7 and 800.9 eV were assigned to Co2p_3/2 and Co2p_1/2, respectively. In the Mn2p spectra, the two Mn2p_3/2 peaks and two Mn2p_1/2 peaks observed at 639.7 and 651.2 eV and 640.7 and 652.4 eV were assigned to Mn(II) and Mn(III), respectively.^25–28 According to the fitting area, the Co(II)/Co(III) and Mn(II)/Mn(III) atomic ratios were 2.2 and 0.33, respectively.

Zoom In Zoom Out Reset image size

The electrochemical behavior of the given microsphere particles was evaluated by galvanostatic charge/discharge profiles (Figs. 5 and S2). The first CV measurement (Fig. S2a) was performed for up to 50 cycles in a voltage range of 0.01–3.0 V (vs Li⁺/Li) at 0.1 mV s⁻¹, to investigate the reversible Li⁺ insertion/desertion behavior of the given CoMn₂O₄ particles as a LIB anode. As shown in the CV curves, the CV profile for the first charge/discharge reaction is substantially different from that of the subsequent cycles. During the first lithiation reaction, the broad peak positioned at 1.26 V might be related to the reduction of Mn³⁺ to Mn²⁺; moreover, the prominent sharp peak at 0.46 V could be ascribed to the further reduction of the divalent transition ions, Mn²⁺ and Co²⁺ to Mn⁰ and Co⁰ metals, respectively, along with the growth of a solid-electrolyte interface (SEI) layer on the active materials surface. The two broad anodic peaks evolving the delithiation reaction were observed at 1.45 and 2.06 V and corresponded to the oxidation of Mn⁰ to Mn²⁺ and Co⁰ to Co²⁺. In the subsequent cycles, the cathodic peak moved to a slightly higher voltage region, ∼0.50 V, and became broader as compared to that in the first cycle, while the anodic peaks remained almost unchanged. The overall CV results are consistent with the previous related reports,^29–32 and we observed that the controlled morphology with pore-embedded structures for the specified CoMn₂O₄ composite electrodes was electrochemically stable enough during the successive electrochemical reactions. From the second cycle onwards, the CV curves overlapped very well, revealing the good reversibility of the electrochemical reactions. Based on the CV results, the electrochemical reactions of the electrode comprising the porous CoMn₂O₄ particles can be expressed as follows:

$$\begin{eqnarray}&&{{\rm{CoMn}}}_{2}{{\rm{O}}}_{4}+8{{\rm{Li}}}^{+}+8{{\rm{e}}}^{-}\to {\rm{Co}}+2{\rm{Mn}}+4{{\rm{Li}}}_{2}{\rm{O}}\end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray}&&{\rm{Co}}+2{\rm{Mn}}+3{{\rm{Li}}}_{2}{\rm{O}}\rightleftarrows {\rm{CoO}}+2{\rm{MnO}}+6{{\rm{Li}}}^{+}+6{{\rm{e}}}^{-}\end{eqnarray} \tag{ 2 }$$

Representative charge/discharge curves from the first to the 150th cycle of the anodes comprising the as-prepared CoMn₂O₄ particles as well as nanoparticles at a current density of 100 mA g⁻¹ are shown in Fig. 5. Although the 1st discharge capacities of porous CoMn₂O₄ electrode, ∼930 mA h g⁻¹, were smaller than that of the electrode comprisng particulate nanoparticles, ∼1130 mA h g⁻¹, the porous electrodes presented a rather stable capacity fading after 150 cycles. However, it is apparent that the electrodes simply composed of nanoparticles easily agglomerated, which finally lead to fast capacity fading and a large irreversible capacity right after the 1st discharge reaction. This is common for the conversion reaction preferred electrode materials; the reduction of metal ions to a metallic state with Li₂O formation accompanied by the SEI layer formation.^33–35 These results indicate that CoMn₂O₄ with an ant-cave-structured spherical shape exhibits a better electrochemical performance by accelerating the charge transport within the electrodes during repetitive cycling. This maximizes the overall electrochemical reactions and thus, a large capacity and good cycling stability are maintained throughout. Corresponding CV and charge/discharge profiles for carbonates, (Co,Mn)CO₃, are presented in Figs. S2b, S2c.

Zoom In Zoom Out Reset image size

To confirm the characteristic cycling behavior of the as-prepared microsphere particles, we have performed a comparative analysis between the capacity retention and the lithium-ion kinetics during the cyclic electrochemical reaction via electrochemical impedance spectroscopy (EIS) measurements. From the supplemental EIS results, we can determine the constituent resistance components (charge transfer, SEI formation etc.) for the specified pore-embedded electrode materials. The EIS spectra shown in Fig. 6 exhibit two prominent semicircles positioned in a high frequency region; the former semicircle is related to the charge-transfer mobility (R₂) and the other one is corresponding to the mass transfer region (R₃). During subsequent cycles, up to 50th discharge reactions, completely different equivalent resistances were obtained for the two electrodes; after first discharge reaction, R₂ decreasd from 60 Ω and stabilized at 50 Ω after 20th cycles for the electrode comprising microsphere CoMn₂O₄ particles; contrarily, R₂ gradually increased for the electrode using particulate nanoparticles. The eventually decreased charge transfer resistance represents stimulating lithium-ion mobility induced by the excellent intimate and effective lithium ions and oxygen diffusion in the porous electrode. This is in good agreement with the fact that the structural relaxation favoring optimized electrochemical reactions for the pore-embedded electrode materials during repetitive cycling enhanced the electrochemical performance, as shown in Fig. 5.

Zoom In Zoom Out Reset image size

The air electrodes were assembled using the CoMn₂O₄ particles, and the catalytic effects of the LAB cathode were evaluated. Herein, we determine the capacity-limited condition to avoid electrolyte exhaustion and the deep discharge state of the oxygen electrode. Figure 7 presents the charge/discharge profiles of different air cathodes measured with the limited capacity at 300 and 1000 mA h g_total⁻¹. Note that the air cathode without catalysts is also presented as a reference sample. Typically, Li₂O₂ formation occurred following the oxygen-reduction reaction (ORR) and decomposition during the charge reaction is referred to as an oxygen-evolution reaction (OER).^7,13,14

Zoom In Zoom Out Reset image size

The discharge plateau voltages of all the cells were similar to 2.75 V for the Li⁺ + O₂(g) + e⁻ → LiO_x(s) equilibrium potential at 2.96 V. Obviously, the reaction voltages during the overall OERs were different enough for the two electrode compositions. The CoMn₂O₄–incorporated electrode exhibited a lower charge overpotential, stable enough with subsequent cycling with ΔV at approximately 1.22 V; however, it was ∼1.55 V for the cathode without catalysts. Moreover, there seemed to be stable catalytic reactions of the CoMn₂O₄–incorporated air cathode in charge process with repetitive cyclic reactions. The cathode presented a stable reversible cycling for up to 30 cycles without detrimental voltage during the repetitive charge/discharge electrochemical reactions. For further comparison, the galvanostatic charge and discharge curve of electrode incorporated with nanoparticulate CoMn₂O₄, which is conducted at the same limited capacities, were plotted in Figs. S3a, S3b. Accordingly, the nanoparticulate sample performs a significant capacity fading out via cycling number which further confirms the enhancement in catalytic behavior of ant-cave structured electrode is strongly affected by the specifically designed pore-embedded morphology. The similar phenomenon also was reported in our previous publication,²⁴ which explains that the designed porous structure could accommodate the detrimental volume exchange during the reversible cyclic reactions as shown in this work.

Interestingly, we found that the charge/discharge profiles were more stable enough when the cathode without binder, PVDF, was used, as shown in Fig. S4, however, the effect of a "binder-free" in air cathode is still ambiguous. Thus, in-depth analysis and discussion considering different cathode conditions regarding the binder incorporation might be needed. Herein, we have only shown the empirical results and expect that they will stimulate the research on "binder-free" cathodes, by indicating new directions in developing "binder-free" cathode with good cycling ability.

The discharge and charge median voltages, which are a function of cycle numbers, are shown in Fig. S5 and summarized in Table SIII. We have found that the cyclic median voltages during discharging remain completely stable during charging at approximately 2.73 V while the median voltage is below 4.0 V within 30 cycles.

The substantially decreased charge overpotential with stable equilibrium charge voltage retention during the Li⁺ + O₂(g) + e⁻ → LiO_x(s) electrochemical reaction might be due to the loose packing of the porous spherical particles with the reactive catalyst nanoparticles. This expedited the formation and decomposition of LiO_x, prevented the pore- structured framework from clogging and finally led to a stable reversible redox reaction in the LAB. We believe that these rationally designed ant-cave-structured, pore-embedded, microsphere particles fully activate the nanoparticle catalysts and prevent agglomeration, which is necessary for high surface-area nanoparticles.

To further understand the catalytic effect of the proposed porous catalysts, we conducted supporting analyses using ex situ XRD/XPS to investigate the reaction products and in situ DEMS to quantify the oxygen-conversion efficiency. The ex situ XRD spectra (Fig. 8a) provide a the solid evidence for the reversible formation/decomposition of the by-product Li₂O₂ on the air cathode sides.^7,13 The reaction product, the Li₂O₂ phase, was clearly observed even after 30 cycles of discharging, while the Li₂O₂ deposits on the cathode surface completely disappeared after full charging. The complete formation/decomposition of Li₂O₂ at the controlled catalyst-incorporated cathode after subsequent cycles proves its superior reversibility of this cathode in the ORR/OER processes. The significantly reduced over-potential with excellent stable cycle retention may be due to the more effective contact between semispherical spinel catalysts, which finally leads to a small charge transfer resistance.

Zoom In Zoom Out Reset image size

XPS analysis also assists in determining the electrochemical reaction and reversibility of cathodes. We examined the CoMn₂O₄ incorporated air cathode after the 30th charge cycle, which showed only one band in the Li 1s region (Fig. 8b) that can be deconvoluted into two peaks positioned at the binding energies of 54.8 and 55.5 eV. In the spectra of the electrodes, a rather small peak attributed to Li₂CO₃ appeared at 55.5 eV along with that of the main product Li₂O₂ (54.8 eV). According to previous studies, Li₂CO₃ mainly originates from a side reaction between Li₂O₂ and CO₂ in air cathodes.^10,11 Thus, the actual concentration of Li₂CO₃ formed via subsidiary reactions such as the decomposition of the electrolyte, might be infinitesimal than that measured via XPS spectra analysis. Nevertheless, significant amounts of Li₂CO₃ were coated on the carbon-based cathode after repetitive cycling indicating that the formation of side products was not negligible even when the porous-structured catalysts were used.

Finally, the oxygen-evolution efficiency during the reversible electrochemical reaction for the specified air cathodes was measured using a real-time in situ DEMS apparatus (Fig. S6). The mutual relation of the DEMS data with the charge/discharge profiles during cycling can more quantitatively evaluate the stability of the air cathode and the flammable electrolytes; contrarily, most of the previously reported studies on carbon-based and/or binder-free cathodes have not quantitatively measured the actual oxygen consumption/evolution efficiency.^3,36–39 We have clearly demonstrated the excellent performance and stability of the CoMn₂O₄–incorporated cathode. The CO₂/O₂ molar ratio and OER efficiency (ratios of the amount of oxygen released during the charge/the amount of oxygen consumed in the discharge)⁴⁰ were 6.3% and 73% for the catalyst-incorporated air cathode, almost ∼10% higher than the OER efficiency for the Ketjen black-only cathode, (9.2% and 64%), respectively. Moreover, as shown in Fig. S6, the evolved O₂ gas was constant during the overall recharge process, indicating that most of the electrochemical reactions were induced by the designated electrochemical reaction Li₂O₂ → 2Li + O₂ with highly efficient oxygen evolution. Some CO₂ (g) evolved towards the end of the charging for both air cathodes. Our previous work and other studies indicated the presence of residual CO₂ for carbon-based cathodes due to parasitic oxidation reactions at high potentials (>4 V) during charging. This may originate from the decomposition of the electrolyte and/or carbonate byproducts on carbon during repetitive cycling.