Intercalation of H3O+ into LMO latticeThe X-ray diffraction (XRD) pattern and scanning electron microscope (SEM) images of the synthesized LMO are shown in Supplementary Figs. 1, 2, respectively. The results indicate that the powdery LMO sample can be well indexed to spinel structure with a space group of Fd − 3 m. For the electrochemical tests of LMO in this work, studies in the aqueous electrolytes (1 M Li2SO4 in ultra-pure water with or without additive) are performed in a three-electrode cell (the cell structure is shown in Supplementary Fig. 3a) and studies in organic electrolytes (1 M LiPF6 in EC:DMC = 1:1) are performed in a coin cell (CR2032, the cell structure is shown in Supplementary Fig. 3b). It is crucial to find a reference electrode with stable potential to study the electrochemical behavior of the working electrode. Unstable voltage plateaus or mismatch of capacity between positive and negative electrodes will cause the charge/discharge curves of the working electrode to deviate from reality, causing the transition of voltage plateaus voltage slope, or leading to the emergence of additional voltage plateaus. Since Li metal has a large theoretical capacity (3860 mAh g−1) and a stable operating voltage plateau, it is commonly used as the negative electrode to provide a consistent voltage reference in organic electrolytes. However, Li metal is sensitive to moist atmosphere. Therefore, we assemble LMO and Li metal negative electrode into well-sealed coin cells to test the electrochemical properties of LMO in an organic electrolyte. For aqueous electrolytes, the drastic reaction between Li metal and H2O hinders the application of Li metal. However, the three-electrode cell with a reference electrode (saturated calomel electrode, SCE for this work) can provide a stable reference voltage for studying the electrochemical behavior of the working electrode (LMO for this work). Therefore, we perform electrochemical tests for LMO in aqueous electrolytes using the three-electrode cell. We first tested the charge/discharge properties of LMO in an aqueous electrolyte (three-electrode cell, 30 mL 1 M Li2SO4) and an organic electrolyte (coin cell, 0.05 mL 1 M LiPF6 in EC:DMC = 1:1) at the same specific current of 0.5 A g−1. According to the standard electrode potentials of SCE (0.245 V vs. SHE) and Li metal (−3.042 V vs. SHE), we converted the charge/discharge potential ranges of LMO in the aqueous electrolyte and the organic electrolyte to 0–1.1 V (vs. SCE) and 3.2–4.3 V (vs. Li/Li+), respectively, to ensure that the potential range of LMO in the two electrolytes is consistent and that the theoretical capacity can be maximized. The charge/discharge curves of LMO (Supplementary Fig. 4a, b) indicate that the LMO electrodes exhibit similar specific capacities and voltage plateaus in both aqueous electrolyte and organic electrolyte at the initial cycle. The initial specific capacities of LMO in an aqueous electrolyte and an organic electrolyte are 122.3 mAh g−1 and 118.25 mAh g−1, and the coulombic efficiencies are 94.5% and 84.3%, respectively (Supplementary Fig. 5 and Fig. 1a). The coulombic efficiency of an electrode is the percentage obtained by dividing the discharge capacity by the charge capacity. For an electrode without side reactions, the coulombic efficiency is theoretically 100%, indicating that the electrochemical reaction is completely reversible. The coulombic efficiency of LMO cycled in the aqueous electrolyte is lower than that of in organic electrolyte (Fig. 1a), indicating that more serious side reations occur in LMO cycling in 1 M Li2SO4 solution. During the subsequent cycles, the capacity retention of LMO in aqueous electrolyte is successively preferable and worse compared with organic electrolyte. As shown in Fig. 1a, in region 1 (1–140 cycles), the capacity retention of LMO in aqueous electrolyte (88.4%) is higher than that in organic electrolyte (79.6%). In Region 2 (140–300 cycles), the LMO electrode shows a quick capacity drop in aqueous electrolyte with a capacity retention of 61.6% after 300 cycles. The coulombic efficiency of LMO (Supplementary Fig. 5 and Fig. 1a) also begins to decrease in Region 2, indicating that the (de)intercalation of Li+ ions in LMO electrode becomes irreversible. To identify the precise charge storage mechanism of LMO in aqueous electrolyte, in situ electrochemical quartz crystal microbalance (EQCM) measurements were performed to monitor the mass variation of LMO during cycling in aqueous electrolyte. Generally, Mn dissolution induced by disproportionation reaction of Mn3+ in a proton environment will decrease the mass of LMO25,26. However, as shown in Fig. 1b and Supplementary Fig. 6, the mass of LMO gradually increases associated with cycling. It should be noted that the electrode tested by EQCM is spray-coated with LMO and does not contain a conductive agent and binder. Therefore, the increase or decrease in electrode mass is the mass change of LMO particles. We found that there is a corresponding relationship between the mass change of LMO and the cycle performance. When the mass increase rate of LMO slows down, the cycle performance also starts to degrade rapidly. The EQCM curves (bottom) and corresponding charge/discharge curves (top) of LMO are shown in Fig. 1c, and the mass variation analyses are shown in Supplementary Fig. 7. During the beginning of the charge process, the LMO exhibits mass loss with ∆m/dq of −7 g mol−1 which can be attributed to Li deintercalation in LMO lattice, and the mass loss with ∆m/dq of −14 g mol−1 at the high voltage of charge progress can be attributed to Li+ deintercalation together with Mn dissolution. During the discharge process, the mass gain ∆m/dq is 13 g mol−1, which exceeds the mass of Li+ intercalation. We consider that the asymmetrical mass change of LMO between charge and discharge in aqueous electrolyte may be attributed to H2O or H3O+ intercalation together with Li+, and the content of crystal H2O or H3O+ tends to be saturated associated with the cycling. The Tourier transform infrared spectroscopy (FTIR) spectra of LMO cycled in aqueous electrolyte (Fig. 1d) show a new peak at 1120 cm−1, which is absent in the FTIR spectra of LMO cycled in organic electrolyte (Fig. 1e). This peak may be attributed to the –OH signal caused by crystal H2O (or H3O+) in LMO lattice27. The O 1 s X-ray photoelectron spectroscopy (XPS) was performed to study the lattice –OH content in LMO surface after different cycles. As shown in Supplementary Fig. 8, LMO surface has a high lattice –OH content (30.02%) after 100 cycles. Moreover, with the further cycling of the LMO electrode, the signal of lattice –OH still increases slowly (35.16% for 200 cycles and 42.48% for 300 cycles). The result shows that the lattice –OH content in LMO surface increases and tends to slow down with the cycling. To determine the intercalated species and their precise sites in LMO lattice, the formation energies (EF) of LMO with H2O or H3O+ in tetrahedral sites and octahedral sites are calculated by density functional theory (DFT) based on the following expression:$${E}^{{{\rm{F}}}}={E}_{{{\rm{Total}}}}^{{{\rm{F}}}}-{\mu }_{{{\rm{LMO}}}}-{\mu }_{{{{\rm{H}}}}_{2}{{\rm{O}}}/{{{\rm{H}}}}_{3}{{{\rm{O}}}}^{+}}$$
(1)
where \({E}_{{\mbox{Total}}}^{{\mbox{F}}}\) is the total energy of the LMO with H2O or H3O+; \({\mu }_{{\mbox{LMO}}}\) and \({\mu }_{{{\mbox{H}}}_{2}{\mbox{O}}/{{\mbox{H}}}_{3}{{\mbox{O}}}^{+}}\) are the chemical potentials of LMO and H2O or H3O+, respectively. For most molecular, chemical potentials are equal to the DFT total energies of their ground states.Fig. 1: The electrochemical analyses, spectral characterizations and theoretical calculations of the phenomenon of H3O+ intercalated into LiMn2O4 lattice.In this figure, all the electrochemical performances of LiMn2O4 (LMO) electrodes in aqueous electrolyte were tested in a three-electrode cell with 30 mL 1 M Li2SO4 and that of in organic electrolyte were tested in a CR2032 coin cell with 0.05 mL 1 M LiPF6 in EC:DMC = 1:1. a Cycle performance of LMO in aqueous electrolyte and in organic electrolyte at a specific current of 0.5 A g−1 (The initial capacity of LMO electrode in organic electrolyte and aqueous electrolyte are 122.3 mAh g−1 and 118.25 mAh g−1, respectively). b The EQCM testing (black curve) of LMO electrode during cycling in 30 mL 1 M Li2SO4 (violet curve). c Analyses of mass changes of LMO during a single charge/discharge curve. The FTIR spectra of LMO at different stages of the charge/discharge progress in (d) aqueous electrolyte and (e) organic electrolyte. f The formation energy of LMO with H2O/H3O+ in various cation sites. The XRD spectra of LMO charged in (g) aqueous electrolyte and (h) organic electrolyte and after immersion in water for 12 h.The calculation models are placed in Supplementary Fig. 9 and the DFT calculation results are shown in Fig. 1f. The EF of H2O in tetrahedral sites and octahedral sites are +1.85 eV and +2.42 eV, respectively, which indicates that the H2O is thermodynamically unfavorable in both tetrahedral sites and octahedral sites. However, the EF of H3O+ in tetrahedral sites and octahedral sites are −2.55 eV and −2.36 eV, respectively, suggesting that H3O+ is thermodynamically favorable in LMO lattice and H3O+ may spontaneously intercalate when Li site vacancies are formed. We immerse the LMO electrodes (fully charged in aqueous and organic electrolytes, respectively) in ultra-pure water for 12 h and perform XRD on them. Figure 1g, h show that the (111) peak of LMO cycled in organic electrolyte significantly shifted to a lower angle after immersion in ultra-pure water. As predicted, the spontaneous insertion of H3O+ causes the expansion of the LMO lattice, resulting in a shift of (111) peak. We monitor the dynamic pH value change of 1 M Li2SO4 during charge/discharge process by using a pH meter. The results are shown in Supplementary Fig. 10, with the pH of the pristine electrolyte is 7.59. In the initial cycles, the electrolyte pH value is 7.66 in the charging state, and the value changes to 7.72 in the discharging state. The observed increase in pH value may be attributed to the insertion of H3O+. After certain cycles, the electrolyte pH values maintain 7.68 and 7.72 during charging state and discharging state, respectively. According to the above studies, we confirm that the intercalation of H3O+ into cation vacancies of LMO lattice will occur during charge/discharge processes in aqueous electrolyte. However, the role of H3O+ in the stability of LMO structure and the rapid capacity drop of LMO when the H3O+ is saturated in LMO require further exploration. We will further address these issues in subsequent chapters.Crystal H3O+ enhances the structural stability of LMOTo investigate the effects of crystal H3O+ on the structure stability of LMO, we perform XRD on the fully charged LMO in aqueous electrolyte and organic electrolyte before and after immersion in ultra-pure water. The LMO containing crystal H3O+ exhibits superior structural stability, as the overall XRD spectrum in Supplementary Fig. 11a does not change before and after immersion in ultra-pure water. Supplementary Fig. 11b shows that the XRD spectrum of LMO electrode charged in organic electrolyte generated additional peaks after immersion in ultra-pure water, indicating that the LMO without crystal H3O+ undergoes structural degradation due to proton electrophilic attack. To understand the effect of crystal H3O+ on the structure evolution of LMO during the charge/discharge process, the Mn 2p XPS spectra of LMO cycled in aqueous electrolyte after different cycles are shown in Fig. 2a. It can be concluded that the ratio of Mn4+ on LMO surface increases with cycling. It can be seen in Supplementary Fig. 12, that in LMO, Li-ions occupy the tetrahedral sites. When crystal H3O+ occupies one octahedral site, the nearby two Li-ions will be subject to repulsive force and are not suitable to occupy the tetrahedral sites, because the H3O+-occupied octahedral site and the two LiO4 tetrahedral sites are face-sharing. With the reduction of Li+ content, to maintain charge balance, the surrounding Mn3+ increases to Mn4+. The Jahn-Taller (J-T) inactive Mn4+ can suppress the cooperative J-T effect and avoid the Mn dissolution. Electron energy loss spectroscopy (EELS) spectra of Mn L3,2-edge and O K-edge are performed from surface to subsurface regions for LMO cycled in aqueous electrolyte. The high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image with EELS line-scan path is shown in Fig. 2b and the EELS spectra are presented in Fig. 2c. In O-K edge spectra, there are three main peaks which are labeled as A1, A2, and A3. Those peaks correspond to transition of electrons from O 1 s to the hybridized bands of O 2p and Mn 3d. The increase in the intensity ratio of the A1 peak to A3 peak, from the surface to the interior phase, indicates a reduction in the Mn valence state of the MnO628. The A2 peak may correspond to the inserted H3O+ in the surface lattice. The O-K edge indicates the H3O+ concentration gradient distribution and the reduction of Mn valence from surface to subsurface of LMO. The peaks of Mn-L3 and Mn-L2 exhibit chemical shifts to higher energy loss and a higher intensity ratio of L2/L3, fully demonstrating the increase of Mn valence states on the surface of LMO28. The etching XPS is also used to analyze the H3O+ content distribution from the surface to the interior phase of the cycled LMO. As shown in Supplementary Fig. 13, the lattice –OH signal decreases from surface to interior phase of LMO, indicating the gradient distribution of H3O+ on the LMO surface. The above analyses of LMO surface and subsurface through XPS and HAADF-STEM with EELS measurements demonstrate the formation of a gradient Mn4+-rich protective shell by crystal H3O+ (Fig. 2d). Since Mn4+ is J-T inert, this protective shell may protect the spinel structure of LMO from H+ attack during cycling in aqueous electrolyte.Fig. 2: The structural evolution of the LMO in different electrolytes.a The Mn 2p XPS spectra of LMO after different cycles. b HAADF-STEM images with EELS line-scan path of LMO. c Selected EELS O K-edge and Mn L3,2-edge spectra at different depths of LMO after cycling in aqueous electrolyte. d Scheme of the surface structure of LMO after a certain number of cycles. e HRTEM image and (f) FFT analysis of LMO cycled in organic electrolyte. g Partial in situ XRD pattern of LMO during the first charge/discharge in organic electrolyte. h HRTEM image and (i) FFT analysis of the LMO cycled in aqueous electrolyte. j Partial in situ XRD pattern of the LMO during the first charge/discharge in aqueous electrolyte. The in situ XRD measurements of LMO electrodes in aqueous electrolyte were tested in a XRD-WB cell with 0.05 mL 1 M Li2SO4 and that of in organic electrolyte were tested in a XRD-LIB cell with 0.05 mL 1 M LiPF6 in EC:DMC = 1:1.The high resolution transmission electron microscope (HRTEM) image and corresponding Fast Fourier Transform (FFT) analysis of the pristine LMO are provided in Supplementary Fig. 14. The as-prepared LMO shows a highly crystallized spinel structure with continuous lattice fringes. After 50 cycles in organic electrolyte, the LMO shows amorphous regions and distorted lattice fringes (Fig. 2e), indicating severe structural degradation. The corresponding FFT pattern (Fig. 2f) shows that the spots in red circles can be indexed to the spinel LMO and the extra spots in yellow circles can be indexed to the Li-rich Li2MnO3. The attack of proton leads to Mn dissolution and further triggers the migration of the Mn in LMO29. This dissolution and migration lead to the generation of Li-rich Li2MnO3. While in aqueous electrolyte (Fig. 2h), the cycled LMO maintains the well-resolved spinel structure. The FFT pattern (Fig. 2i) further confirms the spinel LMO without an extra phase. Even after 200 cycles in aqueous electrolyte, the HRTEM image and FFT analysis of LMO (Supplementary Fig. 15a, b) still display intact cubic spinel phase, suggesting that the spinel structure is well preserved in aqueous electrolyte. The XRD spectra presented in Supplementary Fig. 16 additionally demonstrate that LMO will undergo structural degradation in organic electrolyte during charge/discharge processes, while its structural stability is higher in aqueous electrolyte. We carefully performed in situ XRD coupled with charge/discharge process on LMO in the organic electrolyte and the aqueous electrolyte, respectively. The organic in situ XRD cell (LIB-XRD) structure and aqueous in situ XRD cell (WB-XRD) structure are shown in supplementary Fig. 17a, b, respectively. The electrode parameters, electrolyte dosage, specific current, and other details of the in situ cells are in the methods. The in situ XRD full spectra (Supplementary Figs. 18–20) indicate that the structural evolution of LMO during charge/discharge in organic electrolyte and aqueous electrolyte is similar. To further study the subtle differences in the structural evolution of LMO in different electrolytes, the partial in situ XRD results are enlarged in Fig. 2g. The selected peak at approximately 63° is assigned to the (440) peak of LMO, which is free from the hindrance of the substrate. The (440) peak shifts to a high/low angle, which correlates with the electrochemical (de)intercalation of Li+. The (440) peak shifts smoothly in the low voltage region and splits in the high voltage region, confirming that LMO undergoes a monophasic solid solution reaction at high Li content and a biphasic reaction at low Li content30,31. A significant difference in Mn–O bond lengths between undistorted octahedra and J-T distorted octahedra contributed to the abrupt variation in lattice parameters. Figure 2j shows that the phase transition of LMO in aqueous electrolyte is similar to that in organic electrolyte. Notably, the biphasic transition in the high voltage region is suppressed and even disappears after further cycling (partial in situ XRD pattern of LMO after 200 cycles in Supplementary Fig. 21). The suppression of cooperative J-T effect mitigates the elongation of the Mn–O bonds, leading to a slight variation in lattice parameter. The calculated a lattice parameter change in Supplementary Fig. 22 further demonstrates that the lattice variation of LMO is suppressed during charge/discharge in aqueous electrolyte32. In addition, consistent with existing reports, the strong hydrogen bond between crystal H2O and lattice H could effectively mitigate the H attack on lattice O, thereby alleviating the migration and dissolution of transition metals24.Crystal H3O+ leads to sluggish Li+ diffusion kineticsAs mentioned above, crystal H3O+ effectively improves the structural stability of LMO, but the cycle performance of LMO in aqueous electrolyte is still unsatisfactory. As shown in Fig. 1a, b, although the cycle performance of LMO in aqueous electrolyte is higher than that in organic electrolyte in the early stage, as the cycle number increases, its capacity retention decreases significantly. To reveal the culprit of sudden capacity drops in Region 2 (Fig. 1a), the electrochemical characteristics of LMO were further investigated. Figure 3a, d provide the corresponding charge/discharge curves. It can be seen that similar charge/discharge processes are exhibited upon initial cycles, with two typical charge/discharge plateaus of LMO observed at 4.05 and 4.15 V (vs. Li+/Li) in organic and aqueous electrolytes. During cycling in organic electrolyte, the capacity of LMO gradually decreases and the two typical charge/discharge plateaus are maintained. While in aqueous electrolyte, the two typical plateaus of charge/discharge curves are weakened and the voltage hysteresis is aggravated during cycling. Supplementary Figs. 23, 24 provide the dQ/dV analyses to compare the voltage variation of LMO in the two electrolytes. The enlarged areas of dQ/dV in Fig. 3b−e show that, in aqueous electrolyte, the peaks corresponding to the two plateaus transform into bulges. These results suggest that the LMO suffers from severe electrode polarization and dramatic electrochemical kinetic deterioration in aqueous electrolyte. The rate performance of the pristine LMO and after 200 cycles in the two electrolytes are provided in Fig. 3c, f. The cycled LMO electrode in organic electrolyte exhibits slightly lower capacity but comparable rate performance to the pristine one (Fig. 3c). In comparison, in aqueous electrolyte, the pristine LMO shows a better rate performance due to the high Li+ conductivity of its electrolyte (Fig. 3f). It is noteworthy that although the capacity of LMO dropped significantly after 200 cycles in aqueous electrolyte at a specific current of 0.5 A g−1 (from 106.67 mAh g−1 to 78.7 mAh g−1), it recovered to a high specific capacity of 105.04 mAh g−1 at a low specific current of 0.1 A g−1 (Fig. 3f). The poor rate capability of the electrode indicates that the Li+ diffusion of LMO is challenged after cycling in aqueous electrolyte. The crystal H3O+ is usually firmly anchored in the lattice due to hydrogen bonds and its molecular size, which may block the tetrahedron-octahedron-tetrahedron diffusion pathway of Li+10. We further verified the adverse effect of crystal H3O+ on Li diffusion by removing crystal H3O+. The LMO cycled after 300 cycles in aqueous electrolyte was heated to 150 °C for 12 h to remove the crystal H3O+. The charge/discharge curves of the cycled LMO before and after heat treatment are shown in Fig. 3g, h. Unlike the cycled LMO with low capacity (33.2 mAh g−1) and severe voltage hysteresis, the LMO after removing crystal H3O+ shows redux capacity (107.4 mAh g−1) and reduced polarization. However, the removal of the H3O+ usually accompanies the loss of lattice O, and the stability of LMO is significantly compromised, resulting in a rapid capacity drop in subsequent cycles (Fig. 3g). Therefore, more effective strategies should be implemented to address the unlimited insertion of the H3O+. These electrochemical analyses demonstrate that the capacity drop of the LMO in aqueous electrolyte is mainly dominated by the sluggish Li+ diffusion kinetics. Although crystal H3O+ in the LMO surface lattice can enhance structural stability, its excessive occupation in the interstitial sites hinders the Li+ diffusion, resulting in a rapid capacity decline of LMO. Regulating the H3O+ concentration in LMO in aqueous electrolytes is of paramount importance for enhancing LMO performance.Fig. 3: Electrochemical performance analysis of the LMO in different electrolyte.In this figure, all the electrochemical performances of LMO electrodes in aqueous electrolyte were tested in a three-electrode cell with 30 mL 1 M Li2SO4 and that of in organic electrolyte were tested in a CR2032 coin cell with 0.05 mL 1 M LiPF6 in EC:DMC = 1:1. a Charge/discharge curves and (b) dQ/dV analysis of LMO in organic electrolyte at a specific current of 0.5 A g−1. c The rate capability of pristine LMO and cycled LMO in organic electrolyte. d Charge/discharge curves and (e) dQ/dV analysis of LMO in aqueous electrolyte at a specific current of 0.5 A g−1. f The rate capability of pristine LMO and cycled LMO in aqueous electrolyte. g The cycle performance and (h) the charge/discharge curves of the LMO at a specific current of 0.5 A g−1 before and after heating at 150 °C.Improving the cycle performance of LMO by regulating the hydrogen bond networks of aqueous electrolyteIt has been reported that the addition of solvents that form hydrogen bonds with H2O can decrease the number of H2O in the Li+ solvation shell, thereby broadening the electrochemical stability window of aqueous electrolyte33,34. Whether the reduction of H2O population at electrode/electrolyte interface plays a role in the control of crystal H3O+ content in LMO? For verification, we added different content of acetonitrile (ACN) to 1 M Li2SO4, since ACN can accumulate on the electrode surface and form hydrogen bond interactions with H2O35,36. Fig. 4a displays the FTIR spectra of 1 M Li2SO4 with different content of ACN and the pure ACN. The peak at 2253 cm−1 can be attributed to C ≡ N stretching vibration. The peak is blue-shifted when added to 1 M Li2SO4 electrolyte, indicating the formation of the hydrogen bonds between ACN and H2O. The Raman spectra of 1 M Li2SO4 with different content of ACN and the pure ACN are shown in Fig. 4b, the peaks at ~3250 cm−1 and ~3412 cm−1 are assigned to the asymmetric/symmetric O–H stretching vibration from hydrogen bonds, respectively, and the peak at ~3612 cm−1 corresponds to free O–H in H2O molecules. With the addition of the ACN, the free H2O molecules and the asymmetric/symmetric O–H stretching vibration ratio are reduced. The Raman spectra reveal that more water molecules are included in the hydrogen network, and the hydrogen network is regulated due to the addition of ACN. The linear sweep voltammetry (LSV) curves of 30 mL 1 M Li2SO4 with different content of ACN are tested in a three-electrode cell. Supplementary Fig. 25 shows that the electrochemical windows are gradually broadened with the concentration of ACN. These results indicate that the addition of ACN could regulate the hydrogen network and weaken the activity of H2O. According to XPS spectra of N 1 s and C 1 s in Supplementary Figs. 26, 27, the addition of ACN decomposes on the electrode surface and covers the LMO surface, which may further weaken the activity of H2O on the LMO surface. The O 1 s XPS spectra (Supplementary Fig. 28) are applied to detect the crystal H3O+ in cycled LMO and the qualitative crystal H3O+ content is shown in Fig. 4c. By using ACN as an additive, the lattice –OH content in cycled LMO (after 100 cycles) decreased with increasing ACN content. The electrochemical impedance spectroscopy (EIS) spectra of LMO in 1 M Li2SO4 with different content of ACN are shown in Supplementary Fig. 29. With the excessive insertion of H3O+, H3O+ will block the diffusion channel of Li+ and increase the charge transfer resistance on LMO surface. With the addition of ACN, the charge transfer resistance of LMO surface is greatly reduced. The galvanostatic intermittent titration technique (GITT) measurements in Supplementary Fig. 30 also show that LMO exhibits ion diffusion efficiencies of ~4.51 × 10−10 cm2 S−1, ~1.44 × 10−9 cm2 S−1, and ~1.23 × 10−9 cm2 S−1 for LMO cycled 1 M Li2SO4 with 0%, 10%, and 50% ACN additive, respectively. These results indicate that ACN inhibits the excessive insertion of H3O+, reduces the blocking of the Li+ diffusion channels, and effectively improves the charge transfer on the LMO surface. The cycle performances of LMO in different electrolytes (1 M Li2SO4 with 0%, 10%, 50% ACN and 1 M LiPF6 in EC:DMC = 1:1) are shown in Fig. 4d. The LMO exhibits significantly improved cycle performance in 30 mL 1 M Li2SO4 with 50% ACN compared with 1 M LiPF6 in EC:DMC = 1:1 and 30 mL 1 M Li2SO4 with 0% and 10% ACN. In addition, the coulombic efficiency of LMO electrode increases from ~95% (1 M Li2SO4) to ~99% (1 M Li2SO4 with 50% ACN). We further investigate the long-term cycle stability of LMO electrode in 1 M Li2SO4 with 0% and 50% ACN. As shown in Fig. 4e, after 10,000 cycles, LMO shows a capacity retention of 82.1% with a coulombic efficiency of 99.5% in 1 M Li2SO4 with 50% ACN. The related charge/discharge curves show that LMO cycled in 1 M Li2SO4 with 50% ACN (Supplementary Fig. 31) exhibits a reduced polarization and a superior capacity retention than that of LMO cycled in 1 M Li2SO4 (Supplementary Fig. 32). These results indicate that an appropriate amount of crystal H3O+ can not only stabilize the LMO lattice but also avoid the slow Li+ diffusion kinetics, thus making the cycling performance of LMO in aqueous electrolytes higher than that in organic electrolytes. The rate performance of the LMO in different electrolytes after 200 cycles is shown in Fig. 4f, which once again proves that the crystal H3O+ content directly affects the diffusion of Li+. Figure 4g illustrates that in organic electrolyte, LMO undergoes structural degradation and capacity fading due to side reactions at the electrode/electrolyte interface. In aqueous electrolyte, crystal H3O+ can enhance the structural stability of LMO benefit from the formation of a gradient Mn4+-rich protective shell. However, excess crystal H3O+ will lead to severe polarization caused by poor Li+ diffusion kinetics. With the addition of ACN, the hydrogen network of aqueous electrolyte is regulated, and ACN may decompose on the electrode surface, thereby weakening the reactivity of H2O molecules and achieving a long-term stable cycle of LMO with optimized content of the crystal H3O+.Fig. 4: The characteristic and electrochemical performance of the LMO in aqueous electrolyte with Acetonitrile additive.In this figure, all the electrochemical performances of LMO electrodes in aqueous electrolyte were tested in a three-electrode cell and that of in organic electrolyte were tested in a CR2032 coin cell. a The FTIR spectra of pure Acetonitrile (ACN) and 1 M Li2SO4 with different content of ACN. b The Raman spectra of 1 M Li2SO4 with different content of ACN. c O–H contents on LMO surface after cycling in 1 M Li2SO4 with different content of ACN for 100 cycles. d Cycle performances of LMO in 0.05 mL 1 M LiPF6 in EC:DMC = 1:1 and 30 mL 1 M Li2SO4 with different content of ACN at a specific current of 0.5 A g−1 for 1000 cycles. e Cycle performance of LMO in 30 mL 1 M Li2SO4 with 0% and 50% ACN at a specific current of 5 A g−1 for 10,000 cycles. f The rate capability of LMO in 30 mL 1 M Li2SO4 with different content of ACN after 200th cycles. g Schematic illustration of crystal H3O+ occupancy and its effect on Li+ diffusion in LMO.
Thermodynamically spontaneously intercalated H3O+ enables LiMn2O4 with enhanced proton tolerance in aqueous batteries
