A Lithium-ion Battery Using Partially Lithiated Graphite Anode and Amphi-redox LiMn2O4 Cathode

Abstract

Delithiation followed by lithiation of Li+-occupied (n-type) tetrahedral sites of cubic LiMn2O4 spinel (LMO) at ~4 $${{\bf{V}}}_{{{\bf{Li}}{\boldsymbol{/}}{\bf{Li}}}^{{\boldsymbol{+}}}}$$ (delivering ~100 mAh gLMO −1) has been used for energy storage by lithium ion batteries (LIBs). In this work, we utilized unoccupied (p-type) octahedral sites of LMO available for lithiation at ~3 $${{\bf{V}}}_{{{\bf{Li}}{\boldsymbol{/}}{\bf{Li}}}^{{\boldsymbol{+}}}}$$ (delivering additional ~100 mAh gLMO −1) that have never been used for LIBs in full-cell configuration. The whole capacity of amphi-redox LMO, including both oxidizable n-type and reducible p-type redox sites, at ~200 mAh gLMO −1 was realized by using the reactions both at 4 $${{\bf{V}}}_{{{\bf{Li}}{\boldsymbol{/}}{\bf{Li}}}^{{\boldsymbol{+}}}}$$ and 3 $${{\bf{V}}}_{{{\bf{Li}}{\boldsymbol{/}}{\bf{Li}}}^{{\boldsymbol{+}}}}$$. Durable reversibility of the 3 V reaction was achieved by graphene-wrapping LMO nanoparticles (LMO@Gn). Prelithiated graphite (LinC6, n < 1) was used as anodes to lithiate the unoccupied octahedral sites of LMO for the 3 V reaction.

Introduction

The global market size of lithium ion batteries (LIBs) is dramatically increasing as its application is expanding to electric vehicles and energy storage systems. Needs for high energy densities should be emphasized more than ever before in the large-scale systems. Energy densities of LIBs mostly depend on specific capacities of cathode active materials. Practical cathode materials of LIBs are mainly based on oxidation and reduction of transition metals (Co, Mn, Ni and Fe). Unfortunately, specific capacities of widely used transition-metal-based cathode materials are limited to ~100 mAh g−1 for LiMn2O4 in spinel (LMO), ~150 mAh g−1 for LiNixCoyMnzO2 in layered structure (x + y + z = 1; NMC) and < 170 mAh g−1 for LiFePO4 in olivine structure (LFP). Possibilities for reaching higher energy densities are supported by under-developing alternatives. Nickel-rich NMC (~180 mAh g−1 at x > 0.6 in LiNixCoyMnzO2)1 and overlithiated oxides (OLO; x Li2MnO3∙(1 − x) LiMO2 where M = Ni, Co and Mn; ~200 mAh g−1 at > 4.5 V)2 are the most popular candidates that battery makers are interested in. Recently, we reported that the electrochemistry of Mn4+/3+ of LMO around 3 V (Liy[Mn3+ yMn4+ 2−y]O4 at 1 < y < 2 in mixed cubic/tetragonal phase; equation 1 below) was possibly utilized in addition to the 4 V reaction of Mn3+/4+ used in LIBs (Liy[Mn3+ yMn4+ 2−y]O4 at 0 < y < 1 in cubic phase; equation 2 below) by nanosizing and wrapping LMO with a-few-layer graphene skin (LMO@Gn)3.

LiMn2O4 as a p-type (intrinsically oxidized) cathode around 3 V:

$${{\rm{LiMn}}}^{3+}{{\rm{Mn}}}^{4+}{{\rm{O}}}_{4}+x{{\rm{Li}}}^{+}+x{{\rm{e}}}^{-}={{\rm{Li}}}_{1+x}{{{\rm{Mn}}}^{3+}}_{1+x}{{{\rm{Mn}}}^{4+}}_{1-{\rm{x}}}{{\rm{O}}}_{4}\,\,$$
(1)

LiMn2O4 as an n-type (intrinsically reduced) cathode around 4 V:

$${{\rm{LiMn}}}^{3+}{{\rm{Mn}}}^{4+}{{\rm{O}}}_{4}={{\rm{Li}}}_{1-x}{{{\rm{Mn}}}^{3+}}_{1-x}{{{\rm{Mn}}}^{4+}}_{1+{\rm{x}}}{{\rm{O}}}_{4}+x{{\rm{Li}}}^{+}+x{{\rm{e}}}^{-}\,\,\,$$
(2)

Principally, more than 200 mAh g−1 can be extracted from LMO@Gn by using both electrochemical reactions.

In the practical configuration of LIB cells, lithium ions in n-type cathode materials are utilized for lithiating anode materials during charging: e.g., LiMn2O4 → Mn2O4 + Li+  + e in cathode and at the same time Li+  + e + C6 → LiC6 in graphite as an anode. In other words, reduced species, the reduction potential of which is relatively more positive to that of anode, are required for cathode materials. From a viewpoint of practical usages, however, the Mn4+ of LMO in the above-mentioned 3 V electrochemistry cannot be paired with p-type anodes such as bare graphite because LMO at 3 V is the p-type cathode that cannot be oxidized during the initial charging processes. To use the non-lithiated active materials as cathode materials, additional lithium should be provided to anode sides or lithium metal should be used as anodes. Lithium metal has been adopted as anodes for the oxygen and sulfur cathodes even if dendrite growth problems were not yet solved4,5,6. Moreover, LMO has the amphi-redox characteristics where both oxidation and reduction are possible (equations 1 and 2) due to its mixed oxidation number (Mn4+ and Mn3+) so that both lithium acceptor and donor are required in anodes.

If it were better not to use lithium metal as anode for the amphi-redox LMO, the use of prelithiated anode materials would be the answer. The partially or totally lithiated materials can be obtained by contacting host materials with lithium metal powder directly (or short-circuiting them)7,8,9,10,11. The very negative reduction potential of lithium metal drives lithium oxidation and host material reduction at the same time. The host materials are lithiated via intercalation or alloying reactions to keep charge neutrality. The prelithiation of anode is preferred because the cathode prelithiation might yield safety problems with excessive generation of heat of reaction. The prelithiation of graphite and silicon as anode materials has been reported. In addition to the lithiation7,12,13,14, the solid-electrolyte interphase (SEI) layer is pre-formed during the short-circuit prelithiation process so that a portion of electricity does not need to be consumed for the electrolyte decomposition during the first electrochemical charge15,16,17,18. Therefore, high initial coulombic efficiencies are guaranteed by the prelithiation.

In this work, we demonstrate the first case of the amphi-redox cathode material in full cell operation. The extra amount of lithium required for lithiating LiMn2O4 (amphi-redox species based on Mn3+ and Mn4+ half and half) to Li2Mn2O4 (fully reduced species based on only Mn3+) was supplied from pre-lithiated graphite (Li n C6 where n < 1). ~200 mAh gLMO −1 was reached by the use of amphi-redox LMO@Gn with the Li n C6.

Results

The 4 V electrochemistry of LMO has been popularly used in commercial LIBs, its capacity (Q4V) reaching practically ~100 mAh g−1. Even if the theoretical value of the capacity for the 3 V reaction (Q3V) is equivalent to that of Q4V at ~148 mAh g−1, it has been difficult to extract the 3 V capacity from practically used micrometer-size LMO particles: e.g., only 20 mAh g−1 was obtained for Q3V, which is much less than the theoretical value. In our previous work, we demonstrated that Q4V-level capacity (~100 mAh g−1) was also extracted from the 3 V reaction by nanosizing and wrapping LMO with a-few-layer graphene skin (LMO@Gn, Supplementary Fig. 1)3. ~200 mAh g−1 based on Q4V plus Q3V was realized in a half cell configuration of LMO@Gn||Li metal. In this case, the Li metal anode worked as an amphi-redox material.

If the safety issues caused by the dendrite growth of Li0 on anodes were concerned, the amphi-redox LMO@Gn should be paired with an amphi-redox type anode such as partially lithiated graphite (Li n C6). The cell operation of LMO@Gn||Li n C6 is different from the conventional operation of LMO||C6 (Fig. 1). The lithium ions shuttle between the 4 V reaction region of LMO and the unlithiated empty region of graphite in LMO||C6 (Fig. 1a and b). The conventional operation of typical LMO was named Q operation here because the cell capacity is Q if Q4V of LMO = Q. The conventional cells are initially in discharged states so that the cell operation is initiated by charge. On the other hand, the operation of the amphi-redox cells for doubling capacity (2Q operation) can be initiated by discharge as well as charge (Fig. 1c and Supplementary Fig. 2). In pre-discharging strategy of capacity-doubling operation (Fig. 1c), the empty 3 V reaction region of LMO@Gn is lithiated by lithium ions released from the prelithiated region of graphite to form Li2Mn2O4. The amount of lithium in the LMO@Gn becomes doubled so that its capacity is 2Q because Q4V = Q3V. From the following charging and discharging processes, lithium ions move from fully lithiated LMO@Gn to fully unlithiated C6 during charging; and from fully lithiated C6 to fully unlithiated LMO@Gn during discharging (leaving 0.1Q in the upmost potential region lithiated at N/P = 1.1). The 4 V region of LMO@Gn is paired with the lower potential region of graphite (orange-colored) for lithium transfer while the 3 V region of LMO@Gn is paired with the higher potential region of graphite (blue-colored). In pre-charging operation (Supplementary Fig. 2), graphite is fully lithiated initially and then the following discharge/charge proceeds in the same way as the pre-discharging operation does.

Graphite anodes were prelithiated by pressing passivated lithium metal powder (PLP) loaded on graphite anodes by 1-ton hydraulic press in an Ar-filled glove box (Fig. 2a). Fractional conversion of loaded PLP to lithium available for delithiation (prelithiation yield, Y Li; refer to equation 3 in Methods for definition) was estimated at 65 to 75 % by the PLP sprinkle method. Lithiation of graphite by the press-activated PLP sprinkles was confirmed electrochemically and crystallographically. The potentials of the reaction plateaus and the corresponding dQ/dV peaks along deintercalation were consistent with the previously reported values19,20,21: LiC6 (stage 1) to LiC12 (stage 2) at 105 mV; LiC12 (stage 2) to LiC18 (stage 2L) at 141 mV; and LiC36 (stage 4) to LiC72 (dilute stage) at 223 mV (Fig. 2b and Supplementary Fig. 3 for 74 %-prelithiated graphite (Li0.74C6 = LiC8; L PLP = 100 %); Supplementary Fig. 4a for over-prelithiated graphite obtained by loading excessive amount of PLP (L PLP = ~700 %); Refer to equation 4 in Methods for the definition of L PLP). It indicates that lithium was intercalated to graphite framework by the PLP-prelithiation process. Also, the peaks of graphite intercalation compounds (GICs) at LiC6 and LiC12 were clearly found in the X-ray diffraction pattern of the same graphite electrode that is supposed to be lithiated by PLP7,12,20,22 (Fig. 2c; Supplementary Fig. 4b).

The distribution of lithium ions in graphite changed after the prelithiated graphite anodes were cured in electrolyte (Fig. 2c). Both LiC6 (major) and LiC12 (minor) were clearly found in the XRD pattern of the prelithiated graphite in the absence of electrolyte, indicating that lithium intercalation was localized. After curing in electrolyte for a day, on the contrary, the intensity of LiC12 peak dramatically increased and it was difficult to find LiC6 and C6 peaks. Therefore, we concluded that the intercalated lithium species were delocalized in the presence of electrolyte. The re-distribution of GICs caused by delocalization of lithium species in graphite was monitored by open circuit voltage (OCV) of half cells at Li n C6||Li0 (Fig. 2d). The OCV was dropped, recovered and then saturated. In a dry environment, only the grains or particles of graphite directly contacting lithium metal of PLP are lithiated to form LiC6 or LiC12 (only the two GICs were considered for simplicity). The rest of them are intact in unlithiated phase (C6). A portion of PLP is not used for lithiation, existing as a form of Li0. The potential difference is developed between the GICs and C6 and even between LiC12 and LiC6 throughout electrodes. As the electrodes consisting of C6, LiC6 and LiC12 are in contact with electrolyte, ionic pathways through electrolyte are developed between the lithiated and the unlithiated phases. The initial OCV drop indicates that C6 (unlithiated phase) and LiC12 (partially lithiated phase) are lithiated by lithium ions and electrons from Li0: Li0 + LixCy → Lix+1Cy (step 1 in Fig. 2d and Supplementary Fig. 5). The increase of lithiation amount results in the OCV drop. In the period of OCV increase following the drop, the lithiation delocalization is thought to proceed dominantly between GICs: e.g., 2LiC6 + 2C6 → LiC12 + LiC12 (step 2 in Fig. 2d and Supplementary Fig. 5). Less number of more intercalated GICs at more negative potentials are converted to more number of less intercalated GICs at more positive potential so that the OCV increases. After the lithiation delocalization is completed, the potential shows no more change (step 3 in Fig. 2d and Supplementary Fig. 5). This scenario is confirmed by two extreme cases (Supplementary Fig. 6). The initial OCV drop was not observed in a deficient PLP loading case where all Li0 is completely consumed for the lithiation. On the contrary, PLP overloading case showed only the OCV drop without the OCV recovery and saturation.

Interestingly, the prelithiated graphite anodes by PLP showed improved kinetics for electrochemical delithiation and the following electrochemical lithiation when compared with bare graphite. For strict comparison with the prelithiated graphite anode, the bare graphite electrode was pressed by the same pressure. The overpotentials of prelithiated graphite were lower than those of the bare graphite (Fig. 2e). For example, the first delithiation and the last lithiation peaks were observed at 104 mV and 75 mV in the PLP-prelithiated graphite, respectively (gap = 29 mV). The potential gap of the stage 1/stage 2 transition between delithiation and lithiation was 10 mV less than that of the bare graphite: peak-to-peak gap = 29 mV for PLP lithiation and 39 mV for electrochemical lithiation. Lithium metal conductor that was left after the PLP prelithiation would be partly responsible for the improved kinetics, the existence of which was supported by Li 1s X-ray photoelectron spectra (Supplementary Fig. 7). The reversibility of lithiation/delithiation of the prelithiated graphite was guaranteed (Supplementary Fig. 8).

Prelithiation starts from the highest potential at ~0.25 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$ and then the potential decreases along prelithiation. Therefore, the final potential of the GIC is determined by the extent of prelithiation of graphite (=Y Li x L PLP where Y Li = ~70 %) depending on the loading amount of PLP. The capacity below the final GIC potential is used for accepting lithium coming from LMO at 4 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$. The minimum amount of prelithiation is 1Q used for lithiating LMO at 3 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$ during charging. Also, another 1Q should be guaranteed as the capacity of unlithiated phase of graphite ready for accepting lithium coming from LMO at 4 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$. By increasing the prelithiation amount above 1Q, extra lithium is intercalated into graphite. The lithium reservoir can compensate for the lithium deficiency caused by parasitic Li consumption reaction such as Mn deposition on anode (Mn2+ + 2LiC6 → Mn + 2Li+ + C6)23,24,25,26. However, the prelithiation amount should be limited not to drop the capacity of unlithiated graphite below 1Q. By increasing the amount of graphite at a fixed prelithiation amount, we can design the prelithiated graphite to have spare unlithiated capacity at lower potential as a safety margin for preventing lithium metal deposition. When the graphite amount increases from 2Q to 2.2Q at 1Q prelithiation, the capacity of unlithiated graphite increases from 1Q to 1.2Q (Supplementary Fig. 9a). Therefore, 0.2Q safety margin at lower potential is secured. In the same situation (graphite in 2.2Q) with higher prelithiation at 1.2Q (the prelithiation extent = 1.2Q / 2.2Q = 55 %), the potential during delithiation is kept at lower potential (potential gain) but the safety gain is lost (Supplementary Fig. 9b).

The resultant prelithiated graphite anode realized the doubled capacity of LMO@Gn composed of Q4V + Q3V (+HEMDS* in Fig. 3a and Supplementary Fig. 10a for 50 wt. % and 70 wt. % LMO in cathodes, respectively). The capacity of the amphi-redox LMO batteries (LMO@Gn||Li n C6) was estimated at 200 mAh gLMO −1 based on LMO mass after the initial activation cycles. The capacity at the 100th cycle was kept at > 90 % of the initial value after the 15 activation cycles with the capacity fade rate at ~0.2 mAh g−1 cycle−1 and the coulombic efficiency at ~100 %. The 70 wt. % LMO (equivalently, 76 wt. % LMO@Gn containing @Gn at 8.0 wt. % of LMO@Gn; Supplementary Fig. 10) in cathodes was considered as the Maginot line of the LMO amount. Only 4 wt. % margin was available for carbon black as conducting agent in the maximum LMO loading because 20 wt. % binder was required for the electrode integrity with LMO@Gn.

To secure the capacity of the 3 V reaction (Q3V) and its stability, interestingly, the activation process at low rate 0.1C during the initial 15 cycles was required. The lower cut-off cell voltage was lowered from 2.2 Vcell to 2.0 Vcell during the activation cycles. At the second cycle of the activation, Q3V (excluding capacity based on the tail reaction nearby 2.0 V) was estimated only at ~50 mAh gLMO −1 that was ~50 % of the expected capacity. After the 10th activation cycle, however, the 3 V plateau reaction (at ~2.85 Vcell for discharging and at ~3.0 Vcell for charging) increased to 100 mAh gLMO −1. At the same time, the tail reaction at 2.2 to 2.0 Vcell on discharge and at 3.65 to 3.8 Vcell on charge (Fig. 3b) decreased. The 3 V plateau reaction and the tail reaction were traded off while the sum of the two reactions was fixed at ~ 145 mAh gLMO −1 on discharge. After the activation cycle, the low cut-off voltage was set at 2.2 Vcell to secure the stability of the 3 V reaction.

It should be notified that the aforementioned cyclability of the amphi-redox cell of LMO@Gn||Li n C6 was guaranteed only after relieving problems of Mn dissolution of LMO@Gn (+HEMDS and +HEMDS* versus −HEMDS in Fig. 3a). The Mn dissolution problems of LMO have been well recognized, resulting in structural degradation of LMO and Mn deposition on anodes23,24,27,28. HF generated via the reaction between PF6 and a trace amount of water in electrolyte drives the dissolution of Mn in a divalent cation form. The capacity decrease along cycle was much more serious in the full cell configuration of LMO@Gn||Li n C6 than that in the half cell configuration of LMO@Gn||Li0 (Supplementary Fig. 11). It is probably due to deposition of dissolved Mn2+ on anodes. Mn2+ is reduced to Mn0 while lithiated graphite is delithiated: 2LiC6 + Mn2+ → 2Li+ + 2C6 + Mn0. We implemented multiple strategies to suppress the Mn deposition. Alumina-coated separators were used to scavenge Mn2+ because the metal ions are adsorbed on alumina29. Na-carboxymethyl cellulose (CMC) was used as a binder for the same purpose since Mn ions are captured by ion exchange with Na ions30. 0.1 wt. % of heptamethyl disilazane (HEMDS) was introduced to electrolytes for scavenging HF which causes Mn dissolution31,32. Vinylene carbonate (VC) was used as an additive for making enhanced solid-electrolyte interphase (SEI) layer on anodes for preventing dissolved Mn2+ from penetrating through the SEI and then delithiating graphite33.

The amount of Mn metal detected on anodes after repeated charge/discharge cycles significantly decreased after adopting the former three strategies (Fig. 3c): from 65 ppm for the base electrolyte (without any additives); to 44 ppm after using the ceramic-coated separator and VC; moreover to 32 ppm when the HF scavenger was additionally adopted. CMC binder was reported to capture dissolved Mn2+ by exchanging Na+ with Mn2+ so that it possibly prevents capacity loss caused by Mn deposition30. As the results of these multiple Mn2+-blocking strategies, the capacity retention along cycle was successfully improved as shown in Fig. 3a from −HEMDS to + HEMDS to + HEMDS* (* = CMC). Capacity of LMO@Gn was completely recovered to 200 mA gLMO −1 at 0.15C even after rate capability tests at a series of C rate up to 5C (Supplementary Fig. 10c). Cathode morphologies did not change significantly along repeated cycles of charge and discharge (Supplementary Fig. 12).

Discussion

The energy density (or capacity) gains obtained by using the double-capacity LMO in LMO@Gn||Li n C6 were determined by two factors: N/P and x LMO. N/P is the ratio of the capacity of anode to cathode in LMO@Gn||Li n C6. x LMO is the LMO fraction in cathode. The energy density (or capacity) gains were quantified by the energy density (or capacity) improvement factor, y ED (or y Q). The quantity was defined by the energy density (or capacity) ratio of LMO@Gn||Li n C6 to LMO||C6 at the same condition of N/P and x LMO (Fig. 3d; equations 5 and 6 in Methods). Smaller N/P resulted in larger gains because saving extra graphite mass increased energy densities (or capacities) of the amphi-redox cells more than those of the practical cells based on only 4 V reaction. Smaller x LMO (more amount of binders and conducting agents) decreased energy densities (or capacities) of both amphi-redox and practical 4 V cells but increased gains. The most important point is that the amphi-redox cells were superior to the practical cells over all N/P ratios in terms of energy density and capacity: y ED = ~160 % (~60 % energy density gain) and y Q = ~185 % (~85 % capacity gain) at N/P = 1.1. Also, the amphi-redox cells still showed higher energy densities and larger capacities in a cell level when compared with a practical 4 V cell containing 90 % LMO composition (x LMO = 0.9, N/P = 1.1) (Supplementary Fig. 13). 25 % more energy density and 45 % larger capacity were achieved by the amphi-redox cells at the same N/P and 70 % LMO composition.

The amphi-redox cells were initially designed at N/P = 1.7 and x LMO = 0.5 to maximize its capacity. Large amount of carbon black was used in cathode to compensate low conductivity of the tetragonal phase for the 3 V electrochemistry. However, the LMO content (x LMO) increase to x LMO = 0.7 showed negligible change in cyclability (Supplementary Fig. 10a). Also, the high N/P secures a potential margin of lithium storage. In a typical full cell design for n-type cathode||p-type anode, on the other hand, the N/P is practically set at < 1.1. One of the limitations of the amphi-redox cells of LMO@Gn||Li n C6 is its low operational voltage. To use the 3 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$ reaction as well as the 4 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$ reaction, the cut-off potential of LMO@Gn is set at < 2.5 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$, which is lower than the cut-off potentials of practical LMO cells. Therefore, lower potential region of graphite at < 0.25 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$ should be utilized with 0.05 V margin to secure the cell voltage above 2.2 Vcell. As the overpotential on LMO is developed along repeated charge/discharge cycles, the lower-potential region of graphite should be used. By increasing the N/P ratio to 1.7, 1.4Q below 0.25 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$ is reserved for lithium consumption caused by Mn deposition (1.4Q in Supplementary Fig. 14). When the N/P ratio decreased to the practical value 1.1, capacity retention of the cell at N/P = 1.1 became inferior to that of high N/P cells (N/P = 1.7) after the 50th cycle (Supplementary Fig. 10b). The main reason for the deterioration is the loss of potential margin.

In conclusion, LIB cells based on amphi-redox active materials were presented, where the capacity of LiMn2O4 spinel (LMO) as a cathode material was doubled to be 200 mAh g−1 in full cell configuration. Conventionally, fully lithiated n-type and fully unlithiated p-type electroactive species have been used as cathode and anode materials of LIBs, respectively. LMO is a popular cathode material based on Li+-preoccupied (n-type) tetrahedral sites of cubic phase that is oxidized at ~4 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$ to release lithium ions in ~100 mAh gLMO −1 during charging. Even if LMO has additional 100 mAh gLMO −1 capacity based on unoccupied (p-type) octahedral sites of tetragonal phase at ~3 $${{\rm{V}}}_{{{\rm{Li}}/{\rm{Li}}}^{+}}$$, it has been impossible to utilize the whole capacity at ~200 mAh g−1 in conventional cell and material designs due to instability of the 3 V reaction and the p-type nature of the 3 V active sites. We realized the whole capacity of the amphi-redox LiMn2O4 spinel in a mixed nature of n-type (4 V) and p-type (3 V). Reversible utilization of the 3 V reaction was made possible (1) by graphene-wrapped LMO nanoparticles (LMO@Gn) guaranteeing structural stability and electrical conduction and (2) by additional lithium source in prelithiated graphite (LinC6, n < 1) to lithiate the unoccupied octahedral sites of tetragonal LMO.

Methods

LMO@Gn cathodes

A mixture of LMO (Aldrich, undoped) and graphite (Timcal KS6) at 80:7 w/w was ball-milled in a stainless vial by using a high-energy vibratory ball miller (SPEX 8000D) to yield LMO@Gn. A mixed slurry of LMO@Gn, carbon black (Timcal Super P) and polyvinylidene fluoride (PVDF; Solef 5130) in N-methyl-2-pyrrolidone was casted on the aluminum foil and dried at 120 °C for 1 h. The resultant mass composition of the cathode was fixed at 70 : 6.125 : 3.875 : 20 or 50 : 4.4 : 25.6 : 20 for LMO/Gn/carbon black/PVDF.

Li n C6 anodes

A mixed slurry of natural graphite (Mitsubishi Chemical), carbon black (Timcal Super P) and a binder in N-methyl-2-pyrrolidone was casted on the copper foil and dried at 120 °C for 2 h. Na-carboxymethyl cellulose (CMC; Sigma-Aldrich) or polyvinylidene fluoride (PVDF; Solef 5130) was used as the binder. The resultant mass composition of the anode was fixed at 7 : 2 : 1 for graphite/carbon black/PVDF. The graphite anodes were lithiated by activating passivated lithium powder (PLP, China Energy Lithium Co., Ltd) sprinkles on active layers of anodes. The anodes, on which a proper amount of PLP was loaded, was pressed at 1 ton for 30 s by using a hydraulic press (Specac, AtlasTM manual 15-ton hydraulic press). Two parameters of prelithiation were defined as follows:

$${\rm{Prelithiation}}\,{\rm{yield}}={Y}_{{\rm{Li}}}( \% )=m/{{m}_{o}}^{* }100$$
(3)

where m = the amount of lithium intercalated to graphite during prelithiation process (g) = Q × (3.6 C / 1 mAh) / n F × MWLi; m o = the amount of PLP loaded (g) with Q = delithiation capacity of graphite prelithiated by PLP (mAh); n = 1 mol e (mol Li)−1 for Li+  + e → Li0; F = Faraday constant = 96500 C (mol e)−1; MWLi = 6.94 g (mol Li)−1 and

$$\% {\rm{PLP}}\,{\rm{Loading}}={L}_{{\rm{PLP}}}={m}_{0}/{m}_{100}\times 100$$
(4)

where m 100 = the amount of lithium required for fully lithiating graphite (g).

Amphi-redox battery

LMO@Gn||Li n C6, LMO@Gn||Li0 and Li n C6||Li0 were realized in the 2032 coin cell configuration. Cells were assembled in a glove box of argon atmosphere (H2O < 0.1 ppm and O2 < 0.1 ppm). 1 M LiPF6 in EC/EMC (1:2 v/v; EC = ethylene carbonate, EMC = ethyl methyl carbonate) was used as a base electrolyte. 1 wt. % VC was used as an additive for electrolytes to form the solid-electrolyte-interphase (SEI) layer. 0.1 wt. % heptamethyl disilazane (HEMDS) was used as a HF scavenger. Used as a separator was porous polyethylene separator (Asahi) or two-sided-alumina-coated polyethylene separator (2 μm Al2O3/16 μm PE/2 μm Al2O3; W-Scope).

Improvement factors (y ED and y Q ).

$${y}_{{\rm{ED}}}=\frac{\,\frac{{Q}_{4}{V}_{4}+{Q}_{3}{V}_{3}}{\frac{1}{{x}_{LMO}}+\frac{{Q}_{4}+{Q}_{3}}{{Q}_{Gr}}NP}\,}{\frac{{Q}_{4}{V}_{4}}{\frac{1}{{x}_{LMO}}+\frac{{Q}_{4}}{{Q}_{Gr}}NP}}$$
(5)
$${y}_{Q}=\frac{\frac{{Q}_{4}+{Q}_{3}}{\frac{1}{{x}_{LMO}}+\frac{{Q}_{4}+{Q}_{3}}{{Q}_{Gr}}NP}}{\frac{{Q}_{4}}{\frac{1}{{x}_{LMO}}+\frac{{Q}_{4}}{{Q}_{Gr}}NP}}$$
(6)

where y ED (y Q) = energy density (capacity) improvement factor = the ratio of energy density (capacity) of LMO@Gn||LinC6 to LMO||C6; subscript 4 and 3 = 4 V and 3 V, respectively; Q = specific capacity; Q4 and Q3 = specific capacity of LMO at 4 V and 3 V, respectively; QGr. = theoretical capacity of graphite; V = voltage; NP = capacity ratio of anode to cathode in LMO@Gn||Li n C6; x = x LMO  = LMO fraction in cathode.

Physicochemical characterization

Graphite intercalation compounds (GICs) in lithiated anodes were identified by X-ray diffraction (XRD; Bruker AXS, D8 Advance with Cu–Kα radiation). Lithiated graphite anodes in a circle shape were hermetically sealed by transparent polymer film used for vacuum packaging (Food Keeper, VP-9900) in a glove box to prevent oxidation. The hermetically sealed anode samples were placed on XRD sample holders (not inside the well for powder samples) and fixed on the rims of the holders by tape. The front positions of samples on which X-ray was incident were not perfectly identical over every run so that slight peak position shifts were observed (Supplementary Fig. 15). Lithium metal on lithiated graphite was identified by X-ray photoelectron spectroscopy (XPS) (ThermoFisher K-alpha). Deposited Mn on graphite anode was quantified by using optical emission spectroscopy where samples were atomized by inductively coupled plasma (ICP-OES; Varian, 700-ES). Graphite anode samples were obtained from cells experiencing electrochemical cycles, followed by washing by dimethyl carbonate (DMC) and then drying. Whole graphite anodes were used for quantifying Mn deposits in a cell.

References

1. 1.

Kim, D. et al. Design of nickel-rich layered oxides using d electronic donor for redox reactions. Chem. Mater. 27, 6450–6456 (2015).

2. 2.

Yim, T. et al. 5V-class high-voltage batteries with over-lithiated oxide and a multi-functional additive. J. Mater. Chem. A 3, 6157–6167 (2015).

3. 3.

Noh, H. K., Park, H. S., Jeong, H. Y., Lee, S. U. & Song, H. K. Doubling the capacity of lithium manganese oxide spinel by a flexible skinny graphitic layer. Angew. Chem. Int. Ed. Engl. 53, 5059–5063 (2014).

4. 4.

Aetukuri, N. B. et al. Flexible ion-conducting composite membranes for lithium batteries. Adv. Energy Mater. 5, 1500265 (2015).

5. 5.

Arakawa, M., Tobishima, S.-I., Nemoto, Y., Ichimura, M. & Yamaki, J.-i. Lithium electrode cycleability and morphology dependence on current density. J. Power Sources 43, 27–35 (1993).

6. 6.

Yoshimatsu, I., Hirai, T. & Yamaki, J.-I. Lithium electrode morphology during cycling in lithium cells. J. Electrochem. Soc. 135, 2422–2427 (1988).

7. 7.

Guerard, D. & Herold, A. Intercalation of lithium into graphite and other carbons. Carbon 13, 337–345 (1975).

8. 8.

Nam, K.-W., Ma, S.-B., Yoon, W.-S., Yang, X.-Q. & Kim, K.-B. Novel concept of pseudocapacitor using stabilized lithium metal powder and non-lithiated metal oxide electrodes in organic electrolyte. Electrochem. Commun. 11, 1166–1169 (2009).

9. 9.

Xiang, B., Wang, L., Liu, G. & Minor, A. M. Electromechanical probing of Li/Li2CO3 core/shell particles in a TEM. J. Electrochem. Soc. 160, A415–A419 (2013).

10. 10.

Wang, L., Fu, Y., Battaglia, V. S. & Liu, G. SBR–PVDF based binder for the application of SLMP in graphite anodes. RSC Advances 3, 15022–15027 (2013).

11. 11.

Yoshimoto, Y., Wada, H., Nakaya, H., Yoshida, M. & Nakajima, S. Electrochemical characterization of Li-intercalated pyrolytic graphite. Synth. Met. 42, 2707–2710 (1991).

12. 12.

Sacci, R. L., Adamczyk, L. A., Veith, G. M. & Dudney, N. J. Dry synthesis of lithium intercalated graphite powder and fiber. J. Electrochem. Soc. 161, A614–A619 (2014).

13. 13.

Chevrier, V. L., Zwanziger, J. W. & Dahn, J. R. First principles study of Li–Si crystalline phases: charge transfer, electronic structure, and lattice vibrations. J. Alloys Compd. 496, 25–36 (2010).

14. 14.

Wang, Y. & Dahn, J. Phase changes in electrochemically lithiated silicon at elevated temperature. J. Electrochem. Soc. 153, A2314–A2318 (2006).

15. 15.

Li, Y. & Fitch, B. Effective enhancement of lithium-ion battery performance using SLMP. Electrochem. Commun. 13, 664–667 (2011).

16. 16.

Wang, Z. et al. Application of stabilized lithium metal powder (SLMP®) in graphite anode – a high efficient prelithiation method for lithium-ion batteries. J. Power Sources 260, 57–61 (2014).

17. 17.

Wang, L., Battaglia, V. S. & Liu, G. In The Electrochemical Society 1010–1010 (Honolulu, 2012).

18. 18.

Forney, M. W., Ganter, M. J., Staub, J. W., Ridgley, R. D. & Landi, B. J. Prelithiation of silicon-carbon nanotube anodes for lithium ion batteries by stabilized lithium metal powder (SLMP). Nano Lett. 13, 4158–4163 (2013).

19. 19.

Ko, Y., Cho, Y.-G. & Song, H.-K. Programming galvanostatic rates for fast-charging lithium ion batteries: a graphite case. RSC Advances 4, 16545–16550 (2014).

20. 20.

Ohzuku, T. Formation of lithium-graphite intercalation compounds in nonaqueous electrolytes and their application as a negative electrode for a lithium ion (shuttlecock) cell. J. Electrochem. Soc. 140, 2490–2498 (1993).

21. 21.

Aurbach, D. The study of Li-graphite intercalation processes in several electrolyte systems using in situ X-ray diffraction. J. Electrochem. Soc. 142, 1746–1752 (1995).

22. 22.

Song, X. Y. & Kinoshita, K. Microstructural characterization of lithiated graphite. J. Electrochem. Soc. 143, L120–L123 (1996).

23. 23.

Kim, J.-H. et al. Understanding the capacity fading mechanism in LiNi0.5Mn1.5O4/graphite Li-ion batteries. Electrochim. Acta 90, 556–562 (2013).

24. 24.

Zhan, C. et al. Mn(II) deposition on anodes and its effects on capacity fade in spinel lithium manganate-carbon systems. Nat. Commun. 4, 2437–2444 (2013).

25. 25.

Arora, P., White, R. E. & Doyle, M. Capacity fade mechanisms and side reactions in lithium-ion batteries. J. Electrochem. Soc. 145, 3647–3667 (1998).

26. 26.

Dai, Y., Cai, L. & White, R. E. Capacity fade model for spinel LiMn2O4 electrode. J. Electrochem. Soc. 160, A182–A190 (2012).

27. 27.

Choi, N.-S. et al. Degradation of spinel lithium manganese oxides by low oxidation durability of LiPF6-based electrolyte at 60 °C. Solid State Ionics 219, 41–48 (2012).

28. 28.

Li, B. et al. A novel electrolyte with the ability to form a solid electrolyte interface on the anode and cathode of a LiMn2O4/graphite battery. J. Mater. Chem. A 1, 12954–12961 (2013).

29. 29.

Kim, J.-H., Kim, J.-H., Kim, J.-M., Lee, Y.-G. & Lee, S.-Y. Superlattice crystals-mimic, flexible/functional ceramic membranes: beyond polymeric battery separators. Adv. Energy Mater. 5, 1500954 (2015).

30. 30.

Woo, S. H. et al. Ion-Exchangeable Functional Binders and Separator for High Temperature Performance of Li1.1Mn1.86Mg0.04O4 Spinel Electrodes in Lithium Ion Batteries. J. Electrochem. Soc. 160, A2234–A2243 (2013).

31. 31.

Li, Y., Zhang, R., Liu, J. & Yang, C. Effect of heptamethyldisilazane as an additive on the stability performance of LiMn2O4 cathode for lithium-ion battery. J. Power Sources 189, 685–688 (2009).

32. 32.

Wu, X. et al. Improvement on the storage performance of LiMn2O4 with the mixed additives of ethanolamine and heptamethyldisilazane. Appl. Surf. Sci. 268, 349–354 (2013).

33. 33.

Cho, I. H., Kim, S.-S., Shin, S. C. & Choi, N.-S. Effect of SEI on capacity losses of spinel lithium manganese oxide/graphite batteries stored at 60 °C. Electrochem. Solid-State Lett. 13, A168–A172 (2010).

Acknowledgements

This work was supported by MOTIE (Regional (KIAT): R0004244, R0006515) and MSIP (2014R1A1A1003044), Korea.

Author information

Authors

Contributions

Y.J. and H.K.N. contributed equally to this work. Y.J. electrochemically characterized and optimized graphite prelithiation and LIB cells based on amphi-redox LMO and graphite. H.K.N. characterized prelithiated graphite and investigated graphene-wrapped LMO@Gn cathodes. Physicochemical analyses including XRD, XPS and ICP were carried out by Y.J. and H.K.N. The idea of the amphi-redox LIB cells was initiated by H.K.N. and H.-K.S. H.-K.S. organized this work and supervised all-round progress of this work. Y.J., H.K.N. and H.-K.S. contributed to writing the manuscript.

Corresponding authors

Correspondence to Hyun Kuk Noh or Hyun-Kon Song.

Ethics declarations

Competing Interests

The authors declare that they have no competing interests.

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

Jeon, Y., Noh, H.K. & Song, HK. A Lithium-ion Battery Using Partially Lithiated Graphite Anode and Amphi-redox LiMn2O4 Cathode. Sci Rep 7, 14879 (2017). https://doi.org/10.1038/s41598-017-14741-x

• Accepted:

• Published:

• What do Coulombic Efficiency and Capacity Retention Truly Measure? A Deep Dive into Cyclable Lithium Inventory, Limitation Type, and Redox Side Reactions

•  & Daniel C. O’Hanlon

Journal of The Electrochemical Society (2020)

• Evaluation of nanostructured Nd0.7Co0.3FeO3 perovskite obtained via hydrothermal method as anode material for Li-ion battery

• K.O. Ogunniran
• , R. Thangamuthu
•  & P. Periasamy

Materials Chemistry and Physics (2020)

• Methodologies for Large-Size Pouch Lithium-Ion Batteries End-of-Life Gateway Detection in the Second-Life Application

• Pierrot S. Attidekou
• , Zoran Milojevic
• , Mohamed Ahmeid
• , Simon Lambert
•  & Prodip K. Das

Journal of The Electrochemical Society (2020)

• Electrochemical performance evaluation of carbon nitride synthesized at different temperatures as an anode material for lithium-ion batteries

• Gnanavel Angamuthu
• , Ezhilan Jayabal
•  & Venkatesan Rengarajan

Ionics (2020)

• Prelithiation/Presodiation Techniques for Advanced Electrochemical Energy Storage Systems: Concepts, Applications, and Perspectives

• Kangyu Zou
• , Wentao Deng
• , Peng Cai
• , Xinglan Deng
• , Baowei Wang
• , Cheng Liu
• , Jiayang Li
• , Hongshuai Hou
• , Guoqiang Zou
•  & Xiaobo Ji

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.