Mn3O4 is widely used as a sensor, catalyst, capacitor, and magnetic material because of its non-toxicity, low environmental load, stable physicochemical properties, and abundant Mn resources1,2,3,4,5,6,7. It has also gained considerable attention as an anode material for Li-ion batteries8,9,10, as it permits the following electrochemical conversion reaction: Mn3O4 + 8Li+  + 8e  ↔ 4Li2O + 3Mn. The lithiation process leads to the formation of Li2O and metallic Mn, whereas delithiation gives Mn3O4 by reverse oxidation. The Mn3O4 anode yields a theoretical capacity of 937 mAh/g based on this conversion reaction, which is higher than that of a traditional intercalation-type anode, such as graphite (372 mAh/g). Furthermore, Mn-based oxides including Mn3O4 have obvious advantages for a higher battery potential because of a low operating voltage (0.2–0.4 V vs. Li+/Li) associated with conversion reactions compared to the iron, cobalt, and nickel-based oxides (0.6–1.2 V)10. However, its low electrical conductivity (~ 10−8 S/cm to 10−7 S/cm) and significant volume change during lithiation/delithiation cause particle collapse and discontinuation of the electrical conduction pathway, leading to serious cycle degradation for initial tens of cycles11. Especially in the first cycle, the Coulombic efficiency, defined as the ratio of charge to discharge capacity, tends to be extremely low. This hinders the practical application of Mn3O4 anodes in future high-energy Li-ion batteries12.

Numerous approaches from a microstructure perspective have been proposed to address the intrinsic problems of Mn3O4 active materials. Surface coating or composites with carbon materials such as graphene13,14, graphite15, nanotubes/fibers16,17,18, and nanoparticles19,20 can improve the low electrical conductivity. Meanwhile, particle morphologies consisting of nanosized21, porous22, hollow23, and hierarchical structures24,25 can alleviate volume change. In terms of anode performance, composites of microstructure-controlled Mn3O4 with carbon materials achieve improved cyclability and a higher capacity. Most of the previous efforts implemented a carbon-supported composite structure13,14,15,16,17,18,25. Mn3O4 particles on carbon were exposed to electrolytes, making them prone to side reactions involving electrolyte decomposition, which caused the formation of thick solid-electrolyte interphase (SEI) layers during the charge and discharge cycles26,27. These SEI layers induced particle detachment from the carbon material or the formation of strongly bonded aggregates, diminishing the synergistic effect of the composite structure. To overcome these challenges, Cai et al. constructed Mn3O4/C yolk-shell nanorods as a highly improved composite structure28, wherein the carbon shell minimizes the formation of the SEI layer and the hollow carbon rods restrict the volume change of Mn3O4 rods. Thus, the microstructure of Mn3O4 particles and the composite structure of carbon materials both are important to improve the anode performance.

Herein, we show the carbon nanoparticle-entrapped macroporous Mn3O4 microsphere as a conversion-type anode morphology for improving cycling stability. The fabrication process is composed of a scalable two-step approach involving the thermal decomposition of MnCO3 and mixing in a carbon-dispersed solution, as depicted in Fig. 1. Recently, our group prepared macroporous Mn3O4 microspheres with unique maze-like pore structures by the thermal decomposition of MnCO3 in water vapor29. The Mn3O4 is a porous particle consisting of open macropores. We assessed the applicability of this particle to attain a conversion-type anode material morphology using a single-particle measurement technique30. The intrinsic electrochemical performance of the macroporous Mn3O4 microsphere exhibited a discharge capacity close to the theoretical one, even at an input current equivalent to the 4C-rate, although it decreased gradually due to particle fragmentation. This capacity fading arises from structural changes and insufficient electrical conduction pathways within the Mn3O4 particle. Therefore, we constructed a carbon network within the macroporous Mn3O4 microsphere using its open macropores and maze-like pore structure. We inserted and trapped carbon nanoparticles within the particle walls by mixing macroporous Mn3O4 microspheres in a carbon-dispersed solution. These composites with a simple construction improve cyclability through the following synergistic effects: (1) development of conductive paths within the Mn3O4 particle31; (2) alleviation of volume change by the porous structure; (3) prevention of aggregation by the blocking effect of added carbon nanoparticles32. The carbon nanoparticles perform to prevent contact between adjacent Mn3O4 particle walls due to volume change during charge and discharge cycles. In addition to improving electrical conductivity, the entrapped carbon nanoparticles and their aggregates formed on the composite surface act as buffers during the structural change of Mn3O4 to Li2O and metallic Mn (by the electrochemical conversion reactions) to maintain its unique porous structure. The composite structure and its fabrication process proposed in this study can offer a wide range of applications to other anodes of transition metal oxides due to their simplicity.

Figure 1
figure 1

Schematic illustration of the fabrication process of Mn3O4/C composite.


Synthesis of macroporous Mn3O4 microspheres

Macroporous Mn3O4 microspheres were synthesized via the thermal decomposition of MnCO3 microspheres in water vapor29. The spherical MnCO3 powder (median size, ~ 3 μm) was prepared using the precipitation method33. The prepared MnCO3 spheres were thermally decomposed in a tubular furnace equipped with an evaporator. The powder sample was heated to 750 °C at a rate of 5 °C/min and held for 2 h. Deionized water was pumped at a flow rate of 1 mL/min into the evaporator and heated above 100 °C. The generated water vapor was directly introduced into the tubular furnace without any carrier gas and allowed to flow during the holding time. The furnace outlet was opened into the laboratory atmosphere.

Fabrication of Mn3O4/C composites

The Mn3O4/C composites were fabricated from the macroporous Mn3O4 microspheres and ketjen black carbon (ECP600JD, LION SPECIALTY CHEMICALS Co., Ltd., Japan) nanoparticles as follows. First, carbon powder (3 mg to 30 mg) was added to 2-propanol (30 mL, FUJIFILM Wako Pure Chemical Co., Japan) and dispersed by ultrasonic irradiation for 5 min. Then, the Mn3O4 powder (90 mg) was added to this carbon-dispersed solution and stirred for 1 h. The treated product was collected by centrifugation, dried at 100 °C, and ground using an agate mortar. Next, the collected powder was added to water and kept overnight to isolate the free carbon. The carbon content in the Mn3O4/C composites was estimated by weighing the collected free carbon powder.

Powder characterization

The crystalline phases of the powder samples were characterized by powder X-ray diffraction (XRD, D2 PHASER, Bruker AXS GmbH, Germany) using Cu Kα radiation generated at 30 kV and 10 mA. The diffraction patterns were acquired in steps of 0.02° (2θ) and a counting time of 0.5 s/step. The particle morphologies were examined using scanning electron microscopy (SEM, SU-70, Hitachi Ltd., Japan) and transmission electron microscopy (TEM, JEM-2100F, JEOL Ltd., Japan). The cross-sectional samples were prepared using an Ar ion beam gun (IB-09020CP, JEOL). The particle size distribution of Mn3O4 was determined using the laser diffraction/scattering method (Microtrac MT3300EXII, NIKKISO Co., Ltd., Japan). Small amounts of the sample were dispersed in a solution of sodium hexametaphosphate (0.05 mass%) using an ultrasonic homogenizer. Its specific surface area (Sw) was estimated using N2 adsorption measurements (3Flex, Micromeritics Ltd., USA). Before each measurement, the powder sample was outgassed under vacuum for 3 h at 120 °C. The Sw of the sample was calculated using the Brunauer–Emmett–Teller (BET) method. Its pore structure was determined using a Hg porosimeter (AutoPore V, micromeritics), while its pore size was calculated using the Washburn equation. The surface tension of Hg was 485 dynes/cm. Contact angles of 130° and 154.9° were used in the cases of bare Mn3O4 and carbon composites, respectively34. Raman spectroscopy (LabRAM ARAMIS, HORIBA Jobin Yvon SAS, France) was performed using the 532 nm excitation line of a He–Ne laser.

Electrochemical characterization

The anode performance of the fabricated samples was evaluated by using coin-type cells. Bare Mn3O4 or Mn3O4/C composites were mixed with acetylene black carbon (Denka Co., Ltd., Japan) and sodium carboxymethyl cellulose (FUJIFILM Wako Pure Chemical) in a weight ratio of 80:12:8 and added a small amount of water. The anode slurry was coated onto a Cu foil using a doctor blade and then dried at 100 °C for 12 h in a vacuum. The dried anode was uniaxially pressed at 40 MPa and punched out of the foil. The active material loadings in the electrodes were about 3.4 mg/cm2. A polypropylene membrane (#2400, Celgard LLC, USA) and a 1 M LiPF6 (Kishida Chemical Co., Ltd., Japan) solution (30 μL) in a mixture of ethylene carbonate and diethyl carbonate solvents (50:50 vol.%) were used as the separator and the electrolyte, respectively. The 2032-type coin cells were assembled in a glove box filled with dry Ar. The anode performance was characterized at 20 °C using an electrochemical workstation (VMP3, BioLogic Sciences Instruments, France). Cyclic voltammetry (CV) measurements were conducted at a scan rate of 0.2 mV/s within 0–3 V. The charge–discharge tests were performed within 0.01–3 V at varying rates. Electrochemical impedance spectroscopy (EIS) measurements were taken in the frequency range from 10 mHz to 100 kHz with an amplitude of 5 mV.

Results and discussion

Materials characterization

Figure 2 shows the powder characteristics of the prepared Mn3O4 and carbon nanoparticles used. Open pores enclosed in randomly grown particles are confirmed on the Mn3O4 microsphere surface (Fig. 2a). A cross-sectional view shows the maze-like pore structure formed by the developed particle walls (Fig. 2b). According to the N2 adsorption measurements, the Sw value of the Mn3O4 powder is < 3 m2/g, indicating the existence of macropores solely. The internal pore size is determined to be ~ 400 nm using Hg porosimetry (more details are provided in the next section). The large open pores allow electrolytes easy access into the particles. When measured in a liquid, the Mn3O4 powder exhibits a unimodal particle size distribution with almost the same median size as that of MnCO3 (Fig. 2d, right). By contrast, the carbon powder possesses a spherical hollow shape and an interconnected chain-like structure (Fig. 2c). The average particle size, estimated from TEM observations, is ~ 35 nm (Fig. 2d, left). The hollow shape provides a high Sw of 790 m2/g. These results reveal that the prepared macroporous Mn3O4 microspheres have suitable pores for receiving carbon nanoparticles.

Figure 2
figure 2

(a,b) SEM images of macroporous Mn3O4 microspheres and (c) TEM image of carbon nanoparticles. (d) Particle size distributions of (left, estimated by TEM) carbon and (right, measured in a liquid) Mn3O4 powders.

Fabrication and characterization of Mn3O4/C composites

Initially, we investigated the amount of carbon nanoparticles retained within the macroporous Mn3O4 microspheres. Figure 3 depicts the plot of carbon content in the Mn3O4/C composites against the added amount. For 90 mg of Mn3O4, added carbon of up to 6 mg (6.7 mass%) is fully retained in Mn3O4. A further increase in the amount added exceeds the limit of retention within Mn3O4. However, upon the addition of 30 mg of carbon (33 mass%), the retention amount indicates an upward trend. In the following experiments, three composite samples (Mn3O4/C-3, -7, and -33) fabricated with varying amounts of added carbon (3.3 mass%, 6.7 mass%, and 33 mass%) are compared with bare Mn3O4 particles.

Figure 3
figure 3

Retained amount of carbon nanoparticles compared to the added amount.

The un-inserted carbon nanoparticles in Mn3O4 are trapped and retained on the composite surface aggregates. Figure 4 shows the structures and morphologies of the fabricated composites. After insertion in the liquid, the particle morphology of Mn3O4 remains the same in all the samples (Fig. 4a–c). The carbon nanoparticles are trapped in open macropores on the surface. With an increase in the added amount of carbon, the excess carbon nanoparticles form aggregates on the composite surface due to their chain-like structures (Fig. 4c). This formation of aggregates leads to an increase in the retention amount, as shown in Fig. 3. A cross-sectional view of Mn3O4/C-7 reveals the capture of carbon nanoparticles within the internal walls of the particle (Fig. 4d,e). The XRD analysis shows that all the composite samples retain the high crystallinity of Mn3O4 (Fig. 4f). Since the carbon nanoparticles are only entrapped within the Mn3O4 particles, the crystallinity maintained therein is high compared to that of the yolk-shell28 or core-shell35 structures fully covered with carbon materials. The composites with carbon nanoparticles exhibited increasing specific surface areas: 17.6, 34.9, and 157 m2/g for Mn3O4/C-3, -7, and -33, respectively.

Figure 4
figure 4

(ae) SEM images and (f) XRD patterns of the Mn3O4/C composite samples: (a) Mn3O4/C-3, (b,d,e) Mn3O4/C-7, and (c) Mn3O4/C-33 samples.

Furthermore, the structural evaluation was conducted using Raman spectroscopy (Fig. S1). The Raman bands of the prepared Mn3O4 are consistent with the previous report36. Typical carbon peaks located at 1340 cm−1 (D band) and 1595 cm−1 (G band) are detected in all composite samples. The intensity ratio of the D and G bands (ID/IG) for the used ketjen black carbon is approximately 1.1, suggesting a slightly higher proportion of disordered graphitic form (amorphous). Since these two band positions and the ID/IG ratios in all the composite samples are almost the same, no chemical interaction between Mn3O4 and carbon occurs during the fabrication process.

The pore size of Mn3O4 decreases with the insertion of carbon nanoparticles. Figure 5 shows the pore size distributions of bare Mn3O4 and Mn3O4/C composite samples estimated using Hg porosimetry. The pore region at ~ 1 μm is caused by the interparticle spaces between Mn3O4 microspheres. This pore region disappears for Mn3O4/C-33 since the formation of carbon aggregates on the surface hinders a close packing. Thus, the pore size of Mn3O4 is determined to be ~ 400 nm. The size of the macropores keeps diminishing with the insertion of carbon nanoparticles, finalizing as ~ 300 nm for Mn3O4/C-33. A new pore region appears at 40–50 nm. This nanopore region can be attributed to the carbon aggregates. Since the carbon nanoparticles tend to form interconnected chain-like structures (Fig. 2c), the aggregates spontaneously generate nanopores of several tens of nanometers.

Figure 5
figure 5

Pore size distribution of the bare Mn3O4 and Mn3O4/C composites.

Carbon nanoparticle-entrapped Mn3O4 microspheres, which differ from the commonly constructed composites, were fabricated via a scalable two-step process. Open macropores enable the direct insertion of conductive nanoparticles by mixing in solution. Templates37 or in-situ formations following crystal growth38 have been previously implemented to insert conductive materials into the electrode particles. Our process is probably the simplest and most versatile method.

Electrochemical performance

This study expresses the reduction (lithiation) of Mn3O4 to Li2O and metallic Mn as a charging process and their reverse oxidation (delithiation) to manganese oxides as a discharging process. The cycle performance of the anode is discussed based on its discharge capacities. The C rates (1C = 93.7 mA/g) used in each measurement are calculated from the weight of Mn3O4 in the composite.

Figure 6 shows the CV and charge–discharge curves of bare Mn3O4, Mn3O4/C-7, and Mn3O4/C-33 composite anodes during the selected cycles (first, second, fifth, and tenth cycles). For the bare Mn3O4 anode, the CV curve in the first cycle exhibits only a broad reduction peak at ~ 0.25 V (Fig. 6a). This peak disappears in the second CV curve. Instead, a current drop from 0.3 to 0 V occurs. An oxidation peak at 1.35 V can be observed in the anodic scan. The subsequent cycles show the reduction and oxidation peaks at 0.25 V and 1.35 V, respectively. These irreversible initial CV curves indicate unstable electrochemical reactions of bare Mn3O419. Reflecting this electrochemical instability, the charge and discharge capacities observed at 0.1C decrease drastically with the number of cycles (inset in Fig. 6a). The initial discharge capacity of 873 mAh/g declines to 291 mAh/g after 10 cycles. In contrast, the Mn3O4/C composite anodes exhibit relatively better reversibility and stability (Fig. 6b,c). The first CV curve for the composites shows a broad reduction peak centered at 0.6–0.5 V and a steep one below 0.2 V in the cathodic scans, and an oxidation peak at 1.3 V in the anodic scans. The principal reduction peak is located at ~ 0.25 V in the subsequent cycles, except for the intensive peak below 0.3 V for Mn3O4/C-7 in the second cycle. In contrast, the oxidation peak remains at 1.3 V. A small hump at ~ 2.5 V can be observed in the anodic scan for the Mn3O4/C-33 anode (Fig. S2). In contrast to the electrochemical redox reaction for bare Mn3O4, the Mn3O4/C composites exhibit a reversible one, showing improved cyclability during the charge–discharge runs (inset in Fig. 6b,c). The initial discharge capacities of Mn3O4/C-7 and Mn3O4/C-33 are 753 mAh/g and 787 mAh/g, respectively. These capacities are maintained even after 10 cycles. The electrochemical performance of the Mn3O4/C-3 anode lies between that of the bare Mn3O4 and Mn3O4/C-7 (Fig. S3). We conclude that the electrochemical performance of the Mn3O4/C composites improves with increasing carbon content.

Figure 6
figure 6

CV curves and charge–discharge curves (inset) of (a) bare Mn3O4, (b) Mn3O4/C-7, and (c) Mn3O4/C-33 composite anodes.

The redox reactions during the charge and discharge cycles of Mn3O4 are discussed based on the above results of the initial electrochemical tests. The initial lithiation of unprocessed Mn3O4 leads to its reduction to MnO accompanied by the formation of Li2O, corresponding to the cathodic peak at 0.6–0.5 V in CV. Besides, the decomposition of electrolytes on the particle surface results in the formation of an SEI layer. The initial charging capacity at a potential plateau of ~ 1.0 V is consumed by these reactions. Another steep cathodic peak at a lower potential is attributed to the further reduction of MnO to metallic Mn, indicating a potential plateau at 0.25 V. The delithiation from Li2O and metallic Mn to MnO occurs at 1.3 V. The potential sloping during the discharge process over 2.0 V, especially in the Mn3O4/C composite, indicates the further oxidation of Mn ions18,22,23,25. We have confirmed the crystalline phases of Mn3O4 and MnO2 by XRD from the Mn3O4/C-33 sample after the electrochemical test (Fig. S4). The formation of MnOx (x =  ~ 2) with a higher oxidation state than the initial Mn3O4 has been detected in composites with carbon fibers18 or graphene39. The formation of MnO2 is probably due to the residual oxygen present as defects in the carbon materials39. From the second cycle onwards, the reduction reaction to form Li2O and metallic Mn shifts to a higher potential; thus, its corresponding plateau is observed at 0.5 V. This is attributed to the following reasons: (i) the formed SEI layer with improved conductivity that enhances the reaction kinetics19,23 and (ii) the structural changes during the first lithiation-delithiation process40. According to in operando X-ray studies, the first lithiation occurs in the sequence Mn3O4 → LiMn3O4 → MnO, with the final conversion step involving the reduction of MnO to metallic Mn41. The lithiation after the second cycle is reversible without forming the Li-Mn–O intermediate phase. The structural changes in the active material will cause the difference between the first and subsequent lithiation reactions. Herein, some carbon materials in the composites exhibit Li+ storage reactions18, but the ketjen black carbon used does not show any such behavior. The added carbon nanoparticles contribute to the electrochemical stability of Mn3O4 without reacting themselves.

We tested the cyclability of the fabricated anodes at 0.1C for up to 50 cycles. Figure 7 plots the discharge capacities recorded for all anodes. As shown in Fig. 6a, the bare Mn3O4 exhibits a rapid capacity drop during the second cycle. Its discharge capacity retention is only 18% after 50 cycles. For Mn3O4/C-3, although the discharge capacity stabilizes at ~ 400 mAh/g after the twentieth cycle, its capacity retention eventually drops to 43%. A rapid capacity drop in the initial stage is restrained for the Mn3O4/C-7 and -33 composite anodes, and consequently, their discharge capacity retention increases to 68% and 86%, respectively. The Mn3O4/C-33 anode composed of excess carbon exhibits a discharge capacity of 680 mAh/g after 50 cycles. The Coulombic efficiency exhibited more than 60% for the first cycle and remained over 95% during subsequent cycles (Fig. S5). These results suggest a gradual increase in electrochemically inactive regions for bare Mn3O4 and composites with low carbon contents.

Figure 7
figure 7

Cycling performance of the bare Mn3O4 and Mn3O4/C composite anodes at 0.1C.

Furthermore, we investigated the rate performances of all anodes. Figure 8 plots the discharge capacities at different rates ranging from 0.05C to 5C. The bare Mn3O4 and the Mn3O4/C-3 anodes exhibit relatively low discharge capacities compared to the composite anodes with high carbon contents at most rates. Meanwhile, the discharge capacities of these two anodes are reversed for the applied current density. A similar tendency is observed with the Mn3O4/C-7 and -33 anodes; the highest discharge capacity of ~ 150 mAh/g at 5C is obtained with Mn3O4/C-7. The charge–discharge curves at different rates indicate that the bare Mn3O4 anode suffers from high internal resistance, shown a voltage drop during the charging process with increasing rates (Fig. S6). The Mn3O4/C-7 composite anode exhibited a charging plateau at ~ 0.5 V even at higher rates. Increasing the carbon content does not simply improve rate performance. The cycle and rate performances differ depending on the carbon content of the composites. These results suggest that the carbon nanoparticles within the macroporous Mn3O4 microspheres have both electrochemical and physical merits for improving the anode performance.

Figure 8
figure 8

Rate performance of the bare Mn3O4 and Mn3O4/C composite anodes.

Role of inserted carbon nanoparticles and surface aggregates

The cyclability of the Mn3O4 anode improves with the increasing amount of entrapped carbon nanoparticles within the microspheres. The formation of carbon aggregates on the composite surface further enhances its performance. This improved performance of Mn3O4/C composites can be attributed to the following electrochemical and physical aspects: (1) development of electrical conduction pathways within Mn3O4 microspheres by the insertion of carbon nanoparticles, (2) retention of the porous structure during the charge–discharge process, and (3) prevention of Mn3O4 particle aggregation by the surface carbon aggregates. Figure 9 illustrates the effects of carbon nanoparticles on the electrochemical performance and structural stabilization of macroporous Mn3O4 microspheres during the charge–discharge process.

Figure 9
figure 9

Illustration of the conversion reaction of the (a) bare Mn3O4, (b) Mn3O4/C composite, and (c) composite with an excess amount of carbon during the charge–discharge process.

First, the inserted carbon nanoparticles assume the electrochemical merit as conduction pathways within the Mn3O4 microspheres. The lithiation process from Mn3O4 to Li2O and metallic Mn, and vice versa, involves particle disproportionation. Electrochemically inactive regions appear in the interior of large particles, such as microspheres, due to interruptions in the conduction pathways (Fig. 9a). This expanse in the irreversible region causes internal resistance, leading to cycle degradation. The conversion reaction can be reversible if a conduction pathway is developed within the microspheres (Fig. 9b). EIS measurements were conducted to clarify the electrochemical merit. Figure 10 shows the EIS spectra of bare Mn3O4 and Mn3O4/C-33 composite before and after five charge–discharge cycles. The equivalent circuit is composed of the resistance of the cell system (Rs), including the cell components and electrolyte, the resistance (RSEI) and capacitance (CPE1) of the SEI layer, Faradaic charge-transfer resistance (Rct), double-layer capacitance (CPE2), and the Warburg impedance (Zw)42. Before cycles, the resistances (Rct) associated with a semicircle in the high-frequency region are 78 Ω and 101 Ω for the bare Mn3O4 and the Mn3O4/C-33 composite, respectively. Electrode materials with high surface areas generally exhibit low charge-transfer resistances, but our results show an opposite trend. Such cases are often observed for particles with different surface states43, and yolk-shell structures28,44,45, in which the active material is fully covered. The present composite, wherein the carbon nanoparticles are entrapped both inside and outside the Mn3O4 particle, may cause high charge-transfer resistance. After cycles, the bare Mn3O4 exhibits a high RSEI (214 Ω) and Rct (631 Ω) in the high- and middle-frequency regions, respectively. In contrast, the Mn3O4/C-33 shows low RSEI (10 Ω) and Rct (230 Ω) values. These values indicate activation of the reaction accompanied by improved kinetics. The carbon nanoparticles are incorporated into the SEI layer during the initial charging process, resulting in enhanced electrical conductivity and structural stability of the Mn3O4 particles18,19,28.

Figure 10
figure 10

Nyquist plots of bare Mn3O4 and Mn3O4/C-33 composite anodes before and after five charge–discharge cycles at 1C.

Next, the entrapped carbon nanoparticles exert an anchoring effect to retain the porous structure of the Mn3O4 microspheres. The thick SEI layer formed on the bare Mn3O4 impedes its return to the original particle morphology. A cross-sectional view of the electrode after ten charge–discharge cycles reveals the collapse of the Mn3O4 microspheres (Fig. S7). The segregation of the reaction products continues markedly with the expanse of the electrochemically inactive region, as illustrated in Fig. 9a. In contrast, the porous structure in the Mn3O4/C-33 composite sample is maintained even after 50 cycles (Fig. S7). The carbon nanoparticles within the Mn3O4 microspheres assist in the reversible conversion reactions on each particle wall. Further, the ketjen black carbon itself does not exhibit Li+ storage reactions and thus has the physical merit of restricting the volume change of Mn3O4 inside the pore space (Fig. 9b). Avoiding conversion reactions between adjacent particle walls prevents product aggregation.

Finally, similar to the entrapped carbon nanoparticles within the Mn3O4 microspheres, their aggregates on the composite surface also prevent the aggregation of active particles in the electrode (Fig. 9c). According to the Hg porosimetry (measured while applying high pressure), the surface carbon aggregates serve as buffers between the active particles (Fig. 5). The porous structure can be maintained by limiting the volume change during the conversion reaction of the Mn3O4 anode. The composite structure demonstrated in this study can be applied to other inexpensive transition metal oxide anodes by a further modification with improving electrochemical performance. Liu’s group has designed conversion-type composite anodes with hollow, nanotube, and nanocage structures that exhibited excellent electrochemical performances46,47,48. While allowing for buffering of volume changes, size reduction of porous particles and homogenous carbon coating on the particle wall will attain more improved cycle and rate performances.


We have constructed a composite structure in which carbon nanoparticles are entrapped within macroporous Mn3O4 microspheres with maze-like pore structures. As conversion-type anodes for Li-ion batteries, the Mn3O4/C composites exhibit improved cycle performance with increasing carbon content. The discharge capacity retentions of the bare Mn3O4 and Mn3O4/C-33 anodes after 50 cycles are 18% and 86%, respectively. Carbon nanoparticles are entrapped internally through the open macropores of the Mn3O4 microspheres and form aggregates on the composite surface when added in excess amounts. This morphology of the carbon nanoparticles assists in improving the cycle performance of Mn3O4 both electrochemically and physically. The present composite structure can be fabricated via a scalable two-step process that involves preparing macroporous microspheres of active materials and mixing them in a carbon-dispersed solution. The demonstrated composite structure and its fabrication process can be applied to other conversion-type anode materials that suffer from cycle degradation, including inexpensive transition metal oxides (Mn2O3, Fe2O3, and Fe3O4) and sulfides. Moreover, this study will enable the incorporation of future applied materials in applications such as energy storage, environmental filtration, and selective chemical reaction by utilizing the unique pore structure of the Mn3O4 microspheres.