Here we present that hollow nano-size conductive agent (Ketjenblack (KB), Figure S1a), sulfur blocks (Figure S1b) and nano-Co2O3 (Figure S1c) can be easily composited into microscale sulfur SPs by the above hail-inspired sulfur nanostorm (referred as SP@HSN below). For comparison, sulfur blocks were also composited with KB and Co2O3 by conventional high-temperature melt-diffusion method, and the sample was designated as sulfur AP@HTM. As shown in Fig. 2a the sulfur AP@HTM shows an irregular particle size/shape with a poor coverage of KB against sulfur. In contrast, it is observed that the sulfur SP@HSN shows applesnail-egg-like morphology in Fig. 2b. The structures of sulfur SP@HSN was further characterized by TEM image. Figure 2c reveals that sulfur is encapsulated in partially open hollow KB nano-host (Figure S2). Meanwhile, a number of Co2O3 nanoparticles (see Fig. 2c) are embedded inside the sulfur SP@HSN. Elemental mapping (see Fig. 2d) further confirms that S, C and Co elements are uniformly dispersed throughout the sulfur SP@HSN composite particle. To further investigate the S/C interactions, the nitrogen adsorption-desorption testing was conducted. As shown in Fig. 2e and Figure S3a, the specific surface area of sulfur SP@HSN-5 min sample is 3.2 m2 g− 1, which is far less than that of KB (1400 m2 g− 1), indicating that the porous KB was impregnated by sulfur. Besides, the specific surface area of different S/KB samples is summarized in Table S1. Generally, larger size of SP possesses smaller specific surface area that needs a less amount of electrolyte for wetting, thus boosting the gravimetric energy (Eg) of Li-S batteries.36 However, both sulfur AP@HTM and sulfur SP@HSN particles show similar pore distribution in the range of 5 nm-100 nm (Figure S3b), which is probably related to the stacking pores inside the sulfur SP@HSN. These stacking pores will provide ion-transportation channels inside the secondary particles. The S/C binding interactions were further investigated by the first heating curve of differential scanning calorimetry (DSC) at a scan rate of 10 oC/min. As shown in Fig. 2f, two sharp melting peaks are observed for pure sulfur. In contrast, sulfur SP@HSN presents a single and very weak melting peak as compared with pure sulfur and sulfur AP@HTM. This result further confirms that sulfur is well-confined inside the nano-channels or pores of KB via a vapor-deposition process as demonstrated in Fig. 1b. In other words, the sulfur SP@HSN shows the strongest sulfur/C binding interactions benefited from the hail-inspired sulfur nanostorm technology. Thermogravimetric analysis (TGA) as shown by Fig. 2g reveals that the sulfur content in sulfur SP@HSN composites is 80%, which confirms the S-rich feature of the secondary particles and is consistent with the formula before the compositing.
In addition to the above design flexibility in composition, the size of the sulfur-rich secondary particles can be controlled easily by compositing time. The size of active electrode particles is actually a very critical factor controlling not only the energy density, but also the quality of the final composite electrodes. In particular, it shapes the electrode assembling structures (ion-transport network (ITN) and electron-transport network (ETN)) as introduced previously. Rational size of the active particles is also critical for preparing uniform and crack-free thick electrode.2, 29 Fig. 2h shows that the particle size can be easily regulated by the compositing time. When the compositing time is only 2 min, the particle size of sulfur SP@HSN mainly concentrates in the range of 1–10 um, as shown in the top of Fig. 2h. As the compositing time increased to 3 min or 5 min, the particle size further increased (see Figure S4). According to the compositing time, the samples are defined as the sulfur SP@HSN-3 min (see the middle of Fig. 2h) and sulfur SP@HSN-5 min (see the bottom of Fig. 2h). For these sulfur SP@HSN particles, one can clearly find that they are secondary particles with tightly stacked nanoscale S/KB particles. In practice, the tap density of active particles is widely employed as a parameter to indicate their Ev potential in final electrodes. A high tap density of electrode particles will favor the fabrication of compact electrodes with high Ev.37. From the Fig. 2i, the sulfur SP@HSN-5 min sample can reach a high tap density around 0.83 g cm− 3, which is higher than even pure sulfur particles (0.724 g cm− 3), and much higher than sulfur AP@HTM sample (0.53 g cm− 3). This tap density is at the same level of commercial Si/C secondary particles (0.8 g cm− 3). The results indicate that such a secondary particle is favorable to boost the volumetric-energy-density (Ev) of electrode.
Mechanical properties of S-rich secondary particles are among the foremost properties considering their practical applications. It is important to understand the roles of mechanical properties of active materials (AMs) in controlling the overall quality of the final composite electrodes, which is seldom discussed in literature to our best knowledge. Here, we attempt to understand the mechanical properties of electrodes and their evolution during manufacturing. Firstly, electrode calendering is a well-known essential process that has complicated impacts on the electrode quality and energy density. Therefore, the mechanical properties of individual AM particles have to be strong enough to maintain its structure integrity and the structural stability of the electrode composite. Probably because classic electrode particles, such as LiCoO2 and LiFePO4, are hard inorganic materials, the AM mechanical properties are usually not an obvious issue. However, for new high-capacity electrode materials, such as sulfur and lithium metal, the situations are totally different. To dig out the fundamental principles, we introduce a concept of AM microenvironment below.
The concept of microenvironment in cell biology indicates significant influences on cell metabolism and division. As illustrated in Fig. 3a, the cell exchanges complicated nutrients and signals through the surrounding micro-vessels and neural network, which are the essential physical structures of the cell microenvironment. Similar to this picture in cell biology, the AM particle can also be viewed as an electrochemical micro-cell with its own microenvironment. As illustrated in Fig. 3a, the electron-transport-network (ETN) built by conductive nanoparticles, and the ion-transport-network (ITN) built by pores or voids together generate the physical structures of AM microenvironment for supporting charge (ions and electrons) transport during electrochemical reactions. Similar to the cell microenvironment, the electrochemical performance of individual micro-cells (AM particles) should be dependent on the detail physical structures and their function stability during cycling. Therefore, to regulate and stabilize the AM microenvironment in practical conditions is of great significance for high-performance energy storage devices, but is a challenging task for high specific capacity AMs such as sulfur, silicon and lithium metal.
From the AM microenvironment point of view, the calendering processing not only smoothies the electrode surface, but also shrinks the volume fraction of microenvironment (mainly the ITN part), improving the energy density by decreasing the electrolyte/AM ratio.29, 36, 38 At the same time, calendering is also a critical process regulating the physical structures and structural stability of the AM microenvironment and its assembling structures, that is, the entire ETN and ITN on the electrode level. In this case, the mechanical property of AM is critical for the structural evolution of microenvironment. As illustrated in Fig. 3b, for conventional electrode particles, such as LiCoO2, LiFePO4 etc., they are very hard materials. Therefore, the physical structures of AM microenvironment are mainly controlled by the stacking of AM and conductive agent particles under the compression force during calendering or device assembling. As long as the AM particles stack closely enough to reach a force equilibrium point against the compression, they finally form a strong network to mechanically protect the AM microenvironment, and the INT and ENT on the whole. Therefore, for thick sulfur cathode, the AM particles have to be hard enough or calendering-compatible to achieve a supportive AM microenvironment for efficient ion/electron conduction as illustrated in Fig. 3b. Unfortunately, it has been a big challenge to fabricate calendering-compatible S-rich AM particles. Therefore, the microenvironment issues come from the very beginning of electrode fabrication. As illustrated by Fig. 3c, most of S/C composite particles are weak due to the softness of sulfur in nature, such as the sulfur AP@HTM sample introduced in this study. In this case, calendering process can easily shape the AM particles into a closed stacking configuration, and crush the ITN microenvironment, leading to poor electrolyte wetting. In fact, the above microenvironment issues are not unique to sulfur cathode. With the intensively increasing interests on high-energy-density rechargeable batteries, most of the challenging issues of high-capacity AMs are actually from the microenvironment. For instance, the big volume change of silicon anode can easily destroy its ETN microenvironment, although silicon is a hard material. Therefore, microenvironment issues are generally found in various battery systems and throughout their service life. “Healthy” microenvironment will play a more and more important role in understanding the performance of next-generation LIBs and beyond.
For the sulfur SP@HSN particles, we have particularly investigated their significant advantages in fabrication of thick electrodes with an emphasis on the good calendering compatibility. Firstly, the microenvironment quality of the composite electrodes before calendering treatment was examined for samples with different S-loadings. For the control sample, sulfur AP@HTM, as shown in Fig. 3d, Fig. 3e and Figure S5a, many cracks can be observed when the S-loading is around 5 mg cm− 2. The sulfur AP@HTM even automatically detaches from the Al-foil when the S-loading reaches to 10 mg cm− 2 (Figure S6a). In contrast, the sulfur SP@HSN electrode with 5 mg cm− 2 S-loading (Fig. 3h, Fig. 3i and Figure S5b) displays uniform and crack-free surface. Furthermore, even with S-loading of 15 mg cm− 2, there is no visible crack for the sulfur SP@HSN-5 min electrode (Figure S6b). Electrode cracks can be viewed as uncontrollable big microenvironment defects, which can lead to structural non-uniformity and instability, and thereof non-uniformity of ion and current flux. The above structural differences convincingly confirm that the sulfur SP@HSN effectively improves the microenvironment quality with much better uniformity, even at a high S-loading level.
Importantly, the sulfur SP@HSN active material also demonstrates good calendering compatibility. To investigate the calendering compatibility, the sulfur AP@HTM and sulfur SP@HSN electrodes with S-loading of 5 mg cm− 2 were calendered at different pressures and temperatures. Figure 3f shows the electrode surface morphology for sulfur AP@HTM sample after treatment by 4 MPa at 40 oC. One can find that most of the AM is deformed and the majority of the pores disappear. Furthermore, when the AP@HTM electrode was compressed with 4 MPa at 80℃, the sulfur AP@HTM particles almost deformed into a compact sheet without porous structures (Fig. 3g and Figure S7c). This behavior even happened at a lower temperature around 60 ℃ (Figure S7b). In contrast, for the sulfur SP@HSN electrode, a much better calendering compatibility was confirmed at different temperatures. The particle shape and pores between particles survived at 4 MPa even at 80 ℃ (Fig. 3j, Fig. 3k, Figure S7d, Figure S7e and S7f). The good calendering compatibility at different temperatures indicates that the sulfur SP@HSN is not only mechanically strong, but also effective to confine the sulfur in the KB host at high temperatures. The good mechanical property is likely due to the unique applesnail-egg-like structures with uniformly dispersed hard Co nanoparticles inside. The good calendering compatibility of sulfur SP@HSN means good mechanical properties of the SP@HSN particles. To directly investigate the mechanical properties of sulfur SP@HSN, Atomic Force Microscopy (AFM) probe was randomly performed onto the AM particle as shown by the inset in Fig. 3l. The sulfur SP@HSN sample shows an average apparent Young’s modulus of 5.3 MPa, which is much higher than 4.2 MPa for the control sample. It is noted that this apparent Young’s modulus is not the real modulus of the individual AM particles that is difficult to test. However, it is a useful parameter describing the collective mechanical behavior of the AM particles, which provides critical information for choosing rational compression pressure during calendering processing.
To further study how the microenvironment structures change with compression pressure, the dependence of electrode porosity and thickness on compression pressure was systematically studied for the SP@HSN electrode in comparison with its control sample. Specifically, for the control sample, sulfur AP@HTM electrode, the thickness changes from 210 um to 50 um when the pressure increased from 0 MPa to 30 MPa. For the sulfur SP@HSN electrode, the electrode thickness changes from 207 um to 70 um at the same compression condition (Figure S8). At the same time, the electrode thickness of AP@HTM is much more sensitive to the compression pressure, as compared with that of SP@HSN electrode sample. This result indicates that the spherical SP@HSN helps to form more compact and strong stacking structures during electrode fabrication. Meanwhile, the electrode porosity was also recorded after treatment by different compression pressures. For the SP@HSN electrode, its porosity decreased from 60–40% when the pressure increased from 0 to 30 MPa (Fig. 3m). While, for sulfur AP@HTM electrode, the porosity dropped remarkably from 64–29% at the same condition.
The electrode porosity is a collective parameter describing the total volume fraction of the ion-transport-network (ITN) microenvironment. To build the final ITN, all the pores are filled by liquid electrolytes. Therefore, the liquid electrolyte wetting behavior can also reflect the ability of microenvironment to establish ion-conduction networks around the AM particles. As shown in video S1 and video S2, the sulfur SP@HSN electrode after 4 MPa calendering shows a faster wetting process than that for sulfur AP@HTM electrode, indicating that the sulfur SP@HSN electrode has better ITN microenvironment. The cross-section SEM images of sulfur AP@HTM and sulfur SP@HSN electrode before and after 4 MPa treatment further confirm the above results (Figure S9). As important as that of the ITN, the ETN part of the microenvironment controlling the electronic conduction is mainly determined by the dispersion of conductivity agent and the electronic contact between two different AM particles. To probe ETN part of the microenvironment, Fig. 3n shows the surface resistivity evaluated by conductive micro-probe at 20 different location of the electrode surface. One can find that the resistance and deviation for sulfur SP@HSN electrode is much less than that of sulfur AP@HTM electrode. Such a result indicates that the sulfur SP@HSN helps to form uniform ETN microenvironment across the whole electrode.
From the above results, one can find that the biomimetic S-rich secondary particles (SP@HSN) show several significant advantages in low-cost fabrication of high-quality thick sulfur-electrodes. Especially, all these advantages are fundamentally linked to the capability of microenvironment control during the manufacturing as well as the electrochemical cycling. In order to further investigate the electrochemical performance, the resultant S-cathodes with various sulfur loading were first evaluated with a coin cell using electrolyte/sulfur (E/S) ratio of ~ 10 ul mg− 1. As shown as in Fig. 4a, the sulfur SP@HSN electrode with a sulfur loading of 2 mg cm− 2 delivers a high discharge capacity of 1356 mAh g− 1 at current density of 0.1C, it further delivers discharge capacity of 1019, 877, 760, and 639 mAh g− 1 at current density of 0.2, 0.5, 1 and 2C, respectively, exhibiting an excellent rate capability. Moreover, the voltage profile in Figure S10 shows flat and long plateaus with low polarization even at high current densities, indicating good ITN and ETN in the sulfur SP@HSM electrode. On the contrary, the sulfur AP@HTM electrode (Fig. 4a) delivers obviously lower capacity of 1011, 712, 601, 522 and 430 mAh g− 1 at the current density of 0.1, 0.2, 0.5, 1, and 2C, respectively. Nyqusit plot also shows that the sulfur SP@HSN electrode has a smaller equivalent electrical circuit (R0) and charge transfer resistance (Rct) (Figure S11) as compared with the control sample, further verifying that the sulfur SP@HSN electrode has a more efficient ITN and ETN microenvironment, thus leading to an excellent rate capability. Furthermore, the cycling performances of sulfur AP@HTM and sulfur SP@HSN electrodes with same sulfur loading of 5 mg cm− 2 at a current density of 0.5C were compared in the Fig. 4b. After two cycles of activation at a current density of 0.1C, the sulfur SP@HSN electrode achieves a high initial capacity of 802 mAh g− 1 with good capacity retention of 66% over 110 cycles. In comparison, the sulfur AP@HTM electrode shows rapid capacity decay from 679 to 297 mAh g− 1, indicating a low sulfur utilization and severe “shuttling effect”.
Generally, for practical Li-S batteries, it is critical to develop electrodes with area capacity at least comparable to that of current LIBs. As shown in Fig. 4c, the sulfur SP@HSN electrode with 5 mg cm− 2 delivers a maximum areal capacity of 4.1 mAh cm− 2 at a current density of 0.5C. After 55 cycles, a reversible areal capacity of 3 mAh cm− 2 was still obtained, which is lower than that for the practical level of commercial LIBs (~ 4 mAh cm− 2), but much higher than the level reported in most of the literatures (~ 2 mAh cm− 2). Therefore, to demonstrate its superiority in energy-density as compared with current commercial LIBs, we further prepared thicker electrodes with 8.1 and 9.7 mg cm− 2 of sulfur. These thick electrodes deliver areal capacity of 5.3 and 7.5 mAh cm− 2 at a current density of 0.2 C, respectively, which is much higher than that of the commercial LIBs. To further investigate its possible application in super-high energy density batteries, the sulfur loading was even increased to 14.1 mg cm− 2. As shown by Fig. 4d, a super-high area capacity of 16.5 mAh cm− 2 was achieved at a low current density of 0.5 mA cm− 2 for this super-high S-loading electrode. Finally, a stable areal capacity (~ 12 mAh cm− 2) at a current density of 1 mA cm− 2 was realized over 28 cycles, indicating its promising application in low-cost high-energy-density metal-sulfur batteries.
The above success in coin cells is a good indication that the calendering-compatible biomimetic SP@HSN is able to deliver a high specific capacity even at a high S-loading condition in thick electrodes. It is noted that the good electrochemical performance of the SP@HSN electrode is fundamentally contributed by the ITN and ENT microenvironment as introduced previously. In other words, the superiority in microenvironment control by the SP@HSN in practical conditions is one of the most critical reasons for its good electrochemical performance. To test this advantage in large scale electrodes, pouch-cells with size of 4.7 cm ⋅ 7.7 cm were assembled based on various sulfur loading for the sulfur SP@HSN electrodes. At the same time, a low E/S ratio of ~ 4 ul mg− 1, was used as demonstrated in the Fig. 5a. All the pouch-cells based on various sulfur loading of cathodes from 5 mg cm− 2 to 12 mg cm− 2 exhibited a high discharge specific capacity and good cycling stability (Figure S12a). Meanwhile, all the voltage profiles of pouch-cells (Figure S12b-Figure S12d) are almost identical to those of coin-cells and the average discharge voltage is around 2.1 V even after several cycles, suggesting a lower polarization and a good health state of cell. Specifically, Fig. 5b shows that the pouch-cell with a high sulfur loading of 12 mg cm− 2 delivers a high discharge capacity of 1294 mAh g− 1 at a current density of 0.02C and exhibits a cycling performance over 10 cycles. Subsequently, the energy density of pouch-cell was evaluated by a simplified model based on Eq. 2 and Eq. 3. As shown in Fig. 5c, our pouch-cell with an E/S ratio of 4 ul mg− 1 and sulfur loading of 12 mg cm− 2 shows a high joint Eg (430 Wh kg− 1) and Ev (1002 Wh L− 1) at 0.02C and good cycling stability. Such an energy-density level is much higher than the commercial LIBs and the Li-S batteries reported by most of the literatures,35, 39, 40 as shown in Fig. 5d. It is noted that optimization of mass loading, E/S ratio and other important parameters could be employed to further boost the energy density of pouch-cell.23 For instance, if we further decrease the E/S ratio to 3 ul mg− 1, the energy density of the pouch-cell can improved to be above 500 Wh kg− 1, the goal set by the Department of Energy, USA.
It is noted that the above performance for the calendering-compatible biomimetic SP@HSN electrodes was achieved without additional efforts on electrolyte, binder, separator and lithium anode. The high specific capacity, areal capacity and good cycling stability of sulfur SP@HSN electrode with super-high loading is likely contributed by the following factors. 1) The biomimetic design of the SP@HSN sulfur active material with a hard core-shell structure can help to suppress the shuttling of LiPSs as illustrated in Fig. 1c, which has been widely reported in literature.29, 35, 41 2) The good calendering-compatibility of the SP@HSN can help to build a “healthy”, stable and uniform ITN/ETN microenvironment even for thick and large-scale electrodes. This is critical for the capacity extraction of individual AM particles. 3) The uniform ITN/ETN microenvironment finally assembles into uniform ITN/ENT structures on large-scale electrode level, especially for thick electrode. This can remarkably alleviate the formation of Li2S/Li2S2 precipitates on the interface between cathode and separator, greatly improving the cycling stability of Li-S batteries. 4) A uniform ITN/ETN microenvironment and its assembly structures on electrode level can help to generate uniform Li+-flux onto the lithium metal side that is beneficial to stabilizing Li metal as well.