As is known, lithium-ion batteries (LIBs) (Liu et al., 2010; Dunn et al., 2011; Goodenough, 2014) have been utilized in countless commodities, such as mobile phones and electric vehicles. However, its wider applications are impeded owing to limiting materials, so there is still great potential as long as more novel electrode materials are exploited for LIBs.

So far, various metal selenides, such as FeSe₂ (Kong et al., 2019; Yousaf et al., 2020), MnSe(Li et al., 2016; Liu et al., 2018; Li Z. et al., 2019), and CoSe₂ (Yu N. et al., 2019; Xu et al., 2020) have been deeply studied as active materials in LIBs. Nevertheless, it presents the challenges of inferior rate capability and rapid capacity loss. Some researchers have also focused on phosphorus-based materials such as black phosphorus (Dahbi et al., 2016; Del Rio Castillo et al., 2018), SiP (Duveau et al., 2016), FeP (Jiang et al., 2017), and GeP (Li W. et al., 2019; Li X. et al., 2019; Fan et al., 2019), which also exhibits the inferior cycling performance. Nevertheless, improving the electrochemical performance of metal sulfides and phosphides in terms of rate capability and cycling stability is an enormous challenge due to the large volume change that can lead to crushing and loss of electrical contact.

Recently, a novel ternary metal phosphide sulfide/selenides (MPX₃, M is transition metal, such as Mn, Zn, etc; X is Se or S (Brec, 1986; Pei et al., 2018; Gusmao et al., 2019; Samal et al., 2021), have been investigated in electrocatalysis (Byvik et al., 1982; Dedkov et al., 2020), hydrogen storage (Cabria and El-Meligi, 2018) and toxicological hazards (Latiff et al., 2018). However, the MPX₃ empolyed as active materials in LIBs are rarely reported. The layered MPX₃ owns attractive lithium storage ability for rechargeable ion batteries. Its unique two-dimensional (2D) layered nanostructure, which is composed of weak van der Waals stacking between layers, is deemed to be an ideal framework for fast Li⁺ storage. The layered structure alleviates the volume stress, generates abundant ion diffusion pathways and speedy electron transportation owing to lower energy barrier (Fan et al., 2019; Ding et al., 2020). More importantly, the preferred bandgaps\ of MPX₃ (1.3–3.5 eV) (Wang et al., 2018) and potential ionic conductivity make MPX₃ as superior anode electrode materials. Some groups have reported like-MPX₃, such as MnPS₃ (Sang et al., 2020), CoPS₃ (Jana et al., 2020), FePSe₃ (Xing et al., 2020), NiPS₃ (Dangol et al., 2018) and SnPSe₃ (Ren et al., 2020), exhibit a promising performance of lithium/sodium ions storage.

The manganese phosphorous selenide (MnPSe₃) is one of MPX₃, isostructural with FePSe₃. The MnPSe₃ possesses an interlayer spacing of∼0.32 nm (Li et al., 2014), much larger than the diameter of Li⁺ (∼0.152 nm), providing channels for Li⁺ diffusion in the insertion/extraction process. MnPSe₃ as anodes also have a high theoretical capacity of 746 mA h g⁻¹ by forming Li₃P and Li₂Se alloy (Li et al., 2013; Gusmão et al., 2019; Tang et al., 2020). However, the numerous overlapping layers of bulk MnPSe₃ lead to a decreasing active surface area, slow charge transfer, and even poor resistance to volume expansion in lithiation/delithiation. According to previous studies (Abdelkader and Kinloch, 2016; Chen et al., 2016; Shen et al., 2020), mechanical exfoliation can effectively narrow the size and thickness of bulk materials, causing abundant exposed active sites, highly tunable morphology, reduced diffusion length of charge carriers for Li⁺ and perfect resistance to volume change.

In this work, the lithiation/delithiation processing of bulk MnPSe₃ as the anode in LIBs has been disclosed by half-cell. It occurs to serious volume expansion/contraction for bulk MnPSe₃ in lithiation/delithiation, even the tracking MnPSe₃ electrode is unable to maintain high integrity with serious cracks or pulverization. Thus, we reduced bulk MnPSe₃ to a few layered MnPSe₃ nanoflakes by mechanical exfoliation and comprehensively compared the storage Li⁺ performances of exfoliated MnPSe₃ to that by grinding as LIBs anodes. Moreover, the exfoliated MnPSe₃ electrode shows an initial discharge/charge capacity of 524/796 mA h g⁻¹, and a retention rate of 88% and 72% after 100 and 350 cycles, respectively at 0.2 A g⁻¹. The improved resistance to expansion and pulverization and rapid reaction kinetics indicate the exfoliated MnPSe₃ is able to achieve superb cyclic stability. Ultimately, exfoliated MnPSe₃ is considered a great dynamism and potential anode material with predominant performance in LIBs.

As clearly displayed by the low-magnification TEM in Figure 1A. MnPSe₃ exhibits an ultrathin and transparent lamellar appearance with several micrometers in plane and nanometers in thickness. The typical SAED pattern of MnPSe₃ show the diffraction spots of (-11-2), (-10-1), (0-11) plane with corresponding d-spacings of 4.8, 5.29, and 5.31 Å, respectively viewed along [-111] in Figure 1B, which is commensurate to HRTEM along [-111] in Figure 1C. It indicates that the as-prepared MnPSe₃ possesses high crystallinity and phase purity.

The schematic images of layered MnPSe₃ viewed from the b and c axes are shown in Figure 1D. MnPSe₃ belongs to a hexagonal with the lattice parameters of a = 6.387 Å, b = 6.387 Å, c = 19.996Å; and the angle α, β is 90°, γ is 107.35°. As depicted in Figure 1A, a single layer is composed of the Mn atom’s central shell and the other two shells in PS₃ units. The selenide atoms are located on the two external surfaces of a MnPSe₃ layer (Li et al., 2014; Pei et al., 2018). Moreover, the spacing of two adjacent MnPSe₃ layers is 3.2 Å by Van der Waals (Li et al., 2014), which provides channels and buffers volume expansion/contraction for Li⁺ insertion/extraction.

The chemical compositions of MnPSe₃ nanoflakes were further investigated by XPS in Figures 1E–G. As revealed by the Mn 2p spectrum, the high-resolution Mn 2p profile can be mainly fitted at 640.7 eV (2p3/2) and 651.3 eV (2p1/2), ascribed to the binding energy of Mn²⁺, while the peaks at 642.3 eV (2p3/2) and 657.6 eV (2p1/2) indicate the presence of Mn⁴⁺ (Sang et al., 2020). The narrowly scanned XPS spectrum of P 2p can be contributed to double peaks at 134.5 and 133.4 eV, which are in line with the P 2p3/2 and P 2p1/2, respectively (Edison et al., 2018; Fan et al., 2019). Additionally, there are only a pair of peaks at 166.4 eV (2p3/2) and 160.7 eV (2p1/2) for the Se 2p (Gusmão et al., 2017; Dedkov et al., 2020), indicating only one form of selenium existed in the MnPSe₃, in agreement with aforementioned results.

Electrochemical behaviors of bulk MnPSe₃ have been tested in Figure 2. There are mainly four obvious plateau regions at 2.0–1.8 V, 1.75–1.58 V, 1.53–1.15 V, and 0.75–0.35 V in the first discharge curve, while the charge profile also shows three corresponding three plateaus at 2.07–2.34 V, 1.71–1.94 V, and 1.1–1.38 V, respectively in Figure 2A. Significantly, the bulk MnPSe₃ electrode illustrates the rate capabilities of 0.2, 0.4, 1, 2, and 4 A g⁻¹in Figure 2B. With the increase of current density, the specific capacities decay obviously for the MnPSe₃ electrode. When the current density reaches up to 4 A g⁻¹, the reversible capacity of 35 mA h g⁻¹ is just left. The bulk MnPSe₃ delivers a first 412/550 mA h g⁻¹ discharge/charge capacity with ∼75% initial Coulombic efficiency in a potential of 0.005–3 V at 0.2 A g⁻¹, and an extremely obvious downward trend with a retained capacity of 344 mA h g⁻¹ after 50 cycles in Figure 2C. It indicates the cycling durability for the bulk electrode is really poor. The bulk MnPSe₃ electrode possesses an inferior rate capability and more unstable cycling performance.

In order to further improve the Li⁺ storage performance of MnPSe₃, the bulk MnPSe₃ was refined to nanoflakes by mechanical exfoliation. We compare the morphology between bulk MnPSe₃ and exfoliated MnPSe₃ by SEM in Figures 3A, B; Supplementary Figure S2. Compared to bulk MnPSe₃, exfoliated MnPSe₃ nanoflakes display a smaller and more uniform size. As shown in Figures 3C,D, the size distribution of MnPSe₃ nanoflakes was measured by particle size analysis. The size of bulk MnPSe₃ by hand grinding reaches ∼1.2 μm. However, exfoliated MnPSe₃ nanoflakes have been largely narrowed to tens of nanometers. Moreover, exfoliated MnPSe₃ electrode exhibits more remarkable electrical conductivity than bulk MnPSe₃ in Supplementary Figure S3. According to reported articles (Chen et al., 2016; Dangol et al., 2018; Yu Z. et al., 2019), reducing the size and thinning the thickness of bulk 2D materials can effectively improve abundant exposed active sites and resistance to expansion/shrinkage and shortened diffusion length of charge carriers for Li ions and in the process of Li⁺ insertion and extraction. In addition, as shown in Figure 3E, MnPSe₃ nanoflakes were prepared by two-step method. The liquid-phase ultrasonic exfoliation does not involve in phase transformation and any new phases formation. In addition, this method achieves controllable size nanoflakes and high repeatability. The obtained MnPSe₃ nanoflakes exhibits further enhancement on fast chargeability and long cyclability of Li⁺ storage. Firstly, bulk MnPSe₃ were crumbled roughly by ultrasonic stripping. Secondly, smaller MnPSe₃ nanoflakes effectively were separated by fractional centrifugation.

Furthermore, we examined the discrepancy of the bulk and exfoliation MnPSe³ in morphology and EIS, respectively in Figure 4. The side surface of exfoliated MnPSe³/CNT/CMC electrode shows serious cracking, reaching ∼15 μm due to severe volume expansion/shrinkage after full lithiation/delithiation in Figures 4A,B, which is a key cause of rapid failure for bulk MnPSe³ electrode. While it was found that exfoliated MnPSe³ remained integrity after 100 cycles. This clearly further demonstrates that exfoliated MnPSe³ electrodes resist severe volume expansion owing to possessing excellent mechanical robustness. Interestingly, after 100 charge/discharge cycles, the thinner and smaller layered MnPSe³ electrode obtains a lower transfer resistance than bulk MnPSe³ owing to the contact separation of electrode internal components in Figure 4C, even falling from the current collector, resulting in the decreasing of electrical conductivity and ion transport properties. Thus, the exfoliated layered MnPSe³ electrode facilitates Li⁺ extraction from the insertion region.

To further confirm the phase transformation of MnPSe₃ in lithiation/delithiation, X-Ray Diffraction (XRD) has also been performed on the MnPSe₃ electrode in Figure 5A. The pristine MnPSe₃ electrode exhibits the obvious crystallographic orientations of (003), (006), and (113), and no detectable impurities were found. The ex-situ XRD of the MnPSe₃ anode presents Li₃P peaks at about 26.6°, 33.8°, and 44.3° (Kim and Cho, 2009), and Li₂Se peaks at about 25.1° and 22.6° after the first full lithiation (Liu et al., 2020), which further verify single-crystal MnPSe₃ is entirely alloyed to Li₃P and Li₂Se phase. While upon full delithiation, it presents a new peak at 33.1°, which is caused by the MnSe phase (Xue and Fu, 2007). The marked peaks located at other degrees originate from electrolyte decomposition on the surface of the MnPSe₃ electrode, which is in good agreement with reported results about like-MPX₃. Significantly, the differential capacity profiles display excellent reversibility in a redox reaction and agree well with the ex-situ XRD analysis, which also presents the reduction peak at 1.95, 1.66, 1.30, and 0.6 V, could correspond to the alloying reactions of Li_xMnPSe, Li₂Se/P/Mn Li₃P, and SEI, respectively. Considering the analysis above, phase transformation in first lithiation/delithiation could be summarized as follows:

After First lithiation: M n P S e 3 + 9 L i + + 9 e − → L i 3 P + M n + 3 L i 2 S e (1)

After First delithiation: L i 3 P + M n + L i 2 S e → M n S e + P + 5 L i + + 5 e − (2)

As shown in Figure 6, the exfoliated MnPSe₃ anode is further utilized in a half-cell. Primarily, the typical cyclic voltammogram (CV) curves of the electrode were illustrated in Figure 6A. The initial cathodic sweep displays four distinct reduction peaks at 1.95 V, 1.6 V, 1.2 V, 0.6 V, and 0.35 V, indicating the lithiation/delithiation process is a multiple-step. The prominent peak located at 1.95 V is matched to Li _x MnPSe₃. The peaks centered at 1.6 V and 1.2 V are associated with the formation of Li₂Se, P, and Mn. The peak at 0.6 V is related to the generation of Li₃P. Another weak broad peak located at 0.35 V is attributed to the side reaction (formation of SEI film). In the following anodic sweep. The three strong peaks at 1.38 V, 2.0 V, and 2.27 V are coincident with dealloying of Li₃P, Li₂Se, and the formation of MnSe. The result above is similar to CVs profiles for like-MPX₃ (FePSe₃, NiPS_3, etc) (Sang et al., 2020; Xing et al., 2020; Liu et al., 2021). In addition, the initial three cycles curves exhibit a consistent property of reaction to that of the CV results above in Figure 6B. The charge/discharge curves and CVs in multiple cycles are also nearly overlapped, suggesting the wonderful stability of electrode.

To evaluate the lithium storage properties of the exfoliated MnPSe₃ at _a high rate, the reversible capacities of 616, 562, 458, 348, and 242 mA h g⁻¹ were obtained at current densities of 0.2, 0.4, 1, 2, and 4 A g⁻¹, respectively in Figure 6C. Moreover, the capacity achieves 331, 412, 502, and 568 mA h g⁻¹ with the current density coming back to 2, 1, 0.4, and 0.2 A g⁻¹, illustrating the MnPSe₃ electrode maintains a remarkable rate performance. Compared to bulk MnPSe3, the cycling stability of the exfoliated MnPSe₃ is also effectively improved. It maintains outstanding cycling stability with capacity retention of 709 mA h g⁻¹ after 100 cycles at 0.2 A g⁻¹, and capacity retention of 578 mA h g⁻¹ after 350 cycles at 0.2 A g⁻¹in Figure 6D; Supplementary Figure S4.

In summary, this work researches on performance improvement of the MnPSe₃ as the anode of LIBs in detail by ultrasonic exfoliation, revealing an extraordinary ability to resist volume expansion/shrinkage in full lithiation/delithaition, which provides significant evidence for the research of like-MPX₃. The thinner and smaller MnPSe₃ shows superior performance to the bulk electrode material. When supplied as the anode of LIBs in half-cell, a splendid reversible capacity of 709 mA h g⁻¹ was maintained for the MnPSe₃ within the potential window of 0.005–3 V vs. Li⁺/Li after 100 cycles at 0.2 A g⁻¹. While further improving the cycles, a specific capacity of 578 mA h g⁻¹ was still held after 350 cycles, which benefits from the favorable capacitance kinetics, and resist severe volume expansion. Layered MnPSe₃ as anode materials for LIBs meet the needs of high capacity, rapid charge-discharge, and long cycle.