Li-O₂ battery has received much attention for its high theoretical energy density (3500 Wh/kg) [1, 2]. The reaction principle of typical Li-O₂ battery is Li₂O₂-based reversible formation and decomposition [3]. Li₂O₂, the product of discharge, is dielectric and insoluble, whose accumulation on the cathode surface will impede the transport of oxygen, electrons and lithium ions, thus limiting oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) and causing slow kinetics of electrochemical reaction, high over potential, poor cycling stability, etc. [4, 5]. Over the past few decades, many studies have focused on cathodic catalysts to effectively promote Li₂O₂ decomposition and enhance the performance of Li-O₂ battery [6-10].

MXene is an emerging 2D layered compound material made from metal carbides or nitrides. It is an ideal substrate material for development of high-performance electrocatalysts because of its high electrical conductivity, large surface area and extensive chemical species [11-13]. Domestic and foreign researchers have done much work on it with respect to energy storage and conversion [14-24]. To meet actual demand, we are also in urgent need of electrocatalysts that can catalyze OER and ORR at the same time so as to enhance its performance in Li-O₂ battery.

Transition metal phosphides (TMPs) have been receiving more and more attention for their excellent electrocatalytic activity [25-27]. Systematic research has been done on the use of nickel phosphide, ferrous phosphide, molybdenum phosphide and cobalt phosphide as catalysts. Nonetheless, taking CoP for example, the binding energy of Co-P compound are negative, indicating that it is thermodynamically stable. Density functional theory calculations show that CoP exhibits metallic characteristics, and the bonding behavior between Co and P atoms in Co-P compounds is a combination of covalent and ionic properties [28]. Therefore, CoP can be used as electrocatalyst because of its excellent redox performance. However, its performance is still unsatisfactory mainly for its poor electrical conductivity [29, 30]. Therefore, we may combine it with a high conductive material to enhance its electrical conductivity, thus improving its reactivity and stability.

Here, we develop a simple method to grow CoP nanoparticles (CoP NPs) on the surface of MXene and between the layers, thereby forming the CoP/MXene composite catalyst. It not only prevents CoP-NPs clustering and provides enough active sites, but also significantly improves the accessibility of electrolytes and promotes charge/mass transfer and quick release of gasses during electrocatalysis. The total density of states (TDOS) was used to simulate and evaluate the electrochemical activity. Figs. 1a and c were schematic diagrams of the optimized electronic structure. The TDOS for CoP and CoP/Ti₃C₂T_x composite were investigated as shown in Figs. 1b and d. The TDOS of the CoP/Ti₃C₂T_x remained the metallic properties of Ti₃C₂T_x with the DOS passing the Fermi energy. And CoP/Ti₃C₂T_x demonstrated the high DOS at Fermi energy than that of CoP, indicating that CoP/Ti₃C₂T_x was endowed with more active electron, which were easy to be accepted and lost and facilitated the electrochemical reaction [31].

The preparation process of CoP/Ti₃C₂T_x composite was shown in Fig. 1e. With the influence of HF etching, the Al atom layer of Ti₃AlC₂ was successfully removed. Ti₃C₂T_x was mixed with Co(AC)₂.4H₂O and heated at 120 ℃ for 10 h, and then the obtained CoP/Ti₃C₂T_x composite was prepared successfully by calcinated at 500 ℃ in sodium hypophosphite atmosphere, which was also confirmed by the X-ray diffraction (XRD) results in Fig. 1f. With the effect of HF etching, the characteristic peak at 39.0° disappeared, which proved that Al atom layered in Ti₃AlC₂ was removed and the formation of Ti₃C₂T_x. The accordion-like lamella of Ti₃C₂T_x provides growth sites for the synthesis of CoP/Ti₃C₂T_x composite. The peaks at 31.6°, 36.3°, 46.2°, 48.1° and 56.8° were corresponding to (011), (111), (112), (211) and (301) lattice planes of CoP (JCPDS No. 29–0497), indicating that CoP was grown on the composite. The peak at 60.7°, as the laminate peak of Ti₃C₂T_x, remained in the CoP/Ti₃C₂T_x composite, which proved that the laminate structure of Ti₃C₂T_x was not destroyed with the high-temperature calcination [32, 33]. The decrease of exposure intensity of Ti₃C₂T_x in composite is due to better crystallinity and higher diffraction intensity of CoP sample. The successful preparation of the composite was also verified by Raman spectrum. Fig. S1 (Supporting information) presented a Raman spectrum of CoP. Several detectable signals were shown on small Raman shifts [34, 35], and there were two distinguishable Raman activity patterns in CoP/Ti₃C₂T_x composite, including characteristic peaks of CoP and Ti₃C₂T_x, which proved the correct synthesis of CoP/Ti₃C₂T_x composite.

The microstructure and morphology of CoP/Ti₃C₂T_x were observed through scanning electron microscopy (SEM) and high-resolution transmission electron microscopy (HRTEM). Compared with block Ti₃AlC₂ (Fig. S2 in Supporting information), the layer spacing of Ti₃C₂T_x (Fig. 2a) increased by the influence of HF etching. The lattice spacing of 0.251 nm were attributed to the (012) plane of the accordion-like Ti₃C₂T_x (Fig. S3 in Supporting information). The prepared CoP NPs showed obvious agglomeration phenomenon without adding the substrate of Ti₃C₂T_x (Fig. 2b). It was obvious that the size distribution of CoP NPs was about 100 nm (Fig. S4 in Supporting information). In contrast, with the addition of Ti₃C₂T_x substrate, the structure of MXene layer remained original after hydrothermal and calcination treatment, which provided enough space for loading of CoP NPs uniformly (Fig. 2c). The HRTEM of CoP/Ti₃C₂T_x (Fig. 2d) showed that the lattice spacing of Ti₃C₂T_x (012) was 0.251 nm. Lattice stripes at a spacing of 0.248 nm were attributed to the (111) plane of CoP NPs (JCPDS No. 29–0497). The selected area electron diffraction (SAED) proved that the prepared CoP/Ti₃C₂T_x had good crystallinity and the single crystal electron diffraction of Ti₃C₂T_x (Fig. 2e). The diffraction rings of CoP were consistent with (111), (211) and (301) planes. The uniform distribution of CoP NPs on Ti₃C₂T_x was further demonstrated by energy dispersive X-ray spectrometer (EDS) (Fig. 2f). The above characterization confirmed the successful preparation of CoP/Ti₃C₂T_x composite.

In order to clarify the composition and valence distribution of elements in CoP/Ti₃C₂T_x composite, we analyze the electrocatalyst by XPS. Fig. S5 (Supporting information) showed the presence of Ti, C, O, Co and P elements in the measured spectra. In Co 2p spectra (Fig. 3a), the two peaks located at 778.7 and 793.7 eV were identified with Co-P bond in CoP [36], while 781.3 and 798.0 eV correspond to the oxidized Co (Co²⁺) species, which caused by the partial oxidation process on CoP species [37]. Fig. 3b presented the XPS spectra of P 2p. The two peaks which were located at 129.1 eV and 129.9 eV were subject to the Co-P bond and the binding energy at 133.8 eV corresponded to the PO_x species, respectively [38]. Compared with Ti₃C₂T_x, metal-oxygen bonds were not identified, indicating the surface-functionalization process derived product covered the surface of Ti₃C₂T_x further constructing the composite. In the spectrum of Ti 2p (Fig. 3c), the peaks at 455.1, 456.0, 457.2 and 458.7 eV of Ti 2p_3/2 correspond to Ti-C, Ti²⁺, Ti³⁺ and Ti-O, respectively[39]. Compared with Ti 2p in Ti₃C₂T_x (Fig. S6 in Supporting information), the intensity of Ti-O peak in CoP/Ti₃C₂T_x increased notably, which caused by partial oxidation of Ti atoms in Ti₃C₂T_x. In Fig. 3d, the spectrum of C 1s were matched to four peaks located at 282.2, 284.8, 286.7 and 288.9 eV which belonged to C–Ti, C–C, C–O and C=O bonds, respectively [40].

Moreover, it is well known that the specific surface area and pore size are significant to electrocatalyst performance. They were studied by N₂ adsorption-desorption isotherm. The specific surface area of CoP/Ti₃C₂T_x and Ti₃C₂T_x were 77.31 m²/g and 24.04 m²/g, respectively (Figs. S7a and b in Supoprting information). The N₂ adsorption-desorption isotherm of CoP/Ti₃C₂T_x showed a typical hysteresis loop which caused by the presence of mesoporous. The specific surface area of CoP/Ti₃C₂T_x increased obviously, which was mainly due to the adhesion of CoP NPs between the surface and interlayers of Ti₃C₂T_x. The increased specific surface area was beneficial to fully expose the active site, which can improve the performance of Li-O₂ battery. Figs. S7c and d (Supoprting information) exhibited the pore size of CoP/Ti₃C₂T_x and Ti₃C₂T_x obtained by the BJH method. The pore size of Ti₃C₂T_x was about 2.440 nm, and CoP/Ti₃C₂T_x composite was mostly 2.441 nm. The corresponding pore size distribution of Ti₃C₂T_x was located in the mesoporous range which was due to the aggregation of accordion-like nanosheets [6]. These results suggested that the composition of the CoP and Ti₃C₂T_x could prevent the aggregation of the CoP particles and provide the more and more active sites.

As a new 2D material, Ti₃C₂T_x features a high specific surface area, superior electronic conductivity and adjustable components, etc., which can facilitate oxygen adsorption in a reaction process of Li-O₂ battery and provide a proper space for discharge products storage. With good ORR/OER catalytic activity, CoP has the potential to reduce overpotential on charge process of Li-O₂ battery. Therefore, CoP/Ti₃C₂T_x is expected to become an outstanding cathode material suitable for Li-O₂ battery. To explore electrochemical mechanism of Li-O₂ battery with CoP/Ti₃C₂T_x cathode, electrochemical performance testing was carried out. As can be observed from Fig. 4a, Li-O₂ battery with CoP/Ti₃C₂T_x cathode provided a high discharge capacity of 17, 413 mAh/g in the 1^st cycle at 100 mA/g, and in the 2^nd and 3^rd cycles, their discharge capacity reached 11, 060 mAh/g and 8536 mAh/g, respectively, which was superior to CoP and Ti₃C₂T_x (Fig. S8 in Supoprting information). The reason why specific capacity declines in such three cycles may be that it is difficult for discharge products to be completely decomposed during charge and thus no novel cathodes can be provided in next cycle. In comparison with CoP/Ti₃C₂T_x, specific capacity performance of pure CoP or Ti₃C₂T_x was poor; in the first cycle, their specific discharge capacity were proved to be 10, 560 mAh/g and 9362 mAh/g, respectively. According to Fig. 4b and Fig. S9 (Supporting information), discharge capacity of Ti₃C₂T_x, CoP and CoP/Ti₃C₂T_x in the initial cycle were compared in diverse electric current densities. In the case where CoP/Ti₃C₂T_x was used as a cathode catalyst, the corresponding specific discharge capacity were high up to 17, 413 mAh/g, 9970 mAh/g and 8209 mAh/g at 100, 200 and 500 mA/g, which were above those of pure CoP (10, 560 mAh/g, 9115 mAh/g and 6462 mAh/g) or pure Ti₃C₂T_x (9362 mAh/g, 7215 mAh/g and 5529 mAh/g). This manifests that CoP/Ti₃C₂T_x composite can effectively improve electrochemical performance of Li-O₂ battery. By contrast to CoP/Ti₃C₂T_x, Ti₃C₂T_x with good conductivity was short of highly active sites and thus showed insufficient electrochemical catalytic activity, enabling Li-O₂ battery with Ti₃C₂T_x cathode to fail to run stably and efficiently. Even when the electric current density reached up to 500 mA/g, CoP/Ti₃C₂T_x electrocatalyst that combined high conductivity of Ti₃C₂T_x and catalytic activity of CoP NPs still remained its discharge capacity at 8209 mAh/g. As presented in Fig. 4c, the overpotential of CoP/Ti₃C₂T_x cathode was 1.25 V, lower than pure CoP (1.36 V) or pure Ti₃C₂T_x (1.56 V), which provided that the electric current density was 500 mA/g and the specific capacity limited to 500 mAh/g. In line with such a phenomenon, it is clear that CoP/Ti₃C₂T_x composite possess higher activity of ORR and OER. The cycle performance was tested as shown in Fig. 4d. The CoP/Ti₃C₂T_x cathode was able to stably run for 40 cycles at 500 mA/g. Moreover, the cyclic stability of CoP/Ti₃C₂T_x was significantly better than CoP and Ti₃C₂T_x (Fig. S10 in Supoprting information).

To learn about electrochemical catalysis of CoP/Ti₃C₂T_x in Li-O₂ battery, cyclic voltammetry (CV) based tests were performed, as given in Fig. 4e. The initial redox potential and the peak current intensity of CoP/Ti₃C₂T_x were respectively above those of pure CoP or pure Ti₃C₂T_x, signifying that its ORR/OER catalytic activity was superior to Ti₃C₂T_x. Electrochemical impedance spectroscopy (EIS) is aimed at profoundly investigating electrochemical performance of CoP/Ti₃C₂T_x composite. Nyquist plots of CoP, Ti₃C₂T_x and CoP/Ti₃C₂T_x before the cycle have been depicted in Fig. 4f. All EIS images were semi-circles at high frequency, but become linear at low frequency. Relevant fitting data were listed in Table S1 (Supporting information). In this table, the electrolyte resistance of Ti₃C₂T_x (R_s = 11.80 Ω) was significantly lower than CoP (R_s = 12.26 Ω) and CoP/Ti₃C₂T_x (R_s = 12.03 Ω). With regard to charge transfer impedance, Ti₃C₂T_x (R_ct = 29.74 Ω) was apparently lower than CoP (R_ct = 44.93 Ω) and CoP/Ti₃C₂T_x (R_ct = 30.79 Ω). However, the impedance of CoP/Ti₃C₂T_x was significantly lower than that of CoP, showing that the introduction of Ti₃C₂T_x provided a conductive substrate for CoP NPs. Results described above indicate that Ti₃C₂T_x can effectively improve electronic conductivity of cathodes. Better performance of CoP/Ti₃C₂T_x composite is derived from outstanding electrical conductivity of Ti₃C₂T_x and active catalytic sites of CoP NPs which prevent termination of electrochemical reactions.

To explore the discharge-charge products of Li-O₂ battery for deeply understanding of electrochemical mechanism, the morphology changes of the CoP/Ti₃C₂T_x, CoP and Ti₃C₂T_x electrodes in the first cycle were observed in Fig. 4g and Fig. S11 (Supporting information). The fresh electrode showed that the clear accordion-like surface. After the first complete discharge process (Fig. 4g), a large number of flower-like discharge products deposited by nanosheets were generated on the surface of CoP/Ti₃C₂T_x. The reason for such flower-like discharge product is that the strong interaction between CoP/Ti₃C₂T_x and Li₂O₂ makes it difficult for external Li₂O₂ monomers to aggregate into large Li₂O₂, leading to the isolation of Li₂O₂ on the surface of CoP/Ti₃C₂T_x and easier decomposition during charging process [41]. However, the contact interface between Ti₃C₂T_x and discharge products is unsatisfied, resulting in poor electron and oxygen transport in the subsequent charging process. In addition, the discharge products with small size and discrete distribution lead limited discharge capacity [42]. Therefore, CoP/Ti₃C₂T_x composite exhibited low overpotential and high discharge capacity. According to the XRD analysis (Fig. 4h and Fig. S12 in Supporting information), there were several characteristic peaks after the first discharging. The peaks at 32.8°, 34.8° and 58.5° were attributed to the crystal planes of (200), (201) and (220) in Li₂O₂ (JCPDS No. 73–1640), proving that the characteristic diffraction peaks of Li₂O₂ were very obvious after discharging and peaks of Li₂O₂ had disappeared on the electrode surface after charging. This result was also demonstrated by Raman spectroscopy (Fig. 4i), the characteristic peak of Li₂O₂ at 793 cm⁻¹ is obviously [7]. After the charging process, the flower-shaped discharge product disappeared and the accordion-like CoP/Ti₃C₂T_x structure with nanoparticles grown on the surface recovered, indicating that Li₂O₂ had completely decommissioned during the charging process, and the electrode catalyzed by CoP/Ti₃C₂T_x had good catalytic ability of ORR and OER with the proposed mechanism. Fig. S11 showed the surface topography of the Ti₃C₂T_x electrodes products. It was visibly that the number of toroid-shaped materials on the accordion Ti₃C₂T_x was very small after the first discharge, which meant the Ti₃C₂T_x cathodes didn't have enough capability to catalytic conversion more discharge product. High conductive Ti₃C₂T_x structure was used as conductive network, which promoted the transfer rate of electrons and Li⁺. CoP NPs with high catalytic activity spread evenly between the surface of CoP/Ti₃C₂T_x layers, which increased specific surface area of the material to accommodate more discharge product, thus the capacity of the electrode reaction was improved to achieve long cycle.

To sum up, CoP/Ti₃C₂T_x composite with more active electron was designed through theoretical simulation, and the CoP NPs were uniformly distributed on the surface and interlayer of Ti₃C₂T_x. When using the prepared CoP/Ti₃C₂T_x as an electrocatalyst of Li-O₂ battery, it not only provided high specific discharge capacity (17, 413 mAh/g at 100 mA/g) but also exhibited cyclic stability and low overpotential (1.25 V). As CoP/Ti₃C₂T_x composite possess certain advantages, such as the superior electronic conductivity of MXene together with excellent electrocatalytic activity of CoP, it is appropriate material that can be adopted as electrocatalyst in Li-O₂ battery. In summary, this study provides a prospective strategy for designing MXene-based nanocomposites applied in cathode catalysts of Li-O₂ battery.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

This work was supported by the National Science Foundations of China (Nos. 21871028, 21771024) and China Postdoctoral Science Foundation (No. 2020M680430).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.01.045.