HCl (~37 wt%), HF (~40 wt%), hydrobromic acid (HBr, ~40 wt%), hydroiodic acid (HI 48-57 wt%), isopropyl alcohol (≥ 99.7%) and ethanol (99.5%) were obtained from Shanghai Hushi Laboratory Supplies Co., Ltd. Gallium metal (99.999%), Nafion perfluorinated resin solution was purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Molybdenum carbide (99.5%) was provided by Alfa Aesar (China) Chemical Co., Ltd. These reagents were used as received without further purification. Deionized water (DIW, resistivity > 18 MΩ·cm) was collected from a Milli-Q Biocel system.

Mo₂Ga₂C powder was prepared by mixing and sintering commercial Mo₂C powder with Ga, following the procedure reported in previous literature.

To prepare Mo₂CF_x MXene, 200 mg of Mo₂Ga₂C was slowly added to a 50 mL Teflon-lined autoclave. Subsequently, 20 mL of HF was gradually introduced into the autoclave and stirred for 5 min. The mixture was then left at room temperature for 5 days to facilitate the etching reaction. After reaction, the residual powder was washed with DIW and anhydrous ethanol until reaching a neutral pH. The powder was then freeze-dried for 24 h to obtain Mo₂CF_x.

To prepare Mo₂CT_x MXene, concentrated hydrohalic acid (HX) were used to purify Mo₂CF_x sample. In a typical procedure, 200 mg Mo₂CF_x was slowly added to a 50 mL Teflon-lined autoclave. Subsequently, 20 mL of concentrated HX (X = Cl, Br, I) acid was gradually introduced into the autoclave and stirred for 5 min. The mixture was then left at 180 ℃ for 1 day to facilitate the purification. After purification, the residual powder was washed with DIW and anhydrous ethanol until reaching a neutral pH. The powder was then freeze-dried for 24 h to obtain Mo₂CT_x.

Firstly, the structural models of partial Mo₂CT_x (T = F, Cl, Br, I) MXenes along the c-axis are presented in Figure 1A, with the yellow regions representing surface termination and the blue regions indicating the solvation areas. The enthalpy of formation for halogen-terminated Mo₂CT_x MXene was calculated [Figure 1B]. The enthalpies of formation for Mo₂CF_x, Mo₂CCl_x, Mo₂CBr_x, and Mo₂CI_x are -1.57, -0.93, -0.77, and -0.49 eV/atom, respectively. The bonding strength between terminations is F > Cl > Br > I, similar to the bonding strength of Ti₃C₂T_x MXene surface terminations, suggesting that this order of bonding strength may be universally present in MXenes^[8]. Figure 1C-F illustrates the density of states for Mo₂CT_x, indicating that all four types of Mo₂CT_x MXene are metallic in nature^[23,24]. Mo₂CT_x MXene displays exceptional electrical conductivity, especially in the case of Mo₂CCl_x, Mo₂CBr_x, and Mo₂CI_x. The majority of electrons near the Fermi surface in Mo₂CT_x MXene are supplied by the Mo d-orbital electrons, followed by surface terminations^[23,25]. For Mo₂CF_x, the Mo and C elements almost do not provide electrons, which might be the reason why the F terminations do not participate in redox reactions and thus display inert behavior^[26,27]. However, for other Mo₂CT_x MXenes, the surface terminations provide electrons near the Fermi surface and thus generate additional redox peaks in the redox process, thereby increasing the capacity of the supercapacitor^[17]. Hence, they are potential candidates for supercapacitors. Figure 1G shows the integral specific capacitance in the voltage window of -0.6 to 0.6 V vs. point of zero charge (PZC). The integral capacitances of Mo₂C with halogen terminations are 9.23, 10.10, 11.30, and 12.96 μF/cm², respectively. The order of the integral capacitance size is Mo₂CF_x < Mo₂CCl_x < Mo₂CBr_x < Mo₂CI_x.

Figure 1H shows the quantum capacitance (C_Q) of Mo₂CT_x (T = F, Cl, Br, I) MXene. Since Mo₂CF_x has the fewest electrons at the Fermi level, its C_Q at the Fermi level is also the smallest. The C_Q behavior of Mo₂CT_x MXene is completely different from that of graphene, which exhibits a “V” shape due to the Dirac point^[28,29]. The C_Q of monolayer graphene is about 20 μF/cm² at 0.6 eV, while the minimum C_Q of Mo₂CT_x is 200 μF/cm², far greater than that of graphene^[28]. The quantum charge curves in Figure 2A show that the quantum charging capacity of Mo₂CI_x is the smallest, indicating that the regulation of terminations can indeed effectively control the C_Q behavior of MXene.

To reveal the impact of terminations changes on EDL capacitance, the electronic structure of the EDL at the electrode/electrolyte interface was calculated. Figure 2B shows the net charge distribution of the electrode and the electrolyte along the z-direction at an applied potential of 1 V vs. PZC. Notably, most of the net charge exists at the electrode/electrolyte interface. As halogens belong to the same group, the distribution shape of the excess charge in both the electrode and the electrolyte is almost the same, indicating that the charge distribution shape is independent of the surface terminations. As the atomic radius increases, the positive center of the electrode and the negative center of the electrolyte gradually deviate from the surface. The thickness of the double layer can be obtained by calculating the difference between the two centers. Figure 2B shows that as the electronegativity decreases, the thickness of the double layer decreases. According to the definition of a parallel plate capacitor, capacitance is inversely proportional to thickness, so its capacitance increases, consistent with the EDL charge curves corresponding to Mo₂CT_x (T = F, Cl, Br, I) MXene calculated in Figure 2C. In order to better study the charge distribution on the electrode surface, differential electron densities between charged (1 V vs. PZC) and uncharged Mo₂CT_x MXene were calculated. Interestingly, most of the charge is distributed on the surface of terminations that directly contact the electrolyte, implying the termination’s significant impact on the capacitance of MXene electrodes.

The total capacitance C_Tot can be obtained by connecting the C_Q and the double-layer capacitance in series. Since C_EDL is much smaller than C_Q, the total capacitance is mainly determined by C_EDL. This is similar to metal electrodes^[30]. From Figure 2D, it is found that whether at positive voltage or negative voltage, the total charge curve of Mo₂CF_x is the smallest, and Mo₂CI_x is the largest. The derivative of Figure 2D gives the differential capacitance of Mo₂CT_x MXene. Supplementary Figure 1 shows that when a positive voltage is applied, the differential capacitance of Mo₂CT_x is almost a constant, about 6-9.5 μF/cm². However, in the negative voltage region, the differential capacitance of Mo₂CI_x can be as high as 22 μF/cm². The differential capacitance of graphene is roughly 3-5 μF/cm^2[28]. To probe charge redistribution at the electrode-electrolyte interface, differential electron density iso-surfaces (isovalue = 0.0005 e/ų) were calculated for charged (1 V vs. PZC) vs. uncharged Mo₂CT_x MXenes, shown in Figure 2E-H. Figure 2E (Mo₂CF_x) displays charge mainly on F terminations, indicating strong shielding from high electronegativity, restricting Mo electron delocalization. Figure 2F (Mo₂CCl_x) shows broader density on Cl, with reduced shielding. Figure 2G (Mo₂CBr_x) reveals increased localization on Br and Mo layers, boosting surface electrons. Figure 2H (Mo₂CI_x) exhibits the strongest accumulation on I and Mo, due to low electronegativity, thinning the EDL for better ion access. Overall, decreasing electronegativity from F to I enhances charge redistribution, matching the capacitance order and emphasizing terminations’ role in MXene performance.

Based on the theoretical results, the capacitance behavior of Mo₂CT_x MXene can be effectively controlled through the regulation of terminations. Given that all Mo₂CT_x MXenes are metallic, their capacitance is primarily contributed by double-layer capacitance. When Mo₂C is modified by functional groups with smaller electronegativity, its total capacitance increases, primarily due to the weakening of electrostatic shielding, leading to a reduction in the double-layer thickness. Therefore, the capacitance order is Mo₂CF_x < Mo₂CCl_x < Mo₂CBr_x < Mo₂CI_x. This lays a foundation for understanding the electrochemical behavior of surface terminations of Mo₂CT_x MXene in capacitance.

In order to further prove the theoretical claims, Mo₂CF_x samples prepared by HF etching were further purified using HCl, HBr, and HI through a solvothermal method to obtain samples with specific surface terminations. These four samples were named Mo₂CF_x, Mo₂CCl_x, Mo₂CBr_x and Mo₂CI_x. Furthermore, the specific preparation strategy is described in the Methods section. As shown in Figure 3A, after the exfoliation of Ga atomic layers, Mo₂CT_x (T = F, Cl, Br, I) exhibited similar X-ray diffraction (XRD) results. Compared to the Mo₂Ga₂C precursor, the (002) peak located at around 10° shifted to a lower angle. Besides, the intensities of the remaining diffraction peaks decreased significantly, indicating the successful preparation of MXene. Specifically, the XRD patterns show (002) peak positions at 8.52° for Mo₂CF_x, 8.76° for Mo₂CCl_x, 8.50° for Mo₂CBr_x, and 8.46° for Mo₂CI_x. This overall trend confirms progressive enlargement with decreasing halogen electronegativity and increasing atomic size from F to I, facilitating improved ion accessibility. The exception for Mo₂CCl_x (smaller d-spacing than Mo₂CF_x) can be attributed to incomplete drying of the Mo₂CF_x sample, leading to retained interlayer water that increases its spacing, while the post-processing for Mo₂CCl_x removes such water more effectively, resulting in a more compact structure. Scanning electron microscope (SEM) results reveal the typical layer structure after removed Ga layers [Supplementary Figure 2]. What is more, F, Cl, Br, I terminations are observed clearly via energy-dispersive X-ray spectroscopy (EDS) results [Supplementary Figure 3]. As depicted in Figure 3B, since Mo₂CCl_x, Mo₂CBr_x, and Mo₂CI_x were prepared by solvothermal purification based on Mo₂CF_x, the area ratios of the oxidation peaks Mo⁵⁺ and Mo⁶⁺ increased significantly compared to Mo₂CF_x, implying more pronounced surface oxidation. This can be attributed to the oxidizing properties of HCl, HBr, and HI, which promoted the oxidation process on the surface of Mo₂CF_x. During the solvothermal treatment, these acid molecules acted as oxidants, oxidizing Mo atoms to higher valence states of Mo⁶⁺ and Mo⁵⁺, leading to an increase in the area ratio of these two valence states. In the solvothermal process, the F terminations on the surface of Mo₂CF_x were replaced by -Cl, -Br, or -I. As the electronegativity decreases in the order of Cl, Br, and I, the degree of electron cloud density shifting towards the Mo atoms in the covalent bonds formed with Mo increases accordingly. Consequently, the electron cloud density around the Mo atoms increases, shielding the effective positive charge experienced by the inner electrons and reducing the bond energy of Mo-C. In the X-ray photoelectron spectroscopy (XPS) survey spectrum [Figure 3C], the F terminations on the surface of the Mo₂CF_x sample were replaced by -Cl, -Br, and -I from the HCl, HBr, and HI solutions, respectively. This can be attributed to the solvothermal conditions providing sufficient energy and reaction environment to promote the nucleophilic attack of Cl^-, Br^-, and I^- on the Mo–F bonds, forming the corresponding termination. Intriguingly, the C 1s spectra and termination-specific spectra across the four samples collectively corroborate these conclusions [Supplementary Figures 4 and 5]. While the solvothermal halogen substitution process is designed to replace F with Cl, Br, or I, the subsequent washing with DIW and ethanol may introduce minor oxygen terminations (-O or -OH), as is typical in MXene post-processing. Any trace oxygen presence could marginally enhance pseudo-capacitive contributions (as discussed for -O/-OH in the Introduction), potentially slightly elevating measured capacitances beyond pure EDL effects. Nevertheless, the experimental capacitance hierarchy (Mo₂CF_x < Mo₂CCl_x < Mo₂CBr_x < Mo₂CI_x) aligns closely with DFT predictions for idealized pure-halogen models, indicating limited influence from oxygen on the overall trends.

Furthermore, compared to Mo₂CF_x, the Mo absorption edges of Mo₂CCl_x, Mo₂CBr_x, and Mo₂CI_x shifted significantly towards higher energies, which is consistent with the XPS results, indicating an increase in the valence state of Mo [Figure 3D]. The results of the Mo K-edge extended X-ray absorption fine structure (EXAFS) [Figure 3E] showed that the coordination numbers of Mo–Mo and Mo–C at 2.5 and 1.5 Å followed the order of Mo₂CI_x > Mo₂CBr_x > Mo₂CCl_x > Mo₂CF_x (without fitting). It is evident that due to the electronegativity order of -F > -Cl > -Br > -I, stronger electronegativity leads to stronger covalency and shorter bond lengths, resulting in increased distances between adjacent Mo atoms and weakened interactions between Mo atoms and Mo as well as C. Consequently, the coordination number of Mo decreases with increasing electronegativity of the termination. In the wavelet transform (WT) results [Figure 3F and Supplementary Figure 6], two maxima were observed at around 9.9 and 4.9 Å, corresponding to the chemical bonding environments of Mo–Mo and Mo–C, respectively. In summary, through systematic characterization analysis, it was confirmed that the solvothermal method can effectively control the surface terminations composition of Mo₂CT_x MXene. More importantly, the electronegativity of different terminations significantly affects the structural and electronic properties of the material, such as Mo valence state, Mo–C bond energy, and the coordination environments of Mo–Mo and Mo–C. These results provide important experimental evidence for understanding the structure-property relationship of MXene materials.

To further explore the influence of different surface terminations on the electrochemical performance of MXene materials, Mo₂CT_x (T = F, Cl, Br, I) was used as thin-film electrodes for supercapacitors. The experiments were conducted in 1 M H₂SO₄ electrolytes. As shown in Figure 4A-D, no obvious redox peaks appeared for the four samples at different scan rates, indicating that the electrochemical behavior of the samples was mainly non-Faradaic capacitance. The quasi-rectangular CV shape indicates fast pseudo-capacitive kinetics, where diffusion-controlled intercalation dominates without discrete peaks, as is typical in MXenes^[13,20]. The cyclic voltammetry (CV) curves of the four samples at different scan rates exhibited similar shapes, suggesting that the capacitance behavior had a weak dependence on the scan rate, implying good rate performance, electrochemical stability, and fast charge-discharge capability of the samples. The rapid charge storage and release can be achieved precisely because the charge storage primarily occurs on the electrode surface, and the charge transfer and diffusion paths are short, while the layered structure of MXene provides a structurally stable framework for charge storage. Importantly, at the same scan rate, the area enclosed by the CV curves follows the order of Mo₂CI_x > Mo₂CBr_x > Mo₂CCl_x > Mo₂CF_x, suggesting that the terminations I with lower electronegativity and larger size is beneficial for the adsorption and desorption of electrolyte ions, enhancing the EDL capacitance effect.

As shown in Figure 4E-H, the galvanostatic charge/discharge (GCD) curves of the four samples Mo₂CF_x, Mo₂CCl_x, Mo₂CBr_x, Mo₂CI_x were measured at different current densities. At the same current density, the charge-discharge time of the samples follows the order of Mo₂CI_x > Mo₂CBr_x > Mo₂CCl_x > Mo₂CF_x. Longer charge-discharge times corresponding to higher specific capacitances. This sequence can be attributed to the terminations I with lower electronegativity, which leads to a higher electron cloud density on the surface Mo atoms of MXene and a lower EDL thickness, making it more favorable for the adsorption and desorption of electrolyte ions and enhancing the EDL capacitance effect. Despite a slight decrease in the overlap of the GCD curves at high currents, the shape of the GCD curves for all four samples remained highly consistent across multiple currents, demonstrating the excellent stability of the materials. No significant decay in charge-discharge time or severe deformation of the curve shape was observed during the cycling process, suggesting that the electrode materials can maintain their structural integrity and electrochemical activity during the charge-discharge process. This stability can be attributed to the unique layered structure of MXene and the regulation of the structural and electronic properties of the materials by surface terminations modification. The specific capacitances and capacitance retention rates of Mo₂CT_x (T = F, Cl, Br, I) samples at different scan rates are summarized in Figure 4I and J. As seen, the Mo₂CI_x sample achieves the highest specific capacitance of 76.3 F/g at a scan rate of 8 mV/s. Overall, the specific capacitances of the samples at various scan rates follow the order of Mo₂CI_x > Mo₂CBr_x > Mo₂CCl_x > Mo₂CF_x. Moreover, as the scan rate increases, Mo₂CI_x consistently maintains the highest capacitance retention rate. The capacitive contribution analysis reveals a self-adaptive dual-mode mechanism in Mo₂CT_x MXene [Figure 4K and Supplementary Figure 7]. At low scan rates, diffusion-controlled behavior dominates, analogous to battery-type ion intercalation, while high-rate operation shifts overwhelmingly to capacitance dominated kinetics. This transition originates from MXene’s distinctive 2D layer architecture, which provides exceptionally high specific surface area and abundant active sites for rapid surface redox reactions and modulated ion-transport pathways through interlayer galleries, further enhanced by surface functional group participation in Faradaic processes. Figure 4L illustrates the specific capacitances of the four samples at different current densities. Remarkably, Mo₂CI_x attains an impressive specific capacitance of 74.9 mA·h/g (539.3 F/g) at a current density of 1 A/g, which is 1.35 times that of Mo₂CBr_x, 1.66 times that of Mo₂CCl_x, and 2.16 times that of Mo₂CF_x at the same current density. Similarly, as the current density continues to increase, the specific capacitance gradually decreases.

Through the above systematic electrochemical performance measurements, we confirmed that the Mo₂CT_x (T = F, Cl, Br, I) samples have a charge storage mechanism dominated by non-Faradaic capacitance. The electronegativity and size of different terminations have significant effects on the electrochemical performance. Combining the results of structural analysis, the I termination with lower electronegativity and larger size can result in increasing the electron cloud density of the surface Mo atoms of MXene, reducing the EDL thickness, and increasing the specific surface area and active sites of the materials. These results reveal the intrinsic correlation between the structure and performance of MXene materials and are consistent with our theoretical calculations, may provide useful guidance for the design and optimization of high-performance MXene electrode materials.