With the ongoing development of electric vehicles (EVs), laptops, mobile phones, and other commercial electronics [1–3], the demand for energy storage devices has increased considerably. Moreover, lithium-ion batteries (LIBs) are widely used in modern industrial production because of their high specific energy [4] and long lifespan [5]. However, although LIBs are well-proven in myriad electronic applications, extremely cold working environments can limit their performance considerably [6,7] —such as EVs operating at high latitudes where winter temperatures can fall below −40°C, or in military and aerospace exploration which require LIBs to operate stably under wide temperature conditions (−40°C to 60°C) [7]. Consequently, cryogenic battery research has become a popular topic in the field of electrochemical research to improve the operating temperature range of batteries.

Over the past few decades, LiFePO₄ has been the main commercial cathode material [8–10], used for LIBs, as it offers good performance [11], low cost and safety [12]. However, the performance of LIBs diminishes considerably at cryogenic temperatures [13], which presents a major challenge to their application in various fields. For example, in cold weather conditions during winter, the reduced efficiency of LIBs can hinder their use in EVs [14], portable electronics [15], and energy storage systems [16]. This temperature sensitivity is particularly problematic for critical applications, such as military operations and aerospace exploration [17], where reliable battery performance in extreme environments is essential. Consequently, overcoming the challenges posed by low temperatures is crucial for expanding their practical application in these demanding fields. Owing to their high operating voltage [18] and discharge capacity [19], and particularly their excellent low-temperature cycling performance [20,21], molybdenum (Mo)-doped cathode materials have attracted considerable attention [2,22,23].

In this study, we investigated the optimal molar ratio of Mo in a commercial electrolyte. Our findings revealed that a molar ratio of 0.02 for Mo provided the best performance, delivering a specific capacity of 121.3 mAh/g within an operating voltage range of 3.0–4.3 V at ambient temperature. Based on these results, we developed a novel electrolyte formulation with a volume ratio of 1,3-dioxolane:1,2-dimethoxyethane (DOL:DME) of 8:2. The new electrolyte exhibited a specific capacity of 102.3 mAh/g at room temperature. Even under cryogenic conditions, the electrolyte exhibited impressive stability, with a capacity retention of 42.3% at 223 K, corresponding to a specific capacity of 43.3 mAh/g. Notably, the electrolyte retained 0.11% of its capacity even at 193 K. These results demonstrated that the novel electrolyte not only performed well under standard conditions, but also exhibited remarkable resilience under extremely low-temperature conditions, making it a promising candidate for use in applications requiring robust performance across a wide temperature range.

The XRD pattern of the LVMP is shown in Fig. 1. Compared to the standard cards (JCPDF: 78-5405 and JCPDF: 81-2414), the samples exhibited a monoclinic structure without impurity peaks. However, when Mo was doped at 0.02, no Mo peak was evident owing to the low Mo concentration. The broadening of the peak width indicates an increase in crystal grain size, caused by the incorporation of Mo into the Li₃V₂(PO₄)₃ lattice, leading to changes in the lattice parameters. When the Mo content increased to 0.1, distinctively characteristic Mo peaks were evident, verifying a pure phase with homogeneous V doping in the lattice without any detectable phase separation of MoO₃ [24]. As shown in the inset of Fig. 1, the XRD peaks of the doped samples systematically shift toward higher 2θ positions with increasing Mo doping concentration. This trend indicates a reduction in unit cell volume, suggesting that the larger V³⁺ ions (0.064 nm) have been successfully replaced by the smaller Mo⁶⁺ ions (0.059 nm). And the cell parameters have been listed in Table 1.

The scanning electron microscopy (SEM) image shown in Fig. 2 shows that the particles of LiV_2–xMo_x(PO₄)₃ were relatively evenly distributed. Although the particles were irregular in shape, their sizes were mostly between 0.5 and 1 μm. Almost no large secondary particles were formed through agglomeration. This uniform distribution facilitated better penetration of the electrolyte into the electrode, shortening the lithium-ion diffusion path, thereby enhancing the electrochemical performance of the material.

Fig. 3 presents high-resolution TEM (HRTEM) images of the Li₃V_1.9Mo_0.1(PO₄)₃ batteries, demonstrating that the LVMP sample retained the same layered structure as the LVP sample. The metallic Mo particles on the surface of the LVMP sample were uniformly distributed. The TEM image also revealed a lattice fringe of 0.45 nm, corresponding to the (200) plane of MVO, as noted in reference [24].

The XPS spectra of the Li₃V_2-xMo_x(PO₄)₃ cathode materials are shown in Fig. 4(a,b). After Mo doping, the Li^⁺ valence band shifted from 55.53 to 55.43 eV, indicating that the incorporation of Mo facilitated the insertion and migration of Li^⁺ ions during the charge and discharge processes in the electrode material. This enhancement leads to an improvement in the battery capacity at low temperatures because Mo doping promotes better Li⁺ mobility, thereby increasing the efficiency of the electrode material.

The XPS spectra of the Li₃V₂(PO₄)₃ sample, as shown in Fig. 4(c), indicate a V2p binding energy of 517.18 eV, which closely corresponds to the 517.2 eV [25] observed in the Li₃V₂(PO₄)₃ sample. This suggests that vanadium in the amorphous Li₃V₂(PO₄)₃ precursor existed in the V³⁺ oxidation state. Compared with the Li₃V_1.98Mo_0.02(PO₄)₃ sample, after Mo doping the V2p shifted to 517.16 eV suggesting that the Mo could change the electrode structure and enhance the migration of Li⁺ ions. The band energies and atomic ratios are listed in Table 2 and Table 3.

Fig. 5 shows the CV curves of the commercial STD electrolyte and the novel DDL electrolyte. As shown in Fig. 4, as the voltage increased, the current gradually increased, indicating the onset of decomposition in the STD electrolyte. When the voltage reached approximately 4.25 V, the electrolyte began to decompose, the rate of decomposition increasing with increasing voltage. However, the novel DDL electrolyte used in this experiment maintained a stable current across the entire voltage range of 1–7 V, indicating its superior stability. Moreover, the incorporation of DOL lowered the freezing point of the electrolyte and enhanced its conductivity at low temperatures.

Fig. 6(a) shows the electrochemical cycling curves for different Mo-doping levels in the DDL electrolyte. As shown in the figure, at room temperature an appropriate amount of Mo doping could increase the battery capacity; however, when the Mo-doping level was too high, the battery capacity decreased, leading to reduced performance. This is primarily because Mo, as a transition metal, has good electrical conductivity. Consequently, Mo doping can enhance the electronic conductivity of the electrode materials, allowing faster charge transfer during the charging and discharging processes, thereby improving the rate and overall electrochemical performance of the battery. Moreover, Mo doping can help the formation of a more stable solid electrolyte interface (SEI) layer, reducing side reactions and minimizing electrolyte decomposition. A stable SEI layer contributes to enhanced cycling stability and safety of the battery.

However, excessive Mo doping can introduce too many impurities or defects into the material, obstructing the electron conduction pathways, and thus reducing its conductivity. This can impair the electrochemical performance of the battery, causing lattice distortion or structural instability in the active material, affecting the lithium-ion intercalation and deintercalation. Ultimately, this results in a decrease in the specific capacity and negatively affects the energy density of the battery. In this study, the battery achieved a maximum electrochemical capacity of 121.3 mAh/g when the Mo-doping level was 0.02%. However, when the Mo-doping level was increased to 0.1%, the battery capacity dropped to 117.9 mAh/g.

Fig. 6(b) illustrates the behavior of the Li⁺ ions during the charge and discharge cycles [26]. In the anodic scan, three distinct peaks appear at 3.60, 3.70, and 4.10 V, representing a sequence of phase transitions that occur during the extraction of Li⁺ at various stages from the LiV_2–xMo_x(PO₄)₃ sample (where x = 3.0, 2.5, 2.0, and 1.0). The overall electrochemical reaction can be expressed as follows:

A cyclic voltammogram of the LiV_2–xMo_x(PO₄)₃ obtained at a scan rate of 0.1 mV/s at different temperatures is shown in Fig. 7. Three anodic and three cathodic peaks could be detected during the anodic and cathodic scans in given potential windows (3.0–4.3 V).

Fig. 8 shows three anodic and three cathodic peaks. The potential difference between the anodic and cathodic peaks (ΔV) reflects the insertion and extraction behavior of Li^⁺ during the charging and discharging process of the lithium battery. An increase in ΔV indicates that the extraction of Li^⁺ becomes more difficult during the charging process, meaning that the activation energy for Li^⁺ desorption increases. This phenomenon leads to a reduction in the overall capacity of the battery because Li⁺ cannot desorb smoothly, thereby reducing the effective discharge capacity.

From Fig. 7 and Table 4, for the LVM_0.02P sample, ΔV gradually increased as the temperature decreased. For example, when the temperature decreased from 298 to 243 K, ΔV increased from 0.1 to 0.2 V. This indicates that at lower temperatures, the migration and extraction of Li^⁺ became more difficult, possibly owing to dynamic changes at the interface between the electrode material and the porous medium, leading to an increase in activation energy.

Using the Randles–Sevcik equation, the diffusion coefficient of Li^⁺ (DappLi+) was calculated, the detailed results of which are listed in Table 5. The results showed that at 298 K, the diffusion coefficient of Li⁺ in the LVM_0.02P sample was 3.08×10^–9 cm²/s, whereas in the LVM_0.1P sample, it was 7.44×10^–9 cm²/s. This indicates that the insertion and extraction of Li⁺ were easier in the LVM_0.1P sample at room temperature. However, for the LVM_0.1P sample, as the temperature decreased, ΔV increased from 0.12 V at 298 K to 0.40 V at 253 K, further proving the effect of temperature on the migration of lithium ions and the battery performance.

· i_p = Current maximum in amps (A)

· n = Number of electrons transferred in the redox event (usually 1)

· A = Electrode area in cm²

· F = Faraday Constant in C/mol

· D = Diffusion coefficient in cm²/s

· C = Concentration in mol/cm³

· ν = Scan rate in V/s

· R = Gas constant in J/K/mol

· T = temperature in K

· Constant with a value of 2.69×10⁵ has units of C/mol/V^1/2

Fig. 8 shows a comparison of CV curves for the batteries using the commercial STD electrolyte and novel DDL electrolyte with LVM_0.02P as the anode under ultralow temperature conditions. Fig. 8(a) presents the CV curves of the STD electrolyte at different operating temperatures, whereas Fig. 8(b) shows the CV curves of the novel DDL electrolyte at various temperatures. The data indicate that as the temperature decreased, the performance of all batteries declined. However, the rate of performance degradation of the STD electrolyte was considerably higher than that of the novel DDL electrolyte. At 233 K, the capacity of the battery with the STD electrolyte dropped from 115.0 mAh/g at room temperature (298 K) to 9.0 mAh/g, which was only 7.8% of its initial capacity. Further temperature reductions led to the battery becoming nonfunctional, with the capacity completely disappearing. By contrast, for the novel DDL electrolyte, the battery capacity at 233 K was 61.6 mAh/g, approximately 60.22% of the initial capacity. Even when the temperature was lowered further to 193 K, the battery was still able to charge and discharge, with a capacity of 0.1 mAh/g, about 0.11% of the initial capacity. Specific values are listed in Table 6 and 7, demonstrating that the novel electrolyte exhibited excellent performance under low-temperature conditions, making it suitable for low-temperature LIB applications.

In Fig. 8, the discharge curves of the STD and DDL electrolytes exhibit distinct differences, primarily due to variations in electrolyte composition, ion transport properties, and electrochemical stability. The STD electrolyte, with its conventional formulation, demonstrates limited ion transport at low temperatures, as evidenced by a significant capacity drop at 233 K. In contrast, the DDL electrolyte, tailored for low-temperature applications, likely incorporates additives or solvents that enhance lithium-ion mobility and reduce electrolyte viscosity, leading to improved capacity retention. The discharge profile of the STD electrolyte shows a pronounced decline in capacity and voltage with decreasing temperature, indicative of sluggish ion diffusion and increased internal resistance. Conversely, the DDL electrolyte maintains stable and well-defined voltage plateaus even at 213 K and 193 K, suggesting enhanced electrolyte-electrode compatibility and sustained redox activity. While both electrolytes exhibit comparable performance at 298 K, the STD electrolyte experiences a sharp capacity decline as the temperature decreases. The DDL electrolyte, however, retains capacity more effectively at extremely low temperatures, mitigating key low-temperature challenges such as lithium plating, sluggish kinetics, and electrolyte freezing.

In LIBs, the diffusion coefficient of lithium ions is a critical factor affecting the battery performance. A higher diffusion coefficient generally indicates that ions can move more quickly within the electrode material, thereby enhancing the charge and discharge rates of the battery. By conducting CV experiments at different temperatures, we could calculate the apparent diffusion coefficient of the lithium ions. Additionally, the activation energies for lithium-ion diffusion in different electrolytes could be determined at various temperatures. As shown in Fig. 9, the migration activation energy for lithium ions in the STD electrolyte was 14.89 kJ/mol, whereas it was 9.58 kJ/mol in the DDL electrolyte. This indicated that lithium-ion diffusion was more favorable in the DDL electrolyte, enabling the battery to function properly at low temperatures while maintaining a higher capacity. Consequently, DDL electrolytes are suitable for low-temperature battery applications.

This study investigated the effect of molybdenum (Mo)-doping on the electronic structure of an electrode, thereby improving the electrode conductivity and increasing the battery capacity. This allowed the LIBs to maintain good charge and discharge capacities even at low temperatures, thereby broadening their application range. When the Mo-doping level was 0.02%, the battery exhibited its best specific capacity of 121.3 mAh/g. Next, we introduced a novel DDL electrolyte. Using this electrolyte, the battery capacity at 223 K remained at 61.6 mAh/g, approximately 60.22% of the initial capacity, and at 193 K, the battery capacity was 0.1 mAh/g, approximately 0.11% of the initial capacity. By calculating the activation energies for lithium-ion diffusion in different electrolytes, we found that the lithium ions in the DDL electrolyte exhibited a lower diffusion activation energy of 9.58 kJ/mol.