Boosting High‐Voltage Dynamics Towards High‐Energy‐Density Lithium‐Ion Capacitors

Lithium‐ion capacitors (LICs) are becoming important electrochemical energy storage systems due to their great potential to bridge the gap between supercapacitors and lithium‐ion batteries. However, capacity lopsidedness and low output voltage greatly hinder the realization of high‐energy‐density LICs. Herein, a strategy of balancing capacity towards fastest dynamics is proposed to enable high‐voltage LICs. Through electrochemical prelithiation of Nb2C to be 1.1 V with 165 mAh g−1, Nb2C // LiFePO4 LICs show a broadened potential window from 3.0 to 4.2 V and an according high energy density of 420 Wh kg−1. Moreover, the underlying mechanism between prelithiation and high voltage is disclosed by electrochemical dynamic analysis. Prelithiation declines the Nb2C anode potential that facilitates electron transmission in the interlayer of two‐dimensional Nb2C MXene. This effect induces small drive force for Li+ ions deposition and hence weakens the repulsive force from adsorbed ions on the electrode surface. Benefiting from even more Li+ ions deposition, a higher voltage is eventually delivered. In addition, prelithiation significantly increases Coulomb efficiency of the 1st cycle from 74% to 90%, which is crucial to commercial application of LICs.

Lithium-ion capacitors (LICs) are becoming important electrochemical energy storage systems due to their great potential to bridge the gap between supercapacitors and lithium-ion batteries. However, capacity lopsidedness and low output voltage greatly hinder the realization of high-energy-density LICs. Herein, a strategy of balancing capacity towards fastest dynamics is proposed to enable high-voltage LICs. Through electrochemical prelithiation of Nb 2 C to be 1.1 V with 165 mAh g À1 , Nb 2 C // LiFePO 4 LICs show a broadened potential window from 3.0 to 4.2 V and an according high energy density of 420 Wh kg À1 . Moreover, the underlying mechanism between prelithiation and high voltage is disclosed by electrochemical dynamic analysis. Prelithiation declines the Nb 2 C anode potential that facilitates electron transmission in the interlayer of two-dimensional Nb 2 C MXene. This effect induces small drive force for Li + ions deposition and hence weakens the repulsive force from adsorbed ions on the electrode surface. Benefiting from even more Li + ions deposition, a higher voltage is eventually delivered. In addition, prelithiation significantly increases Coulomb efficiency of the 1st cycle from 74% to 90%, which is crucial to commercial application of LICs.
prelithiation from the perspective of the Li source/anode interfaces by regulating the initial contact state and delivered a clear illustration of the pathogeny for capacity attenuation. Specifically, creating plentiful electron channels is an access to making contact prelithiation with a higher Li utilization, as the mitigated local current density that reduces the etching of Li dissolution and SEI extension on electron channels. [26] Wu et al. studied the mechanism of controllable preilithiaiton by monitoring the electrical conductivity change in the lithiation solution in the duration of its formation. They demonstrated the essential role of lithium radical anions for chemical prelithiation and compared the prelithiation activity of dissociated species and aggregates of lithium radical anions. [40] Unfortunately, it has to be admitted that prelithiated performance (such as increasing the output voltage) is strongly considered, while little attention is paid to the mechanism between prelithiation and high-voltage dynamics.

Results and Discussion
In this work, aiming at solving the main issue of capacity lopsidedness and slow dynamics in LICs (Figure 1a), we optimize the initial capacity of the Nb 2 C anode and further preintercalate Li + ions to the desired voltage with optimal dynamics. Through prelithiation, more Li atoms located at Nb 2 C nanolayers cause a lower potential of Nb 2 C anode near the current collector as shown in the left of Figure 1c. [36] This effect will form a larger potential difference between the surface and the bottom of the Nb 2 C anode, which accelerates the transmission of electrons in the vertical direction of the Nb 2 C anode (the middle panel of Figure 1c). To compare electron transport dynamics between the unprelithiated and prelithiated electrodes, we conducted the block resistance test. As shown in Figure S1, Supporting Information, a higher electrical conductivity in the prelithiated Nb 2 C anode (1.72 AE 0.12 mΩ) is observed than in unprelithiated Nb 2 C anode (3.66 AE 0.24 mΩ). Meanwhile, more electrons will accumulate at the active site of Li + ions deposition, reducing the resistance of Li + ions deposition. Then, the Li + ions adsorbed at the surface defects of Nb 2 C anode tend to deposit inward (the right panel of Figure 1c). Therefore, upcoming intercalated Li + ions are no longer subject to electrostatic repulsion, and more Li + ions can deposit into the anode, which eventually endows LICs with high voltage and high energy density. In contrast, the Nb 2 C anode in unprelithiated LICs are in a low kinetic state (the left panel of Figure 1c). Also, the upcoming intercalated Li + ions suffer from electrostatic repulsion causing low voltage and energy density (the middle and the right panels of Figure 1c). Consequently, the prelithiated Nb 2 C // LFP LICs exhibit a remarkably enlarged potential window from 3.0 to 4.2 V (Figure 1d). Also, a high energy density of 420 Wh kg À1 (based on the weight of a single electrode) at 0.3 C and a high 1st cycle Coulomb efficiency (g) of 90% have been achieved for prelithiated LICs, while they are only 142 Wh kg À1 and 74% for unprelithiated LICs with a potential window of 3.0 V. Note that unprelithiated LICs show a long-time charging plateau at~3.5 V (Figure 1d), meaning excessive polarization induced electrochemical instability. In addition, the resistance shows an apparent decline from 21.9 Ω in unprelithiated LICs to 8.9 Ω in prelithiated LICs ( Figure S2, Supporting Information).
To achieve high voltage of LICs, it is necessary to eliminate the capacity difference between the anode and the cathode. [41] In Nb 2 C // LFP LICs, the initial capacity of the Nb 2 C anode is 270 mAh g À1 , while the LFP cathode shows a much lower specific capacity of 165 mAh g À1 (Figure 2a). On this occasion, the specific capacity of the Nb 2 C anode should be controlled at about 165 mAh g À1 , which correlates with a potential of 1.1 V according to the discharge curve. In theory, Nb 2 C undergoes two main energy storage mechanisms during the discharge process from 2.5 to 0.01 V (Figure 2b): 1) Li + ions adsorption/desorption processes on the defect sites of the MXene layer; and 2) Li + ions intercalation/deintercalation processes in the MXene layers. [42] The adsorption/desorption processes are unsuitable to prelithiation because of low capacity and large voltage variation. On the contrary, intercalation/deintercalation processes with over 90% capacity at low potential are ideal to match with the capacity of LFP. [43,44] In detail, the discharge profiles of Nb 2 C can be divided roughly into five regions as discussed by different electrochemical kinetics under different potentials. In Region I with a high voltage of 2.3 V, the Nb 2 C stores charges through an adsorption/desorption mechanism. With the decrease of voltage, region II exhibits an ion intercalation/deintercalation mechanism, while the reaction kinetics is poor since the interlayer spacing of MXene is relatively narrow. In Region IV and V, the Nb 2 C also displays sluggish kinetic behavior and even poor stability, which are due to the Li + ions occupying most of the active sites of MXene layers, causing limited space and large strain. Moreover, the excessively low reaction voltage will possibly induce lithium dendrite growth and thermal runaway in LICs. On the contrary, the Nb 2 C in Region III delivers desirable electrochemical kinetic behavior due to wide interlayer spacing, abundant available active sites, and low strain. [37] Thus, Region III including 1.1 V is the optimal potential to pair with LFP.
To further reveal the electrochemical kinetic behavior under different voltages, we carried out the electrochemical impedance spectra (EIS), galvanostatic intermittent titration technique (GITT), and relaxation time analysis ( Figure S3, Supporting Information for calculation method). The EIS plots of the Nb 2 C anode in four different voltages of prelithiation are shown in Figure 2c. As expected, the Nb 2 C prelithiated at 1.1 V exhibits the smallest charge-transfer resistance (R ct = 8.8 Ω), indicating its rapid charge-transfer and excellent dynamic behavior. However, the Nb 2 C prelithiated at 0.01, 0.7, and 1.8 V show high charge-transfer resistances of 16.5, 11.2, and 29.4 Ω, respectively. Another evidence to prove the rapid dynamics can be implied by the relaxation time constant (s 0 ). [45,46] According to the equation of s 0 = 1/(2pf), the s 0 values are calculated to be 0.62, 0.28, 0.19, and 1.93 ms for 0.01, 0.7, 1.1 and 1.8 V prelithiated Nb 2 C, respectively ( Figure S4, Supporting Information). The lower s 0 value of 1.1 V prelithiated Nb 2 C reveals its faster dynamic behavior, proving the positive effect of prelithiation. Moreover, the diffusion coefficient calculated by GITT of 1.1 V prelithiated Nb 2 C is the highest. In short, Nb 2 C prelithiated to 1.1 V possesses the minimum charge transfer resistance and the fastest dynamic behavior.
In this regard, prelithiation of the Nb 2 C anode while maintaining matched capacity of 165 mAh g À1 with LFP is significant to achieve the optimal performance of LICs. Therefore, we optimize the structure of Nb 2 C by adjusting the degree of etching, which guarantees optimal data of 1.1 V and~165 mAh g À1 as shown in Figure 2e.
The four different processes were used to prepare Nb 2 C MXenes. As illustrated in Figure  3a, Nb 2 AlC were partially etched by the bath oiling method at 65°C-5 days and 70°C-5 days, while it was severely etched by the hydrothermal method at 180°C-36 h. The optimal etching process is the hydrothermal method at 180°C-30 h. The representative scanning electron microscopy (SEM) images of different etching progresses are displayed in Figure 3b-e. With the increase of the etching degree, the MXene nanolayers show a process from compact unetched completely structure (Figure 3b,c) to complete and obvious 2D layered structure (Figure 3d), and then the layered structure is slightly destroyed (Figure 3e).
To further explore the evolution of the structure, we performed X-ray diffraction (XRD) tests. As showing in Figure 3f, Nb 2 AlC have been etched successfully with four etching measures due to the appearance of Nb 2 C (002), (100), (110) diffraction peaks around at 7.5°, 33.5°, and 59.2°. As consistent with SEM results, there are no additional impurities in Nb 2 C prepared by the hydrothermal method while some additional diffraction peaks of 12.8°, 38.7°, and 59.9°assigned to Nb 2 AlC phase are still existing in Nb 2 C prepared by the bath oiling method. Meanwhile, the (002) peak locates at 7.2°of the hydrothermal method yielded Nb 2 C, which declines by 0.3°in comparison with that of the bath oiling method yielded Nb 2 C, indicating the higher interlayer spacing of the former. The active sites for Li + ions attaching are located between the layers of MXene. Therefore, the capacity of Nb 2 C will be severely affected by the interlayer structure. Nb 2 AlC was partially etched by the bath oiling method at 65°C-5 days and 70°C-5 days. Incompletely etched Nb 2 AlC has a compact structure, which makes it difficult to expose the active sites and embed Li + ions. Nb 2 AlC was severely etched by the hydrothermal method at 180°C-36 h. The layered structure of MXene was destroyed and its mechanical properties deteriorated. It is also accompanied by a decrease in storage capacity. Therefore, a more complete layer structure and appropriate layer spacing will give Nb 2 C a higher capacity. The optimal etching process is the hydrothermal method at 180°C-30 h with the most obvious 2D layered structure and the highest capacity ( Figure S5, Supporting Information) among the four etching measures.
More evidences by X-ray photoelectron spectroscopy (XPS), X-ray energy-dispersive spectroscopy (EDS), Raman, and Transmission electron microscopy (TEM) characterizations support the complete etch of Nb 2 AlC by the hydrothermal method at 180°C-30 h. Deconvoluted XPS spectra of Nb 3d in Nb 2 CT z MXene (Figure 3g) displays three And obvious characteristic peaks of the Raman spectra locating at 120, 231, and 681 cm À1 are all assigned to Nb 2 C MXene ( Figure S7, Supporting Information). [48] TEM studies reveal that the HR-30 h possesses a 2D nanolayer structure ( Figure S8a, Supporting Information). Meanwhile, HRTEM characterization confirms that HR-30 h sample possesses a high degree of crystallization, and the lattice fringe with a distance of 0.825 nm belongs to (006) plane of Nb 2 C ( Figure S8b, Supporting Information).
After eliminating the capacity imbalance, we have conducted GITT, cyclic voltammetry (CV), and EIS tests to demonstrate the effect of prelithiation on broadening voltage and improving dynamic behavior of LICs. Figure 4a shows the GITT diagram of the unprelithiated LICs. The voltage changes greatly after each relaxation, indicating the unprelithiated LICs have large IR drop and high diffusion overpotential. In the charging stage, the voltage first normally increases to 2.2 V, then slowly increases to 2.57 V, and finally stagnates. Voltage stagnation after charging is an abnormal phenomenon. A relaxation process after charging greatly eliminates the electrochemical and concentration polarization in the device, [49] which will cause the Li + ions to be in a low energy state. After charging to a higher voltage, a lot of Li + ions have been deposited in the Nb 2 C interlayer, where the remained Li + ions are in a low kinetic state and possess small deposition driving force. If more Li + ions are inserted further, they are more likely to be adsorbed at the surface defect sites because of lower free energy, which will create an extra electrostatic repulsive force of Li + ions near the surface of the anode. [50] Due to the lack of electrochemical and concentration polarization, Li + ions near the anode surface are difficult to break through the electrostatic repulsion, causing very few penetrated Li + ions and accompanying static voltage.
However, the LICs after prelithiated to 1.1 V can normally charge to the voltage of 4.2 V (Figure 4b). This is due to the lowered potential of the anode after prelithiation that results in a larger voltage difference between the top and the bottom of the anode. And the electric driven force accelerates electron transmission between the nanolayers of the Nb 2 C anode, which provides an additional driving force for subsequent Li + ions deposition. There will be less Li + ions adsorbed at the anode surface that benefits to suppress the electrostatic repulsion and hence increases the kinetic behavior. Eventually, more deposited Li + ions give rise to broadened voltage.
The comparison of diffusion coefficient between prelithiated and unprelithiated LICs is shown in Figure 4c. The diffusion coefficient of unprelithiated LICs decreases sharply at around 120 h, which is consistent with the time as the voltage starting to stagnate. At the later stage of charging, the diffusion coefficient of prelithiated LICs is much higher than that of unprelithiated LICs, proving that prelithiation can improve electrochemical kinetic behavior. Besides, sweep voltammetry sheds light on the charge storage kinetics by preliathiation. To point it out, b = 0.5 determined by fitting the CV redox peaks suggests a diffusion-controlled charge storage process, and b = 1 suggests a nondiffusion-controlled process (so-called capacitive behavior). [51] The closer b is to 1, the better its dynamic behavior will be. [52] In Figure 4d, the log(i) versus log(v) plot for the prelithiation and unprelithiation shows the current dependence on the sweep rate. The b value for the prelithiated and unprelithiated peak current is calculated to be 0.7 and 0.64, respectively. From the perspective of b value, the charge stored in LICs is contributed by both diffusion-limited and non-diffusion-limited process, while the prelithiated LICs possess better dynamic behavior than unprelithiated LICs. To quantitatively distinguish capacitive contribution (non-diffusion-controlled) ratios at prelithiated and unprelithiated LICs, different scan rates for measuring CV curves have been conducted. For example, the capacitive contribution of unprelithiated LICs at 5 mV s À1 is calculated to be 51.3%, which is lower than that of prelithiated LICs with a value of 61.8%.
The EIS and relaxation time analysis have been conducted to further study the promoted electrochemical dynamic behavior by prelithiation. Since the charge and discharge voltage ranges of prelithiated and unprelithiated LICs are different, the same stage of charge (SOC) is adopted to avoid this difference. As shown in Figure S10, Supporting Information, the R ct of unprelithiated LICs increases sharply from 58.8 to 207.6 Ω with increasing the SOC, indicating the sluggish dynamic behavior. However, the R ct of prelithiated LICs is stable at 20.3 Ω, even at a high SOC of 60-80% (Figure 4g,h). Meanwhile, the relaxation times of unprelithiated LICs in 80% are 192.9 and 41.6 ms in 70%, which explains the static voltage after charging to some extent. In contrast, the relaxation time of prelithiated LICs in 80% is as short as 0.62 ms, correlating with rapid dynamic behavior.
After the realization of balancing capacity toward fastest dynamics, we further evaluate the energy density and practical application of prelithiated Nb 2 C // LFP LICs. The prelithiated LICs show much better rate performance as indicated by a specific capacity of 131 mAh g À1 at 0.3 C and 109, 83 mAh g À1 , and 60 mAh g À1 at 0.5, 1, and 2 C, respectively (Figure 5a). However, the unprelithiated LICs only deliver a specific capacity of 80 mAh g À1 at 0.3 C, and then rapidly decrease to 19 mAh g À1 at 2 C. Besides, the initial charge-discharge curve is shown in Figure 5b, where both prelithiated and unprelithiated devices show capacitive dominant curves without an apparent charging and discharging platform. And the 1st cycle Coulomb efficiency largely increases from 74% to 90% with the prelithiation process.
Since the voltage becomes much higher, the energy density is also increased significantly after prelithiation. The 1.1 V prelithiated LICs deliver the maximum high energy density of 420 Wh kg À1 (based on the weight of a single electrode) at 0.3 C, which is about three times higher than that of unprelithiated LICs (142 Wh kg À1 at 0.3 C). As a visualized demo, a "SWJTU" LED panel containing 318 green bulbs can be successfully lightened by 1.1 V prelithiated LICs for more than 2 min without obvious loss of brightness ( Figure 5d and Video S1, Supporting Information). However, the unprelithiated LICs fail to work (Figure S11, Supporting Information). Meanwhile, a smart watch (Figure 5e) and LCD timer (Figure 5f) can also be easily powered by the prelithiated LICs. These results unambiguously demonstrate that the prelithiated Nb 2 C // LFP LICs can be used as a promising candidate for high-voltage and high-energy-density portable electronics and wearable devices.

Conclusion
In conclusion, we optimize the initial capacity of the Nb 2 C anode and further preintercalate Li + ions to the desired voltage with optimal dynamics to solve the main issue of capacity lopsidedness and slow dynamics in LICs. Consequently, a Nb 2 C // LFP LIC with an enlarged potential window of 4.2 V can be successfully realized. Meanwhile, the prelithiated Nb 2 C // LFP LICs demonstrate remarkably enhanced energy density of  . Electrochemical performance of Nb 2 C // LFP lithium-ion capacitors (LICs). a) Gravimetric capacity at different current densities as a function of the discharge rate for prelithiated and un prelithiated LICs. b) The first charge-discharge curves for prelithiated and un prelithiated LICs. c) The energy density of prelithiated and unprelithiated LICs versus different current densities. The visualized demo of 1.1 V prelithiated LICs for d) a LED pattern with more than 300 bulbs, e) smart watch, and f) LCD timer.
Energy Environ. Mater. 2023, 6, e12505 5 of 7 420 Wh kg À1 , which are about three times higher than that of unprelithiated Nb 2 C // LFP LICs. Additionally, the underlying mechanism between prelithiation and high voltage is disclosed from electrochemical dynamic analysis. Prelithiation declines the Nb 2 C anode potential that facilitates electron transmission in the interlayer of twodimensional Nb 2 C MXene. This effect induces small drive force for Li + ions deposition and hence weakens the repulsive force from adsorbed ions on the electrode surface. Benefiting from even more Li + ions deposition, a higher voltage is eventually delivered. Our explanations on the relationship between prelithiation and high voltage based on the electrochemical kinetic behavior pave the way for high-voltage and highenergy-density energy storage devices.

Experimental Section
Preparation of Nb 2 C MXene materials: The Nb 2 C MXene was prepared by a modified solution etching method or hydrothermal etching method with HCl and LiF as the etching agents. The Nb 2 AlC MAX powders were purchased from 11 technology (Jilin, China). As for the modified solution etching method, 2.0 g LiF was dissolved into 30 mL of HCl (9 M) under continuously stirred for 10 min, followed by carefully adding 1.2 g Nb 2 AlC powders. The Nb 2 AlC powders were etched at 65°C or 70°C for 120 h, named as 65°C-5 days and 70°C-5 days, respectively. As for the hydrothermal etching method, 2.0 g LiF was dissolved into 40 mL of HCl (9 M) under continuously stirred for 10 min, followed by carefully adding 2.0 g Nb 2 AlC powders. The Nb 2 AlC powders were then etched at 150°C for 30 or 36 h in an autoclave, named as HR-30 h and HR-36 h respectively. After that, the obtained colloidal solutions were washed with deionized water and centrifuged several times until the pH of the solution reached about 7.0. Finally, the as-prepared Nb 2 CT x powders were collected via vacuum filtration and further dried at 40°C overnight.
Characterization methods: X-ray diffraction patterns were tested with a PANlytical X'Pert Powder diffractometer. TEM was used to characterize the morphologies of the samples by JEOL JEM-2100 F. SEM and EDS (JSM-7800F Prime) were used for morphology detection and elemental mapping analysis. The XPS depth profile was recorded with a Thermo Scientific ESCALAB 250Xi spectrometer. Raman spectra were conducted on RM2000 microscopic confocal Raman spectrometer employing a 532-nm laser beam.
Electrochemical measurements: The active materials (Nb 2 C for anode and LiFePO 4 for cathode) were mixed with Super P (Timical, Switzerland), and polyvinylidene fluoride solution (Arkema, France, 4.0 wt % in N-methyl pyrrolidone solution) at a weight ratio of 85:10:5 in a moderate amount of N-methyl pyrrolidone to obtain a slurry. Then, the slurry was coated on a copper foil (Nb 2 C) or aluminum foil (LiFePO 4 ). The as-obtained electrodes were dried in a vacuum oven at 100°C for 24 h for completely removing the solvents. The mass loading of the active material is 2.0-3.0 mg cm À2 .
The prelithiated progresses of the anodes were realized using CR2032 coin typed cells with Li metal as the counter/reference electrodes. A Neware battery testing system (Neware, China) was used to evaluate the rate and cycling performance of cells with the measured voltage ranging from 0.05 to 2.5 V for half-cell and 2.0 to 4.2 V for full battery. EIS and CV curves were carried out using the electrochemical workstation (CHI660E). For galvanostatic intermittent titration technique (GITT) measurement, the cells were firstly discharged or charged at a constant current pulse of 0.1 C (1 C = 270 mAh g À1 ) for 30 min, followed by an equal duration relaxation time of 3 h, allowing the equilibrium potential of lithium storage at different points to be probed in the whole voltage window of 2.0-4.2 V.