Revisiting Electrolyte Kinetics Differences in Sodium Ion Battery: Are Esters Really Inferior to Ethers?

The ether electrolytes usually outperform ester electrolytes by evaluating sodium‐ion batteries (SIBs) rate performance, which is a near‐unanimous conclusion of previous studies based on an essential configuration of the half‐cell test. However, here we find that contrary to consensus, the ester electrolyte shows better Na storage capability than the ether electrolyte in full cells. An in‐depth analysis of three‐electrode, symmetric cell, and in situ XRD tests indicates that traditional half‐cell test results are unreliable due to interference from Na electrodes. In particular, Na electrodes show a huge stability difference in ester and ether electrolytes, and ester electrolytes suffer more severe interference than ether electrolytes, resulting in the belief that esters are far inferior to ether electrolytes. More seriously, the more accurate three‐electrode test would also suffer from Na electrode interference. Thus, a “corrected half‐cell test” protocol is developed to shield the Na electrode interference, revealing the very close super rate capability of hard carbon in ester and ether electrolytes. This work breaks the inherent perception that the kinetic properties of ester electrolytes are inferior to ethers in sodium‐ion batteries, reveals the pitfalls of half‐cell tests, and proposes a new test protocol for reliable results, greatly accelerating the commercialization of sodium‐ion batteries.


Introduction
Lithium-ion batteries (LIBs) have got enormous success in the past decades, but the limited lithium resources and rising prices significantly restrain the applications of LIBs in large-scale power stations and electric vehicles. [1,2] Therefore, it is urgently necessary to develop new battery technologies to replace LIBs in certain application scenarios. [3,4] Sodium-ion batteries (SIBs) are considered a promising energy storage device for smart grids due to the low cost and abundant natural resources of sodium and the properties similar to LIBs. [5,6] The practical commercial application of SIBs requires the support of high-performance electrode materials and electrolytes. [7][8][9] Recently, hard carbon (HC) has been regarded as the most promising anode material for SIBs due to its high specific capacity (above 300 mAh g À1 ), stability, and low cost. [10][11][12][13] However, the situation on the electrolyte side is not optimistic, especially since numerous studies have shown that HC exhibits poor rate capability in ester electrolytes, which are widely used in commercial LIBs. [14,15] It is usually attributed to the thick solid electrolyte interface (SEI) film with a large impedance and low diffusion coefficient formed by the electrolyte on the HC surface. [16,17] Interestingly, ether electrolytes, less useful in commercial LIBs due to poor anode and cathode passivation, have become highly sought after in SIBs. More and more reports have indicated that ether electrolytes show advantages in almost all important aspects of HC sodium storage performance, especially the rate capability, and are considered alternatives to traditional ester electrolytes. [18][19][20][21][22][23] Not only carbon anodes but the same conclusions are also widely confirmed in the research of other types of anodes such as TiO 2 , [24,25] GeO, [26] Sn, [27] and even cathode materials. [28,29] Nevertheless, these conclusions are all based on a traditional half-cell test (THT) with a rigid cut-off, where one Na metal foil is used simultaneously as the counter/reference electrode. [30] This method is widely used in the basic research of electrodes and electrolytes due to its simple assembly. However, the practical application of SIBs is based on the full-cell system, so its results are more convincing. Moreover, some recent studies have suggested that the poor stability of Na electrodes in ester electrolytes may affect half-cell test results, whereas the situation in ether electrolytes is unclear and has not been reported. [31][32][33] For instance, Ji et al. reported that the HC rate performance was underestimated in traditional half-cell test compared to three-electrode cells. [34] Li et al. found that the HC exhibits superior rate performance in fullcell than half-cell due to the absence of Na electrode interference. [35] Regrettably, to date, no studies have been reported considering the Na electrode interference or using a full-cell system when comparing the sodium storage properties of ether and ester electrolytes. In this context, The ether electrolytes usually outperform ester electrolytes by evaluating sodium-ion batteries (SIBs) rate performance, which is a near-unanimous conclusion of previous studies based on an essential configuration of the half-cell test. However, here we find that contrary to consensus, the ester electrolyte shows better Na storage capability than the ether electrolyte in full cells. An in-depth analysis of three-electrode, symmetric cell, and in situ XRD tests indicates that traditional half-cell test results are unreliable due to interference from Na electrodes. In particular, Na electrodes show a huge stability difference in ester and ether electrolytes, and ester electrolytes suffer more severe interference than ether electrolytes, resulting in the belief that esters are far inferior to ether electrolytes. More seriously, the more accurate three-electrode test would also suffer from Na electrode interference. Thus, a "corrected half-cell test" protocol is developed to shield the Na electrode interference, revealing the very close super rate capability of hard carbon in ester and ether electrolytes. This work breaks the inherent perception that the kinetic properties of ester electrolytes are inferior to ethers in sodium-ion batteries, reveals the pitfalls of half-cell tests, and proposes a new test protocol for reliable results, greatly accelerating the commercialization of sodium-ion batteries.
despite numerous studies that have evaluated/compared the performance of ester and ether electrolytes, research questions of fundamental importance are still unanswered: 1) Is it reliable to evaluate/compare the performance of ester and ether electrolytes in SIBs only by the traditional half-cell test? 2) In a practical full-cell system, which is the better choice between ethers and esters? and 3) What should be the best way to evaluate the performance of electrodes and electrolytes accurately?
Here, cellulose powder-based hard carbon (PHC) was prepared to systematically evaluate/compare the real sodium storage properties in ester and ether electrolytes using different test systems and reveal the interference of the Na electrode. Contrary to the poor results in the half cell, in the full cell, the ester electrolyte exhibits superior rate performance than the ether electrolyte (270 vs 194 mAh g À1 at 1 A g À1 ). Three-electrode cell, symmetrical cell, and in situ XRD test are used to systematically investigate the reasons for the diametrically opposite results of the two electrolytes in full and half cells. The analysis reveals a huge difference in the stability of the Na electrode in the two electrolytes, which is ignored in the half-cell test, interfering with the accuracy of the results. Specifically, the overpotential and impedance of the Na electrode in ester electrolytes are much higher than those in ether electrolytes, making more capacity underestimated and further leading to misjudgment that ester electrolytes are far inferior to ethers. More seriously, this interference is also present in the three-electrode test, which is considered more accurate. To accurately evaluate the rate capability of HC, a "corrected half-cell test" protocol that can shield the Na electrode interference is developed, showing the inherent high-rate capability of PHC in both ester and ether electrolytes (~300 mAh g À1 at 2 A g À1 ). Our report reveals the pitfalls in comparing electrolyte kinetic data for SIBs without considering Na metal electrode interference, breaking the prejudice that the ester electrolyte is worse than the ether electrolyte.

Microscopic Morphology and Structure of PHC
Firstly, the morphology and structure of the obtained PHC were systematically characterized. One can see that all structural characterization results confirm the very typical hard carbon structure of PHC (Figures S1-S5, Supporting Information). A detailed characterization analysis is provided in the Supporting Information.

The Illusion of Ether Superior to Ester Electrolytes
The Na storage performances of PHC in DEGDME and EC/DEC electrolytes were first tested in a traditional two-electrode half-cell, as in many previous reports. The first galvanostatic charge-discharge curves at 0.02 A g À1 are shown in Figure S6, Supporting Information, where PHC displays high reversible capacity and initial Coulombic efficiency (ICE) in both electrolytes. Precisely, the DEGDME electrolyte delivers a slightly higher specific capacity and ICE (377 mAh g À1 , 95%) than the EC/DEC (356 mAh g À1 , 87%). Figure 1a shows the rate performance of PHC at different rates from 0.02 to 2 A g À1 in the voltage range of Figure 1. Traditional half-cell test performance. a) Rate performance of PHC//Na half-cell using EC/DEC and DEGDME electrolyte. Voltage profiles at different rates using b) the EC/DEC electrolyte, c) the DEGDME electrolyte. d) Corresponding capacity contributions attributed to the plateau and slope region at different rates. e) Nyquist plots of the half-cell after the first cycle at 0.02 A g À1 . f) Polarization potential at different rates.
Energy Environ. Mater. 2023, 6, e12523 2 of 8 0-2.8 V. With increasing rate, the reversible capacity rapidly decreases in the EC/DEC electrolyte, especially under the rates >0.1 A g À1 (Table S1, Supporting Information). On the contrary, PHC shows super rate capability in the DEGDME and can still maintain 201 mAh g À1 at even 2 A g À1 , nearly five times higher than 45 mAh g À1 in the EC/ DEC. The rapid capacity degradation at high rates in the EC/DEC electrolyte is mainly due to the loss of the 0.1 V low-voltage plateau region ( Figure 1b). By contrast, it can still be well maintained in the DEGDME electrolyte ( Figure 1c,d). The reaction kinetics of the half-cell was explored using electrochemical impedance spectra (EIS), and the resistances were quantified with an equivalent circuit (Figure 1e, Figure S7, Supporting Information). The results exhibit that the impedance in the DEGDME is significantly lower than that in the EC/DEC, especially the interface impedance related to the SEI film. The lower polarization potential in the DEGDME electrolyte at different rates further confirms the view (Figure 1f, Table S2, Supporting Information). Moreover, the PHC also exhibits more excellent cycling stability in the DEGDME electrolyte at 0.1 A g À1 for 500 cycles ( Figure S8, Supporting Information). The results of the traditional half-cell test all indicate that the PHC anode accesses better kinetics in ether than in ester electrolytes, agreeing with those previous reports. [17,21,36,37] In the full-cell test, the promising layered metal oxide O3-Na (NiFeMn) 1/3 O 2 (NFM) was initially selected as the cathode to match the PHC anode. [38,39] The rate capability of NFM//PHC full cell was examined at different rates from 0.02 to 2 A g À1 (Figure 2a-c). Surprisingly, the full-cell test results overturned the half-cell results. The NFM//PHC full cell demonstrates superior rate capability in the EC/DEC electrolyte and delivers a specific capacity of 237 mAh g À1 at 2 A g À1 , far more than the traditional half-cell test result (45 mAh g À1 ). In contrast, the NFM//PHC full cell retains only 155 mAh g À1 under the same rate in the DEGDME electrolyte (Table S3, Supporting Information). Moreover, the interface impedance of the full cell in the EC/DEC is smaller than in the DEGDME, indicating faster kinetics (Figure 2d). The lower polarization potential in the EC/DEC at different rates further confirms this view ( Figure S9, Table S4, Supporting Information). Furthermore, NFM// PHC full cell exhibits more excellent cycling stability in the EC/DEC electrolyte than in the DEGDME electrolyte ( Figure S10, Supporting Information). In conclusion, the kinetic performance of the ester electrolyte is better than that of the ether electrolyte in the full-cell test, which is contrary to the traditional half-cell test results and has never been mentioned in previous studies. Meantime, the diametrically opposed results further illustrate the necessity of studying the performance of electrodes and electrolytes in a full-cell system.
In other words, the EC/DEC electrolyte, which performs poorly in anode half-cells, exhibits excellent kinetics in full-cell tests, whereas the DEGDME electrolyte is not as ideal as the half-cell test (Figure 2e,f). It is also demonstrated by the strong comparison of EIS results between half-cell and full-cell tests, especially with the EC/DEC electrolyte (Figure S11, Supporting Information). Furthermore, the superior rate capability of the full-cell over the half-cell is also observed in 1 M NaPF 6 -EC/DMC, another commonly used ester electrolyte ( Figure S12, Supporting Information). We would have thought that the complete opposite result was entirely due to the full-cell results being affected by the cathode side. [40,41] However, for the cathode side, the rate performance of NFM//Na half-cells in the EC/DEC electrolyte is also lower than that of DEGDME electrolytes, again in contrast to the full-cell results  Figure S13, Supporting Information). It means that the results of testing the cathode and anode materials separately in half cells are quite different from those assembled into full cells. Now we have to consider whether the half-cell test results reflect the real capability of the electrolyte and electrode material. To confirm the universality of the above results and exclude the influence of cathode type, another promising cathode Na 3 V 2 (PO 4 ) 3 (denoted as NVP) was further adopted to assemble an NVP//PHC full cell. [42] The NVP//PHC full cell also exhibits greater rate performance and smaller impedance in the EC/DEC than in the DEGDME electrolyte, agreeing with the NFM//PHC results (Figure S14, Supporting Information). These results suggest that 1) the traditional half-cell test cannot accurately assess the real performance of electrodes and electrolytes in SIBs and 2) the conclusion in previous studies that ether electrolytes possess better capability than ester electrolytes may be erroneous due to half-cell testing artifacts, while the largest difference of the half-cell compared with the full cell is the use of Na metal electrodes, which is the primary suspect in causing the measurement artifacts.

Revealing the Reason for the Measurement Artifacts
To explore the influence of the Na electrode on the two electrolytes, a three-electrode cell test was carried out with PHC as the working electrode and two separate Na foils as the reference and counter electrode, respectively ( Figure S15, Supporting Information). [34] In the EC/DEC electrolyte, the capacity of the three-electrode cell is much higher than the traditional half-cell test results at all rates, confirming that the traditional half-cell test underestimates the real rate performance of PHC ( Figure S16a, Supporting Information). More direct evidence for this conclusion is shown in the potential curves of the working electrode and the counter electrode (Figure 3a,b). And the intersection of the working electrode and the counter electrode potential curve corresponds to the position of the discharge cut-off in the traditional halfcell test. The Na counter electrode exhibits a high overpotential and increases rapidly with the current rate, reaching 0.167 V at 1 A g À1 , which are all ignored in the traditional half-cell test since the counter and the reference electrode are integrated. Correspondingly, a large part of the sodiation capacity of PHC is collected when its potential is lower than the counter electrode potential, which part capacity will not appear in the traditional half-cell test results. More Seriously, as the rate >0.2 A g À1 , the 0.1 V low potential plateau of the PHC is completely "masked" by the overpotential of the counter electrode, thereby disappearing in the traditional half-cell test. It suggests that in EC/DEC, the high overpotential of the counter electrode at high rates causes the test to reach the rigid cut-off potential prematurely and makes the PHC performance underestimated. By contrast, in the DEGDME electrolyte, the overpotential of the counter electrode is much smaller at all rates, only 0.064 V at 1 A g À1 (Figure 3c,d). Meanwhile, the capacity of PHC in DEGDME is hardly affected by the counter electrode at high rates, in sharp contrast to the results in EC/DEC ( Figure S16b, Supporting Information). It may be related to the different reactivity of Na metal in the two electrolytes. [43] Terribly, the difference between the Na metal electrodes in the two electrolytes is ignored in the traditional half-cell test and crudely attributed to the working electrode. Although the capacity of the PHC is improved by the three-electrode cell, confirming that the Na electrode in the traditional half-cell test interferes with the determination of the capacity, the results of EC/EDC electrolyte are still inferior to that in the full cell (Tables S1 and S5, Supporting Information). Moreover, the rate performance of the ester electrolyte is still inferior to that of the ether electrolyte in the three-electrode cell test, which is inconsistent with the full-cell results. It means that the three-electrode cell test can also not measure the real sodium storage capacity of PHC, especially in ester electrolytes. To further confirm the interference of Na metal electrodes in the traditional half-cell test on the performance evaluation/comparison of different electrolytes, Na//Na symmetric cells were fabricated in EC/ DEC and DEGDME electrolytes, respectively. The time-voltage curves of Na//Na cells at current densities of 0.1, 1, and 5 mA cm À2 are shown in Figure 4a. The DEGDME electrolyte exhibits low overpotential and smooth charge-discharge curves at all current densities, verifying the good compatibility of Na metal with the DEGDME and the stable SEI film formed on the surface. [44] In contrast, the EC/DEC electrolyte exhibits higher polarization potential and voltage fluctuations at all current densities, proving its compatibility with Na metal is worse than that of the DEGDME electrolyte. [45] In addition, voltage spikes are observed in EC/DEC electrolyte at 5 mA cm À2 , indicating that the Na//Na cell cannot normally work now, while it can still run stably in the DEGDME electrolyte. In conclusion, the results show a huge difference in the electrochemical reaction of Na metal in the two electrolytes, especially the overpotential. However, the interference of the Na metal electrode is ignored in the traditional half-cell test, resulting in inaccurate test results. It corroborates well with the above three-electrode cell test results.
Based on the charging/discharging results, the evolutions of the impedance of Na//Na cell over time in the two electrolytes were further explored. Specifically, the fresh Na//Na cells were placed at an open-circuit voltage for 120 h and measured impedance every 4 h. The Na//Na cell in the DEGDME shows a small initial interfacial resistance of only 12 Ω and gradually decreases with time ( Figure 4b). It corresponds to the extremely low polarization and stable voltage distribution in the constant current test. [46] In contrast, the symmetrical cell with the EC/DEC electrolyte shows great differences. First, the initial interface impedance reaches an astonishing 800 Ω, almost 70 times that with the DEGDME electrolyte ( Figure 4c). Furthermore, the interfacial impedance increases significantly with time, indicating that Na metal continuously reacts with the EC/DEC electrolyte and produces an unstable SEI film even without current being applied. [47] The huge contrast between the two electrolytes further indicates that the huge impedance difference in the half-cell mainly originates from the Na electrode side. Regrettably, in the traditional half-cell test, it is often crudely attributed to the working electrode side to explain the inferior EC/DEC rate performance than DEGDME. Meanwhile, the high reactivity with EC/DEC and the unstable SEI film will hinder the Na-metal reference electrode from maintaining a constant potential, explaining the reason the three-electrode cell test cannot measure the real capacity of the PHC.
In the above discussion, it has been demonstrated that the impedance results of half-cells are strongly interfered with by Na electrodes, especially in the EC/DEC electrolyte. To explore the real impedance profile of the working electrodes in the two electrolytes, we disassembled the cycled PHC//Na and NFM//Na half-cells, respectively. The cycled electrodes were assembled into PHC//PHC and NFM//NFM symmetric cells and then subjected to EIS testing ( Figure S17, Supporting Information). The interfacial impedance of HC//HC symmetric cells in the two electrolytes is similar and both are less than 10 Ω (Figure 4d), contrary to the traditional half-cell test results (Figure 1e). The same situation also appears in NFM//NFM symmetric cells, both benefiting from the absence of the Na electrode (Figure 4e). It further explains that the overpotential interference with the traditional half-cell test results is mainly due to the interface impedance between the Na electrode and the electrolyte. Importantly, the EC/DEC electrolyte exhibits low interfacial impedance at both cathode and anode, confirming its potential for application in full cells. It is well known that the electrode interfacial impedance and the composition of the SEI layer are closely related. We carefully investigated the chemistry and structure of SEI layers on different electrodes in EC/DEC and DEGDME electrolytes using X-ray photoelectron spectroscopy (XPS) with depth profiling. Surprisingly, cycled PHC electrodes show roughly similar C1s, O1s, and F1s peaks in EC/DEC and DEGDME electrolytes ( Figure S18, Supporting Information). The analysis displays that the SEI films of PHC electrodes in both electrolytes are composed of a dense inorganic inner layer and a flexible and thin surface organic layer. It is corroborated with the results of interface impedance. Not only the PHC electrode but the NFM cathode also exhibit similar results in both electrolytes ( Figure S19, Supporting Information). In contrast, there is a significant difference in the SEI films produced on Na electrodes in EC/DEC and DEGDME electrolytes. In the DEGDME electrolyte, the interior of the SEI on the Na electrode consists largely of inorganic Na 2 O and NaF species, with a very thin organic layer on the surface ( Figure S20, Supporting Information). While in the EC/DEC electrolyte, the organic layer on the SEI surface is very thick, and the content is still high even after etching, which leads to poor interfacial impedance. In addition, the digital photos of the Na electrodes after cycling show that the surface of the electrodes in DEGDME is significantly flatter and more uniform ( Figure S21, Supporting Information). These results further demonstrate that the stability of Na electrodes is very different in EC/DEC and DEGDME, which will interfere with the half-cell test results.

The Real Na Storage Performance of PHC in Ether and Ester Electrolytes
The preceding discussion indicates that neither the traditional half-cell test with a rigid 0 V cutoff protocol nor the three-electrode cell test can accurately measure the real Na storage performance of PHC. To solve this problem, an in situ XRD experiment was performed to study the discharge curve characteristics of oversodiation ( Figure S22, Supporting Information). For the DEGDME electrolyte, the voltage trace shows an obviously "V" cusp characteristic as the cell voltage approaches À0.03 V, generally considered the Na nucleation growth process. It is confirmed by the simultaneous appearance of Na metal diffraction peaks at 2h % 29°in the in situ XRD pattern ( Figure  5a). [48,49] The EC/DEC electrolyte exhibits a similar result, but the "V" cusp characteristic and Na metal diffraction peaks appear at a lower discharge voltage (À0.06 V) due to the higher overpotential of the Na electrode (Figure 5b). It indicates that once the voltage reaches this "V" cusp nadir, no more Na + can be inserted into the PHC; instead, Na + will be electrodeposited on the carbon surface as a metal. [50,51] That is to say, the "V" cusp can be used as a sign for the sodiation completion of PHC. Based on this, we have developed a "corrected half-cell test" (CHT) protocol, where the half-cell over-discharges until the start of Na plating. Specifically, the discharge cut-off condition can be determined by the appearance of a sharp "V" characteristic or the value of the dV/dt (Figure 5c). And the specific capacity before the "V" nadir is defined as the real specific capacity. The "V" nadirs in both electrolytes are <0 V, further suggesting that the traditional half-cell test with a rigid 0 V cut-off condition underestimates the real Na storage capacity of PHC, especially in the EC/DEC electrolyte.
The rate performance of PHC in the EC/DEC and DEGDME electrolytes was re-evaluated using the corrected half-cell test protocol. The Figure 5. The in situ XRD spectra of the PHC anode during the second cycle, a) with DEGDME electrolyte and b) with EC/DEC electrolyte. While the XRD spectra were being collected, the cell was discharged until Na metal deposition appeared and lasted for 10 min, then charged back to 2.8 V. c) Schematic diagram of working principle of the corrected half-cell test.
Energy Environ. Mater. 2023, 6, e12523 6 of 8 full and zoomed-in charge-discharge curves of PHC at different rates in the EC/DEC electrolyte are shown in Figure 6a,b. As the rate increase, the shape of the discharge curve hardly changes, but the plateau region persists at lower and lower voltages, even completely below 0 V. It exactly explains the illusion that the plateau capacity decays rapidly under high rates in the traditional half-cell test. Furthermore, the excellent real rate capability of PHC in the ester electrolyte is fully released under the corrected half-cell test protocol, and the reversible capacity can still reach 295 mAh g À1 at 2 A g À1 , supporting the full-cell test results (Table S6, Supporting Information). For the DEGDME electrolyte, the capacity is also improved at high rates than the traditional half-cell test protocol, but it is not obvious. In addition, the overpotential at different rates is much smaller than in the ester electrolyte due to the stability of Na in the DEGDME (Figure 6e). Surprisingly, the real Na storage capacity of PHC in the two electrolytes at different rates is almost the same (Table S7, Supporting Information). The difference is that under high rates, in the DEGDME electrolyte, the capacity is mainly concentrated in the area >0 V, while that in the EC/DEC electrolyte is mainly concentrated in the area <0 V due to the different overpotentials of Na (Figure 6f). The above results once again prove that the conclusion that the hard carbon rate performance in the ester electrolyte is far worse than that in the ether electrolyte is a test artifact.

Conclusion
In summary, we evaluated the kinetic properties of ether and ester electrolytes for the first time in full cells and confirmed that the rate performance of ester electrolytes is superior to that of ether electrolytes (270 vs 194 mAh g À1 at 1 A g À1 ), overturning conclusions from previous studies using traditional half-cell tests. A thorough analysis concludes that the traditional half-cell test protocol with a rigid cut-off condition ignores the overpotential and impedance of the Na metal electrode, resulting in an undervalued real capability of the PHC. While the overpotential of the Na electrode in the ester electrolyte is much larger than that in the ether electrolyte, and the capacity underestimation is also more serious, resulting in the illusion that the ester electrolyte is inferior to ether. In addition, the three-electrode cell test, considered more accurate, is also interfered with by the Na electrode leading to inaccurate results. Thus, we develop a "corrected half-cell test" (CHT) protocol that can shield the Na electrode interference. The CHT protocol reveals the inherent high-rate capability of PHC in both ester and ether electrolytes (~300 mAh g À1 at 2 A g À1 ). Finally, although this work focuses on hard carbon for SIBs, we believe that our findings and theories can also apply to other types of anode/cathode for SIBs and other types of battery systems, such as K-ion batteries. In general, our report revisits the real performance of electrodes and electrolytes for SIBs and proposes a new theoretical framework and testing protocol to guide the design and matching between electrolyte and electrode materials for future commercial alkali metal-ion batteries.