Background
The Negative Electrode-Free Lithium Metal Battery (AF LMB) uses an "empty" Cu thin foil current collector where all available LI is stored in the positive electrode. AF LMB has demonstrated the ability to increase the gravimetric energy density of the battery by approximately 35% and the volumetric energy density by approximately 80% compared to graphite-based anode LIBS. The cycling stability of AF LMB can be extended through a range of novel electrolytes, including ether electrolytes, fluorinated electrolytes, and ionic liquids. High static pressure (up to 10 MPa) has been shown to promote the electrodeposition of AF LMBS dissolution stability and improve CE.
While external pressure is known to generally improve electrochemical performance, a limiting aspect of the method is the battery separator. While "dry" polyolefin separators (PE PP) may have strengths of up to tens of MPa, they become very weak when immersed in a liquid electrolyte. As a result, wet diaphragms are prone to tearing under high pressures. This further limits the external pressure that can be used in the actual battery pack. Due to the propensity for dendrite and unstable SEI in the anode-free configuration, achieving an extended cycle of AF LMB under practical pressure remains an extremely challenging task.
Brief introduction of the results
Recently, Liu Wei's team at Sichuan University and D**ID Mitlin from the University of Texas at Oster worked together to explore the correlation between external pressure, solid electrolyte interface (SEI) structure morphology, and lithium metal plating peeling behavior. To model an anode-free lithium-metal battery (AF LMBS), an analysis of the "empty" Cu current collector was performed in a standard carbonate electrolyte. The lower pressures promote organic-rich SEIs and macroscopically inhomogeneous filamentous LIs that are scattered in the pores.
The higher pressure promotes inorganic F-rich SEI with a more homogeneous and denser LI membrane. The "seed layer" (pg@cu) of the lithiated raw graphene favors anion-derived F-rich SEIs and promotes uniform metal electrodeposition, resulting in extended electrochemical stability at lower pressures. State-of-the-art electrochemical performance was achieved at 1 MPa: the PG-enabled half-cell was stabilized after 300 hours (50 cycles) at a rate of 3 mAh cm2 at a capacity of 1 mA, with a cycle coulombic efficiency (CE) of 998%。
An AF-LMB battery with a high-mass loaded NMC622 cathode (21 mg cm2) underwent 200 cycles with a C5 charge and a C2 discharge (1C=178 mAh G1) with a CE of 994%。Density functional theory (DFT) emphasizes the difference between the adsorption energy of solvated Li+ on various crystal planes of Cu(100), (110) and (111) and the adsorption energy of lithiated (0001) graphene, thus providing insight into the role of carrier surface energy in promoting SEI heterogeneity.
The results were published in the top international journal Advanced Energy Materials in an article entitled "interrelation between external pressure, sei structure, and electrodeposit morphology in an anode-free lithium metal battery".
**Reading guide
Figure 1A illustrates a custom-made PEEK cell that allows static pressure to be measured and controlled during electrochemical testing. Two types of experiments were conducted: the one called "lithiation" was designed to probe the structure of the SEI without causing any metal plating, and the lower cut-off voltage was 001v。The experiment, called "plating", is intended to probe the structure of lithium metal and associated SEIs, where the lower voltage is determined by the plating overpotential and the higher voltage is set to 1V relative to the LI Li+. These galvanic distributions at different static pressures are shown in Figure 1b. With pressure from 01 was increased to 1 and 10 MPa, and the nucleation overpotential was reduced from 340 to 290 and 150 mV, respectively.
li||The electrochemical impedance spectroscopy (EIS) results of the Cu half-cell are shown in Figure 1C E. Bilayer SEIs have been previously reported and attributed to a two-step table** formation process, one formed by an impregnation-based chemical reaction of the Cu surface (Cuox) with a LiP6-based electrolyte, and the other more associated with further electrochemical electrolyte reduction. in lithiation to 0At 01V, the ohmic resistance (Rohm) of the battery decreases as the pressure increases.
This is attributed to the improvement in electrical contact inside the battery, as the pressure is not expected to affect the ionic resistance in the electrolyte. Once the LI is plated, the ohmic resistance of the battery actually increases. This is attributed to the degradation of electrical contact due to the metal's poor coverage of the current collector. In the plating state, the effect of increasing pressure becomes very significant. The modeling results will show that the external pressure does more than just flatten the metal surface to achieve better contact with the current collector. External pressures affect metal nucleation and growth energetics, which to some extent compensate for the unfavorable interfacial energetics between LI electrodeposition and unmodified copper foils.
Figure 1] a) Schematic diagram of a custom PEEK cell that allows pressure control during battery operation. b) At 1 mAcm1, the external pressure is or 01 MPa at cu|Voltage distribution of the LI plating on the LI. (i) and (ii) show the extended portion of the nucleation and growth fractions, respectively. (ce) Lithium to 001 V li|EIS spectra of CU.
Figure 2ac compares top-down optical and SEM images, showing the different LI electrodeposition morphologies formed at three pressures. The electrochemical protocol consists of a single plating step of 3 mAh cm2 on an empty Cu current collector at 1 mA. It can be observed that the microstructure and surface topography of electrodeposition are strongly influenced by static pressure. According to 2a, it can be observed that at 0At 1 MPa, the electrodeposition microstructure consists of an interconnected network of metals, SEIs, and micron-sized pores. Electrodeposition covers less than half of the current collector and is black in color. According to 2b, at 1 MPa, electrodeposition is less porosity under the microscope, giving it a visible metallic luster.
Macroscopically, electrodeposition covers a ratio of 01MPa specimen with more current collector surface. According to 2C, at 10MPa, electrodeposition is most advantageous in terms of membrane continuity and morphology. As the pressure increases, the electrodeposition covers more of the current collector surface, has fewer pores in its body, and shows a lower surface roughness. However, an external pressure of 10MPa can be observed to be sufficient to physically tear the PP diaphragm. As shown in 2C, about half of the electrodeposition remains on the copper foil surface, while the other half adheres to the torn PP separator.
Figure 2D L shows the C 1S and F 1S XPS spectra of the li-plated surface at three pressures. As shown by sputter XPS analysis, the LiPoFX product is predominantly present on top of the SEI. The decomposition of the additive FEC produces not only LIF but also organic substances containing Li2CO3+F, while solvents in the electrolyte, such as EC and DEC, decompose to obtain RO CO2 Li and CO substances.
The results show that at higher static pressures, there is a tendency for relatively more PF6 to decompose, resulting in more inorganic-rich SEIs. Increased SEI homogeneity and a more inorganic-rich SEI favor electrochemical stability, especially if the result is a more uniformly distributed LIF phase. As the pressure increases, the decomposition of carbonate solvent (EC DMC EMC FEC) on the LI surface can be inhibited, allowing for preferential accumulation of FEC and anion reduction products. As shown in Figure 2Dl, XPS depth profile experiments verify this trend.
Figure 2] Analysis of lithium plating on a Cu current collector in a half-cell with a capacity of 3 mAh cm2 at 1 mA2. a c) at and 0., respectivelyLow-power optical microscope image and high-power SEM image of the LI metal surface at 1 MPa. D f) Compare the XPS F 1S, Li 1S, and C 1S spectra of SEI structures at 10 MPa. g i) same analysis, but at a pressure of 1 MPa. j–l）0.1 mpa。
This pg@cu half cell at 50 Acm2, at 0 with respect to LiLi+01 In the 3 V voltage range, a lithiation-denitroidation step was undertaken, as shown in Figure 3a. Figure 3b shows the delithiumation of pg@cu after lithiation. It can be observed that the PG layer remains 5 m thick despite being covered by SEI. This layer has been partially separated from the CU carrier, which can occur during sample preparation. As expected, the surface morphology of the copper foil did not change from its receiving state. Although macroscopically flat, the foil is coarse on the micron scale, as shown in the figure.
Figures 3C, D show top-down SEM images and corresponding EDXS plots for C, O, and F, pg@cu and the baseline Cu surface after a single lithiation-deliation. The SEI layer is formed in pg@cu and has a relatively uniform morphology. The EDXS map shows a uniform distribution of C, O, and F. As shown in Figure 3D, the SEI formed at the top of the baseline CU is highly heterogeneous. Figure 3e shows pg@cu after lithiation-deliitation. According to the analysis of F1S, the SEI components containing F were C F, LizpoyFx and Lif. As the sputtering time (depth) increases, the signal strength associated with LIF becomes the most important. The CO and CO peaks were detected at the top surface of the SEI. In general, carbonate solvents, including EC, DEC, EMC, and FEC, have relatively high LUMO (lowest unoccupied molecular orbital). Therefore, Li+-solvent complexes have a strong tendency to be reduced through the ring-opening and decarboxylation reaction pathways. This results in the SEI components Li2O Li2CO3, Roli (lithium alkoxy), RoCO2Li (oxidized lithium alkyl carbonate), and CF compounds. Conversely, the anion Pf6 decomposes to produce LixPfy, the Lif species. In general, Li2O Li2Co3 Fc is reduced in carbonate solvents (EC DEC EMC FEC) and LiF inorganic** is reduced in anions (PF6). This indicates that the highly ordered structure of PG leads to inhibition of solvent decomposition, allowing more Pf6 anion decomposition products (LixPfy and Lif) to precipitate on the surface. These anionic derivatives are known to promote the electrochemical stability of lithium metal anodes, with LIF being a key component in stabilizing the SEI structure.
Figure 3] A) Voltage distribution of pg@cu and baseline Cu half-cells, at 50 A cm01 Single lithiation delithiumation at 3 V. B) Side-view cross-sectional SEM image of the pg@cu electrode. c) Diagram of SEM and EDXS elements of the pg@cu surface. d) Perform the same analysis for the baseline CU.
Figure 4a,b shows a single plating step of pg@cu at a current density of 3 mA2 at a current density of 3 mA2. While there is considerable porosity in the PG below, the polycrystalline Li-metal plated on top of it is almost completely dense. There is some surface roughness in the film, which may be related to the difference in grain growth rate. Figure 4D E highlights LI plated on baseline copper foil under the same conditions. The resulting electrodeposition layer is 52 microns thick and consists of filamentous LIs and SEIs scattered in the pores.
Electrodeposition of CU metals at pg@cu as well as baseline. Electrode sectioning was performed using a combination of argon ion grinding and vacuum sample holders. Figure 4C,F compares the microstructure of the SEI layer formed with the two supports. The associated FFT highlights the crystalline composition of the SEI, i.e., the nano-sized Li2CO3 embedded on the amorphous matrix. It can be observed that pg@cu thin enough to also discern the (110)-oriented FFT from the Li metal. It is not possible to find a region in which both crystalline Li2CO3 and Li metal are discernible in the baseline Cu sample. The discovery of low-temperature TEM further supports the general conclusion that pg@cu support promotes more uniform plating of the Li metal and a more uniform SEI. Figure 4g shows the pg@cu along the line, while 4h shows the analysis of the baseline CU. Comparison of the F, C, and O signals in the two samples confirmed that the pg@cu was dense and did not contain scattered SEIs. For baseline CU, the coated LI is porous and mixed with SEI over the entire thickness.
Figure 4: A, B) Cross-sectional SEM images of the LI coating of the pg@cu half-cell. C) Cryogenic TEM analysis of lithium deposition on pg@cu and fast Fourier transform (FFT) patterns in the LI region (green) and SEI region (red) of bulk metal. d h) perform the same analysis for Li plating on baseline CU. g) EDXS element distribution of pg@cu. h) Baseline CU for the same analysis.
Density functional theory (DFT) calculations were performed to further understand the plating process and the role of PG as a support relative to the baseline CU. As shown in Figure 5E G, the correlation orientations for Cu are <100>, <110>, and <111> and are the standard fiber textures observed in FCC metals. Figure 5h shows the <0001 > orientation of graphene, which is related to the carrier used. Figure 5 d shows a top view of ethylene carbonate (EC) solvated Li+ as an adsorbent, which is a representative LIo solvation structure in a 1 M Lipf6 electrolyte. Figure 5e H shows the binding of anionic PF6 to three Cu crystal planes and graphene. For the adsorbent Li(EC)4, all three Cu planes showed significantly greater negative adsorption energies than graphene, with (100) being the most preferred.
As can be seen from the calculations, the trend of PF6 is reversed. Finally, the role of non-lithiation and lithiation PG was studied. A lithium cluster consisting of 11 LI atoms is used as the initial nucleus. Figure 5i j shows that the adsorption of lithium nuclei on the (0001) surface of the lithiated PG is more favorable compared to the non-lithiated (i.e., raw) surface.
This indicates that embedding LI into graphene should promote the enhanced adsorption of LI clusters, thereby enhancing the wettability of LI layer. After the initial layer of lily has been deposited, the subsequent metal should grow epitaxial at a growth rate dependent on the orientation of the crystals, while additional nuclei are formed as the film thickens. Thus, there may be a "memory effect" in which the initial wetting behavior of the first few layers of LI on PG (or baseline Cu) continues to affect the morphology and texture of the thickened electrodeposition.
Fig. 5] A d) Density functional theory (DFT) analysis. Top view of representative structures of solvated Li+ on (100) Cu, (110) Cu, (111) Cu, and (0001) PG surfaces. e h) Top view of adsorbed anions (PF6) on these surfaces. i j) Li cluster nuclei adsorbed on the surface of PG and lithiated PG, respectively. Li (green), O (red), C (brown), F (silver) and P (light purple).
To illustrate how differences in support structure and associated SEIs affect the electrochemical performance of anode-free lithium-metal batteries, experiments were conducted on half-cell and full-cell configurations using the NMC811 cathode. Figure 6a d compares pg@cu|li and baseline cu|Li half-cell. It can be observed that in the case of baseline CU, the overpotential increases with the number of cycles until the point of failure caused by the impedance rise at 20 cycles. While the baseline CU can only be cycled for 150 hours, pg@cu cells show stable plating stripping behavior over 300 hours. In Figure 6B, it can be observed that the CE of the baseline Cu decays rapidly around 20 cycles, while the pg@cu stabilizes at 95 over the entire 300-hour cycle5%。pg@cu|li and baseline cu|The voltage-capacity curve of the LI cell at the first turn is shown in Figure 6c. According to Figure 6D, the difference in nucleation overpotential persists over the course of cycling pg@cu 68 mV after 30 revolutions compared to 117 mV after 26 cycles at baseline cu.
Figure 6e shows the current density of pg@cu at and 6 mAcm2. Each cycle takes 0Plating time of 5 h with 20 cycles at each current density. For pg@cu voltage distributions at each current density are overlapping, indicating stable electrochemical behavior. Conversely, from 0Starting at 75 mAcm2, the baseline CU is unstable and the peeling capacity is significantly smaller than the plating capacity, as shown in Figure 6F. As shown in Figure 6g, at a high current density of 6 mA cm2, the peel curve becomes highly fluctuating, indicating the onset of battery failure.
Fig. 6] Comparison of the electrochemical performance of half-cells operating at 1 MPa. a–b）pg@cu|li and baseline cu|Li: Current profile and associated CE values at 3 mAh cm Macm2. c d) Current constant curves at 3 mAh cm mA2, 1st and 26th 30 cycles. e) pg@cu|at different current densitiesVoltage capacity curve of li. f-g) at 0Representative sections in 75 m2 and 6 mC2. h) CE analysis, in which 6 mAh cm2 of LI is first deposited as a LI reservoir, then stripped under 1 mA cm2 plating of 1 mAh cm2 for 100 cycles, followed by 100% DOD stripping.
Figure 7 shows the use of pg@cu as well as the baseline Cu as the negative current collector. The battery was tested at an external pressure of 1MPa. The mass load of the NCM cathode is 18 mg cm2, with 34 mAh cm2 commercial horizontal area capacity to match. Figure 7a b shows pg@cu||NCM and baseline CU||The initial cycle of charging and discharging of the NCM battery at C20-C20. Figure 7c shows the cycling performance at various magnifications of the LMB without the anode. At. 5 and 1c pg@cu||NCM batteries are available. 24 and 1Reversible capacity of 57 mAh cm2 with an average operating voltage of . 47 v。In 02 c charge 0pg@cu||The NCM battery exhibited 99 throughout the 30th to 100th cycles68% average CE with a capacity retention rate of 895%。In contrast, the baseline cu||NCM batteries exhibit rapid capacity decay from the beginning of the cycle.
pg@cu||NCM cells use state-of-the-art monocrystalline NCM particles to perform cycling performance tests on the cells. These cells are designated pg@cu||ncm-s。Compared to the commonly known "meatball" (polycrystalline aggregate) NCM, monocrystalline NCM particles can be calendered more aggressively. In addition, for these radical tests, the battery electrolyte volume was reduced from 100 l to 60 l. Figure 7d shows pg@cu||The NCM-S uses a constant current constant voltage (CCCV) cycling protocol. The program consists of 0A galvanostatic charge of 2C, followed by 005 c cut-off and 0Potentiostatic charge with 5 C discharge. pg@cu||NCM-S cells are available at 200 cycles (3.).87 to 273 mAh cm2) with a volume retention rate of 705%。
Figure 7: a b) based on pg@cu||NCM and baseline CU||The initial cycle of charging and discharging of the NCM battery at C20-C20. c) Cycling performance of non-negative LMB at various magnifications. For a c) ncm = 178 mg cm−2，1c=3.4 ma cm−2。d) Based on pg@cu||Cycling performance of NCM-HL, NCM-S=21 mg CM2, 1C=38 ma cm−2。
Summary and outlook
Anode-free lithium metal batteries (AF LMB) use an "empty" negative electrode current collector to facilitate repetitive plating of lithium metal stripping, with the positive electrode being the only one for shuttle ions**. This places stringent demands on the electrodeposition process, making the role of electrolyte interface (SEI) stability even more critical. Using a unique pressure-controlled liquid electrolyte electrochemical cell, combined with density functional theory simulations, the authors analyzed the effect of external pressure on the geometric inhomogeneity and structure of SEI and correlated it with lithium metal morphology and overall electrochemical stability. Three externally applied pressures are used;0.1 MPa is not enough to achieve even minimal electrochemical stability on bare CUs, 1 MPa still results in non-uniform SEI and dendritic metal electrodeposition at higher current densities, and 10 MPa, despite having the most promising electrochemical signature, is practically not feasible for several reasons.
At the same time, the role of the lithium graphene bottom layer as a secondary current collector is considered, demonstrating how and why it allows stable 1MPa plating under other same peeling behavior. The study has shown that for baseline polycrystalline Cu surfaces, organic-rich SEIs are favorable, which is accompanied by filamentous LIs and extensive porosity. The high external pressure favors inorganic F SEI and is accompanied by uniform and dense LI electrodeposition. The results show that the lithium graphene layer favors the same type of F-rich SEI, but at lower external pressures, promotes uniform metal plating peeling and stable electrochemistry in both half- and full-cell cells.
References
wei liu, yiteng luo, yuhang hu et al. interrelation between external pressure, sei structure, and electrodeposit morphology in an anode-free lithium metal battery. advanced energy materials. (2023)