Controlled Large-Area Lithium Deposition to Reduce Swelling of High-Energy Lithium Metal Pouch Cells in Liquid Electrolytes

,

reactions 3,4 , and cell failure mechanisms 5,6 and in increasing cell-level energy and cycling stability 7,8 .Key parameters, including the electrolyte amount, cathode mass loading, and lithium metal thickness, have been identified as crucial factors that dictate the cycling stability of lithium metal coin cells 3 .These parameters must be carefully controlled to achieve meaningful results during coin cell testing and enable fair comparisons 9,10 .However, practical applications of high-energy Li batteries are still hindered by several significant challenges.Despite the great progress on the design and fabrication of practical high-energy pouch cells 8 , the cycle life needs to be further improved.Another significant challenge is the large volumetric change of the anode and subsequent cell swelling.For practical applications, the cell swelling needs to be limited to less than 10% 11,12 .Most modern Li-ion batteries with metal oxide intercalation cathodes and graphite anodes can meet this requirement [13][14][15] .For Li cells, the lithium metal is repeatedly deposited on and stripped from the anode, and the morphology of the electrode becomes increasingly porous, causing significant cell swelling.The degree of cell swelling depends on the cell chemistry and cell parameters, typically ranging from 50% to more than 100% 7 for Li metal pouch cells.Even with a compatible electrolyte, the cell swelling is still between 20% to 50% 8 .Such a large cell swelling can be a key reason for cell failure and cannot be tolerated for practical applications.External pressure has been playing a critical role on the electrochemical processes of lithium metal 16,17 .External pressure is known to significantly enhance the cycling stability of all solid-state lithium metal cells and shape the morphology of the deposited lithium metal 18 .For lithium metal cells using liquid electrolyte, there have been several reports discussing the impact of external pressure on coin cells 19,20 , single-layer pouch cells 21,22 , or small anode-free pouch cells 16,23 .These investigations suggest that the external pressure can influence the structure of the solid electrolyte interphase (SEI) layers and is beneficial for achieving long cycle life.However, to date, there have been no reports to understand the role of pressure on practical, largeformat high-energy Li metal batteries, leaving key questions unanswered.Findings from small areas in coin cells can significantly differ from those in large-format electrodes, as the surface roughness of electroplated Li (and any other metal) becomes worse with increasing plating area 2 .For high-energy Li metal pouch cells, it remains unknown how the pressure stabilizes the cell performance and whether the pressure or electrolyte amount plays the dominant role in determining the cycle life.Achieving a high energy density pouch cell often involves using a lean electrolyte, raising questions about how external pressure affects both the electrolyte distribution and the utilization of lithium metal within pouch cells.This work utilizes three identical 350 Wh/kg lithium metal pouch cells to investigate the fundamental relationship between pressure and the electroplating process of lithium metal.Different external pressures are compared and correlated to the performance of these pouch cells.The long-term cycling cell swelling is reduced from more than 20-40% in the previous studies to less than 6-8% in this work, which is now comparable to that of state-of-the-art Li-ion batteries.A two-step utilization of Li stored in NMC and Li foil anode are discovered which explains the substantial decrease of cell swelling.Additionally, the pressure distribution on the pouch cell surface is mapped to understand the reaction activity across the entire cell surface during cycling, revealing a Li + detour phenomenon leading to preferred plating of Li in the central area of anode.New strategies are needed to extend the region of minimum swelling and mitigate the edge effect to further improve the cell performance.

Results and Discussion
There are two common designs for applying external pressures on Li-ion pouch cells 24 : constant gap or constant pressure.The former involves fixing a cell between two plates, which restricts outward expansion (Fig. S1a and 1d), while the latter typically adds constant spring force between two plates (Fig. S1b and 1e).The "brace" design of constant gap works effectively for Li-ion batteries due to the limited volume change of intercalation compounds used in Li-ion batteries.In the case of lithium metal cells, the volume change during charge is substantial, especially when multiple layers of Li foils are stacked within the same cell.As a result, the screws used to secure the two plates in the constant gap design get pushed up by the expanded cell, creating additional gaps between the plates during cycling.Therefore, more rigid enclosures are required to secure the lithium metal pouch cell 16 .In the constant pressure design, the addition of four springs (Fig. S1b) on top of the two brace plates helps maintain the pressure or force within a fixed range, depending on spring constant.By selecting appropriate springs, the deflection can be minimized, thereby maintaining an almost "constant" pressure on Li-ion pouch cells.However, due to the larger volume expansion of Li metal batteries compared to Li-ion batteries during cycling, the deflection of the spring cannot be minimized by solely using the "constant pressure" fixture depicted in Figure S1b for Li metal pouch cells.In this work, a modified testing fixture is designed (Fig. S1c and 1f) which combines both constant gap and constant pressure.It includes an additional modified bolt in the center of the "constant pressure" brace.This hybrid fixture allows the cell to expand and the spring to deflect while limiting its movement along the length of the bolt.As a result, a relatively consistent pressure is maintained during most of the cycles, potentially preventing significant volume increase at the end of the charge.Three identical Li metal pouch cells were prepared using the same cell parameters, each with a total cell-level energy of 350 Wh/kg (see Table S1, cell design A for details).The three key parameters affecting lithium metal battery cycling stability were kept consistent in all three pouch cells: (1) Li anode thickness of 50 µm on each side of the copper current collector, (2) LiNi0.6Mn0.2Co0.2O2(NMC622) cathode with a mass loading of 19.1 mg/cm 2 corresponding to 3.5 mAh/cm 2 of areal capacity on each side of the aluminum current collector, and (3) an electrolyte amount of 1.9 g/Ah in all cells.Without precise control of these three key parameters, it is challenging to develop a conclusive understanding of the impacts of pressure.In all pouch cells, the self-generated pressures (blue lines in Fig. 1a-1c) increase during charge due to the plating of additional Li metal layers on the anode side.It is important to note that in Li/NMC622 chemistry, Li + ions are stored in two different reservoirs, NMC cathode and the Li foil anode.Li from NMC cathode is used for plating during charge and stripping during discharge, while the original 50 µm Li foil initially functions as a lithium reservoir and current collector before the Li from the cathode is "depleted".The self-generated pressure decreases during discharge as Li is stripped and travels back to the cathode lattice sites.The magnitude of changes in self-generated pressure is inversely related to the externally applied pressure on the fresh pouch cells.When applying pressures of 16 and 26 psi, the peak pressures at the end of the first charge reach 140 and 120 psi, respectively (Fig. 1a and 1b).However, when applying 36 psi to the pouch cell, the highest self-generated pressure is only 70 psi at the end of the first charge (Fig. 1c).
Correspondingly, a relatively high initial pressure leads to the reduced amplitude of maximum pressure observed during each charge (Fig. 1d-1f).During cycling, the pressure at the end of charge increases almost every cycle (Fig. 1d-1f) in all three pouch cells, indicating continuous "thickening" of the anode.Surprisingly, the pressures at the end of each discharge (Fig. 1d-1f) show almost no change in the first 50-100 cycles (under 16 or 26 psi, Fig. 1d and 1e), or even 250 cycles (under 36 psi, Fig. 1f), and usually return to the originally applied value.This result suggests that in the first tens or hundreds of cycles (depending on the external pressure), there is barely cell thickening at the end of discharge.The only explanation for this observation is that during this period, most of the Li being utilized comes from the NMC cathode which are reversible between two electrodes.During the charge and discharge process, the Li from the cathode shuttles between the cathode and anode with minimum Li loss on the anode, similar as what happens in a traditional lithium-ion cell.The higher initial pressure reduces the irreversible loss of Li during each cycle because most of the Li (from NMC) can return to cathode, resulting in almost full recovery of the initial pressure observed at the end of discharge.During this period, the Li foil is almost intact and functions primarily as a current collector (SEM images).At the same time, very limited cell swelling happens.The discovery of "Li shuttling" phase is very important providing a new perspective to lithium metal batteries if such a stable phase can be further extended.For Li + ions stored in original Li foil side, they gradually become involved in the electrochemical and chemical reactions upon cycling.This is reflected in the continuous increase of end-of-charge pressure after each cycle due to anode thickening.Pressures at the end of charge and discharge steadily increase, particularly towards the end of cycling.Nevertheless, the amplitude of pressure increase from the beginning to the end is always lower in the pouch cell tested with higher externally applied pressures.It is worth noting that for liquid cells, applying a high external pressure does not necessarily translate to better performance.Higher external pressure also increases the probability of an internal short (Fig. S2) proportional to the defective sites in the cell, such as those near the tabbing area 25 , especially under elevated pressures.Notably, the highest pressure used in this work, 36 psi, is considerably lower than the range of 2 MPa to 250 MPa (290 psi to 36259 psi) used for solid-state lithium metal batteries [26][27][28] .It is also surprising that the cycling stability of the three pouch cells under different initial pressures does not exhibit significant differences.The cell under 16 psi experiences a capacity degradation to 80% of its original capacity after 282 cycles (Fig. 1g).The pouch cells under 26 and 36 psi reach 80% of their initial capacity after 299 (Fig. 1h) and 318 cycles (Fig. 1i), respectively.To understand this, it is necessary to consider the dominant factors influencing the cycling stability of Li metal batteries at different stages.As mentioned earlier, several parameters impact the cycling of Li metal cells, such as cathode loading, Li thickness, and electrolyte amount.From the current results, limited external pressure (again external pressure cannot be too high in order to avoid internal short) does not significantly extend the cycling performance of lithium metal batteries if the liquid electrolyte almost dries out.In liquid cells, the amount of available electrolyte dominantly determines how long a Li metal battery last.Once the electrolyte is depleted, the nearly "dried" cell does not respond effectively to this minimal external pressure, resulting in similar cycle life (Fig. 1g-1i) among all cells at different pressures.Another 350 Wh/kg pouch cell with an increased amount of electrolyte of 2.2 g/Ah (compared to 1.9 g/Ah in this work) demonstrated approximately 459 stable cycles (Fig. S3 and cell design B in Table S1), confirming that the amount of electrolyte plays a more critical role than external pressure in extending the cycle life of Li metal batteries utilizing liquid electrolyte.In our previous published work using a 350 Wh/kg pouch cell with 50 µm Li, which also contained 2.2 g/Ah electrolyte, 430 stable cycles were achieved 8 .Therefore, external pressure only aids in extending the cycling of Li metal when there is still enough liquid electrolyte available in the cell.The utilization of Li at different stages is summarized in Figure 2.There are two different Li reservoirs in Li/NMC pouch cells, one is in NMC lattice while the second is the original Li foil anode assembled in the cell.In the early stage of the charge-discharge process, Li + ions from the NMC cathode shuttle back and forth between the cathode and anode.There is very limited cell swelling at the end of the discharge state because most of Li + ions from NMC can reversibly go back to cathode (Fig. 2a).As cycling goes on, more Li + ions from NMC are irreversibly lost, Li stored in the foil needs to participate in the electrochemical reaction if there is still enough electrolyte (Fig. 2b).During the second stage of cell cycling, as Li + from the Li foil anode begins to participate in the reaction, the thickness of the entire anode gradually increases because Li foil becomes porous accompanied by dead Li and SEI accumulation.Once the electrolyte is depleted, a rapid capacity decay is typically observed (Fig. 2c).The hybrid constant gap-pressure design greatly reduces cell swelling in all three pouch cells after extensive cycling (Fig. 3).Overall, the pouch cell swelling after extensive cycling is reduced from 20-40% in the previous studies to less than 6-8%, comparable to that of state-of-the-art Li-ion batteries.Prior to cycling, the pouch cell had a measured thickness of 5.45 mm (Fig. 3a).Even under the lowest pressure of 16 psi, cell swelling is only 8.2% after 304 cycles (Fig. 3b).The other two pouch cells under 26 and 36 psi external pressures expand by only 6-7% after more than 300 cycles (Fig. 3c-3d).Appropriate external pressure, combined with a balanced cell design, effectively suppresses the aggressive volume expansion of the Li metal pouch cell during cycling.The cycled lithium metal anodes harvested from three pouch cells are further compared in Figure 3e-3g.The pouch cells were in a discharge state before disassembly, meaning the majority of Li from NMC side returns to the cathode.Figure 3e-3j mainly characterizes cycled Li foil.Intact dense Li is present in all three cycled pouch cells.For consistency, the characterized Li is taken from the same location in the three cells, as indicated in Figure 3k.Some unreacted Li forms column-like structures extending from the current collector to the surfaces of the Li anode, suggesting that the entire Li pillar did not undergo any reaction (Fig. 3e-3g).These unused Li columns are commonly found on the anode side in all three cycled pouch cells (see surface views in Fig. 3h-3j).But the populations of intact Li columns are different on the edges and in the center of the Li foil, which will be discussed in a later section (Fig. 4).It has been observed that after extensive cycling, the thickness of the cycled Li foil also increases, including the intact Li columns, which, in some cases, become even thicker than 50 µm (Fig. 3e and Fig. 3f).These unreacted Li columns are affected by the adjacent Li undergoing dense-to-porous conversion.Under external pressures, intact Li columns elongates beyond their original 50 µm length.After extensive cycling, the SEI and reacted porous Li are entangled forming porous structures in all cycled pouch cells (Fig. S4).The discussion above focuses on the overall response of the entire pouch cell to the external pressures applied vertically on the cell.The change of self-generated pressures are average values detected from the pouch cell upon cycling, which are created through changes of all the layers assembled in the battery.How the pressure is distributed across the surface of the pouch cells and its implications of large-scale electroplating of Li are still unknown.To investigate this question, a pressure mapping system (see Methods for experimental setup) was employed to monitor the pressure distribution on the surface of a 350 Wh/kg pouch cell (Fig. 4 and Fig. S5).After a resting period of 2 hours at open circuit voltage, the average pressure on the cell surfaces was measured to be 32.7 psi (Fig. 4a).At the end of the first charge (Fig. 4b), the average pressure on the cell surfaces increased to 81.4 psi due to the plating of Li (from NMC) on the anode side.The pressure distribution at the end of the first charge does not follow the same pattern at OCV, and more hotspots are observed in the center (Fig. 4b) which is further amplified at the end of the 5 th (Fig. 4c) and 20 th charge (Fig. 4d).At the end of 50 th (Fig. 4e) and 100 th (Fig. 4f) charge, the hotspots propagate to the rest regions of cell surfaces, suggesting the availability of more active areas within the cells.The pressure distribution at the end of discharge (Fig. 4g-4k) follows a similar trend, but the magnitude of pressure increase at the end of discharge is much lower compared to the end of charge, consistent with our earlier discussions.For example, at the end of the first discharge, when the plated Li (from NMC side) returns to the cathode, the average surface pressure reverts back to 34 psi, close to the OCV value of 32.7 psi (Fig. 4g).As cycling progresses, the average surface pressure at the end of discharge increases with small amplitude (Fig. 4g-4k), indicating the presence of residual materials after each cycle, such as entangled SEI and dead Li.The atomic ratio of different elements (C, N, F, O, S, Li, Ni, Mn, Co) and components of SEI on the Li anode at various locations (near-tab area, center, side, bottom) of electrode under different pressures (16 psi, 26 psi, 36 psi) do not exhibit significant differences, suggesting that external pressure, at least within the range discussed in this study, pressure does not impact the SEI formation process notably (Fig. S6-Fig.S7).Each time when reaching the end of the charge, it has been observed that the amplitude of pressure increase in the central part of the pouch cell is consistently higher than in the rest of the regions on the same surface.This indicates that more significant "thickening" occurs in the center of the anode.To understand this phenomenon, the pressure mapping on the cell surface was stopped after about 100 cycles for post-mortem analysis.A representative Li metal anode from the cycled pouch cell is shown in Figure 4l.In all harvested Li metal foils from the cell, the center part of the cycled lithium anode consistently appears shinier than the four edges (Fig. 4l).Unreacted dense Li columns (highlighted in Fig. 4m-4p) are generally observed in all four selected regions from the cycled Li (Fig. 4l), but the population of intact Li columns varies at different locations.The central part of the Li anode contains more intact Li (Fig. 4m) compared to the edges (Fig. 4n-4p), which aligns with the color difference observed (Fig. 4l).This suggests that less Li in the center area is involved in electrochemical reactions, thus maintaining a dense morphology and metallic shine better than the rest of the Li foil.This observation is common when examining both sides of the same cycled Li foil (Fig. S8) as well as different Li foil anodes from the same pouch cell (Fig. S9   and S10).Even after 300 cycles, harvested Li foils from the Li metal pouch cells in Figure 3 exhibit similar attributes to those seen in Figure 4l, with the center part being much brighter than the rest of the Li foils (Fig. S11), confirming that the consumption of Li in the center is slower than at the edges.The presence of more intact and shinier Li in the center part seems contradictory to the significant increase in pressure detected in the center of the pouch cell (Fig. 4e and 4f).Note that the accelerated pressure increase in the central part of pouch cell happens during charge meaning Li from NMC is covering anode surfaces.The pouch cells used for analysis, however, are all disassembled in discharged status with the majority of Li from NMC already traveling back to cathode.The less utilization of center Li foil indicates that the majority of Li involved in the electrochemical reactions occurring in the central part of cell is from NMC side.Li originally "stored" on the edges of the Li foil participates more in the cycling compared with the central Li.Li + detour mechanism is proposed here to explain the observed phenomenon and illustrated in Figure 5a.During charge, Li (from NMC) is not homogeneously plated on the Li foil anode side.Instead, Li + ions (from NMC) preferably plate in the center of Li foil anode leading to the much faster increase of self-generated pressure in the central area of the pouch cell.Ideally, the electroplating of Li (from NMC) should follow the ideal process in Figure 2.However, impacted by external pressure, some of the Li + ions that are supposed to deposit on the edges now detour to the center of Li anode and are plated there, leading to a thicker deposition layer in the central area of anode and thus higher pressure from that region.During the subsequent discharge process, more Li on the edges of Li foil anode participates in the electrochemical reaction due to less coverage by electroplated Li layers compared to the central Li foil that is better "protected" by the Li deposits from the NMC side.After repeated cycling, the utilization and darkening of Li occur earlier on the edges than in the central region of the original Li foil.To prove the Li detour hypothesis, a single layer pouch cell is charged and disassembled at the charge status after 3 cycles (Fig. 5b and Fig. S12a) and 100 cycles (Fig. 5c and Fig. S12b).In contrast to those cells disassembled in discharge status, this time the central area is much darker than the edges of Li anode indicating more nano particles of Li are plated in the center anode during charge.In another parallel 350 Wh/kg Li metal pouch cell, cycling was continued for over 300 cycles (Fig. S13).Significant pressure increase was observed across the entire cell surface, still with the highest pressure concentrated in the central area.This phenomenon is similar to the pressure mapping results observed in Si-based pouch cells 29 , although a different pressure fixture was used.Towards the end of the cycling, the pressure distribution on the surface of the pouch cell stabilized with minimal changes.The pressure mapping images at the 300 th and 309 th cycles (inset of Fig. S13) displayed almost identical pressure distribution and amplitudes, confirming that the electrolyte had almost completely dried out after 300 cycles.Limited external pressure does not extend the cycling of the "dried" cell towards the end of cycling.

Ab Initio Electric Fields:
To gain a deeper understanding of the Li + detour phenomenon during electroplating, we have calculated with ab initio electronic structure methods based on density functional theory, the electrostatic potentials (EP) and electric fields (EF) induced on the neighborhood of the surface of Li metal tablets representing structures at different applied pressures, and thus having different concentrations of Li-nuclei (supplement).We also calculated the EPs and EFs on a larger tablet containing two different nuclei concentrations: a high one around the center of the tablet and a lower one around the peripheral of the tablet (Fig. 5d).The specific movement of chemical species in any type of material (containing nuclei and electrons) can be studied using the electric field (E), i.e., the force per unit of charge, created internally by the same system components in addition to the internal fields.The electric field, mainly, tells the magnitude and direction of the displacements of charged moieties (ions and counterions), as well as the rotational movement of neutral dipolar molecules (such as solvents and additives) to align their electric dipole in the direction of the electric field.To directly demonstrate the effects of a non-uniform pressure, we analyzed a slab having two different concentrations of Li nuclei (Fig. 5d).The full slab has an area of 3L×3L, with 3L = 2.6 nm.The area of the highest concentration site is a square of L×L sitting in the center of the 3Lx3L surface.The remainder of the surface has a lower concentration of Li nuclei.Figure 5d shows the lines of electric field (blue) originated by the Li-metal surface.They only represent the direction of the trajectories.The values of the electric field are indicated with an orange background in the most representative places approaching the anode and the electric potential of the black isopotential contours appear with a green background.The low Li concentration at the left-and rightside surfaces show a repulsive behavior against Li-ions and the central one at high pressure features a much higher attraction, even attracting any Li-ion positioned above the low-pressure sites.We conclude that the driving forces producing the observed trajectory deflection near the anode surface could be determined or controlled by the pressure applied to the pouch cell, increasing the concentration of Li nuclei in the anode, and triggering further electric field changes that drive more Li + ions near the anode to the high-concentration or high-pressure zones of the anode.These induced electric fields in the vicinity of the anode arise from the quantum mechanical effects of the charges carried by the electrons and nuclei of the anode.Therefore, ab initio results provide valuable insights into utilizing external pressure to facilitate preferred deposition in specific anode regions, and such information may be used to optimize the Li deposition and stripping processes.

Conclusions
This study investigates the electrochemical plating of Li + on a large scale in realistic 350 Wh/kg Li metal pouch cells under varying pressures.It has been discovered that a higher external pressure helps to enhance the cycling and significantly minimize the cell swelling after extensive cycling.However, the critical role that external pressure plays in extending cell cycle life is only observable when there is still enough electrolyte present in the battery.Mapping the pressure distribution across the surface of the pouch cell during cycling reveals a preferential plating of Li + in the central area of the anode side.This intriguing phenomenon of Li + detour, influenced by pressure, is explained through precise quantum theory calculations and experimentally validated in a customdesigned Li-free anode-limiting pouch cell, where the cathode area is larger than the anode.These findings shed light on the distinct behaviors of Li + during large-scale electrochemical plating driven by pressure and provide valuable insights for developing effective solutions to address challenges in rechargeable Li metal batteries.

Electrodes/electrolyte preparation
The slurry for the NMC622 cathode (Targray, Canada was prepared by combining 96 wt.% NMC622, 2 wt.% conductive carbon (Super P C65, Timcal), and 2 wt.% polyvinylidene difluoride (PVDF) binder (L1120 from Kureha, Korea) in N-methyl-2-pyrrolidone (NMP) solution.The mixture was thoroughly mixed using a Thinky mixer in a dry room maintained at a constant temperature of 19 °C and a relative humidity of 0.1% (dew point controlled below -55 °C).Additional NMP (Targray, Canada) was added to adjust the solid content to 51 wt%.The slurry was then coated onto both sides of a 10-μm-thick Al foil using a comma coating machine (Mediatech, Korea) to achieve a controlled areal coating weight of 19-20 mg/cm 2 (~3.5 mAh/cm 2 ).or single layer pouch cells and the outer electrodes of multiple layer pouch cells, a single-side coated NMC622 electrode was prepared.The coated cathode was calendared to 2.8 g/cm 3 , punched into rectangular pieces, and dried in a vacuum oven at 60 °C for 24 hours.After drying, the cathode electrode was cut to dimensions of 36.0 mm wide and 54.0 mm long, and any residual NMC622 materials on the Al tab were removed with NMP prior to cell assembly.Free-standing Li foil (50 μm, Li content ≥99.9%) was obtained from China Energy Lithium and used directly in the dry room.The Li foil was laminated onto both sides of a Cu mesh current collector (MTI, USA) in the dry room, and then punched into rectangular pieces.The Li anode electrode was cut to dimensions of 37.5 mm wide and 55.5 mm long prior to cell assembly.Battery-grade lithium bis(fluorosulfonyl)imide (LiFSI) (Nippon Shokubai, Japan) was dried at 120 °C under vacuum for 24 hours before use.1,1,2,2-tetrafluoroethyl-2,2,3,3-tetrafluoropropyl ether (TTE, 99%, SynQuest Laboratories, USA) was dried with 4Å molecular sieves prior to use.LiFSI and 1,2dimethoxyethane (DME, Gotion, Inc. China) were mixed together and then diluted with TTE to create a 1.5 M LiFSI-DME-TTE electrolyte solution (with a DME/TTE molar ratio of 1.2:3) in an argon-filled glove box with oxygen and moisture levels below 0.1 ppm.

Pouch cell assembly
All pouch cells were assembled in a dry room with a semi-automated cell manufacturing line (MediaTech, Korea) at the Advanced Battery Facility Lab of the Pacific Northwest National Laboratory (PNNL).The assembly process involved slurry mixing, comma coating, calendaring, Z-stacking of cathode, anode, and separator, grid trimming, ultrasonic welding for connection with Al (cathode) and Ni (anode) external tabs, packaging foil formation, top and side sealing, and vacuum sealing with electrolyte injection.As seen in the cell design in Table S1, to achieve 350 Wh/kg energy density, the areal capacity of NMC622 cathode was controlled to 3.5 mAh/cm 2 on each side of the Al foil.Each pouch cell consisted of 16 layers of double-sided anodes, 15 double-side coated and two single-side coated NMC622 cathode layers.With the 50 μm Li foil, the N/P ratio or cell balance was 3.0:1, and the E/C ratio was 1.9 g/Ah.The total capacity of each pouch cell exceeded 2.1 Ah.For Li/NMC622 single-layer pouch (SLP) cells, the cathode and anode electrodes were similar as described above for the multi-layer pouch cells.The capacity of the SLP cell was about 160 mAh.The cell configuration (Fig. S12c) comprised of a layer of double-sided Li anode sandwiched with two layers of single side coated NMC622 cathode.Each SLP cell was filled with 1 g of electrolyte.
After resting for 24 hours, the SLP cells were clamped using a compression device (as shown in Fig. S1b) with an initial pressure of 36 psi.The initial pressure was measured using a digital force gauge with a remote load cell (SHIMPO FG-7000L, USA).The testing procedure for the SLP cells was same as that for the multiple-layer pouch cells.

In situ pressure test
For the in-situ pressure test, multiple-layer pouch cells were sandwiched in a stainless-steel clamping device (shown in Fig. S1c) with a target initial external pressure.Two silicon mats were placed between the pouch cell and the stainless-steel plates of the clamp to ensure uniform pressure distribution.The pouch cell, along with the in-situ pressure measurement system (Loadstar Sensors, Model: DI-1000UHS, USA), was set in place, and the initial pressure of the cell was adjusted to the target force and fixed with screws.The applied initial force was calculated based on the target pressure and the battery area.For example, when the targeted initial pressure was 16 psi, so the applied initial force was 16 psi * 3.72 inch 2 (battery area: 4 cm * 6 cm=24 cm 2 =3.72 inch 2 ) = 59.6 lb.The clamped Li metal pouch cells were placed in a gas-detecting safety chamber (Cincinnati Sub-Zero, USA) filled with inert gas (N2) at 25 °C.Galvanostatic cycling tests were conducted within a voltage range of 2.7 V to 4.4 V using Land battery testers (LANHE CT2001B, China).The pouch cells were first charged/discharged at a constant current rate of 0.1 C for two initial formation cycles, followed by charging at 0.1 C and discharging at 0.3 C in subsequent cycles (1 C corresponds to 3.5 mA/cm² or 2.1 A).The in-situ pressure measurement system recorded the force every 10 seconds, and the average pressure was calculated as the average value of all the pressure data points in a cycle.

Pressure mapping test.
For the pressure mapping test, the pouch cell was sandwiched with a clamping device similar to Fig. S1b.Two silicon mats were placed between the pouch cell and the stainless-steel plates of the clamp for uniform pressure distribution.A force mapping sensor (Tekscan, Model 5076, USA) was positioned between the bottom silicon mat and the stainless-steel plate, in contact with the pouch cell.Pressure was applied to the pouch cell until the target pressure of 36 psi or 133.9 lb force was reached.The testing protocol for the pouch cell was the same as the in-situ pressure test.There were slight differences in measurement accuracy between the force mapping sensor and the in-situ pressure measurement system.To make the results from the force mapping sensor relevant to the in-situ pressure measurement system, a calibration was performed.The Tekscan force mapping sensor was calibrated using the Loadstar in-situ pressure measurement system and the SHIMPO digital force gauge.A pressure of 100 psi (or 372 lb force) was applied to the pouch cell, and the average pressure measured by the Tekscan force mapping sensor was recorded.The calibration factor was calculated by dividing the applied pressure (100 psi or 372 lb) by the recorded average pressure from the Tekscan mapping sensor.The average pressure of the pouch cell at a certain voltage during the cycling test was obtained by multiplying the calibration factor by the average pressure provided by the Tekscan force mapping sensor.

SEM/(S)TEM characterization
The surface and cross-section morphology of the cycled Li anodes were characterized using a Thermo Fisher Helios SEM.The cross-sections of the cycled Li anodes were obtained by cutting the anodes with a razor blade in an Ar-filled glove box.The cycled Li anode from the 2 Ah pouch cell at 16 psi was characterized using a 300 kV FEI Titan monochromated (S)TEM equipped with a probe aberration corrector.For the Li metal sample, a Cryo-holder was used to transfer the Li sample to the (S)TEM.XPS Experimental XPS measurements were performed using a Thermo Fisher NEXSA system.The system utilized a focused monochromatic Al Kα (1468.7 eV) source for excitation and a double-focusing hemispherical analyzer with a multi-element input lens and a 128 channel detector.The X-ray beam was incident to the sample at a normal angle, while the photoelectron detector was positioned at a 60° angle relative to the sample normal.High-energy resolution spectra were collected with a pass-energy of 50 eV, a step size of 0.1 eV, and a dwell time of 50 ms.The fullwidth-at-half-maximum (FWHM) for the Cu 2p3/2 peak was measured to be 0.82 eV under the same conditions used for the collection of the narrow scan spectra.

Ab Initio calculation
The contribution of electrons to generate electric fields can be obtained solving the Schrödinger Equation for a system of n electrons located at points ri ( (𝑛) ) and N nuclei located at points Ri ( () ), thus,  ̂( (𝑛) ,  (𝑁) )Ψ( (𝑛) ,  () ) = ( (𝑁) )Ψ( (𝑛) ,  (𝑁) ) (1) In this situation, the array of nuclei positions  (𝑁) is directly determined by the pressure applied to the pouch cell, the more pressure the more concentrated the nuclei (Fig. S14).Solving the above equation with an exact Hamiltonian operator ( ̂) using ab initio methods such as density functional theory (preferred) or a traditional ab initio (such as HF, MP, CC, CI, etc.), we obtain from this eigenvalue problem the wave function Ψ and the energy .The wavefunction Ψ yields a Slater determinant that can be written as Ψ = |  ( 1 )   ( 2 )   ( 3 ) ⋯   (  )| (2) Where the   ( j ) are the crystal or molecular orbitals.In this shorthand notation of the determinant, the rows of the determinant are created running the dummy index i from 1 to n.And the wavefunction is a function of 3n variables.For the sake of simplicity, we use atomic units and ignore the n spin variables and normalization constants, but they are considered in the calculations.Thus, the electron density (  ) at any point  can be obtained from summing all occupied orbitals,   () = − ∑ |  ()| 2  (3) And the corresponding nuclei contribution, at points  i with charges  i , using the Dirac -function can be written as   () = ∑ ( −   i ) i (4) These two densities give us the electrical potential at any point r, V(r), Finally, the electric field E is obtained from the gradient () of the potential,

E(𝑟) = −𝛁𝑉(𝑟)
For a surface under a non-uniform distribution of local pressures, for example a surface with only two of the pressures/concentrations shown in Figure S14, the net electric field in this non-uniform case can be estimated as the vector sum of their corresponding electric fields, since the interactions have been calculated between elementary particles (nuclei and electrons) so their pair interactions are independent of any other particles around.This electric field is obtained through the formalism explained in Equations 1-5 when applied to the nuclei and electrons of the anodes of different densities that can be associated to different applied pressures.Obviously, for this case of having two different pressures on the same surface, the site with the highest pressure will have the highest concentration of lines of force (E).This reasoning can be extended to a surface with any number of different pressures on its surface.We remark that making a larger anode with a given pressure simply adds the electric fields at the center of the anode surface, regardless of the size of the area, and in all cases the electric fields at the lateral boundaries of the anode surface will be mostly repulsive or of reduced attraction with respect to those in the center of the anode.Ideally, the Li foil on the anode is not involved much in the reaction in this early stage.As cycling goes on, some of Li from NMC cathode will be lost irreversibly due to forming "dead" Li.Correspondingly, the Li foil begins to participate in the reaction to compensate for the Li loss and more SEI and dead Li are accumulated.In the final stage, the electrolyte is depleted and the cell capacity falls rapidly.

Figure 1 . 4 V. Figure 2 .
Figure 1.In situ pressure monitoring of three 350 Wh/Kg Li/NMC622 pouch cells (2Ah) tested under different initial external pressure: (a, d, g) 16 psi, (b, e, f) 26 psi and (c, f, i) 36 psi.(a-c) show the evolution of self-generated pressures during charge/discharge for the first five cycles.(d-f) compare the pressures detected at the end of charge/discharge for each cycle, highlighting the differences between the pouch cells with varying initial pressures.The average pressure is also plotted in (d-f) to compare the amplitude of pressure change generated by the pouch cells during cycling.(g-i) represent the cycling stability of the three Li metal pouch cells tested under different initial pressures: (g) 16 psi, (h) 26 psi and (i) 36 psi.All cells were charged at a rate of 0.1 C and discharged at a rate of 0.3 C between 2.7 V and 4.4 V.

Figure 3 .
Figure 3. Li metal pouch cells before and after extensive cycling under different initial pressures.(a) As-prepared 350 Wh/kg Li metal pouch cell with a thickness of approximately 5.45 mm.(b-d) Pictures of pouch cells tested under initial pressures of (b) 16 psi after 304 cycles, (c) 26 psi after 313 cycles and (d) 36 psi after 335 cycles.The swelling rate of all three pouch cells is between 6-8% after extensive cycling.(e-g)show cross-section and surface view SEM images of cycled Li metal anodes harvested from pouch cells tested at initial pressures of (e-h) 16 psi, (f-i) 26 psi and (g-j) 36 psi.For each pouch cell, the 5 th Li anode assembled in the cell was taken out for further SEM characterization.On each harvested Li anode, a small piece of cycled Li was cut from the same location on the long edge near the copper tab, as indicated in (k).

Figure 4 .
Figure 4. Mapping of pressure distribution on the surface of a prototype 350 Wh/Kg Li/NMC622 pouch cell (2Ah) during cycling and correlation to lithium metal morphologies after cycling.(a) The average pressure on the surface of the pouch cell is 32.7 psi at OCV.The pressure applied on the center is slightly higher than the rest area of the pouch cell from the mapping system.Pressure increases continuously on the cell surface after the (b) 1 st , (c) 5 th , (d) 20 th , (e) 50 th and (f) 100 th charge.The central part of the pouch cell demonstrates a faster increase in pressure compared to the rest of the surface.Pressure distribution after the (g) 1 st , (h) 5 th , (i) 20 th , (j) 50 th and (k) 100 th discharge is also shown.The amplitude of pressure increase on the pouch cell surface is much lower at the end of discharge compared to the charge status.(l) is a photo of one of the cycled Li metal anodes harvested from the pouch cell used for pressure mapping.The central part of the cycled Li metal anode maintains its metallic shine and is further analyzed using (m) SEM from cross-section and top views.Additional three locations from the edges are also selected for SEM characterizations.(n) displays SEM images of Li located on the short edges near copper tab.(o) represents the cross-sectional and top view SEM images of Li located in the corner of the cycled Li foil anode, far away from the tabbing.(p) shows SEM images of Li located in the long edge of the same Li metal anode as indicated in (l).The pouch cell is in discharge status before being disassembled to harvest the cycled Li metal anode.

Figure 5 .
Figure 5. Experimental and theoretical study of Li + detour behavior during the electroplating process.(a) Visualization of Li + detour behavior in Li/NMC pouch cell driven by the uneven distribution of externally applied pressure.(b-c) Digital photos of Li anodes harvested from Li metal single-layer pouch