Exploring the Impact of Mechanical Pressure on the Performance of Anode-Free Lithium Metal Cells

Machine-made anode-free pouch cells with NMC532 positive electrodes (i.e. NMC532 vs bare copper foil) with a rolled electrode design were used in this work. Figure 1a shows an anode-free pouch cell (left) compared to a typical Li-ion cell (NMC532/graphite, right). The NMC532/graphite Li-ion cell is a 402035-size (40 mm long, 20 mm wide x 3.5 mm thick) pouch cell. The anode-free cell utilizes the exact same positive electrode and thus has the same capacity as the Li-ion cell of 230 mAh. The absence of negative electrode coating in the anode-free cell results in a total stack thickness of 2.0 mm, and therefore a 402020-size cell. Figure 1b shows single cell stack schematics for an anode-free cell in the discharged state (left), an anode-free cell in the charged state with plated lithium (middle), and a Li-ion cell (right). The stacks are built up of two layers of separator (13.5 μm each, white), an aluminum current collector (15 μm, gray), a double sided positive electrode coating (60 μm each side, black), and a copper current collector (8 μm, orange). Additionally, the anode-free cell in the charged state has lithium plated (18 μm each side, silver) on copper foil, and the Li-ion cell has a graphite negative electrode (80 μm each side, dark gray) coated on the copper foil. Summing up the components of each stack results in stack thicknesses of 170 μm for the discharged anode-free cell, 205 μm for the charged anode-free cell, and 330 μm for the Li-ion cell. The anode-free cell stack is significantly thinner than the Li-ion cell stack, even in the charged state when most voluminous. This, along with the 0.1 V increased average voltage during operation of an anode-free cell results in a theoretical energy density approximately 60% larger than a comparable Li-ion cell (1220 vs 725 Wh/L) as shown in Figure 1b.

Zoom In Zoom Out Reset image size

Pouch cells were received from a reputable manufacturer (Li-Fun Technology Co.) sealed without electrolyte. Cells were opened, dried under vacuum at 100°C for 14 hours, and transferred into an argon-filled glove box for electrolyte filling and resealing. Reagents used for electrolytes include fluoroethylene carbonate (FEC, BASF, purity 99.4%), diethyl carbonate (DEC, BASF, purity >99%), and bis(2,2,2-trifluoroethyl) carbonate (TFEC, HSC Corporation (China), >99%). Two electrolyte solvent systems were used in this work, FEC:DEC 1:2 v:v and FEC:TFEC 1:2 v:v. All electrolyte contained 1.0 M lithium hexafluorophosphate (LiPF₆, BASF, purity >99.9%). Cells were filled with 0.5 mL of electrolyte, corresponding to ca. 0.65 g for the FEC:DEC electrolyte and 0.75 g for the FEC:TFEC electrolyte, resulting in practical electrolyte loadings of ∼3 g/Ah. After filling, cells were held at 1.5 V for 24 hours to allow the electrolyte to fully wet the positive electrode and separator before testing.

To maintain a uniform lithium morphology for stable cycling, it is known that plating (charging) slow and stripping (discharging) fast is beneficial.^16,44 Additionally, cycling with a low areal capacity further aids in maintaining a uniform morphology. However, current densities and areal loadings must still be rationally chosen to yield a useful cell. In this work, we cycle at a charging current density of 0.6 mA/cm² and a discharging current density of 1.5 mA/cm², corresponding to C/5 and D/2 (charge and discharge in 5 and 2 hours, respectively), and operate at an areal capacity of ∼3 mAh/cm². Cells were cycled between 3.6−4.5 V. The NMC532 positive electrode has a ca. 30 mAh first cycle irreversible capacity loss only recoverable below 1.5 V. Therefore, by operating with a 3.6 V lower cut off voltage, a 30 mAh lithium reservoir is formed.¹² This is analogous to operating with a limited lithium metal excess, without the drawbacks of constructing cells with lithium metal since the excess is formed in-situ. Additionally, 30 mAh corresponds to only 15% excess and a ∼2 μm thick lithium excess layer, thus not appreciably reducing the energy density like cell designs with significant excess lithium. As a result, only during the first charge is lithium plating on to bare copper; all subsequent charges begin plating on the ∼2 μm lithium layer formed on the copper current collector due to the irreversible capacity of the positive electrode. Therefore, out of the total 18 μm plated from the positive electrode, 16 μm are cycled reversibly. All tests were performed at 40°C.

Figure 2 shows the apparatus used to constrain pouch cells to different pressures as well as to record operando pressure data, as first described by Louli et al.⁴⁵ Cells were constrained in a rigid enclosure, colloquially referred to as a "superboat" (SB), with a load cell (model LCKD, Omega Engineering), a disc spring, and two force distributing plates dividing the pouch cell, spring, and load cell, as pictured in Figure 2a and schematically shown in Figure 2b. The superboat enclosure has one adjustable wall which can be tightened with screws to exert different uniaxial pressures to the cell stack. Pouch cells were fastened within this enclosure such that any change of volume of the cell stack caused a change of force within the enclosure measured by the load cell. The pressure on the pouch cell was thus calculated as the force recorded by the load cell divided by the area of the pouch cell electrode stack, 6.0 cm² (0.93 in²). The springs were inserted to dampen the change in pressure during cycling. Any gas evolved in the pouch cell during cycling is pushed in to the gas bag of the pouch cell and will not result in a change in pressure measured by the load cell; only uniaxial expansion of the cell stack is probed with this method.

Zoom In Zoom Out Reset image size

The load cells were connected to DP25B-S-A panel meters (Omega Engineering). Operando measurements were performed using an E-One Moli Energy Canada battery testing system. Pouch cells were cycled on Moli channels and the analog 0–10 V output of the panel meter was connected to an adjacent Moli "slave" channel, allowing for simultaneous electrochemical and pressure measurements.

Cells were constrained at the bottom of the first discharge (BOD) with different initial pressures. Tests were performed in the superboat (SB) enclosures shown in Figure 2 at initial pressure loadings of 75 kPa, 375 kPa, 745 kPa, 1260 kPa and 1855 kPa. It is important to note that during cycling this pressure does not remain fixed; as lithium is plated and the cell stack expands, the pressure within the superboat increases until some maximum pressure is reached at the top of charge (TOC). Therefore, an average pressure can be calculated for each pressure loading, listed in Table I. Additionally, the average pressure will evolve as a function of cycle number due to irreversible cell stack expansion concomitant with cycling.⁴⁶ Control tests were also performed at a low pressure loading in plastic enclosures referred to as "normal boats" (NB), in which cells were clamped with rubber blocks. Load cells were not used for these low pressure tests, so operando pressure data was not collected. The pressure within a normal boat is known to be <75 kPa and remains relatively constant throughout testing due to the compliancy of the rubber blocks and plastic enclosure. In total, experiments were performed with initial average pressures of <75 kPa, 485 kPa, 795 kPa, 1205 kPa, 1725 kPa, and 2205 kPa. Hereafter, each pressure experiment will be referred to by these initial average pressures.

In-situ volume measurements were made to validate that operando pressure measurements correlate to the volume expansion of the electrode stack. These measurements were made using the apparatus developed by Aiken et al.⁴⁷ Archimedes' principle was used to infer the change of volume of a pouch cell submerged in fluid by measuring the change in weight of the cell with a thin film load cell. The measured weight changes as the buoyant force acting on the cell changes concomitant with volume expansion. For in-situ measurements, the thin film load cell output was connected to a Keithley 2700 scanning voltmeter and the pouch cells were connected to a Neware BTS 5V-0.1A battery test system (Neware, Neware Technology Limited, 207 Meihua Road, Xiameilin, Shenzhen, China). A minimal pressure of <75 kPa was applied to the pouch cells during in-situ volume experiments by way of constraining clip.

To deconvolute the effect of gas evolution and irreversible electrode expansion on in-situ volume measurements, 0.2 cm³ open-ended tubes were inserted into the gas bag of the tested pouch cells, analogous to the procedure outlined by Krause et al.⁴⁸ With the inserted tubes, gas generation will not result in a measured volume expansion until ca. 0.2 cm³ has been produced and the tube is filled. This was verified with measurements performed with and without an inserted tube, shown in SI Figure S1. Figure S1 shows that gas evolution begins to effect in-situ volume measurements of pouch cells with a 0.2 cm³ inserted tube after 3 cycles.

SEM images were taken to investigate lithium plating morphology. Samples were retrieved from cells at the top of charge (4.5 V) after one charge and after 50 cycles. Cells were cut open inside an argon filled glove box and the negative electrode copper current collector with plated lithium was carefully removed from the cell stack. Samples of ca. 0.75 cm² were cut and prepared on SEM stubs. Samples were then sealed in plastic bags and transferred out of the glove box to a Nano Science Phenom Pro G2 Desktop Scanning Electron Microscope with a backscattered electron detector. During this process, samples were exposed to air for <30 seconds when transferred from the plastic bag in to the SEM. The images of samples were taken with accelerating voltage of 5 kV and current of 0.6 nA.

Figure 3 shows the reversible volume expansion of anode-free cells with FEC:DEC electrolyte due to lithium plating and stripping during charge and discharge. Figure 3a shows voltage vs time for cells that underwent in-situ volume (red) and operando pressure experiments (blue). Figure 3b shows the corresponding in-situ volume data (red, left axis) and operando pressure data (blue, right axis). Cycles two and three are shown here. During charge, lithium ions are de-intercalated from the positive electrode and plated on the copper current collector, thickening the electrode stack resulting in a volume expansion and increase of pressure in these experiments. Cells were cycled with a constant current, thus the thickness of the plated lithium increased linearly. The volume expansion due to lithium plating after one charge is shown to be ca. 0.15 cm³. These cells have a cross sectional area of ca. 6 cm², therefore this volume expansion corresponds to a thickness increase of ∼250 μm. The expected thickness change can be calculated by determining the amount of lithium stored in the positive electrode that is plated during charge. The NMC532 positive electrode has a loading of 21.1 mg/cm² at 96% active loading. The positive electrode capacity is 180 mAh/g over the voltage range cycled here, and the irreversible capacity of the positive electrode is ca. 10%. Thus, the areal capacity of the positive electrode can be calculated:

• (1) 21.1 mg/cm^{2 *} 0.96 ^* 180 mAh/g ^* (1-0.1) = 3.28 mAh/cm².

Using Faraday's number, the molar mass and density of Li, the thickness of 1 mAh of plated lithium can be calculated:

• (2) 26801 mAh/mol/6.941 g/mol ^* 0.534 g/cm^{3 *} 1 cm² = 2062 mAh/cm
• (3) 1 mAh/2062 mAh/cm = 4.85 μm.

One layer of lithium plated from the positive electrode is thus:

• (4) 3.28 mAh ^* 4.85 um = 15.9 μm.

These cells feature a rolled electrode design and have 16 layers of positive electrode. Therefore, the corresponding thickness of 16 layers of plated lithium is:

• (5) 15.9 μm ^* 16 = 254.4 μm

which in good agreement with the in-situ volume results.

Zoom In Zoom Out Reset image size

The first plating/stripping cycle shown in Figure 3b exhibits excellent agreement between the in-situ volume and operando pressure results. Negligible irreversible expansion is observed with either measurement. The second cycle shown in Figure 3b also exhibits no irreversible expansion for the operando pressure measurements; however, there is an irreversible expansion observed by the in-situ volume measurements. This is caused by gas evolution which is detected with the in-situ volume method but not with the operando pressure method. SI Figure S1 demonstrates the significant gassing these cells experience during extended cycling. The agreement of in-situ volume and operando pressure measurements during the first cycle shown in Figure 3 is attributed to reversible volume expansion caused by lithium plating and stripping during charge and discharge. Figure 3 demonstrates that operando pressure measurements correlate well to the cell stack volume and thus electrode thickness evolution.

Figure 4 shows the pressure evolution and capacity retention of FEC:DEC (a,b) and FEC:TFEC cells (c,d) constrained with initial average pressures of 485 kPa (a,c) and 795 kPa (b,d). The operando pressure vs time data (red, left axes) over 100 cycles is shown for each testing condition. The pressure at the top of charge (TOC) is shown with open triangles, the pressure at the bottom of discharge (BOD) is shown with upside-down triangles, and the average pressure is shown with diamonds. The capacity and pressure swing (the difference between the pressure at the top of charge and the pressure at the bottom of discharge) normalized to the third cycle are shown in dark and light blue, respectively (right axes). Figure 4 shows that there is an irreversible cell stack expansion that evolves over extended cycling. Previous works which have measured irreversible expansion in Li-metal cells^40,42 and Li-ion cells^46,49 have attributed this to SEI growth, and in the case of Li-metal cells, dendritic, porous lithium morphology evolution and the formation of dead-lithium. Therefore, increased irreversible expansion is an indicator of poorer cell health. The reversible volume expansion caused by lithium plating and stripping is quantified by the pressure swing. As lithium inventory is lost, less lithium will be plated and thus the pressure swing will decrease concomitant to capacity loss. Figure 4 shows a good correlation between the normalized pressure swing and normalized capacity. Worse performance and cell health, indicated by pressure growth, capacity loss and decreasing pressure swing, is shown to occur at lower pressure (left panels); furthermore, cells with FEC:DEC electrolyte (top panels) exhibit worse performance compared to FEC:TFEC electrolyte (bottom panels).

Zoom In Zoom Out Reset image size

Figure 5 shows cycling data for cells containing FEC:DEC electrolyte (a,c) and FEC:TFEC electrolyte (b,d). Normalized capacity vs cycle no. is shown in the top panels (a,b) and polarization growth (ΔV—defined to be the difference between the average charge voltage and the average discharge voltage) is shown in the bottom panels (c,d). Tests at different pressures are denoted by the symbols as indicated by the legend; repeat tests are plotted with open symbols. Figures 5a and 5b demonstrate the impact of applied pressure between 75–2205 kPa on capacity retention. The most significant gain is achieved by constraining cells in superboats at an average pressure of 485 kPa compared to the low pressure tests in normal boats constrained at <75 kPa. Further increasing the initial average pressure >485 kPa benefits FEC:DEC cells more significantly than cells containing FEC:TFEC electrolyte. Capacity retention over 100 cycles is improved for FEC:DEC cells by increasing the initial average pressure up to 1725 kPa, whereas the benefit for FEC:TFEC cells saturates at a lower pressure of 795 kPa. Some high pressure tests were stopped prematurely due to shunting, e.g. the FEC:DEC 2205 kPa test. Additionally, some of the high pressure repeat tests, not shown in Figure 5, exhibited sudden capacity drops during cycling. This behavior is shown in SI Figure S2. We believe this occurs due to current collector tearing and positive electrode delamination as a result of the high stress exerted on the electrodes at high pressure. The large volume change of lithium metal during charge and discharge results in a large pressure swing exerted on the cell components which can be physically harmful. Such adverse effects must be considered when implementing Li-metal cells.

Zoom In Zoom Out Reset image size

Figures 5a and 5b show that cells initially cycle with little capacity loss and then subsequently begin to lose capacity at an increased rate. The number of cycles achieved before this rapid capacity loss is affected by the initial average pressure as well as the electrolyte. This behavior was identified by Genovese et al.¹² to be caused by the irreversible capacity of the NMC532 positive electrode. Under normal cycling conditions between 3.6−4.5 V, there is a first cycle irreversible capacity of approximately 30 mAh. As such, 30 mAh corresponding to a ca. 2 μm layer of plated lithium remains on the copper current collector when the cell is discharged to 3.6 V, effectively creating a lithium reservoir. Therefore, significant capacity loss is not observed until this lithium reservoir is depleted; 30 mAh of lithium inventory must be depleted to processes such as to SEI formation and accumulation of dead-lithium before the cells begin to exhibit failure. Extended periods of initial minimal capacity loss thus implies a better coulombic efficiency. This effect is somewhat analogous to using limited excess lithium but is without the drawbacks of requiring the manufacturing and handling of lithium metal during cell construction. Instead, excess lithium is achieved in-situ.

Figures 5c and 5d show the effect of pressure on polarization (ΔV) growth. Increasing the average pressure has negligible effect on cells containing FEC:DEC electrolyte; for each pressure, the polarization growth was ca. 50 mV over 100 cycles, except in the low pressure normal boat test which experienced rapid failure over 20 cycles. In contrast, applied pressure had a significant impact on the polarization growth for FEC:TFEC cells. The initial ΔV of FEC:TFEC cells was ca. 30 mV higher than FEC:DEC cells. In addition, the polarization growth appears non-linear for FEC:TFEC cells. For cells constrained up to an initial average pressure of 1205 kPa, the polarization growth was ca. 110 mV over 100 cycles. The polarization growth was more significantly affected at higher pressures, with a 270 mV and 380 mV growth for cells constrained at 1725 kPa and 2205 kPa, respectively. It is also observed that the initial ΔV for FEC:TFEC cells increases as a function of pressure. This contrast of polarization response to applied pressure suggests both an electrochemical and a physical difference exists between these two solvent systems. This was investigated by measuring the ionic conductivity of each electrolyte as a function of temperature, shown in SI Figure S3. At 40°C, the temperature at which the cells were cycled in this work, 1M LiPF₆ in FEC:TFEC has a conductivity of 3.9 mS/cm compared to 1M LiPF₆ in FEC:DEC which has a conductivity of 9.2 mS/cm. The particularly low conductivity of FEC:TFEC coupled with high applied pressure is possibly a contributor to the large polarization growth observed here. This may be due to impeded ionic transport through the separator at high pressure, however, this hypothesis must be investigated in future work.

Figure 6 shows direct comparisons of capacity retention between cells with FEC:DEC and FEC:TFEC electrolyte initially constrained in normal boats at <75 kPa (a), and in superboats constrained at initial average pressures of 485 kPa (b), 795 kPa (c) and 1725 kPa (d). In each case, the cells with FEC:DEC electrolyte begin suffer from severe capacity loss before FEC:TFEC cells; FEC:DEC cells fall below 80% capacity earlier than FEC:TFEC cells as shown in Figure 6e with hatched bars. This indicates that cells with FEC:TFEC electrolyte have a higher initial coulombic efficiency, consistent with previous reports.³³ However, after the depletion of in-situ excess lithium and resulting initial rapid capacity loss, the rate of capacity loss for FEC:DEC cells constrained >75 kPa is not as severe as FEC:TFEC cells. Although FEC:TFEC cells initially cycle with higher capacity retention, the final rate of capacity loss of FEC:TFEC cells result in poorer capacity retention than FEC:DEC cells after approximately 50 cycles. Figure 6e shows the number of cycles to 50% capacity in solid bars. At low (<75 kPa) and moderate (485 kPa) initial pressures FEC:TFEC electrolyte remains superior. Beyond 485 kPa, FEC:DEC electrolyte has matched or surpassed the number of cycles reached to 50% capacity due to the larger final rate of capacity loss of FEC:TFEC cells. This is likely a result of the low ionic conductivity of FEC:TFEC electrolyte possibly contributing to large polarization growth, particularly at pressures >485 kPa.

Zoom In Zoom Out Reset image size

Figures 5 and 6 demonstrate that increasing the initial average stack pressure between 75–2205 kPa is generally beneficial, and that cells with FEC:TFEC electrolyte are superior to cells with FEC:DEC electrolyte up to 50 cycles. However, it is shown here that individual benefits achieved with applied pressure and the choice of electrolyte, when used in unison, are not necessarily synergistic. The performance of cells constrained to different pressures is coupled to the choice of electrolyte, and thus the physical properties of the electrolyte are important to consider in cells with lithium metal in which higher pressures generally increases lithium plating efficiency.

To investigate the effect of applied pressure and choice of electrolyte on lithium plating morphology, samples of plated lithium were retrieved from cycled cells at the top of charge (4.5 V) for SEM. Figure 7 shows SEM images at 5000x magnification of plated lithium after one charge (a-d) and after 50 cycles (A-D) for FEC:DEC cells constrained in normal boats at <75 kPa (a, A) and in superboats at 485 kPa (b,B); and for FEC:TFEC cells constrained in normal boats at <75 kPa (c, C) and in superboats at 485 kPa (d,D). Figure S4 shows the complimentary low magnification (1000x) SEM images. After one charge in a normal boat, the plated lithium for the FEC:DEC cell (a) exhibits larger grain sizes than those for the FEC:TFEC cell (c). In addition, more void space is observed in the FEC:DEC sample—the plated lithium in the FEC:TFEC sample appears more close packed and nodular. In the superboat constrained at 485 kPa, a significant difference is observed as the plated lithium becomes more close packed for both FEC:DEC (b) and FEC:TFEC (d) samples. The FEC:TFEC sample appears to have less void space and thus the smallest surface area. A low surface area is ideal since this results in less contact with the electrolyte thereby reducing the lithium inventory that is lost to SEI formation. Therefore, Figures 7a–7d show that cells constrained at higher pressure and cells containing FEC:TFEC electrolyte exhibit superior plating morphology after one charge. This is consistent with the cycling results which show cells constrained in superboats exhibit superior capacity retention and indicated that FEC:TFEC cells have a higher initial coulombic efficiency.

Zoom In Zoom Out Reset image size

After 50 cycles constrained at low pressure <75 kPa in normal boats, the FEC:DEC (A) and FEC:TFEC (C) samples exhibit a significantly more dendritic, mossy structure; the feature size is dramatically reduced and the morphology is very porous with high surface area. This is a poor morphology for efficient lithium plating, and this is reflected in the cycling results which show that cells constrained in normal boats have lost approximately 80% capacity up to 50 cycles. The samples constrained at 485 kPa (B,D) exhibit a morphology that is much less porous, consistent with results which show that higher pressure improves lithium plating efficiency. Constrained at 485 kPa, the morphology of the FEC:DEC and FEC:TFEC samples appear qualitatively similar, consistent with the cycling data presented in Figure 6b which shows that after 50 cycles, both chemistries have lost approximately the same amount of capacity. The initial benefit of FEC:TFEC electrolyte on plating efficiency, observed in the morphology shown in Figures 7b and 7d and the initial cycling data in Figure 6, is lost by cycle 50.

These results show that fully fluorinating the electrolyte solvent with FEC:TFEC electrolyte improves lithium plating morphology and thus cell performance. This is consistent with previous works which have shown that fully fluorinating the electrolyte alters the nanostructure of the SEI resulting in more uniform lithium deposition and improving electrochemical performance.^5,50 Clearly, the SEI composition and structure, which in large part dictates the lithium morphology, is intimately bound to electrolyte composition, and thus electrolyte development will continue to play a large role in the development of lithium metal cells. Additionally, the role of mechanical pressure to physically thwart the formation of dendrites and to flatten out otherwise high surface area mossy morphology is again demonstrated here.

To put our results in to a broader context, Figure 8 compares the energy density vs cycle no. for the 402020-size anode-free NMC532/Cu cells described in this work with a 402035 Li-ion NMC532/graphite Li-ion cell optimized for capacity retention.⁵¹ The Li-ion cell, shown with black circles, contains 1.2 M LiPF₆ EC:EMC (3:7) electrolyte with the optimized additive blend + 2% vinylene carbonate (VC, BASF 99.97%) + 1% 1,3,2-Dioxathiolane-2,2-dioxide (DTD, Sigma Aldrich, 98%). Anode-free cells with FEC:DEC electrolyte constrained in a normal boat at <75 kPA and in a superboat at 795 kPa are shown with light and dark green squares; FEC:TFEC electrolyte cells constrained in a normal boat at <75 kPa and in a superboat at 795 kPa are shown with light and dark red triangles, respectively. For further comparison, the calculated initial energy density for Li-metal cells using the same cell geometry as presented in this work but with additional excess lithium layers of 15, 80 and 150 μm corresponding to 1x, 5x and 10x Li excess are shown with light, medium and dark blue stars, respectively.

Zoom In Zoom Out Reset image size

The Li-ion cell begins with an energy density of ca. 700 Wh/L, approximately 60% of the energy density of anode-free cells. The Li-ion cell experiences virtually no loss of energy over the course of the 80 cycles shown in Figure 8 (by cycle 2500, this cell had only lost 10% energy). The anode-free cells begin with an energy density of ca. 1150 Wh/L. Although anode-free cells suffer from severe capacity loss, this massive boost to energy density allows them to deliver superior energy density compared to Li-ion cells for some time. The anode-free cells tested in this work with moderate initial average pressures of 795 kPa deliver more energy than the gold standard Li-ion cell for approximately 50 cycles. Although this is still a relatively short cycle life, we believe electrolyte optimization will yield significant improvements, allowing anode-free cells with liquid electrolytes to deliver more energy than Li-ion cells for over 100 cycles in the near future—and hopefully much more as research in this field develops—increasing their commercial viability.

The calculated energy density for lithium metal cells with excess lithium shown with blue stars in Figure 8 demonstrate why lithium metal cells with significant excess lithium are not useful for improving energy density. The lithium excess layer we refer to here corresponds to the single sided lithium layer as shown in Figure 1 since this is how an excess layer would have to be introduced—as such there would be two of these layers per stack. A 15 μm lithium excess, corresponding to 1x excess lithium for this cell design, results in an energy density of ca. 1100 Wh/L, a modest decrease from the anode-free configuration. However, once significant excess lithium is introduced, an 80 μm layer corresponding to x5 excess results in a paltry 10% increase in energy density compared to the Li-ion cell. A 150 μm layer, still less than many standard lithium foils used in lithium metal cells, which corresponds to 10x lithium excess in this cell design results in an energy density of 500 Wh/L—30% lower energy density than the Li-ion cell. Some lithium metal cells reported in the literature utilize this much excess lithium and more, and as a result exhibit artificially prolonged cycle life as there is a vast reservoir of excess lithium which can be depleted before the cathode can no longer be fully lithiated; however, such cells with significant excess result in a cell chemistry that is less safe, more difficult to manufacture and without significant, if any, benefit to energy density compared to conventional Li-ion cells. We believe that to make a useful lithium metal cell, an anode-free or limited excess lithium (<30 μm) configuration, as outlined by Albertus et al.,¹³ must be adopted.