Simplified Pouch Cell Method for 3-Electrode Re-Testing of Harvested Double-Sided Electrodes From Commercial Lithium-Ion Batteries Free

Lithium-ion batteries continue to be the premier chemistry for mobile electrochemical energy storage due to their increased power density and energy density compared with other chemistries [1]. In evaluating commercial lithium-ion cells, destructive physical analysis, or cell dissection, is a common tool used for postmortem analysis of cell materials to garner information about cell materials and determination of failure modes [2–8]. Other than visual inspection and materials analysis, retesting of the electrode materials can also give important insights into electrode capacity decrease or potential characteristic changes as well as show the extent of lithium inventory loss in the cell [9]. Retesting can also get important electrode performance parameters, such as open circuit charge and discharge electrode potential curves, that are necessary to pseudo-2D electrochemical modeling [10–13]. Safe dissection of fully discharged commercial cells and safe charging of the harvested electrode materials in lab-scale cells for calorimetry tests or other analytical testing such as X-ray or electron based measurements is also possible with retesting.

Due to the double-sided electrodes that are used in most commercial lithium-ion cells, particularly wound cells, using coin cells as the rebuild format requires removal of the electrode composite from one side of the electrode in order for the current collector to make electrical contact with the casing of the coin cell [7,14]. The composite removal can be done by mechanical scrapping or using solvents such as N-methyl Pyrrolidone [7,14]. However, this composite removal may carry a risk of damage or contamination to the composite on the opposite side of the metal foil that will be the subject of retesting. Additionally, coin cells do not allow for use of a third electrode [15–19] without feeding a tab through the cell gasket [16,20,21], a process which can compromise the cell seal.

Pouch cells can accommodate double-sided electrodes for retesting, but conventionally require ultrasonic welding to weld tab connections. They also require precision bar-sealers in order seal the pouch cells across tab feedthroughs without forming and electrical connection between the tabs and metal foil in the laminate. An additional challenge with retesting harvested electrodes is the moisture-sensitive nature of SEI on the anode and the electrolyte. This moisture sensitivity would require welding and precision sealing capabilities to be present inside an inert-gas glovebox or a dry room facility for typical pouch cell construction.

In the present work, a simplified pouch cell method for retesting harvested double-sided lithium-ion electrodes in a 3-electrode cell arrangement is developed and demonstrated. The method removes the need for any welding by relying on pressure contact between a feedthrough tab with the double-sided electrode tab. The method also removes the need to house precision sealing equipment inside an inert glovebox or dry room facility, as precision sealing may be performed during the build steps that do not involve the moisture-sensitive components. A cathode and anode from a commercial 3.6 Ah 18650 cell are rebuilt into a 3-electrode pouch cell. The functionality of the pouch cell is tested with C/10 cycling and the quality of the pressure contact of the tabs is evaluated with high frequency impedance measurement.

The pouch laminate used was MTI Corp aluminum laminated film and the separator used was Celgard trilayer separator. The copper foil used was from MTI Corp and 25 µm thick, the aluminum foil used was from MTI Corp and 15 µm thick, and the copper mesh used was from MTI Corp and 9 µm thick. The lithium metal used was 0.6 mm thick lithium metal chips from MTI Corp. The heat seal tape used was hot melt adhesive polymer tape from MTI Corp.

A commercial 3.6 Ah 18650 lithium-ion battery was discharged at a current of 0.68 A to a cutoff voltage of 2.5 V using an Arbin BT200 cycler. Caution: Cell disassembly can lead to inadvertent formation of an internal short circuit. Performing disassembly after the cell was fully discharged minimizes risk of causing the cell to go into thermal runaway or heating significantly if a short circuit develops during disassembly. The cap and bottom of the case were then removed from the cell using an MTI compact disassembling machine. Parafilm was wrapped over the cell top and bottom after removal of the cap and case bottom. The cell was brought into an MBraun argon glove box maintained at <5 ppm H₂O and <5 ppm O₂ by continuous purging of the antechamber. The cell was placed in a vise and the case was cut open using a Dremel rotary tool. The cut was made in the axial direction on the casing. Care was taken to ensure the Dremel tool only cut through the casing and not into the separator or electrode materials. Once the casing was cut open, the jellyroll was removed. The jellyroll was unrolled and sections of the double-sided anode and cathode were cut from the roll. Electrodes were cut from each section with the electrode area measuring 1.0 cm² and featuring a ∼1.5 cm long tab.

The layout of the pouch cells is diagramed in Fig. 1. Figure 2 is a process flowchart of 3-electrode cell build using re-harvested electrodes showing (a) pre-glove box cell assembly steps, (b) inert glove box assembly steps, and (c) final assembly steps. The pouch cells were constructed as follows: the pouch laminate material was cut into 2–5 cm × 12 cm pieces. Heat seal tape was adhered to one of the 5 cm long sides of both pieces of the laminate material by placing the 5 cm end of the onto the lower seal bar heated to 185 °C on a Tester Sangyo TP-701-fB heat sealer and gently pressing the seal tape onto the laminate with a spatula. The adhered seal tape overhung the end of the 5 cm long side of the laminate by about 0.1 cm, as shown in in the top left picture of Fig. 2(a). Next, Cu mesh was cut into rectangles like those shown in the top center picture of Fig. 2(a) that are about 1.5 X 2.5 cm with about 4 cm tabs. The Cu mesh tabs were placed on the heat seal tape and adhered to the sealer tape on a first piece of the laminate by setting the end of the pouch rectangle with the sealing tape on the lower bar of the heat sealer and pressing the mesh with a spatula, as shown in the top right picture in Fig. 2(a). The layout of the copper mesh rectangles was two meshes being placed over one another with their tabs next to each other on one side (hereafter referred to as the “left side”) of the laminate to serve as the counter electrode. The last mesh electrode placed on top with its tab on the opposite side of the laminate (hereafter referred to as the “right side”) to serve as the reference electrode. Next, a metal foil tab about 0.5 × 2.0 cm (Cu in the anode cells, Al in the cathode cells) was placed on the heat seal tape in between the Cu mesh tabs and adhered to the first piece of laminate by placing the end of the laminate rectangle with the seal tape on the heated seal bars and pressing with a spatula (hereafter referred to as the “feedthrough tab”) as shown in the top right picture of Fig. 2(a). The separator was cut into ∼4 × 4 cm pieces and was placed into the open cell as follows: Two separator pieces were placed between the two copper mesh rectangles that had their tabs next to each other on the left side of the laminate as shown in the second and third rows of pictures of Fig. 2(a). The copper or aluminum foil feedthrough tab was between these two pieces of separator. One separator piece was placed between the copper mesh rectangle with its tab on the right side and closest copper mesh to it that had its tab aligned on the left side. Once the separator was in place, the second piece of laminate was aligned with the first, the 5 cm with seal tape adhered to them aligned, and they were sealed together by the heat sealer with the top bar heated to 170 °C, the pressure set to 0.1 MPa, and the seal time set to 4 s. The aligned laminate pieces are shown being sealed in the bottom right picture of Fig. 2(a). The cell was then dried in a VWR 1430D vacuum oven for 2 h at 50 °C. The cell was then removed and transferred into the argon glove box via vacuum cycling of the antechamber.

In the glovebox, the electrode composite material on the tabs of the re-harvested anode or cathode working electrodes that had been prepared previously and stored in the glovebox while the pouch cell was assembled was manually scraped off of both sides using a razor blade as shown in the top left picture of Fig. 2(b). Three lithium chips were placed on the Cu mesh electrodes in an aligned stack as shown in Fig. 1 and the top middle picture of Fig. 2(b). The cell was rolled with a glass vial to press the lithium metal into the mesh until he lithium began to spread out shown in the left picture in the second row of Fig. 2(b). The cathode or anode working electrodes were then placed so that the electrode active area was aligned with the stack of Li metal chips, with the working electrode tab resting on the feedthrough tab as shown in the center picture of the second row of Fig. 2(b). The pouch cell was then rolled again with a glass vial to form the lithium metal around the cathode or anode outline. Two of the remaining 3 open sides of the cell were sealed inside the glove box using a U-line H109 6 in. crimper hand sealer, leaving one side open for electrolyte filling, shown in the bottom row of pictures of Fig. 2(b). Next, about 300 µl (an overfill amount) of 1.0 M LiPF₆ 1:1 ethylene carbonate (EC): diethyl carbonate (DEC) v/v electrolyte was injected into the open side and massaged gently into the cell. The last side of the cell was sealed with the crimper hand sealer and the cells were placed in a ¼ in. aluminum restraint plates with rubber sheets on either side of the cell as shown in Fig. 2(c). The restraint plate bolts were hand tightened and the finished cells were removed from the glovebox for testing.

Galvanostatic cycling of the cells was done with a Bio-Logic VMP-3 cycler at room temperature measuring the potential difference between the working electrode (harvested cathode or anode) and the Li counter electrode (Cu meshes with Li pressed into them with tabs next to each other on the left side) and measuring the potential difference between the working electrode and the Li metal reference electrode (Cu mesh with Li metal pressed into it with tab on the right side). Both tabs of the counter electrode that were next to each other on the left side were contacted by the alligator clip of the VMP-3 counter electrode connection. The constant current was selected based on an estimated C-rate of ∼0.1 C with a 3.0 V to 4.25 V versus Li/Li⁺ voltage window for the cathode cells and a 0.05 V to 1.5 V versus Li/Li⁺ voltage window for the anode cells based on the electrode potential versus the lithium metal reference electrode. Potentiostatic electrochemical impedance measurements were also performed with the Bio-Logic VMP-3 cycler with a 10 mV voltage perturbation magnitude at a frequency of 500 kHz averaging 2 cycles. The impedance was measured between the working electrode and lithium metal counter electrode.

As shown in Fig 1, the extracted electrode tab is lined up to the feedthrough tab and will ultimately be held together by the compression of the pouch cell between sheets of rubber that are compressed between two aluminum plates. This connection does not rely on tab welding of any kind, which is beneficial because no welding equipment, typically ultrasonic welding [16–18], is needed to connect the tabs together. Removal of composite material at the tab of the cut electrodes is necessary for good electrical contact between the electrode tab and feedthrough tab as well as preventing composite material on the tab from participating in the electrochemical activity of the cell. However, only material from the tab area needs to be removed, and mechanically scraping the tabs minimizes the risk of damaging or contaminating the composite electrode area for testing.

The precision sealing necessary at the top of the cell, where the tabs are located, is also done during the assembly of nonmoisture-sensitive pouch cell components outside of the glovebox. Therefore, a dry room facility or housing the precision sealing equipment inside the inert-gas glovebox is not required, and precision sealing equipment can be housed in a standard laboratory environment. The only sealing necessary inside the glovebox after moisture-sensitive materials are incorporated into the cell can be done with a nonprecision, compact hand sealer since no tabs are fed through the other three sides of the pouch cell that need to be sealed. The left side of Figs. 1 and 2 also shows that the Cu mesh counter electrode tabs are placed such that they are side by side to one another when the top of the cell is sealed instead of directly on top of one another. This is done with the intention to limit the possibility of the formation of pathways for air/electrolyte leakage between the two copper mesh tabs during cell operation.

Electrolyte needs to be added to the pouch cell due to the minimal amount of electrolyte in the commercial 18650 cell not allowing for sufficient wetting of separators and lithium metal electrodes. The added electrolyte formulation cannot be matched to that of the cell without identification of the electrolyte formulation from the manufacturer or from experimental procedure, so it is instead approximated in this study by a common lithium-ion electrolyte formulation. Additionally, an overfill amount of electrolyte, as was added in this study, will not allow for analysis of electrolyte consumption-based fade mechanisms such as passivation layer growth in the full commercial cell. Lastly, this cell design using a third Li metal electrode as the reference electrode aims to eliminate the potential measurement inaccuracies, which increase at higher current densities, that come from using a lithium metal electrode as both the counter electrode and reference electrode in common half-cell setups [22].

Restraint plates are necessary to keep uniform pressure on the cell and to ensure proper contact between the internal cell components to prevent inhomogeneous cycling behavior [23–25]. The amount of applied external pressure by the restraint plates has been reported to vary between 0.1 and 1.0 MPa, with the amount of applied pressure having been shown to have an impact on the cycle life and impedance of the cell [23–25]. Thus, the effect of cell pressure is a consideration when using this retesting method. As the reports of applied pressure on lithium-ion ouch cells is quite varied however, a greater importance was placed on maintaining uniform pressure on the cell by inserting rubber sheets into the restraint plates before the plates were tightened.

At rest, the potential difference of the lithium counter electrode and lithium reference electrode was less than 0.05 mV for the anode and cathode cells, indicating that the placement of the reference electrode with the counter electrode between it and the working electrode caused a negligible impact on the potential measurement. During cycling, the magnitude of the potential difference between the lithium counter electrode and lithium reference electrode in the anode cells varied between 0 and 4 mV while for the cathode it varied between 0 and 2 mV. This potential difference is consistent with previous results that show the current applied at the counter electrode changes its potential [22] and therefore measurement of the working electrode potential versus the counter electrode does not give as reliable a value of the electrode potential versus Li/Li⁺ as a third lithium reference electrode.

Figure 3(a) shows the potential profiles measured versus the third lithium reference electrode for five lithium insertion and extraction cycles in the 0.005–1.5 V versus Li/Li⁺ range for the re-harvested anode. The first two insertion and extraction cycles, the anode capacity is lower than in later cycles but increases. The insertion curves of cycles 4 and 5 show a sharp decreasing slope between 1.5 V and 0.2 V versus Li/Li⁺. At about 0.2 V versus Li/Li⁺, there is a short plateau region followed by a decrease to 0.1 V versus Li/Li⁺. Lastly, a flat, plateau-like region below 0.1 V versus Li/Li⁺ is observed. These plateau features are consistent with the insertion stages of graphite anode material [26] and do not show significant plateau like features at potentials > 0.4 V versus Li/Li⁺ that are indicative of solid electrolyte interphase (SEI) formation [27]. This lack of plateau-like features at potentials > 0.4 V versus Li/Li⁺ evidences minimal new SEI formation during cycling in the pouch cell and that the SEI formed during the conditioning and cycling of the commercial 18650 cell before dissection is still present and passivating the anode surface. However, the possibility of evolution of the SEI, particularly when an electrolyte different from that of the commercial cell must be used like in the present case, should be considered when retesting harvested electrodes with this pouch cell design. The capacity increases over the first few cycles and stabilizes in the last two insertion and extraction cycles with an identical capacity of ∼10 mAh in each of the last two cycles. Based on the total insertion capacity and the active area of 1.0 cm² on each side of the electrode, the areal capacity of the anode in the 0.005–1.5 V versus Li/Li⁺ range is ∼5.0 mAh/cm².

Figure 3(b) shows the potential profiles measured versus the third lithium metal electrode for five lithium insertion and extraction cycles for a re-harvested cathode electrode. The first extraction cycle from the cathode shows a higher average voltage than subsequent cycles. The second lithium extraction/insertion cycle through the fifth complete extraction/insertion cycle show uniform, stable cycling, and capacity. The lithium extraction and insertion cycles in Fig. 3(b) show a capacity of ∼12 mAh. Based on the total extraction capacity and the active area of ∼1 cm² on each side of the electrode, the areal capacity of the cathode is ∼6 mAh/cm² in the 3.0–4.25 V versus Li/Li⁺ potential range.

An additional 11 cathode pouch rebuild cells and six anode pouch rebuild cells were constructed and tested. Table 1 shows average areal capacity based on the fifth cycle capacity and the standard deviation. The average and standard deviation for both the anodes and cathodes is nearly identical, at 5.50 ± 0.30 mAh/cm² for the cathode and 5.50 ± 0.31 mAh/cm² for the anode. These results are consistent with the expected areal capacity of 5.2 mAh/cm² based on the cell rated capacity of 3.6 Ah and a total electrode area of 345 cm². They also indicate that the N/P ratio is in the cell design is about 1.0.

The high potential cutoff used for the cathode in this study was estimated based on the manufacturer prescribed voltage range of the commercial cell used, which is 3.0 V–4.20 V, and an estimated anode end-of-charge potential of 50 mV. Although the exact potential ranges of the electrode in the commercial cell are not known, this estimate was based on the assumption of an excess anode capacity design. Based on the results of the electrochemical testing shown in Fig. 3 and Table 1 and using an estimation method of matching the electrode potential profiles [28] the anode potential at the end of charge in the full cell is estimated to be 15 mV versus Li/Li⁺ and the cathode potential at the end of charge in the full cell is estimated to be 4.215 V versus Li/Li⁺. This estimation method still has uncertainty due to the effects [18,19] of an unknown lithium inventory in the full cell on electrode potentials during cycling. Insertion of a reference electrode into the commercial cell and measuring the potential ranges of the electrodes during charge and discharge [29–31] of the full cell would improve the electrode potential range accuracy. Overall however, the data in Fig. 3 and Table 1 shows that this electrode harvesting and retesting technique is effective in achieving stable electrode cycling performance with expected potential plateau features in the anode and similar areal capacities of both electrodes.

The real impedance between the lithium counter electrode and the working electrode was measured at 500 kHz frequency to assess the resistance of the pressure held electrical tab connection. Before the impedance was measured, the anode cell was lithiated to 0.005 V versus Li/Li⁺, while the cathode cell was delithiated to 4.25 V versus Li/Li⁺, respectively. The anode voltage was measured to be ∼85 mV versus Li/Li⁺ and the cathode voltage was 4.2 V versus Li/Li⁺, respectively. This state of each electrode is roughly equivalent to their state in a charged full cell. The real impedance of the anode cell was 1.55 Ohms and the real impedance of cathode cell was 2.17 Ohms. These high frequency real impedances are similar to the high-frequency real impedance of pouch cells with continuous tabs [19] and coin cells [20] of a similar capacity, indicating the electrical contact between the working electrode tab and the feed through tab being held by compression is sufficient.

A simplified pouch cell method for retesting double-sided electrodes from commercial lithium-ion batteries in a 3-electrode cell arrangement for potential measurement against a third reference electrode has been developed and its functionality tested. The method relies on pressure contact between metal tabs and therefore eliminates the need for tab welding and precision heat sealing inside an inert glovebox or dry room environment. Electrochemical cycling of electrodes harvested from a commercial 3.6 Ah 18650 lithium-ion cell and rebuilt with the pouch cell method showed stabilized capacity and potential profile performance after four cycles for the anode and two cycles for the cathode at a targeted C/10 cycling rate. The areal capacity of the cathode and anode electrodes was determined to be 5.50 ± 0.30 mAh/cm² and 5.50 ± 0.31 mAh/cm², respectively, based on the results from galvanostatic cycling of multiple cells. These measured areal capacities indicate the N/P ratio of the commercial cell in this case is close to 1.0 although the potential ranges of the electrodes in the cell are based on estimates and the exact potential ranges are not known. 500 kHz impedance measurements for the anode and cathode cells show a real impedance of 1.55 Ohms and 2.17 Ohms, respectively, which is similar to other coin cells and pouch cells with continuous tab connections of similar capacity. This similarity demonstrates that the tab contact via restraint plate pressure is sufficient. Overall, the present work demonstrates the functionality of a method to retest double-sided electrodes from commercial lithium-ion cells that can be performed in laboratories without dry room capability, welding capability, and/or glovebox space that is too limited to contain precision heat sealing equipment.

The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the Naval Surface Warfare Center, Crane Division or the U.S. Government.

• Naval Innovative Science and Engineering (NISE) program through the Naval Surface Warfare Center, Crane Division.

There are no conflicts of interest.