Improving lithium-ion cells by replacing polyethylene terephthalate jellyroll tape

Polyethylene terephthalate (PET) tape is widely used by well-known lithium-ion battery manufacturers to prevent electrode stacks from unwinding during assembly. PET tape is selected since it has suitable mechanical and electrical properties, but its chemical stability has been largely overlooked. In the absence of effective electrolyte additives, PET can depolymerize into its monomer dimethyl terephthalate, which is an unwanted redox shuttle that induces substantial self-discharge in a lithium-ion cell. This study presents a chemical screening experiment to probe the PET decomposition mechanism involving in situ generated methanol and lithium methoxide from dimethyl carbonate, one of the most common electrolyte solvents in lithium-ion cells. By screening other polymers, it is found that polypropylene and polyimide (Kapton) are stable in the electrolyte. Finally, it is demonstrated that reversible self-discharge of LiFePO4–graphite cells can be virtually eliminated by replacing PET jellyroll tape with chemically stable polypropylene tape. Polyethylene terephthalate (PET) tape is widely used for lithium-ion batteries but its chemical stability has been largely overlooked. Reversible self-discharge is now shown to be virtually eliminated in LiFePO4–graphite cells by replacing PET with polypropylene jellyroll tape.

Polyethylene terephthalate (PET) tape is widely used by well-known lithium-ion battery manufacturers to prevent electrode stacks from unwinding during assembly.PET tape is selected since it has suitable mechanical and electrical properties, but its chemical stability has been largely overlooked.In the absence of effective electrolyte additives, PET can depolymerize into its monomer dimethyl terephthalate, which is an unwanted redox shuttle that induces substantial self-discharge in a lithium-ion cell.This study presents a chemical screening experiment to probe the PET decomposition mechanism involving in situ generated methanol and lithium methoxide from dimethyl carbonate, one of the most common electrolyte solvents in lithium-ion cells.By screening other polymers, it is found that polypropylene and polyimide (Kapton) are stable in the electrolyte.Finally, it is demonstrated that reversible self-discharge of LiFePO 4 -graphite cells can be virtually eliminated by replacing PET jellyroll tape with chemically stable polypropylene tape.
Every commercial lithium-ion battery (LIB) made by stacking or winding the electrode sheets contains tape.This tape holds the assembled cell stack or 'jellyroll' together, before it is inserted into the cylindrical, prismatic or pouch cell casing during manufacturing.Typically, thin tapes are used to avoid wasting valuable space or adding unnecessary weight that would lower the energy density and specific energy of the cell, respectively.After manufacturing, the tape does not serve a function and is generally regarded as an inactive cell component.
Figure 1 shows that many well-known battery manufacturers use polyethylene terephthalate (PET) tape in their cells.Fourier transform infrared spectroscopy (FTIR) of tapes extracted from discarded smartphone batteries show the characteristic absorption bands of PET (Fig. 1a,c), with slight variations in the absorption bands originating from tape degradation (below).Out of ten randomly selected cells from major smartphone original equipment manufacturers (Fig. 1a) and four 18650-sized cylindrical cells from reputable suppliers (Fig. 1b), only two cells did not contain PET tape.Apart from its use as tape, PET has been proposed by various companies, such as Soteria 1 , DuPont 2 and Meta 3 , as a substrate for lighter and safer metal-coated aluminium and copper current collectors.PET is also used as separator material, electrode coating and organic electrode material according to the academic literature 4 .For all these applications, the chemical stability of PET plays an important role, but has not been investigated.
Recently, studies from our group by Logan et al. 5 and Büchele et al. 6 have shown that LIBs with LiFePO 4 (LFP) or LiNi 1-x-y Mn x Co y O 2 (NMC) positive electrodes and graphite negative electrodes are prone to rapid self-discharge at elevated temperature unless effective electrolyte additives are used.The LFP-graphite cells lost roughly 30% of their charge over 500 h of storage at 40 °C, and self-discharged completely when stored at 60 °C for 500 h 6 .Büchele et al. found an unwanted redox shuttle molecule that is generated in situ during formation, cycling or storage of some battery cells at elevated temperature 7 .A redox shuttle can diffuse between the positive and negative electrode of a battery and transport electrons from one side to the other through a reversible redox reaction, accepting an electron from the negative electrode and donating it to the positive electrode.This results in lithium ions being transferred from the negative to the positive electrode without charge being drawn from the cell to power a device in the external circuit.This self-discharge reaction is (in principle) non-damaging and reversible since the lithium inventory is unaffected, but it is of concern https://doi.org/10.1038/s41563-023-01673-3examined.The insets of Fig. 3b,d show that the PET tapes have severely corroded in both cells, and even fully dissolved in some locations.The FTIR spectra in Fig. 3b show that the anode tape has lost some of its characteristic absorption bands between 1,300 and 1,100 cm −1 indicating degradation of the ester bonds found in PET.Similar to the FTIR spectra of PET tapes shown in Fig. 1a, the FTIR spectrum of the anode tape exhibits reduced intensity in the carbonyl region between 1,900 and 1,600 cm −1 , as well as substantial changes between 3,000 and 2,800 cm −1 which are typically associated with a decrease in PET polymer chain length 9 .The wavenumber positions of the remaining PET absorption bands in degraded tapes are unchanged (Figs.1a and 3b), but there are clear modifications in band shapes and intensities, suggesting that structural degradation of the PET polymer chain has occurred.The cathode tape still resembles the PET reference spectrum, indicating that PET degradation reactions primarily occur at the anode.The closure tape of the stacked cell (Fig. 3d) also shows a substantially altered FTIR absorption spectrum as well as substantial dissolution in the middle region of the cell, which indicates PET decomposition just as for the anode tape in Fig. 3b.

The reaction path of PET decomposition
Tanaka et al. 8 have demonstrated in the context of waste recycling that PET can be depolymerized with methanol and lithium methoxide.In lithium-ion cells, both methanol and lithium methoxide can be generated in situ from DMC (Supplementary Scheme 1).(1) Methanol is a product of DMC hydrolysis and a common manufacturing impurity 10 .(2) Lithium methoxide can be formed by DMC reduction at the unpassivated negative electrode 11,12 .When reacting with PET, methanol dissociates into a proton and a methoxide anion, which attacks the electrophilic reaction centre of PET (Supplementary Scheme 2).
(3) This cleaves the ester bonds of the PET polymer, creating the DMT monomer and ethylene glycol from reaction with the adjoining proton, in many applications, for example, consumer electronics.In addition, it is of great concern for cells that are connected in series in a battery module.If the self-discharge rates of such cells differ, the cell balancing algorithm of the battery management system may fail to safely charge the module.Büchele et al. identified the shuttle molecule as dimethyl terephthalate (DMT) 7 .This study will demonstrate that DMT is the product of PET depolymerization with methanol and lithium methoxide 8 , two reactants that can be generated in lithium-ion cells from dimethyl carbonate (DMC), a ubiquitous electrolyte solvent.In addition, this study will show how reversible self-discharge of LIBs can be virtually eliminated by replacing PET tape with polypropylene (PP), an alternative polymer with far superior chemical stability.

Decomposition of PET tape in lithium-ion cells
Figure 2 shows FTIR spectra of the cathode, anode and jellyroll closure tapes of wound LFP-graphite pouch cells either made with (1) PET or (2) PP tapes.To indicate the location of the tapes in the cells, Fig. 2c shows a dissection of one of the 402035-sized pouch cells used in this study.The closure tape is on the outside of the jellyroll (Fig. 2d) and the cathode and anode tapes are on the unrolled electrodes (Fig. 2e,f).

Chemical stability screening of jellyroll tapes
To test the proposed PET decomposition mechanism, PET tape was placed in a pouch bag with mixtures of DMC, 2 wt% lithium methoxide and 10 wt% methanol, sealed and kept at 70 °C for 5 h (Supplementary Fig. 1).Subsequently, the liquid mixture was extracted and analysed by gas chromatography-mass spectrometry (GC-MS) (Fig. 4). Figure 4a-c shows that DMC alone and single combinations of DMC with lithium methoxide or methanol do not depolymerize PET and the only component found by GC-MS is DMC (note that lithium methoxide is extracted with the aqueous phase, and methanol is either all consumed in the reaction or not visible in the retention time range used in this study, Methods).Figure 4d shows that PET tape dissolves and DMT is created when DMC, lithium methoxide and methanol are added.In addition to DMT, EC is generated, further confirming the proposed reaction path shown above.Ethylene glycol bis-(methyl carbonate) (DMOHC) is produced from the in situ generated EC reacting with DMC (Supplementary Scheme 3) 13 .Like the PET depolymerization, this dimerization reaction is also catalysed by lithium methoxide.DMOHC formation consumes EC, which explains why there is less EC than DMT in Fig. 4d, even though they should be produced in equal amounts according to the proposed PET depolymerization mechanism (3). Figure 4e shows that PET also depolymerizes at room temperature when DMC, lithium methoxide and methanol are added, and the pouch bag is stored for 1 week.Similar amounts of DMT, EC and DMOHC are formed as in the experiment at 70 °C, thus PET depolymerization does not need elevated temperature.Note that the analogous depolymerization reaction of PET with ethanol and lithium ethoxide from diethyl carbonate produces diethyl terephthalate, which is another proof that the proposed mechanism is accurate (Supplementary Fig. 2 and Supplementary Scheme 4).
Figure 4f shows an analogous experiment with an LFP-graphite pouch cell that contained PET jellyroll tape (Fig. 2).The cell was filled with 2 ml of 1.5 M LiPF 6 in DMC electrolyte, 0.5 wt% of lithium methoxide and 0.5 wt% of methanol, which are closer (but still very much in excess) to the concentrations of these species in real lithium-ion cells 12 .The pouch cell was kept at 70 °C for 4 days, the time of a standard formation protocol (Methods, note that the temperature is higher than typical formation temperatures).Afterwards, the cell was opened, and the electrolyte was analysed by GC-MS.The obtained spectrum (Fig. 4f) shows the same compounds as the corresponding pouch bag test (Fig. 4d), but in smaller concentrations.A clear signal is obtained for the DMT redox shuttle confirming its in situ generation from PET tape decomposition.In addition, trimethyl phosphate, a compound originating from the reaction of LiPF 6 with DMC and residual water, was produced 7 .This simple screening experiment unambiguously proves that PET depolymerization is a chemical rather than an electrochemical reaction and elevated temperature or voltage are not needed for the reaction to occur (Fig. 4e).Note, however, that (1) methanol generation from DMC hydrolysis does require an elevated temperature above 40 °C to proceed at notable rates 9 , and (2) lithium methoxide generation from DMC reduction is an electrochemical process that can only happen when the negative electrode is not passivated with a well-insulating solid electrolyte interphase (SEI) layer 13 .
Next, the chemical screening method is used to find alternative polymers that do not dissolve in lithium-ion cells.Supplementary Scheme 5 shows that neither PP nor polyimide (Kapton) has ester bonds that could be cleaved by lithium methoxide or methanol.Therefore, tapes made from these polymers should be immune to the depolymerization mechanism described above.To confirm this hypothesis, PP and Kapton tapes were tested in pouch bags with DMC, lithium methoxide and methanol at 70 °C for 5 h, followed by GC-MS analysis of the liquid mixtures.Figure 5a-c shows no DMT redox shuttle in pouch bags with Kapton and PP, but a clear DMT signal with PET.Other peaks in the GC-MS spectra correspond mainly to acrylics or silicon compounds, often used as tape adhesives, for example, the peak at 15.5 min in Fig. 5b,c, which was also seen in Fig. 4d,e (ref.14). Figure 5g-i shows that PP and Kapton tapes are intact, whereas PET tape has fully dissolved after 5 h at 70 °C.Kapton has a substantially higher cost than PP 15 , and its polyimide group has been found to cause a notable increase in irreversible capacity loss when in contact with the negative electrode 16 .PP is already common in many battery separators, hence, PP tape is used in the following to replace PET tape in custom LFP-graphite pouch cells.Figure 5e,f shows GC-MS spectra of electrolytes extracted from LFP-graphite pouch cells with PP and PET jellyroll tape, respectively, after formation at 70 °C with  1.5 M LiPF 6 in EC:DMC (3:7) control electrolyte (CTRL).The DMT redox shuttle was only produced when PET tape was used (Fig. 5f); when PP tape was used no polymer decomposition products were found (Fig. 5f) 7 .DMOHC is found regardless of the jellyroll tape, which indicates that lithium methoxide is generated in both cases, but only in the presence of PET can it yield the DMT redox shuttle.The higher chemical stability of PP tape in lithium-ion cells should lead to lower self-discharge rates and higher coulombic efficiency especially in the absence of additives.

Improving lithium-ion cells by replacing PET with PP tape
Figure 6a-c shows the voltage versus time curves of fully charged LFPgraphite cells with PET or PP jellyroll tape during 500 h of storage at 40 or 60 °C after formation at 40 or 70 °C.The self-discharge rates of these cells can be inferred from the decay in open circuit voltage (OCV) 17 .In the absence of electrolyte additives, cells with PP jellyroll tape (Fig. 6c) show substantially slower self-discharge than cells with PET tape (Fig. 6a,b).When 2% VC is added all cells show very slow self-discharge (dashed lines in Fig. 6a-c).Figure 6d-f shows the reversible (red) and irreversible storage losses (blue) obtained by recharging the cells and comparing the discharge capacity after storage to the initial discharge capacity (black).Cells with CTRL electrolyte and PET tape formed at 70 °C and stored at 40 °C lost one-third of their capacity, and most of the storage losses are reversible since they originate from the in situ generation of the DMT redox shuttle from PET tape.After 40 °C formation and 60 °C storage, the PET CTRL cells have completely self-discharged and 70% of the storage losses are reversible.In addition, these cells exhibited substantial irreversible self-discharge potentially due to the absence of VC and the inferior EC-derived SEI on the negative electrode.The lack of a well-passivating SEI would cause continuous solvent reduction and Li inventory loss, which is a well-known contributor to irreversible capacity loss 17,18 .
By contrast, cells with CTRL electrolyte and PP tape formed at 70 °C and stored at 60 °C (that is, a combination of the harshest conditions) still retained roughly 90% of their initial capacity.The reversible capacity loss is almost completely eliminated in cells with PP tape (only 3%), even without electrolyte additives (Fig. 6f), which underlines the stability of the PP polymer and the absence of a redox shuttle.When 2% VC is added, both, reversible and irreversible losses become very small for all cells, since the well-insulating VC-derived SEI prevents the formation of lithium methoxide and thus the decomposition of PET tape as well as other parasitic reactions 7 .
Figure 6g-i shows ultra-high precision coulometry (UHPC) results for LFP-graphite pouch cells with PET and PP tapes formed at 70 °C and cycled at 40 °C.Cells with CTRL electrolyte and PET tape (black) show a clear slippage of the voltage curves of up to 10 mAh per cycle (Fig. 6g,h), due to the DMT-induced self-discharge during cycling.In cells with PP tape (red) this parasitic reaction is absent, thus there is virtually no charge endpoint capacity slippage and an overall higher coulombic efficiency during cycling (Fig. 6i).Cells with 2% VC (blue and green) show even less slippage and higher coulombic efficiency due to an improved passivation of the negative electrode.The well-insulating VC-derived SEI has also been shown to prevent DMT from shuttling 19 .
Figure 6j-l shows corresponding long-term cycling tests at C/3 and 70 °C.In the absence of additives, cells with PP jellyroll tape show similar capacity retention to cells with PET tape, confirming that the use of PP does not induce new unwanted side reactions.Cells with 2% VC outperform the additive-free cells (note the 10% increase in cycle life when PP tape is used) due to a superior passivation of the negative electrode, which prevents lithium loss: the major ageing mechanism in high temperature LFP-graphite cells as shown by Logan et al. 17 .Since most commercial cells use SEI-forming additives such as VC, one could infer that the use of PET is unproblematic, even if it depolymerizes into the DMT redox shuttle.However, previous studies have shown that VC and other SEI-forming additives can be consumed during cycling 20,21 .It is likely that additive consumption in heavily cycled cells would correlate with a deterioration of the SEI layer, the formation of alkoxides from reduction of linear carbonates, and subsequent decomposition of PET tape (Fig. 3).To eliminate the root cause of redox shuttle generation from tape decomposition, it is recommended that lithium-ion cell producers replace PET jellyroll tape with chemically stable PP tape.

Attenuated total reflectance-FTIR
All FTIR spectra were collected in an Ar-filled glovebox using a Cary 630 FTIR spectrometer equipped with a Ge crystal attenuated total reflectance accessory.MicroLab PC software was used to collect the data with a resolution of 4 cm −1 .The polymer films used in this study were cleaned with dry antistatic tissue to remove any impurities before the measurement.

Pouch bag experiments
Small pockets containing either PET or PP tape were prepared from battery-grade pouch foil with a heat sealer.The pouch bags were filled with 2-3 ml of DMC containing 10% methanol (MeOH) and 2% lithium methoxide (LiOMe) in an Ar-filled glovebox, heat-sealed and placed into a 70 °C temperature box for 5 h.Pouch bags with pure DMC or DMC with just MeOH or LiOMe were also tested, so were full lithium-ion pouch cells with these reactants.Some pouch bags were not heated but stored at room temperature for one week.All pouch bags were opened, the liquid mixture was extracted and analysed in a gas chromatograph connected to a mass spectrometer.

Electrolyte preparation
The CTRL electrolyte in this study was 1.5 M LiPF 6 in a 3:7 (w/w) ratio of EC and DMC.Other electrolyte formulations contained LiFSI, ethyl methyl carbonate (EMC), VC, DTD, MMDS (more than 98.7% Tinci Materials Technology) or TTSPi (more than 95% Sigma-Aldrich).All solvents and salts had less than 20 ppm water, were used as-received from Shenzhen Capchem unless specified otherwise and were mixed in an Ar-filled glovebox.

Pouch cells
The 402035-sized 220 mAh LFP-graphite pouch cells with PET and PP jellyroll tapes were obtained vacuum sealed without electrolyte from LiFUN Technologies.The cells were cut open in an Ar-filled glovebox, dried at 140 °C under vacuum for 14 h, filled with 1 ml of electrolyte and resealed under vacuum at −90 kPa gauge pressure.The stacked NMC622-graphite cell shown in Fig. 3c,d was obtained from a reputable cell manufacturer.

Formation protocol
All cells were charged to 1.5 V and then underwent a 16 h voltage hold to ensure wetting of the electrode pores with electrolyte while avoiding dissolution of the copper current collector.Then the cells completed a single C/20 formation cycle between 2.5 and 3.65 V and a C/10 recharge to 50% SOC on a Maccor 4000 series charger at 40 or 70 °C.

GC-MS
GC-MS protocol and instrumentation was the same as described by Petibon et al. 12 .Pouch cells were filled with 1 ml of methyl acetate, resealed and kept at room temperature for 1 week to ensure full equilibration of electrolyte components, methyl acetate and any reaction products in the jellyroll.Similar to Petibon et al. 12 , 0.2 ml of water was added to the mixture of CH 2 Cl 2 and electrolyte for salt removal.After shaking and resting for 5 min, the aqueous layer was removed with a syringe, MgSO 4 was added to bind residual water and the organic layer consisting of organic electrolyte components in CH 2 Cl 2 was injected into the GC-MS.

Long-term cycling protocols
After formation, the cells were cut open in an Ar-filled glovebox for degassing, resealed under vacuum and brought to a Neware (China) or Novonix (Canada) system for cycling at 40 or 70 °C in constant current constant voltage mode with a C/3 rate for charge and discharge as well as a C/20 'check-up cycle' every 50 cycles.The voltage ranges were 2.5-3.65 and 3.0-4.2V for LFP-graphite and NMC622-graphite cells, respectively.Pair cells were made for every test condition.

Self-discharge experiments
To characterize the rate of self-discharge, OCV storage experiments were conducted at 40 and 60 °C as described by Sinha et al. 17 .After formation, the cells completed two full charge and discharge cycles at C/10 to precisely determine the initial discharge capacity (D 0 ), followed by a full charge to the upper cut-off potential of 3.65 V and a constant voltage hold for 10 h and a 500 h OCV storage period.Subsequently, the cells were discharged (D 1 ), charged, and discharged again (D 2 ) at C/10 rate to determine the irreversible (D 0 -D 2 ) and reversible capacity loss (D 2 -D 1 ) during storage.

UHPC
Charge endpoint capacity slippage and coulombic efficiency were measured on the UHPC system at Dalhousie University described by Bond et al. 22 .The cells were cycled at 40 °C and C/20 from 2.5 to 3.65 V. Pair cells were made for every test condition.

Fig. 1 |Fig. 2 |Fig. 3 |
Fig. 1 | The widespread use of PET tape by original equipment manufacturers.a,b, FTIR spectra of jellyroll closure tapes found in commercial mobile phone batteries (a) and cylindrical 18650 cells from reputable vendors (b).c, Mobile phone batteries from which the tapes were obtained.

Fig. 5 |
Fig. 5 | Chemical screening of Kapton, PP and PET tapes, demonstrating the chemical stability of Kapton and PP tapes.a-c, GC-MS spectra of liquid mixtures extracted from pouch bags with Kapton (a), PP (b) and PET tape (c) in DMC, 10 wt% MeOH and 2 wt% LiOMe.d-f, After 5 h at 70 °C Kapton (d) and PP (e) tapes are intact, whereas PET tape has fully dissolved (f).g-i, Also, GC-MS

TFig. 6 |
Fig. 6 | Demonstration of improvements in self-discharge and cycling performance when PP tape is used instead of PET tape.a-c, Voltage versus time profiles of LFP-graphite pouch cells with PET tape formed at T F = 70 °C and stored at T S = 40 °C (a), with PET tape formed at 40 °C and stored at 60 °C (b) and with PP tape formed at 70 °C and stored at 60 °C (c).d-f, The corresponding bar graphs (d-f) show initial capacity, reversible capacity loss and irreversible capacity loss.g-i, UHPC cycling profiles (g), charge endpoint capacity slippage (h) and Coulombic efficiency (CE) (i) for LFP-graphite cells with either PET or