Tailored PVDF Graft Copolymers via ATRP as High-Performance NCM811 Cathode Binders

High-nickel layered oxides, e.g., LiNi0.8Co0.1Mn0.1O2 (NCM811), are promising candidates for cathode materials in high-energy-density lithium-ion batteries (LIBs). Complementing the notable developments of modification of active materials, this study focused on the polymer binder materials, and a new synthetic route was developed to engineer PVDF binders by covalently grafting copolymers from poly(vinylidene fluoride-co-chlorotrifluoroethylene) (PVDF-CTFE) with multiple functionalities using atom transfer radical polymerization (ATRP). The grafted random copolymer binder provided excellent flexibility (319% elongation), adhesion strength (50 times higher than PVDF), transition metal chelation capability, and efficient ionic conductivity pathways. The NCM811 half-cells using the designed binders exhibited a remarkable rate capability of 143.4 mA h g–1 at 4C and cycling stability with 70.1% capacity retention after 230 cycles at 0.5 C, which is much higher than the 52.3% capacity retention of nonmodified PVDF. The well-retained structure of NCM811 with the designed binder was systematically studied and confirmed by post-mortem analysis.

Nuclear Magnetic Resonance Spectroscopy (NMR). 1 H NMR spectroscopy measurements were performed on a Bruker Advance 300 MHz spectrometer using DMSO-d6 as the solvent.
Gel Permeation Chromatography (GPC).Number-average molecular weights (M n ) and molecular weight distributions (MWD) of the polymers were determined by GPC.The THF GPC system used an Agilent 1260 infinity II pump, a Waters 2414 refractive index detector, a PSS SLD9000 MALS detector, PSS columns (SDV 10 3 , 10 5 , and 10 6 Å) with THF as the eluent at a flow rate of 1 mL/min at 35 °C.The DMF GPC system used an Agilent 1260 infinity II pump, a PSS SLD2020 MALS detector, PSS columns (GRAM 100, 3000, 3000 Å) with DMF as an eluent at 50 °C and the flow rate of 1 mL/min.To obtain absolute molecular weight, dn/dc values were measured by preparing GPC samples with different concentrations of each polymer (1 mg/mL, 2 mg/mL, 3 mg/mL, 4 mg/mL, and 5 mg/mL).Then the injected mass was plotted vs. area.The dn/dc was determined from the slope, which reflects the detector constant multiplied by the dn/dc.The obtained dn/dc values were then used for the calculation of the absolute molecular weight of each polymer.

Fourier-transform infrared spectroscopy (FTIR). FTIR was carried out using a
PerkinElmer Frontier spectrometer with a germanium crystal exploring a 1 cm -1 of resolution over a range of 700-4000 cm -1 .
Thermogravimetric Analysis (TGA).TGA with TA Instruments 2950 was used to measure the thermal stability of the polymers under nitrogen.The data were analyzed with TA Universal Analysis.The heating procedure involved four steps: (1) temperature ramp up to 110 ºC at a rate of 30 ºC/min; (2) hold at 110 ºC for 10 min; (3) ramp up at a rate of 30 ºC/min to 800 ºC; (4) hold for 3 min.

Differential Scanning Calorimetry (DSC).
The glass transition temperature (T g ) of PVDF, PVDF-CTFE, and PVDF-CTFE-g-PEGA-co-PAA were measured by DSC with TA Instrument QA-2000.The samples were first equilibrated at 25.00 °C and isothermal for 1.00 min, following three cooling/heating cycles, each involving the following steps: (1) Ramp 10.00 °C/min to -90.00 °C, (2) Isothermal for 1.00 min, (3) Ramp 10.00 °C/min to 200.00 °C, (4) Isothermal for 1.00 min.The DSC data were analyzed with a TA Universal Analysis instrument, and T g were acquired directly.
Dynamic mechanical analysis (DMA): Bulk film fabrication.The PVDF, PVDF-CTFE, and PVDF-CTFE-g-PEGA-co-PAA polymer solutions in NMP (8 wt%) were drop-casted into 15 mm × 5 mm rectangular Teflon molds.After most of the solvent was slowly evaporated over 24 h at room temperature, molds were transferred to a vacuum oven and continuously evaporated at 120 °C for 48 hours.Transparent free-standing films with thicknesses of 100-200 μm were obtained.
Tensile test.The bulk films were tested in the tensile mode by using DMA (TA RSA-G2).The film thickness was between 100-200 μm.The samples were stretched at a constant tensile rate of 0.05 mm/mm⋅s at room temperature.
Peel test. 1 cm-wide Scotch tape (3M 600) was first firmly pressed and adhered to NCM811 cathodes on Al foil prepared with PVDF, PVDF-CTFE and PVDF-CTFE-g-PEGA-co-PAA binders, respectively.One end of the electrode was clamped and fixed on the bottom geometry of the DMA instrument, and one end of the tape was clamped on the top geometry that can move freely.Then the tape was peeled at an angle of 180 degrees from the electrode at a constant peeling rate of 0.5 mm/s, the adhesion strength of the electrodes was recorded as the force of the geometry (N).

Scanning electron microscope (SEM). SEM was performed for surface morphology
investigation and conducted on a Quanta 600 FEG instrument.Images were acquired through ETD detector at 30 kV with spot size 3.5.

Focused Ion Beam Scanning Electron Microscope (FIBSEM). FIBSEM was conducted on
FEI Helios Xenon PFIB.An accelerating voltage of 5 kV and beam current of 0.34 nA were used for imaging.A plasma beam current of 6.7 nA and 1.8 nA were used for particle milling.
1-1.5 μm of Pt was coated on the samples before ion-milling to enable the clear and fine crosssection images.
Micro Computed Tomography (Micro-CT).Each electrode was cut into rectangular shape with a width of around 1 mm and height of 5 mm, then was mounted on a sample holder using fast-setting epoxy and placed into an Xradia Crystal CT (Zeiss Group).By extending the detector and moving Xray source close to their limits, a Voxel size of 0.74 µm was achieved.
The scan was performed with X-ray source set to 40 kV, 3 A, air filter, bin 1, and 0.5s x 5 exposures per projection using Scout and Scan Control System (Zeiss Group).Reconstruction was done manually to account for large centering shift using Reconstructor Scout-and-Scan Control System program (Zeiss Group).The reconstructed .txmfile was then processed with Dragonfly software (Object Research System).After cropping each sample into 500 µm by 500 µm by electrode height blocks where the images are least distorted and contains only electrodes, they were partitioned into pore space and non-pore space based on the X-ray signal strength of each individual voxel using Dragonfly's Otsu thresholding tool.A threshold of 31000 unit was used on all three samples to ensure consistency and comparability between the samples.

X-ray photoelectron spectroscopy (XPS).
The XPS spectra were collected using a Thermo Fisher ESCALAB 250 Xi instrument with a 650 μm acquisition spot size.A flood gun was used during data collection and each spectrum was referenced to the adventitious carbon C-C peak at 284.8 eV.Each plot represents the average of at least 10 scans and the spectra were fit using Avantage software package.

Inductively coupled plasma mass spectrometry (ICP-MS). After disassembly of the cycled
Li|NCM811 cells, all the other parts except NCM811 cathodes were collected and harvested from the cell, then soaked in 2 mL dimethyl carbonate for 24h. 100 µL of the solution were taken and added into 0.5 mL 70% HNO 3 for digestion.The digested samples were further diluted to 3.5% HNO 3 by ultra-purified water and filtered by a 0.45 μm PTFE syringe filter before analysis by ICP-MS.(Agilent 7700X).

Synthesis of PVDF-CTFE-g-PEGA-co-PtBA
In a 20 mL vial, 1.2 g (0.00103 mol) PVDF-CTFE was dissolved in 8 mL DMF and 1 mL DMSO.Then 1.13 mL (1.24 g, 0.0026 mol) of OEGA480, 2.64 mL (2.31 g, 0.018 mol) of tBA, 110.8 µL (0.00412 mmol, 5 mg/mL solution in DMF) of CuCl 2 , 95 µL (0.412 µmol, 3 mg/mL solution in DMSO) of Eosin Y disodium salt and 142.4 µL (0.0124 mmol, 20 mg/mL solution in DMF) of Me 6 TREN were added to the vial.The vial was then covered with a rubber stopper and the reaction mixture was degassed by nitrogen bubbling for 5 minutes.It was then placed under a green light lamp at room temperature.Once the reaction was complete with the target conversion monitored by 1 H NMR, the reaction mixture was precipitated in 50 vol% methanol/water, redissolved in acetone, and precipitated two more times.The solid product was then dried and collected.

Hydrolysis of PVDF-CTFE-g-PEGA-co-PtBA
The dry product of PVDF-CTFE-g-PEGA-co-PtBA polymer (1.63 g) and 0.2 mL TFA were added to a Schlenk flask and sealed under vacuum, then left in vacuum oven overnight at 60℃.The resultant polymer was then washed with DCM and analyzed using 1 H NMR to monitor the completion of conversion of tert-butyl groups to -COOH group, after drying under vacuum at 50℃ overnight, the PVDF-CTFE-g-PEGA-co-PAA product was obtained and used directly as binder for NCM811 cathode preparation.

Cell assembly and electrochemical testing
Electrochemical characterization was carried out using 2032-type coin cells with lithium metal anodes, celgard battery separator, 0.5 mm stainless steel spacer, and wave spring.1M lithium hexafluorophosphate (LiPF6) in a mixture solution of ethylene carbonate (EC) and ethyl methyl carbonate (EMC) (3:7 v/v) was used as the liquid electrolyte.NCM811 cathodes were prepared with composition of NCM811/carbon black/polymer binders = 85/10/5 and 93/4/3 wt%.The well-mixed cathode slurry was cast onto Al foil (active material loading ~ 5 mg/cm 2 ) and vacuum dried for 8h at ~ 90℃.All Li|NCM811 coin cells were assembled using 600 µm thick Li metal, as-prepared NCM811 cathodes and 60 µL of liquid electrolyte to standardize the testing.All the cells were assembled in a glove box with water/oxygen content lower than 0.5 ppm.Rate tests and long-term testing of Li|NCM811 cell were conducted on Landt Instruments at room temperature.For post-mortem analysis of cycled NCM811 cathodes, coin cells were disassembled in glove box, then NCM811 cathodes were washed by dimethyl carbonate (DMC).After vacuum drying, the cycled cathodes were directly used for XPS, SEM and FIB-SEM analysis.
Table S1.Samples of grafted PVDF-CTFE polymers and their performance as binders in Li|NCM811 cells.
a Absolute molecular weight determined by DMF GPC with MALS detector.
b Absolute molecular weight determined by THF GPC with MALS detector.PVDF-CTFE-g-PEGA-co-PtBA were further exposed to TFA to obtain PVDF-CTFE-g-PEGA-co-PAA to be used as binders for NCM811 cathodes.
c Mean discharge capacities at different C-rates were reported from Li|NCM811 cells in the voltage range of 2.9-4.3Vusing NCM811 cathodes (NCM811/carbon black/binder = 85/10/5 wt%) prepared by corresponding polymer binders to select optimal binder composition.The D app can be calculated using the following equation: n is the charge transfer number, F is the Faraday constant (96485 C mol -1 ), C is the concentration, A is the electrode surface area (cm 2 ), R is the gas constant (8.314J mol -1 K - 1 ), T is the temperature (298 K), D app is the apparent diffusion coefficient (cm 2 s -1 ), and σ w is the Warburg coefficient (ohm•s -1/2 ).

Figure S3 .
Figure S3.DSC curves of a) PVDF; b) derivative of DSC of PVDF; c) DSC curves of PVDF-CTFE; d) derivative of DSC of PVDF-CTFE.

Figure S4 .
Figure S4.Photographs of a) 180° peel test set up.NCM811 cathodes on adhesive tape and Al substrate after peel test with b) PVDF c) PVDF-CTFE and d) PVDF-CTFE-g-PEGA-co-PAA binder.

Figure S7 .
Figure S7.Micro CT of NCM811 cathodes with PVDF binder.a) The raw data of the reconstructed gray-scale image; b) thresholding segmentation of the active material, binder, and carbon black.Porosity: 36.4%.

Figure S8 .
Figure S8.Micro CT of NCM811 cathodes with PVDF-CTFE binder.a) The raw data of the reconstructed gray-scale image; b) thresholding segmentation of the active material, binder, and carbon black.Porosity: 38.0%.

Figure S9 .
Figure S9.Micro CT of NCM811 cathodes with PVDF-CTFE-g-PEGA-co-PAA binder.a) The raw data of the reconstructed gray-scale image; b) thresholding segmentation of the active material, binder, and carbon black.Porosity: 34.6%.

Figure S12 .
Figure S12.Scheme of NCM811 cathodes prepared with a) covalently grafted PVDF-CTFEg-PEGA-co-PAA binder and b) physical mixture of PVDF-CTFE/PEGA/PAA binder using the same molecular weight and ratio of the three components.

Figure S16 .
Figure S16.The equivalent circuits and the fitting results of the EIS in Figure S15.