Use email button above to contact.

Replication Data for: Combined Dynamic Nuclear Polarization and Electron Paramagnetic Resonance at 0.34 Tesla to Investigate Electrochemical Lithium Deposition on Copper

Supporting experimental data and simulation results for the publication DOI: 10.1038/s41598-025-01107-x

In the accompanying publication, plating and dendrite formation in lithium batteries is studied using a newly developed setup for dynamic nuclear polarization (DNP) operating at 0.34 Tesla. Electron paramagnetic resonance (EPR) sensitively detects metallic Li species but misses non-paramagnetic ones. Nuclear magnetic resonance (NMR) is chemically selective, yet exhibits a low sensitivity under low-field conditions. DNP-enhanced NMR overcomes this by transferring electron spin polarization to ⁷Li nuclei. Here, correlative X-band EPR and DNP-enhanced ⁷Li-NMR of plated lithium is demonstrated. DNP experiments were conducted in a pulsed mode to prevent excessive sample heating. The resulting enhanced ⁷Li NMR signal allows the observation of electrochemically plated lithium on copper, harvested from a Cu vs. Li cell, with an enhancement ε > 400. By changing the magnetic field strength by a few Gauss, the saturation of the conduction EPR transition was varied, leading to an altered shift of metallic ⁷Li. The corresponding change of the DNP-polarized ⁷Li chemical shifts in the range from 240 ppm to 80 ppm, in turn, allowed an indirect, saturation-based distinction of EPR species. Moreover, an enhancement ε by a factor of about 2 of the ¹H signal from the surrounding electrolyte of electrochemically deposited lithium was observed, indicating the potential to investigate the solid-electrolyte interface (SEI).

Please, switch to Tree View for the folder structure. The dataset contains the experimental DNP-enhanced NMR spectra and EPR spectra as well as the results from the finite element method (FEM) simulations used for optimizing the coil geometry. All data depicted in the figures of the publication are provided in semicolon-delimited UTF-8 CSV format, organized in two folders: DNP_setup_optimization and Spectra_material_characterization. The former contains the results relevant for hardware development while the latter provides the data obtained from the experiments conducted with the developed setup. The number at the beginning of each file name represents the figure number in the associated publication.

FEM Simulations

The program EMPro 2020 was used to perform FEM simulations. According to the experimental conditions, the simulated coil diameter was set to 1 cm and the coil height to 4.5 cm. The simulated rectangular copper wire cross section was 0.1 mm × 0.1 mm and a current of 1 A was simulated with a frequency of 10 MHz. The magnetic flux density H_x was simulated for an opening angle of θ = 120°. An imaginary and a real component of H_x with x pointing in the direction of the oscillating radio frequency (RF) field were obtained, representing the phase shift of the RF signal. Here we are only interested in the amplitude |H_x| = √(Re(H_x)² + Im(H_x)²), which was calculated and visualized as a heat map using Matlab 2015a.

Nutation Curve

The ¹⁹F NMR nutation curve of LP30 was measured with the console of a Bruker AvanceIII-HD spectrometer, using TopSpin 3.6.1, at the magnet from the EPR spectrometer.

NMR Measurements

Data processing, including SNR determination, was done using MestReNova (version 14.2.3-29241). For ¹H NMR spectra, a background spectrum was recorded and subtracted from the respective measurement. ⁷Li NMR spectra were referenced to Li₂SO₄ (aq) at 0 ppm and ¹H NMR spectra to H₂O at 4.8 ppm.

The ¹H NMR spectra were apodized exponentially with a time constant corresponding to 50 Hz line broadening, and the ⁷Li spectra were apodized exponentially with a time constant corresponding to 200 Hz (disassembled battery) and 100 Hz (complete battery). For plotting, the spectra were exported to OriginPro 2021b (OriginLab Corporation, Northampton, MA, USA). In this work, the signal enhancement is provided as the amplitude enhancement ratio ε_ampl or the integral enhancement ratio ε_int of the NMR signal.

Sample Preparation

TEMPOL (4-hydroxy-2,2,6,6-tetramethyl-piperidin-1-oxyl (aq), 33 mM, 95 % purity, Sigma-Aldrich) was prepared under ambient atmosphere. It was filled into an EPR tube (Wilmad quartz (CFQ) EPR tubes, O.D. 2 mm, L 100 mm), which was positioned centrally in a sample holder with a diameter of 1 cm.

Lithium was plated on copper in a Swagelok cell. The cell was assembled under argon atmosphere using a 7 mm disk of Li foil (0.1 mm thickness, 99.9 % purity, Sigma-Aldrich), a 7 mm disk of Cu foil (10 µm thickness, Evonik), and a glass microfiber separator (Whatman GF/C) of 12 mm diameter and 0.26 mm thickness, soaked with 200 µL of LP30 electrolyte (1.0 M LiPF₆ in EC/DMC=50/50 (v/v), battery grade, Sigma-Aldrich). Lithium was plated onto copper at a constant current density of 2 mA cm^-2 for 1 h. Afterwards, the Swagelok cell was disassembled under argon atmosphere. The Li foil as well as several layers of the separator were stripped off. The remaining separator was cut to a diameter of 8 mm. The resulting sample, consisting of the Cu disk with plated Li and the attached separator, was used for DNP experiments. Prior to conducting the DNP-enhanced NMR measurements, the sample was stored under argon atmosphere in a glovebox for three days. For the DNP measurements, the sample was transferred to a quartz glass cell described by Niemöller et al. This cell was put into the sample holder, which exhibits a cylindrical geometry (diameter: 1 cm, height: 20.5 cm). It was sealed air-tightly using joint grease.

The battery was prepared under argon atmosphere in a glovebox. It was built using a lithium (0.1 mm thickness, purity of 99.9 %, Onyxmet) and a copper electrode, which were punched out from the corresponding plates with a diameter of 7 mm. A separator (Whatman GF/C) with a diameter of 8 mm and 0.26 mm thickness was placed between the electrodes, and 10 µL of LP30 (1.0 M LiPF₆ in EC/DMC=50/50 (v/v), battery grade, Sigma-Aldrich) was added onto the separator. The battery was placed in an in operando cell described by Niemöller et al. This cell was placed inside a battery holder as described in the previous paragraph and each electrode was connected to a platinum wire. At first, Li was plated and stripped at a current density of 4 mA cm^-2 for 50 min each. Afterwards, Li was plated at a current density of 1.3 mA cm^-2 for 3 h. Prior to conducting the DNP-enhanced NMR measurements, the battery was stored under argon atmosphere in a glovebox for nine days, which resulted in the battery cell drying out.

For electrochemical lithium deposition, a copper and a lithium electrode were chosen. The copper electrode was cut from a copper plate (1 cm × 3 cm), and the lithium electrode was prepared by wrapping lithium (0.1 mm thickness, 99.9 % purity, Onyxmet) around a copper wire. They were placed with a distance of roughly 2 cm in a beaker under argon atmosphere in a glovebox. The electrolyte LiPF₆ in PC (1.0 M, battery grade, Sigma-Aldrich) was added into the beaker with a fill height of roughly 2 cm. The electrodes were connected to a 3.6 V battery using copper wires, initiating the deposition of lithium onto the copper electrode. This process continued for approximately 2 h. A mixture of dead and deposited lithium, which was scraped from the copper electrode, was filled into an EPR tube with a diameter of 4 mm (Wilmad quartz (CFQ) EPR tubes) under argon atmosphere in a glovebox. The lithium stuck to the glass wall, and the remaining electrolyte settled at the bottom of the tube. The latter was removed using a needle and a syringe. After closing the tube with a cap, additional parafilm was wrapped around it. The tube was then aligned vertically within the sample holder, which was sealed using joint grease and parafilm.

For further information, please refer to the Methods section of the publication.

Folder Contents

The DNP_setup_optimization folder contains the following data:

• The results from the FEM simulations for different saddle coil opening angles. For each simulated opening angle, a 2a_x_profile_XXXdeg.csv file (with XXX being the respective opening angle) can be found with five columns each: x, z, y, Re(H_x), Im(H_x). Direction x is along the direction of the oscillating RF field, z is along the direction of the static magnetic field, and y is the remaining direction. The amount √(Re(H_x)² + Im(H_x)²) can be plotted as a function of x for each opening angle. The heatmap (Fig. 2(c)) can be generated from the data listed in file 2c_heatmap_xy_120deg with five columns each: x, z, y, Re(H_x), Im(H_x).
• The 2d_Nutation_curve.csv file contains the ¹⁹F NMR nutation curve for different pulse lengths of LP30 (1.0 M LiPF₆ in ethylene carbonate (EC) and dimethyl carbonate (DMC) with EC/DMC=50/50 (v/v)) measured using the saddle coil with three windings per wing at 3960.4 G with an RF power of 30 W and 16 scans each. The file contains two columns: the pulse length (µs) and the corresponding intensities.
• In 3a_TEMPOL_mw_polarization_buildup.csv the amplitude enhancement of the ¹H NMR spectrum of TEMPOL (aq) as a function of the time t is listed. The sample was irradiated for 0.5 s. The data is listed in two columns: the duration of the microwave (mw) irradiation (ms) and the corresponding amplitude enhancement |ε_ampl| of the TEMPOL sample.
• In 3b_TEMPOL_mw_enhancement_decrease.csv the effect of mw irradiation times longer than 0.5 s on the integral enhancement as a function of the mw irradiation time is listed. There are two columns: the duration of the mw irradiation (s) and the integral enhancement of the TEMPOL sample normalized to the first data point at t_mw = 1.4 s.

The Spectra_material_characterization folder contains the following data:

• In 4_7Li_deposited_Li_mw_on_off.csv DNP-enhanced ⁷Li NMR spectra of a Cu disk with electrochemically deposited Li without and with mw irradiation at 15.8 W and a mw frequency of 9.299 GHz can be found. The three columns correspond to the ⁷Li chemical shift (ppm), the intensity without mw irradiation upscaled by a factor of √128, and the intensity of the spectrum with mw irradiation. The data corresponds to the disassembled battery cell.
• In 5_Correlated_EPR_spectrum.csv the EPR spectrum recorded under non-saturating conditions of the Cu disk with electrochemically deposited Li can be found. The two columns correspond to the B field (G) and the intensity (a.u.), respectively. The data corresponds to the disassembled battery cell.
• In 5_Correlated_7Li_B_variation.csv the DNP-enhanced ⁷Li NMR spectra of the Cu disk with electrochemically deposited Li at different magnetic fields corresponding to Fig. 5 in the publication can be found. The 26 columns correspond to 3314.9 G δ ⁷Li (ppm), intensity (3314.9 G), 3315.4 G δ ⁷Li (ppm), intensity (3315.4 G), ..., 3320.9 G δ ⁷Li (ppm), intensity (3320.9 G). The data corresponds to the disassembled battery cell.
• 6_7Li_battery_vs_Li_disk.csv contains ⁷Li NMR spectra of an assembled Li vs. Cu half cell without and with 15.8 W mw irradiation. There are six columns: battery δ ⁷Li (ppm), spectrum battery (mw off), spectrum battery (mw on), Li disk δ ⁷Li (ppm), spectrum Li disk (mw off), spectrum Li disk (mw on). The data corresponds to the complete battery cell.
• In 7_1H_enhancement_Li_electrolyte.csv ¹H NMR spectra of electrochemically deposited lithium wetted with electrolyte can be found. The file contains ten columns: δ ¹H (ppm), spectrum 0 W, spectrum 7.9 W, spectrum 15.8 W, spectrum 0 W (control), spectrum 7.9 W (control), spectrum 15.8 W (control), spectrum 0 W (prior to background subtraction), spectrum 7.9 W (prior to background subtraction), spectrum 15.8 W (prior to background subtraction). The data corresponds to the electrochemically deposited lithium sample.