Surface Chemistry Dependence on Aluminum Doping in Ni-rich LiNi0.8Co0.2−yAlyO2 Cathodes

Aluminum is a common dopant across oxide cathodes for improving the bulk and cathode-electrolyte interface (CEI) stability. Aluminum in the bulk is known to enhance structural and thermal stability, yet the exact influence of aluminum at the CEI remains unclear. To address this, we utilized a combination of X-ray photoelectron and absorption spectroscopy to identify aluminum surface environments and extent of transition metal reduction for Ni-rich LiNi0.8Co0.2−yAlyO2 (0%, 5%, or 20% Al) layered oxide cathodes tested at 4.75 V under thermal stress (60 °C). For these tests, we compared the conventional LiPF6 salt with the more thermally stable LiBF4 salt. The CEI layers are inherently different between these two electrolyte salts, particularly for the highest level of Al-doping (20%) where a thicker (thinner) CEI layer is found for LiPF6 (LiBF4). Focusing on the aluminum environment, we reveal the type of surface aluminum species are dependent on the electrolyte salt, as Al-O-F- and Al-F-like species form when using LiPF6 and LiBF4, respectively. In both cases, we find cathode-electrolyte reactions drive the formation of a protective Al-F-like barrier at the CEI in Al-doped oxide cathodes.

: Co L 3 -edge measurements for a) LiNi 0.8 Co 0.15 Al 0.05 O 2 (NCA) and b) LiNi 0.8 Co 0.2 O 2 (LNC) electrodes charged to 4.75 V at room temperature (24 • C) with no constant voltage (referred to as RT electrode) and charged to 4.75 followed by a 10 hr constant voltage (CV) hold, all at 60 • C (referred to as 60 • C electrode). In the LiPF 6 case, the Co L 3 -edge shows minimal change between the RT and 60 • electrodes (spectra overlaps). This can be contrasted with the LiBF 4 case, where the NCA electrode shows a pronounced increase in lower energy features associated with cobalt surface reduction under thermal stress.
S-2 Figure S2: XPS measurements of the a) C 1s, b) F 1s, c) Al 2p, Ni 3p, Co 3p and Li 1s core region using Al Kα X-ray source (1.486 keV) for LNC, NCA, and LiNi 0.8 Al 0.2 O 2 (LNA) 60 • C electrodes. Normalizing the C 1s to the carbon black peak (C-C/284.5 eV), we find minimal difference lineshape related to the formation of new carbon environments related to electrolyte solvent decomposition. In the F 1s core region, the LNA 60 • C electrode shows the largest amount of fluorine surface species related to LiPF 6 decomposition. The higher binding energy peak at 687.7 eV is associated with P-O-F species while the lower binding energy peak at 685.3 may be associated with residual LiF and the formation of new Ni-F species. The binding energy region from 50 eV to 80 eV contains extensive information on the transition metal (TMs), aluminum, and lithium surface environments. Compared to the NCA and LNC 60 • C electrodes, LNA shows a shift in the Ni 3p core region to higher energy and the presence of a higher binding energy Al 2p peak related to the formation of Ni-F and Al-O-F species, respectively. In all three systems, there are minimal lithium surface species. Figure S3: XPS measurements of the a) O 1s and b) P 2p core regions using Al Kα X-ray source (1.486 keV) for LNC and NCA electrodes held for 10 and 175 hr at 4.75 V and 60 • C. Longer holding results in more pronounced changes for the NCA system than the LNC system, where there is a pronounced buildup of P-O-F species at the surface (ndicated by shading in a and arrow in b). Figure S4: Al 1s core region for commercial NCA (cNCA) and in-house synthesized NCA and LNA electrodes charged to 3.6 V and 4.75 V at 60 • C with a 10 hr CV hold using either LiPF 6 or LiBF 4 . In the 3.6 V case, there is some variation in the asymmetry of the main Al 1s peak (1558-1560.5 eV) which is associated with differences in the amount of Al-Olike environments. For the 60 • C electrodes, we highlight the Al 1s peak at 1562.7 eV and 1564 when using LiPF 6 and LiBF 4 , respectively. These peaks are assigned to Al-O-F-like environments and indicate an aluminum-electrolyte reaction that is dependent on the choice of electrolyte salt.

S-4
Figure S5: Depth-dependent XPS measurements at 0.8 and 6 keV of the a) O 1s and b) Al2p/Ni3p regions for NCA RT and 60 • C electrodes. In the O 1s core region, there is only a slight increase in the surface peaks between the RT and 60 • C when using LiPF 6 . This is much less pronounced than the increase in surface peaks observed for LNA presented in Fig.  6 of the main text. For NCA with only 5% Al-doping, spectral contamination from the Ni 3p leads to difficulty in using the Al 2p core region to examine aluminum surface and bulk environments. In the Ni 3p and Li 1s core regions, we find minimal difference between the RT and 60 • C electrodes associated with the formation of new CEI species. Figure S6: HAXPES measurements of the a) Ni 2p 3/2 and b) O 1s core region using the 9 keV (9.25 keV) Ga Kα lab based X-ray source compared to 6 keV measurements collected at beamline I09 at the Diamond Light Source Ltd. Similar quality data for these oxide cathodes is obtained with the lab based HAXPES while providing probing depths beyond 30 nms.
S-5 Figure S7: O K-edge spectra in surface sensitive total electron yield (TEY) mode for LNA, NCA, and LNC RT and 60 • C electrodes using a) LiPF 6 and b) LiBF 4 . In the LiPF 6 case, the LNA shows a pronounced drop in the Ni 3d-O 2p pre-edge peaks (527 to 533 eV) and a change in the higher energy lineshape associated with the formation of new P-O-F species. In the LiBF 4 case, there is limited change in any of the O K-edge lineshapes except for the LNC electrode that shows a decrease in the pre-edge peak as a result of the 60 • C testing conditions.

S-6
Supplementary Tables Table S1: Aluminum peak fit results Peak fits were conducted on cNCA, NCA, and LNA electrodes with either LiPF 6 or LiBF 4 as the electrolyte salt using the following testing conditions: 1) charged to 3.6 V as a reference, 2) charged to 4.75 V at RT, 3) charged to 4.75 V at 60 • C followed by a 10 hr CV hold, and 4) charged to 4.75 V at 60 • C followed by a 175 hr CV hold. The last testing condition was only used for the cNCA case as using LiPF 6 under these conditions results in a complex exothermic reaction that may influence the cathode-electrolyte reactions. S1

Sample
LiM *May be partially associated with Al 1s satellite peak such as in cNCA prist. with no LiPF 6 contact **Electrodes measured in discharged state (2.7 V) S-7