A study on Ti-doped Fe3O4 anode for Li ion battery using machine learning, electrochemical and distribution function of relaxation times (DFRTs) analyses

Abstract

Among many transition-metal oxides, Fe3O4 anode based lithium ion batteries (LIBs) have been well-investigated because of their high energy and high capacity. Iron is known for elemental abundance and is relatively environmentally friendly as well contains with low toxicity. However, LIBs based on Fe3O4 suffer from particle aggregation during charge–discharge processes that affects the cycling performance. This study conjectures that iron agglomeration and material performance could be affected by dopant choice, and improvements are sought with Fe3O4 nanoparticles doped with 0.2% Ti. The electrochemical measurements show a stable specific capacity of 450 mAh g−1 at 0.1 C rate for at least 100 cycles in Ti doped Fe3O4. The stability in discharge capacity for Ti doped Fe3O4 is achieved, arising from good electronic conductivity and stability in microstructure and crystal structure, which has been further confirmed by density functional theory (DFT) calculation. Detailed distribution function of relaxation times (DFRTs) analyses based on the impedance spectra reveal two different types of Li ion transport phenomena, which are closely related with the electron density difference near the two Fe-sites. Detailed analyses on EIS measurements using DFRTs for Ti doped Fe3O4 indicate that improvement in interfacial charge transfer processes between electrode and Li metal along with an intermediate lithiated phase helps to enhance the electrochemical performance.

Ultrafine Mix-Phase SnO-SnO2 Nanoparticles Anchored on Reduced Graphene Oxide Boost Reversible Li-Ion Storage Capacity beyond Theoretical Limit

“…A similar contribution for the surface film around 1000 Hz has also been reported for other conversion anodes such as Fe 3 O 4 etc. using the DFRT analysis. − In our study, the effective resistances of 30.69 and 6.47 Ω are obtained for the dual peaks in the electrochemically activated cell, whereas the fresh cell showed a single peak resistance of 66.54 Ω. The shifting of the peak toward high frequencies from fresh to activated samples implies the Li-ion migration process in a short span of time.…”

“…As observed from Figure 3f, the impedance data after the CV measurement contain two peaks within the KK compatible regime, one peak around 357 Hz and another one around 1335 Hz, justifying two contributions from the surface film as observed due to Li ion migration through the electrode surfaces. 42,43 However, the pristine cell contains only one peak around 265 Hz with a single contribution of surface film effect. Thus, the electrochemical activation could facilitate the ion movement, and as a result different types of ion migration processes are observed for the former sample.…”

AI for Nanomaterials Development in Clean Energy and Carbon Capture, Utilization and Storage (CCUS)

“…Selecting and designing appropriate electrode materials is of paramount importance in enhancing the performance of an LIB. Critical design criteria include high areal capacity and stability, which are dependent on the intrinsic characteristics (e.g., redox potential, and layer thickness) of electrode materials. − Among them, the crystal structure of an electrode has a significant impact on its physical and chemical properties, and accurate prediction of crystal system is therefore instrumental in discovering and optimizing electrodes in LIBs. , Shandiz et al have utilized several machine learning classifiers including KNN, ANN, SVM, and RF in predicting the three major crystal systems (i.e., monoclinic, orthorhombic, and triclinic) of silicate cathodes with Li-Si-(Mn, Fe, Co)-O compositions. Specifically, with material properties such as formation energy, band gap, number of sites, and volume of the unit cell as the model inputs, the ensemble classifiers (e.g., RF) were found to provide the most accurate predictions among other algorithms (Figure b­(i)), and the volume and number of atoms and volume in a unit cell of the crystal were found to be the dominant descriptors in the classification (Figure b­(ii)).…”

Basic guidelines of first-principles calculations for suitable selection of electrochemical Li storage materials: a review

“…There is also numerous battery research incorporated with machine learning (ML) because ML is playing an increasingly significant role in lithium-ion battery research and development. 27–32 Safety, higher energy/power density, improved electrochemical stability, and a wider electrochemical window have made solid-state electrolytes (SSEs) popular in all-solid-state lithium-ion batteries (ASSBs). SSEs have weaker ionic conductivity, complex interfaces, and unstable physical properties.…”