3.1 Material characterization
Sol-gel synthesized samples were analyzed for structural changes and impure phases using the XRD. In
Fig. 1a, phases corresponding to all major peaks were identified and found to match ICSD-10199(LiNiO
2) suitably. This confirms the layered structure of the LiNi
0.8Co
0.1 Mn
0.1O
2(NCM811) cathode active material with a hexagonal α-NaFeO
2 structure belonging to the R-3m space group. No significant shifts in peaks positions were observed after Al coating indicating that coating occurred on the surface only without modification of the bulk structure. The XRD peak for Al was rendered insignificant by the thin layer of coating. Thin layer coating is essential for efficient surface modification effects. No impurities other than NCM811 phases were observed.
The molar stoichiometry of synthesized NCM811 was further confirmed using ICP-OES data. Based on the ICP-OES data presented in
Table 1, a similar stoichiometry of elements was confirmed as desired, proving the successful synthesis of NCM811 via the sol-gel method. To understand the structural morphology, SEM images were captured, as shown in
Fig. 1 b. The particles were approximately 300–500 nm in size with few agglomerates and rough surface textures. The preparation technique is a critical determinant of the shape and size of the synthesized particle. In contrast to solid state synthesis where the distribution range of the particle size is wide, the solgel synthesis has relatively better control over size range. Similarly sized particles exhibit better capacity than their compact counterparts in many cases [
24,
25].
To further analyze the coating layer thickness and its morphology, TEM images were obtained. The image in
Fig. 1c image affirms the thin layer of Al
2O
3 coating above NCM811 with a uniform thickness of approximately 3.4 nm and lattice fringes were clearly observed with an interplanar distance of 0.47 nm at (003) plane. Notably, the coating was formed ex-situ by ALD deposition. Typically an amorphous and rigid coating increases by approximately 2 in thickness per cycle and can grow slightly larger in the presence of H
2O [
22,
23,
26]. Thus, coating layer thickness observed after the 10 cycles is consistent with that reported in earlier reports. The thickness of the coating is critical, as greater thickness can hinder intercalation activity and lesser thickness can have no effect. The mapping of elements Al, Ni, Co, Mn, and O is presented in d. A high concentration of Al mapping toward outer layers of particles confirms the surface level coating of Al upon NCM.
In
Fig. 2, XPS spectra for coated LiNi
0.8Co
0.1Mn
0.1O
2 and pristine LiNi
0.8Co
0.1Mn
0.1O
2 (inset), shows the surface presence of Al
2O
3 in ALD deposited electrode. The only Al 2p peak located at 74.1 eV confirms [
27] the presence of only Al
2O
3 without other oxide impurities. This confirms the outer coating layer displayed in TEM consisted of amorphous Al
2O
3 only. The samples were further analyzed with respect to their electrochemical behavior.
3.2 Electrochemical characterization
Initial charge/discharge curves are presented in
Fig. 3a and 3b. Charge/discharge tests were conducted at 1 C rate and 180 mA g
−1) in the potential range of 2.8–4.3 V. The initial charge capacity of pristine LiNi
0.8Co
0.1Mn
0.1O
2 (P-NCM) was 195 mAh g
−1 and during initial discharge, P-NCM showed a capacity of 173 mAh g
−1. Meanwhile, ALD coating was performed systematically and with increasing numbers of cycles, the increase in capacity was observed. Significant increase in capacity was observed after 10 cycles of Al
2O
3 coating on the LiNi
0.8Co
0.1Mn
0.1O
2 (A-NCM) sample, yielding a capacity of 200 mA g
−1 during charging and 176 mA g
−1 during discharging. The critical coating thickness was achieved at 10 cycles beyond which the coating acted as a barrier, hindering Li ion diffusion and leading to capacity fade. Upon comparison of the voltage plateau for both samples during discharging, A-NCM showed higher discharging voltage, indicating that the coated sample had more energy density and higher capacity than pristine sample did. This phenomenon is not uncommon and was observed in a previous study [
25]. At 2 C (
Fig. 3b), pristine (P-NCM) displayed an initial charge and discharge capacity of 176 and 152 mAh g
−1, respectively. In contrast, the coated samples could deliver charge and discharge capacities of 183 and 157 mAh g
−1, respectively. Despite the decrease in coulombic efficiency after coating (86.4%–85.8%), the initial charge and discharge capacity increased, as did the charge/discharge at 1 C, where coulombic efficiency decreased from 88.7% to 88.0%.
Differential capacity versus voltage (dQ/dV) was calculated using the initial charge–discharge curves at 1 C, as shown in
Fig. 3c. The peak at 3.73 V in the pristine sample indicated the redox reaction of Ni
2+–Ni
4+. However, in A-NCM, the peak was shifted to a lower voltage, which indicated that this reaction occurred at a lower voltage, resulting in lower polarization and resistance after coating. This result also demonstrated the reason A-NCM showed higher initial capacity and lower discharging voltage plateau at 1 C compared to P-NCM.
Rate capabilities were measured at various C-rates within a voltage range of 2.8–4.3 V at 25°C in
Fig. 3d. The ALD coated sample yielded 196, 193, 182, 169, 153, and 141 mAh g
−1 at 0.1, 0.2, 0.5, 1, 2, and 3 C, respectively; for one back cycle to 0.1 C, a capacity of 188 mAh g
−1 was measured. However, for PNCM, capacities of 181, 176, 162, 147, 129, 113, and 167 mAh g
−1 was measured at 0.1, 0.2, 0.5, 1, 2, and 3 C, respectively, indicating that the coated sample A-NCM showed higher capacity than P-NCM at all C-rate intervals. At a relatively higher C-rate, a significant difference in capacity between pristine and coated samples was observed due to thin and uniform coating, and its synergistic effect on improving the capacity of NCM. When cycled back to 0.1 C, the capacity of A-NCM almost recovered to the previous capacity at 0.1 C, indicating higher reversibility, but in P-NCM, the capacity fade was inevitable. The effect of the coating on discharge capacity is adequately illustrated in
e where the discharge capacity is normalized to the initial discharge capacity at 0.1 C. This yields a magnitude of capacity loss with an increase in C-rate. For P-NCM, the capacity fade from 0.1 to 1.0 C, which is a 10-fold higher C-rate, resulted in significant reduction in capacity (up to 19%) and further worsened to 38% at 3.0 C. Conversely, the coated sample displayed superior Li diffusion with higher capacity at a higher C-rate. Coated sample A-NCM merely lost 14% of its initial capacity from 0.1–1.0 C rate and 28% when the rate was increased to 3.0 C, which is only 2/3 of its pristine counterpart. However, translating it to real capacity values, for the A-NCM with already significantly higher capacity, the resultant capacity at 3 C was significantly higher. Upon cycling back to 0.1 C, the coated A-NCM could retain up to 96% of its initial capacity, whereas the P-NCM could retain a mere 92%. This significant reversibility was possible owing to the Al
2O
3 protection layer which had a synergistic effect with NCM and high capability for faster Li ion diffusion during fast charge–discharge processes with high reversibility.
Cycle performances of P-NCM and A-NCM at 1 and 2 C along with coulombic efficiencies are shown in
Fig. 4. In
Fig. 4a, the discharge capacity of A-NCM maintained a rate of 127 mAh g
−1 after 100 cycles, retaining 72% of the initial capacity. P-NCM could retain only 58% of its initial capacity at 101 mAh g
−1 after 100 cycles. This indicates that the cycle performance of Ni-rich cathode significantly increased after Al
2O
3 coating using ALD. Here the outer layer of Al
2O
3 acts a scavenger by reacting with HF as follows:
This scavenging activity had a tremendous effect on cycle stability of A-NCM. whereas in A-NCM, during the end of the charging process, the Ni
4+ state was present. This Ni
4+ metal state in octahedral 3a sites, moved to Li sites (octahedral 3b) at a delithiated state. Ni
4+ ion is also highly irreversible and thus its capacity loss is inevitable [
19]. Conversely, in pristine P-NCM, the NCM is directly exposed to electrolyte and creates an interface. This interface is highly reactive to the HF that is present in the electrolyte resulting in decomposed products such as LiF, Li
xPF
y and Li
xPO
yF
2 along with other gaseous products such as CO
2 and alkanes [
19,
28]. The thin and uniform coating layer of Al
2O
3 on electrode material, as shown by TEM and XRD images acts as a protection layer from electrolytes, suppressing the side reactions that decay the capacity while cycling. Even at high currents of 2 C, the retention of A-NCM was 72.2%, which was higher than that of P-NCM. This demonstrates the efficacy of ALD coating in electrode material even in harsh conditions such as high C-rates, and its relevance to practical application such as fast charging.
EIS spectra were recorded to observe the changes in resistance after coating and cycling in . The EIS curves are a good fit for the calculated and an equivalent circuit proposed thereof.
Fig. 5a and b show the state before and after 100 cycles at 1 C, the values of R
s (electrolyte resistance) from high to medium frequency range representing mainly the electrode–electrolyte interface, R
sf (surface resistance) and R
ct (charge-transfer resistance). Each fitted value of resistance is listed in
Table 2. Before cycling, R
ct of ALD coated A-NCM was smaller than that of pristine P-NCM. The smaller R
ct could be the main reason behind A-NCM showing higher initial charge/ discharge capacity at 1 C. Logically, the amorphous coating should be an additional resistance component to the pristine surface. However, EIS spectra exhibited antonymous behavior because of the perfect synergistic sync between the NCM surface and the Al
2O
3 coating layer. Smaller R
ct in A-NCM was observed when compared to P-NCM even after cycling. An additional component of surface resistance, R
sf, appeared, indicating the plaguing reactions occurring on the surface leading to capacity decay. A comparison of the EIS spectra of A-NCM in 1 C and 2 C after cycling (
Fig. 5b and c) showed that the R
sf remained almost the same, indicating minimal and consistent surface deterioration in coated A-NCM samples. In contrast, in P-NCM, the R
sf increment was significant, which is in good agreement with the faster and high magnitude capacity decay of P-NCM material. Thus, in addition to acting as a barrier from parasitic reactions, the Al
2O
3 coating layer enhances capacity by improving charge-transfer kinetics and thus improving overall energy density. A schematic depicting the enhanced stability effect of ALD coating in NCM is provided in
Fig. 6.