Introduction

Engineering the surface of battery electrodes is crucial to improve their electrochemical performance as many important reactions happened at this region, including electrolytes decomposition, transition metal dissolution, and Li+ and e− combination [1, 2]. Many inorganic components have been widely explored as the coating agents, like Al2O3 [1, 3], ZnO [4], Al2O3-doped ZnO [5], TiO2 [6], and Li3PO4 [7]. These coatings improve the cycling stability of battery electrodes by alleviating electrolyte decomposition and reducing the generation of irreversible solid electrolyte interphase (SEI). However, simple metal oxides/phosphates coatings could increase impedance over cycling and decrease rate capability because of their insulating nature toward Li–ions [1]. In addition, most of existing coating techniques require high temperature and/or low vacuum conditions that increase the complexity and cost for the battery manufacturing process [8]. Therefore, exploring novel techniques to engineer battery electrodes warrants further investigate.

Electromagnetic (EM) radiation can aid in synthesizing and crystallizing of both inorganic and organic materials due to their ability to reduce the processing temperature and shorten the processing time [9, 10]. One of the most widely studied forms of EM radiation for materials synthesis is microwave radiation (MWR), with frequency in the range of 0.3–300 GHz. Compared to conventional furnace-based synthesis methods, the MWR-assisted technique can synthesize inorganic/organic compounds at much shorter times and with higher yields [11,12,13,14]. Moreover, the MWR-assisted method generates ceramics with unique crystal structures. For example, Nakamura et al. found that TiO2 films grown under MWR exposure contain a different phase composition and increased crystallinity compared to those grown at similar temperatures using a furnace-heating method without EM fields [11].

Despite the widespread exploration of the MWR-assisted method in materials synthesis, applying such filed-assisted techniques in lithium-ion batteries (LIBs) has not attracted much attention. For example, Kim et al. applied the MWR-assisted method to engineer FePO4 on the surface of LiCoO2 cathode electrode and improved its 4.7 V high-voltage cycling stability [15]. However, no follow-up studies were done to explore the effect of other MWR synthesized ceramics or polymers on the electrochemical performance of LiCoO2 or other cathodes. Nevertheless, many inorganic and organic materials have been synthesized using the MWR-assisted technique [16]. Additionally, many experimental parameters, like microwave power and heating rate, affect the final products during the MWR-assisted synthesis, and will therefore affect the electrochemical performance of battery electrodes. Systematically investigating all these parameters and identifying their optimal combination is not trivial. Accord, further investigations are needed to explore the MWR-assisted technique as a surface engineering technique for battery electrodes.

In this study, we applied MWR to engineer TiO2 and ZrO2 on the surface of LiCoO2 electrodes. TiO2 and ZrO2 were selected for the first step of study because of their widely applications [17, 18]. They have also been proved to improve electrochemical performance of many battery electrodes [19]. We chose LiCoO2 because of its high theoretical capacity and good rate performance. In situ synchrotron X-ray diffraction (XRD) technique was applied to monitor the growth of ZrO2 and TiO2 on the surface of LiCoO2 particles. The effect of the coatings on the rate capability and 4.5 V high-voltage cycling stability was then investigated. Our study demonstrates that the MWR-assisted technique can become a versatile tool to engineer the surface of battery cathode electrodes. Exploring other coating materials, including both inorganic ceramics and organic polymers, and applying them to advanced cathode materials warrants further investigations for designing next-generation LIBs with high energy and power density as well as extended lifespan.

Materials and methods

In situ synchrotron X-ray diffraction characterization

MWR-assisted synthesis experiments were performed using an Anton Paar Monowave 300 reactor operating at 2.45 GHz. The picture of the microwave reactor is shown in Fig. S1 in supporting information (SI). LiCoO2 powder was purchased from MTI cooperation. All other chemicals were purchased from VWR cooperation and used without further purification. Three in situ experiments were conducted: (1) MWR heating of LiCoO2 powder, (2) MWR-assisted synthesis of TiO2 coatings on LiCoO2 powder, and (3) MWR-assisted synthesis of ZrO2 coatings on LiCoO2 powder. For the MWR-heating experiment, 200 mg of LiCoO2 powder was dispersed in a 5 mL tetraethyl glycol (TEG) solution and heated up to 250 °C at the rate of 1.5 °C/min (the slowest rate that can be achieved by the microwave reactor), held at that temperature for 10 min, and cooled to 55 °C. The solution was stirred at 650 rpm to avoid hotspot formation. For the other experiments, where TiO2 and ZrO2 coatings were synthesized on LiCoO2 powder, sol–gel solutions based on titanium(IV) butoxide and zirconium(IV) tert-butoxide were used as TiO2 and ZrO2 precursors, respectively [20]. The sol–gel was mixed with TEG in the ratio of 1:4 by volume. 200 mg of LiCoO2 powder was dispersed in a total of 5 ml precursor solution, which was then heated in the microwave reactor using identical experimental conditions as the MWR-heating experiment.

Figure S1 shows a schematic of the in situ synchrotron XRD experiments that were conducted at the 6-ID-D beamline at the Advanced Photon Sources, Argonne National Laboratory. A monochromatic X-ray beam of energy 80 keV was used for diffraction, which allows for X-ray transmission through the thick walls of glass vials. A 2-D PerkinElmer amorphous silicon area detector with cesium iodide scintillator was used to collect 2D diffraction patterns. Because of the high absorption of X-rays by glass vials, an acquisition rate of 1 min was used for data collection. This rate achieved an acceptable signal-to-noise ratio; the ceramic oxide diffraction peaks could be distinguished from the background signal caused by the glass vial and solution while considering the temporal resolution of the chemical reaction. The microwave reactor was modified to include inlet and outlet ports allowing X-rays to pass through and reach the glass vial. Fit2D was utilized to convert 2D detector images to intensity-2 theta angle relationships. High-order polynomial functions were applied to remove the solution and glass vial background via MATLAB [21, 22]. The remaining peaks were then identified and individually fit to Gaussian peaks to obtain peak locations, intensities, and full-width at half maxima (FWHM).

MWR-assisted method to synthesize TiO2 and ZrO2 coating on LiCoO2 surface for LIB electrodes

To engineer TiO2 and ZrO2 on the surface of LiCoO2 powder, 500 mg of LiCoO2 powder was dispersed in a 12 ml solution comprising of 9 ml TEG and 3 ml sol–gel solution. TiO2 synthesis was performed using experimental conditions of 150 °C–200 W–10 min–1 min, where 150 °C represented the maximum temperature, 200 W was the maximum microwave power applied, 10 min was the heating time to reach the maximum temperature, and 1 min was the hold time at the maximum temperature [11]. The vial was cooled down to 55 °C by forced air cooling after the experiment. The experimental conditions for MWR-assisted ZrO2 synthesis were 225 °C–200 W–30 min– min [23]. The obtained powder was washed three times with deionized water and then dried at 80 °C in a vacuum chamber overnight. The dried powder was subsequently annealed at 750 °C for 6 h in air, followed by air cooling to room temperature.

Material characterization

Transmission electron microscopy (TEM) imaging was used to characterize the TiO2 and ZrO2 coatings on LiCoO2 powder. The synthesized powder was dispersed in ethanol, dropped onto TEM grids, and dried overnight before being loaded for imaging. TEM characterization was performed using a FEI Talos F200X operated at 200 kV. Energy-dispersive X-ray spectroscopy (EDS) data were collected by ChemiSTEM technology (X-FED and SuperX EDS) using four silicon drift detectors.

X-ray photoelectron spectroscopy (XPS) measurements were done on a Thermo Fisher ESCALAB 250Xi with an X-ray source of monochromatized Al K α radiation (1486.7 eV). The base pressure was 10−8 Pa, and the spot diameter was 600 μm. Five survey scans with a step size of 1.0 eV were collected, followed by ten high-resolution scans with a step size of 0.1 eV for target elements. The obtained XPS spectra were analyzed by AVENTAGE software with the following parameters: FWHM (eV) = 0.5:3.5 and Lorentzian/Gaussian of 0.3. In the XPS depth profiling analysis, Ar+ ion beam was applied on an area of 3 mm*3 mm for the etching process, followed by XPS data collection for specific elements. The Ar+ gun was operated at 3 keV and the etching time was 15 s for each step. A total of 50 steps were applied on each sample for the depth profiling analysis.

The weight percentages of Co, Ti, and Zr in LiCoO2 powder were measured by inductively coupled plasma mass spectrometry (ICP-MS, Agilent Technologies 7700 Series). Briefly, 7.3 mg of TiO2-coated LiCoO2 powder or 9.6 mg of ZrO2-coated LiCoO2 powder was added into 15-ml centrifuge tube containing 5 ml HCl (37%). The tubes were sealed and rotated on an end-over-end rotator (30 rpm) overnight. Then, the solution was diluted by 2000 times. 6 ml of the solution was filtered through a 0.45-μm filter, and 0.2 ml of trace metal grade HNO3 (69%) was added to acidize the solution. Mixed standards of 0, 10, 20, 50, 100, 250, and 500 ppb for different elements with the same percentage of HNO3 (~ 5%) as the samples were used for calibration.

To study the Co dissolution from LiCoO2 electrodes after cycling, coin cells were disassembled in an argon-filled glove box with O2 and H2O level below 0.5 ppm. All components were immersed in 10 ml dimethyl carbonate (DMC) for 3 days. 2 ml of the solution was then diluted in 6 ml HNO3 (65%). The mixed solution was heated to 120 °C in a vacuum chamber until all liquid disappeared. The remaining white/yellow powder was collected and dissolved in 10 ml deionized water, followed by 30 min of ultrasonic treatment. Finally, 6 ml solution was filtered through a 0.45-μm filter, and 0.2 ml HNO3 (70%) was added to the solution before conducting the ICP-MS measurement.

Electrode preparation, coin cell fabrication, and testing

LiCoO2 powder with and without the TiO2/ZrO2 coating was mixed with carbon black and polyvinylidene fluoride in N-methyl pyrrolidone with a ratio of 8:1:1 by weight. The slurry was then spread on an aluminum foil current collector and dried in a vacuum chamber at 110 °C overnight to remove the solvent and moisture. 14-mm-diameter cathode disks were punched and collected in glass bottles, with a typical mass loading of 4–5 mg. CR-2016 type coin cells were fabricated using these disks as the cathode and lithium chips as the anode. Celgard separators and lithium chips were purchased from MTI Corporation, while electrolytes were purchased from Sigma-Aldrich. The electrolyte was ethylene carbonate/ethyl methyl ethyl carbonate (50/50, v/v) containing 1 M LiPF6. The entire assembly process was carried out in an argon-filled glovebox with O2 and H2O level maintained below 0.5 ppm.

Electrochemical performance of LiCoO2 electrodes was measured using Biologic VMP3 (BioLogic Science Instruments) and LAND battery cyclers (LAND Electronics Co., Ltd.). After assembling, all cells went through a formation process consisting of three continuous cycles at C/10 (1C = 145 mA/g) within 3.0–4.2 V at room temperature. Rate capability was studied by charging cells at C/3 using constant current (CC)–constant voltage (CV) protocol, followed by different discharging rates, including C/10, C/3, 1C, 2C, 5C, and 10C. The cutoff current during the CV charging process was C/100 to ensure that all cells were discharged at the same status. Galvanostatic intermittent titration technique (GITT) measurement was conducted by applying a 30-min galvanostatic charge/discharge pulse (C/10), followed by a 2-h relaxation. Electrochemical impedance spectroscopy (EIS) was tested at room temperature. Before the EIS measurement, cells were charged to 4.0 V using a CC-CV protocol with C/100 as the cutoff current during the CV process. The impedance was then potentiostatically measured by applying an AC voltage of 10 mV amplitude over the frequency range of 100 kHz to 10 mHz. High-voltage cycling test was conducted at room temperature by cycling cells at C/2 within the voltage range of 3.0—4.5 V using a CC protocol for both charging and discharging processes.

Results

Material characterization

Many microwave processing parameters affect the final synthesized products, including maximum temperature, maximum microwave power, heating rate, and hold time at the maximum temperature. The maximum temperature was determined based on the results from the in situ synchrotron XRD experiment, as shown in Fig. 1. Figure 1a shows the temperature and power profiles of the microwave system during the MWR-assisted heating of LiCoO2. The corresponding two-dimensional contour plot for the structure of LiCoO2 powder is displayed in Fig. 1b. At around 115 min, we found significant shifts of two-theta values in the XRD pattern, suggesting a new phase formed because of the microwave heating. Figure S2 suggests the new phase is cubic CoO [24]. The formation of cubic CoO started at 230 °C. Therefore, the maximum temperature for engineering TiO2 and ZrO2 coatings on the surface of LiCoO2 was set below 230 °C to avoid any such phase transformation of LiCoO2 during the MWR processing. Although other parameters, like the heating rate and the hold time, affected the synthesized TiO2 and ZrO2, these parameters were not systematically studied and were chosen based on empirical trials [20]. Optimizing these parameters is not trivial and warrants further investigations.

Figure 1
figure 1

In situ synchrotron X-ray diffraction (XRD) results. a Temperature and power profile during in situ synchrotron XRD experiment for LiCoO2 powder in TEG solution. b The evolution of 2θ values for the diffraction pattern during in situ XRD experiment

Full size image

TiO2 and ZrO2 coatings on LiCoO2 were studied using TEM, scanning transmission electron microscopy (STEM), EDS, and XPS. It has been reported that localized hot spots can form in LiCoO2 due to MWR absorption, which may promote TiO2/ZrO2 nucleation [15]. Figure 2a, b, c shows TEM image of a pristine LiCoO2 particle, a TiO2-coated LiCoO2 particle, and a ZrO2-coated LiCoO2 particle, respectively. No obvious films could be identified in Fig. 2b, while a thin layer is shown in Fig. 2c. However, the thin layer was proven to be carbon rather than ZrO2, as shown in Fig. S3. The EDS results in Fig. 2d, e show that Ti is uniformly distributed on the surface of LiCoO2 powder, while Zr nucleated at a few regions on the edge of LiCoO2 particles. Three different locations were tested to statistically study the distribution of Ti and Zr on LiCoO2 particles, as displayed in Fig. S4 (SI). The result shows that Ti uniformly distributed on the surface of LiCoO2 particles in all cases except one particle (Fig. S4c), while Zr nucleated at a few regions on the edge of LiCoO2 particles in all cases. To further examine the existence of TiO2 and ZrO2, XPS was collected on both TiO2-coated and ZrO2-coated LiCoO2 samples, as shown in Fig. S5. A Ti 2p3/2 peak at 458.4 eV and a Zr 3d5/2 peak at 181.9 eV could be identified, proving the existence of Ti and Zr on the surface of TiO2-coated and ZrO2-coated LiCoO2 powder, respectively. Based on these results, we demonstrate that TiO2 and ZrO2 were successfully synthesized on the surface of LiCoO2 powder using the MWR-assisted technique.

Figure 2
figure 2

TEM images of a a pristine LiCoO2 particle, b a TiO2–LiCoO2 particle and c a ZrO2–LiCoO2 particle. d STEM image and EDS of a TiO2–LiCoO2 particle. e STEM image and EDS of a ZrO2–LiCoO2 particle

Full size image

Table S1 shows that the weight percentage of Co in both TiO2-coated LiCoO2 and ZrO2-coated LiCoO2 powder was very close to the theoretical value of Co in LiCoO2 (60.2%). Additionally, the weight percentages of Ti and Zr in the powder were less than 1%. This result suggests that the weight percentages of TiO2 and ZrO2 in these TiO2 and ZrO2-coated LiCoO2 powders were negligible. Therefore, the TiO2 and ZrO2 coatings should have little effect on the specific capacity of LiCoO2 electrodes.

Influence of MWR-synthesized TiO2/ZrO2 coatings on rate capability of LiCoO2 electrodes

Figure 3 illustrates the effect of TiO2 and ZrO2 coatings on the rate capability of the LiCoO2 electrode. The ZrO2 coating had little effect on the rate capability, while the TiO2 coating improved rate capability. The effect of coatings on the rate performance becomes more obvious at high rate. Figure 3a indicates that the discharge capacity of the LiCoO2 electrode increased by around 50% (from 60 to 90 mA h g−1) after the TiO2 coating at 10C. In comparison, the ZrO2 coating did not affect the discharge capacity of the LiCoO2 electrode in the rate range we measured. Figure 3b compares discharge curves of different LiCoO2 electrodes at C/10, 1C, and 10C. The discharge curves of the ZrO2-coated LiCoO2 and the pristine LiCoO2 electrode almost overlapped each other at all the three C-rates, while the TiO2-coated LiCoO2 electrode behaved differently. At low C-rates (C/10 and 1C), the discharge curves of the TiO2-coated LiCoO2 electrode overlapped the pristine LiCoO2 electrode at the initial stage of the discharging. However, the voltages stayed much higher when it closed to the end of the discharging before quickly dropped to 3.0 V. At the high C-rate (10C), the TiO2-coated LiCoO2 electrode showed higher voltage (~ 0.2 V) during the entire discharging process and 40 mA h g−1 larger specific capacity than the pristine LiCoO2 electrode.

Figure 3
figure 3

Rate performance of different LiCoO2 electrodes. a Specific capacity of LiCoO2 electrodes with respect to discharging rate. Three cells were tested for each case and the result was averaged from all cells. b Comparison of discharge curves of LiCoO2 electrodes at C/10, 1C, and 10C. c Diffusion coefficient of Li in LiCoO2 during lithiation (charging) at around 4.0 V for different electrodes using GITT method. d EIS plots of different LiCoO2 electrodes. The figure inset shows a second-order ECM that is used to fit the impedance

Full size image

The rate performance of the LiCoO2 electrode could be affected by the kinetics of lithium transport in the electrode. We applied GITT to study the effect of TiO2/ZrO2 coating on the transport of lithium in LiCoO2. The details of the measurement process can be found in Fig. S6, which suggests that the diffusion coefficient of lithium in LiCoO2 varied little in the voltage range of 4.0–4.4 V. Additionally, the diffusion coefficients during the charge and discharge processes were similar in value. Therefore, we only measured the diffusion coefficient of lithium in LiCoO2 during the charge process at around 4.0 V for TiO2-coated and ZrO2-coated LiCoO2 electrodes. Figure 3c compares the lithium diffusion coefficient in the three types of LiCoO2 electrodes. The results suggest that the diffusion coefficients were in the same order of magnitude. However, the TiO2-coated LiCoO2 electrode had slightly higher diffusion coefficients (~ 5.5*10−12 cm2S−1) compared with the pristine LiCoO2 electrode (~ 4.0*10−12 cm2S−1). This agreed with the improved discharge performance of the TiO2-coated LiCoO2 electrode seen in Fig. 3b. Figure 4 compares the diffusion coefficient of Li in different LiCoO2 electrodes measured from cyclic voltammetry (CV). The voltage swapping rate was varied from 0.2 to 1.0 mV/s, during which the current was monitored. Figure 4 suggests that the peak current (Ip) linearly increased with the square root of voltage swapping rate (v1/2). By applying Randles–Sevcik equation, we calculated the diffusion coefficient of Li in different LiCoO2 electrodes [6]. The result suggests that both TiO2 and ZrO2 coatings increased the diffusion coefficient of Li. Because the order of magnitude in the difference between these diffusion coefficients was small, more investigations are needed before drawing a clear conclusion.

Figure 4
figure 4

Cyclic voltammetry of a pristine LiCoO2, b TiO2-coated LiCoO2, and c ZrO2-coated LiCoO2 at different scan rates. d The evolution of peak current with respect to scan rate for different samples. Both TiO2 and ZrO2 coating slightly increase the diffusion coefficients of Li in LiCoO2

Full size image

EIS measurements were also conducted to study the effect of these coatings on the impedance of the LiCoO2 electrode. The impedance of a coin cell is very sensitive to the fabrication process. Cells with the same type of LiCoO2 electrode could have different impedance values, as shown in Fig. S7 (SI). However, the overall trend suggests that the TiO2 coating reduced the impedance, while the ZrO2 coatings increased the impedance, as shown in Fig. 3d. To quantitatively analyze these results, the impedances were fitted using a second-order equivalent circuit model (ECM), as shown in the inset in Fig. 3d. The R0 represents the ohmic resistance, the REEI and QEEI represent the resistance and the phase constant element of the EEI layer, and the Rct and Qct represent the resistance and the phase constant element charge-transfer process. The fitted results in Table S2 suggest that the ZrO2 and TiO2 coatings had little effect on R0, but they increased the REEI. Additionally, the ZrO2 coating increased the Rct, while the TiO2 coating decreased the resistance. The effect of different metal oxides coatings on the impedance of the LiCoO2 electrode has been reported before [3,4,5]. The slightly increased REEI was attributed to either poor Li+ transport through the EEI layer after the TiO2/ZrO2 coating or the formation of a thicker EEI layer because of the coating. The decreased/increased Rct of the LiCoO2 electrode after the TiO2/ZrO2 coating could be attributed to different side products on the surface of the electrode, which was generated by reactions between the coatings and electrolytes. The side products could affect the de-solvation process at the interface of EEI and electrolyte as well as affect the charge-transfer process at the interface of EEI and electrolyte, leading to different Rct.

Influence of MWR-synthesized TiO2/ZrO2 coatings on cycling stability of LiCoO2 electrodes

The 4.5 V high-voltage cycling stability of the LiCoO2 electrode was largely improved by the TiO2 and ZrO2 coatings. Figure 5a compares the evolution of specific capacity for the three types of LiCoO2 electrodes during the cycling test. The result shows that the TiO2 and ZrO2 coatings significantly reduced the rate of capacity degradation. The ZrO2 coating improved the capacity remaining from 52.8 to 81.9% after 40 cycles, while the TiO2 coating improved it to 84.4%. Such improvement is battery than many of the existing studies, especially for the TiO2 coating (Table S3). The middle voltage of a cell is the voltage at the half of the discharge capacity, which represents the kinetic properties of the cell. Figure 5b shows that the middle voltage of the pristine LiCoO2 electrode decreased from 3.964 to 3.232 V after 40 cycles, while the ZrO2-coated LiCoO2 electrode decreased from 3.968 to 3.878 V, and TiO2-coated LiCoO2 electrode decreased from 3.968 to 3.924 V. Thus, the TiO2 and ZrO2 coatings reduced the capacity degradation rate and protected the kinetics of the LiCoO2 electrode during the high-voltage cycling test.

Figure 5
figure 5

High-voltage (4.5 V) cycling stability of different LiCoO2 electrodes. a The capacity degradation of LiCoO2 electrodes. Three cells were tested for each type of electrode, and the distribution is shown in the error bar. b The evolution of the middle voltage for different LiCoO2 electrodes. c Discharge curves (C/2) of the LiCoO2 electrodes at new status, after 20 cycles, and after 40 cycles. d Incremental capacity analysis (ICA) for the discharge curves of LiCoO2 electrodes in plot (c)

Full size image

The effect of the TiO2 and ZrO2 coatings on the cycling stability of the LiCoO2 electrode was further investigated by comparing and analyzing discharge curves, as shown in Fig. 5c, d. Figure 5c displays the discharge curves of the three types of LiCoO2 electrodes at different aging statuses. The discharge curves of the pristine LiCoO2 electrode showed a large drop in voltage and specific capacity after 20 (N20) and 40 cycles (N40). Moreover, the discharge plateaus almost disappeared in the discharge curves on N20 and N40. However, both the TiO2 and ZrO2 coatings reduced the decrease in the voltage and the specific capacity of the LiCoO2 electrode after 20 and 40 cycles. Additionally, the coatings maintained the discharge plateaus after the cycling test. Figure 5d compares the incremental capacity analysis (ICA) curves of the three LiCoO2 electrodes at different aging statuses. An ICA peak corresponds to a plateau in a discharge curve, which represents a phase transition process in the LiCoO2 electrode. Before the cycling test, three peaks could be identified for all LiCoO2 electrodes at around 4.15, 4.03, and 3.87 V. The intensity of a peak could be affected by the cathode material and the homogeneous current distribution within the electrode. The higher ICA peak at 3.87 V in the TiO2-coated LiCoO2 electrode suggests that the TiO2 coating improved the homogeneity of current distribution in the LiCoO2 electrode. This further agreed with the better rate performance of the TiO2-coated LiCoO2 electrode seen in Fig. 3. After 20 and 40 cycles, all the ICA peaks of the pristine LiCoO2 electrode disappeared, indicating either LiCoO2 particles were destroyed that led to no phase transformations during the discharge process or largely inhomogeneous current distribution occurred in the electrode during discharge. In comparison, the ICA curves of both the ZrO2-coated and the TiO2-coated LiCoO2 electrodes still showed three ICA peaks after 20 and 40 cycles. However, the locations of these peaks shifted to lower voltages and the intensities decreased after the cycling test, indicating the increase in cell resistance and the inhomogeneity of current distribution in these electrodes.

EIS measurements show that the TiO2 and ZrO2 coatings reduced the increase in the overall impedance of the LiCoO2 electrode during the cycling test. Figure 6a compares the impedances of the three types of LiCoO2 electrodes after 40 cycles. The charge-transfer resistance (Rct) in the pristine LiCoO2 electrode became much larger than that in the TiO2 and ZrO2-coated LiCoO2 electrodes. This phenomenon has been reported inhibit the increase in Rct in LiCoO2 electrodes during cycling tests [25]. A second-order ECM was applied to quantitatively analyze the impedance, and the fitted results are listed in Table S2. The Rct of the pristine LiCoO2 increased by almost two orders of magnitude (80 times) after the cycling test, while the Rct of the ZrO2-coated and TiO2-coated LiCoO2 only increased by around 2.5 and 1.5 times, respectively. This might be from the formation of a passivation film on the surface of LiCoO2, like LiF, LixPFy, and LixPFyOz-type of compounds, which limited the charge-transfer process of Li+ at the cathode-electrolyte interface [26, 27]. The slow increase in Rct in the TiO2/ZrO2-coated LiCoO2 electrode suggested that the TiO2 and ZrO2 coatings could inhibit the formation of these compounds. Table S2 also suggests that the R0 did not increase after the cycling test for all the three LiCoO2 electrodes, indicating that the electron transport pathways were maintained in these electrodes during the cycling test. Additionally, the REEI slightly decreased for all the three types of LiCoO2 electrodes. The decreased REEI could be caused by the formation of cracks on the LiCoO2 electrodes during the cycling test that increased the total surface area of the electrodes [26].

Figure 6
figure 6

Kinetics of LiCoO2 electrode after the cycling test. a EIS results for the three types of LiCoO2 electrodes. b The diffusion coefficient of Li in LiCoO2 electrodes

Full size image

GITT data showed that the TiO2 and ZrO2 coatings helped maintain the kinetics of lithium transport in the LiCoO2 particles during the cycling test. Figure 6b shows that the TiO2-coated LiCoO2 electrode had the highest lithium diffusion coefficient (~ 4.2*10−12 cm2 S−1) after cycling, which changed little from its initial status (~ 5.5*10−12 cm2 S−1). Similarly, the lithium diffusion coefficient in the ZrO2-coated LiCoO2 electrode decreased little after cycling (from ~ 4.0*10−12 cm2 S−1 to ~ 2.3*10−12 cm2 S−1). In comparison, the lithium diffusion coefficient in the pristine LiCoO2 decreased by more than an order of magnitude after the cycling test (from ~ 4.0*10−12 to ~ 2.3*10−13 cm2 S−1). This result suggests that the pristine LiCoO2 particles went through severe damage during the cycling test, such as surface reconstruction caused by Co dissolution and channel blocking caused by the change of crystal structure on the surface [28]. These damages were reduced, if not eliminated, by the TiO2 and ZrO2 coatings according to the well-maintained lithium diffusion coefficients seen in the TiO2/ZrO2-coated LiCoO2 electrodes.

The TiO2 and ZrO2 coating also reduced the Co dissolution from the LiCoO2 electrode into the electrolyte, which is considered as one of the mechanisms for capacity degradation of LiCoO2 during cycling [29]. By applying the treatment described in the methodology section, we measured the concentration of Co element in the final electrolyte solution using ICP-MS technique. After the cycling test, the ratio of the dissolved Co to the total mass of the LiCoO2 electrode was 0.76%, 0.12%, and 0.06% for the pristine LiCoO2 electrode, the TiO2-coated LiCoO2 electrode, and the ZrO2-coated LiCoO2 electrode, respectively. Thus, the TiO2 and ZrO2 coating inhibited the dissolution of Co, which further explained the improved capacity remaining of the LiCoO2 electrode after these coatings.

Discussion

It is reported that the degradation of LiCoO2 during cycling (below 4.5 V) is attributed to surface degradation effects rather than bulk degradation [30, 31]. This suggests the degradation rate could be related to the surface area of LiCoO2 particles, i.e., the radius of the particle. Figure 7 summarizes the capacity fade rate of LiCoO2 cycled at 4.5 V in recent papers. The figure suggests that the capacity fade rate increases with the decrease in the primary particle size. In our study, the LiCoO2 particle size is 2 μm and the degradation rate of pristine LiCoO2 is 1.18%/cycle. Our experiment data follows the trend of LiCoO2 capacity fade rates reported in recent studies.

Figure 7
figure 7

Comparison 4.5 V high-voltage cycling performance of pristine LiCoO2 with different particle sizes. Data are collected from recently published paper, as listed in Table S3. Our experiment data is highlighted by red, where LCO stands for pristine LiCoO2 electrode and ZrO2/TiO2–LCO stands for ZrO2/TiO2-coated LiCoO2 electrode

Full size image

Additionally, our study shows that the degradation rate of LiCoO2 is largely reduced by TiO2 and ZrO2 coatings. Table 1 compares the effect of TiO2 and ZrO2 coatings on LiCoO2 4.5 V cycling performance via different techniques. Compared to existing literature, the ZrO2-coated LiCoO2 in our study shows comparable cycling performance, while the TiO2-coated LiCoO2 shows much better cycling stability.

Table 1 Comparison of LiCoO2 cycling stability at 4.5 V with TiO2 and ZrO2 coatings
Full size table

XPS is a surface-sensitive spectroscopic technique that reveals not only the composition of the sample surface, but also the oxidation states of different elements present on the sample surface. Accordingly, we applied XPS to investigate the effect of the TiO2 and ZrO2 coatings on the electrochemical performance of the LiCoO2 electrode. The TiO2 coating affected the oxidation status of Co on the surface of LiCoO2, while the ZrO2 coating had little effect. Figure 8a compares the binding energy of Co 2p peaks in LiCoO2 powder before and after the TiO2/ZrO2 coating via the MWR-assisted synthesis, followed by the heat treatment. The TiO2-coated LiCoO2 powder before the heat treatment is also shown for comparison. Two XPS peaks at 780.0 eV and 780.46 eV were assigned to Co 2p3/2. The lower binding energy was from Co3+ in LiCoO2, while the higher binding energy was from Co2+ in Co(OH)2 that formed on the surface of LiCoO2 during air exposure [27]. The results show that the ZrO2 coating did not affect the binding energies of Co 2p3/2, while the TiO2 coating increased the Co3+ peak from 780.0 to 780.18 eV and increased the Co2+ peak from 780.46 to 780.63 eV. The increased binding energies for Co 2p3/2 were not found in the TiO2-coated LiCoO2 powder before the heat treatment, suggesting the heat treatment rather than the microwave processing affected the surface state of LiCoO2 in the TiO2-coated LiCoO2 powder. The little effect of the microwave processing on LiCoO2 powder was also supported by analyzing the in situ XRD data in Fig. S9. The data suggests that the lattice parameter of LiCoO2 did not change during the MWR-assisted synthesis of TiO2 on the surface of LiCoO2.

Figure 8
figure 8

X-ray photoelectron spectroscopy (XPS) measurement for different samples. a XPS spectroscopy of Co 2p in pristine LiCoO2, TiO2-coated LiCoO2, and ZrO2-coated LiCoO2. b The XPS spectroscopy of Ti 2p in TiO2-coated LiCoO2 electrodes. The TiO2–LiCoO2* represents the TiO2-coated LiCoO2 without heat treatment. Measurement data (dots) are fitted by several individual spectra (colored regions). The combined spectra from these color shaded regions are shown as an envelope that matches well with experimental data (dots)

Full size image

Figure 8b compares the binding energy of Ti 2p in the TiO2-coated LiCoO2 powder before and after the heat treatment. The results display that the Ti 2p2/3 peak decreased from 458.40 to 458.06 eV after the heat treatment, suggesting that the heat treatment promoted interactions between Ti and Co on the surface of LiCoO2. One possible mechanism is that Ti diffused into the surface of LiCoO2 lattice by either forming Ti interstitial or substituting Co [36]. This process affected the oxidation state of Co and Ti atoms and thus changed the binding energies of Co 2p3/2 in LiCoO2 and Ti 2p3/2 in TiO2. However, the atomic radius of Zr (230 pm) is much bigger than Ti atom (147 pm), making Zr atom harder to diffuse into LiCoO2 lattice. We believe that such a mechanism could also explain the different effects of the TiO2 and ZrO2 coatings on the diffusion coefficient of lithium in LiCoO2 (Fig. 3c) as well as the rate capability of LiCoO2 electrodes (Fig. 3a, b).

To understand the improved high-voltage cycling stability in the TiO2-coated LiCoO2 electrode, XPS depth profiling was conducted for LiCoO2 electrodes with and without the TiO2 coating, after 40 cycles. The results are shown in Fig. 9. The F 1s curve at 0s (no etching) could be roughly allocated into two peaks. The one at higher binding energy (~ 688 eV) was from fluoroorganic species, like C–F in PVDF, P–F in PF6−, as well as other organic side products that contain fluoride, while the one at lower binding energy (~ 685 eV) was attributed to LiF [27]. Compared with the fluoroorganic species peak, the Li–F peak in the pristine LiCoO2 electrode was very weak and only showed up after 450 s of etching. Since many reports show that the passive film formed on battery electrodes is composed of an inner inorganic layer and an outer organic layer, the weak Li–F peak at 0 s indicated the formation of a relatively thick organic passivation film on the surface of the pristine LiCoO2 after 40 cycles [37, 38].

Figure 9
figure 9

XPS measurement of F 1s binding energies for a pristine LiCoO2 electrode, and b TiO2-coated LiCoO2 electrode after 40 cycles. Dotted lines are used to label the possible fluoride compounds

Full size image

In comparison, Fig. 9b shows strong Li–F peak before the etching treatment for the TiO2-coated LiCoO2 electrode, suggesting a much thinner organic layer formed on the TiO2-coated LiCoO2 electrode after the cycling test. Additionally, another peak at ~ 685.5 eV could be seen in the TiO2-coated LiCoO2 electrode, especially after a few steps of the etching treatment. This peak has been assigned to metal fluoride in other studies and could be assigned to Ti–F here because TiO2 is a Lewis base that reacts with trace HF in electrolyte [6]. After 150 s of the etching treatment, the Li–F peak disappeared and left only the Ti–F peak, implying that Ti–F dominated the fluoride components on the surface of TiO2-coated LiCoO2 electrode. The formation of Ti–F consumed HF, which can be harmful to the cycling performance of battery electrolyte by leading to side reactions. Additionally, this metal fluoride layer is believed to have a higher chemical stability than the side products formed on the surface of the pristine LiCoO2 [6]. This explained the phenomenon of less Co dissolution in the TiO2-coated LiCoO2 electrode after the cycling test, which led to much better high-voltage cycling stability of the LiCoO2 electrode after the TiO2 coating.

Conclusions

This study explores the microwave radiation (MWR)-assisted synthesis as a surface engineering technique to grow TiO2 and ZrO2 coatings on the surface of LIB electrodes like LiCoO2. This technique requires relatively short synthesis times and low synthesis temperature compared to traditional sol–gel and hydro/solvothermal synthesis methods. By selecting appropriate coating materials, like TiO2, we altered the surface characteristics of LiCoO2 and improved its electrochemical performance. The capacity remaining was improved from 52.8 to 81.9% and 84.4% after 40 cycles by the ZrO2 and TiO2 coating, respectively. These results suggest that further attention should be given to synthesis under electromagnetic fields, because they provide promising, alternative routes to engineer the surface of battery electrodes for improving their performance. Exploring the synthesis of other materials via this process, including both inorganic ceramics and organic polymers, will provide a novel pathway toward engineering advanced LIBs with high energy and power density, as well as extended lifespan.