Ultrafine MoS₂ Nanosheets Vertically Patterned on Graphene for High-Efficient Li-Ion and Na-Ion Storage

Introduction

The increasing social concern on natural resources (e.g., fossil fuels) and environmental problems (e.g., climate change) has triggered extensive research activities on the development of sustainable and renewable resources of solar energy and wind power (Yang et al., 2011). Accordingly, energy storage systems such as electrochemical energy storage (EES) devices are indispensable to promote the utilization of these intermittent energy resources (Wang et al., 2015). In particular, rechargeable battery technologies, including Li-ion batteries and Na-ion batteries, stand at the forefront of the EES devices due to their high energy density (Yabuuchi et al., 2014). To meet the ever-increasing demands of energy/power densities for applications ranging from portable consumer electronics and electric vehicles to large-scale smart utility grids, it is of great necessity to improve the battery performance by developing advanced electrode materials.

Two-dimensional (2D) nanomaterials have drawn significant attention in recent years and have shown great promise in energy storage applications (Chhowalla et al., 2013; Anasori et al., 2017). Transition metal dichalcogenides (TMDs) represent a class of promising 2D nanomaterials for use in Li-ion/Na-ion batteries (Xu et al., 2014; Yun et al., 2018; Zhu et al., 2021). Among the well-studied TMDs, MoS₂ has been considered to be a prominent candidate due to its high specific capacity and low cost (Wu et al., 2020). MoS₂ shows analogous layered structure to graphite, which is composed of stacked S-Mo-S layers through van der Waals interactions (Liu et al., 2021). The unique layered 2D nanostructure permits easy insertion/extraction of ions along the interlayer channels, provides large surface area for energy-storage sites, and reduces the electron/ion transport distances compared with the bulk counterparts (Wang et al., 2017; Yu et al., 2017). However, the structurally stable 2H-phase of MoS₂ suffers from low electrical conductivity and large volume change during charge/discharge processes, which collectively lead to poor rate performance and short cycle lifetime. To mitigate these shortcomings, the most effective strategy is hybridizing MoS₂ nanostructures with conductive carbonaceous materials, such as graphene (Teng et al., 2016; Sun et al., 2020), carbon nanotubes (Chen et al., 2016; Wang et al., 2016), carbon nanofibers (Zhou et al., 2014; Wang et al., 2017), and carbon nanoboxes (Yu et al., 2015). The carbonaceous materials not only afford high-speed conductive pathways for improving the electrical conductivity of MoS₂, but also function as flexible substrates for alleviating volume excursion, thereby enabling fast electrochemical kinetics and robust structural stability. 2D graphene (nanosheets) is of particular interest due to its large surface area, high electrical conductivity, and strong mechanical strength (Shao et al., 2015). A great deal of research activities have been made in fabricating various MoS₂/graphene nanocomposites for high-performance Li-ion/Na-ion batteries (Geng et al., 2017; Wang et al., 2017; Pan et al., 2018; Li et al., 2019). However, the Li/Na-ion storage properties of these nanocomposites are strongly associated with their hybrid morphology. The uniform anchoring of MoS₂ nanostructures tightly on the planar graphene could facilitate fast electrode kinetics and strong structural stability (Wang et al., 2020). Nevertheless, the negatively-charged graphene oxides are prone to be difficult to have a uniform coupling with the Mo-containing anions (e.g., MoO₄²⁻ and Mo₇O₂₄⁶⁻) due to the electrostatic repulsion albeit with the assistance of cationic surfactants (Wang et al., 2013). Moreover, it is important to point out that the lithiation/sodiation of MoS₂ starts with the intercalation of Li/Na ions into the interlayers due to the large open 2D nanochannels (Shan et al., 2016; Deng et al., 2017). That is to say, edge orientation of MoS₂ on graphene may play a significant influence on the electrochemical performance. Therefore, it is highly desired to engineer the MoS₂ nanosheets with preferential edge orientation and uniform aligning on the graphene to yield a maximum harvest of energy storage.

In this work, a rational strategy is developed for preparing a hierarchical 2D heteroarchitecture of the MoS₂@graphene nanocomposite through a facile solvothermal method. The core issue is the use of acidic Mo precursors that allow for ultrafine MoS₂ nanosheets patterned on the planar graphene surface, thus constructing a unique sheet-on-sheet nanostructure. Impressively, the unique hybrid morphology could achieve a strong synergistic effect of enlarging electrode/electrolyte interfaces, reducing ion/electron transport distances and buffering large volume change of the electrode. These electrochemical characteristics are of great advantage in enhancing the Li/Na-ion energy storage with large specific capacity, superior rate capability, and outstanding cycling durability.

Materials and Methods

Materials Synthesis

The preparation of MoS₂/reduced graphene oxide (MoS₂@rGO) was achieved by a solvothermal treatment. In specific, graphene oxide (GO) was obtained by a modified Hummers method. Then a certain amount of GO powder (20 mg) was dispersed into ultrapure water (40 ml). Subsequently, phosphomolybdic acid hydrate (100 mg) was dissolved into the GO dispersion followed by adding L-cysteine. The resulting mixture was then loaded in an autoclave and allowed to experience a solvothermal treatment of 180°C/15 h. After filtration, water rinse, and drying, black products were obtained. The as-obtained powders were finally treated by an annealing procedure of 800°C/2 h in a tube furnace under an inert N₂ gas flow. For comparison, the Mo precursor of phosphomolybdic acid was replaced by the conventional ammonium molybdate tetrahydrate ((NH₄)₆Mo₇O₂₄·4H₂O, AMT) to prepare the control sample of MoS₂@rGO-C through the similar solvothermal/annealing technique. A pure MoS₂ sample was also synthesized by the same procedure.

Materials Characterization

X-ray diffraction (XRD) (Cu Kα radiation) and a Raman spectrometer (Renishaw inVia) were used to identify the structural properties. The elemental information was recorded by X-ray photoelectron spectrometry (XPS) on an ESCALAB 250Xi. The thermal behavior of the composites was examined by thermalgravimetric analysis (TGA, Mettler Toledo TGA/DSC 3+). A Micromeritics ASAP 2020 was employed to analyze the N₂ adsorption/desorption isotherm and the corresponding porous characteristics. Morphological and structural analysis was conducted by transmission electron microscopy (TEM, Talos F200X) and scanning electron microscopy (SEM).

Electrochemical Measurements

The Li/Na-ion storage properties of the samples were evaluated based on CR 2016 coin cells, which were packed in an argon-filled glove box. A sticky slurry was prepared by dispersing 70 wt% of MoS₂-based materials, 20 wt% of conductive carbon black, and 10 wt% of polyvinylidene fluoride in a N-methyl-2-pyrrolidinone solvent. The above slurry was spread on a copper foil, and circular discs were punched out for working electrodes. The mass loading of the active materials (MoS₂/graphene or MoS₂) is about 1.2–1.5 mg⁻². Coin-type cells were packed by placing the counter electrode of Li foils, a Celgard 2,400 microporous polypropylene separator, and working electrodes together. The electrolyte was 1 M LiPF₆ in a solvent of ethylene carbonate, dimethyl carbonate, and ethylmethyl carbonate with a volume ratio of 1:1:1. The assembly of Na-ion batteries was based on Na discs and the glass fiber filter paper as the separator. The electrolyte was 1 M NaPF₆ in ethylene carbonate/diethyl carbonate (1:1 by volume). A CHI 660E electrochemical workstation was used to measure the cyclic voltammetry (CV) and electrochemical impedance spectra (EIS). NEWARE battery testing channels were implemented to obtain the galvanostatic charge/discharge profiles and the cycling performance at various current rates.

Results and Discussion

Figure 1A presents the panoramic SEM image of the as-prepared MoS₂@rGO nanocomposite. It is noticeable that the nanocomposite complies with the 2D morphology of graphene nanosheets without particle aggregation. A high-resolution snapshot (Figure 1B) indicates that ultrafine MoS₂ nanosheets are tightly and uniformly patterned on the planar graphene surface, thus building a unique 2D/2D sheet-on-sheet architecture. The homogeneous anchoring of 2D MoS₂ nanostructures onto the rGO nanosheets is validated by the even distribution of elemental C, Mo, and S from the energy-dispersive X-ray (EDX) mapping (Figure 1C). TEM imaging (Figure 1D) presents the hybrid sheet-on-sheet configuration, wherein the nanostructured MoS₂ deposits are clearly patterning on the whole rGO nanosheet surface. The associated high-resolution TEM (HRTEM) observation (Figure 1E) reveals vertical anchoring of ultrafine MoS₂ nanosheets with a thickness of 5–10 layers on the rGO. The lattice fringe spacing of 0.68 nm corresponds well to the distance of (002) crystal facets, which surpasses the theoretical value of MoS₂ (0.62 nm). It is important to note that the (002)-edge-oriented and expanded MoS₂ nanosheets are of great benefit in boosting the electrode reaction kinetics because the electrochemical lithiation/sodiation of 2D MoS₂ generally occurs with an initial insertion of Li/Na ions in between the (002) interlayers.

FIGURE 1. (A,B) SEM images of the MoS₂@rGO composite. (C) EDX elemental mapping of C, Mo, and S. (D) TEM and (E) HRTEM images of MoS₂@rGO composite.

XRD was performed to identify the crystal phase of the as-prepared samples. As shown in Figure 2A, the well-defined diffraction peaks with 2θ at around 14°, 33°, 39°, 49°, and 59° belong to the (002), (100), (103), (105), and (110) planes of 2H-type MoS₂ hexagonal phase (JCPDS # 75–1539), respectively (Kong et al., 2014). In addition, there is a broad and low-intensity peak centered at 26°, which can be ascribed to the (002) facet of the amorphous carbon structure (Wang et al., 2021). More structural information of the samples was further examined by the Raman test (Figure 2B). Two sharp Raman peaks are noted at 406 and 376 cm⁻¹, corresponding to the out-of-plane A_1g and in-plane E¹_2g bands of the 2H-MoS₂ phase, respectively (Xie et al., 2017). The Raman-sensitive carbon structure of rGO is clearly confirmed by the disordered carbon-related D band at 1350 cm⁻¹ and the graphitic carbon-induced G band at 1583 cm⁻¹. To determine the weight ratio of MoS₂ in the composite, TGA was measured (Supplementary Figure S1), from which the MoS₂ component is calculated to be ∼82 wt%.

FIGURE 2. (A) XRD pattern. (B) Raman spectra of the MoS₂@rGO composite and the pure MoS₂.

XPS measurement of the MoS₂@rGO was carried out to probe the elemental composition and their bonding environment. The full-scan spectrum (Figure 3A) demonstrates the typical elemental signals of Mo, S, O, and C species. Figure 2C shows the high-resolution spectrum of Mo 3 days having two peaks locating at 232.3 (Mo3d_3/2) and 229.1 eV (Mo 3d_5/2), which agrees well with the Mo⁴⁺ feature of MoS₂. Moreover, the S 2p spectrum in Figure 2D is well resolved into two components of S 2p_1/2 (163.1 eV) and S 2p_3/2 (161.9 eV) stemming from S²⁻ species of MoS₂. The C1s signal (Figure 3C) is readily fitted by three parts containing C-C/C=C (284.6 eV), C-O-C (286.2 eV), and C=O (288.5 eV) bonds. This physiochemical analysis indicates the successful production of the rGO/MoS₂ nanocomposite.

FIGURE 3. (A) XPS survey spectrum of the composite. (B–D) High-resolution XPS spectra of (B) Mo 3d, (C) S 2p, and (D) C 1s.

To highlight the structural advantage of the MoS₂@rGO nanocomposite, a control sample of MoS₂@rGO-C was prepared by a similar solvothermal method but using a different Mo source of ammonium molybdate ((NH₄)₆Mo₇O₂₄). The microstructure of the MoS₂@rGO-C sample is observed by SEM and TEM imaging (Figures 4A,B). Notably, MoS₂ nanoflowers are unevenly grown on the surface of rGO nanosheets, indicating a preferable growth behavior. The difference clearly demonstrates the great importance of the Mo source in tailoring the hybrid morphology. It is well-known that the surface of GO is negatively-charged induced by the functional species, such as carboxylic, hydroxyl, and phenolic groups, thus leading to weak chemical coupling with the Mo-containing anions because of electrostatic repulsion. The literature report established that the negatively charged GO surface can be modified toward a positively-charged state by reducing the pH value of the solution (Choi et al., 2015). In this work, the employment of phosphomolybdic acid as the Mo source significantly reduces the pH of the GO solution to <1, which potentially modifies the GO surface state for better coupling with the phosphomolybdic anions and thus enables the uniform formation of ultrafine MoS₂ nanosheets. As for the ammonium molybdate or sodium molybdate used in many studies, the solution pH does not change and the MoS₂ is prone to nucleate at some preferential sites and then grow into a large flower-like structure. For more comparison, pure MoS₂ submicroflowers (Figures 4D,E) were prepared. It is worth to mention that the petal size of the control MoS₂ nanosheets is much larger than that of the MoS₂@rGO nanocomposite, and the interlayer spacing is relatively smaller (Figures 4C,F).

FIGURE 4. SEM, TEM, and HRTEM images of the (A–C) rGO/MoS₂-C and (D–F) pure MoS₂.

The Li-ion storage performance of the as-prepared samples was examined. The lithiation behavior of the MoS₂@rGO, MoS₂@rGO-C and pure MoS₂ electrodes is first studied by the CV test at a sweep rate of 0.1 mV s⁻¹. As shown in Figure 4A and Supplementary Figure S2, the CV curves of the three electrodes are similar, indicating the identical lithiation process induced by the electroactive MoS₂ components (Ren et al., 2017). In specific, the broad cathodic peak at around 1.4 V in the first cycle is associated with the insertion of Li⁺ into the interlayer space of MoS₂ to form Li_xMoS₂ (i.e., MoS₂+ xLi⁺ + xe⁻ → Li_xMoS₂), while the following peak at ∼0.4 V corresponds to the consecutive reduction of Li_xMoS₂ to generate Li₂S and metallic Mo (i.e., Li_xMoS₂ + (4-x)Li⁺ + (4-x)e⁻ → Mo + 2Li₂S) and the irreversible formation of the solid-electrolyte-interface (SEI) film on the electrode surface. Upon the reverse scans, the dominant anodic peaks at about 2.3 V are reasonably attributed to the oxidation of Li₂S to sulfur, and the 1.9 V peak is associated with the partial oxidation of metallic Mo to MoS₂. From the second and ongoing cycles, two cathodic peaks are observed at around 2.0 and 1.4 V, which comply well with the reversible reduction of sulfur (S + 2Li⁺ + 2e → 2Li₂S) and MoS2, respectively. It is worth to note that the CV curves are almost overlapped from the second cycle, suggesting the superb electrochemical durability and reversibility.

The cycling performance of the electrodes is examined by the galvanostatic charge–discharge test at a current density of 100 mA g⁻¹. It is evident from Figure 4B and Supplementary Figure S3 that the voltage plateaus on the profiles are in good accordance with the redox couples described in the above CV analysis. The cycling comparison of the three electrodes is exhibited in Figure 4C. Remarkably, the MoS₂@rGO electrode delivers high discharge/charge capacities of 1469/1087 mA h g⁻¹, which are much higher than that of the pure MoS₂ electrode (1100/687 mA h g⁻¹). As for the control MoS₂@rGO-C electrode, high capacities of 1428/894 mA h g⁻¹ is also obtained in the first cycle, indicating the effectiveness of the conductive rGO on improving the electrochemical harvest of the electroactive MoS₂. Moreover, in striking contrast to the MoS₂@rGO-C (62.6%) and pure MoS₂ electrode (62.5%), the MoS₂@rGO electrode displays a Coulombic efficiency as high as 74% in the first cycle. The initial capacity loss is primarily stemmed from the inevitable SEI film generation and some undesirable side reactions caused by the structural defects/oxygenated carbons. The Coulombic efficiency substantially rises to 95.6% for the second cycle and preserves over 99% for the subsequent cycles, suggesting excellent reversibility of the lithiation/delithiation processes. More remarkably, there is no noticeable degradation on the specific capacity for the MoS₂@rGO electrode, which retains at 1128 mA h g⁻¹ after 100 cycles, indicating long cycling lifetime. By contrast, a significant capacity decay is observed on both the MoS₂@rGO-C (59%) and the pure MoS₂ (42%) electrodes upon 100-cycling. It is noted that the capacity shows a slight increase from 70 cycles, which can be ascribed to the gradual electrode activation and the reversible formation of a gel-like film (Chen et al., 2014).

Rate capability is an important criterion for the design of high-power LIBs (Zhu et al., 2015). Figure 4E compares the rate capability of the MoS₂@rGO, MoS₂@rGO-C, and pure MoS₂ electrodes at various current densities. The hybrid rGO/MoS₂ is capable of delivering a Li-ion storage capacity of 1122 mA h g⁻¹ at a low current rate of 50 mA g⁻¹. When the current rate elevates to 100, 250, 500, 1000, and 2000 mA g⁻¹, the reversible capacity could keep at 1037, 936, 838, 702, and 559 mA h g⁻¹, respectively. For comparison, pure MoS₂ is almost deprived of Li-ion storage capability at the high current rate of 2000 mA g⁻¹ (43 mA h g⁻¹), while the control rGO/MoS₂-C electrode merely maintains at 268 mA h g⁻¹ at this current rate. This comparison clearly demonstrates the superior rate performance of the rGO/MoS₂ electrode. Moreover, the hybrid electrode could recover the original capacity level when the current density is reduced back to 50 and 100 mA g⁻¹, again suggesting that the electrode is highly reversible and durable. The prominent cycling stability of MoS₂@rGO is further validated by another 2 cells tested at high current densities of 500 and 2000 mA g⁻¹. As shown in Figure 5E, high and stable specific capacities of 759 and 542 mA h g⁻¹ are sustained after 200 cycles at 500 mA g⁻¹ and 400 cycles at 2000 mA g⁻¹. Compared to the MoS₂@rGO-C electrode, the rGO/MoS₂ electrode shows substantial improvement in the cycling stability and high-rate performance. This suggests the significance of a uniform pattern and nanoscaled size of MoS₂ on the rGO nanosheets. It is believed that the unique heterostructure holds great advantage on reducing the electron/ion transport pathways and maintaining the hybrid structural integrity. EIS analysis is carried out to gain an in-depth understanding of the electrode kinetics. Figure 5F displays the resulting Nyquist plots, in which the semicircle diameter represents the charge transfer resistance (R_ct). As expected, compared with the rGO/MoS₂-C (185 Ω) and pure MoS₂ (266 Ω) electrodes, a smaller R_ct is noted for the rGO/MoS₂ electrode (151 Ω). The reduced R_ct value implies an improved reaction kinetics, which is desirable for high power delivery.

FIGURE 5. (A) CV curves and (B) galvanostatic discharge/charge profiles of the rGO/MoS₂ electrode. (C) Cycling performance, (D) rate capability, and (F) Nyquist plots of the rGO/MoS₂, rGO/MoS₂-C, and pure MoS₂ electrodes. (E) Cycling stability of the rGO/MoS₂ electrode at high current densities of 500 and 2000 mA g⁻¹.

Sodium-ion batteries (SIBs) have gained intense research activities as a promising alternative to the LIBs due to their low cost and earth-abundance of sodium in nature (Wang et al., 2018; Wu et al., 2020). Since the two battery systems share similar energy storage characteristics, it is, hence, anticipated that the unique heterostructure of MoS₂@rGO is also of grand benefit for Na-ion storage (Wu et al., 2021). To validate our hypothesis, the Na-ion storage properties of the MoS₂@rGO nanocomposites are examined. The resulting CV curves are displayed in Figure 6A. It is observed that the rGO/MoS₂ electrode exhibits very similar electrochemical behaviors in the SIBs and LIBs. The broad cathodic peak at around 1.0 V in the first cycle is associated with the insertion of Na ions into the interlayer of MoS₂ nanosheets and the formation of SEI layers. The sharp peak below 0.5 V in the following deep cathodic process corresponds to the conversion reaction of MoS₂ to form metallic Mo and Na₂S. The cathodic peaks in the following cycles become inconspicuous, presumably due to the gradual amorphization of crystalline MoS₂ after the first cycle (Wang et al., 2016). During the anodic process, the two peaks can be assigned to the consecutive oxidation of Mo and Na₂S, respectively. In addition, the anodic peaks are almost overlapped, suggesting high reversibility and cycling stability of Na-ion storage in the hybrid electrode.

FIGURE 6. (A) CV, (B) charge/discharge curves, (C) cycling performance, and (D) rate capability of the MoS₂@rGO electrode for SIBs. (E) Energy storage characteristics of MoS₂@rGO for LIBs/SIBs.

Figure 6B presents the galvanostatic charge/discharge curves of the electrode at a current density of 50 mA g⁻¹. The voltage profiles agree well with the sodiation/desodiation reactions of the CV peaks. The hybrid electrode delivers initial discharge and charge capacities of 1056 and 660 mA h g⁻¹, respectively. The huge irreversibility of the first cycle can be attributed to the formation of SEI layers and the side reactions. However, the Coulombic efficiency increases rapidly from the initial 62.5 to 95.4% (2nd cycle) and >99% in the extended cycles. Figure 6C exhibits the cycling performance. It is noted that the MoS₂@rGO electrode sustains a high capacity retention of 88% after 150 cycles, illustrating good cycling stability. Moreover, the hybrid electrode also manifests an attractive rate capability. As shown in Figures 6A,D a high specific capacity of 283 mA h g⁻¹ is retained at a large current density of 1600 mA g⁻¹, revealing a good high-rate capability of the MoS₂@rGO nanocomposite for Na-ion storage. As the current density is returned back to 50 mA g⁻¹, the electrode almost recovers its initial capacity level for extended 40 cycles, confirming the long cycle life. The Na-ion storage performance of the MoS₂@rGO composite is compared with the previously reported MoS₂-based materials. As summarized in Supplementary Table S1, the present performance is superior or at least comparable to those documented in most studies, demonstrating the advantage of the designed MoS₂@rGO in SIBs.

The outstanding Li/Na-ion storage performance of the MoS₂@rGO electrode can be rationally ascribed to the unique hybrid morphology. Figure 6E schematically illustrates the energy storage characteristics of the nanocomposite for LIBs and SIBs. First, the 2D/2D nano-architecture provides sufficient surface area for enlarging the electrode/electrolyte interfaces and thus enables more active sites for Li/Na-ion storage. Second, the nanoscaled size of the MoS₂ nanosheets could not only shorten the ion/electron transport distance but also minimize the volume strain caused by the lithiation/delithiation (or sodiation/desodiation) processes. Third, the preferential edge termination of MoS₂ facilitates fast ion transport through the open 2D nanochannels. Last but not the least, the flexible rGO network plays a role of high-speed conductive pathways for charge collection and transfer and also serves as an elastic substrate to alleviate the repetitive volume change of MoS₂ during cycling operation. As a result, a strong synergistic effect is achieved between the two components to rendering a high specific capacity, a superior rate capability, and long-term cycling endurance.

Conclusion

In summary, a hierarchical MoS₂@rGO heterostructure has been successfully synthesized using a facile solvothermal process. The MoS₂ nanosheets are uniformly and vertically patterning on the surface of graphene with preferential (002) edge orientation. The unique sheet-on-sheet nanoarchitecture could provide a sufficient number of electrochemically active sites, a reduced ion/electron transport distance, and a high-speed electrically conductive network, thereby generating a strong synergistic effect to enable enhanced electrode kinetics and high/stable electrochemical utilization of individual components. As a consequence, the MoS₂@rGO nanocomposite manifests a high specific capacity, superior rate, and cycling performance for use in both LIBs and SIBs. The present work demonstrates the significance of rational engineering 2D architecture of metal sulfide/graphene to serve promising anode material candidates for high-efficient Li/Na-ion storage.

Data Availability Statement

The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding author.

Author Contributions

CW carried out experiments and wrote the manuscript. ZH and HS performed and analyzed experimental results. J-GW designed experiments and contributed to writing the manuscript. All authors contributed to the article and approved the submitted version.

Funding

The work is financially supported by the National Natural Science Foundation of China (51772249 and 51821091) and the Fundamental Research Funds for the Central Universities (D5000210894 and 3102019JC005).

Conflict of Interest

Author CW is employed by Shenzhen Cubic-Science Co., Ltd.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors, and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Acknowledgments

The authors also appreciate the TEM analysis from the Analytical & Testing Center of NPU.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem.2021.802788/full#supplementary-material

References