Microscopic high-resolution optical imaging of metal oxide battery particles
For many metal oxides, including LCO and NMC811, their reflectivity (and thus refractive index)22,23,32 at visible wavelengths changes during insertion of a charge. To investigate the relationship between charge state and reflectivity, ex-situ microscopic reflection spectra of individual LCO and NMC811 particles are shown in Figure 1a as a function of lithiation (x = 0 corresponds to the fully delithiated particles). In both LCO and NMC811, the (surface) reflectivity increases with decreasing x in the near-infrared between 680 and 900 nm for LCO and 800 and 900 nm for NMC811, with changes of up to 20% between x ≈ 0 (0.5 for LCO) and x ≈ 1. Hence, in this wavelength range the magnitude of the particle reflectivity can be used as a proxy of the state-of-charge. However, between 400-680 nm for LCO and 400-800 nm for NMC811, the changes in reflectivity with x are non-monotonic, with spectral shifts consistent with bulk reflectivity measurements of other materials28,33,34. This region should consequently be avoided for microscopic measurements where reflectivity is used to track the lithiation state of particles in an electrode. Furthermore, in some common electrode materials, e.g. LixFePO4, the reflectivity change (400 to 900 nm) at different lithiation stages is <1%, suggesting that microscopic reflectivity may not be a universal tool for tracking particle charge state (supplementary information 1) as recently proposed24. Fitting the reflection spectra in Figure 1a with the Kramers-Kronig relations35 (see supplementary information 2) allows extraction of the state-of-charge dependent refractive index (RI). The real part of the refractive indices (n) for LCO and NMC811 is found to be between 2 – 2.5. This is larger than the RI (1.5 to 1.6) high numerical aperture (N.A.) microscope objective lenses are designed for, hence aberration corrections should be applied to account for the index mismatch (see discussion below).
To understand how ex-situ observations relate to changes in reflectivity during cycling, we perform operando reflection (confocal) microscopy measurements at several wavelengths on a single LCO particle (see methods and below), plotting the total reflected intensity and potential during charging in Figure 1b. At 800 nm (red curve), when light is focussed on/collected from the surface (S) of the particle, the reflectivity decreases linearly as x is increased. Whereas between 380 nm and 600 nm a non-monotonic ‘W’ shaped curve is observed (green, blue, and purple curves) in line with Figure 1a. Interestingly, when focussing/collecting light from inside (I) the particle – an advantage offered by confocal sectioning – the reflectivity at 800 nm decreases linearly with x (Figure 1b, blue dashed line). The difference between the ‘S’ and ‘I’ response at 800 nm can be reconciled by noting that the reflectivity at the interface of two media of RI n1 and n2 will be proportional to the difference, (n1-n2)2 under normal incidence36. At the surface of particles, n1 is equal to that of the electrolyte (~1.4 to 1.6)37,38 and n2 that of the particle 1.8-2.539. However, inside the particle n1 and n2 will both vary between 1.8 and 2.5. Consequently, depending on the focus, different interfacial refractive indices, i.e. electrolyte/particle versus particle/particle (intraparticle), will be resolved. We note that absorption of the particle top surface impacting layers below only results in attenuation of the ‘I’ response and cannot explain the opposing direction of trends between ‘S’ and ‘I’. Indeed, the variation in the imaginary (k; absorptive) part of the refractive index is relatively small with lithiation state at 800 nm. Hence, the different extent to which k (and its changes) contribute to the ‘S’ and ‘I’ signals during (de)lithiation are expected to be less significant than that of n (see supplementary information 2). Our observations potentially explain the contrasting reports of Merryweather et al.24 and Jiang et al. 27 with the former reporting brightening of LCO on delithiation and the latter dimming.
The results above suggest that the choice of the imaging plane plays a critical role when imaging such electrolyte/particle systems. Different combinations of the red and dotted blue curves in Figure 1b might be observed for different particles irrespective of any battery dynamics simply due to defocus or electrode roughness, i.e. inter-particle heterogeneities cannot be established from measurements at a single focal plane (WF). Additionally, interference between different reflection planes on the particle surface will influence the spatial distribution of intensity imaged. In supplementary information 3 and supplementary video 1 we demonstrate (in pseudo-wide field measurements) that even in the absence of any cycling, different intensity patterns – closely resembling patterns that might be ascribed to the motion of phase-fronts or arise from morphology inhomogeneities19,20,40,41 – can be observed across the surface of particles simply by small (20 nm) adjustments in the focus42. The large RI of the battery particles means only a small defocus, which may not be captured, or may even be created by, an auto-focus feedback system, can give rise to such effects. Hence, measuring at a range of focal planes, i.e. in 3D, is key to distinguish spatial intensity patterns modulated by small (high N.A amplified) focal shifts throughout a cycle, and true electrochemical dynamics.
The refractive index, n, also plays a role on the imaged size of an object via the optical path length, (OPL) OPL = n · s, where s is the geometric path length. Hence, before using optical microscopy to also investigate particle size changes during cycling26, the impact of a varying n on the OPL must first be accounted for27. In Figure 1c, we compare the lateral width of a LCO particle extracted from a reflection image at a single focus position (F1) and the lateral width of the same particle calculated from the maximum intensity projection43–45 (MIP) where the influence of OPL on the object size is accounted for (see methods). By comparing the MIP and F1 images we can estimate an error in sizing introduced. For x = 1 (in LixCoO2), LF1 and LMIP (L is the particle length; dashed red and blue lines in Figure 1c) deviate by 0.12%. This deviation increases monotonically with decreasing x, reaching 0.43% at x = 0.48. From error propagations (see supplementary information 4) this will correspond to a deviation in volume of the imaged object (as compared to its ‘true’ size) of ~0.36% to 1.3% as x decreases from 1 to 0.48 (right panel graph in Figure 1c and supplementary information 5). The deviation will be more pronounced for smaller curved particles, with larger n, but generally our results suggest that changes in volume below ~1 to 1.5% are challenging to unambiguously detect with optical microscopy.
Having understood the limitations and conditions under which optical reflection imaging can be performed, we turn to applying these methods. LSCRM, as schematically depicted in Figure 2a, is performed on polycrystalline LCO (average particle diameter, 2.5 to 7.5 μm) and NMC811 ( 5 μm). A laser beam at 820 nm, where absorption and scattering are minimised and changes in reflectivity with lithiation state are monotonic for LCO and NMC811, is scanned across the sample. Adjusting the focus of the beam in the sample, whilst using a pinhole (0.7 AU) to gate out-of-focus light, allows probing of different 2D planes up to the optical penetration length, which is <15 μm in both materials (see supplementary information 6). Stacking the 2D images and correcting for laser attenuation and defocus aberration (supplementary information 6) allows for sub-micrometre 3D information to be obtained on individual particles, as illustrated in Figure 2b,c. A customised battery half-cell with optical access (right side of Figure 2a) allows for galvanostatic cycling during the recording of confocal image stacks of electrodes (~120 s per stack). All data presented herein, is taken from the first 2 to 9 cycles of an electrode.
Operando tracking of single particle volume changes
To first benchmark our LSCRM methodology, we monitor volume changes of individual particles during a cycle. By calculating the number of voxels occupied by particle containing regions within the reconstructed volumes, volumes at the single particle level can be obtained. In Figure 3a-c, the normalised change in single particle volumes (ΔV/V) of LCO and NMC811 are shown as a function of time/charge state and C-rate. The volume of LCO particles in all cases increases during charging and decreases during discharging (standard deviation in maximum volume change (σΔV/V) is 0.42), with a maximum volume change of ~3% ±1% (see supplementary information 6 for error bar estimation). For NMC811, σΔV/V is 0.63 with the volume decreasing during the charge (|ΔV/Vmax| ~12% ±2%), as shown by the graph in Figure 3c. In NMC811 we observe significant changes in volume only between lithium fractions of x = 0.3 and x = 0.7 and as for LCO there are also inactive particles which show a ΔV/V ~0 throughout the cycle. In NMC811 there is a greater scattering of results with a hysteresis in ΔV/V, suggesting irreversible changes, e.g. primary particle cracking or movement within the secondary particle microstructure. However, for polycrystalline secondary particles (NMC811) and polycrystalline particles (LCO) as studied here, contributions from insertion of lithium into the lattice, cracking and rearrangement cannot be disentangled48. Indeed, for a small fraction (~15%) of NMC811 particles measured, positive changes in ΔV/V are observed (see supplementary information 6). Nevertheless, the results in Figure 3b,c are consistent both in terms of magnitude, direction and state-of-charge onset, for volume changes obtained by previous ensemble X-ray diffraction and pressure dependent open circuit voltage measurements of LCO (ΔV/Vx=1→x=0.5 ~ +2%) and NMC811 (ΔV/Vx=1→x=0.2 ~ -8%) 48–51.
Surface-to-core transport
To further understand how structural inhomogeneities might be related to (de)lithiation asymmetries, we examine the transport velocities of Li-ion containing phase fronts from the exterior to the centre of particles. The surface-to-core transport velocity is indeed a parameter of fundamental interest as local inter- and intra-particle heterogeneities in ion transport can limit the overall (dis)charging rate and potentially cause irreversible material changes52–54. To track the motion of phase fronts through polycrystalline NMC811 or LCO particles during the cycle, the normalised time varying reflectivity, a proxy for the state-of-charge, is extracted for each z-plane across the central area of a particle. From this, the depth-dependence of the time-varying reflectivity can be extracted, as shown in Figure 4a, b as a function of cycling rate. For each z-plane, the point at which the reflectivity crosses zero is then determined (dashed line in Figure 4a, b); from the variation of this reference point with depth, a transport velocity for ion containing fronts can be estimated (see supplementary information 7).
For NMC811, in Figure 4a, the delithiation and lithiation both occur from the surface to the core, in a quasi symmetric manner across the particle z-span, with qualitatively similar behaviours from C/2 to 2C. This observation is in-line with a shrinking core type mechanism of ion (de)intercalation1,55,56, where, because of the higher lithium flux on the particle surface as compared to the bulk, ion (de)insertion is diffusion limited, rather than surface limited57. In LCO, delithiation occurs from the surface to the core, as for NMC811. However, lithiation occurs quasi uniformly across the particle volume, with all depths changing reflectivity simultaneously (flat dashed line in Figure 4b). These observations match phase-field modelling58 and several experimental studies1,59,60 which have previously suggested that, in contrast to delithiation, lithiation of LCO is charge transfer limited and occurs via a lithium-poor phase with higher ionic diffusion, resulting in a intercalation wave type mechanism. This differing mechanism of ion transport may explain the uniform surface-to-core lithiation profile and significantly higher diffusivity on lithitation than delithiation for our LCO.
Because the velocity of phase fronts (vp) will depend on the state-of-charge, values extracted in Figure 4c-e represent an average across lithiation/delithiation, albeit at the single particle level. In NMC811 and LCO, vp increases with C-rate, but remains of the same order of magnitude of 2-6 nm s-1 across C-rates for delithiation and lithiation. Note that for LCO, velocities cannot be extracted during lithiation as it is beyond our time resolution (100 s). These values sit at the lower end of those reported previously in the literature (1 nm s-1 to 50 nm s-1)24,61–66. However, in this work, the transport velocities reported are through individual polycrystalline particles i.e. from the surface to the core, which cannot be obtained with other methods such as galvanostatic intermittent titration technique (GITT) or 2D imaging which operate at the ensemble level and/or do not have 3D directional resolution on transport. We note that anisotropy in the transport is not expected due to the random orientation of primary particles and their polycrystalline nature (see supplementary information 7). Finally, the linear scaling of vp with C-rate in both NMC811 and LCO suggests that phase transport is kinetically rather than thermodynamically limited for both layered oxides10,31.
Imaging phase-front dynamics on different surfaces
Having shown that the velocity of phase fronts from the surface to the core of particles can be tracked using LSCRM, we push further the technique to spatially separate (de)lithiation inhomogeneities between the surface and bulk of single particles. For that, we compare using mathematical reconstructions, at different points during a charge-discharge cycle, the spatial distribution of reflected intensity (lithiation state) between the particle exterior, and core for LCO particles cycled at different C-rates (C/2 and 2C). For both the exterior and core the reflection contrast is derived from refractive index changes within the material (see supplementary information 8). LCO particles are first computationally ‘unwrapped’ into shells (~25) of thickness ~500 nm for each time point in the cycle. The minimum volume ellipsoid – a close approximation of the particle shape – enclosing points of the shells is then calculated67,68. The surface of the ellipsoid is projected onto a 2D plane using the Mercator projection69, as show in Figure 5a (see supplementary information 8), and colours correspond to different phases or domains with different lithiation states. In Figure 5b,c, 2D projections of the outer (exterior) and inner (core) most shells are shown for pristine, charged and discharged states during galvanostatic cycles at C/2 and 2C. At C/2, for the particle exterior, the contrast in reflectivity both within and between domains remains small throughout the cycle indicating large area, uniform (de)lithiation. Furthermore, for the pristine and discharged states the distribution of different lithiation domains is near-identical, demonstrating a reversible cycling process. At the particle centre, a spatial rearrangement of lithium phases is observed from the pristine to charged states. This rearrangement remains on discharging, but the differences in lithiation degree between domains, for a given overall state-of-charge, are small. The observation of changes at both the particle exterior and core however indicate that (de)lithiation occurs throughout the entire particle.
For 2C, at the particle exterior, the pristine and discharged states do not show a similar spatial distribution of lithiation phases. For the particle core, no rearrangement of lithium domains at the end of charge is observed, unlike at C/2. Only minimal changes in both the contrast and spatial distribution of lithium domains occur between the pristine, charged and discharged states. This observation suggests that (de)lithiation does not occur throughout the entire particle volume, in agreement with the limited cycling capacity measured at 2C when compared to C/2 (see Figure 3b for instance). For the exterior, regions of contrastingly high (dark blue) and low (yellow) Li-content appear on going from the pristine to charged state, and persist to the discharged state. This indicates non-uniform lithiation on the surface of LCO particles at high C-rate, consistent with several previous observation in layered oxide materials11,70–73.
Our results do suggest that LSCRM, combined with mathematical/computational treatments, can spatially distinguish phase inhomogeneities simultaneously at different surfaces within a single cycling particle. However, for such a non-trivial technique heavily relying on data treatment, more remains to understand the implication of these observation, with for instance NMC811 showing a much more complex behaviour (see supplementary information 9).
Visualisation of electrolyte dynamics
Thus far, optical and other operando imaging methods such X-Ray microscopy have been limited to visualising particles of active material. However, we can show that LSCRM offers the opportunity to image both lithium transport in solids, as demonstrated above, simultaneously with concentration gradients forming in the liquid electrolyte upon polarisation. Indeed, LiPF6, the Li-conducting salt in the electrolyte used, has previously been observed to be a source of fluorescence at visible wavelengths, with the exact origin albeit debated (see discussion in Laurence et al.74). Following these past observations, we find that under two-photon excitation (2PEF), the fluorescence of LiPF6 salt (solid) is significantly more efficient than with one photon (1PEF), particularly above an excitation wavelength of 900 nm (Figure 6a). In Figure 6b we perform ex-situ 2PEF experiments on LiPF6 solutions (in 1:1 vol.% ethylene carbonate/dimethyl carbonate) at different concentrations (ranging from 0.01 to 4.0 mol/L), and find a strong monotonous brightening with concentration. Figure 6c shows the evolution of the electrolyte 2PEF intensity, at different focal planes above a self-standing LCO electrode, during a charge-discharge cycle at 2C. In all cases the electrolyte 2PEF increases near linearly on charging to 4.2 V, i.e. on the release of Li-ions to the electrolyte, with a slight plateauing in the rate of brightening between 4.1 V – 4.2 V. Discharging, i.e. depletion of lithium from the electrolyte, results in dimming of the electrolyte 2PEF. The response is repeatable and of a similar magnitude over 8 cycles (see supporting information 10). Together the ex-situ and operando measurements show the electrolyte 2PEF to be highly sensitive to the Li-ion concentration and indicate that the 2PEF must indeed derive from the LiPF6 salt or a compound bound to it74. Furthermore, the measurements indicated that the 2PEF is unrelated to cycling induced electrolyte degradation75–79. Consequently, it is suggested that the intensity of the 2PEF can be used to qualitatively track the Li-ion concentration in the electrolyte.
Our attention is then turned to examining the spatial distribution of salt concentration upon cycling. To do so, LSCRM experiments are repeated with an excitation of 1020 nm with two separate detectors, one for collecting particle reflectance and another for 2PEF (see Figure 2a; ~250 s per confocal stack). This allows for simultaneous volumetric imaging of the electrolyte and LCO particles, as illustrated in Figure 6d. In Figure 6e, the 2PEF signal of the electrolyte in the plane ~300 nm above the top surface of the electrodes is plotted at selected time points (labelled A to F; correspond to points of equivalent state-of-charge as marked in Figure 6c) and spatial locations during a C/2 cycle (dashed black lines indicate regions of similar 2PEF intensity). Upon charge, the 2PEF intensity increases due to delithiation and an increase in the LiPF6 concentration. A gradient in the 2PEF intensity also becomes present above potentials of 4.0 V around LCO particles. This 2PEF/electrolyte concentration gradient, which decays quasi homogeneously away from particles, originates from the difference between the rate of delithiation at the LCO surface and the rate at which PF6- anions diffuse towards the LCO particles80–83 to balance Li+ ions released and maintain electroneutraility76,84,85. Furthermore, at all potentials the 2PEF is relatively uniform in regions not containing LCO (bottom row) and upon discharge, the concentration gradient around particles disappears. Altogether, our results indicate a reversible process, as would indeed be expected for the formation of a concentration gradient around active material upon cycling. The concentration gradient we observe extends as far as 1.5 μm. This in agreement with previous theoretical and experimental studies which have shown electrolyte concentration gradients around electrode interfaces extending between 500 nm86 and 10s μm76,81,87 depending on the electrode, electrolyte composition (mass to volume ratio) and cycling conditions88,89. We note that on charging the electrode and allowing it to relax to OCP, the polarisation gradient rapidly disappears (see supporting information 10). Switching to a greater C-rate, i.e. 2C, during the initial charge up to 3.9 V (A panels), the 2PEF of the electrolyte is brightest around LCO particles with the intensity decaying quasi homogeneously away from the particles. Increasing the potential to 4.2 V (B, C, D and E) results in a drastic increase in 2PEF intensity but the distribution of electrolyte 2PEF also becomes heterogeneous around both particles and in regions of the electrode ~20 – 50 μm away from LCO particles (bottom row of Figure 6f). The latter observation suggests inhomogeneous electrolyte diffusion within the self-standing electrode, as a result of the geometry and distribution of pores within the particle-carbon/binder matrix90. This influence of the matrix can be expected at higher C-rates where the rate of (de)lithiation at the LCO surface and ionic diffusion rate in the electrolyte are more significantly mismatched91. Finally, on discharging to 3.9 V (panel F) the 2PEF distribution returns to its initial state, confirming that our observations do not stem from electrolyte degradation but are from (de)lithiation and local changes in salt concentration and concentration gradients within the electrolyte. In NMC811, similar behaviour is observed as for LCO but with subtle differences requiring further in-depth analysis (see supporting information 10).