The impact of ultrahigh (dis)charge rates on the phase transition mechanism in LiFePO4 Li-ion electrodes is revealed by in situ synchrotron diffraction. At high rates the solubility limits in both phases increase dramatically, causing a fraction of the electrode to bypass the first-order phase transition. The small transforming fraction demonstrates that nucleation rates are consequently not limiting the transformation rate. In combination with the small fraction of the electrode that transforms at high rates, this indicates that higher performances may be achieved by further optimizing the ionic/electronic transport in LiFePO4 electrodes.

Phase transitions in Li-ion electrode materials during (dis)charge are decisive for battery performance, limiting high-rate capabilities and playing a crucial role in the cycle life of Li-ion batteries. However, the difficulty to probe the phase nucleation and growth in individual grains is hindering fundamental understanding and progress. Here we use synchrotron microbeam diffraction to disclose the cycling rate-dependent phase transition mechanism within individual particles of LiFePO4, a key Li-ion electrode material. At low (dis)charge rates well-defined nanometer thin plate-shaped domains co-exist and transform much slower and concurrent as compared with the commonly assumed mosaic transformation mechanism. As the (dis)charge rate increases phase boundaries become diffuse speeding up the transformation rates of individual grains. Direct observation of the transformation of individual grains reveals that local current densities significantly differ from what has previously been assumed, giving new insights in the working of Li-ion battery electrodes and their potential improvements.

No abstract

Intense interest in the Li-O2 battery system over the past 5 years has led to a much better understanding of the various chemical processes involved in the functioning of this battery system. However, detailed decomposition of the nanostructured Li2O2 product, held at least partially responsible for the limited reversibility and poor rate performance, is hard to measure operando under realistic electrochemical conditions. Here, we report operando nanobeam X-ray diffraction experiments that enable monitoring of the decomposition of individual Li2O2 grains in a working Li-O2 battery. Platelet-shaped crystallites with aspect ratios between 2.2 and 5.5 decompose preferentially via the more reactive (001) facets. The slow and concurrent decomposition of individual Li2O2 crystallites indicates that the Li2O2 decomposition rate limits the charge time of these Li-O2 batteries, highlighting the importance of using redox mediators in solution to charge Li-O2 batteries.

Electrode phase transitions are of large practical importance in Li-ion batteries determining the voltage profile and as a decisive factor in the transition kinetics. In particular in LiFePO4 the impact of defects, particle size, and charge rate on the phase transition in LiFePO4, has been studied intensively. The end-member phases, lithium-rich triphilyte Li1- bFePO4 (LFP) and lithium-poor heterosite LiaFePO4 (FP), demonstrate narrow solid-solution domains however both phase field modelling and DFT predict that the first-order phase transition can be suppressed by applying high rates leading to a solid-solution transition1-3 giving a rationale for the intrinsically fast kinetics of the LiFePO4 material Recent in-situ high-rate diffraction studies by Orikasa et al.4 , 5 revealed a metastable Lix ∼ 0.6FePO4 phase in addition to the thermodynamically stable LiFePO4 and FePO4 phases rather than the predicted solid-solution transformation1-3. To determine the phase transition mechanism dependence on the dis(charge) rate we performed an in-situ synchrotron diffraction study with (dis)charge rates ranging from a very low rate of C/5 up to ultra-high rates of 60C. At C/5 charging diffraction shows the established first-order phase transition that occurs upon delithiating LiFePO4. With increasing charging rate the LiFePO4 and FePO4reflections increasingly shift indicating lattice parameters associated with increasing solubility in both end members. In addition with increasing charge rates considerable diffracted intensity is observed between the Bragg reflections indicating that a fraction of the material undergoes a solid solution transformation bypassing the first-order phase transition. At the early stages of charging the metastable phase reported by Orisaka et al.4 , 5 is observed which can be explained based on the phase diagram of LiFePO4. The distribution in environments observed at high rates, including the charged FePO4 the distribution of LixFePO4 phases and the non-reacted LiFePO4 at high rates indicates that the transformation moves through the electrode as a transformation wave, revealing that improvement of electrode performance should focus on optimization the ionic/electronic transport in LiFePO4electrodes rather than on lowering nucleation barriers. These results challenge theorists to capture the observed non-equilibrium thermodynamics of this material. 1 Malik, R., Zhou, F. & Ceder, G. Kinetics of non-equilibrium lithium incorporation in LiFePO(4). Nature Materials 10, 587-590, doi:10.1038/nmat3065 (2012). 2 Bai, P., Cogswell, D. A. & Bazant, M. Z. Suppression of Phase Separation in LiFePO4 Nanoparticles During Battery Discharge. Nano Letters 11, 4890-4896, doi:10.1021/nl202764f (2011). 3 Cogswell, D. A. & Bazant, M. Z. Coherency Strain and the Kinetics of Phase Separation in LiFePO4 Nanoparticles. Acs Nano 6, 2215-2225, doi:10.1021/nn204177u (2012). 4 Orikasa, Y. et al. Direct Observation of a Metastable Crystal Phase of LixFePO4 under Electrochemical Phase Transition. Journal of the American Chemical Society 135, 5497-5500, doi:10.1021/ja312527x (2013). 5 Orikasa, Y. et al. Transient Phase Change in Two Phase Reaction between LiFePO4 and FePO4 under Battery Operation. Chemistry of Materials 25, 1032-1039, doi:10.1021/cm303411t (2013).

Here we present our recent findings using two novel in-situ techniques1,2. (1) The distribution of Li-ions in LiFePO4 electrodes under operando conditions reveals what charge transport mechanism is rate limiting under what conditions including the impact of particle size and carbon additives. (2) In operando micro beam diffraction is performed from slow to high cycling rates disclosing phase transformation mechanism in individual LiFePO4 for the first time. Neutron Depth Profiling (NDP) offers the possibility to see directly see lithium ions, or atoms, via the capture reaction of neutrons with the 6Li isotope to form an α-particle (4He) and a trition (3H) carrying the  energy that is generated by the reaction according to conservation of energy and momentum. The 4He and 3H particle travel through the surrounding material during which they lose energy. By measuring the particle energy when it reaches the detector, the depth at which the 6Li atom was located can be determined. This allows to reconstruct a Li-atom density profile as a function of depth without significantly influencing the Li concentration due to the low neutron flux making this a non-destructive technique. Advancing the application of NDP from micro batteries3to conventional batteries we show the impact of particle size and charge rate on how the activity in the electrodes is distributed. This gives insight in what transport process dominates the internal resistance under what conditions. Zooming in on the phase transition reaction within single LiFePO4 grains has been achieved by applying a bright and sub-micron sized synchrotron beam. Under these conditions diffraction rings observed with powder diffraction fall apart in individual spots each representing individual LiFePO4 grains in the positive electrode. In this way the phase transition behavior through time of hundreds of individual 140 nm LiFePO4 grains is monitored at the same time while varying the electrochemical conditions ranging from slow (C/5) up to fast rates (2C). Following the phase transformation kinetics of individual grains reveals much slower transformation rates than expected based on the generally accepted mosaic or “Domino Cascade” transformation. In addition we show that the transformation rate of individual grains actually depends on the cycling rate. The appearance of many streaked reflections observed at low rates discloses that the majority of the grains is actively transforming via nanometer thin plate shaped domains that nucleate in specific crystallographic orientations. Also this observations defies the  long believed mosaic or “domino cascade” transformation model. As the (dis)charge rate increases the number of these plate shaped domains decreases and their width increases, driving the local compositions of the coexisting phases towards each other. The observation of a diffuse interface in a single grain at high (dis)charge rates reveals the growth mechanism at high rates, consistent with recent operando powder diffraction probing the average crystalline state over all grains4,5. Thereby, a consistent mechanistic picture is revealed for the phase transformation in individual LiFePO4 grains under realistic operando conditions for the first time. Counter intuition, the existence of well-defined interfaces at low discharge rates, may cause a shorter cycle life as compared to the more diffuse interfaces at higher rates. The electrochemically driven higher transformation rates indicate phase transformation kinetics play a more important role at low rates as compared to higher rates where the diffuse interfaces appear to be responsible for the faster transformation. Consistent with the electrolyte transport limited reactions, a smaller faction of the electrode material appears actively transforming, in accordance to the NDP results, indicating this has to be the focus for high rate electrode improvement. (1) Zhang, X.; Verhallen, T. W.; Labohm, F.; Wagemaker, M. Advanced Energy Materials 2015, 15, 1500498. (2) Zhang, X.; van Hulzen, M.; Singh, D. P.; Brownrigg, A.; Wright, J. P.; van Dijk, N. H.; Wagemaker, M. Nat Commun 2015, 6. (3) Oudenhoven, J. F. M.; Labohm, F.; Mulder, M.; Niessen, R. A. H.; Mulder, F. M.; Notten, P. H. L. Advanced Materials 2012, 23, 4103. (4) Zhang, X.; van Hulzen, M.; Singh, D. P.; Brownrigg, A.; Wright, J. P.; van Dijk, N. H.; Wagemaker, M. Nano Letters 2014, 14, 2279. (5) Liu, H.; Strobridge, F. C.; Borkiewicz, O. J.; Wiaderek, K. M.; Chapman, K. W.; Chupas, P. J.; Grey, C. P. Science 2014, 344. Figure 1

Ideally, the redox activity in Li-ion batteries, which is associated with storage and release of Li ions, is distributed homogeneously throughout the electrode, thereby minimizing detrimental processes while maximizing battery performance. However, it appears that every electrode material displays inhomogeneous redox activity which is expected to play a dominant role in battery performance parameters such as cycle life, efficiency and rate performance.1 For example, the occurrence of localized redox activity in layered electrodes can accelerate degradation reactions through local strain and/or decomposition1-2. Further, inhomogeneous reactions render estimating the true state of charge challenging and can introduce history effects3 that make optimization of battery performance via a battery management system problematic. Inhomogeneous redox reactions in Li-ion battery electrodes can occur on different length scales, starting at the nm length in individual electrode particles progressing up to the dimensions of complete composite electrodes (tens of μm) and can have various origins. Most directly, inhomogeneous reactions in electrodes stem from charge transport limitations, generally determined by the electrode morphology (including electrode thickness, porosity and tortuosity).4-6 A second origin for inhomogeneous reactions can be a difference in insertion potential, through mixing electrode materials5 or due to a distribution of nanoparticle sizes.7 Lastly, the nature of the phase transition can also give rise to inhomogeneous reactions. For example, the particle-by-particle transformation mechanism in LiFePO4 has been associated with a hysteresis effect8 and has been suggested to give rise to a memory effect that can influence the observed potential.3 Despite the above, limited research has been performed on this subject to date and many questions remain, mainly due to the difficulty to probe these heterogeneities in realistic battery geometries and during cycling conditions.Here we present a study of the phase transformation of ten's of individual NMC crystallites under operando cycling conditions in pouch cells using microbeam diffraction (ID11, ESRF, France). Due to the small, bright and sub-micron sized synchrotron beam diffraction rings observed with powder diffraction fall apart in individual spots each representing individual NMC crystallites in the positive electrode. In this way the phase transition behaviour through time of ten's of individual NMC crystallites is monitored at the same time while varying the electrochemical conditions.9,10 Figure 1 shows snapshots of the (108) and (110) reflection of individual NMC111 crystallites during a full C/4 charge-discharge cycle, demonstrated the anticipated continuous shift. For each individual crystallite the evolution of the lattice parameters provide insight in the distribution of the composition and local potential of each crystallite, as well as the individual transformation rate. In addition the average transformation rate can be translated in the act...