Layer spacing gradient (NaLi)1−xCoO2 for electrochemical Li extraction

Abstract

•(NaLi)1−xCoO2 was synthesized by non-Faradaic ion-exchange of Li1−xCoO2 with Na•Li phase and Na phase are connected by an intermediate phase•The chemical compositions and spatial relation of the different phases is critical Electrochemical extraction with insertion materials is a promising method to selectively extract lithium from Na-rich aqueous environments. However, the host structural evolution that accompanies the competitive co-intercalation and the host structure-selectivity relationship is yet to be understood. Here, we designed a core-shell layered cobalt oxide, (NaLi)1−xCoO2, via non-Faradaic ion-exchange of the parent Li1−xCoO2 with Na ions for Li extraction. Using electrochemical intercalation, we perform successful Li extraction from 1:20,000 Li to Na solution to ∼7.6:1 Li to Na, a Li selectivity of 1.5 × 105 over Na. We find the spatial relationship of the Li and Na in the core-shell structure plays an important role in achieving selective Li extraction as revealed by atomic resolution characterizations. Designing materials with high Li to Na selectivity is the key for direct extraction of Li from unconventional sources, such as seawater, without complex separations. However, material discovery is hindered by an insufficient understanding on how a material's composition and structure relate to Li selectivity during electrochemical co-intercalation. Here, we synthesized a core-shell structured (NaLi)1−xCoO2 with a core Li phase (Li0.94CoO2), a shell Na phase (Na0.51CoO2), and a transition intermediate phase via Na ion-exchange of parent Li1−xCoO2. We reveal that the chemical compositions and spatial relation of these phases play critical roles in governing high Li to Na selectivity. The core Li phase restricts large layer spacing expansion and inhibits Na intercalation. The shell Na phase serves to maintain the electrode stability against further Na ion exchange. With the layer spacing gradient design, we achieved a high Li selectivity of 1.5 × 105 with a recovery of ∼7.6:1 Li to Na from 1:20,000 initial ratio. Designing materials with high Li to Na selectivity is the key for direct extraction of Li from unconventional sources, such as seawater, without complex separations. However, material discovery is hindered by an insufficient understanding on how a material's composition and structure relate to Li selectivity during electrochemical co-intercalation. Here, we synthesized a core-shell structured (NaLi)1−xCoO2 with a core Li phase (Li0.94CoO2), a shell Na phase (Na0.51CoO2), and a transition intermediate phase via Na ion-exchange of parent Li1−xCoO2. We reveal that the chemical compositions and spatial relation of these phases play critical roles in governing high Li to Na selectivity. The core Li phase restricts large layer spacing expansion and inhibits Na intercalation. The shell Na phase serves to maintain the electrode stability against further Na ion exchange. With the layer spacing gradient design, we achieved a high Li selectivity of 1.5 × 105 with a recovery of ∼7.6:1 Li to Na from 1:20,000 initial ratio. The demand for lithium has grown tremendously over the past decades due to Li-ion battery development for electric vehicles, portable electronics, and stationary energy storage.1Chu S. Cui Y. Liu N. The path towards sustainable energy.Nat. Mater. 2017; 16: 16-22Crossref Scopus (1984) Google Scholar, 2Dunn B. Kamath H. Tarascon J.-M. Electrical energy storage for the grid: a battery of choices.Science. 2011; 334: 928-935Crossref PubMed Scopus (8981) Google Scholar, 3Armand M. Tarascon J.-M. Building better batteries.Nature. 2008; 451: 652-657Crossref PubMed Scopus (13971) Google Scholar The increasing lithium consumption has been projected to continue into the coming decades with a compound annual growth rate of 15%–30%, which poses a serious risk to the current supply chain in the near to mid-term future.4Crabtree G. The coming electric vehicle transformation.Science. 2019; 366: 422-424Crossref PubMed Scopus (71) Google Scholar,5Ding Y. Cano Z.P. Yu A. Lu J. Chen Z. Automotive Li-ion batteries: current status and future perspectives.Electrochem. Energ. Rev. 2019; 2: 1-28Crossref Scopus (249) Google Scholar Lithium from brines constitutes 66% of the world lithium resource, excluding oceans.6Gruber P.W. Medina P.A. Keoleian G.A. Kesler S.E. Everson M.P. Wallington T.J. Global lithium availability: a constraint for electric vehicles?.J. Ind. Ecol. 2011; 15: 760-775Crossref Scopus (347) Google Scholar This process relies on solar energy and arid climates for water evaporation over a period of 12–18 months, followed by multi-step chemical treatment and the precipitation of Li2CO3.6Gruber P.W. Medina P.A. Keoleian G.A. Kesler S.E. Everson M.P. Wallington T.J. Global lithium availability: a constraint for electric vehicles?.J. Ind. Ecol. 2011; 15: 760-775Crossref Scopus (347) Google Scholar These conditions limit the scaling up of lithium extraction from brines to meet long-term demands.7Kushnir D. Sandén B.A. The time dimension and lithium resource constraints for electric vehicles.Resour. Pol. 2012; 37: 93-103Crossref Scopus (125) Google Scholar,8Flexer V. Baspineiro C.F. Galli C.I. Lithium recovery from brines: a vital raw material for green energies with a potential environmental impact in its mining and processing.Sci. Total Environ. 2018; 639: 1188-1204Crossref PubMed Scopus (170) Google Scholar Supply chain issues may be mitigated by mining other Li sources. Unconventional sources, including seawater, flowback water, and geothermal reservoirs, contain abundant Li. Currently, seawater is not an economically viable source due to the dilute lithium concentrations, 0.177 ppm.9Diallo M.S. Kotte M.R. Cho M. Mining critical metals and elements from seawater: opportunities and challenges.Environ. Sci. Technol. 2015; 49: 9390-9399Crossref PubMed Scopus (83) Google Scholar Efficient and environmentally friendly Li extraction methods enabling access to the 230 billion tons of lithium in seawater could sustain a practically infinite supply. Methods for development of Li extraction have increased recently.10Razmjou A. Asadnia M. Hosseini E. Korayem A.H. Chen V. Design principles of ion selective nanostructured membranes for the extraction of lithium ions.Nat. Commun. 2019; 10: 5793Crossref PubMed Scopus (141) Google Scholar, 11Liu C. Li Y. Lin D. Hsu P.-C. Liu B. Yan G. Wu T. Cui Y. Chu S. Lithium extraction from seawater through pulsed electrochemical intercalation.Joule. 2020; 4: 1459-1469Abstract Full Text Full Text PDF Scopus (33) Google Scholar, 12Yang S. Zhang F. Ding H. He P. Zhou H. Lithium metal extraction from seawater.Joule. 2018; 2: 1648-1651Abstract Full Text Full Text PDF Scopus (101) Google Scholar In light of these developments, electrochemical extraction with Li-ion intercalation electrodes has demonstrated greater selectivity and recovery rates of lithium from brines relative to other methods, such as capacitive deionization and adsorption processes.13Battistel, A., Palagonia, M.S., Brogioli, D., Mantia, F.L., and Trócoli, R. Electrochemical methods for lithium recovery: a comprehensive and critical review. Adv. Mater. 32, 1905440.Google Scholar The mechanism of intercalation could provide selectivity to Li due to the differences in intercalation potential between Li and Na ions in the host material, which partially depends on relative differences of Li-O versus Na-O bond strengths for a given a crystal structure. Selectivity also depends on phase stability and differences in ion migration barriers.14Ong S.P. Chevrier V.L. Hautier G. Jain A. Moore C. Kim S. Ma X. Ceder G. Voltage, stability and diffusion barrier differences between sodium-ion and lithium-ion intercalation materials.Energy Environ. Sci. 2011; 4: 3680Crossref Scopus (1066) Google Scholar Therefore, the efficiency of lithium extraction largely relies on the choice of electrode material. So far, electrodes investigated include olivine LiFePO4,11Liu C. Li Y. Lin D. Hsu P.-C. Liu B. Yan G. Wu T. Cui Y. Chu S. Lithium extraction from seawater through pulsed electrochemical intercalation.Joule. 2020; 4: 1459-1469Abstract Full Text Full Text PDF Scopus (33) Google Scholar,15Pasta M. Battistel A. La Mantia F. Batteries for lithium recovery from brines.Energy Environ. Sci. 2012; 5: 9487Crossref Scopus (136) Google Scholar spinel λ-MnO2,16Kanoh H. Ooi K. Miyai Y. Katoh S. Selective electroinsertion of lithium ions into a platinum/λ-manganese dioxide electrode in the aqueous phase.Langmuir. 1991; 7: 1841-1842Crossref Scopus (57) Google Scholar,17Trócoli R. Erinmwingbovo C. LaMantia F. Optimized lithium recovery from brines by using an electrochemical ion-pumping process based on λ-MnO2 and nickel hexacyanoferrate.ChemElectroChem. 2017; 4: 143-149Crossref Scopus (62) Google Scholar and layered LiNi1/3Co1/3Mn1/3O2 (NMC),18Lawagon C.P. Nisola G.M. Cuevas R.A.I. Kim H. Lee S.-P. Chung W.-J. Li1−xNi0.33Co1/3Mn1/3O2/Ag for electrochemical lithium recovery from brine.Chem. Eng. J. 2018; 348: 1000-1011Crossref Scopus (46) Google Scholar which have mainly demonstrated Li selectivity in brine solutions (Table S1). All these previous works directly adopted conventional battery materials without tailoring the material structures and properties to the complex extraction from an aqueous environment. A fundamental understanding of the structure-selectivity relationship of the materials during competitive co-intercalation of both Li and Na is needed to improve the selectivity and stability of direct Li extraction from more dilute sources. Here, we demonstrate that a new material, from the layered cobalt oxide family, can achieve highly selective Li extraction through structural design. The selection of the layered cobalt oxide family is due to the large difference in stabilization energy between Li and Na, which shows a preference for Li intercalation, and the rich knowledge of its structural information during both Li and Na intercalation.14Ong S.P. Chevrier V.L. Hautier G. Jain A. Moore C. Kim S. Ma X. Ceder G. Voltage, stability and diffusion barrier differences between sodium-ion and lithium-ion intercalation materials.Energy Environ. Sci. 2011; 4: 3680Crossref Scopus (1066) Google Scholar Even though it has been successful in Li-ion batteries,19Mizushima K. Jones P. Wiseman P. Goodenough J. LixCoO2 (0<x⩽1): a new cathode material for batteries of high energy density.Solid State Ionics. 1981; 3–4: 171-174Crossref Scopus (245) Google Scholar layered cobalt oxide has never been investigated for the electrochemical extraction of Li. We discovered for the first time that, when immerged into a Na dominating water solution without any applied current, Li1−xCoO2 particles, with x above a threshold value, transfer into a core-shell structure with most of the core of Li0.94CoO2, a shell of Na0.51CoO2, and an intermediate phase connecting the two. The chemical compositions and spatial distribution of these phases in the (NaLi)1−xCoO2 is significant in governing a high Li selectivity with stable co-intercalation. The shell Na phase and the intermediate phase serve as a sink for further intercalation. The existence and connection of the core Li phase to the shell restricts layer expansion and promotes high selectivity toward Li. The Na0.51CoO2 shell also prevents further Na ion-exchange and maintains the stability of the electrode. During intercalation and deintercalation, the ratio of the Li phase/Na phase oscillates dynamically, which increases during intercalation and decreases during deintercalation, which indicates that Li dominates the charge transfer in the electrode. The (NaLi)1−xCoO2 electrode can extract Li selectively for a range of combined solutions of Li and Na, with a Li to Na mol ratio of 1:100 to 1:20,000. In the lowest Li concentration solution (Li:Na of 1:20,000), we achieved a recovery solution with a Li to Na mol ratio of ∼7.6:1 in a single electrochemical extraction, which is equivalent to a high Li selectivity of 1.5 × 105. Electrochemical extraction of Li works by extracting (electrochemical intercalation) Li from a dilute Li solution and then recovering (electrochemical deintercalation) Li in a fresh solution (schematics shown in Figure S1). The aqueous cycling stability of pure LiCoO2 (LCO) is good between 0.6 and 0.72 V versus Ag/AgCl, as shown in Figure S2. The stable cycling verified that LCO can be cycled in aqueous solution when an oxygen evolution reaction (OER) and a hydrogen evolution reaction (HER) are avoided.20Ruffo R. Wessells C. Huggins R.A. Cui Y. Electrochemical behavior of LiCoO2 as aqueous lithium-ion battery electrodes.Electrochem. Commun. 2009; 11: 247-249Crossref Scopus (211) Google Scholar,21Luo J.-Y. Cui W.-J. He P. Xia Y.-Y. Raising the cycling stability of aqueous lithium-ion batteries by eliminating oxygen in the electrolyte.Nat. Chem. 2010; 2: 760-765Crossref PubMed Scopus (687) Google Scholar For electrochemical extraction with layered cobalt oxide, vacancies are first created using an initial delithiation of parent LCO. An aqueous 1 M Na+ solution instead of Li+ was used as the electrolyte. From X-ray diffraction (XRD) analysis, the (003)LCO-r Bragg peak of the pristine LCO (R3¯m) shifts from 1.344 to 1.342 Å−1 after delithiation at a rate of 0.6 C for 1 h in Na+ solution (Figure S3; Notes S1 and S2). This phase transformation is far from the expected Li0.70CoO2.22Reimers J.N. Dahn J.R. Electrochemical and in situ X-ray diffraction studies of lithium intercalation in LixCoO2.J. Electrochem. Soc. 1992; 139: 2091Crossref Scopus (1425) Google Scholar New peaks appear, with the most notable being at 1.128 Å−1. The peaks match well for a combination of rhombohedral Li0.94CoO2 (R3¯m) and monoclinic Na0.51CoO2 (C2/m). The characteristic peak at 1.128 Å−1 is (003)NCO-m. The smaller Q value of the (003)NCO-m peak than the (003)LCO-r indicates an expansion of the layer spacing due to the intercalation of the Na+ (102 pm) relative to the smaller Li+ (76 pm), as well as a higher vacancy percentage and increased electrostatic repulsion. Once vacancies are formed, the Na/Li exchange occurs quickly without applied current, as characterized by in situ XRD during initial delithiation (Figure S4; Note S3). To maintain the charge neutrality, Na in the solution exchanges with Li in the structure and leaves the material with the same vacancy percentage as (NaLi)1−xCoO2 ((NL)1−xCO). Alkali metal exchange processes allow the creation of electrode materials with metastable structures and mixed compositions of Li and Na.23Mendiboure A. Delmas C. Hagenmuller P. New layered structure obtained by electrochemical deintercalation of the metastable LiCoO2 (O2) variety.Mater. Res. Bull. 1984; 19: 1383-1392Crossref Scopus (83) Google Scholar, 24Berthelot R. Carlier D. Pollet M. Doumerc J.-P. Delmas C. Synthesis and investigations on an O4-LiCoO2 polytype.Electrochem. Solid State Lett. 2009; 12: A207Crossref Scopus (20) Google Scholar, 25Paulsen J.M. Larcher D. Dahn J.R. O2 structure Li2/3Ni1/3Mn2/3O2: a new layered cathode material for rechargeable lithium batteries III. Ion exchange.J. Electrochem. Soc. 2000; 147: 2862Crossref Scopus (49) Google Scholar, 26Heubner C. Matthey B. Lein T. Wolke F. Liebmann T. Lämmel C. Schneider M. Herrmann M. Michaelis A. Insights into the electrochemical Li/Na-exchange in layered LiCoO2 cathode material.Energy Storage Mater. 2020; 27: 377-386Crossref Scopus (13) Google Scholar, 27Xiao B. Wang K. Xu G. Song J. Chen Z. Amine K. Reed D. Sui M. Sprenkle V. Ren Y. et al.Revealing the atomic origin of heterogeneous Li-ion diffusion by probing Na.Adv. Mater. 2019; 31: 1805889Crossref Scopus (16) Google Scholar These materials are formed by the (electro)chemical oxidation of the parent material to extract the original alkali ions, followed by an (electro)chemical reduction to insert new alkali ions. The ion-exchange of layered cobalt oxides has been performed; however, spontaneous ion-exchange with Li partially exchanged with Na after oxidation has not been reported. For this process to be understood, the dependence of the Na/Li exchange on the vacancy percentage was investigated, and the equilibrium of the same Li phase and Na phase was observed. Li1−xCoO2 at different vacancy percentage x was generated using 1 M Li+ solution. Figure 1A shows the shifting (003)LCO-r peak of LCO from delithiation due to the increased vacancy composition. Next, the samples were soaked in a 1 M Na+ solution overnight for cation exchange. As shown in Figure 1B, the (003)NLCO-r peak of Li1−xCoO2, with vacancies of x ≥ 0.06, shifts to 1.342 Å−1 and allows the emergence of the (003)NCO-m peak at 1.128 Å−1. The expected threshold should be greater than 6% of vacancy so that enough vacancy composition is available for both phases. The emergence of a Na phase at the 6% vacancy sample could be due to the heterogeneity, as demonstrated by the (003)LCO-r peak splitting shown in Figure 1A. The (003)NCO-m/(003)LCO-r intensity ratio increases with the vacancy percentage, with 0.104 at 6% vacancy and 1.093 at 30% vacancy, indicating a growing Na phase at higher vacancy percentage. The key material changes are summarized in Figure 1C. The illustration highlights the formation of the (NL)1−xCO from Li1−xCoO2 transforming into a combination of Li0.94CoO2 with a contracted layer spacing and Na0.51CoO2 with an expanded layer spacing and distorted lattice compared with the rhombohedral Li1−xCoO2. Since the spatial distribution of these phases would impact further intercalation, and therefore the selectivity of the electrode, the (NL)0.7CO was characterized to atomic resolution. Individual particles were thinned using a focused ion beam (FIB)/scanning electron microscope (SEM) for imaging using scanning transmission electron microscopy (STEM). Elemental characterization was achieved with energy dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS). The first area of interest was near the side of the particle at the electrode and electrolyte interface where the Na/Li exchange occurs. Pt was deposited on the outer surface of the particle before sample thinning, and its position labeled the electrode and electrolyte interface. Selected area diffraction patterns confirmed the STEM image view along the [110] zone axis (Figure S5). The STEM high-angle angular dark field (HAADF) image in Figure 2A shows four distinct regions by contrast. First, the bottom of the image consists of the protective Pt coating from the particle thinning procedure. Second, above the Pt region, there exists a darker contrast region that would have previously been the electrode and electrolyte interface. Third, a lighter-contrast region makes up the core of the particle. Fourth, a significant number of streaks run parallel to each other within the core region (clearly visible in Figure S6A). The particle surface was confirmed by the EDS data to be Na rich, with thicknesses varying between 10 and 25 nm (Figure 2B). Significant numbers of streaks were also confirmed to be Na, traveling through the core via the 2D planes and diffusing across the Co layers (Figure S6B). As expected, the EDS data show that Co and O are present. The center of the particle accommodates a small content of Na in the form of Na phase streaks; the Na phase and Li phase remain separated. The small phase content of Na indicates that the Li phase dominates the core of the (NL)0.7CO. The core-shell structure of (NL)0.7CO was confirmed by a second sample thinned at the center of the particle at the terminating CoO2 plane. The STEM HAADF images in Figure 2C show three of the four distinct regions in Figure 2A that are confirmed by EDS to be Pt coating, Na shell, and Li core (Figure S7). The surface thickness of the Na phase varies by 35–45 nm. The atomically sharp boundary (Figure S8) in contrast marks the Na phase transition to the Li phase set by a CoO2 layer. From both STEM samples, the Na phase constitutes the particle surface indicating a core-shell structure with Li phase as the core. The compositions of the observed Li and Na regions were further determined. EDS provides evidence that, despite coexisting in the same particle, the Li phase does not contain Na. However, Li might exist in the Na phase or in an additional phase. The Li presence was probed with EELS both in the core near the Na phase streaks (Figure 2D) and in the shell (Figure 2E). The EELS spectra in Figure 2F show the superposition of the Li-K and Co-M2,3 edges in the low-loss region.28Kikkawa J. Mizoguchi T. Arai M. Nagai T. Kimoto K. Identifying lithium K edge anisotropy in LiCoO2.Phys. Rev. B. 2018; 98: 075103Crossref Scopus (7) Google Scholar The spectra reveal three distinct sets of data that are grouped based on overall intensity and the onset edge slope between 59.1 and 61.9 eV. The overall intensity will be larger when Li and Co are present due to the superposition of their signals, and the onset slope will be steeper for a larger Li composition due to those energies being more characteristic of the Li-K edge. The first set (a–c), contains the spectra with the largest overall intensity and the largest onset edge slope, indicating the highest Li composition. Layers (a–c) have a 4.61 Å layer spacing that corresponds well with the expected Li0.94CoO2 layer spacing from XRD (4.62 Å). The second set (f–h) contains the spectra with a lower overall intensity and a flatter onset edge slope, indicating a decrease in the composition of Li. Layers (f–h) have a 4.92 Å layer spacing, indicating an intermediate layer spacing relative to the expected Li0.94CoO2 and Na0.51CoO2 phase layer spacings. The third set (d, e, i, and j) contains the spectra with smallest overall intensity and smallest onset edge slope. Layers (e and d) have a 5.33 Å layer spacing, and layers (i and j) from the shell have a 5.45 Å layer spacing, which correspond well with Na0.51CoO2 phase layer spacing (5.57 Å), indicating minimal presence of Li. The Li phase, Na phase, and their transition was further characterized with high-resolution STEM. Figures 2G and S9A ([110] zone axis) displays a clear image of the planes of Co atoms. Transition across the cobalt oxide layers was accomplished using an intermediate phase with a spacing of 4.92–5.05 Å from the Na phase (5.60 Å) to the Li phase (4.64 Å). The existence of multiple phases in a single layer leads to nonuniformities in layer spacing; particularly in the intermediate phase (Figures 2H and S9B). These nonuniformities are caused by strain from the Li phases and Na phases separated in the same plane and the complexities of the strain fields in a single multiphase particle. With the STEM characterization, we summarize that the (NL)0.7CO structure displays a Na phase (Na0.51CoO2) at the particle surface, a Li core (Li0.94CoO2) with a discrete Na phase, and an intermediate phase for the transition (Figure 2I). The Li0.94CoO2 and Na0.51CoO2 phases are exclusive to Li and Na ions, respectively. Next, we investigated the selectivity of core-shell (NL)0.7CO material to Li. A Na+ electrolyte (1 M) spiked with different concentrations of Li+ was used for the tests. The (NL)0.7CO electrode was cycled 20 times before the final intercalation (extraction). The 20-times cycling was to establish an equilibrium state between the electrode and the solution to avoid recovery of structural Li from the original (NL)0.7CO. Then the electrode was moved to a fresh solution (see the Experimental procedures) for recovery, followed by inductively coupled plasma-mass spectroscopy (ICP-MS) measurement of Li and Na concentrations. As shown in Figure 3A, (NL)0.7CO demonstrated high recoverability of Li+ in these solutions by electrochemical extraction at a rate of 0.6 C. The recovery was consistently high from solutions containing Li to Na ratios of 1:100, 1:1,000, 1:5,000, 1:10,000, and 1:20,000, with recovered mol ratios of Li to (Li + Na) of 89.5% ± 0.6%, 92.7% ± 2.2%, 89.3% ± 3.0%, 87.9% ± 1.1%, and 88.4% ± 0.5%, respectively, corresponding to Li to Na ratios of ∼8.5:1, 12.8:1, 8.3:1, 7.2:1, and 7.6:1 in the recovery solution. The accuracy of ICP-MS measurement of selectivity was confirmed by high agreement between the measured cations and the expected values from the electrochemical extraction reported as percentages of ions measured/ions deintercalated all within 7% of each other (Figure 3B). The ratios were most likely below 100% due to adsorption of deintercalated cations onto the electrode. The (NL)1−xCO electrodes reach Li selectivity factors (Figure S10A) as high as 1.5 × 105 in the 1:20,000 solution, a similar ratio to seawater. These results show promise for high Li recoverability from real seawater using the (NL)1−xCO electrode compared with other electrode materials (Table S1). The electrochemical extraction rates varied for solutions with Li to Na ratios of 1:1,000 and 1:20,000. These rates were important because they are related to the productivity of the mining method for Li. No strong dependence of the Li recoverability on the rate over the same state of charge was observed. For rates between 0.12 C and 1.20 C, 90.4% ± 0.6%, 93.8% ± 2.2%, 92.7% ± 2.2%, and 95.4% ± 0.6% Li to (Li + Na) mol ratios were measured for 1:1,000 Li to Na and 92.7% ± 1.2%, 84.7% ± 1.3%, and 88.4% ± 0.5% were measured for 1:20,000 Li to (Li + Na), as shown in Figures 3C and 3D. The increased overpotential from faster rates increases the probability of OER, which prevented the measurement in the 1:20,000 solution at a rate of 1.2 C. The electrochemical extraction rate can be adjusted without a Li selectivity trade-off (Figure S10B). Therefore, extraction can be accomplished at higher rates as long as the overpotential is not high enough to cause OER. In addition, the high selectivity is retained in seawater-like concentrations of 0.49 M 1:20,000 Li:Na at a rate of 0.60 C (Figure S10C). These results reveal the efficient capability of layer gradient spacing (NaLi)1−xCoO2 for extracting Li from seawater. To understand and confirm the high selectivity using (NL)0.7CO, a series of structure and composition analyses were conducted during 20 cycles in a 1:1,000 solution. In the potential curve (Figure 4A), the first intercalation from states of charges x = (0.300–0.195) for (NaLi)1−xCoO2 reveals a departure from LCO cycling behavior in a pure Li+ electrolyte. The (NaLi)1−xCoO2 potential curves have a greater range and more slope changes than LiCoO2 cycling in a pure Li+ electrolyte due to both phase transformations and a lower half cell potential from a dilute Li+ concentration. The potential reaches a similar voltage to Na0.51CoO2 in a Na-ion battery (∼0.1 V versus Ag/AgCl)29Berthelot R. Carlier D. Delmas C. Electrochemical investigation of the P2–NaxCoO2 phase diagram.Nat. Mater. 2011; 10: 74-80Crossref PubMed Scopus (927) Google Scholar at the end of intercalation, but overall remains at a higher voltage, possibly due to the co-existence of a Li0.94CoO2 phase and the Li ion being the major charge carrier ions. The intercalation curves following the 1st intercalation indicate a permanent change in the structure as they have shifted to higher potentials and are now characterized by two, new, shallower slopes, 0.195 ≤ x≤ 0.245 and 0.245 ≤ x≤ 0.287. As the number of cycles increases, the hysteresis of the intercalation/deintercalation curve decreases, which corresponds to a decrease in the kinetic polarization and free energy barrier representing the interfacial energy and coherency strain energy in a single, multiphase particle.30Van der Ven A. Garikipati K. Kim S. Wagemaker M. The role of coherency strains on phase stability in LixFePO4: needle crystallites minimize coherency strain and overpotential.J. Electrochem. Soc. 2009; 156: A949Crossref Scopus (110) Google Scholar The electrochemical curve shows a good reversibility for the material. To reveal the change of the (NL)1−xCO chemical composition during cycling, parallel samples were sacrificed at different intercalation/deintercalation cycles and the electrodes were dissolved for ICP-MS measurement. In Figure 4B, the LCO transformed into Li0.50Na0.20CoO2 after the initial deintercalation. The Li content is higher than that predicted using phase equilibrium between Li0.94CoO2 and Na0.51CoO2 for the initial deintercalated (NaLi)0.7CoO2. This is due to the formation of the intermediate phase (shown in Figures 2G–2I), which has a high Li content. Over the 20 cycles, the Li to Na ratio in the material was ∼7:3, with an oscillating pattern. Comparing samples in close time series, the Li content was observed to increase upon intercalation and decrease upon deintercalation (Figure 4B). This trend confirms that Li constitutes most of the charge transfer from the Li+-spiked solutions. Li is most likely not intercalated directly into the Li0.94CoO2 phase due to the increase in the migration barrier at low vacancy compositions.31Van der Ven A. Ceder G. Lithium diffusion mechanisms in layered intercalation compounds.J. Power Sources. 2001; 97–98: 529-531Crossref Scopus (212) Google Scholar Even at a low rate of C, the capacity cannot be fully used in a 1:1,000 aqueous solution for a (NL)1−xCO when x = 0.06 or x = 0.04 (Figure S11). As expected, the Na phase does not form in the low vacancy state. Moreover, synchrotron XRD confirms the oscillating pattern of material change during cycling. The spectra in Figure 4C show changes in the relative intensities of the Na phase and Li phase peaks; notably the (003)NCO-m and (003)LCO-r peaks (Figure 4D). The (003)NCO-m/(003)LCO-r ratio remains decreasing during each intercalation until the (003)NCO-m peak almost completely disappears by the 20th intercalation, but continues to partially reverse after deintercalation, when more vacancies are formed. The Na phase has an apparent change in the c-lattice parameter with each intercalation and deintercalation. The (003)NCO-m peak shifts to the same higher Q value (1.137 Å−1) during intercalation; thus indicating a decrease in the c-lattice parameter and a compression of the Na phase layers; potentially due to Na intercalation as well as the increased Li intercalation squeezing the Na to a lower vacancy phase.26Heubner C. Matthey B. Lein T. Wolke F. Liebmann T. Lämmel C. Schneider M. Herrmann M. Michaelis A. Insights into the electrochemical Li/Na-exchange in layered LiCoO2 cathode material.Energy Storage Mater. 2020; 27: 377-386Crossref Scopus (13) Google Scholar,27Xiao B. Wang K. Xu G. Song J. Chen Z. Amine K. Reed D. Sui M. Sprenkle V. Ren Y. et al.Revealing the atomic origin of heterogeneous Li-ion diffusion by probing Na.Adv. Mater. 2019; 31: 1805889Crossref Scopus (16) Google Scholar,32Kim H. Kwon D.-H. Kim J.C. Ouyang B. Kim H. Yang J. Ceder G. Na+ redistribution by electrochemical Na+/K+ exchange in layered NaxNi2SbO6.Chem. Mater. 2020; 32: 4312-4323Crossref Scopus (9) Google Scholar Upon deintercalation, the (003)NCO-m peak shifts back to its original position (1.127 Å−1) and the Na phase layers expand again. The shell Na phase maintains a Na composition, despite a dynamic phase volume, that stabilizes the structure by limiting further ion-exchange between Li in the (NL)1−xCO and Na in the solution. Concurrently, the (003)LCO-r peak remains in the same position and shows no signs of a change in layer spacing. This is consistent, and small layer spacing of the core Li phase prevents Na intercalation. For a L1−xCO material, there is no stabilizing Na phase shell to form the electrode-electrolyte interface. When L0.7CO is used for intercalation in the 1:1,000 solution, the Na ion-exchange and intercalation leads to a decrease in the Li composition of the electrode and a significant increase of the composition of Na (Figure S12). The differences in the 1st intercalation curve are due to differences in the initial material and ongoing Li/Na ion-exchange in the L1−xCO material. This reveals that the Li phase alone will not provide high selectivity during co-intercalation. Interestingly, a new peak appears that varies between 1.255 and 1.267 Å−1 for every intercalation step (Figure 4D, and with better resolution in Figure S13) and then disappears again upon each deintercalation. This peak has a low intensity and has no accompanying peaks; thus providing little information for phase determination. This peak corresponds to a 4.93 to 5.02 Å d-spacing, which matches well with the intermediate layer spacing regions from the core-shell (NL)0.7CO STEM images. The peak growth implies an increasing volume of the intermediate phase and its increased significance in a particle. During the cycling, the preferred Li intercalation may enable the conversion of the Na phase into the intermediate phase, which then precipitates out the Li phase. Because the intermediate phase has a layer spacing well below the 5.57 Å Na phase layer spacing, the energy barrier for layer expansion limits Na intercalation into the intermediate phase. The structural evolution leads to dominant intermediate and Li phases over the Na phase that enable high Li selectivity. The intermediate phase has the important role of storing extracted Li and limiting the large Na ion because of its small layer spacing. Similar trends were observed in the intercalation/deintercalation curve and XRD spectrum (Figure S14) for the 1:20,000 solution; thus supporting the same phase transformations in varying Li+ concentrations. The (NL)1−xCO material enabled the highly selective electrochemical extraction of Li from solutions with initial Li to Na ratios of 1:100 and 1:20,000. The solution with a Li to Na ratio of 1:20,000 was concentrated to a Li to Na ratio of ∼7.6:1, emphasizing the potential to successfully extract Li from real seawater. This high selectivity was made possible by the transformation of the parent Li1−xCoO2 into the core-shell (NL)0.7CO. The key components of this electrode are the shell Na phase, core Li phase, and the intermediate phase. Upon cycling in the dilute Li solutions, the phases transform into an intermediate phase plus a Li phase dominant structure at lower states of charge and a higher Na phase plus Li phase structure at higher states of charge. This complex phase relationship enables high Li extraction by storing Li in the intermediate phase and particle stability by having a core-shell mixed cation composition. These insights showcase the importance of understanding the phase transformation behavior due to cation co-intercalation as well as reveal the opportunities in developing new electrode materials.