Stabilized Electrodes

Abstract

An electrode material for extracting an elemental ion from a liquid medium includes at least one electrode material having at least one ion sieve that is capable of retaining or releasing an elemental ion, or a mixture of such ion sieves, wherein the ion sieve or ion sieves is or are coated with carbon.

Claims

1. An electrode material for extracting an element ion from a liquid medium, comprising at least one electrode material comprising at least one ion sieve capable of intercalating or releasing an element ion, or a mixture of such ion sieves, the at least one ion sieve being coated with carbon.

2. The electrode material as claimed in claim 1, wherein the element ion is a lithium ion.

3. The electrode material as claimed in claim 1, wherein the ion sieve after uptake of the element ion is a lithium-containing metal oxide or a lithium-containing metal phosphate.

4. The electrode material as claimed in claim 3, wherein the ion sieve is a complex oxide containing lithium and at least one further element comprising Co, Mg, Cr, Mn, Ni, Fe, Al, Mo, V, W or Ti or a lithium iron phosphate.

5. The electrode material as claimed in claim 4, wherein the lithium iron phosphate comprises LiFePO.sub.4, Li.sub.xMe.sub.yFePO.sub.4, LiFeMe.sub.yPO.sub.4 or a mixture thereof, where Me is Mn, Co, Mo, Ti, Al, Ni, Nb or a mixture thereof and 0<x<1; and 0<y<1.

6. The electrode material as claimed in claim 1, wherein the carbon layer is obtained by carbonization.

7. The electrode material as claimed in claim 1, wherein the ion sieve is in the form of particles.

8. An electrode for extracting an element ion from a liquid medium, comprising as electrode material as claimed in claim 1.

9. A process for extracting an element ion from a liquid medium, comprising: providing at least one electrode material as claimed in claim 1, and controlling intercalation and release of the element ion by applying a voltage.

10. The process as claimed in claim 9, comprising: a) providing an apparatus for electrodialysis, comprising an electrodialysis cell, dividing the electrodialysis cell into a lithium salt chamber and an alkali chamber by means of an anion-exchange membrane; filling the alkali chamber with salt brine; and filling the lithium salt chamber with a supporting electrolyte solution; b) placing a conductive substrate coated with a depleted ion sieve in the alkali chamber to act as a cathode, placing a conductive substrate coated with a lithium-intercalated ion sieve in the lithium salt chamber to act as an anode, wherein at least one of the two conductive substrates, includes the ion sieve as the electrode material, and wherein, during the performance of the electrodialysis, the ion sieve is capable of intercalating Li.sup.+ in the alkali chamber in order to undergo transformation into another lithium-intercalated ion sieve under an external electric potential, the lithium-intercalated ion sieve being capable of releasing Li.sup.+ into a conductive solution in order to undergo transformation into another ion sieve under the external electric potential, wherein after the intercalation and release of Li.sup.+ by the depleted ion sieve and the lithium-intercalated ion sieve respectively, enrichment occurs in the lithium salt chamber to afford a lithium-enriched solution.

11. The process as claimed in claim 9, wherein the concentration of the oxygen dissolved in the medium is lowered during the intercalation of the element ion, preferably by purging the medium with nitrogen.

12. An apparatus for performing the process as claimed in claim 9.

13. The use of the electrode material as claimed in claim 1 in a process for extracting an element ion from a liquid medium.

14. A process for extracting an element ion from a liquid medium using at least one ion sieve, comprising: controlling intercalation and release of the element ion by applying a voltage, and lowering a concentration of oxygen dissolved in a medium by purging the medium with a non-oxidizing gas.

Description

[0059] The exemplary embodiments are shown schematically in the figures. Identical reference numbers in the individual figures indicate identical or functionally identical elements or elements that correspond to one other in their functions. In the figures are shown, more particularly:

[0060] FIG. 1A-F Transmission electron micrographs of LiFePO.sub.4 (A and B) and LiFePO.sub.4/C (C and D); (E) X-ray diffractograms (F) and Raman spectra of LiFePO.sub.4 and LiFePO.sub.4/C;

[0061] FIG. 2A/B (A) Change in conductivity in the two channels and (B) lithium chloride removal capacity (left axis, squares) and energy consumption of the whole cell in 24 cycles in aqueous 10 mM LiCl (right axis, circles);

[0062] FIG. 3A-F Cyclic voltammograms (1 mV/s) of LiFePO.sub.4 in aqueous electrolytes having a concentration of LiCl, NaCl, MgCl.sub.2, and CaCl.sub.2 of 1 M (A), 100 mM (B), and 10 mM (C). (D and E) Specific capacity of LiFePO.sub.4 for selected LiCl mixtures recorded at 0.1 A/g (D: absolute value; E: relative value). (F) Nyquist plot and equivalent circuit for the system of the mixed electrolyte system;

[0063] FIG. 4A/B (A) Post-mortem X-ray diffraction patterns and (B) Raman spectra of delithiated LiFePO.sub.4 after 100 cycles in various electrolytes;

[0064] FIG. 5A-F (A) Comparison of the stability of LiFePO.sub.4 in 5 mM LiCl+50 mM NaCl with N.sub.2 purging, O.sub.2 purging or no pretreatment, operated in a three-electrode arrangement at 0.1 A/g. (B to D): Outflow concentration of lithium and sodium using LiFePO.sub.4 electrodes in 5 mM LiCl+50 mM NaCl (B) without treatment, (C) with O.sub.2 purging or (D) with N.sub.2 purging. (E) Comparison of lithium extraction capacity and corresponding capacity retention of LiFePO.sub.4 in 5 mM LiCl+50 mM NaCl without treatment, with O.sub.2 purging or with N.sub.2 purging. (F) Energy consumption of LiFePO.sub.4 in 5 mM LiCl+50 mM NaCl without treatment, with O.sub.2 purging, and with N.sub.2 purging;

[0065] FIG. 6A-F Comparison of the stability of LiFePO.sub.4 and LiFePO.sub.4/C in (A) 1 M LiCl and (B) 5 mM LiCl+50 mM NaCl solution in a three-electrode system at 0.1 A/g. (C) Postmortem X-ray diffractogram of LiFePO.sub.4 and LiFePO.sub.4/C after 100 cycles in 1 M LiCl and 5 mM LiCl+50 mM NaCl solution (1: LiFePO.sub.4 after 100 cycles in 5 mM LiCl+50 mM NaCl; 2: LiFePO.sub.4/C after 100 cycles in 5 mM LiCl+50 mM NaCl; 3: LiFePO.sub.4 after 100 cycles in 1 M LiCl; 4: LiFePO.sub.4C after 100 cycles in 1 M LiCl; 5: LiFePO.sub.4 electrode; 6: LiFePO.sub.4/C powder; 8: C PDF 89-8487; 7: LiFePO.sub.4 PDF 81-1173; 9: NaCl PDF 05-0628); (D) Outflow concentration of lithium and sodium using LiFePO.sub.4/C electrodes in 5 mM LiCl+50 mM NaCl with N.sub.2 purging. (E) Lithium separating capacity and retention in 5 mM LiCl+50 mM NaCl of LiFePO.sub.4/C or LiFePO.sub.4 in 5 mM LiCl+50 mM NaCl with N.sub.2 purging. (F) Energy consumption in 5 mM LiCl+50 mM NaCl from LiFePO.sub.4/C or LiFePO.sub.4;

[0066] FIG. 7A/B Calibration curves (relationship of ion concentration and characteristic peak intensity) for lithium (A) and sodium (B);

[0067] FIG. 8 Raman spectra of delithiated LiFePO.sub.4 at various points in a sample;

[0068] FIG. 9 Post-mortem X-ray diffractograms of LiFePO.sub.4 after 100 cycles in 5 mM LiCl+50 mM NaCl with N.sub.2 purging, with O.sub.2 purging, and without pre-treatment; (7: C PDF 89-8487; 8: LiFePO.sub.4 PDF 81-1173; 9: NaCl PDF 05-0628)

MATERIALS AND METHODS

[0069] Synthesis of LiFePO.sub.4, LiFePO.sub.4/C, and Electrode Preparation

[0070] Commercially available LiFePO.sub.4 (SGd-Chemie) having a specific surface area of 16 m.sup.2/g was used. LiFePO.sub.4/C was synthesized by a two-step procedure in analogous manner to Ma, Z.; Fan, Y.; Shao, G.; Wang, G.; Song, J.; Liu, T., In situ catalytic synthesis of high-graphitized carbon-coated LiFePO.sub.4 nanoplates for superior Li-ion battery cathodes. ACS Applied Materials & Interfaces 2015, 7, (4), 2937-2943. First, the LiFePO.sub.4 as delivered (0.4 g) and 0.21 g of ethylene glycol were dispersed in 10 mL of citric acid solution (0.18 g anhydrous citric acid in 10 mL distilled water) while stirring. LiFePO.sub.4 was added last. The suspension was heated to +80? C. for 3 hours to obtain the LiFePO.sub.4 coated with the alkyl ester compound. The alkyl ester coating was then carbonized in two heating steps in argon. First, the sample was heated to +200? C. for 2 h to remove the water in the materials, then the sample was heated to +700? C. and held at this temperature for 6 h to obtain the LiFePO.sub.4-carbon hybrid.

[0071] The electrodes were produced by mixing and grinding the active material (LiFePO.sub.4 or LiFePO.sub.4/C), acetylene black, and polyvinylidene fluoride (Sigma-Aldrich) in a mass ratio of 8:1:1 in 1-methyl-2-pyrrolidinone (Sigma-Aldrich) to form a slurry. The slurry was coated onto graphite paper (SGL, thickness 300 ?m) with a doctor blade (thickness 200 ?m) and then left overnight at room temperature in the hood. The electrode was then dried at +80? C. in a vacuum oven for 24 h.

Electrochemical Measurements

[0072] The electrochemical behavior of LiFePO.sub.4 and LiFePO.sub.4/C was investigated in a custom-built cell in a three-electrode system. Electrodes 12 mm in diameter made of LiFePO.sub.4, LiFePO.sub.4/C or activated carbon electrode (YP-80F, Kuraray) were used as the working electrode (LiFePO.sub.4, LiFePO.sub.4/C) or counter electrode (activated carbon). The electrodes were arranged in a sandwich and separated in the cell body by a glass fiber mat (diameter 13 mm, GF/A, Whatman). An Ag/AgCl reference electrode (3 M KCl, BASi) was mounted laterally in the cell body. Before the test, the assembled cells were filled with various electrolytes. For single salt electrolytes, 1 M LiCl, 1 M MgCl.sub.2, 1 M CaCl.sub.2 and 1 M NaCl were used. For the mixed salt electrolyte, 5 mM LiCl and 50 mM MeCl.sub.x (Me=Na, Ca, Mg, X corresponding to the Me charge) were used.

[0073] The cell was connected to the VSP300 potentiostat/galvanostat (Bio-Logic) and the cyclic voltammogram applied versus Ag/AgCl at a scan rate of 0.1 mV/s and a cutoff potential of ?0.4 V to +0.8 V. To test the performance stability of LiFePO.sub.4 and LiFePO.sub.4/C, potential-limited galvanostatic charge/discharge measurements were performed. Before the electrochemical operation, the electrolyte was continuously purged with nitrogen gas for 1 h to remove the dissolved oxygen in the electrolyte (only when investigating the effect of oxygen). The dissolved oxygen concentration after the gas purge is shown in Table 2.

[0074] In addition, the electrodes were delithiated prior to the electrochemical test to prevent a rise in the lithium concentration in the electrolyte as a result of the LiFePO.sub.4 and LiFePO.sub.4/C charging process. To delithiate the electrodes, the electrodes were charged with a current of 0.1 A/g and with a cutoff voltage of +0.4 V versus Ag/AgCl in a three-electrode system in 1 M LiCl electrolyte. Electrochemical impedance spectroscopy (EIS) was measured vs. Ag/AgCl at the formal potential in the frequency range from 1 MHz to 10 MHz and with an excitation voltage of 5 mV.

[0075] The conductivities of LiCl, MgCl.sub.2, CaCl.sub.2, and NaCl at various concentrations were tested using a microcell electrochemical HC cell with Pt electrodes (RHD Instruments) and a ModuLab electrochemical workstation (Solartron Analytical). 0.9 mL of each electrolyte was added to the measuring cup with a syringe and the Pt electrode crucible was then closed. A heat-conducting paste (Eurotherm 2000) was employed in the center of the closed cell and base unit to improve heat transfer therebetween. The potentiostatic impedance at each temperature was measured once the temperature had stabilized for 10 minutes. The impedance was measured from 1 Hz to 3 MHz at open circuit potential (OCV) at various temperatures from +10? C. to +60? C. in steps of ?10? C. and at +25? C. The conductivity and activation energy values were calculated according to equations 1 and 2.

[00001]$\begin{array}{cc}?=\frac{l}{\mathrm{AR}}& \mathrm{Equation}1\end{array}$

where ? stands for the conductivity (S/cm), R for the resistance (?), A for the area (cm.sup.2), and 1 is the length (cm). The value for 1/A was obtained from a 0.1 M KCl aqueous standard (VWR) having a conductivity of 12.880 mS/cm at +25? C.

[00002]$\begin{array}{cc}?=\frac{A}{T}{e}^{\left(-E_a/kT\right)}& \mathrm{Equation}2\end{array}$

where ? is the conductivity, T the temperature (K), A is as obtained in the experiment, k the Boltzmann constant (1.380649?10.sup.?23 J/K), and E.sub.a the activation energy (kJ/mol).

Lithium Selectivity Experiments

[0076] Lithium extraction experiments were performed in a multichannel cell. Two water channels were created through the seal (area=6.76 cm.sup.2, thickness=500 ?m) filled with glass fiber mat (GF/A, Whatman) and separated by an anion-exchange membrane (FAS-PET-130, Fumatech). The LiFePO.sub.4 or LiFePO.sub.4/C electrodes were situated at the end of the channel and were contacted with the graphite current collector. Before the experiment, one electrode was delithiated by the method described above. The channel with a pretreated electrode is referred to as channel 1 and the other as channel 2.

[0077] A 10 liter tank containing LiCl (10 mM) was used as feed water at a flow rate of 3 mL/min, while a constant current (40 mA/g) was applied at a cell voltage between ?0.4 V and +0.4 V using a VSP300 potentiostat/galvanostat system. Throughout the experiment, the electrolyte was continuously purged with N.sub.2 gas to remove dissolved oxygen. The mass of the electrode was 8 mg and the change in the conductivity and pH of two channels was recorded online by conductivity sensors (Metrohm, PT1000) and pH sensors (WTW SensoLyt 900P) respectively. Lithium is extracted throughout the cycle, accordingly the lithium removal capacity and energy consumption of the cell in one cycle were calculated according to equations 3 and 4.

[00003]$\begin{array}{cc}\mathrm{Lithium}\mathrm{chloride}\mathrm{extraction}\mathrm{capacity}\left({\mathrm{mg}}_{\mathrm{LiCl}}/e_{\mathrm{Electrode}}\right)=\frac{v.Math.M_{\mathrm{LiCl}}}{1000m_{\mathrm{total}}}??c_{\mathrm{Channel}1}\mathrm{dt}+?c_{\mathrm{channel}2}\mathrm{dt}& \mathrm{Equation}3\end{array}$

where v is the flow rate (mL/min), M.sub.LiCl the molecular weight of LiCl (42.4 g/mol), m.sub.total the mass of the electrodes (g), t the time (min), and ?c.sub.channel1 and ?c.sub.channel2 the change in the LiCl concentration (mmol/L) in channel 1 and channel 2 respectively.

[00004]$\begin{array}{cc}\mathrm{Energy}\mathrm{consumption}\left(\mathrm{Wh}/{\mathrm{mol}}_{\mathrm{LiCl}}\right)=\frac{-??E\mathrm{dq}.Math.1000}{3.6.Math.v.Math.?{\mathrm{ac}}_{\mathrm{channel}1}\mathrm{dt}+c_{\mathrm{channel}2}\mathrm{dt}}& \mathrm{Equation}4\end{array}$

where ?E is the cell voltage (V), q the charge (A-s), v the flow rate (mL/min), t the time (min), and ?c.sub.channel1 and ?c.sub.channel2 the change in the LiCl concentration (mmol/L) in channel 1 and channel 2 respectively.

[0078] To investigate the selectivity and stability of LiFePO.sub.4 and LiFePO.sub.4/C, we used an electrolyte containing 5 mM LiCl and 50 mM NaCl and having a volume of 10 L. The outlet of channel 2 was connected to an inductively-coupled plasma optical emission spectrometer (ICP-OES, ARCOS FHX22, SPECTRO Analytical Instruments) to quantify the change in concentration of the cations. The mass of the electrodes was about 18 mg; a low specific current of 30 mA/g and a flow rate of 1.2 mL/min were used to amplify the ICP signal. The calibration curve was constructed according to the correlation between the intensity of the individual wavelength and the concentration of the solution (FIGS. 7A and 7B). The measured intensities from the extracted sample were converted into concentration profiles. The charging and discharging processes had opposite courses and the potential for energy recovery was negligible, therefore the amount of lithium extracted and the energy consumption were calculated according to equations 5 and 6 in a half cycle in order to assess the performance of LiFePO.sub.4 and LiFePO.sub.4/C.

[00005]$\begin{array}{cc}\mathrm{Lithium}\mathrm{chloride}\mathrm{extraction}\mathrm{capacity}\left({\mathrm{mg}}_{\mathrm{Li}}/g_{\mathrm{electrode}}\right)=\frac{v.Math.M_{\mathrm{LiCl}}}{1000m_{\mathrm{total}}}??c_{\mathrm{Channel}2}\mathrm{dt}& \mathrm{Equation}5\end{array}$

where v is the flow rate (mL/min), M.sub.Li the molecular weight of Li (6.99 g/mol), m.sub.total the mass of the electrode (g), t the time over the lithium extraction step (min), and ?c.sub.channel2 the change in the Li.sup.+ concentration (mM) in channel 2.

[00006]$\begin{array}{cc}\mathrm{Energy}\mathrm{consumption}\left(\mathrm{Wh}/{\mathrm{mol}}_{\mathrm{Li}}\right)=\frac{-??E\mathrm{dq}.Math.1000}{3.6.Math.v.Math.?c_{\mathrm{channel}2}\mathrm{dt}}& \mathrm{Equation}6\end{array}$

where ?E is the cell voltage (V), q the charge (A.Math.s), v the flow rate (mL/min), t the time over the lithium extraction step (min), and c.sub.channel2 the Li concentration (mM) in channel 2.

Material Characterization

[0079] The surface morphology of LiFePO.sub.4 and LiFePO.sub.4/C was investigated by scanning electron microscopy (JEOL JSM 7500F) at 1 kV. X-ray diffractometry was performed in a D8 Advance diffractometer (Bruker AXS) with a copper X-ray source (Cu-K?, 40 kV, 40 mA) and a Goebel mirror in point focus (0.5 mm). Raman spectra were recorded with a Renishaw inVia system using a Nd:YAG laser with an excitation wavelength of 532 nm. The spectral resolution was 1.2 cm.sup.?1, and the diameter of the laser spot on the sample was about 2 ?m at a total power consumption of 0.2 mW.

Characterization of LiFePO.sub.4 and LiFePO.sub.4/C

[0080] FIG. 1A shows transmission electron micrographs of LiFePO.sub.4. LiFePO.sub.4 particles are typically about 300 nm long and about 150 nm broad. Discernible at higher resolution (FIG. 1B) are lattice bands with the typical spacing of 0.34-0.35 nm in alignment with the (111) and (021) planes of LiFePO.sub.4. LiFePO.sub.4/C shows the ubiquitous presence of the carbon that coats the LiFePO.sub.4 particles (FIGS. 1C and 1D).

[0081] FIG. 1E shows the X-ray diffractogram of LiFePO.sub.4 and LiFePO.sub.4/C. The diffraction patterns of LiFePO.sub.4/C are the same as those of LiFePO.sub.4, identified as an orthorhombic phase (PDF 81-1173), which indicates that the carbon-coating process does not destroy the inherent structure of LiFePO.sub.4. FIG. 1F shows the Raman spectra of LiFePO.sub.4 and LiFePO.sub.4/C. The band at 952 cm.sup.?1 is attributable to the symmetric stretching of the PO bonds of LiFePO.sub.4. And the peaks at 1338 cm.sup.?1 and 1598 cm.sup.?1 correspond to the D band and G band of carbon respectively. The D band (disordered peak) is attributable to the A.sub.1g vibrational mode and the G band (graphitic peak) to the E.sub.2g vibrational mode of CC bond stretching. Therefore, both commercial LiFePO.sub.4 and LiFePO.sub.4/C show the presence of incompletely graphitic carbon; in the case of the former, carbon is a minority phase.

Lithium Extraction from LiFePO.sub.4 in Aqueous 10 mM LiCl

[0082] The Li removal capacity of LiFePO.sub.4 in aqueous 10 mM LiCl was quantified. When a specific current of 30 mA/g is applied during charging, the conductivity in channel 1 decreases, while the conductivity in channel 2 increases (FIG. 2A). This corresponds to reduction of the cathode with uptake of lithium and oxidation of the anode with release of lithium into the feed water stream (equation 7). The reverse process is observed during the discharge process. The lithium extraction capacity of the cell does not decline during 24 cycles with an average capacity of 79?6 mg.sub.LiCl/electrode, corresponding to 13?1 mg.sub.Li/electrode (FIG. 2B), suggesting that LiFePO.sub.4 is stable in deaerated aqueous 10 mM LiCl. The average energy consumption of the cell is 3 Wh/mol.sub.LiCl, corresponding to 3 Wh/mol.sub.Li.

Li.sub.1-xFePO.sub.4+xLi.sup.++xe.sup.?=LiFePO.sub.4Equation 7

Effect of Other Cations by Comparison with Lithium

[0083] Cyclic voltammetry was carried out to investigate the selectivity behavior of LiFePO.sub.4 toward Li.sup.+ by comparison with other cations. This was done by comparing the electrochemical performance in various single-cation electrolytes. As shown in FIGS. 3A, 3B, and 3C, a pair of redox peaks occurs in all electrolyte types. The measured peak current was lowest when using an aqueous 1 M CaCl.sub.2 electrolyte. With 1 M LiCl, the redox peak currents are highest and the peak potentials are shifted to a more positive range by comparison with the other electrolytes, which could be attributable to the contribution of lithium in the solution leaching from the electrode.

[0084] To further quantify the potential shift, the formal potential (E.sub.1/2) was investigated. As shown by Table 3, the formal potential in LiCl solution decreases with decreasing concentration. During the charging process (the first electrochemical process during cyclic voltammetry), the Li.sup.+ in LiFePO.sub.4 is released into the electrolyte. Therefore, in tests in electrolytes containing other cations, such as NaCl, MgCl.sub.2 or CaCl.sub.2, the cyclic voltammograms and the formal potential are almost identical.

[0085] In order to reduce the loss of additional lithium ions from the electrode to the electrolyte and in order to compare electrolytes having differing concentrations, the LiFePO.sub.4 electrodes were delithiated prior to the galvanostatic charge/discharge test. A significant fall in the initial capacity of LiFePO.sub.4 was observed in all electrolytes (FIGS. 3D and 3E). The capacity decreases most slowly in the mixed solution of Li.sup.+ and Na.sup.+ and most rapidly in Ca.sup.2+-containing electrolytes, with a retention after 100 cycles of 60% and 46% respectively. The capacity in 5 mM LiCl+50 mM MgCl.sub.2 solution initially increases and then decreases; the increase may be attributable to the intercalation of trace amounts of Mg.sup.2+. By comparison with other cations, Mg.sup.2+ intercalates into the structure of LiFePO.sub.4 more readily. However, the intercalation of both Mg.sup.2+ and Li.sup.+ is not fully reversible, consequently capacity continues to decrease. To investigate why the capacity of LiFePO.sub.4 electrodes decreases during cycles, the post-mortem X-ray diffractogram of delithiated LiFePO.sub.4 charging and discharging 100 cycles in various electrolytes was measured.

[0086] In addition to the influence on stability, cations other than Li.sup.+ also influence the kinetics obtained from electrochemical impedance measurements. The Nyquist plots consist of two semicircles and a line (FIG. 3F). The semicircle at high frequency represents the capacity response of the electrolyte, thus it occurs in a significant amplitude only when the electrolytes are investigated with low concentration. The semicircle at medium frequency represents the charge-transfer resistance (R.sub.ct). The oblique line at low frequency represents the Warburg impedance (W.sub.1), which corresponds to the diffusion of lithium ions. The value for R.sub.ct and w.sub.1 is calculated according to the simulated circuit as shown in Table 1. Ca.sup.2+ and Mg.sup.2+ have an adverse effect on the charge-transfer process, with R.sub.ct values of 38.7 ?.Math.cm.sup.2 and 37.3 ?.Math.cm.sup.2 respectively, which are almost twice as high as the R.sub.ct value in pure 5 mM LiCl (20.1 ?.Math.cm.sup.2). The R.sub.ct values in 5 mM LiCl and 50 mM NaCl are similar to those in pure LiCl electrolyte. The effect of other cations on the Warburg impedance follows the pattern observed for the R.sub.ct value, and the Warburg impedances in Mg.sup.2+- and Ca.sup.2+-containing electrolyte are higher than those in Na.sup.+-containing and pure LiCl. This result indicates that lithium ions diffuse into the liquid side less easily in the former electrolytes than in the latter ones. The differing effects have their origin in the differing conductivities of the cations.

[0087] As can be seen in FIG. 4A, the delithiated LiFePO.sub.4 is composed of heterosite FePO.sub.4 and LiFePO.sub.4, since the LiFePO.sub.4 is not fully delithiated. This result is in agreement with the Raman data (FIG. 8). When Raman measurements are carried out at two randomly selected points, there is only one peak at point 1 (952 cm.sup.?1), this being consistent with the Raman spectra of LiFePO.sub.4 (FIG. 1F). In other places, the observed peaks correspond to FePO.sub.4. After 100 cycles in 5 mM LiCl, some diffraction peaks such as the 37.4 ? 2? ((211) plane of heterosite FePO.sub.4) disappear, indicating break-up of the structure of FePO.sub.4. Different behavior is observed after 100 cycles in Ca.sup.2+-containing electrodes, in which the diffraction peaks at 18.13? 2?, corresponding to the (020) plane of heterosite FePO.sub.4, disappear, pointing to significant loss of capacity. FIG. 4B shows the Raman spectra of the electrodes after 100 cycles in various electrolytes. After charging/discharging in NaCl, LiCl, and CaCl.sub.2 electrolytes, the peaks at 488 cm.sup.?1 (Li cage/asymmetric bending of PO.sub.4.sup.3?), 596 cm.sup.?1 (asymmetric bending of PO.sub.4.sup.3?, ?.sub.4), 652 cm.sup.?1 (symmetric bending of PO.sub.4.sup.3?, ?.sub.2), and 691 cm.sup.?1 are not apparent, indicating breakage of the PO bond; this is a further reason for the decline in performance of LiFePO.sub.4, namely the loss of oxygen species.

Effect of the Oxygen Content in Electrolytes

[0088] A mixed electrode containing 5 mM LiCl and 50 mM NaCl as feed water was used to investigate the influence of the oxygen content on the stability of LiFePO.sub.4. The spur for this choice was the observation that sodium ions have the lowest influence on the stability of LiFePO.sub.4. First, the stability of LiFePO.sub.4 at 100 mA/g was tested with a potential range of ?0.4 V to +0.5 V versus Ag/AgCl in 5 mM LiCl+50 mM NaCl solution, which was continuously purged with O.sub.2 and N.sub.2 prior to the electrochemical process. After 100 cycles, LiFePO.sub.4 retains 69% of its initial capacity in the N.sub.2-purged electrolytes, whereas the capacity retention values in the O.sub.2 purged electrolytes and in untreated electrolytes are 43% and 52% respectively (FIG. 5A). This suggests that dissolved O.sub.2 in the electrolyte significantly accelerates the loss in performance of LiFePO.sub.4.

[0089] The outflow solution was continuously analyzed by online monitoring using inductively-coupled plasma optical emission spectroscopy (ICP-OES). As shown in FIGS. 5B, 5C, and 5D, LiFePO.sub.4 retains good selectivity toward lithium throughout, irrespective of the oxygen content. Similar to the trend in the specific capacity (FIG. 5A), the LiFePO.sub.4 tested in feed water with continuous N.sub.2 purging was the most stable, with 70% capacity retention after 10 cycles (FIG. 5E). The lithium extraction capacity is only about half the initial value in feed water without any treatment, which is slightly higher than in the feed water with continuous O.sub.2 purging. At higher oxygen content, the capacity decreases with a greater amplitude. In contrast to the Li.sup.+ extraction/recovery capacity, the energy consumption is stable during the 10 cycles, as shown in FIG. 5F. This result indicates that the oxygen content has little influence on the energy consumption of the system, but a large influence on the structural and chemical stability of LiFePO.sub.4 itself.

[0090] To investigate how dissolved oxygen in the electrolyte influences the stability of LiFePO.sub.4, X-ray diffraction was used. This examined structural changes in the electrode material after 100 cycles in an electrolyte consisting of 5 mM LiCl and 50 mM NaCl (FIG. 9). Unlike in the measurement of the powder only, the cast electrodes show strong reflections from the graphite paper (26.5? 2? and 54.6? 2?), consequently the diffraction pattern was normalized to the peaks not associated with the graphite foil. Compared to the LiFePO.sub.4 and LiFePO.sub.4 electrode, all electrodes exhibit after 100 cycles two additional peaks at 31.7? 20 and 45.5? 2?, corresponding to the (200) plane and the (220) plane respectively; these peaks relate to the presence of a highly symmetrical phase (residual salt). The electrode cycled in the electrolyte with less dissolved oxygen (N.sub.2 purging) shows lower peak broadening compared to the sample with O.sub.2 purging. The higher amount of dissolved oxygen seemingly results in greater degradation of the LiFePO.sub.4 material.

Improved Stability Through Carbon Coating of LiFePO.SUB.4

[0091] To increase the stability of LiFePO.sub.4, LiFePO.sub.4 had layers of carbon bonded onto it to prevent attack by oxygen. FIG. 6A shows the difference between LiFePO.sub.4 and LiFePO.sub.4/C in aqueous 1 M LiCl. As can be seen, the capacity of LiFePO.sub.4/C decreases slightly from 115 mAh/g to 95 mAh/g, with retention of 83% of the initial capacity after 100 cycles. By way of comparison: LiFePO.sub.4 without carbon coating retains only 30% of the initial capacity, with a decline from 92 mAh/g to 27 mAh/g after 100 cycles. Performance at low concentration (5 mM LiCl+50 mM NaCl) is similar (FIG. 6B). LiFePO.sub.4/C still has 85% capacity in the first cycle (41 mAh/g and 48 mAh/g), whereas LiFePO.sub.4 retains only 52% of the initial value.

[0092] FIG. 6C shows the X-ray diffractograms of LiFePO.sub.4/C and LiFePO.sub.4 electrodes after testing in 1 M LiCl and 5 mM LiCl+50 mM NaCl.

[0093] To investigate further whether the carbon coating is able to improve the stability of LiFePO.sub.4 during lithium extraction, the performance of LiFePO.sub.4/C in 10 mM LiCl and 5 mM LiCl+50 mM NaCl solution (continuous N.sub.2 purging) was tested using the rocking chair cell. In the same way as LiFePO.sub.4, LiFePO.sub.4/C also exhibits good selectivity toward lithium (FIG. 6D). However, LiFePO.sub.4/C has a higher lithium extraction capacity (21.0 mg.sub.Li/g electrode compared to 17.8 mg.sub.Li/g electrode in the first cycle) and better stability, with 82% retention after carbon coating, as shown in FIG. 6E. The higher capacity of LiFePO.sub.4/C could be due to the carbon coating improving the electronic conductivity of the LiFePO.sub.4 (Li, J.; Qu, Q.; Zhang, L.; Zhang, L.; Zheng, H., A monodispersed nano-hexahedral LiFePO.sub.4 with improved power capability by carbon-coatings. Journal of alloys and compounds 2013, 579, 377-383), i.e. the active materials can be fully utilized at high current. In addition, the carbon layer is able to block attack by oxygen and OH.sup.? and increase the stability of LiFePO.sub.4 (He, P.; Liu, J.-L.; Cui, W.-J.; Luo, J.-Y.; Xia, Y.-Y., Investigation on capacity fading of LiFePO.sub.4 in aqueous electrolyte. Electrochimica Acta 2011, 56, (5), 2351-2357). In addition, the additional carbon present in nanohybridized form in LiFePO.sub.4/C could effectively reduce the resistance of LiFePO.sub.4, which lowers energy consumption (FIG. 6F). The average energy consumption of LiFePO.sub.4/C in 10 cycles is 3.0?0.5 Wh/mol.sub.Li.

[0094] The influence of the cations and of dissolved oxygen on the stability of LiFePO.sub.4 was investigated in a symmetrical cell (i.e. LiFePO.sub.4 paired with LiFePO.sub.4). Cations other than Li.sup.+ in brine, such as Na.sup.+, Mg.sup.2+, and Ca.sup.2+, influence the stability and electrochemical properties of LiFePO.sub.4. Of these, Ca.sup.2+ has the most adverse effect, and Na.sup.+ shows no apparent influence in the potential range from ?0.4 V to +0.8 V. The stability and electrochemical properties of LiFePO.sub.4 are impaired by Ca.sup.2+. Dissolved oxygen likewise exacerbates the fading of LiFePO.sub.4. Lowering the concentration of dissolved oxygen (N.sub.2 purging) dramatically increases the capacity retention in 10 cycles in 5 mM LiCl?50 mM NaCl from 47% to 70%. After carbon coating, the retention increases further to 82% and energy consumption falls to 3.0?0.5 Wh/mol.sub.Li. These two methods improve the performance stability of LiFePO.sub.4 as a material for lithium extraction. Whereas lowering dissolved oxygen may not be practical in to-scale applications, the carbon coating of LiFePO.sub.4 represents a simple and very promising approach for larger scale uses too.

TABLE-US-00001 TABLE 1 Adjusted values for delithiated LiFePO.sub.4 in various electrolytes according to the equivalent circuit diagram. R.sub.s R.sub.1 R.sub.ct W.sub.1 (? .Math. Electrolyte (?) (?) (?) s.sup.?1/2) ?.sup.2 ?.sup.2/lZl 5 mM LiCl 6.8 1.1 40.0 63.4 315.2 1.67 ? 10.sup.?2 5 mM LiCl + 0.4 1.0 38.7 51.8 210.5 1.36 ? 10.sup.?2 50 mM NaCl 5 mM LiCl + 2.2 0.9 74.2 76.2 265.7 5.98 ? 10.sup.?3 50 mM MgCl.sub.2 5 mM LiCl + 1.1 0.8 77.0 71.4 669.4 7.24 ? 10.sup.?3 50 mM CaCl.sub.2

TABLE-US-00002 TABLE 2 Concentration of dissolved oxygen after O.sub.2 purging and N.sub.2 purging Condition O.sub.2 concentration (ppm) Initial 8.8 O.sub.2 purging for 24 h 12.5 N.sub.2 purging for 24 h 4.0

TABLE-US-00003 TABLE 3 Formal potential of LiFePO.sub.4 in LiCl, NaCl, MgCl.sub.2, and CaCl.sub.2 at a concentration of 1M, 100 mM, and 10 mM. Average potential E.sub.f Electrolyte (V vs. Ag/AgCl) 1M LiCl 0.18 1M NaCl 0.08 1M MgCl.sub.2 0.1 1M CaCl.sub.2 0.07 100 mM LiCl 0.12 100 mM NaCl 0.09 100 mM MgCl.sub.2 0.10 100 mM CaCl.sub.2 0.03 10 mM LiCl 0.09 10 mM NaCl 0.07 10 mM MgCl.sub.2 0.08 10 mM CaCl.sub.2 0.06

CITED LITERATURE

[0095] Kanoh, H.; Ooi, K.; Miyai, Y.; Katoh, S., Electrochemical recovery of lithium ions in the aqueous phase. Separation Science and Technology 1993, 28, (1-3), 643-651. [0096] Pasta, M.; Battistel, A.; La Mantia, F., Batteries for lithium recovery from brines. Energy & Environmental Science 2012, 5, (11), 9487. [0097] Tr?coli, R.; Battistel, A.; La Mantia, F., Nickel hexacyanoferrate as suitable alternative to Ag for electrochemical lithium recovery. ChemSusChem 2015, 8, (15), 2514-2519. [0098] Kim, J.-S.; Lee, Y.-H.; Choi, S.; Shin, J.; Dinh, H.-C.; Choi, J. W., An electrochemical cell for selective lithium capture from seawater. Environmental Science & Technology 2015, 49, (16), 9415-9422. [0099] Ma, Z.; Fan, Y.; Shao, G.; Wang, G.; Song, J.; Liu, T., In situ catalytic synthesis of high-graphitized carbon-coated LiFePO.sub.4 nanoplates for superior li-ion battery cathodes. ACS Applied Materials & Interfaces 2015, 7, (4), 2937-2943. [0100] Li, J.; Qu, Q.; Zhang, L.; Zhang, L.; Zheng, H., A monodispersed nano-hexahedral LiFePO.sub.4 with improved power capability by carbon-coatings. Journal of alloys and compounds 2013, 579, 377-383. [0101] He, P.; Liu, J.-L.; Cui, W.-J.; Luo, J.-Y.; Xia, Y.-Y., Investigation on capacity fading of LiFePO.sub.4 in aqueous electrolyte. Electrochimica Acta 2011, 56, (5), 2351-2357.