Synthesis, Characterization, and Electrochemistry of Layered

16 mins ago - Metrics. Article Views: 0 Times. Received 3 April 2018. Published online 29 May 2018. +. Altmetric Logo Icon More Article Metrics. CURRE...
6 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/IC

Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

Synthesis, Characterization, and Electrochemistry of Layered Chalcogenides LiCuCh (Ch = Se, Te) Martin Valldor,*,† Daria Mikhailova,† Lars Giebeler,† Kwing To Lai,‡ Lena Spillecke,† Hans-Joachim Grafe,† and Bernd Büchner† †

Leibniz Institute for Solid State and Materials Research (IFW) e.V., Helmholtzstraße 20, 01069 Dresden, Germany Department of Physics, The Chinese University of Hong Kong, Shatin, Hong Kong

‡

S Supporting Information *

ABSTRACT: Two novel compounds, LiCuCh (Ch = Se or Te), were synthesized by direct reaction between elements in closed ampules inside corundum crucibles. Both compounds are highly air-sensitive and possess an anti-PbClF crystal structure, which contains CuCh layer analogues to the Fe[As/Se] layers in Fe-based superconductors. In electrochemical battery cells, Li can be almost completely extracted from LiCuSe, but the reverse reaction is only partly successful and Li2Se and Cu2−xSe are formed instead. LiCuSe exhibits a temperature independent and slightly positive magnetic susceptibility. From 7Li NMR measurements, the activation energy of the Li ion diffusion process is about 0.5 eV but is slightly lower for LiCuTe as compared to LiCuSe. Also, the small and almost temperature independent NMR shifts of the 7Li nucleus indicate the absence of Pauli paramagnetism in these compounds, consistent with a 3d10 full valence state of the Cu ions.

1. INTRODUCTION The Fe[As/Se] layers are the key structural feature related to superconductivity in Fe1+xSe (Tc = 8 K) with an anti-α-PbO structure.1 By introducing ions or molecules between the FeSe layers, the superconducting temperature could be raised, e.g., up to 44 K with Li+,2 43 K with Li+−NH2−NH3,3 and 30 K with K+.4 The closely related Li[FeAs], with an anti-PbClF structure and [FeAs] layers, is superconducting below Tc = 16 K5, and its properties can be altered by partial extraction of Li through electrochemical treatments.6 It is also common in Febased superconductors that the tetrahedrally coordinated transition metal can be substituted for, e.g., Co and Ni, as in NaFe1−x(Co,Ni)xAs and BaFe2−x(Co,Ni)xAs2,6 but Cu is only rarely reported in such structures, e.g., in pnictide Li1+xCu2−xP27 and as dopant in BaFe2As28 as well as in NaFeAs.9 However, rather than superconductivity, similar CuCh (Ch = Se, Te) layers can be found in, for example, thermoelectric BiCuSeO10 and in transparent semiconductor LaCuSO.11,12 On the other hand, intercalations of Li ions in layered structures has wide use for the energy storage, as in Li-ion battery cathode materials LiCoO2.13 Therefore, Li sandwiched between CuCh layers might constitute a novel cathode material. In fact, the chemical system Li−Cu−Se already was tested as electrodes in Li-ion batteries, but, in those investigations, only binary Cu−Se compounds were mentioned as identified crystalline species.14,15 Similarly, Li-battery storage properties but with only a binary crystalline constituent were also reported in the Li− Cu−Te system.16 During explorative chemistry investigations in the Li−Cu−Ch−O system (Ch = Se or Te), two novel, layered chalcogenides LiCuCh (Ch = Se or Te) could be © XXXX American Chemical Society

identified having an anti-PbClF structure. Here, we report on their syntheses, basic properties, and lattice dynamics. Further, electrochemical extraction of Li from LiCuSe was performed to determine the flexibility of its Li content and with the aim to induce magnetism on Cu.

2. EXPERIMENTAL DETAILS 2.1. Synthesis. Both samples were obtained by reacting pure elements in stoichiometric amounts in closed vessels under inert conditions; inside an argon-filled glovebox (Glovebox-Systemtechnik), Cu powder (MaTeck 99.8%) was thoroughly homogenized with either Se (Alfa-Aesar 99.5%) or Te (Alfa-Aesar 99.99%) in agate mortars. Subsequently, each mixture was quantitatively transferred into a tubelike corundum crucible. The corresponding amount of Li (MaTeck 99.4%) was added as pieces on top of each powder mixture. Caution! Freshly prepared Li surfaces, e.g., by cutting, might ignite with the mixture and the Li pieces should, for safety reasons, not be f reshly cut. Loaded crucibles were carefully placed in silica tubes that were temporarily shut by a rubber plug, before transporting them out of the glovebox. Each silica tube was melt sealed after lowering the inner pressure to about 0.2 bar. The samples were placed in normal resistance furnaces under fume hoods. Caution! In case of tube breakage during heating, Se and Te will oxidize into poisonous oxides with relatively high vapor pressures. Hence, these reactions should be performed under f ume hoods and broken vessels should only be removed when wearing extra breathing protection. The samples were slowly heated (1° min−1) up to 720 °C and heated isothermally during 15 h before the furnace was shut off, and a cooling at ambient rate followed. After the first heating, both samples appeared to be inhomogeneous and contained dice-like, black Received: April 3, 2018

A

DOI: 10.1021/acs.inorgchem.8b00897 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 1. X-ray diffraction data as compared to Rietveld model (black line) for LiCuCh (Ch = Se (a) or Te (b)), below which the Bragg positions are indicated and the difference between observations (Iobs) and models (Icalc) are plotted. Li2Ch (Ch = Se or Te) crystals, so the samples were homogenized again and re-reacted by the same procedure. Red-lilac LiCuSe needed three heating treatments to be visibly homogeneous but red-brown LiCuTe only two. 2.2. X-ray Diffraction and Synchrotron Investigations. For phase analyses and the test of air sensitivity, powder X-ray diffraction by a Huber camera with a CoKα1 X-ray source (λ = 1.78892 Å) and an image plate detector was used; the samples were placed between two X-ray transparent foils with grease to slow the reaction with air. To obtain data for Rietveld simulations, a STOE STADI_P diffractometer (Debye−Scherrer geometry) with MoKα1 radiation (λ = 0.70926 Å) and a position-sensitive detector (PSD) were used. To completely avoid reactions with air, each powder sample was placed in a glass capillary (Mark tubes, Glass No. 50, Hilgenberg), inside a glovebox, that was subsequently melt sealed. Synchrotron powder diffraction ex situ experiments on pristine cathode material LiCuSe and the Li0.25CuSe (“CuSe”) and Li1.9CuSe (“Li2CuSe”) compositions were performed at the beamline BL04-MSPD17 at ALBA (Barcelona, Spain). For this, electrochemical cells were charged to 3.2 V or discharged to 0.7 V vs Li+/Li and immediately dissembled in the glovebox. The materials were washed with dimethyl carbonate (DMC), dried under vacuum at room temperature, and put into glass capillaries, which were sealed without any contact of materials to air. The initial LiCuSe cathode was also sealed in the glass capillary without contact to air. The wavelength of λ = 0.32510(1) Å was determined from the positions of 8 reflections from a LaB6 reference material. All diffraction pattern simulations (Rietveld calculations) were done with the JANA2006 software.18 2.3. Elemental Analyses. As Li is difficult to detect with X-rays, the samples’ bulk compositions were investigated with inductively coupled plasma-optical emission spectroscopy (ICP-OES) using an iCAP 6500 Duo View (Fa. Thermo Fisher Scientific GmbH). Each sample was tested thrice, and, for each measurement, about 25 mg of powder was digested in dilute nitric acid. 2.4. Electrochemical Investigations. Electrochemical studies on LiCuSe electrodes were performed with the VMP3 (Biologic, France) multichannel potentiostatic−galvanostatic system in standard twoelectrode Swagelok-type cells with metallic Li discs (Chempur, 250 μm thickness) as anode. For the positive electrode, a mixture of LiCuSe, super P carbon (Timcal), and polyvinylidene difluoride (PVDF, Solef 21216) as polymer binder in an 80:10:10 weight ratio was pressed onto copper meshes with 10 mm in diameter and dried under vacuum at 80 °C overnight. As electrolyte, 1 M LiPF6 in DMC/ethylene carbonate (EC) (1:1 v/v) (Selectilyte LP30, BASF) was used. The cells were assembled in an argon-filled glovebox with H2O and O2 contents of less than 1 ppm. Crystalline LiCuSe was first galvanostatically charged up to 3.2 V and then discharged down to 0.7 V vs Li+/Li

with a C/10 rate, corresponding to insertion or extraction of 1 Li during 10 h. 2.5. Magnetic Susceptibility. One pristine LiCuSe sample and one sample that was partly delithiated, with the approximate composition Li0.8CuSe, were placed in polycarbonate capsules inside plastic straws. All preparation steps were made inside an Ar-gas filled glovebox. Both samples were quickly transferred into a magnetometer (MPMS S5, Quantum Design), and their magnetic signals were recorded in the range 2−300 K at a constant field of 1 T. 2.6. Nuclear Magnetic Resonance Spectroscopy. To investigate the mobility of the Li+ ions between the CuCh (Ch = Se or Te) layers, NMR measurements of the Li spectra and the spin−lattice relaxation rates (T1−1) in a temperature range of 100−400 K were performed. Each sample was placed inside silver foil and sealed in an airtight glass tube. All measurements were carried out in a static magnetic field of 7.0453 T (116.579 MHz) using an Apollo Tecmag spectrometer. The Li spectra were obtained by direct Fourier transformation of the free induction decay (FID). For the spin− lattice relaxation rates (T1−1), a saturation recovery pulse sequence (π/ 2 − t − π − τ - echo)19 was used. T1 was obtained by fitting the relaxation curves M(t) with a single stretched exponential function

⎛ ⎛ ⎛ ⎞ λ ⎞⎞ M(t ) t = M 0*⎜1 − f *exp⎜⎜− ⎜ ⎟ ⎟⎟⎟ ⎜ ⎟ M(∞) T ⎝ 1 ⎠ ⎠⎠ ⎝ ⎝

(1)

where f is a fitting parameter and λ describes the stretching of the relaxation curves. Because of the beneficial signal-to-noise ratio of the Li NMR signal, only 10 scans per point with typical π/2 pulse lengths between 3 and 11.5 μs for the Te containing compound and 3−7 μs for the Se containing compound were required, using a repetition time of 4 s for each scan.

3. RESULTS AND DISCUSSION 3.1. Chemical Compositions. According to ICP-OES analyses, the obtained powders had the compositions Li1.0(1)Cu1.04(2)Se1.000(3) and Li0.92(7)Cu1.02(1)Te1.00(1), as normalized to the chalcogen contents. Both compound compositions agree well with the nominal ones, respectively. 3.2. X-ray Diffraction Data. Powder patterns of the title compounds proved relatively pure (Figure 1); all crystallographic data obtained from Rietveld analyses are presented in Table S1 (Supporting Information). The only detectable secondary phases could be identified as 1 vol % Cu2−xSe20 in LiCuSe and 3 vol % Cu1.79Te21 in LiCuTe, respectively. B

DOI: 10.1021/acs.inorgchem.8b00897 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

decomposition (Figure S1, Supporting Information). Obviously, Cu2−xSe forms at the expense of LiCuSe; this can be deduced by comparing absolute diffraction intensities, which are increasing and decreasing, respectively, with time. A tentative decomposition reaction would be 2 LiCuSe + H2O + 1/2 O2 → Cu2−xSe + 1−x Se + 2 Li(OH). However, only the defect anti-fluorite Cu2−xSe20 of the deterioration products could be observed in the diffraction data. 3.3. Electrochemistry and Synchrotron X-ray Diffraction. In a battery set up with Li metal as the anode, the applicability of the cathode material LiCuSe was tested and several intermediate charging stages were investigated by synchrotron X-ray diffraction. The pristine state of the cathode is very similar to the as-synthesized material; only a minor amount of Cu2−xSe has formed during the preparation of the measuring cell (Figure 3). By removing most of the Li

Incoherent scattering of Mo radiation by Se decreases the signal-to-background ratio for LiCuSe (Figure 1), resulting in a higher background. LiCuCh (Ch = Se, Te) have anti-PbClF structures22 and are isostructural to the high-temperature tetragonal LiFeAs,5 which are thoroughly described elsewhere. Here follows only a brief crystal structure description. Cu is tetrahedrally coordinated by Ch (Ch = Se or Te), and together they form 2D layers of edgesharing tetrahedra. The layers are separated by Li that assumes a five-fold, square pyramidal coordination (Figure 2). The resulting structure can be written as Li2∞[CuCh4/4], using the Niggli notations.

Figure 2. Perspective view of a LiCuCh (Ch = Se or Te) crystal structure part, where the coordinations of Li and Cu are emphasized with thick bonds. The Cu layer is marked with thick lines, and the Cu coordination is also shown in polyhedral form. Selected interatomic distances in Å are indicated (Ch = Se above and Te below in bold, italic).

Figure 3. Three synchrotron powder X-ray diffraction patterns of LiCuSe and one pattern (“LiCuSe”) from a lab diffractometer for one complete charging−discharging galvanostatic cycle. The strongest reflections are marked with either (+) for LiCuSe, yellow area for Cu2−xSe, (−) for Li2Se, and (O) for Cu metal.

The scattering factors of Li and Cu are very different, making it possible to conclude that there is no intermixing between the metal sites, although the ionic sizes are relatively similar: Li+ and Cu+ have radii of 0.73 and 0.74 Å for four-fold coordination as well as 0.90 and 0.91 Å for six-fold coordination, respectively.23 Yet, no admixture is observed, which is in contrast to Li1+xCu2−xP27 and in LiCuS24 where Li/Cu intermixing were suggested. However, the obvious cation ordering in both title compounds would emphasize that the different electronic characters matter: Cu+ has a d10, while Li+ has an s0 (or [He]) valence. The ternary sulfide, LiCuS, is known already in the literature,24 but its structure is completely different: LiCuS can be described with a hexagonal close packing of S, in which most of the tetrahedral voids are filled with Li/Cu but columns filled with linearly coordinated Li/Cu are present (structure type anti-Na3AgO225). Noteworthy is the large structural difference, as compared to LiCuSe, occurring despite the fact that the radii of S2− and Se2− differ by less than 10% (r(S2−) = 1.70 Å, r(Se2−) = 1.84 Å).23 Both title compounds are strongly air-sensitive and decompose within a minute outside the inert glovebox environment. By investigating a powder of LiCuSe inside a non-airtight sample holder, it was possible to observe the

electrochemically from the sample (“CuSe” in Figures 3 and 4), an admixture of microcrystalline Cu2−xSe and Cu2Se20 remains, according to X-ray diffraction data (Figure 3). During cell discharging, a gradual Li insertion into the cathode material was observed. However, only parts of the pristine crystal structure were regained and, instead, significant amounts of anti-fluorites Li2Se and Cu2−xSe are observed (Figure 3). On inserting even more Li into the sample, by driving the current until the battery voltage is below 1 V (Figure 4), metallic Cu and Li2Se dominate in the sample (Figure 3). Hence, there is only minor reversibility of the Li content of LiCuSe, during the usage as a cathode material in a Li-ion battery. 3.4. Magnetic Susceptibility of LiCuSe and Li0.82CuSe. By electrochemically removing a minor part of Li from LiCuSe, a delithiated sample was obtained with the approximate composition Li0.82CuSe, as determined from the potentiometric titration. To understand the role of the Cu valence in the slightly Li deficient sample, magnetic measurements were performed on the pristine and delithiated samples (Figure 5). If LiCuSe is purely ionic, only a diamagnetic signal is expected, C

DOI: 10.1021/acs.inorgchem.8b00897 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 6. Relaxation curves with their respective fits (λ describes the stretching of the fit) at multiple temperatures for LiCuSe. LiCuTe shows qualitatively the same relaxation behavior. The data were obtained at 116.479 MHz with a saturation recovery pulse sequence.

Figure 4. Galvanostatic curves of LiCuSe cycled between 3.2 and 0.7 V vs Li+/Li starting with the cell charging. Total Li amount in the sample was calculated from the current flow in the galvanostatic experiment. Four dots mark intermediate states where X-ray diffraction data were extracted (Figure 3) and arrows indicate the process direction.

which leads to the small offset at 400 K compared to the curves at 320 and 200 K. This behavior may be related to the faster relaxation at higher temperatures, and can be accounted for by the fitting parameter f in eq 1. In accordance with ref 26, the relaxation rate T1−1 increases at higher temperatures. The minor stretching of the relaxation curves (λSe/Te < 1) could be attributed to selected 7Li ions starting to move at lower or higher temperatures than the majority due to 7Li defects in the samples (especially in Li0.92(7)Cu1.02(1)Te1.00(1)). This assumption conforms with the broad transition area from low- to hightemperature regime of the FWHM displayed in Figure 7 (main).

Figure 5. Magnetic susceptibility of pristine LiCuSe and a delithiated sample with composition Li0.82CuSe as a function of temperature at a constant field of 1 T.

but pristine LiCuSe exhibits weakly positive, temperature independent magnetic susceptibility (χ), which suggests Pauliparamagnetic or van Vleck contribution that is larger than the diamagnetic signal. Hence, LiCuSe has either polarizable, delocalized electrons and/or quantum dynamically spin excited states. The red-lilac color of LiCuSe could suggest a relatively small activation gap for conductivity, but the Pauli-paramagnetic state is ruled out by spectroscopic observations (see section 3.5). By removing some of the Li ions, the magnetic signal drops and is negative in almost the whole temperature range, corresponding to the more expected diamagnetic behavior (Figure 5). Hence, the Li extraction probably does not increase the oxidation state of Cu above +1; otherwise, a paramagnetic signal would be expected. 3.5. NMR Measurements with Analyses. Figure 6 depicts the relaxation curves fitted using the eq 1 (see section 2.6) for multiple temperatures. Only the curves for LiCuSe are shown because LiCuTe exhibits qualitatively the same relaxation behavior. At higher temperatures (above approximately 340 K), the magnetization cannot be fully saturated,

Figure 7. (inset) Arrhenius plot of spin−lattice relaxation rates T1−1. The solid lines represent the theoretical fits to the data, according to eqs 2 and 3. (main) Temperature dependencies of the line widths for LiCuSe and LiCuTe. The corresponding fits according to eq 4 are drawn as the solid lines.

Figure 7 shows an Arrhenius plot of the spin−lattice relaxation rates T1−1 of LiCuTe and LiCuSe. At lower temperatures, constant background relaxation processes can be observed. As the temperature rises, the spin−lattice relaxation rate T1−1 increases, which signifies the beginning of Li+ ion diffusion. According to Blomberg, Purcell, and Pound’s (BPP) model, a maximum is expected when the condition27ωLτc = 1 is fulfilled, where ωL is the Larmor frequency D

DOI: 10.1021/acs.inorgchem.8b00897 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry and τc the correlation time. As the measurement system as well as the heat stability of the samples is restricted, measurements could only be performed at T ≤ 400 K. Hence, the expected maximum was not reached in the observed temperature range. Yet, the low-temperature activation energy EA,BPP of the diffusion process can be extracted from the curve’s gradient by fitting the relaxation data with the BPP model26 T1−1(T ) ∝

Nevertheless, in both cases, LiCuTe shows a smaller activation energy than LiCuSe. This appears to be a result of the significant Li deficiency of the LiCuTe compared to LiCuSe. Figure 8 shows the line shape of the 7Li spectra at multiple temperatures for both compounds. The red lines depict the

2τc 1 + (ωLτc)2

⎛ EA,BPP ⎞ τc ∝ exp⎜ ⎟ ⎝ kBT ⎠

(2)

(3)

where kB is the Boltzmann’s constant. For the measured data, the activation energies EA,BPP could be obtained as 0.176 eV for LiCuTe and 0.313 eV for LiCuSe (Figure 7, inset). Note that EA,BPP should be regarded as lower limits of the actual values owing to the following three reasons: (i) As the relaxation rate dependence on ωL is unknown, so is the type of relaxation process; by association, the BPP model might be invalid. (ii) The dimensionality of the 7Li migration can only be estimated from the high-temperature regime which means that 2D migration could not be confirmed by the performed T1−1 measurements, even though it could be expected from a layered crystal structure. (iii) The full slope of the lowtemperature flank of the Arrhenius peak might not be reached in the observed temperature range. The Li diffusion activation energy can also be estimated from the temperature dependence of the 7Li spectra line width at half-maximum (FWHM) (Figure 7, main). The data indicate a typical motional line-shape narrowing as described in ref 26 with a constant (LiCuSe) or slightly linearly decreasing (LiCuTe) background. The linearly decreasing component to the line width of LiCuTe might be related to the minor Li deficiency in Li0.92(7)Cu1.02(1)Te1.00(1), which could lead to additional Li motion with a much lower activation energy. Both curves show the predicted drop when the 7Li jump rate becomes comparable to the spectral line width, resulting in a residual linear offset at higher temperatures. To obtain the activation energies, the data can be fitted using the following relation Δν(T ) =

Figure 8. 7Li NMR spectra for the Se (left panel) and Te (right panel) compounds for several temperatures. The red lines describe the simulated curves. The simulation parameters are printed as insets. FWHMc describes the line width of the central peak, FWHMq the line width of the quadrupole satellites, νq the quadrupole splitting, and K the Knight shift.

corresponding powder spectra simulations based on 2nd order perturbation theory assuming a small quadrupole splitting νq in the range of 2 kHz (at 200 K) to 5 kHz (at 400 K) for LiCuSe and 0.5 kHz to 5 kHz for LiCuTe (same temperatures). This splitting is so small that it cannot be resolved clearly in the spectra. At all simulated temperatures (red in Figure 8), the width of the central line (FWHMc) is reduced compared to the satellites (FWHMq), which indicates additional quadrupole broadening instead of pure magnetic broadening. The ratio FWHMc/FWHMq decreases with increasing temperature for both samples. A small, but constant, Knight shift of the Se compound (1 ppm) is observed, which is consistent with the nearly temperature independent susceptibility (Figure 5). However, the Knight shift is so small that it cannot be associated with a Pauli-paramagnetic contribution. Thus, the van Vleck paramagnetic contribution seems to be the more reasonable explanation, which is supported by the temperature independent Knight shift. The observations also coincide with the valences of Cu+ (d10) and Li+ (s0) mentioned in section 3.2. LiCuTe shows a slightly increasing positive Knight shift with rising temperatures from 10 ppm at 200 K to 13 ppm at 400 K. To check whether the NMR data might be influenced by structural phase transitions close to room temperature, a

⎛ ⎛ Δν ⎞ ⎛ E T ⎞⎞ ⎜⎜1 + ⎜ R − 1⎟exp⎜ − A ⎟⎟⎟ + D + a0·T ⎝ B ⎠ ⎝ k B ⎠⎠ ⎝ (4)

where ΔνR describes the constant line width of the rigid lattice regime while B and D represent the temperature independent broadening and the residual line width of the exited ions into account, respectively. Motivated by the linear behavior at low temperatures, the original function is expanded with a linear correction term.28 Using eq 4, activation energies of 0.420 eV for the LiCuTe and 0.649 eV for the LiCuSe compound were estimated, respectively. These values can be compared to ∼0.8 eV for Li3V2(PO4)3,29 which also has been considered as Li battery cathode material. As expected, these values are higher than (about double) the ones obtained from the gradient of the spin−lattice relaxation rates by the BPP model, but the ratio EA,BPP(LiCuTe)/EA,BPP(LiCuSe) ≈ 0.56 is comparable to EA (LiCuTe)/EA (LiCuSe) ≈ 0.65. The reason for this discrepancy could be the mentioned invalidity of the BPP model for the description of our Li migration process. E

DOI: 10.1021/acs.inorgchem.8b00897 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Superconductivity in the PbO-type structure α-FeSe. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 14262−14264. (2) Alekseeva, A. M.; Drozhzhin, O. A.; Dosaev, K. A.; Antipov, E. V.; Zakharov, K. V.; Volkova, O. S.; Chareev, D. A.; Vasiliev, A. N.; Koz, C.; Schwarz, U.; Rosner, H.; Grin, Y. New superconductor LixFe1+δSe (x ≤ 0.07, Tc up to 44 K) by an electrochemical route. Sci. Rep. 2016, 6, 25624. (3) Burrard-Lucas, M.; Free, D. G.; Sedlmaier, S. J.; Wright, J. D.; Cassidy, S. J.; Hara, Y.; Corkett, A. J.; Lancaster, T.; Baker, P. J.; Blundell, S. J.; Clarke, S. J. Enhancement of the superconducting transition temperature of FeSe by intercalation of a molecular spacer layer. Nat. Mater. 2013, 12, 15−19. (4) Guo, J.; Jin, S.; Wang, G.; Wang, S.; Zhu, K.; Zhou, T.; He, M.; Chen, X. Superconductivity in the iron selenide KxFe2Se2 (0 < x < 1.0). Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 82, 180520. (5) Pitcher, M. J.; Parker, D. R.; Adamson, P.; Herkelrath, S. J. C.; Boothroyd, A. T.; Ibberson, R. M.; Brunelli, M.; Clarke, S. J. Structure and superconductivity of LiFeAs. Chem. Commun. 2008, 45, 5918− 5920. (6) Wang, M.; Wang, M.; Miao, H.; Carr, S. V.; Abernathy, D. L.; Stone, M. B.; Wang, X. C.; Xing, L.; Jin, C. Q.; Zhang, X.; Hu, J.; Xiang, T.; Ding, H.; Dai, P. Effect of Li-deficiency impurities on the electron-overdoped LiFeAs superconductor. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 86, 144511. (7) Han, F.; Zhu, X.; Mu, G.; Zeng, B.; Cheng, P.; Shen, B.; Wen, H.H. Absence of superconductivity in LiCu2P2. J. Am. Chem. Soc. 2011, 133, 1751−1753. (8) Grafe, H.-J.; Gräfe, U.; Dioguardi, A. P.; Curro, N. J.; Aswartham, S.; Wurmehl, S.; Büchner, B. Identical spin fluctuations in Cu- and Codoped BaFe2As2 independent of electron doping. Phys. Rev. B: Condens. Matter Mater. Phys. 2014, 90, 094519. (9) Charnukha, A.; Yin, Z. P.; Song, Y.; Cao, C. D.; Dai, P.; Haule, K.; Kotliar, G.; Basov, D. N. Correlation-driven metal-insulator transition in proximity to an iron-based superconductor. Phys. Rev. B: Condens. Matter Mater. Phys. 2017, 96, 195121. (10) Zhao, L. D.; Berardan, D.; Pei, Y. L.; Byl, C.; Pinsard-Gaudart, L.; Dragoe, N. Bi1−xSrxCuSeO oxyselenides as promising thermoelectric materials. Appl. Phys. Lett. 2010, 97, 092118. (11) Ueda, K.; Inoue, S.; Hirose, S.; Kawazoe, H.; Hosono, H. Transparent p-type semiconductor: LaCuOS layered oxysulfide. Appl. Phys. Lett. 2000, 77, 2701. (12) Ueda, K.; Inoue, S.; Hosono, H.; Sarukura, N.; Hirano, M. Room-temperature excitons in wide-gap layered-oxysulfide semiconductor: LaCuOS. Appl. Phys. Lett. 2001, 78, 2333. (13) Mizushima, K.; Jones, P. C.; Wiseman, P. J.; Goodenough, J. B. LixCoO2 (0