Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 4
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Otros (Other)

Magnetic Circular Dichroism at the Oxygen K edge in Microcrystals of Spinels Grown on Ru(0001)

  • Mandziak, Anna
  • Sosa, V.
  • Nita, Pawel
  • Martín-García, Laura
  • Prieto, J. E.
  • Foerster, Michael
  • Niño, Miguel Ángel
  • Aballe, Lucía
  • Granados-Miralles, Cecilia
  • Quesada, Adrián
  • Tejera-Centeno, César
  • Gallego, Silvia
  • de la Figuera, Juan
We have measured the circular magnetic dichroism in the x-ray absorption at the K-edge of oxygen in microcrystals of different spinel oxides. The microcrystals are islands of micrometric size and nanometric thickness, grown on Ru(0001) substrates using high-temperature oxygen-assisted molecular beam epitaxy. The domains observed in the oxygen K-edge dichroism have the same distribution and orientation as those observed in x-ray magnetic circular dichroism at the L edge of the octahedral cations. Integrating the area from a single domain, x-ray magnetic circular dichroic spectra of oxygen were measured and, by the application of the K-edge sum rule, non vanishing orbital magnetic moments aligned with the octahedral cations were found. Density functional theory calculations, which did not show any orbital moment at the oxygen anions, indicate that the energy ranges where oxygen dichroism is observed correspond to those with significant hybridization with the cations d bands. They also show a correlation between the magnitude of the measured value of the oxygen orbital moment and the theoretical one for the cations, and demonstrate that this trend is preserved in the presence of Fe excess in the samples. Our experimental XMCD suggest, following the DFT calculations, that the origin of the oxygen magnetic moment lies in the hybridization of the oxygen unoccupied p-derived bands with the cation bands, mostly with the d-derived ones., This work is supported by the Grants PID2021-124585NB-C31/-C33, PID2021-122980OB-C55, PID2020-117024GB-C43 and TED2021-130957B-C53/-C54 funded by MCIN/AEI/10.13039/501100011033 and by “ERDF A way of making Europe” and by the ”European Union NextGenerationEU/PRTR”. P. N. acknowledges the SciMat Priority Research Area budget under the program Excellence Initiative - Research University at the Jagiellonian University in Krakow - Grant No. 75.1920.2021These experiments were performed at the CIRCE beamline of the ALBA Synchrotron Light Facility. V.S. acknowledges financial support from PIPF-2023/ECO-30425., No
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Magnetic circular dichroism at the oxygen K-edge in microcrystals of spinels grown on Ru(0001)

  • Mandziak, Anna
  • Sosa, Victor
  • Nita, Pawel
  • Martín-García, Laura
  • Prieto, J. E.
  • Foerster, Michael
  • Niño, Miguel Ángel
  • Aballe, Lucia
  • Granados-Miralles, Cecilia
  • Quesada, Adrián
  • Tejera-Centeno, César
  • Gallego, Silvia
  • de la Figuera, Juan
10 pages, 3 figures, 4 tables, We have measured the circular magnetic dichroism in the x-ray absorption at the K-edge of oxygen in microcrystals of different spinel oxides. The microcrystals are islands of micrometric size and nanometric thickness, grown on Ru(0001) substrates using high-temperature oxygen-assisted molecular beam epitaxy. The domains observed in the oxygen K-edge dichroism have the same distribution and orientation as those observed in x-ray magnetic circular dichroism at the L3edge of the octahedral cations. Integrating the area from a single domain, x-ray magnetic circular dichroic spectra of oxygen were measured and, by the application of the K-edge sum rule, non vanishing orbital magnetic moments aligned with the octahedral cations were found. Density functional theory calculations, which did not show any orbital moment at the oxygen anions, indicate that the energy ranges where oxygen dichroism is observed correspond to those with significant hybridization with the cations d bands. They also show a correlation between the magnitude of the measured value of the oxygen orbital moment and the theoretical one for the cations, and demonstrate that this trend is preserved in the presence of Fe excess in the samples. Our experimental XMCD suggest, following the density functional theory calculations, that the origin of the oxygen magnetic moment lies in the hybridization of the oxygen unoccupied p-derived bands with the cation bands, mostly with the d-derived ones and does not correspond to a localized orbital magnetic moment in the oxygen in spinels., This work is supported by the Grants PID2021124585NB-C31/-C33, PID2021-122980OB-C55, PID2020117024GB-C43 and TED2021-130957B-C53/-C54 funded by MCIN/AEI/10.13039/501100011033 and by ‘ERDF A way of making Europe’ and by the ‘European Union Next Generation EU/PRTR’. P N acknowledges the SciMat Priority Research Area budget under the program Excellence Initiative—Research University at the Jagiellonian University in Craków—Grant No. 75.1920.2021. C G-M acknowledges f inancial support from Grant RYC2021–031181-I funded by MCIN/AEI/10.13039/501100011033 and by the ‘European Union NextGenerationEU/PRTR’. These experiments were performed at the CIRCE beamline of the ALBA Synchrotron Light Facility. V S acknowledges financial support from PIPF2023/ECO-30425., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Dynamic Doping of Nickelates with Lithium Reveals a Widely Tunable Insulator–Metal Transition with Charge Filling and Band Renormalization Regimes

  • Zhang, Alan C.
  • Álvarez-Chico, Alejandro
  • Salagre, Elena
  • Gonzalez, Martín
  • Spataru, Catalin D.
  • Sugar, Joshua D.
  • Gross, Adam L.
  • González-Barrio, M. A.
  • Segovia, P.
  • Tallarida, Massimo
  • Dai, Ji
  • Kumar, Suhas
  • Talin, A. Alec
  • Hwang, Harold Y.
  • Michel, Enrique G.
  • Mascaraque, Arantzazu
  • Fuller, Elliot J.
The insertion of electron-donating ions has emerged as a powerful technique to manipulate the electronic structure of correlated oxides. However, the resulting electronic structure remains poorly understood, with challenges in quantifying dopant concentration, unexplained differences with substitutionally doped films, and a poor understanding of how dopant atoms interact with insulator–metal transitions (IMTs). Here, these issues are addressed in the context of the rare earth nickelates, a prototypical correlated oxide family with widely tunable electronic behavior under the insertion of protons and alkali metals as interstitial dopants. RNiO3 (R = Pr, Nd) epitaxial thin films are synthesized, lithium dopants are introduced and quantified using electrochemical and synchrotron-based techniques, and the resulting electronic structure is studied. From electronic transport measurements of LixLiRNiO3, lithium is found to affect the metal–insulator transition, causing more than an order of magnitude reduction in ground-state resistivity at fractions xLi < 0.18, a systematic lowering of transition temperature, and successively smaller ON/OFF ratios over 0.00 < xLi < 0.25. At larger fractions xLi > 0.25, the transition is destroyed, and insulating behavior is observed over T = 5–300 K. Angle-resolved photoemission (ARPES) confirms transport results and reveals band renormalization occurring over 0.10 < xLi ≤ 0.71. ARPES and X-ray absorption spectroscopy (XAS) combined with density functional theory indicate that rigid band filling models are generally insufficient to explain doping from lithium, especially at low temperatures, but could approximate room temperature effects in the low doping regime (xLi < 0.10). Broadly, the results indicate that interstitial dopants lead to complex interactions with metal–insulator transitions and the emergence of an exciting family of correlated electronic phases., Introduction The insertion of ions or molecules as dopants into correlated electron systems has a rich history, especially in the case of 2D materials, where electron-donating ions provide a far greater effective doping concentration relative to gating by electric fields or electric double layers. For example, as early as 1970, it was discovered that the superconducting transition in TaS2 could be tuned through an intercalated layer of pyridine (1) or alkali metals. (2) More recently, a rich field of new correlated phases has emerged from 2D van der Waals materials. (2−5) This includes the extraordinary ability to tune between Bardeen–Cooper–Schrieffer superfluidity and a Bose–Einstein-like superconducting condensate in the 2D superconducting system ZrNCl through the intercalation of lithium. (6) On the other hand, correlated oxide systems have also been explored as ion-insertion hosts, with early work from 1950 to 1980 on tungsten bronzes (7−9) and later work on the rare earth nickelates, (10−14) cuprates, (15) and vanadium oxides. (16) In such systems, a similarly rich landscape of new correlated phases now seems possible. However, oxides lack the weak van der Waals layers between which to insert ions, and doping at interstitial sites has generally been more challenging. For example, the metastable nature of the resulting material coupled with sluggish diffusion kinetics has confined many works to thin-film approaches (as opposed to high-temperature bulk synthesis) in which the species can be introduced across short diffusion lengths but are then limited to thin-film experimental probes. In this case, bulk techniques such as thermogravimetric analysis to quantify dopants or neutron scattering to provide precise structural information are not typically possible. Therefore, new techniques and methodologies by which to synthesize and characterize oxide systems under interstitial doping should enable new discoveries in this fascinating class of emerging materials. In this work, these issues are addressed to reveal the electronic structure at quantified doping concentrations for the rare earth nickelate systems PrNiO3 and NdNiO3. The nickelates are negative charge-transfer insulators, with the chemical formula RNiO3, where R is a trivalent rare-earth (R = La, Nd, Pr, ... etc.), and the interplay between charge and lattice degrees of freedom leads to a bond-disproportionated insulating state at low temperature with a transition to a metal state at elevated temperature. (17) Recently, the insertion of protons or alkali metals has yielded interesting results, such as colossal changes in the room temperature resistivity by up to a factor of 108 and applications in computing devices. (11,18) Nevertheless, the electronic structure of these systems continues to elicit controversy. First, there are substantial differences with substitutionally doped films, which have not clearly been explained. (17,19,20) Second, the number of dopants is often not reported or unknown, making phase diagrams difficult to construct. And third, proposed models can be simplistic (i.e., rigid band filling) and consider only the room temperature behavior without interactions with the transition. (14) On one hand, if there are indeed regimes where rigid band filling is a good approximation, interstitial doping could become a powerful technique to probe the intrinsic electronic structure of nickelates and other oxides. On the other hand, any new electronic phases emerging from interstitial dopants with significant changes to the electronic structure (or interactions with transitions) may be interesting phases in their own right. In this work, quantitative electrochemical cells are constructed to determine dopant concentrations, and cryogenic transport measurements are used to construct phase diagrams across the full temperature range to include the bond-disproportionation transition. To more clearly reveal the resulting electronic structure evolution, including charge filling and band renormalization regimes, synchrotron-compatible doping techniques in situ with angle-resolved photoemission (ARPES) are carried out. In each case, the methodology presented allows the study of the same sample under a continuous doping evolution, within a simple experimental setup, and without the issue of sample-to-sample variability. Results and Discussion PrNiO3 is synthesized as epitaxial thin films, and electrochemical cells are constructed in order to dope them with lithium, quantify the number of lithium dopants introduced, and characterize the resulting effects via electronic transport measurements. Briefly, films with thicknesses ranging from 5 to 100 nm are synthesized using pulsed laser deposition (PLD) on (LaAlO3)0.3(Sr2TaAlO6)0.7 (LSAT) single-crystal substrates. As depicted in the schematic in Figure 1b, the electrochemical cells are constructed by patterning the films into well-defined channels of various sizes (typically 3 mm × 100 μm) using optical lithography and etching unused regions with nitric acid. Electrode contacts (4-wire configuration) are patterned on the film with 50 nm of gold, and LiFePO4 is drop-cast onto a gold current collector to act as a counter electrode (source of lithium). A polydimethylsiloxane (PDMS) well is used to contain electrolyte consisting of either 1 M LiClO4 in propylene carbonate or ionic liquid 1 M LiFSI in PYR-14 TFSI. All electrochemical cycling is either done in vacuum or in an inert argon environment with <5 PPM H2O and O2, and the films are characterized in a vacuum cryostat. More details of the synthesis, fabrication, and electrochemical processes can be found in the Methods/Experimental Section. Figure 1 Figure 1. Electrochemical cycling and transport measurements of LixLiPrNiO3. (a) A schematic of the bond disproportionation transition in PrNiO3 (top panel) and the same transition with lithium interstitial defects (bottom panel). The transition is viewed in the a–b plane, with oxygen octahedra (gray) surrounding Ni sites (dark blue) and lithium ions inserted through a charge transfer reaction (green). (b) A schematic of electrochemical cycling with a thin-film LixLiPrNiO3 working electrode (brown), a LiClO4 electrolyte (blue), a LiFePO4 counter electrode (brown), gold current collectors and electrodes (yellow), and a single-crystal LSAT substrate (tan). (c) Consecutive charge–discharge curves of the working electrode potential EWE (LixLiPrNiO3) vs LiFePO4 (top panel) and corresponding change in the LixLiPrNiO3 resistance ratio (bottom panel). The areas shaded light blue (light red) correspond to lithium insertion into (out of) LixLiPrNiO3. (d) Resistivity plotted against temperature for doping fractions xLi = 0.0–0.27. (e) Resistance ratio at T = 5 K for doping fractions xLi = 0.00–0.27. (f) Activation energy Ea extracted from transport curves at temperatures below the transition. (g) The transition temperature Tc as a function of lithium doping. The films are electrochemically cycled at constant current (chronopotentiometry), and the resistance is monitored simultaneously using a Biologic SP-300 bipotentiostat. A LiFePO4 counter electrode is used to drive Li+ ions through the electrolyte into the LixLiPrNiO3 working electrode. LiFePO4 is chosen because it is a stable, phase-coexistence source of lithium with a flat electrochemical voltage profile, and it more easily provides a quantitative electrochemical probe at lower voltages than pure lithium. (21) At low voltage, side reactions can be reduced, enabling the quantification of electrochemical charge transfer. Figure 1c shows an example of a cycling curve for a 90 nm thick film. At a charging rate of C/8, where C is the amount of charge needed to reach xLi = 1.00 in 1 h, an increase in the working electrode potential EWE to approximately 1 V is observed in the first few minutes (Figure 1c, top panel, blue shaded regions). Next, a shallow but steady increase in EWE to roughly 1.4 V is observed and accompanied by a monotonic but nonlinear increase in the resistance by a factor of roughly 4.5 after 1 h of charging (Figure 1c, bottom panel, blue shaded regions). The large swing in EWE within the first few minutes of charging is ascribed to overpotential, which is related to voltage buildup at the electrolyte–electrode interface due to charge-transfer resistance. The subsequent gradual increase in EWE (to 1.4 V) is ascribed to the lithium insertion process, in which a steady increasing potential suggests a solid solution of LixLiPrNiO3. A subsequent discharge at C/8 returns the resistance to the initial value, although over a shorter duration and a smaller total charge. The discharge process allows for a calculation of the Coulombic efficiency of the cell. Here, the reversible amount of charge is ascribed to the lithium insertion process and the irreversible charge to side reactions. The Coulombic efficiency is then used to calculate the lithium fraction xLi in LixLiPrNiO3 during further cycling and transport measurements. The insertion of lithium into the bulk of RNiO3 through Faradaic electrochemical reactions has been reported elsewhere (13,14,22) and is distinct from ionic liquid gating, which primarily acts as a field effect based upon electric double layers. Cryogenic electronic transport curves are measured as a function of xLi for a 33 nm thick film and are plotted in Figure 1d. In the pristine film (red), intrinsic transport behavior is observed with metallic conduction down to 122 K, followed by a sharp insulator–metal transition with a roughly 1 × 105 ON/OFF ratio and divergent resistivity below 20 K. This is consistent with previous reports of high-quality nickelate films. Next, at doping xLi = 0.16 (dark orange), a sharp drop in the ground-state resistance by more than 1 order of magnitude is observed, accompanied by an increase in the room temperature resistance by a factor of 4.2. The transition temperature, which is defined as the inflection point where the slope switches from negative to positive, is decreased by about 12 K (Figure 1g). This behavior is qualitatively similar to substitutional electron doping at the rare earth site in Nd1–xCeCexCeNiO3, (20) where small fractions xCe of Ce have the same effect, albeit over a much different range in dopant concentration (xCe < 0.01). However, at larger concentrations, lithium doping qualitatively diverges from substitutional doping. First, the trend in the low-temperature resistance reverses (Figure 1e), and a monotonic increase is observed for all doping fractions, with successive suppression of the transition and a smaller ON/OFF resistance ratio. Second, the activated transport gap Ea, which is extracted from the slope in the insulating regime and plotted in Figure 1f, lowers to a minimum but then increases for xLi > 0.24. In contrast, substitutional doping has the effect of completely destroying the bond disproportionation transition and inducing metallic behavior at all fractions xCe > 0.05. Finally, it is observed that between xLi = 0.24 and 0.26, the bond disproportionation is destroyed (Figure 1d, purple and brown curves), giving rise to fully insulating behavior. It is noted that previous reports on proton doping have shown qualitatively similar behavior (e.g., with the xLi = 0.16 transition), although with unknown dopant fractions, (12) suggesting proton-doped films may have similar behavior as lithium-doped films. It is useful to construct and analyze phase diagrams from the transport data of LixLiPrNiO3. These diagrams can provide identification of new correlated phases emerging from the complex interaction of lithium with the lattice. In Figure 2a, we plot the derivative d(ln ρ)/dT (color scale) as a function of doping (vertical axis) and temperature (horizontal axis) to identify the critical regions. We define insulating and metallic regions by negative and positive values of d(ln ρ)/dT, respectively. First, we note the large blue region from about 130 K to 300 K and from xLi = 0.0 to 0.24, corresponding to metallic behavior. Next, the red and yellow regions indicate sharper points in the transition at low doping fractions. Finally, a crossover to fully insulating behavior xLi > 0.24 is noted, where the transition is completely suppressed. While the primary focus in this report is on films which exhibit intrinsic transport properties, often the materials reported in the literature are polycrystalline and defective and exhibit much smaller ON/OFF ratios. This is especially problematic for perovskite nickelates due to the thermodynamic preference for Ni2+ valence states and resulting secondary phases such as Ruddlesden–Popper (RP)-type faults. (23) Therefore, an additional phase diagram is presented in Figure 2b, where the Ni/R ratio had not been optimized in synthesis (Ni/R < 1.0). Here, the xLi = 0.00 ON/OFF ratio is only about 1 order of magnitude (as often reported for polycrystalline films), and a much smaller range in slope is observed over the entire phase diagram. Interestingly, the transition can still be tuned between xLi = 0.00 and 0.08, albeit with much smaller fractions of lithium required to disrupt the metal–insulator transition. In striking contrast to the intrinsic case, the film becomes metallic at all temperatures for fractions 0.08 < xLi < 0.16, although with higher total resistance. Therefore, it is concluded that RP stacking faults have the effect of stabilizing a metallic phase under lithium doping. Figure 2 Figure 2. Phase diagrams of LixLiPrNiO3. (a) Phase diagram of a LixLiPrNiO3 film is generated from the slope d(ln ρ)/dT from the transport curves in Figure 1d. The white-to-blue portion of the color scale represents a positive slope, whereas green-to-red represents a negative slope. (b) Phase diagram of LixLiPrNiO3 with Ruddlesden–Popper stacking faults. The color scale delineates negative and positive slopes as with (a), but the overall slope magnitudes are significantly shallower. ARPES is carried out in situ with lithium doping to measure the resulting band structure. While electronic transport measurements combined with electrochemical cycling can provide quantitative information about metallic and insulating phases at specific doping fractions, ARPES offers direct insight into the evolution of the electronic structure. Briefly, samples are prepared by oxygen plasma etching in an ultrahigh vacuum environment. X-ray photoemission spectroscopy (XPS) is performed to verify the lack of carbon contamination at the surface and to determine lithium fractions. The quality of the surface after cleaning is checked with low-energy electron diffraction (see Supporting Information Figure S16). Lithium doping at various fractions is studied across a single sample by a succession of depositions with a commercial lithium getter from SAES. Deposition steps range between 1 and 90 min, in which a pressure better than 5 × 10–10 mbar is maintained. The lithium content for each doping fraction is then calculated based upon a quartz balance calibration and the measurement of the Li 1s, O 2s, Nd 5p, and Nd 5s core-level intensity. A detailed discussion on the lithium doping process can be found in Supporting Information Figures S8–S11. Measurements of both PrNiO3 and NdNiO3 were performed with qualitatively similar results. The NdNiO3 yielded a higher quality surface with clearer bands, and therefore, we primarily report on NdNiO3 given the similarity in electronic structure and transition temperatures to PrNiO3 (see Supporting Information Figure S12). First, qualitative changes to the valence band from ARPES are described for a single 5.5 nm thick NdNiO3 film after successive lithium evaporation steps. Figures 3a and 3b show the normal emission, long-range valence band as a function of lithium content for hv = 80 eV (Γ point) and for hv = 110 eV (Z point), respectively. Peak B (Fermi edge) originates from Ni 3d eg orbitals, while peak B′ at −1 eV is associated with Ni 3d t2g orbitals. Peak B″, located at a −2.3 eV binding energy, corresponds to a hybridized state involving Ni t2g orbitals and O p-orbitals. In the right insets, the pristine state (red) shows a metallic Fermi edge. With increasing lithium content from xLi = 0.10 to 0.24 (orange to light blue), the Fermi edge decreases, indicating more insulating behavior with spectral weight shifted lower in BE. A larger difference compared with pristine is observed at xLi = 0.71 (dark blue), with a total collapse of the signal near the Fermi level, complete insulating character following this collapse, and significant renormalization of all Ni 3d states. While spectra with lithium content xLi = 0.10–0.24 do not have such a drastic change as xLi = 0.71, closer inspection of the valence band shape and spectral distribution indicates significant renormalization effects. First, deeper levels (observed in the left panels in Figure 3a,b) exhibit a systematic increase in spectral weight at the expense of states closer to the Fermi level. For example, spectral features labeled A and A′ grow in intensity, whereas spectral features labeled B and B′ are found to decrease. These changes are photon-energy-dependent and are observed for both 80 and 110 eV, which indicates that such features are not simply related to surface effects or changes to sample quality. Furthermore, the intensity distribution of the shallow peaks B, B′, and B″ is modified by lithiation as the intensity at the Fermi energy (B) collapses. Figure 3 Figure 3. LixLiNdNiO3 valence band spectra as a function of lithium content and temperature. Panels (a) and (b) are the integrated long-range valence band for different lithium contents (pristine, xLi = 0.1, 0.19, 0.24, and 0.71) taken around 200 K. The photon energy for (a) and (b) is 80 and 110 eV, respectively. Insets on the right for each photon energy are taken near the Fermi level. Panels (c) to (g) are the evolution at each lithium content measured during cooling at temperatures between 20 and 240 K. The temperature scale is indicated by color and vertical displacement of the lines according to the legend (right). The state of the sample in each case is identified by the color bars at the right of each graph, with metallic (red, M), insulating (blue, I) and transition regions (yellow, T). To investigate how the bands change during the metal–insulator transition, the valence band was measured as a function of the temperature. Figure 3c–g shows the valence band spectra for xLi = 0.00–0.71. The spectra are obtained by integrating the acceptance angle (ky = ± 1.15 Å–1) at fixed kx in order to capture changes to both electron and hole pockets. The temperature range of metallic and insulating states and transitions is obtained from the intensity at BE = 0.0 and 0.2 eV, as shown in Supporting Information Figure S14. For the pristine state in Figure 3c, the flattening of the Fermi edge observed between 140 and 160 K is labeled T1, and this is ascribed to the metal–insulator transition in NdNiO3. An additional transition, labeled T2, is found between 40 and 60 K and marked by differences in slope, which is ascribed to the antiferromagnetic insulating transition. The two transitions occur simultaneously in bulk NdNiO3 but appear at different temperatures in the thin film, likely due to epitaxial strain. (17) Next, at xLi = 0.06 (Figure 3d), the intensity at the Fermi edge decreases in the metallic phase, and the metal–insulator transition is lowered to between 110 and 130 K. Interestingly, the T2 transition temperature is raised to between 50 and 70 K compared with pristine. For xLi = 0.10 and 0.19, the metal–insulator transition T1 is lowered further in temperature to between 100 and 120 K with no additional transition T2 observed, potentially indicating that these transitions are now simultaneous. It is noted that the qualitative trend of lowering the T1 transition agrees well with transport data for electrochemically doped PrNiO3 films (Figure 1d), but with systematically lower temperatures for all doping fractions, consistent with differences between PrNiO3 and NdNiO3. A more detailed analysis of the band structure is carried out with constant energy surfaces. The Fermi surface of pristine NdNiO3 is plotted in Figure 4a at around 200 K (metallic state) and presents an electron pocket at the Γ points and hole pockets centered at A points. (24) Figure 4a–f shows cuts through the Fermi surface (intensity distribution at the Fermi energy) near the bulk Γ point (hv = 80 eV) for increasing lithium content (the results for hv = 110 eV can be found in Supporting Information Figure S13). All surfaces have been measured at a temperature between 200 and 250 K except for panel (f) with xLi = 0.71, which was insulating for all temperatures, and a completely unoccupied surface at 12 K is shown as an example. The dashed white lines correspond to the Brillouin zone edges, and the blue and red lines indicate the electron and hole sheets, respectively. In the series, up to xLi = 0.19, the central electron pocket and hole pockets are conserved. There is also a decrease in the feature contrast, as expected from the decrease in slope at the Fermi level seen in the integrated VB spectra in Figure 3. Figure 4a–e shows partial degradation of the quality in the Fermi surface with increasing lithium content, which is partly attributed to the small probing depth of ARPES at the photon energies used and the possibility of some surface lithium terminating the NdNiO3 film. Figure 4 Figure 4. Fermi surfaces and cuts for LixLiNdNiO3. Panels (a) to (f) show constant energy cuts at the Fermi level for each lithium fraction indicated by a label and color-coded dot in the lower right corner of each graph. In panel (a), red (holes) and blue (electrons) lines are intended to be a help to the eye in identifying the bands. Panel (g) corresponds to MDC cuts at the green and yellow dotted lines of panel (a) in order to quantify the size of the surfaces for each different lithium fraction. The results and therefore evolution of the Fermi surface size are plotted in panel (h) for 80 eV photon energy and 110 eV photon energy for both hole (red) and electron bands (blue). To quantify the effects of lithium near the Fermi level, an analysis of momentum distribution curves (MDCs) is carried out. The MDCs are plotted in Figure 4g and correspond to cuts at the horizontal yellow and green dashed lines in Figure 4a, which are the electron and hole pockets, respectively. In Figure 4g, the distance between the two peaks reflects the size of the electron and hole pockets, which are found to increase with the lithium content. The peak distance is plotted in Figure 4h, where both a small increase in the distance between the adjacent hole pockets and an increase in the electron pocket diameter are quantified. Both trends indicate an increase in electron content, consistent with lithium donating an electron to the lattice. The band filling due to lithium insertion is obtained from absolute electron doping, which is calculated from a Luttinger analysis of the volume of the Fermi surface. For a small range, the effect of electron doping could be approximated by band filling. The first lithium doping step, xLi = 0.01, has a corresponding band filling increase of 0.013, and xLi = 0.06 has an increase of 0.071. The evolution of electron filling was also tracked via X-ray absorption spectroscopy at the O K-edge (see Figure S15). A clear decrease in the prepeak intensity with increasing lithium content indicates a progressive weakening of Ni–O hybridization. This prepeak, associated with transitions from O 1s core levels to hybridized Ni 3d–O 2p states, reflects the degree of orbital mixing between nickel and oxygen. Consequently, the reduction in prepeak intensity implies a lower degree of hybridization and a corresponding decrease in the Ni valence. This behavior is consistent with electron doping driven by lithium incorporation, which donates electrons to the system and stabilizes nickel in a more reduced oxidation state and is reflected also in the intensity change of peak B″. (25) At more incorporation steps, the lithium content and band filling values start to diverge further. For xLi = 0.10, the corresponding filling increase is 0.17, which indicates a strong deviation from pure electron filling behavior. This conclusion is supported by the analysis of binding energy vs momentum plots (see Supporting Information, Figure S17), which show that lithium induces a change in the band curvature (i.e., a modification of the effective mass). Ground-state density functional theory (DFT) calculations were carried out to compare with transport measurements in the low-temperature limit. The electronic properties of LixPrNiO3 are calculated using density functional theory (DFT) within the GGA + U approach with U = 2.0 eV for Ni d-states. The calculations are carried out using the VASP code (26) in conjunction with PAW pseudopotentials. (27) Complex S-type antiferromagnetic ordering for the higher-spin Ni ions (28) was assumed, and supercells containing 32 PrNiO3 formula units were considered. Figure 5a (top) shows a schematic of the transition showing the long and short disproportionated NiO6 labeled as NiL and NiS, respectively. The Ni–O bond lengths are 1.89 Å for NiS and 1.95 Å for NiL. The Ni ions are low-spin (0 magnetic moment) for NiS and higher-spin (magnetic moment = 1.15 μB) for NiL, indicative of valence states of +4 and +2, respectively. Lithium atoms were inserted randomly into low-energy intercalation sites. Figure 5a (bottom) shows a lithium intercalation site, with the shortest Li–Ni bond (length = 2.15 Å) taking place between Li and NiS. The initial calculation of the pristine ground state (Figure 5b) shows a bandgap of 460 meV that agrees well with the literature (21) and with the value from transport data of 445 meV in Figure 1f. From DFT, it is found that small fractions of doping xLi = 0.03 are enough to disrupt bond disproportionation (Figure 5c) with a strong reduction in the Eg (Figure 5g). At subsequently larger doping fractions xLi > 0.06, the bandgap opens up again, returning to more insulating behavior, presumably due to Coulomb repulsion due to electron filling, as was proposed previously. (13) We note that the overall trend is precisely what is observed in transport in Figure 1f, with a reduction in the gap, followed by an increase. It is noted, though, that the fractions required to achieve this trend are shifted lower from the experiment, with an increase in gap occurring at smaller lithium fractions. However, the overall trend is captured qualitatively. This might partially be related to the Li intercalation contrasted with effective filling observed in the photoemission measurements. We find that with increasing lithium content, the oxygen octahedra are distorted and bond disproportionation is gradually reduced, while the magnetic moments gradually increase/decrease on the NiS/NiL ions closest to lithium (with magnetic moments of about 0.8 μB for xLi = 0.25), suggesting a gradual change in valency from +4/+2 to +3. The overall decrease in bond disproportionation is also reflected in the reduced difference between NiS and NiL PDOS (Figure 5b–f). We note that at large lithium contents, the insulating behavior captured by DFT agrees with transport and photoemission measurements. This suggests that, even though the DFT calculations target the low-temperature limit, the calculated electronic structure for the highest considered lithium content (e.g. xLi = 0.25) may be partially representative (given the reduced bond disproportionation) for the high-temperature case as well and, in particular, for high enough lithium content where the phase transition is suppressed. Figure 5 Figure 5. Density functional theory calculations of LixLiPrNiO3. (a) Diagram showing long and short disproportionated NiO6 labeled as NiL and NiS, respectively (top). The same diagram is shown for a lithium interstitial defect (bottom). The density of states for xLi = 0.00 (b), xLi = 0.03 (c), xLi = 0.06 (d), xLi = 0.13 (e), and x = 0.25 (f) with contributions from NiL and NiS plotted in blue and red, respectively. (g) The calculated band gap Eg as a function of the fraction xLi in LixLiPrNiO3. Conclusions The effect of interstitial lithium for PrNiO3 and NiNiO3 significantly deviates from both substitutional dopant counterparts (such as Th or Ce (19)), as well as from a pure electron filling model (i.e., rigid band filling models (14)), which does not take into account complex interstitial defect interactions. From transport measurements, ARPES, and DFT, an intriguing picture of interstitial lithium doping emerges, where a strong interruption of the bond disproportionation transition occurs at low temperatures even with small doping fractions x < 0.16, and where more than an order of magnitude reduction in ground-state resistance occurs. At room temperature and at low doping fractions x < 0.10, a rigid band filling model may be a reasonable approximation. However, as doping exceeds x > 0.1, clear renormalization effects are apparent from the resulting changes in spectral weight near the Fermi level and changes to band curvature. Such renormalization does not lead to complete metallization as with substitutional dopants, likely due to the lack of strong steric effects on the Ni–O bond angle as reported by Munoz et al., (19) although structural distortions from the interstitial site occupancy clearly play a role. Next, a discussion of the results in the context of other oxides and interstitial atoms is presented. First, smaller-radius interstitial proton defects or larger-radius alkali metal atoms may exhibit significant similarities. For example, despite their small ionic radius, protons were found to increase the cell volume by as much as 13% at doping fractions approaching x = 1.0, (29) indicating significant structural distortion similar to lithium, and therefore, renormalization effects are likely to play a role. As the ionic radius grows larger with different alkali metals, these distortion effects may increase relative to lithium, potentially leading to phase segregation effects or large structural transformations as observed with intercalated 2D materials. (4) Finally, the methods described in this study can be extended to other oxide systems more generally, either to probe their intrinsic electronic structure (at low fractions) or to create new correlated phases of matter, where interstitial defects interact with correlated transitions. As this study demonstrates, with careful consideration of electrochemistry, interstitial defect densities can be quantified and controlled via quantitative means to precisely map phase diagrams. Further, the in situ doping methods presented allow the band structure to be revealed and at quantified dopant concentrations. Therefore, we anticipate interstitial doping to be of broad interest: first, it is less explored than substitutional doping and may provide a more intrinsic probe of electronic structure at small concentrations and with smaller ionic radii (such as Li, H). Finally, we anticipate the discovery of an exciting new family of correlated phases to emerge more generally in other oxide systems and across a variety of dopant atoms (i.e., H, K, Na, etc.). Methods/Experimental Section Pulsed Laser Deposition PrNiO3 thin films were grown on LSAT (001) substrates by using a Neocera PLD system. The LSAT substrate was chosen because of the small in-plane lattice mismatch; it imposes 1.2% tensile strain on PrNiO3. The stoichiometric and Ni-enriched polycrystalline targets were ablated using a 266 nm Nd:YAG laser operating at 5 Hz. The laser fluence was 3.4 J/cm2 with a spot size of 1.7 mm2. The low-energy outer edge of the laser beam was truncated by imaging through a circular aperture to achieve a more uniform laser spot. The stoichiometric target was purchased from Toshima Manufacturing Co., and the Ni-enriched target was made by sintering a pelletized Pr2O3 and NiO powder mixture with a composition of PrNi1.2O3; the target was sintered in air at 1200 °C for 12 h 4 times with intermediate grinding in between. A combinatorial method was used for synthesis by varying the duty cycling between the two targets with each full cycle totaling 40 shots. The growth substrate temperature was 570 °C in 200 mTorr partial pressure of O2. The films were cooled at 5 °C/min in a 2.5 Torr O2 atmosphere. NdNiO3 thin films were grown on a LSAT (001) by PLD with a 248 nm KrF excimer laser and a polycrystalline NdNiO3 target. The LSAT substrate was prepared ex situ by standard acetone–isopropyl alcohol ultrasonication. Laser fluence was set to 2.6 J/cm2, with a laser spot of 0.77 mm2. Substrate temperature during growth was 580 °C, and partial O2 pressure was set to 150 mTorr. A 30 min postanneal in a partial oxygen atmosphere of 5 × 10–6 Torr at 900 °C was done after growth. Electrical Transport Measurements Electrical transport measurements were performed in the linear Kelvin probe configuration with sputtered Cr/Au electrodes in a Lakeshore ST-500. A Keithley 6221 current source provided a constant DC current, and a Keithley 2182A nanovoltmeter measured the voltage difference across the inner electrodes. Data were recorded using the NHMFL Data Acquisition System v.2.4.1. Electrochemical Cycling The PrNiO3 films were patterned into 3.2 mm long and 0.1 mm wide bars by nitric acid fume etching. The gold counter electrode and contact (working) electrodes were deposited on top of the substrate and PrNiO3. A lithium-ion source was deposited on the counter electrode with electrolyte in a PDMS gasket covering the device and lithium source, as seen in Figure 1b. The electrochemistry was performed at room temperature in an Ar glovebox or in a vacuum chamber using a Biologic SP-300. The first channel performed chronopotentiometry between the counter and working electrodes, and the second channel performed chronoamperometry between the film contact electrodes; this monitored the change in PrNiO3 resistance of lithium-ion intercalation/deintercalation (Figure 1c). After reaching a prescribed resistance change, an electrical transport measurement was taken. The process was repeated on the same device for different intercalation levels. Angle-Resolved Photoemission Spectroscopy ARPES measurements were performed at the LOREA beamline of ALBA synchrotron using an MBS A-1 hemispherical analyzer with high energy resolution (better than 1 meV) and the fast Fermi surface mapping mode. The samples were cleaned in situ with an oxygen plasma source. The cleanliness and structural quality of the samples were confirmed by in situ LEED (Supporting Information Figure S16 and Results and Discussion), valence band UV photoelectron spectroscopy, and XPS of C 1s and O 1s peaks. The angular resolution and total energy resolution were 0.1° and 20 meV, respectively. The experiments were performed in the temperature range 10–200 K and photon energies between 80 and 160 eV. Supporting Information The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsnano.5c06879. Additional characterization and analysis details; structural characterization using XRD, HRTEM, electrical resistivity, and Raman spectroscopy; details about lithium quantification via electrochemical and spectroscopic techniques; additional spectroscopic analysis including PNO data and further analysis of NNO data presented in the main text; LEED pattern of a prepared NNO sample surface; additional ARPES data and band fitting for NNO at different Li doping levels; and atomic force microscopy image of pristine and lithiated samples (PDF) nn5c06879_si_001.pdf (1.57 MB) Dynamic Doping of Nickelates with Lithium Reveals a Widely Tunable Insulator–Metal Transition with Charge Filling and Band Renormalization Regimes 18 views 0 shares 0 downloads Skip to figshare navigation 1 Supplementary Information: Dynamic doping of nickelates with lithium reveals a widely tunable insulator-metal transition with charge filling and band renormalization regimes Alan C. Zhang1†, Alejandro Álvarez-Chico2,3†, Elena Salagre1, Martin Gonzalez4,5, Catalin D. Spataru1, Joshua D. Sugar1, Adam L. Gross1, Miguel A. González-Barrio2, Pilar Segovia7,8, Massimo Tallarida9, Ji Dai9, Suhas Kumar1, A. Alec Talin1, Harold Y. Hwang5,6, Enrique G. Michel7,8*, Arantzazu Mascaraque2*, and Elliot J. Fuller1* 1Sandia National Laboratories, 7011 East Ave. Livermore, CA 94550, United States of America 2Dto. Física de Materiales, Univ. Complutense de Madrid, 28040 Madrid, Spain 3Instituto de Ciencia de Materiales de Madrid, Consejo Superior de Investigaciones Científicas (ICMM-CSIC), 28049 Madrid, Spain 4Department of Materials Science and Engineering, Stanford University, Stanford, California 94305, USA 5Stanford Institute for Materials and Energy Sciences, SLAC National Accelerator Laboratory, Menlo Park, CA 94025, USA 6Department of Applied Physics, Stanford University, Stanford, California 94305, USA 7Dto. Física Materia Condensada, Univ. Autónoma de Madrid, 28049 Madrid, Spain 8IFIMAC (Condensed Matter Physics Center), Univ. Autónoma de Madrid, 28049 Madrid, Spain 9ALBA Synchrotron Light Source, Cerdanyola del Vallès, Barcelona, Spain †contributed equally *corresponding author 2 PrNiO3 film X-ray diffraction XRD of a 33nm thick PrNiO3 grown on LSAT substrate is shown in Figure S1, it shows clear pseudocubic perovskite (00l) peaks of the PrNiO3 film that is epitaxially strained to the substrate. No other reflections were present. The Laue fringes indicate a high-quality single crystal film. Expand Terms & Conditions Most electronic Supporting Information files are available without a subscription to ACS Web Editions. Such files may be downloaded by article for research use (if there is a public use license linked to the relevant article, that license may permit other uses). Permission may be obtained from ACS for other uses through requests via the RightsLink permission system: http://pubs.acs.org/page/copyright/permissions.html. Author Information Corresponding Authors Enrique G. Michel - Dto. Física Materia Condensada, Univ. Autónoma de Madrid, Madrid 28049, Spain; IFIMAC (Condensed Matter Physics Center), Univ. Autónoma de Madrid, Madrid 28049, Spain; Orcidhttps://orcid.org/0000-0003-4207-7658; Email: enrique.garcia.michel@uam.es Arantzazu Mascaraque - Dto. Física de Materiales, Univ. Complutense de Madrid, Madrid 28040, Spain; Orcidhttps://orcid.org/0000-0002-2614-2862; Email: arantzazu.mascaraque@fis.ucm.es Elliot J. Fuller - Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, United States of America; Orcidhttps://orcid.org/0000-0001-7912-8688; Email: ejfull@sandia.gov Authors Alan C. Zhang - Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, United States of America Alejandro Álvarez-Chico - Dto. Física de Materiales, Univ. Complutense de Madrid, Madrid 28040, Spain; Instituto de Ciencia de Materiales de Madrid, Consejo Superior de Investigaciones Científicas (ICMM-CSIC), Madrid 28049, Spain; Orcidhttps://orcid.org/0009-0004-5050-8950 Elena Salagre - Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, United States of America Martin Gonzalez - Department of Materials Science and Engineering, Stanford University, Stanford, California 94305, United States; Stanford Institute for Materials and Energy Sciences, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States Catalin D. Spataru - Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, United States of America Joshua D. Sugar - Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, United States of America Adam L. Gross - Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, United States of America Miguel A. González-Barrio - Dto. Física de Materiales, Univ. Complutense de Madrid, Madrid 28040, Spain; Orcidhttps://orcid.org/0000-0002-4027-3711 Pilar Segovia - Dto. Física Materia Condensada, Univ. Autónoma de Madrid, Madrid 28049, Spain; IFIMAC (Condensed Matter Physics Center), Univ. Autónoma de Madrid, Madrid 28049, Spain Massimo Tallarida - ALBA Synchrotron Light Source, Cerdanyola del Vallès, Barcelona 08290, Spain Ji Dai - ALBA Synchrotron Light Source, Cerdanyola del Vallès, Barcelona 08290, Spain Suhas Kumar - Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, United States of America; Orcidhttps://orcid.org/0000-0002-6772-7250 A. Alec Talin - Sandia National Laboratories, 7011 East Avenue, Livermore, California 94550, United States of America; Orcidhttps://orcid.org/0000-0002-1102-680X Harold Y. Hwang - Stanford Institute for Materials and Energy Sciences, SLAC National Accelerator Laboratory, Menlo Park, California 94025, United States; Department of Applied Physics, Stanford University, Stanford, California 94305, United States Author Contributions A.C.Z. and A.Á.-C. contributed equally. Notes The authors declare no competing financial interest. Acknowledgments This work was supported as part of the Center for Reconfigurable Electronic Materials Inspired by Nonlinear Neuron Dynamics (reMIND), an Energy Frontier Research Center funded by the US Department of Energy (DOE), Office of Science, Basic Energy Sciences., A.Z., E.S., C.S., A.L.G., J.D.S., and E.J.F. were additionally supported by the Sandia Laboratory-Directed Research and Development (LDRD) Program. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International, Inc., for the US Department of Energy’s National Nuclear Security Administration under contract DE-NA-0003525., M.G. and H.Y.H. were supported by the US Department of Energy (DOE), Office of Science, Office of Basic Energy Sciences, Materials Sciences and Engineering Division, under contract DE-AC02-76SF00515. M.A.G., P.S., E.G.M., and A.M. acknowledge the support of the Comunidad de Madrid through the TEC-2024/TEC-380 “Mag4TIC” projects and from the Spanish MICINN through grants PID2021-123295NB-I00 and TED2021-130957B-C53. E.G.M. and P.S. are funded by the “María de Maetzu” Program for Units of Excellence in R&D (Grant No. CEX2023-001316-M). The LOREA beamline is cofunded by the European Regional Development Fund (ERDF) within the “Framework of the Smart Growth Operative Program 014-2020”., With funding form the Spanish government through the “María de Maeztu Unit of Excellence” accreditation (CEX2023-001316-M)., Peer reviewed
Docta Complutense
Publicación Artículo científico (JournalArticle)

Tuning the spin transition and carrier type in rare-earth cobaltates via compositional complexity

  • Zhang, Alan
  • Oh, Sangheon
  • Choi, Byoung Ki
  • Rotenberg, Eli
  • Brown, Timothy D.
  • Spataru, Catalin D
  • Kinigstein, Eli
  • Guo, Jinghua
  • Sugar, Joshua D.
  • Salagre, Elena
  • Mascaraque Susunaga, Arantzazu
  • Michel, Enrique G.
  • Shad, Alison C
  • Zhu, Jacklyn
  • Suhas Kumar1
  • Witman, Matthew D.
  • Kumar, Suhas
  • Talin, A. Alec
  • Fuller, Elliot J.
There is growing interest in material candidates with properties that can be engineered beyond traditional design limits. Compositionally complex oxides (CCO), often called high entropy oxides, are excellent candidates, wherein a lattice site shares more than four cations, forming single-phase solid solutions with unique properties. However, the nature of compositional complexity in dictating properties remains unclear, with characteristics that are difficult to calculate from first principles. Here, compositional complexity is demonstrated as a tunable parameter in a spin-transition oxide semiconductor La1− x(Nd, Sm, Gd, Y)x/4CoO3, by varying the population x of rare earth cations over 0.00≤ x≤ 0.80. Across the series, increasing complexity is revealed to systematically improve crystallinity, increase the amount of electron versus hole carriers, and tune the spin transition temperature and on-off ratio. At high a population (x = 0.8), Seebeck measurements indicate a crossover from hole-majority to electron-majority conduction without the introduction of conventional electron donors, and tunable complexity is proposed as new method to dope semiconductors. First principles calculations combined with angle resolved photoemission reveal an unconventional doping mechanism of lattice distortions leading to asymmetric hole localization over electrons. Thus, tunable complexity is demonstrated as a facile knob to improve crystallinity, tune electronic transitions, and to dope semiconductors beyond traditional means.