Research ArticleAPPLIED SCIENCES AND ENGINEERING

Reversible spin storage in metal oxide—fullerene heterojunctions

See allHide authors and affiliations

Science Advances  20 Mar 2020:
Vol. 6, no. 12, eaax1085
DOI: 10.1126/sciadv.aax1085
  • Fig. 1 Spin-polarized charge trapping.

    (A) Bottom: NEXAFS at the carbon K-edge in a SiO2//Co(30)/Al2O3(1.6)/ C60(15)/MnOx(2.5) junction—film thickness in brackets in nanometers. After a bias is applied and then removed, the LUMO broadens and shifts to 283 eV. This persists until the electrodes are connected to a common ground. Top: π* LUMO states extend up to ~290 eV. In this region, XMCD measurements in the charged state after a bias show spin polarization at room and low temperature. (B) Changes in the carbon K-edge (bottom) and XMCD (top) with the applied voltage or current. XMCD is measured in the floating charged state. The different impedance of the source meter leads to changes in the trapped charge during the measurement and in the LUMO* peak. The higher internal impedance of the voltage source results in better signal-to-noise XMCD measurements.

  • Fig. 2 Interface modeling.

    (A) Top: High-resolution cross-sectional TEM image of a Co(5)/C60(20)/MnOx(10) junction. Inset: Fast Fourier transform (FFT) of the region indicated shows peaks in the diffraction pattern due to nanocrystalline grains in the sublimed C60 film. Bottom: Elemental chemical analysis of the interface—note the Gaussian profile of the e-beam (~10 nm full width at half maximum). The oxygen-to-manganese ratio is ~2:1 close to the C60 interface. (B) Spin-polarized density of states (PDOS) in the neutral (top) and with 0.6 e added to the C60 interface (bottom). The formation of a C60─O bond at the surface leads to a strong interfacial dipole and the generation of half-metallic interface states. (C) Changes in the charge density with respect to the neutral system along the direction perpendicular to the interface plane (z). The profile reveals the formation of a dipole layer density (μσ) owing to accumulation of negative charge at the C60/β-MnO2(110) junction, and positive charge on the opposite side of C60(1st) toward C60(2nd). (D) Computed change in total charge density after 0.6 e are added to the C60 interface (|Δρ| = 10−6 e Å−3) for the oxidized, chemically bound minima for spin-up and spin-down states.

  • Fig. 3 Spin stabilization facilitated by stray fields.

    (A) Wide area TEM image showing the presence of grains some ~50 to 100 nm in lateral size. (B) Schematic for the stray field μoHs generated by Co grains and the buildup of spin-polarized trapped charge via photocurrents. (C) Photocurrent in Co(5)/C60(20)/MnOx(10) devices as a function of the magnetic history. Lower remanent magnetization leads to less efficient charge trapping and therefore a higher photocurrent. (D) Discharge of the open-circuit voltage (VOC) under a series of in-plane magnetic fields μoHx. The discharge time τ is up 30% faster at the coercive field, when the magnetization of the cobalt film is nil. (E) The magnetic field dependence of τ follows the anisotropic magnetoresistance (AMR) of the Co electrode, showing the correlation between Co magnetization and spin-polarized charge storage (blue line is a guide to the eye).

  • Fig. 4 Magnetic field effects and schematic of the trapped spin capacitor mechanism.

    (A) Top: XMCD for the initial, charged (floating), and discharged (ground) states in a Co/Al2O3/C60/MnOx junction showing spin-polarized trapped charge. Bottom: NEXAFS in remanence after OOP magnetic fields (Bz). The 283 eV LUMO* peak is displaced toward the pristine LUMO until the Co electrode is magnetically saturated OOP. (B) Decay time constant τ for optically charged junctions, typically ~30 to 300 s for 100 × 100 μm2 junctions (fig. S11) (35). The dependence of τ on Hz is similar to that of the TEY at the pristine 284 eV LUMO position. (C) Schematic for spin capacitance. Top: Minority electrons are trapped when the quantization axis C60 layer (σ) is parallel to that of the C60/MnOx interface (σ′). Bottom: When σ ≠ σ′, the conductivity G is above zero for both spins. In particular, for σ ⊥ σ′, both spin orientations in C60(2nd) transfer to the interface majority band with the same probability and conductivity (≠0).

  • Fig. 5 LE-μSR detection of local magnetic moments in charged samples.

    (A) Calculated penetration profiles for the voltage-biased sample (46). (B) The μ+@C60[−] frequency is increased in the charged state at 10 and 12 keV due to the enhanced local magnetization at the C60/MnOx interfaces when charged. (C) Increased μ+@C60[−] depolarization rate and oscillation amplitude in the charged state at 10 and 12 keV. (D) Penetration profiles for the photovoltaic sample. (E) The μ+@C60[−] frequency is higher during the photocurrent (light on) and increases further in the charged floating state (light off). (F) Changes in magnetic susceptibility due to spin-polarized trapped charge at the C60/MnOx interface (15.3 keV) are absent in the bulk C60 film (10.4 keV).

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/12/eaax1085/DC1

    Section S1. Sample structure

    Section S2. Junction characterization

    Section S3. DFT simulations

    Section S4. X-ray absorption spectroscopy

    Section S5. Sample preparation and LE-μSR fitting details

    Fig. S1. Transport characteristics of junctions with and without an alumina barrier.

    Fig. S2. Thin film characterization.

    Fig. S3. DFT geometries.

    Fig. S4. DFT optimized geometries for the neutral and charged states.

    Fig. S5. PDOS calculations.

    Fig. S6. Band structure and molecular orbitals around the Fermi energy.

    Fig. S7. NEXAFS experimental details.

    Fig. S8. NEXAFS and XMCD when using a CuOx electrode.

    Fig. S9. Photovoltaic effect in C60/MnOx junctions.

    Fig. S10. Dependence of the photocurrent of the LE-μSR sample with temperature.

    Fig. S11. Discharge for a photovoltaic-charged device.

    Fig. S12. LEM experimental details for electrically charged sample.

    Fig. S13. Fitting of LEM data for electrically charged sample.

    Fig. S14. LE-μSR probing of a photovoltaic sample at 50 K in a transverse field of 300 G.

    Fig. S15. Magnetometry data.

    Table S1. Computed energy differences (eV) between the different adsorption geometries of C60 on β-MnO2(110)-2x2 as a function of the extra electronic charge added to the system.

    Table S2. Computed Bader charges (Q) (63) for the different adsorption geometries of C60 on β-MnO2(110)-2x2 as a function of the extra electronic charge added to the system.

    References (4767)

  • Supplementary Materials

    This PDF file includes:

    • Section S1. Sample structure
    • Section S2. Junction characterization
    • Section S3. DFT simulations
    • Section S4. X-ray absorption spectroscopy
    • Section S5. Sample preparation and LE-μSR fitting details
    • Fig. S1. Transport characteristics of junctions with and without an alumina barrier.
    • Fig. S2. Thin film characterization.
    • Fig. S3. DFT geometries.
    • Fig. S4. DFT optimized geometries for the neutral and charged states.
    • Fig. S5. PDOS calculations.
    • Fig. S6. Band structure and molecular orbitals around the Fermi energy.
    • Fig. S7. NEXAFS experimental details.
    • Fig. S8. NEXAFS and XMCD when using a CuOx electrode.
    • Fig. S9. Photovoltaic effect in C60/MnOx junctions.
    • Fig. S10. Dependence of the photocurrent of the LE-μSR sample with temperature.
    • Fig. S11. Discharge for a photovoltaic-charged device.
    • Fig. S12. LEM experimental details for electrically charged sample.
    • Fig. S13. Fitting of LEM data for electrically charged sample.
    • Fig. S14. LE-μSR probing of a photovoltaic sample at 50 K in a transverse field of 300 G.
    • Fig. S15. Magnetometry data.
    • Table S1. Computed energy differences (eV) between the different adsorption geometries of C60 on β-MnO2(110)-2x2 as a function of the extra electronic charge added to the system.
    • Table S2. Computed Bader charges (Q) (63) for the different adsorption geometries of C60 on β-MnO2(110)-2x2 as a function of the extra electronic charge added to the system.
    • References (4767)

    Download PDF

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article