Design and fabrication of memory devices based on nanoscale polyoxometalate clusters

Journal name:
Date published:
Published online

Flash memory devices—that is, non-volatile computer storage media that can be electrically erased and reprogrammed—are vital for portable electronics, but the scaling down of metal–oxide–semiconductor (MOS) flash memory to sizes of below ten nanometres per data cell presents challenges. Molecules have been proposed to replace MOS flash memory1, but they suffer from low electrical conductivity, high resistance, low device yield, and finite thermal stability, limiting their integration into current MOS technologies. Although great advances have been made in the pursuit of molecule-based flash memory2, there are a number of significant barriers to the realization of devices using conventional MOS technologies3, 4, 5, 6, 7. Here we show that core–shell polyoxometalate (POM) molecules8 can act as candidate storage nodes for MOS flash memory. Realistic, industry-standard device simulations validate our approach at the nanometre scale, where the device performance is determined mainly by the number of molecules in the storage media and not by their position. To exploit the nature of the core–shell POM clusters, we show, at both the molecular and device level, that embedding [(Se(iv)O3)2]4− as an oxidizable dopant in the cluster core allows the oxidation of the molecule to a [Se(v)2O6]2− moiety containing a {Se(v)–Se(v)} bond (where curly brackets indicate a moiety, not a molecule) and reveals a new 5+ oxidation state for selenium. This new oxidation state can be observed at the device level, resulting in a new type of memory, which we call ‘write-once-erase’. Taken together, these results show that POMs have the potential to be used as a realistic nanoscale flash memory. Also, the configuration of the doped POM core may lead to new types of electrical behaviour9, 10, 11. This work suggests a route to the practical integration of configurable molecules in MOS technologies as the lithographic scales approach the molecular limit12.

At a glance


  1. Structure and electrochemical properties of compound 1a.
    Figure 1: Structure and electrochemical properties of compound 1a.

    On the left, the crystal structure of the core–shell cluster [W18O54(SeO3)2]4− (1a) is shown, with the {W18O54} cage shown as black and grey lines. The two Se core dopants are shown as orange spheres. The cluster cage can be reduced multiple times (grey area) and the two Se dopants at the POM cluster core can be oxidized (orange area). On the right, the cyclic voltammetry is obtained from microcrystals of 1a adhered to a glassy carbon electrode (diameter 1.5 mm) in 0.1 M tetrabutylammonium PF6 acetonitrile solution at a scan rate of 200 mV s−1 and a scanning range V of −2.5 V to 1.8 V against a Ag/AgCl reference.

  2. Image of the flash memory device and the drain current behaviour with an applied voltage to the control gate.
    Figure 2: Image of the flash memory device and the drain current behaviour with an applied voltage to the control gate.

    a, A cross-sectional transmission electron microscope (TEM) image (left) of the memory device with an SEM image (right) of the ~5-nm Si nanowire channel with side control gate. b, c, Measurements of the logarithmic (b) and linear (c) drain current versus gate voltage at 0.5 V source–drain voltage: before deposition of the POMs (green dashes), after the deposition of the POMs (orange dashes), after a –20 V pulse (blue line) and a +20 V pulse (red line). Panel b demonstrates depletion of the charge in the nanowire, resulting in a shift of the threshold voltage required to switch on the electrical conduction of the nanowire (orange dashed line) after 1a was deposited around the nanowire. Therefore the control gate voltage required to produce the same drain current in the nanowire has moved to a higher voltage owing to the charge on the molecules of 1a. The control gate was then used to charge and discharge the deposited molecules. −20 V was applied to the control gate, which charged the molecules around the nanowire, further increasing the control gate voltage required to produce the same drain current in the nanowire (solid blue lines in b and c). This effect could be reversed by applying a +20 V pulse to the control gate, which discharged the molecules and returned the control gate voltage for a fixed drain current to the original value with the uncharged 1a molecules around the nanowire (red lines in b and c). The effect is repeatable, demonstrating a clear shift in the threshold voltage of the device when charged. The programming window (the threshold voltage change between the charged and uncharged 1a molecules) is >1.2 V at low gate voltages.

  3. Scheme depicting the formation of the Se(v)-Se(v) bond within the cluster cage.
    Figure 3: Scheme depicting the formation of the Se(v)–Se(v) bond within the cluster cage.

    At the top, a schematic diagram shows the formation of the Se(v)–Se(v) bond in the transformation of 1a to 1b. At the bottom are the results from the DFT analysis, demonstrating the frontier orbitals and the formation of the Se(v)–Se(v) bond. Relevant orbitals delocalized over the Se moieties are highlighted in bold. The HOMO–LUMO gap is the energy gap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO). Although the orbital energies of POM clusters are separated by discrete energies, they can also be viewed as having a pseudoband-like orbital structure, and in this sense the blue box depicts the set of unoccupied tungsten d-like orbitals and the red box the set of occupied oxygen p-like orbitals.

  4. The write-once-erase device.
    Figure 4: The write-once-erase device.

    Conceptual sketch (a) and SEM/AFM images (b) of the fabricated nano-gap electrodes coated with 1a. c, The measurement procedure in terms of the applied voltage. The sample was subjected to a high-voltage push pulse applied between the source and drain electrodes on the surface of the sample, and then measured at a lower voltage level. The data are obtained by sweeping the source–drain voltage (VSD) between the surface electrodes from 0 V to 4 V and back to 0 V with the substrate gate voltage maintained constant at 3 V. ISD is the source–drain current and VG is the gate voltage. d, e, Fowler–Nordheim16, 17 plots of the current–voltage data of the POM-covered nano-gap electrodes, which demonstrate whether trap states which can hold charge for a memory device are present between the two source and drain electrodes (when there is hysteresis between the forward and reverse voltage sweeps, trapped charge is present). For the ‘0’ memory state, the hysteresis in the Fowler–Nordheim plots indicates trapped charge inside the gap between the electrodes. We observe that subjecting the system to excitation with source–drain voltage at 9–10 V changes the nature of the electron transport between the source and drain upon subsequent inspection, removing the hysteresis as shown for the memory state ‘1’. In this measurement, the effect is transient; disappearing after the first post-excitation probe, as shown conceptually in the table in panel f.

  5. Device modelling simulations of compounds 1a and 2.
    Figure 5: Device modelling simulations of compounds 1a and 2.

    a, Schematic diagram representation of a single-transistor non-volatile memory cell, indicating the aimed substitution of the poly-Si floating gate with an array of POM clusters. Tcon is the thickness of the control oxide and Ttun is the thickness of the tunnelling oxide. b, The three-dimensional electrostatic potential in the lower part of the oxide and the substrate, and two-dimensional map of the potential across the plane through the centre of the POMs, arranged in a 3 × 3 regular grid 4.5 nm from the Si–SiO2 interface for the compounds 1a and 2, as schematically illustrated. c, d, Drain current versus gate voltage with a drain bias of 50 mV in logarithmic (c) scale and linear (d) scale for a bulk molecular flash cell: 2× oxidized [W18O54(SeO3)2]2– (1b), [W18O54(SeO3)2]4– (1a), 1× reduced [W18O54(SeO3)2]5– (1c) and 2× reduced [W18O54(SeO3)2]6– (1d), in comparison with [W18O56(WO6)]10– (2), 1× reduced [W18O56(WO6)]11– and 2× reduced [W18O56(WO6)]12–. VT is the threshold voltage.


  1. Joachim, C., Gimzewski, J. K. & Aviram, A. Electronics using hybrid-molecular and mono-molecular devices. Nature 408, 541548 (2000)
  2. Shaw, J. T. et al. Integration of self-assembled redox molecules in flash memories. IEEE Trans. Electron. Dev. 58, 826834 (2011)
  3. Zhu, H. et al. Non-volatile memory with self-assembled ferrocene charge trapping layer. Appl. Phys. Lett. 103, 5310253104 (2013)
  4. Chen, P.-C., Shen, G. & Zhou, C. Chemical sensors and electronic noses based on 1-D metal oxide nanostructures. IEEE Trans. Nanotechnol. 7, 668682 (2008)
  5. Shaw, J., Xu, Q., Rajwade, S., Hou, T.-H. & Kan, E. C. Redox molecules for a resonant tunneling barrier in nonvolatile memory. IEEE Trans. Electron. Dev. 59, 11891198 (2012)
  6. Seol, M.-L., Choi, S.-J., Kim, C.-H., Moon, D.-I. & Choi, Y.-K. Porphyrin-silicon hybrid field-effect transistor with individually addressable top-gate structure. ACS Nano 6, 183189 (2012)
  7. Tans, S. J., Verschueren, A. R. M. & Dekker, C. Room-temperature transistor based on a single carbon nanotube. Nature 393, 4952 (1998)
  8. Long, D. L. & Cronin, L. Towards polyoxometalate integrated nanosystems. Chem. Eur. J. 12, 36983706 (2006)
  9. Lehmann, J., Gaita-Arino, A., Coronado, E. & Loss, D. Spin qubits with electrically gated polyoxometalate molecules. Nature Nanotechnol. 2, 312317 (2007)
  10. Li, H. et al. Layer-by-layer assembly and UV photoreduction of graphene–polyoxometalate composite films for electronics. J. Am. Chem. Soc. 133, 94239429 (2011)
  11. Fleming, L. et al. Surface-mediated reversible electron transfer reactions within a molecular metal oxide nano-cage. Nature Nanotechnol. 3, 229233 (2008)
  12. Bonfiglio, V. & Iannaccone, G. Sensitivity-based investigation of threshold voltage variability in 32-nm flash memory cells and MOSFETs. Solid-State Electron. 84, 127131 (2013)
  13. Vilà-Nadal, L. et al. Polyoxometalate {W18O56XO6} clusters with embedded redox-active main-group templates as localized inner-cluster radicals. Angew. Chem. Int. Ed. 52, 96959699 (2013)
  14. Vilà-Nadal, L. et al. Towards polyoxometalate-cluster-based nano-electronics. Chem. Eur. J. 19, 1650216511 (2013)
  15. Gallon, C. et al. Electrical analysis of mechanical stress induced by STI in short MOSFETs using externally applied stress. IEEE Trans. Electron. Dev. 51, 12541261 (2004)
  16. Simmons, J. G. Generalized formula for the electric tunnel effect between similar electrodes separated by a thin insulating film. J. Appl. Phys. 34, 17931803 (1963)
  17. Fowler, R. H. & Nordheim, L. Electron emission in intense electric fields. Proc. R. Soc. Lond. A 119, 173181 (1928)

Download references

Author information

  1. These authors contributed equally to this work.

    • Christoph Busche &
    • Laia Vilà-Nadal


  1. WestCHEM, School of Chemistry, The University of Glasgow, Glasgow G12 8QQ, UK

    • Christoph Busche,
    • Laia Vilà-Nadal,
    • Jun Yan,
    • Haralampos N. Miras,
    • De-Liang Long &
    • Leroy Cronin
  2. School of Engineering, The University of Glasgow, Glasgow G12 8LT, UK

    • Vihar P. Georgiev,
    • Asen Asenov,
    • Rasmus H. Pedersen,
    • Nikolaj Gadegaard,
    • Muhammad M. Mirza &
    • Douglas J. Paul
  3. Departament de Química Física i Inorgànica, Universitat Rovira i Virgili, Marcel·lí Domingo street, 43007 Tarragona, Spain

    • Josep M. Poblet


L.C. conceived the idea, designed the project and coordinated the efforts of the research team. J.Y. synthesised the clusters and conducted the first electrochemistry experiments and structural characterization with D.-L.L. H.N.M., C.B., L.V.-N., and L.C. helped to characterize the physical properties of the clusters. C.B. did the electron paramagnetic resonance, electrochemistry and spectroscopic measurements. L.V.-N., L.C., V.P.G. and A.A. designed the theory-to-modelling strategy. L.V.-N., with J.M.P., did the DFT calculations. V.P.G. and A.A. did the device simulation. R.H.P. and N.G. fabricated and characterized the electrode arrays, produced the devices, made the measurements and characterized the data. M.M.M. and D.J.P. designed the nanowire arrays and M.M.M. fabricated the electrodes and optimized the data with D.J.P., who helped analyse the results. C.B., L.V.-N. and L.C. co-wrote the paper with input from all the authors.

Competing financial interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to:

Atomic coordinates for the reported crystal structures have been deposited with the Cambridge Structural Database under the accession codes 997534 (compound precursor), 997535 (compound 1a), 997536 (compound 1c) and 997537 (compound 1d), and full synthetic, electrochemical, device theory, device modelling and electronic device data is given in the Supplementary Information.

Author details

Supplementary information

PDF files

  1. Supplementary Information (3.1 MB)

    This file contains Supplementary Text and Data, Supplementary Figures 1-14, Supplementary Tables 1-8 and Supplementary References.

Additional data