Conduction mechanism of nitronyl-nitroxide molecular magnetic compounds

We investigate the conduction mechanisms of nitronyl-nitroxide (NIT) molecular radicals, as useful for the creation of nanoscopic molecular spintronic devices, finding that it does not correspond to standard Mott behavior, as previously postulated. We provide a complete investigation using transport measurements, low-energy, sub-THz spectroscopy and introducing differently substituted phenyl appendages. We show that a nontrivial surface-charge-limited regime is present in addition to the standard low-voltage Ohmic conductance. Scaling analysis allows one to determine all the main transport parameters for the compounds and highlights the presence of charge-trapping effects. Comparison among the different compounds shows the relevance of intermolecular stacking between the aromatic ring of the phenyl appendix and the NIT motif in the creation of useful electron transport channels. The importance of intermolecular pathways is further highlighted by electronic structure calculations, which clarify the nature of the electronic channels and their effect on the Mott character of the compounds.

Figure 1

(a) Schematic representation of the general structure of the nitronyl-nitroxide (NIT-R) radicals, showing the different oxidation states that are available at room temperature and the reduction or oxidation processes from the pristine neutral form. (b) Schematic representation of the five radicals investigated, with the corresponding acronyms below. (c) Crystal structure of radical NIT-PhOMe, with the cylinders highlighting the presence of overlap between the electronic clouds of aromatic appendages and the NIT moiety. Carbon is represented in black, oxygen in red, nitrogen in blue, and hydrogen atoms are omitted for the sake of clarity.

Figure 2

(a) Room temperature $I-V_{sd}$ characteristic curves resulting from averaging on multiple sets of measurements collected on NIT-PhBr (green down triangles), NIT-PhOMe (blue circles), NIT-PhOtBu (red squares), NIT-PhOPh (black diamonds), and NIT-Py (violet up triangles). (b) Dependence of the $I$ vs $V_{sd}$ curves on the sweep rate of the bias voltage, as acquired for the NIT-PhOMe radical. Decreasing the sweep rate capacitive effects can be minimized, as can be seen from the narrowing gap between forward and backward sweeps and decreased zero-voltage residual current.

Figure 3

(a) Scheme of the THz spectrometer used for the measurements (left) and scheme of the multiple internal reflections within the sample, leading to Fabry-Perot interferences (right). (b) Transmission spectra in the 3–23 ${\mathrm{cm}}^{-1}$ range, as measured for the NIT-R compounds on pellets at $T$=10 K. (c) Temperature and frequency dependence of the transmission spectra of NIT-PhOMe in the 3–23 ${\mathrm{cm}}^{-1}$ and 10–300 K ranges. A vertical logarithmic scale is used for clarity.

Figure 4

(a) Schematic representation of the different conduction regimes available for the NIT-R radicals. The Ohmic regime, present at low voltages, gives way, at higher voltages, to surface-charge-limited transport with possible trap effects. The effects of exponentially distributed (violet) and localized-energy (red) traps are also shown. (b) Schematic depiction of the different types of traps available for NIT-R systems, with the localized traps (red lines) and energetically distributed traps (violet curve) placed between the valence and conduction bands of the system.

Figure 5

(a) Bilogarithmic plot of the voltage dependence of the conductance for all NIT-R compounds, showing the constant value for lower voltages, connected to the Ohmic regime, and the raising in conductance at higher voltages, where SCL effects become dominant. (b) Bilogarithmic plots of the current vs source-drain voltage characteristic curves for the different NIT-R compounds. The dashed and continuous lines show the fitting for the Ohmic and SCL regimes, respectively. (See Table 2 and Fig. 6 for the results.) The crossover voltages $V_c$, as obtained from the intersection of the two regimes, are also highlighted with vertical lines.

Figure 6

(a) Example of fitting of the different regimes for a number of measurements performed on NIT-PhOMe, allowing the statistical analysis of the conduction regimes. Details about data sets in Table 3. The resulting distribution of the exponents, as obtained for the low-voltage (blue diamonds) and high-voltage (red dots) regimes, are shown in (b) for NIT-PhBr, (c) for NIT-PhOMe, (d) for NIT-PhOtBu, (e) for NIT-PhOPh, and (f) for NIT-Py. No-count points are shown for clarity, and solid lines correspond to Gaussian fits (values in Table 2).

Figure 7

Left: Calculated DOS for NIT-PhOMe showing the different densities close to $E_F$ and charge density maps obtained by projecting in real space the states in different energy intervals, as highlighted by the shaded areas. Right: Charge density maps obtained for NIT-PhOMe by projecting in real space the states around $E_F$ (bottom panel) and around 2 eV (upper panels).

Figure 8

Temperature dependence of the logarithm of the optical ${\epsilon }_2$ value for NIT-PhBr (green triangles), NIT-PhOMe (blue dots), NIT-PhOtBu (red squares), and NIT-Py (violet triangles), as determined from the fitting of the sub-THz spectra. Solid lines correspond to fits with the expected behavior for small polarons and low-temperature residual conductivity, as described in the text.

Figure 9

(a) Example of the presence of distributed traps within the SOMO-LUMO gap, as observed on a bilogarithmic plot. The black lines represent the exponents that indicate the conductivity regime of each part of the curve. Over $V_t$ = 500 V the effect of distributed traps is clearly visible, as explained in the text. (b) Examples of localized traps for NIT-Py (top) and NIT-PhOMe (bottom).