Missing Atom as a Source of Carbon Magnetism

Atomic vacancies have a strong impact in the mechanical, electronic, and magnetic properties of graphenelike materials. By artificially generating isolated vacancies on a graphite surface and measuring their local density of states on the atomic scale, we have shown how single vacancies modify the electronic properties of this graphenelike system. Our scanning tunneling microscopy experiments, complemented by tight-binding calculations, reveal the presence of a sharp electronic resonance at the Fermi energy around each single graphite vacancy, which can be associated with the formation of local magnetic moments and implies a dramatic reduction of the charge carriers’ mobility. While vacancies in single layer graphene lead to magnetic couplings of arbitrary sign, our results show the possibility of inducing a macroscopic ferrimagnetic state in multilayered graphene just by randomly removing single C atoms.

Figure 1

(a) $17\times 17{\mathrm{nm}}^2$ STM topography, measured at 6 K, showing the graphite surface after the ${\mathrm{Ar}}^{+}$ ion irradiation (for a larger scale overview of the same region, see 21). Data analyzed using WSXM [37]. Single vacancies occupy both $\alpha$ and $\beta$ sites of the graphite honeycomb lattice. Sample bias: $+270\mathrm{mV}$, tunneling current: 1 nA. (b) Schematic diagram of the graphite structure. (c) 3D view of a single isolated vacancy. Sample bias: $+150\mathrm{mV}$, tunneling current: 0.5 nA. (d) STS measurements of the LDOS induced by the single vacancy and of graphite. Black open circles correspond to $dI/dV$ spectra measured on pristine graphite and red solid circles correspond to $dI/dV$ spectra measured on top of the single vacancy, showing the appearance of a sharp resonance at $E_F$. $dI/dV$ measurements were done consecutively at 6 K with the same microscopic tip.

Figure 2

(a) LDOS as a function of sample voltage $V$ and position $x$ along the line drawn in (b). A light central line has been drawn to outline the evolution of the resonance peak height, showing a clear $R3$ modulation. (b) STM topographic image of a single graphite vacancy. Image size $8\times 8{\mathrm{nm}}^2$; sample bias $+200\mathrm{mV}$ tunneling current 0.6 nA. (c) $r^{-2}$ decay of the resonance intensity. Black dots correspond to the maxima of the resonance peak height and the line is parabolic fit to the experimental data.

Figure 3

(a) ($dI/dV$) spectra measured with the same tip on top of a single vacancy in an $\alpha$ (red solid circles), $\beta$ (blue open circles) site and on pristine graphite (black squares). The intensity of the resonance measured on top of the $\alpha$ vacancy is much higher than the one of the $\beta$ one, indicating that removing a C atom from an $\alpha$ site generates a stronger magnetic moment. Solid lines are fits to a Lorentzian function giving a FWHM of $\sim 5\mathrm{mV}$ for both the resonance on the $\alpha$ and $\beta$ site. Small variations (of a couple of mV) in the position of the resonance peak maxima were observed, which we attribute to the local environment of each specific vacancy. (b) Calculated density of states in the atom nearest to vacancy. Red solid line: vacancy in $\alpha$ site; blue dashed line: vacancy in $\beta$ site.