Oxide Growth on Graphene by Pulsed Laser Deposition
The idea of using graphene for long range spin transport has been actively explore by a number of groups around the world [1-4]. As with almost all non-metallic systems, spin injection is only made efficient by engineering a large spin-dependent interface resistance using a thin tunneling barrier [4]. For graphene, the growth of such barriers is challenging due its inertness that causes most films to de-wet into 3D islands. Organic molecular films exhibit weak intermolecular interactions that encourage formation of flat barriers on graphene. However, a more traditional approach of thin oxide tunnel barriers is also important to explore. Our group has initiated a collaboration with Dr. Marc Ulrich at ARO to grow oxide films on graphene by pulsed laser deposition for the purpose of creating tunneling barriers and multiferroic gate dielectrics for spin- field effect transistors. More generally, coupling PLD oxides to graphene could allow integration of the complete range of complex oxide functionality with the unique electronic and transport properties of graphene.
The idea of using graphene for long range spin transport has been actively explore by a number of groups around the world [1-4]. As with almost all non-metallic systems, spin injection is only made efficient by engineering a large spin-dependent interface resistance using a thin tunneling barrier [4]. For graphene, the growth of such barriers is challenging due its inertness that causes most films to de-wet into 3D islands. Organic molecular films exhibit weak intermolecular interactions that encourage formation of flat barriers on graphene. However, a more traditional approach of thin oxide tunnel barriers is also important to explore. Our group has initiated a collaboration with Dr. Marc Ulrich at ARO to grow oxide films on graphene by pulsed laser deposition for the purpose of creating tunneling barriers and multiferroic gate dielectrics for spin- field effect transistors. More generally, coupling PLD oxides to graphene could allow integration of the complete range of complex oxide functionality with the unique electronic and transport properties of graphene.
Our first experiment [5] in this area showed that it is possible to create atomically-smooth MgO films on graphene by PLD at room temperature (Figure 1). This oxide is a very common one for spin injection and can sometimes be used as a band-symmetry-based "spin filter". Roland Kawakami's group showed that it can be grown by oxide MBE on graphene only if a submonolayer Ti film is pre-deposited to inhibit rapid surface diffusion of Mg. Our PLD method achieves the same goal (Fig. 1c) without a Ti seed layer by using a very high instantaneous deposition flux to create a high nucleation density before significant surface diffusion can occur. We have also shown diffraction evidence that such films are non-crystalline (this limits possibilities for active spin filtering) and that they are susceptible to hydroxylation in air (Fig. 1e) that can only be partially removed by annealing in vacuum (due to very strong binding of OH- at submonolayer concentration defect sites in the film) [5]. Nevertheless, they are expected to be suitable for simple tunneling-based spin injection and we are exploring collaborations to test our MgO films in spintronics devices. Our current film growth and characterization efforts in this project are aimed at growing magnetic oxides by this technique to create multifunctional gate dielectrics for spin manipulation in graphene channels [3].
Our first experiment [5] in this area showed that it is possible to create atomically-smooth MgO films on graphene by PLD at room temperature (Figure 1). This oxide is a very common one for spin injection and can sometimes be used as a band-symmetry-based "spin filter". Roland Kawakami's group showed that it can be grown by oxide MBE on graphene only if a submonolayer Ti film is pre-deposited to inhibit rapid surface diffusion of Mg. Our PLD method achieves the same goal (Fig. 1c) without a Ti seed layer by using a very high instantaneous deposition flux to create a high nucleation density before significant surface diffusion can occur. We have also shown diffraction evidence that such films are non-crystalline (this limits possibilities for active spin filtering) and that they are susceptible to hydroxylation in air (Fig. 1e) that can only be partially removed by annealing in vacuum (due to very strong binding of OH- at submonolayer concentration defect sites in the film) [5]. Nevertheless, they are expected to be suitable for simple tunneling-based spin injection and we are exploring collaborations to test our MgO films in spintronics devices. Our current film growth and characterization efforts in this project are aimed at growing magnetic oxides by this technique to create multifunctional gate dielectrics for spin manipulation in graphene channels [3].
Figure 1. a) AFM of clean graphene on SiC(0001); b) line profile from the black line in a); c) AFM of 1 nm MgO film grown on the substrate shown in a); d) Line profile from the black line in c); e) XPS spectra (Al Ka 1486.7 eV) of the O1 s core level as a function of annealing and take-off angle indicating the partial removal or surface bound OH.
Figure 1. a) AFM of clean graphene on SiC(0001); b) line profile from the black line in a); c) AFM of 1 nm MgO film grown on the substrate shown in a); d) Line profile from the black line in c); e) XPS spectra (Al Ka 1486.7 eV) of the O1 s core level as a function of annealing and take-off angle indicating the partial removal or surface bound OH.
References
[1]Tombros et al., Nature 448, 571 (2007).
[2]Cho et al., Appl. Phys. Lett. 91, 123105 (2007).
[3]Semenov et al., Appl. Phys. Lett. 91, 153105 (2007).
[4] Han et al., J. Mag. and Mag. Mater. 324, 369 (2012).
[5] Stuart et al., J. Vac. Sci. Tech B, in press (2013).