Atomic-Scale Metrology is Key in Fabricating Graphene Devices
Graphene is a very interesting nanomaterial with potential for applications in many different fields including nanoelectronics. However, the properties of graphene can vary broadly and depend sensitively on its integration in device structures and the details of its interaction with other materials, such as underlying substrates or gate dielectrics. Unlike other semiconductor electronic devices, where the active layer is buried below the surface and where microscopic details of transport cannot be directly examined, graphene is exposed at a surface and can be directly examined on the atomic scale using scanning tunneling microscopy (STM) and scanning tunneling spectroscopy (STS).
Some of the unique properties of graphene are its chemical stability and a relatively weak interaction with the environment. This is why graphene has high-quality transport properties while being just a monolayer thin. However, while the interaction with the environment is indeed weak, this residual interaction ultimately determines the properties of graphene-based devices and the potential for their applications. To control this interaction, one should first measure it.
Although STM has been recently applied to study graphene devices, understanding the effects of disorder and scattering on transport properties remains a challenging problem as every single location is somewhat unique and it is often hard to identify any common behavior.
New work just published in Nature Physics by researchers at the National Institute of Standards and Technology (NIST), first authored by Suyong Jung and Gregory M. Rutter, provides microscopic details of graphene interaction with a substrate in the most common device structure used so far ("Evolution of microscopic localization in graphene in a magnetic field from scattering resonances to quantum dots").
"From our scanning tunneling spectroscopy measurements, we can determine a wealth of microscopic information on the local properties of graphene devices," Nikolai Zhitenev, Group Leader of the CNST Energy Research Group, tells Nanowerk. "Our work is the first to study magnetic quantization in a graphene device with back gating. In comparison with previous works, we have here, for the first time in STM experiments, the combination of a magnetic field that curves the motion of the carriers and gradually changes the scattering and the localization processes with the control of carrier density in graphene through the backgate. This powerful combination allows us to learn much more about the physics of realistic devices with disorder."
To perform their experiment, the researchers made a 'sandwich' of alternating conducting, semiconducting and insulating layers and structures with a single atomic sheet of graphene and another conductor separated by an insulating layer. When the bottom conductor is charged, it induces an equal and opposite charge in the graphene.
Zhitenev and his collaborators demonstrate that the localization phenomena in graphene contribute to the STS spectra through two distinct set of features.
"First, we see how the density of carriers varies as a function of spatial location" he says. "This density variation is caused by the disorder potential which can be created by charged impurities in the SiO2 gate insulator or trapped at the interface between graphene and oxide during fabrication process, and/or topographical roughness of the substrate casing deformation of graphene or a variety of other sources."
The determined characteristic spatial length of the potential variations is rather small, on the order of 30-50 nm. This explains why an STM is such a capable tool for examining the physical phenomena at these length scales. The density fluctuations can be very significant depending on total carrier density in graphene.
"For example" says Zhitenev, "at low densities, we can never reach zero density of ideal graphene, instead the sign of carriers can change from one location to another and a spatial checkerboard of electron and hole puddles is formed."
He notes that, apart from the density fluctuations, the disorder potential essentially defines how conductive the graphene layer causing carrier scattering and localization. "This statement is true not only for graphene but for any conducting or semiconducting materials. However, graphene is one of the rare cases where the relationship between the disorder and the transport properties can be examined in microscopic detail. Specifically, from the electronic spectra we can infer how the motion of electrons or holes is affected by the disorder."
In the absence of a magnetic field, the disorder potential scatters the carriers, creating local potholes that slow down the average motion. In strong magnetic fields, the carriers can be completely trapped at spatial locations of minima and maxima of disorder potential.
The NIST team points out that the methodology developed in their research can be directly applied to characterize other graphene-based device structures to understand why they work or do not work, and what can be changed or has to be changed to make them work for specific applications.
Image: Reproduced by permission from Macmillan Publishers: Nature Physics, advanced online publication 9 January 2011 (doi: 10.1038/nphys1866).