Beyond Si Oxide: Use of Molecular-Based Ultra-Thin Films for Novel Electronic Devices
Silicon has conquered microelectronics despite the fact that Ge or GaAs have higher electron mobility. The reason is its exceptional natural oxide, SiO2, that is chemically stable with almost no charge traps. Exploiting novel electronic materials (organic, 2D) is similarly hampered by the difficulty to control their surface phenomena. Such surface control is especially critical in non-vacuum fabrication as is often employed for novel applications such as sensors, photo-voltaic, displays etc. Miniaturization further emphasizes the importance of electronic interfaces as it increases their relative weight. The challenge is therefore to control the electronic properties of given interface at our wish. Can we make a 2nm-thick coatings which are hundreds of times more insulating than SiO2? Less prone to defects? Can we invoke sharp potential steps (0.5 eV/ 1nm ∼ 5MV/cm) to manipulate energy alignment at interfaces?
In the last couple of decades, we and others have shown that ‘molecular monolayers’ provide an effective tool toward these goals. A molecular monolayer is a 2D array of molecules (1-3 nm thick), which can homogenously cover infinite (mm2) area of variety of surfaces. It differs from 2D material like graphene in the lack of lateral bonds or a lateral band structure and in its low processing temperature. My study is focused on introduction of molecular monolayers into metal / silicon interfaces, as a generic case of hetero-junctions. I will show that alkyl-based monolayers on Si are superior to Si oxide both in reducing interface traps and blocking leakage current. In addition to their obvious use as gate dielectrics (shown by others) we were able to construct MIS tunnel junction with sharp conductance switching approaching Esaki-like tunnel diodes.
The other unique aspect of molecular monolayers is their ability to induce intense potential gradient. These potential steps easily reach 0.5 to 1 eV – significant values in terms of energy alignment that controls the rectifying / injecting nature of a given interface. This approach which was pioneered in our lab became widely used for optimizing the performance of organic-electronics, and is specifically relevant for 2D materials where introduction of foreign dopants is non-trivial. For the test-case of Hg / monolayer / n-Si junction, we showed that a single atom-modifications in the molecules drives the Si space charge from accumulation into full inversion. Namely, the molecules have induced a p-n junction within an n-doped Si. In my talk I will present both the insulating and surface-dipole effects as well as simple analysis procedure to differentiate between the two contributions to the net transport.