MoS2, which is made of a single layer of molybdenum and sulphur, is a transition metal dichalcogenide (TMDC) semiconductor. The material boasts a direct bandgap (of 1.8 eV), which means that it might be better than indirect bandgap silicon for making certain photonic devices. It might even rival the most famous of all 2D materials, graphene (which does not have a bandgap at all in its pristine state), for use in future electronic circuits.

A direct bandgap is important when it comes to making devices like LEDs, solar cells and photodetectors, and any other photonic devices that exploit electron-hole pair excitation, because devices made with direct rather than indirect gap semiconductors are more efficient. A bandgap also means that a device can be more easily switched on and off – a non-negligible advantage for transistors, for example.

But that is not all: MoS2 also has good charge mobilities - perhaps as high as 500 cm2/Vs. These values compare well to state-of-the-art silicon. It is also a so-called van der Waals crystal (made up of 2D sheets that are weakly bonded to each other), which means that it is compatible with a number of substrates – even transparent or plastic ones. And to top things off, single-layer molybdenite is less than 0.70 nm thick, which means that very thin transistors can be made from it. Such thin devices would be better at dissipating unwanted heat than conventional devices made from silicon.

Efficient hole injection

Now, researchers at the University of California, Berkeley, and the Lawrence Berkeley National Laboratory led by Ali Javey say they have made a p-type MoS2 FET by using molybdenum trioxide (MoOx) contacts. “Instead of using elemental metals, as was previously the case, we use ultrahigh work-function contacts made of MoOx," explains team member Steven Chuang. “MoOx has traditionally been exploited as a hole injection layer for organic materials, but only recently have we begun thinking about putting it to use in inorganic semiconductors.”

So why is MoOx such an efficient hole injection layer for MoS2? “There are two main reasons,” says Chuang. “Firstly, MoOx has a higher work function than any elemental metal, which places its energy level closer to TMDC valence bands so that holes can be injected into the 2D material more efficiently. Secondly, our results suggest that the MoOx/TMDC interface is very different to elemental metal/TMDC interfaces. Indeed, previous research has revealed that, in metal/TMDC interfaces, metal energy levels are pinned far above TMDC valence bands. Our analysis shows that MoOx energy levels are not pinned so far above.”

Clear p-type behaviour

The California team also found that devices made from TMDC contacted with MoOx are better than those contacted with palladium (a popular high-work-function elemental metal contact). “For one, we saw that, when contacted with MoOx, devices made from tungsten selenide (another well known TMDC) show greater 'on' currents and lower 'Schottky barrier' heights than those contacted with Pd. What is more, as mentioned before, MoS2 FETs made with MoOx contacts show clear p-type behaviour while those made with Pd contacts show n-type.”

It is not all plain sailing though. The researchers admit that their carrier-selective oxide contacts are far from perfect and need to be improved. “To this end, we are looking into other oxide contacts to TMDCs to further reduce the resistance between these two constituents for both n- and p-type FETs,” Chuang told nanotechweb.org. “These studies will allow us to make high-performance TMDC logic components and also help us better understand the intrinsic charge transport properties of these 2D dichalcogenides.”

The research is detailed in Nano Letters.

Further reading

Going it alone: new dichalcogenide behaves like a monolayer (Feb 2014)
First light for MoS2 (Apr 2013)
MoS2 in new mobility record (Mar 2013)