[attachment=9474]
GRAPHENE FET’S: PROMISES AND CHALLENGES
Abstract:
We present a critical review of graphene for field-effect transistor (FET) application. By using graphene sheet as the channel material, the new FET is expected offer significant advantages in certain aspects over and above that of the present Si-based counterpart. There are many challenges as well since the overall operating principle of the FET channel is different. We present our perspective with regard to these challenges. The most exciting properties of graphene today people are interested in is its high mobility. It has been reported that the mobility could be as high as 200,000cm2V-1s-1, which is much higher than silicon and other semiconductor materials. Graphene is considered to be a promising candidate for future nanoelectronics due to its exceptional electronic properties. Unfortunately, the graphene field-effect transistors (FETs) cannot be turned off effectively due to the absence of a band gap, leading to an on/off current ratio typically around 5 in top-gated graphene FETs. All these advantages of graphene are the reasons one wants to use graphene as a channel material.
KEYWORDS: Graphene, field-effect transistors, on/off current ratio, transport band gap, ultracapacitors.
1. Introduction:
Graphene is a two-dimensional material composed of carbon atoms. All the carbon atoms are arranged in a hexagonal lattice structure, which is similar to that of carbon nanotubes (CNTs) and fullerenes (C60). Figure 1 shows the band structure (E-k relationship) of graphene. At the six corners of its Brillouin zone, the E-k relationship becomes nearly linear and the gandgap vanishes. So graphene is a zero-bandgap material, unlike the common semiconductor materials.
Graphene was not discovered until 2004 when people used scotch tape to pill off one layer of graphene from a graphite block. After its discovery, the research on utilizing graphene in FETs has never been interrupted. Typical schematic diagrams of a top-gated graphene FET (gFET) and a back-gated gFET’s are shown in Figure 2. The structure of gFET's is similar to that of conventional Si-based FETs, with the Si channel being replaced by a graphene sheet. And the conductance of the graphene channel is controlled by the voltage applied on gate in exactly the same way as in conventional FETs
. Promises of Graphene for gFET Applications:
As mentioned earlier, the high mobility of graphene is the main reason for the anticipated high performance of gFETs. In addition to mobility, its two-dimensionality promises reduced short-channel effects (SCEs), one of the most challenging hurdles to continued scaling of Si MOSFET technology. gFETs may also benefit from the higher saturation velocity vsat of graphene compared to that of silicon, to gain a higher fT. Silicon has a lower energy of transverse phone, which is around 60meV, resulting in a vsat of 107cm/s, while graphene, in contrast, has a vsat of 7×107 cm/s, attributing to its lowest lying optical phonon branch where the out-of plane optical phone energy phonon energy is about 110 meV and longitudinal and optical phonon energies are about 200meV. All these advantages of graphene are the reasons one wants to use graphene as a channel material. Some people started to dream that gFETs could be next generation transistors when the current Si-based FET technology runs out of streams. However, to evaluate this possibility, not only the advantages of graphene, but also the challenges that graphene will inevitably bring in should be considered.
3. Challenges of gFET Applications:
FETs are the main elements in radio frequency (RF) and digital applications. It is not surprised that gFET, with the channel material replaced to graphene, has different properties compared with those of conventional FETs. However the replacement leads to some challenges that need to be solved to truly fabricate a working gFETs with better performance.
The first challenge that gFET has to face is that graphene is a zero band gap material (Figure 1). This leads to the gFET not able to be truly turned off and hence gFET cannot compete the conventional FET in terms of the Ion/Ioff ratio. Recently it was found that the band gap could be opened up by confining the graphene into graphene nano-ribbons (GNRs) whose bandgap is dependent on the ribbon width following the rule (Figure 3)
Eg = 0.8/W (nm) -- [1]
Thus GNRs may be good channel materials if only the Ion/Ioff ratio is concerned. Electrostatic simulations of gate capacitances estimated that the hole mobility of sub- 10nm GNRs is around 100-200cm2v-1s-1, which is still comparable to the conventional silicon devices, therefore forfeiting the advantages of graphene discussed above. But the value is too low compared with the 200,000 cm2v-1s-1, and a question arises whether the low mobility is intrinsically associated with GNRs, or a better experimental fabrication process can improve it.
Another challenge with graphene in gFETs is the lack of saturation or week saturation on its IDS-VDS curve (Figure 4). It is not a problem in Si-based FETs in that either channel pinch off or carrier velocity saturation causes IDS saturation. However, graphene is a special material with near zero bandgap and it can be switched from n-type to p-type at a certain VDS, leading to the superimposition of the p-type linear region onto n-type saturation region. Consequently, no saturation region is found on the output IDS-VDS curve. It hinders the RF performance of gFETs, because most RF FETs are working in the saturation region. However, no experimental evidence is available at this time with regard to IDS saturation due to electron velocity saturation.
The linearity of a gFET is necessary for the RF applications. Graphene has nearly constant charge carrier mobility except near the Dirac point (Figure 5) and hence the transconductance gm is near constant. As a result, the linearity of gFET could be realized.
We have reviewed the perspective challenges of gFETs. On the other hand, challenges exist in the process of fabricating gFET too, which will be discussed next section.
4. Graphene FET (gFET) Technology: the Building Blocks of gFET:
In gFET building blocks, the contact resistance between graphene and the source/drain metal is of big importance because graphene itself has a very low resistance. Surprisingly, the contact resistance between graphene and metal is dominated by contact edge, not contact area and that all carriers are only injected at the edge of the contact metal, not through the whole contact area. In addition, selecting proper contact material is a direct method to lower the contact resistance. For example, Ni has a lower contact resistivity (RcW~500Ωμm) with graphene compared to Ti/Au and Cr/Au (103~106Ωμm). A metal plate shields the external quasistatic electric fields completely by redistributing the internal electrons. The situation is different for graphene channel in a gFET as it is a genuine two-dimensional system, which is only able to shield part of the penetrating electrical field because of its lower density of states. The incomplete screening manifests itself as a quantum capacitor CQ (Figure 6) that has a value comparable to that of gate oxide Cox.