Graphene just doesn’t bond …

12 11 2010

From what we have been told, people putting contacts down onto graphene samples are having issues with wire bonding. This is a serious problem, as without the ability to put wires on, it’s difficult to connect it up to the rest of the circuit and do something interesting with it. Testing in the lab is not going to be affected, as everybody makes contacts with probes, but this is only useful for short term tests! Longer term, the ability to connect the devices up with wires is an absolute necessity…

Recently, we’ve had this problem as well and we’ve made a discovery.

It all seems to depend on the metalisation stack people use, prior to attempting the bond. A lot of the work in graphene seems to be based on the use of Chromium, followed by gold, which is then bonded using gold wire. The Chromium is a good diffusion stop for gold and ensures that the gold is kept away from the graphene surface (where it will act as a p-type dopant) However, the weak binding energy of Cr on graphene (0.17eV) and the very low diffusion energy barrier (0.01eV) means that it doesn’t form a chemically bonded contact or a complete layer, instead it forms little islands and the gold overlayer is then in direct contact with the graphene. Gold is notoriously poor at adhering to anything and so it pulls off during the bonding process.

The trick appears to relate to the use of a Titanium / Gold stack instead. Ti has a greater binding energy (1.74eV) and so it forms a chemical bond to the graphene and the diffusion energy barrier is 0.17eV. Hence, Ti forms a complete layer (there is evidence of TiC formation at high temperatures) and so sticks far better. Whilst this appears to be a step in the right direction, it still doesn’t produce a perfect result every time, so we are looking at optimising the gold thickness (this a balance between enough thickness to allow deformation during the bonding against too much intrinsic stress) in future work.

Our findings (and the other data supporting our observations on the technology behind graphene contacts) is being prepared for a submission to Applied Physics Letters, so more details on this later.

Advertisements

Actions

Information

4 responses

30 11 2010
Simon Kim

Dear blogger,

I have read the binding energy with graphene in other paper.
Cu is 0.033eV, Ni 0.125eV, Ti 0.327eV in Khomyakov’s paper in 2009.
However I can not get the result for Chrome.
So your experimental result is Cr 0.17eV, that is right?

Thank you for your labor.

Simon

30 11 2010
Alton

Our value is based on that reported in Valencia, et al (J. Phys. Chem C, Vol. 114, (2010), 14141). They also quote the value of binding energy for chemisorbed Ti as being 1.74eV

Hope that is of use

30 11 2010
Kartheek

@ Simon
An adsorbed metal atom on graphene can have three sites to bond: the hollow (H) site at the centre of hexagon, the bridge (B) site at the midpoint of C-C bond and the top (T) site at directly above a carbon atom. Each site has different Binding Energies and the most stable site is usually considered. As mentioned by Alton, our values are based on theoritical calculations by Valencia, et al. According to them (and also from other calculations), Ti is highly stable at H site with a binding energy of 1.74 eV, where as Cr has a equal Binding Energy at all the three sites with 0.17 eV. Which ever the method is used to calculate, the BE of Ti is>>>Cr.
Hope this helps.

3 03 2011
Graphene bonds! « Resilient Technology Blog

[…] identified in a previous post that wire bonding to contacts on a graphene device is tricky if the underlying metals have a low […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s




%d bloggers like this: