how to construct a Titanium electrode

[njuxyh]

Hi:

i am trying  to construct a Titanium electrode, i am puzzled that when i used a "add"---->"from database", it shows three choice: alfa, beta, omega. so i do not know which to choose? which is often chosen as a electrode? as i know for gold, its (1,1,1) surface is ofter chosen as a electrode.
Ti electrode is used widely in molecular electronic ,  i have found a reference http://pubs.acs.org/doi/pdfplus/10.1021/nn401357p,(The file is too large, and i can not attach it) it is metal-GNRs-metal connection, how to construct the Ti-GNRs-Ti device in ATK?

i need your help!

thanks very much!

[Umberto Martinez]

Indeed, in your reference they do not specify the form.
I presume is the alpha one which is the most stable at standard conditions.

they do specify the surface plane Ti(0001).

[njuxyh]

Yes, the Titanium electrode is built following (0001) cristal orientation,
but another question is the electrode is Titanium cluster,

i know in ATK, the electrode can be a period system, so i am wondering whether the cluster can be dealt correcltly?


i need your help!

thanks very much!   

[Umberto Martinez]

yes, you can build a cluster from your system (which definitely is NOT a device).
In this case just use a MoleculeConfiguration and calculate the density of states as in your reference.

However, you can easily build a proper device (with electrodes) and calculate all the relevant properties.
see for example http://quantumwise.com/publications/tutorials/builder/174-graphene-nickel-interface

So,  I do not see any reason to use the cluster approximation above in ATK.

[njuxyh]

Thank you

in fact, i don not understand why the author used the Ti cluster as two electrodes also.

[Anders Blom]

That's because they used a code which cannot handle proper electrodes (with open boundary conditions). So it's a highly approximate treatment, compared to ATK which would be more correct from the perspective of the transport model, although NWChem is very accurate to estimate the band gap, but there are ways in ATK to do that too, e.g. using LDA+U, and extended Huckel model, or MGGA.

[njuxyh]

Yes, the cluster model can not be a practice device.

i know. 

thank your for your reply.

[Umberto Martinez]

Only for an hexagonal crystals you can specify 4 Miller indices h,k,i,l
the third one, i, being fixed by the constrain h + k + i = 0

Titanium alpha has an hexagonal lattice.

[Anders Blom]

Good point, we can improve the cleaver to reflect this on the stash item name after the cleaving.