Main Content Region

A bridge between nano- and bio-science

Modelling electron transport at the atomic level

Dr Stefano Sanvito, School of Physics, Trinity College Dublin

Hybrid devices at the boundary between physics and biology are becoming increasingly important over a wide spectrum of applications. These range from high-end revolutionary computer architectures, to disposable electronics, to diagnostic tools for genetically driven medicine, and they will soon change the quality of several activities of our life.

From the electronics side fully functional molecular memories and logic gates have been already demonstrated [1] suggesting a possible roadmap for the post-silicon era. At the same time molecular devices have found a place in multifunctional sensors designing, showing a sensitivity orders of magnitude superior to that achievable with conventional methods. These molecular devices include carbon nanotubes based NO2 and nerve agent detectors [2], nanowires-based virus detectors [3], and nanowire-based chemical sensors [4]. In a near future we can envision on-chip nanolabs able to sense a particular signature of gene or protein expression and therefore able to diagnose various diseases. These will be a formidable tool for systems biology and preventive medicine [5]. Finally some of these devices may offer the possibility of quantum coherence over long time scales, therefore providing a way to the physical implementation of quantum computing [6].

Recent advances of order-N methods [7] and the availability of large-scale computational facilities are changing the landscape of ab initio modeling of complex systems. A few thousand atoms are currently the threshold of accurate electronic structure simulations, but there are good expectations that this boundary will be soon expanded. This is the playground for systems relevant for both nanotechnology and biology.

In the last few years the Dublin team, headed by Dr. Sanvito, in conjunctions with groups from the Universities of Lancaster and Oviedo, has developed a new tool for calculating the I-V characteristics of atomic-scaled devices. This is the code Smeagol [8], which implements the non-equilibrium Green’s function method within density functional theory. The main goal of Smeagol is to achieve the prediction capabilities for nanoscaled devices, that quantum chemistry codes have for structural and electronic properties of molecules.

Smeagol has also been designed with an eye to flexibility and universality, and can tackle extremely different problems. These include transport through H2 molecules [9] and atomic point contacts [10], molecular spin-valves [8] and more recently magnetic tunneling junctions and DNA strands.

Magnetic tunnel junctions are a crucial ingredient for data storage industry, since they are the core element of hard-drive read-heads. One of the challenges for the design is that of preserving large sensitivity with ultra-thin tunnel barriers and therefore small layer resistances. Our Smeagol investigation has established a theoretical limit for the sensitivity of such devices, when the thickness of the tunnel barrier is only two atomic layers. We have demonstrated that large magneto-resistance, i.e. the change in resistance when a magnetic field is applied, persists at this tiny length scales, when the tunnel barrier is defect-free, whereas the signal is largely deteriorated for defective systems. This is the result of the rather different transmission properties of majority and minority electrons in typical transition metals.

Fig.1: Conductance as a function of the k-vector in the Brillouin zone for majority (left) and minority (right) spins for a 4 monolayer thick Fe/MgO(001) tunnel junction. The color code represents the magnitude of the transmission: red=high transmission, blue=low transmission.


Fig. 2: Current distribution across a poly(G)-poly(C) DNA strand sandwiched between gold current-voltage probes. Note that most of the current flows around the DNA backbone and not across the base pairs.

Establishing the conducting state (metallic or insulating) of DNA strands is currently an hot topic in the nano-science community. If demonstrated metallic, DNA can open uncountable possibilities for nano-design. In fact its biological functionalities, such has recorgnition, can be combined with the ability of conducting current. This can pave the way for constructing self-assembled nano-circuits interconnecting other molecular devices. Alternatively it can be used as templating material for the growth of inorganic nanosctructures, thus avoiding complicated and expensive lithographic methods.

Despite these tantalizing prospects, the conducting nature of DNA remain elusive. Experiments show evidence for any type of electronic properties, ranging from metallic to semiconducting, to even superconducting. This calls for first principles transport calculations. We have tackled the problem with Smeagol, which was efficiently optimized for running massive-parallel simulations. In particular we consider poly(G)-poly(C) strands, an artificial sequence of DNA where only two bases are repeated. Our main result is that, although the band-structure suggests metallic conductance through electronic states delocalized across the base pairs, actual transport calculations demonstrate that these are extremely inefficient transport channels. In contrast backbone-related states can conduct current pretty efficiently, although they are positioned several eV below the Fermi level of the current voltage electrodes. This results in typical I-V characteristics with a gap of about 3V around zero bias, independently on the sequecing.


  1. P.J. Kuekes, J.R. Heath, and R.S. Williams, US Patent, number 6128214 (Hewlett-Packard), October 2000; C.P. Collier et al., Science 285, 391 (1999); Y. Huang, X. Duan, Y. Cui, L.J. Lauhon, K.-H. Kim, C.M. Lieber, Science 294, 1313 (2001).
  2. P. Qi et al., Nano Lett. 3, 347 (2003); J.P. Novak et al., Appl. Phys. Lett. 83, 4026 (2003).
  3. F. Patolsky, et al., Proc. Natl. Acad. Sci. USA 101, 14017 (2004).
  4. Y. Cui, W. Qingqiao, P. Hongkun, and C.M. Lieber, Science 293, 1289 (2001).
  5. J.P. Heath, M.E. Phelps and L. Hood, Mol. Im. Biol. 5, 312 (2003).
  6. D.P. DiVincenzo, The physical implementation of quantum computation, in Fort. der Physic. 48, 771 (2000)
  7. P. Ordejon, Comp. Mat. Sci. 12, 157 (1998).
  8. A.R. Rocha, V. Garcia-Suarez, S.W.Bailey, C.J. Lambert, J. Ferrer and S. Sanvito, Nature Mat. 4, 335 (2005); Phys. Rev. B 73, 085414 (2006).
  9. V. Garcia-Suarez, A.R. Rocha, S.W. Bailey, C.J. Lambert, S. Sanvito and J. Ferrer, Phys. Rev. B. 72, 045437 (2005).
  10. V. Garcia-Suarez, A.R. Rocha, S.W. Bailey, C.J. Lambert, S. Sanvito and J. Ferrer, Phys. Rev. Lett. 95, 256804 (2005).