Tal Group
SINGLE MOLECULE CONDUCTORS
Home Research People Publications Contact Gallery

Simulation of a Pt/benzene molecular junction by J. C. Cuevas

The field of molecular electronics was developed following a nice idea of two researchers (Aviram and Ratner).  They suggested that  a  single-molecule  can serve  as an ultimately  small electronic-device, a diode,  while the  function of  the device (may it  be a diode, memory switch or transistor) can be controlled by the structure of the molecule.

By  using  the  rich  structural  variety  of  organic  and  inorganic  molecules  we  are looking for  the relation between the structure of a suspended molecule between two metallic electrodes and the properties of electronic transport through this structure.

From  the point of  view of  chemical physics, molecular junctions provide the opportunity to learn about the  structure  and  dynamics of molecules  not  as isolated  entities but as a part of an open system  ( an  electronic   circuit ) out  of   equilibrium.  For   example,  the  interaction  of   molecular vibrations with the electrons that cross the molecule is a major issue in this context.

In a general perspective  we use the  molecular junctions as a test-bed for  exploring new ways for manipulating  electronic  current at  the atomic scale. Molecules are good candidates for this task thanks to the structural diversity offered by chemical synthesis as well as their unique mechanical and electronic properties.
Mechanically Controllable Break Junction (MCBJ) Technique
We use the mechanically-controllable break-junction technique which works in the following  way. A metal wire with a notch is attached to a flexible substrate. Then a chamber around the sample is pumped and  the system  is cooled  by liquid helium to ~4.2 K. At  this stage most  of the unwanted contaminations freeze on the chamber surface (cryogenic vacuum).

After cooling, the wire  is broken at  the notch by bending the substrate and we  have two  freshly exposed  electrodes  in  ultra-clean  conditions.  At  this  point molecules can  be introduced to the metallic junction.

Thanks to the bending mechanism, the distance between the electrodes can be controlled in a sub-atomic accuracy, and we can adjust the gap to have a molecule size. In case that a molecular bridge is formed it is possible to  stabilize it or even  stretch the molecule in a controlled way (e.g. D. Djukic et al. PRB 71. 161402 (2005), O. Tal et al. PRB 80, 085427 (2009) ).

The two metallic electrodes can be squeezed towards each other by relaxing the bent substrate. If this  procedure  is  done  properly a  new molecular junction  can be  formed  by  breaking  again the contact  between  the electrodes. Repeating  this procedure many times when measuring electron transport   properties   allows  us  to collect  statistical   data  on  hundreds  or  even   thousands  of different molecular junctions.


Mechanically Controllable Break Junction


Gold Conductance traces  


Conductance Histogram of an Au junction
Conductance Traces
Conductance traces are obtained by measuring the junction conductance  while slowly increasing the electrode distance. For example, several conductance traces of a gold junction are presented on  the  left  figure.  A gold atom  is  known  to  have a conductance of  1G0  where  G0= 2e2/h  is the conductance  of  a fully open  single  conductance  channel. For a  gold  atom suspended between two gold  electrodes the current is carried  by a single conductance  channel associated with the s valence orbital of the  gold atom. The abrupt  jumps to lower conductance values observed in the top figure to the  left appear when  the number of  gold atoms in the narrowest part of the wire is reduced during  stretching. The long plateau in the end of the traces (near 1G0) takes place when the wire cross-section is reduced to a single atom and a chain of gold atoms is formed.

As we  break and reform  the junction, the conductance traces are  varied. The high stability o f the MCBJ  allows  us to measure  thousands  of such  traces and  use  them  to construct  conductance histograms  that  indicate  the number  of  times  that  a certain conductance  value was measured. These   histograms  can   tell  us  what   the  most    probable  conductance   values  are  ( e.g.  when stretching  the junction). For  example, the fingerprint  of a  clean  platinum junction is a histogram with a  main peak at 1.6 G0 and  low contribution  between  this  peak and  the tunneling tail at low conductance (left figure, solid curve). Here, unlike in the case of gold, contributions from both the s and d valence  orbitals  lead  to  total  conductance  which  is  higher  than  1G0 for  the narrowest metallic constriction.

When  a  molecule  is  introduced  between  the  electrodes  the   conductance  histogram  changes substantially,   allowing   us  to  identify   the  event   and  learn   about  the   typical   most-probable conductance  of  the  molecular  junction.  For  example, the figure to the left shows the difference between a  conductance  histogram  of a  platinum  junction,  to  the one produced when benzene molecules are introduced (gray).
Inelastic Spectroscopy
During    the    measurements   of   inelastic   spectroscopy  the differential conductance  (dI/dV) is  measured as a function of the applied voltage (V) across a  molecular  junction. Inelastic spectroscopy  provides information  about   inelastic electron transport   processes   that   can   take    place   as  a   result   of excitation  of   molecular   vibrations,  spin   flips,  etc.  The  top figure   on   the  right   provides  an   example   of   a   measured spectrum. In the presented case, when the voltage difference between the metal electrodes reaches a  threshold given  by a vibration mode energy () the electrons have enough energy to excite  the  molecular  vibration  mode.  At  this  voltage (eV = ) a step in the differential conductance appears due to the change  in  the  overall  transmission probability (e.g. at 40 mV in  the   right  figure ).   The  interpretation   of  the  step   in  the differential  conductance  signal  in  terms of inelastic changes induced  by  a molecular  vibration  mode can be  confirmed by measurements of  isotopes derivatives.  The change in mass is expressed as a shift in vibration energy.

For each type of molecular junction we can  collect many dI/dV spectra  where each  measurement  is  performed  on& nbsp; a newly    formed   junction.   The   figure   on   the   bottom   right presents  the number  of   times  that   a vibration   mode   with  certain   energy   was    detected    for   four   different   types  of molecular junctions. The  unique  distribution  of  the  vibration-mode  energies  for the   four   cases    reveals   an   individual   signature   for   each molecular junction.


Differential conductance (benzene on Pt)
M. Kigushi et al. PRL 101, 046801 (2008)


dI/dV Peak Histograms for different molecules
O. Tal et al. PRB 80, 085427 (2009)


                                       Length histogram of an Au junction

Length Histograms
When breaking a metal  junction of platinum,  iridium or gold, chains  of  single   atoms   can   be   formed   before  complete   detachment.  These     chains can be observed in length histograms. Such histograms present         the   number  of   times   that  a   certain   plateau   length  appears  in  an ensemble  of   conductance   traces.  Here  the  peaks   in  the  left  figure represent  chains  of 1,2 and 3 gold  atoms.  Length  histograms and the combination   of  length   and   conductance   histograms   can  teach  us  about the structure of molecular junctions.

Shot Noise Measurements
Shot-noise  measurements   are  used to  analyze the  electronic transmission   channels   available   for   electrons   to   cross   an atomic or molecular junction. So by measuring the level of shot noise,  one  can  gain  insight  into electronic properties that are not accessible by standard conductance measurements.

Current shot noise results from  time-dependent fluctuations in the electrical current that take place because the current  is not smooth  but  it is a  flow of  discrete  electrons that can either be transmitted or backscattered. The noise level of a point contact ( e.g.   atomic   or   molecular   junction)    is   determined   by   the number of transmission channels in which  electrons  can   cross the  junction  and  their  transmission  probabilities.  In  the  right figure  three  sets  of  noise  measurements  vs.  current  bias are presented  together  with  a  fitting  to  the  expression for noise power  for  an  arbitrary  set  of  transmission  probabilities. Each set was measured on a different Pt/water junction.

Since shot noise provides information about the decomposition of  the  total  conductance  in  terms  of   individual  transmission channels it  can help  us to  determine  the molecule orientation, the  nature  of  its   binding  to   the  electrodes  and  can  be  very useful in relating theory to measurements.

 

 

 



Shot noise measurements of a Pt/H2O molecular junction
O. Tal et al. PRL 100, 196804 (2008)