AbstractIn this thesis we aim to develop further understanding of the mechanisms leading to the nonlocal manipulation of adsorbed aromatic molecules on a Si(111)-7 × 7 surface. The injection of either electrons or holes into the surface above a bias threshold results in the manipulation of molecules up to tens of nanometers away from the injection site. We use the tip of a scanning tunnelling microscope (STM) as both a tunable source of hot charge carriers, as well as an imaging tool with atomic-scale resolution. While far slower than the timescale of the manipulation process at room temperature, this allows for observation of the distribution of aromatic molecules on the surface before and after nonlocal charge injection.
The nonlocal manipulation process can be separated into three distinct steps, each of which occurs identically and independently for each individual charge carrier; charge
injection, charge transport, and molecular manipulation. In this thesis we focus on the charge transport step, the link between the nonlocal process and light emission from
the same system, and the manipulation step itself.
To isolate the charge transport process, the radial extent of nonlocal manipulation has been measured at two injection biases and across a range of charge injection times
from 1 to 500 s (i.e., varying the number of injected charge carriers). For sufficient injection duration, the manipulated spot-size is described by purely diffusive transport
of the injected charge carriers. This sets a practical limit on the maximum size of the nonlocal effect, which is constrained by the surface diffusive properties and feasible
injection duration. Conversely, at the lower limit of injection times, the radius of manipulation is observed to plateau with a radius between 6 and 10 nm, conforming to
an initial non-manipulative ballistic transport regime prior to the diffusive transport.
Secondly, we consider similarities in the onset biases between nonlocal manipulation and light emission from the same system, alongside constant emission spectra, which
suggest a common origin for the two mechanisms. We demonstrate that the nonlocal manipulation, and hence photon emission, is not mediated by a localised surface plasmon decay process, and instead that light emission occurs after the charge transport step of nonlocal manipulation. The probability of both outcomes follows a similar exponential increase with increasing bias above the shared threshold, between +2.0 and +2.7 V, suggesting that both processes follow an identical pathway prior to manipulative relaxation or light emission.
To observe the final manipulation step, we consider the outcome, rather than only the distribution, of nonlocal manipulation. Previous investigation into the outcome of local manipulation demonstrated a bias invariance of the final manipulation step. Here we expand this to nonlocal manipulation, thus demonstrating that the molecular manipulation occurs from a common energy state for each injected charge carrier, independent of the injection bias above or below the nonlocal threshold. For electron injection, the nonlocal branching ratio is observed to be a similar order of magnitude to the local branching ratio. Additionally, we observe a similar probability of manipulation per charge-molecule interaction across both local and nonlocal manipulation. This suggests that the manipulation step occurs identically for both local and nonlocal manipulation, with only the addition of a charge transport step in nonlocal manipulation altering the outcome.
Finally, we discuss the effect of varying the adsorbate molecule on the manipulation step. While changing the adsorbate from a lighter (toluene) to a heavier (bromobenzene) aromatic molecule has no effect on the manipulation for electron injection, for hole injection the relative probability of desorption is suppressed. This is in agreement with the differing manipulation processes between the injected charges; surface-mediated electron injection proceeds mostly independent of the adsorbate molecule, while molecule-mediated hole injection is dependent on the mass of the adsorbate.
|Date of Award||26 May 2021|
|Supervisor||Peter Sloan (Supervisor) & Peter Mosley (Supervisor)|