CORDIS - EU research results

Dopant-surface interactions in silicon nanoclusters

Final Report Summary - SINANOTUNE (Dopant-surface interactions in silicon nanoclusters)

The use of silicon nanocrystals (Si-NCs) is one of the most promising strategies for light generation, capture and amplification. The reduced dimensionality of the Si-NC structures gives us access to certain physical properties which are inhibited or frustrated in human-size materials (band-folding, quantum-confinement, multiple-exciton generation, etc.). However, the processes used for engineering the electronic structure of silicon nanoparticles have to be different from those used for bulk silicon. Namely, impurity doping of nanoparticles suffers from difficulties not present in the bulk material, such as quantum confinement, screening by image charges and segregation.

In the SINANOTUNE project (see online), first-principle calculations were used to understand the energetics, ionisation energies and diffusion of group-III and group-V dopants in Si-NCs, showing how they differ from bulk silicon, and how do they depend on the surface chemistry. Further, two alternative strategies to enhance or replace traditional doping were presented: surface-driven offset engineering and surface doping with an organic molecule. Finally, the structure of Si-NCs / polymer interfaces for bulk heterojunction (BHJ) solar cells was investigated to understand electron transfer effects. The main conclusions of the project are highlighted below.

(1) Doping with group-III and group-V impurities

While previous density-functional theory studies have brought insight into some of the phenomena hindering the doping efficiency, most studies have focused on hydrogen-terminated nanoparticles, in which the segregation behaviour is different from the oxidised silicon nanoparticles. One of the most relevant results of this project was to clarify what is the preferred location of the two most important dopants, B and P, in oxidised Si-NPs or silicon nanoparticles embedded in a silicon oxide matrix, and clarify why they differ from hydrogen-passivated NPs. Simulations of the relative energies of B and P in realistic models of partially oxidised Si-NPs have shown that positively charged P is most stable in the Si rich region, while B is predicted to have no energetic preference for the Si or SiO2 regions. This is in agreement with the segregation behaviour observed experimentally for Si nanocrystals embedded in an SiO2 matrix (1). Presumably, during growth phosphorus has enough thermal energy to diffuse to the most favourable dopant locations, and once trapped at the surface, release hydrogen forming a stable electrically inactive defect. While boron may also diffuse, its three-fold coordinated structure at the surface is not stable enough, suffering a strong competition from four-fold coordinated sub-surface structures.

In contrast, in H-terminated Si-NPs, the formation energy of P+ is nearly independent on the lattice position it occupies in the nanocrystal, provided that it binds to four Si neighbours. For nanocrystals 1.2 nm in diameter, the energy changes less than 0.1 eV with the lattice position. The formation energy of four-fold coordinated B- decreases slightly towards the surface (less than 0.5 eV for nanocrystals 1.2-nm diameter). Neither P+ nor B- is stable at the surface with coordination four. This is similar to what has been found for silicon nanowires with H-terminated surfaces (2). However, at the surface both B and P are stabilised by losing one hydrogen, becoming three-fold coordinated. For the same hydrogen chemical potential, the energy advantage of the three-fold coordinated surface configurations is greater for phosphorus.

(2) Alternative doping strategies

As a consequence of the research reported in the previous section, it was possible to suggest innovative strategies which can be used to change the electronic properties of small Si-NCs in conjunction with, or in replacement to, traditional doping with group-III or group-V elements. It is anticipated that a possible route is based on surface doping (remote doping). This was proposed theoretically using an organic oxidising agent, F4-TCNQ (7,7,8,8-tetracyano-2,3,5,6-tetrafluoroquinodimethane). The proximity of the lowest unoccupied level of F4-TCNQ to the highest occupied level of the Si-NPs promotes the formation of an empty hybrid state overlapping both nanocrystal and molecule, reducing the excitation energy of the combined system to about 0.8 - 1 eV in vacuum. For coverage of three molecules per nanocrystal, transfer of one electron effectively takes place.

Another possibility is to shift the electronic energy levels of the nanocrystals up or down in the energy scale by using a suitable surface termination. For example, chlorinated nanocrystals have higher electron affinity, higher ionisation energy and lower optical absorption energy threshold. This can be used in the design of heterojunctions (for example, hybrid nanocrystal-polymer solar cells) to modify the alignment with the bands of the other material.

(3) Optimisation of Si-NC-polymer blends for BHJ solar cells

Another highlight has been the theoretical investigation on the injection time for electron transfer in oligo / polythiophene / Si-NP blends, a class of hybrid materials which are currently of interest for BHJ solar cells. We found that the injection time for transfer of an electron from photoexcited dodecathiophene and polythiophene to a Si-NC (2.2-nm diameter), calculated by computing the retarded Green's function for the system from the Hamiltonian and Kohn-Sham states produced by density functional calculations, can be of the order of 10-100 fs if the thiophene chain lies approximately parallel to the silicon surface. However, the electron injection time is 1 - 2 orders of magnitude longer if the oligothiophene chain lies perpendicular to the silicon surface. Chemisorption interaction between the thiophene chain and the nanocrystal provides a relatively small improvement (decrease) of injection times, much weaker than that achieved by enforcing the parallel arrangement of the chain with respect to the nanocrystal.

(4) Impact

The results obtained in the SINANOTUNE project have been published in several prestigious peer-reviewed publications, presented in international conferences and shown in seminars and informal meetings at collaborating institutions. The contribution of the funding of the Marie Curie programme to the expenses of the researcher was crucial for the participation in so many and varied dissemination activities, and resulted in a visibility that would be difficult otherwise. The first target group was experimentalists in the same field, who can make use of the theoretical results to guide further experiments to develop material technologies suitable for subsequent use by industry. The research related to the SINANOTUNE project was also presented in simplified terms to the local community, primarily as a result of the publicity obtained by the Gulbenkian Foundation prize. An effort was made to improve the awareness of the people in general for the importance of performing research in physics and materials science and how it can lead to the development of practical applications that can be used in everyday life.

(5) References

(1) Michele Perego, Caroline Bonafos, and Marco Fanciulli, Nanotechnology 21, 25602 (2010).
(2) Hartwin Peelaers, Bart Partoens, and Francois M. Peeters, Nano Lett. 6, 2781 (2006).