Boron-Incorporating Silicon Nanocrystals Embedded in SiO2: Absence of Free Carriers vs. B-Induced Defects

Abstract

Boron (B) doping of silicon nanocrystals requires the incorporation of a B-atom on a lattice site of the quantum dot and its ionization at room temperature. In case of successful B-doping the majority carriers (holes) should quench the photoluminescence of Si nanocrystals via non-radiative Auger recombination. In addition, the holes should allow for a non-transient electrical current. However, on the bottom end of the nanoscale, both substitutional incorporation and ionization are subject to significant increase in their respective energies due to confinement and size effects. Nevertheless, successful B-doping of Si nanocrystals was reported for certain structural conditions. Here, we investigate B-doping for small, well-dispersed Si nanocrystals with low and moderate B-concentrations. While small amounts of B-atoms are incorporated into these nanocrystals, they hardly affect their optical or electrical properties. If the B-concentration exceeds ~1 at%, the luminescence quantum yield is significantly quenched, whereas electrical measurements do not reveal free carriers. This observation suggests a photoluminescence quenching mechanism based on B-induced defect states. By means of density functional theory calculations, we prove that B creates multiple states in the bandgap of Si and SiO2. We conclude that non-percolated ultra-small Si nanocrystals cannot be efficiently B-doped.

Correlated Structural and Luminescence Analysis of B‐Doped Si‐Nanocrystals Embedded in Silica

“…[9,18,19] Moreover, it confirms the recent experimental investigation of Hiller et al based on a PECVD elaboration. [11] It is also consistent with recent work based on other investigation method. In fact, by combining X-ray diffraction experiments with Monte Carlo simulation, Hunter et al have evidenced the location of B at the Si-ncs surface in Si-ncs of barely 5 nm of diameter doped with 2.5-10% B.…”

“…Then, the samples were implanted with 29 Si at 50 keV at a dose of 6 Â 10 16 cm À2 . For the B-doped sample, to match with the projection range of 29 Si, 11 B was implanted at 17 keV at a dose of 5 Â 10 15 cm À2 . Finally, the thin films were annealed at 1100 C for 4 h in pure N 2 to form Si-ncs and for impurities diffusion.…”

“…[10] By consequence, from an experimental point of view, Hiller et al have shown that B doping of Si-ncs can be hardly performed by plasma-enhanced chemical vapor deposition (PECVD). [11] In this case, only a minority of Si-ncs are doped and contain less than two impurities that may be located at the Si/SiO 2 interfaces. In this article, we propose to investigate the compositions of B-doped Si-ncs embedded in SiO 2 thin layer, elaborated by ion beam synthesis.…”

“…Both of these effects on PL luminescence of B-doped Si-ncs agree with the observations of other authors whether for the case of freestanding Si-ncs or Si-ncs embedded in SiO 2 matrix. [11,[22][23][24][25][26] APT analyses evidenced that B atoms are only detected in the biggest Si-ncs. As the presence of a single B atom may lead to the deactivation of the Si-nc, we can assume that only the smallest Si-ncs exhibit a PL signal.…”

Absence of free carriers in silicon nanocrystals grown from phosphorus- and boron-doped silicon-rich oxide and oxynitride

“…The overall P-distribution corresponds very well to previously observed trends for P in Si NCs [ 23 – 26 ]. We note that the ≈20% O-concentration in the NC-interior is an artefact from local magnification effects (LME) [ 27 – 28 ] which is generally observed in this material system [ 24 – 26 29 – 30 ]. Inevitably, this artefact also influences the exact values of the P-concentration, but since both samples are subject to the same LME the comparison discussed above is not influenced.…”