Access Official Technical PDF

Share Summary:LinkedInEmail

Silicon quantum dot LEDs offer a route to light-emitting devices built from silicon-compatible materials, but their emission efficiency remains a central engineering challenge. One approach is to place metal nanoparticles near the light-emitting quantum dots. A study of layered silicon nitride devices found that gold nanoparticles (AuNPs) produced the strongest electroluminescence improvement when separated from silicon quantum dots (SiQDs) by a thin, non-radiative silicon nitride spacer. The result points to the importance of layer design—not simply adding a plasmonic material.

How the layered silicon quantum dot LEDs were designed

The devices used a metal-insulator-semiconductor structure on a p-type silicon substrate. Silicon-rich silicon nitride contained the SiQDs that served as the radiative material. A transparent conductive aluminum-doped zinc oxide contact was deposited above this active layer, while aluminum formed the bottom electrode.

Researchers compared four configurations to distinguish the effects of the gold and spacer layers. Two reference devices contained no AuNPs: one placed the SiQD film directly on the silicon substrate, and the other included a 10 nm non-radiative silicon nitride (SiNx) layer beneath the SiQDs. The corresponding gold-containing devices placed AuNPs at the substrate interface, either directly below the SiQDs or below the 10 nm SiNx spacer.

The SiQDs and non-radiative SiNx had different compositions. The silicon-rich film supported quantum dot formation, while the higher-nitrogen SiNx film had a wider optical band gap and acted as an insulating spacer. Both types of silicon nitride were deposited by remote plasma-enhanced chemical vapor deposition. The AuNPs were deposited by sputtering.

Why the spacer changed electroluminescence

Gold nanoparticles can support localized surface plasmon resonances (LSPRs), collective oscillations of electrons that interact with nearby optical emitters. In a properly arranged structure, this interaction may affect light emission. But proximity alone is not enough: the spacing and properties of the intervening dielectric can influence whether the metal enhances or suppresses emission.

The device without a spacer showed only a modest increase. At a similar injected current, its integrated electroluminescence (EL) was 1.14 times that of the corresponding reference without gold. By contrast, the configuration with AuNPs beneath a 10 nm SiNx spacer achieved an integrated EL enhancement of 4.7 at approximately 27.7 mA. Its EL turn-on voltage was 11 V, compared with 14 V for the matching reference device.

The spectral comparison also showed that the enhancement depended on wavelength. For the spacer-containing device, the EL enhancement factor peaked at 7 around 510 nm and was 5.4 near the main emission peak at approximately 600 nm. The strongest spectral enhancement was close to the reported gold nanoparticle absorption peak near 538 nm, consistent with a possible plasmonic contribution. It does not, by itself, establish that plasmonic coupling is the only mechanism.

Carrier transport is part of the performance story

The electrical measurements suggest that AuNPs may influence more than the radiative process. In the spacer-containing structure, current injection was higher for the gold-containing device under certain bias conditions, although the reference device conducted more current at lower electric fields. The authors proposed that the nanoparticles and layer interfaces could affect carrier distribution and injection toward the luminescent regions.

This distinction matters when interpreting EL. A brighter device under electrical drive may reflect changes in the efficiency of light generation, the delivery of carriers to emitting centers, or both. The study found that the EL enhancement in the spacer-containing structure was greater than the photoluminescence enhancement previously observed for the same configuration. Because photoluminescence is excited optically while electroluminescence depends on injected carriers, this difference supports the possibility that electrical transport contributes alongside plasmonic effects.

The emission mechanisms also require careful interpretation. The EL peak near 600 nm was red-shifted from the previously measured photoluminescence band near 505 nm. The researchers noted that defect states or interface states in addition to quantum-confined SiQDs could contribute to the longer-wavelength EL. The measured enhancement therefore should not be treated as proof of a single emission pathway.

Implications for thin-film device engineering

For engineers developing silicon-based emitters, the practical lesson is to evaluate the entire stack. The position of the metal nanoparticles, the spacer thickness, the composition of each nitride layer and the resulting electrical behavior all affect device output. Comparing structures with and without a spacer helped separate these effects and identified the AuNP–SiNx–SiQD arrangement as the strongest configuration in this study.

  • Layer order matters: the strongest improvement occurred when a 10 nm non-radiative SiNx layer separated the AuNPs and SiQDs.
  • Electrical and optical data should be considered together: EL intensity, turn-on voltage and current injection provide complementary information.
  • Enhancement can have multiple causes: plasmonic interaction is plausible, but carrier transport and material interfaces may also contribute.

The films were produced using deposition methods described as compatible with established silicon microelectronics processes. That compatibility is relevant to silicon photonics development, although the reported work demonstrates experimental devices rather than a finished manufacturing platform.

Conclusion

In these silicon quantum dot LEDs, gold nanoparticles alone produced limited EL improvement, while adding a thin silicon nitride spacer yielded a substantially larger increase and a lower turn-on voltage than the corresponding reference. The findings show why material placement and carrier behavior must be treated as linked design variables. For device developers, the next step is not simply to maximize nanoparticle loading, but to understand how the complete multilayer structure governs both charge injection and light emission.

Access Official Technical PDF

About The Author