(201h) Efficient Luminescence From Silicon Nanocrystals: Role of the Plasma Afterglow

The pursuit of efficient and sustainable lighting technologies has led to great interest in the optical characteristics of nanoscale silicon. Silicon is abundant and largely nontoxic, and silicon nanocrystals (SiNCs) exhibit efficient and tunable light emission that can be used in the fabrication of visible and IR light-emitting devices (LEDs)¹, as well as in other applications.  In order to maximize the performance of SiNCs in LED structures, it is important to study and understand the parameters of SiNC synthesis and processing that lead to optimal light-emission.

For synthesis of high-quality SiNCs, the nonthermal plasma reactor provides excellent reliability and control. Following a surface functionalization step, these SiNCs have been shown to exhibit photoluminescence (PL) quantum yields (QYs) of greater than 60%².  The current study examines the effects of the plasma afterglow on SiNC PL efficiency.  We held constant the flowrates of argon and silane (5% bal. He), the reactor pressure, and rf power.  The only parameter changed was the injection of gas through a downstream sidearm of the reactor into the plasma afterglow. We chose hydrogen, helium, and argon as sidearm injection gases, keeping the partial pressure of injected gas constant. 

Using FTIR analysis, we saw a change in the Si-H_x stretch between the H₂, He, and Ar samples, with decreasing contributions from Si-H₃, respectively.  This indicates a change in surface based on gas injection.  For PL measurements, we surface-functionalized the SiNCs in a thermal hydrosilylation reaction, exchanging some of the surface hydrogen with dodecyl ligands.  This improves PL QY and allows clear colloidal suspsensions of SiNCs in nonpolar solvents to be formed.  When comparing the PL QYs of these samples, we find a surprisingly strong dependence of the PL QY on the type of the injected gas.  Argon injection leads to the lowest QY of only ~12%, while hydrogen leads to the highest QY with almost 55%.  Helium injection yields QYs between these two extremes.

After synthesis, when a SiNC passes through the plasma afterglow, interactions with energetic gas species could cause desorption of H from the SiNC surface.  One possibility is that sidearm gas has a quenching effect on this H desorption.   H₂ and He have similar thermal conductivities, while Ar has a lower thermal conductivity and also easily absorbs energy through production of excited states.  This hypothesis may explain why argon sidearm gas leads to the poorest PL performance. 

However, hydrogen also has the special capability to further passivate the SiNC surface in the plasma effluent.  To test the possibility of injected hydrogen reacting with the SiNC surface, we  injected deuterium into the afterglow.  Chemically, deuterium and hydrogen should be have alike—however, if the deuterium binds to the SiNC surface, we should be able to distinguish Si-H_x peaks from Si-D_x peaks in FTIR.  We saw that the hydride-covered surface we see for other injection schemes is almost fully exchanged for a deuterated surface.  Quantum yields for SiNCs produced with deuterium and hydrogen injection are comparable.  This result demonstrates that hydrogen injection into the plasma afterglow leads to an exchange or replacement of surface species, and may create a more defect-free surface, resulting in PL enhancement.  We confirmed the reduced defect density of hydrogen-injection SiNCs compared to argon-injection SiNCs using EPR analysis.

In conclusion, we have shown that the injection of gas into the afterglow of the plasma leads to measurable changes in PL performance of the nanocrystals.  We hypothesize that hydrogen injection yields SiNCs with the most efficient photoluminescence by a combination of quenching of surface hydrogen desorption and additional hydrogen passivation of the SiNC surfaces.   The results of this work will hopefully enable more groups to achieve high-QY SiNCs in aerosol or plasma reactions, and possibly allow better control of SiNC luminescence efficiency, leading to higher-performance light-emission technologies such as LEDs.

Acknowledgements: This work was primarily supported by NSF through the MRSEC grant DMR-0819885.

References:

1. Cheng, Anthony, Kortshagen, and Holmes. Nano Letters 11(articles ASAP) 2011.

2. Jurbergs, Rogojina, Mangolini, and Kortshagen.  Applied Physics Letters, 88 (233116) 2006.