Experiments locate light source in silicon nanocrystals

29 Feb 2008

A long-standing controversy surrounding the source of light emission in silicon nanostructures has been settled by new experiments in high magnetic fields.

Silicon nanostructure

Ever since porous silicon was found to emit light in 1990, scientists have debated the origins of photoluminescence from silicon nanostructures. Now, Manus Hayne of the University of Lancaster in the UK, along with a number of co-workers from across Europe, have concluded that most of the emitted light originates from defects in the nanocrystals, contradicting earlier studies that attributed light emission to quantum confinement effects (Nature Nanotechnology doi:10.1038/nnano.2008.7).

According to Hayne, understanding the underlying mechanisms for light emission is crucial for researchers who are attempting to make efficient light-emitting devices from silicon. Bulk silicon's indirect bandgap means that it emits almost no light, but efficient photoluminescence has been observed from different silicon nanostructures.

"Even from silicon nanocrystals the luminescence is not very bright, compared to nanostructures made from compound semiconductors, for example," Hayne told optics.org. "It is very important to understand where the light is coming from if you want to improve the efficiency."

But under normal circumstances the light emission from defects and from quantum confinement effects look very similar, which makes it hard to distinguish between the two. The solution adopted by Hayne and co-workers was to make their photoluminescence measurements in a strong magnetic field.

"The magnetic field has a fundamental length-scale associated with it, called the magnetic length, which decreases with the square root of the applied field," explained Hayne. "When a magnetic field is applied it squeezes the wavefunction of the state responsible for the luminescence by an amount that depends on the relative size of the wavefunction and the magnetic length, increasing the energy."

Light emission from a highly localized state, such as a defect, is largely unaffected by the applied field, while the photoluminescence from a quantum confined state – which has a similar size to the nanocrystal – is shifted to a slightly higher energy. Because the nanocrystals are so small, however, a very large magnetic field is needed to see the resulting differences in the photoluminescence spectra.

Experimental set-up

Hayne and colleagues used a pulsed magnetic field of up to 50 T. "Even at 50 T the energy shift of luminescence for quantum confinement was only expected to be about 0.5% of the full-width at half maximum of the luminescence line," said Hayne. The relatively weak light emission from the nanocrystals also dictated the use of an extremely sensitive CCD detector, while a significant background signal from the optical fibres used in the experiment also had to be removed to ensure accuracy.

In the experiments, the team first used electron spin resonance to detect the presence of paramagnetic defects in the sample, and then measured the photoluminescence in the applied magnetic field. "When large numbers of defects were present we saw no shift of the luminescence in magnetic field, implying that the luminescence was of defect origin," said Hayne. "But when we removed the defects we saw a very clear shift in magnetic field, exactly as we had hoped."

What's more, the researchers could switch between the two regimes by either eliminating the defects by passivating them with hydrogen, or re-introducing them by irradiating the nanocrystals with UV light, which removed the hydrogen. "We showed that both quantum confinement and defects can be responsible for the luminescence, but that the defect luminescence dominates," said Hayne.

So does Hayne think that these experiments will put a stop to the controversy? "We think that using a magnetic field is a pretty definitive test, but the answer could vary from sample to sample," he said. "It is still possible that in other work quantum confinement was the dominant mechanism. Indeed we studied several samples, a minority of which seemed to have fewer defects and showed quantum confinement in their virgin state."