Hello,

I am doing SHG experiment using a fs-laser (140fs pulse width) on a probe
on silicium substrate.
As a preliminary test, I measured a reflection SHG spectrum from a bare
silicium wafer, expecting to get no signal. However I get two peaks
separated by 15nm, typically like that:

impinging Fundamental: 1300nm
outgoing SHG: 650nm
secondary peak: 665nm

impinging Fundamental: 1400nm
outgoing SHG: 700nm
secondary peak: 715nm

I made various tests, and the result is that all these come definitively
from the probe. But at these wavelength, there should be no suitable
energy levels for SHG.

Do you have any idea where the peaks can come from? From the break of
symmetry at the surface? How?

Best regards,

Mat'

Surface second harmonic generation has been studied since at least
the > late 1980's. The surface breaks the inversion symmetry and makes

The secondary peak might be due to a Raman shift, though the
wavelengths that you give don't exactly correspond to the same Raman
shift; the 665 nm peak corresponds to a 348 cm^-1 while the 715 nm
peak corresponds to a 298 cm^-1 shift. If the accuracy of the
wavelengths is not good to better than ±1 nm then the data might
be > consistent with a Raman process.

Hello,

Thank you for this reply!

Surface second harmonic generation has been studied since at least
the > late 1980's. The surface breaks the inversion symmetry and makes it
allowed.

Do you know exactly how? I am new in the field, and I do not understand
how this can happen when the band structure do not match?!

The secondary peak might be due to a Raman shift, though the
wavelengths that you give don't exactly correspond to the same Raman
shift; the 665 nm peak corresponds to a 348 cm^-1 while the 715 nm > peak
corresponds to a 298 cm^-1 shift. If the accuracy of the > wavelengths is
not good to better than ±1 nm then the data might be > consistent with a
Raman process.

The accuracy of my measurement in within this range.
How can I check the Raman character of the peak?

Best regards,

Mathieu

A Raman peak will tune in frequency with the pump frequency ( in your case
as twice the pump frequency) to keep a constant energy difference between
the pump and Raman shifted photons.

If I recall correctly YR Shen at the University of California at Berkeley
was one of the pioneers in surface second harmonic generation. Google
"surface second harmonic generation" and you will get many hits including
one in Wikipedia.

Bret Cannon