Oxidized porous silicon planar waveguides'

1
Experimental results and modelling of gain in
ErSi-NC waveguides
Nicola Daldosso Nanoscienze Lab, Università di
Trento, Italy
D. Navarro-Urrios, L. Pavesi NL-Department of
Physics, Università di Trento Italy F.
Gourbilleau, R. Rizk SIFCOM, UMR CNRS 6176,
ENSICAEN - France
2
Outline

• Introduction
• Rate equations
• Emission and absorption cross sections
• Lifetime and up-conversion coefficients
• Excitation cross section
• Modelling
• Signal enhancement measurements

3
Er3 coupled Si-NC

• Higher index contrast for light confinement
• Broad band absorption (UV-VIS)
• Most efficient excitation for Er3 sexc from
  10-21 to 10-16-10-18 cm2
• Fast ( 1ms) and efficient (?) energy transfer
  from Si-NC to Er3
• Possibility of electrical injection (? CMOS
  compatibility)

4
Gain favouring

• As much Er3 as possible
• Enough Si-NC to excite all the Er3 ions
• Good modal confinement
• Low waveguide losses

5
Is gain (net) achievable in such systems ?
Is it possible to efficiently coupled Si-NC to
Er3 ?
If yes, how ?
6
Introduction
4I11/2, 4I9/2
4I13/2
4I15/2
Er3
Nexc
Excitons
Steady state
Nexc density of excitons NNC density of Si-NC
?NC absorption cross section ? intrinsic
lifetime of the exciton kt average coupling
rate Cind percentage of Er3 coupled
to Si-NC
For low F (ltlt1020 ph/cm2s)
7
Introduction
4I11/2, 4I9/2
4I13/2
4I15/2
8
The samples
ErSi-nc produced by Reactive Magnetron
co-Sputtering and successive annealing to get
phase separation and reduction of non radiative
defects
Annealing time
F. Gourbilleau et al., JAP, 94, 3869 (2003) JAP
95, 3717 (2004).
9
Absorption Losses
10
Determination of sabs and sem
Mc Cumber relation
From transmission measurements
sabs and sem
Daldosso et al., APL 88, 161901 (2006)
11
Lifetime and up-conversion
Quantitative measurements of the photon flux
emitted from the samples It is so possible to
correlate the number of emitted photons with N2
td and Cup
12
Excitation cross section
?exc is orders of magnitude higher than that of
Er3 in pure silica (10-21 cm2), for samples B
and C, resonant (488 nm) and non-resonant (476
nm) result in the same ?exc10-16 cm2
but seems to be flux dependent, the slope is
changing with increasing pump flux ?
sexc
13
Fraction of excited Er3 vs pump flux
We get all the parameters for the rate equation
sabs sem
td Cup
sexc
We have the experimental data
but
Photon flux (ph/cm2s)
14
Modelling
Er3 ions near the Si-NC are efficiently coupled
to them, whereas Er3 ions far away behave as
Er3 in SiO2 (or in SiOx according to the
annealing time) that can be excited only
directly.
Model for sexc
We consider that the first Er to be excited and
therefore the strongest coupled would be the
closest to the Si-Nc The coupling diminishes
with the distance It can be defined like shells
of probability around each nanocluster
is then solved for each R Thus, by integrating
over all the shells, we get the temporal
dependence of the total excited state population.
Garrido et al., APL in print
15
Simulations
td3.8 ms, Cup2x10-17cm3s-1, so3x10-16cm2,
sd5x10-21cm2, Rnc4nm , Ro0.5nm,
NNC1x1017cm-3.
And this means that only 2-3 of the whole erbium
population can be excited trough transfer from
Si-NC. In any case it is about 10-100 excitable
Er3 per Si-NC
16
Signal enhancement
SEgt1
SE?1
Probe
17
Signal enhancement
but only 2-3 is excited via Si-NC
18
Conclusions

• We have measured
• Absorption and emission cross section
• Total lifetimes and cooperative up-conversion
  coefficients
• Excitation cross section
• Values of internal gain using a pump and probe
  technique

• But due to the low of Er3 excitable through
  Si-NC
• not yet net optical gain
• although numbers show that it is possible
• by reducing carrier absorption
• by increasing the coupling efficiency

in progress
19
Thank You