Trimming of silicon ring resonator by electron beam induced compaction

Rating : Rate It:
Post a comment
    Post Comment on Twitter

1 : Trimming of silicon ring resonator by electron beam induced compaction Jonathan Schrauwen, Dries Van Thourhout and Roel Baets ECIO 2008
2 : Silicon Photonics 5 µm 5 µm 25 µm 1.2 µm 500 nm 500 nm
3 : 248 nm excimer laser Lithography 193 nm excimer laser Lithography 1 nm change in linewidth corresponds to 1 nm wavelength change Even with most advanced CMOS mass fabrication tools 1 nm is a challenge!! Deep UV fabrication on CMOS pilot line
4 : Example of problematic device Polarization insensitive filter needs identical ring resonators on different locations
5 : Example of identical rings Not good enough!! s = 8 nm s = 2 nm 248 nm DUV litho 193 nm DUV litho
6 : Outlook About trimming / tuning Trimming experiment: In situ trimming of silicon ring with electron beam Physical mechanisms
7 : Need for trimming / tuning Tuning = time dependent Temperature tuning with heaters: 2 mW/nm Carriers: more complex devices ? high power consumption if many devices are integrated on one chip ? OVERKILL Trimming = irreversible Low index contrast systems: common practice UV or electron beam SiO2 or Si(O)N compaction Refractive index changes of core material Resonance wavelength changes up to 10 nm High index contrast silicon-on-insulator: Silicon can not be compacted SiO2 cladding confinement too low ? stress in Si
8 : Experiment: Equating the resonance of two non-identical resonators In situ experiment: Fibers glued to sample Vertical grating couplers Vacuum fiber feedthroughs 1550 nm light from SLED into chamber Optical Spectrum Analyzer Vacuum fiber Feedthroughs Non-identical racetrack resonators
9 : Experiment: Equating the resonance of two non-identical resonators
10 : In detail 4.91 nm red shift: In three stages Blue shift at start due to carriers (0.1-0.2 nm) Subsequent redshift (idem) Final dose 2.8 x 1023 keV/cm3or 0.157 C/cm2 Equating the resonance of two non-identical resonators Schrauwen et al, Optics Express, 16(6), p.3738-3743 (2008) Resonance wavelength of trimmed ring (nm)
11 : Physical mechanisms: index change?? Irradiation with 2 keV electrons: Penetration in Si and SiO2 about 70 nm Oxide next to waveguide compacts Volume compaction due to ionization events Up to 10% volume fraction Up to 3% refractive index change Silicon does not compact Higher refractive index of the oxide cladding Limited confinement in cladding (< 1.5%) Expected resonance shift < 0.5 nm Too small!!
12 : Physical mechanisms: stress!! Deformation of finite element mesh (with scale factor of 2) Stress and strain in Si: Stress mostly in horizontal direction (1-3 GPa tensile) Finite elements calculation 5% volume compaction = shift of 5 nm Theoretical estimation of resonance wavelength shift (nm)
13 : Conclusions Fabrication imperfections: Need for trimming Tuning = Overkill Trimming!! Trimming experiment: 5 nm trimming range Physical mechanisms: Stress is dominant
14 : Thank you for your attention!


Views 873 | Favourite 0 | Total Upload :1

Embed Code:

Flag as inappropriate

Related  Most Viewed

Free Powerpoint Templates


Medical PowerPoint Templates | Powerpoint Templates | Tags | Contact | About Us | Privacy | FAQ | Blog

© Slideworld