laser integration challenges - eexu.home.ece.ust.hk
TRANSCRIPT
![Page 1: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/1.jpg)
Laser Integration challenges
For on-chip optical interconnects
Date 2016 – OPTICS workshop March 18th, 2016
Y. Léger, C. Cornet, R. Tremblay,
I. Lucci, O. Durand
Foton Laboratory, CNRS unit,
Rennes, France
![Page 2: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/2.jpg)
Who are we? 2
Date 2016 – OPTICS workshop March 18th, 2016
Optical
communications
& sensors
Materials for
photonics
![Page 3: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/3.jpg)
Who are we? 3
Date 2016 – OPTICS workshop March 18th, 2016
Optical
communications
& sensors
Materials for
photonics
Monolithic silicon photonics
GaPp
Active area
GaPn
GaP buffer
Si substrate
![Page 4: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/4.jpg)
Integrating lasers in microprocessors 4
Date 2016 – OPTICS workshop March 18th, 2016
Where?
How?
What?
The best place of the photonic layer in a
microprocessor
The best integration approach of active photonic
devices
The best laser design for a given MP architecture
![Page 5: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/5.jpg)
Where? 5
Date 2016 – OPTICS workshop March 18th, 2016
Very different approaches by the electronics giants
IBM: Co-integration of CMOS
and photonics
Combined front-end
integration
Intel: Front side bonding of III-V lasers on
Si
![Page 6: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/6.jpg)
Where? 6
Date 2016 – OPTICS workshop March 18th, 2016
At the fab line level:
Thermal management
Fast communication
Process ease
Off-chip communication
Test costs
![Page 7: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/7.jpg)
Where? 7
Date 2016 – OPTICS workshop March 18th, 2016
In 3D architectures:
Inspired from N. Solomon patent
US8136071 B2
٥New issues for thermal
management
٥ Is the previous classification
still relevant?
Always production-line-related
considerations
Consequences at different abstraction
levels of the architecture design
٥Huge issues related to the integration method…
![Page 8: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/8.jpg)
How? 8
Date 2016 – OPTICS workshop March 18th, 2016
٥Group IV photonics (Si, Ge, GeSn)
٥Heterogeneous integration of III-V
٥Metamorphic integration of III-V
٥Pseudomorphic integration of III-V
Laser electric drive
In some cases the wavelength
III-V wafer issue
(large scale integration, costs)
Pseudo-substrate issue
Optical property immaturity
CMOS compatibility
Process and architecture
Best(?)yields
Flexibility
Large scale integration
Good yields
Thermal management
Large scale integration
Best choice for combined
front-end
Thermal management
Nat. Phot. ‘16
Nat. Phot. ‘15
Gent Univ.
Foton
![Page 9: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/9.jpg)
What? 9
Date 2016 – OPTICS workshop March 18th, 2016
Four kinds of laser designs:
Different demands in PICs:
Ridge lasers Microdisks/rings Ph. C. cavities
VCSELs
III-V lab Gent - CEA NTT
DTU
Corona archi.
-External optical drive
- high power laser (0.8W)
- WDM source
- no footprint issue
Chameleon archi.
-thousands of on-chip micro-lasers
-Low power source
- tunability?
- huge footprint constraint
![Page 10: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/10.jpg)
What? 10
Date 2016 – OPTICS workshop March 18th, 2016
D.Miller, Proc. IEEE, 97, 1166 (2009) S. Le Beux et al., ATD Europe (2014)
Calculated from Chameleon data.
![Page 11: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/11.jpg)
What? 11
Date 2016 – OPTICS workshop March 18th, 2016
Ridges
VCSELs
microdisks
Ph. Cs
![Page 12: Laser Integration challenges - eexu.home.ece.ust.hk](https://reader035.vdocuments.mx/reader035/viewer/2022071613/61570b39a097e25c76508081/html5/thumbnails/12.jpg)
Any conclusion?
1) Integration issues at the chip scale
2) The ultimate monolithic integration might
not be so unreachable
3) No laser solution so far : PhCs, microdisks or power
sharing of ridges or VCSELs
12
Date 2016 – OPTICS workshop March 18th, 2016
All the issues, no solutions in …
Integrated lasers on Silicon (C. Cornet, Y. Léger, C. Robert)
To be published (ISTE/Elsevier)