Full Transcript
https://www.youtube.com/watch?v=oAoH8mEymjA
[00:00] hello welcome to your practices webinar
[00:05] hello welcome to your practices webinar series on advanced photonics packaging
[00:09] series on advanced photonics packaging I'm Ramsey Salim your practice lead to
[00:13] I'm Ramsey Salim your practice lead to Tyndale National Institute where we
[00:15] Tyndale National Institute where we focus on system integration and advanced
[00:17] focus on system integration and advanced photonics packaging this webinar
[00:20] photonics packaging this webinar continues our series where we step into
[00:23] continues our series where we step into the world of advanced with onyx
[00:25] the world of advanced with onyx packaging so far we've had five episodes
[00:28] packaging so far we've had five episodes and we've covered various topics with me
[00:31] and we've covered various topics with me today is my co-host dr. Francesco Flores
[00:34] today is my co-host dr. Francesco Flores see head of the training programs and
[00:40] see head of the training programs and Lucas Agana who is helping develop our
[00:43] Lucas Agana who is helping develop our grating coupler technology here at
[00:45] grating coupler technology here at Tyndall I'm Francesco I look at and so
[00:50] Tyndall I'm Francesco I look at and so five episodes how did we get here yeah
[00:54] five episodes how did we get here yeah well five episodes we started with the
[00:58] well five episodes we started with the photonics packaging rules are
[01:00] Photonics packaging rules are interesting because we shall have to put the devices on a pica on a photonic integrated circuit and make it in the smartest way possible in order to be the peak let's say automatically compatible with our packaging rules which means easy and cheaper and then we decided to move on.
[01:28] And in the other episode we saw the physics behind photonic packaging so for example we introduced the coupling scheme so mainly each calendar so we show you the concept of the mode mismatch how you can use a spot size converter and lenses in order to match the notes between fibre and we're kind.
[01:50] And then the second way to couple light in and out of a PDF that is the grating cut Linga so we show you the basic concept so the so-called standard rating coupler which means pitch constant and
[02:03] coupler which means pitch constant and then we move further in order to
[02:06] then we move further in order to increase the cutting efficiency so we
[02:08] increase the cutting efficiency so we talked about the scattering strength we
[02:10] talked about the scattering strength we increase the scattering self by adding
[02:13] increase the scattering self by adding a polysilicon overlay on top of the
[02:16] a polysilicon overlay on top of the grating coupler and then we introduce
[02:19] grating coupler and then we introduce the concept of the apodization
[02:22] the concept of the apodization polarization is you do not have a single
[02:26] polarization is you do not have a single pitch anymore you change the pitch along
[02:29] pitch anymore you change the pitch along the longitudinal direction of the
[02:31] the longitudinal direction of the reading coupler in order to achieve
[02:33] reading coupler in order to achieve impedance matching and we saw that in
[02:37] impedance matching and we saw that in that case we were able to increase the
[02:38] that case we were able to increase the coupling efficiency from let's say 50 to
[02:41] coupling efficiency from let's say 50 to 60 to 70 75 percent so we are getting
[02:46] 60 to 70 75 percent so we are getting always better and better in terms of
[02:48] always better and better in terms of performances but another question is how
[02:51] performances but another question is how can you do better so okay can i for
[02:55] can you do better so okay can i for example learn a way to optimize and
[02:58] example learn a way to optimize and create my own design well the question
[03:01] create my own design well the question obviously it's interesting but it's
[03:02] obviously it's interesting but it's interesting because the answer is yes
[03:05] interesting because the answer is yes and today Luca will show you our method.
[03:08] and today Luca will show you our method our approach that is based on the so
[03:10] our approach that is based on the so called piezo so the particle swarm
[03:13] called piezo so the particle swarm optimization algorithm and we
[03:15] optimization algorithm and we implemented this inside an fdtd
[03:17] implemented this inside an fdtd software really really interesting so
[03:20] software really really interesting so look at so I can share my screen okay so
[03:34] look at so I can share my screen okay so as Francesco Redi said today I'm gonna
[03:37] as Francesco Redi said today I'm gonna talk about our design routine that we
[03:40] talk about our design routine that we use to optimize and design creating
[03:42] use to optimize and design creating couplers so the the content of today's
[03:47] couplers so the the content of today's talk is divided into four main parts the
[03:50] talk is divided into four main parts the first part is just a quick overview on
[03:52] first part is just a quick overview on the current challenges that we have in
[03:55] the current challenges that we have in coupling the light inside a photonic
[03:57] coupling the light inside a photonic integrated circuit or pic then just a
[04:00] integrated circuit or pic then just a general introduction on grating coupler
[04:02] general introduction on grating coupler structure and then we shall directly
[04:04] structure and then we shall directly into the design routine so how the
[04:06] into the design routine so how the particles were optimization or PSO.
[04:09] particles were optimization or PSO algorithm works and how we can implement.
[04:11] algorithm works and how we can implement inside the fdtd software how can we.
[04:14] inside the fdtd software how can we extract the final result and build our.
[04:16] extract the final result and build our three-dimensional grating cutter inside.
[04:19] three-dimensional grating cutter inside a 3d fdtd.
[04:20] a 3d fdtd how we can validate the design routine.
[04:23] how we can validate the design routine and then.
[04:23] and then an application so the fundamental.
[04:28] an application so the fundamental problem that we have in coupling the.
[04:29] problem that we have in coupling the light inside a peak is that there is.
[04:31] light inside a peak is that there is different mismatch between the mode of.
[04:34] different mismatch between the mode of the sources that we use to couple the.
[04:36] the sources that we use to couple the light inside a peak and once obtained by.
[04:39] light inside a peak and once obtained by the waveguides inside a peak so here to.
[04:43] the waveguides inside a peak so here to overcome this issue we can use a.
[04:45] overcome this issue we can use a structure that is able to couple but.
[04:49] structure that is able to couple but efficiently the light and to adapt these.
[04:51] efficiently the light and to adapt these two mode one example can be grating.
[04:54] two mode one example can be grating cutters as you can see a grating.
[04:56] cutters as you can see a grating Cadbury's of three-dimensional structure.
[04:58] Cadbury's of three-dimensional structure that has a plane of symmetry if we.
[05:00] that has a plane of symmetry if we consider the projection onto this plane.
[05:02] consider the projection onto this plane of symmetry of the 3d structure we get.
[05:05] of symmetry of the 3d structure we get the two-dimensional cross-section and.
[05:07] the two-dimensional cross-section and here we can see the different layers of.
[05:09] here we can see the different layers of the SOI so the silicon substrate the.
[05:13] the SOI so the silicon substrate the bottom oxide layer and the silicon.
[05:15] bottom oxide layer and the silicon grating coupler layers and also the.
[05:17] grating coupler layers and also the grating coupler structure itself with.
[05:19] grating coupler structure itself with its parameters that can mediate in depth.
[05:22] its parameters that can mediate in depth the pitch and these are the parameters.
[05:24] the pitch and these are the parameters that defines the quality of the.
[05:27] that defines the quality of the diffraction process that couples the.
[05:29] diffraction process that couples the light inside the waveguide and then if.
[05:32] light inside the waveguide and then if we consider the 2d cross section and if.
[05:34] we consider the 2d cross section and if we want to rotate for example it about.
[05:37] the axis of rotation for a certain angle.
[05:39] we want to rotate for example it about the axis of rotation for a certain angle we are able to rebuild the.
[05:41] we are able to rebuild the three-dimensional structure just using.
[05:44] three-dimensional structure just using the 2d cross section so that sudha cross.
[05:46] the 2d cross section so that sudha cross section is the starting point of our.
[05:48] section is the starting point of our design within here you can see the core.
[05:52] design within here you can see the core concept of the design routine so the.
[05:56] concept of the design routine so the part how the particles worm algorithm.
[05:58] part how the particles worm algorithm works and how it implemented in study.
[06:00] works and how it implemented in study fdtd we use the fdtd in order to wrestle.
[06:04] fdtd we use the fdtd in order to wrestle the athlete the electromagnetic problem.
[06:06] the athlete the electromagnetic problem and to get the coupling efficiency that.
[06:08] and to get the coupling efficiency that is basically the function that we want.
[06:10] is basically the function that we want to optimize and the particles were is.
[06:13] to optimize and the particles were is the algorithm that optimize the.
[06:15] the algorithm that optimize the structure so first we want to define the.
[06:18] structure so first we want to define the parameters that we want to optimize that.
[06:21] parameters that we want to optimize that can be in this case the pitch the.
[06:23] can be in this case the pitch the etching depth and the height of the.
[06:25] etching depth and the height of the talks or top oxide layer and then what.
[06:30] talks or top oxide layer and then what we can do is to use the particle in.
[06:32] we can do is to use the particle in swarm to randomly choose set of these.
[06:35] swarm to randomly choose set of these parameters inside the parameters.
[06:37] parameters inside the parameters and what we had is basically general.
[06:41] and what we had is basically general structures these general structures are.
[06:43] structures these general structures are agents are treated as agents or.
[06:46] agents are treated as agents or particles inside the parameter space by.
[06:48] particles inside the parameter space by the PSO algorithm at the beginning what.
[06:51] the PSO algorithm at the beginning what we can do is to simulate its general.
[06:53] we can do is to simulate its general structure extract the coupling.
[06:55] structure extract the coupling efficiency for each agent and then we.
[06:58] efficiency for each agent and then we have our starting position inside the.
[07:01] have our starting position inside the parameter space then the particles won't.
[07:03] parameter space then the particles won't start the iteration process so what it.
[07:07] start the iteration process so what it does is basically change the parameters.
[07:11] does is basically change the parameters rebuild the general structures inside
[07:14] rebuild the general structures inside the fdtd and get as a feedback the
[07:16] the fdtd and get as a feedback the coupling efficiency at each iteration
[07:18] coupling efficiency at each iteration and after a certain amount of iterations
[07:21] and after a certain amount of iterations what we see is that all the agent after
[07:25] what we see is that all the agent after scanning the parameter space collapsed
[07:28] scanning the parameter space collapsed to the addressed global solution and we
[07:31] to the addressed global solution and we can extract these parameters from the
[07:34] can extract these parameters from the algorithm and rebuild the 2d
[07:36] algorithm and rebuild the 2d cross-section then we can use the 2d
[07:40] cross-section then we can use the 2d cross-section to rebuild the 3d
[07:43] cross-section to rebuild the 3d structure and export it inside a
[07:46] structure and export it inside a three-dimensional external simulation so
[07:49] three-dimensional external simulation so the 3d fdtd is quite time-consuming so
[07:53] the 3d fdtd is quite time-consuming so we want to be sure that we are able to
[07:56] we want to be sure that we are able to look at all the parameters that defines
[08:00] look at all the parameters that defines the diffraction process and the
[08:02] the diffraction process and the propagation of delighted interaction
[08:04] propagation of delighted interaction with the grating router and to do that
[08:06] with the grating router and to do that we can use a set of monitors for example
[08:09] we can use a set of monitors for example the exact cross section monitor is used
[08:12] the exact cross section monitor is used to look at the quality of the diffraction process.
[08:17] the XY cross section is used to look at the focusing effect of the grating cutter.
[08:22] so the ability of focusing the light at the entrance tip of the waveguide and the white cross section is able to give information on the quality of the coupling process.
[08:32] so if this structure is able to couple the light inside the fundamental mode of the wave and then we have other three monitors the transmittance monitors and reflective monitors that are used to get the losses of the structure and and of course the public efficiency monitor to get the amount of light that it's coupled inside the wave.
[08:50] so after we have set up these three-dimensional simulation and we have run it we are able to validate our design routine.
[08:57] how well in this case we use the particles world to read to optimize to a structure known in literature.
[09:06] in this case we can see a uniform crating cobbler with a silicon overlay on top of each tooth and then I
[09:14] overlay on top of each tooth and then I put a sweetened coupler with a silicon.
[09:16] put a sweetened coupler with a silicon grating copper layer of 260 nanometers.
[09:18] grating copper layer of 260 nanometers so would appear so we can look at the.
[09:21] so would appear so we can look at the parameters the geometrical parameters.
[09:23] parameters the geometrical parameters that defines is to structure and we can.
[09:25] that defines is to structure and we can compare them with the one reported in.
[09:29] compare them with the one reported in the papers but also we can use the three.
[09:31] the papers but also we can use the three monitors due to for the losses and the.
[09:34] monitors due to for the losses and the one for the buffing efficiency to look.
[09:35] one for the buffing efficiency to look at the electromagnetic spectrum of our.
[09:39] at the electromagnetic spectrum of our structure and what we see is that at the.
[09:41] structure and what we see is that at the working way length the difference.
[09:42] working way length the difference between the structure that the particles.
[09:45] between the structure that the particles were found is just three percent respect.
[09:47] were found is just three percent respect to the measure one and reported in the.
[09:49] to the measure one and reported in the paper and one percent for the appetizer.
[09:52] paper and one percent for the appetizer 18 cobbler plus using the losses we are.
[09:57] 18 cobbler plus using the losses we are able to see also the behavior of the.
[10:00] able to see also the behavior of the interaction between the electromagnetic.
[10:02] interaction between the electromagnetic field and the structure for example what.
[10:04] field and the structure for example what we can see is that there is a difference.
[10:06] we can see is that there is a difference between the transmission transmittance.
[10:09] between the transmission transmittance channel sorry of the uniform grating.
[10:11] channel sorry of the uniform grating copper and the transmittance of the.
[10:13] copper and the transmittance of the appetizer it in copper.
[10:14] appetizer it in copper why because in this case we have better impedance matching due to the fact that we have a firing pitch throughout the grating copper and basically we are filtering better or we are creating a better than pass filter effect that is able to boost the coupling efficiency even further respect to the uniform rating Cutler why the particles were can be interesting well we can apply the particles work to optimize our structures for a different light coupling scheme for example the fdtd can be used to mimic the emission of a vertical fiber plain fiber and in my practical bench or a movie so in this case we have a laser as a source we use the bow lens to collect the light and to reshape the laser in order to get a focal spot that has the same features in terms of mode field diameter of the fibers mission and the priests in order to get the right
[11:14] priests in order to get the right impinging angle on the top surface of the grating.
[11:16] impinging angle on the top surface of the grating so we can use the fdtd to extract is it the emission of this free object import inside the design routine.
[11:19] the grating so we can use the fdtd to extract is it the emission of this free object import inside the design routine.
[11:21] extract is it the emission of this free object import inside the design routine.
[11:24] object import inside the design routine and optimize our creating cutters for is free capping schemes.
[11:26] and optimize our creating cutters for is free capping schemes so that's it for today.
[11:30] free capping schemes so that's it for today I hope that the webinar world was interesting and clear.
[11:33] today I hope that the webinar world was interesting and clear.