Full Transcript
https://www.youtube.com/watch?v=fBuDLJ4Mqzs
[00:03] Great.
[00:05] Hi everyone and uh welcome to our webinar series titled Introduction to Phutonics Packaging.
[00:10] Uh my name is Mark Renzing and I'm a program manager within the photonix packaging and system integration group here in Tindle National Institute based in Cork and Ireland and this is our first webinar in the series and uh we'll concentrate on the more practical engineering side of packaging uh where we will introduce and discuss the critical critical aspects of designing a package and namely the mechanical thermal electrical and optical considerations.
[00:38] So this series has been developed under uh the your practice platform by my colleague Dr. Camille Gratowski.
[00:46] Camil is a senior researcher here in Tindle and is responsible for the development of new packaging technologies.
[00:52] His main research interests uh focus on pluggable optical connections using microoptics uh as well as general photonic packaging processes.
[01:01] So in our first webinar today, Camil
[01:03] So in our first webinar today, Camil will provide an introduction to photonics packaging and package design.
[01:07] photonics packaging and package design.
[01:10] Webinar two, which will be next Tuesday, we'll discuss then electrical packaging and thermal management.
[01:15] and thermal management.
[01:17] And in webinar three, we'll look at the core topics of optical packaging and specifically focusing on two of the main uh coupling methods uh for coupling to a photonic integrated circuit or pick.
[01:28] photonic integrated circuit or pick.
[01:30] In our final webinar, then we'll focus on automation and design rules and we'll take a look at some examples of the packages that we've developed in the packaging uh group here in Tindle.
[01:39] And as you can see here, the series will run every Tuesday at 1500 hours Brussels time uh for the month of October.
[01:46] Each webinar is approximately 40 minutes and we've allocated 15 minutes for questions and answers at the end.
[01:53] So please feel free to use the chat to ask any questions and we'll address those at the end of the talk.
[02:00] So I hope you enjoy our first webinar and with that I'll hand over to my
[02:03] and with that I'll hand over to my colleague uh Camille to begin this presentation.
[02:10] Perfect.
[02:11] Perfect. Hello everybody. Good morning, good afternoon wherever you are in the world.
[02:17] I hope I think we have some people even calling in from uh from as far as India.
[02:20] So my name is Kami Gratkovski and as Mark mentioned I'm a senior researcher here at Photonix packaging group and in this first webinar we're going to essentially introduce not only photonic packaging but also a little bit I'll talk about the packaging group itself so that uh you can know why you know what what we do and what is our remit and how do we approach things.
[02:46] Um yes so we are located in Tindle National Institute which is uh as you can see from the map on the very eastern northeastern part of Europe in Cork in Ireland in the Republic of Ireland.
[02:59] So this is where where we're located and we're essentially spreading our you know spreading our wings from here. Uh Tindle
[03:06] spreading our wings from here.
[03:08] Uh Tindle Institute is technically part of Institute is technically part of University College Cork and we're
[03:10] University College Cork and we're located in a very picturesque building
[03:12] located in a very picturesque building at the edge of the river.
[03:14] It actually at the edge of the river. It actually was a a brewery back in the day about
[03:17] was a a brewery back in the day about 100 years ago.
[03:19] Now it's turned into 100 years ago. Now it's turned into high, you know, high stakes, high
[03:21] high, you know, high stakes, high profile platonic and micro electronic
[03:23] profile platonic and micro electronic facilities.
[03:26] facilities. Uh, as you can see from the picture over here down below.
[03:28] Perhaps I'll change to laser pointer so that you
[03:31] can see it. This facility is our main
[03:34] fab which is 35 and silicon fab that uh
[03:36] that's state-of-the-art construction here in Cork.
[03:38] So we can do per many things.
[03:41] uh Tindle as a whole concentrates on photonics and micro
[03:43] concentrates on photonics and micro electronics integration where
[03:45] electronics integration where essentially our motto is from atoms to
[03:47] essentially our motto is from atoms to systems.
[03:50] So if we do everything from fury growth of uh photonic structures
[03:51] fury growth of uh photonic structures processing
[03:54] processing uh integration here on our group and
[03:56] uh integration here on our group and then uh does deliver to photonic systems
[03:58] then uh does deliver to photonic systems as well as biopotonics for for integration and for demonstration of uh
[04:07] integration and for demonstration of uh of useful working devices.
[04:11] Our group has two two labs and the primary lab that you can see here depicted on the picture.
[04:15] we have everything that we need to do for prototyping of uh photonic packages.
[04:20] So here we have uh our workstations for design mechanical thermal electronic that can all be uh designed over here on over here uh down to simulations quite sophisticated simulations.
[04:34] We have several machines dedicated for micro electronic packaging.
[04:38] We have wire bonders, flip chip bonders, automated uh uh flip chip bonders as well.
[04:44] We have solder solder jetters, everything that's necessary for basic micro electronic packaging as well as some instruments for packaging of the photonics part.
[04:54] So attachment of fibers in various forms.
[04:57] So this is our essentially our lab that's dedicated for uh prototyping.
[05:03] However, uh that's fairly low scale.
[05:06] we can prototype up to several packages uh
[05:10] Can prototype up to several packages uh for any individual collaboration.
[05:13] So in order to proliferate the photonics and make it more useful uh in a high volume manufacturing scheme, we have a secondary lab onto which we have uh several machines which are more dedicated to the automated uh part of packaging.
[05:26] So whatever prototypes we develop in one lab, we work with a company in this case FCON tech uh on automation of those uh of those packaging processes that then can be then taken over standardized and scaled up to higher volume manufacturing and that's essentially our imitate.
[05:45] We're working with uh many collaborators from around the world from universities in Europe, America and Asia uh through uh equipment manufacturers at various levels.
[05:56] So micro uh so uh for example here we have microoptics uh manufacturers.
[06:03] We have uh collaborations with uh uh optical connector manufacturers uh foundaries like IMAC.
[06:12] Manufacturers, uh, foundries like IMAC, like, uh, where is like Global Foundries.
[06:16] Like, uh, where is like Global Foundries, also Amphotonix on on that level.
[06:19] Also Amphotonix on on that level, and then we take those, uh, processes and we.
[06:22] Then we take those, uh, processes and we work with, uh, pretty major players around.
[06:24] Work with, uh, pretty major players around the world like Samsung, Boeing, Intel, but.
[06:27] Also Striker, which is a medical company.
[06:30] Also Striker, which is a medical company, which has its head headquarters in.
[06:32] Which has its head headquarters in Dubai.
[06:34] So we're working on many things in the world of photonix, and when you, uh.
[06:37] In the world of photonix, and when you, uh, and why photonix packaging is so.
[06:41] and why photonix packaging is so critical at the moment, the best way I.
[06:43] Can describe it is if we compare, if we.
[06:46] take the trace, the routed, and micro.
[06:50] Electronics has taken over the past 70.
[06:53] Odd years from 1947, where the first.
[06:55] Germaine transistor was demonstrated, and.
[06:58] This is how it looked like.
[07:00] It's essentially the size of a palm of your.
[07:03] Hand, and, uh, through many, many years of.
[07:07] Processing development, uh, we have now.
[07:09] Reached a a state where the complexity.
[07:12] reached a a state where the complexity of photo electronic integrated circuits is quite high.
[07:15] of photo electronic integrated circuits is quite high.
[07:17] You can cram a huge amount of uh transistors onto a very small piece of silicon and have to working in you know in some sort of coherent uh way.
[07:20] amount of uh transistors onto a very small piece of silicon and have to working in you know in some sort of coherent uh way.
[07:21] small piece of silicon and have to working in you know in some sort of coherent uh way.
[07:23] working in you know in some sort of coherent uh way.
[07:26] coherent uh way. However, these are just disparate circuits and to actually have a useful device, you have to combine often several uh or several tens of those circuits on some sort of substrate and combined in such a way that you have a useful working device and this is uh this is where the packaging comes in.
[07:29] disparate circuits and to actually have a useful device, you have to combine often several uh or several tens of those circuits on some sort of substrate and combined in such a way that you have a useful working device and this is uh this is where the packaging comes in.
[07:31] a useful device, you have to combine often several uh or several tens of those circuits on some sort of substrate and combined in such a way that you have a useful working device and this is uh this is where the packaging comes in.
[07:34] often several uh or several tens of those circuits on some sort of substrate and combined in such a way that you have a useful working device and this is uh this is where the packaging comes in.
[07:37] those circuits on some sort of substrate and combined in such a way that you have a useful working device and this is uh this is where the packaging comes in.
[07:38] and combined in such a way that you have a useful working device and this is uh this is where the packaging comes in.
[07:42] a useful working device and this is uh this is where the packaging comes in.
[07:44] this is where the packaging comes in. So the packaging of micro electronics have been developing for the past 50 years.
[07:46] the packaging of micro electronics have been developing for the past 50 years.
[07:49] been developing for the past 50 years. So now it's pretty well developed and we're pretty pretty satisfied where we're going.
[07:50] So now it's pretty well developed and we're pretty pretty satisfied where we're going.
[07:52] we're pretty pretty satisfied where we're going. And of course uh the the innovation here uh over the last 10 years has been um pretty staggering.
[07:54] we're going. And of course uh the the innovation here uh over the last 10 years has been um pretty staggering.
[07:56] innovation here uh over the last 10 years has been um pretty staggering.
[07:59] years has been um pretty staggering. We now can essentially have computers in the palms of our hands.
[08:01] We now can essentially have computers in the palms of our hands.
[08:05] The same or very similar uh path can be traced for photonics where in the early 80s we started with a simple bar semiconductor lasers and uh early photonics was just
[08:07] similar uh path can be traced for photonics where in the early 80s we started with a simple bar semiconductor lasers and uh early photonics was just
[08:09] photonics where in the early 80s we started with a simple bar semiconductor lasers and uh early photonics was just
[08:11] started with a simple bar semiconductor lasers and uh early photonics was just
[08:14] Lasers and uh early photonics was just that.
[08:17] You had your sources, you had your fibers, and you had your detectors.
[08:19] And those were very simple networks.
[08:20] And right now, many foundaries around the world are working on putting as much functionality as they can onto again small pieces of silicon or lithium nioate or silicon nitride, whatever platform they're using for phosphite.
[08:33] So those are by uh by comparison, those are called photonic integrated circuits.
[08:39] And in order to turn those photonic integrated circuits again, you need packaging in order to turn them into working photonic devices.
[08:48] However, since this has started relatively uh late in the process, we're active for the last 10 years when we start developing photonic packaging advanced photonic packaging techniques.
[08:58] We are not yet as advanced as microeleronic manufacturing.
[09:01] So we're at the stage where we're trying to uh come into line with what electronics has uh is capable of and that's why that's why packaging is important right now.
[09:13] And of course packaging uh when people think about
[09:15] packaging uh when people think about photonics is not only dataccom and telecom which were let's say the the most obvious uh uh places where you would where you would go looking for photonic packages but also very recently uh over the past less than a decade we have started looking into putting photonic sensors into use uh for example lidars or very small biomedical devices uh which can help with uh very early stage detection.
[09:43] ction of some diseases or even help with with uh uh with diagnosis and of course AI uh is a major driver right now and all of these possibilities are essentially using light uh as their as their main as as their main uh as the core.
[10:03] So photonic packages is essentially the way that we understand it is centered around the photonic integrated circuit or pick and uh that's essentially the hard that's the the processor that that uh uh that does all the hard work and
[10:17] uh uh that does all the hard work and the uh plotonic package has to provide
[10:21] the uh plotonic package has to provide the mechanical stability but also
[10:23] the mechanical stability but also connectivity over several uh over
[10:26] connectivity over several uh over several critical dimensions.
[10:29] several critical dimensions. Uh mostly electronic connectivity.
[10:31] electronic connectivity. You have to power the device and you have to read
[10:32] power the device and you have to read out some uh electrical signals or
[10:35] out some uh electrical signals or provide some electrical signals.
[10:39] provide some electrical signals. Uh also however uh which is a difference from
[10:41] however uh which is a difference from electronics you have to provide optical
[10:44] electronics you have to provide optical connectivity and also thermal control is
[10:46] connectivity and also thermal control is quite important for photonic packages as
[10:49] quite important for photonic packages as unlike electronics photonics is quite
[10:51] unlike electronics photonics is quite sensitive to variations in temperature.
[10:54] sensitive to variations in temperature. So from this you can see that photonic
[10:56] So from this you can see that photonic packaging is quite multi-disiplinary
[10:58] packaging is quite multi-disiplinary uh engineering science which requires
[11:00] uh engineering science which requires knowledge in several key aspects
[11:02] knowledge in several key aspects mechanical engineering thermal
[11:04] mechanical engineering thermal engineering electrical electronic
[11:07] engineering electrical electronic electronic engineering as well as a
[11:09] electronic engineering as well as a little bit of knowledge in the optics.
[11:12] little bit of knowledge in the optics. However uh typically in packaging
[11:16] However uh typically in packaging uh is essentially limited effectively to
[11:21] uh is essentially limited effectively to the interfaces i.e.
[11:25] We don't really from the interfaces i.e.
[11:27] We don't really from the photonic packaging perspective, it's the photonic packaging perspective, it's not really critical what's on the chip.
[11:29] not really critical what's on the chip itself.
[11:31] What's critical from the itself. What's critical from the packaging perspective is the interfaces,
[11:33] packaging perspective is the interfaces, how many electrical connections, optical.
[11:35] how many electrical connections, optical connections you have, what is the size.
[11:37] connections you have, what is the size of uh of the chip and also how much heat.
[11:40] of uh of the chip and also how much heat does it generate and how it needs to be.
[11:42] does it generate and how it needs to be handled. So if you think about photonic.
[11:45] handled. So if you think about photonic packaging, you have to think about.
[11:46] packaging, you have to think about engineering of the interfaces. However,
[11:49] engineering of the interfaces. However, of course, photonic packaging is uh.
[11:51] of course, photonic packaging is uh critical if you want to have a working.
[11:52] critical if you want to have a working device because you can have the best.
[11:54] device because you can have the best chip in the world, but without putting.
[11:55] chip in the world, but without putting it into some sort of, you know, a device.
[11:57] it into some sort of, you know, a device that can provide all those uh all those.
[12:01] that can provide all those uh all those interfaces, uh it's essentially a.
[12:03] interfaces, uh it's essentially a useless piece of silicon.
[12:07] useless piece of silicon. There are several uh critical uh issues.
[12:11] There are several uh critical uh issues with photonic packaging right now.
[12:13] with photonic packaging right now. Mostly they stem from the fact that the.
[12:15] Mostly they stem from the fact that the photonic packages are typically with few.
[12:18] photonic packages are typically with few exceptions actually uh they're uh very.
[12:22] exceptions actually uh they're uh very customs they're very application specific.
[12:24] you have a chip and you build a package around that chip so uh it's it's not very standardized discipline.
[12:30] we're only recently started working on standardization of some uh of some processes and neural practice is one of those uh one of those avenues through which we're doing it.
[12:37] also optical packaging as you will see in the third webinar requires very high precision uh which means that it's quite difficult to align optical fibers to the waveguides and it takes a long time and it's also package by package process.
[12:58] It's a very serial process which is difficult to automate and scale up and it all means that if you look at the cost breakdown it's not surprising that the assembly packaging and testing takes about would take 70 to 80% of the overall cost of the package.
[13:15] So part of our remit is also uh decreasing this uh uh this cost fraction so that all the packaging can be scaled up for high volume.
[13:23] packaging can be scaled up for high volume manufacturing and become cheaper.
[13:25] volume manufacturing and become cheaper that's coming down the line over the next several years.
[13:30] If you look at for packaging technologies you can see they're essentially divided into two domains.
[13:33] Uh the domain on the right is is the part that can be leveraged from the developments in the micro electronics industry and those are essentially mechanical uh uh pack mechanical aspects of packaging, thermal management as well as electrical aspects of packaging.
[13:51] So interposers, PCB design, wire bonding, flip chip bonding, all those technologies are legacy technologies that we can incorporate into photonic packaging.
[13:58] However, the optical part uh which require source integration and typically fiber attachment, those are the parts that require new solutions and right now they're essentially tacked on they're essentially a separate pipeline from the micro electronic packaging.
[14:16] So in the future what we are going towards is a full integration of these two disparate domains into a single pipeline where we
[14:23] domains into a single pipeline where we can leverage the legacy techniques and can leverage the legacy techniques and use them for optical and uh uh for optical packaging as well.
[14:31] So that is the the ultimate vision.
[14:34] So uh photonic packaging is quite a broad uh topic.
[14:42] Uh therefore for the purposes of this course uh I I have prepared several key topics uh package design, electrical packaging for management.
[14:55] Those two will be actually the the the topics of the next webinar.
[15:00] Optical packaging is the big topic that requires luminar of its own and uh there are several other topics which are not which are which were excluded from this due to the time constraints but also some of them are covered by other Europractice tutorials for example source integration and additive manufacturing those are covered by other tutorials like micro transfer printing and 3D printing.
[15:20] So uh you can refer to
[15:24] and 3D printing.
[15:24] So uh you can refer to those.
[15:27] However, the full the full course of autonomic packaging would take about 15 hours.
[15:32] So we're limiting to to several key topics.
[15:34] So let's go on to the package design.
[15:37] This is uh this is going to be a little bit of a of a broad view uh because as I mentioned previously uh packaging right now is application specific.
[15:48] So it's difficult to to to say uh to say definitely what's going how do you need to design your package.
[15:56] Of course when standardization kicks in that's going to be a different topic but however we're now working mostly with prototypes.
[16:02] So this is going to provide with some key ideas.
[16:04] So how can you lay out and design your prototype.
[16:08] So the purpose of design this is the most critical aspect.
[16:11] You can say that actual packaging process is uh the easy part at the tail end of the design and the purpose of design is to provide the stable platform for optical connection.
[16:20] As I mentioned optical uh connectivity is a very precise uh
[16:25] connectivity is a very precise uh endeavor.
[16:27] So we need to be sure that at endeavor.
[16:29] So we need to be sure that at the end of the process you can actually the end of the process you can actually provide a stable platform uh for those
[16:32] provide a stable platform uh for those fibers to to stick onto the the end of
[16:34] fibers to to stick onto the the end of the chip.
[16:37] Of course uh we have to electronically connect the chip to the
[16:39] electronically connect the chip to the to the system uh around it using some
[16:42] to the system uh around it using some redistribution layers or RDL's typically
[16:45] redistribution layers or RDL's typically also have to provide thermal
[16:46] also have to provide thermal stabilization and uh make from the
[16:49] stabilization and uh make from the mechanical point of view we have to
[16:50] mechanical point of view we have to ensure mechanical compatibility with the
[16:52] ensure mechanical compatibility with the rest of the system.
[16:53] So this is typically the size of the package has to conform
[16:56] the size of the package has to conform whatever restrictions there are uh
[16:58] whatever restrictions there are uh within the broader system.
[16:59] But however, if you're working with prototypes,
[17:01] if you're working with prototypes, that's typically not as critical.
[17:04] that's typically not as critical.
[17:07] And also uh something that's often overlooked uh in the prototyping process
[17:10] overlooked uh in the prototyping process is that you have to be conscious about
[17:13] is that you have to be conscious about what the packaging machine and the
[17:15] what the packaging machine and the processes can deliver.
[17:17] And without that knowledge, you might uh go through those
[17:21] knowledge, you might uh go through those four steps and still not be able to
[17:24] four steps and still not be able to package your device because uh what you
[17:27] package your device because uh what you did is incompatible with the packaging.
[17:29] did is incompatible with the packaging processing techniques that you can find.
[17:31] processing techniques that you can find uh within uh uh with working with for.
[17:34] uh within uh uh with working with for example us but also with other.
[17:36] example us but also with other providers.
[17:38] providers. So there are several design layers uh of.
[17:41] So there are several design layers uh of course for platonic packaging uh optical.
[17:44] course for platonic packaging uh optical packaging. So the design of the chip and.
[17:46] design of the chip and design of the fiber layer those are.
[17:48] those what comes into the optical uh.
[17:50] optical uh design part. Thermal design is the heat.
[17:53] design part. Thermal design is the heat spreaders as well as thermal electric.
[17:55] spreaders as well as thermal electric coolers typically. Uh electrical design.
[17:58] coolers typically. Uh electrical design covers any PCBs, interposers, electronic.
[18:01] covers any PCBs, interposers, electronic integrated circuits which are sitting.
[18:02] integrated circuits which are sitting close by on the interposers as well as.
[18:05] close by on the interposers as well as electronic connectors whether DC or RF.
[18:08] electronic connectors whether DC or RF and of course everything has to sit very.
[18:09] and of course everything has to sit very stably on a submount or within some sort.
[18:12] stably on a submount or within some sort of housing. And also strain relief is a.
[18:15] of housing. And also strain relief is a key component typically that that you.
[18:17] key component typically that that you need also to provide for stab for.
[18:19] need also to provide for stab for long-term stability.
[18:21] long-term stability. So this is an example of a very old.
[18:23] So this is an example of a very old package that we have designed about.
[18:26] package that we have designed about seven or eight years ago. So here you.
[18:27] seven or eight years ago.
[18:29] So here we can see all those layers.
[18:33] So here we have the base uh housing the stable platform onto which uh the interposer and the chip that sits on top are placed.
[18:41] There is electronic uh layer made of PCB and several DC connectors which uh are actually I think for this package we had close to 200 electrical connections.
[18:50] So that required quite a lot of design.
[18:53] And then we have our fiber array which will connect it ultimately for the chip and then a strain relief layer to prevent any accidental mishandling mishaps.
[19:03] Uh picks uh can come in variety of shapes, sizes and interfaces.
[19:09] However, from the packaging perspective, you're typically interested from the mechanical side in all the three sizes.
[19:15] So length, width and thickness of the chip because they will design your mechanical uh constraints.
[19:21] Uh from the electrical side, you're looking at number of DC interfaces and how they're positioning the chip.
[19:25] And also for some
[19:27] the chip.
[19:29] And also for some applications, not all of them, you also need some sort of RF or radio frequency interfaces.
[19:35] Uh which can also have to be positioned correctly for packaging to work properly.
[19:41] And so if we take take just these three domains we have essentially an electronic integrated circuit.
[19:49] However uh if we have any optical couplers in this case we have edge couplers that gives us puts us into the realm of photonic packaging.
[19:54] So in this case we have for example eight edge couplers pitched at 127 microns.
[19:59] And this is a very simple RX chip that was developed in Pixab pilot line.
[20:05] Uh it's very simple.
[20:08] However, of course, picks can come all shapes, sizes, functionalities, and levels of complexity.
[20:14] So, this is uh not the most complicated chip that we worked on, but you can see there are dozens of optical interface connections, edge couplers, and there are also multiple electrical connections which are in this case they're notware bonded, flip chip bonded.
[20:27] So, we can see several bond
[20:28] bonded. So, we can see several bond paths appearing throughout the surface
[20:30] paths appearing throughout the surface area of the chip. So the picks can be
[20:33] area of the chip. So the picks can be very complicated and the design of the
[20:35] very complicated and the design of the pick actually also has to uh take into
[20:38] pick actually also has to uh take into consideration
[20:40] consideration uh the other uh other uh considerations
[20:43] uh the other uh other uh considerations including also your packaging machines.
[20:47] including also your packaging machines. The thermal layer sounds very simple and
[20:49] The thermal layer sounds very simple and it's typically very underappreciated um
[20:52] it's typically very underappreciated um and it's typically composed of your heat
[20:54] and it's typically composed of your heat sources. So in this case we have a 3D
[20:56] sources. So in this case we have a 3D integration where we have a photonic
[20:58] integration where we have a photonic chip with electronic integrated circuit
[21:00] chip with electronic integrated circuit uh assembled on top of it and this is a
[21:03] uh assembled on top of it and this is a major heat source in this case. Then we
[21:04] major heat source in this case. Then we have a heat spreader which is placed on
[21:07] have a heat spreader which is placed on top of the thermmoelectric cooler and
[21:09] top of the thermmoelectric cooler and the base is our essential heat sink. So
[21:10] the base is our essential heat sink. So the path of the heat essentially comes
[21:12] the path of the heat essentially comes from EIC through the pick heat spreader
[21:16] from EIC through the pick heat spreader and is extracted through the tech. So
[21:18] and is extracted through the tech. So the the path uh of the heat actually
[21:21] the the path uh of the heat actually through the pick is not the optimum one.
[21:23] through the pick is not the optimum one. However, in some cases, it's the only
[21:25] However, in some cases, it's the only thing that you can do to uh decrease the
[21:28] thing that you can do to uh decrease the uh the path length between your EIC and
[21:30] uh the path length between your EIC and the pick
[21:32] the pick electrical layer. Again, a very simple
[21:34] electrical layer. Again, a very simple example here. You have photonic
[21:36] example here. You have photonic integrated circuit. That's the very
[21:38] integrated circuit. That's the very small piece here. Uh in the end, uh you
[21:40] small piece here. Uh in the end, uh you can also see some electronic circuits uh
[21:44] can also see some electronic circuits uh placed on the interposer. So that's
[21:46] placed on the interposer. So that's let's say that's one sub assembly of the
[21:48] let's say that's one sub assembly of the electrical layer. Then everything is
[21:51] electrical layer. Then everything is again flip chip bonded onto a PCB.
[21:54] again flip chip bonded onto a PCB. That's another distribution layer that
[21:57] That's another distribution layer that connects all the DC lines to this uh
[21:59] connects all the DC lines to this uh large connector and RF lines to these uh
[22:03] large connector and RF lines to these uh to these connectors on the other side.
[22:05] to these connectors on the other side. So the electrical layer again this is
[22:08] So the electrical layer again this is something that we'll discuss in the next
[22:10] something that we'll discuss in the next webinar. Mechanical layer also sounds
[22:12] webinar. Mechanical layer also sounds simple but it has to take into
[22:14] simple but it has to take into consideration uh several aspects. In
[22:17] consideration uh several aspects. In this case, of course, we have the submod
[22:19] this case, of course, we have the submod that has to provide mechanical support
[22:21] that has to provide mechanical support for the entire assembly. Uh there is a
[22:24] for the entire assembly. Uh there is a strain relief that provides uh support
[22:27] strain relief that provides uh support for fibers. So any yanking action during
[22:30] for fibers. So any yanking action during handling will not translate to the
[22:32] handling will not translate to the optical interface. It will not lead to
[22:34] optical interface. It will not lead to disconnect or loss of signal. Uh and of
[22:37] disconnect or loss of signal. Uh and of course we have some in this case it will
[22:38] course we have some in this case it will be mounted on optical table. So in this
[22:40] be mounted on optical table. So in this case, we have some mounting holes as
[22:42] case, we have some mounting holes as well as a uh plastic cover uh for uh any
[22:47] well as a uh plastic cover uh for uh any to prevent any debris from falling down.
[22:49] to prevent any debris from falling down. This is typically what what we also do
[22:52] This is typically what what we also do and it's useful to to shield your
[22:54] and it's useful to to shield your optical interface from any any accidents
[22:57] optical interface from any any accidents whether mishandling or any uh any things
[23:01] whether mishandling or any uh any things falling down on it.
[23:03] falling down on it. The packaging machine is the critical
[23:05] The packaging machine is the critical thing that you have to be conscious of
[23:07] thing that you have to be conscious of when you're even at the stage when
[23:09] when you're even at the stage when you're designing your pick because of
[23:11] you're designing your pick because of course uh in many cases you have to
[23:13] course uh in many cases you have to attach the fiber to to your chip at the
[23:16] attach the fiber to to your chip at the end. Uh there are some some companies I
[23:19] end. Uh there are some some companies I know they they package uh the other way
[23:23] know they they package uh the other way around. First they attach the fiber to
[23:25] around. First they attach the fiber to the chip and then they build a package
[23:26] the chip and then they build a package around it. The way we work is we
[23:28] around it. The way we work is we typically assemble uh the this entire
[23:33] typically assemble uh the this entire uh sub assembly and then last thing we
[23:35] uh sub assembly and then last thing we do is we attach the fiber. There are
[23:37] do is we attach the fiber. There are both ways both have their pluses and
[23:41] both ways both have their pluses and minuses. Uh so a packaging machine uh is
[23:46] minuses. Uh so a packaging machine uh is the key aspect of autonomic packaging
[23:48] the key aspect of autonomic packaging and the design has to be compatible with
[23:51] and the design has to be compatible with the machine that you'll be working on.
[23:52] the machine that you'll be working on. So what we typically do is we build our
[23:56] So what we typically do is we build our uh package in a virtual space that this
[23:58] uh package in a virtual space that this design that goes uh that we design a 3D
[24:01] design that goes uh that we design a 3D CAD model uh then goes onto this uh more
[24:05] CAD model uh then goes onto this uh more sophisticated 3D CAD model that you see
[24:07] sophisticated 3D CAD model that you see here which is essentially our packaging
[24:10] here which is essentially our packaging machine. So here we have a packaging
[24:12] machine. So here we have a packaging stage that the package is going to
[24:14] stage that the package is going to positioned on. We have two alignment
[24:17] positioned on. We have two alignment stages both six axis with a with a
[24:20] stages both six axis with a with a gonometer that defines our center of
[24:22] gonometer that defines our center of rotation. We have a single camera in
[24:24] rotation. We have a single camera in this case and this cone that you can see
[24:26] this case and this cone that you can see this is actually the uh the distance the
[24:28] this is actually the uh the distance the working distance of the camera that you
[24:30] working distance of the camera that you can also build. It's actually good idea
[24:32] can also build. It's actually good idea to build into your into your model so
[24:34] to build into your into your model so you can see how far your camera sees and
[24:36] you can see how far your camera sees and are there any clearance issues. We also
[24:38] are there any clearance issues. We also have some other equipment such as beam
[24:41] have some other equipment such as beam profilers in this specific machine. We
[24:43] profilers in this specific machine. We also have here on the side ready to be
[24:45] also have here on the side ready to be implemented the model our set of tools
[24:47] implemented the model our set of tools that you can use to handle either fibers
[24:50] that you can use to handle either fibers or some other components that you'll be
[24:52] or some other components that you'll be using in packaging. So this is extremely
[24:55] using in packaging. So this is extremely useful uh for for the design process.
[24:59] useful uh for for the design process. Uh this is another example of another
[25:01] Uh this is another example of another machine. This one is geared more towards
[25:03] machine. This one is geared more towards automation but again the same principle
[25:04] automation but again the same principle applies. We have a packaging stage upon
[25:07] applies. We have a packaging stage upon which actually a package can be placed
[25:09] which actually a package can be placed as a treated model and we can see if
[25:11] as a treated model and we can see if there are any clearance or any other
[25:13] there are any clearance or any other mechanical obstructions that will uh
[25:16] mechanical obstructions that will uh prevent packaging process from taking
[25:18] prevent packaging process from taking place. So this is very useful if you are
[25:21] place. So this is very useful if you are interested in more let's say more
[25:23] interested in more let's say more serious uh uh packaging
[25:28] serious uh uh packaging for prototyping. What we found is that
[25:31] for prototyping. What we found is that uh visibility is quite important. you
[25:33] uh visibility is quite important. you have to see what you're doing because if
[25:36] have to see what you're doing because if you can if you can't see what you're
[25:37] you can if you can't see what you're doing then you'll have a very hard time
[25:40] doing then you'll have a very hard time aligning fibers or fiber arrays to the
[25:42] aligning fibers or fiber arrays to the chips. Uh and uh since many of the
[25:45] chips. Uh and uh since many of the automated procedures require machine
[25:47] automated procedures require machine vision if you can't see it yourself then
[25:51] vision if you can't see it yourself then the machine most likely won't be see it
[25:53] the machine most likely won't be see it as well and therefore you cannot
[25:54] as well and therefore you cannot automate those processes. So the design
[25:57] automate those processes. So the design has to incorporate some features for
[25:59] has to incorporate some features for viewing. So in this case we have a very
[26:00] viewing. So in this case we have a very simple example where there's a photonic
[26:03] simple example where there's a photonic chip that's placed on a stage here. Uh
[26:06] chip that's placed on a stage here. Uh we have a in this case we have only
[26:07] we have a in this case we have only single top view camera. So we can see
[26:09] single top view camera. So we can see from the top we can align the fiber to
[26:12] from the top we can align the fiber to the chip from the top. However we have
[26:13] the chip from the top. However we have no visibility essentially with a single
[26:15] no visibility essentially with a single camera from the sides. That's why in
[26:17] camera from the sides. That's why in this case for example we have a prism
[26:19] this case for example we have a prism and a mirror system. to to turn the few
[26:22] and a mirror system. to to turn the few 90° in order to see from the side and we
[26:24] 90° in order to see from the side and we can also place a mirror on the back uh
[26:28] can also place a mirror on the back uh to see from the front and align all the
[26:30] to see from the front and align all the angles. So these are the features that
[26:32] angles. So these are the features that are useful for prototyping
[26:35] are useful for prototyping and they have they can also that's very
[26:38] and they have they can also that's very useful incorporated into the package
[26:40] useful incorporated into the package design. So for example, if we need to
[26:41] design. So for example, if we need to have a side view visibility, what we
[26:43] have a side view visibility, what we typically do is we require some features
[26:45] typically do is we require some features that you can see here. These are cutouts
[26:47] that you can see here. These are cutouts in the PCB which allow us to place a
[26:51] in the PCB which allow us to place a mirror a prism on one side and a mirror
[26:53] mirror a prism on one side and a mirror on the other side so we can actually see
[26:54] on the other side so we can actually see what we're doing from the side. Again,
[26:56] what we're doing from the side. Again, for the same reason, if we cannot see
[26:58] for the same reason, if we cannot see it, we're going to have very hard time
[27:00] it, we're going to have very hard time actually aligning it. So your design
[27:02] actually aligning it. So your design features have to incorporate, especially
[27:04] features have to incorporate, especially when you're working on a prototyping
[27:06] when you're working on a prototyping stage, incorporate those features for
[27:07] stage, incorporate those features for viewing. Otherwise, it
[27:10] viewing. Otherwise, it a packaging this is going to be a living
[27:13] a packaging this is going to be a living nightmare. Uh the package orientation
[27:16] nightmare. Uh the package orientation also is uh quite important. You have to
[27:19] also is uh quite important. You have to be conscious that while many packages
[27:22] be conscious that while many packages actually come uh are assembled with the
[27:24] actually come uh are assembled with the pick facing up. So we can see actually
[27:26] pick facing up. So we can see actually your wave guides and your either in this
[27:29] your wave guides and your either in this case these are edge couplers here but
[27:31] case these are edge couplers here but you can also have grating couplers. So
[27:32] you can also have grating couplers. So if you can see them from the top that's
[27:35] if you can see them from the top that's you know that's one problem less.
[27:36] you know that's one problem less. However, many modern packages require as
[27:40] However, many modern packages require as we'll discuss in the next webinar uh
[27:42] we'll discuss in the next webinar uh flip chip bonding uh instead of wire
[27:45] flip chip bonding uh instead of wire bonding for uh for more scalability and
[27:48] bonding for uh for more scalability and therefore they are actually placed face
[27:50] therefore they are actually placed face down. So what you see here this cutout
[27:52] down. So what you see here this cutout over here is actually so that we can
[27:55] over here is actually so that we can place the package upside down in the
[27:57] place the package upside down in the packaging machine and we can actually
[27:58] packaging machine and we can actually have a single top view camera a preview
[28:01] have a single top view camera a preview of of of uh of the chip passet. So there
[28:06] of of of uh of the chip passet. So there is a little bit of an overhang of the
[28:08] is a little bit of an overhang of the order of 100 microns. This is what the
[28:10] order of 100 microns. This is what the right side image shows here. You have
[28:13] right side image shows here. You have have a very very limited u view range.
[28:17] have a very very limited u view range. So this the uh the dashed line actually
[28:19] So this the uh the dashed line actually shows the edge of the mechanical
[28:22] shows the edge of the mechanical submount and these very faint features
[28:24] submount and these very faint features these are your edge couplers. So these
[28:27] these are your edge couplers. So these design features have to be incorporated
[28:29] design features have to be incorporated especially critically when the package
[28:31] especially critically when the package orientation is non-facing up.
[28:36] orientation is non-facing up. Then we go to a more broad view when we
[28:39] Then we go to a more broad view when we zoom out from just the package itself.
[28:41] zoom out from just the package itself. The typically the packages have to uh
[28:43] The typically the packages have to uh perhaps not that very early stage
[28:45] perhaps not that very early stage prototyping. However, at some stage you
[28:47] prototyping. However, at some stage you have to conform to whatever your system
[28:49] have to conform to whatever your system uh you know mechanical constraints is.
[28:51] uh you know mechanical constraints is. So typically for example if you look at
[28:54] So typically for example if you look at QSFP packages uh which are used in data
[28:58] QSFP packages uh which are used in data centers they're very standardized very
[29:00] centers they're very standardized very small form factor but they have to fit
[29:02] small form factor but they have to fit everything. So you have to fit your
[29:04] everything. So you have to fit your silicon photonix transceiver. Uh you
[29:07] silicon photonix transceiver. Uh you have to have your micro electronics uh
[29:10] have to have your micro electronics uh drivers. You have to have your micro
[29:12] drivers. You have to have your micro electrical interfaces on one side and
[29:14] electrical interfaces on one side and optical interfaces on the other. So this
[29:16] optical interfaces on the other. So this is a very defined uh volume uh that you
[29:19] is a very defined uh volume uh that you have to fit everything in. Uh even if
[29:23] have to fit everything in. Uh even if you look at say packages that are more
[29:25] you look at say packages that are more let's say butterfly type again you have
[29:27] let's say butterfly type again you have limited area that you have to fit in you
[29:29] limited area that you have to fit in you have to work work with. However, as I
[29:31] have to work work with. However, as I mentioned at the very beginning,
[29:32] mentioned at the very beginning, packages typically are non-standardized.
[29:34] packages typically are non-standardized. So, it's very difficult right now to
[29:36] So, it's very difficult right now to come up with a standardized package. In
[29:38] come up with a standardized package. In Europractice, one of the limits is to is
[29:40] Europractice, one of the limits is to is to solve that issue. So, when you're
[29:42] to solve that issue. So, when you're designing a package, especially for for
[29:45] designing a package, especially for for integration into a bigger system, you
[29:47] integration into a bigger system, you have to be conscious of the housing
[29:49] have to be conscious of the housing restrictions as well as all the optical
[29:51] restrictions as well as all the optical electrical interfaces that that system
[29:53] electrical interfaces that that system allows you to interface with.
[29:56] allows you to interface with. Finally, we come to uh uh to
[30:01] Finally, we come to uh uh to the actual process of packaging. Uh as
[30:05] the actual process of packaging. Uh as you saw, we have a full 3D CAD model
[30:08] you saw, we have a full 3D CAD model built of our packaging machines and
[30:09] built of our packaging machines and those come with tools
[30:12] those come with tools uh that we have on the side that we use
[30:14] uh that we have on the side that we use for packaging. So for example, uh for uh
[30:18] for packaging. So for example, uh for uh handling fiber rays, we use let's say
[30:20] handling fiber rays, we use let's say wide area grippers. for handling smaller
[30:23] wide area grippers. for handling smaller elements like in this case a micro optic
[30:25] elements like in this case a micro optic we have these narrow tipped uh tweezer
[30:28] we have these narrow tipped uh tweezer grippers to handle them within the
[30:30] grippers to handle them within the packaging machine. So you have to always
[30:33] packaging machine. So you have to always find the optimum tool for for the job
[30:37] find the optimum tool for for the job and that's also part of design process.
[30:39] and that's also part of design process. It's not let's say the package itself
[30:42] It's not let's say the package itself design of the package itself but it has
[30:44] design of the package itself but it has to be uh the design of the process also
[30:47] to be uh the design of the process also is very good to have when you're uh to
[30:50] is very good to have when you're uh to incorporate very early into the into the
[30:52] incorporate very early into the into the overall package. How is this package
[30:53] overall package. How is this package going to be assembled when it's actually
[30:55] going to be assembled when it's actually sitting in the machine there of course
[30:58] sitting in the machine there of course uh much more sophisticated grippers. You
[31:01] uh much more sophisticated grippers. You can get ones with PZO mechanisms uh
[31:04] can get ones with PZO mechanisms uh which used for delicate small elements
[31:07] which used for delicate small elements and you have very precise control about
[31:10] and you have very precise control about the open and close positions. However,
[31:12] the open and close positions. However, these come at the cost of a much smaller
[31:14] these come at the cost of a much smaller gripping force. So again you have to be
[31:17] gripping force. So again you have to be conscious which grippers are going to
[31:19] conscious which grippers are going to use for which to for which purpose. Uh
[31:22] use for which to for which purpose. Uh there are also vacuum grippers uh that
[31:23] there are also vacuum grippers uh that you can have at the end at one end you
[31:26] you can have at the end at one end you have you input this uh vacuum tube and
[31:29] have you input this uh vacuum tube and on the other end there is actually a
[31:31] on the other end there is actually a small uh small hole through which the uh
[31:35] small uh small hole through which the uh the you can grip your in this case these
[31:38] the you can grip your in this case these are mainly used for ball lenses for
[31:39] are mainly used for ball lenses for example but uh any man of m of vacuum
[31:44] example but uh any man of m of vacuum grippers can also be used for other
[31:45] grippers can also be used for other jobs. This is again uh designing a tool
[31:48] jobs. This is again uh designing a tool for a specific job.
[31:50] for a specific job. A little bit of history. This is
[31:52] A little bit of history. This is something that we don't use anymore, but
[31:54] something that we don't use anymore, but it was used in about about 10 years ago.
[31:56] it was used in about about 10 years ago. Again, as an example of a specialized
[31:59] Again, as an example of a specialized use uh uh of of grippers in this case
[32:02] use uh uh of of grippers in this case for handling very tiny microoptical
[32:05] for handling very tiny microoptical components. This is a lens which I
[32:08] components. This is a lens which I believe is about 300 microns in length.
[32:12] believe is about 300 microns in length. So, it's a very very tiny component. You
[32:14] So, it's a very very tiny component. You can see it's the scale is difficult to
[32:16] can see it's the scale is difficult to appreciate. However, this is sitting
[32:19] appreciate. However, this is sitting just at the tip of these grippers. So,
[32:21] just at the tip of these grippers. So, what these are, this is a wire, very
[32:23] what these are, this is a wire, very thin wire shaped appropriately. And in
[32:26] thin wire shaped appropriately. And in this case, what you do is you dip this
[32:28] this case, what you do is you dip this wire in wax that wax that's typically
[32:31] wire in wax that wax that's typically used for b for temporary bonding uh of
[32:34] used for b for temporary bonding uh of wafers, for example, and you run current
[32:37] wafers, for example, and you run current through the leads on the other end. And
[32:40] through the leads on the other end. And uh when uh the wax is molten the Tg of
[32:43] uh when uh the wax is molten the Tg of this wax is about 140° but of course you
[32:46] this wax is about 140° but of course you can have different different types and
[32:48] can have different different types and once sorry when it's the molten state
[32:51] once sorry when it's the molten state you can actually make contact with your
[32:53] you can actually make contact with your micro optic shut off your current and
[32:55] micro optic shut off your current and when the wax device you create a
[32:57] when the wax device you create a temporary bond between your gripper and
[33:00] temporary bond between your gripper and the micro optical component. So when
[33:01] the micro optical component. So when you're dealing with a very very tiny
[33:03] you're dealing with a very very tiny elements, this sort of mechanism, this
[33:06] elements, this sort of mechanism, this sort of gripper can be used when no
[33:08] sort of gripper can be used when no other mechanical gripper will be able to
[33:10] other mechanical gripper will be able to handle such small elements. So general
[33:13] handle such small elements. So general the takeaway message from the design
[33:15] the takeaway message from the design part again this is quite broad because
[33:18] part again this is quite broad because you know there's a breadth of uh of
[33:20] you know there's a breadth of uh of different applications biomedical
[33:22] different applications biomedical applications will have different
[33:23] applications will have different considerations from data center which
[33:26] considerations from data center which will have different considerations from
[33:28] will have different considerations from uh uh from telecoms from sensors. So all
[33:31] uh uh from telecoms from sensors. So all these come uh all these have their
[33:34] these come uh all these have their specific requirements. However in
[33:36] specific requirements. However in general the design has to be seen as the
[33:39] general the design has to be seen as the most important part of any packaging
[33:41] most important part of any packaging project. because it assumes all aspect
[33:44] project. because it assumes all aspect of the package uh uh at which the pick
[33:46] of the package uh uh at which the pick is the center. So mechanical
[33:48] is the center. So mechanical integration, electronic connectivity,
[33:49] integration, electronic connectivity, optical connectivity and management you
[33:52] optical connectivity and management you it's very good to be uh to be conscious
[33:56] it's very good to be uh to be conscious of the capabilities of your packaging
[33:58] of the capabilities of your packaging machine uh and tooling that that you're
[34:01] machine uh and tooling that that you're going to use and sometimes tooling is
[34:02] going to use and sometimes tooling is even part of design. You cannot also
[34:05] even part of design. You cannot also forget about the system and environment
[34:07] forget about the system and environment onto which the package will operate
[34:08] onto which the package will operate especially if you're working towards a
[34:11] especially if you're working towards a some sort of a product. uh and uh
[34:14] some sort of a product. uh and uh something to be also conscious of that
[34:16] something to be also conscious of that it can take uh quite a long time uh from
[34:21] it can take uh quite a long time uh from from the idea to arrive at the final
[34:24] from the idea to arrive at the final design of the package in in more than
[34:27] design of the package in in more than two years is not uncommon especially for
[34:29] two years is not uncommon especially for quite sophisticated modern packages
[34:31] quite sophisticated modern packages which require numerous hundreds of uh uh
[34:34] which require numerous hundreds of uh uh DC connections, tens of optical
[34:36] DC connections, tens of optical connections. These are quite difficult
[34:38] connections. These are quite difficult to design and fit everything into a
[34:40] to design and fit everything into a small or comfortably small form factor.
[34:43] small or comfortably small form factor. So the design process is not easy and
[34:45] So the design process is not easy and has to be taken with uh seriousness and
[34:49] has to be taken with uh seriousness and uh however if it's taken such seriously
[34:53] uh however if it's taken such seriously then well-designed package will be more
[34:55] then well-designed package will be more resistant to any surprises. Essentially
[34:57] resistant to any surprises. Essentially you can we you can test your packaging
[35:00] you can we you can test your packaging process even as part of your uh of your
[35:03] process even as part of your uh of your design. If you have for example as we do
[35:05] design. If you have for example as we do the pack the packaging machine in the 3D
[35:08] the pack the packaging machine in the 3D CAD model. So this is uh it for the
[35:12] CAD model. So this is uh it for the first webinar in in regarding the
[35:14] first webinar in in regarding the package design uh