Expert Session: Optical Fiber Coupling to Photonic Chips

uh welcome everybody good afternoon and thank you henning for this introduction and i'm proud to begin this series of webinars on glass with trust and the topic of the today's webinar is optical fiber coupling to photonic chips and i will show you what is the challenge in it and to begin with i will present the some theoretical basics of coupling efficiency i will introduce the quantity of mod field diameter and at the end i will show experimental methods how to measure the mod field diameter and experimental methods how to efficiently couple five optical fiber chips into optical photonic chips and i will restrict myself to um bad coupling like in this picture in the background but first let us begin with an overview with the historical overview how the integration density of electro-optical systems evolved over time um so in the 19th century and in the in throughout the most of the 20th century the electro-optical systems have been built on optical benches they were they are large and heavy but they are very stable and actually they are still very useful if you want to build a demonstrator or a prototype and i have here such a demonstrator just to show you what is an optical bench and this is a magenta interferometer well known from from the school from the um from the physics laboratories and you see that it's quite large and in the last decades or years also in our groups at the front of variety time we made some progress in miniaturizing such electro-optical systems and we do it by integrating optical systems into micro benches on glass and i have also here a demonstrator this is also a martender interferometer integrated onto a optical micro bench and you see how much smaller it is compared to the large optical bench but the future of micro integration belongs to optical chips well this is a silicon photonic chip which also um contains a maxender interferometer and the phase difference of the between the arms of this interferometer is controlled electrically and it can be used to switch the the output of the chip on and off and with a rate of up to 40 billion times per second this is uh of the application is obviously the telecommunication and since these chips are so tiny um there's a challenge to electrically and optically interconnect them to surrounding devices and optical interconnections will be the topic of my talk today so first let's have a look at the physical dimension of the problem so optica here we see a cross-section of optical fiber and of waveguide waveguide on the chip the fiber is 120 micrometer uh in diameter and the light is guided in a 10 micrometer with a thin core and uh however the white waveguide on silicon chip is only 400 nanometer wide and we want to couple one to another and to be consistent let me throw the wave guide on the chip uh in a proper scale this is how small it is so there is a large dimensional mismatch between the optical fiber and the chip integrated waveguides and the questions that arise are what is the coupling of each efficiency in case of edge but coupling what does it depend on and how can we can it be optimized and to answer these questions we need a little bit of mathematics and let's begin with a light with a light distribution in inside of these waveguides so suppose that the light is propagation in z direction which is perpendicular to the screens to to your screens and um i suppose that we know the field distribution of both the fiber and the waveguide in the xy plane it is e1 and e2 of x and y and then the coupling efficiency is given by this mod field overall overlap integral it looks ugly but let's have a look what is uh what is it composed of so in the denominator we have just a product of intent of light intensities and this is the normalization factor and in the numerator up there there's an overlap integral of the electric fields and this gives us an idea of how good the fields are overlapped now let's look at some properties of the overlap integral um first the overlap integral regards the electric fields and not the intensities this means that the mode matching the coupling efficiency depends not only on the intensity distributions but also on the optical phase so in case when the phase profile is not completely flat and this is the case when we don't have the bad coupling but some offset then the up the efficiency will be um affected by this not flat phase profile uh how to uh solve this integral so if we have a good computer or a good student this integral can be solved numerically and it can be solved numerically for any field distributions but to make life easier there are analytic solutions for some special cases and this special case is for instance a gaussian field profile so let's look at gaussian field profiles let's assume that the e1 and e2 are both gaussians they are characterized by mod field diameters here it is a general case of elliptic field profiles and so we have multiple diameters in x and y directions and um actually this assumption that the filter our gaussian is a very good approximation for the ground modes of most of the single mod wave single mod wave guides and then for for the gaussian fields the mod field integral is analytically solvable and the result is that the ether the coupling efficiency depends on the transverse offset on the longitudinal offset between the coupling partners and on the mod field diameters so the simplest cases is the bat coupling without any offsets so we have delta x data y delta z equal to zero and then the eta zero this coupling efficiency depends on only on the mod field diameters one and two and you can easily verify that if you set mfd1 equal to mfd2 the eta0 will be one now if we have an offset delta x then eta0 will be corrected by an exponential factor e to the minus delta x square and if you are interested in further cases uh uh including the longitudinal offset and also a tilt angle between the waveguides i refer to this to this paper but now let's go back to the um to the transverse offset and this is this transverse offset gives us also a basic for the first measurement measurement method so we couple the peak with a source fiber we detect with the detection fiber and we have a perfect coupling but coupling so delta x delta y delta z is equal to zero and now we begin to scan the source fiber in the x y plane and we record the intensity profile with the detecting fiber and so we have this intensity profile which is actually measured eta and now if the mod field diameter of the source fiber is known and we measured eta then from the formula that i've shown in the last slide we can determine the unknown mod field diameter of the waveguide on the chip this is a actually quite a complicated method and quite advanced but it is a very practical since the measured eta is in the same times your tolerance curve tolerance it gives you an idea of how tolerant is your systems for misalignments that can occur during assembly process for instance and here you see an example of such coupling curves and you can recognize that just as one micrometers of micrometer of misalignment is enough to drop the coupling efficiency by 20 percent so next method is if we don't know the mod field diameter of the fiber of the source fiber or or both of the fiber and then on the of the peak we must determine it with some other method and one of these method is a far field angular angular scan methods and it is also implemented at front hoover isotherm the method base is basing on the angular scan of the intensity profile far away from the tip so we measure the intensity as a function of the angle theta you can do it when in both direction here you see an example of such a measurement and once you have the e of theta you can calculate the mod field diameter by using this uh known uh peterman two integral and this method is actually very popular in fiber community it was in invented for the fiber industry and it is recommended by the telecom industry association and this is described by an industrial norm named here but since it was developed for the fibers it works only for cylindrical symmetric field distributions and now we know both mode fields diameters of the fiber and the core and the question is what is the coupling efficiency in this particular case and it is no surprise that it is very small it is actually less than one percent so we need to think now of how can we increase it so the first step to increase the um the mod field diameter of the waveguide on the chip happens on the chip itself so there are chip structures called inverse papers and this inverse taper expands the the beam up to five to six micrometers and with this and it expanded such that the face front of the expanded beam is perfectly flat on the chip facet and with this the coupling efficiency can be increased up to 50 percent and we then have to bridge the residual 50 percent with some methods outside of the chip so if you think of changing the beam size you usually take lenses and because we have here tiny fiber and tiny waveguides on the chip we don't we are very limited in space so we need to necessarily take micro lenses so the advantage of taking lenses to change their beam size is that first with lenses with the system of two lenses you can create a short section of a collimated beam and in this collimated beam you can place a other optical element like a filter or like optical isolator and this is necessary for some applications another advantage of lens is the of micro lenses is their skull scalability micro lenses are available as lens arrays and uh which means that with a with a single adjustment step you can couple a parallel multitude of parallel waveguides into parallel pitch of waveguides on a tip but the drawback is that this alignment process is quite sophisticated now the second method is to modify the facet of the fiber and the facet of the fiber can be modified in such way that the fiber itself is a lens these are so called lens fibers this lens fiber can focus the light to a spot size spot size down to 2 to six micrometers and the working distance the distance between the fiber facet and the focus is typically five to twenty micrometers so these fibers cannot be bad coupled to the chip but they have to work with with the gap and this is air gap because the fiber is a lens and it is optimized to work in a surrounding with a refraction index um which is usually different than that of the refraction index of glass so you cannot fill the space between the fiber and the chip with adhesive to fix the fiber so in order to fix the fiber we in our group we developed some special holders these holders are made of glass and with these holders you can hold the fiber you can adjust it before the coupling and finally to fix it so this lens fibers are are available as polarization maintaining fibers so which is important in some applications now next point is modified fibers ultra high and a fibers these fibers have a very thin core um thinner than a standard that the standard fiber does have and other in other words they have a high numerical aperture that's why they are called ultra high and a fibers and the mode field diameter of this fiber is between 3 and five micrometers which fits very good to the silicon botanic chips but these fibers are quite expensive so you don't want to use kilometers of them to to send the data uh however uh there is it's enough that a short piece of this fiber as short as one centimeters is enough to adiabatically convert the the mode field from a standard fiber to the one fitting to the chip and uh connection between the standard fiber and ultra high na fiber is done with in a special process of fusion splice and the splice loss in this process is as low as 0.1 db which is acceptable for most of applications these fibers are also available as single mod fibers a single mod fiber arise however they are not polarization maintaining and this is missing so and we at front of our isa time we constructed a custom ri of polarization my timing and ultra high ni fibers so we spliced a short piece of ultra high and a fiber to the polarization maintaining fiber and finally we built an ri out of it and the last possibility that i present is to use some external chip spot size converter chip these chips contains wave guides and these wave guides are tapered which means that on the entrance facet they are matching the mode profile of a single mod fiber of a standard fiber they are adiabatically converting the beam profile down to three to four micrometers which are matching the mod profile of the chip and these chips can be used both are as spot size converters and as pitch converters you can begin with a broad pitch and you can adiabatically narrow the distance between the waveguides and these are this kind of spot size converters are available commercially and we used these waveguides to coupling these two two couplings to photonic chips and we first attached a standard fiber right to this spot size converter and finally this assembly was held in in a special glass holder and all together was attached to a silicon chip so to summarize uh we learned that a transverse offset method delivers information about the mode profile on the chip facet and if the mod profile of the probe fiber is known if this is not the case we need to use other methods like for instance far field angular scan method but this method can be used only for cylindrically symmetric beam profiles and finally if we if we know the mod field diameters of the fiber and on the chip then we can use diverse methods to match them together so and the methods are internal chip internal or external spot size converting structures this can be also micro lenses or modified fibers so finally i make a loop and i go to the title of this webinar series in glass we trust and i would like to advertise the future webinars uh which will give you a more detailed overview of how we use glass for photonic packaging some examples you saw already on in this webinar like i showed you some tiny glass holders but i would like to announce already now the future permit webinar which will show you details of photonic assembly with and on glass and the second important topic is glass integrated waveguides i was focused on wave guides in silicon but we in our group are integrating waveguides into glass and i have here a piece of glass which looks probably transparent but you have to believe me that there are waveguides on this glass which means that we can build photonic chips out of glass and this photo these waveguides together with some electric functionality they are basic of the so-called electro-optic circuit boards and also we are happy to present it to you in more details in one of the future webinars so stay with with us in contact we are happy to um to have you as an audience of our future webinars i'm happy to hear questions now in the live discussion or you can send me or to hang shredder and email in anytime after the webinar thank you

Channel: Fraunhofer IZM

Views: 9,269

Keywords: photonics, optical waveguides, fiber-to-chip coupling, technology, microelectronics

Id: rHcfemtfP2s

Length: 23min 54sec (1434 seconds)

Published: Wed Nov 04 2020