Integrated chip

(54) Title: Integrated chip (56) Published publications: WO 2004 034 530, US Al 2002 006 4201, US Al 2004 010 5476, US A 5 987 050 (47) Abstract:

Integrated chip for data, telecommunications and optical analysis for at least one, two or more wavelengths there, the chip can be used both to emit laser light and to detect laser light.

The invention is characterized in that the chip (1) comprises a waveguide (2) having a first port (3), that the waveguide (2) is expanded from the first port (3) in the direction of at least one second waveguide (4) and a third waveguides (6) having at least one second port (5) and a third port (7), respectively, which are placed parallel or at an angle and at a distance from one another and that the different components of the chip (1) are monolithically integrated.

25,

<img file="SE528653C2_D0001.tif" />

528 653

Summary

Integrated chip for data, telecommunications and optical analysis for at least one, two or more wavelengths there, the chip can be used both for transmitting laser light and for detecting laser light.

The invention is characterized in that the chip (1) comprises a waveguide (2) having a first port (3), that the waveguide (2) is exion panned from the first port (3) in the direction of at least one second waveguide (4) and a a third waveguide (6) having at least a second port (5) and a third port (7), respectively, which are placed parallel or at an angle and at a distance from one another and that the various components of the chip (1) are monolithically integrated.

Figure 1 is desired to be published.

h: \ docwoik \ corrected final submission.inl.doc, 2006-10-20

528 653

The present invention relates to an integrated chip for data, telecommunications or analysis applications where the chip can be used both to emit light from narrow band light sources such as lasers or broadband light sources such as LEDs and to detect laser light or light from other light sources, e.g. from LEDs or luminescent light from biological samples or the like.

io

In the case of data and telecommunications, lasers and LEDs are used to transmit light and photodetectors to detect light transmitted through light conductors. Typically, laser chips are used to emit laser light, which is modulated with the ice information to be transmitted, and chip photodetectors, respectively, to detect received laser light. Each such functionality usually includes different components.

There are also composite chips with several functions such as transmitting and receiving laser light of a number of wavelengths. In existing solutions, it is forced to use costly fine-mechanical alignment of several discrete components, such as separate laser diodes, LEDs, photodiodes, wavelength selective prisms, waveguides, etc. on a common carrier of e.g.

silicon to achieve a multi-function chip.

Thus, such known chips are therefore expensive, complicated to manufacture.

The present invention solves this problem and offers a chip that can be used for multiple functions, but which is simpler and cheaper to manufacture than known chips.

The present invention thus relates to an integrated chip for data, telecommunications and optical analysis for h: \ docwork \ rattat shitförelägging.inl.doc, 2006-10-20

528 653 at least one, two or more wavelengths where the chip can be used both to emit light and to be able to detect light, the chip comprising a waveguide with a first port, the waveguide being expanded from the first port in the direction of at least a second waveguide. and a third waveguide having at least a second port and a third port, respectively, which are placed parallel or at an angle and at a distance from one another and characterized by the chip having a basic structure, which is equal from a carrier and upwards across the entire surface of the chip, that said waveguides are formed at the top surface of the chip by etching the base structure so that protruding waveguides are formed and that the various components of the chip are monolithically integrated.

The invention is described in detail below, partly in connection with an embodiment of the invention shown in the accompanying drawings, in which

Figure 1 shows a sketch of a monolithic integrated chip according to the invention

Figure 2 shows a cross section through a basic structure according to the invention

Figure 3 schematically illustrates a function of an optical system in which the invention is applied

Figure 4 schematically illustrates another function of an optical system in which the invention is applied

Figure 5 shows an alternative connection

Figure 6 shows two interconnected components according to the invention.

The invention thus relates to a monolithic integrated chip in which several functions such as transmitting and receiving a number of wavelengths are integrated into a unit.

h: \ docwork \ corrected final submissionand.inl.doc, 2006-10-20

528 653

As a basic principle for achieving a method of monolithic integration directly in the material from which the sub-components are made, some of the performance of each component is sacrificed in the form of, for example, increased internal losses or limitations in modulation speed due to the common basic structure being adapted to all components included. Probably the basic structure is not optimized for the function of any individual subcomponent.

However, the added losses etc. do not completely eliminate the functionalities, but only limit them somewhat so that the remaining performance is sufficient to meet technical requirements in a variety of applications.

The present invention thus relates to an integrated chip for data, telecommunications and optical analysis for at least one, two or more wavelengths where the chip can be used both to emit light and to be able to detect light in a wide range, e.g. 1480-1600 nm for data and telecommunications or other spectrum for optical analysis.

Figure 1 shows schematically a monolithic integrated chip according to the invention, prior to metallization, with switching waveguides, MMI couplers (Multi Mode Xnter ferrometer) and Distributed Feedback (DFB) lasers. According to the invention, the chip 1 comprises a waveguide 2 with a first port 3 into or out of which light is intended to be guided. The waveguide 2 is expanded from the first port 3 in the direction of a second waveguide 4 with a second port 5 and a third waveguide 6 with a third port 7, which ports 5, 7 are placed parallel or at an angle and at a distance from one another. The gates 5, 7 are arranged to guide light in and out. According to the invention, the various components of the chip 1 are monolithically integrated.

h: \ docwork \ nittat finalization.ml.doc, 2006-10-20

528 653

According to the invention, the expanded portion of the waveguide 2 after the coupler can open in more than two ports.

According to a preferred embodiment, each of the second 4 and the third 6 waveguides together with a respective grid 9, 10 is a respective RWG-DFB laser, which lasers are tuned to different or equal wavelengths.

To detect incoming light, it is preferred to provide photodetectors 25, 26, such as photodiodes, instead of lasers or LEDs, or photodetectors located after the DFB lasers that monitor the status of the lasers or LEDs over the life of the component.

According to another preferred embodiment, LEDs are provided at the first waveguide 2, the second 4 and / or the third waveguide 6.

According to a preferred embodiment of the invention, the expanded portion 8 of the waveguide comprises a wavelength selective coupler to achieve simultaneous bidirectional functionality. In this case, wavelength selective switches of the type MMI, evanescent, array waveguide couplers etc. can be used, which theoretically gives no power loss since all light of a certain wavelength goes to the port where it is to be detected and vice versa for transmission.

However, a Multi Mode Interferometer (MMI) coupler is preferred because it is production stable and can be made polarization independent. Furthermore, it can handle multiple wavelengths and the number of ports can be varied. It is also possible to manufacture at low cost. Even an evanescent coupler, with or without a distance between the input waveguides, has similar advantages as the MMI coupler, but is also completely without internal h: \ docwork \ corrected final presentation.inl.doc, 2006-10-20

528 653 reflection from the gates. The evanescent coupler may consist of two or more waveguides.

An MMI waveguide 8 allows the incoming fundamental mode to expand to a combination of several higher order modes and the fundamental mode. The interference pattern that arises depends on the thicknesses and refractive indices in the basic structure and other materials added in the process as well as the physical dimensions of the MMI waveguide. It is possible to design in a 1x2 MMI in such a way that at a certain distance from the input (corresponding to the length of the MMIzn), the two incoming wavelengths have super positions maximum at different locations (relative to the MMIzn width) where the respective output port is placed. The MMIzn and the other switches may have a symmetrical function, ie if the output ports function as optical inputs, the light from both input ports will search for the only output port. MMIzn and the evanescent coupler can be designed as polarization independent which is of great importance. Furthermore, they can be made compact, which is important for low-cost applications and to reduce optical absorption.

In an alternative embodiment, the expanded portion 8 of the waveguide comprises a so-called star coupler. A star coupler expands the optical mode received from the fiber so that two ports can be placed next to each other. Variants of DFB filters or other filters are placed at the parallel ports so that only one wavelength can pass per port. This is a compact solution but much of the incoming optical power is lost in losses.

In the case where the expanded portion 8 of the waveguide includes a star coupler, a wavelength selective filter not shown is formed at the second and third port 7, respectively.

h: \ docwork \ rated final submission.inl.doc, 2006-10-20

528 653

Figure 2 shows a basic structure according to the invention. In the example in Figure 2, the basic structure is processed into an RWGlaser.

On a substrate of InP (or GaAs), epitaxial layers are made with MOCVD or MBE.

From the top down, the base structure comprises a Ti / Pt / Au p contact 11, a p-InGaA contact layer 12, a SiNx or SiO_x two layers 13, one p-InP cladding layer 14, one layer of p-SCH: InGaAsP or AlInGaAs 15, one layer 16 of quantum wells 24 surrounded by barrier layers or both replaced by a bulk layer

23rd To the right of Figure 2, layer 16 is shown on a larger scale, where it can be seen that layer 16 comprises 1 to 24 alternating quantum wells and barrier layers for generating a wavelength, for example 1310 nm or alternatively 1 to 24 alternating quantum wells and barrier layers or both layers replaced by a bulk layers for generating or detecting a second wavelength, for example 1550 nm in the band 1480-1600 nm. In order to obtain a polarization-independent photodetector, a bulk layer can be used with great advantage for equal sensitivity to TE and TM modes of the incoming signal. Alternatively, a polarization-independent photodetector can be obtained using stress optimization of the quantum well package or bulk layers so

The TE and TM modes are similarly absorbed in the photodetector. Further, the basic structure comprises a layer of n-SCH: InGaAsp or AlInGaAs 17, followed by a buffer layer 18 of nInP, for example, the thickness of 0.5 µm, and an n-cladding layer, for example with the thickness of 1 µm. Below is an n-InP substrate 19, an n-contact 20 of, for example, 500 Å Ti-W, followed by Au with, for example, a thickness of 1000-4000 Å. Further, there is a solder 21 and a carrier 22 of, for example, CuW, A1N or similar. The thickness of layers 15 and 17 can be 1.1 pm, at h: \ docwork \ corrected final presentation.inl.doc, 2006-10-20

528 653 stored 19 80-120 pm. The width of the waveguides 2, 5, 7 can be 5-15 pm.

The SiNx / SiOx layer can be excluded and the etch depth can vary from β-cladding to n-cladding and re-cultured or filled in with PCB ™ or similar polymers. For an MMI, the same process is used but the waveguide width can be up to 100 µm wide.

A basic structure according to the invention can be manufactured by conventional methods such as Metal Organic Chemical (MOCVD)

Vapor Deposition), MBE (Molecular Beam Epitaxy) and the number of subsequent process steps can be kept to very few if sufficient functionality of the subcomponents can be reached with a common etch depth. More etch depth is of course applicable and can improve the system's performance, but expensive manufacturing something which can be an obstacle if you are interested in extremely low manufacturing costs on a par with eg. CDROM lasers. However, deeper etching depth may be required for optical and electrical insulation between the included subcomponents.

Said basic structure is covered with resist and a pattern is defined by E-beam, lithography or interference pattern and then dry or wet etched into a waveguide structure.

The etching can be terminated in p-contact layers, p-cladding, active waveguide or n-cladding.

The structure can then possibly undergo a re-culture with resistive InP, GaAs, InGaAs, InGaAsP, InAlGaAs or possibly covered with PCB or similar polymer partially up to or above the edge of the waveguides or after new resist coverage where contact surfaces (pads) are defined, covered with about 500 Å Ti, 500 Å Pt and 1000-4000 Å Au via evaporation or sputtering and then undergo metal lift-off processing.

h: \ docwork \ nittat finalization.inl.doc, 2006-10-20

528 653

The structure can optionally be covered with silicon nitrides, silicon oxides or other dielectric material and then the silicon nitride etc. is opened so that subsequent metallization directly contacts the semiconductor in selected areas but is blocked from contact in other areas.

The structure can then be thinned down from the back and polished to a total thickness of 80-120 µm before the back is metallized with sputtering (evaporation) with 500 Å Ti-W alloy and io 1000-4000 Å Au. Optionally, the back may be provided with a 40008000Å thick eutectic alloy of Au-Sn by evaporation or sputtering.

In principle, there are three options for the chip's execution.

A first alternative is based on a basic waveguide, the so-called ridge wave guide (RWG), which gives a weak connection of the light to the waveguide. The advantage is that the method is used for mass production of e.g. 980 nm pump lasers and that the sensitive active layer is not etched through. In addition, there are good results with a variant of RWG lasers that do not have SiN_x or SiO_xdielectric between contact metal and semiconductor and where the current can in principle flow from metal to semiconductor in etched areas, ie on the side of the waveguide. In practice, however, the current only enters a high-doped β-contact layer at the top of the waveguide, ie the part which is not etched.

Furthermore, it is likely that the stress that is often found in SiN_x

- or SiO_x- over time, the semiconductor material breaks down and degrades components. Very stable components have been presented without intermediate dielectrics. A major advantage of achieving low manufacturing costs is that you avoid delicate process steps where openings must be made in SiN_xh: \ docworic \ corrected final submission.inl.doc, 2006-10-20

528 653 or SiO_x-materials to enable contact with the semiconductor in the right place. Alternatively, only a significantly more insensitive step is left, namely that contact pads are defined with an accuracy of +/- 2-10pm in a lithography step compared to an accuracy of +/- l-2pm if contact openings are to be made.

In addition, the number of steps where there is resistance to the surface to be covered later with Ti-Pt-Au or the like, where resistance-related residues often provide increased contact resistance or even deteriorate the yield drastically because they are difficult to detect during manufacture. they are transparent and only about 10-50 Å thick.

A second alternative is a standard RWG process according to a 980 nanometer type of pump laser with SiNx / SiOx or PCB or other polymer as a dielectric. This option is more expensive than the first option, but gives slightly higher performance.

A third alternative is to make a deeper etch through the active waveguide and re-cultivate the components with resistive InP by MOCVD or HVPE. This option provides better component performance, but is also the most expensive manufacturing option, but also provides the highest performance.

The present integrated chip can be manufactured according to all alternatives.

Etching can be done with dry or wet methods and the mask definition can be done with lithography or electron beam drawing. DFB and Distributed Bragg grating (DBG) grids can be defined by electron beam drawing or optical interference. The most direct manufacturing method is to use electron beam drawing to define both waveguide types and DFB and / or DBG grids simultaneously and combine it with a dry: \ doework \ corrected final presentation.inl.doc, 2006-10-20

6γ ·

Rets ο legal process etching out both waveguides and DFB and / or DBR grids in the same process step.

However, it cannot be ruled out that suitable wet methods can be used.

In terms of design, the MMI and the evanescent coupler are flexible and can be manufactured in a variety of variants, with varying numbers of ports and wavelength sensitivity. E.g. we can imagine the commercially interesting case where our case is added with a port receiving signals at 1490 nm and a photodiode is placed at that port, in the same way a triplexes are built up and usually sit in homes and receive data at 1550 nm, transmit data at 1310 nm (both digitally coded) and also receive analog coded cable TV at 1490 nm. Similarly, the chip can be manufactured for the second link page, which transmits at 1550 and 1490 nm and receives information at 1310 nm.

The combination of facet-free lasers with cavities completely defined by lattice and light that is directly passed to variants of MMI waveguides, directly aligned to each other in the production, opens up opportunities for other applications as well. E.g. with an IxN MMI with N outputs connected to N pcs. lasers with varying laser wavelength λ_Ν, according to e.g. The ITU channels are probably more stable map wavelength operation compared to alternative solutions. In addition, all wavelengths are always available, and thus the time for changing wavelengths is basically zero in tunable applications. Alternatively, N channels can be used simultaneously, or N / 2 channels simultaneously and N / 2 channels act as backup options. MMI and evanescent couplers etc. should also be able to be cascaded if there are advantages in separating wavelength or h '. \ docwork \ corrected final order.inLdoc, 2006-10-20

528 653 waveguide management in units that can be designed / manufactured for more specific requirements than a larger MMI or other coupler.

Figure 3 shows an example of the use of the present chip 1. A single mode fiber 27 conducts light of two wavelengths

1310 nm (or the like in the band 1260-1360 nm) respectively

1550 nm (or similar in the band 1480-1600 nm) to and from the first port 3. The light is selected in the MMI 1 so that the wavelength 1310 nm comes to the waveguide 4 while the wavelength io 1550 nm comes to the waveguide 6. Similarly, the wavelength is generated. 1310 nm with the laser in the waveguide 4 and light with the wavelength 1550 nm can be generated with the laser in the waveguide 6. The most common use should be transmission on a wavelength and reception on a wavelength.

In Figure 4, the same is shown in principle as in Figure 3, but in Figure 4, the MMI 1 has been performed with three parallel waveguides 28,

29, 30 designed as lasers or photodetectors. For example, incident light in the single mode fiber 27 may have two different wavelengths, for example 1490 and 1550 nm. By means of the MMI, the three different wavelengths are selected for each of the three waveguides 28, 29 and 30 of which transmission is at 1310 nm (alternatively transmitting at 1490 and 155 nm and receiving at 1310 nm).

Figure 5 illustrates a connection where incident light, in a single mode fiber 2 having four different wavelengths, is divided into a first MMI 30 without gratings in the waveguides 34, 36 such that two of the wavelengths are selected to a first waveguide 34 and the remaining two wavelengths are selected to a second waveguide 30. A waveguide 32 directs the light to a second MMI 31 and a waveguide 33 directs the light to a third MMI 32. In the respective MMI 31 and 32, the light is divided so that light of only one wavelength is selected for each waveguide 37 - 40. The waveguides 37 - 40 may include lasers for the current wavelength. Because h: \ docwork \ drawn final submission.inl.doc, 2006-10-20

653 the switches are recursive, each port can be configured as transmitter or receiver.

Figure 6 shows two interconnected components according to the invention. This can be accomplished by utilizing two components 1 manufactured into one unit prior to splitting for control measurement directly on the wafer, which lowers the manufacturing cost.

A number of embodiments have been described above. It will be obvious that these can be modified by one of ordinary skill in the art to be adapted to the current application.

One application is biological analysis of samples, where optical illumination takes place with a broadband light source (LED 11002000nm) or one or more narrowband (laser) sub-spectra where reception of luminescent light from the sample is conducted via a wavelength selective coupler to unique photodetectors. Other types of couplers may conceivably be applicable as multi-channel wavelength dividers so that the wavelength selection increases to more ports. The reception range for photodetection and light emission via LED or laser (1100-2000 nm) can be extended by subband absorption of the photodiode and cascade coupling in the active region of LED and lasers (&gt; 2000 nm).

Therefore, the present invention is not to be considered limited to the above embodiments, but may be varied within the scope of the appended claims.

h: \ docwoik \ rated final submission.ml.doc, 2006-10-20

528 653