GB2372831A

GB2372831A - Integrated optical switch

Info

Publication number: GB2372831A
Application number: GB0105188A
Authority: GB
Inventors: Ivan Evans
Original assignee: Bookham Technology PLC
Current assignee: Lumentum Technology UK Ltd
Priority date: 2001-03-01
Filing date: 2001-03-01
Publication date: 2002-09-04
Also published as: WO2002071141A1; GB0105188D0

Abstract

An integrated optical switch comprises a pair of waveguides 14,16 on a substrate 10, a p-n junction across the pair of waveguides, being defined by p and n-doped regions 18,20 in the substrate, one on each side of the pair of waveguides, the arrangement being such as to enable an asymmetric current density to be applied across the respective waveguides. Typically, this is achieved by defining said regions substantially asymmetrically about the pair of waveguides. The arrangement can be asymmetric in respect of the location and/or the shape of the doped regions. The asymmetry can exist longitudinally with respect to the waveguides, or transverse thereto, or in other ways. Suitable asymmetries include differences in the length of the doped region on the substrate as between the n and p regions, measured longitudinally relative to the waveguide, or differences in the depth as between the n and p regions, or differences in the width of the doped region on the substrate as between the n and p regions, measured transversely relative to the waveguide, or differences in the horizontal or vertical location of the doped region on the substrate as between the n and p regions, and/or combinations of these. A series of p and n-doped regions can be provided, potentially in a symmetric layout, and driven selectively and asymmetrically so as to achieve the same or a similar effect. An an integrated optical switch is also referred to which comprises a pair of waveguides on a substrate, a p-n junction across the pair of waveguides, being defined by p and n-doped regions in the substrate, one on each side of the pair of waveguides.

Description

Integrated Optical Switch The present invention relates to an integrated optical switch. This includes digital optical switches, often referred to as DOSs.

Digital optical switches are (generally) integrated optical devices formed by at least two waveguides, one or both of which can be perturbed. When the structure is perturbed, the optical signal can be routed between the output waveguides. A structure that allows a signal to exit via one of two output waveguides is referred to as a 1 x2 switch. In this way, 1 x3 and 2x2 switches can also be formed; a 1 x3 switch routes a signal to one of three output waveguides whereas a 2x2 switch has 2 inputs and 2 outputs and, in effect, selectively exchanges signals between two waveguides.

This application relates to the general case of a DOS structure, ie that formed (typically) by two gradually separating waveguides, one or both of which can be perturbed, hence enabling the signal to be routed.

A straightforward way of perturbing one half of a twin waveguide is to provide a doped region on top of the waveguide which then forms part of a p-i-n junction which can be used to inject carriers, perturbing one of the waveguides.

The other half of the junction is most likely located in the slab area. The doped region on the waveguide can however introduce losses due to absorption of light in the doped region and the effect of the metal contact layer leading to the doping.

Doping between the waveguides can also be difficult and is potentially lossy.

The present invention seeks to provide structures in which perturbations of this type can be introduced, without requiring doping on or near the waveguide.

In this application, a"switch"is an arrangement adapted to cause at least part of an optical mode to be transferred between waveguides.

The present invention therefore provides an integrated optical switch comprising a pair of waveguides on a substrate, a p-n junction across the pair of waveguides, being defined by p and/or n-doped regions in the substrate, one on each side of the pair of waveguides, the arrangement being such as to enable an asymmetric current density to be applied across the respective waveguides.

Typically, this is achieved by defining said regions substantially asymmetrically about the pair of waveguides. The arrangement can be asymmetric in respect of the location and/or the shape of the doped regions. The asymmetry can exist longitudinally with respect to the waveguides, or transverse thereto, or in other ways.

Suitable asymmetries include differences in the length of the doped region on the substrate as between the n and p regions, measured longitudinally relative to the waveguide, or differences in the depth as between the n and p regions, or differences in the width of the doped region on the substrate as between the n and p regions, measured transversely relative to the waveguide, or differences in the horizontal or vertical location of the doped region on the substrate as between the n and p regions, or differences in the waveguide surroundings such as by selective etching of the slab or waveguide region, and/or combinations of these.

A series of p and n-doped regions can be provided, potentially in a symmetric layout, and driven selectively and asymmetrically so as to achieve the same or a similar effect.

Alternatively, or in addition, the relevant area can be additionally doped with gold in a selective manner. Gold doping limits the diffusion length and, if asymmetric, can create the required asymmetry.

The p and n regions should be separated by less than the diffusion length, typically 20pm. If their spacing is significantly greater then the effect of the asymmetry is likely to be insignificant at the waveguides.

The present invention also relates to a an integrated optical switch comprising a pair of waveguides on a substrate, a p-n junction across the pair of waveguides, being defined by p and n-doped regions in the substrate, one on each side of the pair of waveguides. This novel arrangement permits effects such as the above to be obtained without the need to dope between the waveguides.

Embodiments of the invention will now be described by way of example, with reference to the accompanying figures, in which; Figure 1 is a view from above of an optical switch according a first embodiment of the present invention; Figures 2 to 5 are vertical sections through optical switches according a second to fifth embodiments of the present invention, respectively ; and Figure 6 is a view from above of an optical switch according a sixth embodiment of the present invention.

The figures illustrate a number of potential arrangements which fall within the present invention. It should be noted that the asymmetries illustrated can be combined as between figures, for example. In the drawings, a width is the dimension transverse to the waveguide, a length is the dimension along the waveguide, and a depth is the dimension perpendicular to the surface of the generally planar substrate. The associated conductors bringing current to the n and p doped areas are not shown for clarity reasons, but in practice they will need to be provided.

In each figure, an silicon on insulator (SOI) substrate 10, comprising an epitaxial layer of silicon separated from a support substrate (which may be of silicon) by an insulating layer 11, eg of SiO2, is provided in which one or more optical modes 12 propagate, guided by rib waveguides 14,16. In general, the majority of the mode lies in the epitaxial Si layer of the SOI substrate 10 beneath the ribs of the waveguides 14,16. A p-n junction is defined by a p doped area 18 and an n-doped area 20; when a current is passed through the p-n junction (or more properly, p-i-n junction) thus formed in the SOI substrate, the free carriers injected into the Si layer affect the refractive index and thus perturb the optical mode.

When properly controlled, this can act to switch the mode between waveguides.

Such an arrangement is described, for instance, in US 5757986.

The use of this effect to perturb waveguides is well characterised and models exist for predicting the behaviour of a proposed device. However, it is relevant to this invention that the perturbation of each waveguide must be different. Normal practice is either to form a p-n junction on the waveguide, ie with one doped region on the rib and the other beneath it or adjacent, or to form separate p-n junctions straddling each waveguide. The former practice leads to optical losses since the dopant on the rib on the metal contact lead to absorption of light. The latter is difficult to fabricate accurately since the waveguides are (at this point) close to each other so even if the doping process is accurate the centre region lies in close proximity to the optical mode. Therefore, any doping here is still likely to lead to optical losses. If the doping is less accurate, some of the dopant intended for the area between the waveguides may fall on one or both waveguides and again lead to absorption.

In figure 1, the p region 18 is a similar width to the n region 20 but is of lesser length. Thus, carriers flowing between them and spreading as they travel will present a different distribution in terms of shape and density at each waveguide.

In figure 2, the doped regions are of substantially the same length but the n region is of a greater width. This will distort the current flow pattern and again result in a different current density across the two waveguides.

In figure 3, the doped regions are of similar lengths and widths but the n region 20 is of a greater depth. As a result, the current flow pattern will be deeper in the Si epilayer beneath waveguide 16. The current density across the optical mode 12 associated with the waveguide 16 nearer the n region 20 will therefore be lower.

In figure 4, an arrangement similar to figure 2 is shown but the p region 18 is buried in the Si epilayer. This will draw the current density down in a similar fashion to the arrangement of figure 3 so the current density across the optical mode 12 associated with the rib waveguide 14 nearer the p region will therefore be lower.

In figure 5, the p region18 is formed by doping the appropriate area of the substrate 10, but after that area has been etched back to form a pit 22. Thus, the p region is deeper than would otherwise be the case since it extends downwards from the base of the etch pit 22. This will therefore have a similar effect to the arrangement of figure 4, but with easier access to the p region 18.

Figure 6 shows a sixth embodiment in which 4 p-doped regions 24a-24d are provided on one side of the pair of rib waveguides and 4 n-doped regions 26a-26d are provided substantially symmetrically on the other side of the pair of rib waveguides. Separate connections are made with each region, to allow each of the p regions (for example) to be driven independently. Thus, this arrange permits an asymmetric current density to be applied across the respective waveguides, by (for example) driving p regions 24b and 24c but not p regions 24a and 24d, and all n regions. In this case, the driven p regions will be asymmetric compared with the driven n regions.

In this example, there are 4 of each type of region and they are arranged in a substantially symmetric fashion. Other arrangements are of course possible, such as a greater or lesser number of regions on one or both sides, or a substantially asymmetric layout which nevertheless has a plurality of n and/or p-doped regions on either side of the pair of waveguides.

It will of course be appreciated that many variations may be made to the above described embodiments without departing from the scope of the present invention. Some such variations are described above; others will be apparent to the skilled person. For example, the regions either side of the structures could be etched back to the insulating layer 11 in order to confine the carriers. The relative device dimensions are illustrative only and can be adjusted to the specific requirements of the example. The overall dimensions have not been shown as these will typically be reduced as far as possible to optimise performance, the reduction improving with time.

Further, the specific example shown is based on a p-n junction, although other structures are possible such as FET (field effect transistors) and the like, all of which enable a current to be passed through the optical modes.

Claims

CLAIMS 1. An integrated optical switch comprising a pair of waveguides on a substrate, a p-n junction across the pair of waveguides, being defined by p and/or n doped regions in the substrate, one on each side of the pair of waveguides, the arrangement being such as to enable an asymmetric current density to be applied across the respective waveguides.
2. An integrated optical switch according to claim 1 in which said regions are defined substantially asymmetrically about the pair of waveguides.
3. An integrated optical switch according to claim 2 in which the arrangement is asymmetric in respect of at least one of the location and the shape of the doped regions.
4. An integrated optical switch according to claim 2 or claim 3 in which the asymmetry exists longitudinally with respect to the waveguides.
5. An integrated optical switch according to any one of claims 2 to 4 in which the asymmetry exists transversely with respect to the waveguides.
6. An integrated optical switch according to any one of claims 2 to 5 in which the asymmetry is at least one of; a difference in the length of the doped region on the substrate as between the n and p regions, measured longitudinally relative to the waveguide; a difference in depth as between the n and p regions; a difference in the width of the doped region on the substrate as between the n and p regions, measured transversely relative to the waveguide; a difference in the horizontal location of the doped region on the substrate as between the n and p regions; a difference in the vertical location of the doped region on the substrate as between the n and p regions.
7. An integrated optical switch according to any preceding claim in which a series of p and n-doped regions are provided.
8. An integrated optical switch according to claim 1 in which a series of p and n-doped regions are provided in a symmetric layout, adapted to be driven selectively and asymmetrically.
9. An integrated optical switch according to any preceding claim in which the area of the doped regions is selectively doped with gold.
10. An integrated optical switch according to any preceding claim in which the p and n regions are spaced by a distance less than the diffusion length in the material.
11. An integrated optical switch according to any preceding claim in which the p and n regions are spaced by a distance less than 20, um.
12. An integrated optical switch comprising a pair of waveguides on a substrate, a p-n junction across the pair of waveguides, being defined by p and n-doped regions in the substrate, one on each side of the pair of waveguides.
13. An integrated optical switch substantially as any one described herein with reference to and/or as illustrated in the accompanying drawings.