Wavelength Selective Switching System .

This invention relates to the field of optical communication systems. In particular this invention relates to a dynamically reconfigurable wavelength selective switching system suitable for optical add-drop multiplexing within an optical communication system.

Intelligent next-generation dense wavelength division multiplexing (DWDM) optical communication networks are evolving towards mesh topologies . with flexible nodes that are remotely and automatically controlled for the dynamic allocation of bandwidth. The key component central to the realisation of these future networks is the wavelength selective switch (WSS) . A WSS may be used to add or drop data or spectral channels (each characterised by a distinct centre wavelength and ^' associated bandwidth) at a node within an optical DWDM network in a fully reconf igurable manner. When used for channel drop for example, a IxN WSS typically employs one input port, and M output ports. In a metro ring, one output port is typically used as an express output for traffic travelling through the node, while N-I output ports are used to drop traffic at the node.

Crucially, when used for channel drop, the WSS also enables higher degree switching of channels to fibres travelling e.g. north-south as well as fibre rings disposed in the east-west direction for the realisation of dynamic mesh networks.

When used for channel add, an NxI WSS has N input ports and one output port. In a metro ring, one input port is typically used as an express input for traffic travelling through the node, while N-I input ports are used to add traffic at the node. Crucially, when used for channel add, the WSS also enables higher degree switching of channels from fibres travelling e.g. north-south as well as fibre rings disposed in the east-west direction for the realisation of dynamic mesh networks.

A WSS is strictly defined as a device capable of switching any channel from the designated input port to any output port or switching any channel from any input port to the designated output port. It is this flexibility that forms the key competitive advantage of all WSS solutions. Thermo-optic planar lightwave circuit (PLC) technology is not well suited to the manufacture of WSSs and the majority of commercial prior art WSS solutions to date rely on free-space optics and liquid crystal (LC) technology, or on free-space optics and diftractive micro-electromechanical systems (MEMS) or analogue MEMS tilting micro-mirrors.

Central to all of these solutions is the use of power monitoring on the different WSS input/output ports to enable closed-loop control of the LC or MEMS beam- steering element and so maintain the coupling efficiencies between the different ports in the presence

of component misalignment post-assembly, or subsequently during operation in the presence of environmental change. Power monitoring also enables independent control of the optical power levels in the different channels via the beam-steering element e.g. channel equalisation.

LC technology has been used for the manufacture of variable optical attenuators (VOAs) , Dynamic Channel Equalizers (DCEs) , and 2x2 switches that all operate by controlling the amplitude of the optical field through polarisation change on application of an electric field. Most of the prior art WSS solutions that employ LC technology rely upon a diffractive optics beam-steering approach based on pure phase modulation using a reflective Liquid Crystal on Silicon (LCoS) spatial light modulator (SLM) . In a IxM WSS of this type, multiple channels emerge from a single fibre input port. Bulk optics are employed to condition the beam and the different channels are dispersed via a bulk diffraction grating. The individual beams (each associated with a different channel) are then expanded in one dimension and are directed onto the SLM such that each beam "sees" one or more pixels of the SLM. _: The SLM pixels seen by a given channel are electronically addressed to adjust the optical phase imparted and effect beam-steering on a per- channel basis such that different channels can be directed towards different output ports .

Due to imperfect dielectric mirrors within the LCoS SLM, index mismatches between the mirror and the liquid crystal and diffraction from the underlying metal electrodes, the intrinsic ^' insertion loss (IL) measured in the zero deflection state is typically between 0.4 and 1.OdB, depending on the device and assuming the incident

light is linearly polarised with the optimum orientation. Deflections of only a few degrees can result in an additional IL penalty of up to 3dB. Furthermore, depending on the type of LC employed, the intrinsic polarisation dependent loss (PDL) may also be of such significance that non-optimum input polarisation will result in a further IL penalty.

The PDL can be suppressed by the insertion of a quarter- wave plate in the optical path, but this will result in an additional IL contribution. Alternatively, complex expensive polarisation diversity optics can be introduced into the optical path to suppress PDL effects.

Perhaps the most potentially catastrophic consequence of poor SLM IL performance however is the poor optical cross-talk performance resulting from poor far-field side-lobe suppression. If a side-lobe were to be coupled into one of the output ports, this could result in an unacceptable level of optical cross-talk (OXT) . Alternatively, if the different mechanisms that contribute to the IL as described above result in homogeneous scattering, then some of the scatter could potentially be collected at the output ports resulting in unacceptable cross-talk performance. A further complication when using diffractive optic SLMs is that beam-steering is only intrinsically continuous over a small angular range. Techniques have however been developed that make the beam-steering appear continuous over a wider angular range.

Finally, the anisotropy (and therefore the birefringence) of liquid crystals is intrinsically temperature dependent. Thus, the phase imparted to light incident on

a specific pixel, and the resulting angle of deflection achieved using a LC SLM is also found to display a significant temperature dependence.

Beam-steering using dif fractive MEMS is often found to . suffer from many of the same issues associated with beam- steering using LC, i.e. high IL, high PDL, OXT, and discontinuous beam-steering.

In contrast to beam-steering using either LCs or diffractive MEMS, MEMS mirrors may be used for beam- steering with low-loss (limited only by intrinsic reflectance of gold and scatter from residual surface roughness ~3nm) , low PDL, low cross- talk, continuous angle, non-dispersive, temperature-independenc performance. For these reasons it is perhaps- not surprising, that the most widely-reported market-leading WSS solutions are based on MEMS micro-mirrors. By way of example, US Patent No. 6,687,431 describes a WSS solution based on MEMS micro-mirrors. In particular, this document teaches of a WSS that employs a diffraction grating to resolve a multi-channel optical signal into its constituent spectral channels. These resolved signals are then passed through a quarter wave plate, which is employed to mitigate polarisation sensitive effects, before being focused by a focusing lens onto an array of corresponding channel micro-mirrors. The channel micro-mirrors are individually controllable so as to allow reflection of the spectral channels into selected output ports.

A significant drawback to all of the known MEMS mirror- based WSS solutions resides in the fact that they are not readily ^' scalable to higher port counts, they do not

facilitate the conversion of a single WSS to smaller independent WSSs, and they often require the employment of large expensive 2D MEMS mirror arrays. Furthermore, they all require the use of additional fibre taps and optical power or optical channel monitoring modules for closed-loop control. Some of the incumbent solutions are also limited in their ability to perform hitless switching i.e. switching of a channel between a first port and a second port that does not require the switched channel to traverse a third port of the device.

It is therefore an object of an aspect of the present invention to provide a Wavelength Selective Switching System that overcomes one or more of the above outlined deficiencies of the prior art.

Summary of Invention According to a first aspect of the present invention there is provided a wavelength selective switching system comprising a first dispersive element and a first focussing element that defines an optical axis of the switching system, the switching system further comprising:

a first port array located along the optical axis such that the first dispersive element is located between the first port array and the focusing element; a first collimating element array located along the optical axis such that the first dispersive element is located between the first collimating element and the focusing element; and an optical element array located along the optical axis such that the first focusing element is located

between the f irst dispersive element and the optical element array; the f irst port array comprising at leas t first and second port sub-arrays conf igured so as to define at least a f irst wavelength selective switch and an independent wavelength dependent optical device ; wherein a relative off set in a plane substantially orthogonal to the optical axis of the switching system exists between the first port sub-array and the collimating element array .

The introduction of a relative offset between the first port sub-array and the collimating element array has the effect of converting the wavelength selective switching syscem into independent wavelength dependent optical devices e.g. opcical switches or an optical switch and an optical tap, the number of devices corresponding to the number of separate port sub-arrays present. Although the devices share some common optical elements, they all operate quite independently of each other i.e. a particular channel propagating within a switch can only exit via a single output port of the switch. This provides for an integrated system that is suitable for employment as a multiple IxN, a multiple NxI switch or as a combination of IXN and NxI switches.

Preferably the first collimating element array is located between the first port array and the first dispersive element. Alternatively, the first collimating element array is coplanar with the first port array.

Optionally the first and second port sub-arrays are configured collinearly such that common port elements of the first and second port sub-arrays are aligned with the

1 same collimating element of the first collimating element

2 array. 3

4 Preferably the first collimating element array comprises

5 at least first and second collimating element sub-arrays

6 corresponding to the at least first and second port sub-

7 arrays . 8

9 Preferably the wavelength dependent optical device

10 comprises a second wavelength selective switch.

11 Alternatively the wavelength dependent optical device

12 comprises an optical tap. 13

14 Optionally the relative offset in a plane substantially

15 orthogonal to the optical axis of the switching system

16 comprises a displacement of the first port sub-array

17 relative to the optical axis. Preferably the relative

18 offset in a plane substantially orthogonal to the optical

19 axis of the switching system further comprises a

20 displacement of the second port sub-array relative to the

21 optical axis. ">2

23 Most preferably the displacement of the first port sub-

24 array and the second port sub-array relative to the

25 optical axis have equal but opposite magnitudes. This

26 arrangement produces a system that is symmetric about the

27 optical axis. As a result the associated insertion

28 losses are balanced between the two WSSs. 29

30 Alternatively, the relative offset in a plane

31 substantially orthogonal to the optical axis of the

32 switching system comprises a displacement of the first

33 collimating element sub-array relative to the optical

34 axis. Preferably the relative offset in a plane

substantially orthogonal to the optical axis of the switching system further comprises a displacement of the second collimating element sub-array relative to the optical axis.

Most preferably the displacement of the first collimating element sub-array and the second collimating element sub- array relative to the optical axis have equal but opposite magnitudes.

Preferably the optical element array comprises a first one dimensional optical element array optically aligned with the first port sub-array.

Preferably the optical element array comprises a second one dimensional optical element array optically aligned with the second port sub-array.

Most preferably one or more elements of the one dimensional optical element arrays comprise reflective beam-steering elements. In such an embodiment the wavelength selective switching system is of a folded configuration such that the input ports and the output ports are co-planar.

Preferably one or more of the reflective beam-steering elements comprise two axis reflective beam steering elements .

Alternatively one or more of the elements of the one dimensional optical element arrays comprise transmissive beam-steering elements. In such an embodiment the wavelength selective switching system is of an unfolded configuration .

Preferably the transmissive beam-steering elements comprise two axis transmissive beam-steering elements.

Preferably one or more elements of the second one dimensional optical element array comprise a photodetector . In such an embodiment the photodetectors can be employed within an optical tap path of the wavelength selective switching system.

Optionally the wavelength selective switching system further comprises a second focusing element located along the optical axis such that the optical element array is located between the first focusing element and the second focusing element .

Optionally the wavelength selective switching system further comprises a second dispersive element located along the optical axis such that the second focusing element is located between the optical element array and the second dispersive element.

Optionally the wavelength selective switching system further comprises a second collimating element array located along the optical axis such that the second dispersive element is located between the second focusing element and the second collimating element array.

Optionally the wavelength selective switching system further comprises a second port array located along the optical axis such that the second collimating element array is located between the second dispersive element and the second port array.

Most preferably the first dispersive element comprises a transmission grating. In this embodiment the wavelength selective switching system further comprises a reflective element located such that the transmission grating is employed in a double pass arrangement.

Preferably the wavelength selective switching system further comprises a quarter-wave plate located between the transmission grating and the reflective element. In this embodiment the quarter-wave plate acts to substantially compensate for any polarisation dependent loss associated with the grating such that the PDL between the input port and the optical element array, as well as between any input port and any output port of the switch, is substantially suppressed.

Alternatively, the wavelength selective switching system further comprises a quarter-wave plate located between the transmission grating and the first focusing element.

Alternatively the first dispersive element comprises a reflective grating. Preferably the wavelength selective switching system further comprises a quarter-wave plate located between the reflective -grating and the first focusing element .

According to a second aspect of the present invention there is provided a port array suitable for use in the wavelength selective switching system of the first aspect of the present invention, the port array comprising an irregular arrangement of ports.

Preferably, the port array comprises a first regular port sub-array and at least one further port offset with

respect to the first regular port sub-array. Most preferably the first regular port sub-array comprises all ports other than the at least one further offset port.

Preferably, the at least one further offset port is not collinear with both a row and a column of the first regular port sub-array.

Optionally, the port array comprises a first regular port sub-array and a second port sub-array offset with respect to the first regular port sub-array. The second port sub-array may be regular.

Preferably the first regular port sub-array comprises a first planar lightwave circuit wherein each port comprises a first waveguide core layer having a first waveguide optical axis and a first re-direction mirror that tends a first angle relative to the first waveguide optical axis and which is suitable for reflecting an optical signal into or out of the first waveguide core layer .

Preferably the second port sub-array comprises a second planar l ightwave c ircui t wherein each port comprises a second waveguide core layer having a second waveguide optical axis and a second re-direc tion mirror that tends a second angle relative to the second waveguide optical axis and which is sui table for ref lec ting an optical signal into or out o f the second waveguide core layer .

Most preferably the f irs t and second planar l ightwave circui ts are integrated on a s ingle subs trate .

Most preferably the first angles tended by the first re- direction mirrors have substantially equal magnitudes and orientations. Most preferably the second angles tended by the second re-direction mirrors have substantially equal magnitudes and orientations.

Optionally the second angles and the first angles also have substantially equal magnitudes and orientations.

Optionally the first and second re-direction mirrors are profiled such that optical signals reflected out of the first and second waveguide core layers are substantially collimated.

Optionally at lease one second waveguide core layer is arranged to optically tap at least one first waveguide core layer of a port of the first regular port sub-array.

Optionally the at least one second waveguide core layer is arranged as an optical input tap for the at least one first waveguide core layer of the port of the first regular port sub-array.

Alternatively the at least one second waveguide core layer is arranged as an optical output tap for the at least one first waveguide core layers of the port of the first regular port sub-array.

Optionally the second angle of the second re-direction mirror of the optical output tap comprises an equal magnitude and an opposite orientation to the first angle of the first re-direction mirror.

Optionally the optical output tap further comprises a tap waveguide core layer and a tap mirror.

Preferably the tap mirror is located between the tap waveguide core layer and the at least one second waveguide core layer.

According to a third aspect of the present invention there is provided a collimating element array suitable for use in the wavelength selective switching system of the first aspect of the present invention, the collimating element array comprising an irregular arrangement of collimating elements.

Preferably, the collimating eiemenc array comprises a first regular collimating element sub-array and at least one further collimating element offset with respect to the first regular collimating element sub-array. Most preferably the first regular collimating element sub- array comprises all collimating elements other than the at least one further offset collimating element.

Preferably, the at least one further offset collimating element is not collinear with both a row and a column of the first regular collimating element sub array.

Optionally, the collimating element array comprises a first regular collimating element sub-array and a second collimating element sub-array offset with respect to the first regular collimating element sub-array. The second collimating element sub-array may be regular.

According to a fourth aspect of the present invention there is provided a method of converting a wavelength

selective switching system, having an optical axis, into at least a first wavelength selective switch and an independent wavelength dependent optical device the method comprising the steps of: 1) dividing a first port array into at least a first port sub-array and a second port sub-array; 2) introducing a relative offset in a plane substantially orthogonal to the optical axis of the switching system between the first port sub-array and the collimating element array.

Preferably the method of converting a wavelength selective switching system further comprises the step of dividing a first collimating element array into at least a first collimating element sub-array and a second collimating element sub-array corresponding to the at least first port sub-array and the .second port sub-array;

Optionally the introduction of a relative offset in a plane substantially orthogonal to the optical axis of the switching system comprises the introduction of a displacement of the first port sub-array relative to the optical axis.

Preferably the relative offset in a plane substantially orthogonal to the optical axis of the switching system further comprises a displacement of the second port sub- array relative to the optical axis.

Most preferably the displacement of the first port sub- array and the second port sub-array relative to the optical axis have equal but opposite magnitudes.

Alternatively, the introduction of a relative offset in a plane substantially orthogonal to the optical axis of the switching system comprises the introduction of a displacement of the first collimating element sub-array relative to the optical axis. Preferably the relative offset in a plane substantially orthogonal to the optical axis of the switching system further comprises a displacement of the second collimating element sub-array relative to the optical axis .

Most preferably the displacement of the first collimating element sub-array and the second collimating element sub- array relative to the optical axis have equal but opposite magnitudes.

According to a fifth aspect of the present invention there is provided a method of converting a wavelength selective switching system into at least a first wavelength selective switch and an independent wavelength dependent optical device the method comprising the steps of: 1) dividing a port array into a first port sub-array and at least a further port; 2) offsetting the at least further port with respect to the first port sub-array.

According to a sixth aspect of- the present invention there is provided a method of converting a wavelength selective switching system into at least a first wavelength selective switch and an independent wavelength dependent optical device the method comprising the steps of:

1) dividing a collimating array into a first collimating element sub-array and at least a further collimating element; 2) offsetting the at least further collimating element with respect to the first collimating element sub-array.

Brief Description of Drawings Aspects and advantages of the present invention will become apparent upon reading the following detailed description and upon reference to the following drawings in which:

Figure 1 presents a schematic representation of a generic 9-port w ^" 3S N. B. for clarity purρυ≤e$ the principal rays corresponding to return beams are not shown.

Figure 2 presents a schematic representation of the effects of a vertical offset in the 2D collimating element array on a generic 9-port WSS in accordance with an aspect of the present invention N. B. principal rays corresponding to return beams are again not shown to assist clarity.

Figure 3 presents a schematic representation of an integrated Ix2-port WSS and a Ix5-port WSS in accordance with an aspect of the present invention

Figure 4 presents an orthographic schematic representation of the integrated 1x2 -port WSS and Ix5-port WSS of Figure 3 in:

( a ) plan view; (b) view from behind port array looking along optical axis towards array of collimating elements; and (c) view looking along optical axis towards array of beam-steering elements; all with the ports and corresponding collimating elements arranged in two different blocks. All the ports are located on a uniform grid but each block of collimating elements has a different vertical offset.

Figure 5 presents an orthographic schematic representation of an alternative embodiment of che incegrated Ix2-port WSS and 1x5-pore W53 of Figure 3 in: (a) plan view; (b) view from behind port array looking along optical axis towards array of collimating elements; and (c) view looking along optical axis towards array of beam-steering elements; all with the ports and corresponding collimating elements arranged in two different blocks . All collimating elements are located on a uniform grid but each block of ports has a different vertical offset.

Figure 6 presents a preferred embodiment of 3x3 port array for the independent 1x2-port WSS and 1x5- port WSS of Figure 5 based on a PLC with out- of-plane re-direction mirrors in: (a) plan view; and (b) section on XX.

The first column of ports is displaced downward as per Figure 5.

Figure 7 presents an orthographic schematic representation of the integrated Ix2-port WSS and Ix5-port WSS of Figure 5 that incorporates a double pass transmission grating as the dispersive element in: (a) plan view with quarter-wave plate located between grating and folding mirror; (b) alternative plan view with quarter-wave plate located between grating and focussing element; (c) view from behind port array looking along opcical axis cowards array of coilitnaL iriy elements; and (d) view looking along optical axis towards array of beam- steering elements.

Figure 8 presents an embodiment of a 3x3 port array based on a PLC with out-of -plane re-direction mirrors for 1x5 WSS with an integrated input tap port and two integrated output tap ports in: (a) plan view; (b) section on XX; (c) section on YY; and (d) section on ZZ.

Figure 9 presents an alternative embodiment of a 3x3 port array based on a PLC with out-of -plane re- direction mirrors for 1x5 WSS with an integrated input tap port and two integrated output tap ports in:

(a) plan view; (b) section on XX; (c) section on YY; and (d) section on ZZ.

Figure 10 presents a further alternative embodiment of 3x3 port array based on a PLC with out-of-plane re-direction mirrors for 1x5 WSS with an integrated input tap port and two integrated output tap ports in: (a) plan view; (b) section on XX; (c) section on YY; and (d) section on ZZ;

Figure 11 presents an orthographic schematic representation of the Ix5-port WSS with an integrated input tap port and two integrated output tap ports of Figures 8 to 10 in: (a) plan view; (b) view from behind port array looking along optical axis towards array of collimating elements; (c) view looking along optical axis towards ^• array of beam-steering elements. In particular a double pass transmission grating is employed as the dispersive element. All collimating elements are located on a uniform grid. The input tap port is offset upward while the output tap ports are offset downward requiring two ID arrays of beam- steering elements with downward and upward offsets respectively.

Figure 12 presents an orthographic schematic of independent 1x2-port and Ix5-port WSSs, with a reflection grating employed as the dispersive element, and with their ports and corresponding collimating elements arranged in two different blocks and in: (a) plan view; (b) view from behind port array looking along optical axis towards array of collimating elements; and (c) view looking along optical axis towards array of beam-steering elements. All collimating elements are located on a uniform grid. Each block of ports has a different vercical offset requiring two ID arrays of beam-steering elements each with a different vertical offset.

Figure 13 presents an orthographic schematic of Ix5-port WSS with an integrated input tap port and two integrated output tap ports, with a reflection grating employed as the dispersive element, in: (a) plan view; (b) view f rom behind port array looking along optical axis towards array of collimating elements ; and ( c ) view looking along optical axis towards array of beam-s teering elements . All collimating elements are located on a uni form grid . The input tap port is of f set upward while the output tap ports are of f set downward requiring two ID arrays of beam- steering elements with downward and upward of f se ts respec tively .

Figure 14 presents an alternative embodiment of 3x3 port array for independent 1x2-port and Ix5-port WSSs analogous to Figure 6 and based on a PLC with profiled out-of-plane re-direction mirrors with a first column of ports displaced downward as per Figure 5 in: (a) plan view; and (b) section on XX.

Figure 15 presents an orthographic schematic representation of an alternative embodiment of the integrated Ix2-port WSS and Ix5-port WSS of Figure 5 in: (a) plan view; (b) view from behind port array looking along optical axis towards array of collimating elements; and (c) view looking along optical axis towards array of beam-steering elements; all with the ports and corresponding collimating elements arranged in two different blocks. All collimating elements are located on a uniform grid but each block of ports has a different vertical offset.

Figure 16 presents a schematic ^• representation of two integrated Ix2-port WSSs with collinear ports and collinear collimating elements in accordance with an aspect of the present invention.

Figure 17 presents an orthographic schematic representation of the two integrated Ix2-port WSSs of Figure 16 in: (a) plan view; (b) view from behind port array looking along optical axis towards array of collimating elements ; and (c) view looking along optical axis towards array of beam- steering elements; All collimating elements are collinear and all ports are collinear but each block of ports has a different vertical offset.

Figure 18 presents a schematic representation of a 1x6 port array suitable for use within the two integrated independent Ix2-port WSSs of Figures 16 and 17. The port array ^' is based on a PLC with the port positions defined by the waveguide end faces : (a) plan view; and (b) end elevation.

For consistency and clarity purposes the various features of the described WSS are referred to by the same reference numerals- throughout the specification. Where appropriate, those reference ^{' '} numerals employed to describe the features common to alternative embodiments of the WSS are also maintained within the specific description.

Detailed Description In order to assist understanding of the present invention. Figure 1 and Figure 2 present schematics of a generic 9 -port WSS. In particular, Figure 1 presents a

1 schematic of a generic Ix8-port ( i . e . 1 input port and 8

2 output ports ) WSS 1 capable of switching any one of four

3 spectral channels ( each characterised by a distinct

4 centre wavelength and associated bandwidth) , to any of

5 its output ports . The WSS 1 can be seen to comprise a

6 3x3 port array 100 , a 3x3 collimating elements array 101 ,

7 a dispersive element 102 , a focusing element 103 and a

8 1x4 , 2-axis reflective beam-steering elements array 104 ,

9 all symmetrically disposed about a common optical axis

10 105 . 1 1

12 For the purposes of clarity, a port is defined simply as

13 a point of ingress or egress of a light beam to or from

14 the optical system. Light is generally confined within

15 an optical waveguide, such as an optical fibre or an If) integrated channel waveguide in a PLC before it enters or

17 exits the free-space WSS optical system via a port. A

18 collimating element is defined as any combination of

19 lenses and/or mirrors that collimates a beam diffracting 0 from corresponding ports. The dispersive element 102 is 1 any component capable of angularly dispersing light 2 including, but not limited to, a grating or a prism. It

23 should be noted that the dispersive element 102 can work

24 in transmission or reflection, as detailed further below.

25 The focussing element 103 is defined as any combination

26 of lenses and/or mirrors that focuses the angularly

11 dispersed collimated beams emerging from the dispersive

28 element 102 thus forming a ID array of images (one image

29 per wavelength or channel) such that each different image

30 substantially coincides with a different element of the

31 ID array of 2-axis reflective beam-steering elements 104. 32

33 For minimum insertion loss, the 3x3 collimating element

34 array 101 is disposed along optical axis 105 such that

1 the 3x3 port array 100 is substantially located in the

2 back focal plane and the dispersive element 102 is

3 substantially located in the front focal plane.

4 Similarly _/ for minimum insertion loss the focussing

5 element 103 is disposed along optical axis 105 such that

6 the dispersive element 102 is substantially located in

7 the back focal plane and the 1x4, 2-axis reflective beam-

8 steering element array 104 is substantially located in

9 the front focal plane. 10

11 In Figure 1, light is arbitrarily shown entering port

12 100-12 (as represented by the corresponding principal ray

13 shown as a solid line) and the associated beam propagates

14 and diffracts through free-space to the corresponding

15 collimacing element 101-12. The substantially collirnated lfi beam continues to propagate through the dispersive

17 element 102. At this point the four different spectral

18 channels are angularly dispersed to form four

19 substantially collimated beams propagating at different 0 angles, each of which is imaged by focussing element 103 1 onto a different 2-axis beam-steering element 104-1 2 through 104-4. Each 2-axis beam-steering element may be 3 independently activated to reflect any of the four beams

24 (i.e. spectral channel) to any output port. Note that,

25 for clarity purposes, the principal rays corresponding to

26 the beams in the return path are not shown. Importantly,

27 from Figure 1 it can be seen, that multiple spectral

28 channels can be directed to any port, but that a given

29 channel can itself only appear at any one output port at

30 any given time. 31

32 It will be appreciated by those skilled in the art that,

33 in general, the optical axis 105 may not be a straight

34 line as represented in Figure 1 for clarity, but rather

that it may be re-directed by any of the components 101, 102, 103 or 104.

It will also be appreciated by those skilled in the art, that the 2-axis reflective beam-steering elements 104 need not be reflective at all but rather transmissive such that the optical axis 105 is not folded at the plane of the beam-steering elements. Following transmission through the beam-steering elements 104 such an embodiment requires that the light must then traverse in order, a second focussing element substantially similar to 103, a second dispersive element substantially similar to 102, and a second collimating elements array substantially similar to 101 before exiting the syscem via an output port array.

It will further be appreciated by those skilled in the art, that the 2-axis reflective beam-steering elements 104 may be realised by employing any technology including the exemplary case of analogue MEMS tilting mirrors, diffractive MEMS or LC SLMs. Furthermore, the WSS 1 could employ 1- or 2-axis beam-steering elements, however the use of 2-axis beam-steering elements enables hitless switching since output beams can then, in general, be switched between two ports without having to trace out a path in space that would result in the detrimental unwanted coupling with a third port. ^■ The beam steering elements need not necessarily be on a constant pitch and can be arranged to optimise the spectral response of the WSS device.

Figure 2 presents a schematic representation of a WSS 2 that illustrates the effect of introducing a vertical offset in the 2D collimating elements array 201 with

1 respect to the 2D ports array 200. As can be seen, when the collimating elements 201 are offset in the upward

3 direction (shown dashed) , the four different spectral channels are angularly dispersed by dispersive element

5 202 and imaged by focussing element 203 onto a ID array

6 of 2-axis beam-steering elements 204-11 through 204-14

7 (shown dashed) that is offset in the upward direction

8 with respect to the optical axis 205. Each 2-axis beam-

9 steering element 204-11 through 204-14 may be 0 independently activated to reflect any of the four beams ! (i.e. any spectral channel) to any output port. 2 3 Alternatively, when the collimating elements 201 are 4 offset in the downward direction (shown docted) , the four 5 different spectral channels die angularly dispersed by ft dispersive element 202 and imaged by focussing element 7 203 onto a ID array of 2-axis beam-steering elements 204- 8 31 through 204-34 (shown dotted) that is offset in the 9 downward direction with respect to the optical axis 205. 0 Each 2-axis beam-steering element 204-31 through 204-34 1 may again be independently activated to reflect any of 2 the four beams (i.e. any spectral channel) to any output 3 port. 4 5 Note that in Figure 2 the principal rays corresponding to 6 the beams in the return paths are not shown for clarity 7 purposes. Multiple spectral channels can again be 8 directed to any port, but a given channel can only appear 9 at any one output port at any particular time. 0 Furthermore, it will be appreciated that rather than 1 shifting the 2D collimating elements array 201 2 upwards /downward, the 2D ports array 200 could have been 3 offset by an equal amount in a downward/ upward direction, 4 respectfully, with an identical outcome.

Figure 3 presents a WSS 3 in accordance with an aspect of the present invention that harnesses the power and versatility of vertically offsetting different collimating elements of the 3x3 collimating elements array 301, as described above in connection with Figure 2. In this embodiment the offsetting of the different collimating elements of the 3x3 array 301 results in the integration of independent 1x2 -port and Ix5-port WSSs.

Both WSSs employ the same dispersive element 302 and the same focussing element 303 however, their ports are arranged within two different blocks, which correspond to the two blocks on the collimating element array 301 that exhibit a relative vertical offset i.e. the block containing collimating elements 300-11, 300-21 and 300-31 exhibits an upward vertical displacement with respect to the optical axis 305. Corresponding to each WSS is a ID array of 2-axis reflective beam-steering elements each with a different vertical offset. These ID arrays of 2- axis reflective beam-steering elements are most advantageously monolithically- integrated onto the same substrate to form a 2D array of 2-axis reflective beam- steering elements with one row of elements per WSS.

The Ix2-port WSS has input port. 300-11 and output ports 300-21 and 300-31. The corresponding block of collimating elements 301-11, 301-21 and 301-31 is shown dashed and is offset upward with respect to the optical axis 305. The corresponding path of the principal ray is also shown dashed and the individual beams corresponding to the four different spectral channels are focussed onto the uppermost ID array of 2-axis beam-steering elements 304-11, 304-12, 304-13 and 304-14. Each 2-axis beam-

steering element 304-11 through 304-14 may be independently activated to reflect any of the four beams (i.e. any spectral channel) to either of the two output ports 300-21 or 300-31.

It will be appreciated that in an alternative embodiment that rather than shifting the block of 2D collimating elements 301-11, 301-21 and 301-31 upward, the ports 300- 11, 300-21 and 300-31 can be offset by an equal amount in a downward direction, 305 with an identical outcome.

The Ix5-port WSS has input port 300-12 and output ports 300-13, 300-22, 300-23, 300-32 and 300-33. The corresponding block of collimating elements 301-13, 301- 22, 301-23, 301-32 and 301-33 are shown solid and are not offset with respect to the optical axis 305. The corresponding path of the principal ray is also shown solid and the individual beams corresponding to the four different spectral channels are focussed onto the middle ID array of 2-axis beam-steering elements 304-21 through 304-24. Each 2-axis beam-steering element 304-21 through 304-24 may be independently activated to reflect any of the four beams (i.e. any spectral channel) to any output port 300-13, 300-22, 300-23, 300-32 or 300-33.

Note that,- for clarity, the principal rays corresponding to the beams in the return path for both the 1x2 WSS and the 1x5 WSS have been omitted. As described before it should be noted that multiple spectral channels within either WSS can be directed to any of the corresponding output ports, but that a given spectral channel can only appear at a single output port at any particular time.

Figure 4 presents an orthographic representation of the same integrated independent 1x2 -port and Ix5-port WSSs shown in Figure 3 with corresponding component numbering. All the ports are located on a uniform grid but the two blocks of collimating elements corresponding to the two WSSs have different vertical offsets, i.e. the block of collimating elements corresponding to the 1X2 WSS, namely elements 401-11, 401-21 and 401-31, has a vertical upwards offset introduced with respect to the block of ports corresponding to the 1X5 WSS. The principal rays are shown in the plan view Figure 4 (a) for light entering the WSS system via the input ports 400-11 and 400-12. Note that, for clarity, the principal rays in the return paths are not shown .

Figure 5 is identical to Figure 4 except for the face that now all the collimating elements' are located on a uniform grid but the blocks of ports corresponding to the two WSSs are vertically offset. In particular, the block of ports corresponding to 1X2 WSS, namely 500-11, 500-21 and 500-31 have a vertical, downward offset introduced with respect to the block of ports corresponding to the 1X5 WSS. The same component numbering scheme is used once again. The principal rays are shown in the plan view Figure 5 (a) for light entering the WSS system via the input ports 500-11 and 500-12.

Figure 15 presents a yet further alternative embodiment of the integrated 1x2 port and 1x5 port WSSs device of Figure 3. A similar component- numbering scheme is used once again.

All of the collimating elements 1501-11. through 1501-33 are again located on a uniform grid and it is the block

of ports corresponding to the two WSSs that are again vertically offset. However, in this embodiment the block of ports corresponding to the 1x2 WSS, namely 1500-11, 1500-21 and 1500-31 have a vertical, downward offset introduced while the block of ports corresponding to the 1x5 WSS, namely 1500-12, 1500-13, 1500-22, 1500-23, 1500- 32 and 1500-33 have a vertical, upwards offset introduced. The principal rays are shown in the plan view Figure 15 (a) for light entering the WSS system via the input ports 1500-11 and 1500-12.

As can be seen from Figure 15 (c) different spectral channels of the light entering the 1x2 WSS system via the input ports 1500-11 are dispersed by dispersive element 1502 and thereafter focussed by focussing element 1503 onto the beam steering elements array 1504. In particular, the spectral channels are incident on the uppermost ID array of 2-axis beam-steering elements 1504- 11, 1504-12, 1504-13 and 1504-14.

In a similar manner, the different spectral channels of the light entering the 1x5 WSS system via the input ports 1500-12 are dispersed by dispersive element 1502 and thereafter focussed by focussing element 1503 onto the beam steering elements array 1504 and in particular the lower ID array of 2-axis beam-steering elements 1504-21, 1504-22, 1504-23 and 1504-24.

It is noted that the vertical offset of the uppermost ID array .is of an opposite sense to the offset introduced to the block of ports corresponding to 1x2 WSS. The magnitude of the required vertical offset of the uppermost ID array is dependent upon the offset introduced to the block of ports corresponding to 1x2

WSS. An optical magnification generally exists within the system such that the ID array requires to be offset by a lager magnitude than the offset introduced to the corresponding block of ports. An analogous condition exists for the block of ports corresponding to 1x5 WSS and the lower ID array.

The presently described embodiment, where every port has an offset relative to the optic axis, is particularly advantageous because the system is then vertically symmetric about the optical axis. ^' As a result the associated insertion losses are balanced between the two WSSs.

Ic will be appreciated by those skilled in the art that the concepts presented above are not limited to a 3x3 arrangement of ports and collimating elements but that they can easily be generalised to any number of ports and any ID or 2D arrangement of ports and collimating elements. Also, there is no specific requirement for ports corresponding to the same WSS to be arranged in a contiguous block; the only condition is that there exists the same relative offset between each port and its corresponding collimating element. Furthermore, although the combination of two independent WSSs was described for the purposes of illustration, it will be understood that, in general, two or more independent WSSs can be combined as described above. In addition, a mixture of independent IxN WSSs (i.e. 1 input port and N output ports) and NxI WSSs (i.e. N input ports and 1 output port) can be combined using the concepts outlined. Clearly, the concepts outlined above are also independent of the number of spectral channels W employed. Four spectral channels have been employed for illustrative

purposes; however the number of channels W is limited only by the number of elements that can be employed to make up each ID array of 2-axis beam-steering elements 304-χW. A further significant advantage of the present invention is that independent IxN and NxI WSSs can optionally be combined in the aforementioned manner thus achieving a fully-flexible ROADM solution with both add- drop functionality and East-West separation as is most desirable for use within metro-area optical networks.

Figure 6 shows a preferred embodiment for the 3x3 port array 600 compatible with the independent Ix2-port and Ix5-port WSSs of Figure 5. This port array 600 is based on PLC technology using downward cotal internal refleccion (TIR) re-directiυn mirrors monolithically- integrated with buried channel waveguides. A full description of these redirection elements can be found within the author's earlier UK Patent Application No. GB 052155.0 entitled "Improved Optical Excursion Device".

In summary, the waveguide layers comprise a substrate 610, an optional lower clad layer 620, a waveguide core layer 630 having an optical axis 650, and an upper clad layer 640. The channel waveguides terminate at substantially vertical facets that -constitute ports 600- 11 through 600-33. These ports are displaced with respect to corresponding out-of-plane re-direction mirror facets 660-11 through 660-33. For improved manufacturability, the out-of-plane re-direction mirrors 660 are fabricated simultaneously after the deposition of the waveguide core and clad layers. The re-direction mirror surfaces 660-11 through 660-33 fully intersect the upper clad layer 640 and at least partially intersect the lower clad layer 620 for efficient re-direction. The

displacement of the waveguide facets 600-11 through 600- 33 with respect to the mirror facets 660-11 through 660- 33 eliminates any wavefront distortion associated with propagation of the beam up or down through the waveguide rib, while the presence of the waveguide facets 600-11 through 600-33 in the optical path has only a marginal effect on the device return loss for waveguide index contrasts up to 1.5%. To suppress back reflections at higher index contrasts, the normal of the waveguide facets can be angled at 8° with respect to the waveguide optical axis when viewed in plan; the re-direction mirrors 660-11 through 660-33 should also then be rotated in sympathy when viewed in plan to achieve downward reflection in a direction that is substantially normal co the PLC plane. Furthermore, assuming the upper clad 640 and lower clad layers 620 are of a similar composition, the re-direction mirror facets 660-11 through 660-33 are etched through homogeneous material ^' with the result of a smoother mirror surface and a more uniform mirror angle. For a substantially smooth mirror, TIR at the mirror facets 660-11 through 660-33 results in zero theoretical IL, PDL and WDL penalties on reflection.

In the 3x3 port array 600 light is most advantageously edge-coupled to/ from the left-hand edge of the PLC, ^' as shown in Figure 6, via a fibre V-groove array (not shown) . Light is launched into the waveguide terminating at facet 600-11 that serves as the input port to the" 1x2 WSS and into the waveguide terminating at facet 600-12 that serves as the input port to the 1x5 WSS. Light exits the PLC via the other waveguides terminating at the other facets that serve as the output ports 600-21 and 600-31 for the 1x2 WSS and 600-13, 600-22, 600-23, 600-32 and 600-33 for the 1x5 WSS.

Depending on the exact processes used to fabricate the re-direction mirrors, different mirrors on the same PLC chip may however be subject to different imperfections such as absolute angle error, uniformity of mirror angle across the mirror facet (intra-mirror angle uniformity), uniformity of mirror angle between different mirror facets on the same chip (inter-mirror angle uniformity), surface roughness etc. The most significant error for an optimised micromachining process is likely to be in the absolute magnitude of the mirror angle rather than in the intra-mirror or inter-mirror angle uniformity. To ensure that the out-of-plane excursion (defined as that portion of the WSS optical path from the input port through the free-space optical system of the VlSS and back to any of the output ports) is tolerant to the absolute re- direction mirror angle, the out-of-plane re-direction mirrors 660-11 through 660-33 (that will all have substantially similar absolute angles) should all critically be fabricated with the same orientation, or sign, of angle.

For any port array based on PLC technology such as that shorn in Figure 6, an anti-reflection (AR) coating may be optionally applied at any dielectric-air interface to reduce back-reflections and optimise IL. In Figure 6 the AR coating is represented by the dotted line. It is also obvious that for efficient put-of-plane downward re- direction, the substrate should be substantially transparent at the wavelength of interest and, in addition, the beam diameter at the substrate underside will be a function of the substrate thickness.

It will be again be appreciated by those skilled in the art that the concept of port array 600 presented in Figure 6 is not limited to a 3x3 arrangement, but that it generalises to any number of ports and any ID or 2D arrangement of ports. Also, there is no requirement for ports corresponding to the same WSS to be arranged in a contiguous block; the only condition is that there -exists the same relative vertical offset between each port and its corresponding collimating element. Furthermore, two or more independent WSSs can be combined in the manner described. In addition, a mixture of independent IxN WSSs (i.e. 1 input port and N output ports) and NxI WSSs (i.e. N input ports and 1 output port) can be combined using the concepts outlined above.

In the exemplary case illustrated in Figure 6, downward TIR from the re-direction mirrors is assumed without loss of generality; however, it will be understood that the concepts presented above apply to any type of re- direction mirror.

Figure 7 shows the same independent 1x2-port WSS and 1x5- port WSS of Figure 5 with a double pass transmission grating 702 employed as the dispersive element 302. The principal rays are shown in the plan view, Figure 7 (a) , and (b) for light entering the WSS system via the input ports. For clarity, the principal rays in the return paths are again omitted. The system now comprises a port array 700, a collimating element array 701, a transmission grating 702, a quarter-wave plate 706 (or 706'), a plane mirror 705 disposed at an angle of θ with respect to grating 702, a focussing element 703 and an array of beam-steering elements 704. Light is launched into input port 700-11 for the 1x2 WSS and input port

700-12 for the 1x5 WSS. These beams diffract in free- space and are collimated by corresponding collimating elements 701-11 and 701-12.

In the embodiment of Figure 7 (a) after the first pass through transmission grating 702, these beams are angularly dispersed and pass through quarter-wave plate 706. On reflection from plane mirror 705, the beams pass through quarter-wave plate 706 once again- and undergo dispersion augmentation via a second pass through transmission grating 702. As will be appreciated, from this point on, the function of the WSS system is identical to that outlined previously except that now the return paths must also pass twice through the transmission grating 702 and the quarter-wave plate 706.

When located between the grating 702 and the plane mirror 705, the quarter-wave plate acts to substantially compensate for the polarisation dependent diffraction efficiency of the grating such that the PDL between the input port and the array of beam-steering elements (as well as between the input port and any output port) is substantially suppressed. This has profound consequences when implementing low PDL power taps on the input and outputs of the WSS when a photodetector array used to measure power in each spectral channel is located in the plane of the beam-steering array, as discussed in more detail below.

Figure 7 (b) shows the same WSS system as Figure 7 (a) in plan view except that now the quarter-wave plate 706' is shown in an alternative position located between the grating 702 and the focussing element 703. In this position, the quarter-wave plate only acts to

substantially compensate for the polarisation dependent diffraction efficiency of the grating such that only the PDL between the input port and any output port is substantially suppressed.

The use of a double pass transmission grating 702 as the dispersive element confers several advantages. In the first instance the dispersion is enhanced facilitating larger beam separations in the plane of the beam-steering array 704 for a given focal length ^" of the focussing element 703. Alternatively the focal length of the focussing element 703 can be reduced for a given beam separation in the plane of the beam-steering array 704. Secondly, the fill factor associated with the area illuminated of focussing element 703 is not unduly increased on the addition of extra columns of ports resulting in reduced optical aberration and therefore reduced loss for high port count WSS systems. Furthermore, when located between the grating 702 and the plane mirror 705, the quarter-wave plate 706 acts to substantially compensate for the polarisation dependent diffraction efficiency of the grating such that the PDL between the input port and the array of beam-steering elements is substantially suppressed.

Figure 8 shows an alternative embodiment of a 3x3 port array 800 based on a PLC with out-of-plane re-direction mirrors for a 1x5 WSS arbitrarily chosen to have an integrated input tap port 800-11 and two integrated output tap ports 800-21 and 800-31. The 1x5 WSS again has input port 800-12 and output ports 800-13, 800-22, 800-23, 800-32, 800-33 consistent with the 1x5 WSS of Figure 3 through Figure 7. It should be noted that all the WSS re-direction mirrors 860-12, 860-13, 860-22, 860-

23, 860-32, and 860-33 and the input tap port re- direction mirror 860-11 all share the same orientation while the output tap port re-direction mirrors 860-21 and 860-31 have an opposite orientation. Integrated channel waveguide taps are used to tap a fraction of the input or output power as shown and route it to the appropriate re- direction mirror 860-11, 860-21 or 860-31. The input tap port 800-11 is offset upward with respect to the corresponding position on the uniform WSS port grid while the output tap ports 800-21 and 800-31 are both offset by an equal amount downwards with respect to their corresponding positions on the uniform WSS port grid.

Figure 9 shows a second embodiment of a 3x3 port array 900 based on a PLC with ouc-of-plane re-direction mirrors for a 1x5 WSS arbitrarily chosen to have an integrated input tap port 900-11 and two integrated output tap ports 900-21 and 900-31. The 1x5 WSS has input port 900-12 and output ports 900-13, 900-22, 900-23, 900-32, 900-33 consistent with the 1x5 WSS of Figure 3 through Figure 8. In contrast to the preferred PLC port array embodiment of Figure 8, in this embodiment all re-direction mirrors 960 corresponding to WSS ports 900-12, 900-13, 900-22, 900- 23, 900-32, 900-33, the input tap port 900-11 and the output tap ports 900-21 and 900-31 now share the same orientation. This is enabled by the use of bi- directional integrated channel waveguide taps 980-21 and 980-31 that rely upon reflection from substantially vertical HR-coated waveguide facets 970-21 and 970-31, respectively. Power is tapped from the WSS input waveguide in an identical manner to Figure 8, but power from output ports 900-23 and 900-33 is tapped and is first reflected from substantially vertical HR-coated waveguide facets 970-21 and 970-31, respectively, before

being routed to output tap ports 900-21 and 900-31, respectively. Note that the waveguides are normal to the substantially vertical HR-coated waveguide facets 970-21 and 970-31. This will result in light being reflected back along the same waveguide path back to the integrated channel waveguide tap and the corresponding output tap ports 900-21 and 900-31, respectively.

Figure 10 shows a third embodiment of a 3x3 port array IOOO based on a PLC with out-of-plane re-direction mirrors for a 1x5 WSS arbitrarily chosen to have an integrated input tap port 1000-11 and two integrated output tap ports 1000-21 and 1000-31. The 1x5 WSS has input port 1000-12 and output ports 1000-13, 1000-22, 1000-23, .1000-32, 1000-33 consistent with the 1x5 WSS of Figure 3 through Figure 9. As for che embodiment of Figure 9 all the WSS and the input and output tap port re-direction mirrors share the same orientation. In this embodiment this is enabled by the use of bi-directional integrated channel waveguide taps 1080-21 and 1080-31 that rely upon reflection from substantially vertical HR- coated waveguide facets 1070-21 and 1070-31, respectively. Power is tapped from the WSS input waveguide in an identical manner to Figure 8 and Figure 9, but power from output ports 1000-23 and 1000-33 is tapped and is first reflected from substantially vertical HR-coated waveguide facets 1070-21 and 1070-31, respectively, before being routed to output tap ports 1000-21 and 1000-31, respectively. In contrast to the embodiment of Figure 9, the waveguides now impinge on the substantially vertical HR-coated waveguide facets 1070-21 and 1070-31 at a non-zero angle of incidence resulting in light being reflected back along a different waveguide

path to the corresponding output tap ports 1000-21 and 1000-31, respectively.

Within the embodiments of Figures 8 -to 10 it will be appreciated that the integrated channel waveguide taps may be of any type or design including but not limited to integrated directional couplers or Mach-Zehnder interferometers. It will also be appreciated that, in general, the number of WSS tap ports is not restricted to three but that every WSS input port can have an associated input tap port and every WSS output port can have an associated output tap port.

Figure 11 presents an orthographic representation of a Iκ5-ρort WSS with an integrated input tap port and two integrated output tap ports, and with a double pass transmission grating 1102 employed as the dispersive element, that is compatible with any of the PLC port array embodiments of Figures 8 to Figure 10. Note that the numbering convention ^' used for the components is the same as used for Figure 7 with the only difference now being that 1104-31 through 1104-34 represent a ID array of photodetector elements for measurement of the power associated with the different spectral channels appearing at the WSS input port, and 1104-11 through 1104-14 now represent a ID array of photodetector elements for measurement of the power associated with the different spectral channels appearing across the tapped WSS output ports (any one spectral channel can only appear at any one output port at any particular time) . For clarity, the principal rays are shown in Figure 11 (a) from the WSS input port 1100-12 and the input tap port 1100-11 but the principal rays are not shown for any of the return paths. It will be obvious from earlier

comments in relation to the position of the quarter-wave plate 1106 between the transmission grating 1102 and the plane mirror 1105, that the PDL associated with diffraction from the grating in either the input tap path or the output tap paths is substantially suppressed.

It will be further appreciated that with the PLC port array embodiment of Figure 8, the fact that the output tap port re-direction mirror orientation is different from the orientation of the other re-direction mirrors, should not unduly affect the loss in the optical path from the tap ports to the photodetector elements in Figure 11. This is in contrast to the re-direction mirrors for the WSS output ports in the same PLC port array embodiment which crucially must have the same orientation as the corresponding WSS inpuc port.

Figure 12 presents . an alternative embodiment for implementing the independent 1x2-port WSS and 1x5-port WSS of Figure 5 based on a reflection grating 1202 being employed as the dispersion element 502. This WSS embodiment is analogous to the WSS embodiment of Figure 7 that explicitly showed a double pass transmission grating 702 as the dispersive element 502. Note that the embodiment of Figure 12 employs the same numbering convention and is compatible .with any of the PLC port array embodiments of Figure 8 to Figure 10.

Figure 13 illustrates a Ix5-port WSS with an integrated input tap port and two integrated output tap ports and with a reflection grating 1302 shown as the explicit dispersive element. This WSS embodiment is analogous to the WSS embodiment of Figure 11 that employed a double pass transmission grating 1102 as the dispersive element.

Note that the embodiment presented in Figure 13 uses the same numbering convention and is compatible with any of the PLC port array embodiments of Figure 8 to Figure 10. For clarity, the principal rays are shown in Figure 13 (a) from the WSS input port 1300-12 and the input tap port 1300-11 but the principal rays are not shown for any of the return paths .

In Figures 12 and 13, the quarter-wave plate (1206 or 1306) is located in the optical path between the reflection grating (1202 or 1302) and the focussing element (1203 or 1303) such that the PDL between the input port and any output port is still substantially suppressed. However, it should be noted that in both embodiments the PDL between the input ports and the array of beam-steering elements is no longer subscantially suppressed. Accordingly, a low PDL reflection grating is required in order to suppress the tap path PDL for these embodiments.

For the embodiments shown in Figures 11 to 13 -it will be appreciated that as with the PLC port array embodiment of Figure 8, the fact that the output tap port re-direction mirror orientation is different from the orientation of the other re-direction mirrors, should not unduly effect the loss in the optical path from the tap ports to the photodetector elements. This is in contrast to the re- direction mirrors for the WSS output ports in the same PLC port array embodiment which crucially must have the same orientation as the corresponding WSS input port .

As previously stated the number of WSS tap ports is, in general, not restricted to three but that every WSS input port can have an associated input tap port and every WSS

output port can have an associated output tap port. Thus, without introducing any significant additional manufacturing complexity or unduly adding to the cost of the materials, the concepts presented above enable the integration of a IxN WSS with an independent NxI WSS both with power monitoring on all ports. This is most advantageous as it enables an electronic closed-loop control system to maintain the coupling efficiencies between the different ports in the presence of component misalignment post-assembly, or subsequently during operation in the presence of environmental change. Power monitoring also enables independent control of the optical power levels in the different spectral channels via the beam-steering element for e.g. differential channel equalisation.

From the standpoint of manufacturability , it is particularly advantageous to have the electrically- activated beam-steering elements and the photodetector arrays located in the same plane. Thus, the beam- steering elements and the photodetector arrays can optionally be bonded onto a common substrate and packaged together and, in addition, all electronics necessary for operation of the beam-steering elements and the photodetector arrays can be located in close proximity and mounted on a single board.

It will be appreciated ^" that in the WSS embodiment of Figure 11, beam-steering elements could alternatively be substituted for the arrays of photodetector elements 1104-11 through 1104-14 and 1104-31 through 1104-34, and that the photodetector arrays could then be located in the plane of the PLC. The beam-steering elements inserted in these tap paths could be activated to direct

the beams associated with the corresponding spectral channel to the appropriate photodetector element located at the PLC substrate underside or at the PLC upper surface .

In a similar manner, within the WSS embodiment of Figure 13 the beam-steering elements could alternatively be substituted for the arrays of photodetector elements 1304-11 through 1304-14 and 1304-31 through 1304-34, and the photodetector arrays could then be located in the plane of the PLC. This alternative embodiment overcomes the limitation associated with the embodiment of Figure 13 in that the PDL in the tap path associated with a single pass from reflection grating 1302 is now substantially suppressed as a result of the double pass from the gracing. The beam-sceering elements inserted in these tap paths could be activated to direct the beams associated with the corresponding spectral channels to the appropriate photodetector element located at the PLC substrate underside or the PLC upper surface.

In a variant of these alternative embodiments, additional PLC photodetector ports could be created by adding re- direction mirrors: one re-direction mirror (with the same orientation as the input tap port re-direction mirror) for every spectral channels in the input tap path, and one re-direction mirror (with the same ^" orientation as the output tap port re-direction mirrors) for every spectral channels in the output tap path. ' Then, the beam-steering elements inserted in the tap paths could be activated to direct the beams associated with the corresponding spectral channels to the appropriate PLC photodetector port. ^• In these variants of the alternative embodiments, - waveguides could be employed to route light from the PLC

phobodetector ports to arrays of photodetector elements bonded onto one or more of the PLC edges, or waveguides could be employed to route light from the PLC photodetector ports to further re-direction mirrors that reflect the light onto arrays of photodetector elements bonded onto the PLC substrate underside or the PLC upper surface.

As an alternative to any of the foregoing WSS embodiments, the quarter-wave plates can be removed, and polarisation diversity optics can be used instead to substantially suppress PDL between the input pore and any other point in the WSS system.

It will be appreciated that in addition to the integrated channel waveguide taps on the PLC, other integrated waveguide devices could, in principle, be integrated on the PLC to perform a wide range of different optical processing functions including but not limited to: alteration or modulation of the optical field amplitude or phase; WDM MUX or DEMUX; optical routing, branching, switching or coupling; optical tap-detection for power monitoring and channel OSNR monitoring; wavelength conversion and dispersion compensation.

In a further alternative the 3x3 PLC port array concept presented in Figure 6 could be realised by- employing optical fibres arranged in a 3x3 fibre array with the optical axes of the individual fibres all aligned substantially parallel to the optical axis 650. Furthermore, it will be obvious that the PLC port array concepts with integrated channel waveguide taps presented in Figure 8 to Figure 10 could alternatively be realised by employing optical fibre taps and optical fibres

arranged in a 3x3 fibre array with the optical axes of the individual fibres all aligned substantially parallel to the optical axis.

A further alternative to the 3x3 PLC port array concept presented in Figure 6 is now presented in Figure 14. In this embodiment the re-direction mirrors 1460 have been profiled to simultaneously effect re-direction and beam collimation and thereby combine the functions of the port array 500 and the collimating element array 501 of Figure 5 in a single component. Note that the profiled re- direction mirrors 1460-11 through 1460-33 are profiled or curved as shown in Section XX with a radius of curvature about the x axis and curved with a radius of curvature about the z axis (not shown explicitly) i.e. the re- direction mirrors are curved about two axes. The crucial point here is that the port array 1400 is strictly the array of waveguide facets while the array of collimating elements is now coincident with the array of re-direction mirror facets 1460-11 through 1460-33. Hence, to achieve an offset between any port and the corresponding collimating element in this embodiment, it is necessary to offset the curved re-direction mirror in the x direction relative to the corresponding waveguide facet (as shown for 1400-11, 1400-21 and 1400-31) or vice versa .

Various representations of two integrated Ix2-port WSSs in accordance with an aspect of the present invention are presented in Figure 16 and Figure 17. Note that, for clarity, the principal rays corresponding to the beams in the return path for both the 1x2 WSSs have been omitted. In addition the corresponding reference numerals within Figure 17 have been prefixed with the number "17" rather

than "16" when compared with Figure 16. Both WSSs employ the same dispersive element 1602 and the same focussing element 1603 as described previously. The respective ports 1600-la, 1600-2a and 1600-3a correspond to the first 1x2 WSS while ports 1600-lb, 1600-2b and 1600-3b correspond to the second 1x2 WSS. The ports are again arranged within two distinct blocks.

However, in the presently described embodiment there is only a relative displacement ^' between the ports and the collimating elements 1601-1, 1601-2 and 1601-3 along a single axis, namely a vertical axis. In particular, the respective ports 1600-la, 1600-2a and 1600-3a, corresponding to the fix's t 1x2 WSS, have a vertical, upwards offset such that the spectral channels of the input light interact with the lower 1-D array of beam steering elements 1604-al, 1604-a2, 1604-a3 and 1604-a4. Similarly, the respective ports 1600-lb, 1600-2b and 16O0-3b, corresponding to the second 1x2 WSS, have a vertical, downwards offset such that the spectral channels of the input light interact with the upper 1-D array of beam steering elements 1604-bl, 1604-b2, 1604-b3 andl604-b4.

The respective upwards and downwards offsets of the ports of the 1x2 WSSs are chosen such that corresponding ports of the WSSs share a common collimating element. Such an arrangement has the particular advantage that it provides for further miniaturisation of the device. As described previously it should be noted that multiple spectral channels within either of the 1x2 WSS can be directed to any of the corresponding output ports, but that a given spectral channel can only appear at a single output port at any particular time.

Figure 18 presents a schematic representation of a 1x6 port array 1800 suitable for use within the two integrated independent Ix2-port WSSs of Figures 16 and 17. The port array is based on a PLC with the port positions defined by the end faces 1800-la to 1800-3b of the waveguides 1830-la to 1830-3b.

With reference to Figure 18 (a) light propagates from left to right through the input waveguides 1830-la and 1830-lb and onto the 1x3 array of collimating elements. After propagating through the remaining components of the system, light that has entered the system via input port 18O0-la propagates from right to left via waveguide 1830- 2a and/or 1830-3a before exiting the system. In a similar, manner light that has entered the system via input port 1800-lb propagates from right to left via waveguide 1830-2b and/or 1830-3b before exiting the system.

A significant advantage of employing the 1x6 port array 1800 resides in the fact that its simplified design removes the requirement for integrated redirection mirrors within the waveguides 1830-la to 1830-3b. As a result the manufacture of the 1x6 port array is simplified thus making such an array design cheaper to produce .

The above described embodiments of the present invention have predominantly been described in relation to input port arrays and collimating element arrays comprise 3 x3 elements. However, it will be readily apparent to those skilled in the art that the aspects of the present invention are not limited to arrays comprising such

arrangement of elements . In essence the apparatus and methods can be easily extended so as to convert a IxN Wavelength Selective Switch (WSS) into an integrated, but independent, first IxM WSS and an Ix(N-M-I) node optical device. The optical device may take the form of a second WSS and/or may provide a facility for optically tapping any of the input or output channels of the first IxM WSS. What is required to achieve these results is the introduction of a relative offset between a first port sub-array of the port array of the device and corresponding elements of the optically connected collimating element array. Furthermore IxN optical device can be converted to multiple (Z) IxN WSSs where this is dictated by the number of unique Z port sub-array off-sets. The sub-array can define other optical devices such as optical taps. A number of ways of incroducing this relative offset have been described e.g. the employment of an irregularly arranged port array or the employment of an irregularly arranged collimating element array.

Finally, it will be appreciated by those skilled in the art that the array of beam-steering elements appearing in all the WSS embodiments can be based on any technology including but not limited to MEMS analogue tilting mirrors, diffractive MEMS or liquid crystal spatial light modulators .

The foregoing description of the invention has been presented for purposes of illustration and description and is not intended to be exhaustive or to limit the invention to the precise form disclosed. The described embodiments were chosen and described in order to best explain the principles of the invention and its practical

application to thereby enable others skilled in the art to best utilise the invention in various embodiments and with various modifications as are suited to the particular use contemplated. Therefore, further modifications or improvements may be incorporated without departing from the scope of the invention as defined by the appended claims.