Optical device and methods

Abstract

Methods and devices for manipulating optical signals. In one example, a LCOS (liquid crystal on silicon) device includes a surface bearing an anti-reflection structure. The anti-reflection structure includes i) a physical surface having a topography with features having lateral dimensions of less than 2000 nm and having an average refraction index which decreases with distance away from the surface; and ii) a configuration of the topography, averaged over lateral dimensions of greater than 2000 nm, varies with lateral position on the surface.

Claims

1. An LCOS (liquid crystal on silicon) device comprising a surface bearing an anti-reflection structure, wherein: i) the anti-reflection structure comprises a physical surface having a topography with features having lateral dimensions of less than 2000 nm and having an average refraction index which decreases in a direction that is perpendicular to said surface; and ii) a configuration of said topography comprising a stepped pyramid, averaged over lateral dimensions of greater than 2000 nm, varies with lateral position on said surface.

2. An LCOS device as claimed in claim 1 wherein said surface is a front surface of said device.

3. An LCOS device as claimed in claim 1, combined with a controller to display a hologram on said LCOS device to deflect first light into a first diffractive order of said hologram, wherein said hologram is further configured to deflect second light into a second diffraction order of said hologram, wherein said LCOS device has an interface generating unwanted reflected light, and wherein said second light is in antiphase with said unwanted reflected light.

4. An LCOS device as claimed in claim 3 wherein said second light has substantially the same power as said unwanted reflected light.

5. The LCOS device of claim 1 in an optical system comprising a light source having at least two different wavelengths, 1 and 2, wherein a first lateral region of said surface is adapted by said topography for anti-reflection at 1 and a second lateral region of said surface is adapted by said topography for anti-reflection at 2.

6. The LCOS device of claim 1 in an optical system comprising a wavelength-selective demultiplexer configured to direct light of at least two different wavelengths, 1, and 2, towards different spatial regions of said LCOS device, wherein a first lateral region of said surface is adapted by said topography for anti-reflection at 1 and a second lateral region of said surface is adapted by said topography for anti-reflection at 2.

7. A method of suppressing an unwanted reflection using the LCOS device; wherein the LCOS device comprises a surface bearing an anti-reflection structure, and i) the anti-reflection structure comprises a physical surface having a topography with features having lateral dimensions of less than 2000 nm and having an average refraction index which decreases in a direction that is perpendicular to said surface; and ii) a configuration of said topography comprising a stepped pyramid, averaged over lateral dimensions of greater than 2000 nm, varies with lateral position on said surface, the method comprising configuring said anti-reflection structure such that, for at least two different wavelengths, 1, and 2, in a first lateral spatial region of said surface, reflection from said surface at 1 is greater than at 2, and in a second, different lateral spatial region of said surface, reflection from said surface at 2 is greater than 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) In the following embodiments of the invention will be described, by way of example only, with reference to the accompanying drawings in which:

(2) FIG. 1 shows a schematic cross section of an optical routing device;

(3) FIG. 2 shows a graph illustrating the behaviour of a blue phase structure under applied electric field;

(4) FIG. 3 shows the phase shift of light transmitted through a cell as a function of voltage for three wavelengths and both polarisations;

(5) FIG. 4 shows the cell and detected intensity of light detected;

(6) FIG. 5 shows voltage schemes for driving liquid crystal material on top of an LCOS backplane;

(7) FIG. 6 shows a phase/voltage curve for a cell;

(8) FIG. 6A shows an example phase/voltage curve for a cell;

(9) FIG. 7 a driving scheme for a cell having a controllable variable amplitude square wave drive with a separately defined offset;

(10) FIGS. 8 and 9 show a system diagram for a multiple wavelength polarisation independent router;

(11) FIG. 10 shows a switching device with a segmented front electrode;

(12) FIG. 11 shows a switching device mounted on a substrate;

(13) FIG. 12 shows a cross section of an optical device;

(14) FIG. 13 shows an optical device in use;

(15) FIG. 14 shows an apparatus for modifying the phase of a beam of light;

(16) FIG. 15 shows an adaptive optics system applied to a telescope;

(17) FIG. 16 shows a system for modifying the phase of a beam of light in a transmissive configuration;

(18) FIG. 17 shows a graph of phase shift as a function of angle of incidence for normal and oblique incidence;

(19) FIG. 18 shows the maximum variation of the phase shift as a function of the incidence angle;

(20) FIG. 19 shows a schematic diagram of a LCOS SLM;

(21) FIG. 20 shows an illustration of a problem with reflection from a front coverplate surface;

(22) FIG. 21 shows an example of light being dispersed across the face of a LCOS SLM;

(23) FIGS. 22A to 22C illustrate three potential coating schemes, showing (A) thin film dielectric layers, (B) a graded index layer, and (C) patterned microstructures;

(24) FIG. 23 shows an illustration of an effective medium representation of a structured anti-reflective surface; and

(25) FIG. 24 shows a block diagram of an SLM controller/driver configured to drive an SLM with a reflection-compensation hologram.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

(26) FIG. 1 shows a schematic cross section of an optical routing device 100. The device is formed on a substrate 102. A plurality of optically reflective pixel electrodes 104 are arranged on the substrate 102. A layer of adjustable refractive index material 106 is arranged over the pixel electrodes 104. A transparent counter electrode 108 is arranged over the layer of adjustable refractive index material 106.

(27) The adjustable refractive index material 106 has a refractive index which varies with the electric field applied across it. An example of such a material is blue phase liquid crystal. The adjustable refractive index material 106 has an isotropic refractive index in the direction perpendicular to the applied electric field.

(28) In the device shown in FIG. 1, by applying a spatially varying pattern of voltages between the electrodes 104 and the counter electrode 108, the refractive index of the adjustable refractive index material 106 can be caused to vary spatially. In order to direct a beam of light incident on the upper surface of the counter electrode 108, a first pattern of voltages is applied across the adjustable refractive index material 106 that cause its refractive index to vary spatially to form a first diffraction grating. The first diffraction grating deflects the incident beam of light in a first direction. To switch the beam of light to a second direction, the applied pattern of voltages is changed to a second pattern of voltages that cause the refractive index of the adjustable refractive index material 106 to vary spatially to form a second diffraction grating.

(29) By switching between different patterns of voltages as described above, an incident beam of light can be switched between different outputs.

(30) If a material that has an isotropic refractive index in the direction perpendicular to the applied electric field is used, the switching is independent of the polarisation of the incident beam.

(31) In the following, the wording isotropic perpendicular to the applied electric field is used to describe that there is substantially no difference in refractive index in any direction perpendicular to the applied electric field, the refractive index may vary in any other direction.

(32) FIG. 2 shows the behaviour of a blue phase structure under applied electric field. The blue phase liquid crystal has symmetrical domains of liquid crystal molecules arranged in a structure which, with no applied electrical field, is optically isotropic. A blue phase liquid crystal material is a liquid crystal material in a blue phase. Examples of blue phase liquid crystal materials stable over a wide temperature range can be found in US2009/0115957, to which reference may be madethis describes materials stable over a range of greater than 35 C. and/or which are stable at room temperature.

(33) Deforming the domains (e.g. by applying an electric field) along the propagation direction of the light can be used to modulate polarised light, since the deformed domain undergoes a change in optical refractive index and becomes optically anisotropic. If the deformation is along the light transmission axis so that the deformed domain is still symmetrical normal to the light transmission, a localised change in refractive index occurs which is independent of the incident polarisation angle, and can be used to produce SLMs capable of polarisation independent phase modulation.

(34) The behaviour discussed above was experimentally demonstrated as follows. A liquid crystal LCM-PSBP.1328UV (from LC Matter Corp.) was used for the experiment. A 12 micron thick glass cell with ITO electrodes and no other alignment layers was filled with the mixture. Square waveform voltage with frequency of 1 kHz was applied across the cell in order to introduce anisotropy in the Blue Phase.

(35) FIG. 3 shows the phase shift of light transmitted through the cell as a function of voltage for three wavelengths and both polarisations. Phase shift was identical for both polarisations within the range of applied voltage for any wavelength. For wavelength of 480 nm, a 68 volt square wave at 1 kHz gave a phase shift as shown of greater than 2.

(36) Phase shift was measured using a cell as a Fabry-Perot interferometer. The cell and detected intensity are shown in FIG. 4. The spectrum of the light transmitted through the cell showed a characteristic interference pattern due to the direct and double-reflected light in the cell. The positions of the interference peaks depend on the refractive index of the material and move as voltage applied across the cell.

(37) The high voltage required to get 2 phase shift reflects the fact the material has low Kerr constant (K) of 1.6 nm/V2 at 600 nm. The drive voltage and cell thickness should be reduced considerably by using materials with the large Kerr constant (K up to 12.7 nm/V2) that have been recently developed for in-plane switching display applications. For this material a 2 phase shift at 600 nm wavelength should be obtained with a 10 m reflective cell and driving voltage of 14V, well within the range of LCOS backplanes.

(38) Thus, the inventors have demonstrated a method for phase modulation of randomly polarised light.

(39) It is envisaged that the liquid crystal material may be chosen in order to get a wide temperature range blue phase. Mixtures of a liquid crystal and a polymer have been shown to stabilise the blue phase in a temperature range from 60 C to 10 C degrees. In order to stabilise the blue phase in a temperature range, the blue phase material may be doped with nanoparticles. This improves the sensitivity of the blue phase material to electric fields.

(40) Another possibility is the use of bimesogenic liquid crystals.

(41) The Blue Phase has the additional advantage that no alignment or rubbing layers are required. This makes possible simplified polarisation-independent telecoms routing.

(42) FIG. 5 shows two ways of driving the liquid crystal (LC) material on top of the LCOS backplane. The first driving scheme is an analogue variable amplitude square wave drive. Here the counter electrode which is the front electrode (FE) of the LCOS device is maintained permanently at half the supply voltage of the pixel driver transistors, and a variable amplitude square wave is applied to the pixels to give a symmetrical dc-balanced square wave of adjustable amplitude across the pixels. The mean RMS voltage determines the state of the LC.

(43) The second method, digital variable mark/space ratio drive, involves applying a square wave to the front electrode of the same amplitude as the supply voltage, and the pixels are switched at different times between the supply or ground. This allows a dc-balanced but asymmetrically timed waveform across the LC which has a variable RMS value, to which the LC responds, giving an average response similar to the fixed FE version, but with small variations as the LC follows the waveform.

(44) The latter method is satisfactory for image projection applications, but causes unacceptable phase flicker which decreases the efficiency of phase holograms. However it allows twice the drive voltage across the liquid crystal from the same backplane process, or for the same voltage requires a lower voltage process, allowing smaller transistors and therefore higher integration, lower power, and lower cost.

(45) Since rows of pixels are usually loaded sequentially down the pixel array a combination of the two causes progressive loss of drive and increased phase flicker down the pixel array.

(46) The blue phase liquid crystal responds more rapidly to the applied electric field than a nematic phase liquid crystal. As shown in FIG. 3, the refractive index change is proportional to the square of the applied field. This means that for a given backplane voltage swing the phase modulation may be quite small, and may require a very high drive voltage.

(47) FIG. 6 shows a phase/voltage curve for a cell. If the backplane drive swing can be pushed further up the response curve, it is possible to get an improved phase modulation for the same backplane swing. As can be seen in FIG. 6, by offsetting the B/P Drive voltage, the phase response can be increased from the range shown as Phase no offset to the range Phase with offset. This can be implemented by overswitching the front electrode above the supply rail and below ground to produce an offset added to the backplane drive voltage, but only by using the second drive scheme described above. A combination of the first and second schemes allows this; rather than switch between the supply rails, the pixels are switched at the FE transition time to different analogue voltages, thus giving a controllable variable amplitude square wave drive with a separately defined offset.

(48) FIG. 6a shows a graph of phase modulation in steps of pi against drive voltage across a cell. The phase modulation is proportional to the square of the voltage applied across the cell. FIG. 6a shows example values for the voltage and phase modulation to demonstrate the effect of offsetting the voltage on the phase change.

(49) As can be seen in FIG. 6a, from 0 volts a drive voltage of 82 volts 602 is required to give a phase shift of 2pi 604. However, if an offset of 82 volts is applied, if takes only 35 volts 608 of change to give the same differential phase modulation 606. If the front electrode is switched above and below the supply rail by this voltage, the modulation required to drive the pixel is reduced from around 80 volts to around 35 volts.

(50) FIG. 7 shows such a driving scheme. The front electrode (FE) voltage is overswitched above and below the VLSI supply voltage and Ground. This allows the full VLSI voltage range to be used for phase control further up the phase/voltage curve shown in FIG. 6.

(51) However for this to work on a large 2-dimensional pixel array, all the rows of pixels have to be loaded on the FE transition, otherwise phase flicker will be reintroduced which will get progressively worse down the rows of refreshed pixels.

(52) The chip can be designed so that rows of rectangular pixels forming 1-dimensional arrays, suitable for grating-type holograms, can be driven, with a horizontally segmented FE. This allows a separate driver for each FE segment, ensuring that the FE switches at the same time that the pixels are refreshed, eliminating the phase flicker, and allowing the LC drive to be offset up the response curve.

(53) For a multiple wavelength routing device a system can be designed to ensure that the separated wavelengths fall onto the required segments. Such a system requires a reasonably polarisation independent wavelength splitting grating and allows a single-device polarisation independent router.

(54) FIGS. 8 and 9 show a system diagram for a multiple wavelength polarisation independent router 800. FIG. 8 is a side-view and FIG. 9 is a top view.

(55) The router 800 switches signals between fibres of a fibre array 802. An input signal is received in an input fibre 802. This is switched to one or more output fibres depending on the state of the switch. The input signal is passes through a lens 806 and is wavelength demultiplexed by wavelength demultiplexer 808. Different wavelengths are directed to different areas of the LCOS device 812. The LCOS device 812 applies a hologram/grating which deflects the wavelength incident on that particular area in the direction orthogonal to the wavelength splitting axis. The beam is reflected through a wavelength multiplexer 814 which collects together all the wavelengths which have been deflected by a given angle and these are then focussed by a lens 816 into the corresponding output fibre 818. The angle of deflection (inversely proportional to grating pitch) determines the destination fibre.

(56) This LCOS device can be segmented by having strips of ITO on the cover glass each corresponding to a separate wavelength channel, so that each wavelength can be driven independently of other wavelengths.

(57) FIG. 10 shows a switching device 1000 with a segmented front electrode. The LCOS device 1010 has a front electrode divided into strips 1012. Each strip 1012 covers a plurality of electrodes 1014 on the opposing side of the variable refractive index material. The wavelength demultiplexer directs different wavelengths of the input signal 1018 to different strips. The pixel dimensions on the LCOS backplane are between 1.6 and 15 microns.

(58) Beneath each front electrode strip there is a one-dimensional strip having the same width as the front electrode strip. The pixels may be rectangular as shown in FIG. 10, or may be square.

(59) The segments which form the diffraction grating can be up to 20 mm high and unto 100 microns wide.

(60) FIG. 11 shows a switching device 1100. The device is mounted on a substrate 1102. A temperature control may be attached to the substrate 1102 which controls the temperature of the device so that the liquid crystal material is maintained in the blue phase. A flexible PCB 1104 is mounted on the substrate 1102. An active backplane 1106 is also mounted on the substrate 1102. The active backplane 1106 runs under a layer of blue phase liquid crystal material. Bonding wires 1108 connect the active backplane 1106 to the flexible PCT 1104. A front glass cover 1110 is located over the layer of blue phase liquid crystal material. The glass cover carries stripe front electrodes 1112 which run across the blue phase liquid crystal material. The flexible PCB runs beside the edge of the blue phase liquid crystal material and the glass over overlaps the flexible PCB 1112, this allows connections between the flexible PCB 1104 and the front electrodes 1112 to be made.

(61) The front electrode connections are made by silver loaded epoxy from the ITO front electrodes on the front glass cover.

(62) FIG. 12 shows a cross section of an optical device 1200. The device has a transparent substrate 1202. The transparent substrate is formed from glass. A transparent electrode formed by transparent conducting oxide layer 1204 is coated on the substrate 1202. The transparent conducting oxide layer is, for example, formed from Indium tin oxide. Beneath the transparent conducting oxide layer 1204 there is a layer of adjustable refractive index material 1206. The layer of adjustable refractive index material 1206 is, for example a layer of blue phase liquid crystal as described above. Beneath the adjustable refractive index layer 1206, there is a reflective layer 1210. the reflective layer is formed from for example, 15 layers of zinc sulfide and magnesium fluoride. Beneath the light blocking layer 1210 there is a layer of a photoconductive material 1208. The photoconductive layer 1208 is, for example, formed from hydrogenated amorphous silicon or cadmium sulfide. Beneath the photoconductive layer 1208, there is a second transparent electrode which is formed by transparent conducting oxide layer 1212 coated on a second transparent substrate 1214. The first and second electrodes are planar and occupy planes that are parallel and spaced apart with the photoconductive layer and the adjustable refractive index material are disposed between the first and second electrodes.

(63) FIG. 13 shows the optical device 1200 in use. A voltage waveform 1220 is applied between the electrodes formed by the transparent conducting oxide layers 1204 and 1212. A write beam 1230 of light which varies laterally across the device is applied to the device through the substrate 1214. The write beam 1230 causes the conductivity of the photoconductive layer 1208 to vary spatially. Since the conductivity of the photoconductive layer 1208 varies spatially, the voltage across the adjustable refractive index layer 1206 will vary spatially corresponding to the spatial variation in the write beam 1230.

(64) A read beam 1240 is incident on the substrate 1202 on the side of the device including the adjustable refractive index layer 1206. The variation in the voltage across the variable refractive index layer 1206 causes the refractive index of the variable refractive index layer to vary spatially. Therefore, the phase of the outgoing light 1242 reflected from the reflective layer 1210 can be modified. The reflective layer 1210 stops the write beam from passing through the device and the read beam from activating the photoconducting layer.

(65) The voltage waveform 1220 may take the form described in reference to FIGS. 5 to 7 above. It is noted that an optically driven device as shown in FIGS. 12 and 13 may be driven with higher voltages than those available from an LCOS backplane.

(66) In place of or in addition to the reflective layer, the device may also comprise a light blocking layer. Such a light blocking layer could be arranged below the reflective layer and configured to absorb the 1-5% of residual light and stop it hitting the photoconductive layer.

(67) In an alternative embodiment, the device may be used in a transmissive configuration. In this configuration, there is no reflective layer and the read beam of light passes through the device. In such an embodiment, different wavelengths of light are used for the read and write beams of light or each beam is pulsed in a different part of the voltage waveform.

(68) FIG. 14 shows an apparatus for modifying the phase of a beam of light using the device 1200 described above with reference to FIGS. 12 and 13. The apparatus has a light source 1260 such as a laser or high brightness light emitting diode. The light from the light source is collimated by a lens 1262 onto a beamsplitter 1264. The beamsplitter directs the light onto a microdisplay 1266. The microdisplay 1266 may be a liquid crystal on silicon display or a digital light processing display. The microdisplay 1266 sets the pattern of the write beam which will be incident on the device 1200. The signal from the microdisplay 1266 passes back through the beamsplitter 1264 and then through a relay lens 1268. The signal is imaged on the transparent substrate 1214 of the device 1200. The spatial variation of intensity of the write signal causes the resistance of the photoconductive layer 1208 to vary according to the electrical signal that is input to the microdisplay 1266. This variation causes the voltage across the variable refractive index material 1206 caused by the voltage between the electrodes 1212 and 1204 to vary. A read light, for example a signal to be routed or a signal having its wavefront modified is incident upon the substrate 1202 on the opposite side from the side on which the write signal is incident. The read signal is modified by the refractive index of the variable refractive index layer 1206, however since the refractive index of the variable refractive index material is isotropic in the direction perpendicular to the applied electric field, the modification of the read signal is polarisation independent.

(69) The device described above has applications in adaptive optics. Light travelling through a medium can be disrupted by variations in refractive indexe.g. in the atmosphere due to turbulence arising from pressure changes or convection caused by temperature changes. For any object viewed in the far field, this alters the shape of the plane wave across the entry optics, and limits the resolution of the optics below the theoretical limit. E.g. for an astronomical telescope this gives a blurred jittering image of a star which should be a point source. For planetary or surveillance images this also gives instantaneous spatial distortion.

(70) Adaptive optics can partially compensate for the disturbance, usually by using a deformable mirror with an array of electro-mechanical actuators. These can correct the wavefront deformation of a few microns and partially restore the wavefront.

(71) FIG. 15 shows an adaptive optics system applied to a telescope. The telescope 1500 has a concave mirror 1502 and a convex mirror 1504. The waveform 1506 of light coming into the telescope 1500 has fluctuations due to atmospheric disturbance. The fluctuating waveform is focussed by the mirrors 1502 1504 to give an input waveform 1508.

(72) The input waveform 1508 is partially split by a beamsplitter 1510. Part of the beam is incident upon a wavefront sensor 1512 and the remaining part of the beam is incident upon an active phase correction device 1514 which comprises a layer of variable refractive index material such as blue phase liquid crystal as described above. The output of the wavefront sensor 1512 is used by an adaptive optics correction controller 1516 to control the spatial variation of the voltage applied across the variable refractive index material and thereby control the phase correction. The corrected beam 1518 is output to imaging optics of the telescope.

(73) Adaptive optics can also be used to correct for aberrations in optical systems, improving vision to a greater extent than lens correction, and e.g. for laser surgery, when the imperfections of the eye's lens must be corrected for retinal welding, etc.

(74) Many adaptive optics applications require polarisation independence and a rapid (millisecond) response time, the devices described above are suited to these applications since they allow for polarisation independent phase correction without a requirement for moving parts such as adjustable mirrors.

(75) FIG. 15 shows an open-loop system, however, those of skill in the art will recognise that the devices described above could also be applied to a closed loop system.

(76) FIG. 16 shows a system for modifying the phase of a beam of light which has an active phase correction device in a transmissive configuration. A beam of light from a source 1702 is broadened and deflected by atmospheric interference to give waveforms 1704 with fluctuations 1704. The phase correction device 1706 corrects the phase of the beam to give a phase flat beam.

(77) The devices described above provide polarisation independent phase modulation for light beams of normal incidence. For beams with a non-zero angle of incidence, some polarisation dependence is introduced into the phase change.

(78) FIG. 17 shows phase shift as a function of the polarisation angle for normal and oblique incidence. There is a phase change of up to .sub.max for light beams incident at a non-zero angle of incidence.

(79) FIG. 18 shows Maximum variation of the phase shift as a function of the incidence angle. For normal incidence the polarisation ellipsoid is seen along its long axis as a circle, therefore, the phase shift is independent of the polarisation angle. For any oblique incidence the polarisation ellipsoid is seen at some angle as an ellipse, therefore, the phase shift is now depends on the polarisation angle.

(80) As can be seen from FIGS. 17 and 18, the angle of incidence is preferably kept low to ensure that the phase shift is relatively polarisation independent. If a variation of /8 is the accepted variation then the angle of incidence of the incoming beam should be kept within 5 degrees.

(81) Anti-Reflection Structures/Coatings

(82) We will now describe how the performance of blue phase and other LCOS SLMs can be enhanced using structured surfaces on the front SLM cover-plate to reduce reflections. This is particularly applicable to wavelength selective switches, where a very low reflectivity over a large bandwidth is desirable.

(83) The example LCOS SLM of FIG. 19 has three components: a silicon backplane, a coverplate with a common electrode on one surface, and a liquid crystal layer. The coverplate is typically float glass with a thin-film anti-reflection coating on the front surface, and an ITO layer. Optionally, a liquid crystal alignment layer may be provided on the back surface, although this is not required for a blue phase device. Ideally, both the front coverplate surface and the coverplate/liquid crystal interface (comprising glass, ITO, optional alignment layer, and liquid crystal layers) have zero reflectivity. Thus the wavefront exiting the SLM is solely affected by the phase delay imparted across the liquid crystal layer.

(84) For most applications, the front surface reflectivity can be sufficiently reduced through the use of standard thin film coatings. However, for telecom applications, such as the implementation of wavelength selective switches based on beam-steering, the reflectivity of the front surface should preferably be reduced to R.sub.f<0.01% over a wide wavelength range in order to minimize crosstalk. To illustrate this consider FIG. 20, where a plane wave at an angle of .sub.i is incident on a LCOS SLM that displays an ideal blazed grating of period T. A fraction of the signal beam (m=1) is reflected back towards the blazed grating which is re-diffracted such that it travels in the same direction as the m=2 order. The m.sup.th diffraction order of the grating is diffracted through an angle of .sub.m according to

(85) $\begin{array}{cc}\sin {}_m=\sin {}_i+\frac{m}{T}& \left(1\right)\end{array}$
where is the wavelength of the light and the angles are the values measured in air. For an ideal blazed grating, all the light is diffracted into the m=+1 order. However, due to spatial and phase quantization of the grating, some light ends up in higher and symmetric (m1) orders. For certain switch geometries, light in these m1 orders can couple into output ports leading to crosstalk. For practical telecom applications the crosstalk should preferably be suppressed to <40 dB.

(86) Consider the case where .sub.i=0. As mentioned, ideally all the light is diffracted into the +1 order through an angle of .sub.1. However, if the front surface of the SLM has a finite reflectivity of R.sub.f, a portion of the +1 order will be reflected back towards the SLM at an angle .sub.1 with respect to the normal. This order will be re-diffracted by the grating such that it propagates at an angle of 2.sub.1 to the normal, which, for small diffraction angles, corresponds to the angle of propagation of the m=+2 order of the original diffracted beam, .sub.2. Assuming a coverplate with refractive index of n=1.5, and no anti-refection coating, the Fresnel reflection coefficient of the front coverplate surface is approximately 4%. Let us assume an SLM which has 100% diffraction efficiency (no absorption losses and a perfect blazed grating). The power in the m=+1 and m=+2 directions will therefore be given by 0.96P.sub.in and 0.04P.sub.in respectively (ignoring multiple reflections). As a result, the theoretical crosstalk is 13.8 dB. This compares well to values of 14 to 18 dB measured experimentally. If we reduce the front face reflectivity to 1% (typical of single layer thin film coatings), the crosstalk reduces to approximately 20 dB. To reduce the crosstalk to <40 dB the front face reflectivity should preferably be reduced to <0.01%. Such a low reflectivity is challenging, particularly as we should preferably ensure this value is maintained across the C-band, L-band, or C and L band. For example, the C-band extends from 1530 nm to 1570 nm, and L-band from 1565 nm to 1625 nm, and an optimized multi-layer coating can expensive to fabricate.

(87) FIG. 21 shows the functional operation of a LCOS SLM used in conjunction with a wavelength de-multiplexer. Such a configuration may be used in wavelength selective switches, where separate sub-holograms deflect specific optical wavelengths, and in pulse shaping systems, where light is dispersed across the face of the SLM and each wavelength band is modified to produce the desired output pulse. Ideally we desire an anti-reflection coating that has continuously varying optical properties to match the incident wavelength.

(88) Although particularly important for telecom applications, other applications that employ the optimization of complex hologram patterns also benefit from a reduction of front face reflectivity. The quality of the replay field in display holography and optical tweezers will improve as reducing R.sub.f ensures that the output wavefront more closely matches the designed wavefront.

(89) It is desirable to provide an anti-reflection layer with the following properties: a. Sufficient angular independence b. Good polarization independence c. Substantially no wavelength dependence over the desired operational bandwidth d. Low temperature sensitivity e. Is cheap and robust f. Can be optimized across the surface of a coverplate to match operational wavelength range

(90) There are four potential techniques which may be employed for reducing crosstalk in practical LCOS SLM applications: geometric, thin film dielectric coatings, graded index coatings, and patterned nano-structures: 1) GeometricalRather than using a coverplate with parallel surfaces, we use a wedged coverplate that reflects light away from the main replay field. The applicability of this approach depends on the actual optical geometry, and may not be practical in some WSS (wavelength selective switch) architectures. 2) Thin film dielectric coatingsThis technique for implementing anti-reflection coatings is illustrated in FIG. 4a. Typical C-band coatings available off-the-shelf can have a reflectivity <0.1%. To reduce R.sub.f to <0.01% on a complex substrate (ITO layer and optional alignment layers on the reverse side) could be done with an expensive custom coating run. However, as thin film coatings exhibit polarization dependence if used off-axis, it will be important to factor in the beam steering angles and alignment tolerances when designing the coating. Finally, the flexibility of thin film coatings is limited by the number of suitable deposition materials available (refractive indices), temperature stability, and suffer potential delamination problems if environmental conditions change. 3) Graded index coatingsThis refers to a class of coatings that use an impendence matching layer with a graded change in refractive index as illustrated in FIG. 4b. By continuously increasing the refractive index from n.sub.air to n.sub.s, where n.sub.s is the substrate refractive index, one can avoid sharp dielectric boundaries that cause Fresnel reflection. The profile can be, for example, parabolic or cubic, and can be implemented by varying the packing density of the films or by using inclined nano-rods. 4) Patterned nano-structuresThis refers to a technique based on patterned nano-structured surfaces, an example of which is illustrated in FIG. 4c. The dimensions of nano-structures are <, so the incident light sees an average refractive index at a specific value of z that depends on the ratio of air and substrate. Thus there is an effective gradually change in refractive index through the patterned layer from n.sub.air to n.sub.s. The structures can be designed and optimized using an effective medium theory (EMT) where the refractive index depends on the topology of the material. There are two main approaches for making such layers. The first is by imparting porosity to the film, with the porosity decreasing as we approach the substrate, for example by sol-gel processing. The second approach involves fabricating arrays of nano-structures that either suppress reflections by light trapping and multiple-internal reflections (surface texture) or by using an anti-reflection grating where the grating topology is engineered to ensure that only the m=0 order of the grating propagates, and that it produces a continuously varying refractive index. FIG. 23 illustrates this approach, with a patterned nano-structure with N=6, showing the structured surface on the left and a corresponding effective equivalent medium model on the right. The effective index of the six layers are set such that n.sub.air=n.sub.a<n.sub.1<n.sub.2<n.sub.3<n.sub.4<n.sub.5<n.sub.6<n.sub.s=n.sub.substrate.

(91) Graded Index/Patterned Nano-Structures for Telecoms Wavelengths

(92) Raguin and Morris showed theoretically that a multi-level approximation to a pyramidal anti-reflection structure fabricated on a GaAs substrate (N=2, 4, and 8 phase-levels) and optimized for operation at 10.6 m could have a reflectivity <310.sup.3% for all values of N for a randomly polarized normally incident beam (D. H. Raguin and G. M. Morris, Antireflection structured surfaces for the intrared spectral region, Appl. Opt. 32, 1154-1166 (1993)). This exceeds the above target of 0.01%. Moreover, their model predicted a T=99.9% transmission when N=8 even if the wavelength range varied by 10.6 m10%, and the incident half angle varied by 30.

(93) Experimental results for graded index coating and patterned nano-structures have been presented for a variety of surface topologies, material systems, wavelengths and operating conditions (see the references in P. Lalanne and M. Hutley, The optical properties of artificial media structured at a subwavelength scale). Surface fabrication and replication techniques allow such surfaces to be fabricated for the visible. For example, Hutley and Gombert have used UV embossing into plastic using a nickel master to produce AR coated Fresnel lenses for overhead projectors (M. Hutley and A. Gombert, Moth-eyes: the tortuous path from a glint in the eye to a commercial reality, Photonics Science News 6, 35-39 (2000)). Anti-reflection nano-structures may be produced for the optical C- and L-bands by similar techniques.

(94) Referring again to FIG. 21, one can further vary the graded or nano-structured surface topology across the coverplate surface (varying pitch and/or shape, and the like), to thereby tailor the coating to match specific the wavelength at that section of the device. This is not possible using standard thin film deposition processes. In addition, a surface relief profile is more rugged than complex thin film dielectric stacks that can de-laminatea potential problem during operation at elevated temperatures (telecom systems typically run at 60 C. to simplify temperature stabilization), and during LCOS SLM device fabrication.

(95) Example Nanostructured Surfaces

(96) Broadly speaking, the aim of using a nano-structured anti-reflection coating is to make a more effective broadband coating for telecoms, and one that can be tailored so that the reflectivity is <0.01% at each wavelength location across the SLM in a system that uses a wavelength de-multiplexer. This will allow simplification of the hologram design and optimization by reducing the front surface reflections (so they may be neglected), particularly coupled with the fact that blue phase SLMs do not require an alignment layer, thereby simplifying calculation and measurement of the coverplate/electrode /liquid crystal interface reflectivity.

(97) For details of how to implement a wavelength-optimised coating reference may be made to A. Deinega, I. Valuev, B. Potapkin, and Y. Lozovik, Minimizing light reflection from dielectric textured surfaces, J. Opt. Soc. Am. A, Vol. 28, No. 5, pp. 770-776, May 2011. In this paper simulations of the reflectivity from a variety of nano-structured surfaces are modelled using finite difference time domain analysis (FDTD). This technique gives a numerical solution to Maxwell's equations for complex surfaces by probing the structure with a plane wave in the form of a short pulse of light. As the pulse has a certain wavelength spread, on analysis of the reflected field the reflectivity as a function of wavelength can be derived. They model pyramidal surface relief structures which have a depth of d, a width of 2 L and a period of A over a full range of wavelengths on a glass substrate (n=1.5). We are interested in the long wavelength limit (<) as this minimizes scattering losses as the light sees an effective graded index interface. For and > we would get scattering and diffraction losses as the light would interact with effectively a periodic macro-prism structure on the surface of the SLM. This regime is considered in the paper, and in particular Deinega et al analyse square pyramids with linear, cubic and quantic profiles where =2 L. The two key results from this paper are in FIGS. 4 and 8. FIG. 4 shows that for a given , as we increase d, the depth of the structure, the reflectivity reduces (ratio of d/ increasing). This reduction depends on the pyramid profile (five different types analysedlinear pyramid, cubic profile, quantic profile, cones, and an ideal integral profile); our target is <0.01% or 10.sup.4. FIG. 8 shows the reflectance map as a function of (/d) and (d/) for a square linear pyramidal structure.

(98) As mentioned we are interested in the long wavelength region where > in order to minimize scattering. There are two ways to use this graph. Firstly for a varying nano-structure depth, d, and secondly for a fixed nano-structure depth. Let us assume that we have two discrete wavelengths, .sub.1 and .sub.2. a. For a fixed value of d/, if .sub.2<.sub.1, then (/.sub.1)>(/.sub.2). Thus R.sub.1>R.sub.2. To reduce R.sub.1, we increase d/. As the graph tilts down as d/ increases, R.sub.1 decreases. Thus the pyramid depth for .sub.1 should be greater than for .sub.2 for the same reflectivity. b. For a fixed value of d, let us consider points on the surface map that have the same reflectivity. The aim would be to choose the correct value for .sub.1 and .sub.2 that give the same values of R.sub.1 and R.sub.2 for given values of .sub.1 and .sub.2. For example, if .sub.1>.sub.2, then (d/.sub.1)<(d/.sub.2), then we have two lines that run parallel to the (d/) axis. We choose them for the given wavelengths, .sub.1 and .sub.2 so that R.sub.1=R.sub.2. This uses knowledge of the reflectivity map as both axes depend on .

(99) As blue phase liquid crystal based SLMs do not need alignment layers, this simplifies the suppression of crosstalk from the coverplate/ITO/liquid crystal surface as it is easier to calculate the theoretical reflectivity of this surface and, optionally, incorporate this in the hologram design.

(100) Coverplate/Liquid Crystal Interface

(101) We now consider the reflectivity due to the coverplate/liquid crystal (comprising glass, electrode (for example ITOindium tin oxide), optional alignment layer, and liquid crystal layer(s)). This is more complicated because potentially four interfaces are present, and the alignment layer/liquid crystal layer reflectivity depends on the state of the liquid crystal. However, it is possible to tune out the coverplate/liquid crystal reflectivity by employing a hologram pattern which sends an equal amount of power into the +2 order 180 out-of-phase with the power due to the coverplate/liquid crystal reflectivity. As the liquid crystal layer is very thin, temperature changes have negligible effect on the path length, so the reflection-cancellation hologram needs only be determined once. This approach is possible but not so easily implemented when dealing the coverplate front surface reflectivity, which is 1 mm thick, as the same temperature change induces a larger optical path length change. To illustrate this, the optical path length change, opl, is given by opl=ndT, where n is the index, d is the material thickness, is the coefficient of thermal expansion, and T is the temperature change. Thus opl is proportional to T, and maintaining the 180 out-of-phase condition becomes progressively harder as d increases.

(102) FIG. 24 takes the system of FIG. 8 as an example (like elements are indicated by like reference numerals) and shows, in block diagram form, an SLM controller/driver 2400 configured to drive the LCOS SLM with such a reflection-compensation hologram. The input data to the controller/driver 2400 may comprise hologram data or higher level data such as beam switching/routing data (which may be converted into data defining a diffraction pattern for display on the SLM by, for example, a lookup table).

(103) The reflection-compensation, H.sub.R, hologram is computer generated, for example by any of a range of standard techniques (for example a Gerchberg-Saxton algorithm) and stored in non-volatile memory. This is then added to the target hologram H.sub.T to provide a combined hologram H.sub.C=H.sub.R+H.sub.T for display.

(104) One way to compensate for the coverplate/ITO/alignment layer/liquid crystal interface reflectivity is, if the fundamental blazed grating has a period of T, to add a weak blazed grating (a compensating grating) of period 2T. This generates a +1 order beam that travels in the same direction as the +2 diffracted beam of the fundamental blazed grating (which also includes the unwanted reflected light as shown in FIG. 20) with the same field value. If we move this weak blazed grating sideways through a distance of 2T the phase of the +1 order from the compensating grating will continuously change from 0 to 2 due to the Fourier shift theorem. In this way it is possible adjust the phase of the compensating grating such that the +1 order from this grating is approximately in anti-phase with the unwanted reflected light. However, this is not a good approach in practice as the compensating grating intended to suppress the reflection itself may introduce new noise orders that could lead to crosstalk at other positions. A more preferable approach, particularly if one knows in advance the reflectivity of the coverplate/ITO/liquid crystal layer and the front coverplate reflectivity is minimized, is to globally optimize a hologram (for example along the lines described in WO2012/123713).

(105) As previously mentioned, applying this cancellation technique to the front coverplate surface is more difficult due to the relative thickness of the coverplate. Temperature changes can cause the phase of the reflected beam to change with respect to the cancellation beam. However this may be compensated for by correcting the hologram for temperature, using a very thin coverplate, and/or by using a low thermal expansion coefficient glass.

(106) We have described, in embodiments, a phase-only LCOS device employing liquid crystal in a blue phase. It will be understood that the invention is not limited to the described embodiments and encompasses modifications apparent to those skilled in the art lying within the spirit and scope of the claims appended hereto.