GB2410339A

GB2410339A - Three lens arrays optical system, light source and projection display

Info

Publication number: GB2410339A
Application number: GB0401190A
Authority: GB
Inventors: Grant Bourhill; Bronje M Musgrave; Marina Khazova; Estelle K Tidey; Emma J Walton
Original assignee: Sharp Corp
Current assignee: Sharp Corp
Priority date: 2004-01-21
Filing date: 2004-01-21
Publication date: 2005-07-27
Also published as: GB0401190D0; JP2005208644A

Abstract

An optical system comprises a first array (33) of polarisation sensitive lenses and a second array (37) of polarisation sensitive lenses. The lenses of the first array act as converging lenses for one polarisation but have substantially no effect on another orthogonal polarisation. Conversely, the lenses of the second array (37) act as converging lenses for the other polarisation but have substantially no effect of the first polarisation. A third array (44) of lenses is located at or adjacent the focal plane (36) of the lens arrays (33, 37). The first and second lens arrays (33, 37) may be offset laterally from each other by half the lens pitch so that spots of S-polarised light are interspersed with spots of P-polarised light at a patterned retarder (38). The patterned retarder (38) changes at least one of the polarisations so that light exits the retarder with a substantially single uniform polarisation.

Description

24 1 0339 OPTICAL SYSTEM, LIGHT SOURCE AND PROJECTION DISPLAY The present

invention relates to an optical system. Such a system may be used, for example, in a polarisation conversion optical system (PCOS), for example for converting input unpolarised light to output light of a substantially single or uniform polarisation. Such a system may, for example, be used in a light source for a projection display, in a projection display or in a direct view display. The present invention thus also relates to a light source including such a system, to a projection display and direct view LC display device including such a light source and to a direct view display including such a system.

It is known that certain types of projection displays include a polarisation conversion optical system for receiving unpolarised light from a lamp and for converting this into uniformly polarised light. The illumination system of the projector illuminates a spatial light modulator (SLM), such as a liquid crystal device (LCD), so that, by converting the polarization of the illuminating light to the substantially uniform polarization, the efficiency or brightness of the projection display is improved. Figure 1 of the accompanying drawings illustrates a typical example of this type of display, for example as disclosed in US5,978,136.

The projection display shown in Figure 1 comprises a lamp 1, for example comprising a miniature arc lamp with a parabolic reflector, which emits a substantially collimated beam of unpolarised light with a non-uniform intensity distribution. Light from the lamp 1 is supplied to a polarization conversion optical system (PCOS) and homogeniser 2, comprising first and second microlens arrays 3 and 4, a polarising beam splitter (PBS) 5, an array of half wave phase plates 6, and a field lens 7. Output light is directed onto and illuminates a liquid crystal panel 8.

The first microlens array 3 forms an array of spatially distributed bright spots of unpolarised light at its focal plane. The second microlens array 4 is disposed at the focal plane of the first array 3 and the lenses of both arrays have the same focal lengths. With the field lens 7, the second array 4 acts as part of a beam homogenizing system for imaging a uniform beam of light at the plane of the liquid crystal panel 8. The aspect ratio of the microlenses of the first array 3 is substantially the same as the aspect ratio of the panel 8. The light beam from the lamp 1 is thus re-shaped such that the homogenised beam which illuminates the panel 8 has substantially the same aspect ratio as the panel.

For homogenizing systems of this type to work correctly, the focal lengths of the microlenses in each array must be equal and the second array must be placed at the focal plane of the first array (see Stupp and Brennesholtz "Projection Systems", publ. Wiley, 1999).

The polarising beam splitter 5 is disposed after the second array 4 and comprises an array having a pitch of elements which is approximately half the pitch of the microlenses of the array 4. Thus, focused light beams are incident on alternate elements or sections of the polarising beam splitter 5. Polarisation separating films such as 9 transmit light of one polarisation, such as the P polarisation, in the "forward" direction.

Light with the orthogonal polarisation, such as the S polarisation, is reflected to a reflective film such as 10 in an adjacent section of the beam splitter, where it is reflected into the forward direction. Thus, each unpolarized beam incident on the beam splitter 5 generates two spatially separated forward-propagating beams with orthogonal polarizations.

The element 6 comprises an array of individual phase plates or strips of retarder film arranged such that only beams of one polarisation are incident on the retarder film.

The retardation is chosen so that the incident polarised light emerges from the retarder with its polarisation changed to the orthogonal polarisation. Thus, all forward- propagating beams have substantially the same polarisation state. These beams are condensed by the field lens 7 to illuminate uniformly the panel 8. Jo

The f-number of the microlenses of the arrays 3 and 4 is chosen to match the characteristics of the lamp 1, which generally emits slightly diverging light. In order to reduce the length of the polarisation conversion optical system 2 without substantially affecting its lightcollecting efficiency, it is necessary to reduce both the focal length and the pitch of the microlenses in order to maintain the desired f- number.

The use of microlenses with a smaller pitch for any given lamp is advantageous because the illuminating beam at the panel 8 can be better homogenised, resulting in better illuminance uniformity at the panel. Improved homogenization is achieved because there are more lenses in the first array 3 to sample the non-uniform beam emitted from the lamp 1. Conversely, the use of microlenses of smaller pitch allows light emitted by a light source having a smaller form factor to be effectively homogenised, thus allowing miniaturised "projection engines" to be provided.

A consequence of reducing the pitch of the microlenses of the arrays 3 and 4 is that the pitch of the beam splitter array 5 also has to be reduced. The beam splitter is typically the most expensive element within a PCOS and reducing its pitch substantially increases the cost of manufacture because more polarisation-separation and mirror coatings are required. Also, when producing a smaller pitch element, tolerancing becomes more critical and light wastage increases. As a result of these constraints, it is not currently viable commercially to manufacture a polarization beam splitter array with a pitch of less than about 0.5 mm so that the total length of a PCOS employing such an array cannot be made less than about 7 mm.

The PCOS 2 shown in Figure 1 comprises four separate units, namely the two microlens arrays 3 and 4, the beam splitter array 5 with attached retarder strips and a field lens 7. In order to reduce light loss caused by Fresnel reflections out of the surfaces of these units, the interfaces between the units must be index-matched or anti reflection coatings must be provided. The former measure increases the system length and the latter measure increases the system cost.

US6621533 discloses a PCOS as illustrated in Figure 2 of the accompanying drawings. This system comprises microlens arrays 3 and 4 and a polarization rotating element 11, which functions similarly to the arrays 3 and 4 and the phase plate 6, respectively, shown in Figure 1. However, the PBS 5 is omitted and its function is performed by a polarization splitting element 12 comprising an array of optically anisotropic microprisms such as 13 in contact with an array of optically isotropic microprisms such as 14. Each microprism 13, 14 has a wedge-shaped crosssection.

This arrangement allows microlenses of the required f-number for collecting light efficiently to be provided with very small pitch and focal length since the pitch of the microlenses is no longer controlled by the necessity of using a glass polarising beam splitter array 5. Consequently, the system has a minimum length of approximately 4 mm. Perfectly collimated unpolarised input light at normal incidence on the polarization splitting element 12 encounters the inclined interfaces such as 15 between the anisotropic and isotropic prisms 13 and 14 and is angularly separated into two orthogonally polarised beams. The angle between the beams is determined by the angle of incidence of light on the interface 15 and the change in refractive index encountered by the light at the interface 15.

When the splitting element 12 is illuminated with divergent light as occurs in real optical systems, some of the light is incident on the vertical walls or interfaces such as 16 and is lost to the PCOS because it is refracted in the wrong direction. The light- efficiency of such a system is therefore reduced. A further disadvantage of this arrangement is that it is composed of relatively thick layers of polymer or liquid crystal materials, for example with thicknesses of approximately 60 to 100 micrometres. This thickness is generally required to achieve the desired splitting angle without compromising diffractive loss in the system, which loss happens if thickness and pitch are reduced. Such thick polymer or liquid crystal materials may degrade as a result of the exposure to the bright or intense light supplied from high pressure arc lamps which are typically used as light sources in projection engines. Thus, the working life of the PCOS may be unacceptably low. Also, for optimum light efficiency, the outer surfaces of three units of the PCOS must be provided with anti-reflection coatings and this increases the cost.

US4,446,305 discloses a light polarising sheet material as illustrated in Figure 3 of the accompanying drawings. This material or device comprises a functionally isotropic layer 20 having a relatively low refractive index nil, a highly birefringent polymeric layer 21 having ordinary and extraordinary refractive indices no, ne, and a functionally isotropic layer having a relatively high refractive index n2. The materials of the layers 21 to 22 are chosen such that no substantially matches no and n2 substantially matches ne Unpolarised light such as 23 is incident on the upper surface of the layer 20.

The interfaces between the birefringent layer 21 and the isotropic layers 20 and 22 form lens structures 24 and 25 such that the lens structure 24 is effective only for light of a first polarisation whereas the lens structure 25 is effective only for light of the orthogonal polarization. The P and S polarised components of the incoming light 23 are thus focused to different interlaced images.

The birefringent material of the layer 21 is a transparent solid plastic film and optical anisotropy results from molecular order induced by applying a mechanical stress to the material. The lens structures must be formed by modifying the surfaces of the layer 21 by embossing, grinding or polishing. Molecular rearrangement is caused by the application of heat or solvents necessary to soften a material before it can be embossed. Grinding and polishing introduce stress birefringence. Thus, all of these processes have the effect of altering the refractive index profile of anisotropic material, thus reducing the efficiency of the device.

In order to fabricate the device, it is also necessary to identify and select materials with the appropriate refractive indices and, in particular, with the appropriate matching conditions indicated hereinbefore. For broadband or white light, it is difficult to identify three materials which meet these index-matching conditions across a substantial range of wavelengths. This device thus has reduced broadband efficiency.

JP030005707 discloses an arrangement similar to that disclosed in US4 446 305 but with refractive lenses in the form of Fresnel lenses.

US5,671,034 discloses techniques for making polarization dependent refractive optical elements from birefringent materials.

JP03-294803 discloses a polarization dividing system, which uses a birefringent lens 17 to spatially divide the random polarization from a light source into two orthogonal linear polarizations, a system 18 to convert one of the polarisation states to be the same as the other polarisation state and a hologram lens 19 to combine the substantially uniformly-polarised light into a collimated beam. The system is illustrated in the form of an array in Figure 19.

In order to collimate the light focused by the birefringent lens 17, the hologram lens 19 (focal length fH) must be placed such that it is confocal with the birefringent lens (focal length fg). In other words, the birefringent lens is separated from the hologram lens by a distance equal to the sum of their focal lengths (fH + fB).

Owing to the fact that the output light is collimated, this known system cannot be used as an homogeniser in a polarisation conversion optical system. As the light is collimated, it is not possible to overlay the images of the individual polarisation sensitive microlenses and, therefore, it is not possible to form an homogenized beam of uniform illuminance at the panel plane. This system suffers further disadvantages because it is limited to the use of a hologram lens 19. The performance efficiency of a transmission holographic optical element is strongly wavelength-dependent. Hence, this system cannot be made efficient in white light. Additionally, the hologram lens 19 is not polarisation- sensitive and cannot act to change the polarisation of incident light.

According to a first aspect of the invention, there is provided an optical system comprlsmg: a first array of polarisation sensitive lenses for converging light of a first polarisation with a first convergence and for transmitting light of a second polarisation orthogonal to the first polarisation with a second convergence less than the first convergence; a second array of polarisation sensitive lenses for receiving light from the first array, for converging light of the second polarisation with a third convergence and for transmitting light of the first polarisation with a fourth convergence less that the third convergence, the first and second arrays converging light of the first and second polarizations to first and second images substantially at first and second image planes, respectively; and a third array of lenses disposed substantially no further from the first and second arrays than the one of the first and second image planes further from the first and second arrays.

The first and second polarizations may be linear polarizations.

At least one of the second and fourth convergences may be substantially zero.

Light of the second polarization may be substantially undeviated by the first array.

Light of the first polarization may be substantially undeviated by the second array.

The lenses of the first array may be substantially identical to each other.

The lenses of the second array may be substantially identical to each other.

The first and second image planes may be substantially coplanar. The third array may be substantially at the first and second image planes.

The first and second image planes may not be substantially coplanar and the third array may be substantially at one of or between the first and second image planes.

The first and second arrays may have pitches which are substantially equal to each other. The lenses of the second array may be offset laterally with respect to the lenses of the first array. The lateral offset may be substantially equal to half the pitch of each of the first and second arrays.

The lenses of the first and second arrays may comprise an optically anisotropic material. The lenses may be in contact with an optically isotropic material whose refractive index is substantially equal to one of the refractive indices of the anisotropic material. The lenses of the first and second arrays may be of the same material and the lenses of the first array may have an optic axis which is orthogonal to that of the lenses of the second array.

The lenses of the first and second arrays may be Fresnel lenses.

The lenses of the first and second arrays may be positive and negative doublets.

The first and second arrays may have substantially flat surfaces facing each other.

The lenses of the first and second arrays may be graded refractive index lenses.

The system may comprise a fourth array of polarisation changing elements for receiving light from the second array and for changing at least one of the first and second polarizations so that light from the fourth array is of substantially uniform polarization. The fourth array may be arranged to change one of the first and second polarizations substantially to the other of the first and second polarizations. The polarization changing elements may be polarization rotators. As an alternative, the polarization changing elements may be retarders.

Each of the lenses of the third array may be optically aligned with a respective lens of the first or second array. The focal length of each lens of the third array may be substantially equal to the focal length of the respective lens of the first or second array.

The third array may comprise the fourth array and at least some of the lenses of the third array may comprise the polarization changing elements. The at least some lenses may be optically anisotropic.

The system may comprise a field lens for converging output light from the system.

According to a second aspect of the invention, there is provided a light source comprising a system according to the first aspect of the invention and at least one light emitter.

According to a third aspect of the invention, there is provided a display comprising a source according to the second aspect of the invention.

The display may comprise a spatial light modulator arranged to be illuminated by the source. Each lens of the first array may have an aspect ratio substantially equal to that of the modulator.

According to a fourth aspect of the invention, there is provided a direct view display comprising a system according to the first aspect of the invention disposed between a backlight and a spatial light modulator.

Each lens of the third array may be optically aligned with a respective pixel of the modulator.

The modulator may be a liquid crystal device.

It is thus possible to provide a polarization conversion optical system of very small length. For example, lengths of the order of as little as 3 mm are possible. It is also possible to provide the lens arrays with very small pitches and this provides improved beam homogenization. Such systems may be used to provide efficient homogenization of very small form factor light sources, such as individual light emitting diodes (LEDs) or small arrays of LEDs, and this allows miniature projection engines to be provided. Relatively thin layers of polymers may be used and this provides systems of good endurance, even in the very intense light used in projection systems.

Polarisation conversion optical systems of relatively low complexity may be made without requiring expensive polarising beam-splitters and with relatively few anti- reflection coatings so that the cost of such systems may be relatively low.

Alternatively, such a polarization conversion system may be used to convert unpolarised light emitted by a narrow angle backlight in a direct view display into substantially polarised light. More of the light from the backlight may thus be used and this improves the efficiency of, for example, a direct view liquid crystal device. In this case, microlenses in the third array act as field lenses for microlenses in the first and second arrays, beneficially changing the convergence of light rays.

The invention will be further described, by way of example, with reference to the accompanying drawings, in which: Figure 1 is a cross- sectional diagram illustrating a known type of projection system including a known type of polarization conversion optical system; Figure 2 is a cross-sectional diagram illustrating another known type of polarization conversion optical system; Figure 3 is a cross-sectional diagram illustrating a known type of polarization sensitive optical system; Figure 4 is a cross-sectional diagram illustrating a polarization conversion optical system constituting an embodiment of the invention; Figure 5 is a cross-sectional diagram illustrating an example of a detail of the system of Figure 4; Figure 6 is a cross-sectional diagram illustrating a technique for making a system of the type shown in Figure 4; Figures 7a and 7b illustrate examples of details of the system illustrated in Figure 4; Figures 8a and 8b illustrate a technique for making the example illustrated in Figure 7a; Figure 9a to 9c illustrate techniques for making the example illustrated in Figure 7b; Figures lea to 1 to illustrate other examples of lens arrays which may be used in a system of the type shown in Figure 4; Figures 12a and 12b illustrate passage through a Fresnel lens for light incident normally and at a small angle; Figure 13 illustrates the effect of chromatic aberration in a lens; Figures 14a to 14d illustrate other examples of lens arrays which may be used in a system of the type shown in Figure 4; Figures 15a and 15b illustrate techniques for making examples of lens arrays for use in a system of the type shown in Figure 4; Figure 16 is a cross-sectional diagram of a polarisation conversion optical system constituting another embodiment of the invention; Figure 17 illustrates diagrammatically a GRIN lens which may be used in the system of Figure 16; Figures 18a and 18b illustrate the effect of chromatic dispersion and beam divergence of a microlens illuminated with divergent light; Figure 19 is a cross- sectional diagram showing a known polarising system; Figures 20a and 20b illustrate the focusing effects of lenses acting on divergent incident light; Figure 21 is a cross-sectional diagram of a projection engine constituting an embodiment of the invention; and Figure 22 is a cross- sectional diagram of a direct view display constituting an embodiment of the invention.

Like reference numerals refer to like parts throughout the drawings.

The polarization conversion optical system shown in Figure 4 may be used, for example, in the illumination system of a projection system for projecting images displayed on a spatial light modulator, such as a liquid crystal device. The system receives unpolarised light illustrated at 30 from a lamp with orthogonal P and S polarizations being illustrated at 31 and 32, respectively. The unpolarised light 30 is incident on a first array 33 of substantially identical polarization sensitive microlenses made of a birefringent material with the optic axes of the lenses being parallel to the direction of the S polarization 32. The material of the lenses of the array 33 has positive birefringence, i.e. the extraordinary refractive index ne is greater than the ordinary refractive index no. In the embodiments described herein, all anisotropic materials have positive birefringence. However, anisotropic materials of negative birefringence may alternatively be used and the changes to accommodate such material are apparent to a person of ordinary technical skill.

The lenses of the array 33 are set in isotropic material as illustrated at 34 and 35 with the refractive index n of this material substantially matching the ordinary refractive index no of the lenses. In one example, the lenses of the array 33 are made of a reactive mesogen known as RMM34 available from Merck UK with no = 1.53 and ne = 1.69 for green light and the material 34, 35 comprises a Norland optical adhesive known as NOA68 having n = 1.54 for green light. In another example, the lenses of the array 33 comprise a liquid crystal material known as MLC6697 available from Merck UK with no = 1.50 and ne = 1.66 for green light and the material 34, 35 comprises an adhesive known as SK9 available from Summers Optical and with n = 1.51 for green light. The lenses of the array 33 are of piano-convex type and are "optically positive".

Alternatively, an array of positive refractive lenses may be provided by index-matching the isotropic material to the extraordinary refractive index ne of the anisotropic material and forming an array of negative lenses in the anisotropic material.

The S-polarised light 32 is focused at a series of spots substantially at the focal plane 36 of the first array 33 by the lenses of the array 33. However, the P-polarised light 31 passes through the lenses of the array 33 with substantially no convergence and with substantially no deviation. The second array 37 has substantially the same pitch as the first array 33 and is laterally offset with respect to the first array 33 by substantially half the pitch of the lenses of each array. The lenses of the second array 37 are made of the same material as the lenses of the first array 33 but the optic axes of the lenses of the second array are substantially orthogonal to the optic axes of the lenses of the first array.

The lenses of the second array 37 pass the S-polarised light substantially without converging or deviating the light. The lenses of the second array 37 focus the P- polarised light 31 to the focal plane of the microlens array 37, which is substantially coincident or coplanar with the focal plane 36 of the first array 33.

The focused spots of light formed by the microlens arrays 33 and 37 at the plane 36 form a substantially regular pattern at half the pitch of each of the arrays 33 and 37.

A patterned retarder 38 is disposed adjacent the plane 36 and comprises elements having the same pitch as and disposed at the locations of the focused spots of light formed by the microlens arrays 33 and 37. The elements of the patterned retarder 38 arranged to act on the S and/or P polarizations such that light leaving the patterned retarder 38 has substantially the same uniform polarization. For example, as illustrated in Figure 5, the retarder 38 comprises elements such as 39 acting as half wave plates with their optic axes oriented so that the incident Spolarised light 40 is converted to P polarised light 41. The elements 39 are interleaved with elements 42 which pass the incident P-polarised light 43 without substantially affecting the polarization. Thus, substantially all of the incident light exits the patterned retarder 38 with the P polarisation. t

Light from the elements 39, 42 of the patterned retarder 38 passes through respective lenses of an isotropic microlens array 44. The array 44 forms part of the beam homogenization system and the light exiting the array is, for example, converged by a field lens of the type illustrated at 7 in Figure 1 so as to produce a uniform homogenised beam at an input plane of a spatial light modulator, such as of the type illustrated at 8 in Figure 1. The pitch of the lenses of the array 44 is substantially equal to half the pitch of the lenses of each of the arrays 33 and 37. Each microlens of the third array 44 collects polarised focused light from a single corresponding lens of one of the arrays 33 and 37 and has the same focal length as the corresponding lens. Thus, the lenses of the array 44 are not necessarily identical to each other.

The patterned retarder 38 is shown as being disposed before the microlens array 44 but may alternatively be disposed after it. The microlens array 44 is located substantially at the common focal plane 36 of the microlens arrays 33 and 37.

The pitch of the lenses of each of the arrays 33 and 37 may be selected in accordance with the requirements of the system. For example, pitches of between 80 micrometres and 500 micrometres may be provided. Also, the F-number of the lenses may be chosen so as to collect light from a light source used with the system. By minimising the pitch of the lenses of the arrays 33 and 37 while preserving the F- number of the lenses, a system of small length may be provided which has high efficiency and provides a high degree of homogenization.

The lenses of the arrays 33, 37 and 44 may be spherical or parabolic. The aspect ratio of the lenses may be chosen so as substantially to match the aspect ratio of a spatial light modulator forming part of a projection system including the polarization conversion optical system. A very compact "projection engine" may be provided with high efficiency of light utilization and with high uniformity of illumination throughout the image.

For use in a direct view liquid crystal display, the lenses of the arrays 33, 37 do not require their aspect ratios to be matched to that of the panel. The lensless of the arrays 33, 37, 44 may be two-dimensional or they may be cylindrical (lenticular array).

The polarisation sensitive microlenses may be formed from various different types of optically anisotropic materials. Such materials include low molecular mass liquid crystal, photopolymerisable liquid crystal(reactive mesogen), polymer liquid crystal and anisotropic crystalline solid. Various examples are described herein.

Figure 6 illustrates a technique which may be used during manufacture of the system to ensure that the individual elements or regions such as 39 and 42 are correctly aligned. The elements 33 to 37 are formed or assembled as illustrated in Figure 6 and a layer 50 of reactive mesogen having a thickness corresponding to a half wave plate is formed adjacent or at the common focal plane 36 of the microlens arrays 33 and 37.

The array 44 of isotropic microlenses 51 is shown as being formed on the layer 50 but may in practice be formed after the layer 50 has been processed.

S-polarised ultraviolet (W) light 52 is incident on the microlenses of the array 33 and is focused to a series of spots on the layer 50, which is planar- aligned by being disposed on a suitable alignment layer 53. The angular range of incidence of the ultraviolet light 52 is selected so as to create a spot with a diameter substantially corresponding to half the pitch of the lenses of the array 33.

The intensity of the incident ultraviolet light 52 is such that the reactive mesogenic material of the layer 50 is cured where the ultraviolet light is incident on it.

The direction of incidence of the ultraviolet light 52 may be altered so as to create a series of stripes instead of spots. As another alternative, a substrate carrying the layer 50 may be translated laterally to achieve substantially the same effect.

Following exposure to the ultraviolet light, the uncured material of the layer 50 is removed to leave the regions such as 39 which act as half wave plates to visible light and regions such as 42 which do not substantially affect the polarization of incident light.

A two layer (Pancharatnam) patterned retarder may be used as the retarder 38 and may be made by a similar technique. For example, after forming the first layer as described hereinbefore with reference to Figure 6, a new alignment layer may be formed on the existing layer and another thin layer of reactive mesogen may be formed on the new alignment layer. The second layer may then be cured in the same way by exposure to the polarised ultraviolet light 52. Uncured material may then be removed to provide a two layer patterned retarder having more efficient polarization conversion performance for white light than a single layer retarder.

The patterned retarder 38 may be made in accordance with an alternative technique from a photoisomerisable reactive mesogen as described, for example, in US2002/0187282. According to this technique, the reactive mesogen layer 50 is exposed to the polaTised ultraviolet light 52 as described hereinbefore but, in this case, the exposure causes photoisomerisation to occur. The photoisomerisation causes the optical retardation of the reactive mesogen in the exposed areas to become zero whereas the unexposed areas retain the original retardation. Subsequently, the whole layer 50 is polymerized by heating or photocuring. Thus, the regions exposed during the photoisomerisation exposure do not substantially affect the polarization of incident visible light in the final system whereas the initially unexposed regions act as half wave plates to incident visible light.

It is also possible to form the patterned retarder 38 separately from the rest of the polarization conversion optical system. For example, a birefringent liquid crystal material may be deposited on a substrate carrying a patterned alignment layer. Such a patterned alignment layer may be provided by: a multi-rubbing technique, for example as disclosed in EP0887667; use of a microstructured grating, for example as disclosed in GB2384318; or use of a photo-alignment technique, for example as disclosed in "Photofabrication of micro-patterned polarising elements for stereoscopic displays", Matsunaga et al, 14 pp 1477-1480, 2002.

The reactive mesogen may be spun in solution onto the alignment layer. As the solvent evaporates, it leaves behind a layer of nematic liquid crystalline material. The optic axis of the liquid crystal aligns with the alignment direction of the underlying alignment layer so as to produce the retarder layer. The reactive mesogen is cross linked by exposure to ultraviolet light. In order to improve white light performance of the patterned retarder, a second uniform retarder layer may be provided, for example between the polarisation sensitive microlenses of the arrays 33 and 37 and a spatial light modulator of a projection system.

The performance of the retarder may be optimised to compensate for wavelength-dependent performance elsewhere in the polarisation conversion system.

For example, refractive index dispersion in the microlenses leads to chromatic aberration, which may result in better polarisation conversion at blue wavelengths than red, for example. In this example, the performance of the retarder may be designed to be more efficient for red light, in order to improve the white light or achromatic performance of the system.

Another example of the patterned retarder 38 comprises alternate stripes of planar-aligned and homeotropically-aligned reactive mesogen. The planar-aligned stripes act as half wave retarders for incident polarised light. The homeotropically aligned stripes are optically isotropic and do not substantially affect the polarisation of the incident light. Such a patterned retarder may be made using photo-alignment techniques.

As a further alternative, the patterned retarder 38 may comprise an array of strips of retarder film of width substantially equal to half the pitch of the polarisation sensitive microlenses of the arrays 33 and 37.

In a further example, the patterned retarder 38 may comprise a cell containing regions of twisted nematic liquid crystal or reactive mesogen operating in the Mauguin regime alternating with regions of isotropic material, such as an isotropic polymer, or a liquid crystal or reactive mesogen with homeotropic alignment. The regions containing twisted nematic material rotate the plane of polarisation of incident light whereas the other regions do not substantially affect the polarisation state of the incident light.

Figure 7a and 7b illustrate arrangements in which the microlens array 44 also performs the function of the patterned retarder 38, which may therefore be omitted so as to simplify the system. In the arrangement shown in Figure 7a, microlenses such as 55 made of optically isotropic material alternate with microlenses such as 56 made of twisted nematic liquid crystal material operating in the Mauguin regime. When these lenses are immersed in an optically isotropic dielectric medium that is index matched to the ordinary refractive index nO of the liquid crystal material of the lenses such as 56, the refractive index of the isotropic microlenses 55 is substantially equal to the extraordinary refractive index ne of the liquid crystal material of the lenses such as 56.

In this arrangement, the twisted nematic liquid crystal of the lenses 56 rotates the plane of polarization of incident light from the Spolarisation to the P-polarisation.

The isotropic microlenses such as 55 do not substantially change the Ppolarisation of the incident light. Thus, light of substantially uniform P-polarisation emerges from the microlens array 44. Such an arrangement has advantages of reduced size and complexity because the functions of polarization conversion and homogenization are combined in the single element constituted by the microlens array 44. Further, the number of elements which have to be correctly aligned is reduced and this results in easier manufacture of the system.

The arrangement shown in Figure 7b differs from that shown in Figure 7a in that the isotropic microlenses 55 are replaced by liquid crystal microlenses 57 in a planar or homeotropic alignment. Thus, the microlenses 57 do not substantially affect the polarization of incident light. However, the lenses 56 and 57 may be made of the same nematic liquid crystal material so that the dispersions of the lenses are substantially matched.

Figures 8a and 8b illustrate an example of a technique for making the microlens array 44 shown in Figure 7a. A substrate 60 comprising an isotropic medium of refractive index nO is provided and the isotropic microlenses 55 are formed as an array in an isotropic medium of refractive index ne with gaps therebetween. A uniform alignment layer 61 is formed on the microlenses 55 and the upper surface of the substrate 60.

A counter substrate 62 of isotropic medium is provided with its lower surfaces formed in the shape of the whole microlens array, without any gaps. The lower surface of the counter substrate 62 is provided with a uniform alignment layer 63 oriented such that the alignment directions of the alignment layers 61 and 63 are orthogonal. The substrate 60 and the counter substrate 62 are brought together to form voids for the microlenses 56. These voids are filled with a nematic liquid crystal material, which is aligned by the alignment layers 61 and 63 to form twisted nematic liquid crystal lenses such as 56.

Figure 9a and 9b illustrate a technique for making the microlens array illustrated in Figure 7b. A patterned alignment layer 65 is formed on the optically isotropic substrate 60 as an array of regions such as 66 and 67 of orthogonal alignment directions.

The alignment layer 65 may be formed by, for example, multi-rubbing techniques, use of a grating microstructure or photopatterning as described hereinbefore.

The optically isotropic counter substrate 62 is similar to that shown in Figure 8a and has formed on its lower surface an unpatterned or uniform alignment layer 63 whose alignment direction is parallel to that of the regions such as 66 of the patterned alignment layer 65.

The substrates 60 and 62 are brought together so as to define an array of voids for forming the microlenses of the array 44. The voids are filled with nematic liquid crystal material to form the microlenses such as 57 of planar alignment and the microlenses 56 of twisted nematic alignment.

Figure 9c illustrates an alternative arrangement in which the curved surface for the microlenses is formed, such as by stamping, in the substrate 60 and provided with the unpatterned alignment layer 63 whereas the counter substrate 62 has a plane lower surface provided with the patterned alignment layer 65.

In order for the microlens arrays 33 and 37 to form the light spots in substantially the same common focal plane 36, the radii of curvature of the microlenses of the second array 37 may be made smaller than those of the lenses of the first array 33, as illustrated in Figure 4, so as to reduce the focal length of the microlenses of the second array 37. Alternatively, as illustrated in Figure lea, the microlenses of both i arrays 33 and 37 may have the same radii of curvature and a birefringent medium 70 may be disposed between the microlenses of the second array 37 and the patterned retarder 38. The birefringent medium 70 has ordinary and extraordinary refractive indices nO2 and net, respectively.

As illustrated at 71, S-polarised light "sees" the extraordinary refractive index nC2 of the medium 70 whereas, as illustrated at 72, Ppolarised light "sees" the lower ordinary refractive index nO2 of the medium 70. This has the effect of "shortening" the focal lengths of the microlenses of the array 37.

Figure 1 Ob illustrates an alternative arrangement in which the microlenses of the arrays 33 and 37 are made of birefringent materials having different birefringences. The microlenses of the array 33 are made of a birefringent material having an ordinary refractive index not which matches that of the isotropic medium 34. The microlenses of the array 37 are made of a birefringent material having an ordinary refractive index nO2 which matches that of the isotropic medium 35. The effects of this on S- and P- polarised light are illustrated at 71 and 72, respectively. Suitable selection of the materials and the refractive indices results in the microlenses of the second array 37 having a shorter focal length than the mircrolenses of the first array 33. The differences between the focal lengths determine the distance between the microlens arrays 33 and 37 in the direction of light propagation through the system.

In practice, it is not essential for the focal planes of the microlens arrays 33 and 37 to coincide provided these focal planes are sufficiently close to each other. This may be achieved by using substantially identical microlenses in the arrays 33 and 37 provided the arrays are sufficiently close to each other in the direction of light propagation through the system. For example, as illustrated in Figure 11 a, the microlenses of the array 37 may be reversed such that the plane sides of the microlenses of the arrays 33 and 37 face each other and can be separated by a relatively thin layer 75 of isotropic material, such as adhesive or a thin substrate. A larger separation between identical polarisation sensitive microlenses may be advantageous in systems in which the polarization conversion efficiency is required to be substantially constant with wavelength. r

Figure I lb illustrates another example in which the microlenses of the arrays 33 and 37 are Fresnel lenses. The thickness of a Fresnel lens of given diameter and focal length depends upon the number of sections but is less than that of an optically equivalent conventional lens. The Fresnel lenses shown in Figure 1 lb comprise two sections and are approximately half the thickness of the lenses shown in Figure 11 a.

Figure tic illustrates a further example in which the lens arrays 33 and 37 comprise Fresnel lenses but with the array 37 facing in the opposite direction so that the plane surfaces ofthe arrays 33 and 37 face each other and are adjacent.

When light is incident on a Fresnel lens parallel to its optical axis as illustrated in Figure 12a, substantially all of the light passes through the lens. However, when light is incident on such a lens at an angle to its optical axis as illustrated in Figure 12b, some of the light, such as that incident on the flat surfaces such as 76, is not focused by the lens so that the efficiency of the system is decreased. However, for small angles of incidence, such losses are relatively small.

Fresnel lenses are more susceptible to diffraction because of the larger number of divisions reducing the pitch. Also, their ridged shape may affect the alignment of liquid crystal material during manufacture as described hereinbefore. For a relatively small number of divisions, such as one or two, there is a substantial reduction in lens thickness without the smallest width y of a division becoming less than acceptable limits, such as 10 to 20 micrometres, to cause significant diffraction or to cause liquid crystal molecules to align with the shape of the lens. For example, for a lens of 200 micrometre diameter, y is substantially equal to 28 micrometres for two divisions and 17 micrometres for three divisions. The lens shown in Figure 12a and 12b has two divisions and its thickness is half that of a conventional equivalent lens.

A possible problem with single element lenses is that of chromatic aberration resulting from dispersion of the refractive index and hence the focal length of a lens varying with wavelength of light. This causes the minimum size of a focused spot of light on the patterned retarder to increase as illustrated in Figure 13. If the diameter of f the spot is greater than half the pitch of the lens and thus greater than the size of a pattered retarder element, some light will be "lost" from the system because its polarisation will not be changed correctly.

Figure 18a illustrates that the position and minimum diameter of a spot of white light converged by a lens is determined by the extreme red and blue rays from the edges of the lens. For simplicity, figure 18b shows only the red and blue rays that define the minimum diameter of the focused spot of white light. The diameter of the beam increases rapidly just after the plane containing the minimum focused spot or circle of least contusion. This feature should be taken into account when designing an efficient polarisation conversion system. Hence, in cases where the focal planes of the arrays 33, 37 do not lie in the same plane, it is advantageous to place an array of polarisation changing elements closer to the focal plane of the first microlens array 33 than to the focal plane of the second microlens array 37. Furthermore, the preferred position for the array of polarisation changing elements may not lie in the plane of the circle of least confusion, where the red and blue focused spots overlap most closely. While such placement would give the highest average polarisation conversion efficiency, the blue light polarisation conversion efficiency may be significantly higher than the red light polarisation conversion efficiency, which may be undesirable.

In order to reduce the effects of chromatic aberration, either or both of the microlens arrays 33 and 37 may be replaced by pairs of negative and positive microlenses such that each microlens of each "compound" array comprises a doublet.

In each of Figure 14a to Figure 14d, the pair of arrays equivalent to the array 33 are indicated at 33a and 33b whereas the pair of arrays equivalent to the array 37 are indicated at 37a and 37b. Arrays of positive lenses have the suffix "a" whereas arrays of negative lenses have the suffix "b".

If the microlens array 37 is imperfectly index-matched to the surrounding optically isotropic material 35, then array 37 can act as an array of very weak negative lenses for the light that has been converged by microlens array 33. This can reduce the i effect of chromatic aberrations in the microlens array 33, by moving the focal planes of the red and blue light closer together.

A similar effect may be provided by the birefringent medium 70 shown in Figure 10a. For example, its dispersion properties may be selected so as to minimise chromatic aberration and the distance between the microlens arrays 33 and 37 may be varied so that their focal planes substantially coincide.

Figure 15a illustrates a technique for making the microlens arrays 33 and 37. A polymer layer 80 is formed on a substrate 81 and the upper surface of the polymer 80 is shaped to define the microlens profiles. For example, the polymer 80 may comprise an ultraviolet curable optical adhesive, such as Norland NOA61 or Summers Optical SK9, which has the lens profile formed in it by means of a lens-shaped mold and is then cured. An alignment layer 82, such as rubbed polyimide, a grating microstructure or a photoalignment layer, is formed on the upper surface of the polymer 80.

A counter substrate 83 has formed thereon a similar alignment layer 84. The alignment layers 82 and 84 have uniform alignment directions and are oriented such that the alignment direction of the alignment layer 84 is antiparallel to that of the alignment layer 82.

The substrate 81 and the counter substrate 83 are brought together but spaced apart by glue spacers 85 to define a cell gap which is filled with a liquid crystal material, such as MLC6647, or a reactive mesogenic material, such as RMM34. 1h the case of a reactive mesogenic material, the material is then cured by ultraviolet irradiation. For relatively thin lens arrays, it may only be necessary to provide a single alignment layer on one side of the reactive mesogen or liquid crystal material 86. For example, as illustrated in Figure 15b, only the counter substrate 83 has an alignment layer 84 and this is sufficient to ensure that the material 86, in this case a reactive mesogen, is correctly aligned.

When the material 86 has been cured, the counter substrate 83 with the alignment layer 84 may be removed. The thickness of the system may therefore be reduced and, in the case of a rubbed polyimide alignment layer 84, the absence of the alignment layer from the finished system avoids any potential problems with limited endurance of such alignment layers in projection systems.

Figure 16 illustrates a polarisation conversion optical system which differs from that shown in Figure 4 in that the microlens arrays 33 and 37 comprise arrays of graded refractive index (GRIN) microlenses. Such lenses may be manufactured from a polymer network liquid crystal or from a reactive mesogen and examples of suitable methods are disclosed in US5, 671,034. Light of one polarisation state is focused by each GRIN lens of one of the arrays because it encounters a radial variation in refractive index from n2 at the centre of the lens to nil < n2 at the edge of the lens. The orthogonal polarisation experiences a substantially isotropic refractive index profile and is not substantially deviated as it propagates through the polarization sensitive GRIN lens layer.

In one example of such GRIN microlens arrays, the refractive index profile (nL) as a function of lens radius r is given by: nL = n2 (l-Ar2) where A is a constant defined hereinafter and n2 is the maximum refractive index of the lens. For a given optically anisotropic medium with refractive indices no and ne, the maximum lens refracted index n2 cannot be greater than net The constant A is related to the lens thickness I, the lens focal length f and n2 as follows: A= 1 2tfn2 The thickness of a GRIN lens layer is determined by the required F-number or focal length of the lens and by the material and processing parameters. For example, to the achieve a lens diameter of 200 micrometres and a focal length in air of 1260 micrometres (F/6.3), where the lens profile varies from n2 = 1.69 at r = 0 to nil = 1.53 at r = 100 micrometres, the required thickness of the anisotropic material would be 24.8 micrometres.

A given F-number can be achieved for lenses of different thicknesses using the same optically anisotropic material by varying the manufacturing process parameters so as to change the refractive index profile and the values of the maximum and minimum refractive indices nil and n2.

As an alternative, it is possible to make the GRIN lenses from an inorganic birefringent substance such as lithium niobate, which displays a definable variation in birefringence as a function of ion concentration. An arrangement of this type is illustrated in Figure 17. The use of inorganic substances may improve the endurance of the system.

An advantage of GRIN lenses is that there is no curvature across the lens so that a polarisation sensitive GRIN lens aligned so as to focus Spolarised light will not act like a lens at all on P-polarised light, even though the lens material may not be index matched to its surroundings. To the P-polarised light, the GRIN lens appears to be an isotropic plane parallel-walled sheet of dielectric material and its effect on P-polarised light is therefore very slight.

It is also possible to use non-polarisation sensitive GRIN lenses in place of the conventional lenses in the array 44. Use of such lenses allows the length of the system to be reduced.

The polarization conversion system may be used to convert the unpolarised light emitted by a directional backlight in a direct view display into substantially polarised light. As the second polarization is re-used, this substantially improves the efficiency of the direct view liquid crystal device.

In the polarization conversion system for a direct view display, the microlenses in the fourth array act as field lenses for the microlenses in the polarisation-sensitive arrays, beneficially changing the convergence of the rays. Figure 20a illustrates the r focusing effect of a lens acting on divergent light. Figure 20b illustrates the effect on divergent polarised light of an optically anisotropic lens and an additional lens placed at the focal plane of the anisotropic lens. The additional lens has the same focal length as the anisotropic lens. The second lens acts as a field lens for the first lens, bending the S rays towards the optic axis of the system and reducing the divergence of the light.

Figure 20b shows two birefringent microlens arrays 33 and 37 and an isotropic microlens array 44, placed at the focal plane 36 of the anisotropic arrays. When an array of polarisation rotating elements is placed adjacent and aligned with the isotropic microlens array 44, the system converts the unpolarised light from the backlight into substantially polarised light and acts as a polarisation conversion system for a direct view display. In a preferred embodiment, the individual lenses of the isotropic array 44 are aligned with the individual corresponding pixel apertures of a direct view liquid crystal display. The isotropic array 44 reduces the divergence of the light, thus allowing the liquid crystal layer of the panel to be placed at a further distance from the polarisation conversion elements without loss of light at the inter-pixel gaps.

Figure 21 illustrates a projection engine comprising a lamp 1, for example of the same general type as illustrated in Figure 1, and a polarisation conversion optical system and beam homogeniser 2, which may for example be of any of the types described hereinbefore and shown in Figures 4 onwards. In particular, the sub-system 2 comprises birefringent microlens arrays 33 and 37, a patterned retarder 38 and isotropic microlenses 44. In addition, a field lens 7 is provided.

The lamp 1 and the sub-system 2 illuminate a liquid crystal panel 8 with substantially homogeneous illumination intensity across the panel and with substantially uniform single polarisation matched to the input polarisation requirements of the panel 8. The panel 8 modulates the light with image data and the modulation light from the panel 8 is supplied to a projection lens 90, which projects the or each displayed image onto a projection screen (not shown).

Figure 22 illustrates a direct view display which comprises a polarisation conversion optical system 2 as described hereinbefore and as shown in Figures 4 r 1 onwards. The system 2 is disposed between a narrow angle backlight 100 and a liquid crystal panel 8 having a layer of pixels such as 95. A polariser 101 is provided on the output side of the panel 8. The system 2 converts light from the backlight 100 to a substantially single uniform polarisation matched to the input requirements of the panel 8.

Claims

CLAIMS: 1. An optical system comprising: a first array of polarisation

sensitive lenses for converging light of a first polarisation with a first convergence and for transmitting light of a second polarisation orthogonal to the first polarisation with a second convergence less than the first convergence; a second array of polarisation sensitive lenses for receiving light from the first array, for converging light of the second polarisation with a third convergence and for transmitting light of the first polarisation with a fourth convergence less than the third convergence, the first and second arrays converging light of the first and second polarizations to first and second images substantially at first and second image planes, respectively; and a third array of lenses disposed substantially no Farther from the first and second arrays than the one of the first and second image planes farther from the first and second arrays.
2. A system as claimed in claim 1, in which the first and second polarizations are linear polarizations.
3. A system as claimed in claims 1 or 2, in which at least one of the second and fourth convergences is substantially zero.
4. A system as claimed in any one of the preceding claims, in which light of the second polarisation is substantially undeviated by the first array.
5. A system as claimed in any one of the preceding claims, in which light of the first polarisation is substantially undeviated by the second array.
6. A system as claimed in any one of the preceding claims, in which the lenses of the first array are substantially identical to each other.
7. A system as claimed in any one of the preceding claims, in which the lenses of the second array are substantially identical to each other.
8. A system as claimed in any one of the preceding claims, in which the first and second image planes are substantially coplanar.
9. A system as claimed in claim 8, in which the third array is substantially at the first and second image planes.
10. A system as claimed in any one of claims 1 to 7, in which the first and second image planes are not substantially coplanar and the third array is substantially at one of or between the first and second image planes.
11. A system as claimed in any one of the preceding claims, in which the first and second arrays have pitches which are substantially equal to each other.
12. A system as claimed in claim 11, in which the lenses of the second array are offset laterally with respect to the lenses of the first array.
13. A system as claimed in claim 12, in which the lateral offset is substantially equal to half the pitch of each of the first and second arrays.
14. A system as claimed in any one of the preceding claims, in which the lenses of the first and second arrays comprise an optically anisotropic material.
15. A system as claimed in claim 14, in which the lenses are in contact with an optically isotropic material whose refractive index is substantially equal to one of the refractive indices of the anisotropic material.
16. A system as claimed in claim 14 or 15, in which the lenses of the first and second arrays are of the same material and the lenses of the first array have an optic axis which is orthogonal to that of the lenses of the second array.
17. A system as claimed in any one of the preceding claims, in which the lenses of the first and second arrays are Fresnel lenses.
18. A system as claimed in any one of the preceding claims, in which the lenses of the first and second arrays are positive and negative doublets.
19. A system as claimed in any one of the preceding claims, in which the first and second arrays have substantially flat surfaces facing each other.
20. A system as claimed in any one of the preceding claims, in which the lenses of the first and second arrays are graded refractive index lenses.
21. A system as claimed in any one of the preceding claims, comprising a fourth array of polarization changing elements for receiving light from the second array and for changing at least one of the first and second polarizations so that light from the fourth array is of substantially uniform polarization.
22. A system as claimed in claim 21, in which the fourth array is arranged to change one of the first and second polarizations substantially to the other of the first and second polarizations.
23. A system as claimed in claims 21 or 22, in which the polarization changing elements are polarization rotators.
24. A system as claimed in claim 21 or 22, in which the polarization changing elements are retarders.
25. A system as claimed in any one of the preceding claims, in which the third array has a pitch which is substantially equal to half the pitch of the first and second arrays.
26. A system as claimed in any one of the preceding claims, in which each of the lenses of the third array is optically aligned with a respective lens of the first or second array.
27. A system as claimed in claim 26, in which the focal length of each lens of the third array is substantially equal to the focal length of the respective lens of the first or second array.
28. A system as claimed in any one of claims 21 to 24 or in any one of claims 25 to 27 when dependent on any one of claims 21 to 24, in which the third array comprises the fourth array and at least some of the lenses of the third array comprise the polarization changing elements.
29. A system as claimed in claim 28, in which the at least some lenses are optically anisotropic.
30. A system as claimed in any one of the preceding claims, comprising a field lens for converging output light from the system.
31. A light source comprising a system as claimed in any one of the preceding claims and at least one light emitter.
32. A projection display comprising a source as claimed in claim 31.
33. A display as claimed in claim 32, comprising a spatial light modulator arranged to be illuminated by the source.
34. A display as claimed in claim 33, in which each lens of the first array has an aspect ratio substantially equal to that of the modulator.
35. A direct view display comprising a system as claimed in any one of claims 1 to 29 disposed between a backlight and a spatial light modulator.
36. A display as claimed in claim 35, in which each lens of the third array is optically aligned with a respective pixel of the modulator.
37. A display as claimed in any one of claims 33 to 36, in which the modulator is a liquid crystal device.