CN1163000A - Display panel with electrically controlled waveguide routing - Google Patents

Abstract

A flat panel display is based on a new switching technology for routing laser light among a set of optical waveguides and coupling that light toward the viewer. The switching technology is based on poled electro-optical structures. The display technology is versatile enough to cover application areas spanning the range from miniature high resolution computer displays to large screen displays for high definition television formats. The invention combines the high brightness and power efficiency inherent in visible semiconductor diode laser sources with a new waveguide electro-optical switching technology to form a dense two-dimensional addressable array of high brightness light emissive pixels.

Description

Display panel with electrically controlled waveguide routing

The present invention relates to devices (particularly some optical arrangements) for controlling the delivery of energy (particularly light beams) using electric field control. Further, the present invention relates to polarized structures (including periodically polarized structures) and electrodes that allow controlled transmission of light energy by providing a controlled electric field between the electrodes. The invention also relates to an entirely new class of flat panel optical displays.

Figure 1 depicts a prior art electro-optic switchable grating. In this structure, periodically shaped electrodes are used as the elements that can form the grating. The structure of the electrode without the underlying material will be described below. An input light beam 12 is coupled to an electro-optically active material 2 comprising an optically controllable permanent grating 6 . When the voltage source 10 coupled to the grating electrodes is switched off, the input beam continues to propagate through the material to form an output beam 16 . When the grid electrode voltage source is turned off, the input beam continues to propagate through the material to form the output beam 16 . When the grating control voltage source is switched on, an index modulated grating is created in the material and a portion of the input beam is coupled to a reflected output beam 14 . The material has an electro-optically active polarized region 4 (with a single domain) with the same polarity throughout the poled structure. A first electrode 6 and a second electrode 7 are formed into an interdigitated structure on a common surface 18 of the substrate. When a voltage is applied between the electrodes, the vertical component of the electric field (along the path of the light beam 12) alternates with opposite signals to produce alternating positive and negative index transitions to form a grating. Two conductors 8 couple between the two electrodes voltage sources which control the intensity of the grating.

A second general problem with EO and prior art piezoelectric devices using uniform substrates and shaped electrodes is that the excited electric field decays rapidly with increasing distance from the electrodes. The pattern is almost completely washed out when its distance from the electrode is equal to the characteristic size of the pattern. This problem is exacerbated in gratings because of their small feature size. Existing gratings formed by interdigitated electrodes can only provide modulated effects in shallow surface layers. The EO structure interacts weakly with waveguides whose diameter is larger than the characteristic size. In higher order interactive devices where longer grating periods are available, this lack of resolution again severely limits efficiency. The minimum grating length for effective interaction of the prior art is about 10 microns. A method is needed to maintain the efficiency of EO devices based on small structures despite high aspect ratios (the resulting ratio of the width of the beam to the characteristic size). There is a need for a switchable shaped structure that lasts the full width of the waveguide and is even larger than the unguided beam.

At present, there are some related prior art that use a light source coupled to a waveguide structure for application in displays.

J. Viitanem and J. Lekkala, SPIE No. 1976, pp. 293-302, 1993, "Ray Liquid Crystal Displays", "High-Resolution Imaging" and their references) review the use of guided wave Principles Coupled liquid crystal switching features of flat panel displays, where a number of designs are discussed, all of which share the common design principles described below. A modulated light source is mechanically scanned through an array of electro-optical active waveguides that form the column elements of the display. A column of parallel electrodes forms the row positions of the display, light is guided from the waveguide and scattered to the viewer at a row position using the electro-optic effect, thus forming a two-dimensional array of pixels is formed.

In this prior art, light is confined in a waveguide consisting of a core material having a higher refractive index than the surrounding cladding material. Typically, light confined in the core is forced to "leak" out of the core of the waveguide at desired spatial locations. Since the difference in refractive index between the core and cladding material decreases electro-optically along a certain distance thus destroying the waveguide effect, electro-optical active materials could in principle be in the core (to lower the refractive index) or in the cladding material Inside (to increase the refractive index, the coating material is active in Wittnam et al.). In some literature, the technique of breaking the waveguide effect is called "waveguide tap".

The "leaked" light diffracts through a free space to a scattering center where it is directed toward the viewer to form the image cable of the display. The light leaking from the broken waveguide is then spatially confined, but instead expands in the region according to the standard diffraction theory, because the light propagates outward from the broken waveguide segment. The two-dimensional expansion of the light causes three problems.

First, the long interaction length (a significant component of the light energy has to leave the core before it can scatter toward the viewer) due to the small diffraction angle of light previously confined within the waveguide will limit the spacing of the scattering centers to Above one millimeter, this effect leads to low pixel packing densities in low resolution displays.

Second, the two-dimensional expansion of the beam makes it virtually impossible to collect a larger component at the scattering center and direct it toward the observer, resulting in low power efficiency.

Thirdly, the two-dimensional expansion of the beam creates large scattering centers, thereby increasing the pixel size and reducing the resolution of the display.

A consequence of the large spacing of the pixels is that long waveguide lengths must be used to cover enough pixels for the display, and then the display must operate in regions with large losses due to waveguide effects, again resulting in reduced efficiency. Therefore, existing designs suffer from low pixel packing density, large pixel size, and low power efficiency.

What is needed to solve these problems is the development of a compact, efficient, low-loss electro-optical waveguide switch that orients and pushes the entire beam from the column waveguide into a narrow solid angle so that the switched beam can be efficiently directed A pixel scatter center or enter another waveguide (connecting to another scatter center), which concentrates the beam at the scatter center and maximizes display efficiency and pixel packing density.

U.S. Patent No. 5,106,181 (April 1992) and U.S. Patent No. 5,009,483 (April 1991), "Optical Waveguide Display System" to Rockwell III suggest another approach using the "Waveguide Plug" approach. Example. Rockwell III suggests a display that uses waveguides to direct light in simultaneous columns. Light is connected from the waveguide to the cladding, and the electro-optical effect is used within the cladding. Although this structure is different from that suggested by Wittnen et al. above, the disadvantage of this display is also low pixel density and poor efficiency.

U.S. Patent No. 5,045,847 (September 1991) to Tarui et al. entitled "Flat Display Panel" suggests another version of the "waveguide plug" approach using a planar waveguide structure with a Layered core material composed of interspersed layers of silicon. Light from a laser diode source is confined within a waveguide until a voltage is applied across the core of the waveguide. The core refractive index is lowered by the voltage, thus allowing light to escape from the waveguide structure. Such displays suffer from the disadvantages mentioned above and are not efficient because all pixels are illuminated simultaneously, but only one pixel is activated at a time, so only a fraction of the light is directed toward the viewer at any one time.

Nelson in US Patent No. 5,083,120 (January 1992) entitled "Flat Panel Display Using Leaky Light Guides" shows a leaky light guide used as a display background light. Light from a light source, such as a laser diode, is directed into the light guide to provide uniform illumination of a row within the display. The backlight is combined with an array of ferroelectric liquid crystal shutters to create display pixels. At this time, the waveguide is only used to replace the fluorescent "waveguide plug" or switch commonly used in LCD displays.

Nishimura et al. in U.S. Patent No. 4,640,592 (February 1987) entitled "Optical Displays Using Thermoformed Bubbles in Liquid Core Waveguides" propose a line display with multiple liquid-filled lines or waveguides , these waveguides are positioned above a series of heating electrodes. When the heater is energized, bubbles form within the waveguide core, thus scattering light towards the observer. When the modulated laser mechanically scans each row of waveguides sequentially, a display image is formed. Even if this approach solves some of the problems caused by the "waveguide plug" approach by placing small scattering centers in the core, the timing of the heating process will not allow the display to have a frame rate fast enough to display a moving image, which is not present in this design. Use waveguide switches.

Existing technologies all use parallel input waveguides directly excited by the light source, because the switching mechanism known in the prior art cannot effectively switch the light from the supply waveguide into the desired traveling waveguide, resulting in Wittnan et al. and Nishimura et al. Bulky waveguide beam illumination scanning setup, simultaneous waveguide beam illumination by Rockwell III, planar waveguide excitation by Tarui et al., and simultaneous diode array illumination by Nelson and Nishimura et al. Yet the simplest structure of a display (where light is coupled into a single waveguide and directed to the pixels) requires at least two consecutive switches to illuminate the full two-dimensional array of pixels, one of which selects the row waveguide and the other selects the pixel. Although it is relatively difficult to manufacture pixel switches in the prior art, the row switches that must be connected with waveguides cannot have the above-mentioned waveguide plugs. The coupling efficiency between the waveguides is low, because the light diffracts and diffuses in two dimensions after the waveguide is broken in the switching region, so a switch is needed to efficiently condense the switching light into another waveguide.

The integrated optical frequency waveguide of Becker and Chang cannot be used as a display because it does not include a "waveguide plug" or other means to direct the light propagating in the waveguide to the viewer in front of the pixel format. Furthermore, the switch was fabricated using interdigitated electrodes as previously described (shown in conjunction with FIG. 1 ) and did not include any shaped polar structures. This switch design has a high switching insertion loss, so it cannot be used in applications that require a large number of switches on the same waveguide, such as displays or switching data buses for communication applications.

The problems of the prior art can be summarized as follows: (1) small high-resolution displays cannot be fabricated due to large pixel spacing, (2) inefficient "waveguide plugs" limit the brightness of the display, (3) low power efficiency due to optical losses , and (4) the power cannot be effectively switched to another waveguide.

The flat panel display according to the invention is based on a novel switching technique to direct the laser light in a set of optical waveguides and couple the light to the viewer, the switching technique is based on polarized electro-optic structures. This display technology is versatile enough to be used in many applications from small high resolution electronic computer monitors to large screen displays in high definition television format. The present invention combines the high brightness and power efficiency of visible semiconductor diode laser sources with novel waveguide-to-optic switching techniques to form dense two-dimensional addressable arrays of high-brightness light-emitting pixels.

The present invention provides an all-solid-state, full-color, high-resolution display, which can be used to display information generated by electronic computers and full-motion HDTV, so as to obtain a high-brightness display with high performance and can be driven by batteries. The invention also provides a portable small high-resolution display.

With the cooperation of the following detailed description and the accompanying drawings, the present invention can be further understood.

Illustration

Figure 1 is a modulator with interdigitated electrodes according to the prior art;

Figure 2 is a generalized embodiment of a switchable grating for interacting with a volume beam according to the present invention;

Figure 3 is an embodiment of a waveguide retroreflector utilizing a switchable grating;

Figure 4 is an embodiment of an electrode configuration for a retroreflective device having three electrodes disposed on the same surface of the crystal;

Figure 5 is an example of an electrode configuration for the same device, wherein two electrodes are arranged on the same surface of the crystal;

Figure 6 is an example of an electrode configuration for the same device, where three electrodes with tapered spacing are placed on the same surface of the crystal;

Figure 7 is a T-shaped embodiment of a polarized cross-waveguide coupler;

Figure 8 is an X-shaped embodiment of a polarized cross-waveguide coupler;

Figure 9 is an embodiment of a polarized waveguide output coupler outputting out of the plane of the waveguide;

Figure 10 is an embodiment of a parallel waveguide polarized directional coupler;

Figure 11 is a top view of an X-crossed waveguide coupler showing alternate input and output mode distributions;

Figure 12 is an embodiment of an X-crossed waveguide coupler having a tapered coupling region geometry excited with a tapered electrode gap;

Figure 13 is an embodiment of an X-cross waveguide coupler with generalized coupling region geometry and electrode pattern;

Figure 14 is a bulk optical embodiment of a tunable frequency polarized electro-optic retroreflector;

Figure 15 is a waveguide embodiment of a tunable frequency polarized electro-optic retroreflector;

Figure 16 is a bulk optical embodiment of an individually actuated frequency tunable electro-optic retroreflector with electro-optic coatings and polarizing gratings and coatings;

Fig. 17 is the waveguide embodiment of the electro-optic retro-reflector of a frequency doubling polarization;

Figure 18 is an explanatory diagram of a phase-shifted polarization grating;

Figure 19 is an embodiment of a multi-period grating reflector;

Figure 20 is an illustration of the frequency response curves of two devices with multiple periods and different free spectral ranges;

Figure 21 is an embodiment of a double grating tunable reflector;

Figure 22 is a schematic illustration of an integrated collimator consisting of dual gratings with adjustable optical path length;

Figure 23 is an embodiment of a double grating switchable Y-shaped joint with a phase shifter;

Figure 24 is an embodiment of a polarized waveguide mode converter;

Figure 25 is an embodiment of a waveguide router using waveguide mode converters;

Figure 26 is an embodiment of a switchable parallel waveguide resonator;

Fig. 27 is the embodiment of a three-arm waveguide etalon;

Figure 28 is an embodiment of a ring waveguide calibrator;

Figure 29A is an embodiment of a modulator/attenuator with a controllable polarization intermediate structure;

Figure 29B is an embodiment of an adjustable lens structure;

Figure 30 is an embodiment of a polarized total internal reflection (TIR) waveguide switcher with switchable polarized waveguide stubs;

Fig. 31 is the embodiment of a double TIR waveguide switcher;

Figure 32 is an embodiment of a TIR electrically switchable beam director with switchable non-polarized waveguide stubs;

Figure 33 is an embodiment of a two-position polarized waveguide router without TIR;

Figure 34 is an embodiment of a polarized TIR switch array with 50% switch packing density;

Figure 35 is an embodiment of a polarized TIR switch array with 100% switch density;

Figure 36 is an embodiment of a dual waveguide structure for a high density packed structure with permanently tuned mirrors and asymmetric lossy intersection regions;

Figure 37 is an embodiment of a switchable waveguide array with a TIR switch;

Figure 38 is an embodiment of a switchable waveguide array with a grating switcher;

FIG. 39A is an embodiment of an m×m communication switching array with system control lines;

FIG. 39B is an embodiment of a 3×3 switch array with WDM capability;

Figure 40 is an embodiment of a two-dimensional via array with pixel elements;

Figure 41 is an embodiment of a one-dimensional via array with pixel elements coupled to data traces;

Figure 42 is an embodiment of a switchable optical spectrum analyzer utilizing selectable grating reflector sections and a detector array;

Figure 43 is a schematic diagram of a polarized acoustic multilayer interferometer structure;

Figure 44 is an embodiment of a polarized acoustic transducer;

Figure 45 is an embodiment of a tunable coherent detector for multi-frequency light waves;

Figure 46 is an embodiment of a low-loss switchable waveguide beamsplitter utilizing a single polarization region;

Figure 47 is an embodiment of a low-loss switchable waveguide beamsplitter utilizing multiple polarization regions;

Figure 48 is a schematic diagram of key design elements for a 1 x 3 waveguide beam splitter;

Figure 49 is a multilayer stack of active waveguide devices shown as adjustable phased array modulators;

Figure 50 is an embodiment of a prior art adjustable waveguide attenuator;

Figure 51 is an embodiment of a multi-polarization segment adjustable waveguide attenuator;

Figure 52 is an embodiment with a structure that utilizes an angularly broadened polarization grating to broaden the frequency band;

Figure 53 is an embodiment with a structure that utilizes a curved waveguide to broaden the frequency band;

Figure 54 is an embodiment of an electrically controllable polarizing lens.

Figure 55 is an embodiment of a laser feedback device using a periodically polarized reflector.

Figure 56 is an embodiment of a laser feedback device using periodically polarized waveguide reflectors.

Figure 57 is an embodiment of a laser feedback device using multiple switched feedback gratings.

Figure 58 is an embodiment of a wavelength tunable tunable focusing system.

Figure 59 presents a diagram of a generalized embodiment of a display device according to the present invention.

FIG. 60 is a cross-sectional view of the generalized embodiment of FIG. 59. FIG.

Fig. 61 is a diagram of an embodiment of a three-color beam combiner.

Figure 62 is a diagram of an embodiment of an out-of-plane reflector where the mirror is reflected through the substrate.

Figure 63 is a diagram of an embodiment of an out-of-plane reflector with divergent and diffuse reflection through a substrate.

Figure 64 is a diagram of an embodiment of an out-of-plane reflector where the reflection is off the surface of the substrate.

Figure 65 is a diagram of an embodiment of an out-of-plane reflector where the divergent and diffuse reflections are away from the substrate surface.

FIG. 66 is a diagram of another embodiment of a pixel switch structure combined with the attenuator structure in FIG. 50 or FIG. 51 .

Figure 67 is a diagram of an embodiment of an out-of-plane grating reflector.

68 is a diagram of an embodiment of an absorber/deflector light shield for a display pixel showing the relative placement of the light shield and the pixel.

Figure 69 is a diagram of an embodiment of an absorber/deflector light shield that deflects ambient light away from the substrate surface.

Figure 70 is a diagram of an embodiment of an absorber/deflector light shield that polarizes background light into a substrate.

Figure 71 is a diagram of an embodiment of an absorber/deflector light shield that substantially absorbs background light.

FIG. 72 is a diagram of an embodiment of a part of an electrode pattern for driving a normal two-dimensional via array similar to FIG. 34 .

Figure 73 is a diagram of an embodiment of an out-of-plane switch in a multilayer stacked waveguide distribution structure that directs reflected light into the substrate.

Figure 74 is a diagram of an embodiment of an out-of-plane switch in a multilayer stacked waveguide distribution structure that directs reflected light into the plane of the substrate.

Fig. 75 is a diagram of an embodiment of a -90° grating waveguide switch, which is introduced in the form of a three-input waveguide.

Figure 76 is a diagram of an embodiment of a tilting display.

Figure 77 is a diagram of an embodiment of a light directing structure with multiple parallel supply waveguides, multiple partial reflectors, and tuners.

Figure 78 is a structural diagram of a display supplied by a laser array.

Figure 79 is a diagram of an embodiment of a multi-laser array coupled to an interdigitated waveguide.

Figure 80 is a diagram of a coupling embodiment of an out-of-plane waveguide.

Fig. 81 is a diagram of an embodiment of a dual-polarized display structure used in a three-dimensional display.

The claims of this application relate specifically to the structures described in FIGS. 59-81. Referring to Figure 2, there is shown a generalized embodiment of the device 11 of the present invention, which is a patterned polarized dielectric device. The device is basically an electrically controllable stacked dielectric light energy redirector, or more simply, an electrically switchable mirror. In a preferred embodiment, the present invention is a bulk photoreflector in a lithium niobate ferroelectric crystal 20, the electronically controlled switching element is a polarized grating 22, which is composed of two types 36 and 38 composed of alternating polarized domains.

A domain—which can be of any shape or size—is a physical region in which certain material properties are approximately constant. A polarized domain is a region in a material in which molecular groups have a directionality and are substantially aligned (or partially aligned), or approximately aligned, in a direction known as the polarization direction direction. There are many types of domains, they include: domains of atomic structures aligned in different directions, domains of molecular or atomic structures aligned with various parameters like nonlinear activity or electro-optical coefficients, domains of non-linear Domains of preferentially oriented atomic structure, defined by regions activated by different electrodes, as occurs in the case of polymers and fused silica polarized with local electrodes, where the polarization direction changes systematically over the entire region domains, domains of randomly oriented molecules, and extended-range, random domain structures: domains of randomly polarized subdomains within a domain. A polarized structure is a collection of individual domains. A patterned polarized region is a region in a material in which the domains within the region are polarized into more than one domain type in a spatial pattern. Depending on the nature of the poling method, there may be systematic deviations between the polarized pattern and the applied pattern used during the poling process. The borders of the pattern may also be somewhat irregular and may not be exactly the same as the applied pattern, especially if the polarization process is not fully under control. Since the device is controlled using an electric field, the device is described as a patterned polarized dielectric, so the material must be dielectric in order to withstand the required electric field without damage. The typical poling process is also done with an electric field that the material must be able to withstand. Generally speaking, we use dielectric to indicate the ability of a material to withstand the minimum electric field required in use.

In operation, an input beam 40 is incident on the crystal along an optical axis and passes through the crystal. The optical axis is perpendicular to the phase front of the beam and is determined by the average position of the brightness distribution of the propagating beam on the wavefront. The optical axis is straight in homogeneous materials, but may be curved in several cases including curved waveguides, inhomogeneous media, and in reflective or diffractive structures. The input beam 40 preferably has a sufficiently small spot size 21 over the entire crystal length so that it is not perforated by the crystal, causing undesired power loss and mode conversion. In a bulk interaction device as shown in FIG. 2, the domains 36 and 38 must penetrate the substrate 20 a sufficient distance so that they overlap each other with at least a portion of the input beam 40. The grating 22 is positioned transversely with respect to the input beam 40 . This means that the plane 34 of the grating 22 is transverse to the axis of the input beam 40 . We say that two lines (or a line and a plane, or two planes) are transverse to each other meaning that they are not parallel. Since the grating is transverse to the light beam 40 , the light beam passes through at least a part of the structure of the grating 22 .

Beam 40 is generated from an optical frequency source (not shown) and has a wavelength such that the beam is substantially not absorbed by the crystal and such that photorefractive effects do not distort the beam significantly. The optical frequency source arrangement may include one or more optical actuators capable of providing sufficient brightness at wavelengths within the range accepted by the grating reflector 22 to produce a useful switched output beam 44 . The output beam may be coupled into other components on the same substrate, or it may be coupled into an external device, in which case the output surface from which the beam emerges is preferably non-reflectively coated. The non-reflective coating can be a multi-layer dielectric coating, a single quarter wave coating with a well-suited refractive index material, or a sol-gel coating. The exciter can be any light source, including lasers, light-emitting diodes, arc lamps, discharge lamps, or even incandescent lamps, as long as the appropriate spectral brightness is achieved. Appropriate spectral brightness may be provided directly by one or more exciters, indirectly by one or more frequency converting (frequency doubling, mixing, or parametric amplification) exciters, or by a combination of the above combination provided. Absorption effects limit the wavelength to a range of approximately 400 to 4000 nm. The effect of photorefractive phenomena varies with configuration, wavelength, dopant, and polarization configuration, and here we assume it is under control so that any beam distortion is within acceptable limits.

The grating 22 is formed or defined by boundaries 34 of two types of domains arranged alternately. Domains 36 of the first type have a different electro-optic (E-O) coefficient than domains 38 of the second type, so a uniform electric field applied between electrodes 24 and 26 results in a difference in the refractive index in the two types of domains. Variety. Since the refractive index changes the wave's phase velocity, there is an impedance mismatch in regions of different refractive index or phase velocity. Regions 36 have opposite orientations in the material relative to the other domain types 38 and the original wafer 20, as indicated by the polarization direction arrows 39, 41, contributing to the material's purpose of this refractive index change. We use opposite pointing to indicate the direction of polarization opposite to some reference direction. (An alternative method of controlling the grating with an electric field is realized in a radiation-masked polymer film whose E-O coefficient is disrupted inside or outside region 36.) A uniform electric field applied to structure 22 produces a modulated refractive index. The index-modulated pattern is added to a pre-existing index profile; the simplest configuration has no index modulation in the absence of an applied electric field, and produces an index grating in linear response to an applied electric field. A period 48 of the grating 22 is the distance between two domain boundaries completely encompassing a region corresponding to each domain type.

An alternative method of realizing a refractive index grating is obtained by applying a stress field to the polarized region. The photoelastic response of the material produces different refractive index changes in different polarization regions. A permanent stress field can be applied, for example, by laying a thin film on top of the substrate at elevated temperature and then cooling to room temperature. For example, stress concentrations can be created by etching away a strip of film.

Polarized elements 36 and 38 alternately fill the entire grating 22 with no spaces between them. If additional types of domains are available, more complex alternating patterns can be formed using domains separated by variable distances of different types of domains. For some applications, grating 22, as shown in FIG. 2, is a uniformly periodic grating such that domain types contained in one period along the length of grating 22 recur in other periods. For further applications, it is helpful to vary the period to gain advantages such as multiple spectral peaks or wider spectral bandwidth. We use the term grating to mean arrays that include all possible kinds of geometric shapes and periodic recognizable structures.

Periodic index gratings can provide virtual photons in the interaction between light beams. This suggests that the grating structure can provide momentum, but not energy, to the interaction. In order to interact, both energy and momentum must be conserved, and gratings are useful when momentum needs to be increased to satisfy both conservation relationships. The periodicity of the grating determines the momentum available for the interaction. The strength of the grating determines the "brightness" of the virtual photon beam. The number of periods in the cross-section of the grating that the beam traverses determines the bandwidth of available virtual photon momentum. Due to bandwidth limitations, the interaction occurs only within a specific spectral range (or multiple spectral ranges) of optical frequencies. The grating arrangement is therefore inherently frequency selective and typically operates around one nominal wavelength.

For example, during simple reflection at an angle as shown in FIG. 2 , photons of input beam 40 have the same optical frequency as photons of output beams 44 and 42 , thus energy conservation is observed. However, the photons in the input beam 40 and the redirected output beam 44 have different momentums; in order for the reflection process to occur, the grating 22 must provide a change in momentum, as shown in the vector diagram 43 associated with FIG. 2 . The grating 22 provides virtual (momentum but no energy) photons for the interaction so that momentum can be conserved. The momentum vector associated with the i ^th mode, _ki = 2πn _i /λ _i , is equal _{to the product of 2π times the effective index ni of that mode divided by the wavelength λ i of} that wave, and it points in the direction of propagation. The magnitude of the momentum vector is also known as the propagation constant. In the case of a single periodic grating, the momentum vector k ₁ =2π/Λ is directed perpendicular to the grating surface, and it can have any wavelength value Λ that exists in the Fourier transform of the grating. The optical spacing (width of grating lines and spaces) related to the propagation constant _ki of a 50% duty cycle grating is thus Λ/2. The frequency of the interaction can also be tuned by adjusting, for example, the reflectivity of the beam, or changing the period of the grating using thermal expansion or other means. Depending on how a given device is used, a refractive index structure can have a spectrum of wavelengths and vector directions available for interaction. Multiple virtual photons can also be used for interactions in so-called "higher order" grating interactions. A "higher order" grating is one that has a period related to the period required for momentum preservation divided by an integer. The required momentum virtual photons are obtained from the harmonics of the "higher order" gratings. The conditions for momentum that can be preserved by this method are generally referred to as Bragg conditions, so the grating of the present invention is a Bragg grating, and the angle of incidence on the grating is the Bragg angle for the in-band or resonant frequency components. This dual conservation of energy and momentum is necessary for any energy beam interaction, whether that energy beam is a light beam, a microwave beam, an acoustic beam, or any other wave-like energy form that includes a time-varying energy field . Only the use of gratings can be varied to generate impedance modulations for different energy forms so that patterns of structures can be coupled with wave-like energy forms.

In Figure 2, the refractive index grating acts as a frequency-selective optical energy router or reflector. A beam of characteristic frequencies (capable of interacting with one or more virtual photons) within the interaction bandwidth is called an in-band beam, and energy beams of other frequencies are called out-of-band beams. The grating 22 has a bandwidth corresponding to the full width of half the maximum value of the reflective efficiency of the grating as a function of optical frequency. When a refractive index grating is present (grating "on"), light beams having optical frequencies within the bandwidth of the grating reflect from the grating at an angle 46 around a normal 47 to the grating structure. The out-of-band beam is emitted through the crystal along the same optical axis as the input beam and in the same direction as the input beam, forming a partially emitted output beam 42 . An electric field applied in the region containing the grating controls the intensity of the refractive index modulation (which can also be thought of as the intensity of virtual photons), adjusting the ratio of power in the emitted output beam 42 and reflected output beam 44 .

For a weakly retroreflective grating (which does not completely exhaust the input beam), the full-width half maximum bandwidth Δλ is given by: $\mathrm{\Δ\λ}=\frac{{\mathrm{\λ}}^2}{2.24\mathrm{nL}}---\left(1\right)$

in:

λ = vacuum wavelength of the input beam.

n = refractive index of the beam, and

L = length of the grating.

For highly reflective gratings, the effective length is less than the total length of the grating, thus increasing the bandwidth.

The two types of domains can exhibit a difference in refractive index before an electric field is applied. In this case, a permanent index grating accompanies a polarization switchable index grating. When an electric field is applied, the net modulation of the refractive index (grating strength) may increase or decrease depending on the polarity. At a certain value of the applied electric field a "grating off" condition occurs (refractive index grating value approaches zero). The grating can then be "turned on" by applying any other field strength. If the polarity of the applied electric field is reversed, for example, a refractive index grating with twice the strength of the original permanent grating will be produced.

The polarized grating structure of the present invention has two main advantages over the prior art. First, the polarized domain structure can have very sharp boundaries, providing a strong Fourier coefficient for the virtual photon momentum, which is several times larger than the momentum corresponding to the fundamental grating period. This is useful where photolithography is not possible due to the small form factor required. Second, strongly index-modulated gratings can be produced even at large optical waveform sizes compared to the grating period. This is not possible in uniformly polarized substrates excited with patterned electrodes because the electric field modulation decreases exponentially with distance from the plane of the electrode array, losing most of the modulation when the distance is equal to the grating period. The poling method can produce poling profiles with very large aspect ratios, or extremely large domain depth-to-width ratios. Using electric field polarization technology, aspect ratios exceeding 250:1 have been achieved. Since we use substantially uniform electrodes, we achieve good electrostatic penetration; thanks to deep domain walls, good modulation of the entire beam is possible.

The grating can also be a two-dimensional array of refractive index variations, in which case the grating has periodicity in two directions. Therefore, the virtual photons provided by the grating can provide momentum in two directions. This can be used, for example, to output several beams from a single grating.

In the preferred embodiment, the ferroelectric crystal is a commercially available Z-cut lithium niobate monolith. Other cuts including X-, Y-, and angle-cuts can also be used, depending on the polarization method and desired orientation of the polarized domains. The fabrication steps include initial poling and electrode fabrication. Crystals are cleaned (eg, ashed with oxygen plasma) prior to processing to remove all hydrocarbons and other contaminants from the polishing and handling process. To control polarization, masks and process electrodes are used to create a pattern of applied electric fields at the surface of the wafer and across the wafer, as described in U.S. Patent Application Serial No. 08/239,799, filed May 9, 1994 like that. During application of the polarizing electric field, the polarization pattern is adjusted to produce polarization domain inversion in region 36 . Briefly, a layer of silicon dioxide a few microns thick is deposited on the +Z surface 23 of the wafer 20 . Thinning or removing the silicon dioxide film on the region 36 where domain inversion is desired, using a liquid electrode or deposited metal film to create a good equipotential surface on the patterned silicon dioxide, and then applying a The electric field exceeds about 24 kV/mm and makes the +Z surface 23 at a higher potential than the -Z surface 25. Using this technique, a ferroelectric crystal of lithium niobate has been polarized, producing a pattern of two types of domains with opposite polarities (domain inversion). The magnitudes of the electro-optic coefficients for the two types of domains are the same although they have opposite polarities.

In addition to the present preferred technique, domain inversion has also been achieved in ferroelectric materials using indiffusion, ion exchange, and alternating electric field polarization techniques. Domain formation in lithium niobate has been achieved using titanium through a thermally enhanced indiffusion technique. The triangular shape of the inversion region limits the interaction efficiency of small-sized domains, but can be mainly used in long-period waveguide devices. Figure polarization in KTP by ion exchange is shown in a salt bath containing rubidium and barium ions, where potassium ions in the crystal are replaced by rubidium ions. Magnetic field polarization using the alternating magnetic field technique is the preferred method, which has also shown success in both lithium niobate and lithium tantalate. All solid ferroelectric materials including KTP and barium titanate can be polarized by electric field domain inversion technique. (The so-called solid means that it can maintain its structure for a certain period of time, such as cooled fluid, glass, cross-linked polymer, etc.)

Different techniques are used to produce gratings with different characteristics. Electric field polarization aligns the domains in the crystal without substantially changing the refractive index, but ion exchange and diffusion techniques do cause refractive index changes in the polarized regions. A permanent index grating accompanies a switchable polarization grating when using these latter techniques.

In general, there are two different types of domains, at least the first type of domains being polarized. Although only two types of domains are required, more complex switchable grating structures can be fabricated with additional types of domains. The second domain type can be polarized in the opposite direction, unpolarized, or polarized at other angles, and can be distinguished by a specific electrical activity coefficient (eg, electro-optic or calender-optical coefficient) it possesses. For example, in some applications, devices can be fabricated at lower cost from unpolarized lithium niobate wafers, in which case the substrate wafer includes a plurality of randomly oriented domains. Polarized domains will have uniform directionality, while directionality in other domains will be random. Depending on the type of device, the details of the random pattern will affect the performance of the device. As another example, domains of the second type can be oriented perpendicular to the first type or at another angle, and the difference in electrical response can still yield useful electrically controlled structures. Polarized domains can also form in a material that has not been previously polarized and is randomly oriented at the molecular level, for example in fused silica or polymers. The poling process orients the material structure to form a first domain type, while the second domain type is made up of unpolarized or randomly oriented regions in the material.

In an alternative technique, the polarized structure can be formed by selectively altering or destroying the electrical activity coefficient in regions corresponding to the second domain type. There is no need to change the orientation of the atomic structure in these regions: if you change the electrical activity in the second domain region, the domains are different. In nonlinear polymers, for example, irradiation can be used to deactivate the electro-optic coefficients and generate electroactive regions in radiation-shielded regions. The same effect was shown in lithium niobate, where proton exchange disrupted the nonlinear coefficients. Electro-optic coefficients can also be altered in many other materials—including most nonlinear materials such as KTP and lithium tantalate—by photoirradiation, electron bombardment, and/or ion bombardment techniques.

In lithium niobate, an electric field _E3 applied along the z-axis of the crystal causes a change in the extraordinary refractive index _δne , _which is given by: ${\mathrm{\&delta;\; n}}_e=\frac{r_{33}{\mathrm{E.}}_3{\mathrm{no}}_{\mathrm{<figure-callout\; id="3"\; label="e"\; filenames="A9519499001521.png"\; state="\{\{state\}\}">e</figure-callout>}}^3}{2}---\left(2\right)$

where _r33 is the appropriate electro-optical nonlinear optical coefficient. Since _r33 is the largest nonlinear constant in lithium niobate, it is preferable to exploit the extraordinary refractive index variation in practical devices. (The nonlinear constant _r13 , which produces a change in the ordinary refractive index due to the applied electric field E3, is a factor 3.6 smaller than _r33 .) To take advantage of the change in the extraordinary refractive index, the light wave must be polarized along the z-axis of the material. In a Z-cut crystal, this polarization is called TM. (In TE polarization, the electric vector lies in the plane of the crystal surface. The only other non-linear coefficient of importance is r ₁₅ , which couples TE and TM waves when an electric field E ₁ or E ₂ is applied.)

Since the refractive index change caused in the polarized structure is quite small (10V/μm electric field applied along the z-axis of the lithium niobate substrate, the refractive index change _δne is only 1.6×10 ^-3 ), the Grating reflectors have a strong angular dependence. The Brewster angle for a weak index change is 45°, so when the plane of the grating is at 45° from the beam phase front, the grating will fully emit any TE polarized wave. Thus, the device can be used as a polarizer. The reflected beam will always be polarized at a 45° angle of incidence. If the reflection coefficient of the TM wave is high, (sufficient grating period and high applied electric field can be arranged to increase the reflection coefficient), the extinction ratio of the polarizer can also be very high in the forward direction. Of course, at normal incidence, the reflection between the two polarizations does not differ due to this effect (although there are differences due to other effects, eg different electro-optic coefficients as described above). Total internal reflection devices operating at grazing incidence are far from the Brewster angle and there is only a small difference in reflection due to this effect.

The wafer material can be any polarizable solid dielectric material, including ferroelectric materials, polymer film materials, and some amorphous materials, such as fused silica, which can also be polarized to produce many useful device. The polarized material may also be a thin film deposited on a substrate of a second material. Many polarizable films have been successfully deposited on substrates, such as fused silica, lithium niobate, potassium niobate, barium titanate, zinc oxide, II-VI materials, and various polymers, etc. . A wide variety of substrates have been used, including MgO, silicon, gallium arsenide, lithium niobate, and various glasses including quartz and fused silica. For domains to be electrically switchable, they must consist of electro-optic materials that have a refractive index change induced by an applied electric field.

After the poling step, the liquid electrode and the silicon dioxide masking film are preferably removed. Referring again to FIG. 2, the first electrode 24 and the second electrode 26 face the dielectric material so as to provide a means for generating an electric field for controlling the grating. (Facing a material is meant to be positioned close to the material but not necessarily touching, nearly aligned to the surface of the material but not necessarily with a fixed gap size, and includes the case of additional material of various sizes placed on top of the material.) Made of conductive material The constituent electrodes 24 and 26 are preferably applied in a spaced apart manner on opposite surfaces of the crystal using standard deposition techniques. The electrodes are said to be on opposite planes even though the surfaces may be curved and/or non-parallel as part of a larger geometry. These electrodes may be formed of any material that provides sufficient charge transfer to achieve the appropriate field strength and thus excite the polarized grating for a time consistent with the time of application. For example, electrodes can be made of metal materials such as aluminum, gold, titanium, chromium, conductive paint, epoxy resin, semiconductor materials, or light-transmitting materials such as indium and tin oxides, and liquid conductors such as saline solutions. of. They may also face surfaces 23 and 25 with gaps filled with air, light-transmissive buffer layers, and/or other materials. Only one electrode is required, since a potential voltage difference can be established between this electrode and any potential reference, such as an external ground plane, a second electrode, or multiple electrodes. An electrode is an electric field establishing device since the voltage applied to the electrode establishes an electric field pattern determined by the electrode. Of course voltage and current sources are also required. The electrodes are arranged in such a way that the control electric field is applied through the active volume of the invention, which may consist of a patterned polarized area or a grating.

In the case of metal electrodes, it is preferable to incorporate a coating deposited below the electrode to reduce the optical loss that occurs when a portion of the guided wave mode extends to the metal electrode. Where multiple electrodes are mounted on the same surface, the coating should be thin enough to sustain high electric fields, but thick enough to reduce optical loss. It is also possible to use another coating over the electrodes to reduce the possibility of breakdown.

A voltage control source 32 (or a potential source) provides potential to drive the electrode-activated grating via connecting lines 30 . The activated electrodes are polarized relative to each other according to the polarity of the applied voltage. The voltage of the voltage source produces an electric field large enough across the polarized region to switch a substantial amount of light into the switched output beam 44 . The voltage of the voltage source is variable to provide a means to control the power ratio in the two output beams. In fact, if the electric field is high enough, a long grating can be used to reflect all the input beam, forming an electrically active mirror. For lower electric fields, the grating forms a partial reflector. The voltage controlled source may be a battery, an electrical transformer, a gas driven generator, or any other type of controllable source of current and potential. Control unit 32 may also include a controller that generates a time-varying voltage and supplies current to vary the voltage across electrodes 24 and 26 at the frequency required by the application. The control device 32 can also have multiple outputs that can control multiple devices, and which can be temporally ordered according to certain graphics. Source 32 may have control inputs from a computer or other instrument for manually or electronically controlling its functions.

In order to avoid unnecessary repetition, it should be understood that the various changes described with reference to FIG. 2 are applicable to the embodiments described below, and the various changes described with reference to the following figures are also applicable to FIG. 2 .

Referring now to Figure 3, there is shown a waveguide embodiment of the present invention. In particular, this embodiment is an electronically controlled, frequency-selective waveguide retroreflector. All light beams in the device are confined in two dimensions by an optical waveguide 64 passing through a surface of the polarizable dielectric material of the substrate 60 forming the device 61 .

A waveguide is any structure that allows waves to propagate through its length despite diffractive effects and is bendable. An optical waveguide is defined by an extended region having an increased refractive index relative to the surrounding medium. The strength of wave guidance or confinement depends on wavelength, refractive index difference, and guidance width. Stronger limits generally result in narrower waveforms. A waveguide can support multiple optical modes or only a single mode, depending on the strength of the confinement. In general, a light wave type is distinguished by its electromagnetic field two-dimensional geometry, its polarization state, and its wavelength. The polarization state of a waveguide or asymmetric waveguide in a birefringent material is typically linear polarization. However, common polarization states can contain non-parallel polarized components, as well as elliptical and unpolarized components, especially when the waves have large bandwidths. If the refractive index difference is small enough (eg, Δn = 0.003) and the waveguide is sufficiently narrow (eg, W = 4 μm), the waveguide will only confine a single transverse mode (lowest order mode) in a certain wavelength range. If the waveguide is placed on the surface of a substrate so that there is an asymmetry of refractive index above and below the waveguide, there is a cutoff value in the refractive index difference or waveguide width below which there is no mode limitation. A waveguide can be placed in the substrate (e.g., with indiffusion), on the substrate (e.g., by etching away the surrounding area, or by applying a coating and etching away all but the strips defining the waveguide ), within a substrate (for example, by bonding or gluing together several processed substrates). In all cases we speak of waveguides passing through the substrate. An optical mode propagating in a waveguide has a lateral dimension that is related to all limiting parameters, not just the waveguide width.

The substrate is preferably a lithium niobate single crystal, formed as a chip with two opposing surfaces 63 and 65 separated by the thickness of the wafer. The two opposing surfaces need not be parallel, or planar. The waveguide is preferably fabricated by existing techniques, such as annealed proton exchange (APE) at the surface 63 . Alternatively, ions, rather than protons, can be inwardly discretized or ion-exchanged into the substrate material. The APT waveguide increases the extraordinary refractive index of the crystal, forming a waveguide for polarizing light along the z-axis. For Z-cut crystals, this corresponds to the TM polarization mode. Waveguides formed by alternative techniques, such as inward diffusion of titanium into lithium niobate, can support both TM and TE polarization.

The waveguide is preferably designed to support only a single lowest order transverse mode to eliminate the complexities associated with higher order modes. Higher-order transverse modes have different propagation constants than the lowest-order modes, and higher scattering losses, which can cause problems in some applications. However, multimode waveguides may be preferable for certain applications, such as for high power propagation.

An alternative configuration is to actuate the grating by applying pressure, rather than directly applying an electric field. The effect of applying pressure is indirectly the same: By virtue of the piezoelectric effect, the applied stress generates an electric field, which in turn changes the refractive index of the domains. However, no holding energy needs to be applied to maintain the stress, for example if the structure is compressed mechanically. This alternative, like the others described here, is also applicable to other similar implementations of the invention described below.

Once the dimensions of the waveguide are determined, a photomask for the waveguide can be created and the pattern transferred to the mask material on the substrate using one of a number of known photolithographic processing techniques. The mask material can be SiO ₂ , tantalum or other metals, or other acid resistant materials. To fabricate an APE waveguide, the masked substrate material is dipped in molten benzoic acid to exchange protons from the acid for lithium ions in the crystal. The resulting step-index waveguide can then be annealed at around 300°C for several hours to diffuse the protons deeper into the crystal and produce a low-loss waveguide with a high electrical activity coefficient.

In addition to indiffused and ion-exchanged 2D waveguides, planar and 2D ridged or striped waveguides can also be formed. Planar waveguides can be formed by depositing electroactive materials on a low refractive index substrate. Deposition techniques for waveguide fabrication are known and include liquid phase epitaxy (LPE) molecular beam epitaxy (MBE), flame hydrolysis, spinning, and sputtering. Ridge waveguides can be fabricated from these planar waveguides using fabrication methods such as lift-off, wet etching, or dry etching such as reactive ion etching (RIE). Planar waveguides can also be used in the present invention, particularly in devices utilizing variable diffraction off gratings.

The grating 62 in this embodiment is arranged perpendicular to the optical waveguide 64 which passes through the substrate. The grating is made up of domains of one type 66 and domains of a second type 68, which do not necessarily extend across the substrate. For example, when using in-diffusion or ion exchange polarized active materials, inversion domains 66 generally extend to a finite depth into the material. Partial domains can also form when ion bombardment or UV irradiation is used to disrupt the activity of the material (or reduce the electro-optic activity) to achieve polarization.

The input light beam 80 is incident on the waveguide and is coupled into the waveguide. Coupling is the transfer of power from one region to another across some generalized boundary, such as across an interface, or between two parallel or angled waveguides, or between Between a planar waveguide and a striped waveguide, or between single-mode and multimode waveguides, etc. When the grating is on, a portion of the input beam is coupled back into a retroreflected output beam 82 . The retroreflection of the grating need not be perfect, i.e. the grating can reflect within a few degrees in the opposite direction, while the waveguide captures most of the light and forms a perfectly retroreflective beam. Imperfections in back reflections cause coupling losses for the back reflected light beams as they enter waveguide 64 . When the grating is switched off (when the control electric field is adjusted to the "off" position, where the refractive index of the grating has an almost zero minimum, typically at zero electric field), the input beam continues to propagate through the waveguide in the same direction to form An emitted output beam 84 . In bulk devices, the intensity of the grating can be varied with voltage source 76 to control the power ratio in the two output beams.

The first electrode 70 and the second electrode 72 face opposite surfaces of the dielectric material 60 . The substrate is dielectric in that it can withstand an applied electric field without damage, but it need not be a perfect insulator as long as the current flow does not adversely affect the performance of the device. The electrodes may be formed from any conductive material. There must be a means to use the first electrode structure to establish an electric field across the dielectric material.

The electrodes connect at least two elements of a first type of polarization structure forming a grating. This indicates that the electric field generated by the electrodes penetrates to at least two elements. Therefore, these elements can be activated by an electric field. Two wires 74 connect a voltage control source 76 to the two electrodes to provide an electric field in the region formed by the junction of waveguide 64 and polarization structure 62 . The wires may be formed of any material, and may be of any shape, which is sufficiently conductive at the operating frequency to charge the electrodes to the extent required for the application. Conductors can be round, flat, coaxial, or integrated pilot pattern conductors, and they can be resistors, capacitors, semiconductors, or leakage dielectrics.

Alternatively, the electrodes may be arranged in any manner that allows an electric field to be applied across the electroactive material. For example, the electrodes may be alternately disposed in different layers on the substrate with the active material disposed between the electrodes. This configuration can generate high electric fields with low voltages and is particularly useful for amorphous active materials that can be deposited on electrode materials, such as silicon dioxide and some polymers.

The poling structure 62 is preferably deeper than the waveguide, such that the interface between the waveguide 64 and the poling structure 62 has the lateral dimension of the mode in the waveguide and the longitudinal dimension of the grating.

Figures 4, 5 and 6 show alternative electrode configurations in which the electrodes are mounted on a common plane of dielectric material 189 . Because coplanar electrode configurations allow high electric fields at low voltages, these configurations are particularly useful for embodiments of the invention that use waveguides 180 to guide light beams. These electrode configurations are of particular interest for low voltage control of the grating 182 due to the close proximity of the electrodes to the waveguide section through the grating. In the electrode configuration 186 shown in FIG. 4, the first electrode 170 and the second electrode 172 face the dielectric material on the same surface. These electrodes are said to be on the same plane even though the plane may be a curved part of a larger geometry. The first electrode is arranged above the part of the waveguide comprising several grating elements, each grating element consisting of alternating regions of one domain 184 of the first type and one domain 185 of the second type. The second electrode is disposed around the first electrode. The distance between the electrodes along the waveguide is approximately constant along the axis of the waveguide to provide the desired uniform electric field along the axis of the waveguide. The electrode spacing can also be varied to create a tapered field strength that tapers off, as shown in device 188 of FIG. 6 . A voltage source 174 connected between two electrodes placed as shown in FIG. 4 can generate an electric field between the electrodes. The largest component of the electric field vector 176 is normal to the material surface in the region of the electroactive waveguide. For ferroelectric crystals like lithium niobate, this electric field configuration activates the largest electro-optic coefficient r ₃₃ , establishing a refractive index change for TM polarized beams. For an applied electric field of 10 V/μm in lithium niobate and a beam of wavelength 1.5 μm, the intensity of the first order grating is 40 cm ⁻¹ .

For each electrode configuration, a device 178 is required to connect the electrodes to a voltage source. To form such a device, a conductive material, such as a wire, is electrically connected between the terminals of the potential source of the electrodes of the device. In all electrode configurations, each electrode generally has a region, or pad, or contact to which the power supply wire contacts. The spacers are preferably of sufficient size to reduce placement tolerances on the electrical connections and facilitate bonding. The wires can then be attached to the pads using, for example, ultrasonics, heat, or wire bonding techniques such as conductive epoxy. Alternatively, a spring-loaded piece of electrical conductor can be used to directly contact the electrodes to achieve the desired electrical connection to the voltage source. In the drawings, the electrodes are generally large enough to serve as contact pads themselves.

Fig. 5 shows another embodiment of the coplanar electrode structure, in which the first electrode 171 and the second electrode 173 are arranged on either side of the optical waveguide. When a potential is applied across two electrodes so positioned, the largest component of the electric field vector 177 is parallel to the substrate surface. For Z-cut ferroelectric crystals, the electro-optic coefficient establishing the refractive index change and applied electric field for TM polarized light waves is _r13 . For an applied electric field of 10 V/μm in lithium niobate and a beam of 1.5 μm wavelength, the first order grating coupling constant is 12 cm ⁻¹ .

Alternatively, for TE waveguides, the active electro-optic coefficient can be switched for both configurations. For the case where the electric field vector is perpendicular to the chip surface, the appropriate coefficient is r ₁₃ , and for the case where the electric field vector is parallel to the chip surface, the electro-optic coefficient used is r ₃₃ . The same situation can be used for X- or Y-cut crystals, or intermediate-cut crystals.

As another variation on the configuration of Fig. 5, the electrodes are arranged asymmetrically so that one of the electrodes nearly covers the waveguide 180, while the other electrode is placed at the side. In this configuration, a strong vertical electric field induced under the edges of adjacent electrodes passes predominantly through the waveguide region under one of the electrodes.

In FIG. 6, electrodes 175 and 179 are spaced apart from central electrode 181 to form a cone. When a voltage is applied across these electrodes, this configuration produces a cone of electric field strength, with strong fields to the right and weak fields to the left. "Tapered" means any parameter that has a generalized spatial variation from one value to another without specifying whether the variation is linear or monotonic; such parameter may be gap, width, density, refractive index, thickness , duty cycle, etc. Thus the refractive index change induced in the polarized domains towards the left of the waveguide 180 is weaker than the refractive index change induced towards the right. This may be useful, for example, to obtain very narrow bandwidth total reflectors where it is necessary to extend the length of the interaction region. In non-normal incidence angle devices, as shown in Figures 7 and 8, the taper can be used to preferentially couple a specific input mode into a specific output mode.

In all electrode configurations, the applied voltage can range from a constant value to rapidly changing or pulsed signals, and any polarity can be applied between the two electrodes. The voltage value is chosen appropriately to avoid sudden damage to the electroactive material and surrounding materials in a given application.

When a steady electric field is applied across a material such as lithium niobate, charge buildup on the electrodes can cause a DC drift in the field strength over time. Changing the polarity of the voltage source from time to time cancels out this variation, thus returning the electric field strength to its full value. If the electric field between averaging times is close to zero, the net charge drift will also be close to zero. For drift-sensitive applications, the photorefractive sensitivity of the material has to be carefully reduced, for example by inward diffusion of MgO, and is preferably operated without a DC field.

The surface layer helps prevent electric field breakdown and loss of optical contact with the electrodes. Losses are especially important for waveguide devices since light beams propagate on or near surfaces. This also applies to the polarization of active materials, as well as to electro-optic switching. The largest vector component of the electric field between two coplanar electrodes is parallel to the material surface. The problem of breakdown and light loss can be greatly reduced by depositing a layer of light-transmitting material with high dielectric strength between the waveguide region and the electrodes. Silicon dioxide is a good example of such a material. Since breakdown can also occur above the surface between the two electrodes and in the air itself, a similar high dielectric strength material can also be deposited on top of the electrodes.

Figures 7 and 8 show two embodiments of electrically controlled frequency-selective waveguide couplers. In FIG. 7, a pair of two-dimensional waveguides pass through one side of a dielectric material and intersect each other at an angle 118 to form a T-shaped structure, forming a three-way device. A grating 100 consisting of a domain of the first type 104 and a domain of the second type 102 forms an angle between two waveguides where they intersect (the volume jointly occupied by the light modes in the two waveguides). The peak refractive index change in the intersection region is preferably equal to the peak refractive index change in the waveguide. This can be done if the fabrication of the T-shaped structure is done in one step (by means of in-diffusion, ion exchange, etching, etc.). In an alternative approach where the two waveguides are placed in a subsequent step, this is most convenient for making the intersecting waveguide of Figure 8, but the peak index change in the intersecting region is the index change in the waveguide twice, which is not needed. Usually, the periodicity and angle of the grating are chosen such that the phase matching of the reflection process is done by the virtual photon momentum within the bandwidth of the grating. For optimal coupling between the in-band input beam in the first waveguide and the output beam 114 in the second waveguide 108, the angle of incidence of the input beam is equal to the angle diffracted off the grating. At this point, the bisector of the angle between the two waveguides is perpendicular to the domain boundary of the grating in the plane of the waveguides.

A light beam 112 is incident on the first waveguide 106 and is coupled into the first waveguide 106 . First electrode 120 and second electrode 122 are disposed on the same surface of the dielectric material such that when a voltage source 124 connected by conductor 126 to both electrodes is turned on, an electric field is established in the intersecting region of the waveguides. The electric field couples the in-band beam from the first waveguide into the second waveguide to form a reflected output beam 114 by virtue of the electro-optic effect controlling the intensity of the grating in the intersection region. When the grating is switched off, the input beam continues to propagate predominantly down the first waveguide section, forming the emitted output beam 116 with little loss. Alternatively, a counterpropagating beam may be used in the waveguide, such that the input beam enters the second waveguide 108 and is switched into the output waveguide 106 by interaction with the grating.

In a single mode type system, the grating intensity is preferably distributed in a non-uniform manner in space such that the lowest order Gaussian wave entering waveguide 106 couples to the lowest order Gaussian wave of waveguide 108 . The grating intensity can be modulated by adjusting the electrode geometry, adjusting the electrode gap, and adjusting the duty cycle of the grating. The bandwidth of the grating can also be increased by one of many known techniques, such as chirping, phase shifting, and using multi-period structures.

The size of the coupling region is limited, in the geometries of Figures 7 and 8, by the size of the intersection region between the waveguides where their modes overlap. In order to obtain a high net interaction strength for a given electric field strength, it is necessary to oversize the waveguide to produce a larger intersection area. But large waveguides are multi-mode, which may not be suitable for some applications. If adiabatic expansion and contraction are used, the advantages of a large intersection area and a single-mode waveguide can be obtained simultaneously. The input waveguide 106 starts out as a narrow waveguide and adiabatically increases in width as the intersection region is approached. The output waveguide 108 has a large width in the intersection region to capture most of the reflected light, and it tapers adiabatically in width to a narrow waveguide. The concept of adiabatic tapering of the input and/or output waveguides can be used for many of the intersection regions described herein.

Referring to Figure 8, the two waveguides 136 and 138 meet at an angle 158 to create an X intersection, forming a four-way arrangement. This device is a multifunctional waveguide switcher because it produces two switching operations simultaneously (beam 142 entering beams 146 and 148, and beam 144 entering beams 148 and 146). A grating 130 composed of domains of the first type 134 and domains of the second type 132 is disposed in the intersection region between the two waveguides at an angle to the two waveguides. The angle of the grating is chosen such that the bisector of the angle between two waveguides in the plane of the waveguides is perpendicular to the boundary of the grating.

A first input beam 142 is incident on and coupled to the first waveguide 136 and a second input beam 144 is coupled to the second waveguide 138 . The first electrode 150 and the second electrode 152 are disposed on a dielectric material such that when a voltage source 154 connected between the two electrodes is turned on, an electric field is established in the intersecting region of the waveguides. The strength of the refractive index grating in the intersection region is controlled by the electric field through the electro-optical effect. When the grating is on, a portion of the in-band component of the first input beam is coupled from the first waveguide into the second waveguide to form a first output beam 146 . At the same time, a portion of the in-band component of the second input beam from the second waveguide is coupled into the first waveguide to form a second output beam 148 . In addition, the out-of-band components of the two beams, and any unswitched components of the in-band beams, continue to propagate down in their respective waveguides to form additional portions of the appropriate output beams. Thus, for two beams of multiple optical frequency components, a single frequency component in the two output beams can be switched between the two output beams.

The waveguide can be just one segment, in which case it is connected to other optical components either external to the substrate or integral with the same substrate. For example, waveguide segments can be connected to pump lasers, optical fibers, crossing waveguides, other switchable gratings, mirror devices, and other components. An array of crossed waveguide switches shall consist of an optical switching network.

In Figure 9, another embodiment of a waveguide coupled switch. The domain walls of the grating are now arranged non-perpendicular to the surface 157 of the crystal 158 so that the input beam 159 in the waveguide 160 is reflected out of the plane of the crystal to form a reflected output beam 161 . As before, the unreflected beam continues to propagate through the waveguide to form an emitted output beam 162 . A light transmissive first electrode, which may be composed of indium tin oxide, is placed on one surface of the dielectric material 158 and over the portion of the grating that spans the waveguide. A second electrode structure 164 - which may be light absorbing - is placed on the material. As described herein, the second electrode is arranged in any of a number of alternative configurations: around the first electrode as shown in FIG. 7 , on the opposite side of material 158 as shown in FIG. 2 , as shown in FIG. 6 Tapered construction. The two electrodes are connected by two wires 156 to a voltage source 154 which controls the power split ratio of the in-band beam between the transmitted beam 162 and the reflected beam 161 . Alternatively, the electrode configuration can be as shown in Figure 5, where both electrodes are opaque.

Referring again to FIG. 9 , domain walls are preferably formed by electric field polarization of a ferroelectric crystal cut at an angle to the z-axis 165 . Since the electric field polarizes domains preferentially down the z-axis, polarizing an angle-cut crystal with this technique results in boundaries parallel to the z-axis at the same angle as the plane. The crystal cut angle 166 is preferably 45° so that the propagation of light in the plane of the crystal can be reflected out of the substrate perpendicular to the material surface (any angle can be used). The domains shown in Figure 9 are planar, but can be structured as more general structures. A planar grating will produce a flat output phase front from a flat input phase front. If the device shown is used as a large reflector without a waveguide, a collimated input beam will produce a collimated output beam. Such devices are useful as bulk reflectors, for example if the beam is incident on the outside of the device, or if the waveguide is at one end of the device and there is a distance between the end of the waveguide and the polarized reflector. However, since in some applications focusing is required - for example reading data from an optical disc - it is preferable to be able to produce a curved output phase front from a collimated beam. By patterning, a set of curved domains is formed on the upper surface of the substrate as shown in Figure 9, and since domain inversion preferentially propagates along the z-axis, a set of curved domains can be polarized into the bulk of the material. Thus, when excited by an electric field, a set of concave (or convex) domains can be formed to create a cylindrical lens. Wedge-shaped and more complex three-dimensional structures intersecting the surface at an angle can be formed in the same way.

In an alternative approach, Z-cut crystals can be used as substrates if the poling technique allows domain boundaries to propagate at an angle to the z-axis. For example, inward diffusion of titanium (Ti) into a Z-cut crystal of lithium niobate creates triangular domains that can be used to reflect light beams away from the crystal surface. The angle of the domains formed by inward diffusion with respect to the surface is typically about 30° such that an input beam incident on the grating will be reflected off the surface at an angle of about 60° to the crystal surface. The output beam can then be extracted with a prism, or output from the rear surface (which can be polished at an angle) after total internal reflection from the top surface.

The electrode configuration shown excites an _E3 component, and either an _E1 or an _E2 component. A TM polarized input wave undergoes a refractive index change that is a combination of extraordinary and ordinary refractive index changes.

In FIG. 10, an embodiment of a switchable waveguide directional coupler is shown. A first waveguide 204 is substantially parallel to a second waveguide 206 over a certain length. As the beams propagate close to each other and in similar directions, their central axes are shifted. The central axes are never coaxial so that the waveguides do not intersect. However, the waveguide segments are fairly close together at the positions defined by the coupler length, so that the lateral profiles of the optical modes of the two waveguides more or less coincide. The propagation of the two modes is thus at least evanescently coupled (meaning that the refractive index tails coincide). The evanescent portion of the mode field is an exponentially decaying portion outside the high index region of the waveguide. The propagation constants associated with the modes of each of the two waveguides are determined by the equation k = 2πn _eff /λ for the direction of propagation. The effective refractive index n _eff is the ratio of the speed of light in vacuum to the group velocity of propagation, which varies according to the mode in the waveguide. The value of n _eff is determined by the overlap of the mode distribution with the guided wave structure.

The widths of the two waveguides are preferably different, and thus the propagation constants of the modes in the two waveguides are also different, so that when the grating is cut, the coupling between the modes is not phase matched. (The refractive profiles of the two waveguides can also be tuned to produce different propagation constants.) When the grating is switched off, any input beam 210 in the first waveguide will continue to propagate in the waveguide to form an output first waveguide 204 The output beam 214 of . When the grating is turned on, the grating compensates for the difference in the propagation constants of the two waveguides such that the coupling between the two modes is phase matched and outputs an in-band output beam 212 out of the second waveguide 206 . To optimize the coupling, the grating period Λ is chosen such that the propagation constant in both waveguides is equal to the grating constant (within tolerance). The propagation constants of the two waveguides can be chosen to be equal so that coupling occurs between the two waveguides when the grating is cut. In this case, turning on the grating will reduce the coupling between the two waveguides.

The strength of the grating determines the coupling constant, which defines the degree of coupling between the two waveguides. Along the length of the interface region of the two waveguides, the power propagates back and forth between the waveguides as a sinusoid, so that the coupling is first from the first waveguide to the second waveguide and then back to the first waveguide. The distance between the two locations of maximum power in a particular waveguide mode is known as the best length of the coupled waveguide. The length of this pulse depends on the strength of the grating.

A first electrode 220 and a second electrode 222 are disposed on the surface of the material such that when an electric field is applied between the two electrodes, an electric field is established across the grating region 202 . A voltage source 226 is connected to both electrodes by means of a conductive material 224 . The strength of the grating, and thus the pulse length between the two waveguides, is controlled by the voltage applied to the grating.

The propagation constants of the two waveguides are very wavelength dependent. Since the momentum of the virtual photon is substantially or predominantly fixed (i.e., it is determined by a parameter that does not change in the application), power is propagated into the second waveguide only around a single frequency, and the bandwidth of the single frequency is Determined by the length of the coupling region. Depending on the grating strength, an adjustable portion of the in-band input beam exits the second waveguide as a coupled output beam 212, while an out-of-band portion of the input beam emerges from the first waveguide as a transmitted output beam 214 along with the remaining portion of the in-band beam. output.

The coupling between two modes can be electro-optically controlled in several ways, including the strength of the coupling between the modes, increasing the overlap of the modes, or changing the effective refractive index of one of the waveguides. Electro-optic controlled coupling as described above is the preferred method. For efficient coupling between the modes in the two waveguides, the input beam is forward scattered, which requires a minimum grating period.

Alternatively, the coupling grating may be a combination of permanent and switchable gratings, as described above in connection with FIG. 2 . Here we detail how to accomplish this. After forming the desired periodic domains, the substrate can be chemically etched to form a concave-convex grating whose period is exactly the same as the structure to be poled. For the preferred lithium niobate material, etching can be done without further masking steps because different types of domains etch at different rates. For example, hydrofluoric acid (HF) etches the -z domains of lithium niobate significantly (>100x) faster than the +z domains. Thus, by immersing a z-cut crystal in a 50% HF solution, regions composed of domains of the first type are etched while regions composed of domains of the second type remain substantially unetched. This step produces a permanently coupled grating, which itself can be used to create a coupling between the two waveguides. After application of the electrodes, the polarized grating can be actuated to produce an additive index of refraction grating superimposed on the grating of the etched substrate. The depth of etching can be controlled so that the effective refractive index change induced by permanently etching the grating can be partially or fully offset by the electro-optic induced grating when the electrode is excited in one polarity, while the index grating is multiplied in the other polarization. This results in a push-pull light barrier, whereby the light barrier can be switched between an inactive state and a strongly active state.

Etched gratings are also useful when the etched region is filled with an electro-optic material, such as a polymer or a light-transmitting liquid crystal, that has a high electro-optic coefficient and a refractive index close to that of the substrate. The filled etch region preferably extends down to the beam. When a voltage is applied across the fill etched region, the refractive index of the fill material also changes around that of the remainder of the waveguide.

Alternatively, the mode overlap in the two waveguides can be modified electro-optically. For example, the refractive index of the region between two waveguides can be increased. This reduces the confinement of the waveguide and expands the amount of space for the individual modes to face each other, increasing the overlap. To implement this method, the region between the two waveguides can be polarized oppositely with respect to the polarity of the substrate transverse to the waveguides. If the electrodes extend across both waveguides and their intermediate regions, an applied voltage will increase the refractive index in the region between the waveguides while decreasing the refractive index in both waveguides. A consequence of the reduced mode limitation is increased overlap and coupling between the two modes. Care must be taken not to induce unwanted reflection or mode coupling losses in the waveguide, which can occur at the edges of the polarized regions. These losses can be minimized, for example, by tapering the polarization region or electrode geometry so that the mode change occurs adiabatically along the waveguide to minimize reflections. By adiabatic change is meant a very slow change compared to the equilibrium maintenance process that occurs at a finite rate. In this context, it means that the change is slow compared to the rate of energy redistribution due to diffraction within the waveguide and maintains the wave-mode characteristics of the light in the waveguide.

A third way to vary the coupling between two waveguides is to vary the effective index of refraction of one waveguide relative to the other. Therefore, the propagation constant of the waveguide is changed, which in turn changes the phase matching condition. This effect can be maximized by polarizing one of the waveguides so that the sign of its electro-optical coefficient is opposite to that of the other waveguide. In this case, the coupling grating can be a permanent or switchable grating. A first electrode covers the two waveguides and the area between them, and the second electrode can be arranged on both sides of the first electrode. An electric field applied between two electrodes causes the propagation constant of one waveguide to increase and that of the other to decrease, thus maximizing the difference between the two propagation constants. The grating coupling process has maximum efficiency when the difference in propagation constant is a certain value. By tuning the applied voltage, the phase matching can be adjusted as desired. This effect can be used to create a wavelength tunable filter.

The parallel waveguides shown in Figure 10 may not be parallel, and the waveguides may not even be straight. For example, if it is desired to vary the strength of the spatial interaction between the waveguides, this can be accomplished by adjusting the spatial separation between the waveguides. Of course, these modifications are applicable to the parallel waveguide embodiments described below.

Referring to Figures 12 and 13, there are shown alternative embodiments of crossed waveguide couplers for controlling reflected beam distribution. In each embodiment, the area covered by the grating does not extend completely across the interface region of the two waveguides. See Figure 11 for the motivation of these grating structures. Depending on how it is configured, the power coupling structure 282 can distort the spatial distribution of the modes 284 that it couples into the output waveguide. A power coupler that is spatially uniform and uniformly completely covers the intersection region 280 between two waveguides disposed at a large included angle, such as 90°, will produce an output beam distribution, such as an asymmetric distribution 286 . As the input beam passes through a power coupling structure or grating, the power in the input beam is reduced. At right-angle intersections, the near field profile of the reflected beam matches the monotonically decreasing power in the input beam. A disadvantage of the asymmetric distribution 286 is a single mode structure where only part of the coupled power will remain in the waveguide. Most of the power will be lost from the waveguide.

For a single mode device, there needs to be a structure to couple power into the Gaussian spatial configuration 288 of the lowest order mode of the output waveguide. To achieve this, region 282 must extend outward into the evanescent tail of the guided wave mode, and must modulate - either shape modulation or spatially modulate the local intensity of the power-coupled grating - the net interaction. Figures 12 and 13 show how the geometrical arrangement of the gratings can be used for this purpose. It is also possible to spatially modulate the "duty cycle" of the grating in the power coupling region 282, by varying the order of the grating in the selected region, and in the case of electrically controlled coupling, by tapering the strength of the applied electric field (e.g. This is accomplished by adjusting the spacing of the electrodes as shown in Figure 6, or by adjusting the electrode duty cycle in the case of a grating electrode structure). The duty cycle of a grating is the fraction of each period occupied by a given domain type; the duty cycle can vary with position.

In FIG. 12 , a device 300 with a modified grating is shown, wherein the grating region 310 covers part, but not all, of the rectangular interface region of two orthogonal waveguides 316 and 318 . When the grating is inactive, input beam 302 passes through waveguide 316 undeflected for output as output beam 308 . The dimensions of the interface region match the widths 304 and 305 of the two waveguides. The presence of a small area of power coupling structure at any point in the interface region will cause local coupling of a given transverse segment of the beam profile in an input waveguide into a given transverse segment of the beam profile in an output waveguide. The reflected beam distribution is constructed from the propagated sum of these phase-coupled distributions. The grating area 310 is triangular in shape with triangular vertices 311 , 312 and 313 . The grating area shape can be changed to something other than triangular, and the local grating intensity can be modulated. The precise shape of the grating regions that optimize the single-mode coupling characteristics between the waveguides is calculated using established waveguide propagation techniques, such as beam propagation methods.

FIG. 13 illustrates yet another embodiment of a single-mode coupled grating device 340 . Grating region 350 is a biconvex shape in which a point at corner 351 is shared with waveguides 346 and 348 and beams 330 and 342 and a point at the opposite corner 352 is shared with waveguides and beams 342 and 332 . The advantage of this configuration is that most of the power is reflected in the middle of the beam, where the light intensity is highest, thus better coupling the power between the lowest order modes in the two waveguides 346 and 348 . The preferred shape of the grating region is also determined by the grating coupling constant.

Referring to Figures 12 and 13, a first electrode 320 is disposed on the same surface of the substrate as the waveguide, above the grating region, and a second electrode 322 is disposed around the first electrode on the same surface. The distance between the two electrodes is constant as shown in FIG. 13 , or can be gradually reduced along one direction as shown in FIG. 12 . A voltage control source 324 is connected by two wires 326 to the two electrodes. An electric field can thus be applied across the grating region to activate one of the electro-optic coefficients and change the coupling between the input and output beams.

Figure 12 also shows a tapered input waveguide section 287 and a tapered output section 289 for purposes of illustration. An input beam of light 285 is adiabatically expanded through tapered section 287 to increase the interface area and thus the total reflection from grating 310 . The grating may reflect the now expanded beam 285 towards the output beam 308 . If desired, the output waveguide may also include a tapered section 289 to reduce the width of the output beam. (Alternatively, the width of the output beam can be maintained for later beam switching interactions, if desired.)

The grating may extend beyond the interface region of the two waveguides. A grating extending along the input waveguide allows residual emission behind the interface region to be removed from the waveguide, typically into the radiation mode. The extended grating minimizes crosstalk between optical channels in a switched array, where a single waveguide can have more than one signal channel propagating along its length.

A particular concept of the invention is a device for tuning a grating. Several embodiments in which such tuning can be accomplished are shown in Figures 14-17. Referring to FIG. 14 , a bulk light device 400 in which the intensity and central wavelength of a normal incidence reflective grating is controlled by a single voltage source 426 is shown. This device consists of a patterned polarization grating electro-optically activated by two electrodes on opposite surfaces of the material and connected by conductor 424 to a voltage source 426 . The intensity and central frequency of the grating are tuned simultaneously by a single voltage applied between the two electrodes of the device. The average refractive index of the grating varies with the applied electric field, causing the central wavelength of the grating to change proportionally to the electric field. In a periodic grating, the Average index of refraction over a single grating period. The weighting factors are the physical lengths 416 and 418 of each type of domain along the optical path of the input beam 404 . The condition for frequency modulation is that the weighted sum is not equal to zero, so that the average refractive index change is the result of the electric field change.

The product of the index of refraction and the actual distance traveled by the beam is called the optical path length. (For waveguide devices, substitute effective index of refraction for index of refraction). A 50% duty cycle is obtained in a grating with two types of domains if the average optical path lengths across the two types of domains are substantially equal (approximately equal within error limits determined by application requirements). The average is taken over many subsequent domains to allow for the possibility of chirped, aperiodic, or other more general types of gratings. In general, domains can have different indices of refraction as well as different electro-optical coefficients. General conditions for tuning are expressed in terms of actual distance traveled in different types of domains. For each domain, the all-optical phase advance is given by the optical path traveled by the light (multiplied by 2π/λ). However, the change in phase advance is given by the product of the applied electric field, the appropriate electro-optic coefficient, and the actual distance (by 2π/λ). The average change in refractive index experienced by this wave is equal to the sum of the phase advance changes in all domains that the light wave passes through in a segment of material length 1 (times λ/2π1). The average refractive index change determines the change in peak interaction wavelength according to δλ/λ=δn/n. The grating intensity varies simultaneously with the wavelength in the structure, but this simultaneous variation may be undesirable. The structure is designed such that the operating point at which tuning is done maintains a sufficiently high grating strength for applications spanning the entire wavelength tuning range. Alternatively, a separate tuning structure as described below with reference to Figures 16 and 17 may be used.

The change in average refractive index can be achieved by different methods. One way is to randomly arrange non-electro-optic domains 414 and electro-optic domains 412 alternately. Electro-optic active regions are polarized domains, while non-electro-optic active domains can be randomly polarized or non-polarized or radiation-forbidden. Thus, the electric field causes the average value of the refractive index across the grating to increase Δn _avg . In the randomly polarized structure of FIG. 14, Δn _avg is equal to the product of the refractive index change in the active domain 412 and the duty cycle. The duty cycle is equal to length 418 divided by the sum of lengths 418 and 416 . The tunability achieved using this technique in a randomly polarized structure is λΔn _avg /n, where λ is the wavelength of light and n is the intrinsic (effective) index of refraction of the material. Assuming a wavelength of 1.55 μm in lithium niobate and an electric field of 10 V/μm, the tuning range is 1.1 nm for a 50% duty cycle structure.

The grating couples the beam into the retroreflected output beam 402 when the input beam 404 is within the bandwidth of the grating; otherwise, the input beam forms a transmitted output beam 406 . If the two types of domains have the same electro-optic coefficient but opposite polarity, as in the case of domain inversion, then the behavior of the grating with a 50% duty cycle is reversed. In the latter case, the average refractive index is not changed since the change in the refractive index of the domains of the first type is offset by the change in the refractive index of the domains of the other type. The inverse grating of the 50% duty cycle domain does not tune its center frequency.

Another way to achieve average effective index change in a domain inversion grating is to use a non-50% duty cycle of unequal length 416≠418 polarized domains. The tunability obtained using this technique is (2D-1)Δnλ/n, where D is the duty cycle of the largest domain type (D > 0.5). For example, for a 75% duty cycle, a wavelength of 1.55 μm in lithium niobate, and an electric field of 10 V/μm, the tuning range is 0.54 nm. Domain inversion gratings are also stronger than the second type of gratings in which the domains are non-electro-optic.

Figure 15 shows a waveguide device 440 using the same average refractive index effect. In this case, the average effective index of refraction of the waveguide 442 changes in the grating region 450 as a function of the applied electric field causing the central wavelength of the grating to change. A voltage control source 466 is used to apply an electric field between the first electrode 460 and the second electrode 462, preferably disposed on the same surface of the material. The average effective index can be achieved by various geometries including non-electro-optic domains or domain inversion gratings with a 50% duty cycle. When the input beam 455 is within the bandwidth of the grating, the grating couples the beam into the retroreflected output beam 444; otherwise, the input beam forms a transmitted output beam 446.

One way to enhance the tunability of the grating in the waveguide device 480 is to overlay the waveguide with a second electro-optic material 482 to form a cladding layer, as shown in FIG. 16 . The cladding should be transparent to waves propagating in the waveguide and should be sensitive to electric fields so that its refractive index can be tuned. The average effective refractive index is determined in part by the refractive index of the cover layer. The second material may have a higher electro-optical coefficient than the substrate. Liquid crystals and polymers are good examples of cover layer materials. The refractive index of the cladding is preferably close to that of the guiding region to guide most of the light beam propagating in the cladding.

For this embodiment, the first electrode 502 is surrounded on the substrate by a second electrode 504 for applying an electric field across the polarization grating 490 . The electrodes are preferably arranged directly below the cover layer on the substrate. If the first electrode 502 is placed directly above the waveguide 484, as shown in FIG. 16, it must be made of a light-transmitting material. The electrodes may also be disposed on either side of the waveguide 484, in which case the material is not required to be transparent. A third electrode 506 is disposed on top of the cover layer, above the waveguide and the first electrode. For this embodiment, the center wavelength and intensity of the grating can be controlled separately. The grating intensity is controlled by a first voltage source 510 connected to the first and second electrodes with two wires 513, 514, and a second voltage source 512 connected between the first and second electrodes with two wires 514, 515. Controls the center wavelength of the grating. In another electrode structure, only two electrodes are used, both electrodes are preferably placed on top of the cover material so that their sensing fields are transmitted through the cover material on the grating, and the grating structure itself. Control the center wavelength and grating intensity with one voltage source instead of independently.

The amount of tunability achieved with an electro-optic coating depends on which part of the guided beam propagates in the coating. If the two indices are relatively close such that 10% of the beam propagates in the cladding, the average change in the effective index of the guided mode is then equal to 10% of the change in the cladding index. The tunability is on the order of 7nm for a change in the refractive index of the capping layer of 0.1.

FIG. 17 shows an embodiment of an individually tunable grating device 520 consisting of independently controllable gratings 530 , 532 , 534 . This grating is in series with all gratings in the path of the input beam 522 , the forward beam 523 and the reflected beam 524 . Each independent grating in this structure is also continuously tuneable within a small range. Each grating in FIG. 17 has a first electrode 542 and a second electrode 544 wired to a voltage control network 552 . One grating can be switched on at a time to reflect only one wavelength at a time in a small passband, or multiple gratings can be switched on simultaneously to produce programmable optical filtering where the center wavelength and bandwidth can be controlled separately device. The gratings themselves can be implemented with the above mentioned improvements including the possibility of multiperiods in each grating.

The structure can be realized in bulk or as a catheter device. In the latter case, an optical waveguide 528 is fabricated on the substrate such that the waveguide intersects the polarization grating. The grating domains 536 may only extend through the waveguide and need not extend through all directions of the material. The two electrodes are preferably (for higher field strengths) deposited on the same surface of the substrate as the waveguide. The second electrodes of all gratings can be connected as shown to minimize the number of electrical connections.

Alternatively, the stand-alone addressable grating structure could be a bulk device, in which case waveguide 528 could be omitted, and polarization regions 530, 532, and 534 suitably fabricated with sufficient depth to overlap in the propagating optical modes. The two electrodes controlling each grating are suitably positioned at preferred locations on opposite sides of the material so that transmission of the electric field is preferred, as shown in the example of a single grating in FIG. 12 . Cross-excitation between adjacent gratings caused by electric field edges between electrodes can be minimized by spacing the grating-electrode groups to a range comparable to the thickness of the substrate, or by adding alternating fixed potential electrodes.

Another way to tune the grating is to vary the temperature of the active material. The tuning occurs due to two effects: thermal expansion and the thermo-optic effect. For different materials, either of the two effects may govern thermal induction tuning. In lithium niobate, the larger effect is thermal expansion, with a maximum (a-axis) expansion coefficient ΔL/L of +14×10 ^-6 ℃ ^-1 and a thermo-optic coefficient of ordinary axis Δn ₀ /n of +5.6 ×10 ^-6 °C ^-1 . The combination of these two effects gives a total wavelength tuning range of 2.6 nm for a temperature range of 100°C.

For many purposes it is desirable to generate polarized gratings with generalized frequency content. Examples include the need for multiple interaction peaks, or simply broadening the bandwidth of an interaction. To achieve this, some way is required to determine the figure corresponding to the boundary of the polarization region given the mathematical function containing the desired frequency. Figure 18 shows the processing results in the case of a single frequency involving arbitrary phase shifts. Referring to FIG. 18, optical phase shifts 564 and 565 may be combined at one or more locations along the sinusoidal function 560 to improve its wavelength structure. The average level of this function is given by straight line 561 . Also shown is a corresponding square wave function 562 with the same phase shift, as achievable by general polarization processing. To achieve the conversion from a continuous function to a square wave function, the region 570 where the curve 560 exceeds the mean value 561 of the sine wave corresponds to one type of domain, and the region 572 where the curve 560 falls below the mean value 561 corresponds to a second type of domain. The Fourier transform of the square wave curve 562 will have the same frequency components as the transform of the sinusoidal function 560 in the low frequency range below the harmonics of the sine wave frequency. The method works with any type of generalized frequency distribution as long as the bandwidth does not exceed a fraction of the carrier frequency.

Where a phase shift grating can be implemented with, for example, any device such as that described in FIG. 2, the position of the domain walls 34 in the grating 22 can be determined by the graph 562 of FIG. 18 rather than by a periodic function. The phase shift pattern can be controlled with a polarization mask that introduces the desired pattern.

Any number of periodic gratings can be specified using a similar technique. Each period present in the grating is represented in a Fourier series (or integral) by a corresponding sine wave of desired amplitude. All waves add together to form a synthetic wave. The positive-valued portion of the synthesized wave corresponds to one type of domain, while its negative-valued portion corresponds to the second type of domain. The number of stacked gratings can in principle be increased to any number, but in practice is limited by the smallest achievable size.

FIG. 19 shows another method of fabricating a stacked multi-period grating device 580 . Therein a two-grating waveguide structure is disclosed, in which there is a switchable mono-periodically polarized grating 582 and a permanently released grating 584 which interacts with a single beam in the waveguide. A coating 588 is deposited on top of the release grating to reduce losses that occur when the evanescent tail of the guided wave mode overlaps the metal electrode. This coating is an important design optimization for all components described here and should be applied between each electrode structure and the adjacent optical waveguide. A coating may also be used on the electrodes in all elements described here to reduce the possibility of breakdown.

In a superperiodic configuration, an electrically controllable grating is switched by a pair of electrodes 602 and 604 connected by a lead 606 to a voltage control source 608 . The first electrode 602 is preferably in the middle above the waveguide, while the second electrode 604 is on either side of the waveguide and parallel to the first electrode. The device is a waveguide device with a waveguide 586 that confines an input beam 590 and transmits an output beam 592 and reflects an output beam 594 .

Multi-period gratings can be constructed in many ways. For example, multiple independent peaks in the spectrum can be used as a multi-frequency feedback mirror. Both operations (such as phase matching and reflection) can be achieved in a single grating combining two appropriate periods of the activation process. As a final example, a grating can be fabricated by adjusting the phase and amplitude of its elements so that the two polarization modes are equivalent, making a polarization-insensitive element.

Another beneficial modification of the periodic structure is the chirp period. The period gradually increases or decreases along the length of the grating structure so that the center frequency changes from one end of the grating to the other. Therefore, the wavelength bandwidth of the grating is widened beyond that of a constant-period grating. A chirp across a grating is not necessarily linear: many different wavelength reflection distributions in frequency space (eg square wave, Lorentzian type) can be achieved depending on the chirp rate. As mentioned above, the duty cycle and/or strength of the excitation electric field can also be adjusted spatially to improve the strength of different parts of the chirped raster. The duty cycle of the grating can be controlled via a mask as required. The electric field strength was controlled by adjusting the electrode spacing as shown in the example in FIG. 6 .

Broad spectrum tunable devices can be realized in a split grating structure comprising two multi-peak structures, as shown in Figs. 21 and 22 . Figure 20 illustrates the basic principles of these devices and plots the multi-peak comb transmission (or reflection) distributions 620 and 622 as a function of the optical frequency of the two gratings. The first grating distribution 620 has transmission peaks separated by a first period 626, while the second grating distribution 622 has peaks separated by a second period 624, which differs slightly from the first period. The main idea of the device is to make the device operate only at frequencies determined from the superposition of the peaks of the two curves (frequency ν1). Tuning is achieved by comb teeth that tune the transmission peaks of the grating relative to each other. The different transmission peaks in the two combs will overlap each other at different ranges of relative frequency shift, so that the net transmission of the combined grating jumps individually to a much wider wavelength range than can be achieved by a single thermal or electro-optical tuning. wavelength range. In the example of FIG. 20, the peak spacing difference is 10%, if the frequency of the first grating increases by 10% of the frequency spacing 626, the next higher frequency peaks will overlap, resulting in an effective frequency shift that is ten times larger than the tuning amount.

A guided wave embodiment of the device is shown in FIG. 21 , where two gratings 650 and 652 are disposed on a single waveguide 642 . Output beam 644 is partially reflected to beam 643 and transmitted as beam 645 . A first electrode 666 and a second electrode 668 are disposed around the first grating 650 such that a first voltage source 662 connected to the electrodes triggers the grating. The third electrode 664 is disposed around the second grating 652 together with the second electrode. The second grating is controlled by a second voltage source 660 connected to the second and third electrodes. In the preferred embodiment, each grating is a multi-peak structure as described in Figure 20, and the arrangement forms a frequency hopping tuned reflector. According to the curves of Fig. 20, the grating is formed as a broadband reflector, mainly reflecting all frequencies of the incident radiation except the equally spaced frequency comb teeth which transmit higher. Thus, the cascaded grating will reflect all frequencies in the frequency range shown in Figure 20, except where the two transmission peaks at ν1 overlap. If the reflections of the two gratings are arranged so as to increase the phase in the reflected beam 643, the transmission spectrum will be substantially equal to the product of the two transmission curves 620 and 622. When tuning the center frequency of one of the gratings, a single transmission peak at ν1 will jump to the next adjacent peak, then the next, and so on. This structure is particularly useful as an electrically tuned receiver in, for example, wavelength division multiplexing (WDM) communication systems. The receiver can be made to detect only specific frequency bands of output light, while being insensitive to light of other frequencies.

As seen above, assuming an electric field of 10 V/μm in a domain inversion grating with a 75% duty cycle, the grating structure can be shifted by about 0.5 nm. If the width of a single frequency peak 628 is narrower than 1/100 of the frequency interval, this continuous tuning range can be used to produce discrete tuning in the 50 nm range in a structure 640 spanning perhaps 100 frequency bands.

It should be noted that the device can only be realized with a single grating structure having a transmission spectrum of curve 622, mainly using the Moire effect, if it is known that the frequency of the input light is, for example, only in the transmission band of curve 620 in Fig. 20 . By tuning the center frequency of spectrum 622, any one of the desired frequency bands can be selected while the rest are reflected. The T-shaped structure of FIG. 7 has its specific significance in this context: the input beam 112 containing multiple frequency components is divided by the grating structure 100 (configured for tuning as described) into components that can be detected or otherwise processed. A single transmitted beam 116, and a reflected beam 114 containing all other frequency components. The power contained in the light beam 114 is not lost, but can be routed, for example, to other nodes in a communication network.

Other variations can form this basic structure, where, for example, the spectrum of FIG. 20 is the reflection curve of a grating alone rather than the transmission curve. In this case, when the frequencies of the reflected polarizations are aligned with each other, the structure acts as an etalon by reflectivity according to the relative phase of the reflected waves. Otherwise, the net reflection of the composite structure is primarily the sum of the reflection curves of the two individual structures.

It may be important to optimize the relative phase of the two reflections by adjusting the optical path length 653 between the two gratings. The optical path length 653 can be adjusted by controlling the relative phase between the two grating entrances 654 and 655 using an electro-optic structure (as shown for example in FIG. 22 ). For a lithium niobate crystal and an input wavelength of 1.5 μm, the trigger distance between the gratings needs to be at least 250 μm to adjust the relative phase between the two beams up to ±π, (using the z-axis with an applied 10 V/μm electric field). If the grating is not designed for tuning (making its mean refractive index independent of the applied voltage), the strength of one of the gratings (rather than its frequency) is arbitrarily controlled via an electric field applied at its electrodes. If both gratings are tuned together, it will result in narrow range continuous tuning. As an alternative or supplement to electronic excitation, both the phase of the two reflections and the peak wavelength of the grating can be changed together by thermal or mechanical control of the chip.

Figure 22 schematically shows two grating reflectors 633 and 634 separated by a phase shifter section 635 and forming an integrated etalon 640 with a characteristic free spectral range (FSR). (Structure 630 is essentially the same as that in Figure 21, but with the addition of a phase shifter section consisting of electrodes capable of exciting regions of electro-optic material transverse to waveguide 636). For simplicity, we consider the case of uniform single-period gratings, but individual gratings are usually more complex structures. The grating can be fixed or electronically activated. By adjusting the voltage applied to phase shifter section 635, reflections off two gratings can be made to increase the phase of the beam at the reference frequency. If the frequencies of two gratings are separated by multiple FSRs, the beam at the second frequency will also increase in phase. Since FSR is inversely proportional to the optical path length between two gratings, the choice of optical path length determines the density of etalon reflection peak structures. For example, two short highly reflective gratings separated by 220 μm in lithium niobate can have grating reflection peaks separated by multiples of 1 nm. Each of the multi-peak structures 620 or 622 depicted in FIG. 20 may be implemented as an integrated etalon.

FIG. 23 shows a dual grating Y-junction embodiment in which two gratings 690 and 692 extend across two separate waveguides 682 and 684 . A Y-junction typically has one input and multiple output waveguides that may be in one plane or one space. The two waveguides are connected to the first waveguide 686 by a Y-junction 688 . The power in the optical input beam 691 is split between the second waveguide 682 and the third waveguide 684 such that approximately 50% of the input beam 691 is incident on each grating. Both gratings can have a simple reflective structure, or can have a series of highly reflective peaks. The grating may be permanent, or may also be electronically adjustable, in which case electrodes 694 and 696 are provided to actuate the grating. A common electrode 698 is then placed across the wafer (or alternatively on the same plane as the waveguides, similar to FIG. 21 adjacent to other electrodes).

The relative optical path lengths of the two branches of the waveguide can be adjusted by electrodes 689 disposed above the electro-optic active region. By adjusting the voltage on the phase adjustment electrode 689, the two counterpropagating reflected beams can be adjusted to have the same phase when they meet at the Y-junction. The reflected modes superimpose according to the relative phases of the two waves and form a wavefront profile with a phase discontinuity in the center. As the combined wave propagates, the phase shift strongly affects the spatial concentration of the light modes in the guided region. If they have the same phase, their distribution forms a symmetric mode that effectively couples to the lowest order mode of the input waveguide to form the back-reflected output beam 693 . Adding the two reflected beams out of phase at the Y-junction will have very little coupling into the waveguide 686 of any symmetric modes (eg, the lowest order modes). If the waveguide 686 is a single mode, this reflected energy will be rejected from the waveguide. Thus, by adjusting the optical path length of one of the Y-arms with electrode 689, the reflection can be quickly adjusted from almost 100% to a value almost near zero. Furthermore, if the grating is implemented as an electronically tuned reflector in one of the tunable structures described here, the modulated reflection characteristics can be shifted into different spectral regions.

Referring to Figure 24, a switchable waveguide mode converter 720 using a polarization grating 722 is shown. The waveguide 730 preferably supports both input and output modes, which may be two transverse modes or two polarized modes (eg, TE and TM). The two modes in a waveguide typically have different propagation constants determined by the mode's effective refractive index. Grating 722 is electrically actuated by electrodes 740 and 742 that are coupled to a source of potential 744 by lines 746 . The grating period Λ (724) is chosen such that the magnitude of the difference in propagation constants in the two waveguides is equal to the grating constant 2πn/Λ. When the grating is turned on, the grating compensates for the difference in the propagation constants of the two waveguides so that the coupling phase between the two modes is matched. The grating strength and the interaction length of the device in the grating should be set so that the power flow is optimized from the input mode to the output mode. The net rate of power conversion from one mode to the other is determined by the strength of the electro-optical coefficient (γ ₅₁ in lithium niobate) and the electric field strength.

For two transverse modes, the coupling depends on the spatial superposition of the two modes and the grating strength in the presence of a grating structure. The two modes can be made orthogonal through symmetry, so that even though the modes are phase matched, there is no conversion in a symmetrical structure. In this case, the phase matching structure itself can be made asymmetric to eliminate this problem. In the preferred embodiment of Figure 24, the asymmetry can be introduced by an electric field that excites the polarized structure. The vertical component of the electric field reverses sign between the two electrodes 740 and 742 . The electrodes are preferably placed in the center of the waveguide so that mode switching between different symmetrical transverse modes is preferred. Its inversion is true when the same symmetrical transverse mode is coupled: the phase-matching structure should now be made symmetrical to make the transition preferable. Several different methods can also be used. A three-electrode structure has a symmetrical vertical component and an asymmetrical horizontal component of the electric field. This horizontal electric field can be used in conjunction with one of the electro-optic coefficients for horizontal coupling to couple different symmetric modes. Alternatively, the polarized structure may have an anti-phase plane that substantially bisects the waveguide, in which case a symmetrical component of the electric field may be used to couple the different symmetrical modes (vertical field in the three-electrode case, horizontal electric field in the two-electrode case).

Since the propagation constants of the two modes are strongly wavelength-dependent, the beat length of their interaction is also wavelength-dependent. Thus, for a given length of the coupling region between the two modes, the power coupled to the second mode is frequency sensitive. This coupling has a frequency bandwidth associated with it. For a given grating strength, a portion of the in-band input beam is coupled to an output mode that exits as a coupled output beam, while the remainder of the input beam exits the first waveguide as a transmitted output beam.

The structure shown in Figure 24 can also be used to couple between TE and TM polarization modes. For example, the electro-optical coefficient γ ₅₁ enables coupling between two orthogonal polarizations in lithium niobate crystals. As before, the grating period is chosen such that the grating constant is equal to the difference in propagation transmission between the two modes. The interaction length is chosen to optimize power transfer.

A waveguide such as an indiffused titanium waveguide supporting both TE and TM modes is used in applications where polarization can enter or exit the transducer. A waveguide such as a proton exchange waveguide that supports only one polarization (TM in z-cut lithium niobate substrate or TE in x-cut or y-cut) can be used in applications requiring only a single polarization. This single polarization waveguide can act as a very effective filter for other polarizations. The wrong polarization component dissipates rapidly from the waveguide due to diffraction, leaving only the guided polarization in the waveguide. For example, the proton exchange waveguide 731 can function to guide only the input polarization or the output polarization, as desired. If the grating coupling is strong and the interaction length and electric field are chosen correctly, the device can be used as a light modulator with excellent transmission and absorption properties. A modulator constructed with a proton-exchanged waveguide transmits essentially all correctly polarized input light and produces very low transmitted light coupled to vertically polarized modes. Alternatively, the input waveguide can be indiffused titanium to accept either polarization at the input. The refractive index profiles of the waveguides forming the two beams are preferably similar to allow good overlap of the profiles of the TE and TM modes and to maximize coupling efficiency.

To trigger the γ ₅₁ coefficient, an electric field is applied along the crystal Y or X axis. Electrode configuration to achieve proper electric field orientation depends on crystal cut. For a z-cut crystal with a waveguide oriented along the x-axis, the first and second electrodes can be placed on either side of the waveguide. On the other hand, for a y-cut crystal with a waveguide oriented along the x-axis, the first electrode can be placed directly above the waveguide, while the second electrode can be placed on either side of the waveguide, parallel to the first electrode.

Since the polarization domains in the grating 722 can extend through a solid substrate (eg 0.5mm or thicker), the structure of Figure 24 can also be used for a controllable bulk polarization converter. In this case, the waveguide 730 is not required and the electrodes are optimally formed on either side of the monolithic sheet of polarizing material.

Referring to FIG. 25 , a switched beam director 700 comprising a Y-shaped power splitter 702 and a transverse mode converter 704 is shown. The wave mode converter works in a similar manner to the transverse mode converter described above in relation to FIG. 24 . The grating structure 706 phase matches the energy conversion from the lowest order (symmetric) mode incident on the waveguide 708 to the next higher order (asymmetric) mode of the waveguide. The length and strength of the interaction region where the waveguide and grating structures overlap are chosen so as to convert approximately half the input power of a single symmetric mode into a higher order asymmetric mode. In addition, the optical path length between the grating mode converter section 704 and the Y-type phase splitter 702 is selected so that the phase of the two modes increases constructively in one of the Y-type branches 712 and destructively in the other branch 713. Increase. The result is that power is primarily routed to the constructively interfering waveguide 712 with very little power leaking into the other branch 713 . In this case, the counterpropagating power in waveguide 713 is substantially excluded from the counterpropagating mode in waveguide 708 after coupling to mode coupler 704 . The device forms an effective power router in the forward direction and an isolation structure in the reverse direction.

By adjusting the optical path length between the grating mode converter section 704 and the Y-type phase splitter 702, the output power can be switched from the waveguide 712 to the waveguide 713. This is done by adjusting the relative optical path lengths of the lowest and higher order modes so that the two modes are phase shifted by π relative to each other, thereby producing constructive interference in waveguide 713 and destructive interference in waveguide 712. By exciting the electrode pair 711 and 709 with a voltage source 714, the refractive index under the electrode 711 is changed through the electro-optic effect in the substrate 703 to achieve relative optical path length adjustment in the optical path length adjustment part 705. The substrate 703 is preferably niobium Lithium Oxide (but can be any electro-optic material that is transparent to waves such as Lithium Tantalate, KTP, GaAs, InP, _AgGaS2 , crystalline quartz, etc.). The propagation distance of waveguide 708 under electrode 711 is selected, along with the excitation voltage, so as to change the relative phase of the two modes by at least a desired amount.

Grating 706 may be a permanent grating fabricated by any technique known in the art. However, for optimal functioning of the device, it is desirable that the power in the symmetric and asymmetric modes be nearly exactly equal. It is difficult to achieve sufficient control to achieve this in existing manufacturing techniques, so some adjustments in the grating intensity are required. This adjustability can be achieved using at least some portions of the polarization grating excited by electrodes 709 and 710, which are driven by a power source 715 and which can themselves be used to perform the desired mode conversion, or to adjust the combined polarization permanent The strength of the raster.

The input waveguide 708 is preferably implemented as a single mode waveguide including a (preferably adiabatic) taper 701 to allow two modes to be guided between the transverse wave mode coupler 704 and the Y-type phase splitter 702 . Both waveguides 712 and 713 are preferably of the single mode type. Other modes can also be used in this device as long as their symmetry is reversed. For interconnection purposes it is most desirable to work with the lowest order modes at the input and output leads. The intermediate excitation mode is of less importance and could be, for example, a higher asymmetric mode.

FIG. 26 shows a parallel waveguide switchable resonator 750 where an input waveguide 752 is coupled to a parallel waveguide 754 along an interaction region 753 . Grating reflectors 755 and 756 are arranged across waveguide 754 in such a way that the retroreflected light propagates in the waveguide. The pair of spaced reflectors and waveguide 754 form an integrated etalon coupled to input waveguide 752 . The length of the coupling region 753 and the spacing of the parallel waveguides in the coupling region are selected so as to couple a desired fraction T of the input beam 757 into the waveguide 754 . Light coupled to etalon structures 754 , 755 and 756 resonates between reflectors 755 and 756 and couples out to two main output channels: forward propagating wave 759 and counter propagating wave 758 in waveguide 752 . The same fraction of T power circulating in the etalon is coupled to each of the two output channels 758 and 759 .

As with any etalon, an integrated etalon has a frequency-receiving structure with a width dependent on resonator losses, including multiple peaks within frequency control, and spaced equal to the free spectral range. If the optical frequency of the input beam 757 matches one of these resonant frequencies, the power circulating in the etalon will increase to a value P _circ determined by P _cir =P _inc T/(T+Γ/2) ² , where P _inc is the incident power in the waveguide 752, Γ is the etalon loss, excluding the output of the forward propagating wave 759 and the counter propagating wave 758 coupled into the waveguide 752, and assumes weak coupling and low loss. The output coupled wave from the etalon propagating in the waveguide 752 forms a reflected wave 758 . The reflected power in beam 758 is equal to _Pref =P _inc /(1+Γ/2T) ² at the resonance peak. When T>Γ/2, all incident power is basically reflected. The output coupled wave from the etalon propagating forward in waveguide 752 is out of phase (at the cavity resonance) with the uncoupled portion of input wave 757 , and the two beams interfere destructively, producing low amplitude output beam 759 . Due to the unequal amplitude of the two beams, the residual power P _trans =P _inc /(1+2T/Γ) ² in the output beam 759 is not equal to zero, but can be very close to zero. If the coupling Γ is made very large compared to the etalon loss, the transmission of the device is greatly suppressed (26dB for T=10Γ). The structure then acts as a very low loss reflector at the frequency comb teeth separated by the FSR.

The device is switchable by changing the optical path length between the two reflectors 755 and 756 . Electrodes 761 and 762 are arranged to generate an electric field across waveguide 754 between mirrors 755 and 756 . The electrodes are excited by the power supply 763, and the effective refractive index of the substrate under the electrodes 761 is changed by the electro-optical effect, thereby changing the optical path length between the mirrors and shifting the resonance of the integrated etalon. If the resonance shift is greater than one of the resonance width or the bandwidth of the incident beam, the reflection will drop to zero and the transmission will rise to essentially 100% as the circulating power in the etalon is suppressed to about _Pinc T/4.

As shown in the previous figures and description, gratings 755 and 756 may be permanent gratings, or they may be polarized gratings excited by electrodes. If the grating 756 is a polarizing grating, the device can be switched by switching off the grating. With the grating 756 switched off, ie not reflected, the loss of the incident wave 757 is equal to the coupling constant T, but now the comb structure is eliminated and only frequency shifted by the electrode 761. The difference in the switching function between two operating modes can be significant, for example for wideband input signals, where reflections must be cut off rather than just changed in frequency. For a single frequency input beam, the reflection can be switched uniformly by changing the optical path length with electrode 761 or by breaking the Q of the resonator by cutting mirror 756 . However, if the reflectivity of mirror 756 is maintained and only the spectrum of the etalon is shifted with electrodes 761, other frequency components of the broadband input wave will be reflected, which may be highly undesirable in some applications.

If T and Γ are small, the power P _circ built up in the etalon can be quite large and can be used, for example, in applications such as second harmonic generation. In this application, a quasi-phase-matched (QPM) periodic poling structure in a portion of the lithium niobate substrate is introduced between the mirror 756 and the interaction region 753, or possibly in the interaction region itself. One of the resonant frequencies of the standard is then tuned to coincide with the phase matching frequency of the QPM doubler. The resulting power build-up enhances the frequency conversion efficiency of the device due to the square of the settling factor P _circ /P _inc . High reflections occurring at this frequency can also be used to injection lock the pump laser at the desired frequency if the FSR is large enough that other resonant modes are not simultaneously injection locked. The geometry of the linear integrating etalon described above with reference to Figures 21 and 22 can also be used to accomplish the same purpose.

To optimize the power build-up between reflectors 755 and 756 in the etalon, losses in the resonator must be minimized. The coupling of Fig. 26 cannot be "impedance matched", similar to the process known in the art to build up a cavity in its entirety, adjusting the input coupled to the resonator by destructively decoupling the portion of the incident beam that is not coupled into the cavity Interfere and offset. This is the condition for the etalon to transmit the interference peak. As mentioned above, what happens in the integrated structure is that the transmitted beam is almost canceled out, while at the same time power is built up in the coupled resonator, but a strong reflected wave occurs. This reflected wave can be canceled in the ring waveguide structure, as illustrated in FIGS. 27 and 28 .

The output 751, which is proportional to the power cycle in the etalon, can be passed through grating 756, or alternatively through grating 755, if desired.

Figure 27 shows a three-arm etalon 760 comprising an input waveguide 752, a parallel waveguide coupling region 753, a ring resonator consisting of three waveguide segments 764, 765 and 766, three Grating reflectors 767, 768 and 769. The optical path length adjustment section formed between the electrodes 761 and 762 is optional. The grating reflector 767 is arranged to preferentially reflect incoming power from the waveguide 764 to the waveguide 765 . In a single mode system, the grating (and its electrodes, if any) are spatially structured such that the lowest order mode from waveguide 764 is coupled to the lowest order mode of waveguide 765. Gratings 768 and 769 are similar in structure so that power is preferentially passed from waveguide 765 to waveguide 766 and then to waveguide 764, forming a Fabry -Perot resonators. Impedance matching is now possible and achieved when the coupling coefficient T is equal to the total travel loss coefficient of the resonator less than the output coupling loss, mainly in the coupling region 753 . If the phase-matching frequency doubler is placed in the resonator, the converted power outside the fundamental frequency beam circulating in the resonator is not counted as a loss in the total travel loss.

If the incident beam 757 is incident on the device at a frequency equal to one of the resonances of the three-arm etalon, power will couple to the etalon across the parallel waveguide interaction region and establish a circulating power P _circ =P _inc T/(T+Γ ) ² . Power circulates mainly in one direction from waveguide 764 to waveguides 765, 766 and back to 764 due to the ring structure. There is only one single coupled wave from the etalon to the waveguide 752 and it is propagating in the forward direction. The output coupled wave interferes destructively with the remainder of the input wave 757 to form a weakly transmitted wave 759 . The transmitted power P _trans in the output beam 759 is given by P _trans =P _inc (1-Γ/T) ² /(1+Γ/T) ² , which is zero if Γ=T, as an impedance matching condition. In this case, all incident power flows into the resonator. In the impedance-matched state, the amplitudes of the two beams are equal and the transmitted power is reduced to zero. There is essentially no reflected power in beam 758 except from reflections from waveguide 752 discontinuities, which can be minimized by good design.

The grating 767 or any other grating can constitute a switchable grating, in which case the quality Q of the etalon can be destroyed by switching off the grating, eliminating the entire comb structure but leaving some optical loss due to power coupled to the waveguide 764 . Output beam 751 may be directed to transmit through grating 768 , and/or through grating 767 or 769 .

FIG. 28 shows a ring waveguide etalon 770 . As before, input waveguide 752 is coupled to waveguide 772 at parallel interaction region 753 . Interaction region 753 includes a grating in FIG. 28 (although not required) to emphasize that grating coupling is a useful option in the etalon geometries of FIGS. 26, 27 and 28. The waveguide 772 feeds back a portion of the power emerging from the section 753 to the interaction region 753 following a curved closed path (of any geometry, including potential multiple loops with intersections). As before, electrodes 761 and 773 are provided to allow adjustment of the optical path length and thus the FSR, although in this case they are disposed on the same side of the substrate. Flat areas 771 are provided in which certain reusable functional elements can be fabricated depending on the application of the etalon structure. If the etalon device 770 is used for frequency doubling, it is helpful to insert the frequency doubling structure into a flat region, eg portion 771 of the ring, but provision must be made to couple the frequency converting light out of the ring waveguide.

Device 770 otherwise functions similarly to device 760 . While device 760 may occupy less surface area on the substrate, device 770 is less lossy in etalons, especially if the diameter is 1 cm or larger.

Devices 760 and 770 can function to create a frequency doubling cavity in which feedback to the light source is minimized. They can also switch the transmission of a given frequency without back reflection for use in applications including optical communications.

In WDM communications, a single fiber can carry many communication channels separated by its optical wavelengths. In order to detect channels, the light in the desired wavelength region must first be separated from the remaining channels routed to other destinations. This separation function is performed by a channel dropping filter. Channel drop filters are communication devices used in wavelength division multiplexing (WDM) environments. There is a need to multiplex channels across a single transmission fiber by carrying the channels on different wavelengths. A reused element in this system is a channel drop filter that allows individual channels to be decimated for routing or detection purposes. An ideal fiber draws essentially all the light in the channel with a good extinction ratio so that the same wavelength can be used in the subsequent network without crosstalk. Since multiple channel drop filters can be installed on any given line, the insertion loss of out-of-band components must be low. It is preferably switchable so that the channel can go offline at the destination, and after the communication is over, the channel can continue through the location to another destination. Corresponding to the channel down-line filter is the channel up-line filter, which adds channels to the optical fiber without significantly affecting the power propagation in other channels. Transmission and reflection filters have been analyzed in detail in [HL91, KHO87]. Various configurations above can be used for channel drop filters, including the arrangements described with reference to FIGS. 7, 10, 26, 27 and 28.

The T-type grating-coupled waveguide in Fig. 7 is a channel down-line filter with low loss for out-of-band components. Compared to prior art gratings, this structure has difficulties with crosstalk due to the very long grating required to achieve 99.9% outcoupling for in-band components. The coupling strength of the periodically polarized gratings of the present invention is significantly improved over the prior art due to the ability to use higher order gratings with sharp interfaces spanning the entire extent of the waveguide. While the prior art is limited by the use of shallow waveguides to optimize the overlap between shallow gratings and waveguides, since our grating structures extend across the entire depth of the waveguide, lower loss waveguide structures with substantially equal depth and width can be used. This structure can also be used as an inline filter on the channel.

The device in Figure 10 is also a channel drop filter if the grating is constructed as described by Haus et al. Our contribution in this case lies solely in the polarization grating coupling technique, which enables strong coupling between waveguides over short distances, thereby reducing the difficulty of fabricating higher-order gratings with high efficiency.

By tuning the resonance of the etalon to the channel frequency extracted from the input waveguide 752, the devices 750, 760 and 770 can be used as channel drop filters. If the integrated etalon is close to impedance matching, essentially all the power at the resonant frequency is transferred to the etalon. In the ring geometry of Figures 27 and 28, the transmitted and reflected power in the waveguide 752 can be reduced to any desired degree, minimizing crosstalk. The light corresponding to the desired channel is completely extracted (off line) from the input waveguide, leaving neither reflection nor transmission. In the linear geometry of Figure 26, some light is lost to reflection, not significantly reducing detection efficiency, but potentially causing crosstalk problems in the communication network. The signal carried in the light is detected by placing a detector on the waveguide portion of the etalon and coupling to the light in the waveguide. Alternatively, the detector may be coupled to, for example, one of output waveguide 754 in FIG. 26 , 764 , 765 , or 766 in FIG. 27 , and 794 in FIG. 28 . In the case of device 760, external coupling is accomplished by adjusting the reflection of one of the resonator grating reflectors 767, 768, or 769 such that a fraction of the circulating power is coupled to a continuous portion of the waveguide as shown for output beam 751. coupling. Those continuous waveguide sections can also be connected to ports of other devices, which can be discrete devices or integrated on the same substrate. In the case of device 770, a parallel waveguide output coupler (with or without grating) may be placed in the flat region 771 of the ring. Although only a small fraction of the circulating power can be outcoupled to these ports, the total outcoupled power can be very close to 100% of the channel power entering the waveguide 752 due to the set-up generated in the etalon. The output coupling is represented by an adjacent waveguide 794 , resulting in output beam 751 .

Ring geometries have advantages in terms of extinction ratio (high extinction ratio for high optical separation efficiency) and low crosstalk, since they can be tuned so that nearly all power is delivered into the etalon. All etalon devices can be designed with very low insertion loss for out-of-band beams. All devices of Figures 26-28 are switchable by phase shift electrodes 761 and 762 (and 763 in Figure 28).

As previously described, the optical path length can be adjusted with electrode 761 to shift the resonant frequency of the etalon. The desired channel can be selected directly in this way. Alternatively, multiple channels can be selected by this technique using the method described above with reference to Figures 20, 21, and 22; the Moire effect can be used to select widely spaced channels with minimal continuous tuning if the FSR of the selection standard differs slightly from the channel spacing . (A good choice is to make FSR equal to the channel spacing plus the frequency bandwidth obtained when the channel bandwidth is wound with the resonant bandwidth of the etalon).

As a variation on structures 750, 760, and 770, coupling region 753 may be implemented as a grating-assisted coupler described above with reference to FIG. 10 . This has the advantage that, in a polarized grating implementation, the coupling coefficient T can be adjusted. Especially for ring resonator designs 760 and 770, adjustable coupling helps achieve impedance matching. As a further variation, electrodes can be placed on the same side of the substrate as described above to obtain lower voltage actuation.

For example, if the signal applied to output beam 759 is directed onto waveguide 766, or if it is coupled to planar region 771 via waveguide 794, the structures of Figs. 27 and 28 can also be used as effective channel on-line filters. These input interactions are preferably impedance matched.

Referring to FIG. 29A, a waveguide modulator/attenuator 800 using a polarizing section 806 is shown. The function of the polarizing section 806 is to (switchably) collect light emitted from the input waveguide section 802 and emit it to the output waveguide section 804 when the output waveguide section 804 is switched on. In this device, an input beam 820 is coupled to an input waveguide 802 . Polarization section 806 is located between the input waveguide section and output waveguide section 804 . The input and output waveguide segments are preferably permanent waveguides manufacturable by any standard technique including in-diffusion and ion exchange. Segment 806 is preferably a reversely polarized region within a uniformly polarized substrate such that there is substantially no difference in refractive index and no waveguide effect when the electric field is switched off. Segment 806 is the waveguide segment as shown in Figure 29A. (It can be constructed alternately in several different geometries, for example, a convex lens structure, a concave lens structure, or a composite structure that transmits light between many such elements: see Figure 9). It is switched on by applying an electric field through segment 806 . The electric field changes the refractive index of the polarized segment and the surrounding region. Since the segment 806 is polarized differently (preferably oppositely polarized) according to the substrate material, by applying the correct polarity of the electric field, the refractive index of this segment can be increased relative to the surrounding material, forming a waveguide. The refractive index inside the waveguide boundary can be increased, or can be decreased outside the boundary. When the polarized section is turned on, a continuous waveguide connecting the input and output sections is formed. This is accomplished by joining the waveguides together, aligning them on the same axis, and adjusting the width of the polarized sections so that their transverse mode profiles preferably match those of the input and output waveguides 802 and 804 .

With the polarization section switched off, the input beam is not confined in the polarization region, allowing the beam to expand sufficiently by diffraction before reaching the output waveguide section. If the separation between the input and output waveguide sections is much larger than the Rayleigh range of the unguided beam, so that the beam expands to a size much larger than the output waveguide, only a small fraction of the input beam will couple to the output waveguide section to An output beam 822 is formed. By adjusting the length of segment 806 relative to the Rayleigh range, the amount of power transmission in the off state can be reduced to a desired degree.

The position of the ends of the polarizing section 806 is adjusted relative to the position of the ends of the input and output waveguides to minimize loss due to the discontinuity. Since the permanent waveguide has a diffuse boundary, the polarized waveguide has a discontinuous boundary, and the refractive index change in the switching section adds to the pre-existing refractive index, there is a need for A small gap equal to half the diffusion length is left in between. To further reduce reflections and losses at the junction between waveguides 802 and 806, the initiation of the refractive index change in segment 806 is also facilitated by making excitation electrode 810 slightly shorter than segment 806 or by tapering the electrode width near its end. Cone-shaped, in both cases exploiting the reduction of the electric field through edge effects.

A notable aspect of this structure is that it minimizes reflected power in both the on and off states. With the cut off, the reflection is dominated by the residual reflection at the end 803 of the waveguide 802 . This reflection can be minimized by tapering the refractive index difference along the length of the waveguide. The reflection from the end 805 of the waveguide 804 is suppressed by the square of the "cut off" transmission. In the "on" state, this reflection can be minimized by forming a smooth boundary rather than a sharp interface, again by gradually reducing the refractive index difference of the structure 806 along the direction of propagation.

When a polarized region is excited due to a refractive index increase within the boundary, the boundary of the excited polarized region laterally confines the light beam. If the depth of the polarized region is equal to the depth of waveguides 802 and 804, the polarized segment boundaries also confine the beam in the vertical direction. However, it is difficult to control the polarization depth in z-cut lithium niobate wafers. It is easiest to polarize a deep domain and obtain confinement on the vertical scale using one of several different measures. A preferred method is to arrange the electrodes so that the magnitude of the electric field decreases in the vertical scale. This is accomplished by the configuration of electrodes on the same side as shown in Figure 29A, rather than electrodes disposed on opposite sides of the substrate. The penetration depth of the electric field can be reduced by narrowing the gap between the two electrodes and by reducing the width of the entire electrode structure.

Additionally or alternatively, a weak permanent waveguide can be created in the space between the input and output waveguides, which by itself is insufficient to transfer much energy, but which in combination with the rise in the refractive index produced in the polarizing section 806 may be preferable The light is confined in two dimensions so that substantially all light is transmitted to the output waveguide 804. This is accomplished by, for example, adjusting the permanent refractive index change within the segment (relative to the substrate) to about 0.6 times the refractive index change in waveguides 802 and 804 . If the "on" index change in segment 806 is adjusted to about 0.5 times the same value, the combined index change is sufficient for reasonable guidance, while the permanent index change is not. In the "on" state, the mode is confined in two lateral dimensions, even though the resulting switching index change in the poling region can be much deeper than the required waveguide dimensions: the effective depth of the "on" waveguide is mainly determined by The permanent refractive index change is determined. The weak waveguide can be fabricated in a second mask step, or can be fabricated in the same step with a narrower weak mask section defining the weak waveguide section.

As a relative alternative, the region between the input and output waveguides could be a planar waveguide, in which case the propagating mode would be at least diffracted in one dimension. Switching on a polarized region will increase the required lateral confinement despite deeper index changes than planar waveguides. Since the definition of the waveguide in two dimensions in both cases is achieved by two independent techniques, switchable waveguides of essentially any aspect ratio (ratio of waveguide width to depth) can be formed. Planar and channel waveguides can be fabricated using the same technique, preferably an annealing proton exchange process. Planar and channel waveguides can be defined using separate proton exchange steps. The waveguide fabrication process is completed by annealing, during which the refractive index change is diffused down to the desired depth and records the optical activity of the material. Both sets of waveguides are preferably annealed for the same length of time, although a partial anneal of one set can be made deeper before the second proton exchange step is performed.

One re-use alternative is to use a full, uniform permanent waveguide transverse to the polarized section 806, and use electronically energized sections to cut the waveguide. In this case, the polarity of the electric field is chosen to suppress the refractive index in the polarized region, which can be very deep (actually, this has its advantages in terms of mode dispersion). Such switched waveguides are usually conducting (ie, transmissive) and require the application of an electric field to switch them off. Normally closed and normally open switch configurations have advantages in terms of their behavior during power failures, so it is important that the present invention can provide both modes. To sever the waveguide in section 806 requires a refractive index change approximately equal and opposite to the induced refractive index change in the permanent waveguide. The effect of the variation with field depth in the "off" state is very small, since it is sufficient to suppress most of the waveguide so that the light diffuses strongly.

Planar waveguides are not required, and confinement can be achieved in two dimensions by finite depth polarization techniques. Several poling techniques (e.g. such as titanium indiffusion in lithium niobate and lithium tantalate and ion exchange in KTP) produce poling to a finite depth through which a particular depth may be optimized to form a poling channel waveguide . However, these techniques, along with polarization, produce a refractive index change that slightly forms some permanent waveguides depending on the processing parameters. Polarized waveguide sections can be fabricated in either "normally closed" or "normally open" configurations depending on the intensity of the refractive index change.

The electric field is preferably generated by applying a voltage across two electrodes arranged on the same crystal plane as the polarizing waveguide segments. The first electrode 810 is arranged on the polarized area, and the second electrode 812 is arranged near one or more sides of the first electrode. For z-cut crystals, this constitutes the d33 electro-optic coefficient of the excitation substrate. A voltage source 816 is basically connected to the electrodes via two wires 814 to provide the drive voltage for the device. The device can be used as a digital or nonlinear analog modulator. Full turn-on voltage is defined as the voltage with the lowest losses across the polarized region. The break-off voltage is defined as the voltage that reduces the coupling to the output waveguide section to the desired level. Continuously varying between on and off voltages allows the device to be used as an analog modulator or variable attenuator.

In another configuration, structure 806 forms a switched curved waveguide, again aligned with input waveguide 802 and output waveguide 804 . This structural mode is called a "whispering gallery" mode, and in rare cases the curvature is less and the mode definition on the inner edge becomes independent of the waveguide inner waveguide edge. For larger curvatures, the wave pattern is a modified whispering gallery wave pattern with some confinement provided by the inside edge of the waveguide. This polarized structure offers another advantage besides switchability, namely that the sharp refractive index on its outer wall greatly improves the confinement of the improved whispering gallery mode propagating in the curved waveguide. In this case, the input and output waveguides do not have to be coaxial or parallel, possibly increasing forward isolation in the off state. If the input and output waveguides are aligned at an angle to each other along their axes, the structure 806 may be a curved waveguide segment with a single radius of curvature or a tapered radius of curvature to preferably couple between them when the curved waveguide structure 806 conducts. even power.

Fig. 29B shows another structure 801, which is a switched lens modulator/attenuator, in which the prismatic structure of segment 806 is modified into a lens-type structure in which the product of the local optical path length and the local (signed) refractive index varies with The lateral distance from the axis of waveguides 802 and 804 decreases quadratically. The lens type structure is arranged so that it converges or refocuses the light beam 821 emanating from the end 803 of the input waveguide 802 to the end 805 of the output waveguide 804 . Light waves are allowed to diffract off end 803 and pass through lens-type structure 807 . It should be noted that in this configuration, multiple elements may be placed adjacent to each other to increase the net focusing effect. The refractive index in region 807 is increased to obtain a focusing effect. The spacing between lenses can also act as a focal zone if the surrounding zone is polarized in the opposite direction to zone 807, or if the electro-optic coefficient of the surrounding zone is opposite to that of zone 807. (The regions between lenses 807 form concave lenses actuated to lower refractive index values as converging lens structures). Electrode 810 is disposed on structure 806 and electrode 812 is disposed on the outside of the structure adjacent to electrode 810 but separated by a desired gap. When the electrodes are not excited, the beam diverges continuously and only little power is refocused to the end 805 of the waveguide. When the switch is on, the beam is refocused and a fraction of the power is continuously passed through the waveguide 804 . In the on state, vertical confinement is required for active power harvesting, but not in the off state. For example, vertical confinement can be provided as desired by providing a uniform planar waveguide 835 across the entire surface of the structural pattern. Vertical confinement can also be provided by the lens type structure 806 if the polarization goes deep into the substrate, the electric field reduction as a function of depth is suitable for harvesting and refocusing energy back to the end 805 of the waveguide. Of course, the structure of FIG. 29B can also be used in other situations where neither or both waveguides 802 and 804 are present.

Referring to FIG. 30, a polarized total internal reflection (TIR) optical energy redirector 830 using a polarized waveguide segment is shown. The figure illustrates the combination of both a polarized TIR reflector for high switching reflection and a polarized waveguide section for low insertion loss. The input waveguide 832 extends across the entire device. The polarized region 836 extends across the waveguide at an angle 848, forming a TIR interface for propagating light beams in the waveguide when the polarized region is electro-optic excited. A portion of the polarized region also forms a polarized waveguide section 837 connected to the output waveguide section 834 . The polarized waveguide section and the output waveguide section are arranged at twice the angle 848 relative to the input waveguide. A voltage source 846 provides electrical excitation for the switcher and is connected to the switcher via two wires 844 .

According to the figure, the polarization region 836 is bounded by six vertical faces, one of which is transverse to the waveguide 832 at a small angle 848 equal to the TIR angle and less than the critical angle for total internal reflection for desired electrode excitation. This face is the TIR reflective interface. The next three consecutive vertical planes of the polarized region surround a protrusion on the outside of the waveguide 832 . This protruding part is the switchable waveguide segment. The next two vertical planes are not critical and follow the waveguide boundary and intersect it at 90°.

The domains (836 and regions outside the substrate 836) are characterized by a static refractive index profile, the spatial distribution of the refractive index in the absence of an applied electric field. When an actuating electric field distribution is applied across these domains, they will have a different actuated refractive index distribution than the corresponding static distribution. The excitation profile also has a range according to the reachable range of the applied electric field. An advantage of juxtaposing the two types of domains close to each other is that the electric field response in the two domains can be reversed, providing a transition of twice the refractive index change across the juxtaposed region. In the case of a change in refractive index or refraction, this transition forms a reflective boundary whose reflectance is greater than that obtained with a single type of domain.

When the switch is on, the input beam coupled to the waveguide reflects off the TIR interface, propagates down the polarized waveguide section, and enters the output waveguide section 834 to form the deflected output beam 854 . When the switch is off, the input beam propagates across the polarization interface and continues through the input waveguide, forming an undeflected output beam 852 . Reflection is low in the off state due to the low refractive index change at the TIR interface. Since the permanent waveguide segments 834 are separated by several wave-mode exponential decay lengths from the waveguide 832, the power loss due to scattering of the light beam passing through the switching region is also very low. The "cut off" switch is essentially invisible to the waveguide, producing very low losses in the input waveguide. The additional loss of the switching region in the off state compared to the same length of the unexcited waveguide is called the insertion loss, low insertion loss is especially required when the input waveguide is a bus with many polarization switches.

The angle θ(848) of the polarization interface with respect to the input waveguide must be less than the maximum or critical TIR angle θc, derived from Snell's law: $\mathrm{\&theta;}\&le;{\mathrm{\&theta;}}_c=\mathrm{co}{\mathrm{the\; s}}^{-1}(1-\frac{2\left|\mathrm{\&Delta;n}\right|}{\mathrm{no}})\&infin;2{\sqrt{\frac{\left|\mathrm{\&Delta;\; n}\right|}{\mathrm{no}}}}^{\left(3\right)}$ where θ = TIR angle (angle between waveguide and polarization interface) n = refractive index of the waveguide region, and Δn = electro-optical change in refractive index on each side of the polarization boundary

Since the refractive index change occurs on each side of the polarization boundary with opposite signs, the effective refractive index change is 2Δn. This expression assumes that the refractive index changes away from the boundary slowly (adiabatically). Since the effective refractive index change is doubled, the maximum switching angle achievable with a polarized TIR switcher increases to that of a prior art switcher with a pair of electrodes and no polarized interface. times. This is a significant improvement since it increases the maximum storage density of the switch array achievable using one TIR switch.

Since the refractive index change Δn depends on the polarization, the critical angle θc depends on the polarization of the input beam. For example, in z-cut lithium niobate, with a vertical electric field E ₃ , TM waves are sensitive to changes in the extraordinary refractive index via r ₃₃ , while TE waves are sensitive to changes in the ordinary refractive index via r ₁₃ . Since r ₃₃ >r ₁₃ , it is easier to switch TM waves. The use of annealed proton exchange waveguides is convenient since they only arrive at waves that are polarized in the z direction. On the other hand, in x-cut y-propagation (or y-cut x-propagation) lithium niobate, the TE wave has a higher refractive index change. It should be noted that in this case the electrode configuration must be changed so that the electric field component is generated in the z-direction of the substrate plane rather than perpendicular to it.

The design angle of a practical TIR switch must be selected after optimizing several factors. The modes to be switched include two angular distributions (in and out of the plane of waveguide fabrication), which can be different if the waveguide widths are different in the two planes. The angular content of the wave modes in a given plane nearly covers δΦ=±λ/πw _O , where w _O is the 1/e ² mode convergent fraction in that plane. We expect most of the light to be reflected at the TIT interface, since the angle of incidence must be smaller than the critical angle _θc by approximately the angular content εΦ in the plane of the switched waveguide. The angular content δΦ is inversely proportional to the size of the convergent part, but is thus the storage density we wish to prefer. The angular content of the mode in directions out of the plane of the waveguide must also be considered, since it also contributes to the effective angle of incidence, although its geometry is more complex.

Another way to produce a TIR switch is to replace the electric field with a strain field, or to add a strain field in addition to the electric field. Strain fields are most convenient to implement in a permanent fashion; electric fields are most effective when producing reflective changes. A directional strain field applied at the domain boundary produces a differential change in the index of refraction, which in the two domains produces an index of refraction at the refraction interface through the photoelastic effect. As described above with reference to Figure 2, the strain field can be generated by heating the sample to high temperature, depositing a film with a different coefficient of thermal expansion, and cooling to room temperature. A pattern applied to the film by etching out, for example, stripes will create a strain field around the gaps in the film. This strain field can then be used to stimulate changes at the domain boundaries. If the applied film is a dielectric, an electric field can be applied through the dielectric to the polarized regions as long as the deposition of the electrodes does not undesirably alter the strain field. The film is preferably a low absorbance film so that it is in direct contact with the substrate rather than being separated by a buffer layer.

The polarized region includes a portion of the input waveguide and an interface perpendicular to the propagation axis of the waveguide. The section of the input waveguide that contains the TIR intersection interface defines the switcher length as follows: $L=W\cot \left(\mathrm{\&theta;}\right)\&infin;\frac{W}{\mathrm{\&theta;}}---\left(4\right)$

where θ is defined as before

L = switcher length measured along the input waveguide, and

W = waveguide width

Therefore, to minimize the size of the switch, the waveguide width must be made as small as possible. For space critical applications, the waveguide section is preferably of the single mode type. For example, if the width of the single-mode waveguide is 4 μm, the maximum refractive index change Δn is 0.0015, the refractive index is 2.16, the TIR angle θ is 3° and the switch length L is 76 μm.

The polarized waveguide section forms an angle equal to 2θ with the input waveguide, which is the deflection angle of the TIR interface. In order to effectively mode-match the reflected beam off the TIR interface to the polarized waveguide section, the polarized section should have nearly the same transverse mode distribution as the input waveguide. Effective mode matching can be achieved by choosing an appropriate combination of polarized waveguide width and refractive index difference. The polarized waveguide intersects the input waveguide along the rear half of the waveguide occupied by the switch interface. The exact dimensions and positions of the waveguides are determined such that the near-field mode distribution emanating from the total internal reflection process preferably matches that of the waveguide in terms of propagation direction and transverse distribution. The same is true for matching between polarized waveguide segments and permanent waveguide segments 834, similar to that described above with reference to FIG. 29A.

The permanent waveguide section is basically a continuation of the polarized waveguide section. The polarized section length depends on the preferred losses in the input waveguide and the switched waveguide. To avoid scattering effects between undeflected beams in the input waveguide when the switcher is switched on and off, the permanent waveguide section must be separated from the input waveguide by a distance (at least one optical wavelength). For a bus waveguide with many switches, the loss in the input waveguide must be reduced to a value that is inversely proportional to the number of switches. The profile of the beam in the input waveguide extends a certain distance beyond the undiffused edge of the waveguide, where it decays exponentially. If the permanent segment is separated from the input waveguide by several exponential decay constants, the loss can be reduced to an acceptable level for the bus waveguide.

The length of the polarizing section also affects the losses in the reflected beam. Due to the high wall roughness, the loss per unit length of the polarized waveguide section is higher than that of the undiffused waveguide. In addition, there is the above-mentioned mode conversion loss at each end of the waveguide, which can be minimized by preferably matching the mode distribution. If the polarization segment is short (equivalent to the Rayleigh range of the beam), the transmitted beam is substantially not converted into the mode of the polarization segment, thereby reducing coupling loss. The preferred length of the polarizing section depends on the relative losses that can be tolerated in the input waveguide and the beam within the switching waveguide.

In the case of the waveguide segment modulator/attenuator shown in FIG. 29A , vertical confinement of the wave modes is required in the switched waveguide 837 . The same options described above can be implemented here. Figure 30 shows a planar waveguide 835 confining the beam in a plane parallel to the substrate surface. Since the planar waveguide is uniform, its presence has no effect on the loss of the junction of the waveguide switcher in its cut-off state. Other alternatives to planar waveguides or some combination thereof may also be implemented, including engineering the depth of the electric field for vertical confinement, using short depth polarizations, using portions of the waveguide that are enlarged by electric field-induced refractive index changes, and using field-excited polarized regions that are closed. fully permanent waveguide. The latter two alternatives have the disadvantage of higher loss of the light beam passing through the waveguide 832 due to adjacent refractive index discontinuities.

Horizontal constraints are also an issue in optimizing handover regions. If high switching efficiency is required, it is preferable to have a larger TIR reflection angle. The left half of the input wave 851 first reflects off the interface 838 to form the right half of the reflected wave. However, after reflection, the right half of the reflected wave is unconfined in lateral dimension until it reaches waveguide section 837 . The beam power coupled to the output waveguide 834 is reduced by a small fraction during its unconfined pass through due to diffraction expansion. This effect reduces the efficiency of the switcher in its on state. However, the average unguided distance is approximately limited to the waveguide width divided by the four times the sine of angle 848 . The right half of the input wave remains confined after passing through waveguide section 837 until it is reflected off interface 838 due to the permanent refractive index change caused by the right half of waveguide 832 . Then, it is better matched to the output waveguide 834 . Both parts of the input beam 851 undergo unwanted reflections from the sides of the waveguide 832 after reflecting from the TIT surface 838 . Since this surface is at the same angle as the beam propagation axis and surface 838, but only a small fraction of the index of refraction is different, there is only partial rather than total reflection from this surface, which also increases switcher losses.

To optimize reflector efficiency and minimize waveguide losses, electrode design is an important aspect of the switch. Preferably two electrodes are used to trigger the switch. A first electrode 840 is disposed above the TIR interface 838 and a second electrode 842 is disposed beside the first electrode, adjacent to the interface. The main parameters of optimization are the distance between the two electrodes and the distance between the edge of the first electrode and the polarization boundary. The electrode edge and the polarization boundary may or may not overlap. The spacing between the two electrodes affects the voltage required to activate the device, as well as the width of the electric field pattern that penetrates the substrate and produces a distribution of refractive index changes. Further spaced electrodes require higher voltages, but generate electric fields that extend deeper into the substrate than closely spaced electrodes.

The depth of penetration of the electric field is important to obtain a large net reflection. Since the electric field becomes weaker the farther away from the electrode, the induced refractive index change at the polarization boundary also decreases with depth, as does the TIR angle. At a certain depth, called the effective depth, the refractive index change becomes insufficient to maintain total reflection of the central ray of the beam at the angle of the switching structure. TIR mirrors are essentially ineffective at this depth because reflection values below the minimum TIR value drop off rapidly with refractive index changes. For high net reflections into waveguides 837 and 834, the device design can be tuned to create an effective depth below the dominant electric field distribution in waveguide 832.

The second major operating parameter affected by electrode design is the penetration of the evanescent field of reflected waves beyond the TIR interface 838 . Although no power can be transmitted beyond the TIR interface in the on state, the electromagnetic field penetrates the TIR surface for a distance equivalent to one wavelength. There is also a spatial dependence on the applied electric field outside the TIR surface, with the electric field strength decreasing (actually reversed) in regions close to the other electrodes 842 . Therefore, the refractive index change decreases outside the TIR interface. Care must be taken that the evanescent field decays to a negligible value before a substantial change in the electric field occurs, or power leaks across the TIR interface. It is preferred that the first electrode overlaps the polarization interface by a distance selected for maximum refractive index change and sufficient electric field stability beyond interface 838 .

The first electrode also extends across the polarized waveguide segment 836, and possibly into adjacent regions. The shape of the two electrodes that excite this region 836 is determined by optimizing the power flowing through the waveguide segment and into the permanent waveguide 834 . Other electrodes can be used to improve the electric field strength in the polarized region. For example, if the second electrode extends around to form a U shape, the electric field under the first electrode increases on average, but it forms somewhat like a raised waveguide, which may not provide the desired refractive index profile.

The TIR switcher is a light energy router and also acts as a modulator. If the voltage source is continuously variable, the modulator is analog, with a non-linear relationship between voltage and reflectivity. As the applied voltage increases, the depth of the total reflection interface increases, producing a continuously adjustable reflected incoming wave 854 from wave 851 . This modulator can be used in either reflected or transmitted mode, depending on whether the transmission should go to zero or 100% when the voltage is removed. Non-linearities in reflection and transmission coefficients as a function of voltage may be useful for special nonlinear applications, such as receivers that are logarithmically sensitive to signal levels.

Figure 31 shows a TIR switcher with two TIR reflectors. If it is desired to increase the angle between output waveguide 834 and input waveguide 832, a second TIR interface 839 can be added. The angle between the input waveguide 832 and output waveguide 834 is double that in Figure 30 and can be doubled again and again by adding additional TIR interfaces. Interface 839 is established at an angle 849 to opposing interface 838 equal to twice angle 848 . (If a TIR interface is added, subsequent TIR interfaces should be added at the same angle 838 relative to the previous TIR interface). The switched waveguide section 837 of FIG. 30 is no longer needed because the dual TIR mirrors keep light away from the input waveguide 832 so that the permanent waveguide 834 can rest directly on the end of the polarized region 836 without significant loss to the waveguide 932 . In addition, vertical confinement is provided in the polarized region 836 . Polarized region 836 and output waveguide 834 are constructed and aligned such that the electric field distribution propagating in the TIR chain and waveguide section preferably matches the local lowest order mode electric field distribution of input waveguide 832 . After the TIR reflector, the deflected beam is matched to a permanent waveguide 834 to form an output beam 854 when the switch is turned on.

The shape of the inner boundary of the polarized region outside the input waveguide 832 is defined by the input waveguide reflections through one TIR mirror after the other. This definition of the inner boundary achieves a preferred guidance of the inner edge of the waveguide mode when reflected from the two TIR mirrors.

Figure 32 shows a TIR switched beam guide with a polarized TIR switch 831 having an electronically switched waveguide section. In this configuration, region 836 is oppositely polarized, located behind interface 838 , and is energized as previously described by a pair of electrodes 840 and 842 , which are energized by a voltage source 846 connected via conductor 844 . Again the excitation polarity is chosen to produce a negative index change from the direction of the input beam 851 . When the switch is on, the light beam leaves the TIR interface 838 and reflects to the permanent waveguide 834, but unlike FIG. 31, there is no polarized waveguide segment between the two connecting them. Instead an electrode 842 extends over the intermediate region between the input waveguide and the output waveguide 834 . The coupled waveguide segments may be formed by applying an electric field to the region between the lateral boundary of the input waveguide 832 segment and the input boundary of the output waveguide 834 containing the TIR reflective boundary. As before, the three-dimensional distribution of the electric field is determined with the help of electrodes and Maxwell's equations. The electric field generated by this electrode produces a positive refractive index change through the electro-optic effect, providing the required switched waveguide segments. This waveguide segment is also configured and aligned such that the input mode is preferably coupled to the output mode 854 as described above. As an alternative to this and any TIR switcher implementation, the output waveguide may start at the input waveguide with negligible gap. This alternative has higher insertion loss when the switch is in an off (through) configuration.

Referring to Figure 33, a two-position waveguide router using polarized segments is shown, not based on total internal reflection. The polarized region 866 forms an electrically excitable waveguide segment that spans the input waveguide 862 at a small angle. When an electric field is applied, the index of refraction in segment 866 increases, while the index of refraction in adjacent regions in the input waveguide decreases. The input light beam 880 is thus at least partially coupled to the polarized waveguide section. When the switch is off, the input beam continues to propagate through the input waveguide to form the unswitched output beam 882 . If it is desired to switch all or most of the input light to the output waveguide 864 as the switched output beam 884 exits the device, the small angle can be tapered adiabatically, forming a low loss waveguide bend.

At least two electrodes are used to apply an electric field across the polarized region to excite the waveguide. A first electrode 870 is disposed over the polarized waveguide segment, and a second electrode 872 is disposed adjacent to the first electrode. The second electrode 872 is adjacent to the first electrode and can be located on both sides of the polarized waveguide section to achieve a high power split ratio. Vertical confinement of switched propagating modes is obtained by exciting the electrodes via conductor 844 from power source 846, using planar waveguide 835, or electric field drop with depth, or one of the other methods described herein, as previously described.

Referring to FIG. 34 , several polarized TIR switches are placed side by side to form an array 900 . Polarized regions 912 and 914 forming the TIR interface are arranged one after the other along the waveguide 910 . Each polarized region has the same crystal orientation, with the crystal z-axis being opposite in regions 912 and 914 from the rest of the crystal. Other aspects and many variations of this configuration have been described above with reference to FIG. 30 .

Each switch is individually energized by a multiple output voltage control source 926 connected to the electrodes by leads 928 . When all switches are switched off, the input beam 902 propagates down the input waveguide 910 to form an unswitched output beam 904 with negligible loss. If the first switch is on, the input beam is reflected off the TIR interface to form a first reflected output beam 908 in the waveguide 916 . If the first switch is off and the second switch is on, the input beam is reflected off the second TIR interface to form the second reflected output beam 906 in the waveguide 918, and so on for subsequent switches. The multi-switch structure can be extended to n switches.

Electrodes are arranged at each TIR interface as described above. One or more of the electrodes 920, 922, and 924 act as the cathode of one switch and the anode of the other switch. For example, a voltage is applied between the second electrode 922 and the first and third electrodes 920 and 924 to energize the second switch to form the output beam 906 . The electrodes 922 acting as anode and cathode should preferably extend near the TIR interface of the pre-polarized segment 912 while covering the TIR interface of one polarized region 914 and one waveguide segment of the polarized region 914 . Only a portion of the structure is shown, in which there are two complete polarized sections 912 and 914 and one complete electrode 922 . The structure can replicate n switches by aligning and replicating complete electrodes and polarized segments.

To avoid crosstalk in the channel, a voltage is applied to the electrodes in such a way that no electro-optic index change is seen by the input beam until it enters the activated switch. For example, to excite the TIR interface of the second polarization region 914, a voltage can be applied between electrodes 922 and 924, keeping electrodes 920 and 922 at the same potential as the previous electrodes.

Although the total length of the polarized regions is longer than L, the distance occupied by a given region along the waveguide is equal to L as defined. A linear array of TIR switches with a storage density of 100% therefore has a new polarization region at every distance L initially. This is referred to as 100% storage density because at this density adjacent regions only touch each other at the inside corners of the polarized regions within the waveguide. It is disadvantageous for adjacent regions to touch each other since some of the light guided in a previously polarized structure leaks to the next polarized structure in contact with that structure.

We have pointed out above that the corners in contact with the preceding polarization zone are formed by the two perpendicular faces of the polarization zone and their position is not important. By moving these faces to make the width of the polarized region thinner on the side of the inner corner, the regions can no longer be arranged in contact with each other, thereby reducing light energy leakage. For example, the inside corner can be moved to the middle of the waveguide by halving the length of the 90° face transverse to the waveguide. The face that used to be parallel to the waveguide is now parallel to the TIR interface and becomes an important location surface. We refer to polarized regions with this geometry as "dense storage" polarized regions. (There are other ways to achieve the goal of minimizing light leakage, such as adding a seventh vertical plane between two unimportant planes, but this variation has another advantage in terms of dense storage as previously stated).

Figure 35 shows a structure in which the linear density of switches is doubled by using a dense storage geometry for polarized regions and reversing the polarization of adjacent polarized regions. The interface of the polarized region transverse to the waveguide is now the same, but only for transfer along the waveguide axis. The polarized regions will thus be firmly stacked along the waveguide, doubling the switcher density. In fact, since the other regions are in the same direction of polarization as the substrate (in the preferred case the substrate is fully polarized), only the opposite region is spatially fully delimited. FIG. 35 shows two oppositely polarized regions 952 and 954 . The TIR interface can be considered to be the first or input face and the second or output face of the polarized region, and unswitched light propagating in waveguide 950 is likely to enter or leave the unexcited polarized region, respectively.

A TIR interface for the output beam 946 is formed between the polarized substrate and the first (input) face of the oppositely polarized region 952 , excited by the electrode 966 . A TIR interface for the output beam 947 is formed between the second (output) face of the reverse polarized region 952 and the polarized substrate, excited by the electrode 967. A TIR interface for the output beam 948 is formed between the polarized substrate and the first face of the reverse polarized region 954 , energized by the electrode 968 . A TIR interface for the output beam 949 is formed between the second face of the reverse polarized region 954 and the polarized substrate, excited by the electrode 969 . The electrodes extend along the switched waveguide segments connected to the permanent output waveguides 956, 957, 958 and 959 at the respective TIR interfaces. One or more of the electrodes 966, 967, 968, 969 and 970 preferably act as the cathode of one switch and the anode of the other switch. Thus, each electrode extends in parallel along the entire length of the TIR interface of the preceding switch.

Each switch is individually switchable by applying an electric field from a voltage source 926 through a conductor 928 . When all switches are switched off, the input beam 942 propagates down the bus waveguide 950 forming an unswitched output beam 944 . When the first switch is turned on, the input beam is reflected off its corresponding TIR interface and coupled to the first output waveguide segment 956 to form the first reflected output beam 946 . The input beam is reflected off the corresponding subsequent TIR interface to couple to waveguide segment 957 , 958 , or 959 to form reflected output beam 947 , 948 , or 949 corresponding to a subsequent switch. A typical setting of the voltage on the electrodes is that there is no light interference from adjacent switches: all previous switches are off. This can be achieved, for example, by keeping all previous electrodes at the same potential as switched electrodes. The multi-switch structure can be extended to n switches.

It is desirable to extend the upstream end of the dense storage polarization region significantly beyond the edge of the input waveguide 950, maintaining the angle of the vertical surface with respect to the waveguide. This extension captures the full exponential tail of the input waveguide mode and pushes the remaining unimportant positioning surfaces of the extended densely stored polarization region out of the waveguide 950, thereby eliminating optical losses. (Upstream and downstream are defined as directions of propagation relative to the input beam 942).

If the switched waveguide of the polarization region is designed as described above with reference to Figure 30, the spacing of the output waveguides becomes the same as its width in the highest density storage so that it falls under the planar waveguide. For some applications planar output waveguides may be useful, which can be separated using a second polarized TIR interface in each switch. The use of two TIR interfaces in one switcher has been described with reference to FIG. 31 . It should be noted that in the case of Figure 35, the geometry of the polarized regions is slightly different to achieve stacking. The "output waveguide" portion extending the dense storage polarization region is rotated by an angle 3Θ relative to the input waveguide 942 around the end of the first TIR interface, keeping its faces parallel. This "output waveguide" section thus becomes the second TIR reflector section.

The width of the second TIR reflector segment is about 50% larger than the input waveguide. Mode propagation within the second TIR reflector segment is not limited to a distance of about 2W/sin[theta] inside it, where W has been defined as the waveguide width. Any diffraction that occurs on this side will result in a reduction in the power coupled to the output waveguides 956-959. This distance should be kept less than the Rayleigh range. In the case of a 4 μm wide waveguide operated at a 4.5° TIR angle, the total unconfined distance is about 100 μm, approximately equal to the Rayleigh range of the blue beam. One solution to optimize the performance of this switcher array is to add a permanent decrement of the refractive index (without degrading the electro-optic coefficient) at strategic locations within the second TIR reflector segment. The critical location is the region bounded by the inner walls of the extended dense storage polarization region and by the inner walls of the polarization region 836 as defined with reference to FIG. 31 . The permanent index decrement defines a permanent waveguide boundary at the mode-confined preferred location as it reflects from two consecutive TIR mirrors. The added index decrement tapers off to zero as it reaches the input waveguide, and the loss added to the input waveguide can be significantly reduced by truncating the index decremented region far enough from the waveguide. Nor does the index decrement interfere with the TIR function of the previous TIR interface (it actually helps).

Thus, the switched beam is reflected from two consecutive TIR interfaces, doubling the full deflection angle of the switcher to 4θ. By doubling the output angle, space is available for output waveguides with a width equal to that of the input waveguides, spaced equal to their width in their densest configuration.

The output waveguide connects to the polarized region in Figure 35 at the final corner of the second TIR reflector, at an angle Θ relative to the second TIR interface, and is preferably aligned so as to collect light reflections off the second TIR interface. Preferably the two TIR reflectors for a given switch are connected without intervening waveguide sections. This minimizes the path length of the deflected beam that must travel in the polarized waveguide, which may have higher losses than a permanent channel waveguide due to wall roughness and asymmetry.

In another polarization boundary structure, the boundary between two adjacent polarization regions may be a curved TIR structure. The mode of this structure is also a whispering gallery mode, possibly modified by some confinement on the inner boundary of the waveguide. Make the radius of curvature of the polarization boundary small enough to have a good match between the whispering gallery mode and the larger power waveguide mode coupled between the two types of waveguides, within the mode for the angular distribution within a sufficiently large practical full Internal reflection occurs.

Figure 36 shows a double cross waveguide structure 980 for higher storage density. The structure has two modifications: an asymmetric lossy waveguide cross 997, and 90° mirrors 976 and 977. The density increases with the addition of a second input waveguide 982 parallel to the first input waveguide 984, the first and second waveguides being on the same surface of the substrate 981, effectively doubling the storage density. Switcher elements 983 and 985 have been schematically illustrated as one of the variations of the polarized TIR switcher described above, but could be replaced by any of the integrated optical switches described herein, and therefore we do not describe this switcher in detail here or in FIG. 36 . (The switch can also be implemented in other ways described herein, such as the raster switch described with reference to FIG. 7, the coupler described with reference to FIG. 10, the distributor described with reference to FIG. 25, and the guide switch described with reference to FIG. 33. )

The first input beam 992 propagates down to the first waveguide, while the second input beam 994 propagates up to the second waveguide. The two beams can be emitted from different light sources, or from the same light source through an active or passive splitter. When the associated switch is off, input beams 992 and 994 propagate therethrough to form undeflected output beams 993 and 995, respectively. If the associated switch is on, the first input beam 994 is deflected into the output beam 996 while the second input beam 992 is deflected into the output beam 998 .

In an asymmetrical waveguide crossing 997, two waveguides cross each other, adjusting the refractive index profile to minimize loss in one waveguide at the expense of slightly higher loss in the other waveguide. The two intersecting waveguides are arranged at a relatively large angle (here illustrated as 90°) relative to each other in order to minimize the crossing loss. Referring to the geometry of Figure 36, the second deflected beam 998 spans the first waveguide 984 (in this case, the switched output beam may propagate parallel to the output waveguides 986 and 988). Waveguide 988 is separated from waveguide 984 at the intersection point, leaving gaps 990 and 991 . This minimizes the loss in waveguide 984, creating an asymmetric lossy structure in which the loss in waveguide 988 is higher than the loss in waveguide 984 in the crossover region. For the convenience of the following description, we call the asymmetrical intersecting waveguide with lower loss as "fixed point". An asymmetrical crossover 997 is positioned along the waveguide 984 . If gaps 990 and 991 are wider than several exponentially decaying lengths of the modes in waveguide 984, the crossover structure provides substantially no additional loss to waveguide 984. Multiple asymmetric crossing structures can then be positioned sequentially along the waveguide 984 to create a low loss waveguide that crosses many waveguides. Gaps 990 and 991 will cause some reflection and scattering of beam 998 propagating in broken waveguide 988, the gap width being minimized within the combined constraints of required low loss in both waveguides. To minimize optical loss at the crossed configuration from beam 998 propagating in waveguide 988, the refractive index profile transverse to the waveguide propagation axis can be modulated or tapered along the waveguide axis. The purpose is to keep losses very low in waveguide 984 while minimizing losses in waveguide 988. This can be achieved if the change in the refractive index in the region adjacent to the waveguide 984 is small and slow compared to the change in the refractive index of the waveguide 984 itself. (All refractive index changes referred to herein are relative to the substrate).

The loss in the second waveguide has two parts: one is due to reflections caused by the refractive index discontinuity, and the other is due to diffractive expansion. The return loss is determined by the magnitude of the refractive index change in the waveguide and its taper distribution at the ends and sides of the waveguide. For example, if the refractive index variation in the two waveguide cores is the same, Δn = 0.003, the net reflection losses at the four interfaces will be less than 5%, and the correct refractive index profile that reduces reflections will cause negligible corrections. Since the gap width is typically much smaller than the free-space Rayleigh range, the diffraction loss is even lower. For example, if the depth of the narrowest mode dimension is 2 μm, assuming a material with a refractive index of 2.2 and a wavelength of 0.5 μm, the Rayleigh range would be 55 μm. Assuming a gap width of 3 μm, the diffraction loss per gap is less than 1%. If the waveguide depth is 4 μm, the diffraction loss is substantially smaller. Diffraction loss can be minimized by increasing the waveguide size relative to the gap size.

Typically, "gap" 900 has a refractive index profile adjacent to the intersection region. The refractive index profile is defined relative to the substrate refractive index. The index of refraction in the gap may taper from a value equal to the index profile of the waveguide 988 to another value adjacent to the intersection region. The significant portion of the intersection region is the volume in which the light modes of the waveguide 984 propagate. To minimize losses in waveguide 984, the refractive index near the crossover region in this significant portion is much lower than the index profile within waveguide 984.

Crossed waveguide geometries with asymmetric optical loss can be combined by many geometric variations. For example, three or more input waveguides may be used for multiple intersections where the switched output waveguides are transverse to the input waveguides. The choice of the preferred waveguide can also be done in many ways, preferably in the sense that it minimizes losses at the intersection points. We have already discussed an example where it is preferred that the waveguides be parallel. However, in more complex systems it is possible to have preferred waveguides that cross each other, and preferred waveguides that cross non-preferred waveguides. The choice of how to achieve the crossover of the preferred waveguides depends on its application. The waveguide crossing structure in a device can be any combination of asymmetric lossy crossings and symmetric lossy crossings with a gap width of zero.

For switches that deflect beams at small angles (such as TIR switchers), additional beam rotation devices such as 976 and 977 can be provided to obtain large intersection angles at waveguide intersections. The beam rotation devices 976 and 977 are preferably vertical micromirrors mounted in fixed positions. Each waveguide is formed by moving the substrate material within its volume, leaving a flat vertical surface (preferably low roughness) adjacent to the waveguide and oriented at an angle that directs the reflected light preferably downward toward the output waveguide 986 or 988. a tiny mirror. Micromirrors can be fabricated using conventional processing techniques, including laser ablation with, for example, high power excimer lasers or ion beam etching, both of which define the mirror geometry with the help of masks. The volume can be filled with a low-index, low-loss material, such as alumina or silica, to prevent contamination of the mirror surface and maintain the total internal reflection properties of the mirror.

The angle of the micromirror relative to the input of one of the waveguides is preferably adjusted to provide total internal reflection. The thickness of the micromirror volume perpendicular to its reflective surface is preferably much larger than the wavelength of light to minimize leakage of the evanescent tail of the reflected light wave through the micromirror. The angles relative to the other waveguides are adjusted so that the mid-propagation direction of the reflected beam is parallel to the mid-axis of the other waveguides. The position of the micromirrors is adjusted to optimize optical coupling from one waveguide to the other. The position of the mirror in the junction area is preferably adjusted so that the "center of gravity" of the two light beams illuminating the surface of the mirror is at the same location. The length of the mirror transverse to the incident and reflected beams is greater than twice the width of the waveguide in order to fully reflect the entire mode, including the exponentially reduced intensity in the beam tail. Light input from one of the waveguide modes diffracts through the waveguide junction to the micromirror, where it is reflected and diffracted back through the waveguide junction at a reflection angle before being coupled to the output waveguide mode. The junction region between the two waveguides near the mirror is preferably kept small compared to the Rayleigh range of an unconfined beam, achievable with waveguide widths in the range of 2 to 5 microns.

The structure of Figure 36 enables a large interdigitated array of switched light distribution waveguides. The entire structure 980 is replicated multiple times along a pair of input waveguides, producing a set of interleaved output waveguides with small patterns of alternating sources (pattern rentage) (herein, a source refers to the optical power derived from a particular "source" input waveguide). Each input waveguide can be connected to a large number of output waveguides as long as the switching element has little insertion loss, as is the case for the elements listed above and described here. Due to the asymmetric crossover structure, adding more input waveguides above the other waveguides (including additional switches, micromirrors, asymmetric waveguide crossovers, and interleaved output waveguides) does not significantly increase the loss of the lower input waveguides or affect their long-term The ability to distribute light over distance to many output waveguides. It will modestly increase the source power required for each additional input waveguide in order to deliver the same power to the end of its corresponding output waveguide. As many input waveguides as required can be used in parallel to distribute as much total optical power as possible. Its output waveguides can be interleaved in many different patterns using the method of FIG. 36 . The same result can be obtained by using a grating reflector instead of a TIR switch. If the grating reflector is oriented at a large angle relative to the input waveguide, the micromirror is no longer required.

The structure described in the previous paragraph is a one-to-many structure where one output per switch is accompanied by multiple switches per input. It is not possible to wire many inputs to the same output. A many-to-one construct is required. Many-to-many constructions are obtained by combining one-to-many and many-to-one constructions.

Figure 37 shows a waveguide array 1060 with TIR switches arranged in a many-to-one configuration. In the configuration shown, two input waveguides 1072 and 1074 switch two input beams 1062 and 1064 into one output beam 1070 in one output waveguide 1076 . Input TIR switches 1090 and 1092, and output switches 1094 and 1096 have been described above with reference to FIGS. controller, vertical confinement device, polarization zone depth, output waveguide confinement type). As described with reference to FIG. 36, the input TIR switch is arranged to forward propagate the beam and the output TIR switch is arranged to reverse propagate the beam. Switches 1090 and 1092 switch substantially simultaneously, as do switches 1094 and 1096 , since both are required to complete injecting power into output switch 1076 . As shown with reference to FIG. 36, when switch 1090 or 1094 is turned on, a fraction of light beams 1062 and 1064 are switched to waveguide 1078 or 1084, respectively. The remaining input beam propagates along the extension of the input waveguide to an output path as beam 1066 or 1068, which can be used by other components or its beam dump to absorb or scatter out of the system. Micromirrors are arranged to reflect the light beams from waveguides 1078 and 1084 to waveguides 1080 or 1086, respectively. In its on state, TIR switch 1092 or 1096 receives the beam propagating in waveguide 1080 or 1086 respectively, forming output beam 1070 . If it is desired to switch beam 1062 to output beam 1070, obviously switch 1096 and all subsequent switches must be switched off. (Otherwise, reflect as much of the beam as needed out of the waveguide 1076). Similar restrictions apply to all other switched beams in a multi-switcher array.

Substrate 1098 was processed as described herein to produce the illustrated structures. When the switch 1090 or 1094 is switched off, the input beam propagates through the switch region 1090 or 1094 with negligible loss, then is transverse to the waveguide 1076 (in an asymmetric crossover if desired) and emerges as the output beam 1066 or 1068 respectively , which can be used as an input for an additional switcher.

Additional input waveguides may also be provided, coupled to waveguide 1076 (or uncoupled if desired), repeating the structure in a modified fashion in the direction of output beam 1070 . Additional output waveguides may also be provided, if desired coupled to input waveguides 1072 and/or 1074, repeating the structure in a modified fashion in the direction of beams 1066 and 1068.

Figure 38 shows a grating reflector array 1210 in a many-to-many configuration. In the configuration shown, two input waveguides 1222 and 1224 switch two input beams 1212 and 1214 into two output beams 1220 and 1221 in two output waveguides 1226 and 1228 that are close to or touch the input waveguides. The grating TIR switches 1230, 1232, 1234, and 1236 including the gratings 1238, 1240, 1244, and 1246 have been described above with reference to FIGS. Components (such as electrodes, contacts, power sources, controllers, vertical confinement devices, polarization zone depth, polarization zone taper, or electrode spacing). When switch 1230 or 1232 is turned on, a fraction of beam 1212 is switched to output beam 1220 or 1221, respectively. The remaining input beam propagates along the extension of the input waveguide to an output path as beam 1250, which can be used by some other element or its beam dump to absorb or scatter out of the system. When switch 1234 or 1236 is turned on, a fraction of beam 1214 is switched to output beam 1220 or 1221, respectively. The remaining input beam propagates along the extension of the input waveguide to an output path as beam 1252, which can be used by some other element or its beam dump to absorb or scatter out of the system.

Substrate 1248 was processed as described herein to produce the illustrated structures. When the switch is off, the input beam propagates through the switch region (where the waveguides can be configured as asymmetrical crossovers if desired) and emerges as output beam 1250 or 1252 respectively, which can be used as input to an additional switch. The waveguides may join each other at simple large angles, or may be asymmetrical intersections that do not substantially affect the position of the gratings 1238, 1240, 1244, and 1246. It should be noted that the actual grating may be part of a single large grating covering the substrate and being actuated by the required electrodes only at different switch regions. For example, if the grating is formed of polarized domains, this allows the entire substrate to be polarized for the grating, making production simpler. Alternatively, the gratings may be arranged in stripes or other groups.

Additional input waveguides may also be provided, coupled to waveguides 1226 or 1228 (or uncoupled if desired), repeating the structure in a modified fashion in the direction of output beams 1220 and 1221 . Additional output waveguides may also be provided, if desired coupled to input waveguides 1222 and/or 1224, repeating the structure in a modified fashion in the direction of beams 1250 and 1252.

Fig. 39A schematically shows another application example of switch array in nxn communication routing application. In this application, the optical power in the input optical channel is routed to the output optical channel with minimum loss and minimum crosstalk. The controller establishes an addressable path between the two channels. A simple square array is formed by repeating the structure of Figure 38 until n inputs are arranged on the left and n outputs are arranged on the bottom, with switches at all ⁿ² intersections of the waveguides. The angle of intersection may be any suitable angle. In this configuration, switching from one channel to the other is achieved by energizing one of the switches. The beams cross each other at waveguide intersections with a small amount of crosstalk, which can be reduced by optimizing the waveguide geometry. The structure is capable of independent one-to-one connections between any input and any output. It should also be noted that the connection can be bi-directional, so that in fact the through channel can be used equally in both directions at the same time. The switch is shown as a specific application of an optical grating, but a dual TIR switch as described with reference to Figure 37 can also be used to form n x n inputs and outputs by replicating the structure of Figure 37, or any other optical switch known or discovered technology to implement it. It should be noted that in the case of a TIR switch, the optical data path does not pass through the apex of the intersection between the input and output waveguides, but through another waveguide near the intersection. Depending on the particular geometry of the switcher, the input and output waveguides may intersect at large angles as shown in Figures 37, 38, and 39, or at an oblique angle. The fixed reflectors 1088 and 1082 in the dual TIR switching geometry are not required in the case of oblique intersection of the waveguides.

In this simple square geometry with n parallel input waveguides, one input waveguide is connected to the nearest output waveguide through a single switch, forming the preferred case with the lowest loss. At the other extreme, there is one waveguide that must cross 2(n-1) waveguides that switch to the farthest output waveguide transversely. This worst case connection will have much higher losses than the best case connection. To reduce the maximum insertion loss of the switch array structure, an asymmetrical cross-connect as described with reference to FIG. 36 can be used. An asymmetric crossing at each waveguide crossing at its point transverse to the direction of light propagation along one of the input or output waveguides preferably aids in losses of the worst case connection. Clearly, this structure cannot be generalized in internal waveguides, since the use of asymmetric switches in the intermediate junction will help some switching paths at the expense of others. What is needed is an algorithm for selecting the preferred direction of the asymmetric crossing. A good way to configure asymmetric crossings is to spot about half the crossings in each direction. Observe that the n(n-1) intersections on the upper left of the diagonal (not including the diagonal) are mainly used to distribute energy to the right. This intersection point should therefore be positioned in the direction of the input waveguide and the lower right intersection point should be positioned in the direction of the output waveguide. In a bidirectional structure, the intersections on the diagonal shall be simple symmetric intersections, referred to herein as a simple diagonal arrangement of asymmetrical intersections. Other arrangements of graphics may be used depending on the application, but this is the preferred arrangement for general purposes.

An nxm (where n>m) arrangement allows only fully connected lines between n "input" lines and m "output" lines. Here, "in" and "out" are used for identification purposes only, since all lines are bidirectional. Additional n-m "system" lines are available for system control in supervisory and broadcast functions. For example, if line A wants to connect to line B, it will send a system request for that function until answered. For example, line m+3 can be used to scan all "incoming" lines requested by the system. (To provide similar lines to monitor "outgoing" lines, a large matrix of lines is required, such as the n x n matrix shown in Figure 39A, where m lines in m x m line subgroups are provided to the subscriber. A line, such as line n-2 can be used to monitor "output" lines). In monitoring, the system will switch on the continuous light barrier corresponding to the "input" or "output" line and detect whether the line is running. If any of the lines being monitored is active, some power can be switched to the supervisory detectors by sequentially switching on the light grids on the same line as the supervisory detectors. A running line has an active reflector connecting it to another selected line. However, some power will leak through the excited reflector, forming a beam detectable by the monitor detector. When the supervisory detector connected to line m+3 in this example switches switch 1255 (drawn as a raster switch for concreteness) to conduction and receives a request from line A, the control system will detect that line B is busy . When this connection goes through switch 1253 to line n-2, the residual beam leaking through the line B connection switch will alert the system that line B is in service. If no operation is detected, a system request will be sent to both lines A and B (possibly over the same monitor line if it has multiple send/receive capabilities, or possibly over a separate system line) and switch 1254 can be closed to establish the connection .

Since even a switch that partially switches on the desired row corresponding to all outputs from a given input will interfere with some established and possibly operational communication connections between other channels, the basis for a one-to-one connection m× mThe wiring inside the switching unit is not suitable for the broadcast function. The system line "outside" from the mxm switching unit illustrated in FIG. 39A is most suitable for broadcasting. (The "inner corner" of the geometry is the preferred case for the lowest loss waveguide connection between Line 1 on the "Input" side and Line 1 on the "Output" side). Using the broadcast example, line C is shown operatively connected to most or all of the "out" lines in FIG. 39A via a light barrier 1256 . Switch 1256 on line C must be only partially on so that sufficient power is delivered to each "output" line. A similar protocol can be used to prevent collisions between channels in the case of simple communication connections when broadcasting. To establish a broadcast connection with only inactive channels, the system can group channels together and/or wait for individual channels to allow broadcasting to them.

To improve switching efficiency, the waveguides may be larger multi-mode waveguides, in the case of a single-mode communication network connecting the single-mode input and output ports 1 to m with an adiabatic extender as described elsewhere.

The entire structure described above with reference to FIG. 39A can be used as an asynchronous transmission mode switcher, or any point-to-point communication application. An efficient variation of this structure is for multi-wavelength operation in WDM networks. Wavelength selective optical switches may be implemented as described herein using polarized grating switches, or using tunable fixed gratings tuned in and out of specific communication frequency bands. In WDM networks, what is required is to switch a particular wavelength between channels without affecting other wavelengths that may be transmitted (bidirectionally) in the same channel. For a tunable switch that can selectively reflect frequencies while transmitting substantially other sets of frequencies in the WDM spectrum, the simple geometry of Figure 39A is suitable. However, if a switched grating with a single operating frequency is used, separate connection paths are required for each wavelength.

Figure 39B shows a switched WDM communication network 1260 with separate paths for each frequency used in the network. This example is for a dual frequency WDM network, but can be generalized to any number of communication frequencies. Figure 39B shows an "input" waveguide 1276 connected to three ports 1a, 2a, and 3a, and an "output" waveguide 1276 connected to three ports 1b, 2b, and 3b. The waveguides form 9 intersections. At each intersection, there are three additional optical paths connected to each "input" and each "output". The additional paths are the same in this example, and include three types. The first type of optical path 1266 includes a pair of fixed frequency switching reflectors that reflect the first of the two signal frequency bands of the WDM system. The reflector is preferably a grating transverse to the "input" and "output" waveguides associated with the intersection point and reflects power in a first frequency band between the corresponding waveguides and additional waveguide segments connected to the two gratings. The second type of optical path 1268 includes a pair of fixed frequency switching reflectors that reflect the second of the two signal frequency bands of the WDM system. In addition, the reflector is preferably a grating placed transversely to the respective waveguide and reflects power in a second frequency band between the respective waveguide and an additional waveguide segment connected to the two second gratings. The third type of optical path 1270 includes a pair of frequency independently switched reflectors, the two reflectors can reflect the two signal frequency bands of the WDM system. A third type of optical path can be implemented by a TIR reflector pair connected by a waveguide and a fixed mirror (described with reference to Figure 37).

In this case, ports 1a, 2a, 1b, and 2b plus associated waveguides 1276, 1277 form a 2x2 switching network that simultaneously switches dual frequency channels between any "input" port and any "output" port . System control ports 3a and 3b with associated waveguides 1276, 1277 provide monitoring and system communication functions. For example, if it is necessary to switch the first frequency of the WDM system between ports 2a and 1b, the two switches connected to ports 2a and 1b at the intersection of the waveguides associated with the first type of optical path 1266 conduct, through the waveguide connecting the two switches Optical power is routed at a first frequency between ports 2a and 1b. If all of the frequencies associated with a given port are routed to another port, the switch and type 3 optical path 1270 conduct at the intersection corresponding to the two ports. Since switching between two WDM frequencies between any two channels triggers two corresponding optical paths 1266 and 1268, optical path 1270 is actually redundant in a 2x2 network. However, in more advanced communication networks with many WDM frequencies, a single full frequency connection is required as it will have minimal losses.

A two-dimensional one-to-many routing structure is shown in FIG. 40 . The first row of waveguide routing switches connects an optical power from the input waveguide to the columns of pixel waveguides. The details of the switch are not shown here, only in the form of a grating, but can be implemented in several different ways. A two-dimensional array of "pixel" switches directs the work leaving a pixel at a "pixel location". (What happens to the power at the pixel location depends on the application). Double layer switches are used to reach all pixels. This structure can be used to display, start or control a process or device, or read certain forms of data. In the latter case, the power flow is reversed and the device operates as a one-to-one directional configuration.

An input beam 1342 is transmitted in an input waveguide 1352 and coupled by a two-dimensional array of switching elements 1356 to one of many pixel waveguides 1354 . Switching element 1364 can be used as a raster switcher, as described above in connection with FIGS. 7, 8, 12, 13, and 38, or as a TIR switcher, as described above in connection with FIGS. for other switchable elements. The beam 1344 is switched by switching element 1358 to a pixel waveguide where it is switched a second time by switching element 1360 to form beam 1346 which passes to pixel element 1362 . The pixel element 1362 can be separated from the waveguide 1354 by the waveguide section shown, or it can be a small distance above the waveguide so that a small fraction of unswitched light passes through the pixel element.

In the case of a display application, the pixel elements may be radiation for generating light 1342 from a substrate 1348 . The pixel elements may be rough patches on the surface of the substrate 1348, or beveled micromirrors, or rough beveled micromirrors for light scattering, or fluorescent pits, or other devices that generate visible light. In the case of a display, the input light beam 1342 may contain several colors, in which case the waveguide can be directed to all colors and the switch can couple all colors. The waveguide switches are scanned sequentially to generate the image for the display. Grating switchers can be implemented as multi-period gratings, but TIR switchers require little modification for this purpose. If single-mode waveguides, they must efficiently wave toward the shortest wavelength beam. The input beam 1342 is preferably externally modulated (including all its color components) so that the switching element is a pure on-off device. Note that if the pixel elements are arranged in approximately linear fashion and are electrically connectable along the row, then a single row of electrodes can be placed across each column waveguide to actuate a row of pixel switches.

In the case of projection displays, an additional lens is required to collect the light emitted by all the pixels in the array and refocus them onto the screen at a (large) distance from the lens. The lens should preferably have a good off-axis characteristic so that the focal plane is reasonably flattened on the screen, and it should have a numerical aperture (NA) large enough to collect a large portion of the light radiated by the pixel array. part. It would be advantageous to couple a matrix of lenses to the pixel structure to reduce the scattering of the beams produced by the individual lenses, thereby reducing the (expensive) NA requirements on the projection lenses. Another way to achieve this is to taper the waveguide again to the largest possible size of the pixel. It is relatively easy to scale down pixels to large lateral dimensions, but it is difficult to obtain a deep waveguide. Large pixels can be formed by coupling a wide waveguide with a long grating coupler.

The light distributed in the routing structure can also be used to activate the process, for example in the case of DNA readers or allergy readers, or protein readers. In each of the specific cases described above, separate arrays of DNA or allergies or proteins are prepared using fluorescent labels that can be activated by light. One molecule or one preparation of molecules can be arranged for activation during each pixel. The light is electrically scanned during the different pixels, the speed and sequence of the scanning being determined from the results. Fluorescence is collected for detection through an external lens and detector. However, for some applications, the collection and waveguide of the emitted radiation to an optical energy detection device and the control of the radiation from the source light are advantageous both for the image speed (and its lens) and for the waveguide structure itself. Depending on the desired illumination and collection modality, this lens can be a collimating lens, a refocusing lens, or even a conceivable lens that produces a diverging beam. The collimating lens is separated from one end of the waveguide by the focal length of the lens so that the emitted (and collected) beams are substantially parallel. Collimating lenses are useful if there is a large amount of material that will be traversed by the interrogation beam. The refocusing lens is separated from one end of the waveguide by an object distance, the inverse of which is related to the difference between the inverse of the image distance and the inverse of the focal length, where the image distance is the distance from the lens to the desired imaging beam spot. If it is necessary to focus the sample into a small spot and illuminate and/or read it from the waveguide, a refocusing lens is used. The scattered beam is generated using a lens separated from the waveguide by a distance less than its focal length. If the scattering of waves approaching a lens is different in the two planes, the output beam from the lens need not be circumscribed. The easiest way to produce beam wrapping (for the smallest spot after refocusing) is to achieve a wraparound beam start at one end of the waveguide, which can be achieved by designing the waveguide, or by tapering the waveguide. The lens preferably has a suitable numerical aperture to admit the entire wave from the waveguide and focus it to a diffraction limited spot or collimated beam depending on the application.

Pixel element 1362 can be any of the elements described above, and it can be bonded directly to the material to be activated, or indirectly by cooperating with an external sheet to which the material is already bonded. Each image velocity element may contain a collimating lens as described above, so that a switcher array may be coupled to a lens array through imaging beam spots in a substantially common focal plane. (In this case, substantially common means above and below in the Rayleigh region of the true focal plane, which may be greatly distorted due to aberrations). If the routing structure is also used to detect fluorescent radiation, a reflector is preferably used in the pixel element 1362 instead of a scatterer, which couples the radiation back to the waveguide from which it came. This coupling is maintained as long as the switch for a given pixel is activated. If desired, the light source can be switched off before switching to another pixel element in order to account for radiation attenuation.

When used as a data reader, the direction of light propagation is reversed from that shown in FIG. 40 . Light, including data, from a device is collected at each pixel element and coupled to a routing waveguide structure that directs its wave back to the input waveguide 1352 . Connected to the waveguide 1352 is a detector for reading data. The detector can be simultaneously connected to the waveguide 1352 via a beamsplitter located between the waveguide 1352 and the light source for data medium illumination. Pixel elements 1366 (or just "pixels") are preferably lens-coupled to each data spot to collect light routed through structure 1350 and deliver it to the data medium. A coupled lens is also used to collect light reflected or emitted from the data medium and refocus it onto the end of the waveguide coupled to the pixel element. The data can be in the target volume, in which case the lens can be configured parallel to the beam, the data can be on the target surface, in which case the different pixel elements can correspond to the rotating magneto-optical disk data storage Different tracks on a surface (eg CD). The lenses are constructed to refocus light from the pixels onto the data spots in a diffraction limited manner. By combining different pixels with different tracks, track-by-track switching can be achieved electronically with essentially no delay.

Different pixels may also be coupled to different sides of the data medium. This is useful for reading data that has been recorded in multiple sides on the medium to increase the overall storage capacity. Switching between planes can also be accomplished electronically by switching between pixels coupled to different planes.

In addition, several different pixel elements may be grouped into locations separated by (preferably orthogonal to) fractions of the track division cross-section of a given track. When the track drifts, positive tracking can be achieved electronically by switching between pixels instead of mechanically. A sensor and electronics are needed to detect the track offset and a controller is needed to switch to the desired pixel. The signal strength or signal-to-noise ratio (SNR) in different channels can be detected to determine the preferred (most accurate) channel. If the transducer along the waveguide 1352 is made as a 4-way junction instead of a 3-way junction, with the fourth leg exposed at the edge of the substrate, the detector array 1368 can be positioned in alignment with the fourth leg. Each detector 1367 is individually aligned with each column for detecting return power from each column. If detector 1367 is used, to maximize return power from the data medium on detector array 1368, the preferred reflectivity for the grating positioned along waveguide 1352 is about 50%. If a single beamsplitter is placed upstream of the router structure, in the waveguide 1352, its preferred reflection is also 50%.

Note that partial activation of different pixels can be achieved by partial activation of switches along the input waveguide or pixel waveguide. Switching elements 1364 can be tuned by applying an electric field to change their reflection coefficient. Some of the beams may be routed through the desired partial active switch for simultaneous use in another pixel. Multi-pixel activation has special significance in the case of track offset correction, since multiple detectors can also be implemented in router 1350 . For example, if three different pixels on three different columns of routing structure 1350 were to be activated simultaneously, their corresponding pixel column switches would need to be partially activated. The controller needs to perform calculations to determine the appropriate number of switches to activate. Neglecting the losses of the switches, for optimal SNR, in order to produce equal intensities on their respective detectors, the first switch corresponding to the first column of pixels should be activated to reflect about 3/16 of the incident light, with the first The second switch corresponding to the second pixel column should be activated to reflect about 1/4 of the remaining light that has passed through the first switch, and the last switch corresponding to the third pixel column should be activated to reflect About 1/2 of the remaining light that has passed through the first two switchers. Assuming 100% reflection from the media and 100% light collection efficiency, approximately 15% of the incident light is reflected to each detector. This is quite good compared to the preferred 25% of the beam received on a single detector in the case of a single pixel (preferred switch activation = 50% reflectivity). Indeed, more light is collected with three beams than with only one. Electronic tracking will result in lower cost, faster and more reliable data read/write devices.

Any combination of these methods (electronic tracking switching, electronic data plane switching, and electronic tracking) can be used to improve the performance of the data storage device. There is also a need for a way to achieve electronically variable focus for potentially removing all mechanical motion (other than media rotation) from the drive. As described below with reference to FIG. 54, electronically variable focusing can be achieved by changing the wavelength of light beam 1342 using a zone plate lens.

As shown, the routing structure of Figure 40 is a polarized structure, and the 90-degree grating switch only reflects the TM mode. As a result, only intensity-dependent spectroscopic methods can be used. It would be very advantageous to be able to use a polarizing beam splitter as this would produce a factor of 4 increase in signal strength for a given light intensity. However, there is a need for a switching structure capable of shifting and then separating these two polarizations. While the polarization relationship of each TIR switch can be achieved with a negligible substantially tangential TIR angle, there is a storage density penalty in using very low angle switching geometries.

FIG. 41 shows a linear array of strongly polarized switches associated with a data reader 1370 configured. The switch is activated by beam 1342 which is TM polarized and highly reflected in excitation switch 1372. Waveguides for both polarizations are waveguided using waveguides 1376 and 1378, such as titanium scattering waveguides in lithium niobate. The pixel elements are realized by micromirrors 1347 combined with integrated lenses 1380 . Data spots (eg, 1382 ) are arranged in tracks on a disc 1386 that rotates about an axis 1388 . Vertically polarized light reflected from birefringent data spots (or separators) on the data track is collected by lens 1380, refocused back into waveguide 1378, and reflected by micromirrors back onto the waveguide surface with TE polarization. Since the TE mode is both polarized at the Brewster angle of the grating and has a different non-phase matched transmission constant for reflection, it is transmitted through the switch and not reflected back to the detector 1367 of the detector array 1368 . (Alternatively, if the switch is a TIR switch, the TE waves are less reflective than the TM waves, and most of the TE waves are sent through the switch to impinge on the detector.) If another switch 1373 Activated instead of the switcher 1372, the light beam will travel to a different pixel 1375 and be focused to either another data track, or another data plane, or the same track but with Lateral deviation in fractions of track width (depending on whether the pixel 1375 is used for track switching, data plane switching, or tracking control).

Many variations on the structure described with reference to Figures 40 and 41 are evident, such as any switch in the router that can be oriented to change the direction of light propagation in the plane; a variety of type switchers, and higher switching levels can be attached. Other variations abound.

FIG. 42 shows a switchable integrated spectrum analyzer 930 . The input beam enters the input waveguide 932 and stops after a certain distance. The input beam 912 can be relayed in another waveguide, or it can be a free space beam aligned and mode matched so that the power reaching the waveguide 932 is optimized. The device 930 is provided with a planar waveguide 835 for confining the retransmission to a plane. The light beam 927 exiting the end face of the input waveguide is scattered in one plane within the planar waveguide until it passes the integrated lens element 925 . The integrated lens has an increased refractive index relative to the planar waveguide within a boundary defining an optical thickness that decreases approximately squarely from the optical axis. (If it has a reduced index, the optical thickness increases approximately quadratically.) The lens can be fabricated by mask diffraction or ion exchange, or it can be a reverse poled segment excited by electrodes. segment).

Lens 925 collimates the light beam, which then goes to at least one of the three grating sections 929 , 931 , and 933 . The grating is formed of individual cells, each of which is a domain, which are distinguished from the background material by varying amounts of spacing depending on the application. The cells have a constant or adjustable index of refraction different from the substrate, and different cells may have different domain types. Invariant domain types include, for example, scattering regions, ion-exchanged regions, etched regions, particle-irradiated regions, and regions generally formed by modification of any refractive index type. The grating portions may be formed by etching, ion exchange, or inward diffusion, in which case these gratings are permanent, but in the preferred embodiment the gratings are shown formed from poled domains . Electrodes 932, 934, and 936, along with common electrode 938, are used to drive the gratings, respectively. For simplicity, common electrodes may be placed on opposite sides of the substrate as shown, or around electrodes 932, 934, and 936 for low voltage excitation.

The elements in individual gratings can be arranged in a pattern to create the desired periodicity in the desired direction to provide virtual photons with the desired momentum. They can be arranged in rows to define certain planes with virtual photon momentums perpendicular to the plane of momentum defined by the row spacing. At this point, there will also be virtual photons with momentum along the plane with momentum defined by the spacing between cells in the row. To phase match the backreflection, the virtual photon has exactly twice the momentum of the incident photon and is waveguided in the opposite direction. Any other reflection process has less momentum and is directed transverse to the axis of incidence. Thus, the period Λ of the row spacing is partly related to the incident wavelength λ, where Λ is a fraction of the quantity λ/2n _eff . In general, cells can be separated by a distance distribution that varies with position through the grating such that the virtual photon momentum along any axis of incidence is determined by the spatial spectrum of the cells along that axis (via Fourier transformation determined) determined.

At least one of the gratings 929, 931, or 933 is turned on by adjusting the potential state of the corresponding electrode. In Figure 42, the grating 929 is activated. The energized grating provides virtual photons to incident photons that phase match the scattering process to an output direction to form multiple output beams 935 and 937 with different wavelengths, the output beams being angularly separated according to their wavelengths. The output beam from the energized grating 929 passes through a lens 939 which refocuses the output beam onto a detector array 941 . The detector is a set of sensors arranged to receive a portion of the output beam for detection and is preferably glued to the edge of the device 930 as shown. However, if device 930 is to be integrated on a larger substrate, the substrate need not have an edge at this location. At this point, other methods of beam extraction, such as vertically refracting mirrors, can be used to refract a portion of the beam 935, 937 into the detector array. The sensing device is positioned approximately within a Rayleigh region of the focal plane of the output lens 939 where the input beam angle can be mapped to the output beam position. Since the grating maps the input wavelength to the output beam angle, collimating the input beam causes different wavelengths to be mapped to different locations in the focal plane, and the spatial resolution of the wavelengths depends on the properties of the grating. The detected power (as a function of the position of the detectors in the array 941 ) is related to the power spectrum of the input beam 921 . Thus, device 930 is a spectrum analyzer. It may also be a multi-channel detector when the input beam is divided into channels occupying several different channels, and the arrangement is configured to distribute the channels to predetermined detectors or a group of detectors.

By switching different gratings, the device acts in different frequency ranges. For example, if grating 931 or 933 is activated, the scattered light is focused by lens 939 to a different detector array 943 or to a different portion of extended detector array 941 . The frequency range of the grating is determined by the angle of the grating to the beam and the periodicity of the grating. The grating 931 is shown having a shallower angle to the light beam so that it is excited to select a higher optical frequency range. The gratings 933 have multiple periodicities transverse to each other to enable selection of multiple overlapping frequency ranges. Multiple frequencies may be mapped to polar region boundaries, as described above with reference to FIG. 18 . The polarizing elements of the grating 933 can generally be arranged on a plane perpendicular to the two main virtual photon momentum directions. The phasing of the plane is determined by the process of transcribing the desired grating component frequencies to the domain boundaries. In general, however, gratings can have momentum components in all directions, in which case the resulting domain boundaries will not organize into the plane, except perhaps in one dominant direction.

A transmitted light 913 is refocused by an integrated lens 907 to an output waveguide section 909 to form an output beam 911 that includes at least a portion of the out-of-band portion of the input beam 921 that did not interact with the grating.

A useful variation of the switched-range spectrum analyzer is combined with the elements of Figures 42 and 30-35, the basic concept of which stems from the fact that the spectral range of a grating can be changed by changing its angle or equivalent to the source point displacement. In this variation, a waveguide routing structure is used to allow the source to be switched. Waveguide switchers are placed at one or more locations on the input waveguide 923 (and possibly the launch waveguide) to create an array of parallel source waveguides in which the input beam 921 can be switched. All waveguides terminate in the same plane, preferably at the focal plane of the input lens 925 . The rest of the spectrum analyzer remains the same, although it has multiple inputs, it does not necessarily have the additional gratings 931 and 933 . The separation of the multiple switched waveguides is adjusted according to the application to achieve the desired switchable spectral range of the analyzer 930 .

Figure 43 shows a polarized acoustic multilayer interference structure 953. The incident acoustic wave 972 may be a bulk or surface acoustic wave. The poled structure is fabricated in a piezoelectric substrate 965 in region 955, which includes domains 963 and 964 of two morphologies. It is well known that polarity reversal causes partial reflection of sound waves (eg US Patent 4,410,823 to Miller et al.). Reflection to beam 973 and transmission to beam 961 is affected by the spacing of the interface between the polar regions. If high reflection and low transmission are required, the interval between adjacent interfaces should be equal to an integer multiple of 1/2 the acoustic wavelength. If it is desired to pass through a structure with high transmission and low reflection, the spacing should be equal to 1/4 the acoustic wavelength plus an integer multiple of 1/2 the acoustic wavelength. Non-reflecting (AR) structures can be fabricated by providing an appropriate number of polar regions near an interface where the acoustic impedance changes, provided that the phase of the reflected wave is chosen to be out of phase but of the same magnitude as the reflected wave from the interface.

Figure 44 shows a bulk acoustic wave transducer 971 that is polarized. An incident acoustic beam 972 is incident on the polar region of a piezoelectric substrate 965 comprising a pair of electrodes 974,975. The pole region comprises two domains 963 and 964 which are preferably opposite. By reversing the polarization direction every 1/2 acoustic wavelength, the electric field induced by the acoustic wave at each polar region can be chosen to be the same. At this time, a single electrode can be used to pick up the induced voltage instead of the interdigitated electrodes of the prior art. The presence of an incoming wave 972 is detected using electrodes 974 and 975 . The output voltage (derived from conductor 979 and seen from controller 978) varies sinusoidally (for a narrow frequency band) and is a function of time and amplitude relative to the amplitude of the sound wave. As noted above, if the polarized interfaces are separated by 1/2 wavelength, the structure also acts as a high reflector, which may not be required in a given implementation. This feature can be eliminated by alternately spacing the interface at 1/4 and 1/3 of the wavelength shown in FIG. 44 . In this case, the structure is a non-reflective coating to eliminate unwanted reflections. The structure 971 is an effective acoustically tuned detector since almost the entire sound wave penetrates the polarized structure, where its energy can be almost completely absorbed into the detection electronics. The bandwidth of the structure is inversely proportional to the number of acoustic cycles that fall within the polarized structure covered by the electrodes. Its efficiency is related to the acoustic path length under the electrode. Therefore, the bandwidth is related to the efficiency of the detector and can be adjusted by changing the size of the detector.

Structure 971 acts as a sound generator, which basically runs the reverse process. A time-dependent electrical signal is provided between the two electrodes at the frequency of the acoustic wave to be excited. The piezoelectric coefficients of the substrate create a periodic tension at the acoustic frequency and generate a pair of acoustic waves, one 961 traveling in the forward direction and the other 973 traveling in the opposite direction. If it is desired to generate only one wave, an efficient unidirectional generator can be made by combining devices 953 and 971, while device 953 can be made as a total reflector for the unwanted waves. If the total reflector is oriented at a 90 degree angle to the unwanted wave, and the reflected wave is phase-selected to be in phase with the desired sound wave, then the two waves will appear in the same direction as a single wave.

A variation on the structure in Figure 44 is a strain-activated light interaction device. In this device, poles 964 and 963 are excited by a strain field, which produces a change in refractive index through the photoelastic effect. Structure 975 is now a strain-sensitive damper that can be deposited onto substrate 965 at elevated temperatures such that the different coefficients of thermal expansion of the film and substrate create a strain field at room temperature. After the mechanical strain field is acted upon by the photoelastic stretcher, the domain-to-domain refractive index change in the substrate again produces a substrate with a shaped refractive index, which can also be used in other applications described herein. The electric field using the electro-optic effect can be combined with the photoelastic effect if the deposition process of the electrodes does not affect the required strain field.

The structure 890 of Figure 45 is a tuned coherent detector for a pair of light waves. It is tuned in such a way that it can only detect frequency differences between light waves within a certain bandwidth (approximately the desired central "resonance" frequency difference). In the simplest case, the device is constructed with equal spacing between the interdigitated electrodes 885 and 886, which form a periodic structure with a period Λ. At a given instant, the two input frequencies present in the input beam 887 produce an interference pattern of electric fields in the waveguide 888, the spatial period of which depends on the difference in optical frequencies and the refractive index of the substrate 889 at the optical frequency. At a frequency difference (where the spatial period of the interference pattern is equal to the period Λ) the electrode structure is in resonance, and due to the induced displacement charge at the top of the waveguide, the electrodes will be excited to a potential difference.

The frequency response characteristic is related to a sinusoidal square function, and its resonance frequency is determined by the optical frequency difference, and the phase shift of the two optical waves in the polarized grating period is 2π. After the electrode structure is formulated, the buffer layer 891 is needed to reduce the loss of transmitted light waves. If its thickness is much smaller than the period Λ, the induced potential intensity cannot be reduced substantially. The interference pattern has a low frequency component that vibrates at the frequency difference between the two light waves. Therefore, the electronic signal picked up by the electronic controller 978 via the waveguide 979 also vibrates at this frequency difference. The amplitude of the electronic signal is large at the resonance frequency difference and drops to other frequency differences depending on the device bandwidth, which is related to the inverse of the number of pulsation cycles in the interdigitated electrode structure.

Interdigitated electrodes can also be configured with multi-frequency components, in order to have several resonant frequencies, or to modify the response bandwidth. Note that this device can also be multi-order sensitive. If the electrode is narrow compared to 1/2 period, there will be a significant response at odd harmonics of the resonant difference frequency. Responses to even harmonics can be generated by having asymmetry along the waveguide axis by moving the fingers relative to each other. Higher order responses can only be improved by reducing the first order response, which can be minimized by centering the electrodes relative to each other and increasing their bandwidth. Finally, the waveguide 888 is not strictly necessary. It can be omitted, but the detected wave should be brought very close to the electrodes to optimize signal pickup.

Figure 46 shows a low-loss switched waveguide splitter 780, the device has a permanent Y-shaped waveguide splitter 774, which consists of an input waveguide section that widens into a Y-junction and branches into two Output waveguide sections 775 and 776, 775 and 776, may serve as paths for light incident on the input section. The width and index of the input and output sections are preferably equal. The splitter 780 also has a polarized structure 778 whose electro-optic coefficient lies in the region of the Y-shaped splitter 774 . Pole region 778 can be a thin layer near the top of the substrate, it can have multiple layers, or it can extend across the entire substrate. Other parts of the substrate may be polarized or non-polarized. A pair of planar electrodes 777 , 779 are disposed adjacent to each other over the waveguides, with one electrode 777 covering a portion of one output waveguide 775 and the other electrode 779 covering a portion of the other output waveguide 776 . The electrodes are planar for ease of manufacture and for functionality: if the surface on which they are placed is flat or curved, they are identical. Edge 781 of electrode 777 crosses waveguide 775 at a shallow angle and forms a smooth continuous inner edge of waveguide 776 at the Y-junction. Likewise, edge 783 of electrode 779 spans waveguide 776 at a shallow angle and forms a smooth continuous inner edge of waveguide 775 at the Y-junction. When the electrodes are energized with one polarity relative to each other, the index of refraction under electrode 777 decreases and the index of refraction under electrode 779 increases. As a result, an excitation region below the electrode edge 781 forms a waveguide boundary that diverts the input beam 789 almost completely into the output beam 784 with very little power leaking into the other output beam 782 . The increased index of refraction below electrode 779 helps direct light energy away from boundary 781 . The input beam diverts almost completely into the other output beam 782 when the opposite electrode is applied between the electrodes. If no voltage is applied, the input power is divided evenly between the two output ports, assuming a symmetrical structure. Thus, the structure is a 3dB splitter that can be electrically switched into one of the directions with a light guide at low loss.

The electrodes 777, 779 taper away from the Y-shaped structure 774 at the structure input to form a progressive low-refractive region towards the waveguide to minimize optical loss. The smoothing effect of the electrostatic field distribution produces a very smooth refraction transition refractive index under the two electrodes. Electrodes across the output waveguide away from the Y-branch region are preferably clamped at a 90 degree angle to the waveguide to reduce losses.

The Y-splitters can be arranged in an asymmetric manner to produce a split ratio different than 3dB when the electric field is turned off. This can be done by increasing the offset angle of one of the waveguides and/or decreasing the offset angle of the other waveguide. The switching function operates almost as if there is an asymmetric structure and a symmetric structure, assuming a sufficiently large electric field is applied to the electrodes. Despite the large asymmetry, the extinction ratio (the ratio between the power in the on waveguide and the power in the off waveguide) can be maintained very large. However, the optical loss will be slightly different at the two legs of the asymmetric switched waveguide splitter. Thus, device 780 can be configured as a splitter with any desired split ratio, and can be switched with good efficiency and high absorbance ratio.

This device can be cascaded to allow switching between more than two output waveguides. For example, if the output waveguide 775 is connected to the input of a second device like 780, its power can be passively or actively switched into an additional pair of waveguides. Sixteen switched output lines can be accomplished with four banks of switches (1, 2, 4 and 8 respectively) similar to the 780. The power distribution ratio between these lines is equal in the unswitched state, or any other power distribution ratio. When these switches are activated, a single output waveguide can be switched on, a single output waveguide can be switched off, or any combination of output waveguides can be switched on and off.

The direction of light propagation in the device can be reversed, at which point the input on either of the output ports 775 and 776 can be switched to emerge from the input port. If no voltage is applied, the power at each output port is coupled to the input port with an attenuation (3dB in the case of a symmetrical device). When the electric field is switched on, the power in the "on" waveguide is routed to the input port with very low loss, while the power in the "off" waveguide diffracts away from the input waveguide very efficiently. The "cut off" waveguide is substantially spaced from the input port.

Alternatively, a mirror device can be connected back-to-back to the switch so that the input waveguides are joined together to form a 2x2 switch or router. Inputs on any pair of waveguide ports can be switched to any waveguide of the other pair of ports. Likewise, cascading can be done to create an nxn switch/router.

Figure 47 shows an alternate embodiment 790 of a switched waveguide splitter using multiple pole regions. In this structure, the switched index difference along the boundaries of the waveguides in the Y-shaped region is enhanced, thereby better confining the optical mode to a narrower region and reducing the residual coupling into the severed output waveguide. . Along each side of the input waveguide 774 of the Y-shaped separation region there are two polar regions 785 and 786 whose boundaries 787, 788 cross the output waveguides 775 and 776 at shallow angles and form the waveguide 776 at the Y-junction , the smooth continuous inner edge of 775. The boundaries of the polar regions taper slowly from the input waveguide to allow slow initiation of electrically actuated refractive index changes, and they cross the output waveguide away from the Y-junction (where the electric field is substantially reduced) to reduce optical loss. Electrodes 791 , 792 are disposed substantially over pole regions 785 , 786 .

A potential difference is applied across the electrodes to excite an electric field, electrostatic in form, across the space between and around the electrodes. This electric field penetrates the pole region and the surrounding area, causing a light-responsive form of refractive index change. The local optical refractive index change is related to the product of the local electric field direction and the local electro-optical coefficient. The polar region is preferably surrounded by regions of opposite polarity, in which case the sign of the electro-optic coefficient is opposite to that of the surrounding region. There is a sharp index change at interfaces 787 and 788, and on one side of the waveguide, the index of refraction decreases at the interface to create a waveguide tendency away from the low index region. The opposite is true for the other side of the waveguide. If the applied electric field is large enough, the interface with reduced refractive index forms the waveguide boundary. The input light beam is waved into the output waveguide due to the smooth connection of the waveguide interface as an extension from the inner boundary of the output waveguide spanned by the pole region. If the curvature of the guide boundary of the waveguide is gradual, light leakage into the "cut" waveguide is low. The loss at the input is low because the polar region approaches the waveguide slowly. Losses at the Y-junction are low because the portion of the pole extending beyond the junction reduces the waveguide effect for the "off" output waveguide and enhances the waveguide for the "on" output waveguide.

As an alternative, the polar region can be surrounded by unpolarized material, where there is still a sudden change in the refractive index at the interface 787, 788 for the device to function, but when the polar region is surrounded by an oppositely polarized material When surrounded, its refractive index is only half of the value obtained, so the applied electric field must be higher. The various modifications described above are also applicable to this device.

Figure 48 shows the main design elements of a 1x3 switch, the design elements shown here show how the device 780 of Figure 46 can be transformed into a 1x3 switch with a single polar region and shaped electrodes. The device consists of a permanently branched waveguide with the desired number of branches n (n=3). The waveguide passes through a polar region that extends deeper than the waveguides (for good light absorption ratio) and well beyond the junction region where the waveguides are separated by a large space (e.g. three times their width) . Several regions are defined by the waveguide boundaries, by which they extend smoothly back into the input waveguide boundaries, and by which they span the normal boundaries of the output waveguides at a distance outside the junction. There are (n ² +2n-2)/2 regions defined in this way. It is useful to have the outermost region extend beyond the outermost waveguide in order to taper the input. Separate electrodes are provided over each zone, with a small gap between all electrodes, but sufficient to prevent electrical breakage during actuation.

To operate the device, an electric field is applied to each region independently with a polarity depending on whether the region concerned is confined within the desired waveguide. For example, the five fields in Figure 48 are activated according to Table I. As before, the magnitude of the electric field is adjusted to produce a good waveguide boundary along the boundary of adjacent regions excited with different polarities.

district Number of electrodes top middle the bottom 1 + - - 2 + + - 3 - + - 4 - + + 5 - - +

Table I

Alternatively, the design elements of Fig. 48 also show how the device in Fig. 47 can be transformed into a 1 × 3 switch with multiple poles, the device comprising a permanently branched waveguide with the required number of branches n (n=3 ). Several regions are defined by the waveguide boundaries, by which they extend smoothly back into the input waveguide boundaries, and by which they span the normal boundaries of the output waveguides at a distance outside the junction. It is useful to have the outermost region extend beyond the outermost waveguide in order to taper the input. Each zone is polarized in the opposite direction to adjacent zones that share a common boundary. Regions with the same polarization direction can share at most one vertex. The input waveguide region is preferably polarized opposite to the innermost region (ie the region closest to the input waveguide). In Figure 48, the innermost zones are labeled Zone 2 and Zone 4, and the zone-based polarity selection procedure causes only Zones and 4 to be polarized in opposite directions, while Zones 1, 3, 5 (which are The output waveguide region) is forward polarized (same direction as the surrounding region, if the surrounding region is polarized). If 4 output waveguides are used, there are 9 regions, 6 of which are polarized in reverse, including all output waveguide regions. Thus, splitters with an even number of output waveguides have some advantages because only the output waveguide region of an even splitter is polarized opposite to a voltage substrate polarization, with the advantage of a limited increase in the final split point and higher transmission for In the "on" state, and in the "off" state there is better reverse isolation. Separate electrodes are provided above each zone.

To operate the device, an electric field is applied to each region independently and its polarity is reversed. Polarity is determined by two factors: whether the region of interest is confined within the desired waveguide, and the polarity of the underlying polar region. For example, if applying a positive polarity to a positively polarized region produces an increase in the refractive index, the selection rule is: if a region is positively polarized, its electronic excitation polarity is selected to be positive (if the region is in the desired waveguide inside) or negative (if the region is outside); if a region is reversely polarized (negative), its polarity is chosen to be negative (if the region is inside the desired waveguide) or positive (if the region is outside) . Table II shows the preferred polarization direction of the zone when n=3 and its three output ports are shown in FIG. 48 . The design of 1xn and nxn switches can be derived from the description with respect to FIGS. 46 , 47 and 48 .

district Polarization direction top Central bottom 1 - - + + 2 + + + - 3 - + - + 4 + - + + 5 - + + -

Table II

The planar constituent elements described herein can be stacked into multilayer three-dimensional structures including electro-optic controlled devices and waveguide assemblies. The three-dimensional structure of the above-mentioned planar waveguide and switch is manufactured by alternately laying or disposing electro-optical active polarizable films (preferably polymers) and buffer isolation layers (which may be dielectric materials or conductive materials). Advantages of the laminated structure include crosstalk isolation caused by further spaced waveguide elements. The stacked structure can also achieve higher power handling capability because more optical power can be distributed among the layers. Independent layers can also be used if it is desired to allocate independent waveguides in the display device.

Once placed on a suitable substrate, the polarization of the active optical waveguide/switching layer can be accomplished using the techniques described above. In order to isolate one active layer from other active layers, a buffer layer with a low index is necessary and designed to create the required guidance in a direction perpendicular to the above-mentioned plane. For example, a buffer layer of silicon dioxide may be used. This is followed by a ground plane and a thick buffer layer, which can be made with a metal layer since it is isolated from the above optically active layer. The above-mentioned buffer layer must also be able to withstand the voltage applied between the above-mentioned different electrode layers and the ground plane. In polymers, a large area can be polarized, and by UV de-radiation techniques as described above, the desired area is selectively depolarized to produce waveguide properties, even in a light-transmitting buffer layer ( For example after silica has been laid). Alternatively, polarization may be performed electrically. In the case of polymers, depolarization of one layer with UV de-radiation will not affect the layer following it because of the shielding provided by the underlying metal ground plane mentioned above. Standard masking and coating techniques are then used to lay down metal electrodes and conductive paths, followed by another dielectric buffer layer and the next active layer. The aforementioned buffer layer should be planarized to minimize losses in the following active optical waveguide/switching layer. The above-described process of adding layers can be repeated for as long as required by a given device.

A variation of the fabrication technique used to fabricate the active pathways and electrodes for the polarized device stack described above is to coat the electro-optic layer with an insulating layer, and then dope or implant impurities into the insulating layer to allow the use of standard photoetching masks. Membrane technology creates a conductive pattern within the aforementioned buffer layer. Incorporating the above-mentioned electrodes into the above-mentioned buffer layer can minimize the thickness of the stacked devices.

Hybrid devices composed of different electro-optically active materials are used to improve fabrication complexity. For example, the first electro-optical active layer comprising the waveguide device can be fabricated in a lithium niobate substrate, which also serves as a support substrate. Next, a buffer layer and an electrode layer are provided for the lithium niobate device. Subsequently, two insulating buffer layers sandwiching a conductive plane are applied to the device, followed by the next active layer, which may be a polarizable polymer. The next several layers are built, polarized and patterned as described above. The conductive planes between the buffer layers can act as electrodes to allow the area polarization of each polymer layer and can make the previous layers unaffected by the polarization process.

For example, stacked waveguide arrays can be used as steering devices for free-space beam steering. Electronically activated and individually addressable waveguide elements are tightly stacked together and aligned with an array of sources to form a controllable phased array to emit optical radiation. As previously mentioned, the relative phase of the beams can be adjusted by varying the voltage across the polarized regions. By adjusting these phases with linear scaling, the light emitted from the waveguide array can be swept rapidly in the plane of the upper array. Thus, a linear array of devices on a plane can only be scanned within that plane. However, when the polarized waveguide array plane is vertically integrated into a three-dimensional monolithic device, the beam emitted from the device can be directed in two dimensions.

An extension of this concept is the use of a laminated waveguide grating reflector for mode control of an array of multimode laser bars. The waveguide stacks are spatially aligned to interface with a laser diode array. By controlling the phase of each of the elements described above, the emission mode form of a multi-element laser rod can be controlled. For example, in devices that do not require confinement to single-mode waveguides, multiple modes or entire arrays can be layered together to increase the power handling capability of switched polarized devices.

Figure 49 shows an embodiment of a phased array waveguide stack section 1630, showing only a single column of waveguides for clarity. Optical radiation 1640 enters the stack 1630 through a number of waveguides 1638 made in an electro-optically active film 1650 such as a polarizable polymer. Here, the input beams 1640 are shown spaced apart to represent beams of the same wavelength but with different phases. The light travels along the waveguide 1638 where they encounter a polarized region 1634 where the index of refraction can be electronically modified using the techniques described herein. Beam 1642 represents the output of the phased array after the phase of each light wave has been individually adjusted to produce output component beams whose phases are aligned.

Many other input and output wave situations are possible. For example, a single-mode laser beam with a flat phase wavefront can illuminate a waveguide element region, which then imposes an arbitrary phase delay across the spatial mode of the laser beam, thereby allowing the beam to be electronically manipulated in free space. Directional beam steering devices utilizing this method will be faster and more compact than current mechanical or A-O devices. The presence of phase differences or multi-frequency components within or outside the stack can be detected using opto-electrical pickup devices described herein or known in the art to provide instantaneous information for a feedback loop.

The device segments 1630 shown here are fabricated on a substrate 1632 (eg, silicon dioxide) by alternately arranging electrodes, buffer layers, and polarizable material in a manner to be described later. A wide area planar electrode 1654 (made of opaque metal film or transparent conductive material, such as indium tin oxide) is deposited, followed by an electrically insulating buffer layer 1652 (such as silicon dioxide), which also as a boundary layer for the waveguide 1638 constructed in the underlying polarizable material 1650 . On top of this layer of polarizable material 1650, another buffer layer 1652 is added to form the upper waveguide boundary, followed by the application of shaped electrodes 1636 for exciting the polarized structure. Another buffer layer 1652 is then added, this time electrically insulating the shaped electrode from the next layer, another wide area planar electrode 1654 . The shaped electrodes 1636 are separated from one planar electrode only by a thick buffer layer, and from the other planar electrodes by a buffer layer and polarizable material. Because of the need to apply an electric field across the polarizable material, the electrical isolation across the polarizable material should be smaller than across the buffer layer alone. The layering sequence between the two wide area planar electrodes is repeated until the last layer of polarisable material 1650, after which only the buffer layer 1652, shaped electrode 1636 and optical final insulating layer 1652 need be added to complete the stack. Electrical leads 1646 and 1648 contact electrodes 1636 and 1654, respectively, and are connected to voltage distribution control unit 1644 by integration and bonding techniques known in the art.

The voltage control unit 1644 has a dual purpose: individually actuating the aforementioned polarized devices, and isolating them from the electric field used to control adjacent actuation element layers. In essence, the voltage control unit 1644 can be a collection of coupled floating power supplies where the voltage between the electrodes 1636 and 1654 sandwiching one active layer can be controlled without changing the voltage difference across any other active layer.

Region 1634 represents a polarized region with one or more domains, and electrode 1636 represents an unbroken or segmented or patterned region with one or more insulating elements. The waveguide stack 1630 is described as a device for phase control, but the waveguide stack structure may include any number of combinations of polarized devices (optical in series or otherwise configured) described herein.

Figure 50 shows a prior art adjustable attenuator 1400. An input waveguide 1402 is transverse to an electro-optical active region of substrate 1404 . An input beam 1406 propagates along the input waveguide into an output waveguide 1408 forming an output beam 1410 . Electrodes 1412, 1414 and 1416 are arranged above the waveguide so that when electrode 1414 is excited at a given polarity (positive or negative) relative to electrodes 1412 and 1416, due to the electro-optic effect, There is a refractive index change within segment 1418, below and adjacent to the electrode. The electrode structure is arbitrary and can be different and can be more complex than the prior art structure shown in FIG. Increases the refractive index of the surrounding area.

In the absence of an applied electric field, the losses of the above-mentioned waveguide segments are low, mainly determined by scattering caused by roughness along the waveguide walls. However, the above losses can be increased to a very large value when an electric field is applied. A three-electrode pattern allows a negative index change to occur within the waveguide while a positive index change occurs outside the waveguide, significantly flattening and broadening the index profile. The modified portion of the waveguide 1418 below the electrodes has a broad distribution of the lowest order modes from the waveguide's input 1402 and output 1408 when an electric field is applied. As a result, modal coupling losses occur both when the input beam 1416 passes into the portion 1418 and when the light in the portion 1418 couples back into the output waveguide 1408 . If the refractive index change is large enough, the lowest order mode drops below the cutoff, and the light emitted from the end face of the waveguide 1402 diffracts almost freely onto the substrate, causing a large coupling loss at the beginning of the waveguide 1408.

When a given mode enters the modified portion 1418 of the waveguide, the change in the refractive index profile of the modified segment reduces the overlap between its intensity profile and any mode profile of the modified portion 1418 . If segment 1418 is multimode, several transmission and radiation modes will be excited. If segment 1418 is single mode, multiple modes will be activated. The combination of these modes is then propagated to the far end of the segment 1418 and coupled to the output waveguide section 1408 where only a fraction of the light is returned to one mode of the waveguide to form the output beam 1410 . By controlling the voltage applied to the electrodes, the losses in device 1400 can be tuned from very low to very high.

The maximum loss that can be achieved depends on the magnitude of the refractive index change, the size of the excited regions, their length, and on whether the input and output waveguides are single-mode or multi-mode. In a variation of its geometry, it is only possible to place two electrodes on the waveguide section 1418, thereby reducing the refractive index in this waveguide section and increasing the refractive index to one side instead of both. It functions again as an attenuator, but the rejected radiation field will tend to leave the device towards the side of the increased refractive index. This ability to direct the lost radiation is advantageous in some systems where it is desired to control the rejected light. An absorber could also be placed downstream (on one or both sides) of section 1418 to prevent rejected light from interfering with other functions elsewhere in the system.

Figure 51 shows a polarization switched attenuator 1420. The device is a modification of the prior art device shown in Figure 50, in which the polarized regions are used to increase the resolution of the refractive index change and to increase the discontinuity of the refractive index, thereby increasing the functional gain in a single stage. attenuation. Regions 1422 and 1424 are electro-optical polarized in the opposite direction to the surrounding material. (Alternatively, the surrounding material may not be polarized, or have no photoelectric coefficient, or it may only be polarized in the opposite direction to regions 1422 and 1424.) A central electrode 1426 overlies the polarized region and the surrounding material. It is energized in relation to electrodes 1428 and 1430 to produce a refractive index change in polarized regions 1422, 1424 and the surrounding material. Device 1420 operates in a similar manner to device 1400 described above. The applied voltage decreases and broadens the refractive index profile of the waveguide segment 1418 , reducing the coupling between the modes of the output waveguide 1408 and the modes excited by the input beam in the segment 1418 . In this configuration, at the beginning of the modified waveguide region 1448, the change in the refractive index profile is sharp and thus lossy. The number and shape of the polarized waveguide sections 1422 and 1424 can be varied as long as the modes coupled to the excited waveguide section 1418 are different from the modes coupled to the unexcited waveguide sections. The device can be made to have high losses when not electrically energized, and adjusted to low losses when electrically energized. In this case, the electro-energized region and/or the polarized region forms part of the structure of the waveguide segment 1418 . The waveguide section 1418 itself can be constructed in a number of ways, most notably completely unexcited if it is not present, in which case the device is identical to the switched waveguide modulator of FIG. 29 .

As mentioned above, these devices can be cascaded to increase maximum attenuation.

The apparatus of Figures 50 and 51 can also be considered as a variable intensity localized ("point") light source. Light propagating in waveguide 1402 is confined to following only the path of the waveguide until a voltage is applied to the electrode structure. When the waveguide effect is reduced or destroyed by changing the refractive index, some or all of the previously confined beam will propagate according to the theory of free-space diffraction. This diffracted beam will continue to propagate in a forward direction, with the beam region expanding in two dimensions much larger than the core of the waveguide 1408 . At a suitable distance from the electrode structure, the beam region may occupy a substantial portion of the substrate aperture, appearing to the observer as a point source of light from a spatial location close to the electrode structure.

If desired, this technique can be used to create a one-dimensional localized light source. The waveguide segment 1418 in Figures 50 and 51 can be embedded in a planar waveguide structure fabricated using prior art techniques so that when an appropriate voltage level is applied to the electrode structure, the lateral confinement of the modes is destroyed and the The vertical confinement in the planar waveguide is preserved. Thus, the beam region expands in one dimension, confining the light to a narrow plane.

FIG. 52 shows a polarizing device 1500 with a widening angle polarizing grating. The method shown for widening the bandwidth is an alternative to the bandwidth modification scheme described with reference to FIG. 18 et al. As shown, a periodic structure 1500 has several poling regions 1502, which are preferably polarized to a poling region of the structure 1504, and other structures such as waveguides and electrodes and gratings are provided as desired. The pattern of domains 1502 spanning the central axis of input beam propagation may be strictly periodic with a 50% duty cycle. The top surface of the polarized region is aligned on all sides along a line drawn from an alignment point. At some distance into the material, each polarized region replicates its surface shape. The result is a structure with polarization whose period changes linearly with lateral position in the polarized substrate. The input beam 1508 passing through the polarized region can be a freely propagating Gaussian beam (if the domains are deeply polarized) or it can be confined in the waveguide 1512 . Depending on the function of the grating, the input beam may be coupled into a filtered or frequency converted output beam 1510, or coupled into a retroreflected beam 1514. The extent of the period in the grating structure (and thus its bandwidth) depends on the width of the beam and the spacing of points 1506 from the axis of the beam. By adjusting these quantities, the bandwidth of the polarized structure can be increased substantially above a minimum determined by the number of first-order periods suitable for the grating described above. There is a limit to the maximum required angle of a polarized boundary, therefore, the structure shown in Fig. 52 cannot be extended without limit. However, a long interaction region can be obtained by cascading together several segments. In order to maximize the coherence between the segments, the period of the domains along the central axis of the beam should not be modified at the junctions between the segments.

Although increasing the bandwidth of the grating reduces the strength of the interaction, it makes the device using the grating substantially less sensitive to small frequency drifts. For example, frequency doubler devices using angularly widened gratings are more tolerant to temperature drift. An example of another application is a channel drop filter, which tends to have a narrow bandwidth because it must use a strong raster. The use of an angle widening grating enables widening of the passband to accept high bandwidth communication signals. The angular widening grating can also be applied in other grating structures discussed above.

There are alternative embodiments implementing an angular widening grating that do not follow the above pattern. For example, the relationship between the angle of the grating period and the distance along the axis of propagation may be more complex than linear. For some applications where most of the interactive power exists at one end of the grating, a quadratic or exponential variation may be more suitable. The angular broadening technique is also applicable to prior art grating types such as diffuse, ion exchanged, and etched gratings.

Figure 53 shows another angular widening device using curved waveguides. In this case, the polarized regions 1522 have parallel faces whose corners are inclined with respect to the local direction of propagation within the waveguide. Second, the above-mentioned bandwidth is broadened with different components of the wave passing through the different Fourier components of the grating. The curved waveguide has higher losses than straight waveguides, but does not require a large curvature. Several of the sections shown in Figure 53 may be joined together, for example to form a waveguide structure that meanders forward and backward along a substantially straight line.

FIG. 54 shows a controllable polarizing lens 1530 . The concentrically disposed domains 1532, 1534, 1536, and 1538 are polarized into an electro-optical substrate 1540 with a polarity opposite to that of the substrate. Two transparent electrodes 1542 and 1544 are placed on opposite sides of the device (above and below the polarized area). When an electric field is applied between the two electrodes, the refractive index of the polarized region increases or decreases depending on the polarity. The geometry of this polarized region is determined by the diffraction requirements to focus light waves of a given color. The spacing between boundaries varies approximately squarely with the radius. For example, if the application required focusing a plane wave onto a circular spot, the polarized region would be circular (for equal focusing on both planes), and as the diameter of the polarized region increases, The number of the above-mentioned polarized regions is reduced, and thus the above-mentioned polarized regions are divided. The boundaries of the polarized regions are determined by the phase of the output wave relative to the input wave on the face of the lens structure. Every time the relative phase of the waves changes by π, a polarized region boundary appears. For example, if the incoming wave is a plane wave, its phase is constant along the surface. If the outgoing wave is a convergent wave (will focus on a point away from the face), then it is essentially a spherical wave and phase changes in the spherical wave determine the boundaries. The lens 1530 is a phase plate with an adjustable phase retardation according to the applied voltage, and the above-mentioned domains occupy the Fresnel domains of the target.

To focus a plane wave of a given color, a voltage sufficient to cause the plane wave to lag (or lead) by π is applied. Each different frequency has a different focal length defined by a different Fresnel domain structure of the polarizing lens 1530 . Higher frequencies have longer focal lengths. Each wavelength is preferably focused, if not scattered, at the same voltage. This voltage can be adjusted to compensate for scatter in the substrate material 1540 . If this voltage is adjusted away from the preferred value, the amount of light that is focused onto the spot is reduced because the phases of the light from the different domains no longer add optimally. They will interfere partially destructively, reducing the pure intensity.

Figure 55 shows a laser feedback device 1450. The laser source consists of amplifier region 1452, back reflector 1454, and a low reflective output region 1456 which may be, for example, a non-reflective coated window. While a normal laser would have another high reflector, in the present invention this high reflector is eliminated so that a grating feedback arrangement can control the oscillation of the laser. Reflections from output region 1456 and coupler 1458 are low enough that the laser does not lase without additional feedback from an external source. The external feedback source consists of an optical coupling system 1458 and a polarizing material 1460 that reflects the beam from the optical amplifier when excited by an electric field. Since the reflection spectrum of the polarizing material 1460 can be very narrow in frequency space, it is possible to select a narrow region in which the laser can operate around a single frequency or multiple frequencies that determine the grating according to the distribution of the grating period . If the cavity is long enough so that the FSR is on the same order as the width of the reflection spectrum, the combined device will oscillate in a single longitudinal mode.

Means 1458 for coupling optical energy between the optical amplifier and material 1460 collect the output mode of the laser and refocus it onto the polarized material. Couplers can be composed of a variety of alternative configurations, including one or more of the following components: high numerical aperture lenses, such as GRIN (graded-index), aspheric, diffractive, or multi-element spherical lenses; tapered waveguides; Proximity adjusters and aligners for use in waveguide-to-waveguide to-head coupling situations. The surface of the coupler 1460 is preferably non-reflectively coated. The AR coating can be a multilayer dielectric coating, or a sol gel coating, or a quarter wave layer of material with a suitable refractive index (geometric mean of two adjacent media). If the material is globally polarized, the preferred focus in the material 1460 has a Rayleigh region approximately equal to the length of the polarized region. If the material has a waveguide for confining the propagating beam, the optical coupling system should preferably convert the laser mode into a mode distribution at the entrance of the waveguide that, in terms of phase front angle, radius of curvature, and lateral dimension, The distribution matches the desired mode of the waveguide. The polarized structure consists of at least two types of domains, which are preferably polarized opposite each other. The polarized material has electrodes 1462 and 1464 that extend across the region to be polarized and are electrically energizable by a power source 1466 . When a voltage is applied to the two electrodes, the induced electric field in the material causes a change in the refractive index, which will be changed in part according to the polarization direction and the strength of the electric field. By inducing a periodic structure in the polarization, an electrically controllable periodic modulation can be generated as a function of the refractive index.

Amplifier 1452 is provided with the necessary attachment and excitation to produce an optical gain factor over an extended region characterized by a central optical axis. The optical bandwidth of the amplifier is limited by the process by which the gain is generated. The bandwidth is the width of the gain distribution (typically 3dB full bandwidth): the gain as a function of optical frequency. Semiconductor diode technologies (such as InGaAs, AlGaAS, AlGaInP, InGaAsP, ZnSe, GaN, InSb) are advantageous for providing large bandwidths, although they do not provide high power capabilities. Optical reflector 1454 is a feedback mirror, which may be an integral mirror whose radius is aligned and matched to the phase front of the mode, to reflect the mode propagating out the back of the amplifier back to itself. Or, in the case of waveguide amplifiers, (Nd:YAG, Er:YAG, Nd:LiNbO ₃ , Er:LiNbO ₃ , and various combinations of rare oxide ions and crystalline or glass matrices), which can be the A face that is attached or polished perpendicular to the waveguide. If the cavity geometry is a ring (allowing propagation of light in a single direction), the light reflector is a multi-element structure consisting of at least two elements to collect light reflected from material 1460 that does not pass through the amplifier , and aligning and refocusing light returning through the amplifier onto material 1460, it will again have the same mode characteristics as it had on its previous pass.

Interaction of several beams with the periodically poled material 1460 is possible. If the period is chosen to be a multiple of the period required to not reflect light within the gain of amplifier 1452, then the device will act as a (higher order) field controlled feedback mirror. When the voltage 1466 is turned on, the laser can be turned on, thereby creating no reflection within the bandwidth of the grating. Then by modulating the voltage, the laser output is amplified because the laser oscillation varies in direct proportion to the electric field strength. Modulation control 1466 provides the voltage and current required to establish the desired electric field in material 1460 as a function of time. The laser can also be mode locked by operating the modulation control at a frequency equal to a multiple of the round trip frequency of light between the material 1460 and the laser reflector 1454 . Because the reflectivity of the polarized structure 1450 is modulated at the same frequency, the beam resonating between the two feedback mirrors 1450 and 1454 tends to split into one (or more) pulses, which of course orbit at the round-trip frequency pulsation. If the frequency is a multiple of the round trip frequency (1x, 2x, 3x, . . . ), the reflectivity will be high each time the pulse approaches reflector 1450 . At higher multiples, the reflectivity remains high for a shorter time, thus producing a shorter pulse, but some means may be needed to suppress additional pulses that tend to form at other high reflective multiples during the round trip transit time. This additional pulse can also be suppressed by applying a component of the signal to reflector 1450 at the round trip frequency, by modulating amplifier 1452, or by other means including common additional components. A light output can be extracted into beam 1468 or 1649.

The laser frequency is stabilized by using the feedback device 1450, since the periodic reflector only operates at a certain frequency. Outside the incident frequency, the bandwidth of the polarized structure is not reflected. In a simple construction, the bandwidth is determined by the reciprocal of the number of grating periods of the first order, which is adapted to the length of the polarization region containing the frequency components. In a more complex structure with multiple periods, the bandwidth is determined by the Fourier transform of the polarized structure along the bisected angle of the incident and reflected beam propagation directions. Because feedback occurs only during a restricted frequency region, the output frequency of device 1450 can be much narrower than that of a free-running laser oscillator in which the polarization structure has been replaced by a single mirror . If the bandwidth of this reflection is comparable to the spacing of the longitudinal modes of the extended cavity formed by reflector 1454 and the polarized structure, then the device will operate in a single frequency mode.

The stability characteristics are particularly important in the case of semiconductor diode lasers, where the gain is high and broadband. As with diode lasers, all unwanted internal reflections, such as from the output region 1456, should preferably be kept very low (eg, below 10 ⁻³ ). Depending on the depth of the field and the preferred voltage applied, the electrodes 1462 and 1464 can be on the same or opposite sides of the structure. The resonator including the laser and the switched reflector can also be a ring resonator instead of the linear resonator shown in FIG. 5 . As is known in the art, some additional optical elements are required to form the ring resonator, and the reflections from the polarizing material 1460 are not perpendicular to the incident light. The period and angle of the grating must always be tuned so that the virtual photons added to the interaction create a conservation of momentum between the incoming and outgoing photons. This confinement determines the angle and period of the polarized grating.

Figure 56 shows a feedback device 1470 with a waveguide. A waveguide 1470 can be placed into the polarizing material 1460 to confine the light beam to a long distance. This is particularly useful in devices that require interaction lengths to produce large reflections, and in integrated devices where all light travels in a waveguide. As shown in FIG. 56, a waveguide laser 1474, such as a semiconductor diode laser or a diode vacuum solid-state laser, can be butt-coupled to the waveguide for robust and efficient operation. In butt coupling, the optical coupling system 1458 is AR coated on surfaces 1475 and 1477, which along with the alignment and mounting structures must maintain alignment. The waveguides of the optical amplifier 1474 and the polarizing substrate 1460 are aligned so that the phase front of the optical field exiting the optical amplifier towards the substrate has the same degree of preference as the phase front of the modes propagating in the waveguide 1472: the same angle , radius, and lateral dimension. The two waveguides must be spaced within a Rayleigh region, and their deviation from axial alignment should be less than a fraction of the lateral dimension. One of the waveguides 1472 or the waveguide in the amplifier 1476 may be tapered to optimize this stacking. In a waveguide device, it is not necessary for the polarized regions 1478 and 1480 to extend entirely through the substrate 1460 . Electrodes 1482 , 1484 and 1486 are disposed over the polarized regions intersected by waveguide 1472 . When electrode 1484 is energized relative to electrodes 1482 and 1486, a refractive index pattern is generated in the waveguide, the structure of which is essentially determined by the structure of the polarized substrate. The index pattern may act as a reflector as described in Figure 55, and/or may act as a coupler to other waveguides as described above. At the opposite end via port 1488 or amplifier 1489, an optical output can be taken from the device.

A frequency doubler can be provided in the substrate 1460 if the substrate material is a nonlinear optical material such as lithium niobate, Lithium tantalate, or KTP. Quasi-phase-matched frequency doublers can be installed as part of, before or after the feedback grating structure. If the grating structure sets multiple reflection frequencies, the optical amplifier 1452 or 1474 can be induced to oscillate at two or more frequencies within its gain bandwidth. In this case, the nonlinear frequency converter could be a summing frequency mixer instead of a frequency doubler, or several of the above devices could be cascaded to form a multi-frequency combination of multi-frequency outputs.

Variations of the above relating to the polarization structure, its excitation, and the modes it uses can also be applied in combination with external optical amplifiers. In particular, frequency tunable lasers can be realized by combining the structures of FIGS. 55 and 56 with the tunable gratings of FIGS. 14 and 15, respectively. As previously mentioned, tuning is achieved by arranging the polarized grating structure so that the average refractive index changes with the applied field. The operating frequency of the optical amplifier 1452 or 1474 is determined by the frequency of the feedback from the polarized structure 1460, thus, the output frequency can be modulated by modulating the average refractive index of the polarized structure. Changing the average index of refraction changes the momentum vector of the photon, without changing the momentum vector of the virtual photon contributed by the grating. After the average refractive index has changed, the old reflection frequency is no longer optimally phase matched to the reflection; the peak reflectivity has shifted to a new frequency.

Using the configuration described with reference to Figures 55 and 56, with the addition of changes to the average refractive index as described in Figures 14 and 15, a frequency modulated (FM) laser can be constructed. By the word "modulation" is meant that the change is a function of some parameter, in this case time, such as a pulse with high or low duty cycle, a sinusoidal change, or a change with an arbitrary temporal relationship. A control system can be provided to control the voltage and apply the required current to accommodate the desired instantaneous changes in the electric field.

Typically, preferred feedback for semiconductor lasers requires a reflectivity of the grating to be less than 10%. The remaining light is available for output. The laser can be made to operate in either TM or TE polarization, depending on the beam confinement on the grating sheet in the waveguide. Scattering in the grating, and the relative gain of the polarizations in the gain element. Because the intensity of the grating is controllable, the reflectivity can be adjusted to maximize the output coupling of the laser and thus the output power.

Similarly, the gratings can be used to form passive or cavity fitted reflectors. Since the coupling of a laser to a cavity depends on the ratio of the relative reflectivity of the input coupler to the cavity loss, variable reflectivity input couplers provide a means of optimizing this parameter and thus provide to match the impedance of the resonator.

In a cavity, the present invention can also be used for single-pulse switching, mode locking, or cavity ejection, with little or no chirp for lower power CW sources (e.g., semiconductor lasers). In addition, the tuning potential enables The laser is used as a source for communication, beam splitting, and remote sensing.

Figure 57 shows a selectable wavelength laser 1490, preferably a diode laser 1474, controlled by an array of switchers, with a waveguide 1476 butt-coupled to a waveguide 1472 in a substrate 1461. Facets 1475 and 1477 are preferably coated so that optical amplifier 1474 will not emit according to the reflectivity of its own facets. The substrate can be any substrate capable of supporting a switch 1492, which can be implemented in a variety of ways, including the TIR switch of Figures 30-32 and 34-35, the grating switch of Figures 7-8 and 12-13 10 and 26-28 couplers, the splitters of FIGS. 23, 25, 33 and 46-48, or any other optical waveguide switching structures currently known or to be discussed. The TIR switch 1492 has been fully described above and is only shown schematically in this figure. When they are in the on position, the switches route the light energy from the amplifier down to one of the waveguides 1494 associated with that switch. An array of non-reflectors 1496, here shown as a grating, is disposed within the waveguide. The grating reflects incident light at a specific frequency, and the laser emits laser light within the bandwidth of the grating. The grating elements are shown pointing towards a point 1498 at a somewhat distance, so that the period of the grating further away from the laser is progressively shorter. The frequency at which the laser is required to operate is selected by selecting the switcher in relation to the desired grating period. The optical frequency is determined by the geometry, and linearly separated optical wavelengths are obtained with a constant switching interval. Any wavelength interval within the storage density capacity of the arrangement can be selected if desired. Due to the low insertion loss of TIR switches and their high storage density, a large number of switches can be placed along the waveguide 1472 . The sloped output waveguides are also very compactly stacked.

The output beam can be taken out from the rear surface of the optical amplifier 1474 as beam 1489, or it can be taken out of the waveguide 1472 as beam 1488 since the TIR switch will leak a small portion of the laser's light along the waveguide 1472. A number of alternative structures such as those described with reference to Figure 56 also involve this structure. For example, the reflector array may consist of fixed gratings, or switching gratings fabricated using a number of techniques. It may consist of a uniform grating structure where different optical path lengths to the grating are selected for different FSRs of the laser cavity, yielding single-mode operation with a selectable array of closely spaced spectral peaks. It could even consist of an array of fixed mirrors along the waveguide 1494, which could be coated for high or variable wavelength reflection. Furthermore, the different spacing of the mirrors provides the opportunity to tune the switchable laser cavity path over a large range.

Part of the structure of FIG. 57 is a modulator utilizing an adjustable optical energy redirector 1492, and one of the feedback reflectors in one of the waveguides 1494. If the waveguide 1472 is made non-reflective, such as for example having its width tapered to zero, followed by a redirector 1492, the laser is reflected again by feedback borne by the reflector 1496. By adjusting the amount of light energy fed back through the excitation of the redirector 1492, the output characteristics of the laser can be controlled. In this way the laser can be modulated, in which case the reflector 1496 can be a fixed grating or even a broadband fixed mirror. Using a grating reflector has the advantage of a fixed frequency, so that the laser can be deeply modulated without frequency shift, resulting in almost pure amplitude modulation. If the redirector 1492 is modulated by an integer multiple of the cavity round-trip travel time, the device is a mode locker and produces a pulsed output. The round trip time is the time taken by the coaxial alignment pulse to return to its original position and orientation within the cavity. By using different switches, the length of this cavity can be varied to vary the spacing of the pulses. By using two different switches at the same time, it is also possible to discriminate against IF pulses which tend to grow with the mode-locked frequency by a high order of magnitude multiple of the round-trip travel time. Frequency modulation is obtained by modulating the center frequency of the reflector. In this case, the reflector 1496 is preferably formed as a tunable grating as described in FIGS. 14-22.

Figure 58 shows a wavelength tunable tunable focusing system 1550. The combination of a diffractive focusing element 1522 such as the zone plate lens of FIG. 54 (or such as an opaque body or etched zone plate) with a tunable light source 1554 provides new capacity in the data storage field. When an area plate is combined with an adjustable frequency light source, the distance to the focal point is adjusted by tuning the light source. This capability is useful for multi-layer data storage devices where data is read from and written to the data side 1556 stacked at various distances into the data storage medium 1558 . If the waveguide of the light source is tuned (as we have described above by various means), the distance from the area plate to the focal point is adjusted accordingly.

The frequency tunable laser chosen to drive the multifaceted data storage system is a semiconductor diode laser 1560 based laser system, tuned based on feedback from an electrically tunable grating 1562 such as we have described. The laser can also be frequency doubled in a quasi-phase matching section 1564, in which case the use of an angular broadening polarization grating is the preferred method to make the frequency doubler's reception broad enough to accommodate significant tuning of the source laser. Lens system 1566 aligns and surrounds the laser output in preparation for final focusing on zone plate lens 1552 .

A lens system is not necessary since the zone plate can also refocus a diverging beam. However, it is desirable to form a surrounding beam through the zone plate as this will produce the smallest spot size and thus the highest density data read/write capacity. Devices 1562, 1564 and 1566 can be fabricated in waveguides on the same substrate, and if desired (and in combination with one of the out-of-plane reflectors) can be integrated with an area plate lens on the rear of the substrate to Be a small, light-weight unit capable of booting quickly in a data storage system.

Above, the present invention has been described with reference to specific embodiments. Other embodiments will be apparent to those skilled in the art. Accordingly, the invention is not to be restricted to the parts forming this invention, except as indicated by the appended claims. A general description of the display

The main field of application of the switching technique according to the invention is in the field of optical displays, especially emissive optical displays. FIG. 59 depicts a block diagram of a display device 1001 , and FIG. 60 is a cross-sectional view taken along line A-A of FIG. 59 . The intensity of the optical frequency source device (or light source) that can generate visible light or infrared light from the light source 1000 is modulated by a beam modulator 1002 . Visible light information is thus encoded into a time-varying intensity of said beam. This modulated light beam is connected by the above-mentioned coupling device 1004 to an optical waveguide 1006 (hereinafter referred to as distribution waveguide), which can be transmitted in or on the surface of a suitable substrate material 1028 in various well-known ways. processing to produce the optical waveguide 1006 described above.

The distribution waveguide 1006 abuts at least one optical waveguide 1014 (hereinafter referred to as a pixel waveguide) at an angle sufficient to separate the pixel waveguide from the distribution waveguide by a suitable distance. Figure 59 shows that the angle is 90° in the preferred embodiment. A parallel array of pixel waveguides forms the scan lines for the above-described display.

The distribution waveguide and the pixel waveguide are connected by a beam guiding structure (composed of a row of light energy guides 1008 ) at the junction area. The optical energy director may be any active or passive means of directing or switching the propagation of light in the distribution waveguide into a desired pixel waveguide. Examples of passive optical energy directors include directional waveguide couplers, dispersive waveguides, grating couplers, star couplers, mode separators, and Bragg grating reflectors. Active light energy directors operate on electro-optic, acousto-optic or thermo-optic principles. Some examples of active optical guides include voltage activated directional couplers, total internal reflection (TIR) switches, dispersive waveguide switches, Bragg grating reflective switches, various types of electro- Optical switches, polarization-controlled switches (such as TE-TM mode converters combined with passive polarization mode separators), acousto-optic switching devices (such as SAW filters and modulators) and any Polarized switching structures are discussed as prototypes. The preferred embodiment described above uses an active beam director based on a variation of the TIR switch described above. A fixed mirror (shown in Figure 32) may be included in the beam guide to make the angle between the distribution waveguide and the image waveguide 90°. Such a beam guide is hereafter referred to as a distribution switcher.

Once the light has propagated within the aforementioned pixel waveguide, the light is directed to a desired point on the substrate 1028 (hereinafter referred to as pixel switcher) by activating an electronically controllable light energy redirector 1016 (hereinafter referred to as pixel switcher). are pixel locations or simply referred to as pixels). The above-mentioned pixel switcher is an active switching device (passive energy diverter) using the above-mentioned operation principle. Examples of pixel switchers include voltage-activated directional couplers, total internal reflection (TIR) switches, dispersive waveguide switches, Bragg-grating reflective switches, various types of electro-optical switches that operate on the principle of mode conversion , Polarization-controlled switches (such as TE-TM mode converters combined with a passive polarization mode separator), acousto-optic switching devices (such as SAW filters and modulators), cut-off switches, liquid crystal waveguide switches any other waveguide plug method discussed in the prior art and any polarized switching structure discussed in this specification as a template . The preferred embodiment of the above-mentioned pixel switcher is based on a modification of the above-mentioned TIR switcher.

The light beam propagating in the pixel waveguide 1014 is guided by the pixel switcher 1016 into an output waveguide section 1017 (hereinafter referred to as reflector waveguide) to reach a desired pixel position on the surface of the substrate. The reflector waveguide may comprise a two-dimensional waveguide, a planar waveguide, or several short or long segments of free space propagation in the bulk substrate.

The optical energy (or light) is then directed out of the substrate surface towards the viewer using an optical reflective device 1018 (hereinafter referred to as a non-planar reflector) disposed in or on the surface of the substrate. The non-planar reflector may be one of many micromirrors whose function is to guide a portion of the light previously propagating in the output waveguide away from the plane of the waveguide. The light transmitted from each of these small reflectors 1018 towards the viewer constitutes a pixel of information leaving the display, and the term 'pixel' will hereinafter be used interchangeably with the reflector 1018. In FIG. 59, the inventors illustrate a pixel 1019 that is turned on and a pixel that is turned off at the same time. The array of pixels is organized into a common column and grid arrangement. Each pixel can be addressed by its column and row location. This array of pixels constitutes the above-mentioned display screen.

After leaving the surface of the above-mentioned substrate, the light impinges on an optical scattering screen 1026, which converts the forwardly propagating light beam into a more dispersed light beam having a propagation angle (relative to the above-mentioned normal to the substrate) so that an observer can observe at a polar angle θ (relative to the normal to the substrate surface above). If the light beams emerging from the above-mentioned display screen can be distributed close to the Lambertian angle, the above-mentioned display screen will appear equally bright in all polar angles from 0 to 90°, regardless of the azimuth angle. Other angular assignments can be used to achieve different viewing angle specifications. The diffusing screen 1026 described above can be made as a separate optical element.

The present invention activates the above distribution (or column) switcher and the above pixel (or row) switcher in a certain timing sequence to form an image on the above display, so that the light emitted by the above reflector forms a "scan field scan". The invention produces this scanning field in a manner similar to the way an electron beam is scanned in a standard cathode ray tube (CRT). The distribution switcher and the pixel switcher are coupled to a field generating device by an electronic structure. Such field generating means may be any form of voltage generator sufficient to generate a voltage to activate the switch, eg a DC power supply, a battery, an electrical circuit (known as drive electronics). The electrodes of the distribution switcher 1008 are connected to column switcher drive electronics 1010 by which the distribution switcher 1008 is driven. Several pixel electrode structures 1020 are connected to some column driving electronic components 1012 , and the pixel switcher 1016 is driven by the column driving electronic components 1012 .

During the field scan, the invention sequentially modulates the intensity of the light source and the pixel switching is activated so that each pixel receives an intensity corresponding to its spatial position in the image. Any standardized video format can be used as an example of the format of the display information of the present invention. In the simplest configuration, the positions of the above-mentioned active pixels 1019 (from which light is emitted) are scanned from left to right, top to bottom one column at a time, thus generating the framed information, scanning all the pixel positions required The reciprocal of time is called the "frame rate".

The display of the present invention differs from a standard CRT display in that the light from the emitting pixels does not last as long as the pixel switcher is activated. For a high information content display with a pixel format of 640x480 and a frame rate of 80 Hz, this time is less than 50 ns. When the refresh rate is higher than 80Hz, the eye's sensitivity to flash drops rapidly, even in the more sensitive parts of peripheral vision.

One embodiment of the invention is a full color display. The light source of the present invention achieves this effect by comprising 1000 three intensity sources operating at different central wavelengths (colors). For optimal effect, the above light sources are based on diode lasers to generate light near 630nm, 530nm and 470nm. Waveguides 1006 and 1014 can be designed to support the propagation of these different wavelengths, just as switches 1008 and 1016 described above can support the propagation of these different wavelengths. The present invention adds a feature to the light source 1000 described above, namely a connection device to combine the outputs of the three lasers into one waveguide so that they can then be manipulated by a single set of waveguides and switches. Details of the added features will be described below.

Please note that the above-mentioned laser light enters the display of Fig. 59 from the bottom right, because a frame of information is scanned from the top left. In a frame of information, the present invention scans the individual columns in a sequence, the switch closer to the laser being activated next. The rise time of the switch described above quickly redirects the light into the newly selected waveguide without being affected by any charge that may be left on the previous electrode. The advantage of this is that the falling time of the switcher does not affect the pixel rate, allowing the falling time to be longer than the rising time. Indeed, a long fall time can be coordinated with energy recovery from the above electrodes to minimize power consumption. For similar reasons, the order of stimulation in the aforementioned trains is the same (towards the aforementioned laser light). If the fall time is greater than the rise time, the present invention may employ a display scan delay at the end of each column to allow all distribution switches to reach their low loss state, allowing light to pass through the first pixel in that column. This delay should be several times the abatement time of the above-mentioned distributor. Similarly, the present invention may employ a delay at the end of the frame information to provide sufficient time for the aforementioned pixel switcher to relax.

Since the inventor may use many different versions of each of the elements shown in Figure 59 to construct a particular embodiment of the display of the present invention, the inventor will describe variations of these elements in detail and indicate which variations may be selected Take it as a preferred embodiment of the present invention. B. Light source

Any light source that can be connected to a waveguide can be used as a light exciter for use in the displays of the present invention. The aforementioned light source may have any wavelength in the visible, ultraviolet, infrared region of the spectrum. Examples of single emitter light exciters include any type of laser, light emitting diode (LED), and high brightness small volume incandescent and fluorescent light sources. Examples of multi-emitter light sources include arrays of lasers or LEDs, incandescent and fluorescent light sources arranged in a line or distributed in two-dimensional light distribution areas. The solid-state laser used by the preferred embodiment of the invention is a semiconductor diode laser source (hereinafter "diode laser") for several reasons. First, a laser light source can be connected to a waveguide with the highest efficiency of any light source, thereby minimizing the electronics required to operate the display. Second, the diode laser light source is an efficient converter that converts electrical energy into light energy, again minimizing the power requirements of the display. Third, the inventors fabricated the above-mentioned diodes with semiconductor processing technology, thus allowing the above-mentioned displays to incorporate low-cost light sources.

Commercially available diode lasers can generate light with a wavelength greater than 620 nm in the visible region, and such diode lasers can be used in red monochrome displays. To obtain a full-color display, wavelengths in the green region (nearly 530nm) and in the blue region (nearly 470nm) are also required. Using nonlinear optics (referred to in this technique as a frequency doubler), wavelengths in the green and blue regions are obtained from the infrared diode laser. Embodiments of the above frequency doubler include bulk nonlinear crystals such as KDP, LiNbO ₃ , KTP, LiTaO ₃ , SBN, and BaTiO ₃ and nonlinear waveguide structures fabricated in the above bulk frequency doubler crystals or using nonlinear Polar polymer waveguides. In a preferred embodiment of the present invention, the above-mentioned frequency doubler is a periodically polarized structure, and the material used to manufacture the regularly polarized structure is a periodically polarized material of nonlinear optics, such as LiNbO ₃ , KTP , LiTaO ₃ , SBN, and ferroelectric crystals of BaTiO ₃ . Since the substrate materials of the above-mentioned displays can be fabricated from these materials, the above-mentioned displays and frequency doublers can be integrated into a single chip.

Using such a frequency doubler, for example, a 1060nm diode laser can be frequency doubled to obtain 530nm green light, and a 940nm diode laser can be frequency doubled to obtain 470nm blue light. The above-mentioned doubling process has been shown to be more than 50% efficient in the laboratory. The inventors can use this technique to obtain the full range of visible light wavelengths required for displays.

Diode solid-state lasers can be used as light sources for the displays of the present invention. Dense green diode lasers are commercially available, and blue diode lasers have been demonstrated in laboratories. Diode sources include frequency-multiplied fiber lasers, up-conversion lasers, frequency-multiplied diode YAG, YLF, and other host laser materials with impurities.

Directly visible solid-state semiconductor lasers (made from semiconductor materials such as ZnSe (and Gallium Neonide)) operating in the blue and green regions now exist in laboratories and if their lifetimes can be extended And if these light sources are commercialized, they can become the preferred light source for the display of the present invention.

In a preferred embodiment of the present invention, the laser and the frequency doubler are provided on a substrate separate from the display. The inventors then used fiber optic connections or butt connections to connect this combination to the waveguide in the display. Separating the two subcombinations increases the efficiency of each unit. The laser can be butt-connected to the wafer or flip-chip in etched grooves on the wafer. The butt coupling is robust, requires no additional components and is compatible with self-alignment techniques. In a preferred embodiment, depth and width waveguide tapers are used in the frequency doubler chip (or the diode laser chip) to match the laser mode size to the frequency doubler waveguide size for optimal connection efficiency. The depth taper in the doubler can be obtained with two waveguide fabrication steps. The first step produces a deep waveguide with a taper, while the second step produces a shallow waveguide (joining the aforementioned taper). In this way, the aforementioned laser light can be effectively coupled to the aforementioned single mode frequency doubler waveguide. As an alternative, the laser diodes described above can be directly connected to the same substrate as the displays described above. This allows the aforementioned frequency doubler, feedback stabilization element, wavelength combiner and other components to be integrated in the same die. If the efficiency is high, the direct connection scheme described above has a cost advantage. C. Modulator

The beam modulator 1002 can vary the intensity of the laser light to obtain a frequency bandwidth that is always sufficient for the display. The frequency bandwidth in which the above-mentioned preferable effect can be obtained is larger than the above-mentioned pixel scanning rate. This bandwidth exceeds 25MHz for a 640x480 pixel VGA display with an 80Hz update rate. The aforementioned modulator is preferably capable of reproducing many gray scale intensity levels. The aforementioned modulator is preferably fabricated on the same die as the aforementioned frequency multiplier.

Several different beam modulators are known to meet the above requirements. The first method is to modulate the intensity of the diode laser by modulating the current supplied to the diode laser. Since the above-mentioned diode laser gain spectrum also varies with current, the frequency variation accompanies such modulation. If the above-mentioned diode laser is used to directly illuminate the above-mentioned display (like a red AlGaInP diode), such a frequency change is not a problem, because the eye cannot perceive such light. However, if the output power of the above-mentioned diode laser drives a frequency doubling unit to produce a green or blue beam, such frequency modulation is problematic because of the narrow acceptance of the multiplication efficiency curve as a function of laser frequency (at 0.1 nm within a factor of about 100, depending on the construction and design parameters of the frequency doubler). In the latter case, the diode laser operating frequency must be stabilized within the acceptable range of the frequency doubler, although not necessarily operating in a single frequency mode. (Multi-mode operation does increase conversion efficiency.) Frequency stabilization can be achieved using external grating feedback (or any other method known in the art) combined with an electronic feedback control loop. The grating is integrated into the waveguide using methods such as known mask etching or bulk diffusion. The above-mentioned grating center frequency is electro-optical or thermal vibration, and a photodiode is used to monitor the strength of the electric power whose frequency is doubled. The frequency component of the photodiode signal vibrating at the high frequency is used as an error signal and input into the electronic feedback loop to control the frequency of the diode laser. Visible light is thus produced by matching the aforementioned laser frequencies to the preferred wavelengths required for efficient frequency multiplication. In addition to grating stabilization (sufficient suppression of spurious reflections back into the diode laser), it is possible to use modulation techniques that vary the diode laser current as it is feedback controlled by the grating at the frequency. Being able to modulate the diode laser current is an important advantage because this technique minimizes power consumption in the laser.

A second modulation method is to use known modulator designs to fabricate a separate modulator (possibly in the substrate 1028 described above). An example of a standard modulator design is the Mach-Zehnder interferometric modulator. Any of the electro-optic switches described above may also serve as the modulation device 1002 described above. The preferred embodiment of the present invention uses the internally reflective optical switch depicted in FIG. 30 as the modulator described above because it is compact, uses the same polarization technique as the TIR switches 1008 and 1016 described above, and has an input strength equal to The voltage relationship works well to provide digital grayscale modulation. When the output beam is the reflected output of the switch, the gray scale of the pixel has almost a linear relationship with the input voltage of the switch. D. Optical connection equipment

The optical connection device 1004 connects the light from the light source 1000 to the distribution waveguide 1006 . Many different techniques are known to achieve a high connection efficiency. Some examples of these known techniques include bulk lens beam trains, docking of the laser diodes to the substrate 1028, and fiber connections of the laser sources to the substrate. A preferred embodiment is a fiber connection because it is efficient and inexpensive and allows the diode laser source 1000 and modulator 1002 described above to be manufactured and packaged independently. Thus, the light source package can be detached from the substrate 1028, separating the features that limit the efficiency in generating the laser light from the features that limit the efficiency of the display panel. The light output of the light source 1000 enters a fiber output end for connection of the light source to the display panel. The fiber output can be connected to the distribution waveguide 1006 using known standard techniques.

In the preferred embodiment of the full-color version of the invention, the three laser sources are packaged together and connected using a three-color beam combiner (preferably a grating-assisted coupler as depicted in Figure 61). to a waveguide. The beam combiner connects the above-mentioned red laser beam 1050, the above-mentioned green laser beam 1052 and the above-mentioned blue laser beam 1054 to the same waveguide 1044 with minimal loss of light already propagating in the waveguide 1044 . This combiner results in a combined laser output 1056 which is then passed into the waveguide 1006 described above. The red laser beam 1050 is connected from the waveguide 1042 to the waveguide 1044 using a temporary waveguide connected to a fixed index, depressurized or polarized grating 1056 to match the propagation parameters of the input mode 1050 in the waveguide 1042 to the red wavelength of the waveguide 1044 . The blue laser beam 1054 is connected from the waveguide 1046 to the waveguide 1044 using a temporary waveguide connected to a fixed index, decompressed or polarized grating 1058 so that the propagation parameters of the input mode 1056 in the waveguide 1046 are matched at Waveguide 1044 for blue wavelengths. This is a known standard technique that can be used to efficiently combine power in three waveguides into one waveguide. Further discussion of grating assisted parallel waveguide couplers having a polarized grating has been provided earlier in this specification with reference to FIG. 10 . The bandwidth of the grating-assisted coupler is usually much narrower than the bandwidth of the separation of the three color wavelengths, so that the power in each color beam can be independently controlled and optimized. The above-mentioned grating-assisted coupler is suitable for efficiently combining three waveguide color beams into the same waveguide.

The three-color grating-assisted coupler 1040 is preferably implemented on the same substrate chip as the frequency multiplier for the blue and green light sources and the modulator for the blue and green light sources. All three diode laser sources are connected to the same die. In addition, there may be another modulator placed after the aforementioned beam combiner (possibly also on the same die) to supply bias correction to the three colors simultaneously. Bias correction is an intensity modulation and must be used to compensate for the different lengths of wavelengths blocked by the beam when stimulating different pixels for the optical power drop between the combiner and each pixel. E. Substrate material

Substrate 1028 for display 1001 is a solid state material, which is any electro-optic active material that can be made flat enough to maintain photoetching tolerances for the entire display area. The aforementioned distribution (and pixel) waveguides and switches are fabricated in or on this sheet of substrate material. In the preferred embodiment, the electro-optic TIR switch is formed by a polarization process, so the substrate or a film on the substrate must be a polarizable electro-optic active material. Some examples of suitable electro-optically active solid materials are wafers of bulk ferroelectric crystals (e.g. LiNbO ₃ , KTP, LiTaO ₃ , SBN, BaTiO ₃ ), coated on suitable support materials (e.g. glass or plastic). Polarized electro-optic polymers, thin films of ferroelectric crystalline material (eg _LiNbO3 , KTP, _LiTaO3 ) on a suitable support material (eg bulk crystalline substrate structured to allow single crystal film growth).

A preferred embodiment of the invention uses a wafer 1028 of lithium niobate ( _LiNbO3 ). _LiNbO3 is an excellent electro-optic active material and is the lowest cost of all large diameter crystalline materials except silicon when grown in a mesa to a diameter of 6 inches. F. Allocation and pixel waveguide structure

In a preferred embodiment of the invention, a single distribution waveguide 1006 connects all the pixel waveguides, and distribution switches are used to route light from the distribution waveguides to the pixel waveguides. In general, a connection means is any means by which light can pass from the distribution waveguide to the pixel waveguide. In other embodiments, multiple distribution waveguides may be used to form a busbar of distribution waveguides. The pixel waveguides are parallel to each other and occupy the entire display area. There is a large angle (eg 90°) between the distribution waveguide and the pixel waveguide, and it is close to one side of the display screen.

The waveguide parameters (width, depth and index change Δn) are chosen such that the waveguide supports a single mode for the three colors. The shortest beam wavelength (470nm) is only 25% shorter than the longest beam wavelength (630nm), so that the waveguides can be chosen to support the lowest sub-mode for the three colors in the same waveguide. Multimode waveguides can also be used, but the associated higher radiation losses of these waveguides require higher input power and increase the degree of background blur (flare) of the display screen.

The waveguides used in the present invention are fabricated in an electro-optical material, such as the substrate material as defined above. In a preferred embodiment of the invention, the waveguides are fabricated in _LiNbO3 by a standard annealed proton exchange process (APE). Another available waveguide fabrication process in _LiNbO3 is the titanium bulk diffusion process. Either or both of these waveguide fabrication processes can be used to fabricate the switch shown in FIG. 30 . Other substrate materials can be obtained using various known waveguide fabrication processes. G. Distribution switcher

The main requirement of the distribution switcher 1008 is that it must have near zero loss when it is switched off. This requirement is for display applications since the directed light beam should pass through hundreds of distribution switches without any appreciable loss to deliver enough power to the final distribution switch in the distribution waveguide. The preferred embodiment of the present invention achieves this requirement using a TIR reflective switch with a polarized waveguide section as shown in FIG. 30 . The other main specification is that the distribution switcher must bend the beam by nearly 90° to allow the manufacture of large display areas in a small wafer size. An almost 90° planar micromirror reflector (also known as a fixed mirror) is shared with the switcher of FIG. 30, as shown in the part of FIG. Waveguide 986 (the asymmetric lossy waveguide cross 997 and the extra switch 983 and waveguide 982 are not necessary in the simplest display configuration). When the TIR switch 985 is cut away, the polarized portion is completely transparent and the beam continues along the distribution waveguide. Light is sent from waveguide 984 to pixel waveguide 986 with only low losses when a voltage is placed on the TIR switch electrode as 30 . The specification has previously described the fabrication techniques used for the preferred embodiment of the micromirror 977 with reference to FIG. 36 .

Other switching methods can be used to enable light energy to enter or exit waveguides and to transmit between waveguides. The TIR switches of Figures 30-32 or Figures 34-36 or the guided wave switches of Figures 33, 46 and 47 may be used. The raster switchers of Figures 7, 12 and 13 are also available alternatives. Alternatives in the prior art, such as electro-optic switches with interdigitated electrodes that are non-polarizable, can be used in place of the grating. The four-block switch of Figures 8, 12 and 13 can be used in some special configurations, such as the non-planar switch 168 of Figure 9 or the waveguide coupler switch of Figures 10 and 26-28. The reflective switch of Figures 2-6, 14-17, 19 and 21-23 may be used with an upstream splitter to direct a portion of the reflected light into an output waveguide. Other switching techniques (including those in the prior art or those not yet invented) can also be used as long as these techniques transfer a substantial portion of the power propagating in a distribution waveguide (or a pixel waveguide) into a pixel waveguide (or a non-planar reflector). H. Pixel switcher

The pixel switcher 1016 forms a two-dimensional array in the flat panel display. The main requirement of the pixel switcher is similar to that of the distribution switcher in that it has almost zero penalty when it is switched off. Therefore, the preferred embodiment of the present invention uses switchable total internal reflection (discussed previously with reference to Figures 30-32) to achieve the desired low loss operation. This switch is used to take optical power from the pixel waveguide and direct it into the reflector waveguide and the non-planar reflector. An alternative embodiment could use a switchable grating reflector (discussed previously with reference to Figures 7 and 8). I. Non-planar reflector

A non-planar reflector (also known as an optical reflector device) means a mechanism that changes the direction of propagation of light from the above-mentioned surface to any direction (including a component at right angles to the surface) and thus causes the light to leave the above-mentioned surface. The basic elements of a nonplanar reflector include a light energy reflecting boundary that has a reflectivity discontinuity in a plane at right angles to the path of the light beam. The non-planar reflectors described above may also include materials that can scatter or convert the wavelength of light beams. Examples of such non-planar reflectors include mirrors, scattering pits, and connecting gratings and wavelength-converting fluorescent objectives.

In order to cause only low losses for a light beam from the pixel switcher to the non-planar reflector, an output waveguide 1017 (hereinafter referred to as reflector waveguide) can be used. Since the waveguides are short (at most from one pixel waveguide to an adjacent pixel waveguide), the reflector position can be optimized to minimize reflected background light, thereby minimizing contrast ratio. If the non-planar reflectors are close to the waveguide, they will subtend the larger solid angle of the scattered light and produce a "flare" or background illumination at the pixel, even when the pixel is not positioned by the waveguide switching system The smallest flicker occurs at the mirror position halfway through the pixel waveguide, where most of the light scattered by the waveguide has been transmitted below the level of the surface waveguide and the reflector and into the substrate. , the contribution of two adjacent waveguides to the above flash is equal.

62 to 67 show several embodiments of the non-planar reflector 1018 described above. The aforementioned non-planar reflection is a permanent structure; the actuated pixels are selected by the aforementioned two-dimensional array of pixel switches. Figure 62 depicts a cross-sectional view of a TIR non-planar reflector 1100 whose output passes through the substrate described above. The 90° turning mirror 1108 reflects the input light beam 1110 to become the output light beam 1104 , which passes through the substrate 1101 and is emitted to the rear surface 1112 . The above mirror 1108 can achieve total internal reflection or it can act as a good reflector with a suitable reflective paint. A diffusing screen 1026 or fluorescent layer 1113 may be coated on the rear surface 1112 as described above. The pixel beam 1110 exiting the pixel switcher and incident on the turning mirror propagates in a reflector waveguide 1017 or 1102 . The waveguide makes the beam two-dimensional to create a beam spot at the turning mirror (patterned to match the waveguide). In the vicinity of the turning mirror, the light beam must be confined by a waveguide, which may be confined in a planar waveguide (in the surface of the crystal). In the case of 100% waveguide packing density, permanent channel waveguides cannot be formed. In this case, the inventors would like to have a planar waveguide, or that the input beam 1110 can propagate directly from the polarized TIR switch without being confined by the waveguide in the intermediate region.

The turning mirror is preferably flat, and the input beam is reflected from the mirror at 45° and enters the substrate at right angles to the surface of the crystal. The reflected light beam diverges according to the principle of Gauges beam propagation to generate a light spot on the bottom surface 1112 of the substrate. A 4 micron diameter spot in the above turning mirror produces a 36 micron diameter spot on the backside of a layer of 0.5 mm thick lithium niobate crystals.

The preferred embodiment of the present invention overlays the diffusing screen 1026 with a 1/e ² intensity beam diameter per pixel. If the substrate 1100 is too thin to achieve this overlap, the substrate 1100 may be placed on a light-transmissive support material of appropriate thickness to achieve the desired overlap. Examples of the above support materials include glass, plastic and silicon dioxide. Alternatively, if the mirror surface 1108 is not flat, but has a suitable radius of curvature to expand the 1/ ^e2 diameter to the desired size, or if the folded mirror is intrinsically rough enough to scatter the light beam in the above If a light spot with a desired size is produced on the curtain 1026, the above-mentioned 1/ ^e2 spot size overlap can be formed on the lower lithium niobate surface. Another embodiment has the bottom surface 1112 coated with an optical fluorescent or pigmented layer 1113 which can emit the desired display color after the waveguides and non-planar reflectors transmit light to stimulate the fluorescent layer 1113. For example, the phosphor layer also removes the need for a separate scattering screen.

The turning mirror 1108 shown in FIG. 62 can be fabricated using a number of techniques including laser ablation, reactive ion etching (RIE), reactive ion beam etching (RIBE), ion beam milling, and wet etching. Wet etching is preferred by the inventors and will be described below. The inventors transferred a chrome mask onto the turning mirror 1108 material using standard photolithography techniques. The lithium niobate layer is then soaked in hydrofluoric acid for several hours at about 80°C. Please note that the stirring of hydrofluoric acid should be strictly limited during the etching process. The low mass transport at the bottom of the etch pit limits the etch rate while the etch rate is faster at the edges because of the higher mass transport of etched ions at the edges. Many materials have some nice properties to etch along natural planes, so if a bent mirror is aligned with one of these natural planes, the resulting etched surface will have a lower residual roughness.

As an alternative technique, a mask etched at the same time as the sample described above can control the slope of the etched walls. By sloping the characteristics of the mask at the edge of the mirror, the relative slope can be transferred to the substrate during etching. The characteristic geometry can be controlled by varying the composition and stirring rate of the mask and etch described above.

The shape of each etch pit can be controlled by changing the composition of the region to be etched using proton exchange or polarization. The inventors have known that a proton-exchanged region in lithium niobate etches at a rate greater than that of non-proton-exchanged crystals. By firing the proton-exchanged substrate, the etch rate varies with the cross-section of the exchanged properties such that the etching process produces a sloped surface. The inventors can correct the correct shape of the surface by adjusting the amount of proton exchange and firing temperature.

The mirrors described above can also be formed using pulsed laser ablation. The substrate material is masked by a layer of reflective material such as aluminum. The angle between the aimed laser light and the surface of the substrate is 45° to form a curved facet by masking. The laser can form a small spot (covering only one mirror at a time) or a line (the laser is swept across an array of mirrors at a time). The duration and energy in each laser pulse can be chosen to minimize the roughness of the interface material to produce a reflective surface, and then the time and energy in each laser pulse can be chosen to make the interface material The roughness of is minimized to produce a mirror, which can then be etched using hydrofluoric acid to smooth the surface of the mirror.

In a similar manner, active ion milling or active ion beam milling can be used to form the above mirror. The aimed ion source must be moved a distance from the sample such that the angle between the sample and the laser is 45°. This type of etching also creates a rough surface, which the inventors can smooth out with hydrofluoric acid if necessary. The inventors may use any combination of the techniques described above to obtain better control and fastest fabrication of the reflectors described above.

The mirrors and waveguides shown in Fig. 62 are preferably protected by a light transmissive layer 1106 of a low index material such as aluminum oxide. The material has a sufficiently low index of refraction such that the angular channel width of the reflection of the mirror for all three colors of the display is sufficiently internally reflective. For a lithium niobate substrate covered by alumina, light with an incident angle greater than 36° will produce total internal reflection. If the above-mentioned surface 1108 is not a TIR reflector, it can be highly reflective of conventional technology.

Figure 63 presents an alternative embodiment of a non-planar reflector where light propagating in the waveguide 1102 is scattered by a pit in the crystal as described above. The scattering pits 1116 combine the functions of the mirror and the scattering screen described above. In such an embodiment, light from each pixel is scattered into the substrate away from the opposite side of the crystal. In a preferred embodiment, the normal cone angle 1114 of the scattered light does not exceed the aforementioned TIR angle of 36° so that most of the light exits the substrate towards the viewer. The above-mentioned lower surface can be covered with a layer of non-reflective material to reduce the loss of light energy due to reflection.

In a preferred embodiment, a stirred wet etch process forms the aforementioned scattering pits 1116 . The scattering pits are 45° at their edges, so that when the guided light beam 1110 is scattered, the angle between the center of the scattered angle and the surface of the substrate is 90°. The convexity of the pits in the plane containing the waveguide and the normal surface must be tuned to a preferred condition to obtain a preferred range of viewing angles outside the substrate. Since the pits are circular or elliptical pits in the surface of the substrate, light scattered from the sides of the waveguide 1102 is also diverged in a plane perpendicular to the plane of the portion. The curvature of the pits in the surface of the substrate can be adjusted regardless of the convexity, thereby obtaining a substantially square pixel area. If the surface of the pits is rough, the extra emission caused by the scattering of light reduces the above two preferred curvatures. Since the desired cone angle 1114 of the scattered light is large, the surface of the pits can be roughened by the stirred etching process (or by increasing the etching temperature), thereby forming the desired Divergent reflective surfaces. As an alternative, the inventors can use laser or ion beam ablation to create the pits described above. An additional absorbing layer 1168 is placed behind the substrate to absorb low grazing incidence background light to improve contrast ratio. The absorbing layer 1168 absorbs less than 50% of the normal incident light beam 1104 or 1114 so as not to bring too high brightness.

Fig. 64 depicts another embodiment of the non-planar reflector in which the input beam is reflected out of the top surface of the crystal. The mirror 1126 consists of a reflector structure comprising a flat 45° sloped wall and a vertical wall between which material is removed from the substrate. The input beam 1110 enters the reflector structure at the vertical interface 1122 and propagates across the reflector structure to reflect off the 45° sloped wall away from the reflector structure. The reflected light beam 1124 passes through the top surface of the substrate to illuminate the diffusing screen. All the methods of making this reflector structure can be the same as those used to make the structure shown in FIG. 62 . For optimal reflection of surface 1126, surface 1126 may be coated with a reflective paint, such as aluminum. This layer of reflective paint should be applied at an angle 1121 of less than 90°, so that the surface 1122 is shaded and thus not covered with reflective paint. In a preferred embodiment, the normal surface 1122 should be coated with a non-reflective material. To achieve this, the substrate 1100 can be held at an angle during the application of the coating so that the surface 1122 is exposed.

Finally, the sample is covered with a light-transmitting layer 1106, such as silicon dioxide or aluminum oxide. Surface 1126 may be rough and/or curved to enhance the divergence of light emitted from this portion.

Figure 65 illustrates an embodiment of directing light away from the plane of the crystal, in which a scattering pit directs light away from the top surface of the crystal. As in the case shown in Figure 64, the input beam 1110 penetrates a vertical interface into the reflector structure and scatters out of the curved scattering interface on the opposite side of the etched region. The scattered light beam 1134 (through any coating 1106) preferably has a Lambertian distribution to provide a full cone of 2π steradians. The requirements for the angle of incidence of the light beam and the shape of the curvature are the same as for light entering the scattering pit shown in FIG. 63 .

FIG. 66 depicts an alternate embodiment of the pixel switcher 1016 shown in FIG. 59 in combination with scattering pits 1018 . In such an embodiment, a variable intensity light source as described herein with reference to Figures 50 and 51 is used to create a light emitting pixel. Beam 1406 propagating in waveguide 1402 (equivalent to pixel waveguide 1014 in FIG. 59 ) enters pixel switching region 1420 (or 1400), pixel 1420 (or 1400) in FIG. 1016 peers. If the voltage supplied to the switching region 1420 is turned off, the light beam is guided by the output waveguide 1408 . Then, the output beam 1410 propagates to the next pixel switching area. When the voltage is activated, the two-dimensional waveguide in the switching region 1148 is destroyed and the light beam propagates in the planar waveguide 1409 as before. As shown at 1441 in FIG. 66, the above-mentioned light beam expands freely in one dimension. Within the optimum propagation distance from the switching region 1148, the diameter of the beam expands sufficiently that a substantial portion of the beam propagates out through the scattering pit 1018 to reach the observer. These scattering pits may be any of the pit shapes described in this specification above in conjunction with Figures 63-65. Figure 66 depicts two scattering pits 1018 that maximize efficiency. Such an embodiment can work with a single scattering pit at the cost of half the efficiency.

One requirement of this switching configuration is that the distance between the waveguide and the sides of the scattering pit be large enough for the output beam 1410 to pass through the scattering pit 1018 without suffering appreciable scattering losses. This requirement establishes a minimum packing density and limits the efficiency of such embodiments. Since a smaller waveguide width results in a larger diffraction angle θ of the light beam 1411, by reducing the width of the waveguide 1402, the density of the above-mentioned packets can be increased.

The above-mentioned etching pits may also be occupied by a fluorescent target device. Any material, solid or liquid, that can absorb one or more photons from the photo-stimulator beam with the same or different width of the waveguide 1402, can increase the density of the aforementioned packets.

The above-mentioned etching pits may also be occupied by a fluorescent target device. Any material, solid or liquid, that absorbs one or more photons from the photo-exciter beam and re-emits light energy at the same or a different wavelength can be used as a fluorescent target device. For a three-color display, each color may be produced in a different phosphor target in a different etch pit. It is preferred to use a single wavelength of illumination in one system to excite all fluorescent targets. A fluorescent target comprising etched pits is optically configured such that a substantial portion of the incident beam is absorbed by the fluorescent target and a remainder of the input beam is not reflected from the pits or toward a viewer. The constraints on the curvature and roughness of the scattering pits discussed above are independent of the configuration and roughness of the fluorescent target pits. In a preferred embodiment, a reflector is incorporated behind the phosphor away from the observer, so that the backward fluorescence is reflected back to the observer.

Direct illumination and fluorescent target arrangements can also be used to produce the three-color displays described above. For example, a fluorescent target can be excited by a color of light (eg, blue light) to produce red or blue light. Thus, blue light can be displayed directly (using one of the non-planar mirrors shown in Figures 62 to 67) together with the other two colored lights formed in the two phosphors. In another way, when a phosphor is excited by blue light or red light or a light beam with a third frequency to generate green light beams, the red light beams and blue light beams are directly displayed.

An alternative embodiment of the present invention includes the use of up-converting phosphor target materials or up-converting materials at the pixel locations described above. Infrared light is transmitted and switched through the waveguide of the present invention to the pixel locations described above, which are coated or filled with a wavelength up-conversion material, referred to herein as an up-conversion medium or up-converter. The above-mentioned up-conversion materials include standard phosphors, such as crystal-based rare earth ions, sulfur oxides, dyes, and rare earth-doped glasses. These media convert infrared to visible light through direct excitation, using single or multiple wavelengths, by absorbing two or more infrared photons. At low intensities, the emitted intensity is thus proportional to the second or higher power of the laser intensity. When the aforementioned phosphor is saturated (typically above a few watts/cm2), the intensity of the pixel may be directly proportional to the laser power. Phosphors (such as Yb ³⁺ -Er ³⁺ , Yb ³⁺ -Ho ³⁺ , Yb ³⁺ -Tm ³⁺ in the crystal nucleus) are up-conversion solutions that can generate blue, green and red light from infrared light some examples of . It was found that the infrared-to-visible light conversion efficiency in these phosphors exceeds 10% and the excitation band overlaps well with commercially available infrared laser diode wavelengths. As mentioned above, the excitation beam may be spatially and/or temporally multiplied to converge on the pixel.

In the present invention, infrared rays are directed to specific pixel locations where an up-converting medium is located. Standard coating techniques (such as spin coating or sputtering) and patterning techniques (such as mask etching) can be used to place the above-mentioned up-conversion media. In one embodiment of the invention, infrared light is delivered to the pixel locations through waveguides. Here, the above-mentioned waveguide ends in some areas or recesses (the above-mentioned up-conversion medium is placed in these areas, and the intensity of light is highest in these areas. These areas or recesses can be applied with paint to modify the surface reflectivity, thereby improving Guidance of the excitation light entering the phosphor and changing the emission angle away from the phosphor. Alternatively, a non-planar reflector can be used to reflect the guided infrared to the etched or unetched portion of the display surface, where the above-mentioned phosphor is made into a continuous point or uniform surface layer, and its pixel size is determined by the size of the refracted infrared beam on the above-mentioned surface. For all embodiments, an opaque film can be added to Fills non-pixel areas to better define pixel areas and remove IR and visible background scatter.

The up-converter requires two or more photons to generate visible light, and because the excitation light is most confined in a certain pixel location where the intensity of the excitation light is greatest, the visible light background exiting the waveguide Scattering and switching losses are strongly suppressed, resulting in high contrast values of the display. Since the degree of excitation in these media is quite high, any laser wavelength stabilization is required. Problems associated with laser speckle are also eliminated because the illumination of the above-mentioned up-conversion medium is incoherent. A related advantage of guiding infrared light through waveguides is that most electro-optic active materials have a higher damage threshold in the near infrared and thus can tolerate greater infrared intensities than visible light.

One or more phosphors and/or pigments may be used in the above-described displays to produce monochrome or color displays. Laser excitation can work with a single wavelength beam or with multiple lasers and laser wavelengths. A single phosphor can emit multiple colors, depending on the intensity, timing and wavelength of the light beam that excites it. The color of a pixel comprising such a phosphor can thus be selected according to the intensity of the crossover product incident on the pixel comprising such phosphor. For example, if an up-conversion medium is excited by two IR photons from beam 1 to emit blue light and excited by photons from beams 1 and 2 to emit green light, the intensity of the blue light and the intensity of the green light are respectively proportional to I ₁ ² and I ₁ ×I ₂ , where I ₁ and I ₂ are the intensities of beam 1 and beam 2, respectively. This approach eliminates the inventor's need for a frequency doubler and generates red light directly from a red (AlGaInp) diode laser.

In all display embodiments, the excitation laser can be continuous wave (CW) or pulsed, as long as the pulse repetition frequency is a multiple of the above pixel switching rate (or the pulse repetition frequency is much higher than the pixel switching rate). In the case of non-linear excited phosphors, however, the use of pulsed lasers has advantages in terms of conversion efficiency. Pulse excitation also allows further temporal control of pixel intensity and color. Phosphors have intermediate energy levels that persist for microseconds after initiation of single-photon excitation, thus allowing timed multiplexing to control pixel radiance. In the fluorescence excitation scheme, the energy level originally generated by one or more infrared photons can be directly raised to a higher energy level, which can then emit (or moderate and emit) visible light, or their Can be relaxed to a lower energy level after initial excitation and then excited to another energy level (this energy level emits at a different wavelength, or relaxes and emits). Pulsed excitation also allows higher peak powers to drive the medium into saturation, thereby easing signal-to-optical processing requirements.

Color displays using only such phosphor targets require at least one laser and multiple phosphor targets with different emission wavelengths, multiple lasers and a single phosphor target with multiple excitation and emission wavelengths, or multiple lasers and phosphors body target. Multiple phosphor targets in a color display array can be combined into a single location controlled by a single waveguide or divided into adjacent regions controlled by different waveguides to form a combined single pixel element.

In a preferred embodiment of a color display based on fluorescent targets, one laser is used for each of the three primary colors to excite one phosphor. The waveguide and switching structures allow any of the three laser wavelengths to pass through separate distribution arrays to their phosphor pixel targets, but this results in an effective packing density of that of a monochrome display. The preferred approach is to distribute multiple laser wavelengths through the same distribution array and select phosphors excited by only one wavelength or a combination of wavelengths to generate red, blue and green light. The above lasers can be combined into a single wavelength by the beam combiner shown in FIG. 61 or multiplied in time into a common line by the TIR switcher shown in FIG. 30 .

Figure 67 depicts another alternative embodiment of the non-planar reflector described above. In this embodiment, a non-planar coupler 1146 is used to direct the pixel beam 1144 substantially perpendicular to the substrate surface. The single-mode input waveguide 1102 may have adiabatic expansion 1142 coupled to a multi-mode waveguide 1143 where only the lowest order mode is excited. A permanent grating 1146 placed above the multimode waveguide is phase matched to scatter light normal to the surface of the crystal. The grating is a refracting grating structure produced by manufacturing a reflective discontinuity-like refractive index in the core or skin of the waveguide by any known method. Discontinuities in reflectance may be formed, for example, by a tangible ion implantation process in the core or skin of the waveguide as described above, or by the arrangement of a tangible dielectric structure such as the skin described above. If the multimode waveguide is excited with multiple spatial modes, the period of the grating can be tuned along its length to allow phase matching for most of the modes, thereby improving coupling efficiency. The intensity of the grating can be varied along its length so that the intensity of the scattered light beam can be adjusted to a desired pattern.

The grating can be fabricated by etching the surface of the substrate or the surface of the waveguide, in which case the inventors can obtain variable grating strength by varying the depth modulated by the etched groove along the length of the grating. The inventors have achieved this by varying the duty cycle of the mask in each period across the grating. The etch described above is preferably a wet etch without agitation so that the lower quality transfer results in a shallower etch in narrower areas. Furthermore, the width of the multimode waveguide is chosen to match the length of the grating such that the size and divergence of the beam diffracted off the grating is substantially the same in both dimensions. As noted above, a diffusing screen may be placed at a desired distance from the non-planar lenticular mirror to produce the desired divergence of the illumination. The above-mentioned grating preferably has a protective layer made of a transparent material 1106, such as silicon dioxide. To fabricate the above-mentioned gratings, the inventors can use direct writing techniques such as holography or e-beam lithography to mask the above-mentioned substrates.

To reduce glare and increase display contrast, refractors or absorbers can be used. An optical refractor or absorber can be placed around the non-planar reflector to refract background light that would otherwise hit the reflector and cause flare, reducing the contrast ratio. Each refractor or absorber is optimally arranged so that it shadows at least one non-planar reflector. Since the main source of background light described above is waveguide leakage, much of the background light travels in the same direction as the light in the pixel waveguide 1014 . The above-mentioned refractor or absorber is thus arranged at least upstream of the main propagation direction of the above-mentioned backlight, so that a substantial part of the above-mentioned backlight is blocked. The pixel region 1151 shown in FIG. 68 has an optical refractor or absorber 1153 enclosing the non-planar mirror structure 1152 and one end of its reflector waveguide 1149 . Light in pixel waveguide 1147 is switched into reflector waveguide 1149 in pixel switching region 1148 as described above. The refractor or absorber region 1153 is deep enough to obscure the entire non-planar reflector region 1152 . If it is an absorber layer, the absorptivity and thickness of said absorber is adjusted to substantially attenuate light traveling towards said non-planar reflector. The design of the refractors and absorbers depends on the orientation of the non-planar reflectors since neither should refract the background light into the image direction. The refractor is designed to refract light rays leaving the image in opposite directions.

FIG. 69 depicts an optical refractor 1157 that refracts the aforementioned ambient light away from the surface of the device 1155 . In this example, non-planar reflectors 1152 are arranged in an orientation to direct the aforementioned image light through the substrate shown in FIGS. 62 and 63 . The refractor 1157 is similar in design to the non-planar reflector 1130 or 1140 depicted in FIGS. 64 and 65 . Refraction from the vertical face 1154 of the above-mentioned refractor is preferably minimized, since any reflected fraction of the background light 1156 can cause a glint at any preceding pixel one. The curved surface 1158 is painted to reflect and/or absorb light passing through the vertical surfaces. In this way, no light can pass through refractor 1157 . By making the beam refractor 1157 deeper and wider than the non-planar reflector 1152, and by placing the reflector in its shadow (but not so close as to interfere with the illumination of the reflector by the switched display beam), Minimizes flare on adjacent pixels. An opaque layer 1168 is laid behind the substrate to reduce background light propagating in the substrate. Because of its short path length, the display light beam 1114 is only partially absorbed in the opaque layer 1168, thus the background light (propagating obliquely) is strongly absorbed in the longer path length through the absorber.

FIG. 70 is a cross-sectional view of an optical refractor 1161 that refracts the aforementioned background light into the substrate of device 1163. This configuration may be used when non-planar reflector 1152 is arranged in an orientation to direct image light 1124 away from the surface of device 1163 . The refractor 1161 is similar in design to the non-planar reflectors 1100 and 1120 depicted in FIGS. 64 and 65 . Again, the non-planar reflector 1152 is placed in the shadow 1159 of the refractor 1161, and a back surface absorbing layer may be used.

FIG. 71 depicts a cross section 1167 of an optical absorber 1166 placed upstream of a non-planar reflector 1152 . The overall optical density and length of the absorber must be high enough to substantially reduce the intensity of any light waves passing through the absorber. The refractive index of the absorber relative to the substrate should be low to avoid reflections at the interface between the intact substrate and the absorber region. By orienting the input face of the absorber region to reflect in directions away from the display and other pixel directions, the inventors can minimize the effect of residual reflections.

The above-mentioned absorber layer.

Optical refractors or absorbers may be placed between the above-mentioned distribution waveguides and the above-mentioned pixel array. The refractors/absorbers should not be interleaved with the pixel waveguides such that they leave gaps through which background light generated by the distribution waveguides can pass. However any of these refractors/absorbers can be angled at an obtuse angle to the distribution waveguide so that their shadows overlap each other without gaps. The remainder of the ambient light generated by the distribution waveguide passes through the substrate to impinge on the rear or far side thereof. The absorber plane behind can be configured such that obliquely propagating background light is effectively eliminated. The far side of the substrate may be coated with a non-reflective material (and/or absorbing material) such that the background light is absorbed from the substrate where the background light is efficiently captured and absorbed.

A portion of the light leaking from the waveguide bounces back from the opposite side 1112 of the substrate causing glare problems. In embodiments where the unused face 1112 is used as a scattering layer or is coated with a layer of fluorescent or pigmented light contact, it may be advantageous to apply a layer of absorbing paint to the face to reduce glare. Figures 63, 64, 65, 69 and 70 depict such an absorbent layer 1168. However, the relative coefficient change of the coating must be low to minimize reflections off the interface. Furthermore, when the display direction is through the substrate, the display light should not be greatly attenuated. These three objectives are simultaneously achieved by using an in-body diffused chromium layer behind the above-mentioned substrate. The above-mentioned chrome diffusion layer can strongly absorb light in the entire visible light region with only a small spectral change in its absorption intensity. If the bottom layer of the above-mentioned lithium niobate substrate is covered with a 10% by weight chromium layer in the first 8 microns in the form of internal diffusion, the absorption of light shown in the entire visible light spectrum is roughly 6% of light loss, which is very small. small and can be ignored. However, the aforementioned background light (traveling within 3° of the wafer plane) passes through an absorber of about 300 microns and is attenuated by about 90%. The production of the required chromium layer requires evaporation of about 500 nm of chromium on the above-mentioned backside, followed by baking at about 1100° C. for about 30 minutes. These numbers can be changed by two or three on either side, depending on offset adjustments in other manufacturing parameters.

Finally, if display light is extracted from the surface of the substrate as shown in FIGS. Non-planar reflectors leave openings. This has the advantage of eliminating glare from the areas between the aforementioned reflectors, such as waveguides and switches. J. Electrode

Figure 72 depicts a schematic diagram of a pixel electrode structure for use in a preferred embodiment of the present invention. Each electrode 1164 is made up of a plurality of pixel sections, each similar to electrode 922 depicted in FIG. In FIG. 72, the above array is a linear column and many electrodes are arranged close to each other to form a square array of excitation electrodes. Each pixel electrode is arranged over a polarization boundary forming a TIR interface and over a switched waveguide section. Furthermore, each electrode extends beyond said polarization boundary into an extended portion which may form a second electrode for an adjacent pixel. None of the polarized regions, the reflector waveguides, and the non-planar reflectors are depicted in FIG. 72 . One pair of electrodes is used to excite all TIR reflectors in a row. Most of the above reflectors may be grounded at a given time leaving only one connected to a drive voltage, in which case the excited electrode and the two adjacent electrodes determine the electric field. The electrodes have an irregular structure such that all of the switches in a single row are in electrical communication without adjacent electrodes intersecting.

In order to reduce the generation of radio waves from the electrodes on the above-mentioned panel surface, a ground plate for reducing RF is added. In order to achieve this, the treated surface of the aforementioned substrate is coated with a highly conductive layer of separating material. If the non-planar reflector directs the light away from the top surface of the substrate, then the spacer material must also be light transmissive or have holes at the location of the non-planar reflector. The spacer material can be made thinner than the space between the electrodes to reduce capacitance and coated with a conductive material that forms a ground plane. As another alternative, the above-mentioned conductive material may be light-transmissive, such as indium-tin-oxide. K. Electronic control

Specially designed electronic components will be used to drive the above displays. The electronics required for the preferred embodiment of the invention are categorized into four groups: (1) laser diode drivers and optical power control, (2) modulators, (3) waveguide steering electronics, and (4) central controller.

The laser diode driver described above is responsible for maintaining a constant level for the red, green and blue light source outputs. If the above-mentioned laser light itself is modulated, a constant level of power represents the power level in a particular gray-scale excitation. The inventors need about two volts to turn on the semiconductor diode laser, and the required current depends on the required output power, which is a function of the size and illumination intensity of the display screen. Photodiodes are used to measure power at all three diode laser wavelengths (and any wavelengths that are frequency doubled) and at any point of interest (such as after the integral wavelength modulator). The control electronics will periodically assume the maximum output power of the diode lasers and adjust the maximum supply current to maintain their output power from any slow aging effects. Since it is necessary to reconcile power measurements in the operation of the displays, the functions described above are delivered (and possibly actually performed in) the central controller. The laser driver can also control the optical frequency of the diode lasers (related to the acceptance of their frequency doublers). If the laser frequency or central frequency of the frequency doubler drifts, the central wavelength of the feedback grating is adjusted in relation to the central wavelength of the frequency doubler until the maximum multiplied power is optimal. If the laser is driven with a continuous unmodulated current, the relative frequency can be controlled by a feedback loop. If the laser power is modulated, the frequency controller must be more complicated because the frequency sensing must be done without significant disturbing modulation. In the latter example, frequency control is preferably done during "dead" times (when the laser light is not being displayed, eg between consecutive lines or boxes on the display screen). When less than maximum brightness is desired, the overall current into the three laser diodes is adjusted with a dimmer control. Furthermore, if the ratio of power at different wavelengths is not preferred, the laser diodes can be adjusted relative to each other by a concentration and chromaticity control and/or a second feedback loop.

Three modulators (one for each display) control power and color in all pixels across the display. The best configuration of these modulators is one modulator with one color channel. Since the preferred embodiment of the present invention uses electro-optical modulators with a nonlinear voltage-reflectance relationship (the technique of FIG. 30) the voltage required for each bit is determined from a look-up table. However, since the voltage-reflectivity relationship approaches the gray scale intensity of the coefficient over most of its range, the modulator provides an almost linear relationship between voltage and gray level. The pixels of the display are scanned in succession, with each color modulator simultaneously providing the appropriate intensity to its color components at each pixel. In a preferred embodiment of the invention there is also a third modulator, this third modulator is called a bias modulator and modulates the intensity of the combined three-color light. The purpose of the bias modulator described above is to compensate for the loss of the varying length of the waveguide between the laser light and the pixel. The data (obtained from waveguide length variation analysis) can be compensated according to the look-up table to obtain the third modulator. The above key data can also be derived from measurements on a given production unit to allow compensation for loss of irregularities due to symmetry and asymmetry effects (including optical imperfections). Alternatively, the bias voltage modulation described above can be performed in separate modulators in each of the three color channels (separate from the image modulator described above), or in combination with the image modulator in the image color modulator described above .

The above-mentioned waveguide-guiding electronic components control the execution order of the above-mentioned switches. Light propagating in the above-mentioned distribution waveguides is switched one by one into the above-mentioned pixel (row) waveguides at a pixel rate of 24.6 MHz for a simple VGA architecture with a frame rate of 80 Hz. The row electrode that is activated determines which row of pixels is illuminated when scanning a column. Line switching is relatively slow, about 38.4KHZ for a simple VGA configuration. Because the capacitance of short distribution electrodes is much lower than that of long pixel electrodes, the distribution controller has been selected to switch at the fastest rate, minimizing the drive current required for the display. This is why the distribution electrodes are preferably used to switch the above grid. Higher capacitance pixel row electrodes operate at low rates. The switching sequences used for both controls are preferably sequential, although alternative row scanning could also be achieved with similar controls.

The above-mentioned central controller preferably has micro-processing capability and is responsible for coordinating the functions of the above-mentioned electronic component modules. It preferably has control lines so that the same electronics modulus can be used to display different formats such as VGA (640x480), 800x600, 1024x768, 1600x1288, HDTV, etc. If the number of pixels is less than the number of pixels implemented in a particular display, pixels outside the range are not activated. As higher pixel computing display devices hit the market, the same electronic drivers can still be used to drive them, tying up a few input control lines. L. Diffusion screen

The diffuser 1026 is a dispersing surface that allows the screen to have a preferred large stereoscopic viewing angle of 2π radians. The light beam reflected from the specular non-planar reflector in FIG. 62 is split from the surface containing the reflector until it reaches the diffuser 1026. The degree of separation is determined by the distance between pixels, that is, adjacent pixels must overlap by more than 30%, thus forming a continuous image. The diffusing surface may be the rear surface 1112 of the substrate comprising the waveguide, or a separate substance such as silicon diodes placed or attached to the substrate. The surface of the screen is roughened to form dispersed particles of a size distribution that achieves the desired distribution of angular spread.

As shown in Figure 63, in a preferred embodiment, the non-planar reflector is a non-specular diffuser and the array 1116 of non-planar reflectors constitutes the screen. In this way, the diffusion screen is not necessary. In this case 1116 make up the pixels and disperse the light rays roughly into a stereoscopic view of 2π radians in the Lambertian intensity range.

For some applications it is desirable to narrow the solid angle to the point where the display light disappears. For a given display light energy, concentrating the light energy improves brightness when viewed from an angular area encompassed by the display. For example, reducing laser drive light energy by narrowing the display solid angle to preserve brightness is a useful stopgap measure when battery light energy needs to be conserved. M. Control markings

Speckle is an observable phenomenon in which coherent laser light in space forms a fine-grained image reflecting or propagating from a diffuse surface. Because interference causes a uniformly illuminated disc with a laser to appear as speckles of light and dark areas, the size of individual speckles in this display is larger than the pupil due to the small pixel size, and in terms of perfect coherence, when the viewer moves the head, the pixels randomly Alternately, each pixel has its own independent speckle group that produces 100 percent modulation of the pixel. Such a visual experience is a distorted display, which is uncomfortable to watch. There are several known prior art techniques to eliminate the speckle phenomenon, most of which are to move the speckle group in space to be uniform.

There are several ways to eliminate the perceptual impact of speckles. First, if the laser is operated in a single-length frequency mode, the laser wavelength can be scanned at a rate that changes the speckle penetration over a speckle area in less than 50 milliseconds. Secondly, the laser can operate in multiple length and frequency modes. Each length mode produces different speckles when combined with other length modes. Between different modes, due to the oscillation interference of different frequencies and different modes, interlaced speckles are also formed. Usually, the speed Much faster than the eye can see, these interlaced markings are perceived entirely in a single instance. The speckle formed by the interference of the speckle itself is stable in space, but its peaks and troughs appear randomly. In multiple modes, these speckles overlap in the field of vision and average out quickly as the number of modes increases. . In this case, the eye perceives only a single light zone. Third, displays can be designed to use multi-mode waveguides instead of single-mode waveguides, since each waveguide mode has a different spatial allocation which forms different speckle groups, about ten groups of waveguide modes are sufficient to alleviate the speckle problem. The preferred embodiment uses a multiple length mode visible laser to reduce speckle. N. Hierarchy of distribution arrays

It is useful to think of a stacked waveguide structure as a color video display, where each layer assigns one of three separate constituent colors to each pixel location. In a preferred embodiment, Figures 72 and 74 illustrate the use of a three-dimensional stacked array structure in a color video display using one of the pixel units 1600 and 1601 as an example. This display is shown with a polymer film 1602, silicon dioxide 1604, the waveguide surface 1606, and the electrodes 1607 stacked on the soluble substrate 1620. The pixel directional switch of each layer is individually addressed and simultaneously controlled by the multiple output voltage control interface as shown in FIG. 34 . The laminated waveguide 1608 can be terminated at a surface at a 45 degree angle at the pixel location by a non-planar reflector or, as previously described, by a diffuser to make a permanent non-planar TIR unit reflector as shown in FIG. 73 1610, passing light 1616 through the substrate. Alternatively, the stacked waveguide 1608 can be terminated at a surface perpendicular to the waveguide 1612 of the external phosphor or at a light reflection 1618 of the coating surface 1614 away from the substrate, similar to that shown in Figures 64 and 65 . O. Conclusion of display making steps

The plane making steps of the preferred embodiment are as follows:

1 Electric field electrodes for the substrate of the TIR switching structure, and a switchable grating.

2. Proton exchange and heat treatment for making waveguides and integrated lenses.

3. Etching of microlenses and other etched gratings.

4. Insertion of silicon diodes in the outer layer of the waveguide.

5. Metal deposition used to make electrodes that provide micromirror reflection and excitation switching and modulators.

6. Drive electronics integration.

7. Laser merging and input coupling. P. Alternate Embodiment

An alternative embodiment of the present invention is to use raster switching to combine three color beams as in FIG. 75 . This structure can be connected to component 1004 in FIG. 59 or made as substrate 1028 in FIG. 60 or used to generate source 1000. The three waveguides can be parallel and together form a distribution conveyor to form the multi-waveguide distribution switch 1008, or all three can be used in other locations and in other environments. The first waveguide 1172 confines red input light rays 1182 , the second waveguide 1174 confines green input light rays 1184 , and the third waveguide 1176 confines blue input light rays 1186 . This distribution switch consists of three voltage-activated gratings similar to those shown in Figures 7, 8, 11, 12 and 13. The gratings are switchable and have actuating electrodes. During the manufacture of each grating, the light propagating to the waveguide must be divided These gratings reflect light perpendicular to the distribution waveguide conveyor in a narrow band, so beams of other wavelengths beyond the grating are not reflected. The light emitted from the three distribution waveguides is collected and refocused by an integrated optical lens system 1178, which focuses the beams onto the pixel waveguides 1180 to form pixel beams 1188. This lens system should properly couple the reflected light to the pixel waveguides, Since the length of the reflective grating is much longer than the width of the waveguide, the presented beam must be refocused to the output waveguide. Component 1170 is a two-color beam combiner.

The shape, duty cycle and interval of the grating should be designed so that the reflection mode profile can coincide with the preferred mode profile of the waveguide 1180 after passing through the lens system. Coupled to a single mode pixel waveguide, for optimal light energy coupling, the planar waveguide should be placed between the waveguides 1172, 1174, 1176 and the lens 1178 and waveguide 1180, the planar waveguide can alternately cover the entire spanning structure 1170 substrate.

An integrated lens is a permanently index-modulating structure that can be realized by proton exchange, non-diffusion, or etching. The technique of proton exchange is preferred because of the realization of high index refraction and lower dispersion than etched lenses.

Figure 36 shows an alternative embodiment of a waveguide distribution structure in which two or more beams are split to confine one set of distribution waveguides and one set of pixel waveguides, more than two sets of distribution waveguides and pixel waveguides can be arranged in this way To implement this arrangement, the additional distribution waveguides are added in parallel and on top of the other distribution waveguides 982 and 984, the additional pixel waveguides are interleaved with the previous asymmetric loss distribution waveguides, similar to minimizing losses in the distribution conveyor 997 The way.

This embodiment is particularly useful for high density displays where the length of the switches 983 and 985 is greater than the required pixel space, and multiple colors can be assigned to each of the aforementioned waveguides or a group of three waveguides, each representing a color. If an output waveguide is used to carry the laser light energy from the pixel switch to the non-planar reflector, then asymmetric loss waveguide interleaving should be used to avoid extreme losses in the pixel waveguide.

Conveyors that distribute waveguides are also quite useful in large or high optical power displays where optical disruption of the waveguides is inevitable. If the laser density in the waveguide exceeds the damage limit (about 10MW per square centimeter in lithium niobate), the laser energy will severely damage the waveguide to form pits or fragments or metal areas, so the intensity must be kept at the damage limit with a double safety factor Under such limitations, the propagation of the total light energy in the waveguide is limited (approximately 800 milliwatts in a 4 square micron lithium niobate waveguide), and if more energy is required, the waveguide conveyor can be used to carry the total required energy.

Another embodiment is a display that uses a passive optical energy router for pixel switch 1018 in FIG. It leads into a reflector waveguide, which can be of any length with any curvature, which ends in a non-planar reflector. The spatial position of the array of non-planar reflectors forms a static image. Each pixel waveguide of these non-planar reflectors has its own unique set, with a distribution switch connecting its position to a distribution waveguide, with a given pixel The image associated with the waveguide can actuate the distribution switch to open and close. Q. Electric tile display

FIG. 76 shows a schematic diagram of a display 1200 divided into tiles 1202, which are sub-arrays of pixels, similar to those shown in FIG. 59, that can be independently controlled. Several tiles can be arranged together and controlled collectively to form a display. In this embodiment, each tile is fabricated on a single substrate to maintain registration of pixels between tiles because the display is partitioned into pairs of tiles. Should not be visible to the viewer, each electrical watt is placed in its own laser source from the diode laser arrays 1204 and 1206 using the distribution waveguide 1206 so that the display emits more light than the single laser source embodiment of FIG. 59 Greater laser light energy is also particularly advantageous for a projection display.

In general, tiles can be distinguished in different ways, for example, they can have different substrates, different electrode stimuli, and different laser sources. In Figure 76, electric tiles are distinguished by emitting light through a different waveguide. Diode laser array 1204 is connected to waveguide feeders 1206 and 1208 , distributor 1008 and pixel 1016 switches and electrodes are not shown in FIG. 76 .

At least one waveguide carries light from each conveyor to the array tiles. The fiber waveguide distribution can be arranged in a variety of ways, for example, in Figure 76 the first waveguide of the lower order conveyor carries light to the lower order right tile, below One waveguide is carried to the left, as are the other waveguides on the bottom line, a conveyor 1208 can be used to similarly carry light to the next row of electrical tiles, only partially shown in the figure. Other conveyors can be constructed for additional tiles, or a single laser array can be used to provide light to many waveguides and then to many conveyors into tiles, or a combination of these methods.

The laser vision display that feeds each waveguide needs to modulate it with a video message, and this excites the electrodes of the display to select rows and columns to emit light at a given time. For a full-color display, several waveguides can emit a given electrical watt, or several laser beams can be combined into one waveguide that emits each electrical watt. Conveyors 1206 are considered parallel waveguides because they are located at the edge of the display and their space is available. The light from the waveguides in each conveyance enters the waveguides of electrical tiles from switches 1008 and 1016 and the electronics. Conveyor 1208 Shown is a waveguide that includes a packing density optimized region where separation of the waveguide is mitigated before traversing the display.

Since the total space of the waveguides between tiles is smaller than the separation of the pixels, the waveguides need to be tightly packed together, for example, if all the space is available, maybe ten 4 micron waveguides can effectively fit into the 100 micron inner pixel gap. Since there is unused space between the non-planar reflector 1018 and an adjacent pixel waveguide, activation of the distribution conveyor along the row direction is necessary. However, between the rows of a dense display, only four or five waveguides can be used. In the latter case, the conveyors can be separated into sub-conveyors, which work along the space of adjacent rows until they reach the electrical tiles. . If the conveyor operates in the column direction, at least some of the reflector waveguides 1017 are interleaved (because of their skew angles), resulting in an additional loss for compensation to create a single display, these interleave with reflector waveguides having asymmetric loss interleaving, But the losses at reflector waveguides interlaced with many transmitter boundaries can vary greatly. While the light energy per watt is limited by the distributed waveguide light energy, the display light energy will increase as the number of waveguides to couple the laser light increases.

As shown, each distribution waveguide in the tile is coupled to a light source that is modulated according to the tile's pixel activation group, which is also segmented according to the display tile. The selection switch of row and column is periodically repeated between the display tiles, and the rows and columns of each tile can be eliminated at the same time. For example, if a VGA has 16 tiles, each tile contains 160×120 pixels, Row 160 and column 120 on the display are activated simultaneously. The 160 rows are scanned continuously and repeated, and the same is true for the columns. To achieve a planar rate of 80 Hz, the rows are scanned at 1.53 MHz and the columns at 9.6 KHz. In a square array, the tiles can be stacked in direct alignment with each other or Stacked next to each other, or possibly in a staggered array.

Electric tile displays can also be used to make large area displays. In this embodiment, multiple pixel arrays are fabricated on separate substrates and connected to a second substrate of glass or plastic. Each electric tile is fabricated as shown in FIG. 59 , and operating as a sub-display area, laser light can be connected to each electrical tile from the outer edge of the laser light using fibers connecting it from the light source to the appropriate electrical tile.

FIG. 77 shows an alternative embodiment of a three-color display in which distribution reflectors 1272 and 1274 reflect only part of the light from distribution waveguides 1262 and 1264 to pixel waveguides 1292 and 1294 . The distribution reflector is not a switch but a permanent passive optical energy router that acts like a beam optical energy splitter. The preferred method of manufacturing this type of optical energy splitter is to etch a narrow and shallow groove inclined at 45 degrees to the distribution waveguide at the junction of the two distribution waveguides. Such display applications can preferably be used in conjunction with adjusting the reflection coefficient of each distribution reflector so that approximately equal power is introduced into each of the waveguides 1292 and 1294 connected to the distribution waveguides. The adjustment of reflectivity must be based on the number of reflectors bonded to the distribution waveguide and the length and loss coefficient of the waveguide itself. The reflectivity of each reflector is adjusted by changing the width or depth of the etched groove. If asymmetry loss occurs due to interleaving of the distribution waveguides as described by element 997 in FIG. 36, then multiple distribution waveguides can be combined into a single structure. Distribution waveguides of this class are preferred and low-loss waveguides with the same interleaving. Other distributed waveguides also conform to the above.

In addition, each pixel waveguide contains a separate modulator 1282 and 1284. Since there is an array of modulators in this structure, light energy abandonment control for each modulated beam is necessary to remove unnecessary light and prevent background light. Such light energy abandonment control can be implemented in several alternative ways. It may be a waveguide capable of propagating light away from the modulator, through the pixel and off (or through) the edge of the light energy absorbing substrate. If the loss of the waveguide is small enough, it is a good choice, but if the loss is large, the waveguide will discard most of the light energy as the length of the light energy absorbing substrate increases. The light energy abandonment control can also be a kind of beam refractor, similar to the structures mentioned in FIG. 69 and FIG. 70 . As mentioned above, light energy other than the display direction must be discarded, and the discarded light energy should be absorbed as much as possible. An alternative light energy rejection control is an absorbing substance placed near each modulator. The waveguide can guide the removed light into a groove filled with an absorbing substance, or through a non-diffuse absorbing region, such as the chrome-painted region described above. Pixel switches are also preferred embodiments as are non-planar reflectors, as shown in FIG. 59 . One of the advantages of this configuration is that the high current drag required to drive the electronics at high speeds can be eliminated in favor of low frequency (38.4 kHz) modulator arrays. As before, the pixel switch row is driven at 384 kHz for a simple VGA with a scan rate of 80 Hz. R, projection display

Figure 78 is an embodiment of a laser bar-pumped monochrome display. Laser 1302 is a light source that includes multiple semiconductor diode laser groups on a wafer and is coupled to a pixel waveguide array on the side of substrate 1300 in the display. A distribution waveguide does not require such an embodiment. During the manufacture of the laser beam and the display, the laser amplifier is adjusted so that the separation degree of the laser beam is lower than the separation degree of the waveguide of the display until the waveguides are approximately equal. An array of lasers on the beam is aligned with an array of pixel waveguides. Coupling can be achieved using an edge coupling that allows light to couple to the waveguide through the end of the waveguide on one side of the lithium niobate wafer (bump coupling or lens coupling). There is another planar coupling method to achieve coupling, that is, the light energy is coupled to the waveguide through the surface of the lithium niobate wafer. Each pixel waveguide preferably spans a modulator 1306 and a frequency selective reflector 1304 that is frequency stabilized as a diode laser. A set of pixel transfer arrays 1016 are fabricated along each waveguide. Each pixel switch reflects the beam to a non-planar reflector 1018 . Simultaneously, a series of non-planar reflectors 1019 are driven such that all modulators act simultaneously.

Multiple laser beams can be simply changed in this structure, and can be used as the light energy of a single display, and can be used to increase the total effective light energy. The modulator can be operated alternately in the laser beam if the laser excitation electrodes can be fabricated differently and easily accessed independently, however compared to gallium arsenic modulators, the modulator on the lithium niobate substrate reaches the control electronics is relatively easy, because the routing work is easy from the immobilized electrons from the surface to the control electrode of the modulator under the same interlayer. If the laser frequency is doubled, the polarization frequency doubled area will also be processed periodically along with the waveguide, support and the above-mentioned control mechanism. Finally, a laser amplifier can be coupled to an array of waveguides that form a distribution delivery device connected to a tile-covered display. In fact, the display shown in Figure 78 is composed of electrical tiles with a single row and multiple columns.

The light power requirement of the display under the structure of Fig. 78 is very high, because there are multiple laser light sources and no one waveguide limits the light power demand of many other waveguides. High optical power displays are quite specific for use in projection displays, and the present display considerations are projected onto a diffused screen separate from the substrate using standard projection optics techniques well known in the art. This projection optics technology collects most of the light emitted from the display and focuses it on the screen so that it is directly visible to the viewer, for example, the phosphor screen can be used as a large projection video screen or as a screen in a movie theater. For a projection display, the fixed angle at which rays are emitted from individual pixels should be as small as possible, which implies that smooth surfaces should be used for non-planar reflectors. There is also an advantage to having a larger interleaving area for reflective waveguides in that it reduces loss and tapering the pixels or reflective waveguides also reduces loss.

The display device in Fig. 78 can be increased to full color, that is, adding pixel waveguides of each color to get final product. The dimensions of the waveguide appearance must be such that three separate pixel waveguides (one for each color) are tightly and efficiently merged together. For example, if laser beam 1302 produces blue, an array of lasers producing green can be coupled to the other side of the display using the same coupling technique as the blue laser beam, but with green pixel waveguides interleaved with blue pixel waveguides , and its propagation direction is relatively opposite. As mentioned above, blue and green laser arrays may include infrared laser arrays with variable frequency regions to generate visible light. If the red laser array is coupled to the display waveguides using a different technology, then the red laser array can also be coupled to the interleaved array of pixel waveguides at the same time. For example, if the blue and green light beams are edge-coupled, the red beam can be surface coupled. In the surface coupling process, the blue and green waveguides can pass the red laser beam without substantial loss.

Figures 79 and 80 show a surface coupled processor 1310 in which a laser beam is coupled to a waveguide of a planar substrate. Three sets of laser beams 1312, 1314, and 1316 are coupled to the third set of waveguides of waveguide array 1318, one waveguide array for each laser beam, and the spacing is equal to the spacing between surface coupling mirrors designed to receive the output of the laser beams. In the illustration, the laser beam 1312 has two sets of waveguides 1320 and 1322, but the entire set of laser beams can have more laser amplifiers, 1320 and 1322 are arranged side by side with waveguides 1324 and 1326 on the substrate, and the laser beam 1312 is arranged on the surface coupling On the coupled mirror arrays 1328 and 1330 , such that the beam emitted from the laser waveguides 1320 and 1322 is reflected from the coupled mirrors 1328 and 1330 into a reflected beam, which propagates coaxially to the waveguides 1324 and 1326 . The waveguides 1320, 1322, 1324, and 1326 are sized to approximately the maximum such that reflected beam coupling occurs. The residual waveguide away from the surface coupling mirrors can be attenuated for other purposes such as amplification performance of the laser beam 1312 and subsequent frequency doubling or modulation or switching or display in the substrate 1332 .

Other diode bars 1314 and 1316 can also be coupled to the waveguides on the substrate in a similar manner as shown and the waveguides and different diode bars can be interleaved in different ways.

FIG. 80 shows an expanded vertical region passing through the member located in one of the rotating mirrors 1330 in FIG. 79 . The fabrication of the surface coupling mirror 1330 may refer to any of the techniques described in FIGS. 62-67 . The coating 1334 on the waveguide 1326 is a transparent material with a lower refractive index, such as silicon or rock, and its thickness is sufficient to contain several coefficient attenuation lengths under 1326 propagation. If there is damage to the coating The structure of the waveguide 1326 will not have a significant impact on the loss. Reflective coating 1331 is moderately applied to the angled surfaces in 1330, vertical surfaces 1333 cannot be coated with any reflective coating, so coating 1331 must be applied at an angle. In order to reduce the Fresnel coupling loss of 1333, non-reflective material can be coated on the surface of 1333. Coating 1336 is preferably applied to the front facet of the diode laser so that the laser facet in contact with substrate 1332 is not damaged. If it is easy to install, then the fixing seat 1338 can be pasted on the surface of the 1312 and one side of the substrate 1332 through the coupler 1340. If the 1338 has a good thermal conductivity and is connected with a thermal conductor for heat dissipation, it is also suitable for the heat dissipation of the diode laser. quite useful. S. Stereoscopic display

Three-dimensional stereoscopic image display can be achieved using the array technology described herein. In stereoscopic display, each pair of eyes observes a slightly different image. This has been done in the past related prior art, for example, projecting two orthogonal The polarized image is displayed on the screen, and then crossed polarized glasses are worn to look at the screen. This technology causes problems due to head movement, because the head movement will put the rotating parts of the two images into each other, replacing the existing technology. Glasses that use a set of circular lenses that include light valves that switch independently for both eyes, that is, when the light valve for the left eye is open, the light valve for the right eye is closed, and vice versa. A synchronized display that produces different alternating images for the two eyes creates the stereoscopic effect. This technique is limited by the switching speed of the display, as the previous afterimage is superimposed on the new image in the other eye to create a chaotic image. Further methods that have been used to construct circular lens glasses include Two separate small displays (LCD or CRT), which simultaneously display two images for each eye, present the same weight and price problems as the prior art. In each display, only one image is seen by each eye because the other image is obscured by some objects.

The same effects can be achieved by the embodiments described below. In the preferred embodiment illustrated in Fig. 59, the light traveling from the display to the viewer can be designed to be polarized without the use of thin-film polarizers, thus enabling the implementation of polarized technology for stereoscopic displays without the initial use of CRT or LCD displays The resulting loss of efficiency because they must use thin-film polarizers to form the polarizing effect. If a TIR switch is used, the two sets of polarized light can be routed through the binary polarized light support waveguide, but the reflection coefficient of the TIR switch must be much larger than that of the TM polarized light (Z-section of the lithium niobate substrate), so that it can be purely TM designs an efficient display. If the switch is a knife switch, only one polarized light can be reflected from the 90 degree waveguide junction (Brewster's angle), along with the beam that is TM polarized (the Z profile is the preferred profile in the lithium niobate substrate).

Because waveguide switches have a short interaction length, a pixel footprint of less than 100 microns is possible in the technique described here. In this way, VGA high-resolution displays with a screen diagonal of less than three inches can be manufactured, and these small displays can be installed on the above-mentioned prior art glasses with circular lenses.

The advantage of this display over a CRT or LCD in this application is that the substrate is actually transparent because the opaque area of the non-planar reflector creates a pixel, so the transparent substrate can make the area between pixels 100 times smaller Shrinking, that is, the area of the opaque display is less than one percent. If a diffuse screen is fabricated as part of the non-planar reflector described in Figure 63, the display will appear transparent if the viewer focuses his eyes on a point outside the image plane. In this case, a head-up or "view-through" head-mounted display can be realized, and this technology can be applied to a variety of displays, including virtual reality scenarios, home entertainment and military aircraft.

Figure 81 illustrates an embodiment of a compact stereoscopic display comprising two sets of pixel arrays 1706 and 1708 fabricated using the same design and oriented on two separate substrates 1702 and 1704 as indicated by the axes in Figure 81 Note that array 1708 has been rotated 90 degrees relative to array 1706 so that the light emitted by array 1706 to the viewer has been polarized orthogonally to the light emitted by array 1708 . The corners of the display are right angles as indicated by line 1710, which allow for good registration of the two images. Because the substrate thickness is less than 0.5 mm, the light emitted by the combined display will appear to come from nearly the same plane, as long as array 1708 uses non-planar reflectors as shown in FIGS. With non-planar reflectors, it is possible to achieve nearly identical emission planes for both arrays. Although a square array of pixels is illustrated, any aspect ratio may be used provided that the arrangement of the non-planar reflectors for each array allows for orthogonal polarization of light emitted by arrays 1706 and 1708.

The operation of the two arrays is similar to the technique used for the display in FIG. 59 . The viewer wears a pair of glasses with polarizers arranged orthogonally so that a pair of eyes can only see the light emitted by one array, and the image displayed by each array is formatted using known techniques to achieve stereoscopic For visual effects, this embodiment can be used for direct viewing or a stereoscopic projection application if a lens is used to project the image plane onto a larger viewing screen.

The present invention is explained with reference to the specific embodiments described. Other embodiments are well within the ordinary skill of the relevant art, therefore, the present invention is not intended to be described except as specified in the appended claims (which form a part of the description of the invention). No other restrictions apply.

Claims (56)

1. A beam guiding element comprising: A solid material that transmits light; at least one material of a first conductivity type forming a first electrode contacting said solid state material; a pixel waveguide segment transverse to the solid material along a plane of the solid material; a plurality of output waveguide sections transverse to the solid material along the plane, at least one of the output waveguide sections transverse to the pixel waveguide section at a first junction; an active light redirector, disposed adjacent to the first interface region, the active light redirector reflects light from the image waveguide section to the output waveguide section, on the first electrode selectively applying an electric field; and light reflector means positioned at selected locations on said plane and in-line with said output waveguide segment to project light out of said one face. 2. A component according to claim 1, wherein said light redirector comprises a switched internal total reflector. 3. The component of claim 1, wherein said light redirector comprises a switched grating reflector. 4. The component of claim 1, wherein said light redirector comprises a waveguide switch. 5. A beam directing structure used to form an array using at least one element according to claim 1 to provide at least one second junction region, further comprising: a distribution waveguide abutting at least one pixel waveguide segment of said element at said second junction, each second junction having an active light director disposed at said second junction to direct light into the pixel waveguide segment. 6. The beam directing structure of claim 5, further comprising a stationary mirror disposed between said light director and said light redirector to direct light in said plane. 7. The element of claim 1, wherein said light reflecting means comprises: A light-reflecting boundary in the solid state material, the reflecting boundary having a discontinuous refractive index across a surface transverse to a light path and at an angle to incident light. 8. The element of claim 1, wherein said light reflector comprises: In the solid material and there is a grating structure of refractive index, the grating structure is formed by a discontinuous pattern of refractive index, and its refractive index is transverse to the incident light transmission axis and in the incident light plane. 9. The element according to claim 1, wherein at least one light reflector means comprises: A fluorescent target device disposed on the plane at a pixel location collinear with the output waveguide segment. 10. The device of claim 1, further comprising an electrode associated with said light redirector, said electrode disposed in said first interface region to control the path of light. 11. The element according to claim 5, further comprising a plurality of electrodes arranged in rows and columns and connected to said first and second junction areas, a common column of said electrodes is connected to a common conductor and said conductor The body is then coupled to a device for generating a field. 12. The component of claim 5, further comprising an optical frequency source means comprising an optical exciter coupled to said distribution waveguide. 13. A component according to claim 12, further comprising a frequency doubling device coupled to said optical actuator and actuated by said optical actuator. 14. The device of claim 13, wherein said frequency doubling means is a periodically polarized structure in said solid state material, and wherein said solid state material is a non-linear periodically polarizable optical material. 15. The component of claim 12, further comprising a beam modulator located at the output of said optical actuator. 16. A component according to claim 15, wherein said light modulator is integral with said solid state material. 17. The component of claim 12, comprising a plurality of distribution waveguides, wherein said optical actuator is a solid state laser. 18. The component of claim 12, comprising a plurality of distribution waveguides, wherein said optical actuator is an array of solid state lasers integrally provided with a single substrate. 19. The device of claim 12, wherein said solid-state laser is aligned to direct energy transverse to said plane of solid-state material, said plane comprising mirrors to direct injected radiation into said distribution waveguide. 20. The component of claim 19, wherein said solid state laser is alternately coupled to said distribution waveguide. 21. The component according to claim 5, further comprising an optical frequency source device comprising a plurality of laser exciters, each exciter being used to generate beams of different optical frequencies, a beam combiner coupling each laser An exciter is used to combine light beams of different optical frequencies, and the output of the beam combiner is coupled to the distribution waveguide. 22. The component according to claim 5, further comprising a waveguide intersection configuration having an asymmetric loss for a plurality of distribution waveguide sections in parallel in said second interface region, wherein at least one distribution waveguide section has a the first index of refraction distribution in the intersection region; and where The at least one pixel waveguide section is transverse to the at least one distribution waveguide section, and a second refractive distribution ratio is substantially less than the first refractive distribution ratio near the intersection region with the distribution waveguide section. 23. The element of claim 1, further comprising: a waveguide intersection structure having asymmetric loss for use in a plurality of parallel pixel waveguide segments in said first junction region, wherein at least one pixel waveguide segment has a first refractive split in an intersection region; and among them At least one of said output waveguide segments is transverse to said at least one pixel waveguide segment, and a second refractive distribution index is substantially less than said first refractive distribution ratio near said intersection region with said pixel waveguide. 24. The element of claim 1, further comprising: means positioned adjacent said reflector means to block false light illumination around said light reflector to prevent false light illumination from being projected along a viewing axis. 25. An element according to claim 24, wherein said shielding means comprises a reflector to direct false light illumination out of said viewing axis. 26. A component according to claim 24, wherein said shielding means comprises a light absorber to intercept spurious light exposure. 27. The element of claim 1, further comprising: means disposed along the surface of the solid state material and transverse to a viewing axis to absorb spurious light illumination not directed out of the solid state material by the light reflector. 28. The element of claim 1, further comprising: A device arranged along the surface of the solid state material and transverse to a viewing axis to block spurious light that is not directed out of the solid state material by the light reflector. 29. A beam directing structure comprising: a solid material for transmitting light; a distribution waveguide section transverse to said solid material along a plane of said solid material; at least one pixel waveguide section transverse to said solid state material along a plane of said solid state material, and each pixel waveguide section abuts said distribution waveguide section at an interface; at least one passive light director disposed at said interface region, said passive light director configured to reflect a portion of said light from said distribution waveguide into said image waveguide section; and A light modulator is used for each of said at least one pixel waveguide segment to modulate said portion of said light respectively. 30. The beam directing structure of claim 29, further comprising: an output waveguide that absorbs light from the at least one pixel waveguide segment; and A light reflector device disposed on the plane at a selected location and in-line with the output waveguide segment to project light out of the plane. 31. The beam directing structure according to claim 29, wherein a plurality of passive light directors are arranged along said distribution conduit section, each passive light director passing a selected amount of light and reflecting a complementary amount thereof, said selected amount Quantities are established according to a desired dispensing pattern. 32. A display device comprising: a solid material used to transmit light; at least one pixel waveguide segment transverse to the solid state material along a plane of the solid state material; a plurality of output waveguide segments transverse to the solid material along the plane, the output waveguide segments resting on the at least one pixel waveguide segment at a plurality of junctions and coupled to receive signals from the at least one pixel waveguide segment light energy from the pixel waveguide segment; and A light reflector device disposed on the plane at a selected location and in-line with the output waveguide segment to project light out of the plane. 33. The apparatus of claim 32, further comprising: a distribution waveguide segment abutting said at least one pixel waveguide segment at selected junctions, each junction having an active light director disposed at said junction to redirect light to said at least one pixel in the waveguide section. 34. The device of claim 33, comprising a plurality of electrodes, selected electrodes being positioned at said interface to control the path of light. 35. The apparatus of claim 32, further comprising: a first distribution waveguide section abutting said first set of pixel waveguide sections at selected first junctions, each first junction having an active light director at the selected junction to redirect light one of the first set of pixel waveguide segments; and a second distribution waveguide section abutting the second set of pixel waveguide sections at selected second junctions, each second junction having an active light director at the selected junction to redirect light One of the second set of pixel waveguide segments. 36. The apparatus of claim 35, wherein said first waveguide segments are interleaved with said second waveguide segments; and wherein The light reflector devices are arranged in a two-dimensional array corresponding to image elements of a display. 37. The apparatus according to claim 35, wherein: a first set of said light reflectors associated with said first distribution waveguide in a row and column arrangement in a first tile of a display; and The second set of light reflectors associated with the second distribution waveguide is arranged in rows and columns in a second electrical tile of the display. 38. The device of claim 32, wherein said light reflector device comprises: A light-reflecting boundary in the solid material, the reflecting boundary having a refractive index that is discontinuous across a surface that is transverse to a light path and at an angle to incident light. 39. A display device comprising: a solid material used to transmit light; at least one first conductivity type material forming a first electrode contacting the solid state material; at least one pixel waveguide segment transverse to the solid state material along a plane of the solid state material; a plurality of output waveguide segments transverse to the solid material along the plane, the output waveguide segments abutting the pixel waveguide segments in a plurality of junction regions; a plurality of electrically controlled light redirectors each transverse to one of the interface regions, the light redirectors being capable of reflecting light from the pixel waveguide section to the output waveguide section; and A fluorescent target device is arranged at selected pixel positions on the plane and is collinear with the output waveguide segment to re-emit light out of the plane under optical excitation. 40. The display device according to claim 39, further comprising: A light reflector device disposed at each pixel location on the plane and in-line with selected output waveguide segments to project light out of the plane. 41. A three-dimensional display device comprising: a solid material having first and second surface regions aligned with a common viewing axis, said solid material for transmitting light; Each of the first and second surface regions includes an array consisting of the following elements: at least one first conductivity type material forming a first electrode contacting the solid state material; at least one pixel waveguide segment transverse to the solid state material along a plane of the solid state material; a plurality of output waveguide segments transverse to the solid material along the plane, the output waveguide segments abutting the pixel waveguide segments in a plurality of junction regions; and a plurality of electrically controlled light redirectors, each light redirector being transverse to one of the interface regions, said light redirectors being capable of reflecting light from a pixel waveguide section to said output waveguide section; Each of the separated arrays is constrained to redirect light of a respective light polarization. 42. A display device according to claim 41, wherein said separate arrays are polarized at orthogonal angles transverse to the viewing axis. 43. The display device of claim 41, wherein the pixels of the separated array are directly aligned along the viewing axis. 44. The display device of claim 41, wherein the pixels of the spaced apart array are aligned off-center along the viewing axis. 45. A display device comprising: a solid material for transmitting light; at least one first conductivity type material forming a first electrode contacting the solid state material; An optical frequency source device comprising an optical exciter of a first frequency and a frequency doubling device coupled to said optical exciter and driven by the exciter driven by the optical exciter; at least one pixel waveguide segment transverse to the solid dielectric material along a plane of the solid material; a plurality of output waveguide segments transverse to the solid material along the plane, the output waveguide segments abutting the pixel waveguide segments in a plurality of junction regions; and A plurality of electronically controlled light redirectors, each light redirector is transverse to one of the interface areas, the light redirectors can reflect light from the pixel output waveguide section to the output waveguide section. 46. The display device according to claim 45, further comprising: fluorescent target devices positioned at selected pixel locations on the plane and in-line with the output waveguide segments to redirect light out of the plane when stimulated by light, with a light redirecting system for each Fluorescent target set-up. 47. The display device according to claim 45, further comprising: A light reflector device is positioned at each pixel location on the plane and in-line with selected output waveguide segments to project light in the vicinity of the phosphor target device out of the plane. 48. A display device according to claim 45, wherein said frequency doubling means is a periodic polarized structure in said solid state material. 49. A device for distributing light within a volume comprising: light path means for transmitting light; a plurality of light coupling regions for said light path arrangement arranged in a pattern within said volume; and Controllable energy field light coupling means coupled to said light routing means at said light coupling region for controllably distributing said light in said volume. 50. A method of distributing light in a volume comprising: transmitting light along a plurality of light paths within the volume; and The light is controllably distributed in the volume, the distribution being selected according to the use of energy fields in a plurality of coupling regions of the light path. 51. The method according to claim 51 , wherein said distributing step includes switching the energy field between a first state and a second state, wherein in said first state, substantially in a selected optocoupler region All light in a first optical path is directed in the second state, and substantially all light in the selected optical coupling region is directed in a second optical path. 52. The method of claim 51, further comprising the step of modulating said light onto said light path at an input. 53. The method of claim 51 , further comprising the step of modulating said light in a plurality of parallel regions, wherein said parallel regions are between an input to said energy path and a plurality of outputs from said light path between. 54. A beam steering element comprising: a plurality of sheets of solid material for transmitting light beams, said sheets of solid material forming at least a first device plane and a second device plane; In each of said first and second device planes: a first conductivity type material forming a first electrode; a pixel waveguide segment transverse to the solid state material; At least one output waveguide section transverse to the solid material, the at least one output waveguide section transverse to the pixel waveguide section at a first junction. An active light diverter, which is arranged near the first junction area, under the selective application of an electric field from the first electrode, the active light diverter reflects the light from the pixel waveguide section to the output waveguide segment; and A light reflecting device positioned at a selected location and in-line with the output waveguide segment to project light out of the plane. 55. The element according to claim 54, further comprising at least one layer of energy-permeable material disposed between said sheets of solid material, said layer having wave interaction properties at boundaries with said sheets of solid material, An energy wave is thus guided along the sheet of solid material. 56. The element of claim 54, further comprising a planar electrode opposite said first electrode to maintain an electric field across individual sheets of said equal sheets.

Images