JP2012123398A - Interference modulator and method of controlling array thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an interference modulator and a method of controlling an array of the interference modulator to realize a color system having excellent flexibility with modulation of incident light by manipulating optical properties of a micromechanical device.SOLUTION: The interference modulator includes a plurality of interference modulators. At least one of the interference modulators emits colored light, and at least one of the interference modulators emits white light.

Description

  This application is a continuation-in-part of US patent application Ser. No. 08 / 744,253, filed Nov. 5, 1996, which is filed on May 1, 1995. International application No. PCT / US95 / 05358, filed on May 5, 1994, which is a continuation-in-part of US patent application Ser. No. 08 / 238,750 filed on May 5, 1994. This US patent application is currently issued as US Pat. No. 5,835,255.

The present invention relates to branching interferometric modulation.
An interferometric modulator (IMod) modulates incident light by manipulating the optical properties of a micromechanical device. This is accomplished by using various techniques to change the spurious interference characteristics of the device. IMod is suitable for many applications ranging from flat panel displays and optical computing to fiber optic modulators and projection displays. Different IMod designs can be utilized to address different applications.

In general, in one aspect, the invention provides an IMod-based display having an anti-reflective coating and / or a microfabricated auxiliary illumination source.
In general, in one aspect, the invention provides an efficient drive scheme for an IMod matrix addressing array or other micromechanical device.
In general, in one aspect, the invention provides a color scheme that provides excellent flexibility.
In general, in one aspect, the present invention provides electronic hardware that is field reconfigurable to accommodate various display formats and / or application functions.
In general, in one aspect, the invention provides an IMod design that decouples the electromechanical behavior of the IMod from the optical behavior of the IMod.

In general, in one aspect, the present invention provides an IMod design with alternative actuation means, any one of which can be hidden from view.
In general, in one aspect, the invention provides an IMod or IMod array that is fabricated and used in conjunction with a MEMS switch or switch array and / or MEMS type logic.
In general, in one aspect, the invention provides an IMod that can be used for optical switching and modulation.

In general, in one aspect, the invention provides an IMod having 2-D and 3-D photonic structures.
In general, in one aspect, the invention provides various applications for light modulation.
In general, in one aspect, the invention provides a MEMS manufacturing and packaging system that utilizes a continuous web feed method.
In general, in one aspect, the invention provides an IMod that is used as an inspection structure for the evaluation of residual stress in a deposited film.

FIG. 6 is a cross-sectional view of a display substrate having an antireflection film and integrated auxiliary illumination. It is a figure which shows another system of auxiliary illumination. It is a figure which shows the detail of the process of a micromachining arc lamp source. It is a figure which shows the bias center type drive system of the array of IMod in a display. 1 is a schematic diagram showing a color display system using the concept of “base + pigment”. 1 is a block diagram of a system that provides a field reconfigurable display-centric product. FIG. It is a figure which shows the concept applied to a general purpose display center product. FIG. 6 is a schematic diagram illustrating an IMod geometry in a non-actuated state that decouples light behavior from electromechanical behavior. It is a figure which shows the same IMod in an operation state. FIG. 6 is a plot showing the performance of the IMod design in black and white states. FIG. 6 is a plot showing the performance of several color states. 1 is a schematic illustration of an IMod that similarly separates light behavior from electromechanical behavior, but with the support structure hidden. It is a figure which shows the same design in an operation state. FIG. 6 illustrates an IMod design that utilizes an anisotropically stressed membrane in one state. It is a figure which shows the same IMod in another state. 1 is a schematic diagram illustrating an IMod that utilizes rotational actuation. It is a figure which shows the process sequence of rotation IMod design. It is a block diagram of a MEMS switch. It is a block diagram of the row drive device using a MEMS switch. It is a block diagram of the column drive device using a MEMS switch. It is a block diagram of a NAND gate using a MEMS switch. 1 is a block diagram of a display system having logic and driver components utilizing MEMS. FIG. 1 is a schematic diagram showing the structure, processing method and operation of a MEMS switch. FIG. 5 illustrates two alternative switch designs. 1 is a schematic diagram illustrating an example of a 2-D photonic structure using a microring. 1 is a schematic diagram of a periodic 2-D photonic structure. 1 is a schematic diagram illustrating an example of a 3-D photonic structure. 1 is a schematic diagram showing an IMod having a microring structure in a non-operating state. It is a figure which shows the same IMod in an operation state. It is a figure which shows IMod which has a periodic 2-D photonic structure. FIG. 5 shows an IMod design that acts as an optical switch. FIG. 6 shows a variation of this design that acts as an optical attenuator. Figure 2 is a schematic diagram of an IMod design that functions as an optical switch or optical decoupler. It is a figure which shows how the combination of IMod can act as a NxN optical switch. It is a figure which shows the process sequence of a wavelength variable IMod structure. It is a figure which shows how a wavelength variable IMod structure can be integrated in a wavelength selective switch. FIG. 3 shows how a wavelength selective switch can have a solid state device. It is a figure which shows how a bump bonded component can be integrated. 2 is a schematic diagram of a two-channel equalizer / mixer. FIG. 6 illustrates how an equalizer / mixer can be implemented using an IMod-based component. 1 is a schematic diagram showing a processing process using a continuous web. It is a figure which shows how an IMod utilization test | inspection structure can be used as a stress measurement tool. It is a figure which shows one form of a discontinuous film. It is a figure which shows the film coated on the board | substrate. It is a figure which shows another form of a discontinuous film.

Anti-reflective coating One IMod design described above (inductive absorber design described in US patent application Ser. No. 08 / 554,630 filed Nov. 6, 1995) (Description of such US patent application specification) The characteristic of which is cited here as forming part of this specification is its dark state efficiency, in which case IMod can absorb as much as 99.7% of the incident light. . This is useful for reflective displays. In the design described above, IMod reflects certain colors of light in the inactive state and absorbs light in the active state.

Since the IMod array is provided on the substrate, the possibility of absorption is reduced by the intrinsic reflectivity of the substrate. For glass substrates, the amount of reflection is generally about 4% for the visible spectrum. Thus, despite the absorption performance of the IMod structure, the dark state is only dark enough to allow frontal reflection from the substrate.
One method for improving the overall performance of the IMod-based display is to provide an antireflection film (AR film). These anti-reflective coatings may have one or more layers of dielectric film deposited on the surface of the substrate and are designed to reduce reflection from the surface. A wide variety of forms are conceivable for such films, and design and fabrication are well known techniques. One simple film design is a coating of magnesium fluoride that is about a quarter wavelength thick. Another example utilizes a lead fluoride quarter-wave film deposited on glass, then a magnesium fluoride quarter-wave film deposited, and a third example. Is a film in which a zinc sulfide film is interposed between these two films.

  FIG. 1A illustrates one way to incorporate an AR film into an IMod display to improve the performance of the display system. In FIG. 1A, as described above, the AR film 100 composed of one or more thin films is deposited on the surface of the glass layer 102 bonded to the glass substrate 106, and an IMod array 108 is formed on the opposite side of the glass substrate. Is provided. The presence of the AR film 100 reduces the amount of incident light since most of the incident light 109 reflected from the surface is coupled into the glass layer 102. As a result, since much of the incident light is affected by the IMod array, a darker display state can be obtained when the IMod is operating in the absorption mode. The AR film 100 may be directly deposited on the surface of the glass substrate 106 on the side opposite to the IMod array.

Integrated Lighting FIG. 1A also shows how auxiliary lighting can be supplied to such a display. In this case, an array 104 of microscopic arc lamps is made in the glass layer 102. An arc lamp is an efficient means of supplying light. Historically, arc lamps have been made using techniques associated with making regular light bulbs. A typical variation of such a lamp is described in US Pat. No. 4,987,496. A glass container is made and a separately made electrode is contained in the container. After filling with a suitable gas, seal the container. Although such bulbs can be made small, this manufacturing method may not be suitable for making large monolithic arrays of such bulbs.

  Techniques used in the manufacture of micromechanical structures can be applied to the production of microscopic discharges or arc lamps. Because these “microlamps” are of microscopic size, the voltages and currents that drive them are considerably lower than the voltages and currents needed to supply arc lamps made using conventional means and sizes. In the example of FIG. 1A, the array is fabricated such that light 113 emitted by the lamp is directed to the IMod array 108 by a unique reflector layer 111 described below.

  FIG. 2 provides details on how one of such lamps optimized for flat panel displays can be made. The sequence is as follows. As can be seen in step 1, the glass layer 200 is etched to form the reflector bowl 201 using wet or dry chemical etching. The depth and shape of the bowl is determined by the required illumination area for each lamp. Shallow bowls cause a broad reflected beam spread, while parabolas tend to collimate the reflected light. The diameter of the bowl varies from ten to several hundred microns. This dimension is determined by the amount of display area that can be acceptablely hidden from the viewer's view. This is also a function of the density of the microlamp array. Reflector / metal halide layer 204 and sacrificial layer 202 are deposited and patterned using standard deposition techniques, such as sputtering and standard photolithography. The reflector / metal halide layer may be a film stack of aluminum (reflector) and a metal halide such as thallium iodide, potassium iodide and indium iodide. Metal halides (which are not essential) can improve the properties of the light produced. The sacrificial layer may be a layer such as silicon.

  Next, an electrode layer 206 is deposited and patterned, thereby forming two separate electrodes. This material may be a refractory metal such as tungsten, and the material is thick enough to provide mechanical support on the order of thousands of angstroms. Next, the sacrificial layer 202 is removed using a dry release method. The assembly (in the form of an array of such lamps) is sealed by bonding to a glass plate such as substrate 106 (shown in FIG. 1A) so that the reflector faces the plate. A gas, such as xenon, is used to backfill the cavity formed by the lamp during the sealing process to a pressure of about 1 atmosphere. This can be achieved by performing the sealing process in an airtight chamber that has been conventionally filled with xenon.

  As a result of the application of a sufficient voltage to the electrodes of each lamp, a discharge occurs in the gas between the ends of the electrodes, causing a light emission 205 in a direction away from the reflector 204. This voltage is as low as several tens of volts if the gap interval is on the order of several hundred microns or less. If the electrode material is deposited with minimal stress, the sacrificial layer 202 will determine the position of the electrode within the bowl. In this case, the thickness is selected to position the discharge at the focal point of the bowl. Should there be a residual stress that causes the electrode to move on release, the thickness is selected to compensate for this movement. Generally, the thickness is a fraction of the depth of the bowl, i.e. a few microns to a few tens of microns.

Referring again to FIG. 1A, light is shown traveling along path 113. Thus, light is emitted towards the IMod array where it is acted upon by the array and then reflected along path 110 and directed to interface 107 and viewer 111.
A lamp can be made without a reflector layer, thus allowing the lamp to emit light in all directions.

  Lamps made with or without reflectors can be used in a variety of applications that require microscopic light sources or light source arrays. These applications include projection displays, backlights for emissive flat panel displays, or conventional light sources for indoor (residential, building) or outdoor (automobile, flashlight) applications.

  Referring to FIG. 1B, another auxiliary lighting scheme is shown. The light guide 118 has a glass or plastic layer bonded to the substrate 112. Light sources 116, which may consist of any number of light emitting sources, such as fluorescent tubes, LED arrays or micro lamp arrays as described above, are mounted on opposite sides of the light guide. The light 122 is coupled into the light guide using a collimator 120 so that most of the light is captured in the guide by total internal reflection. Scatter pad 124 is an area of a light guide that has been roughened using wet or dry chemical means. The scatter pad is coated with a material or thin film stack 126 that includes a reflective surface facing the substrate 112 and an absorbing surface facing the viewer 128.

  When light trapped in the guide is incident on the scatter pad, the condition of total reflection is broken and a portion 129 of light is scattered in all directions. Normally, scattered light that escapes to the surrounding medium, for example, the viewer 128, is reflected into the substrate 112 because the reflective side of the film 126 exists. Similar to the microlamp described above, the scatter pads are made in an array, and each pad is sized so that the portion of the display that is not visible from direct view is hardly noticeable. Although these dimensions are small on the order of tens of microns, these dimensions can provide sufficient auxiliary illumination due to the inherent optical efficiency of the underlying IMod array 114. The shape of the scatter pad may be circular, rectangular, or any shape that can minimize these perceptions by the viewer.

In order to operate the array of addressing elements IMod in the array in an emphasized manner for display purposes, a series of voltages is commonly referred to as a “line-at-a-time” manner in the array columns (vertical Applied to rows) and rows (horizontal rows). The basic concept is that a sufficient voltage is applied to a particular row and the voltage applied to the selected column causes the corresponding element on the selected row to operate or release depending on the column voltage. Is to do so. The threshold and applied voltage must be such that only elements on the selected row are affected by the application of the column voltage. By sequentially selecting the set of rows that make up the display, the entire array can be addressed over a period of time.

  One simple way to accomplish this is shown in FIG. The hysteresis curve 300 is an ideal representation of the electro-optic response of the reflective IMod. The x axis indicates the applied voltage, and the y axis indicates the amplitude of the reflected light. IMod indicates hysteresis. This is because if the voltage is increased beyond the pull-in threshold, the IMod structure is activated and the absorption is increased. When the applied voltage is decreased, the applied voltage must be below the release threshold in order for the structure to return to the inoperative mode. The difference between the pull-in threshold and the release threshold causes a hysteresis window. The hysteresis effect is described in US patent application Ser. No. 08 / 744,253, filed Nov. 5, 1996, along with other addressing schemes, and the contents of such US patent application specifications are described herein. Is quoted here as forming part of By maintaining the bias voltage Vbias, the hysteresis window is always utilized to allow the IMod to remain in whatever state it is driven or released. The voltages Voff and Von correspond to the voltages required to operate or release the IMod structure. The array is driven by applying voltages to the columns and rows utilizing electronics known as column and row drivers. The IMod was made with a pull-in threshold of 6 volts and a release threshold of 3 volts. For such devices, typical values for Vbias, Voff, and Von are 4.5 volts, 0 volts, and 9 volts, respectively.

  In FIG. 3, a timing diagram 302 shows the types of waveforms that can be used to activate an array of IMods that exhibit a curve 300 that resembles a hysteresis curve. A total of 5 volts is required, ie 2 volts in columns and 3 volts in rows. Such a voltage is chosen such that Vcol1 is exactly twice the value of Vbias and Vcol0 is 0 volts. The row voltage is selected such that the difference between Vsel F0 and Vcol0 is equal to Von and the difference between VselF0 and Vcol1 is equal to Voff. Conversely, the difference between VselF1 and Vcol1 is equal to Von, and the difference between VselF1 and Vcol0 is equal to Voff.

  Addressing occurs with frames 0 and 1 alternating. In a typical addressing sequence, data for row 0 is loaded into the column driver during frame 0, so that the voltage level of either Vcol1 or Vcol0 depends on whether the data is binary 1 or 0 Will be applied. When the data settles, the row driver 0 applies a selection pulse with a value of VselF0. As a result, the IMod on the column where Vcol0 exists is activated, and the IMod on the column where Vcol1 exists is released. Data for the next row is loaded into the column and a selection pulse is applied sequentially to that row and so on until it reaches the end of the display. Next, addressing begins again at row 0, but at this point addressing occurs within frame 1.

  The difference between frames is that the correspondence between data and column voltage is switched, the binary 0 is now displayed as Vcol0, and the row select pulse is now at the level of VselF1. With this method, the overall polarity of the voltage applied to the display array alternates with each frame. This is particularly useful for MEMS-based displays. This is because it allows compensation for DC level charge accumulation that may occur when only a single polarity voltage is applied. Charge accumulation in the structure can significantly impair the electro-optic curve of an IMod or other MEM device.

Since the color display system IMod is a general purpose device with various potential light response systems, many different color display systems can be implemented with various characteristics. One potential approach takes advantage of the fact that there are binary IMod designs that can achieve color, black and white states in the same IMod. By using this function, a color system that can be called “base + pigment” can be achieved. This terminology is used because this method is similar to the method of creating a paint color by adding a pigment to a white base to achieve the desired color. Using this scheme, a particular paint can achieve any color and any saturation level in the spectrum by controlling the content and amount of pigment added to the base. The same is true for displays containing colored pixels and black and white pixels.

  As shown in FIG. 4A, the pixel 400 includes five subpixels 402, 404, 406, and 408, and each subpixel can reflect red, green, blue, and white, respectively. All subpixels can achieve a black state. Control of the brightness of each sub-pixel can be achieved using a technique related to pulse width modulation as described in US Pat. No. 5,835,255. In conjunction with a properly selected relative subpixel size, this results in a pixel that can exert a very large degree of control on brightness and saturation. For example, a very saturated color can be achieved by minimizing the overall brightness of the white sub-pixel. Conversely, bright black and white modes can be achieved by minimizing the brightness of the color subpixels or by maximizing them in conjunction with the white subpixels. It is clear that all these changes can be achieved.

User Control of Color Scheme The color scheme described above provides flexibility in display performance, along with the inherent characteristics of IMod-based displays in terms of resolution, gray scale and refresh rate. Given this range, it is useful to give the user of a product containing such a display control of its overall properties. As a variant, it would be advantageous if the display could automatically adapt to various visual requirements.

  For example, if a user wants to see only text for some reason, he may want to use the product in black and white mode. However, in other situations, the user may want to see a high-quality color still image and still want to watch live video in another mode. Each of these modes is potentially within the scope of a given IMod display configuration, but requires a trade-off relationship in certain attributes. Tradeoffs include a low refresh rate when high resolution images are required, or a high gray scale depth can be achieved when only black and white are required.

  To give the user this kind of on-demand flexibility, the controller hardware can be reconfigured to some extent. The trade-off is the result that the display is essentially limited by the response time of the pixels and thus has only a certain amount of bandwidth that defines the amount of information that can be displayed at a given time.

  One display architecture that provides such flexibility is shown in FIG. 4B. In this block diagram, the controller logic 412 is implemented using one of a variety of IC technologies, such as a programmable logic device (PLA) and a field programmable gate array (EPGA), These IC technologies allow the functionality of the component to be changed or reconfigured after the component leaves the factory. Such devices, traditionally used for dedicated applications such as digital signal processing or image compression, can provide the high performance required for such processing and can provide flexibility during the design phase of products incorporating such devices. .

  Component 412 provides signals and data to driver electronics 414, 416 for display 418 addressing. Conventional controllers utilize ICs or application specific integrated circuits (ASICs), which are effectively “programmed” by their design during manufacture. The term “program” in this case refers to an internal chip layout having many basic and high level logic elements (an assembly of logic gates and logic modules or gates). By using a field programmable device such as PLA or EPGA, different display configurations can be loaded from component 410 into a display controller component in the form of a hardware application or “hardapp”, which component 410 It may be a memory device or a microprocessor with a conventional memory device. The storage device may be EEPROMS or other reprogrammable storage device, and the microprocessor is in the form of a simple microcontroller whose function is to load hard-up from the storage device into the EPGA. It is good to be. However, no matter what processor is associated with the general functioning of the product, this is not performed by such processor. This scheme is advantageous because a wide variety of display performance forms and mixed display scan rates can be achieved with the possibility of combining them with relatively simple circuitry.

  For example, a part of the screen may be operated as a low-resolution text entry area, and a high-quality display of an email coming from another part may be performed. This can be achieved within the constraints on the overall bandwidth of the display by varying the refresh rate and the number of scans for the various segments of the display. The low resolution text region can be scanned only once or twice, corresponding to one or two bits of grayscale depth quickly. The high-display email area can be scanned quickly and with 3 or 4 passes corresponding to 3 or 4 bits of grayscale.

Configurable Electronic Product This concept does not include the functionality of the display controller, but can be generalized to include overall product functionality. FIG. 4C shows the configuration of an equivalent general purpose portable electronic product 418 at the programmable logic device or its core 420. In many display-centric personal electronic products, such as PDAs (personal digital assistants) and electronic organizers, the central processor is a variation of the RISC processor that uses a reduced instruction set. Although RISC processors are more efficient variants of the CPU that powers most personal computers, these processors are still general purpose processors that consume a large amount of energy to perform repetitive tasks, eg, retrieval instructions from a storage device. is there.

  In personal computers, power consumption is not an issue, and users generally want to run many complex software applications. The opposite is true for typical display-centric / personal electronic products. These require low power consumption and relatively few simple programs to provide. Such a scheme is advantageous for the execution of a dedicated purpose program, such as a web browser, calendar function, drawing program, telephone / address database and handwriting / speech recognition, among others. Thus, whenever a particular functionality mode, eg, a program, is required by the user, the core processor is reconfigured with the appropriate hard-up and the user interacts with the product. Thus, a typical example of a hard-up processor, i.e. a field programmable gate array, has a hard-up (i.e. program) revealed in its internal logic and connections, which are loaded with new hard-ups. Will always be rearranged and rewired. Many suppliers of these components also provide application development systems that allow a dedicated programming language (hardware description language) to be reduced to a logical representation of the appropriate processor. A great effort is also underway to simplify the process or reduce the high-level programming language to this form. One way to realize such a processor is the paper “Plastic Cell Architecture: Towards Reconfigurable Computing for General-Purpose” by Kouichi Nagami, et al, Proc. IEEE Workshop on FPGA-based Custom Computing Machines, 1998. Are described in detail.

  Referring again to FIG. 4C, the hard up processor 420 is located at the center of a group of I / O devices and peripherals that it uses, modifies or ignores based on the nature and function of the currently loaded hard up. It is shown in a state. Hardups can be loaded from an external source via a storage device 422 or RF or IR interface 424 present in the product, such an interface can be loaded with the contents of a specific hardup application from the Internet, cellular network or other electronic device It can be pulled out. Other examples of hard-up include a speech recognition or speech synthesis algorithm for the audio interface 432, a handwriting recognition algorithm for the pen input 426, and an image compression and processing mode for the image input device 430. Such products can perform more functions due to their main components, the display as the main user interface and the reconfigurable core processor. The total power consumption for such devices is on the order of tens of milliwatts, compared to hundreds of milliwatts consumed by existing products.

Decoupling electromechanical features from an optical point of view US patent application Ser. No. 08 / 769,947 filed on Dec. 19, 1996 and U.S. patent application Ser. No. 09/056 filed on Apr. 4, 1998. 975 describes an IMod design that proposes decoupling the electromechanical performance of IMod from its optical performance. The contents of these U.S. patent application specifications are cited herein as forming part of this specification. Another way in which this can be achieved is illustrated in FIGS. 5A and 5B. This design uses electrostatic forces that change the geometry of the branching interference cavity. An electrode 502 is fabricated on the substrate 500 and is electrically isolated from the membrane / mirror 506 by an insulating film 504. The electrode 502 serves only as an electrode, not as a mirror.

  An optical resonator 505 is formed between the membrane / mirror 506 and the auxiliary mirror 508. The support for the auxiliary mirror 508 is obtained by a transparent superstructure 510, which may be a thick organic deposit, such as SU-8, polyimide or an inorganic material. When no voltage is applied, the membrane / mirror 506 maintains a certain position relative to the auxiliary mirror 508 as shown in FIG. 5A, which is determined by the thickness of the sacrificial layer deposited during manufacture. For an operating voltage of about 4 volts, a thickness of several thousand angstroms is suitable. If the auxiliary mirror is made of a suitable material, such as chrome, and the mirror / membrane is made of a reflective material, such as aluminum, the structure will reflect a certain frequency of light 511 that can be perceived by the viewer 512. Become. In particular, the structure can have a wide variety of optical responses when the chrome is translucent and thin, about 40 angstroms, and the aluminum is sufficiently thick to be opaque, at least 500 angstroms. FIGS. 5C and 5D show examples of black and white responses and color responses, respectively, all of which depend on the length of the resonator and the thickness of the constituent layers.

  FIG. 5B shows the result of the voltage applied between the main electrode 502 and the membrane / mirror 506. The membrane / mirror is displaced in the vertical direction, thus changing the length of the optical resonator and thus the optical properties of the IMod. FIG. 5C shows one type of reflective optical response possible in two states, showing a black state 521 when the device is fully activated and a white state 523 when the device is not. FIG. 5D shows the optical response with color peaks 525, 527, 529 corresponding to blue, green and red colors, respectively. Thus, the electromechanical behavior of this device can be controlled independently of the optical performance. The material and form of the main electrode affecting the electromechanical behavior can be selected regardless of the material constituting the auxiliary mirror. This is because these materials do not play a role in the optical properties of IMod. This design example can be made using surface micromachining methods and techniques, such as those described in US patent application Ser. No. 08 / 688,710 filed Jul. 31, 1996. It should be noted that the description of such US patent application is cited here as forming part of this specification.

  In another example shown in FIG. 6A, the IMod 606 support structure is positioned to be hidden by the membrane / mirror 608. In this way, the amount of rest area is effectively reduced. This is because the viewer sees only the minimum space between the area covered by the membrane / mirror and the adjacent IMod. This is different from the structure of FIG. 5A where the membrane support is visible and constitutes rest and inaccurate areas from a color standpoint. FIG. 6B shows the same structure in the activated state.

  In FIG. 7A, another geometric shape used for the IMod structure is shown. This design is similar to that shown in US Pat. No. 5,638,084. This design utilizes an opaque plastic membrane that is isotropically stressed and is naturally positioned in a curled state. By applying a voltage, the membrane is flattened, whereby a MEMS-based optical shutter is obtained.

  The functionality of the device can be improved by making it branching coherent. A variation of IMod is shown in FIG. 7A, where the thin film sack 704 is an inductive absorber IMod described in US patent application Ser. No. 08 / 688,710 filed Jul. 31, 1996. Similar to the dielectric / conductor / insulator stack on which the design is based, the contents of such US patent application are hereby incorporated by reference as if forming part of this specification.

  By applying a voltage between the aluminum membrane 702 and the stack 704, the membrane 702 hits the stack and is positioned in a flat state. During fabrication, aluminum 702, which may further comprise other reflective metals (silver, copper, nickel) or a dielectric or organic material covered with a reflective metal, is deposited on a thin sacrificial layer, It can be released using wet etching or gas layer release. The aluminum membrane 702 is further mechanically secured to the substrate by support tabs 716 that are directly attached to the optical stack 704. For this reason, light incident on the area where the tab and stack overlap is absorbed, making this mechanically dormant area optically dormant. This method eliminates the need for a separate black mask for this and other IMod designs.

  Incident light 706 is completely absorbed, or light 708 of a specific frequency is reflected depending on the spacing of the stack layers. The optical behavior is similar to the optical behavior of the inductive absorber IMod described in US patent application Ser. No. 08 / 688,710 filed Jul. 31, 1996. Are hereby incorporated by reference as forming part of this specification.

  FIG. 7B shows the shape of the device when no voltage is applied. Due to the residual stress in the membrane, the membrane curls into a tightly wound coil. Residual stress can be applied by depositing a thin layer 718 of material on top of the membrane with very high residual tensile stress. Chromium is one example where high stress can be achieved with film thicknesses as small as a few hundred angstroms. With the membrane no longer hiding its path, the light beam 706 passes through the stack 704 and intersects the plate 710. The plate 710 may be located in a state of high absorption or high reflectivity (for a particular color or white). When the modulator is used in a reflective display, the optical stack 704 causes the device to reflect a particular color (if the plate 710 is absorptive) or absorb (if the plate 710 is reflective) when the device is activated. Designed to be (if sex).

Rotation Operation As shown in FIG. 8A, another geometric IMOD utilizes rotation operation. Using the method described in US patent application Ser. No. 08 / 688,710, filed Jul. 31, 1996, an electrode 802, an aluminum film about 1000 angstroms thick, was fabricated on substrate 800. ing. A support shutter 812 to which a support post 808, a rotary hinge 810, and a pair of reflective films 813 are attached is provided. The support shutter is preferably an aluminum film having a thickness of several thousand angstroms. The XY dimension is preferably on the order of several tens to several hundreds of microns. The film is bifurcated and is designed to reflect a specific color. A fixed bifurcation interference stack in the form of an inductive absorber similar to that described in US patent application Ser. No. 08 / 688,710 filed Jul. 31, 1996 is sufficient, and such US patent application. The contents of the specification are hereby incorporated by reference as forming part of this specification. These films may consist of a polymer without color pigments or may be aluminum or silver to provide broadband reflection. The electrode 802 and the shutter 812 are designed such that when a voltage (for example, 10 volts) is applied between the two, the shutter 812 rotates partially or once around the hinge axis. Only the shutter 818 is shown rotated. However, typically, all shutters for a given pixel are driven simultaneously by a signal on a common bus electrode 804. Such shutters take the form of electromechanical hysteresis when the distance between the hinge and the electrode is designed to overcome the spring tension of the hinge at some point during the rotation of the electrostatic attraction force of the electrode. Become. Thus, the shutter will have two electromechanically stable states.

  In the transmissive mode of operation, the shutter blocks incident light or allows incident light to pass. FIG. 8A shows a reflection mode in which incident light 822 is reflected back to the viewer 820. In this mode, in one state, the shutter reflects white light when metallized, and reflects a particular color or set of colors when covered with a branched interference film or pigment. Typical thicknesses and resulting colors for a branched interference stack are also described in US patent application Ser. No. 08 / 688,710 filed Jul. 31, 1996, and such US patent application. The contents of the specification are hereby incorporated by reference as forming part of this specification. In the other state, light will enter the substrate 800 and be absorbed therein if the opposite side of the shutter is covered with one or more absorbing films 722. These films may consist of another pigmented organic film or an inductive absorber stack designed to be absorbent. On the other hand, the shutter may have a high absorption, for example, black, and the opposite side of the substrate 800 may be coated with a highly reflective film 824, or a pigment or a color reflective film as described above. It is selectively coated with a branched interference film that reflects the color along the line.

  By adding an auxiliary electrode 814, the operation of the apparatus can be further enhanced, and such an auxiliary electrode can charge additional torque to the shutter when charged to a potential that induces electrostatic attraction between the auxiliary electrode 814 and the shutter 812. give. The auxiliary electrode 814 includes a combination of a conductor 814 and a support structure 816. The electrode may be made of a transparent conductor having a thickness of about 1000 angstroms, for example ITO. All of the structure and associated electrodes are deposited on the surface of a single substrate, i.e. deposited monolithically, and thus easily produced with good control of the electrode gap space and operate reliably. Machined from the material to be processed. For example, when such an electrode is attached to an opposing substrate, variations in the surface of both the device substrate and the opposing substrate may combine to produce a large deviation of several microns or more. Thus, the voltage that needs to influence the change in the specific behavior has a large value variation of several tens of volts or more. A monolithic structure accurately follows variations on the surface of the substrate and is less susceptible to such variations.

  FIG. 8B shows a rotation modulator manufacturing sequence in steps 1 to 7. In step 1, the substrate 830 is covered with an electrode 832 and an insulator 834. Typical electrode and insulating materials are aluminum and silicon dioxide, each 1000 Å thick. In step 2 these are patterned. A sacrificial spacer 836 (which is made of a material several microns thick, eg, silicon) is deposited and patterned in step 3 and covered with post / hinge / shutter material 834 in step 4. . This may be an aluminum alloy or a titanium / tungsten alloy having a thickness of about 1,000 angstroms. In step 5, material 838 was patterned to form bus electrodes 844, support posts 840 and shutters 842. In step 6, a shutter reflector 846 was deposited and patterned. In step 7, the sacrificial spacer was etched to produce a complete structure. Step 7 also shows a plan view of the structure, showing the details of the hinge consisting of support post 848, torsion arm 850 and shutter 852.

For IMod, a switching element binary device, the number of voltage levels required to address the display may be small. The driver electronics need not generate the analog signals necessary to achieve gray scale operation.
Thus, the electronics can be embodied using other means as suggested in US patent application Ser. No. 08 / 769,947 filed on Dec. 19, 1996. The contents of this document are hereby incorporated by reference as forming part of this specification. In particular, the driver electronics and the logic function can be implemented using a switching element using MEMS.

  9A-9E illustrate this concept. FIG. 9A is a schematic diagram of a basic switch unit or block, where input 900 is connected to output 909 by application of control signal 902. FIG. 9B shows how the row driver is implemented. The addressing type row driver described above requires outputs at three voltage levels. One of the input voltage levels can be selected for output 903 by applying an appropriate control signal to the row driver. The input voltages are Vcol 1, Vcol 0 and Vbais corresponding to reference numerals 906, 908 and 910 in the figure. Similarly, for the column driver shown in FIG. 9C, selection of one or the other input voltage level sent to output 920 is made as a result of providing the appropriate control signal. The input voltages are Vsel F1, Vsel F0 and ground potential corresponding to reference numerals 914, 916, and 918 in the figure. FIG. 9D shows how the logic device 932, in this case a NAND gate, can be implemented using basic switch building blocks 934, 936, 938, 940. FIG. All of these components can be configured and combined in a manner that allows for the creation of the display subsystem shown in FIG. 9E. The subsystem includes a controller unit 926, a row driver 924, a column driver 928, and a display array 930, and uses the addressing system described above with reference to FIG.

By fabricating the switch element as a MEMS device, the entire display system can be fabricated using a single method. The switch fabrication method is a sub-process of the IMod fabrication method and is illustrated in FIG. 10A.
Step 1 shows both a side view and a plan view of the initial stage. An arrow 1004 indicates the direction of the side view. The substrate 1000 has a sacrificial spacer 1002 deposited and patterned on its surface, ie a silicon layer with a thickness of 2,000 angstroms. In step 2, a structural material, ie, an aluminum alloy with a thickness of a few microns, is deposited and patterned to form a source beam 1010, a drain structure 1008, and a gate structure 1006. Hundreds of angstroms of non-corrosive metal, such as gold, iridium or platinum, may be plated onto the structural material at this point to maintain low contact resistance throughout the life of the switch. A notch 1012 is formed in the source beam 1010 by etching to facilitate the movement of the beam in a plane parallel to the plane of the substrate. The perspective view of the drawing is different in step 3 and step 4, which now compare the front view with the plan view. An arrow 1016 indicates the direction of the front view. In step 3, the sacrificial material is etched away, after which the source beam 1010 remains intact and can move freely.

  When a voltage is applied between the source beam and the gate structure, the source beam 1010 is deflected towards the gate 1006 and eventually comes into contact with the drain 1008, thereby bringing the source and drain into electrical contact. . The mode of operation is parallel to the surface of the substrate, thus allowing the use of fabrication methods that are compatible with the main IMod fabrication method. This method also requires fewer steps than those used to make a switch that operates in a direction perpendicular to the substrate surface.

FIGS. 10B and 10C show two alternative design examples for a flat MEM switch. The switch of FIG. 10B differs in that the switch beam 1028 helps to allow contact between the drain 1024 and the source 1026. In the switch of FIG. 10A, the current that must flow through the source beam to the drain switches the threshold and complicates the circuit design. This is not the case for switch 1020. The switch of FIG. 10C shows another modification. In this case, insulator 1040 electrically isolates switch beam 1042 from contact beam 1038. This insulator may be a material that can be deposited and patterned using conventional methods, such as SiO ₂ . By using such switches, it is not necessary to electrically isolate the switch drive voltage from the logic signals in the circuits that make up these switches.

Multidimensional photonic structures In general, IMods have useful optical properties and are characterized by comprising elements that can be moved by themselves relative to themselves or other electrical, mechanical or optical elements. .
Thin film assemblies for making bifurcated interference stacks are a subset of a larger class of structures called multidimensional photonic structures. In a broad sense, a photonic structure is defined as a structure that can change the propagation of electromagnetic waves due to the geometrical shape of the structure and the associated change in refractive index. Such a structure has dimensional characteristics because it interacts with light primarily along one or more axes. Multidimensional structures are also called photonic band gap structures (PBGs) or photonic crystals. The text “Photonic Crystals” by John D. Joanopoulos et al. Describes photonic structures with periodicity.

  One-dimensional PBG may occur in the form of a thin film stack. As an example, FIG. 16 shows the fabrication and final product of IMod in the form of a dielectric Fabry-Perot filter. An IMod structure having thin film stacks 1614, 1618, which are alternately arranged with silicon and silicon dioxide layers each having a quarter wavelength, fabricated on a substrate and having a central resonator 1616. Is formed. In general, the stack is continuous in the X direction and the Y direction, but is composed of alternating layers of high and low refractive indexes, so that it is optical in the Z direction due to variations in the refractive index of the material. It has periodicity in a general sense. This structure can be considered one-dimensional. This is because the effect of periodicity is greatest for waves propagating along one axis, in this case the Z axis.

  11A and 11B show two displays of a two-dimensional photonic structure. In FIG. 11A, microring resonator 1102 may be made from one of many well-known materials, for example, an alloy of tantalum pentoxide and silicon dioxide, using well-known methods. Typical dimensions for devices optimized for wavelengths in the 1.55 um range are w = 1.5 um, h = 1.0 um, r = 10 um.

  When the structure is fabricated on a substrate 1100 (glass is one possible material, many others), the refractive index and dimensions w, r, h are the frequencies of light propagating therein. And an essentially circular waveguide that determines the mode. Such a resonator, if correctly designed, can act as a frequency selective filter for broadband radiation coupled to it. In this case, the radiation propagates in the XY plane indicated by the symbol 1101 indicating the direction. The one-dimensional analog of this device is a Fabry-Perot filter made with a single layer mirror. Although no device exhibits higher order optical periodicity due to a single layer “boundary” (ie, mirror), these can be considered in a broad sense as photonic structures.

  A more conventional PBG is shown in FIG. 11B. The columnar array 1106 shows a periodic change in refractive index in both the X direction and the Y direction. The electromagnetic wave or electromagnetic radiation propagating in this medium is most significantly affected if propagating in the XY plane indicated by the direction symbol 1103.

  The array of FIG. 11B shares attributes with the one-dimensional thin film stack, due to its periodic nature, with the exception of its higher dimensionality. The array is periodic in the sense that the refractive index varies between the refractive index of the column material and the refractive index of the surrounding material, usually air, along several axes passing through the array in the XY plane. . Using the same principles applied to the design of thin film stacks, the appropriate design of this array is based on a wide variety of optical responses (mirrors, bandpass filters, edge filters, etc.) acting on radiation propagating in the XY plane. ).

  The array 1106 in the case shown in FIG. 11B includes singularities or defects 1108 in the form of columns having different dimensions and / or refractive indices. For example, the diameter of this row may be partly larger or smaller than the remaining rows (which is a quarter-wavelength in diameter) or of another material (perhaps air And silicon dioxide). The overall size of the array is determined by the size of the optical system or component that needs to be manipulated. Defects may also occur in the absence of one or more columns (rows) depending on the desired behavior. This structure is similar to the dielectric Fabry-Perot filter of FIG. 16, but only works in two dimensions. In this case, the defect is similar to resonator 1616. The remaining columns are similar to adjacent two-dimensional stacks.

  Appropriate dimensions for the structure of FIG. 11B are represented by column x, spacing sx, column y spacing sy (any of which may be considered a lattice constant), column diameter d, and array height h. . As with the quarter-wave stack, the one-dimensional equivalent, row diameter and spacing should be approximately a quarter wavelength. The height h is determined by the desired propagation mode, and few wavelengths greater than one-half wavelength are used for single mode propagation. The equations relating structure size to these effects on light are well known and are described in the article “Photonic Crystals” by Joannopoulos et al.

  This type of structure can also be made using the same materials and methods used to make the resonator 1102. For example, using conventional methods, a single film of silicon is deposited and patterned on a glass substrate and etched using reactive ion etching to produce high aspect ratio columns. For a wavelength of 1.55 um, the row diameter and spacing are on the order of 0.5 um and 0.1 um, respectively.

  The photonic structure also allows radiation to be directed under constraints on restrictive geometric shapes. Thus, these photonic structures are very useful in applications where it is desirable to redirect and / or select a particular frequency or frequency band of light where dimensional constraints are very severe. A waveguide capable of guiding light propagating in the XY plane and rotating the light 90 degrees in a space shorter than the wavelength of the light can be produced. This can be achieved, for example, by creating a column defect in the form of a linear row that can act as a waveguide.

  A three-dimensional structure is shown in FIG. A three-dimensional periodic structure 1202 acts on radiation propagating in the XY plane, YZ plane, and XZ plane. Various optical responses can be obtained by appropriate design of the structure and selection of its constituent materials. The same design rules apply, but these rules apply here in three dimensions. Defects occur in the form of lines or regions with respect to points and lines, and these defects differ in size and / or refractive index from the surrounding medium. In FIG. 12, the defect 1204 is one point element, but may be a line element or a combination of a line element and a point element or region. For example, a “straight” or “meandering” array of point defects may be created that follows an arbitrary three-dimensional path through the PBG and acts as a strictly conditional waveguide that propagates light. . Defects are generally located inside, but are shown on the surface for illustrative purposes. All suitable dimensions and shapes of this structure are shown. The PBG diameter and spacing and material are completely application dependent, but the above design criteria and equations also apply.

  A three-dimensional PBG is complicated to make. Conventional means of creating one-dimensional or two-dimensional features, when applied in three dimensions, must be subjected to multiple deposition, patterning, and etching cycles to achieve the third dimension in the structure. I must. As a fabrication technique for forming a periodic three-dimensional structure, a holography in which a photosensitive material is exposed to a standing wave and the photosensitive material replicates the standing wave in the form of a change in refractive index within the material itself. Self-assembled organic or self-assembled material, one-time control, utilizing the inherent adhesion and orientation properties of certain copolymer materials to create arrays of rows or spheres during material deposition Once a spherical structure of a given size and shape is solidified, it can be introduced into a liquid suspension that organizes the structure and can be dissolved or removed at high temperature, a ceramic system, a combination of these systems, and the like.

  Copolymer self-assembly methods are of particular interest. This is because such methods are low temperature, require minimal photolithography, and do not require photolithography. In general, in this method, a polymer, such as the polymer polyphenylquinoine (PPQmPSn), is dissolved in a solvent, such as carbon disulfide. After applying the solution on the substrate to allow the solvent to evaporate, a tightly packed hexagonal structure of air-filled polymer spheres results. If this method is repeated many times, multiple layers can be formed and the period of the array can be controlled by manipulating the number of repeating units of the polymer components (m and n). Introducing nanometer-sized colloids containing metals, oxides or semiconductors that can have the effect of further reducing the array period and increasing the refractive index of the polymer.

  Defects may be introduced on a submicron scale through direct manipulation of the material using tools such as a focused ion beam or atomic force microscope. With a focused ion beam, a very narrow selected area of material can be removed or added, or the optical properties of the material can be altered. Material removal occurs when a high energy particle beam, such as a particle beam used by a focused ion beam tool, sputters away material in its path. The addition of material occurs when the focused ion beam is passed through a gas, for example a volatile metal containing tungsten hexafluoride (for tungsten conductors) or silicon tetrafluoride (for insulating silicon dioxide). This gas decomposes and deposits components where the beam contacts the substrate. Using atomic force microscopy, you can move matter around on a molecular scale.

  Another approach uses a method that can be referred to as microelectrodeposition, which is described in detail in US Pat. No. 5,641,391. In this manner, a single microscopic electrode can be utilized to construct sub-micron resolution 3D features using a variety of materials and substrates. Subsequent oxidation of such deposited metal “defects” can form dielectric defects and a PBG array can be created around the dielectric defects using the methods described above.

  The presence of surface features in the form of patterns of other materials on the substrate on which the PBG was made may also serve as a template for creating defects in the PBG during the formation of the PBG. This is particularly suitable for PBG methods that are sensitive to substrate conditions, mainly self-assembly. These features can promote or prevent PBG “growth” in regions that are highly localized around the seed, depending on the particular nature of the method. In this manner, a pattern of defect “seed” can be created, after which the PBG can be formed and defects can be created therein during the PBG formation process.

  Thus, a class of devices called IMod can be further expanded by introducing a large family of multi-dimensional photonic structures into the modulator itself. Any kind of photonic structure that is inherently a static device can now be created dynamically by changing its geometry and / or changing its proximity to other structures . Similarly, a micromechanical Fabry-Perot filter (shown in FIG. 16) having two mirrors each having a one-dimensional photonic structure can be tuned by changing the resonator width by the action of static electricity. .

  FIG. 13 shows two examples of an IMod design that incorporates a two-dimensional PBG. In FIG. 13A, the cut-away view shows a free-standing membrane 1304 that is made with the microring resonator 1306 attached to the side facing the substrate. Waveguides 1301 and 1302 located within the body of substrate 1303 are flat and parallel to each other and can be fabricated using known methods. In FIG. 13A, IMod is shown in a non-driven state with an air gap (number) having a limit positioned between the microring and the substrate. The microring is made so that its position overlaps and aligns with a pair of waveguides in the underlying substrate. The dimension and shape of the microring are the same as in the above example. A cross section 1305 shows the dimensional shape of the waveguide, and it is preferable that w = 1 μm, h = 0.5 μm, and t = 100 nm of the waveguide. In the undriven state, light 1308 propagates unimpeded through waveguide 1302 and output beam 1310 has the same spectrum as input 1308.

  By driving the IMod to bring the microring in intimate contact with the substrate and waveguide, the optical behavior of the device is altered. The light propagating through the waveguide 1302 can now be coupled into the microring due to the phenomenon of evanescence. When correctly dimensioned, the microring acts as an optical resonator that couples selected frequencies from the waveguide 1302 and injects them into the waveguide 1301. This is illustrated in FIG. 13B, where the light beam 1312 is shown propagating in a direction opposite to the light direction 1308. Such a device can be used as a frequency selective switch that extracts specific wavelengths from the waveguide by applying the voltage or other driving means necessary to bring the structure in intimate contact with the underlying waveguide. . Static variations of this geometric shape are described in B.C. A paper “Vertically Coupled Microring Resonator Channel Dropping Filter” by B. E. Little et al., IEEE Photonics Technology Letters, vol. 11, no. 2, 1999.

  Another example is shown in FIG. 13C. In this case, a pair of waveguides 1332 and 1330 and resonators 1314 are fabricated on the substrate in the form of a columnar PBG. A PBG is a uniform array of columns, a waveguide constructed by removing two rows (one for each waveguide) and a resonator constructed by removing two columns. It has. Plan view 1333 gives details of the structures of waveguides 1330, 1332 and resonator 1314. The dimensions will depend on the wavelength of interest as well as the material used. When the wavelength is 1.55 um, the diameter and interval of the rows are 0.5 um and 1 um, respectively. The height h determines the propagation mode that is supported, and this propagation mode needs to be slightly larger than half of the wavelength if only a single mode is propagated.

  Formed on the inner surface of the membrane 1315 are two isolated rows 1311 directed downward, which are the same size and shape as the rows on the substrate and the same material (or Optically equivalent). The resonator and the column are designed to complement each other, and there is no corresponding column in the resonator where the column on the membrane is positioned.

  When the IMod is in an undriven state, there is a vertical gap 1312 with a limit of at least several hundred nm between the PBG and the membrane row, and thus no optical interaction occurs. The absence of a column in the resonator acts like a defect and causes coupling between waveguides 1330 and 1332. In this state, the device acts like the device shown in FIG. 13B, with the selected frequency of light propagating along the waveguide now being injected into the waveguide 1332 and the light 1329 of Propagate in the opposite direction in the form.

However, by driving the IMod and bringing it into contact with the PBG, the column is placed in the resonator, changing its behavior. Resonator defects are eliminated by placing membrane rows. In this state, the device acts like the device shown in FIG. 13A and light 1328 propagates without interference.
Static variations of this geometric shape are described in H.C. A. A paper by HA Haus, “Channel drop filters in photonic crystals”, Optics Express, vol. 3, no. 1, 1998.

In FIG. 14A, a device using an inductive absorber has a self-supporting aluminum membrane 1400 of about several tens to several hundreds of microns, and this aluminum membrane is composed of a combination of metal and oxide, It is provided in a floating state on a stack 1402 of material patterned on a transparent substrate. The film described in US patent application Ser. No. 08 / 688,710, used in inductive absorber modulators and filed on July 31, 1996, serves this purpose. It should be noted that the contents of such US patent application specifications are hereby cited as forming part of this specification. The film on the substrate can also constitute a transparent conductor, such as IPO. This structure may have a lossy metal film, such as molybdenum or tungsten, on the underside, with a thickness of a few hundred angstroms.

  These materials are configured so that, in the undriven state, the device reflects within a specific wavelength region but is very absorbent when brought into contact with the membrane. Side view 1410 shows a view of the device looking at the side of the substrate. The light beam 1408 propagates through the substrate at some arbitrary angle and is incident on the IMod 1406 shown in an undriven state. Assuming that the frequency of the light coincides with the reflection region of the non-driven IMod, the light is reflected at a complementary angle and propagates away. Side view 1414 shows the same IMod in the driven state. Since the device is now very absorbing, the light incident on it is no longer reflected and is absorbed by the material in the stack of IMod.

  Thus, in this form, IMod can act as an optical switch for light propagating within the substrate on which it is fabricated. The substrate is machined to mold a surface that is very polished, very parallel (up to 1/10 of the wavelength of light of interest) and many times the wavelength of light (at least a few hundred microns) Has been processed. This allows the substrate to act as a substrate / waveguide in that the light beam propagates in a direction that is parallel to the substrate, but reflects multiple times from surface to surface. Light waves in such a structure are often called substrate waveguides.

  FIG. 14B shows a modification of this method. The membrane 1420 is no longer rectangular but is patterned to taper towards one end. The mechanical spring constant of the structure remains constant along this length, but the electrode area is reduced. Thus, the amount of force that can be exerted by the action of static electricity is small at the narrow end of the taper. When increasing voltage is applied, the membrane begins to operate first at the wide end, and operation proceeds along arrow 1428 as the voltage increases.

For incident light, IMod acts as an absorption region, and the area of this absorption region is determined by the value of the applied voltage. Side view 1434 shows the effect on the substrate propagating beam when no voltage is applied. A reflection region 1429 corresponding to IMod viewed from the direction of the incident beam indicates the “footprint” 1431 of the beam superimposed on the reflection region. Since the entire region 1429 is non-absorbing, beam 1430 is reflected from IMod 1428 in the form of beam 1432 (with minimal loss).
In the side view 1436, an intermediate voltage value is applied to attenuate the reflected beam 1444 to some extent. This is because the reflective region indicated by 1437 is now partially absorbent. Figures 1438 and 1429 show the results of full operation and complete attenuation of the beam.

Thus, by using a tapered geometry, a variable optical attenuator can be formed, the response of which is directly related to the value of the applied voltage.
Another type of optical switch is shown in FIG. 15A. The support frame 1500 is made from a metal, such as aluminum, that is several thousand angstroms thick in such a way that it is electrically connected to the mirror 1502. A mirror 1502 is located on the transparent light standoff 1501. The mirror 1502 may consist of a single metal film or a combination of metals, oxides or semiconductor films.

The standoff is made from a material having a refractive index that is the same as or greater than the refractive index of the substrate. This may be SiO ₂ (same refractive index) or a polymer capable of changing the refractive index. The standoff is machined so that the mirror is supported at an angle of 45 °. Standoff machining can be accomplished using a technique known as analog lithography utilizing a photomask whose features are continuously variable in terms of their light density. By appropriately varying this density on specific features, a three-dimensional shape can be formed in the photoresist exposed using this mask. This shape can then be transferred to another material by reactive ion etching. The entire assembly is floated on a conductor 1503 that is patterned to provide an unobstructed “window” 1505 into the underlying substrate 1504. That is, the bulk of the conductor 1503 is etched away so that the window 1505 made of bare glass is exposed. Other IMod-like switches may be activated to bring the entire assembly into contact with the substrate / waveguide. Side view 1512 shows the optical behavior. Beam 1510 propagates through the substrate at an angle of 45 ° from the vertical that prevents it from propagating across the substrate boundary. This is because 45 ° is larger than the angle known as the critical angle, which minimizes the beam at the interface 1519 between the substrate and the external medium by the principle of total internal reflection (TIR). It can be reflected without loss or loss.

  The principle of TIR uses Snell's law, but the basic requirement is that the medium outside the substrate has a refractive index smaller than the refractive index of the substrate. In side view 1512, the device is shown with switch 1506 in an undriven state and beam 1510 propagating unobstructed. When the switch 1506 is activated to contact the substrate as shown in side view 1514, the beam path is changed. Since the standoff refractive index is greater than the refractive index of the substrate, the beam no longer produces TIR at the interface. The beam propagates from the substrate into the optical standoff where it is reflected by the mirror. The mirror is at an angle of 45 ° so that the reflected beam now travels at an angle perpendicular to the plane of the substrate. As a result, this light no longer meets the TIR criteria, so it can propagate through the substrate interface and is captured by the fiber coupler 1520, which is attached to the opposite side of the substrate / waveguide. It has been. The same concept is It is described in the paper “Waveguide Panel Display Using Electromechanical Spatial Modulators” by X. Zhouet al, SID Digest, vol. XXIX, 1998. This particular device was designed for light emitting display applications. The mirror can also be embodied in the form of a reflective grating, which can be etched into the standoff surface using conventional patterning methods. However, this method shows wavelength dependency and causes loss due to many diffraction orders that are not a problem in a thin film mirror. In addition, other optical structures can be used in place of mirrors with their respective characteristics and disadvantages. These structures can be categorized as refractive, reflective and diffractive, these structures include microlenses (both transmissive and reflective), concave or convex mirrors, diffractive optical elements, holographic optical elements, prisms and microfabrication methods. It may include any other type of optical element that can be made using. If another optical element is used, the standoff and the angle it gives to the optical component may not be necessary depending on the nature of the micro-optical component.

  This variation on IMod acts as an optical decoupling switch. If the broadband radiation mirror is correctly designed, broadband radiation or natural frequencies can be arbitrarily coupled from the substrate / waveguide. Side view 1526 shows a more elaborate embodiment in which an additional fixed mirror at an angle of 45 ° is fabricated on the side of the substrate opposite the side of the decoupling switch. This mirror differs from the switch in that it cannot be activated. By carefully selecting the angles of the mirrors on both structures, light 1522 that has been effectively decoupled from the substrate by switch 1506 can be coupled back into the substrate by recombination mirror 1528. However, by making recombination mirrors with various orientations in the XY plane, the combination of mirrors can be used to redirect light in any new direction within the substrate / waveguide. The combination of these two structures is called a directional switch. In addition, when a recombination mirror is used, light propagating in the substrate in a direction perpendicular to the surface can be coupled.

  FIG. 15B shows a specific example of an array of directional switches. When the substrate 1535 is viewed from above, the linear array 1536 is an array of fiber couplers that direct light into the substrate at an angle perpendicular to the XY plane. An array of recombination mirrors (not visible) is positioned directly opposite the fiber coupler array to couple light to the substrate parallel to beam 1530. On the surface of the substrate 1535, an array of directional switches is formed, one of which is indicated by 1531. The switch is positioned such that light coupled into the substrate from any one of the input fiber couplers 1536 can be directed to any one of the output fiber couplers 1532. In this way, the device can act as an N × N optical switch that can switch any one of any number of different inputs to any one of any number of different outputs.

Wavelength Tunable Filter Referring to FIG. 16, an IMod in the form of a wavelength tunable Fabry-Perot filter is shown. In this case, conductive contact pads 1602 are deposited and patterned with dielectric mirrors 1604 and 1608 and sacrificial layer 1606. This may consist of a silicon film whose thickness is a multiple of one-half wavelength. The mirror may consist of a stack of materials with alternating low and high refractive indices, two examples of materials being TiO ₂ (high refractive index) and SiO ₂ (low refractive index), One of the layers may be air. An insulating layer 1610 is deposited and patterned so that only the second contact tab 1612 contacts the mirror 1608. Next, a mirror “island” 1614 is formed by patterning the mirror 1608 and subsequently connected by a support 1615. The lateral dimension of the island is mainly determined by the size of the light beam with which it interacts. This is usually on the order of tens to hundreds of microns. The sacrificial layer 1606 is partially chemically etched, but is left with a standoff that is large enough to provide mechanical stability, perhaps several tens of microns square. When the top layer of mirror 1608 and the bottom layer of mirror 1604 are lightly (low impurity) doped to be conductive, a voltage is applied between contact pads 1602 and 1612 to displace the mirror island. Thus, the optical response of the structure can be tuned.

FIG. 17A shows an application example of this tunable filter. On the top surface of the substrate 1714, a wavelength tunable filter 1704, a mirror 1716, and an antireflection film 1712 are formed. The mirror 1717 is also made on the bottom surface of the substrate, for example from a metal, for example gold with a thickness of at least 100 nm. An optical superstructure 1706 is provided on the top surface of the substrate, the inner surface of which is at least 95% reflective, for example by adding a reflective film made of gold, which is also angled. The mirror 1710 is supported. In this device, the light beam 1702 propagates through the substrate at an angle greater than the critical angle, which is about 41 ° for glass substrates and air media. Therefore, the mirror 1716 needs to keep it confined within the substrate / waveguide interface. This configuration can provide great flexibility in selecting the angle at which light propagates.
A beam 1702 is incident on the Fabry-Perot 1702, which transmits light 1708 of a specific frequency while reflecting the remaining 1709. The transmitted frequency is incident on and reflected by the reflective superstructure 1706 and is reflected again by the mirror 1716 and is directed to the angled mirror 1710. The mirror 1710 is tilted so that light is directed to the anti-reflective coating 1712 at an angle perpendicular to the substrate and travels through the external medium into it. Thus, the device as a whole acts as a wavelength selective device.

  Many methods can be used to fabricate the superstructure. One method includes bulk microfabrication of a silicon slab to form a cavity or resonator of precise depth, eg, about the thickness of the substrate, ie, at least a few hundred microns. Angled mirrors are made after cavity resonator etching and the entire assembly is bonded to a substrate, such as glass, using any one of a number of silicon-glass bonding techniques.

  FIG. 17B shows an elaborate variation. In this example, a second tunable filter 1739 is added to form an additional frequency selection channel. That is, two separate frequencies can now be selected separately and independently. A detector 1738 has also been added in anticipation of a high degree of overall functionality.

  FIG. 17C includes an integrated circuit. Light beam 1750 is coupled into substrate 1770 and is incident on tunable filter 1752. This filter differs from the filters of FIGS. 17A and 17B in that it has a recombination mirror 1756 made on the surface of the movable mirror of the filter. The mirror angle is such that the frequency selected by filter 1752 is now coupled back directly into the substrate at a right angle in the form of light beam 1758. The remaining frequencies contained in the light beam 1750 propagate until they hit the recombination mirror 1760, which is tilted to have a surface perpendicular to the propagating beam 1756. Thus, the beam travels back in its path and exits the device, where it can be used by other devices that are optically coupled. Light beam 1758 is incident on IC 1764, which can detect and decode information in the beam. The IC may be in the form of an integrated circuit utilizing an FPGA or other silicon, silicon / germanium or gallium arsenide device that is advantageous when coupled directly to the information-transmitting light. For example, a high bandwidth interconnect may be formed between the ICs 1764 and 1762 by a bidirectional optical path (optical path) 1772. This is formed by the combination of mirror 1766 and recombination mirror 1768. If the IC has an element such as a vertical cavity surface emitting laser (VCSELS), light can be emitted by the IC or by a light emitting diode (LED). Light can be detected by any number of photosensitive elements, and the properties of such elements depend on the semiconductor technology used to make the IC. Light incident on the IC can also be modulated by an IMod created on the surface of the IC that is exposed to substrate propagating light.

Optical mixer (optical mixer) using substrate waveguide
18A and 18B are schematic views of a two-channel optical mixer embodied using a substrate / waveguide TIR variation. FIG. 18A shows a schematic of this device. Light that includes multiple wavelengths has two specific wavelengths 1801, 1803 that are split and directed to two separate and independent variable attenuators 1805. These wavelengths are then output to several possible channels 1807 or output into the light stop 1813.
FIG. 18B shows a specific example. The input light is directed into the device through the fiber coupler 1800 and the antireflection film 1802, and is coupled into the substrate using the recombination mirror 1806. The recombination mirror directs this light to the tunable filter 1808, splits the frequency λ1 (beam 1815), directs all unselected frequencies to the second tunable filter 1809, and this second tunable filter. Splits the frequency λ2 (beam 1817) so that the remaining frequency beam 1819 propagates further downstream via the TIR. Following the path of the beam 1815 transmitted by the tunable filter 1808, the direction of light is returned through the mirror 1810, through the AR film, back into the substrate waveguide, and recombined back into the substrate. The recombination mirror 1811 directs the beam 1815 to the attenuator 1812, where the beam continues to propagate along a path parallel to the beam 1817 selected by the second tunable filter 1809. These two beams are shifted in position by a beam repositioner 1816.

  This structure yields the same result as a recombination mirror, except that the mirror is parallel to the surface of the substrate. Since the mirror is suspended at a certain distance on the substrate surface, the position of the incident point on the opposite substrate interface is shifted to the right. This deviation is directly determined by the height of the repositioner. A beam 1819 containing a non-selected wavelength is also shifted by the repositioner 1818. As a result, all three beams are equally separated when they enter the array of decoupling switches 1820, 1824. They work selectively to direct the beam orientation into one of two optical combiners (1828 is one of these combiners) or detector / absorber 1830. The optical combiner can be manufactured by various methods. A polymer film patterned using reactive ion etching into the form of pillars with the top formed in the lens is one way. An absorber / detector having a semiconductor device bonded to the substrate serves to allow measurement of the output of the mixer. The optical superstructure 1829 supports external optical elements and provides a hermetic package for the mixer.

  The combination of planar IMod and substrate waveguides provides a family of optical devices that are easily fabricated, configured, and coupled to the outside world. This is because the device is located on the waveguide and / or on the superstructure and can act on light propagating in the waveguide and between the waveguide and the superstructure. Since all of the elements are made in a planar form, it is possible to achieve economies of scale by bulk manufacturing over a wide area, and different parts can be easily and accurately aligned and joined. In addition, since all active components are active in a direction perpendicular to the substrate, these components are relatively simple to make and drive compared to more sophisticated non-planar mirrors and beams. Active electronic devices may be joined to the superstructure or substrate / waveguide to improve functionality. Alternatively, the active device can be made as part of the superstructure, especially if it is a semiconductor, such as silicon or gallium arsenide.

Printing Style Fabrication Since these devices are planar and many of the layers do not require electrical semiconductor properties that require a dedicated substrate, IMod and many other MEM structures are manufacturing methods in the printing industry. A similar manufacturing method can be used. Such methods generally use a “substrate” that is flexible, for example in the form of a continuous sheet of paper or plastic. Such a method, called the web feed method, typically feeds a continuous roll of substrate material into a series of tools, each of which selectively coats the substrate with ink to sequentially produce a full color graphical image. Such a method is of interest because the product can be manufactured at high speed.

  FIG. 19 is a schematic diagram of a sequence that can be adapted to create a single IMod, and thus an array of IMod or other microelectromechanical structures. Web source 1900 is a substrate material, such as a roll of transparent plastic. The representative region 1202 on the portion of material from the roll has only a single device for purposes of this description. An embossing tool 1904 imprints the dent pattern on the plastic sheet. This can be achieved by a metal master in which protrusions of appropriate patterns are provided by etching.

  The metal master is attached to a drum that is pressed against the sheet with sufficient pressure to deform the plastic to form a recess. FIG. 1906 illustrates this. A coater 1908 deposits a thin layer of material by well-known thin film deposition methods, such as sputtering or evaporation. The result is a four film stack 1910 comprised of oxide, metal, and oxide sacrificial film. These materials correspond to the induction absorber IMod design. Tool 1912 dispenses, cures and exposes the photoresist to pattern these layers. Once the pattern is defined, film etching is performed with tool 1914. Alternatively, patterning may be accomplished using a method known as laser ablation. In this case, the laser is scanned against the material in such a way that the laser can be synchronized with the moving substrate. The frequency and power of the laser is such that the laser evaporates the material of interest and has a size in the micron range. The frequency of the laser is tuned so that it only interacts with the material on the substrate and not the substrate itself. Since evaporation occurs so quickly, heating of the substrate is minimal.

  In this device example, the films are all etched using the same pattern. This can be seen at 1918 where the photoresist is stripped after use of the tool 1916. Tool 1920 is another deposition tool for depositing what will be the structural layer of IMod. Aluminum is one candidate material for this layer 1922. This material should further include organic materials that exhibit minimal residual stress and can be deposited using various PVD and PECVD methods. The layer is then patterned using tools 1924, 1926, 1928, respectively, etched, and the photoresist is stripped therefrom. The sacrificial layer is etched away using tool 1930. If the layer is silicon, this can be accomplished using XeF2, a gas phase etchant used for such purposes. As a result, the free-standing membrane structure 1932 forms the IMod.

  Packaging of the resulting device is accomplished by bonding the flexible sheet 1933 to the top surface of the substrate sheet. This is also procured by a continuous roll 1936 coated with an airtight film, such as metal, using a coating tool 1934. The two sheets are joined using a joining tool 1937, resulting in a packaged device 1940.

  Residual stress is a factor in the design and fabrication of MEM structures. In IMod and other structures where structural members are mechanically released during fabrication, the residual stress defines the resulting geometric shape of the member.

  IMod as a branching interference device is vulnerable to variations in the resulting geometric shape of the movable membrane. The reflected or transmitted color in the case of other designs is a direct function of the cavity spacing of the cavity resonator. As a result, this distance variation along the length of the cavity may result in an unacceptable variation in color. On the other hand, this property is a useful tool in determining the residual stress of the structure itself. This is because the variation and degree of deformation of the membrane can be obtained by using the color variation. By knowing the deformation state of any material, the residual stress in the material can be determined. Computer modeling programs and algorithms can determine this using two-dimensional data regarding the deformation state. Thus, the IMod structure can be a tool for this evaluation.

  20A and 20B show examples of how IMod can be used in this way. IMod 2000, 2002 is shown from the side and bottom perspective (ie, looking through the substrate). These IMods are of the double cantilever form and the single cantilever form, respectively. In this case, the structural material has no residual stress and both membranes show no deformation. When viewed through the substrate, the device exhibits a uniform color defined by the thickness of the spacer layer in which they are formed. IMod 2004, 2006 is shown with a stress gradient that has a greater compressive force when it is provided at the top than at the bottom. The structural membrane shows the resulting deformation and the bottom view shows the nature of the resulting color change. For example, if the color area 2016 is green, the color area 2014 is blue because it is close to the substrate. Conversely, the color region 2018 (shown in a double cantilever configuration) is red because it is further away from the substrate. IMod 2008, 2010 is shown in a state where the stress gradient shows higher tensile stress at the top than at the bottom. The structural membrane deforms appropriately and the color area changes as a result. In this case, the region 2020 is red and the region 2022 is blue.

  FIG. 20B shows a system that can be used to quickly and accurately assess the residual stress state of a deposited film. Wafer 2030 has an array of IMod structures consisting of both single and double cantilever membranes of various lengths and widths. The structural membrane is made from a material with well-characterized mechanical and residual stress properties. Many materials are available, subject to very low required reflectivity limitations if IMod is not used for display purposes. Good candidate materials can be made or can be made to be compatible from a manufacturing point of view, exhibit some degree of reflectivity, and mechanical properties can be determined with high precision or of a determined crystalline form. Examples include materials (silicon, aluminum, germanium). These “test structures” are made and released to be self-supporting. If the material is stress free, the structure will not show color variations. However, although this is not the case, the color state or color map can be recorded using the high-resolution imaging device 2034, and such a high-resolution imaging device can obtain a high magnification image via the optical system 2032. Can do.

  The imaging device is connected to a computer system 2036, and hardware capable of recording and processing image data is provided on the computer system. The hardware should have a readily available high speed processing board that performs numerical calculations at high speed. The software may comprise a collection routine that collects color information and calculates surface deformations. The core routine can use the deformation data to determine an optimal combination of uniform stress and stress gradient across the thickness of the membrane, which can result in an overall shape.

  One mode of use is to have a detailed record of the stress state in the non-deposited state, resulting in a batch of “unused” test wafers that are retracted for later use. When a demand arises for the residual stress of the deposited film, a test wafer is selected and the film is deposited on top. The deposited film changes the geometric shape of the structure and consequently changes these color maps. Using software present on the computer system, the color map of the test wafer can be compared both before and after to accurately evaluate the residual stress in the deposited film. The test structure should also be designed to be activated after deposition. Observing these behaviors in operation with a newly deposited film can provide information on residual stress conditions and information on changes in film properties over many operating cycles.

  This method can also be used to determine the stress of the film as it is being applied. With a suitable design modification of the deposition system, an optical path can be created that allows the imaging system to see the structure and track these colormap changes in real time. This facilitates the construction of a real-time feedback system that controls deposition parameters in an attempt to control residual stress in this way. Software and hardware “questions” the test wafer to allow the deposition tool operator to change conditions as the film grows. The overall system is superior to other methods of measuring residual stress, which use only electromechanical actuation or expensive and complex branching interference that measures deformation of the fabricated structure. You are using the system. Methods that utilize only electromechanical actuation have the disadvantages of providing drive electronics to a large array of devices and inaccuracies in measuring displacement electronically. The latter method using expensive and complex branching interference systems has the disadvantages of being the optical properties of the film under observation and the complexity of the required external optical components and hardware.

Discontinuous film Another class of materials with interesting properties are films that are not homogeneous in structure. This can occur in several forms, which are collectively referred to as a discontinuous film. FIG. 21A shows one form of discontinuous film. The substrate 2100 may be a metal, a dielectric, or a semiconductor, and contours 2104, 2106, and 2108 are formed on the surface thereof by etching. The contour, including the individual structural profile having a height 2110 that is a fraction of the wavelength of light of interest, is etched using photolithography and chemical etching techniques to indicate 2104 (triangle) 2106 ( A profile similar to that shown by the (cylindrical) 2108 (klopfenstein taper) has been achieved. The effective diameter of the base 2102 of any individual profile is also approximately the height of the pattern. Each contour is slightly different, but all of these contours share the property that the effective refractive index gradually transitions from the refractive index of the incident medium to the refractive index of the film substrate 2100 itself as it travels from incidence into the substrate. Yes. This type of structure serves as an excellent antireflective coating compared to an antireflective coating made from a combination of thin films. This is because these structures do not have much of the angular dependence drawback. Thus, these structures remain highly non-reflective from a wide range of incident angles.

  FIG. 21B shows a coating 2120 that is deposited on a substrate 2122 and may also be a metal, dielectric or semiconductor. In this case, the film is still in the initial molding stage and has a thickness of 1000 Angstroms or less. During most deposition processes, the film undergoes a gradual nucleation process, where the films begin to bond together and at some point form progressively larger material sites until a continuous film is formed. Reference numeral 2124 shows a plan view of the film. The optical properties of the film at the initial stage are different from those of the continuous film. In the case of metals, films tend to exhibit greater losses than their continuous equivalents.

  FIG. 21C shows a third form of discontinuous film. In this case, film 2130 has been deposited on substrate 2132 to a thickness of at least 1000 angstroms, and this film is considered continuous. A pattern of “sub-wavelength” (ie, smaller diameter than the wavelength of interest) holes 2134 is formed in the material using techniques similar to the self-assembly method described above. In this case, the polymer can serve as a mask that transfers the etching pattern into the underlying material, and the holes are etched using a reactive ion etching method. Since the material is continuous but perforated, it does not act like the early stage film of FIG. 21B. Instead, its optical properties differ from non-etched films in that incident radiation results in low losses and can exhibit transmission peaks based on surface plasmons. In addition, manipulating the hole geometry and the incident angle and refractive index of the incident medium can control the spectral characteristics of the transmitted light. Reference numeral 2136 shows a plan view of this film. Such films are described in the article “Control of optical transmission through metals perforated with subwavelength hole arrays” by Tae Jin Kim. These are regular in structure but different from PBG.

All three of these types of discontinuous films are candidate materials that can be incorporated into an IMod structure. That is, these discontinuous films can serve as one or more of the material films in the static and / or moving parts of the IMod structure. All three show unique optical properties that can be manipulated primarily in a manner that utilizes the structure and geometry of the individual films, rather than a combination of graded thickness films. These films can be used in conjunction with other electronic, optical and mechanical elements of the IMod that they constitute. In the very simple case, the optical properties of each of these films can be altered by bringing them in direct contact with or in close contact with other films via surface conduction or optical interference. This can occur by directly changing the conductivity of the film and / or changing the effective refractive index of its surrounding medium. Thus, a more complex optical response of an individual IMod can be obtained with a simple structure that is less complex.
Other embodiments are within the scope of the present invention as set forth in the claims.

Claims (17)

A plurality of interferometric modulators, the plurality of interferometric modulators comprising:
At least one interferometric modulator adapted to emit colored light;
An interferometric modulator comprising: at least one interferometric modulator configured to emit white light. The plurality of interferometric modulators,
At least one interferometric modulator adapted to emit red light;
At least one interferometric modulator adapted to emit green light;
The interferometric modulator according to claim 1, further comprising at least one interferometric modulator configured to emit blue light.   The number of the plurality of interferometric modulators is twice the number of interferometric modulators that emit the white light, and the number of interferometric modulators that emit the white light is the red light. The interferometric modulator according to claim 2, wherein the number of interferometric modulator diagrams adapted to emit is equal to the number of interferometric modulator diagrams.   Between at least one interferometric modulator configured to emit at least one color light of red light, green light, and blue light, between the first modulator and the second modulator configured to emit the white light. The interferometric modulation apparatus according to claim 2, wherein   2. The interference according to claim 1, wherein the plurality of interferometric modulators are electrically connected to a controller adapted to control light intensity and saturation of light emitted from the interferometric modulators. Modulation device.   2. The stop of at least one interferometric modulator adapted to emit white light is twice the diaphragm of at least one interferometric modulator adapted to emit colored light. The interferometric modulator according to 1.   At least one interferometric modulator adapted to emit white light has at least one interferometric modulator adapted to emit a plurality of white lights, and emits the plurality of white lights. 2. The interferometric modulator according to claim 1, wherein the at least one interferometric modulator has a collective stop that is twice the stop of the at least one interferometric modulator adapted to emit the colored light.   The plurality of interferometric modulators have a movable thin film and a reflective surface to form an optical cavity therebetween, and the thin film moves relative to the reflective surface to change the height of the optical cavity. The interferometric modulation apparatus according to claim 1, wherein   The interferometric modulator according to claim 1, wherein the interferometric modulator further includes a plurality of MEMS switches, and at least a plurality of the MEMS switches are connected to the interferometric modulator. At least some of the MEMS switches are
The first terminal and the second terminal;
Electrodes,
The refracting arm is responsive to a voltage across at least a part of the refracting arm and the electrode, and the first terminal and the second terminal can be selectively connected. Item 10. The interferometric modulation device according to Item 9.   The interferometric modulator according to claim 10, wherein the refractive arm has a first portion and a second portion separated by an insulator.   The interferometric modulator according to claim 9, wherein the plurality of interferometric modulators and the plurality of MEMS switches are integrally formed on a single substrate. A method of controlling an array of interferometric modulators, the method comprising:
Bringing at least one interferometric modulator in the array into a first state by a first electrical signal and forming a charge in the interferometric modulator;
Placing at least one interferometric modulator in the array in a second state by a second electrical signal and removing at least a portion of the formed charge from the interferometric modulator.   The array of interferometric modulators comprises at least one interferometric modulator adapted to emit colored light and at least one interferometric modulator adapted to emit white light. 13. The method according to 13.   The method of claim 13, wherein the first electrical signal and the second electrical signal have voltages having controlled waveforms.   The at least one interferometric modulator is actuated by a hysteresis curve that is at least partially changeable in response to an applied voltage, the hysteresis curve being varied by the first electrical signal or the second electrical signal. The method of claim 13.   The method of claim 13, wherein the polarity of the first electrical signal is opposite to the polarity of the second electrical signal.

Images