JPH0756532B2 - Method of manufacturing spatial light modulator - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to spatial light modulators (light valves), and more particularly to spatial light modulators having electronically addressable deflectable beams.

[0002]

BACKGROUND OF THE INVENTION A spatial light modulator (SLM) is a converter that modulates incident light with an electrical input or a spatial pattern corresponding to the optical input. It is possible to modulate the phase, intensity, polarization or direction of the incident light, this modulation of the light being made by different materials with different electro-optical or magneto-optical effects, as well as by materials which modulate the light by surface deformations. You can do it. SLMs have numerous applications in the fields of optical information processing, projection display devices and electrostatic printing. 30 I
L.E. in EEE Transactions on Electronic Devices, page 539 (1983). See the reference cited in Hornbeck's article "128 x 128 Deformable Mirror Device".

A well-known SLM used in large bright electronic displays is the Eidophor.
This is a device that uses an oil film that can be recessed by electrostatic action as an active optical element. 20J. SMPTE magazine 3
E. 51, p. 51 (1953). Baumann's paper "The
See Fisher Large Screen Projection System (Eid Fore). In this device, a continuous oil film is raster-scanned with a modulated electron beam to create a spatially periodic distribution of deposited charge in each resolvable pixel area on the oil film. Due to this charge distribution,
The electrostatic attraction between the surface of the oil film and the supporting substrate held at a constant potential creates a phase diffraction grating in each pixel. Due to this attractive force, the film surface is deformed by an amount proportional to the amount of accumulated charges. The deformed oil film is illuminated with spatially coherent light from a xenon arc lamp. The light incident on the modulated pixel on the oil film is diffracted by the local phase grating into a discrete set of regularly spaced orders which, by a portion of the optics, alternate between transparent and opaque bars. It is incident on a schlieren stopper composed of a periodic array of. The spacing of the Schlieren stopper bars is chosen to match the spacing of the diffracted signal orders in the stopper plane so that the efficiency of light passage is high. Light incident on the unmodulated area of the light valve is blocked by the opaque bar of the Schlieren stopper from reaching the projection lens. Therefore, the image formed on the projection screen by the Schlieren imager from the unmodulated area of the light valve is dark, but the phase perturbation introduced by the modulated electron beam causes the screen to be projected by the Schlieren projector. It is converted into a bright light spot at. Despite the numerous technical difficulties associated with the polymerization of oil by electron irradiation and the contamination of the cathode with organic vapors, this type of oil film device can be used when the screen requires thousands of lumens of total light. , Has been successfully developed to the point that it becomes a device that is used almost all over the world. However, such devices are expensive, bulky and have short component life.

A number of non-oil film SLMs have also been developed,
There are a type using a deflectable element, a polarization plane rotating type, and a light scattering type. Each of these types of SLM is a modification of a reflective layer made of metal, elastic body or elastic body-photoconductor, ferroelectric, P
It uses various effects such as polarization and scattering of LZT ceramics and liquid crystals. For example, 299 Proc. SPI
R.E., page 68 (1981). Sprague et al., "Linear Total Internal Reflection Spatial Light Modulator For Laser Printing," and 299 Proc. R.S., page 76, SPIE (1982). Sprague et al., "Integrated Total Internal Reflection (TIR) Spatial Light Modulator for Laser Printing," and U.S. Pat. An apparatus is described for non-impact printing on a light sensitive surface such that it is imaged on a light sensitive medium after passing through a linear array of vessels. This array is configured as a total internal reflection spatial light modulator with electrodes and drive circuitry made on an integrated drive element that is placed against a total reflection surface of an electro-optic crystal such as lithium niobate. The local change in the refractive index generated by the fringe electric field between the two electrodes is read out using the Schlieren reading optical system. This optics images the TIR interface onto the photosensitive medium. This is a one-dimensional image, where a photosensitive medium is rotated on a drum under a linear array of images to produce a two-dimensional image for printing (eg page 1 text).
To occur. However, SLMs (light valves) are very susceptible to manufacturing problems due to their mixed molding nature. The strength of the fringe field, and thus the amount of light diffracted by the modulated pixel, is dependent on variations in the thickness of the air gap between the address electrodes and the electro-optic crystal plane of less than 1/10 micron. Therefore, even the smallest particles trapped between the crystal and the electrode structure can lead to non-uniform illumination problems at the light sensitive surface. The optical response of the device to pixels at the boundary between the modulated and unmodulated areas of the light valve is also significantly lower than the response to pixels near the center of the modulated region due to the addressing nature. No printer based on this technology has hitherto appeared on the market.

Proc. SID Symp. 250 (1
April 982), M. Little et al., "CCD Addressed Liquid Crystal Light Valve," describes an SLM with a CCD area array on the front side of a silicon chip and a liquid crystal array on the back side of the chip. CCD until a complete frame of analog charge data is introduced
The electric charge is input to. This charge is then released to the backside of the chip, where it modulates the liquid crystal. This device
Since the charges are spread by the transfer from the front side to the back side, it suffers from a significant constant pattern noise and deterioration of resolution.

Another type of SLM that can be made in both one-dimensional and two-dimensional arrays is a deformable mirror. The deformable mirror can be divided into three types. That is, an elastic body, a diaphragm, and a cantilever. In the elastic body method, when the metallized elastic body is addressed by a spatially varying voltage, this voltage causes surface deformation through compression of the elastic body. Elastic bodies are not good candidates for integration with high density silicon address circuits because they require address voltages on the order of 100 or 200 volts. In general, 24. IEEE Transactions on Electronic Devices, page 930 (1977), A. Lacatos and R.A. See Bergen's paper "TV Projection Display with Amorphous Se RUICON Light Bulbs".

There are various types of diaphragm-type deformable mirrors. One type replaces the oil film of the idfall device described above. In this apparatus, a thin reflection film is attached to a face plate of a cathode ray tube (CRT) by a supporting grid structure. The addressing operation is performed by a raster scanning electron beam as in the case of the idle fall. CRT
The electric charges deposited by the electron beam on the glass face plate of the above attract electrostatically the film kept at a constant voltage. This attractive force causes the film to sag in the well formed by the lattice structure, thus forming a tiny spherical mirror at each modulated pixel location. Light diffracted from this type of modulated pixel is concentrated in a relatively thin cone that is rotationally symmetrical about the specularly reflected beam. For this reason, this type of light valve is used with a schlieren stopper. This schlieren stopper
After specular reflection from the unmodulated area of the light valve, it consists of a single central shield positioned and sized to block the image of the light source formed by the optics. The modulated pixel produces a circular patch of light in the plane of the schlieren stopper, which is larger than, but centered on, the central obscuration. Stopper efficiency,
That is, the portion of the modulated pixel energy that is not blocked by the schlieren stopper is generally somewhat lower in a deformable film-based projector than in an oil film idfall projector. Moreover, such a deformable mirror device of the membrane has at least two major problems. High voltages are required to address the relatively rigid reflective film, and slight misalignment between the electron beam raster and the pixel support grid structure causes addressing problems. This misalignment causes blurring of the image as well as non-uniformity of display brightness.

Another type of membrane-shaped deformable mirror is three.
0 IEEE Transactions on Electronic Devices, page 539 (1983). Hornbeck article and US Pat. No. 4,441,7
No. 91, a hybrid integrated circuit composed of an array of metallized polymer mirrors bonded to a silicon address circuit. The lower analog address circuit is separated from the mirror element by an air gap,
The electrostatic attraction displaces the array of mirrors at the selected pixel. The resulting two-dimensional displacement pattern produces a corresponding phase modulation pattern for the reflected light. This pattern can be converted into an analog intensity change by the Schlieren projection method, or can be used as an input converter for an optical information processing device. However, membrane-shaped deformable mirrors suffer from manufacturing problems because they are susceptible to defects such as those that occur when very small, micron-sized particles are trapped between the membrane and the underlying support structure. . The membrane forms a tent over such trapped particles, the lateral extent of which is much larger than the size of the particles themselves, which are imaged by the Schlieren imager as bright spots.

A deformable mirror of a cantilever beam is a micromechanical array of deformable cantilever beams, which are electrostatically individually deformed by some addressing means to make incident light into a linear pattern or It can be modulated with an area pattern. When used with appropriate projection optics, the cantilevered deformable mirror can be used for display, optical information processing and electrophotographic printing. An early version using a metal cantilevered beam made on glass by vacuum deposition was described in US Pat. No. 3,600,798. This device has a structure that is not integrated,
There are manufacturing problems including alignment of the front and back glass substrates.

There are two cantilevered deformable mirror devices.
2 R. R., published in IEEE Transactions on Electronic Devices, page 765 (1975). Thomas et al. "The Mirror Matrix Tube: A Nobel Light Bulb For Projection Display" and U.S. Pat. No. 3,886,3.
10 and 3,896,338. This device is made as follows. A thermal silicon dioxide layer is grown on the silicon on the sapphire substrate. The oxide forms a pattern of four cantilevered clover leaf-shaped arrays connected at the center.
The silicon is isotropically wet etched until the oxide is undercut, leaving four oxide cantilevers in each pixel supported by a central silicon support column. Next, the clover leaf-shaped array is metallized with aluminum in order to have a reflectance. Aluminum deposited on the sapphire substrate is the reference grid electrode, which holds a DC bias. A scanning electron beam is used to address the device. This electron beam deposits a charge pattern on the leaf-shaped beam of clover,
The electrostatic attraction causes the beam to deform toward the reference grid. Erasure is accomplished by applying a negative bias to a closely spaced outer lattice to flood the device with low energy electrons. A Schlieren projector is used to transform beam deformations into changes in brightness on the projection screen.
An important feature of this device is that the cloverleaf shape causes the beam to deflect 45 ° from the opening between the beams. Therefore, even if a fixed diffraction background signal is blocked by using a schlieren stopper having a simple cross section, the modulated diffraction signal is not attenuated. The device is manufactured with a pixel density of 500 pixels per inch and the beam can be flexed up to 4 °. The optics used a 150 watt xenon arc lamp, reflective Schlieren optics, and a 2.5 x 3.5 foot screen with a gain of 5. 35ft lumen screen brightness, 15: 1 contrast ratio, and 48%
The beam diffraction efficiency demonstrated a resolution of 400 television scan lines. A write time of less than 1/30 second was achieved and the erase time was as short as 1/10 of the write time. However, this device suffers from reduced resolution due to scanning errors, poor manufacturing yield, and lack of advantages over conventional projection cathode ray tubes. That is, the positioning accuracy for each scan is not so high as to obtain the reproducibility of writing of individual pixels. As a result, the resolution is reduced, so that it is necessary to increase the number of pixels by at least four times in order to maintain the same resolution as compared with a lighter body having the same degree of writing. In addition, there is no etch stopper for the leaf-shaped support columns of the clover, the failure of the beam due to the wet etching of the beam, and the normal tensile stressed aluminum at zero stress on the oxide beam. The need for vapor deposition limits the yield of the device. Furthermore, this device is a standard projection CRT.
There are no obvious cost or performance advantages compared to.

A modification of the cantilever without the need for electron beam addressing and a vacuum envelope using the high voltage circuitry of the cantilever device previously described, integrated on silicon with the addressing circuit. The mirror to be obtained is 31 Applied Physics Letters, p. 521 (1977).
The K. Petersen's paper "Micro-Mechanical Light Modulator Arrays on Silicon" and U.S. Pat. No. 4,22.
No. 9,732. The first of these documents describes a 16 × 1 array of jumper plate cantilevers, which is made as follows. An epitaxial layer of <100> -oriented silicon (p-type or n-type) about 12 microns thick is grown on the p + substrate (or buried layer). The epitaxial layer is oxidized to a thickness of about 0.5 micron and covered with a Cr-Au coating having a thickness of about 500Å.
Cr-Au is removed by etching to form contact pads and address lines to form a dipping plate type metallized portion. In a second mask step, the oxide is etched away in a comb pattern around the metallization. Finally, the silicon itself is etched at 120 ° C. with a solution of ethylenediamine and catechol. If the mask is properly oriented with respect to the crystallographic axis, the metal-coated oxide studs are undercut by etching and freed from silicon. This etch is anisotropic,
Further lateral etching is stopped by the <111> planes that define the rectangular envelope of the comb pattern.
Furthermore, the etchant is suppressed by the p + material, which limits the depth of the well below the dip plate by the thickness of the epitaxial layer. When a DC voltage is applied between the substrate and the dip plate metallization, the thin oxide dip plate is electrostatically deflected downward toward the etched well.
The 106 micron long and 25 micron wide dip plate had a threshold voltage of about 66 volts.

The second cited document, US Pat. No. 4,229,732, describes a diving plate device (as an etch stopper for forming a well under a metallized silicon dioxide cantilever). Device, but with a different configuration is described. That is, the cantilever is in the form of a square flap that is hinged at one corner. The flaps form a two-dimensional array rather than a one-dimensional, one-column array like the diving plates, the wells under the flaps are unconnected, and the address lines for the flaps are between the row and column flaps. Can be formed on the upper surface of silicon. Of course, hingedly joining the flaps at the corners is described in US Pat. No. 3,886,310.
No. 3,896,338, the leaf-shaped structure of the clover, but the leaf-shaped structure of the complete clover cannot be used. This is because the clover leaf-shaped flaps are then hinged to the central post, which is isolated from the silicon surface, thus leaving no surface address lines. Moreover, such devices have poor resolution and low efficiency due to density constraints and small fractional active areas,
There are problems that the manufacturing yield is low, the contrast ratio is lowered due to the diffraction effect of the address circuit, and there is a residual image due to the effect of charging the oxide flap. More specifically, the address circuit is tucked around the active area (flap). This is because the wells are formed by etching away the epitaxial layer down to the p + type etch stopper, which leaves no choice to place the address circuitry under the active area. This causes the active area to be substantially reduced, which in turn reduces the diffraction efficiency. This means that more light power is needed for the same screen brightness. Since the address circuitry requires extra area, the pixel size is significantly increased over the flap area, resulting in a reduction in the achievable resolution. The wet etching required to form wells leads to low electrical and mechanical yields. In fact, wet cleaning, such as after dicing into chips, destroys the flaps and dive plates. This is because during the spin-wash / dry cycle the water trapped under the beam breaks the beam as it spins off the surface. Alternatively, when water is allowed to evaporate from the surface, the surface residue that remains behind increases the surface leakage current, which leads to variable operation of the device. In addition, the addressing circuitry on the silicon surface is exposed to and modulated by the incident light, causing unwanted diffraction effects from the transistor gate and lowering the contrast ratio. In addition, the light leaking into the address structure produces a charge generated by the light, which shortens the storage time. Finally, the oxide / metal flaps, with the insulating side facing the well, are charged due to the strong electric field existing before and after the well.
This produces a residual ("burn-in") image. The AC drive required to eliminate this residual image problem is described in N
It cannot be supplied from the MOS drive circuit. In addition, flexing the flap through the maximum stable deflection causes the flap to break and stick to the bottom of the well. Therefore, a voltage higher than the collapse voltage must be absolutely avoided.

A modified cantilever system is disclosed in K. K. of the 24 IBM Journal of Research and Development, page 631 (1980). Pietersen's paper "Silicon Twisting Scanning Mirrors" and 4 IEEE Electronic Devices Letters, page 3 (1983).
The M. Cadman et al., "A New Micromechanical Display Using Thin Metal Films". This scheme works by forming a metal flap connected to the surrounding reflective surface at two opposite corners and twisting the flap along the axis formed by this connection. The flap is not monolithically formed with the underlying addressing substrate, but is adhered to the substrate, similar to the previously described deformable membrane device.

All references cited above for cantilever beams describe the use of Schlieren projection optics with a cantilever device. However, such devices are limited in the optical performance they can achieve. First, the aperture diameter of the imaging lens must be larger than required to pass only the signal energy. Therefore, in order to pass all the signal energy before and after the shield at the center of the schlieren stopper, the speed of the lens must be relatively high (or, equivalently, the f-number is relatively Must be small). Moreover, in this imaging format, the signal passes through the outer portion of the lens pupil. Light rays emanating from any given point on the SLM and passing through the outermost area of the imaging lens pupil bring to a well-corrected focus, whatever the optical design of the imaging lens. Is the most difficult ray of light. With good control of the outer rays, the rays passing through the center of the imaging lens are automatically well corrected. Therefore, a higher level of optical design complexity is required for the imaging lens. Second, the angle of view is also limited such that the imaging lens can form a fully corrected image of the off-axis pixels of the cantilever SLM. Any lens design task is a trade-off between lens speed and angle of view that can be covered with good image quality. Fast lenses tend to work over a small field of view,
Wide-angle lenses tend to be relatively slow. The Schlieren imager must be well-corrected throughout its aperture, and the diameter of this aperture is larger than that needed to pass the image-forming light, so that the signal is of unobstructed diameter. More than if you could devise a different imaging format to pass through the center of a smaller lens,
The angle of view that can be covered by the lens is smaller. Finally, with a finite velocity imaging lens, the use of the Schlieren stopper format also limits the size of the light source available. This limits the level of radiant flux density that can be delivered to the projection screen or photoreceptor at the deflected pixel image. The level of this radiant flux density, ie the energy delivered per unit area, is related to the product of the radiance of the light source, the transmission of the optical system and the solid angle of the cone of rays forming the image. The radiance of a light source is determined only by the particular lamp used. The transmittance of an optical system is related to the stopper efficiency and surface transmission loss of a particular SLM / Schlieren stopper type. However, the solid angle of the light cone forming the image is directly proportional to the area of the pupil of the imaging lens which is filled with signal energy. The use of a schlieren stopper to block the central area of the pupil of the imaging lens limits the available pupil area and thus the image plane emission that can be achieved with a given velocity lens and a given radiance source. The level of bundle density is limited. This is in addition to the fundamental constraint on radiant flux density that the maximum light cone available has an aperture angle equal to the deflection angle of the beam.

For this reason, the known cantilevered SLM limits the fractional active area of the pixel by the address circuit, lowers the yield due to processing steps, is affected by the stress of the beam film, and Includes insulator charging effects, no overvoltage protection against beam collapse, incompatibility between low cost optics design and performance, and low contrast ratio for non-planar address circuits to surfaces. There's a problem.

[0016]

SUMMARY OF THE INVENTION The present invention relates to a spatial light modulator having a flexible beam and an electronic addressing system, and, generally as a final step, a method of making the beam free from a processing spacer. . This solves the problem of the known method of damaging the beam during processing. The preferred embodiment releases the beam by plasma etching the spacers. This plasma etch can be performed after dicing the substrate into chips, each chip forming an SLM and posts on the posts. Another preferred embodiment also plasma etches the spacer, but only halfway, thus allowing the spacer to be used as a structural element of an SLM. Such a method allows for an SLM structure that incorporates a substrate addressing scheme, and solves the diffraction and packing problems of known methods.

[0017]

DETAILED DESCRIPTION OF THE INVENTION A flexible beam spatial light modulator (SLM) of the present invention is typically formed with a linear or area array of pixels. Each pixel is individually addressable and has a deflectable reflective cantilever. Pixels are combined in the form of monolithic silicon-based chips. In the process of this invention, chips are manufactured by treating a silicon wafer, dicing the wafer into chips, and then treating the individual chips. Chip dimensions vary depending on the application. For example,
A linear array of 2400 x 1 pixels (which can be part of a printer with 300 dots per inch) can be made on a chip of about 1300 x 250 mils, with pixels about 12 microns square. SLMs operate by pixels reflecting light, and the reflected light is modulated by changing the deflection of a deflectable beam. Although the following description is primarily concerned with individual pixels of the SLM, the drawings are shown schematically for clarity.

One pixel of the first preferred embodiment of a spatial light modulator using a deflectable beam manufactured by the method of the first preferred embodiment of the present invention is shown in a simplified perspective view in FIG. 1A. , A side cross-sectional view is shown in FIG. 1B, and a plan view is shown in FIG. 1C. The pixel is shown generally at 20, but is basically a flap over a shallow well and includes a substrate 22, spacers 24, a reflective layer 26 and a flap 28 formed in layer 26. The flap 28 has a plasma etch access hole 30. Typical dimensions of the pixel 20 are as follows. The flap 28 is a square with sides of 12 to 25 microns, and the spacer 24 has a thickness of 1 to 2.5 microns (so that the distance from the bottom of the flap 28 to the substrate 22 is 1 to 2.5 microns. Layer 26 is 0.12 microns (1200 Å) thick, holes 30 are 2 microns square, and plasma etch access gap 32 is 2 microns wide.

The substrate 22 is <100> oriented silicon having a resistivity of 5 to 10 ohm cm. The spacer 24 is a positive photoresist that is an insulator. Layer 26 is 4
% Aluminum alloyed with copper. The coefficient of thermal expansion of this alloy is not significantly different from the spacer 24, which, as will be explained later, reduces the stress between layers 24, 26 caused by depositing layer 24 on layer 26. To be the smallest.

Pixel 20 is activated by applying a voltage between layer 26 and substrate 22. The exposed surfaces of flap 28 and substrate 22 form the two plates of the air gap capacitor, and the opposite charges induced in the two plates by the voltage exert an electrostatic force that attracts flap 28 to substrate 22. This attractive force causes flap 28 to bend in hinge area 34 and flex towards substrate 22. FIG. 2 shows this deflection exaggeratedly, and shows that the electric charges are concentrated in the region of the smallest gap. For voltages in the range of 20 to 25 volts, the deflection is in the range of 1 ° to 2 ° (1 deflection is
In the case of °, a flap with a size of 20 microns has a hinge 3
The furthest corner of flap 28 from 4 moves vertically about 0.5 microns). Note that this deflection is a significant nonlinear function of voltage. This is because the restoring force generated by bending the hinge 34 is a roughly linear function of deflection,
This is because the electrostatic attraction increases with the logarithm of the reciprocal of the distance between the nearest corner of the flap 28 and the substrate 22 (capacitance increases with decreasing distance, and thus the amount of induced charge is Note that as it increases, it gets closer. FIG. 3 shows the dependence of deflection on voltage. The voltage at which the flap 28 becomes unstable and bends completely to contact the substrate 22 is called the collapse voltage. At voltages just below the collapse voltage, the deflection is roughly a linear function of voltage (see dotted line in Figure 3), which is the analog operating region.

Applying a voltage between the flap 28 and the substrate 22 (in other words, addressing the pixel 20) occurs when considering other pixels of the cantilevered spatial light modulator. Since it is illustrated regardless of complexity, pixel 2
0 is extremely simple. The circuit describes another preferred embodiment which will be described later. Here, a manufacturing method of the pixel 20 will be described first.

The manufacturing process of the pixel 20 is as follows.
(1) Initially <100 with a resistivity of 5 to 10 ohm cm
> Oriented silicon substrate is used (typically the substrate is in the form of a 3 inch diameter circular wafer). (2) If the layer is to be about 2 microns thick, a chlorobenzene-insoluble positive photoresist layer (eg, a resist based on a non-barrack resin) is spun on. This rotational attachment is
It should be done in three steps to avoid surface waves generated in the thick spin deposit. That is, about 0.7 micron resist is spun on and baked, another 0.7 micron resist is spun on, baked again, and the last 0.7 micron resist is spun on and finally baked. To do. (3)
In order to minimize the temperature mismatch and the resulting stress between the metal layer and the resist layer, 4% at a temperature close to room temperature.
A 0.12 micron aluminum layer alloyed with copper is deposited by sputtering. (4) Apply a positive photoresist and define its pattern by photolithography to define plasma etch access holes and gaps. (5) Plasma exposed aluminum alloy
Etch (for example, chlorine-3 boron chloride-4 silicon chloride etch gas can be used) to form plasma etch access holes and gaps. (6) Spin-on a PMMA (polymethylmethacrylate) layer that acts as a protective layer during subsequent steps. (7) Dice the substrate into chips (each chip becomes an SLM). (8) Dissolve PMMA by spraying chlorobenzene, and immediately remove the diced pieces by centrifugation.
Note that the positive resist does not dissolve in PMMA developer. (9) Perform an isotropic plasma etch of the chip to remove the spacer (positive photoresist) from under the flap, thus forming a well. This etch also removes the resist layer on top of the flaps,
The chip processing is complete. Protects during die cutting work P
Note that during MMA cleaning, the flaps are not directly exposed to the diced debris. The convenient plasma etch of step 9 is oxygen based. Oxygen is PMM
Rapidly etches A and photoresist, but neither silicon nor aluminum. Spacer is about 120 ℃
This plasma etch must be at a low temperature as it softens at.

Of course, the plasma etch of step 9 must be monitored or timed. This is because if we continue to remove the spacers beyond the portion of the spacer between the flap and the substrate, the support for the reflective layer will eventually be removed,
This is because it enters the adjacent pixel. This is the reason for providing the plasma etch access holes in the flap. Figure 4
As shown in FIGS. 4A-4D, the removal of spacers is orientation independent and undercuts layer 26 at the same rate that undercuts flap 28. FIG. 4A shows the initial stage of the etch. The left part is a plan view, and the horizontal range of the undercut is indicated by a broken line. The right part is a side cross-sectional view (the cross-section is along the diagonal of the flap through the hinge), showing the effect of the plasma etch access hole 30. 4B and 4C show successive steps and FIG. 4D shows when the etch is complete. As is apparent from FIG. 4C, the holes 30 are shallow wells or thin spacers 24.
Is most effective against

The precise monitoring of the plasma etching step 9 can be performed as follows. Each etch control structure at the corner (s) of the chip has the pattern shown in FIG. The top is shown in a plan view and the bottom is shown in a side sectional view. The rectangular portion 36 of the layer 26, which is isolated from the rest of the layer 26, is called a hollow beam and has a series of increasing widths. During the plasma etch removal of the spacers, the control corners of the chip are periodically observed under a bright field microscope. FIG. 6 shows a series of positions taken by the hollow beam as the plasma etch progresses. When viewed under a bright field microscope, the hollow beam 36 initially appears bright and continues to appear bright until undercuts from the plasma etch from both sides of the hollow beam 36 are encountered. When they meet, the hollow beam 36 is released from the spacer 24 and tilts by a sufficient angle. At this time, the beam 36 looks dark. Continuing the plasma etch removes the crescent-shaped spacers 24 that support the beam 36, causing the beam 36 to re-bright through the midtones. By examining the etch control structure as a function of time, the extent and rate of undercut can be precisely determined. Once the beam 36 is tilted, the beam no longer has useful information because the etching of the crescent shaped spacers 24 supporting the beam 36 is done from only one side. Therefore, the extra time required to bring beam 36 back into the bright-looking state does not represent an undercut rate for the flap of the SLM tip. The critical width of the hollow beam (the width at which the beam tilts so that the spacer is just entirely under the flap, see FIG. 4D) is empirically determined. The plasma etch undercut rate is related to the area of the exposed spacer material and is not a simple shape. For convenience of monitoring the progress of plasma etch,
Beams with some distributed width around this critical width are chosen.

Flap 2 using the same process but different shapes
It is also possible to form a pixel 20 having 8 pixels. FIG. 7 is a plan view of various alternatives. Of course, the number of plasma etch access holes is related to the size of the flap or dip and the depth of the well below it. The flaps bend primarily at the hinged joints, while the diversion plates bend along the entire beam. Since the flap bends at a substantially constant angle, it has a higher diffraction efficiency than a diving plate of comparable size. The flap hinges may be in the corners or along the sides. The flap corner hinge uses a 45 ° schlieren stopper or 45 ° darkfield discrimination against the plasma etch access gap that defines the flap's perimeter to provide a high contrast projected image of the flexed flap. Can occur.
With this discrimination method, the opening (plasma
Design an optical stopper that blocks all on-axis diffracted light from the etch access gap) but allows off-axis light from the 45 ° flex of the flap to pass through. Some of this light passes around the schlieren stopper because the plasma etch access holes diffract the light with almost isotropicity. However, each hole diffracts only a small fraction of the total light energy incident on each pixel, and only a small fraction of this energy passes through the stopper,
The resulting degradation is negligible.

As shown in two of the alternatives shown in FIG.
By extending the hinge area, the flexure sensitivity of the flap can be increased.

Pixels 20 of various modifications are made in substantially the same manner.
Can be manufactured. For example, FIG. 8A shows the conductive substrate 4
2. Spacer 4 which may be either insulating or conductive
4 is a side sectional view of a pixel 40 including a dielectric layer 46 and a metal reflection layer 48. FIG. A flap 50 is formed in layers 46 and 48, similar to flap 20 in layer 26, and has plasma etch access holes 52 and hinges 54. The only constraint is that the spacer 44 must be able to be etched by a plasma that does not etch the substrate 42 or the dielectric 46. Metal 48 can be deposited after spacers 44 have been removed by plasma etching to form a well under flap 50, which results in a metal deposit 56 at the bottom of the well, as shown in FIG. 8A. Remain. However, the metal deposit 56 is benign, and depositing the metal 48 as the last processing step ensures a very clean, stress-free surface. This makes it possible to use pure aluminum, so that a very high reflectance is obtained. Metal 48
If the spacers 44 can be etched without etching, then both the dielectric 46 and the metal 48 can be deposited and patterned before the spacers 44 are removed. This avoids the metal deposit 56. Further, if the substrate 42 is etched by the plasma etch used to remove the spacers 44, an etch stop layer is formed over the substrate 42 before the spacer layer 44 is formed and the same process steps are used. You can The etch stopper layer may be insulative. For example,
Both the etch stop layer and the dielectric layer 46 may be silicon dioxide and the spacer 44 may be polysilicon.

The composite flap of the pixel 40 (the metal being above the dielectric) causes a charging action at the dielectric-air interface due to the strong electric field present in the well. In order to avoid this charging action, the electric field before and after the well must be periodically reversed by placing an address signal on the AC carrier. This method has no noticeable drawbacks for a linear array of pixels, but for an area array of pixels with active switching elements at each pixel location, the AC addressing scheme becomes quite complicated, and such metal / Composite flaps of dielectric should be avoided. The same considerations need to be made in the case of diving boards as well as in the use of an insulative insulative etch stop layer on the substrate 42.

In the pixels 20 and 40, although not explicitly stated, the signal (address) is applied to the beam (flap or jumper plate) because the substrate is conductive and common to all pixels of the SLM. Need to be applied. Each beam is connected to an electrode formed by defining the pattern of layers 26,48. This patterning can be done in the same step as patterning layers 26, 48 to form plasma etch access holes and gaps. See FIG. 9, which shows a plan view of a few adjacent pixels with connection electrodes. Of course, this means that the electrodes are also undercut during the plasma etch that removes the spacers. If the electrodes are formed by patterning layers 26, 48 after a plasma etch to form wells under the beam, a significant yield loss would be expected due to beam breakage. However, as previously described for pixel 40, plasma etch access holes and plasma etch removal of spacers 44 to define the dielectric layer 46 for both the gaps and electrodes. The metal 48 can then be deposited. Due to the undercutting effect of the dielectric during the plasma etch, the deposited metal does not form a short circuit electrode, but instead leaves a deposit on the substrate between the electrodes, as shown in Figure 8B. Such a beam addressing system is an SLM having a linear array of pixels.
However, in an SLM having an area array, too much of the area of the SLM is assigned to electrodes and too little to pixels, and therefore cannot be used. Furthermore, the gap between the electrodes (see FIG. 9) diffracts the incident light (which cannot be entirely focused on the pixel), and combined with the large linear dimension of the gap,
Due to the flare of the lens of the optical system, even a schlieren stopper cannot completely prevent this diffracted light from appearing as noise in the fixed pattern of the SLM output.

The pixel 60 of the second preferred embodiment is shown in FIG.
A side cross-sectional view is shown in FIG. That is, all the beams in the array of pixels are electrically interconnected and the signal electrode is at the bottom of the well under the flap. This can eliminate the diffraction caused by the electrode gap described with reference to FIG. Pixel 60
Has a silicon substrate 62, an insulating layer 64, an electrode layer 66, a spacer 68 and a reflective layer 70, and a flap 72 has a plasma etch access hole 74 and a hinge 7 in the reflective layer.
It is formed together with 6. Insulating layer 64 may be silicon dioxide grown on substrate 62. The electrode layer is LP
It may define a pattern of polysilicon deposited by CVD. The spacer may be a positive photoresist by spin deposition. The reflective layer can be aluminum alloyed with 4% copper and deposited on the spacer layer 68 by sputtering. The manufacturing process of the pixel 60 (shown in FIGS. 11A to 11D) is the same as that of the pixel 20, and is as follows. <1
A thermal oxide layer having a thickness of about 2000 Å is grown on a substrate of 00> orientation. Polysilicon is deposited to a thickness of 3000 Å and is doped with phosphorus to have an area resistivity of about 50 ohm / square. Plasma etch the polysilicon using the electrode pattern. The positive photoresist is then spun on in three separate applications and baked to a total thickness of about 2.4 microns. In order to avoid the surface wave of the resist that can occur when spin-coating a very thick layer,
The resist is applied in three times to build up this thickness. Baking to about 180 ° C. between each application is necessary to prevent the previous layer from dissolving in the resist solvent. Even after the last layer, baking at 180 ° C. is required to drive out excess solvent from the spacers. This last bake avoids the formation of solvent bubbles under the metal of the beam when firing the photoresist for photolithography of the beam pattern. Then, about 1200Å of 4% Cu: Al alloy is deposited by sputtering at a substrate temperature as close to room temperature as possible. In general, organic (material) spacers have a larger coefficient of expansion than aluminum, so when the wafer cools from the sputtering temperature, the metal of the beam becomes compressed, causing buckling of the beam at its pivot point. There is. (Note that this extraneous compression induced by heat is separate from the inherent compression that results from the deposition process itself.) As a spacer material, positive resist is more preferred than aluminum. It is preferable to PMMA because it has a close expansion coefficient. The copper-aluminum alloy was chosen because it resists the formation of bumps during firing for photoengraving and is less prone to fatigue after long-term operation of the beam. A positive resist is spun onto the metal of the beam and developed in the pattern of the beam. Plasma etching of the metal of the beam (chlorine added with a recombination agent to etch aluminum)
Forming a plasma etch access hole and a gap. A resist is defined on the wafer to define the metal of the beam and to define the pattern. A protective layer of PMMA about 1.5 microns thick is spun on and the chips are diced using a diamond saw. Each tip is placed on a spinner and the surface is flooded with PMMA solvent while periodically pulsing the spinner from a low spin rate (100 rpm) to a high spin rate (8000 rpm). This pulsing operation has the advantage of improved removal of die cutting debris, and during the low rotational speed portion of the cycle P
The MMA layer softens. Then, when the rotation speed is suddenly increased to a high rotation speed, the softening layer and the embedded die cutting fragments are thrown away by the strong centrifugal force. All PMMs
Repeating this cycle of rotation until A is removed ensures a die-debris-free surface that is likely to get caught in the plasma etch access holes and gaps and affect the nearby plasma etch rate. It The tip is plasma etched in oxygen on the temperature controlled cathode of a planar plasma etcher. Temperature is 60 to 100
Control within the range of ℃. Above 100 ° C, the spacers may soften and relieve stress, and when cooled, the metal of the beam will be in a compressed state. If the temperature is lower than 60 ° C, the undercut speed is remarkably reduced. The energy density and pressure are chosen to provide isotropic etching of the spacers to minimize the undercut time. During the initial step of undercutting the beam, the overlying positive resist layer is etched away. The result is a clean, undercut tip with no mechanical damage to the beam and tension on the metal of the beam.

The operation of pixel 60 is similar to that of pixel 20, but a signal is applied to electrode 66 at the bottom of the well, with flap 72 and substrate 62 both grounded or DC biased. Again, the charge induced in the plates of the air gap capacitor formed by the flaps 72 and the electrodes 66 exerts an electrostatic force which causes the flaps 72 to flex and the restoring force generated by bending the hinges 76. Will counter it.

There are many deformations of the pixel 60, which are readily conceivable immediately, but there are the following. (a) The shape of the beam may be a flap or a dive plate, as shown in FIG. Of course, the number and location of plasma etch access holes (if any) depend on beam size, well depth,
And the tolerances under which the beam may be undercut. (b) The beam may be a composite of metal on insulator (see, eg, pixel 40 and its associated description), or a composite of two metals such as aluminum on refractory ( Aluminum has a reflectivity, and refractory has a high yield stress). Note that the refractory metal pattern can be defined and the aluminum deposited after the plasma etch. This means that aluminum is deposited at the bottom of the well (Fig. 8
And its description), which avoids the formation of aluminum patterns, thereby avoiding the growth of ridges or other stress relieving effects that may occur during photoresist baking. (c) Spacer 68 may be any spin-on insulator such as PMMA. The spacer 68 uses polysilicon for the electrodes 66 and thermally grows oxide (LOCOS) between the electrodes for isolation and leveling with the electrodes 66.
Alternatively, when the address electrode 66 has a planar shape so as to be obtained when the SWAMI method is sufficient), it may be an insulator having a uniform deposit such as silicon nitride. (d) The electrode 66 is n + in the substrate 62 which is p-type silicon
It may be a shape diffusion part. The junction is reverse biased to isolate the electrodes. The insulator layer 64 can be omitted when the insulating spacer 68 is used. (e) Between the spacer 68 and the reflective layer 70 (as in the composite beam or pixel 40 of metal on insulator in (b)).
Alternatively, if an insulating layer is formed between the spacer 68 and the electrode 66, the spacer 68 may be a conductor such as polysilicon. For example, in the diffusion type electrode (d), the planarization oxide 78 (LOC) is formed before the spacer 68 is deposited.
OS process) can be formed. See Figure 10B. (f) The transistor for switching the electrode 66 is
It can be formed under the conductive spacer 68 by various methods. For example, as shown in FIG. 10B, the electrode 66 is divided into two parts as shown by a broken line on the left side of the drawing to form the drain and source of the field effect transistor,
Oxide 78 can be the gate oxide and spacer 68 can include the gate.

Many of the variants described above are based on pixels 20 and 4.
The same applies to 0.

A third preferred embodiment pixel, indicated generally by the reference numeral 80, is shown in side cross-sectional and plan views in FIGS. 12A and 12B. Pixel 80 is substrate 82, insulating layer 8
4, including electrodes 86, spacers 88, and a reflective layer 90 having flaps 92 and hinges 94 formed therein. Electrode 86 is shown in phantom in FIG. 12B and has a triangular hole under flap 92 farthest from hinge 94. This hole in the electrode 86 allows the flap 92 more flexibility before instabilities and collapses occur. This is because the induced charge that exerts an attractive force is now closer to the center of the flap 92, so that the force changes as a function of deflection than it does with the charge farthest from the hinge. Is not so rapid. K.
Jump plate type beam model developed by Petersen (25
The article "Dynamic Micromechanics on Silicon: Technics and Devices," published in IEEE Transactions on Electronic Devices, page 1241 (1978), shows maximum stability under uniform load. The deflection is stated to be about 44% of the well depth (spacer thickness), regardless of the length of the beam. Therefore, by using an electrode with a hole below the free end of the beam, the effective length of the beam when using this model is only a fraction of the length of the beam, and the end of this effective length is the well. It can be flexed stably up to 44% of the depth. This means that the ends of the beam will flex more greatly.

The pixel 100 of the fourth preferred embodiment is shown in FIG.
3A and 13B are shown in cross section and plan view, respectively. The pixel 100 includes a substrate 102, an insulating layer 104, and an electrode 10.
6, the spacer 108 and the reflective layer 110, and the electrode 10
6 is provided with a triangular hole shown by a broken line in FIG.
2 is formed. Pixel 100 is manufactured by the same process steps as making pixel 80. Pixel 100 operates by twisting flap 112 along an axis that passes through two twist hinges 116. Torsion torque is electrode 10
Caused by the signal applied to 6. This is the electrode 106
13A, only the left side portion of the flap 112 is attracted as seen in FIG. 13A. Note that FIG. 13A is a view along the twist axis, and the twist is counterclockwise. The twist hinge compliance is the twist hinge 116.
It can be adjusted by changing the length-width ratio and the thickness of the reflective layer 110.

Pixel 100 has 45 to the plasma etch access gap that defines the perimeter of flap 112.
Deflection at an angle of °, which causes the pixel 10 to
For optical processing of reflected light from an SLM composed of pixels similar to 0, a 45 ° Schlieren stopper or 4
5 ° dark field discrimination can be used. Just flap 1
Pixel 10 has a larger deflection angle than that of the flap of the same size, because 12 bends above and below the reflection layer 110.
0 has a more efficient diffractive action than the beam of a bent hinge. As mentioned previously, the maximum stable deflection is determined by how close the flap corner is to the electrode, which is 44% of the well depth in the springboard model described for pixel 80. . Finally, the pixel 100 is fitted with both ends of the flap 112 so that rotation about the twist axis is not affected by stress gradients or compression in the reflective layer 110. For this reason, the pixel 100 has a layer 110 thickness greater than that of a cantilever pixel having only one hinge.
Can operate under a wider range of deposition conditions.

One pixel of the fifth preferred embodiment for a flexible beam SLM is shown in cutaway perspective view in FIG. 14A, in cross-sectional view in FIG. 14B, and in plan view in FIG. 14C. ing. The pixel is shown generally at 120, which basically consists of a clover leaf consisting of four flaps hingedly coupled to a central post on the substrate, but with a substrate 122, an insulating layer 124, Field plate 126, flap 128, support post 130 and flap
It has an address diffusion unit 132. The support column 130 has a cylindrical shape and symmetrically supports the four flaps 128 by the hinges 134 at the corners. (See Figure 14C). Pixel 1
The curved dimensions of 20 are as follows. Flap 128
Is a square with a side length of 12 to 25 microns, and the gap between the flaps is about 1 micron wide.
30 has a diameter of 1 to 2 microns and a height of 1 to 2.5
It is micron. The flap 128 has a thickness of about 0.12 micron (1200Å) and the insulating layer 124 has a thickness of 0.15 micron (1500Å).
The thickness of the plate 126 is about 0.1 micron (1000
Å).

Substrate 122 is <100> oriented silicon and has a resistivity of 5 to 10 ohm cm. Flap 12
8 and support column 130 is a piece of aluminum alloyed with 4% copper. This alloy minimizes aluminum ridge growth as a stress relief during metal deposition and processing, but still has a relatively high reflectivity. The insulating layer 124 is silicon dioxide and the field plate 126 is aluminum.

By applying a voltage between the flap 128 and the field plate 126, the pixel 120
Works. The flap 128 and field plate 126 form the two plates of the air gap capacitor, and the opposite charges induced in the two plates by the voltage exert an electrostatic force that attracts the flap 128 to the field plate 126. This force flaps 128 at hinge 134
To bend the flap toward the substrate 122. Figure 15
Is an exaggerated view of this flexure, together with the charge concentrated in the smallest gap area. For voltages in the range 20 to 25 volts, the deflection is in the range 1 to 2 °. (For 1 degree deflection, for flaps of size 20 microns, flap 12 farthest from hinge 134)
The vertical displacement of the eight corners is about 0.5 micron. Although the restoring torque produced by the bending of hinge 134 is a roughly linear function of deflection, the torque applied by the electrostatic force is the same as field plate 126 and hinge 134.
Farthest from the corner of the flap 128, that is, the field
Note that this deflection is a very linear function of voltage, as it increases approximately as the logarithm of the distance between the corners closest to plate 126. FIG. 16 shows the dependence of the deflection on the applied voltage. The voltage at which the flap 128 becomes unstable and bends completely to contact the field plate 126 is called the collapse voltage. At voltages slightly below the collapse voltage, the deflection is approximately a linear function of voltage (see dotted line in FIG. 16), which is the analog operating region of pixel 120.

The manufacturing process of the pixel 120 is shown in FIGS.
It is shown in Figure 7F and is as follows. (1) A silicon substrate having an initial specific resistance of 5 to 10 ohm cm <100> is used. (Typically, the substrate is in the shape of a circular wafer having a diameter of 3 inches.) (2) Diffusion line 132, insulating layer 124
And field plate 126 are formed by standard stamping, deposition and photolithographic methods. (3)
A planarizing spacer 140, such as a positive photoresist, is spun on to a thickness of 1 to 2.5 microns. (This is the distance from the flap 28 to the field plate 26.) The pattern is defined so that holes 142 are drilled. (4) Deposit 4% copper-aluminum alloy to a thickness of 0.12 micron on the spacers by sputtering. (5) Apply photoresist 144,
The pattern is determined by photoengraving, and the flap 12
The gap between 8 and its periphery is limited. (6) Plasma etch the exposed aluminum (eg, in a mixture of chlorine, boron trichloride and silicon tetrachloride) to form the gap between flaps 128 and its perimeter. (7)
Polymethyl, which acts as a protective layer during the subsequent steps
A layer 146 of methacrylate (PMMA) is spun on. (8) Dice the substrate into chips. (Each chip becomes an SLM.) (9) PMMA is dissolved by spraying chlorobenzene, and the die cut pieces 148 are immediately removed by centrifugal action. Note that the positive photoresist does not dissolve in chlorobenzene. (1
0) Plasma etch the chip (in oxygen) to remove the positive photoresist, including spacers 140, thus forming the pixel.

Adjacent flaps 128 on support posts 130
The width and length of the gap (as well as the diameter of the post 130 and the thickness of the metal forming the flap 128 and the post 130) are
Determine the rigidity of hinge 134. Therefore, the sensitivity of the pixel 120 can be adjusted. Since the field plate has a common voltage for all pixels and does not require isolation, pixel 12 is shown as shown in broken lines in FIGS. 14C and 17.
0s can be packed closely in the array.

The support posts 130 allow the substrate 122 (at the diffuser 132) to be formed by the aluminum wetting the silicon surface.
Adhere firmly to. However, other pixel shapes are readily conceivable and provide wider contact with the base of the support column 130. For example, FIG. 18 shows a pixel 160 in cross section. The pixel 160 is a p-type silicon substrate 162, n
It has a + source region 164, an n + drain 166, a p + channel stopper 168, a gate oxide 170, a metal gate 172, a metal contact 174, a support post 176 and a beam 178. A beam 180 of adjacent pixels is also shown. Pixel 160 is manufactured by essentially the same process as pixel 120, but with field effect transistors for addressing (source 164, drain 166 and gate 172).
The difference is that there are extra steps involved. But gate 17
Metal contact 174 due to the metal deposit for 2
Is also deposited, and the subsequent deposition of the support post 176 is not to the source 164, but to the metal contact 17
It is performed for 4. Note that gate 172 is also a field plate.

Typically, the metal deposition to form the flap 128 of the pixel 120 and the flap 178 of the pixel 160 leads to the metal being compressed. (There is both intrinsic compression at the interface with the spacer and, in some cases, extraneous compression due to the metal having a smaller expansion coefficient than the spacer.) However, complete removal of the spacer causes There is no possibility of metal buckling at the interface.

Alternate support columns for the flaps of pixels 120 and 160 are shown in FIGS. 19A to 19D show a manufacturing process of the support column 190. Spacers (spin-on for planarization, or similar if substrate 190 surface has planarization circuitry only)
, At a location relative to the support post 190. See Figure 19A. In this figure, the spacer is shown at 194,
Circular holes 196 are shown that define a pattern. A conductive material for the support posts 190 is conformally deposited or deposited on the spacers 194 to define a pattern and define the posts 198. See Figure 19B. The thickness of this deposition is chosen to adequately cover hole 196. Beam metal 200
Are deposited to cover both spacers 194 and columns 198. See Figure 19C. The rest of the process is the same as for pixel 120, with support pillars 190 and flaps shown in Figure 19D. The structure in which the supporting column 190 is composed of two kinds of materials 198 and 200 makes it possible to make the supporting column extremely rigid without having the rigidity corresponding to the flap, and at the same time, to the edge of the spacer hole. The cover of the material column of the support pillar above the base is improved.

20A to 20D show a manufacturing process of the support column 210. In contrast to the support posts 130, 176 and 190, the support posts 210 are limited prior to spacer deposition. First, a conductive material is deposited and patterned to form columns 212. See FIG. 20A. The pillar material can be polysilicon, metal, or other conductor that is not etched in the final plasma etch that removes the spacers. Next, a spacer 214 for planarization is rotationally attached to a thickness corresponding to the height of the column 212. The planarization effect of the spacer material reduces the thickness above the posts 212. See Figure 20B. Next, a spacer mask 214 is deposited, a pattern is defined, and holes 21 are formed on the pillars 212.
Open eight See Figure 20B. Next, plasma the exposed spacers until the column 212 is clear of spacer material.
To etch. See Figure 20C. The spacer mask 216 is then removed and the beam material 220 is deposited. The remaining steps are the same as previously described, which results in FIG.
The support pillar 210 shown in FIG.

Note that support pillars 190 and 210 both have large effective diameters due to the tolerances of mask alignment that require the flanges and holes 218 on the top of pillars 198 to be oversized. I want to be done. However, the support posts 130 require that the holes 142 be sufficiently tapered so that the beam metal can cover the steps of the holes. If the holes 142 have to be etched with a large taper such that the diameter of the holes is too large, the support posts 130 can be changed to support posts 230. The manufacturing process of the support pillar 230 is shown in FIGS.
Shown in 1D. First, spacers 232 are spun on to define the pattern of holes 234. See Figure 21A. Next, a thick metal layer 236 is deposited over the spacer 232 and the holes 234. See Figure 21B. This is followed by an anisotropic etch of metal 236, leaving columns 238 shown in Figure 1C. When the metal 236 is aluminum, a plasma etch in a mixture of chlorine, boron trichloride and silicon tetrachloride can be used to achieve the desired anisotropy. Finally,
The beam metal 240 is deposited and the process described above results in the support columns 230 shown in FIG. 21D. Support pillar 2
Reference numeral 30 is a composite of a thick metal cylindrical column 238 and a beam metal layer 230. The posts 238 assist in connecting the thinner beam metal to the underlying addressing structure, both mechanically and electrically.

A sixth preferred embodiment pixel, generally designated 250, is shown in cutaway perspective views in FIGS. 22A-22C, respectively.
It is shown in a side sectional view and a plan view. The pixel 250 has a substrate 252, a spacer 254 and a metal layer 256 in which a flap 258 and a hinge 260 are formed. The flap 258 has a plasma etch access hole 262 that helps to quickly remove the spacer 254 from under the flap 258, as described below. Flap 2 for the address circuit for the pixel 250
It can be provided in the substrate 252 below 58 and the metal layer 256 including the flap 258 is kept at a common voltage for all pixels in the SLM. Alternatively, the address operation may be performed by setting the substrate 252 to a common voltage and dividing the metal layer 256 into one electrode for each flap, as shown by the broken line in FIG. 22C. Details of the address circuit, including the insulator between the metal layer 256 and the substrate 252, have been omitted for clarity.

As shown in the right portion of FIGS. 22A and 22B, the pixel 250 has a spacer 2 covered with a metal 256.
A surrounding portion of 54 can be formed. Pixel 250
The pixel is manufactured by the same process as that for manufacturing the pixel 120, but the spacer in the surrounding portion is not removed by plasma etching in oxygen because the metal 256 overlying the spacer protects the spacer. 23A to FIG.
3D shows a side sectional view and a plan view of a processing step for manufacturing the pixel 250. First, the substrate 2 with the circuitry formed on its surface (not shown for clarity)
52 is covered with a spin-on spacer layer 254 for planarization. Note that layer 254 has a planar surface even if the circuitry on the surface of substrate 252 is not planar. Layer 254 may be a positive photoresist, similar to the process described above for pixel 120, and is typically 1 to 2.5 microns thick. Then, the pattern of the layer 254 is defined and covered with the metal layer 256 by sputtering deposition or the like. See Figure 23B. Then the metal layer 2
56 patterns are defined to define the perimeter of flap 258 and plasma etch access holes 262. This pattern is determined by a process using a multi-layered resist in multiple steps because a large step is inserted by the spacer 254. If the metal layer 256 is an aluminum alloy, it is convenient to use a chlorine-based plasma etch to define this pattern. See Figure 23C. Finally, as in the case of the pixel 120, the substrate 252 is diced into chips, and then oxygen-based plasma etching is performed to cover the spacer 25 from below the flap 258.
Remove 4. See Figure 23D. Note that the spacer 254 shown in Figure 23D is not removed in the final etch. This is because it is protected by the metal layer 256 and thus can be used as an insulator or other structural element of an SLM.

Some alternate shapes of pixel 250 are shown in FIG.
4A through 24C, the patterned spacers 254 are not on the steps as in the case of the pixel 250.
2 shows the pattern of flaps 258 over the flat portion of the. Due to such a pattern, the metal edge 264 does not adhere to the substrate 252, and if the metal 256 is in a compressed state (the inherent compression due to the non-epitaxial deposition, and the thermal expansion coefficient of the metal 256 and the spacer 254). Both the difference and the extraneous compression due to the high temperature of the deposit) may cause buckling when deposited. However, the flap 258 is restrained only by the hinge 260, and the range of the flap 258 is too small to prevent the buckling of the lowest mode.
58 does not buckle. Further, the hinge 260 itself is the substrate 252.
, And thus cannot buckle freely, so that flap 258 remains flat and parallel to substrate 252.

Another variation of pixel 250 is shown in FIG. 25, which includes two levels of metal and insulator on substrate 252. The first level metal 255 is patterned to have holes corresponding to the areas that form the flaps (shown in phantom in the opposite view). Next, the flattening spacer 25
4 is attached by rotation, and a hole pattern for hinges is defined therein. Metal 256 is then deposited over spacer 254 to define a pattern and define flap 258. Finally, the spacer 254 is removed. This variation of pixel 250 is that pixel 120 is in that all spacers 254 are removed.
Very similar to.

Pixels 120, 160 and 250 all have corner hinges and therefore bend along a line rotated 45 ° with respect to the axis formed by the perimeter of the flap. In this configuration, a 45 ° Schlieren stopper or a 45 ° darkfield discrimination can be used to generate a high contrast projected image of the deflected flap. This discriminating method blocks all on-axis diffracted light coming from the flap's peripheral aperture and, if present, the surrounding covered spacer portion, but allows it to pass off-axis light coming from the flap's 45 ° deflection. A simple optical stopper. plasma·
The etch access holes and hinges are substantially isotropic and diffract the light so that some of this light passes around the schlieren stopper. However, the resulting degradation is negligible because the holes and hinges diffract only a small fraction of the total light incident on each pixel.

A seventh preferred embodiment pixel, generally indicated at 270, is shown in cutaway perspective, side sectional and plan views in FIGS. 26A-26C. The pixel 270 is the substrate 27
2, a metal layer 274 having a flap 276 and a hinge 278,
And an address electrode 280 isolated from the substrate 272 by an insulating layer 282. Pixel 270 is pixel 12
0 and 250 are manufactured by similar process steps as previously described. Twisting flap 278 along an axis passing through hinge 278 operates pixel 270. Torsional torque results from the signal applied to electrode 280. This attracts only the left side portion of flap 276 as viewed in FIG. 26B. That is, FIG. 26B is a view taken along this twist axis, and the torque generated by the signal causes counterclockwise rotation. The metal compression of flap 276 caused by the deposition is eliminated by bending hinge 278 as well as keeping flap 276 parallel to substrate 272.

27A and 27B show a modification of the field plate of pixel 120, it can also be used for pixels 160, 250 and 270, as described below, for a given spacer thickness. (Ie, the distance from the flap to the field plate), the deflection angle of the flap can be made larger. From the simple model of electrostatic force and restoring force to the flap with the hinge in the corner, the signal voltage is changed so that the farthest tip of the flap bends by 51% of the distance to the field plate. It can be seen that as the voltage increases, further increases in voltage destabilize the flap and cause it to collapse against the field plate. Recall the description of FIG. 16 which labeled this maximum stable flap voltage as the collapse voltage. From the same simple model, it can be seen that when electrostatic force is applied to only the half of the flap closest to the hinge, the tip of the flap can be flexed stably by 83% of the distance to the field plate. . (Of course, the collapse voltage for this half actuated flap is higher than the collapse voltage for the entire actuated flap.) Therefore, the portion of field plate 126 below the tip is shown in FIGS. 27A and 2.
By removing it as shown in 7B, the flap 128 only has a partial effect, which results in a greater and more stable deflection. Furthermore, for the same breakdown voltage,
A thinner spacer can be used. In the above description, various changes in dimensions, materials, configurations and the like can be easily considered.

Further, the following matters will be disclosed.

(1) (a) A plurality of pixels formed in a layered structure, (b) the layered structure is a substrate, a spacer layer on the substrate, a reflective layer on the spacer layer, and an electrical address circuit. (C) each of the pixels is (i) an electrostatic deflectable element formed in the reflective layer,
And (ii) a well formed in the spacer and disposed below the deflectable element and an adjacent portion of the reflective layer, the well being between the substrate and the reflective layer. A spatial light modulator characterized by being formed by plasma etching of spacers.

(2) In the spatial light modulator described in the first item, (a) the spatial light modulator which is a rotationally adhering material for making the spacer flat.

(3) In the spatial light modulator described in item 1, (a) the deflectable elements of the pixels are electrically interconnected (b) the address circuit is provided for each pixel. A spatial light modulator including an electrode on the substrate at the bottom of the well.

(4) The spatial light modulator as described in the item (1), wherein: (a) the reflective layer has no insulating material.

(5) In the spatial light modulator described in the paragraph 1, (a) the deflectable element has at least one hole, and the plasma etching of the spacer is advanced through the hole. A spatial light modulator designed to be able to do.

(6) (a) having a plurality of pixels formed in a layered structure, (b) the layered structure having a substrate, a spacer layer on the substrate, a reflective layer on the spacer layer, and an electrical address circuit. (C) each pixel is (i) an electrostatic deflectable element formed in the reflective layer,
And (ii) a well formed in the spacer layer and below the deflectable element and an adjacent portion of the reflective layer, and (d) rotation for planarizing the spacer layer. A spatial light modulator that is a deposition material.

(7) (a) having a plurality of pixels formed in a layered structure, (b) the layered structure having a substrate, a spacer layer on the substrate, a reflective layer on the spacer layer, and an electrical address circuit. (C) each pixel includes (i) an electrostatic deflectable element formed in the reflective layer,
And (ii) formed in the spacer layer and including a well below the deflectable element and an adjacent portion of the reflective layer, (d) the addressing circuit for each pixel. , On the substrate at the bottom of the well,
A spatial light modulator including an electrode located away from a portion of the bottom closest to the deflectable element during electrostatic deflection.

(8) (a) having a plurality of pixels formed in a layered structure, (b) the layered structure having a substrate, a spacer layer on the substrate, a reflective layer on the spacer layer, and an electrical address circuit. (C) each of the pixels is (i) formed in the reflective layer and has a substantially rectangular shape, and the pixel of the reflective layer is formed at each of two diagonally opposite corners. An electrostatic deflectable element having a connection to the rest, and (ii) a well formed in the spacer layer below the deflectable element and an adjacent portion of the reflective layer. (D) the address circuit is on the substrate at the bottom of the well for each of the pixels,
A spatial light modulator including an electrode disposed below a portion of the deflectable element on one side of a diagonal line passing through the connection.

(9) (a) having a plurality of pixels formed in a layered structure, (b) the layered structure having a substrate, a spacer layer on the substrate, a reflective layer on the spacer layer, and an electrical address circuit. (C) each pixel includes: (i) an electrostatic deflectable element formed in the reflective layer,
And (ii) includes a well formed in the spacer layer and below the deflectable element and an adjacent portion of the reflective layer, (d) the bottom of the reflective layer and the bottom of the well. Spatial light modulators that are both electrical conductors.

(10) (a) having a plurality of pixels formed in a layered structure, (b) the layered structure having a substrate, a spacer layer on the substrate, a reflective layer on the spacer layer, and an electrical address circuit. (C) each pixel includes: (i) an electrostatic deflectable element formed in the reflective layer,
And (ii) includes a well formed in the spacer layer below the deflectable element and an adjacent portion of the reflective layer, and (d) the reflective layer comprises at least a first material. The first partial layer is provided on the second partial layer, and (e) the well is formed into the second partial layer after the well is formed.
A spatial light modulator including a deposit of a first material characterized by a substantially uniform deposition of the first material on a partial layer of.

(11) (a) having a plurality of pixels formed in a layered structure, (b) the layered structure having a substrate, a spacer layer on the substrate, a reflective layer on the spacer layer, and an electrical address circuit. (C) each pixel includes: (i) an electrostatic deflectable element formed in the reflective layer,
And (ii) a well formed in the spacer layer below the deflectable element and an adjacent portion of the reflective layer, and (d) the deflectable element being formed by a hinged portion. Characterized by a substantially square portion connected to the remaining portion of the reflective layer, the hinge portion is a spatial light whose length measured from the square portion to the remaining portion is larger than the width of the hinge portion. Modulator.

(12) In the method of making the spatial light modulator,
(a) forming an electrical address circuit on the substrate, (b) depositing a spacer layer on the substrate and on the circuit, (c) depositing a reflective layer on the spacer layer, and (d) ) Defining a pattern of the reflective layer to define a plurality of pixel elements connected to the rest of the reflective layer, (e)
The spacer layer may be plasma etched to form a well below the pixel element, thus freeing the pixel element from the spacer layer and electrostatically deflecting the pixel element in response to a signal from the address circuit. A method comprising the steps of:

(13) In the method of making the spatial light modulator,
(a) Deposit a spacer layer on the conductive substrate, and (b)
Depositing a reflective layer on the spacer layer, (c) defining a pattern of the reflective layer to define a plurality of pixel elements and electrodes, and (e) plasma etching the spacer layer to form the pixel element. Forming a well underneath, thus releasing the pixel element from the spacer layer and allowing the pixel element to be electrostatically deflected by a signal applied between the electrode and the substrate.

(14) In the method of making the spatial light modulator,
(a) deposit a spacer on a substrate having an electric address circuit, (b) define a pattern of pillar holes reaching the substrate on the spacer, and (c) deposit a metal layer on the spacer and on the pillar holes. , (D) defining a pattern of the metal layer to define deflectable elements, each element being hingedly bonded to the metal deposited in one of the pillar holes, (e)
A method comprising the step of removing the spacer.

(15) In the method described in Item 14, the removal is performed by plasma etching.

(16) In the method of making the spatial light modulator,
(a) defining a deflectable beam and a support structure in a combination of layers of reflective material and spacer, said reflective material being on said spacer on a substrate, (b) protecting on said limited beam Applying a layer, (c) dicing the substrate and layer into chips, each chip being a spatial light modulator, (d) alternately applying a solvent to the protective layer and forming the chip. By rotating, of the protective layer, to remove the portion softened or dissolved by the solvent, (e)
Removing the at least a portion of the spacer to release the deflectable beam from the spacer.

(17) In a method of monitoring the plasma etch undercut of a spacer between a substrate and an upper layer, (a) defining a pattern of a substantially rectangular region in the upper layer, the rectangular region being Having a sequence of known width, (b) periodically observe the area under light, (c) undercut,
Determining the width of the widest rectangular region whose reflectivity for the light has changed to be about half the width.

(18) In a method of manufacturing a spatial light modulator having a deflectable beam, (a) a spacer is applied on a substrate having an electronic address circuit, and (b) a pattern of the spacer is defined. To limit the pillar holes reaching the substrate by (c) depositing a metal layer on the spacer and on the pillar holes, (d)
Limiting the deflectable beams to the metal so that each beam is hingedly bonded to the metal in one post, and (e) removing the spacer.

(19) In the method described in Item 18, the removing is performed by plasma etching.

(20) In a method of manufacturing a spatial light modulator having a deflectable beam, (a) a spacer is rotationally attached on a substrate having an electronic address circuit, and (b) the spacer pattern is formed. Define the columnar holes reaching the substrate, (c)
Deposit metal on top of the spacers and in the pillar holes, (d)
Removing the metal from the spacers except near the pillar holes, (e) applying a second metal layer on the spacers and on the metallized pillar holes, (f) in the second metal layer Limiting the deflectable beam to (g) and removing the spacer (g).

(21) (a) A plurality of pixels are provided on a substrate having an electronic address circuit, (b) each pixel is (i) a reflection element capable of electrostatic deflection, (ii) the element And (iii) the element and the support are both electric conductors, and the support is electrically connected to the address circuit.

(22) (a) A plurality of pixels are provided on a substrate having an electronic address circuit, and (b) each pixel is 1 pixel.
A metal piece, the metal piece including a support portion attached to the substrate, a hinge portion connected to the support portion, and a deflection portion connected to the hinge portion and held away from the substrate. A spatial light modulator having a possible portion, and (c) the supporting portion is electrically connected to the address circuit and the deflectable portion is above an electrode connected to the address circuit.

(23) In the spatial light modulator described in the paragraph 22, (a) the one metal piece has the shape of four square flaps, and each flap is with respect to the central column. Spatial light modulator connected at the corner.

(24) In the spatial light modulator described in paragraph 22, (a) a space in which the electrode extends only under a portion of the deflectable portion near the support. Light modulator.

(25) (a) a plurality of pixels on a substrate having an electronic address circuit, (b) each pixel includes a pillar and a hinged flap, (c) A spatial light modulator in which the pillar has a first conductive portion attached to the substrate and a metal portion adjacent and overlapping the hinged flap.

[Brief description of drawings]

FIG. 1 is a diagram showing an SLM pixel of a first preferred embodiment, in which A is a simplified perspective view, B is a side sectional view, and C is a plan view.

FIG. 2 is a diagram illustrating an operation of the pixel of FIG.

FIG. 3 is a graph showing a response curve of the pixel of FIG.

4 is a diagram showing a plasma etch step of the method of the first preferred embodiment for manufacturing the pixel of FIG.

5 is a hollow-beam plasma for the pixel of FIG.
The figure which shows control of the range of etching.

6A and 6B are views showing the operation of the hollowed-out beam in FIG.

FIG. 7 is a diagram showing various shapes of a beam used for the pixel of FIG.

FIG. 8 is a simplified side sectional view of a modification of the pixel of FIG.

FIG. 9 is a simplified plan view showing three adjacent pixels in a linear array SLM.

FIG. 10 shows a pixel of the second preferred embodiment, where A is a simplified side sectional view and B is a simplified side sectional view of a modification of the pixel of A.

FIG. 11 is a series of simplified side cross-sectional views showing the steps of the method of the second preferred embodiment.

FIG. 12 shows a pixel of a third preferred embodiment, where A is a simplified sectional view and B is a plan view.

FIG. 13 shows a pixel of a fourth preferred embodiment, where A is a simplified sectional view and B is a plan view.

FIG. 14 shows a pixel of the fifth preferred embodiment, where A is a simplified cutaway perspective view, B is a cross-sectional view and C is a plan view.

FIG. 15 is a cross-sectional view of the pixel of FIG. 14 illustrating the flexure of a beam.

16 is a graph showing a response curve for the pixel of FIG.

FIG. 17 is a diagram showing a manufacturing process of the pixel of FIG.

FIG. 18 is a cross-sectional view of a modification of the pixel of FIG.

FIG. 19 shows process steps for the deformation of the pixel of FIG.

20 illustrates process steps for another variation of the pixel of FIG.

FIG. 21 shows process steps for yet another modification of the pixel of FIG.

FIG. 22 shows a pixel of a sixth preferred embodiment, where A is a simplified perspective view, B is a sectional view and C is a plan view.

FIG. 23 is a diagram showing process steps for manufacturing the pixel of FIG. 22.

FIG. 24 is a diagram showing a modification of the pixel of FIG. 22.

FIG. 25 is a diagram showing a modification of the pixel of FIG. 22.

FIG. 26 is a diagram showing a pixel of a seventh preferred embodiment.

FIG. 27 is a diagram showing modifications applicable to the pixels of the fifth, sixth, and seventh preferred embodiments.

[Explanation of symbols]

 20 pixels 22 substrate 24 spacer 26 reflective layer 28 flap

Claims (1)

[Claims] 1. A method for producing a spatial light modulator, conductive
Forming a conductive material made of a support pillar onto the substrate, the spacer layer is deposited on the substrate and the support pillars, and depositing a reflective layer on the spacer layer except for the upper portion of the support column, wherein Pattern the reflective layer to remove plasma
Forming a recess, removing the spacer layer above the support column to expose the head of the support column , connecting the head of the support column and the reflective layer with a beam material ,
The spacer is formed by etching through the access hole.
A method of manufacturing a spatial light modulator comprising the step of removing a layer .