Multi-level deformable mirror device

TECHNICAL FIELD OF THE INVENTION

0001This invention relates to deformable mirror devices of the type disclosed in EP-A-0 332 953 or US-A-4 710 732 and more particularly to such devices where the torsion hinges are constructed in a plane different from the reflective surface.

BACKGROUND OF THE INVENTION

0002Deformable mirror devices are finding many uses in light deflecting technologies. In operation, the devices are akin to small mirrors which rotate about a fixed axis. The rotation causes light to be deflected under control of the rotation. Thus, in some applications an array of such DMD's are positioned such that as the individual DMD devices, called pixels, are selectively rotated, patterns can be created for various purposes.

0003The DMD devices are used, typically, in a dark field projection optic arrangement and can be used, for example, in a HDTV application where a large array of pixels are necessary for the desired clarity. These large pixel arrays require close packing of the individual DMD pixels and a high modulation efficiency to maximize brightness as well as low light scattering to maximize contrast.

0004Light scattering is controlled, in part, by the amount of non-rotatable (dead) surface area exposed to the light. This dead area arises because it is used to provide the torsional hinges which support the rotatable DMD pixels. One such arrangement of such torsional hinges and the construction thereof is shown in U.S. Patent No. 4,662,746, issued to me on May 5, 1987.

0005Thus, there exists in the art a need for a DMD matrix which couples high density with high modulation efficiency and which has low inherent light scattering.

0006There is a further need in the art for such a structure which maximizes the light reflective area for a given density of DMD pixels.

0007There is a further need in the art for a DMD architecture which will support HDTV applications.

SUMMARY OF THE INVENTION

0008The spatial light modulator device of the invention as defined in claim 1 has been constructed with the torsion hinges under the rotatable mirror surface at a location within the substrate, for example, between the address electronics and the surface, such that the "dead" area of the surface is significantly reduced. Using this structure, then, a DMD device can be constructed following the method of the invention as defined in claim 7, the structure having several technical advantages. First, this structure achieves lower light scattering (higher contrast), because the light scattering caused by the torsional hinges is eliminated. Another technical advantage of this structure is the fact that the surface area occupied by the rotatable mirror is increased, leading to higher modulation efficiency (higher brightness). A further technical advantage of this invention is the creation of several different configurations of pixel elements, all of which are controllably deflective, and which are on a reflective surface of a substrate with the mechanically linked rotational supports contained in a surface other than the reflective surface.

0009The multilayered mirror surface and hinge are constructed in several steps by a combination of spin depositing, sputter depositing, coupled with several plasma etching and photolithography steps.

BRIEF DESCRIPTION OF THE DRAWINGS

0010For a more complete understanding of the present invention, and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which: <ul id="ul0001" list-style="none" compact="compact"><li>FIGURE 1 is an array of DMD pixels having the prior art architecture;</li><li>FIGURE 2 is an array of DMD pixels having the multi-level hinge architecture;</li><li>FIGURE 3 shows in broken line format the underlying hinge, electrodes and supports for the new architecture;</li><li>FIGURE 4 shows the layout of the hinge, electrodes and support mechanism of the underlying layers;</li><li>FIGURES 5a and 5b show sectional side views of the multi-layered device taken along section 5a-5a and 5b-5b, of FIGURE 4, respectively;</li><li>FIGURE 6 shows the operation of the beam in two states under control of address electrodes;</li><li>FIGURES 7a-7d snow the steps in the process of manufacturing the multi-layered device; and</li><li>FIGURES 8a-8d show alternate physical configurations of the multi-layered device.</li></ul>

DETAILED DESCRIPTION OF THE INVENTION

0011FIGURE 1 illustrates a conventional prior art 45-degree bistable torsion beam array 10, with a 12 micron pixel pitch. Each pixel 100 is supported for rotational purposes by hinges 101, supported on the substrate by supports 102.

0012The prior art arrangement is to be compared with the structure shown in FIGURE 2 showing a hidden hinge array 20, also having a 12 micron pitch. When the hidden hinge architecture is viewed from the top, as in FIGURE 2, only the square torsion beam reflective surface 200 of each pixel and the beam support post 201 are visible. A comparison of element 100 (FIGURE 1) with element 200 (FIGURE 2) clearly shows that the hidden hinge provides a greater rotatable reflective surface for a given pixel size.

0013The underlying hinges, address and landing electrodes are shown as dotted lines in FIGURE 3. Beam support post 201 rigidly connects beam 200 to underlying torsion hinge 401. Details of the underlying hinge and electrodes are shown in FIGURE 4. Beam support post 201 allows beam 200 to rotate under control of hinges 401 which in turn are connected to posts 406. This allows rotatable surface (beam) 200 to rotate under control of an electrode supported by posts 403. Beam 200 lands in contact with landing electrode 405. Contact 402 extends through the substrate and is in contact with the underlying address electronics. The construction and operation of this device will be discussed hereinafter.

0014Cross sections along the hinge are shown in FIGURES 5a and 5b. FIGURE 5a is a cross-section along the hinge showing hinge support post 406, hinge 401, and beam 200. FIGURE 5b is a cross-section taken at right angles to hinge 401 and shows address electrode 404 and electrode support post 403. Also shown are protective oxide layer 501, METAL2 layer 502, and CMOS circuit and substrate layer 503.

0015FIGURE 6 illustrates beam 200 rotation 200a to landing angle -θ_L, and rotation 200b to landing angle +θ_L. Also shown are address electrodes 404 which control the movement (200a, 200b) and landing electrodes 405 positioned at the other end of the see-saw swing of beam 200. By way of example it might be assumed that the address voltage applied to the beam is OV or +5V and the required differential bias to achieve bistability is 15V. The negative bias electrode is then biased at -15V and the positive bias electrode is set at +20V.

0016When the beam is addressed to OV, there is a +15 volt potential difference between the beam and the negative bias electrode and a -20 volt difference between the beam and the positive bias electrode. The torque produced by the bias electrodes depends only on the absolute value of the potential difference. Because of the five volt larger potential difference between the beam and the positive bias electrode, the beam rotates to the positive landing angle.

0017When the beam is addressed to +5V, there is a +20 volt potential difference between the beam and the negative bias electrode and a -15 volt difference between the beam and the positive bias electrode. As a result, the beam rotates to the negative landing angle.

0018Oxide pads are placed at the tips of bias electrodes to prevent electrical discharge between the beam and the electrode when the beam lands. Conventional landing electrodes can also be employed. However, the landing electrodes cannot be connected in common, as they are in the prior art. For this proposed driven beam address scheme the beams are electrically isolated and the landing electrodes between neighboring pixels must also be electrically isolated and connected to their respective beams. This bussing complication between the beams and their respective landing electrodes makes an oxide landing pad a more attractive approach.

0019Resonant reset is accomplished in the usual fashion except the reset pulses are applied to the two bias electrodes.

0020By placing hinge 401, as shown in FIGURE 4, between the address electrodes 404, less space is available for the address electrodes. However, this has a negligible effect on the address voltage requirement. The area that is lost to the address electrodes has a near zero moment arm, and therefore, the torque exerted by the address electrodes is only slightly reduced.

0021The hidden hinge architecture is a two spacer process consisting of the hinge spacer and the beam spacer. The hinge spacer is a thin (∼0.5 µm) layer. When it is removed, the resulting air gap allows the hinge to rotate freely. The beam spacer is thicker (∼1.5 µm) and determines the final angular rotation of beam 200. The thin hinge and thicker electrode are formed with a single plasma aluminum etch using the buried hinge process as discussed in above-identified U.S. Patent Number 4,662,746.

0022The process sequence for the hidden hinge architecture is shown in FIGURES 7a-7d and consists of five layers (hinge spacer, hinge, electrode, beam spacer, and beam). This is to be compared with the conventional four layer process (electrode, spacer, hinge, beam).

0023The process begins with a completed address circuit 503 including contact openings formed in protective oxide 501 of the address circuit. The address circuit is typically a two metal layer/poly CMOS process. The contact openings allow access to the second level metal (METL2) 502 bond pads and to the METL2 address circuit output nodes.

0024Hinge spacer 701 is spin-deposited over the address circuit and patterned with holes 702 that will form the hinge support posts and electrode support posts and contacts. This spacer is typically 0.5 µm thick and is a positive photoresistant deep UV hardened to a temperature of 200°C to prevent flow and bubbling during subsequent processing steps.

0025As shown in FIGURE 7b, the next two layers 703 and 704 are formed by the so-called buried hinge process. An aluminum alloy that forms the hinge is sputter-deposited onto the hinge spacer. This alloy is typically 75 nm (750 Angstroms) thick and consists of 0.2% Ti, 1% Si and the remainder Al. A masking oxide is plasma-deposited and patterned in the shape of hinges 401. This hinge oxide is then buried by a second aluminum alloy layer 704 that is to form the electrode (typically 300 nm (3000 Angstroms) thick).

0026Next, a masking oxide is plasma-deposited and patterned in the shape of the electrodes 404, the electrode support posts 406 and the beam contact metal 405.

0027Finally, a single plasma aluminum etch is used to pattern the hinges, electrodes, support posts and beam contact metal. The electrode metal overlying the hinge region is etched away, exposing the buried-hinge oxide which acts as an etch stop. When the plasma aluminum etch is complete, regions of thin hinge metal 703 and thick electrode metal 704 have been simultaneously patterned. The masking oxide is then removed by a plasma etch.

0028Next as shown in FIGURE 7c, beam spacer 705 is spin-deposited over the hinges and electrodes and patterned with holes that will form beam support posts 201. Spacer 705 determines the torsion beam angular deflection and is typically 1.5 microns thick and is a positive photoresistant. It is deep UV hardened to a temperature of 180°C to prevent flow and bubbling during subsequent processing steps. Note that no degradation of hinge spacer 701 occurs during this bake, because the hinge spacer was hardened to a higher temperature (200°C).

0029Finally, an aluminum alloy that is to form beam 200 (typically 400 nm (4000 Angstroms)thick) is sputter deposited onto beam spacer 705. Next, masking oxide 707 is plasma-deposited and patterned in the shape of the beams. The beam is then plasma etched to form the beams and beam support posts.

0030This completes the process at the wafer level. Masking oxide 707 on beam 200 is left in place. The wafers are then coated with PMMA, sawed into chip arrays and pulse spin-cleaned with chlorobenzene. Finally, the chips are placed in a plasma etching chamber, where masking oxide 707 is removed and both spacer layers 701 and 705 are completely removed to form the air gaps under the hinges and beams as shown in FIGURE 7d.

0031Determination of the necessary amount of undercut to free the hinge is made by the use of modified dropout structures. Dropout structures were first disclosed in U.S. Patent Number 4,566,935 which is hereby incorporated by reference herein. The modified dropout structure consists of the structure of FIGURE 5a of this application, except the hinge is not connected to either hinge support post. When beam spacer 705 has been removed and a sufficient amount of hinge spacer 701 removed to free the hinge for rotation, beam 200 will drop to one side, with its edge resting on one of the electrodes. Alternately, correlation of the amount of hinge and electrode undercut can be made by using conventional dropout structures and removing the metal layers to observe the amount of residual spacer.

0032The hidden hinge architecture has several advantages, a few of which are obvious, and some, more subtle. First, as mentioned previously, by placing the hinges under the beam, the surface area occupied by the rotatable mirror is increased, leading to a greater modulation efficiency and hence a brighter display. Second, the hidden hinge is shielded from the projector light so it cannot scatter light into the pupil of the projection lens. Thus, the contrast ratio is increased. Third, for the same size pixel, the hidden hinge architecture allows torsion hinges to be made longer without reducing brightness and contrast, thus increasing the compliance and lowering the address voltage requirement.

0033Fourth, because the hinge spacer planarizes the underlying address circuit, no special planarization of the final address circuit metalization is required, (e.g. resist etch-back planarization of the protective oxide). Planarization of the underlying topography is necessary prior to the anisotropic hinge/electrode etch to prevent residual metal filaments in the electrode metal and to planarize the aluminum hillocks that have formed in METAL2.

0034Fifth, the landing electrodes are air isolated from the address circuit, thus reducing feed-through of the high-voltage reset pulse to the underlying address circuit. This coupling, if uncontrolled, can produce voltage spikes on the gates of the address transistors causing the transistors to momentarily turn on at the wrong time. Also, this air isolation minimizes the possibility of dielectric breakdown through the protective oxide.

0035Sixth, because the address electrodes are air isolated, they have reduced capacitance. It is therefore possible with the hidden hinge architecture to build a frame-addressed DMD (U.S. Patent Number 4,615,595) where the address electrode rather than the beam is charge addressed. This would permit the fabrication of frame addressed torsion beam architectures without the need to electrically isolate the beams.

0036The hidden hinge has been illustrated for the 45 degree pixel architecture, but other pixel architectures are possible, as shown in FIGURES 8a-8d by changing the architecture of the hinge. When the hidden hinge is used for the flexure beam pixel, (FIGURE 8c) beam support post 201 moves vertically through a distance of λ/4 (typically 150-200 nm (1500 to 2000 Angstroms)). The 500 nm (5000 Angstrom) hinge spacer air gap more than accommodates this vertical travel. The flexure beam pixel can be used as a phase-only modulator. The effect of hiding the hinges is to reduce the nonactive background which reduces the amplitude modulation component and makes the flexure beam closer to an ideal phase-only modulator.

0037Although this description describes the invention with reference to the above specified embodiments, the claims and not this description define the scope of the invention. Various modifications of the disclosed embodiment, as well as alternative embodiments of the invention, will become apparent to persons skilled in the art upon reference to the above description. Therefore, the appended claims will cover the modifications that fall within the scope of the invention as defined in the claims.