WO2002005008A1 - Microelectromechanical deformable grating for binary optical switching - Google Patents

Abstract

A deformable grating switch having low actuation voltage and high bandwidth is disclosed. The grating device is made of metal and has alternately addressed diffracting bars (411) suspended above and below patterned electrodes (405, 407). The actuated grating efficiently diffracts light into the first diffraction order having large angle with respect to the zero order. In this manner, high extinction between 'on' and 'off' signals is realized. Instead of relying on a stiff hinge or spring to provide the restoring force, high bandwidth is achieved by actively driving alternate grating bars (411) up or down by a distance equal to 1/8 wavelength of the incident light.

Description

MICROELECTROMECHANICAL DEFORMABLE GRATING FOR BINARY

OPTICAL SWITCHING

Related Applications

This application incorporates (1) U.S. Patent Application No. 60/183,793, filed on February 22, 2000, entitled "Point Probe Memory With Light Modulator Readout," naming Dr. Charles Hester and Charles Whitehead as inventors, which application is incorporated entirely herein by reference; (2) U.S. Provisional Patent Application No. 60/142,931, entitled "Analog Compressive Network," naming Charles F. Hester and Marshal K. Quick as inventors, filed on July 9, 1999, which application, along with the entirety of its attached Appendix, is incorporated entirely herein by reference; (3) the U.S. Patent Application entitled "Deformable Grating Modulator Capable Of Both Phase And Amplitude Modulation," naming Charles F. Hester as inventor, filed on July 10, 2000, which application is incorporated entirely herein by reference; (4) the U.S. Patent Application entitled "Grating Type Spatial Light Modulators And Method Of Manufacturing Gratmg Type Spatial Light Modulators," naming Charles F. Hester and Charles A. Whitehead as inventors, filed on July 10, 2000, which application is incorporated entirely herein by reference; and (5) the U.S. patent application entitled "Adaptive Compressive Network," naming Charles F. Hester and Marshall K. Quick as inventors, filed on July 10, 2000.

Background Of The Invention

Field Of The Invention The present invention relates to digital microelectromechanical optical switches having high modulation rates and requiring low actuation voltages. Such switches can be used for the routing of fiber-optic signals or arrayed to form spatial light modulators or display projectors.

Description Of The Prior Art The spatial light modulator is a frequently used component in optical processing and computing. It is a structure that allows a light beam to be controllably spatially (as opposed to frequency) modulated. There are many different types of programmable spatial light modulators, including, for example, modulators employing liquid crystal devices, magneto optical materials, and acousto optical materials. Various types of spatial light modulators are described in "Two- Dimensional Spatial Light Modulators: A Tutorial" by John A. Neff et al., Proceedings of the IEEE, Vol. 78, No. 5 (May 1990), pages 826-855, which is incorporated entirely herein by reference.

Spatial light modulators are used for a variety of purposes, including inputting information into an optical system, modulating carrier light during an optical computation, such as in a Fourier filter, and as a neural network interconnect. Spatial light modulators are particularly useful as digital optical switches. Digital optical switches are employed in, for example, fiber-optic routing networks, printer heads, and displays.

One category of spatial light modulators used as digital optical switches is the micromechanical mirror devices, such as the digital micromirror device (DMD™) of Texas Instruments. These devices electrostatically actuate one or more miniaturized mirrors to switch light from being reflected at one angle to being reflected at another angle. Microelectromechanical switches are easily manufactured using conventional semiconductor techniques, can be very small (several tens of microns), and have very simple moving parts. They can operate in either digital or analog fashion, and typically have a hinge or actuator that is electrostatically or piezoelectrically driven.

One application of microelectromechanical switches to a spatial light modulator is in a two-dimensional grating type spatial light modulator, which uses a set of parallel reflecting bars that form a grating. To provide a better understanding of the features of the invention, one conventional type of spatial modulator will now be described in detail. Referring to Figs. 1-3, this type of modulator 101 has a number of reflective parallel bars 103 suspended above a reflective surface 105. Each bar 103 is supported on either end by a support post 107. The modulator 101 also has one or more electrodes 201 corresponding to the bars 103.

Referring now to Fig. 2, when the electrodes 201 are inactive, the distance d between the bars 103 and the reflective surface 105 is λ/2 (or an integral multiple of λ/2), where λ/2 is the wavelength of the light to be modulated. Thus, the light 203 reflected from the upper reflective surface of the bars 103 is in phase with the light 205 reflected from the reflective surface 105, so the modulator 101 acts as a mirror. This is the "off' state of the modulator 101. To modulate light, a charge is applied to the bars 103, and an opposite charge is applied to the electrodes 201. The charge on the electrodes 201 attracts the charge on the reflective bars 103, and pulls all of the reflective bars 103 (i.e., the micromirrors) toward the reflective surface 105, as can be seen in Fig. 3. If all of the reflective bars 103 are pulled to a distance d of λ/4 (or an integral multiple of λ/4) from the reflective surface 105, then the light 203 reflected from the reflective surface of the bars 103 is opposite in phase from the light 305 reflected from the reflective surface 105. Thus, the destructive interference between light 203 and light 205 then prevents the modulator 101 from reflecting light to any distance. This is the "on" state of the spatial light modulator 101.

As previously noted, each bar 103 is supported at either end by a support post 107. Accordingly, the end portions of the bars 103 serve as hinges and flex in order to allow the height of the center portion of the bar 103 to change. While these end portions must flex, however, they must also be stiff enough to support the reflective bar 103. Accordingly, both the bars 103 and the electrodes 201 must carry a large amount of electric charge in order to sufficiently flex the bars 103.

Those of ordinary skill in the art will understand that, in order to achieve high extinction coefficients between the on and off states of the modulator, these types of micromirrors require high tilt angles and large actuation voltages. Moreover, in order to achieve stable, predictable response, the micromirrors are usually driven at a frequency below their first natural mechanical resonant frequency. Because the resonant frequency is proportional to the square root of the spring constant of the micromechanical hinge or actuator, a stiffer hinge (i.e., a hinge with a larger spring constant) can be used to increase the bandwidth of the switch. This increase in operational bandwidth comes, however, at the expense of higher actuation voltages.

Another limitation with conventional micromirror devices is stiction. The separation between moving and static components in a microelectromechanical switch is usually more than an order of magnitude smaller than the overall switch dimensions. Intentional or unintentional contact between the moving and static parts can lead to an undesirable sticking force between the parts (i.e., stiction) that often renders the switch inoperable. Not only do the static parts inherently stick to the static parts, but the contact short circuits the two, preventing opposite charges from being applied to the switch and electrode to operate the switch. Various methods employed to overcome the problem of stiction include the use of anti-sticking flourinated monolayer coatings and fabrication techniques designed to minimize the surface areas that come into contact with each other. Also, insulative layers are applied to the portions of the switch, the electrode, or both to prevent a short circuit between the two. These methods add non-standard manufacturing steps to the foundry process, however.

In view of these limitations with conventional micromirror devices, it would be desirable to have a microelectromechanical switch that can achieve a high operational bandwidth with a low actuation voltage, and which overcomes the problem of stiction.

Summary Of The Invention

Accordingly, the invention is directed to a micromechanical switch that can be embodied to provide a high operational bandwidth while requiring only a low actuation voltage. Various embodiments of the invention also overcome the problems associated with stiction. Instead, various embodiments according to the invention can advantageously employ stiction to provide a micromechanical switch with a high operational bandwidth and low actuation voltage.

In one preferred embodiment of the invention, a deformable grating is fabricated using a process that forms three conductive layers, such as layers of metal or indium tin oxide (ITO). The upper and lower conductive layers are patterned to form electrodes, while the middle metal layer is patterned to form optically reflective grating bars. In addition, a sacrificial layer is formed between each metal layer. To complete the device, portions of both sacrificial layers are removed, severing all connections to the bars. The upper and lower electrodes then control the position of each independently movable bar.

Without hinge connections, the operational bandwidth of the reflective bars is limited only by the switching frequency of the electrodes. Moreover, if a reflective bar sticks to an upper electrode, it can easily be displaced by the pull from a lower electrode. Similarly, if a reflective bar sticks to a lower electrode, it can easily be displaced by the pull from an upper electrode. This, a micromechanical mirror device according to the invention can overcome the problem of stiction associated with conventional micromechanical mirror devices.

Brief Description Of The Drawings

Fig. 1 is a top view of a conventional micromechanical mirror type spatial light modulator.

Fig. 2 is a cross-sectional view of the spatial light modulator shown in Fig. 1 taken along section line 2-2'. Fig. 3 is a cross-sectional view of the spatial light modulator shown in Fig. 1 taken along section line 3-3'.

Fig. 4 is a cross sectional view of a first embodiment of the invention in an "off state.

Fig. 5 is a cross sectional view of the first embodiment of the invention in an "on" state.

Fig. 6 is a top planar view of another embodiment of the invention.

Figs. 7-17 illustrate the cross-sectional side views of various steps in the construction of still another embodiments of the invention.

Detailed Description Of Preferred Embodiments

Figs 4-6 illustrate a first embodiment of a spatial light modulator according to the invention. As seen in these figures, the modulator 401 includes a substrate 403 with a series of pedestals 405 formed on the substrate. The modulator 401 also has a first set of lower electrodes 407 and a second set of lower electrodes 409. The lower electrodes 407 are electrically connected together, and are arranged on the substrate 403 such that an electrode 407 is positioned between each pair of adjacent pedestals 405. The lower electrodes 409 are also electrically connected together, and are arranged on the pedestals 405 such that an electrode 409 is positioned on each pedestal 405. In this fashion, electrodes 407 and 409 form alternating series of electrodes over the surface of the substrate 403.

As can be seen in Figs. 4 and 5, the modulator 401 further includes a reflective bar 411 positioned above each electrode 407 and 409. The reflective bars 411 are independently movable. That is, they are not physically attached to any other structure of the modulator 401, and thus can move freely in any direction. The modulator 401 also has a first set of upper electrodes 413 and a second set of upper electrodes 415. The electrodes 413 are electrically connected together and are arranged such that each electrode 413 is directly above a bar 41 IB and an electrode 407. Likewise, the electrodes 415 are electrically connected together and are positioned such that each electrode 415 is directly above a bar 411A and an electrode 409. Both sets of electrodes 413 and 415 can be formed of a transparent conductive material, such as indium tin oxide (ITO).

With the conventional microelectromechanical switch (see, e.g., U.S. Patent 5,311,360 to Bloom et al., which patent is incorporated entirely herein by reference), the hinge of the microelectromechanical switch serves the dual purposes of providing charge for electrostatic actuation and restoring the switch to its equilibrium position once the actuation voltage is released. As will be appreciated from the above description, however, the micromechanical switch (i.e., the bar 411) according to the invention has no hinge, and is therefore free to move until it contacts either its upper electrode (413 or 415) or its lower electrode (407 or 409). Thus, each independently movable bar 411 has an electrode directly underneath it and an electrode directly above it.

Once a bar 411 contacts either of its associated electrodes, stiction holds the bar 411 in place. The charge on the contacting electrode can then be removed without any loss in diffraction efficiency (i.e., without movement of the bar 411). To subsequently move the bar 411 away from the contacting electrode, a charge is supplied to the bar 411 through the contacting electrode. A same charge on both the contacting electrode and the bar 411 repels the bar 411 from the contacting electrode, thus overcoming the stiction and releasing the bar 411 for travel. If an opposite charge is provided to the noncontacting electrode, this opposite charge will attract the bar 411 to the noncontacting electrode. With no hinge to slow it down, the bar 411 moves rapidly toward the noncontacting electrode until it contacts the electrode. The bar's charge is again bled off and subsequently refreshed when return to its complementary position is required. Of course, those of ordinary skill in the art will appreciate that, with some embodiments of the invention, a bar 411 may be suspended between its upper and lower electrodes by balancing the charge difference between the electrodes.

In conventional microelectromechanical switches, a current source must be constantly employed to maintain the position of the switch. Thus, intimate contact between the charged conductive material of the switch and oppositely charged conductive material of an electrode will destroy the switch by shorting the power supply. With these switches, then, shorting is typically prevented with a dielectric layer insulating the areas of possible contact between the switch and electrodes. While this solution prevents shorting, it also requires a larger actuating voltage for a comparable actuating force.

With the switch according to the invention, the bars 411 are connected to a current source. Accordingly, contact between a bar 411 and either of its associated electrodes will short circuit a power supply, obviating the need for an insulating layer between the bar 411 and its electrodes. This reduces the actuation charge needed to move the bar 411. Further, because the bar 411 is being repelled by one electrode while it is being attracted by the opposite electrode, the actuation voltage required for either electrode is lower than the actuation voltage required with a conventional microelectromechanical switch. Still further, because a hinge does not limit the movement of each bar 411, the operational speed of the switch according to the invention is limited only by the speed at which charge can be applied to and removed from the electrodes and the bar's inertia, which is extremely small.

Referring back now to the embodiment of the invention shown in Fig. 4, each of the pedestals 405 are formed from the same layer of material to a thickness (in the direction perpendicular to the substrate surface) of 1/8 the wavelength λ of the light to be modulated by the modulator (i.e., l/8λ). Each of the electrodes 407 and 409 are then also formed from a same layer of conductive material so as to each have a thickness of l/8λ in the direction perpendicular to the substrate surface. Further, each of bars 411 is formed from a same layer of reflective material to have a thickness of l/8λ in the direction perpendicular to the substrate surface. The upper electrodes 413 and 415 are then also formed from a same layer of conductive material to have the same thickness of l/8λ in the direction perpendicular to the substrate surface.

Moreover, each electrode 413 is positioned above its corresponding lower electrode 407 such that the intermediate bar 41 IB travels at least l/8λ from resting on the upper surface of the lower electrode 407 to contact the lower surface of the upper electrode 413. Similarly, each electrode 415 is positioned above its corresponding lower electrode 409 such that the intermediate bar 411 A travels at least l/8λ from resting on the upper surface of the lower electrode 409 to contact the lower surface of the upper electrode 415.

With this arrangement, when the bars 411 A contact their lower electrodes 409 and the bars 41 IB contact their upper electrodes 413, the upper surfaces of the bars 411 A and the upper surfaces of the bars 41 IB are substantially coplanar. In this "off state, the bars 411 appear to form a single reflective surface, thereby reflecting the incident light at the zeroth order. On the other hand, when the bars 411 A contact their upper electrodes 415 and the bars 41 IB contact their lower electrodes 407, the upper surfaces of the bars 411 A are ^lλλ higher than the upper surfaces of the bars 41 IB. Thus, in this "on" state, incident light is reflected off at the first order. As the grating pitch (i.e., the spacing between the bars 411 A and 41 IB) of the bars decreases, the angle between the diffracted orders increases. This increase in the angle of diffracted orders provides higher extinction ratios for the same modulation depth, unlike tilted micromirrors which require larger modulation distances. While the embodiment of the invention employs the component thicknesses described above, those of ordinary skill in the art will appreciate that many other variations in thickness and arrangement are possible. For example, rather than using pedestals 405, electrodes 407 may be formed in depressions in the substrate 403. Alternately, rather than using pedestals 405, the electrodes 403 may be formed with a thickness of Vzk instead of ^lAX. Preferably, however, the variations will still result in a configuration such that, when the bars 411 are in the "off state, the upper surfaces of the bars 41 IB are substantially coplanar with the upper surfaces of the bars 411 A, and in the "on" state, the upper surfaces of the bars 411A are VΛK higher (or an integral thereof) than the upper surfaces of the bars 41 IB.

With the above-described embodiment of the invention, the maximum mechanical motion necessary for any bar 411 to achieve modulation (i.e., a state transition between the "on" state and the "off state) is a distance of l/8λ, because alternating bars 411 A and 411 B are driven in opposite directions simultaneously. This is only one half the distance required for each bar in with a conventional grating type spatial light modulator, like that disclosed in U.S. Patent No. 5,311,360 to Bloom et al. Thus, a spatial light modulator according to the invention may provide an immediate doubling of the modulation rate over an equivalent conventional spatial light modulator. This arrangement also provides an advantage over conventional switches in that that its small full-scale deflection described above allows, for the same actuation voltage on a conventional switch, electrostatic forces larger by a factor of four (Force oc 1/d²) to overcome stiction and decrease actuation time. Further, as previously noted, additional increases in modulation rate are realized with the switch according to the invention due to the absence of a hinge.

Another embodiment of the invention will now be described with reference to Fig. 6. In this figure, those shown Figs. 1-4 will be identified with the same number. Thus, as seen in this figure, the spatial light modulator 401 includes a series of bars 411 A alternating with a series of bars 41 IB. Unlike the previously described embodiment, however, a single transparent electrode 415 or 413 does not cover the upper surface of each bar 411, respectively. Instead, only the ends of each bar 411 are covered by a pair of electrodes 415 or 413. In particular, each bar 411 A is covered only at one end by a first upper electrode 415 A, and is covered only at its opposite end by a second upper electrode 415B. A conductive line 601 then connects the electrodes 415A and 415B, so that they carry a balanced charge simultaneously. Similarly, each bar 41 IB is covered only at one end by a first upper electrode 413 A and at its opposite end by a second upper electrode 413B, with the electrodes 413A and 413B then being connected by a conductive line 603. (As shown in Figure 6, the line 601 connects all of the electrodes 415, while the line 603 connects all of the electrodes 613. As will be understood by those of ordinary skill in the art, however, other arrangements, such as an individual connection line for each pair of electrodes, are possible.) This embodiment of the invention allows the electrodes 413 and 415 to be formed from opaque material, rather than transparent material. In doing so, it obviates any need to account for the absorption effect of a transparent electrode material.

The construction of yet another embodiment of the invention will now be described with reference to Figs. 7-17. As shown in Fig. 7, a layer of insulative material 701 is formed on the base substrate 403. Preferably, the insulative layer 701 is formed of a thickness λ/8, where λ is the wavelength of light to be modulated by the modulator. Next, as shown in Fig. 8, the insulative layer 701 is patterned and etched to form a recess 801. The insulative layer 701 may be patterned and etched according to any conventional lithographic process known in the art.

Then, as shown in Fig. 9, a layer of conductive material 901 is formed over the substrate 403 and the remaining portions of the insulative layer 701. The layer 901 may be formed of metal, but it may also be formed of doped polysilicon, implanted silicon, or other suitable material. Preferably, the layer 901 has a thickness of λ/8 (measured in the direction perpendicular to the surface of substrate 403), but other thicknesses may be used as described in detail above. As shown in Fig. 10, the layer of conductive material 901 is then patterned and etched to form electrodes. Specifically, a portion of the conductive layer 901 is left remaining in the recess 801 to form an electrode 407, while a portion of the conductive layer 901 resting on the resistive layer 701 is left remaining to form an electrode 409.

As may be seen in Fig. 11, a sacrificial layer 1101 then is formed over the resistive layer 701, the exposed portion of the substrate 403, and the electrodes 407 and 409 in. As will be explained below, the entire sacrificial layer 1101 is removed to release the bars 411. Accordingly, this sacrificial layer 1101 is preferably formed of any material that can easily be removed in a subsequent process. For example, the sacrificial layer be formed from silicon dioxide, which can easily be removed using a wet chemical release process, or removed using an organic material, like polyimide, during a plasma release process. The sacrificial layer may conveniently have a thickness of λ/16. Next, as shown in Fig. 12, a second layer 1201 of conductive material is formed over the sacrificial layer 1101. Preferably, the layer 1201 of conductive material also is formed with a thickness of λ/8, and again may be formed from metal, doped polysilicon, implanted silicon, or other suitable material. The layer 1201 of conductive material is then patterned and etched to form the bars or switches. Specifically, as shown in Fig. 13, a portion of the layer 1201 is left remaining directly over electrode 407 to form bar 41 IB, while a portion of layer 1401 is left remaining directly over electrode 409 to form bar 411 A. Preferably, the spacing between the grating bars 411 should be kept to a minimum, to increase the overall reflectivity of the modulator. After the bars are formed, a second layer of sacrificial material 1401 is formed over the first layer of sacrificial material 1101 and the bars 41 la and 41 lb (see Fig. 14). Again, the sacrificial layer 1401 may conveniently have a thickness of λ/16, and may be formed of any material that may easily be removed in a subsequent process, such as silicon dioxide.

^' After the second sacrificial layer 1401 is formed, vias (not shown) are formed through sacrificial layers 1401 and 1101 to at least remaining portions of conductive layer 901 that do not form electrodes. Preferably, one or more of these vias may extend to the remaining portions of layer 701 or even to the substrate 403. These vias should not contact any of electrodes 407 or 409, however, or any of bars 411. As shown in Fig. 15, a third layer 1501 of conductive material is then formed over the second sacrificial layer 1401, such that the layer 1501 is anchored to the substrate through posts formed in the vias. The layer 1501 preferably has a thickness of λ/8, and may be formed from metal, doped polysilicon, implanted silicon, or other suitable material.

As shown in Fig. 16, the third layer 1501 of conductive material is patterned and etched to form electrodes. Specifically, the layer 1501 is patterned and etched to leave a portion of it remaining over bar 41 IB to form electrode 413. Similarly, a portion of layer 1501 is left remaining over bar 411 A to form electrode 415. The electrodes 413 and 415 remain anchored in position relative to the electrodes 407 and 409 by the posts formed in the vias. As previously noted, the third conductive layer 1501 may be formed of a transparent conductive material if the electrodes 407 and 409 are to extend over the entire length of the bars 411, and may be formed of an opaque conductive material if the electrodes 407 and 409 are to extend only partially over the bars 411.

Preferably, the layer 1501 is patterned and etched such that electrodes 413 and 415 include overhanging portions 413' and 415', as illustrated in Fig. 16. These overhanging portions will provide a sidewall to entrap the movable grating bars 411 A and 41 IB, and prevent wind currents or other external forces from removing the bars 411 from the modulator. As seen from Figs. 7-17, if the sacrificial layer 1401 is non-planarized, then the topology caused by electrodes 407 and 409 will create the sidewalls 413' and 415'. In the sacrificial layer 1401 is planarized, then this layer can be patterned and etched to form the topology to produce the sidewalls 413' and 415'.

Once the electrodes 413 and 415 have been formed, then the sacrificial layers 1101 and 1401 are removed to completely release the bars 411, as shown in Fig. 17. That is, the bars 411 will have no static contact with any other structure of the modulator. The position of the bars 411 can then be controlled by charges applied to the electrodes 407, 409, 413 and 415 and to the bars themselves through the electrodes, as explained in detail above.

According to still yet another embodiment of the invention, the bars 411 may be manufactured to have a permanent dipole moment. With this arrangement, the permanent moment can be used to actuate the switch without a charge transfer to the bars 411. Ferroelectric materials, for example, are excellent materials to provide a permanent dipole moment, and could be combined with a coating of a thin reflective layer to form the reflective bars 411 according to various alternate embodiments of the invention.

While the switch devices disclosed above have been described as applied to a spatial light modulator, those of ordinary skill in the art will appreciate that the switch devices may also be employed in a variety of other structures, particularly where high modulation speed and/or low actuation voltage is desired. Still further, while the switch device according to the invention has been described with reference to specific exemplary embodiments, it will be evident to those of ordinary skill in the art that various modifications and changes may be made to these embodiments without departing from the broader scope and spirit of the invention as set forth in the claims. For example, while the above-described embodiment is a digital switching spatial light modulator (i.e., switching between an "on" state and an "off state), various embodiments of the invention can be employed where the switch can be balanced at any position between the two opposing electrodes controlling the switch. Accordingly, this specification and the drawings are to be regarded in an illustrative rather than restrictive sense.

Claims

I claim:

1. A microelectromechanical switch device, comprising: a first electrode, a second electrode opposite the first electrode, and a hingeless switch positioned between the first electrode and the second electrode such that the switch contacts the first electrode in a first position and contacts the second electrode in a second position, but is not physically connected to either the first electrode or the second electrode.

2. The microelectromechanical switch device of claim 1, wherein the first electrode, the second electrode and the switch are formed of conductive materials.

3. The microelectromechanical switch device of claim 2, wherein the first electrode, the second electrode and the switch are formed of metal.

4. The microelectromechanical switch device of claim 1, wherein at least one of the electrodes has sidewalls to prevent the switch from moving from between the first electrode and the second electrode.

5. The microelectromechanical switch device of claim 1, wherein the switch is reflective.

6. The microelectromechanical switch device of claim 5, wherein the switch is formed of reflective material.

7. The microelectromechanical switch device of claim 1, wherein the switch includes a layer of reflective material along a surface of the switch.

8. A spatial light modulator comprising an array of switch devices as recited in claim 5.

9. The spatial light modulator of claim 5, wherein the array of switch devices includes a set of first switch devices and a set of second switch devices, the set of first switch devices and set of second switch devices being arranged such that a first switch device is positioned between each adjacent pair of second switch devices.

10. The spatial light modulator of claim 9, wherein the array of switch devices includes a set of first switch devices and a set of second switch devices, the set of first switch devices and set of second switch devices being arranged such that, when the switches of the first switch devices are in the first position and the switches of the second switch devices are in the second position, upper surfaces of the switches of the first switch devices are coplanar with upper surfaces of the switches of the second switch devices, and when the switches of the first switch devices are in the second position and the switches of the second switch devices are in the first position, the upper surfaces of the switches of the first switch devices have a height difference from the upper surfaces of the switches of the second switch devices of an odd integral of λ/4, where λ is a wavelength of light to be modulated by the spatial light modulator.