Silicon substrate as a light modulator sacrificial layer

Description

FIELD OF THE INVENTION

The present invention relates to (micro-electro mechanical) MEM devices. More particularly, the present invention relates to controlled etch processes for fabricating optical MEM devices.

BACKGROUND OF THE INVENTION

Optical MEM devices are used to modulate one or more wavelengths of light. Optical MEM devices can have applications in display, print and electrical device technologies. Examples of optical MEM devices which utilize suspended micro-ribbon structures to modulate light are disclosed in the U.S. Pat. Nos. 5,311,360, 5,841,579 and 5,808,797, all issued to Bloom et al., the contents of which are hereby incorporated by reference.

Briefly, an optical MEM device described in the above referenced patents has one or more sets of movable ribbons that comprise a support layer and a reflective top-layer. The support layer is preferably a silicon nitride layer and the reflective toplayer is preferably an aluminum layer. The ribbon structures are typically secured to a substrate through opposite ends of the ribbon, whereby center portions of the ribbons, referred to herein as the active portions, move up and down to modulate an incident light source.

The optical MEM devices described above, and methods for making the same, suffer from several shortcomings. Firstly, fabrication can involve numerous time consuming steps, raising the cost of the devices. Secondly, the dielectric materials typically used to fabricate these devices (silicon nitride and silicon oxide) result in devices which exhibit a propensity to charge. Charging can perturb or shift the switching bias voltage required to operate the devices possibly leading to unreliable or irreproducible optical signals. Thirdly, fabrication steps of these devices are often incompatible with materials and/or fabrication steps used to make integrated circuits. Accordingly, optical MEM devices and integrated circuits generally need to be fabricated separately and then laboriously combined in subsequent steps.

What is desired is an optical MEM device which exhibits reduced charging characteristics and which can be fabricated with fewer processing steps. Further what is desired is a method of fabricating optical MEM devices which can be compatible with standard (complementary metal oxide semiconductor) CMOS materials and/or (integrated circuit) IC wafer processing.

SUMMARY OF THE INVENTION

Embodiments of the invention are directed to a device comprising a substrate structure having a cavity and support regions. The device has one or more micro-structures suspended over the cavity coupled to the support regions of the substrate. The device is preferably an optical MEM device and the micro-structures are preferably ribbons configured to modulate light having a wavelength in a range of about 0.1 to about 2.0 microns. The ribbons preferably have lengths in a range of about 5 to about 2000 microns and widths in a range of about 0.5 to about 20 microns. The ribbons preferably comprise a layer of silicon nitride in a range of about 200 to about 2000 Angstroms thick and a reflective layer of aluminum in a range of about 200 to about 2000 Angstroms thick.

In accordance with the embodiments of the invention, the ribbons are formed by depositing a device layer directly onto a single crystal silicon wafer substrate. The device layer is then cut into ribbons with the preferred dimensions, and a cavity is formed in the silicon wafer under the center portions, or active portions, of the ribbons such that the ribbons remain attached at opposing ends to the wafer and are suspended over the cavity. The substrate can have any number of layers including silicon dioxide and silicon nitride layers, but preferably the sacrificial material is the single crystal silicon substrate that is either doped, undoped or otherwise patterned, as explained below. The cavity is preferably formed by a controlled etch process utilizing xenon difluoride.

Before etching the cavity into the silicon wafer substrate, the wafer can be modified to control the location, the direction and/or the rate of the etching process. In accordance with the embodiments of the invention, the single crystal wafer substrate is selectively modified in etching regions of the wafer and/or in the support regions of the wafer. The wafer is selectively modified by doping the etching regions and/or the support regions of the wafer with a dopant, such as boron, phosphorous and/or other suitable dopants. The wafer is selectively modified with one or more dopants using an ion implant process or other suitable process, whereby the chemical composition or structure of the wafer is modified.

In still further embodiments of the invention, access holes and/or access trenches are formed through the ribbons prior to etching the cavity in the wafer substrate. The access holes and/or access trenches are formed in order to increase the local etching rate of the silicon wafer under the center portions, or active portions, of the ribbons.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGS. 1a-b show cross-sectional representations of a micro-structure comprising a plurality of moveable ribbon structures, in accordance with the embodiments of the invention.

FIGS. 2a-b show cross-sectional representations of a micro-structure comprising two sets of ribbon structures, in accordance with the embodiments of the invention.

FIG. 3a shows a cross-sectional view of a ribbon structure, in accordance with the embodiments of the invention.

FIG. 3b shows a cross sectional view of a plurality of ribbons having the ribbon structure, shown in FIG. 3a, suspended over a substrate in accordance with the embodiments of the invention.

FIG. 4 illustrates a cross-sectional representation of a micro-device comprising a multi-layer ribbon structure,.in accordance with the embodiments of the invention.

FIGS. 5a-e illustrate steps of making a device comprising a plurality of ribbon structures are suspended over a cavity, in accordance with the embodiments of the invention.

FIGS. 6a-c illustrate steps for selectively modifying a silicon substrate, in accordance with the embodiments of the invention.

FIG. 7 is a block diagram outlining the steps of selectively etching a cavity into a silicon wafer, in accordance with the embodiments of the invention.

FIG. 8 is a schematic diagram of a chamber apparatus, or processing station, configured to etch a cavity into a silicon wafer, in accordance with the embodiments of the invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

Referring to FIG. 1a, an optical MEM device can have a plurality of movable ribbons 100 with reflective surfaces 104 spatially arranged over a substrate 102 and a plurality of reflective surfaces 105 spatially arranged on the substrate 102 between movable ribbons 100, such that adjacent reflective surfaces 104 and 105 modulate light, as described below. The surfaces 104 and 105 are made to be reflective by depositing a thin film of reflective material, such as silver or aluminum, on the ribbons 100 and substrate 102, respectively. The ribbons 100 and the substrate 102 are preferably formed from silicon-based materials such as silicon nitride. A height difference 103 between the reflective surfaces 104 and 105 is preferably configured to be λ/2 when the ribbons 100 are in an up position, as shown in FIG. 1a. When light having a wavelength λ impinges on the compliment of reflective surfaces 104 and 105, the light that is reflected from the surfaces 104 and 105 is in phase, thereby generating the conditions for constructive interference. Light which strikes the reflective surfaces 105 travels λ/2 further than the light striking the reflective surfaces 104 of the ribbons 100. The light that is reflected back from the reflective surfaces 105 travels an additional λ/2 for a total of one complete wavelength λ. Therefore, the compliment of the reflective surfaces 104 and 105 function as a mirror to the incident light source with a wavelength λ.

By applying an appropriate bias voltage across the ribbons 100 and the substrate 102, the ribbons 100 move towards the substrate 102, as shown in FIG. 1b. A thickness T_r of the ribbons 100 can be designed to be λ/4 such that the distance 103′ is also λ/4. When light having a wavelength λ impinges on surfaces 104′ and 105 with the ribbons 100 in a down position, as shown in FIG. 1b, the light reflected from the surfaces 104′ of the ribbons 100 will be out of phase with the light reflected from the surfaces 105, thereby generating conditions for destructive interference. By alternating the ribbons 100 between the positions for constructive interference, as shown in FIG. 1a, and the positions for destructive interference, as shown in FIG. 1b, the optical MEM device is capable of modulating light having a wavelength λ.

FIGS. 2a-b illustrate cross sectional views of an optical MEM device, in accordance with an alternative construction. The optical MEM device can have pairs of alternating ribbons 206 and 207 that are approximately in the same reflective plane, wherein adjacent ribbons 206 and 207 form sets or ribbon pairs. Referring to FIG. 2a, the ribbons 206 and 207 are suspended over a substrate structure 202 by a distance 203. The ribbons 206 and 207 are provided with reflective surfaces 204 and 205, respectively. The substrate 202, or a portion thereof, can also have a reflective surface 208. In accordance with the embodiments of the invention, the reflective surface 208 of the substrate 202 and the reflective surfaces 204 and 205 of the ribbons 206 and 207 are configured to be separated by a distance approximately equal to a multiple of λ/2 of the impinging light source. Thus, a portion of light that is reflected from the compliment of surfaces 204, 205 and 208 are all in phase and a maximum intensity of reflected light is observed. In operation, the optical MEM device alternates between the conditions for constructive and destructive interference by moving the first set of ribbons 206 or the second set of ribbons 207 relative to each other by a distance corresponding to a multiple of λ/4.

In one mode of operation, light is modulated by moving one set of alternating ribbons relative to a stationary set of alternating ribbons. The ribbons that are moved are referred to as the active ribbons and the stationary ribbons are referred to as the bias ribbons. The active ribbons are moved by any number of means including mechanical means, but are preferably moved by applying a sufficient bias voltage across the active ribbons and the substrate to generate Coulombic attractions between the active ribbons and the substrate.

Now referring to FIG. 2b, when a sufficient bias voltage is applied across the active of ribbons 207 and the substrate 202, the ribbons 207 are displaced relative to the bias ribbons 206 by a distance 211 that is approximately equal to a multiple of λ/4. Accordingly, the light that is reflected from the surfaces 205′ of the active ribbons 207 destructively interferes with the light that is reflected from the surfaces 204 of the bias ribbons 206. It will be clear to one skilled in the art that an optical MEM device may be configured to modulated an incident light source having a wavelength λ in other operative modes. For example, both sets of ribbons 206 and 207 may be configured to move such that they are separated by multiples of λ/4 in order to alternate between conditions for constructive and destructive interference to modulate light having a wavelength λ.

The ribbons of the MEM devices, described in FIGS. 1a-b and FIGS. 2a-b, are preferably hermetically sealed within a die structure. Methods and materials used for providing a hermetically sealed die are described in the U.S. patent application Ser. No. 09/124710, filed Jul. 29, 2001, entitled “METHOD OF AND APPARATUS FOR SEALING AN HERMETIC LID TO A SEMI CONDUCTOR DIE”, now U.S. Pat. No. 6,303,986, the contents of which are hereby incorporated by reference.

The present invention is directed to optical MEM devices and systems which comprise ribbon structures that are preferably formed directly on a silicon wafer substrate, wherein the silicon wafer substrate is selectively etched in the regions corresponding to the center portions, or active portions of the ribbons. In accordance with the embodiments of the invention, an optical MEM device is formed without requiring the deposition of a sacrificial layer, such as poly-silicon, prior to etching step(s). Further, the number of dielectric surfaces in the resultant device can be reduced and, therefore, devices can be made which exhibit reduced charging characteristics. Also, because the time consuming step of depositing a sacrificial layer is not required for fabricating optical MEM devices, in accordance with the embodiments of the invention, optical MEM devices can be fabricated using a process flow having greater compatibly with standard CMOS materials and/or IC wafer processing steps.

FIG. 3a shows a cross-sectional view of a micro-structure 300 formed in accordance with the embodiments of the invention. The device layer 300 has a silicon based layer or support layer 305 that is preferably silicon nitride with a thickness in a range of about 800 to about 1200 Angstroms. The micro-structure 300 also has a reflective layer 301 that is preferably formed from a metal and has a thickness in a range of about 500 to about 1500 Angstroms. The reflective layer 301 can be formed from any number of metals and metal alloys, but is preferably formed from aluminum or some other metal that can be deposited using sputtering techniques at relatively low temperatures.

Still referring to FIG. 3a, the device layer 300 can also have an oxide layer 303, that is preferably a silicon dioxide layer with a thickness in a range of about 800 to about 1800 Angstroms. The oxide layer 303 is preferably interposed between the reflective layer 301 and the layer 305. Alternatively, or in addition to the oxide layer 303, a second oxide layer (not shown) can be formed below the support layer 305.

FIG. 3b shows a portion of a micro-device 325, in accordance with the embodiments of the invention. The micro-device 325 preferably has a plurality of ribbon structures 332 and 332′ suspended over a substrate 326. The ribbon structures 332 and 332′ are preferably formed from a multi-layer structure device layer, such as described above, wherein the ribbon structures 332 and 332′ comprise a nitride layer 335, a metal layer 331, and one or more oxide layers 333. The plurality of ribbons 332 and 332′ can comprise an alternating first set of ribbons 332 and second set of ribbons 332′ which are moved relative to each other, such as described above. In accordance with the embodiments of the invention, the ribbons 332 and 332′ move in opposite directions. In alternative embodiments, one set of the ribbons 332 or 332′ moves while the other set of ribbons remains stationary. In yet other embodiments, the micro-device 325 is configured to operate in two or more modes of operation, whereby the micro-device 325 is configured to move one set of ribbons 332 and 332′ in one of the modes of operation and both sets of the ribbons 322 and 322′ in another of the modes of operation. Also, the ribbons can be configured to operate in a contact mode, whereby ribbons contact the substrate (such as shown in FIGS. 1a-b), or the ribbons can be configured to operate in a non-contact mode (such as shown in FIGS. 2a-b), wherein the ribbons do not contact the substrate.

In accordance with the embodiments of the invention, the substrate 326 can have a layer 325 of reflective material. The ribbon structures 332 and 332′, shown in FIG. 3b, all have uniform widths W₁ and W₂ and spacings S₁. However, any number of ribbon constructions, configurations or arrangements are considered to be within the scope of the present invention. For example, a micro-device can be fashioned with ribbons having varied widths W₁ and W₂ and varied spacings S₁. Examples of micro-devices with ribbons having varied widths W₁ and W₂ and varied spacings S₁ for optimized contrast are further described in U.S. patent application Ser. No. 09/802,619, filed Mar. 08, 2001, entitled “HIGH CONTRAST GRATING LIGHT VALVE”, the contents of which is hereby incorporated by reference. Also, while the preferred micro-structure(s) comprise the silicon nitride layer 335, the reflective metal layer 331 and one or more oxide layer(s) 333, it is understood that the compositions and thicknesses of the layers 331, 333 and 335 can be varied without departing from the spirit and scope of the embodiments. For example, the metal layer 331 can be formed from an alloy of metals, and the nitride 335 and/or oxide layer(s) 333 can contain impurities and/or dopants, such a boron, phosphorus and the like.

FIG. 4 illustrates a cross-sectional representation of a micro-device 400 comprising a multi-layer ribbon structure 408, in accordance with the embodiments of the invention. The micro-device 400 comprises a substrate 402, which can comprise a silicon layer 401 and an oxide layer 403, or, alternatively, the substrate 402 is a silicon wafer without an oxide layer. The ribbon structure 408 preferably comprises a layer of silicon nitride 411 and a reflective layer 415 of aluminum, as previously described. In further embodiments, a layer of oxide 413 is provided between the nitride layer 411 and the reflective layer 415. The preferred thickness of the oxide layer 413 varies depending on the application at hand, but a thickness of about 500 to about 2000 Angstrom is suitable for most applications.

Still referring to FIG. 4, the ribbon structure 408 is preferably suspended over the substrate 402 such that there is one or more gaps 409 between the ribbon 408 and the substrate 402. Preferably, the ribbon structure 408 is coupled to the substrate 402 through securing regions 423 and 427 comprising doped or un-doped silicon, such as described below. The securing regions 423 and 427 can form posts and/or anchors to support the ribbon structure 408, as is well known in the art. The device 400 preferably comprises a plurality of ribbons (not shown), wherein ribbon pairs are configured to move relative to each other and modulate light, as described above.

FIGS. 5a-e will now be used to illustrate steps for making a micro-device, in accordance with the embodiments of the invention. Micro-structures are formed on a suitable substrate 502. The substrate 502 comprises a wafer layer 503, that is preferably a layer of single crystal silicon.

Referring to FIG. 5b, a device layer 505 is formed directly on the wafer layer 503. The device layer 505 preferably comprises one or more layers of silicon nitride, oxide and a reflective layer, as explained above. After the device layer 505 is formed on the wafer layer 503, the device layer 505 is cut into ribbons 511 such as shown in FIG. 5d.

The device layer 505 is cut or patterned into ribbons 511 using any suitable technique, but preferably the device layer 505 is cut or patterned using conventional photo-lithography techniques and/or steps. For example, a photo-resist layer (not shown) is formed on the device layer 505. The photo-resist is patterned and developed to form a patterned photo-resist mask. Portions of the pattern photo-resist layer are removed using conventional techniques leaving mask patterned with the ribbon cuts 507. The ribbon cuts 507 are then etched to form the patterned device layer 505′ patterned with electrically isolated ribbons 511, as shown in FIG. 5d. FIG. 5d shows the top view of the device 500 shown in FIG. 5c, wherein a line 513 shows an x-axis of the device 500 and a line 515 shows the y-axis of the device 500. A z-axis 514 of the device 500, shown in FIG. 5d, is normal to the view shown in FIG. 5c.

Referring to FIGS. 5c-d, in accordance with further embodiments, the device structure 500 is also provided with a plurality of access holes and/or trenches 502 and 504 in addition to the ribbon cuts 507. The access holes and/or trenches 502 and 504 are provided to facilitate the local etching rate under central portions 531 of the ribbons. The access holes and/or trenches 502 and 504 are preferably formed with average diameters in a range of about 0.4 to about 1.5 microns and more preferably with average diameters in a range of about 0.6 to about 0.8 microns. The device layer 505 (FIG. 5b) is patterned with the access holes and/or tenches 502 and 504 in the same step as the ribbon cuts 507 or in a different step.

Now referring to FIG. 5e, after the device layer 505 is patterned with ribbon cuts 507 and/or access holes and/or trenches 502 and 504 to form the patterned device layer 505′, then the intermediate device structure 500 is treated with a selective etching process to etch a cavity 509 under the center portions 531 of the ribbons 511, while end portions 520 and 521 remain attached to the etched wafer layer 503′. The remaining portion of the etched wafer layer 503′ residing underneath the end portions 520 and 521 act as a support structure for the ribbons 511. The cavity 509 is preferably formed by etching the device structure 500 with an enchant comprising noble gas fluoride. More preferably, the etchant comprises xenon difluoride. Advantages of using xenon difluoride etchant are described by Pister in U.S. Pat. No. 5,726,480, the contents of which are hereby incorporated by reference.

In accordance with the embodiments of the invention, a substrate structure is modified prior to forming a patterned device layer, such as described. Referring to FIG. 6a, a substrate structure 602 preferably comprises a single crystal silicon wafer layer 604. To form a modified substrate structure 602′, as shown in FIG. 6b, the substrate 602 is selectively exposed to an ion source 611. The ion source 611 is configured to emit one or more ionic species 603 and 605, with a sufficient energy to implant the ions 603 and 605 into the single crystal silicon layer 604 through the outer regions 620 and 621 and/or through the central region 609 of the wafer layer 604. Preferably, the implanted ions render outer portions 627 and 623 of the modified layer 604′ less reactive to the etch process than a central portion 625, render the central portion 625 of the modified layer 604′ more reactive to etch process or a combination thereof.

Still referring to FIG. 6b, after the modified substrate structure 602′ is formed with the desensitized outer portions 627 and 623 and/or the sensitized central portion 625, then a patterned device layer 605 is formed directly over the modified layer 604′. Referring to FIG. 6c, the modified layer 604′ is etched to form a cavity 619, wherein the central portion 631 of the patterned device layer 605 is suspended over the etched substrate 602″ and attached to the etched substrate 602″ through the end portions 627 and 623 of the etched silicon layer 604″. The device 600 preferably comprises a plurality of ribbons (not shown), wherein ribbon pairs are configured to move relative to each other and modulate light, as described above.

FIG. 7 is a block diagram 700 outlining the preferred method of etching an intermediate device structure patterned device layer patterned with ribbon cuts and/or access holes, such as illustrated in FIGS. 5c-d and FIG. 6b. Referring to FIG. 7, in the step 701 the intermediate device structure is place under a vacuum within a chamber, wherein the partial pressure of water is preferably about 5×10⁻⁴ Torr or less. In the step 703, xenon difluoride crystals are preferably sublimed at a pressure in a range of about 0.1 to about 100 Torr, more preferably in a range of about 0.5 to about 20 Torr and most preferably at approximately 4.0 Torr. In the step 705, a controlled stream of xenon difluoride is provided to the chamber, which is preferably a chamber of a cluster tool equipped with a load lock for transferring the device in and out of the chamber and/or between processing chamber within the cluster tool. The chamber is preferably maintained at a pressure lower than the sublimation pressure of the xenon difluoride crystals to ensures a positive flow of the xenon difluoride to the chamber. The pressure in the chamber is preferably maintained in a range of about 0.1 milliTorr to about 10.0 Torr, more preferably in a range of about 1.0 milliTorr to about 100 milliTorr and most preferably at approximately in a range of about 60 to about 80 milliTorr (0.06-0.08 Torr) during the etching steps.

FIG. 8 illustrates a schematic representation of an etching station 800 in a cluster tool for carrying out the etching steps described above. The etching station 800 is preferably coupled to a load lock 845 for transferring a device 820 in and out of the chamber environment 805′ and/or between processing stations (not shown) within the cluster tool. The etching station 800 preferably comprises a vacuum source 807 that is capable of drawing a vacuum in the chamber environment 805′. The etching station 800 preferably includes a pressure measuring device 809 that allows a user to monitor the pressure within a chamber 810. A container 808 of an etchant source (e.g. crystals of xenon difluoride) is coupled to the chamber 810 through a pressure or flow controller 813. The container 808 can have a pressure measuring device 811 coupled to the container 808 to allow the user to monitor the pressure within the container 808.

In operation, the device-structure 820 comprising a patterned device layer 861 and a single crystal silicon layer 803 with or without an etch stop layer 821, is placed in the chamber 810. A vacuum control valve 840 is opened and the vacuum source 807 draws a vacuum reducing the partial pressure of water within the chamber 810 to about 5.0×10⁻⁴ Torr or less. It is preferable that the partial pressure of water within the chamber 810 remains at about 5.0×10⁻⁴ Torr or less during the etching steps in order to maintain a controllable etching process. Under known conditions, the xenon difluoride crystals at room temperature form a vapor pressure of XeF₂ of approximately 4.0 Torr, as determined by the pressure measuring device 811. The pressure controller 813 is adjusted to change the pressure of the chamber environment 805′ to be in a range of 60 to about 80×10⁻³ Torr. The structure 820 is etched for a time sufficient to form a cavity 840 and to release a central portion or portions 831 of the patterned device layer 861. The etching conditions provided above are intended to provide a description of the etching process. The preferred conditions of the etching process depends on the etching pressure(s) chosen, the materials used to fabricate the device 820, the size to the cavity 840 needed, physical details of the structure 820 and flow dynamics of the chamber 810.

During the etching step, the device 820 can be cooled or maintained at a temperature of approximately 20 degrees Celsius, through heat sink 827. The heat sink 827 is preferably coupled to a cooling means 829, such as a refrigeration unit for controlling the temperature of the heat sink 827 at or below 20 degrees Celsius. The heat sink 827, is preferably formed from metal or another material suitable for absorbing heat from the device 820, while the device 820 is coupled to the heat sink 827.

After the etching step is complete, a suitable storage environment may then be provided and the device 820 is hermetically sealed die (not shown), such as described in the U.S. patent application Ser. No. 09/124,710, filed Jul. 29, 2001, entitled “METHOD OF AND APPARATUS FOR SEALING AN HERETIC LID TO A SEMI CONDUCTOR DIE”, now U.S. Pat. No. 6,303,986, previously referenced. Alternatively, after reestablishing a low pressure vacuum, the chamber may be backfilled with a noble gas. Accordingly, a noble gas source 815 may be coupled to the control chamber 810 through a control valve 812. The chamber environment 805′ is flushed with a noble gas by opening the gas valve 812 prior to sealing the device 820 in a die structure.

Because devices are preferably fabricated in accordance with the embodiments of the invention by first forming a patterned device layer directly on a doped or undoped silicon wafer and etching the wafer to release selected portions of the device layer, optical MEM devices can be fabricated without requiring the deposition of a sacrificial layer, such as poly-silicon. Further, the number of dielectric surfaces can be reduced, thereby reducing the charging characteristics of the devices formed. Also, by eliminating the deposition steps required to form a sacrificial layer, the fabrication of MEM devices in accordance with the embodiments of the invention is more compatible with standard CMOS materials and/or IC wafer processing steps than previous methods which typically require deposition of a sacrificial layer.

The above examples have been described in detail to illustrate the preferred embodiments of the instant invention. It will be clear to one of ordinary skill in the art that there are many variations to the invention that are within the scope of the invention. For example, while the embodiments of the invention are directed to optical MEM devices and method for making the same, other micro-structures formed with pattern device on a doped or undoped wafer substrate are contemplated. Accordingly, such references to specific embodiments and details are not intended to limit the scope of the claims appended hereto.