Vacuum-cavity MEMS resonator

[0001] This application is a divisional of U.S. patent application Ser. No. 09/949,368 filed on Sep. 7, 2001, now U.S. Pat. No. 6,808,954, incorporated herein by reference.

[0002] The present invention relates to microelectromechanical systems (MEMS), and in particular to vacuum cavity packaging of MEMS resonators.

[0003] A resonator is an electronic device used for setting up and maintaining an oscillating electrical signal of a given frequency. Conventional resonators typically include electronic circuitry in combination with a mechanical oscillator element (e.g., a quartz crystal, a ceramic element or a resonance circuit). Resonators are used in many electronic devices, such wireless radio frequency (RF) equipment, for generating outgoing signals of a particular frequency, and filtering incoming signals.

[0004] In most electronic devices that require signal generation and filtering, conventional resonators are used. Such resonators have a high Q-factor (i.e., a sharp resonance peak) good frequency stability and are generally very reliable. However, conventional resonators tend to be relatively large (i.e., on the order of 1 cm), so that alternatives are preferred when trying to fabricate a compact electronic device.

[0005] One alternative to conventional crystal-based resonators is a microelectromechanical systems (MEMS) resonator. Generally, a MEMS device is a microdevice that integrates mechanical and electrical elements on a common substrate using microfabrication technology. The electrical elements are typically formed using known integrated circuit fabrication technology. The mechanical elements are typically fabricated using lithographic and other related processes to perform micromachining, wherein portions of a substrate (e.g., silicon wafer) are selectively etched away or added to with new materials and structural layers.

[0006] FIG. 1 shows one type of prior art MEMS resonator 10 formed on a substrate 12. MEMS resonator 10 has a cantilever-type beam 16 arranged between a lower electrode 20 and an upper electrode 26. Beam 16 is electrostatically driven by the upper and lower electrodes to oscillate at a given frequency. FIG. 2 shows another type of prior art MEMS resonator 40 similar to MEMS resonator 10 but having a bridge-type beam 46 and an optional bridge-type upper electrode 50. Beam 46 is anchored to the substrate at its ends so that the center portion of the beam can be driven to oscillate by being electrostatically deflected between the upper and lower electrodes. FIG. 3 shows yet another prior art MEMS resonator 70 called a “breathing bar resonator.” MEMS resonator 70 includes a bar-type beam 76 fixed to substrate 12 with a single central support member 80. Side electrodes 84 and 86 are located on either side of beam 76 with small gaps 88 in between. Electrodes 84 and 86 drive beam 76 to expand and contract (i.e., resonate) along its long axis in a manner that resembles breathing.

[0007] MEMS resonators are desirable for many miniaturized electronic devices because they can be made smaller than conventional resonators by an order of magnitude or more. However, because a MEMS resonator relies on the mechanical oscillation of a very small (i.e., micron-sized) beam as opposed to the vibration of a relatively large oscillation element (e.g., a centimeter-size crystal), the resonator must be packaged in a vacuum to eliminate air damping of the beam's oscillation. Vacuum packaging is also necessary to avoid the adsorption of contaminants, which can alter the resonant frequency of the beam.

[0008] A challenge in fabricating MEMS resonators is the vacuum packaging step. Various techniques for vacuum packaging a MEMS resonator are available, such as wafer bonding, flip-chip, and thick membrane transfer techniques. However, these techniques require dedicated alignment/bonding technologies that are relatively complicated to apply to MEMS packaging. Another technique for MEMS vacuum packaging involves using a permeable polysilicon release process. While conceptually simple, such a process has proven very difficult to control and has yet to lead to a manufacturable MEMS resonator vacuum packaging process.

[0017] In the Figures, like reference numbers refer to like elements.

[0018] The present invention provides vacuum cavities for MEMS resonators by a film deposition and release process that does not add a significant number of additional steps to the MEMS resonator fabrication process.

[0019] In the following detailed description of the embodiments of the invention, reference is made to the accompanying drawings that form a part hereof, and in which is shown by way of illustration specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention, and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the present invention. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is defined only by the appended claims.

Cantilever-beam MEMS Resonator (FIGS. 4A–4K)

[0020] In FIG. 4A, the method begins by providing a substrate 100 having an upper surface 102. In an example embodiment, substrate 100 is a standard semiconductor wafer. A nitride layer 110 (e.g., silicon nitride) is formed on the upper surface 102 of substrate 100. In an example embodiment, nitride layer 110 has a thickness ranging anywhere from about 1000 Angstroms to 5000 Angstroms. Nitride layer 110 serves as a passivation layer upon which the MEMS resonator of the present invention is formed. Atop nitride layer 110 is formed a polysilicon layer 120, which in an example embodiment may have a thickness of up to one micron or so.

[0021] In FIG. 4B, polysilicon layer 120 is selectively etched to form a lower electrode 126, a beam base 130, and upper electrode contacts 136 and 140. In an example embodiment, electrode contacts 136 and 140 and lower electrode 126 are connected to bond pads (not shown), which in turn are connected to a programmable voltage source (not shown) for providing power to the device. Also, beam base 130 may be connected to ground.

[0022] The selective etching of layer 120 may be performed, in an example embodiment, by using standard lithography techniques, i.e., coating layer 120 with a layer of photoresist (not shown), exposing the photoresist layer with a pattern using photolithographic imaging of a patterned mask (not shown), etching the structure (e.g., with a reactive-ion etch) to form the pattern in the layer, and then removing the remaining photoresist.

[0023] In FIG. 4C, a release layer 150 is formed atop lower electrode 126, electrode contacts 136 and 140, and the exposed (i.e., uncovered) portions of nitride layer 110. Release layer 150 may be, for example, an oxide such as silicon dioxide, or a polymer such as polyimide or photoresist. Release layer 150 is formed to a predetermined thickness (typically, a few hundred to a few thousand Angstroms) to set the spacing between lower electrode 126 and the yet-to-be-formed beam (FIG. 1). To ensure release layer 150 has the proper thickness, chemical-mechanical polishing (CMP) of the layer may optionally be performed.

[0024] In FIG. 4D, release layer 150 is selectively processed (e.g., using standard lithography techniques) to define an opening 160 that extends down to beam base 130, and to define openings 162 and 164 that extend down to upper electrode contacts 136 and 140.

[0025] In FIG. 4E, a second polysilicon layer 170 is formed atop the processed release layer 150. In an example embodiment, the thickness of layer 170 can be anywhere from about 1 to 2 microns. Material from layer 170 fills opening 160 to form an anchor 172 connected to beam base 130. Material from layer 170 also fills openings 162 and 164 and connects to upper electrode contacts 136 and 140, respectively.

[0026] In FIG. 4F, layer 170 is selectively processed (e.g., using standard lithographic techniques) to define a cantilever beam 174 with opposite sides 176 and an upper surface 178. The selective processing of layer 170 also defines upper electrode bases 180 and 182 connected to electrode contacts 136 and 140 and separated from beam 174 by openings 190 and 192. Openings 190 and 192 extend down to release layer 150.

[0027] In FIG. 4G, a spacer layer 200 is selectively formed atop upper surface 178 of beam 174 and over sides 176. Spacer layer 200 is used to space apart beam 174 and the yet-to-be-formed upper electrode (FIG. 1). In an example embodiment, spacer layer 200 has a thickness that ranges from about 100 Angstroms to about 5000 Angstroms. Also in an example embodiment, spacer layer 200 is formed from release material such as an oxide (e.g., SiO₂) that is selectively grown on the structure by, for example, masking the structure (e.g., with a layer of photoresist), selectively etching the oxide, and then removing the photoresist masking layer. The portion of spacer layer 200 covering sides 176 of beam 174 constitute sidewall spacers 202 that connect with release layer 150.

[0028] In FIG. 4H, a third polysilicon layer 210 is formed atop electrode bases 180 and 182, and atop spacer layer 200. Layer 210 fills openings 190 and 192 so that the layer also covers sidewalls 202. In an example embodiment, layer 210 has a thickness anywhere from about 1 micron to about 5 microns.

[0029] In FIG. 4I, layer 210 is selectively etched to form a release hole 216 that extends through to release layer 150. The selective etching of layer 210 further defines from layer 210 an upper electrode 220, which may be cantilever-type (FIG. 1) or bridge-type (FIG. 2).

[0030] In FIG. 4J, release layer 150 and spacer layer 200 are removed through release hole 216 by etching, dissolving or other techniques suitable to the material of the release layer. This leaves an internal contiguous cavity 224 in the resulting structure, the cavity being defined by first nitride layer 110, lower electrode 126, upper electrode contacts 136 and 140, beam 174, and upper electrode 220. Cavity 224 is open to outside the structure via release hole 216. Beam 174 resides within cavity 224 so that it is free to resonate when electrostatically engaged by lower electrode 126 and upper electrode 220.

[0031] In FIG. 4K, a nitride (e.g., SiN) layer 230 is formed over the structure shown in FIG. 4J. Nitride layer 230 at least partially fills release hole 216, thereby sealing the structure, including cavity 224. The formation of nitride layer 230 is performed under vacuum so that internal cavity 224 is vacuum-sealed. The result is a vacuum-cavity MEMS resonator 250 with a cantilever beam. Note that upper electrode 220 is designed to serve as a support structure to facilitate the vacuum packaging of MEMS resonator 250.

Bridge-beam MEMS Resonator (FIGS. 5A–5D)

[0032] The method of forming a vacuum-cavity bridge-beam MEMS resonator is similar to the method of forming a cantilever-beam MEMS resonator, as described above. Accordingly, only the essential differences between the two methods are discussed.

[0033] FIG. 5A shows a structure similar to that of FIG. 4B described above, except that an extra beam base 262 is formed atop nitride layer 110 from polysilicon layer 120 between lower electrode 126 and upper electrode contact 140. Also, the structure in FIG. 5B is similar to that formed up through FIG. 4E, except that second polysilicon layer 170 fills an additional opening formed in release layer 150 over beam base 262 that results in the formation of an anchor 272 that connects to beam base 262. Thus, layer 170 is connected to nitride layer 110 via bases 130 and 262 located on opposite sides of lower electrode 126.

[0034] Further, the structure of FIG. 5C is similar to that formed up through FIG. 4G, except that instead of cantilever beam 174, there is a bridge beam 300 with opposite sides 304 and an upper surface 306. Spacer layer 200 with sidewalls 202 is formed on beam 300 (including over sides 304) in the manner described above in connection with FIG. 4G, and is connected to release layer 150.

[0035] The resulting vacuum-cavity bridge-beam MEMS resonator 350 is shown in FIG. 5D. Internal cavity 224 is contiguous because beam 300 does not contact upper electrode bases 180 or 182 or other structural elements in the MEMS resonator in the direction out of the plane of the FIG. 5D. Thus, beam 300 resides within cavity 224 so that it is free to resonate when electrostatically engaged by lower and upper electrodes 126 and 220.

[0036] As with MEMS resonator 250, upper electrode 220 of MEMS resonator 350 is designed to serve as a support structure to facilitate vacuum packaging. MEMS resonator 350 can have a cantilever-type upper electrode or bridge type upper electrode (FIGS. 1 and 2; FIG. 4K).

Breathing-beam MEMS Resonator (FIGS. 6A–6K)

[0037] The method begins as described above in connection with FIGS. 4A and 4B. However, as illustrated in FIG. 6A, two electrode bases 400 and 406 and a beam base 410 are formed from layer 120. In an example embodiment, electrode bases 400 and 406 are connected to bond pads (not shown), which in turn are connected to a programmable voltage source (not shown) for providing power to the device. Beam base 410 may be connected to ground.

[0038] In FIG. 6B, release layer 150 is formed over electrode bases 400 and 406, beam base 410 and the exposed or uncovered portions of nitride layer 110. Planarizing of layer 150 (e.g., using CMP) to ensure the layer has the required thickness (e.g., a few hundred to a few thousand Angstroms) may optionally be performed.

[0039] In FIG. 6C, release layer 150 is selectively processed to define openings 420 and 424 that extend down to electrode bases 400 and 406. Also formed in release layer 150 is an opening 430 that extends down to beam base 410.

[0040] In FIG. 6D, a second polysilicon layer 440 is formed (e.g., conformally deposited) over release layer 150. A portion of layer 440 fills openings 420 and 424 down to electrode bases 400 and 406. Further, a portion of layer 440 fills opening 430, forming a beam support member 460 fixed to beam base 410.

[0041] In FIG. 6E, layer 440 is selectively processed to define a center beam 470 with opposite sides 472. Center beam 470 is supported by beam support member 460 as well as by the portion of release layer 150 between the beam and beam base 410. Also, the selective processing may optionally include adjusting the thickness of electrode bases 400 and 406.

[0042] In FIG. 6F, sidewall spacers 480 made of a release material (e.g., an oxide) are formed over sides 472 of beam 470, e.g., as described above in connection with FIG. 4G. Sidewall spacers 480 serve to space apart the yet-to-be-formed side electrodes from the beam (FIG. 3).

[0043] In FIG. 6G, a relatively thick (e.g., 2 to 3 microns) polysilicon layer 500 is then formed (e.g., conformally deposited) atop the structure of FIG. 6F, with polysilicon filling openings 420 and 424 and covering sidewall spacers 480.

[0044] In FIG. 6H, the structure of FIG. 6G is planarized (e.g., via CMP) down to at least sidewalls 480 in order to form side electrodes 510 and 514, which are connected to respective electrode bases 400 and 406 and spaced apart from sides 472 of beam 470 by sidewall spacers 480. At this point in the process, various interconnections (e.g., bond pads, electrical connections to the electrodes, etc.) for powering the MEMS resonator and connecting the MEMS resonator to other devices can be formed using standard lithographic techniques.

[0045] In FIG. 6I, another release layer 520 is selectively formed atop beam 470, sidewall spacers 480, and a portion of side electrodes 510 and 514, in a manner similar to that discussed in connection with FIGS. 4C and 4D.

[0046] In FIG. 6J, a polysilicon layer 530 with an upper surface 532 is then formed over release layer 520 and is planarized, if necessary. A release hole 536 is then formed in polysilicon layer 530 that connects upper surface 532 to release layer 520. Polysilicon layer 530 provides the structural support for isolating beam 470 and for vacuum-packaging the MEMS resonator.

[0047] In FIG. 6K, the release material making up layers 150 and 520 and sidewall spacers 480 are removed through release hole 536. When the removal process is complete, an internal cavity 546 is formed within the resulting structure that includes release hole 536. Then, as discussed above in connection with FIG. 4K, nitride (e.g., SiN) layer 230 is deposited over the structure, with some of the nitride at least partially filing release hole 536. The result is the formation of a MEMS resonator 550 having a vacuum-sealed cavity 546 in which resides “breathing beam” 470 located between side electrodes 510 and 514 that electrostatically engage the beam.

Breathing-beam MEMS Resonator-Alternative Method (FIGS. 7A–7E)

[0048] An alternative method of forming a vacuum-cavity breathing-beam MEMS resonator is now described with reference to FIGS. 7A through 7E. The alternative method adds a step of forming a release layer to the first method, but also eliminates a CMP planarization step.

[0049] With reference to FIG. 7A, the alternative method begins with the structure of FIG. 6E described above, and further includes forming (e.g., conformally depositing) a second release layer 600 atop the structure of FIG. 6E. Then, in FIG. 7B, release layer 600 is selectively processed (e.g., using standard lithographic techniques) to leave an island 606 of release layer material atop beam 470.

[0050] In FIG. 7C, sidewall spacers 612 made of release material (e.g., oxide) are formed over sides 472 of beam 470, e.g., in the manner described above in connection with FIG. 4G.

[0051] In FIG. 7D, a polysilicon layer 620 is then formed (e.g., conformally deposited) atop the entire structure of FIG. 7C, thereby defining side electrodes 640 and 644, which extend down to and are fixed to respective electrode bases 400 and 406. Release holes 650 and 652 are then formed in polysilicon layer 620 down to release layer 150.

[0052] In FIG. 7E, the release material making up release layer 150, sidewalls 612 and island 606 is removed from the structure through release holes 650 and 652, thereby forming an internal cavity 666 within the structure. As discussed above in connection with FIGS. 4K and 6K, nitride layer 230 is then deposited under vacuum over the structure, with some of the material from the layer at least partially filling release holes 650 and 652. The result is a vacuum-cavity breathing-beam MEMS resonator 700 having a vacuum cavity 666 in which resides central “breathing” beam 470 between side electrodes 640 and 644.

MEMS Resonator Circuit

[0053] With reference to FIG. 8, there is shown a MEMS resonator circuit 800 that includes a vacuum-cavity MEMS resonator 806 according to the present invention. For example, MEMS resonator 806 can be the vacuum-cavity cantilever beam MEMS resonator 250 (FIG. 4K), the vacuum-cavity bridge-beam MEMS resonator 350 (FIG. 5D), the vacuum-cavity breathing-beam MEMS resonator 550 (FIG. 6K) or the vacuum-cavity breathing-beam MEMS resonator 700 (FIG. 7E).

[0054] MEMS resonator 806 includes a beam 810 with a beam contact 812 (e.g., an extension of beam base 130 of FIG. K), and first and second electrodes 820 and 824 (e.g., lower electrode 126 and upper electrode 220 of MEMS resonator 250 or side electrodes 510 and 514 of MEMS resonator 550). Beam 810 is biased with a DC polarization voltage V_P from voltage source 840. First electrode 820 is connected to an input voltage source 844 providing an input voltage V_I. Input voltage source 844 may be, for example, an antenna. An inductor 848 and a voltage output line 852 for measuring an output voltage V_O are connected in parallel to second electrode 824.

[0055] In an example embodiment, beam 810 has a resonant frequency ranging anywhere from a fraction of a megahertz to hundreds of megahertz to a few gigahertz, and a Q-factor ranging anywhere from a thousand to several tens of thousands (generally, the lower the resonant frequency, the higher the Q-factor).

[0056] If the frequency of the input voltage V_I is not close to the resonant frequency of beam 810, then there is no output voltage V_O from second electrode 824. When the frequency of the input voltage V_I is at or near the resonant frequency of beam 810, the beam starts to resonate (how close the input voltage V_I must be to the resonant frequency is determined by the Q-factor). This causes a capacitance change with respect to electrode 824, which causes an AC current I=(V_P)dC/dT to pass through inductor 848 at or near the resonant frequency of the beam, thus providing a mechanism for filtering an input voltage.

[0057] The present invention includes several different types of MEMS resonators that have vacuum cavities, and methods of forming same. In each MEMS resonator, the vacuum cavity surrounds the beam, allowing the MEMS resonators to maintain a high resonant frequency by eliminating air-damping of the oscillation. The vacuum cavity also prevents adhesion of contaminants to the beam, which can alter the beam's resonant frequency.

[0058] Further, the upper electrode is designed to serve as a support structure for vacuum packaging, which keeps the number of process steps reasonable from a manufacturing viewpoint. Also, the methods disclosed can utilize standard lithography techniques, which further limits the need for specialized process steps.

[0059] While the present invention has been described in connection with preferred embodiments, it will be understood that it is not so limited. On the contrary, it is intended to cover all alternatives, modifications and equivalents as may be included within the spirit and scope of the invention as defined in the appended claims.