CN111422827A

CN111422827A - Wafer manufacturing process flow of high-performance MEMS inertial sensor

Info

Publication number: CN111422827A
Application number: CN202010283685.0A
Authority: CN
Inventors: 郭述文; 何凯旋; 刘磊; 王鹏; 郭群英
Original assignee: China North Industries Group Corp No 214 Research Institute Suzhou R&D Center
Current assignee: No 214 Institute of China North Industries Group Corp
Priority date: 2020-04-13
Filing date: 2020-04-13
Publication date: 2020-07-17

Abstract

The invention discloses a wafer manufacturing process flow of an MEMS (micro-electromechanical system) inertial sensor, which comprises the following steps of: providing a first substrate having a top plane and a bottom plane, an oxide layer disposed below the bottom plane; providing a second substrate having an upper planar surface; etching a portion of the second substrate from the upper planar surface to form a plurality of protrusions and shallow cavities, each protrusion having an upper planar surface; bonding a top planar surface of the first substrate to the upper planar surface of the protrusion to form an anchor portion; etching off the oxide layer at a certain position of the first substrate; a portion of the first substrate is etched from the bottom surface and/or the top surface to a total thickness of the first substrate to form a sensitive structure having a structure that is free to rotate. The method has the advantages that icon alignment is not needed during silicon-silicon bonding, so that the bonding process is simple, the thickness of the formed beam is equal to that of the mass block, no error is generated, and the method has better structural symmetry.

Description

Wafer manufacturing process flow of high-performance MEMS inertial sensor

Technical Field

The invention relates to a manufacturing process of an inertial sensor, in particular to a wafer manufacturing method of an MEMS inertial sensor.

Background

High performance accelerometers and gyroscopes with near microgravity resolution, high sensitivity, high linearity and low bias drift are needed for a wide variety of applications, particularly aerospace applications such as inertial navigation systems, guidance systems and airborne data measurement systems. The thermo-mechanical brownian noise of the sensor limits the resolution of high performance accelerometers and gyroscopes, which is determined by the damping coefficient and mass of the structure and the reading electronics.

Fabrication techniques play a crucial role in ensuring that simultaneously large masses, large capacitances and small damping are obtained, as well as achieving microgravity resolution. Previously, many high performance silicon accelerometers and gyroscopes have been reported. These devices incorporate capacitive, resonant or tunneling current sensing schemes that can achieve high sensitivity with greater detection quality. Of all these, silicon capacitive accelerometers have many advantages that make them very attractive in a number of applications ranging from low cost, large volume automotive accelerometers to high precision inertial grade microgravity devices. The silicon capacitance type accelerometer has high sensitivity, good direct current response and noise performance, low drift, low temperature sensitivity and low power consumption.

Capacitive accelerometers are generally vertical and lateral structures. Some designs use a teeter-totter structure with a proof mass such as a flat plate suspended by a torsion beam. The structure is typically asymmetrically shaped so that one side has a greater mass than the other, resulting in the centre of mass being offset from the axis of the torsion bar. When the acceleration forces generate a moment about the torsion bar axis, the plate rotates freely, constrained only by the spring constant of the torsion bar.

The sensitivity of these types of accelerometers is defined as the ratio of deflection to acceleration. The mass of the plate, the distance from the center of mass to the torsion bar axis and the stiffness of the torsion bar determine the sensitivity. To increase the offset of the center of mass, the plate structure is designed to have an asymmetric shape. For example, the width of one side of the plate may be greater than the width of the other side of the plate, or the length of one side of the plate may be greater than the length of the other side. However, increasing the center mass shift by the asymmetric shaping method described above may result in an increase in the total mass of the plate, which results in a decrease in the resonant frequency and a decrease in sensitivity. Increasing the center mass offset by asymmetric shaping also results in sacrificing some of the dynamic g-range, which is determined by the spacing between the fixed sensing element and the pendulum acceleration sensor plate. Another method of increasing center mass offset involves lengthening a portion of the pendulum sensor plate. The center mass offset is proportional to the length of the plate extension. However, one side of the extension plate may cause unbalanced gas damping, resulting in performance degradation. The gas damping can be balanced by perforating the expansion plate. However, such perforations also reduce center mass shift and therefore sensitivity. In addition, extending one side of the board may result in an increase in the overall chip size. It is defined by the distance between the fixed sensing element and the pendulum acceleration sensitive structure. Another method of increasing center mass offset involves lengthening a portion of the pendulum sensor plate. The center mass offset is proportional to the length of the plate extension. However, one side of the extension plate may cause unbalanced gas damping, resulting in performance degradation. The gas damping can be balanced by perforating the expansion plate. However, such perforations also reduce center mass shift and therefore sensitivity. In addition, extending one side of the board may result in an increase in the overall chip size. It is defined by the distance between the fixed sensing element and the pendulum acceleration sensitive structure. Another method of increasing center mass offset involves lengthening a portion of the pendulum sensor plate. The center mass offset is proportional to the length of the plate extension. However, one side of the extension plate may cause unbalanced gas damping, resulting in performance degradation. The gas damping can be balanced by perforating the expansion plate. However, such perforations also reduce center mass shift and therefore sensitivity. In addition, extending one side of the board may result in an increase in the overall chip size. Another method of increasing center mass offset involves lengthening a portion of the pendulum sensor plate. The center mass offset is proportional to the length of the plate extension. However, one side of the extension plate may cause unbalanced gas damping, resulting in performance degradation. The gas damping can be balanced by perforating the expansion plate. However, such perforations also reduce center mass shift and therefore sensitivity. In addition, extending one side of the board may result in an increase in the overall chip size. Another method of increasing center mass offset involves lengthening a portion of the pendulum sensor plate. The center mass offset is proportional to the length of the plate extension. However, one side of the extension plate may cause unbalanced gas damping, resulting in performance degradation. The gas damping can be balanced by perforating the expansion plate. However, such perforations also reduce center mass shift and therefore sensitivity. In addition, extending one side of the board may result in an increase in the overall chip size. This can lead to performance degradation. The gas damping can be balanced by perforating the expansion plate. However, such perforations also reduce center mass shift and therefore sensitivity. In addition, extending one side of the board may result in an increase in the overall chip size. This can lead to performance degradation. The gas damping can be balanced by perforating the expansion plate. However, such perforations also reduce center mass shift and therefore sensitivity. In addition, extending one side of the board may result in an increase in the overall chip size.

Other conventional structures utilize a deeper gap below the extended plate portion to increase the maximum angle of rotation while maintaining balanced gas damping. This structure can increase the dynamic g range to some extent. However, the extended portion of the plate increases the size of the overall chip size, resulting in unbalanced gas damping and lowering the resonant frequency of the rotating structure, which again results in reduced performance of the accelerometer.

US patent application No. US7736931B1 discloses a method for manufacturing a pendulum accelerometer, wherein a first substrate is provided by first etching a recess with a depth from the top plane (lower surface of the mass), then performing direct pattern-aligned silicon bonding with a second substrate, and then etching a mass with a certain cavity depth from the upper surface of the mass, the sum of the etching depth of the lower surface and the etching depth of the upper surface is just the thickness of the mass, that is, the etching depth of the upper surface is just enough to etch through the entire mass, and the entire structure is released. And the thickness of the beam connecting the central anchor point and the mass blocks on the two sides is equal to the thickness of the mass block minus the etching depth of the lower surface. However, the above process requires the first substrate to etch a groove with a certain depth, and then performs pattern alignment bonding, which makes the bonding process complicated; and the thickness of the beam is smaller than that of the mass block, and the gravity center of the structure is inclined upwards, so that certain error is generated in the output symmetry of +/-1G of the accelerometer sensor.

Disclosure of Invention

In view of the above technical problems, the present invention aims to: the wafer manufacturing process flow of the high-performance MEMS inertial sensor is provided, icon alignment is not needed during silicon-silicon bonding, the bonding process is simple, the thickness of the formed beam is equal to that of the mass block, no error is generated, and the wafer manufacturing process flow has better structural symmetry.

The technical scheme of the invention is as follows:

a method of manufacturing a MEMS inertial sensor, comprising the steps of:

s01: providing a first substrate having a top plane and a bottom plane with an oxide layer disposed below the bottom plane, the bottom plane being substantially parallel to the top plane;

s02: providing a second substrate having an upper planar surface;

s03: etching a portion of the second substrate from the upper planar surface to a first predetermined depth to form a plurality of protrusions and shallow cavities, each protrusion having an upper planar surface;

s04: bonding a top planar surface of the first substrate to the upper planar surface of the protrusion to form an anchor portion;

s05: etching off the oxide layer at a certain position of the first substrate;

s06: etching a portion of the first substrate from the bottom surface and/or the top surface to a second predetermined depth at least equal to the total thickness of the first substrate to form a sensitive structure on the first substrate having a structure that is freely rotatable, the sensitive structure including a structured mass around the anchor portion.

In a preferred technical scheme, the total thickness of the first substrate is 50-150 μm.

In a preferred embodiment, the first predetermined depth is 3 to 10 μm.

In a preferred embodiment, after step S03, the method further includes:

forming fixed electrodes in the shallow cavities at the two sides of the middle bulge;

and depositing an oxide layer on the surface of the fixed electrode.

In a preferred embodiment, after step S08, the method further includes the following steps:

providing a cover wafer; and

etching a portion of the cap wafer to form a top recess; and

the capping wafer is bonded to the second substrate such that the freely rotatable sensitive structure is enclosed within the recess of the capping wafer.

In a preferred technical scheme, the method further comprises the step of depositing a thin film on a part of the surface of the groove through getter sputtering.

In the preferred technical scheme, the method further comprises the steps of printing glass slurry on the contact surface of the cover wafer by using glass frit printing;

and completing vacuum bonding of the cover wafer and the second substrate under high vacuum condition.

In a preferred embodiment, two sides of the anchoring portion of the first substrate are etched to form a beam structure for connecting the proof mass.

In a preferred technical scheme, the sensitive structure is a hollow structure, a comb tooth structure or an annular structure.

Compared with the prior art, the invention has the advantages that:

the top plane (lower surface of the mass block) of the first substrate is not pre-etched, but a light plate without patterns is directly bonded with the second substrate through silicon and silicon, then a structural pattern is etched from the upper surface of the mass block by adopting a double-sided photoetching technology, a part needing to be etched with grooves is covered with silicon dioxide with a certain thickness, and because the silicon dioxide is much slower than the etching rate of silicon, in the process of etching the next structure release deep silicon, when the whole mass block is ready to be etched through (namely the structure is released), the area covered by the silicon dioxide is not etched through, and a mass block with grooves is formed. The method has the advantages that:

1. the silicon bonding does not need to carry out icon alignment, so that the bonding process is simple;

2. the alignment precision of the double-sided photoetching is higher than that of silicon-silicon bonding alignment, so that the structural symmetry is better;

3. the thickness of the formed beam is equal to that of the mass block, so that certain error cannot be generated in the output symmetry of the positive and negative 1G of the accelerometer sensor.

Drawings

The invention is further described with reference to the following figures and examples:

FIG. 1 is a flow chart of a wafer fabrication process for a high performance MEMS inertial sensor of the present invention;

FIG. 2 is a cross-sectional view of an accelerometer structure made by the method of manufacture of the invention;

FIG. 3 is a schematic view of a first substrate according to the present invention;

FIG. 4 is a schematic structural diagram of a second substrate according to the present invention;

FIG. 5 is a schematic structural diagram of a second substrate after shallow trench etching in accordance with the present invention;

FIG. 6 is a schematic structural diagram of a second substrate after etching of the fixed electrode according to the present invention;

FIG. 7 is a schematic diagram of the structure of the invention after silicon-silicon bonding;

FIG. 8 is a schematic view of the structure of the present invention after removal of the substrate;

FIG. 9 is a schematic diagram of the structure of the present invention after etching of the oxide layer;

FIG. 10 is a schematic diagram of a structure of the present invention after etching of the structural layer;

FIG. 11 is a schematic diagram of the packaged structure of the present invention;

FIG. 12 is a cross-sectional view of an accelerometer of the symmetrical comb configuration of the present invention;

FIG. 13 is a schematic diagram of the sensitive structure of the ring gyroscope of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings in conjunction with the following detailed description. It should be understood that the description is intended to be exemplary only, and is not intended to limit the scope of the present invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.

Example (b):

as shown in fig. 1, the wafer fabrication process flow of the high-performance MEMS inertial sensor includes the following steps:

providing a first substrate S01: providing a first substrate having a top plane and a bottom plane with an oxide layer disposed below the bottom plane, the bottom plane being substantially parallel to the top plane; the first substrate is silicon-on-insulator (SOI) and the top silicon has a thickness of 50-150 μm.

Providing a second substrate S02: providing a second substrate having an upper planar surface; the second substrate is silicon-on-insulator (SOI) having a silicon thickness of about 5-15 μm.

Shallow cavity etching S03: the second substrate is lithographically patterned, a shallow cavity etch is performed, a portion of the second substrate is etched from the upper planar surface to a first predetermined depth (3-10 μm) to form a shallow cavity of 3-10 μm, and a plurality of protrusions, each protrusion having an upper planar surface, are then Deep Reactive Ion Etched (DRIE) to form the fixed electrode.

Silicon-silicon bonding S04: melt bonding a top planar surface of the first substrate to the raised upper planar surface to form an anchor portion;

removing the substrate silicon of the first base plate S05: removing the substrate of the first substrate, leaving an oxide layer (silicon dioxide);

oxide layer etch S06: etching off the oxide layer at a certain position of the first substrate;

structural layer etching S07: the first substrate is photo-patterned and released by DRIE etching to release the rotatable sensitive structure. A structural pattern is etched from the upper surface of the mass block by using a double-sided lithography technique, a part of the first substrate is etched from the bottom surface and the top surface to a second predetermined depth which is at least equal to the total thickness of the first substrate, and since the oxide layer is etched at a speed much slower than that of silicon, when the oxide layer is etched through (i.e. structure is released) in a place not covered by the oxide layer, the mass block with the groove is not etched through in a place covered by the oxide layer, so that a sensitive structure with a structure capable of freely rotating is formed on the first substrate, wherein the sensitive structure comprises the mass block with a structure at the periphery of the anchoring part. The sensitive structure thus allows the formation of asymmetric proof masses during the top electrode release etch without the addition of an additional etch step. The released device wafer is then metallized by shadow mask sputtering followed by sintering. The wafer is bonded and sealed to the top wafer by a glass frit at a partial pressure of about 50-100 mTorr.

The method can be used for manufacturing the asymmetric torsional pendulum type accelerometer, and can also be used for manufacturing various accelerometers or gyroscopes with symmetric comb tooth structures or annular structures. The sensitive structure of the asymmetric pendulum accelerometer comprises a substantially hollow mass on the first and second sides of the anchoring portion or a solid mass on the first side and a substantially hollow mass on the second side. The sensitive structure of the accelerometer or gyroscope with the symmetrical comb tooth structure comprises a plurality of mass blocks which are carved on two sides of an anchoring part and connected through a mass block frame forming the comb tooth structure. The sensitive structure of the accelerometer or gyroscope of the annular structure comprises a plurality of ring-shaped harmonic oscillators respectively at the periphery of the anchoring part.

The pendulum type accelerometer is taken as an example and is described in detail with reference to the attached drawings:

a cross-sectional view of a fabricated pendulum accelerometer 100, as shown in figure 2, the accelerometer 100 may comprise a first substrate 102, a second substrate 104 and a capping wafer 106.

First substrate as shown in fig. 3, the first substrate 102 comprises a silicon-on-insulator (SOI) material. The first substrate 102 may include a top silicon layer 108 bonded to an oxide layer 110, the oxide layer 110 in turn bonded to a bottom silicon layer 112. In one exemplary embodiment, the oxide layer 110 is made of silicon dioxide. The bottom silicon layer 112 may be used as a handle wafer during the fabrication process. The top silicon layer 108 includes a top planar surface 114 and a bottom planar surface 116 substantially parallel to the top planar surface 114.

Second substrate 104 as shown in fig. 4, the second substrate 104 may be a silicon-on-insulator (SOI) substrate including an upper silicon layer 124, a buried oxide layer 126, and a lower silicon layer 128. The second substrate 104 may further include a handle oxide layer 142 formed on the silicon substrate, the handle oxide layer 142 being on a lower surface of the lower silicon layer 128. The upper silicon layer 124 of the second substrate 104 includes an upper planar surface 130, which may be significantly less than the thickness of the upper silicon layer 108 of the first substrate 102. The upper silicon layer 124 of the second substrate 104 has a thickness of about 5-15 μm and the upper silicon layer 108 of the first substrate 102 has a thickness of about 50-150 μm.

As shown in fig. 5, which may include applying a nitride mask layer using low pressure chemical vapor deposition (L PCVD) to select areas of the upper planar surface 130, the portions of the upper silicon layer 124 not covered by the nitride layer will be etched to a predetermined depth, which may be 3-10 μm, an etch process may be formed using a KOH etch to form a plurality of shallow cavities 134 having a rectangular cross section and a plurality of protrusions 132, each second protrusion 132 having a rectangular cross section and a planar upper surface 136 substantially parallel to the wafer, the shallow cavities 134 overlying the buried oxide layer 126.

The second substrate 104 is further lithographically patterned as shown in fig. 6. Selected portions of the upper silicon layer 124 are etched to expose the buried oxide layer 126. The etching step may include applying a nitride mask using Deep Reactive Ion Etching (DRIE) to deposit a silicon layer and selectively remove portions of the upper silicon layer 124. The portion of the upper silicon layer 124 within the shallow cavity 132 is etched down to the buried oxide layer 126 of the second substrate 104 to form a plurality of protrusions, i.e., fixed electrodes 140, having rectangular cross-sections extending upward from the buried oxide layer 126. An oxide layer 148 is deposited on the surface of the fixed electrode 140. To protect the fixed electrode from etching during subsequent structure release.

As shown in fig. 7, the first substrate 102 is silicon bonded to the second substrate 104. The top planar surface 114 of the top silicon layer 108 of the first substrate 102 is bonded to the upper planar surface 136 of the bump 132. The middle protrusion 132 is bonded with a middle portion of the top planar surface 114 to form an anchor portion 152 of the sensitive structure of the accelerometer 100.

As shown in fig. 8, polishing and etching may be used to remove the bottom silicon layer 112 of the first substrate 102 to expose the oxide layer 110. The oxide layer 110 is a buried oxide layer inherent to the SOI silicon wafer, the thickness of the buried oxide layer is generally fixed to 2 μm, and the thickness of the oxide layer needs to be determined according to the depth of a groove to be etched since the thickness of the oxide layer is related to the depth of the groove to be etched later. Typically less than 1 μm of oxide is required and therefore thinning of the buried oxide layer is required to reduce the thickness of the buried oxide layer. Or the original buried oxide layer can be etched away, and a new oxide layer with a certain thickness can be grown again.

As shown in fig. 9, the first photolithography of the structural layer etches away the oxide layer at a certain position of the first substrate, i.e. removes the oxide layer 122 that needs to be etched through the structural portion.

As shown in fig. 10, a double-sided lithography is used to pattern a structure from the top surface of the upper silicon layer 108 of the first substrate 102, and a portion of the upper silicon layer 108 is etched simultaneously from the bottom surface and the top surface to a predetermined depth, which is at least equal to the total thickness of the first substrate, since the oxide layer is much slower than the silicon etch rate, and when etching through where there is no oxide layer (i.e., structure release), there is no etching through where there is an oxide layer, resulting in a grooved proof mass. The result is a sense structure 154 having a solid mass 156 above the left fixed electrode surface and a substantially hollow mass 158 above the right fixed electrode surface. The solid mass 156 and the substantially hollow mass 158 are connected by the anchoring portion 152 such that the sensitive structure 154 can rotate freely around the anchoring portion 152.

In order to allow the sensitive structure 154 to rotate freely around the anchoring portion 152 better, parts of the surface are etched away on both sides of the anchoring portion 152 of the first substrate 102, respectively, to form beam structures 153 for connecting the masses, i.e. narrowing where the anchoring portion 152 connects the masses on both sides.

In an exemplary embodiment, each half of the sensitive structure 154 has substantially the same dimensions. That is, the solid mass 156 and the substrate hollow mass 158 have substantially the same length and width, respectively. The substantially hollow mass 158 may include a substantially planar bottom 160 with sidewalls 162 extending upwardly from the bottom 160 to form one or more cavities 164. The side wall 162 may include at least two walls that intersect at a substantially right angle to form a plurality. In one exemplary embodiment, the substantially hollow mass 158 includes four rectangular cavities 164 separated by sidewalls 162 that are substantially perpendicular to the bottom 160. Other shapes and sizes of the cavity 164 are also within the scope of the present invention.

In another embodiment, the sensitive structure 154 may comprise a mass forming a substantial hollow on both sides of the anchoring portion, and the masses of the two sides are not equal.

Preferably, the substantially hollow mass 158 allows for center of mass shift while maintaining equal surface area on the bottom surface of each side of the sensitive structure 154, which may balance gas damping. The center of mass offset can be adjusted by adjusting the thickness of the sensitive structure 154 and/or adjusting the depth of the cavity 164.

The anchor point can be located in the middle of the left solid mass block and the right hollow mass block, or can be located in other positions, or a plurality of anchor points can be distributed in other positions, depending on the needs of the sensitive structure (the mass block can be hollow or can also be a comb tooth structure).

As shown in fig. 11, the capping wafer 106 is made of silicon and etched to a predetermined depth to form a top recess 170. A layer 172 of sputter deposited by getter may be applied to a portion of the surface of the cap wafer 106 within the top recess 170 and a frit bond 174 may be applied to the surface of the cap wafer 106 to be bonded to the structural wafer. The capping wafer and the structural wafer are bonded using frit under high vacuum conditions to complete the vacuum bonding.

In another embodiment, the accelerometer or gyroscope with the symmetrical comb-tooth structure is manufactured by the same process as the above method, but the manufactured sensitive structure is different and is the comb-tooth structure. As shown in fig. 12, the sensing structure 200 includes an anchoring portion 201, and a plurality of mass frames 202 formed by cutting through the middle portions of two sides of the anchoring portion 201, wherein the mass frames 200 at two sides are staggered to form a comb-tooth structure. The structure is shielded by photoresist at the position where the mass block frame needs to be formed through a photoetching process, and is etched through at the position where the mass block frame does not exist.

In another embodiment, the method of the present invention can also be used for manufacturing an accelerometer or gyroscope with a ring-shaped structure, and the manufacturing process is the same as that of the above, except that the manufactured sensitive structure is different and is the ring-shaped structure. As shown in fig. 13, the sensing structure 300 includes an anchoring portion 301, and a plurality of ring-shaped mass frames 302 are respectively formed around the anchoring portion 301, the mass frames 302 may be circular rings or regular polygons, and the mass frames 302 are connected by spokes 303. The structure is masked by photoresist at the place where the mass frame and the spokes need to be formed through a photoetching process.

It is to be understood that the above-described embodiments of the present invention are merely illustrative of or explaining the principles of the invention and are not to be construed as limiting the invention. Therefore, any modification, equivalent replacement, improvement and the like made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention. Further, it is intended that the appended claims cover all such variations and modifications as fall within the scope and boundaries of the appended claims or the equivalents of such scope and boundaries.

Claims

1. A method of manufacturing a MEMS inertial sensor, comprising the steps of:

s02: providing a second substrate having an upper planar surface;

s05: etching off the oxide layer at a certain position of the first substrate;

2. The manufacturing method according to claim 1, wherein the total thickness of the first substrate is 50 to 150 μm.

3. The manufacturing method according to claim 2, wherein the first predetermined depth is 3 to 10 μm.

4. The manufacturing method according to claim 1, further comprising, after the step S03:

and depositing an oxide layer on the surface of the fixed electrode.

5. The manufacturing method according to claim 1, further comprising, after the step S08, the steps of:

providing a cover wafer; and

etching a portion of the cap wafer to form a top recess; and

6. The method of manufacturing according to claim 5, further comprising depositing a thin film on a portion of the surface of the recess by getter sputtering.

7. The method of manufacturing of claim 5, further comprising printing a glass paste on the contact surface of the cap wafer with glass frit printing;

8. A method of manufacturing according to claim 1, wherein portions of the surface are etched away on both sides of the anchoring portion of the first substrate to form beam structures for connecting the proof mass.

9. A method of manufacturing according to any of claims 1-8, wherein the sensitive structure is a hollow structure, a comb-tooth structure or a ring structure.