JPH10270715A - Manufacture of semiconductor inertia sensor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a semiconductor inertia sensor with the low cost which does not require laser processing for a wafer and is suitable for high volume production, which is low in parasitic capacitance and excellent in size accuracy. SOLUTION: A polysilicon layer 23 is formed on opposite surfaces of a laminate comprising a second oxide film 42, a single crystal silicon layer 20, a first oxide film 41, and a second silicon wafer 22. The polysilicon layer connected with the second oxide film is selectively removed to form a spacer layer 25 composed of the polysilicon layer and the second oxide film. The exposed single crystal silicon layer is selectively removed to form a movable electrode 26 comprising the single crystal silicon and fixed electrodes 27, 28 comprising the single crystal silicon. A structure 43 including the movable electrode, the fixed electrodes, and the spacer electrode is joined with a glass substrate 10 through the spacer layer. The second silicon wafer and the polysilicon layer on the first oxide film are removed, and then the first oxide film is removed to configure a semiconductor inertia sensor 30.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of manufacturing a semiconductor inertial sensor suitable for a capacitance type acceleration sensor, angular velocity sensor and the like.

[0002]

2. Description of the Related Art Conventionally, as this kind of semiconductor inertial sensor, a resonance angular velocity sensor having a structure of a glass substrate and single crystal silicon has been proposed (M. Hashimoto et al.,
"Silicon Resonant Angular Rate Sensor", Techinica
l Digest of the 12th Sensor Symposium, pp.163-166
(1994)). This sensor has a movable electrode with a tuning fork structure that floats on both sides with a torsion bar. This movable electrode is excited by electromagnetic drive. When the angular velocity acts, a Coriolis force is generated on the movable electrode, and the movable electrode causes torsional vibration around the torsion bar to resonate. The sensor detects an angular velocity that acts due to a change in capacitance between the movable electrode and the detection electrode due to resonance of the movable electrode. When this sensor is manufactured, a structure such as a movable electrode portion is manufactured by etching a single crystal silicon substrate having a thickness of about 200 μm and having a crystal orientation of (110) perpendicular to the substrate surface. In order to vertically etch this relatively thick silicon substrate, anisotropic dry etching using SF ₆ gas is performed, or after drilling a YAG laser at the corner of the base of the torsion bar to the movable electrode portion, ,
Wet etching is performed with KOH or the like. The etched silicon substrate is integrated with the glass substrate by anodic bonding.

As another semiconductor inertial sensor, after a sacrificial layer is patterned on a silicon substrate by etching,
A micro-gyro (K. Tanaka et al., "A
micromachined vibrating gyroscope ", Sensors and A
ctuators A 50, pp. 111-115 (1995)). This microgyro has a structure using a so-called surface micromachining technology. In particular,
Forming a detection electrode by impurity diffusion on a silicon substrate,
After forming and patterning a phosphate glass film serving as a sacrificial layer thereon, a polysilicon film is formed, and a process such as vertical etching is performed to form a structure. Finally, by removing the sacrificial layer by etching, the movable electrode portion is cut off to create a gap with respect to the detection electrode, and the movable electrode is brought into a floating state.

As another semiconductor inertial sensor, a gyroscope having a structure of a glass substrate and single crystal silicon has been proposed (J. Bernstein et al., "A Microma
chined Comb-Drive Tuning Fork Rate Gyroscope ", IEE
E MEMS '93 Proceeding, pp.143-148 (1993)). This gyroscope has a glass substrate on which a detection electrode is formed,
After the etching, high-concentration boron diffusion is performed to form a movable electrode, a fixed electrode, etc., on a single-crystal silicon substrate, where the boron-diffused portion is used as a bonding surface, and further, the silicon substrate portion where boron is not diffused. Is removed by etching.

[0005]

The above-mentioned conventional techniques for manufacturing a sensor have the following disadvantages. In the method of manufacturing the resonance angular velocity sensor described above, the silicon active portion which should be a structure floating with respect to the glass substrate sometimes adheres to the glass substrate due to electrostatic attraction during anodic bonding and does not become a movable electrode. In order to prevent this sticking, the movable electrode and the detection electrode are short-circuited and anodic-bonded in a state where electrostatic force does not work, and then the short-circuited electrodes are separated using a laser. In addition, it is necessary to perform laser-assisted etching after bonding to a glass substrate to form an island-shaped fixed electrode. These laser processes are extremely complicated and unsuitable for mass production of sensors.

[0006] In the micro gyro, since a silicon wafer is used as a substrate, the parasitic capacitance of the sensor is large, and it is difficult to increase sensitivity and accuracy. Further, in the gyroscope manufacturing method, since the entire structure is formed by using the portion where boron is diffused as an etch stop portion, if the etch stop effect is incomplete, the thickness of the movable electrode and the fixed electrode is reduced by overetching. Therefore, there is a problem that the dimensional accuracy is inferior. Further, since the gap between the movable electrode and the detection electrode depends only on the control by the etching time, there is a problem in the accuracy of forming the gap between the electrodes.

An object of the present invention is to provide a method of manufacturing a low-cost semiconductor inertial sensor which does not require laser processing and is suitable for mass production. Another object of the present invention is to provide a method of manufacturing a semiconductor inertial sensor with low parasitic capacitance, high sensitivity and high accuracy. Still another object of the present invention is to provide a method for manufacturing a semiconductor inertial sensor having excellent dimensional accuracy.

[0008]

The invention according to claim 1 is
As shown in FIG. 1, a first silicon wafer 2 having a first film 41 that can be etched without eroding silicon on both surfaces.
Bonding a second silicon wafer 22 to 1;
Polishing one surface of the first silicon wafer 21 to a predetermined thickness to form a single-crystal silicon layer 20, and forming a second film 42 on the single-crystal silicon layer 20 that can be etched without eroding silicon. A step of forming the polysilicon layer 23 on the second film 42;
The second film 42 and the polysilicon layer 23 sequentially laminated thereon
Forming a spacer layer 25 comprising a second film 42 and a polysilicon layer 23 laminated thereon on the single crystal silicon layer 20 by selectively etching away the single crystal silicon layer. 20 is selectively removed by etching, whereby a movable electrode 26 made of single crystal silicon connected to the first film 41 and a pair of fixed electrodes 27 and 28 made of single crystal silicon connected to the first film 41 and the spacer layer 25 are formed. Forming the second silicon wafer 22, the first film 41, the movable electrode 26, and the fixed electrode 2
Bonding the structure 43 having the spacers 25 to the glass substrate 10 via the spacer layer 25 so that the movable electrode 26 faces the glass substrate 10; A step of etching and removing the film 41 as an etch stop layer, and a step of etching and removing the first film 41 and a movable electrode sandwiched between the pair of fixed electrodes 27 and 28 and floating above the glass substrate 10. Semiconductor inertial sensor 3 having
And a step of obtaining 0.

According to a second aspect of the present invention, as shown in FIG. 4, a step of forming a detection electrode 12 on a glass substrate 10 and a first step capable of etching both sides without eroding silicon.
A step of bonding the second silicon wafer 22 to the first silicon wafer 21 having the film 41, a step of polishing one surface of the first silicon wafer 21 to a predetermined thickness to form the single-crystal silicon layer 20, A step of forming a second film 42 that can be etched without eroding silicon on the silicon layer 20, a step of forming a polysilicon layer 23 on the second film 42, and a step of sequentially laminating on the single crystal silicon layer 20 The removed second film 42 and polysilicon layer 23 are selectively removed by etching, whereby the spacer layer 25 composed of the second film 42 and the polysilicon layer 23 laminated thereon is formed on the single crystal silicon layer 20. Forming step and selectively removing the exposed single-crystal silicon layer 20 by etching, thereby forming a movable electrode 26 made of single-crystal silicon connected to the first film 41. A step, a second silicon wafer 22
Bonding the structure 43 having the first electrode 41, the first electrode 41, the movable electrode 26, and the spacer layer 25 to the glass substrate 10 via the spacer layer 25 such that the movable electrode 26 faces the detection electrode 12; The silicon wafer 22 is placed on the first film 41
Etching away as an etch stop layer,
The movable electrode 26 floating opposite to the detection electrode 12 on the glass substrate 10 by etching away the first film 41.
Obtaining a semiconductor inertial sensor 40 having the following formula:

According to a third aspect of the present invention, as shown in FIG. 5, a step of forming a detection electrode 12 on a glass substrate 10 and a first step capable of etching both sides without eroding silicon.
A step of bonding the second silicon wafer 22 to the first silicon wafer 21 having the film 41, a step of polishing one surface of the first silicon wafer 21 to a predetermined thickness to form the single-crystal silicon layer 20, A step of forming a second film 42 that can be etched without eroding silicon on the silicon layer 20, a step of forming a polysilicon layer 23 on the second film 42, and a step of sequentially laminating on the single crystal silicon layer 20 The removed second film 42 and polysilicon layer 23 are selectively removed by etching, whereby the spacer layer 25 composed of the second film 42 and the polysilicon layer 23 laminated thereon is formed on the single crystal silicon layer 20. Forming step and selectively removing the exposed single-crystal silicon layer 20 by etching, thereby forming the movable electrode 26 made of single-crystal silicon connected to the first film 41 and the first film Forming a pair of fixed electrodes 27 and 28 made of single crystal silicon to be connected to the first and spacer layers 25, respectively, a second silicon wafer 22, a first film 41, a movable electrode 26, fixed electrodes 27 and 28, and a spacer. Bonding the structure 43 having the layer 25 to the glass substrate 10 via the spacer layer 25 so that the movable electrode 26 faces the detection electrode 12, and etching the second silicon wafer 22 to the first film 41. A step of etching and removing the layer as a layer, and a step of etching and removing the first film 41 to form a pair of fixed electrodes 27 and 28 and the fixed electrodes 27 and 28.
Semiconductor inertial sensor 5 having a movable electrode 26 which is sandwiched between and floats on a glass substrate 10 so as to face the detection electrode 12.
And a step of obtaining 0.

In the manufacturing method according to the first to third aspects, since laser processing of the wafer is unnecessary and suitable for mass production, a semiconductor inertial sensor can be manufactured at low cost. The structural members of the structure 43 are a second silicon wafer 22 having a sufficient thickness and the first silicon wafer 21 bonded thereto.
Of the structure is polished to a predetermined thickness, so that the thickness of the structure can be freely selected. Accordingly, handling when the structure 43 is bonded to the glass substrate 10 becomes easy, and the thickness of the movable electrode 26 and the like can be set freely. Also, since a glass substrate is used for the substrate,
The resulting semiconductor inertial sensor has low parasitic capacitance. Furthermore, the gap between the electrodes is not controlled by the etching time, but is determined by the thickness of the spacer layer 25 composed of the second film 42 and the polysilicon layer 23, so that the dimensional accuracy is excellent. For this reason, a highly accurate semiconductor inertial sensor is manufactured.

In this specification, "a film which can be etched without eroding silicon" means that an etchant which does not corrode silicon when etching the film can be selected. . Examples of the film having such properties include an oxide film and a nitride film. In the present invention, the crystal orientation of the second silicon wafer 22 is preferably a (110) orientation in consideration of the etching rate.

[0013]

Embodiments of the present invention will be described in detail with reference to the drawings. As shown in FIGS. 1 and 2, a semiconductor inertial sensor 30 according to a first embodiment of the present invention is an acceleration sensor, and a second film 42 and a polysilicon which can be etched on a glass substrate 10 without eroding silicon. Fixed electrode 27 fixed via spacer layer 25 composed of layer 23
And 28 have a movable electrode 26. Movable electrode 26,
The fixed electrodes 27 and 28 are each made of single-crystal silicon, and opposing portions of the electrodes 26 and 27 and the electrodes 26 and 28 are formed in a comb shape. Movable electrode 26
Are supported at both ends by beams 31, and are floating with respect to the glass substrate 10. The base 31 a of the beam 31 is fixed on the glass substrate 10 via the second film 42 constituting the spacer layer 25 and the polysilicon layer 23. Although not shown, the beam base end 31a, the fixed electrode 2
7 and 28 are individually provided with electrical wiring. In the semiconductor inertial sensor 30, when a horizontal acceleration perpendicular to a line connecting the beam base ends 31a and 31a acts on the movable electrode 26 as shown by arrows in the figure, the movable electrode 26 applies the beams 31, 31. Vibrates around the shaft. Movable electrode 26
When the distance between the movable electrode 26 and the fixed electrodes 27 and 28 increases or decreases, the capacitance between the movable electrode 26 and the fixed electrodes 27 and 28 changes. The acceleration that acts is obtained from the change in the capacitance.

Next, a method of manufacturing the semiconductor inertial sensor 30 according to the first embodiment of the present invention will be described. As shown in FIG. 1, the first silicon wafer 21 is thermally oxidized to form first oxide films 41 on both surfaces thereof. Next, the second silicon wafer 22 is bonded to the first silicon wafer 21 via the first oxide film 41. The surface of the first silicon wafer 21 to which the second silicon wafer 22 is not bonded is ground and polished to a predetermined thickness using a grindstone and a polishing cloth together with the first oxide film 41 formed thereon. A crystalline silicon layer 20 is formed. A second oxide film is formed on the second silicon wafer 22 and the single-crystal silicon layer 20 by thermal oxidation in the same manner as in the case of the first oxide film 41. After the second oxide film 42 on the second silicon wafer 22 is removed by dry etching using a gas such as CF ₄ , the second oxide film 42 remaining on the single crystal silicon layer 20 and the exposed second silicon wafer 22 A polysilicon layer 23 is formed on the substrate by a CVD method. After a nitride film 35 is formed on both of the polysilicon layers 23 by a CVD method, the nitride film 35 is patterned and formed on the second oxide film 42.
Expose a predetermined portion of The exposed polysilicon layer 23 and the second oxide film 42 formed thereunder are removed by etching, whereby the spacer layer 25 composed of the second oxide film 42 and the polysilicon layer 23 laminated thereon is removed.
Is formed on the single crystal silicon layer 20. Spacer layer 2
An Al layer 36 is selectively formed on the surface of the single crystal silicon layer 20 by sputtering and patterning, followed by low-temperature anisotropic dry etching using SF ₆ gas. As a result, the single crystal silicon layer 20 is selectively etched using the first oxide film 41 as an etch stop layer. As a result, the movable electrode 26 made of single crystal silicon is formed on the first oxide film 41, and the movable electrode 26 A pair of fixed electrodes 27 and 28 made of single crystal silicon are formed on both sides of the substrate with a slight gap. After removing the Al layer 36, the structure 43 having the polysilicon layer 23, the second silicon wafer 22, the first oxide film 41, the movable electrode 26, the fixed electrodes 27 and 28, and the spacer layer 25 is replaced with the movable electrode 26 made of glass. Anodically bond to the glass substrate 10 via the spacer layer 25 so as to face the substrate 10. Subsequently, the polysilicon layer 23 and the second silicon wafer 22 are etched and removed by an etchant such as KOH using the first oxide film 41 as an etch stop layer. Next, the first oxide film 41 is etched away by wet etching using an etchant such as hydrofluoric acid or dry etching using an etchant such as CF ₄ . As a result, a semiconductor inertial sensor 30 in which the movable electrode 26 made of single-crystal silicon is sandwiched between the pair of fixed electrodes 27 and 28 made of single-crystal silicon and formed above the glass substrate 10 so as to float is obtained.

FIGS. 3 and 4 show a semiconductor inertial sensor 40 according to a second embodiment. The semiconductor inertial sensor 40 is an acceleration sensor, and has a second film 42 and a polysilicon layer 23 which can be etched on the glass substrate 10 without eroding silicon.
Frame 29 fixed via a spacer layer 25 consisting of
And a movable electrode 26 between them. Movable electrode 26 and frame 2
9 are made of single-crystal silicon,
Are accommodated in a window frame-shaped frame body 29 at intervals. The movable electrode 26 is supported at both ends by beams 31, 31.
It is floating with respect to the glass substrate 10. Beam 31
The base end 31a is located in the recess 29a of the frame 29 and is fixed on the glass substrate 10 via the spacer layer 25 composed of the second film 42 and the polysilicon layer 23. The detection electrode 12 having a thickness smaller than the thickness of the spacer layer 25 is formed on the glass substrate 10. Although not shown, the beam base end 3
Electrical wiring is individually made to the 1a and the detection electrode 12.
In this semiconductor inertial sensor 40, beam base ends 31a and 31a are
When a vertical acceleration perpendicular to the line connecting is applied, the movable electrode 26 vibrates around the beams 31 and 31 as a supporting axis.
The distance between the movable electrode 26 and the detection electrode 12 increases,
When the width is reduced, the capacitance between the movable electrode 26 and the detection electrode 12 changes. The acceleration that acts is obtained from the change in the capacitance.

Next, a method of manufacturing the semiconductor inertial sensor 40 according to the second embodiment of the present invention will be described. As shown in FIG. 4, first, a detection electrode 12 made of a thin film of a metal selected from Au, Pt, Cu, or the like is formed on a glass substrate 10 by sputtering, vacuum deposition, or the like. On the other hand, the first silicon wafer 21 is formed in the same manner as in the manufacturing method of the first embodiment.
A first oxide film 41 is formed on both surfaces of the first oxide film. Next, the second silicon wafer 22 is bonded to the first silicon wafer 21 via the first oxide film 41. Second silicon wafer 2
First silicon wafer 2 on the side where 2 is not bonded
The surface of 1 is ground and polished to a predetermined thickness using a grindstone and a polishing cloth together with the first oxide film 41 formed thereon to form the single crystal silicon layer 20. A second oxidation is performed on the second silicon wafer 22 and the single crystal silicon layer 20 by thermal oxidation.
The oxide film 42 is formed in the same manner as the first oxide film 41. The second oxide film 42 on the second silicon wafer 22 is
After being removed by dry etching with a gas such as F ₄ , the second
A polysilicon layer 23 is formed on the oxide film 42 and the exposed second silicon wafer 22 by a CVD method. After a nitride film 35 is formed on each of the two polysilicon layers 23 by the CVD method, a predetermined portion of the polysilicon layer 23 formed on the second oxide film 42 is exposed by patterning. This exposed polysilicon layer 2
3 and the second oxide film 42 formed thereunder are removed by etching, whereby a pair of spacer layers 25 composed of the second oxide film 42 and the polysilicon layer 23 laminated thereon are formed of single crystal silicon. Formed on layer 20. An Al layer 36 is selectively formed on the surfaces of the spacer layer 25 and the single-crystal silicon layer 20 by sputtering and patterning, followed by low-temperature anisotropic dry etching using SF ₆ gas. As a result, the single crystal silicon layer 20 is selectively etched using the first oxide film 41 as an etch stop layer. As a result, the movable electrode 26 made of single crystal silicon is formed on the first oxide film 41, and the movable electrode 26 A frame 29 made of single crystal silicon is formed with a gap on both sides of the frame 29. After removing the Al layer 36, the movable electrode 26 is moved to the detection electrode 12 by the structure 43 having the polysilicon layer 23, the second silicon wafer 22, the first oxide film 41, the movable electrode 26, the frame 29, and the spacer layer 25. Is anodically bonded to the glass substrate 10 with the spacer layer 25 interposed therebetween. Subsequently, the polysilicon layer 23 and the second silicon wafer 22 are etched and removed by an etchant such as KOH using the first oxide film 41 as an etch stop layer. Next, the first oxide film 41 is etched away by wet etching using an etchant such as hydrofluoric acid or dry etching using an etchant such as CF ₄ . As a result, a semiconductor inertial sensor 40 in which the movable electrode 26 made of single-crystal silicon is sandwiched between the frame bodies 29 made of single-crystal silicon and formed above the detection electrode 12 so as to float is obtained.

FIGS. 5 and 6 show a semiconductor inertial sensor 50 according to a third embodiment. The semiconductor inertial sensor 50 is an angular velocity sensor, and the second film 42 and the polysilicon layer 23 that can be etched on the glass substrate 10 without eroding silicon.
A pair of movable electrodes 26, 2 having a tuning fork structure is fixed between fixed electrodes 27, 28 fixed via a spacer layer 25 composed of
6. The movable electrode 26, the fixed electrodes 27 and 28
Each of the electrodes 26 and 27 is made of single crystal silicon.
Opposite portions of the electrode 26 and the electrode 28 are formed in a comb shape. The movable electrodes 26, 26 have a U-shaped beam 3.
Both ends are supported by 1 and 31 and float with respect to the glass substrate 10. The base 31 a of the beam 31 is fixed on the glass substrate 10 via the spacer layer 25 including the second film 42 and the polysilicon layer 23. The detection electrode 12 having a thickness smaller than the thickness of the spacer layer 25 is formed on the glass substrate 10. Although not shown, electric wires are individually provided to the beam base 31a, the fixed electrodes 27 and 28, and the detection electrode 12, and an alternating voltage is applied to the fixed electrodes 27 and 28 so that the movable electrode is excited by electrostatic force. Has become. In the semiconductor inertial sensor 50, the movable electrode 2
When the angular velocities act on the lines connecting the beam base ends 31a and 31a to the movable electrodes 26 and 26, the movable electrodes 26 and 26
Generates a Coriolis force, causing a torsional vibration around the center line and resonating. The angular velocity acting due to the change in the capacitance between the movable electrode 26 and the detection electrode 12 at the time of the resonance is detected.

Next, a method of manufacturing the semiconductor inertial sensor 50 according to the third embodiment of the present invention will be described. As shown in FIG. 5, first, as in the second embodiment, a detection electrode 12 made of a thin film of a metal selected from Au, Pt, Cu, or the like is formed on a glass substrate 10 by sputtering, vacuum deposition, or the like. On the other hand, in the same manner as in the manufacturing method of the first embodiment,
First oxide films 41 are formed on both surfaces of the silicon wafer 21. Next, the second silicon wafer 22 is bonded to the first silicon wafer 21 via the first oxide film 41. The surface of the first silicon wafer 21 to which the second silicon wafer 22 is not bonded is ground and polished to a predetermined thickness using a grindstone and a polishing cloth together with the first oxide film 41 formed thereon. A crystalline silicon layer 20 is formed.
A second oxide film is formed on the second silicon wafer 22 and the single-crystal silicon layer 20 by thermal oxidation in the same manner as in the case of the first oxide film 41. After the second oxide film 42 on the second silicon wafer 22 is removed by dry etching using a gas such as CF ₄ , the second oxide film 42 remaining on the single crystal silicon layer 20 and the exposed second silicon wafer 22 A polysilicon layer 23 is formed on the substrate by a CVD method. The two polysilicon layers 2
After a nitride film 35 is formed on the substrate 3 by the CVD method, a predetermined portion of the polysilicon layer 23 formed on the second oxide film 42 is exposed by patterning. The exposed polysilicon layer 23 and the second oxide film 42 formed thereunder are removed by etching, whereby the spacer layer 25 composed of the second oxide film 42 and the polysilicon layer 23 laminated thereon is removed. Is formed on the single crystal silicon layer 20. An Al layer 36 is formed on the surfaces of the spacer layer 25 and the single crystal silicon layer 20 by sputtering and patterning.
Is formed selectively, followed by low-temperature anisotropic dry etching using SF ₆ gas. Thereby, the first oxide film 4
1 as an etch stop layer, the single crystal silicon layer 20 is selectively etched, and as a result, a movable electrode 26 made of single crystal silicon is formed on the first oxide film 41, and a slight gap is formed on both sides of the movable electrode 26. A pair of fixed electrodes 27 and 28 made of single crystal silicon are formed. A
After removing the first layer 36, the structure 4 including the polysilicon layer 23, the second silicon wafer 22, the first oxide film 41, the movable electrode 26, the fixed electrodes 27 and 28, and the spacer layer 25.
3 is anodically bonded to the glass substrate 10 via the spacer layer 25 so that the movable electrode 26 faces the detection electrode 12. Subsequently, the polysilicon layer 23 and the second silicon wafer 22 are etched and removed by an etchant such as KOH using the first oxide film 41 as an etch stop layer. Next, the first oxide film 41 is etched away by wet etching using an etchant such as hydrofluoric acid or dry etching using an etchant such as CF ₄ . Thus, the movable electrode 26 made of single-crystal silicon is sandwiched between the pair of fixed electrodes 27 and 28 made of single-crystal silicon, and the semiconductor inertial sensor 50 is formed above the detection electrode 12 so as to float.
Is obtained.

FIG. 7 shows a semiconductor inertial sensor 3 according to the first embodiment.
0 illustrates another embodiment of a semiconductor inertial sensor 60. First
The difference from the sensor 30 of the embodiment is that the polysilicon layer 2
In the spacer layer 25 including the third and second oxide films 42,
That is, a part of the second oxide film 42 constituting the spacer 25 is removed to form the opening 42a. This opening 42
Since the fixed electrode 27 joined to the spacer 25 is electrically connected to the polysilicon layer 23 constituting the spacer layer due to the presence of a, the electrode (not shown) on the glass substrate 10 side is connected to the spacer 25. It is possible to electrically connect to the outside via the power supply.

[0020]

As described above, unlike the conventional method of manufacturing a semiconductor inertial sensor by laser processing of a wafer, according to the present invention, laser processing of a wafer is unnecessary and a low-cost semiconductor inertial sensor suitable for mass production. Can be manufactured. By using a glass substrate instead of a silicon substrate for a sensor that performs detection using capacitance, the parasitic capacitance of the element is reduced, and a highly sensitive and accurate semiconductor inertial sensor can be obtained. Since the single-crystal silicon layer forming the movable electrode, the fixed electrode, the frame, or the like is bonded to the silicon substrate while being supported by the second silicon wafer having a desired thickness, a sticking phenomenon as in the related art occurs. And the movable electrode can be provided at a predetermined gap with respect to the detection electrode and the silicon substrate. Further, since the movable electrode and the like are made of single-crystal silicon, they have excellent mechanical properties as compared with polycrystalline silicon and metal. Further, since the gap between the movable electrode and the substrate is defined by the thickness of the spacer layer without being controlled by the etching time, the gap can be formed with high precision.

[Brief description of the drawings]

FIG. 1 is a cross-sectional view showing a semiconductor inertial sensor according to a first embodiment of the present invention, corresponding to a main part of line AA in FIG. 2, and a manufacturing process thereof.

FIG. 2 is an external perspective view of the semiconductor inertial sensor according to the first embodiment of the present invention.

FIG. 3 is an external perspective view of a semiconductor inertial sensor according to a second embodiment of the present invention.

FIG. 4 is a sectional view showing a semiconductor inertial sensor according to a second embodiment of the present invention corresponding to a main part of line BB in FIG. 3 and a manufacturing process thereof.

FIG. 5 is a sectional view showing a semiconductor inertial sensor according to a third embodiment of the present invention, corresponding to a main part of line CC in FIG. 6, and a manufacturing process thereof;

FIG. 6 is an external perspective view of a semiconductor inertial sensor according to a third embodiment of the present invention.

FIG. 7 is a sectional view showing another embodiment of the semiconductor inertial sensor according to the first embodiment of the present invention.

[Explanation of symbols]

 Reference Signs List 10 glass substrate 12 detection electrode 20 single-crystal silicon layer 21 first silicon wafer 22 second silicon wafer 23 polysilicon layer 25 spacer layer 26 movable electrode 27, 28 pair of fixed electrodes 41 first film 42 second film 30, 40, 50,60 Semiconductor inertial sensor

Claims (3)

[Claims] 1. a step of bonding a second silicon wafer (22) to a first silicon wafer (21) having a first film (41) that can be etched without eroding silicon on both sides; A step of polishing one surface of (21) to a predetermined thickness to form a single-crystal silicon layer (20); and a second film that can be etched on the single-crystal silicon layer (20) without eroding silicon. Forming a polysilicon layer (23) on the second film (42); and forming the second layer sequentially on the single-crystal silicon layer (20).
The film (42) and the polysilicon layer (23) are selectively removed by etching, whereby the spacer layer (25) comprising the second film (42) and the polysilicon layer (23) laminated thereon is formed. )
Forming a step on the single crystal silicon layer (20) and selectively etching away the exposed single crystal silicon layer (20),
Thereby, a movable electrode (26) made of single-crystal silicon connected to the first film (41) and a pair of fixed electrodes (single-crystal silicon) connected to the first film (41) and the spacer layer (25) are formed. Forming the second silicon wafer (22), the first film (41), the movable electrode (26), the fixed electrodes (27, 28), and the spacer layer.
(25) bonding the structure (43) having the movable electrode (26) to the glass substrate (10) via the spacer layer (25) such that the movable electrode (26) faces the glass substrate (10); Etching the second silicon wafer (22) using the first film (41) as an etch stop layer; and etching and removing the first film (41) to form the pair of fixed electrodes (27, 28). The movable electrode (2) sandwiched between the fixed electrodes (27, 28) and floating above the glass substrate (10)
6) obtaining a semiconductor inertial sensor (30) having the following steps: A step of forming a detection electrode on a glass substrate; and a step of etching a first film on both sides without eroding silicon.
1) bonding a second silicon wafer (22) to a first silicon wafer (21) having (1); polishing one surface of the first silicon wafer (21) to a predetermined thickness to form a single-crystal silicon layer (20); Forming a second film (42) that can be etched without eroding silicon on the single crystal silicon layer (20); and forming a polysilicon layer on the second film (42). Forming a (23); and the second layer sequentially laminated on the single-crystal silicon layer (20).
The film (42) and the polysilicon layer (23) are selectively removed by etching, whereby the spacer layer (25) comprising the second film (42) and the polysilicon layer (23) laminated thereon is formed. )
Forming a single crystal silicon layer on the single crystal silicon layer (20), and selectively removing the exposed single crystal silicon layer (20) by etching to thereby form a single crystal silicon layer connected to the first film (41). Forming a movable electrode (26); a structure (4) having the second silicon wafer (22), the first film (41), the movable electrode (26), and the spacer layer (25).
3) bonding the second silicon wafer (22) to the glass substrate (10) via the spacer layer (25) such that the movable electrode (26) faces the detection electrode (12). Etching the first film (41) as an etch stop layer; and etching the first film (41) to float on the glass substrate (10) in opposition to the detection electrode (12). Obtaining a semiconductor inertial sensor (40) having the movable electrode (26). 3. A step of forming a detection electrode (12) on a glass substrate (10), and a first film (4) which can be etched on both sides without eroding silicon.
1) bonding a second silicon wafer (22) to a first silicon wafer (21) having (1); polishing one surface of the first silicon wafer (21) to a predetermined thickness to form a single-crystal silicon layer (20); Forming a second film (42) that can be etched without eroding silicon on the single crystal silicon layer (20); and forming polysilicon on the second film (42). Forming a layer (23); and forming the second layer sequentially laminated on the single crystal silicon layer (20).
The film (42) and the polysilicon layer (23) are selectively removed by etching, whereby the spacer layer (25) comprising the second film (42) and the polysilicon layer (23) laminated thereon is formed. )
Forming a single crystal silicon layer on the single crystal silicon layer (20), and selectively removing the exposed single crystal silicon layer (20) by etching to thereby form a single crystal silicon layer connected to the first film (41). Forming a movable electrode (26) and a pair of fixed electrodes (27, 28) made of single crystal silicon connected to the first film (41) and the spacer layer (25), respectively; (22), the first film (41), the movable electrode (26), the fixed electrodes (27, 28), and the spacer layer
Bonding a structure (43) having a (25) to the glass substrate (10) via the spacer layer (25) such that the movable electrode (26) faces the detection electrode (12), Etching the second silicon wafer (22) using the first film (41) as an etch stop layer; and etching and removing the first film (41) to form the pair of fixed electrodes (27, 28). Semiconductor inertial sensor (5) having the movable electrode (26) sandwiched between the fixed electrodes (27, 28) and floating on the glass substrate (10) in opposition to the detection electrode (12).
0) obtaining a semiconductor inertial sensor.