JP2000161964A - Semiconductor device and its manufacture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the amount of oxygen flowing into an airtight space in an anodic bonding operation and to form an airtight space, whose vacuum is high in a semiconductor device comprising the airtight space between a silicon substrate and a glass lid. SOLUTION: In this semiconductor device, a nearly vacuum airtight space S is formed between a substrate 10 and a glass lid 60 and a bonding preventive film 70 is formed on a bonding face to a frame body 20. Since the bonding preventive film 70 prevents the glass lid 60 from being bonded to the frame body 20, the bonding area of the glass lid 60 to the frame body 20 is reduced, and the amount of oxygen which is generated from their bonding part is reduced. The generated oxygen flows into a gap SS, which is formed of the bonding preventive film 70 situated between the inside face and the outside face in the width direction of the bonding part. Thereby, the vacuum in the airtight space S can be kept high.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor device in which a glass lid is anodically bonded to a silicon substrate to form a substantially vacuum sealed space between the silicon substrate and the glass lid, and a method of manufacturing the same.

[0002]

2. Description of the Related Art Conventionally, for example, Japanese Unexamined Patent Publication No.
As disclosed in Japanese Patent Application Laid-Open No. 8 (1999) -1996, in order to simultaneously manufacture a large number of semiconductor pressure sensors having a sealed space between a silicon substrate and a glass lid, a large number of semiconductor pressure sensors having a concave portion for the sealed space are formed of silicon. A large number of semiconductor pressures are formed by forming a glass substrate on a substrate wafer, overlaying a glass substrate on the silicon substrate wafer, and then performing anodic bonding, and then dicing the semiconductor wafer along a predetermined dicing line. The sensor was cut out. In this case, when the glass substrate and the silicon substrate wafer are stacked before the anodic bonding, dicing is performed to avoid manufacturing a pressure sensor having non-uniform characteristics due to the close contact between the glass substrate and the silicon substrate wafer. A groove having a width equal to or greater than the dicing width is formed in advance on the silicon substrate wafer at a position corresponding to the line.

[0003]

However, in a semiconductor device manufactured by the above-described conventional method, oxygen generated at the time of anodic bonding is also accommodated in the groove, and the oxygen generated in the closed space in the pressure sensor is reduced. Although the amount of oxygen decreased somewhat, the degree of vacuum in the enclosed space could not be sufficiently increased. Further, JP-A-7-113708 does not pay any attention to the fact that the degree of vacuum in the closed space is reduced by oxygen generated by anodic bonding.

[0004]

SUMMARY OF THE INVENTION The present invention has been made with a view to reducing the degree of vacuum in a closed space formed between a silicon substrate and a glass lid due to oxygen generated by anodic bonding. Another object of the present invention is to provide a semiconductor device in which the amount of oxygen flowing into a closed space during anodic bonding is reduced to form a closed space having a high degree of vacuum.

To achieve the above object, a feature of the present invention is a semiconductor device in which a glass cover is anodically bonded to a silicon substrate to form a substantially vacuum sealed space between the silicon substrate and the glass cover. A bonding prevention film is provided between each bonding surface of the silicon substrate and the glass lid and between the inner surface and the outer surface in the width direction of the glass lid to prevent bonding between the glass lid and the silicon substrate. I have done it.

In the present invention having the above-described structure, at the time of anodic bonding, the bonding stopper film prevents the bonding between the bonding surfaces of the silicon substrate and the glass lid, which are in contact with the blocking film. The bonding area between the silicon cover and the silicon substrate is reduced, and the amount of oxygen generated by anodic bonding between the glass lid and the silicon substrate is reduced. In addition, gaps are formed between the bonding surfaces of the silicon substrate and the glass lid, and the generated oxygen flows into the formed gaps. Therefore, the amount of oxygen flowing into the closed space can be reduced, and the degree of vacuum in the closed space can be kept high.

Another feature of the present invention is a method of manufacturing a semiconductor device in which a glass cover is anodically bonded to a silicon substrate to form a substantially vacuum sealed space between the silicon substrate and the glass cover. A bonding prevention film that prevents bonding between the glass lid and the silicon substrate is interposed between the bonding surfaces of the substrate and the glass lid and between the inner surface and the outer surface in the width direction of the glass lid. Anodic bonding. According to such a manufacturing method, the same operation and effect as described above can be obtained.

Another feature of the present invention is a method of manufacturing a semiconductor device in which a glass cover is anodically bonded to a silicon substrate to form a substantially vacuum sealed space between the silicon substrate and the glass cover. At the time of joining, a needle electrode is brought into contact with the upper surface position of the glass lid facing the center position of the upper surface of the closed space, and a voltage is applied between the needle electrode and the silicon substrate.

According to such a manufacturing method, the joining of the glass lid and the silicon substrate is started from the inner end in the width direction of the glass lid and proceeds outward in the width direction. Therefore, the oxygen generated by the anodic bonding between the glass lid and the silicon substrate is pushed outward in the width direction, so that the amount of oxygen flowing into the enclosed space located inside is reduced, so that the degree of vacuum in the enclosed space is kept high. Can be.

[0010]

DETAILED DESCRIPTION OF THE INVENTION a. First Embodiment A first embodiment of the present invention will be described with reference to the drawings.
1A is a schematic plan view of the semiconductor device according to the same embodiment, and FIG. 1B is an end view of the same device along line B1-B1.

This semiconductor device comprises a substrate 10 formed in a square shape of silicon, a rectangular frame 20 having a predetermined width and fixed to a peripheral portion of the upper surface of the substrate 10 and formed of p-type silicon. It has. Note that the substrate 10 and the frame 20 constitute the silicon substrate of the present invention.

On the inside of the frame 20 and on the substrate 10,
A vibrator 30 is provided floating above the upper surface by a small predetermined distance. The vibrator 30 is formed of n-type silicon in a substantially rectangular shape, is connected to the vibrator 30 at each inner end, and has an inner surface 20a of the frame 20 at each outer end.
Are supported by the beams 11a to 11d connected to the inside of the frame body 20 so as to be able to vibrate in the X-axis direction (left-right direction in FIG. 1A) and the Y-axis direction (vertical direction in FIG. 1A). I have. The beams 11a to 11d are formed in the substantially L-shape from the same material as the vibrator 30, and float from the upper surface of the substrate 10 by the small predetermined distance similarly to the vibrator 30. The vibrator 3
0 has a plurality of through holes.

A substrate 1 is provided on each outer side of the vibrator 30 in the X-axis direction.
A plurality of electrode fingers 40a and 40b are provided, each of which is fixed on the first electrode 0, and each of the plurality of electrode fingers 40a and 40b extends in the X-axis direction and is arranged at equal intervals in the Y-axis direction. It has. Further, each of the comb-tooth electrodes 40a, 40
Pad portions 41a and 41b fixed to the substrate 10 and connected to the comb-shaped electrodes 40a and 40b, respectively, are provided on the outer sides of the b portion in the X-axis direction, respectively.
Electrode pads 42a and 42b formed in a rectangular shape with a conductive metal (for example, aluminum) are provided on a and 41b, respectively.

A vibrator 3 is provided on the side of the vibrator 30 in the X-axis direction.
Similarly to the case 0, comb-shaped electrodes 31a and 31b are provided, each of which is floating above the substrate 10 by a predetermined distance and formed integrally with the vibrator 30. Each of the comb-shaped electrodes 31a and 31b includes a plurality of electrode fingers extending outward in the X-axis direction and arranged at equal intervals in the Y-axis direction. The electrode 40b has entered the center position in the width direction (Y-axis direction) between the electrode fingers. The comb-shaped electrodes 40a and 40b constitute a driving unit for the vibrator 30 together with the comb-shaped electrodes 31a and 31b.
0 is excited in the X-axis direction by electrostatic attraction when a driving signal is applied to the comb-shaped electrodes 40a and 40b.

A substrate 1 is provided on each outer side of the vibrator 30 in the Y-axis direction.
A plurality of electrode fingers 50a, 50b are provided, each of which is fixed on the first electrode 0, and each of the plurality of electrode fingers 50a, 50b extends in the Y-axis direction and is arranged at equal intervals in the X-axis direction. It has. Further, each of the comb-tooth electrodes 50a, 50
Pad portions 51a and 51b fixed on the substrate 10 and connected to the comb-shaped electrodes 50a and 50b, respectively, are provided on the respective outer sides in the Y-axis direction of b.
Electrode pads 52a, 52b formed of a conductive metal (for example, aluminum) in a rectangular shape are provided on a, 51b, respectively.

A vibrator 3 is provided on the side of the vibrator 30 in the Y-axis direction.
As in the case of 0, comb-shaped electrodes 32a and 32b are provided, each of which is floated from the substrate 10 by a predetermined distance and formed integrally with the vibrator 30. The comb-shaped electrodes 32a and 32b each include a plurality of electrode fingers extending outward in the Y-axis direction and arranged at equal intervals in the X-axis direction. 50b intrudes into the center position in the width direction (X-axis direction) between the electrode fingers. The comb-shaped electrodes 50a and 50b together with the comb-shaped electrodes 32a and 32b constitute a detection unit for the vibrator 30.
0 is used to detect vibration in the Y-axis direction.

On the substrate 10, there is provided a pad portion 12 which is fixed on the substrate 10 and connected to the beam 11c. The pad portion 12 is formed of a conductive metal (for example, aluminum) in a rectangular shape. Electrode pad 13 is provided.

At the inner peripheral edge of the upper surface of the frame 20, beams 11a to 1
1d, vibrator 30 and comb-shaped electrodes 31a, 31b, 32
A rectangular glass cover 60 formed of a glass material and having a substantially U-shaped cross section is fixed by anodic bonding so as to cover a and 32b.
A substantially vacuum sealed space S between the substrate 10 and the glass lid 60
Are formed. At this time, the frame 20 and the glass lid 60
Between the frame 20 and the glass cover 6
In order to prevent the anodic bonding, the bonding preventing film 70 is provided between the inner side surface and the outer side surface of the frame 20 and the glass lid 60 in the width direction of the bonding portion and at the center thereof along the entire circumference of the bonding portion. It is provided. The junction blocking film 70 is made of silicon nitride (SiN).
, Having a predetermined width smaller than the bonding width of the glass lid 60 and a small predetermined thickness (for example, about 0.5 μm)
have. Therefore, both end portions of the joining surface of the glass cover 60 are joined to the frame body 20 and the joining prevention film 7 is formed.
A gap SS is formed between the frame body 20 and the glass lid 60 by the thickness of 0. Note that the thickness of the bonding prevention film 70 is set to a value such that the volume of the gap SS can be secured to a predetermined value or more.

Next, a method of manufacturing the semiconductor device configured as described above will be described with reference to FIG.

(1) First Step A single-crystal silicon substrate having a thickness of about 10 μm is formed on an upper surface of a lower layer a of single-crystal silicon as a substrate material via an intermediate layer b of a silicon oxide film having a thickness of about 1 μm. An SOI (Silicon-On-Insulator) substrate provided with an upper layer c is prepared. Note that the upper layer c is formed of p-type silicon.
The vibrator 30 (excluding the through-hole) on the upper surface of the upper layer c, and the comb-shaped electrodes 31a, 31b, 32a, 32b, 4
0a, 40b, 50a, 50b, pad portions 12, 41
The portions other than the portions corresponding to a, 41b, 51a, 51b and beams 11a to 11d are masked, and the unmasked portions are doped with an impurity such as phosphorus to convert the portions into n-type silicon. Then, foreign matter directly adheres to the upper surface of the upper layer c and the n
In order to prevent conduction between the p-type silicon portion and the p-type silicon portion, an oxide film is formed on the entire upper surface of the upper layer c.

A portion corresponding to the frame 20 (pad portions 12, 41a, 41b, 51a, 51b, and these pad portions 12, 41a, 41b, 51a, 51b and the beam 11)
c, excluding the portions corresponding to the portions electrically connecting the comb-tooth-shaped electrodes 40a, 40b, 50a, 50b) to the portions corresponding to the inside of the frame 20 and the pad portions 12, 41a. , 41b, 51a, 51b, these pad portions 12, 41a, 41b, 51a, 51b and the beam 11
(c) The portions corresponding to the portions electrically connecting the comb-shaped electrodes 40a, 40b, 50a, 50b may be doped with an impurity such as phosphorus to convert the portions into n-type silicon.

(2) Second Step An oxide film formed on the upper surface of the upper layer c and comprising an electrode pad 13,
After contact holes (not shown) are formed in portions corresponding to 42a, 42b, 52a, and 52b by photolithography etching, an aluminum film is formed by sputtering or the like, and electrode pads 13, 42a, 42b, and 52 are formed.
a and 52b are respectively formed.

(3) Third Step After a silicon nitride layer is formed on the upper surface of the upper layer c, a portion of the upper surface of the silicon nitride layer corresponding to the junction preventing film 70 is masked with a resist film. Is etched to form a junction blocking film 70 on the upper surface of the frame 20 (upper layer c).

(4) Fourth step The upper surface of the upper layer c, the vibrator 30 (excluding through holes), the comb-shaped electrodes 31a, 31b, 32a, 32b, 40a, 4
0b, 50a, 50b, beams 11a to 11d and frame 20
Is masked with a resist film. And
The upper layer c and the oxide film on the upper layer c are etched by RIE (reactive ion etching) or the like to form the comb-like electrodes 40a, 40b, 50a, 50b and the frame body 20 on the intermediate layer b, 30, comb-shaped electrodes 31a, 31
b, 32a, 32b and portions corresponding to the beams 11a to 11d are left.

(5) Fifth Step Transducer 30, comb-shaped electrodes 31a, 31b, 32a, 32
b and the intermediate layer b sandwiched between the beams 11a to 11d (upper layer c) and the substrate 10 (lower layer a) are removed by etching with a hydrofluoric acid aqueous solution, and the vibrator 30, the comb-shaped electrode 3
1 a to 31 d and beams 11 a to 11 d are formed so as to float above the substrate 10. This is because when the intermediate layer b is etched with a hydrofluoric acid aqueous solution, not only the portion sandwiched between the upper and lower layers a and c but also the portion sandwiched between the upper and lower layers a and c is close to the outer surface. This is because the part is also removed. Thereafter, the resist film provided as a mask in the fourth step is removed.

(6) Sixth Step The glass lid 60 is fixed to the upper surface of the frame 20 by anodic bonding in vacuum, and a sealed space S is formed between the substrate 10 and the frame 20 and the glass lid 60. The gap SS is formed between the bonding surfaces of the frame body 20 and the glass lid 60 because the bonding prevention film 70 is interposed therebetween. At the time of this anodic bonding, oxygen is generated from the bonding portion, and the generated oxygen flows into the closed space S and the gap SS.

In the above-described semiconductor device, at the time of anodic bonding, the bonding prevention film 70 is formed by the glass cover 60 and the frame 20.
In order to prevent the joining of the portions where the blocking film 70 is in contact with each other, the joining area between the glass lid 60 and the frame 20 is reduced, and the anodic bonding between the glass lid 60 and the frame 20 is generated. The amount of oxygen is reduced. In addition, the glass cover 60
A gap SS is formed between each of the bonding surfaces of the frame body 20 and the generated oxygen flows into the formed gap SS.
Therefore, the amount of oxygen flowing into the closed space S can be reduced, and the degree of vacuum in the closed space S can be kept high.

Next, in using the semiconductor device configured as described above, each of the electrode pads 13, 42a, 42
b, 52a and 52b are connected to an electric circuit device (not shown). The electric circuit device sets the vibrator 30 to its natural frequency f _0.
In order to oscillate at a constant amplitude in the X-axis direction, drive signals having phases opposite to each other are supplied to the electrode pads 42a and 42b, respectively. Further, in order to detect the vibration of the vibrator 30 in the Y-axis direction, detection signals having opposite phases to each other are applied to the electrode pads 52a,
52b. According to this, the vibrator 30 vibrates in the X-axis direction at a constant amplitude and a natural frequency f ₀ due to an electrostatic attraction generated in the driving section by a driving signal from an electric circuit device.

In this state, when an angular velocity about the Z-axis orthogonal to both the X and Y axes acts on the vibrator 30, the vibrator 30 also vibrates in the Y-axis direction with an amplitude proportional to the angular velocity by Coriolis force. With the vibration of the vibrator 30 in the Y-axis direction, the comb-shaped electrodes 32a and 32b connected to the vibrator 30 also vibrate in the Y-axis direction. Thereby, the comb-shaped electrodes 32a, 50a
And the capacitances of the comb-shaped electrodes 32b and 50b change in opposite directions. A signal indicating this change in capacitance is used as a capacitance signal as the electrode pad 13.
Is input to the electric circuit device via the. The electric circuit device
The angular velocity around the Z axis is derived using the capacitance signal. In this case, since the degree of vacuum in the closed space S covered by the glass lid 60 is kept high, the vibration of the vibrator 30 is favorably vibrated without being suppressed by the resistance of the gas. As a result, the detection accuracy of the angular velocity is improved.

In the first embodiment, the junction blocking film 70 is formed of silicon nitride. However, instead of silicon nitride, the bonding blocking film 70 has good adhesion to the upper layer c on which the bonding blocking film 70 is provided. Materials that do not oxidize or oxidize, such as silicon oxide (SiO ₂ ), polysilicon (Poly-Si) formed on silicon oxide, titanium (T
i), tungsten (W), aluminum (Al), chromium (Cr), gold (Au), or the like. The case where the junction blocking film 70 is formed of silicon oxide or polysilicon formed on silicon oxide is the same as in the second step, except that the junction blocking film 70 is formed of titanium, tungsten, aluminum, chromium, or gold. Is the upper surface of the upper layer c instead of the second step.
A portion other than the portion corresponding to 0 is masked, and a joining prevention film 70 is formed of titanium, tungsten, aluminum, chromium, and gold on the upper surface of the frame 20 (upper layer c) by a sputtering method. With this, the same operation and effect as described above can be expected.

In the first embodiment, the joining preventing film 70 is provided on the joining surface of the frame 20.
In order to prevent anodic bonding between the frame body 20 and the glass lid 60,
The bonding prevention film 70 may be provided on the bonding surface of the glass lid 60. Also in this case, the joining prevention film 70 made of silicon nitride (SiN) and having a predetermined width smaller than the bonding width of the glass lid 60 and having a small predetermined thickness (for example, about 0.5 μm) is formed on the frame 20 and the glass. A cover is provided between the inner surface and the outer surface in the width direction of the joint of the lid 60 and at the center thereof along the entire periphery of the joint. Therefore, the glass lid 6
In addition, both ends of the bonding surface 0 are bonded to the frame 20, and a gap SS is formed between the frame 20 and the glass lid 60 due to the thickness of the bonding prevention film 70.

In manufacturing such a modified example, the second step in the first embodiment is omitted, and the bonding surface of the glass lid 60 is formed before the sixth step of anodically bonding the glass lid 60 to the frame 20. Then, a junction blocking film 70 is formed. That is, after a silicon nitride layer is formed on the lower surface of the glass lid 60, a portion of the upper surface of the silicon nitride layer corresponding to the bonding prevention film 70 is masked with a resist film, and the silicon nitride layer is etched. A bonding prevention film 70 is formed on the lower surface of the glass lid 60.

In the above-described modification, the bonding prevention film 70 is formed of silicon nitride. However, instead of silicon nitride, the bonding prevention film 70 has good adhesion to the glass cover 60 on which the bonding prevention film 70 is provided. It may be made of a material that does not oxidize, for example, polysilicon, titanium, tungsten, aluminum, chromium, or the like. When the junction blocking film 70 is formed of polysilicon, the manufacturing method is the same as that described above. However, when the bonding blocking film 70 is formed of titanium, tungsten, aluminum, or chromium, the bonding blocking film 70 A portion other than the portion corresponding to 70 is masked, and a joining prevention film 70 is formed using titanium, tungsten, aluminum, and chromium by a sputtering method. With this, the same operation and effect as described above can be expected.

B. Second Embodiment A second embodiment of the present invention will be described with reference to the drawings.
FIG. 2A is a schematic plan view showing a state where a needle-shaped electrode is brought into contact with the semiconductor device according to the embodiment, and FIG. 2B is an end view of the device taken along line B2-B2. This semiconductor device
This is the same as the semiconductor device of the first embodiment except that the junction blocking film 70 is omitted. Therefore, the entire surface of the bonding surface of the glass lid 60 is bonded to the frame 20. The other configuration is the same as that of the first embodiment, and the description thereof is omitted.

In manufacturing this apparatus, the third step is omitted from the manufacturing steps of the first embodiment, and the glass lid 60 and the substrate 10 are heated to a high temperature (for example, 350 to 450 ° C.) instead of the sixth step. To the upper surface of the closed space S (ie, the glass lid 60)
The needle electrode 80 is brought into contact with the upper surface position of the glass cover 60 facing the center position, and the needle electrode 80 and the substrate 1
0, a voltage (for example, about 1 kV) is applied for a predetermined time (for example, 15 to 30 minutes), and the glass cover 60 is attached to the frame 20.
To form a closed space S between the glass cover 60 and the substrate 10 and the frame 20. The joining of the glass lid 60 and the frame 20 starts from the widthwise inner end of the joint between the glass lid 60 and the frame 20 and proceeds outward in the widthwise direction. Therefore,
Oxygen generated by anodic bonding between the glass lid 60 and the frame body 20 is pushed outward in the width direction, so that the amount of oxygen flowing into the enclosed space S located inside is reduced, so that the degree of vacuum of the enclosed space S is increased. Can be kept.

In the first and second embodiments and the modifications, the present invention is applied to a semiconductor device for detecting an angular velocity. However, the present invention is closed with a glass lid to detect a physical quantity. , For example, a semiconductor device for detecting acceleration. In this case, the amount of displacement of the vibrator with respect to the substrate may be detected based on the acceleration acting on the substrate.

[Brief description of the drawings]

FIG. 1A is a schematic plan view of a semiconductor device according to a first embodiment of the present invention, and FIG. 1B is an end view of the same semiconductor device of FIG. 1A along line B1-B1.

FIG. 2A is a schematic plan view showing a state where a needle electrode of a semiconductor device according to a second embodiment of the present invention is in contact with the semiconductor device;
(B) is an end view along the B2-B2 line of the semiconductor device of (A).

[Explanation of symbols]

10: substrate, 11a to 11d: beam, 12, 41a, 41
b, 51a, 51b... pad portion, 13, 42a, 42
b, 52a, 52b ... electrode pad, 20 ... frame, 30 ...
Oscillator, 31a to 31d, 40a, 40b, 50a, 5
0b: comb-shaped electrode, 60: glass cover, 70: bonding prevention film, 80: needle electrode, S: closed space, SS: gap.

Claims (3)

[Claims] In a semiconductor device in which a glass lid is anodically bonded to a silicon substrate to form a substantially vacuum sealed space between the silicon substrate and the glass lid, a bonding surface of each of the silicon substrate and the glass lid is formed. A semiconductor device, wherein a bonding prevention film for preventing bonding between the glass lid and the silicon substrate is interposed between the inner side and the outer side in the width direction of the glass lid. 2. A method for manufacturing a semiconductor device in which a glass lid is anodically bonded to a silicon substrate to form a substantially vacuum sealed space between the silicon substrate and the glass lid. The anodic bonding is performed between the bonding surfaces and at a position between the inner side surface and the outer side surface in the width direction of the glass lid with a bonding prevention film for preventing bonding between the glass lid and the silicon substrate interposed therebetween. A method for manufacturing a semiconductor device, characterized in that: 3. A method of manufacturing a semiconductor device in which a glass lid is anodically bonded to a silicon substrate to form a substantially vacuum sealed space between the silicon substrate and the glass lid. A semiconductor device, wherein a needle-shaped electrode is brought into contact with an upper surface position of the glass lid facing a center position of an upper surface of a space, and a voltage is applied between the needle-shaped electrode and the silicon substrate. Production method.