JP2002318244A - Semiconductor dynamic quantity sensor and method for manufacturing the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a semiconductor dynamic quantity sensor capable of ensuring a large movable range of a beam structure while suppressing a trouble caused by a micro-loading effect, and a method for manufacturing the same. SOLUTION: A laterally extending cavity and a longitudinally extending groove are formed in a silicon substrate. The beam structure 6 has a movable electrode 9 and a fixed electrode 12 is arranged in opposed relation to the movable electrode 9. The beam 7c of the beam structure 6 is bent and the interval A between beams is almost equal to the interval B between the movable electrode 9 and the fixed electrode 12. The movable range of the beam structure 6 is set to a predetermined range larger than the interval B between the movable electrode 9 and the fixed electrode 12.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a semiconductor dynamic quantity sensor provided with a beam structure having a movable electrode and for detecting a dynamic quantity such as acceleration, angular velocity, and vibration.

[0002]

2. Description of the Related Art In a process of forming various physical quantity sensor elements (elements) in a substrate by etching a semiconductor substrate to form various structures (mechanical structures) on the substrate, usually, a plasma is used. Dry etching using ions or radicals is applied. Then, after covering one surface of the semiconductor substrate with a mask, a dry etching is performed from an opening of the mask to form a structure on the semiconductor substrate.

[0003] An angular velocity sensor having a general shape will be described with reference to FIG. A primary vibrator 101 and a secondary vibrator 102 are formed on a substrate 100, and excitation electrodes 103 and 10 are provided.
The primary oscillator 10 is generated by generating an electrostatic force on
The first weight 105 is forcibly vibrated in a certain direction (excitation direction) in the substrate plane. This vibration is caused by the primary oscillator beam 106.
Is possible depending on the degree of freedom. In this state, when an angular velocity around an axis perpendicular to the substrate surface is applied, the weight 107 is applied to the direction perpendicular to both the excitation direction and the axial direction of the angular velocity (= direction perpendicular to the excitation direction in the substrate surface). Coriolis force Fc proportional to the magnitude of the angular velocity occurs.

Fc = 2 mvΩ m: Mass of weight 107 of secondary oscillator 102 v: Excitation speed Ω: Angular speed Displacement of weight 107 by this Coriolis force changes secondary oscillator electrode 108 in primary oscillator 101, The angular velocity is detected by detecting the change in capacitance at 109. This weight 107
Can be displaced by the degree of freedom of the beam 110 for the secondary vibrator.

Here, the magnitude of the Coriolis force is very small, and the displacement due to the Coriolis force is also very small (for example, 0.
About 1 nm). Therefore, it is important to design the sensor so as to obtain the highest possible sensitivity. The sensitivity of the angular velocity sensor is proportional to the excitation speed, and the excitation speed is proportional to the amplitude at the time of excitation when the frequency of excitation is constant. Therefore, the larger the amplitude at the time of excitation, the better.

In order to obtain such a large amplitude at the time of excitation, it is necessary to make the interval between the weight portion 105 and the supporting portion 111 larger than the amplitude of the excitation. In the excitation electrodes 103 and 104, the excitation force is almost inversely proportional to the electrode interval. Therefore, it is necessary to reduce the electrode interval in order to generate as large an excitation force as possible. , 109, the detection sensitivity of the change in capacitance is almost inversely proportional to the square of the electrode interval. Therefore, it is necessary to make the electrode interval as narrow as possible.

[0007] Under these circumstances, the patterns of the angular velocity sensor cannot be all equally spaced, and inevitably have a groove pattern with a large opening width (opening area) and a groove pattern with a small opening width. These structures often have a uniform thickness in the depth direction. The structure having a uniform thickness is formed by covering one surface of the semiconductor substrate 100 with a mask, dry-etching these groove patterns having different opening widths to a uniform depth, and further lowering the lower part by some means. Formed by removing the substrate material.

However, in ordinary dry etching, a difference occurs in an etching rate between a groove pattern having a large opening width and a groove pattern having a small opening width, and a groove pattern having a large opening width can obtain a deeper groove. A microloading effect occurs. Therefore, in order to etch a plurality of groove patterns having different opening widths to a uniform depth, for example, in the case of a silicon substrate, an oxide film whose etching rate is extremely slow is buried in the silicon substrate in advance. Both the groove pattern having a wide opening width and the groove pattern having a narrow opening width are required to have a special substrate structure in which the progress of etching is stopped when the oxide film surface is exposed. However, an increase in cost is unavoidable for providing such a structure on the substrate.

On the other hand, Japanese Patent Laid-Open No. 2000-150462
In the gazette, a mask having a pattern in an opening portion which divides the opening into a plurality of divided areas having the same area is used as a mask at the time of etching, and after covering one surface of a silicon substrate as a semiconductor substrate with the mask, A groove is formed by a micro-loading effect by a process including a first step of performing etching, and a second etching step of removing the mask and then removing the remaining side wall corresponding to the divided pattern by etching. To prevent the depth of the window from being different.

However, this method has a problem that the process becomes very complicated. In addition, since it is premised that etching of about 0.5 to 2 μm is performed in the lateral direction in advance, it is necessary to use a mask with a narrow opening width in anticipation of the etching, but even if it does so, it is completely desired. It is difficult to get a shape.

[0011]

SUMMARY OF THE INVENTION The present invention has been made under such a background, and an object of the present invention is to secure a large movable range of a beam structure while suppressing a problem due to a microloading effect. An object of the present invention is to provide a semiconductor dynamic quantity sensor and a method of manufacturing the same.

[0012]

According to a first aspect of the present invention, a beam of a beam structure is bent so that a distance between the bent beam and the beam is substantially equal to a distance between a movable electrode and a fixed electrode. The movable range of the beam structure is set to a predetermined range larger than the distance between the movable electrode and the fixed electrode.

When it is desired to increase the movable range of the beam structure to some extent, in a conventional beam without bending or a beam bent only once, a groove having a width larger than a desired movable range is formed between the beam structure and the fixed portion. Will be needed. For example, when it is desired to take a predetermined movable range at a certain one-sided amplitude (= total amplitude / 2), a groove having a size of the entire amplitude / 2 is required. Usually, this movable range is considerably larger than the electrode interval. Then, there is a problem that the depth of the groove differs due to a microloading effect (an effect that a groove pattern having a larger opening width can obtain a deeper groove than a groove pattern having a smaller opening width). On the other hand, by adopting a device as claimed in claim 1, it is possible to (i) secure a large movable range of the beam structure, and (ii) to make the groove width of the beam portion and the interval between the electrodes the same so that the microloading effect can be obtained. Can be realized at the same time.

This effect is particularly effective when the beam is used as an excitation beam of an angular velocity sensor because the movable range is large, but it can also be used as a beam of an acceleration sensor.
According to the second aspect of the present invention, the microloading effect can be reduced by making the opening width of the tip and the facing portion of the beam equal to the opening width of the electrode.

According to the third aspect of the present invention, the microloading effect can be reduced by making the opening width of the portion facing the side surface of the beam equal to the opening width of the electrode.
According to the fourth aspect of the invention, if the number of bending of the beam is increased, the movable range can be increased as the number of bending is increased. However, increasing the number of bendings of the beam causes the beam to move in a direction other than the direction in which the beam should originally move (particularly in a direction perpendicular to the direction). If the number of bendings n of the beam is determined by the calculation expression according to claim 4, the minimum number of bendings for obtaining a necessary movable range can be obtained.

According to the fifth and eighth aspects of the present invention, a groove having a width larger than a desired movable range is usually required between the movable part and the fixed part of the beam structure. For example, when it is desired to take a predetermined movable range at full amplitude / 2, at least the full amplitude / 2
A groove of the size was required. In this case, the microloading effect becomes a problem, but this problem can be reduced by inserting a dummy electrode according to the fifth and eighth aspects.

This effect is effective especially when this electrode is used for exciting the angular velocity sensor because the movable range is large, but it can also be used in the detection direction of the acceleration sensor.

According to the sixth and ninth aspects of the present invention, the distance between the movable and fixed electrodes acting as electrodes and the distance between the movable and fixed electrodes acting as dummy electrodes are made substantially equal to each other. The opening widths at the portions become equal, and the microloading effect can be further reduced.

According to the invention of claim 7, according to claim 1,
In addition to the functions and effects of the invention described in the above, in the electrode (detection electrode) of the beam structure for the secondary vibrator, conventionally, the movable electrode and the fixed electrode alternately move, the movable electrode → the fixed electrode → the movable electrode → the fixed. In general, the electrodes are arranged in the order of the electrodes, and the distance between the movable electrode and the fixed electrode is generally narrow → wide → narrow → wide. The microloading effect can be further reduced by setting the order of movable electrode → fixed electrode → fixed electrode → movable electrode → fixed electrode → fixed electrode, and making them all equally spaced.

According to the tenth aspect of the present invention, the distance between the tip of the beam and the opposing portion of the beam structure for the secondary vibrator is made substantially equal to the electrode distance in the beam structure for the secondary vibrator. ,
The opening width of these portions becomes equal, and the microloading effect can be further reduced.

According to the eleventh aspect of the present invention, the distance between the tip of the movable electrode and the facing portion in the beam structure for a secondary vibrator is also
The microloading effect can be further reduced by making the electrode interval substantially equal to that of the secondary vibrator beam structure.

According to the twelfth aspect of the present invention, the distance between the tip of the fixed electrode and the opposing portion in the beam structure for a secondary vibrator is also different from the electrode (detection electrode) in the beam structure for a secondary vibrator. By making the distance approximately equal, the microloading effect can be further reduced.

According to the thirteenth aspect, especially in an angular velocity sensor, the excitation amplitude is usually 5 μm to 100 μm.
On the other hand, the interval between the excitation electrodes and the interval between the detection electrodes are very small, typically about 2 μm, and the effect of the microloading effect is large. Therefore, the effect of reducing the microloading effect is large.

According to the fourteenth aspect of the present invention, as an excitation method other than the electrostatic force, there are conventionally an electromagnetic method and a piezoelectric method. For example, the electromagnetic method requires extra parts such as permanent magnets and electromagnets. Become. In the case of the piezoelectric type, the formation of a piezoelectric thin film is necessary, and the process becomes complicated. Therefore, the excitation using the electrostatic force has advantages that the process is simple, the number of parts is small, and the size can be reduced.

According to the fifteenth aspect of the present invention, the arrangement of the electrodes in the beam structure for the primary vibrator is at equal intervals, so that the exciting force (electrostatic attraction) is straight in the direction of the comb-tooth electrode. appear. As a result, excitation (vibration) can be generated in a desired direction. In addition, it is possible to avoid a problem that the movable electrode and the fixed electrode in the primary oscillator beam structure collide with each other (this causes a very large noise).

According to the invention of claims 16 and 17,
In the excitation electrode in the primary oscillator beam structure, the interval between the fixed electrode and the fixed electrode and the interval between the movable electrode and the movable electrode are considered to be the opening width.
Since the distance between the movable electrodes is larger than twice the distance between the fixed and movable electrodes, in order to prevent the microloading effect, the distance between the fixed and fixed electrodes and the distance between the movable and movable electrodes is made as small as possible. Want to. At the same time, the distance between the tip of the movable electrode and the facing part and the distance between the tip of the fixed electrode and the facing part should be as short as possible. At least, this distance needs to have the same value as the amplitude at the time of excitation. Therefore, the microloading effect can be prevented by using the minimum value approximately.

According to the eighteenth aspect of the present invention, the distance between the fixed electrode and the movable electrode in the electrode in the beam structure for the primary vibrator is such that the distance between the fixed electrode and the movable electrode in the electrode in the beam structure for the secondary vibrator is increased. Are approximately equal to each other, the opening widths of these portions become substantially equal, and the microloading effect in these portions can be prevented.

According to the nineteenth aspect of the present invention, in the electrodes of the beam structure for the primary vibrator, the interval between the fixed electrodes and the interval between the movable electrodes are considered to be the opening width. Since the value is larger than twice the distance between the fixed and movable electrodes, to prevent the microloading effect,
I want to make this interval as narrow as possible. For that purpose, it is desired to make the width of the fixed electrode and the movable electrode as narrow as possible. On the other hand, when the width is reduced, there is a problem that the strength of the excitation electrode is reduced. In order to increase this strength, it is desired to shorten the length of the electrode. Generally, it is sufficient for the length of the electrode to be twice the excitation amplitude, so that there is no merit in making it longer. According to the present invention, the width of the comb teeth can be reduced, and the microloading effect can be further reduced.

The structure of the invention described in claim 20 means that the space between the fixed and fixed electrodes and the space between the movable and movable electrodes in the description of the functions and effects of the inventions of claims 16 and 17 are reduced. And is effective in reducing the microloading effect.

According to the twenty-first aspect of the present invention, since the interval between the beam structure and the fixed portion is substantially equal to the electrode interval between the detection electrodes, the opening width of this portion becomes substantially equal,
The microloading effect can be further reduced.

According to a twenty-second aspect of the present invention, there is provided a means for monitoring an amplitude at the time of excitation. By monitoring the amplitude of the excitation, the amplitude of the excitation changes due to, for example, a temperature change. In such a case, the sensitivity can be kept constant by making the excitation constant or correcting it by a circuit.

Therefore, it is possible to reduce the temperature drift of the sensitivity (change of the zero point of the angular velocity output value due to the temperature change). Incidentally, the amplitude of the excitation greatly depends on the viscosity coefficient of the atmospheric gas, and the viscosity coefficient of the gas is generally very sensitive to the temperature. Therefore, the influence of the temperature on the sensitivity of the angular velocity sensor (temperature drift of the sensitivity) is very small. The effect of the present invention is great.

According to the twenty-fourth aspect of the present invention, other types of vibration monitors include strain gauges utilizing the piezoelectric effect and those utilizing induced electromotive force. The strain gauges are formed by the number of masks. In addition, the sensor using the induced electromotive force increases the number of masks and requires an external permanent magnet or electromagnet, so that the size of the sensor itself increases and the cost increases. On the other hand, a capacitance detection method that does not increase the size without increasing the number of masks is most advantageous.

According to the twenty-fifth aspect of the present invention, since an electric potential is generated between the fixed electrode for vibration monitoring and the movable electrode, the same electrostatic force acts as the excitation electrode, but the distance between the monitoring electrodes is reduced. If non-uniform, this electrostatic force has a component in the direction perpendicular to the excitation direction, which causes noise. In contrast, noise can be prevented by making it uniform. Also, making the opening width as constant as possible leads to a reduction in the microloading effect.

According to the invention described in claims 26 and 27,
Also in the vibration monitoring electrode, similarly to the excitation electrode, the interval between the fixed electrode and the fixed electrode and the interval between the movable electrode and the movable electrode are considered to be the opening width, and this value is more than twice the interval between the fixed to the movable electrode. To prevent the microloading effect because it is large, we want to make this interval as narrow as possible. At the same time, the distance between the tip of the movable electrode and the facing part and the distance between the tip of the fixed electrode and the facing part should be as short as possible. This distance is
At a minimum, the amplitude must be the same as the excitation,
By using approximately the minimum value, the microloading effect can be reduced.

According to the twenty-eighth aspect, the microloading effect can be reduced by equalizing the opening widths of the vibration monitoring electrode and the electrode of the secondary vibrator beam structure.

According to the invention described in claim 29, in the vibration monitoring electrode, the interval between the fixed electrode and the fixed electrode and the interval between the movable electrode and the movable electrode are considered to be the opening width. Since the spacing is greater than twice the spacing, it is desirable to make this spacing as narrow as possible to prevent the microloading effect. For that purpose, it is desired to make the width of the fixed electrode and the movable electrode as narrow as possible. On the other hand, when the width is reduced, there is a problem that the strength of the vibration monitoring electrode is reduced. In order to increase this strength, it is desired to shorten the length of the electrode. Generally, it is sufficient for the length of the electrode to be twice the excitation amplitude, so that there is no merit in making it longer. According to the present invention, the width of the electrode can be reduced, and the microloading effect can be reduced.

The structure of the invention described in claim 30 means that the space between the fixed and fixed electrodes and the space between the movable and movable electrodes in the description of the functions and effects of the inventions of claims 26 and 27 are reduced. And is effective in reducing the microloading effect.

According to the thirty-first aspect of the present invention, no electrostatic force is generated in the dummy electrode, but keeping the opening width as constant as possible leads to a reduction in the microloading effect.

According to the invention described in claims 32 and 33,
In the case of the dummy electrode, similarly to the excitation electrode, the distance between the fixed electrode and the fixed electrode and the distance between the movable electrode and the movable electrode are considered to be the opening width. In order to prevent the loading effect, we want this interval to be as narrow as possible. At the same time, the distance between the tip of the movable electrode and the facing part and the distance between the tip of the fixed electrode and the facing part should be as short as possible. This distance, at a minimum,
Since the same value as the amplitude at the time of excitation is required, it is possible to prevent the microloading effect by using the minimum value substantially.

According to the thirty-fourth aspect, the microloading effect can be reduced by making the openings of the dummy electrode and the electrode of the secondary vibrator beam structure substantially equal.

According to the invention described in claim 35, in the dummy electrode, the distance between the fixed electrode and the fixed electrode and the distance between the movable electrode and the movable electrode are considered to be the opening width. Since the interval is larger than twice the interval, in order to prevent the microloading effect, it is desirable to make this interval as narrow as possible. For that purpose, it is desired to make the width of the fixed electrode and the movable electrode as narrow as possible. On the other hand, when the width is reduced, there is a problem that the strength of the dummy electrode is reduced. In order to increase this strength, it is desired to shorten the length of the electrode. Generally, it is sufficient for the length of the electrode to be twice the excitation amplitude, so that there is no merit in making it longer. According to the present invention, the width of the electrode can be reduced, and the microloading effect can be prevented.

The structure of the invention according to claim 36 means that the spacing between the fixed and fixed electrodes and the spacing between the movable and movable electrodes in the description of the functions and effects of the inventions of claims 32 and 33 are reduced. And is effective in reducing the microloading effect.

According to the thirty-eighth aspect of the present invention, the semiconductor physical quantity sensor according to the first aspect can be actually manufactured.

[0045]

Embodiments of the present invention will be described below with reference to the drawings. 1 and 2 show an angular velocity sensor according to the present embodiment. FIG. 1 is a plan view of the angular velocity sensor. FIG. 3 shows a cross section taken along line II of FIG.

FIG. 2 is a plan view showing a state in which the wiring of the angular velocity sensor has been removed. That is, FIG. 1 is a plan view including the wiring of the present sensor, and FIG. 2 does not show the wiring.

In FIG. 3, a cavity 2 is formed inside a silicon substrate 1 as a single-layer semiconductor substrate, and the cavity 2 has a predetermined thickness t and has a lateral direction (horizontal direction). Extends to. A portion of the substrate 1 below the cavity 2 is a base plate portion 3. That is, the base plate portion 3 is defined by the cavity 2 and the base plate portion 3
Is located below the cavity 2. Further, at a portion of the substrate 1 above the cavity 2, as shown in FIGS.
a, 4b, 4c, 4d are formed, and the grooves 4a, 4b,
4 c and 4 d extend in the vertical direction and reach the cavity 2. The rectangular frame portion 5 and the primary vibrator beam structure 6 are defined by the cavity 2 and the grooves 4a, 4b, 4c, 4d. The square frame portion 5 includes the cavity 2 and the grooves 4a, 4a.
b, 4c, 4d, and are erected on the upper surface of the base plate portion 3. This square frame portion 5 is constituted by the side wall of the substrate 1. The beam structure 6 is located above the cavity 2 and extends from the rectangular frame 5. At this time, the beam structure 6 is disposed at a predetermined interval t from the upper surface of the base plate 3.

In FIG. 2, the beam structure 6 includes beams 7a and 7a.
b, 7c, 7d, a weight 8, and movable electrodes 9, 10, 11. Beams 7a, 7b, 7c, 7d extend from the square frame portion 5, and a weight portion 8 is connected to the end thereof. The beam structure 6 can be displaced in the X direction (left-right direction) in FIG. The movable electrodes 9, 10, 11 extend from the weight 8 in the X direction. The movable electrode 9 is an excitation electrode, the movable electrode 10 is a vibration monitoring electrode, and the movable electrode 11 is a dummy electrode.

On the other hand, the fixed electrodes 12, 1
3 and 14 extend in the movable direction X of the beam structure 6. The fixed electrodes 12, 13 and 14 are formed by the cavity 2 and the grooves 4a and 4a.
b, 4c, 4d, located above the cavity 2,
It is arranged to face the movable electrodes 9, 10, 11 of the beam structure 6. For details, fixed electrode for excitation (primary fixed electrode)
Numeral 12 is arranged to face the movable electrode 9 for excitation, and fixed electrode 13 for vibration monitor is arranged to face the movable electrode 10 for vibration monitor.
The dummy fixed electrode 14 is arranged to face the dummy movable electrode 11. Further, the aforementioned movable electrodes 9, 10, 11;
The corresponding fixed electrodes 12, 13, and 14 are electrodes having a comb structure.

In this embodiment, the beam structure 6 and the fixed electrodes 12 to 1 are formed by the base plate 3 and the rectangular frame 5.
4 constitutes a supporting portion. In the beam structure 6 of FIG. 2, the excitation movable electrode 9 and the vibration monitoring movable electrode 10 are formed of insulating materials 30 and 31 at the base thereof, and are electrically insulated from the weight portion 8. The fixed electrodes 12 for excitation and the fixed electrodes 13 for vibration monitoring have their roots made of insulating materials 32 and 33, respectively.
And are electrically insulated. Note that the dummy movable and fixed electrodes 11 and 14 do not need to be electrically insulated.

In the beam structure 6 shown in FIG. 2, the weight 8 has grooves 21a, 21b, 21c, 21 extending in the vertical direction.
d is formed to penetrate the weight portion 8. The grooves 21 a, 21 b, 21 c, and 21 d define the secondary vibrator beam structure 20 inside the weight portion 8. The beam structure 20 includes beams 22 a and 22 b and a square frame 23.
And a weight 24 and a movable electrode 25. Beams 22a and 22b extend from the square frame portion 23, and a weight portion 24 is connected to the tip thereof. The beam structure 20 can be displaced in the Y direction in FIG. A movable electrode 25 extends in the weight portion 24 in a direction X orthogonal to the movable direction Y of the beam structure 20. The movable electrode 25 is an electrode for detecting angular velocity, and has a comb-tooth structure.

On the other hand, the fixed electrodes 26,
27 extends in a direction X orthogonal to the movable direction Y of the beam structure 20. The fixed electrodes 26 and 27 are defined by vertically extending grooves 21 a and 21 b in the beam structure 6.
It is located above the cavity 2. The fixed electrode 26 is the beam structure 2
The fixed electrode 27 is arranged in a comb shape opposite to the other surface of the movable electrode 25. In the beam structure 20, the fixed electrodes 26, 27 have insulating materials 34,
35, and is electrically insulated from the square frame 23. The beams 22a and 22b of the beam structure 20 are
The base is made of an insulating material 36, and the square frame 23
And are electrically insulated. Therefore, all the angular velocity detecting movable electrodes 25 are set to the same potential as the weight portion 24.

Next, wirings for these electrodes will be described with reference to FIG. In FIG. 1, reference numerals 40 to 45 denote wirings. The wiring 40 of FIG. 1 is for the fixed excitation electrode 12 of FIG. 2, and the wiring 41 of FIG. 1 is for the movable movable electrode 9 of FIG. The wiring 4 in FIG.
2 is for the vibration monitoring fixed electrode 13 in FIG. 2, and the wiring 43 in FIG. 1 is the angular velocity detecting movable electrode 25 in FIG.
Is for 1 is for the fixed electrode 26 for detecting angular velocity in FIG. 2, and the wiring 45 in FIG. 1 is for the fixed electrode 27 for detecting angular velocity in FIG.

Referring to FIG. 3, the excitation movable electrode 9 as a representative of the cross-sectional structure of the wiring will be described. The potential of the movable electrode 9 is once pulled up from the contact hole 52 in FIG. 3 by the wiring material 53 to the upper wiring above the insulating films 50 and 51, and then the lower wiring 5 between the insulating film 50 and the insulating film 51.
By 4, it is taken out above the beam 7 b of FIG. 2. Here, for example, the potential of the weight portion 24 of the secondary vibrator is similarly pulled up from the contact hole onto the insulating film 51 by the wiring material 53 in FIG.
As it is pulled out through the beam 7a in FIG.
The two wires do not intersect (while maintaining another potential)
Pulled out. The other electrodes are similarly drawn out by combining the upper wiring 53 and the lower wiring 54 of FIG.

Next, the operation of the angular velocity sensor according to the embodiment will be described. The primary oscillator 6 is excited in the X direction by applying a periodic voltage such as an offset sine wave (or rectangular wave) to the excitation fixed electrode 12 with the excitation movable electrode 9 shown in FIG. I do. Of course, the secondary vibrator 20 also vibrates in the X direction together with the primary vibrator 6. Here, voltages of opposite phases are usually applied to the left and right excitation fixed electrodes 12, and the frequency of the voltage applied as the drive voltage is generally applied at the resonance frequency of the system. This is because the sensitivity of the sensor is proportional to the amplitude at the time of excitation, but the amplitude can be made very large by using resonance. As a method of vibrating at this resonance, a method called self-excited oscillation is often used. That is, by generating a drive voltage from the voltages detected by the movable and fixed electrodes 10 and 13 for vibration monitoring, the system is automatically vibrated at the resonance frequency of the system.

The method called auto gain control, that is, the movable and fixed electrodes 10 and 1 for vibration monitoring are used.
3, a method of simultaneously examining the vibration amplitude and adjusting the value of the drive voltage so that the vibration amplitude is always constant is used. This is because, in the vibration caused by resonance, the amplitude is proportional to the Q value of the vibration system. Here, in many cases, the Q value is almost inversely proportional to the viscosity coefficient of the atmospheric gas, but the viscosity coefficient greatly changes with temperature. For example, the viscosity coefficient of air at 1 atm is 14.6 × 10 ⁻⁶ Pa · s at −50 ° C. 17.1 × 10 ⁻⁶ Pa · s at 0 ° C. 19.3 × 10 ⁻⁶ Pa · s at 50 ° C. It is 21.6 × 10 ⁻⁶ Pa · s at 100 ° C. That is, when the auto gain control is not used, the temperature change of the vibration amplitude is large, and the temperature change of the sensitivity is also large. By using the auto gain control, it is possible to suppress the temperature change of the vibration amplitude, that is, the temperature change of the sensitivity.

The vibration monitoring electrodes 10 and 13 are designed to extract the displacement of the amplitude due to the excitation as a change in the capacitance of the comb electrodes (10 and 13). In this way, the sensor is excited. In this situation, it is assumed that an angular velocity Ω is applied around an axis in a direction Z perpendicular to the surface of the substrate 1. Then, a Coriolis force represented by the following equation is applied to the secondary vibrator 20 in the Y direction (Although the Coriolis force is also applied to the primary vibrator 6,
It is not directly related to the operation of the sensor).

Fc = 2 mvΩ Fc: Coriolis force m: Mass of weight 24 of secondary vibrator 20 v: Excitation velocity Ω: Angular velocity The secondary vibrator 20 vibrates in the Y direction by this Coriolis force Fc. Here, when vibrating in the Y direction, the capacitance of the movable and fixed electrodes 25, 26, 27 for detection changes. Since the amount of change in the capacitance is proportional to the angular velocity,
It is converted into a voltage by a CV conversion circuit and output.

More specifically, the detection fixed electrode 2
6, 27 are electrically separated into two. For example, when the secondary vibrator 20 is displaced in the Y direction, an electrode having a larger capacitance (a smaller distance from the movable electrode) and an electrode having a smaller capacitance (a larger distance from the movable electrode). ).
That is, when one capacity increases, the other capacity decreases. The circuit is configured to detect the angular velocity by taking the difference between the capacitance changes of these electrodes.

Next, the characteristic components of this embodiment will be described. FIG. 4 is an enlarged view of a main part of the angular velocity sensor according to the present embodiment. The sensor structure of the present embodiment is different from the conventional example shown in FIGS.

That is, almost all the grooves 4a, 4b, 4
The widths of c, 4d, 21a, 21b, 21c and 21d are equal to the electrode spacings B, W1, W2, E as shown in FIG. 10 and the electrode spacings M and N in FIG. 4 (B = W in FIG. 10).
1 = W2 = E = dn). Only the width dw is between the movable movable electrode 9 and the movable electrode 9 for excitation, between the fixed electrodes 12 and 12, between the fixed electrodes 13 and 13 for vibration monitoring, and between the movable movable electrodes 11 and 11 for the dummy. (Eg, dn = 2 μm, dw = 15 μm). In addition,
FIG. 10 shows only one movable electrode 10 for vibration monitoring and one fixed electrode 14 for dummy.
In the case where there are a plurality of
The same applies between 4,14.

That is, almost all openings of the beam structures 6 and 20 are made equal so that the microloading effect does not occur, and therefore, the beams 7a to 7d,
A predetermined (sufficiently large) amplitude is obtained at the time of excitation by adjusting the number of bends of 22a and 22b, and the groove width is set to a predetermined value equal to the interval M, N between the fixed electrode and the movable electrode at the detection electrode. The dummy electrodes 11 and 14 have a comb-like shape, and the groove widths of other portions are the same.

The points for realizing this are as follows. Since the sensitivity is proportional to the amplitude of the primary vibrator 6, it is desirable to increase the sensitivity as much as possible. For example, 5 μm to 10
It is about 0 μm. For example, let's assume that the size is 15 μm for simplicity. When the beam 106 bent once (or a beam that is not bent) as in the conventional example shown in FIG.
A space of μm is required. Also, in FIG. 4, between the movable movable electrode for excitation and the fixed electrodes 9 and 12, between the movable movable electrode for detection and the fixed electrodes 25, 26 and 27, and the movable movable fixed electrode 1 for vibration monitoring.
The interval between 0 and 13 should be as small as possible. this is,
According to the following relational expression.

Sensitivity of sensor ∝1 / (interval between excitation movable and fixed electrodes) Sensor sensitivity ∝1 / (interval between detection movable and fixed electrodes) ² Sensitivity of vibration monitor∝1 / (vibration monitor movable to fixed Here, the distance between the movable and fixed electrodes is, for example, 2 μm.

In this embodiment, the beam 7 of the primary vibrator is used.
By increasing the number of bendings of the beams 7a, 7b, 7c and 7d, the vicinity of the beams 7a, 7b, 7c and 7d is also 2 μm.
It is possible to take an amplitude of 15 μm, despite having only a space.

Specifically, the number of bends is n, and the beam 7
When the space around a, 7b, 7c, 7d is dn (see FIG. 10), the maximum possible amplitude value is {(n + 1) / 2} .dn. , Beams 7a, 7b, 7c,
The entire space near 7d can be set to 2 μm.

That is, the beams 7a, 7b, 7 of the beam structure 6
The number of bends of c and 7d is “n”, the interval B between the movable electrode 9 and the fixed electrode 12 and the interval A between the bent beams are “dn”, and the tip of the movable electrode during the operation of the beam structure. By determining the value of n so that δ = {(n + 1) / 2} · dn−Δ, where the minimum distance between the opposing portions is “Δ” and the required total amplitude / 2 is “δ”. , Beam spacing,
In addition, the distance between the movable electrode and the fixed electrode can be made equal.

FIG. 13 shows a conventional example. The maximum amplitude Amax is determined by the distance a between the beams 106. In the present embodiment shown in FIG. 10, the case where the beam is bent eight times and the interval between the beams is dn, the maximum amplitude is 4.5 · dn. Generally, as shown in FIG. 11, when the number of times of bending is n and the interval between beams is dn, the maximum amplitude Amax = ｛(n + 1) /
2｝ · dn.

As described above, the beams 7a, 7b,
The distance A between the bent beams (see FIG. 4) is made substantially equal to the distance B between the movable electrode 9 and the fixed electrode 12 by bending the beams 7c and 7d one or more times. The movable range is the movable electrode 9 and the fixed electrode 1
The predetermined range was larger than the interval B of 2. This has the following effects. When it is desired to increase the movable range of the beam structure to some extent, a beam having a width larger than a desired movable range is required between the beam structure and the fixed portion in a conventional beam without bending or a beam bent only once. . For example, when it is desired to take a predetermined movable range at a certain one-sided amplitude (total amplitude / 2), a groove having a size of the one-sided amplitude is required. Usually, this movable range is considerably larger than the electrode interval. Then
Due to the microloading effect (the effect that a groove pattern having a larger opening width can obtain a deeper groove than a groove pattern having a smaller opening width), there is a problem that the groove depth is different. In contrast, in the present embodiment,
At the same time, (i) securing a large movable range of the beam structure, and (ii) making the groove width of the beam portion equal to the distance between the electrodes so as to be less susceptible to problems due to the microloading effect can be simultaneously realized (microloading effect , And a large excitation amplitude can be obtained). This effect is particularly effective when the beam is used as the excitation beam of the angular velocity sensor because the movable range is large, but it can also be used as the beam of the acceleration sensor.

The distance C between the tip of each of the beams 7a, 7b, 7c, and 7d and the facing portion 12 in the direction Y orthogonal to the movable direction X of the beam structure 6 in FIG. It was almost equal to B. As described above, the microloading effect can be reduced by making the opening width between the tip of the beam and the facing portion equal to the opening width of the electrode.

Further, in the movable direction X of the beam structure 6, the distance D between the side surfaces of the beams 7a, 7b, 7c, 7d and the opposing portions 5, 23 was made substantially equal to the distance B between the movable electrode 9 and the fixed electrode 12. As described above, by making the opening width formed by the portion facing the side surface of the beam equal to the opening width of the electrode, the microloading effect can be reduced.

Further, as described above, the number of times of bending of the beams 7a, 7b, 7c, 7d of the beam structure 6 is set to "n", the distance B between the movable electrode 9 and the fixed electrode 12, the bent beam, and the like. And the minimum distance between the tip of the movable electrode and the opposing portion during the operation of the beam structure is “Δ”.
By determining n so that δ = {(n + 1) / 2} · dn−Δ when the required total amplitude / 2 is “δ”, the distance between the beams,
The distance between the movable electrode and the fixed electrode was made equal. This has the following effects. If the number of bending of the beam is increased, it is possible to increase the movable range as the number is increased. However, increasing the number of times the beam is bent causes the beam to move in a direction other than the direction in which it is originally supposed to move (especially in a direction perpendicular thereto). If the number of bendings n of the beam is determined by the above formula, the minimum number of bendings for obtaining a necessary movable range can be obtained.

Further, the dummy movable electrode 11 extending in the movable direction X is provided on the beam structure 6, and the dummy movable electrode 11 extends in the movable direction X of the beam structure 6 from the rectangular frame portion 5, and the dummy movable electrode 11 extends in the movable direction X of the beam structure 6. A dummy fixed electrode 14 facing the electrode 11 was provided. This has the following effects. Usually, a groove having a width larger than a desired movable range is required between the weight portion 23 and the fixed portion. For example, when it is desired to take a predetermined movable range with one-sided amplitude (total amplitude / 2), a groove having at least the one-sided amplitude is required. In this case, the microloading effect becomes a problem.
This problem can be eliminated by adding 1,14. This effect is effective especially when this electrode is used for exciting the angular velocity sensor because the movable range is large, but it can also be used in the detection direction Y of the acceleration sensor.

Further, the distance E between the dummy movable electrode 11 and the dummy fixed electrode 14 was made substantially equal to the distance B between the movable electrode 9 acting as an electrode and the fixed electrode 12. As a result, the distance between the movable-fixed electrodes 9 and 12 acting as electrodes and the distance between the movable-side comb teeth 11 and the fixed-side comb teeth 14 of the comb-teeth structure acting as dummy electrodes are made substantially equal. The opening widths at the portions become equal, and the microloading effect can be further reduced.

On the other hand, as shown in FIG. 4, the beams 7a, 7b, 7c, 7 of the beam structures 6, 20 for the primary and secondary oscillators.
By bending d, 22a, and 22b, the distance A, σ2 between the bent beams is made substantially equal to the distance B between the primary fixed electrode 12 and the movable electrode 9 of the primary vibrator beam structure 6, and The movable range of the vibrator beam structure 6 is set to a predetermined range larger than the distance B between the primary fixed electrode 12 and the movable electrode 9 of the primary vibrator beam structure 6, and further, the secondary vibrator beam structure 20 The arrangement of the movable electrode 25 and the secondary fixed electrodes 26 and 27 in the order of the movable electrode 25, the fixed electrode 26, the fixed electrode 27, the movable electrode 25, the fixed electrode 26, and the fixed electrode 27 as shown by a reference symbol G in FIG. Arranged at intervals. As a result, in the detection electrode, conventionally, the movable electrode and the fixed electrode are alternately arranged in the order of the movable electrode → the fixed electrode → the movable electrode → the fixed electrode, and the interval between the movable electrode and the fixed electrode is narrow → wide → narrow → wide. In general, the order is as follows: movable electrode 25 → fixed electrode 26 → fixed electrode 2
By setting 7 → movable electrode 25 → fixed electrode 26 → fixed electrode 27 in this order and making them all equally spaced, the microloading effect can be further reduced.

The distance σ 1 between the beam 22 a of the secondary vibrator beam structure 20 and the opposing portion 23 in the movable direction Y of the secondary vibrator beam structure 20 is determined by the following equation. The distances M and N between the movable electrode 25 and the fixed electrodes 26 and 27 were substantially equal. As described above, the interval between the beam 22a and the opposing portion 23 of the secondary oscillator beam structure 20 is changed.
By making the electrode intervals substantially equal to each other, the opening widths of these portions become equal, and the microloading effect can be further reduced.

Further, the distance H between the tip of the beam 22a of the secondary vibrator beam structure 20 and the facing portion 23 in the direction X orthogonal to the movable direction Y of the secondary vibrator beam structure 20 is determined by the secondary vibration. The movable electrode 25 and the fixed electrodes 26, 2 of the sub beam structure 20
The intervals M and N were substantially equal to each other. In this way, by making the distance between the tip of the beam 22a of the secondary vibrator beam structure 20 and the opposing portion 23 substantially equal to the electrode distance in the secondary vibrator beam structure 20, the openings of these portions are opened. The width becomes equal, and the microloading effect can be further reduced.

Further, the distance J between the tip of the movable electrode 25 of the secondary vibrator beam structure 20 and the facing portion 23 in the direction X orthogonal to the movable direction Y of the secondary vibrator beam structure 20 is determined by the following equation. Movable electrode 25 and fixed electrode 2 of vibrator beam structure 20
The distances M and N were substantially equal to each other. in this way,
The gap between the tip of the movable electrode 25 and the opposing portion 23 in the secondary oscillator beam structure 20 is also substantially equal to the electrode interval in the secondary oscillator beam structure 20, thereby further reducing the microloading effect. can do.

Further, a fixed electrode 26 disposed opposite to the movable electrode 25 of the secondary vibrator beam structure 20 in a direction X orthogonal to the movable direction Y of the secondary vibrator beam structure 20,
The distances K and L between the tip of 27 and the opposing portion 24 were set substantially equal to the distances M and N between the movable electrode 25 and the fixed electrodes 26 and 27 of the secondary vibrator beam structure 20. As described above, the tips of the fixed electrodes 26 and 27 in the beam structure 20 for the secondary vibrator and the opposing portion 2
The microloading effect can be further reduced by making the interval 4 substantially equal to the interval between the detection electrodes.

Also, in the angular velocity sensor, the excitation amplitude is very large, usually about 5 μm to 100 μm, whereas the interval between the excitation electrodes and the detection electrode is usually very small, about 2 μm. Since it is large, the effect of reducing the microloading effect is great.

The movable electrode 9 and the primary fixed electrode 12 of the primary vibrator beam structure 6 were of an electrostatic drive type using comb-teeth electrodes. This has the following effects. Excitation methods other than electrostatic force include electromagnetic and piezoelectric types.
For example, the electromagnetic type requires extra parts such as permanent magnets and electromagnets. In the case of the piezoelectric type, the formation of a piezoelectric thin film is necessary, and the process becomes complicated. Therefore, the excitation using the electrostatic force has advantages that the process is simple, the number of parts is small, and the size can be reduced.

Further, the movable electrode 9 and the primary fixed electrode 12 of the primary vibrator beam structure 6 were formed in a shape arranged at equal intervals. As described above, since the arrangement of the electrodes in the primary oscillator beam structure 6 is at equal intervals, the excitation force (electrostatic attraction)
Occur straight in the direction of the comb electrode. Thus, vibration (excitation) can be generated in a desired direction. In addition, it is possible to avoid a problem that the movable electrode 9 and the fixed electrode 12 collide with each other in the primary oscillator beam structure 6 (this causes a very large noise).

Further, the opening width dw (FIG. 10) different from dn is always required near the excitation electrode, vibration monitoring electrode, and dummy electrode.
See). Since the opening width dw is represented by dw = 2 · dn + electrode width (dn is equivalent to the distance between fixed to fixed electrodes and the distance between movable to movable electrodes in this embodiment), the narrower the electrode width is, the smaller the opening width dw is. This is advantageous because dw can be reduced. However, when the electrode width is reduced, there is a problem that the strength of the electrode is reduced. Since the strength of the electrode increases as the electrode length decreases, it is necessary to shorten the electrode length as much as possible within a range where the required excitation amplitude can be obtained. Usually, it is sufficient that the electrode length is about twice the excitation amplitude. Therefore, as the electrode lengths S and T, lengths approximately twice the excitation amplitude are adopted. That is, the length of the portion where the fixed electrode and the movable electrode face each other = excitation amplitude. In consideration of this, as shown in FIG. 4, in the movable direction X of the primary vibrator beam structure 6, the tip of the movable electrode 9 of the primary vibrator beam structure 6 and the facing portion 5
Is substantially equal to the amplitude at which the primary vibrator beam structure 6 is excited. In addition, the primary oscillator beam structure 6
In the movable direction X, the distance P2 between the tip of the primary fixed electrode 12 and the opposing portion 23 was set substantially equal to the amplitude when the primary vibrator beam structure 6 was excited.

That is, in the excitation electrode in the primary oscillator beam structure 6, the opening width is considered to be the distance between the fixed electrodes (dw in FIG. 10) and the distance between the movable electrodes (dw). Since this value dw is larger than twice the value of dn, in order to prevent the microloading effect, it is desirable to make this interval dw as narrow as possible. At the same time, the distance between the tip of the movable electrode and the facing part and the distance between the tip of the fixed electrode and the facing part should be as short as possible. Since this distance needs to be at least the same value as the amplitude δ (> dn) at the time of excitation, the microloading effect can be prevented by using the minimum value substantially. Actually, the distance P1 between the tip of the primary movable electrode 9 and the facing portion 5 and the primary fixed electrode 1
The distance P2 between the tip of No. 2 and the opposing portion 23 is smaller than the excitation amplitude.
Make it slightly larger than the excitation amplitude to prevent contact.

The distance B between the primary fixed electrode 12 and the movable electrode 9 in the primary vibrator beam structure 6 in FIG. 4 is set to be equal to the secondary fixed electrodes 26 and 27 and the movable electrode 9 in the secondary vibrator beam structure 20. The spacing M, N was approximately equal to 25. As described above, the interval between the fixed electrode and the movable electrode in the electrode of the primary oscillator beam structure 6 is substantially equal to the interval between the fixed electrode and the movable electrode in the electrode of the secondary oscillator beam structure 20. The opening widths of these portions become substantially equal, and the microloading effect in these portions can be prevented.

Further, the widths U and V between the primary fixed electrode 12 and the movable electrode 9 of the primary vibrator beam structure 6 are set to about と to 3 times the distance B between the primary fixed electrode 12 and the movable electrode 9. did.
This means reducing the value of dw, which is effective for reducing the microloading effect.

The distance between the primary vibrator beam structure 6 and the fixed member 5 on the outer peripheral side thereof is set to the distance B between the primary fixed electrode 12 and the movable electrode 9 in the primary vibrator beam structure 6 or the secondary vibration. Secondary fixed electrodes 26 and 2 in sub beam structure 20
The distance between the movable electrode 7 and the movable electrode 25 was substantially equal to M and N. As described above, since the interval between the beam structure and the fixed portion is substantially equal to the electrode interval of the detection electrode, the opening width of this portion is almost equal, and the microloading effect can be further reduced.

On the other hand, means 10 and 13 for monitoring the amplitude at the time of excitation of the primary vibrator beam structure 6 are provided. By monitoring the amplitude at the time of the excitation in this way, when the amplitude of the excitation changes due to, for example, a temperature change, the sensitivity is kept constant by making the amplitude of the excitation constant or correcting it by a circuit. be able to. Therefore, sensitivity temperature drift (change in the zero point of the angular velocity output value due to temperature change) can be reduced. Incidentally, the amplitude of the excitation greatly depends on the viscosity coefficient of the atmospheric gas, and the viscosity coefficient of the gas is generally very sensitive to the temperature. Therefore, the influence of the temperature on the sensitivity of the angular velocity sensor (temperature drift of the sensitivity) is very small. The effect of monitoring is great.

Further, the amplitude at the time of excitation of the primary vibrator beam structure 6 is controlled to be constant by negative feedback using the monitoring results of the monitoring means 10 and 13. In addition, a capacitance detection method using the comb electrodes 10 and 13 was used as a vibration monitor. Other types of vibration monitors include strain gauges using the piezoelectric effect and those using induced electromotive force.However, creating a strain gauge leads to an increase in the number of masks, and those using induced electromotive force. Since the number of masks is increased and a permanent magnet or an electromagnet is required outside, the size of the sensor itself becomes large, resulting in an increase in cost. On the other hand, a capacitance detection method that does not increase the size without increasing the number of masks is most advantageous.

Further, the vibration monitoring electrodes 10 and 13 were shaped such that fixed electrodes and movable electrodes were arranged at equal intervals. This is because a potential is generated between the fixed electrode and the movable electrode for vibration monitoring, and the same electrostatic force acts as the excitation electrode.However, if the interval between the monitoring electrodes is not uniform, the electrostatic force is different from the excitation direction. It has a component in the vertical direction, which causes noise. In contrast, noise can be prevented by making it uniform. Also, making the opening width as constant as possible leads to a reduction in the microloading effect.

Further, the interval α1 between the tip of the vibration monitoring movable electrode 10 and the facing portion 5 was made substantially equal to the amplitude at the time of excitation. Further, the interval α2 between the tip of the fixed electrode 13 for vibration monitoring and the facing portion 23 was set substantially equal to the amplitude at the time of excitation. This is done for the following reason. In the vibration monitoring electrode, similarly to the excitation electrode, the distance between the fixed electrode and the fixed electrode (dw) and the distance between the movable electrode and the movable electrode (d
w) is considered to be the opening width, and since this value dw is larger than twice dn, in order to prevent the microloading effect, it is desirable to make the interval dw as narrow as possible. At the same time, the distance between the tip of the movable electrode and the facing part and the distance between the tip of the fixed electrode and the facing part should be as short as possible. Since this distance needs to be at least the same value as the amplitude δ (> dn) at the time of excitation, the microloading effect can be reduced by using the required minimum value δ substantially.

The distances W1 and W2 between the vibration monitoring movable electrode 10 and the vibration monitoring fixed electrode 13 are determined by the distance B between the primary fixed electrode 12 and the movable electrode 9 in the primary vibrator beam structure 6, or the secondary vibration. The distances M and N between the secondary fixed electrodes 26 and 27 and the movable electrode 25 in the sub beam structure 20 were set substantially equal. As described above, the microloading effect can be reduced by equalizing the opening widths of the vibration monitoring electrode and the electrodes of the beam structures 6, 20.

Further, the lengths β1 and β2 of the vibration monitoring movable electrode 10 and the vibration monitoring fixed electrode 13 were set to be approximately twice or less the amplitude at the time of excitation. This is because, in the vibration monitoring electrode, the distance between the fixed electrode (dw) and the distance between the movable electrode and the movable electrode is considered to be the opening width. Since this value dw is larger than twice dn, the microloading effect is not obtained. In order to prevent this, it is desirable to make this interval dw as narrow as possible. For that purpose, it is desired to make the width of the fixed electrode and the movable electrode as narrow as possible. On the other hand, when the width is reduced, there is a problem that the strength of the vibration monitoring electrode is reduced. In order to increase this strength, it is desired to shorten the length of the electrode.
Generally, it is sufficient for the length of the electrode to be twice the excitation amplitude, so that there is no merit in making it longer. Therefore, by setting the β1 and β2 values to approximately twice or less the excitation amplitude, the width of the electrodes can be reduced, and the microloading effect can be reduced.

Further, the widths γ1 and γ2 of the vibration monitoring movable electrode 10 and the vibration monitoring fixed electrode 13 are determined by the distance W between the vibration monitoring movable electrode 10 and the vibration monitoring fixed electrode 13.
1, about 1/3 to 3 times W2. This means reducing the value of dw, which is effective for reducing the microloading effect.

The dummy movable electrode 11 and the dummy fixed electrode 14 were formed in a shape arranged at equal intervals. Therefore, no electrostatic force is generated in the dummy electrode, but keeping the opening width as constant as possible leads to a reduction in the microloading effect.

Further, the distance λ1 between the tip of the dummy movable electrode 11 and the facing portion 5 was made substantially equal to the amplitude at the time of excitation. Further, the distal end of the dummy fixed electrode 14 and the facing portion 23
Is made substantially equal to the amplitude at the time of excitation. This is done for the following reason. In the dummy electrode, similarly to the excitation electrode, the distance (dw) between the fixed electrodes is fixed.
In addition, the distance (dw) between movable electrodes is considered to be the opening width. Since this value dw is larger than twice dn, in order to prevent the microloading effect, it is desirable to make the distance dw as narrow as possible. At the same time, the distance between the tip of the movable electrode and the facing part and the distance between the tip of the fixed electrode and the facing part should be as short as possible. Since this distance needs to be at least the same value as the amplitude δ (> dn) at the time of excitation, the microloading effect can be prevented by using the required minimum value δ.

Further, the distance E between the dummy fixed electrode 14 and the dummy movable electrode 11 is changed to the distance B between the primary fixed electrode 12 and the movable electrode 9 in the primary vibrator beam structure 6, or
The secondary fixed electrode 26 in the beam structure 20 for the secondary vibrator,
The distance between the movable electrode 27 and the movable electrode 25 was substantially equal to M and N. Thus, the microloading effect can be reduced by making the openings of the dummy electrode and the electrode of the beam structure substantially equal.

The length λ3 of the dummy fixed electrode 14 and the dummy movable electrode 11 was set to be approximately twice or less the amplitude at the time of excitation. This is because, in the dummy electrode, the distance (dw) between the fixed electrode and the fixed electrode and the distance (dw) between the movable electrode and the movable electrode are considered to be the opening width. Since this value dw is larger than twice dn, the microloading effect is obtained. In order to prevent this, it is desired to make the interval dw as narrow as possible. For that purpose, it is desired to make the width of the fixed electrode and the movable electrode as narrow as possible. On the other hand, when the width is reduced, there is a problem that the strength of the dummy electrode is reduced. In order to increase this strength, it is desired to shorten the length of the electrode. The length of the electrode is usually
Since twice the excitation amplitude is sufficient, there is no merit in making it longer. Thus, by setting the λ3 value to approximately twice or less the excitation amplitude, the width of the electrode can be reduced, and the microloading effect can be prevented.

Further, the width λ4 of the dummy fixed electrode 14 and the dummy movable electrode 11 was set to be about 1/3 to 3 times the interval E between the dummy fixed electrode 14 and the dummy movable electrode 11.
This means reducing the value of dw, which is effective for reducing the microloading effect.

Further, the distance between the beam structures 6 and 20 and the fixing members 5 and 23 around the beam structures 6 and 20 is determined by the distance B between the primary fixed electrode 12 and the movable electrode 9 in the primary vibrator beam structure 6 or the secondary vibrator. Fixed electrodes 26 and 2 in the beam structure 20 for use
7 and the distances M and N between the movable electrodes 25 are substantially equal to each other.

Next, a method of manufacturing the angular velocity sensor will be described with reference to the process diagrams shown in FIGS. In the description of the manufacturing process, the insulating structure (supporting structure) of each fixed electrode and beam structure is the same, and therefore, the description of the other portions will be omitted by describing the II section of FIG.

First, as shown in FIG. 5, a single-crystal silicon substrate 1 as a single-layer semiconductor substrate is prepared, anisotropic etching is performed from the upper surface of the silicon substrate 1, and electrodes are electrically connected from a rectangular frame or the like. A first groove 37 extending in the vertical direction for insulation is patterned. Then, a silicon oxide film is formed on the silicon substrate 1, the trench 37 is filled with an insulating material (oxide film) 30, and the substrate surface is covered with the oxide film 50.

Further, as shown in FIG.
Is formed and patterned, and then an oxide film 51 is formed to cover the wiring pattern 54. Subsequently, as shown in FIG. 7, a part of the oxide films 50 and 51 is removed to form a contact hole 52, and further, a wiring material 53 is formed and patterned.

Then, as shown in FIG. 8, a mask photo for forming a structure is performed on the substrate 1, and the oxide films 50 and 51 are etched with the mask 70. Then, the mask 7
0, anisotropic etching (trench etching) is performed from the upper surface of the silicon substrate 1, and the space A is formed by bending the support portion (base plate portion 3 + square frame portion 5).
7 whose distance is substantially equal to the distance B between the movable electrode and the fixed electrode
a, 7b, 7c, 7d, a beam structure 20 having beams 22a, 22b whose interval σ2 is substantially equal to the interval B by bending, and a fixed electrode 1
Grooves 4a to 4d and 21a to 21d extending in the vertical direction for forming the sections 2, 13, and 14 are formed.

In this trench etching step, in this embodiment, the width of the opening (groove 4a
4d, 21a to 21d), making it difficult to receive the microloading effect.
4d and 21a to 21d can be made substantially equal in depth. Therefore, when a silicon substrate in which an oxide film is buried as an etching stopper is used to have a uniform groove depth, the cost increases because the substrate has a special structure, but in this embodiment, The groove depth can be made substantially equal using a single-layer silicon substrate, which is advantageous in cost. Also, Japanese Patent Application Laid-Open
In the case where the groove depth is made uniform using the method disclosed in Japanese Patent Application Laid-Open No. 150462, there is a problem that the process becomes complicated, but in the present embodiment, the groove depth is increased without complicating the process. Can be made substantially equal.

Returning to the description of the manufacturing process, a protective film 60 is formed on the inner wall surface (side surface and bottom surface) of the groove subjected to anisotropic etching during or after the anisotropic etching to protect the side wall. Further, the protective film attached to the bottom of the groove is removed. Thus, the protective film 60 is formed on the side walls of the grooves 4a to 4d except for the bottom surfaces of the grooves 4a to 4d.

Here, it is necessary to select a protective film 60 suitable for the manufacturing process. Specifically, a polymer or the like is formed at the time of trench etching, a thermal oxide film is formed, an oxide film is formed by CVD or the like, O ₂
A thin oxide film may be formed with plasma or the like, or an oxide film may be formed with a chemical solution or the like. In addition, the wiring material is appropriately selected according to the method of forming the protective film to be employed, and the mask is devised. That is, when the heat process is not applied, the wiring materials 53 and 54 may be a metal wiring such as aluminum or a material such as polysilicon, and there is no problem even if a photoresist or the like is left for the mask for forming the structure. Absent. On the other hand, when a heating process is applied, the wiring materials 53 and 54 may be made of a material such as tungsten or a compound thereof, which is a high melting point metal, or polysilicon. After peeling, a structure is formed using an oxide film mask.

As described above, an oxide film can be used as the protective film 60, and in particular, a thermal oxide film can be used as the oxide film. If the oxide film is formed by oxygen plasma treatment, the side wall protective film can be formed more easily than thermal oxidation, and a heat process is not added. Manufacturing methods can be used. Further, as the protective film, a film generated at the time of etching for forming a groove (a side wall protective film generated at the time of etching) may be used.

Returning to the description of the manufacturing process, next, as shown in FIG. 9, isotropic etching is performed on the silicon substrate 1 from the bottom surfaces of the grooves 4a to 4d and 21a to 21d to form the cavity 2 extending in the lateral direction. Form. Thereby, the support portion (the base plate portion 3 located below the cavity 2, the cavity 2 and the groove 4).
Square frame portions 5) located beside a, 4b, 4c, and 4d, beam structures 6, 20, and fixed electrodes 12, 13, 14 are defined. In FIG. 9, only the silicon under the structure and the fixed electrode is removed by etching. In particular, the structure and the base plate portion 3 are completely separated from each other, and a gap having a predetermined interval t is formed below.

In the isotropic etching, it is necessary to select a combination that does not etch the sidewall protective film 60 described above. In addition, in this isotropic etching, by using a plasma etching process using a gas material such as SF ₆ or CF ₄ , the yield of structure formation after etching is improved as compared with a wet process. Can be.

Finally, when the etching mask 70 is removed, the angular velocity sensor shown in FIG. 3 can be formed.
Through such a step, the movable structure can be formed of single crystal silicon without using a substrate bonding step or a bonded wafer, so that an inexpensive and highly reliable sensor can be formed. . More specifically, a single-layer semiconductor substrate (single-crystal silicon substrate) can be used as a start wafer at the time of manufacture in an angular velocity sensor of a type that is electrostatically driven and electrostatically detected on one silicon substrate. The manufacturing cost of the sensor can be significantly reduced without using a substrate bonding step.

[Brief description of the drawings]

FIG. 1 is a plan view showing a semiconductor angular velocity sensor according to an embodiment.

FIG. 2 is a plan view of the semiconductor angular velocity sensor of FIG. 1 from which electric wiring is removed.

FIG. 3 is a longitudinal sectional view taken along line II of FIG. 1;

FIG. 4 is an enlarged view of a main part of the semiconductor angular velocity sensor according to the embodiment;

FIG. 5 is a cross-sectional view for explaining a manufacturing process.

FIG. 6 is a cross-sectional view for explaining a manufacturing process.

FIG. 7 is a cross-sectional view for explaining a manufacturing process.

FIG. 8 is a cross-sectional view for explaining a manufacturing process.

FIG. 9 is a cross-sectional view for explaining a manufacturing process.

FIG. 10 is a diagram showing a bent state of a beam.

FIG. 11 is a diagram showing bending of a beam.

FIG. 12 is a plan view of a conventional angular velocity sensor.

FIG. 13 is an enlarged view of a main part of a conventional angular velocity sensor.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Single crystal silicon substrate, 2 ... Cavity, 3 ... Base plate part, 4a-4d ... Groove, 5 ... Square frame part, 7a, 7b, 7
c, 7d: beam, 9: movable electrode for excitation, 10: movable electrode for vibration monitoring, 11: movable electrode for dummy, 12: fixed electrode for excitation, 13: fixed electrode for vibration monitoring, 14: fixed electrode for dummy, 22a, 22b: beam, 25: fixed electrode for detection, 26: fixed electrode for detection.

Continued on the front page (72) Inventor Yukihiro Takeuchi 1-1-1, Showa-cho, Kariya-shi, Aichi F-term in DENSO Corporation (Reference) 2F105 BB01 BB15 CC04 CD03 CD05 CD13 4M112 AA02 CA12 CA24 CA36 DA02 DA03 EA03 EA06 EA11

Claims (38)

[Claims] 1. A support section (3, 5) defined by at least a laterally extending cavity (2) formed in a semiconductor substrate (1), and a vertical section formed in the cavity (2) and the semiconductor substrate (1). Beam structure defined by grooves (4a, 4b, 4c, 4d) extending in the direction, located above the cavity (2), extending from the support (3, 5) and having a movable electrode (9) (6)
And is defined by the cavity (2) and the grooves (4a, 4b, 4c, 4d), is located above the cavity (2), and is movable from the support portion (3, 5) to the beam structure (6). A fixed electrode (12) extending in the direction (X) and opposed to the movable electrode (9) of the beam structure (6); By bending the beams (7a, 7b, 7c, 7d) of the body (6), the distance (A) between the bent beams and the distance (B) between the movable electrode (9) and the fixed electrode (12) is changed. And a movable range of the beam structure (6) is set to a predetermined range larger than a distance (B) between the movable electrode (9) and the fixed electrode (12). . 2. The beam (7a, 7b, 7) in a direction (Y) orthogonal to a movable direction (X) of the beam structure (6).
The distance (C) between the tip of (c, 7d) and the facing part (12) is substantially equal to the distance (B) between the movable electrode (9) and the fixed electrode (12). Semiconductor dynamic quantity sensor. 3. A distance (D) between a side surface of the beam (7a, 7b, 7c, 7d) and an opposing portion (5, 23) in a movable direction (X) of the beam structure (6) is determined by the movable electrode. 3. The semiconductor dynamic quantity sensor according to claim 1, wherein the distance (B) between (9) and the fixed electrode (12) is substantially equal. 4. A beam (7a, 7b, 4) of said beam structure (6).
7c, 7d), the number of bends is "n", the distance (B) between the movable electrode (9) and the fixed electrode (12) and the distance (A) between the bent beams are "dn". When the minimum distance between the tip of the movable electrode and the opposing portion during the movement of the body is “Δ”, and the required total amplitude / 2 is “δ”, δ = {(n + 1) / 2} · dn−Δ By determining n, the distance between the beams,
The semiconductor dynamic quantity sensor according to any one of claims 1 to 3, wherein a distance between the movable electrode (9) and the fixed electrode (12) is equalized. 5. Supports (3, 5) defined by at least a laterally extending cavity (2) formed in the semiconductor substrate (1), and a vertical portion formed in the cavity (2) and the semiconductor substrate (1). Beam structure defined by grooves (4a, 4b, 4c, 4d) extending in the direction, located above the cavity (2), extending from the support (3, 5) and having a movable electrode (9) (6)
And is defined by the cavity (2) and the grooves (4a, 4b, 4c, 4d), is located above the cavity (2), and is movable from the support portion (3, 5) to the beam structure (6). A fixed electrode (12) extending in the direction (X) and facing the movable electrode (9) of the beam structure (6), comprising: A dummy movable electrode (11) extending in the movable direction (X) is provided in (6), and extends in the movable direction (X) of the beam structure (6) from the support portion (3, 5), and the beam is provided. A semiconductor dynamic quantity sensor comprising a dummy fixed electrode (14) facing a dummy movable electrode (11) of a structure (6). 6. An interval (E) between the dummy movable electrode (11) and the dummy fixed electrode (14) is substantially equal to an interval (B) between the movable electrode (9) acting as an electrode and the fixed electrode (12). The semiconductor dynamic quantity sensor according to claim 5, wherein 7. Supports (3, 5) defined by at least a laterally extending cavity (2) formed in the semiconductor substrate (1), and a vertical portion formed in the cavity (2) and the semiconductor substrate (1). A primary vibrator defined by grooves (4a, 4b, 4c, 4d) extending in the direction, located above the cavity (2), extending from the support (3, 5) and having a movable electrode (9) Beam structure (6), the cavity (2) and the grooves (4a, 4b, 4c, 4d), which are defined on the cavity (2) and from the support portions (3, 5). A primary fixed electrode (12) extending in the movable direction (X) of the structure (6) and facing the movable electrode (9) of the beam structure (6); and a beam for the primary vibrator. The structure (6) is defined by grooves (21a, 21b, 21c, 21d) extending in the vertical direction. A beam structure for a secondary vibrator (20) having a movable electrode (25), and a groove (21a, 21b) extending in the longitudinal direction in the beam structure for a primary vibrator (6), and the secondary vibration The secondary beam structure (20) extends in the direction (X) orthogonal to the movable direction (Y) of the secondary beam structure (20), and is arranged to face the movable electrode (25) of the secondary vibrator beam structure (20). A secondary fixed electrode (26, 27), comprising: a beam (7a, 7b, 7c, 7d, 22a) of the beam structure (6, 20) for the primary and secondary oscillators. , 22b), the distance between the folded beams (A,
σ2) is substantially equal to the distance (B) between the primary fixed electrode (12) and the movable electrode (9) of the primary vibrator beam structure (6), and the primary vibrator beam structure (6) The movable range is set to a predetermined range larger than the distance (B) between the primary fixed electrode (12) and the movable electrode (9) of the primary oscillator beam structure (6), and further, the secondary oscillator beam structure is provided. As the arrangement of the movable electrode (25) and the secondary fixed electrode (26, 27) in (20), the movable electrode (25), the fixed electrode (26), the fixed electrode (27), the movable electrode (25), the fixed electrode ( 26) A semiconductor dynamic quantity sensor which is arranged at regular intervals in the order of the fixed electrode (27). 8. A primary vibrator beam structure (6) is provided with a dummy movable electrode (11) extending in a movable direction (X), and a primary vibrator beam structure is provided from a support portion (3, 5). A dummy fixed electrode (14) extending in the movable direction (X) of (6) and facing the dummy movable electrode (11) of the primary vibrator beam structure (6) is provided. The semiconductor dynamic quantity sensor according to claim 7. 9. The distance (E) between the dummy movable electrode (11) and the dummy fixed electrode (14) is substantially equal to the distance (B) between the movable electrode (9) acting as an electrode and the fixed electrode (12). 9. The semiconductor dynamic quantity sensor according to claim 8, wherein: 10. A beam structure for a secondary vibrator (20) in a movable direction (Y) of the beam structure for a secondary vibrator (20).
The distance (σ1) between the tip of the beam (22a, 22b) and the opposing portion (23) is determined by the distance between the movable electrode (25) and the secondary fixed electrode (26, 27) of the secondary vibrator beam structure (20). The interval (M,
The semiconductor dynamic quantity sensor according to any one of claims 7 to 9, wherein N) is substantially equal to N). 11. The tip of the movable electrode (25) of the secondary vibrator beam structure (20) in a direction (X) orthogonal to the movable direction (Y) of the secondary vibrator beam structure (20). And the distance (J) between the opposing portion (23) and the beam structure for secondary vibrator (2).
0) movable electrode (25) and secondary fixed electrode (26, 27)
The semiconductor dynamic quantity sensor according to any one of claims 7 to 10, wherein the distance is substantially equal to the interval (M, N). 12. A movable electrode (25) of the secondary vibrator beam structure (20) in a direction (X) orthogonal to the movable direction (Y) of the secondary vibrator beam structure (20). The tip of the disposed secondary fixed electrode (26, 27) and the facing portion (2)
The distance (K, L) of 4) is changed to the beam structure (2
0) movable electrode (25) and secondary fixed electrode (26, 27)
The semiconductor dynamic quantity sensor according to any one of claims 7 to 11, wherein the distance (M, N) is substantially equal to the distance (M, N). 13. The semiconductor physical quantity sensor according to claim 7, wherein the semiconductor physical quantity sensor is operated as an angular velocity sensor. 14. The primary vibrator beam structure (6), wherein the movable electrode (9) and the primary fixed electrode (12) are of an electrostatic drive type using comb-teeth electrodes.
4. The semiconductor physical quantity sensor according to 3. 15. The device according to claim 14, wherein the movable electrode (9) and the primary fixed electrode (12) of the primary oscillator beam structure (6) are arranged at equal intervals. Semiconductor dynamic quantity sensor. 16. An interval (X) between the tip of the movable electrode (9) of the primary oscillator beam structure (6) and the facing portion (5) in the movable direction (X) of the primary oscillator beam structure (6). P1),
16. The semiconductor dynamic quantity sensor according to claim 14, wherein the amplitude when exciting the beam structure for the primary vibrator is substantially equal to the amplitude. 17. The distance (P2) between the tip of the primary fixed electrode (12) and the facing portion (23) in the movable direction (X) of the primary vibrator beam structure (6) is determined by changing the distance between the tip (P2) and the primary vibrator beam structure. The semiconductor dynamic quantity sensor according to claim 14 or 15, wherein the amplitude when exciting the body (6) is substantially equal to the amplitude. 18. The distance (B) between the primary fixed electrode (12) and the movable electrode (9) in the primary vibrator beam structure (6) is determined by the secondary in the secondary vibrator beam structure (20). 2. The distance (M, N) between the fixed electrode (26, 27) and the movable electrode (25) is substantially equal.
The semiconductor physical quantity sensor according to any one of items 4 to 17. 19. The length (S, T) of the primary fixed electrode (12) and the movable electrode (9) of the beam structure for a primary vibrator (6).
19. The amplitude of exciting the beam structure (6) for the primary vibrator (6) is substantially twice or less.
The semiconductor dynamic quantity sensor according to any one of the above items. 20. The widths (U, V) of the primary fixed electrode (12) and the movable electrode (9) of the beam structure for a primary vibrator (6).
20. The semiconductor dynamic quantity according to any one of claims 14 to 19, wherein the distance is set to about 1/3 to 3 times the distance (B) between the primary fixed electrode (12) and the movable electrode (9). Sensors. 21. The distance between the primary vibrator beam structure (6) and the fixing member (5) on the outer peripheral side thereof is set to a secondary fixed electrode (26, 27) in the secondary vibrator beam structure (20). 21. The semiconductor physical quantity sensor according to claim 14, wherein a distance between the movable electrode and the movable electrode is substantially equal. 22. The apparatus according to claim 14, further comprising means (10, 13) for monitoring the amplitude of the primary vibrator beam structure (6) during excitation. Semiconductor dynamic quantity sensor according to the above. 23. The apparatus according to claim 13, wherein the amplitude at the time of excitation of said primary vibrator beam structure (6) is controlled to be constant by negative feedback using the monitoring result of said monitoring means (10, 13). The semiconductor dynamic quantity sensor according to claim 22, wherein: 24. A comb-shaped electrode (1) as the vibration monitor.
24. The semiconductor physical quantity sensor according to claim 23, wherein a capacitance detection method according to (0, 13) is used. 25. The vibration monitoring electrodes (10, 13).
26. The semiconductor dynamic quantity sensor according to claim 24, wherein the shape of the fixed electrode and the movable electrode are arranged at equal intervals. 26. The apparatus according to claim 22, wherein an interval (α1) between the tip of the vibration monitoring movable electrode (10) and the facing portion (5) is substantially equal to the amplitude at the time of excitation. 2. The semiconductor dynamic quantity sensor according to claim 1. 27. An apparatus according to claim 22, wherein an interval (α2) between a tip of said fixed electrode for vibration monitoring and said opposing portion is substantially equal to an amplitude at the time of excitation.
The semiconductor dynamic quantity sensor according to any one of the above items. 28. The distance (W1, W2) between the vibration monitoring movable electrode (10) and the vibration monitoring fixed electrode (13).
The distance (M, N) between the secondary fixed electrodes (26, 27) and the movable electrode (25) in the secondary vibrator beam structure (20)
The semiconductor dynamic quantity sensor according to any one of claims 22 to 27, wherein: 29. Lengths (β1, β2) of the vibration monitoring movable electrode (10) and the vibration monitoring fixed electrode (13)
The semiconductor dynamic quantity sensor according to any one of claims 22 to 28, wherein is set to be approximately twice or less the amplitude at the time of excitation. 30. The width (γ1, γ2) of the vibration-monitoring movable electrode (10) and the vibration-monitoring fixed electrode (13) is determined by the following:
The semiconductor dynamic quantity sensor according to any one of claims 22 to 29, wherein the distance (W1, W2) in step 3) is set to about 1/3 to 3 times. 31. The dummy electrode according to claim 5, wherein the dummy movable electrode and the dummy fixed electrode are arranged at equal intervals. Semiconductor dynamic quantity sensor according to the above. 32. The distance (λ1) between the tip of the dummy movable electrode (11) and the facing portion (5) is substantially equal to the amplitude at the time of excitation. The semiconductor dynamic quantity sensor according to any one of the above items. 33. An apparatus according to claim 5, wherein the distance (λ2) between the tip of the dummy fixed electrode (14) and the facing portion (23) is substantially equal to the amplitude at the time of excitation. The semiconductor dynamic quantity sensor according to any one of the above items. 34. The distance (E) between the dummy fixed electrode (14) and the dummy movable electrode (11) is set to be equal to the distance between the secondary fixed electrodes (26, 27) in the secondary vibrator beam structure (20). The semiconductor dynamic quantity sensor according to any one of claims 8 to 33, wherein the distance (M, N) between the movable electrodes (25) is substantially equal to the distance. 35. The length (λ3) of the dummy fixed electrode (14) and the dummy movable electrode (11) is set to approximately twice or less the amplitude at the time of excitation. The semiconductor dynamic quantity sensor according to any one of the above items. 36. The width (λ4) of the dummy fixed electrode (14) and the dummy movable electrode (11) is set to one of the distance (E) between the dummy fixed electrode (14) and the dummy movable electrode (11).
36. The method according to claim 8, wherein the ratio is about 程度 to 3 times.
The semiconductor dynamic quantity sensor according to any one of the above items. 37. The distance between the beam structure (6, 20) and the fixing members (5, 23) around the beam structure (6, 20) is set to be equal to the distance between the secondary fixed electrodes (26, 27) in the secondary vibrator beam structure (20). 37. The semiconductor physical quantity sensor according to claim 7, wherein the distance (M, N) between the movable electrodes (25) is substantially equal to the distance (M, N). 38. A support (3, 5) and a movable electrode (9).
A method for manufacturing a semiconductor dynamic quantity sensor, comprising: a beam structure (6) having: a fixed electrode (12) arranged to face a movable electrode (9) of the beam structure (6). Anisotropic etching is performed from the upper surface of the substrate (1),
A beam structure (6) having at least a support (3, 5) and a beam (7a, 7b, 7c, 7d) whose distance (A) is substantially equal to the distance (B) between the movable electrode and the fixed electrode by bending. And grooves (4a, 4b, 4c, 4d) extending in the vertical direction for defining and forming the fixed electrode (12), and grooves (4a, 4b, 4b, 4c, 4d) excluding the bottom surfaces of the grooves (4a, 4b, 4c, 4d). , 4b, 4c, 4d) on the side walls of the protective film (6
0), and isotropically etching the semiconductor substrate (1) from the bottom surface of the groove (4a, 4b, 4c, 4d) to form a cavity (2) extending in the lateral direction and support the semiconductor substrate (1). A step of partitioning the part (3, 5), the beam structure (6), and the fixed electrode (12).