JP5067295B2 - Sensor and manufacturing method thereof - Google Patents

Description

  The present invention relates to a sensor capable of detecting a physical quantity and a manufacturing method thereof, and more particularly to a sensor of a type that detects acceleration in a plurality of directions using a piezoresistive element as a detection element and a manufacturing method thereof.

  In recent years, a sensor using a piezoresistor as a detection element has been put to practical use as a sensor having a small and simple structure using a MEMS (Micro Electro Mechanical Systems) technology. Such a sensor is a sensor that detects physical quantities such as acceleration and pressure. Piezoresistive acceleration sensors are used in a wide range of fields such as HDD (hard disk drive) drop detection devices, in-vehicle airbag devices, mobile terminals such as mobile phones, and game machines.

As a conventional acceleration sensor, there is one described in Patent Document 1, for example. A flexible part having two pairs of beam structures orthogonal to each other, and a thicker body thicker than the flexible part and a fixed part disposed around the flexible body are connected by the flexible part. There are four piezoresistive elements for each axis on the beam so that the piezoresistive elements formed on the same beam in the axial direction and the piezoresistive element formed on the other beam orthogonal to the Y-axis direction are detected. A formed sensor structure is shown. Note that, as described in FIG. 1 of Patent Document 1, the piezoresistive element is connected to a lead electrode arranged to form a bridge circuit for each axis and an external connection pad connected to the lead electrode. .
JP 2003-101033 A

  When the conventional acceleration sensor is actually used, the acceleration acting on the weight body is detected by measuring the output caused by the distortion of the flexible portion in a state where the voltage is applied to the bridge circuit. In the piezoresistive acceleration sensor, Joule heat is generated from the piezoresistive element in order to detect acceleration in an energized state. In general, the resistance characteristic of a piezoresistor changes depending on the operating temperature. Therefore, there is a problem that the output characteristic of the sensor fluctuates due to thermal fluctuations every energization time.

  Therefore, in view of the above, an object of the present invention is to provide a sensor that suppresses fluctuations in output characteristics due to energization and a method for manufacturing the sensor.

  The sensor according to the present invention includes a weight part, a frame part positioned around the weight part, a diaphragm-like or beam-like flexible part connecting the frame part and the weight part, the flexible part, And / or a strain detection unit disposed on the frame unit for detecting a physical quantity in at least one axial direction, and the strain detection unit connected to form a bridge circuit on the flexible unit and the frame unit. A plurality of external connection pads connected to the strain detection unit via the wiring unit, and the external connection pad connected to the strain detection unit via the wiring unit. Another dummy pad is arranged on the frame portion.

  A sensor manufacturing method according to the present invention is a method of manufacturing a sensor that detects a physical quantity in at least one axial direction using a semiconductor substrate, and diffuses impurities on one surface of the semiconductor substrate to provide a strain detection unit. Forming a weight part, a frame part located around the weight part, and a diaphragm-like or beam-like flexible part connecting the frame part and the weight part from the semiconductor substrate; A step of forming a wiring portion for connecting the strain detection portion so as to form a bridge circuit on the flexible portion and the frame portion; and a connection with the strain detection portion via the wiring portion. Forming a plurality of external connection pads and a dummy pad connected to the strain detection section via the wiring section and different from the external connection pads on the frame section. And

  ADVANTAGE OF THE INVENTION According to this invention, the sensor which suppressed the output characteristic fluctuation | variation by electricity supply, and its manufacturing method can be provided.

Hereinafter, a semiconductor triaxial acceleration sensor which is an aspect of the sensor according to the present invention will be described with reference to the drawings. FIG. 1 is an overall perspective view of an acceleration sensor according to the present invention.
As shown in FIG. 1, the acceleration sensor 1 is a substantially rectangular parallelepiped, and is composed of a sensor body 2 made of a semiconductor substrate and a support substrate 3 made of glass or the like. In the figure, two axes (X axis and Y axis) perpendicular to the plane of the acceleration sensor are set, and a direction perpendicular to the two axes is defined as the Z axis. The sensor body 2 is composed of an SOI (Silicon On Insulator) substrate 110, and is formed by sequentially laminating a silicon film 120, a silicon oxide film 130, and a silicon substrate 140. A weight body (weight part 142) is arranged at substantially the center, and a frame (frame part 121 and frame part 141) having an opening around the weight body is positioned to connect the weight body and the frame. It is the structure which has the beam (flexible part 123) which has. The support substrate 3 has a function as a pedestal for supporting the sensor body 2 and a function as a stopper substrate for restricting excessive displacement of the weight body in the downward direction (Z-axis negative direction). When the sensor body 2 is directly mounted on a package substrate (not shown), the support substrate 3 is not necessarily required.

  FIG. 2 is an exploded perspective view of the acceleration sensor. The silicon film 120 includes a fixed frame part 121 (the upper part of the frame), a weight joint part 122 disposed in the frame part 121, and two pairs (four in total) that connect the frame part 121 and the weight joint part 122. The flexible part 123 is provided. The frame portion 121, the weight joint portion 122, and the flexible portion 123 are defined by the opening 124. The frame part 121 is joined to the frame part 141 (the lower part of the frame) via the silicon oxide film 130. Further, the weight joint portion 122 is joined to the weight portion 142 having a substantially crowbar shape in the vertical view through the silicon oxide film 130. The weight part 142 is disposed in the frame part 141 so as to be separated.

  The support substrate 3 is made of, for example, a glass substrate and is bonded to the sensor body 2 by anodic bonding. The support substrate is not limited to a glass substrate, and a metal plate (stainless steel, an invar made of Fe-36% Ni alloy, etc.), an insulating resin plate, a semiconductor substrate such as Si can be used. As a joining method, it can select suitably from direct joining, eutectic joining, joining by an adhesive agent, etc.

FIG. 3 is a plan view and a cross-sectional view of the acceleration sensor. FIG. 3A is a plan view of the acceleration sensor, and strain detectors Rx to Rz for detecting accelerations in the triaxial (XYZ) directions are arranged on the four flexible portions 123. The strain detectors Rx to Rz are arranged in a region where the flexible part 123 is connected to the frame part 121 and the weight joint part 122. In the drawing, strain detection units Rx1 to Rx4 and Rz1 to Rz4 are arranged in a pair of flexible parts arranged in the direction along the X axis in order to detect acceleration in the X direction and the Z direction. On the other hand, strain detection units Ry1 to Ry4 for detecting acceleration in the Y direction are arranged in a pair of flexible parts arranged in the direction along the Y axis. Note that the strain detectors Rz1 to Rz4 may be arranged in a pair of flexible parts arranged in the direction along the Y axis. Although an example in which the strain detection unit R is formed of a linear piezoresistive element is illustrated, it may have a folded structure at least once and have a meandering shape. This will be described later.
FIG. 3B is a cross-sectional view of the acceleration sensor along A-A. The lower surface of the weight portion 142 is higher than the lower end of the frame portion 141, and the Z-negative direction is caused by a gap between the acceleration sensor and the glass substrate 3. It is set so that a certain amount of displacement is possible. FIG. 3C is a cross-sectional view of the acceleration sensor along BB, and the flexible portion 123 is a self-supporting thin film having flexibility.

  In FIG. 3A, the strain detection part R is arranged with Rx1 to Rx4 on one side and Rz1 to Rz4 on the other side with the width center line of the flexible part 123 as a boundary, but this is limited to this arrangement. It is not something. For example, the strain detection unit Rx (which may be Ry or Rz) may be arranged point-symmetrically with respect to the center of the weight part 142, and in that case, the torsion of the flexible part 123 caused by acceleration (the weight 142) The detection of physical quantity fluctuations in the direction not desired to be detected (rotation sensitivity) due to rotation) can be suppressed. Therefore, the accuracy of the detection signal can be improved.

  In addition, although the strain detection unit Ry is arranged in the flexible portion 123 along the Y axis, the strain detection unit Ry and the dummy strain detection unit are arranged so as to have the same layout as the flexible portion 123 along the X axis. And may be arranged symmetrically with respect to the width center line of the flexible portion. By disposing the dummy strain detector, the layout of the strain detector R and the wiring 152 on the flexible portion can be improved in symmetry, and the offset voltage can be reduced.

  FIG. 4 is a diagram for explaining the details of the flexible portion. FIG. 4A is a plan view of a flexible part in which a strain detection unit is arranged, and FIG. 4B is a cross-sectional view taken along a line CC in the plan view of FIG. The strain detection part R is a piezoresistive element placed near the stress concentration part of the flexible part 123. One end of the strain detection unit R is disposed so as to be in contact with the boundary where the flexible portion 123 and the frame portion 121 are connected and the boundary where the flexible portion 123 and the weight joint portion 122 are connected. Note that the distortion detection unit R may be arranged so as to cross the boundary, and the arrangement position can be changed according to the output difference for each axis.

  An insulating layer 150 is formed on the strain detection portion R, and the insulating layer 150 has a contact hole 151 at a location where the wiring 152 is connected. In order to lower the connection resistance, a portion exposed by the contact hole 151 may be provided with a high concentration diffusion region (an impurity region whose concentration is higher by about one digit than the piezoresistive element region). Note that a total of twelve strain detection units, Rx1 to Rx4, Ry1 to Ry4, and Rz1 to Rz4, are connected for each detection direction to form a bridge circuit. An external connection pad P and a dummy pad P ′ (described later) for connecting to an external circuit, which will be described later, are provided on the frame portion 121, and an electrical signal accompanying acceleration is connected to the wiring 152 and the external connection pad P. It is taken out to the external circuit. In order to protect the wiring 152, a protective layer 153 is formed over the wiring 152.

  FIG. 5 is a diagram illustrating a strain detection unit having a folded structure. The strain detection unit R is a mode in which the piezoresistive elements are folded (see FIG. 5A), or a mode in which a plurality of piezoresistive elements are connected in series by metal wiring or high-concentration impurity diffusion layers to form one strain detection unit. (See FIG. 5B). The strain detector R has a structure that is folded at least once (folded portion 170). The strain detector is efficiently arranged in the stress concentration portion and has a predetermined resistance value (corresponding to a resistance length). It is possible to provide a sensor that can be obtained and can cope with high sensitivity and low power consumption.

In FIG. 5A, the folded portion 170 is formed of a layer having a concentration region (10 ¹⁷ to 10 ¹⁹ atm / cm ³ ) substantially the same as that of the piezoresistive element. In FIG. 5B, the folded portion 170 is constituted by a metal wiring or a high concentration impurity diffusion layer. The high-concentration impurity diffusion layer refers to a low-resistance layer whose impurity concentration is one digit or more higher than that of the piezoresistive element.

The length of the strain detection portion R can be set as appropriate in the range of 10 μm to 100 μm, the width is in the range of 0.2 μm to 10 μm, the length of the folded portion 170 is 2 μm to 20 μm, and the width is in the range of 0.2 μm to 10 μm. However, the width of the strain detection part is preferably 2 μm or more, more preferably 3 μm or more, and preferably 1/10 or less of the width of the flexible part. If the thickness is less than 2 μm, defects in the lithography process when forming the strain detection portion tend to appear, leading to variations in sensor characteristics. It has been confirmed by the present inventors that the output fluctuation can be further suppressed when the thickness is 3 μm or more. If the width of the strain detection part is larger than 1/10 of the width of the flexible part, a resistance change due to twisting is detected greatly.
Moreover, it is preferable that the space | interval between adjacent strain detection parts is 5 micrometers or more. This is because when the thickness is less than 5 μm, the impurity diffusion regions overlap between adjacent strain detection portions, and current leakage occurs.

Next, the wiring layout according to the present invention will be described in detail with reference to FIG.
FIG. 6 shows a wiring layout. In addition to the external connection pad P for connecting to an element such as an IC on the mounting board (not shown) side, a dummy pad P ′ different from the external connection pad is connected to the wiring 152 on the frame portion 121. Is arranged. When the external connection pad is connected to an element such as an IC, the dummy pad P ′ is not normally used for external connection. The external connection pad P and the dummy pad P ′ are made of a conductive material, for example, a metal material. Since the external connection pad P is used for connection to an external element by means such as wire bonding, the protective layer 153 is opened to expose the surface, and is in contact with the atmosphere. The dummy pad P ′ is not connected to an external element, but may be exposed to the atmosphere with its surface exposed. The surface of the external connection pad P and the dummy pad P ′ may be protected by a conductive barrier conductive layer (not shown) to prevent surface deterioration.

  The dummy pad P ′ is connected to the strain detector R via the wiring 152. In the drawing, the dummy pad is connected to the strain detection unit by a wiring 152 ′ branched from the wiring 152 connecting the external connection pad P and the strain detection unit. Therefore, the heat generated in the strain detection unit R can be effectively released to the outside through the dummy pad P ′. In particular, in the case of a small acceleration sensor (for example, the outer diameter is 1 to 3 mm), the width / length of the strain detection unit must be designed to be small as the sensor housing is downsized. For this reason, since the resistance value of the strain detection unit is reduced, the power consumption is increased and the amount of heat generation is increased. By disposing the dummy pad P ′, Joule heat generated in the strain detection unit can be effectively released to the outside. The dummy pads P ′ are preferably arranged on the frame part 121 with as wide an area as possible, and the total surface area of the dummy pads P ′ is 1.5 times or more of the total surface area of the external connection pads P. It is preferable.

In FIG. 6, among the external connection pads P, the input pad for supplying the input voltage is Vd, the ground terminals are XG, YG, ZG, and the output pads for detecting the output voltage in the XYZ directions are respectively (X ₊ , X ₋ ), (Y ₊ , Y ₋ ), (Z ₊ , Z ₋ ), and the dummy pad P ′ is expressed as dm. FIG. 6 illustrates the case where all the dummy pads P ′ are connected to the input pad Vd.

FIG. 7 is a circuit diagram showing a detection circuit (bridge circuit). Each pad and the strain detection unit are connected via a wiring unit to form a bridge circuit. In the bridge circuit, the output voltage Vout (the voltage between | X ₊ −X ₋ |, the voltage between | Y ₊ −Y ₋ |, and the input voltage Vin (Vd) represented by the following expressions (1) to (3): The acceleration can be detected from the relationship of the voltage |) between | Z ₊ -Z ₋ .
Vxout / Vxin =
[Rx4 / (Rx1 + Rx4) −Rx3 / (Rx2 + Rx3)] (1)
Vyout / Vyin =
[Ry4 / (Ry1 + Ry4) −Ry3 / (Ry2 + Ry3)] (2)
Vzout / Vzin =
[Rz3 / (Rz1 + Rz3) −Rz4 / (Rz2 + Rz4)] (3)

  FIG. 8 is a diagram for explaining another aspect of pad connection. The dummy pad P ′ may be connected to the strain detector by a wiring 152 ′ branched from the wiring 152 connected to the output pad in each direction. The dummy pad P ′ is arranged in parallel to the external connection pad, and can be used as a substitute pad when the external connection pad is defective. Therefore, the defect rate can be reduced.

  FIG. 9 illustrates an acceleration sensor having a diaphragm-like flexible portion. The acceleration sensor according to the present invention may have a structure in which the weight part 142 disposed substantially at the center is connected to the frame part 121 by the thin diaphragm-like flexible part 123. FIG. 9A is a plan view of the acceleration sensor. A strain detection unit for detecting the X and Y axis directions is arranged along two orthogonal axes, and a strain detection unit for detecting the Z axis direction is arranged in a direction forming 45 ° with respect to the XY axis. FIG. 9B is a (DD) cross-sectional view along the X-axis in FIG. The weight part 142 is integrated with or joined to the diaphragm-like flexible part 123.

Next, a method for manufacturing the acceleration sensor according to the embodiment will be described with reference to FIG. FIG. 10 is a drawing showing a manufacturing method according to an embodiment of the present invention.
§Acceleration sensor manufacturing method (1) Preparation of SOI substrate (see FIG. 10A)
An SOI substrate 110 in which a silicon film 120, a silicon oxide film 130, and a silicon substrate 140 are stacked is prepared. As described above, the silicon film 120 is a layer constituting the frame part 121, the weight joint part 122, and the flexible part 123. The silicon oxide film 130 is a layer that joins the silicon film 120 and the silicon substrate 140 and functions as an etching stopper layer. The silicon substrate 140 is a layer constituting the frame part 141 and the weight part 142. The SOI substrate 110 is produced by SIMOX or a bonding method. In the SOI substrate 110, the silicon film 120, the silicon oxide film 130, and the silicon substrate 140 have thicknesses of 5 μm, 2 μm, and 600 μm, respectively. Note that a plurality of acceleration sensors 1 having an outer circumference of about 1 to 2 mm are arranged on a wafer having a diameter of 150 mm to 200 mm in a multifaceted manner.

(2) Formation of strain detection unit (see FIG. 10B)
A mask for impurity diffusion is formed on the silicon substrate 120 side of the SOI substrate 110 (not shown). As this mask material, for example, a silicon nitride film (Si ₃ N ₄ ), a silicon oxide film (SiO ₂ ), a resist, or the like can be used. Here, after a silicon oxide film is formed on the entire surface of the silicon film 120 by thermal oxidation or plasma CVD (Chemical Vapor Deposition), a silicon nitride film is formed, and a resist pattern (see FIG. An opening corresponding to the detection portion R is formed in the silicon nitride film and the silicon oxide film by wet etching such as RIE (Reactive Ion Etching) and hydrofluoric acid. The diffusion mask has a two-layer structure of a silicon oxide film and a silicon nitride film from the silicon film 120 side. The silicon nitride film is used for preventing diffusion of a diffusing agent, which will be described later.

A diffusion agent containing B (boron) or the like is applied on the silicon film 120. Then, heat treatment (about 1000 ° C.) is performed to diffuse (drive in) the silicon film 120 to form a piezoresistive element. Unnecessary diffusing agent is removed using hydrofluoric acid or the like. Thereafter, the silicon nitride film (Si ₃ N ₄ ) is removed by etching with hot phosphoric acid. In addition to the above-described coating diffusion method, a gas diffusion method using a dopant gas such as BBr ₃ can also be used.
Note that the piezoresistive element may be formed by an ion implantation method other than the thermal diffusion method. In the case of ion implantation, a resist can be used as a mask.

  In the case where the strain detecting portion R is continuously formed by the piezoresistive element by the folded portion 170, the impurity may be diffused into the silicon film 120 using a mask having a meandering opening. When the strain detection unit R is formed by connecting a plurality of piezoresistive elements in series with metal wiring, a plurality of piezoresistive elements are formed in this step, and then a wiring, an external connection pad, and a dummy described later are formed. The folded portion 170 may be formed by forming a film and patterning in the pad forming process. When the strain detection unit R is formed by connecting a plurality of piezoresistive elements in series with impurity diffusion wirings, the folded portion 170 is implanted by injecting impurities using a mask having an opening corresponding to the folded portion 170. Can be formed.

(3) Formation of insulating layer and contact hole (see FIG. 10C)
An insulating layer 150 is formed on the silicon film 120. For example, by thermally oxidizing the surface of the silicon film 120, the insulating layer can be formed of a SiO ₂ layer. The insulating layer may be formed by a plasma CVD method. A contact hole 151 is formed on the insulating layer 150 by RIE using a resist as a mask. Note that the contact hole 151 may be formed after the silicon substrate 140 is processed.

(4) Gap formation (see FIG. 10D)
The gap 160 is formed by etching the silicon substrate 140 using a mask having an opening along the inner frame of the frame portion 141. The gap 160 is an interval necessary for the weight portion 142 to be displaced downward (on the glass substrate 3 side), and is, for example, 5 to 10 μm. The value of the gap 160 can be appropriately set according to the dynamic range of the sensor.

(5) Processing of silicon substrate (see FIG. 10E)
Next, a mask for defining the frame part 141 and the weight part 142 is formed on the lower surface of the silicon substrate 140. Using this mask, the silicon substrate 140 is etched until the lower surface of the silicon oxide film 130 is exposed. It is preferable to use DRIE (Deep Reactive Ion Etching) for the etching.

In DRIE, an etching step of digging while eroding the material layer in the thickness direction and a deposition step of forming a polymer wall on the side wall of the carved hole are alternately repeated. Since the side wall of the hole that has been dug is protected by forming a polymer wall in sequence, it is possible to advance erosion almost only in the thickness direction. An ion / radical supply gas such as SF ₆ can be used as an etching gas, and C ₄ F ₈ or the like can be used as a deposition gas.

(6) Creation of wiring, external connection pads, and dummy pads (see FIG. 10F)
The wiring 152, the external connection pad P, and the dummy pad P ′ are formed. These are obtained by depositing a metal material such as Al, Al-Si alloy, Al-Nd alloy or the like to a thickness of, for example, 0.1 μm to 1.0 μm by vapor deposition or sputtering, and patterning it. It is done. The individual areas of the external connection pads and the dummy pads assume values that can be appropriately set according to the size of the sensor. Metal materials such as Cu and Au can also be used. However, when Cu is used, there is a possibility of metal diffusion, and when Au is used, the manufacturing cost increases. An Al-based material is preferable in terms of product reliability and manufacturing cost. Note that heat treatment (380 ° C. to 400 ° C.) is performed in order to form an ohmic contact between the wiring 152 and the strain detection portion R. The dummy pad P ′ is connected to a predetermined strain detector R by wirings 152 and 152 ′.

  By providing TiN or the like as a barrier conductive layer on the external connection pad P and the dummy pad P ′, surface deterioration can be prevented. The barrier conductive layer is formed with a thickness of 0.1 μm to 0.5 μm by sputtering.

Then, a film such as a silicon nitride film (Si ₃ N ₄ ) is provided as a protective layer 153 on the wiring 152. A portion covering the external connection pad P and the dummy pad P ′ is opened by RIE or wet etching. The protective layer 153 is preferably set as thin as possible for high sensitivity. For example, the thickness is preferably 0.3 μm to 0.6 μm.

  The acceleration sensor is formed by forming various layers (insulating layer, protective layer, wiring) on a Si substrate. Along with the Joule heat generated by the piezoresistor, stress is applied to the sensor housing due to the difference in thermal expansion coefficients of these layers. When a dummy pad that effectively radiates Joule heat is arranged in the frame part, the frame part is a thick part thicker than the flexible part, so it is suitable in that distortion due to the heat radiation of the dummy pad is difficult to occur It is. Although it may be difficult to provide an electrode for heat dissipation on the flexible part or the weight part due to the design rule of the wiring, the frame part is not easily restricted by the design rule. Further, since the wiring is not designed to be thick, a heat dissipation effect can be obtained without increasing the wiring resistance.

(7) Processing of silicon film (see FIG. 10G)
The silicon film 120 is etched by RIE until the upper surface of the silicon oxide film 130 is exposed to form an opening 124, thereby defining the frame part 121, the weight joint part 122, and the flexible part 123.

(8) Removal of unnecessary silicon oxide film (see FIG. 10H)
The unnecessary silicon oxide film used as an etching stopper is removed by RIE or wet etching. Thus, the silicon oxide film 130 exists between the frame portion 121 and the frame portion 141, and the weight joint portion 122 and the weight portion 142.

(9) Bonding of glass substrates (see FIG. 10I)
The sensor body 2 and the glass substrate 3 are joined. The glass substrate 3 is so-called Pyrex (registered trademark) glass containing movable ions such as Na ions, and anodic bonding is used for bonding to the SOI substrate 110. In order to prevent the weight 142 from sticking to the upper surface of the glass substrate 3 due to electrostatic attraction during anodic bonding, a sticking prevention film (not shown) such as Cr is formed on the upper surface of the glass substrate 3 by sputtering. You may keep it. Thereby, the sensor main body 2 and the glass substrate 3 are joined, and the acceleration sensor 1 is comprised.

(10) Individualization The acceleration sensor 1 is diced with a dicing saw or the like, and is divided into individual acceleration sensors 1. In the present specification, the “acceleration sensor” arranged in a multi-face manner on the wafer and the “acceleration sensor” separated into pieces are referred to as the acceleration sensor 1 without any particular distinction. The above is an example of the method of manufacturing the acceleration sensor according to the present invention, and the order can be changed as appropriate, and is not limited to the above.

  Although the acceleration sensor 1 is distributed as a single chip, it is distributed as an electronic component in combination with a package substrate / circuit substrate on which an active element such as an IC is mounted. The acceleration sensor 1 can be used in various applications such as a game machine and a mobile terminal (for example, a mobile phone). The external connection pad P of the acceleration sensor 1 and the package substrate / circuit substrate are electrically connected by a method such as wire bonding or flip chip.

Example 1
A 1.5 mm square acceleration sensor was fabricated on a SOI substrate (5 μm / 2 μm / 625 μm) with multiple faces. In addition to 11 external connection pads made of Al (each area: 100 μm × 100 μm), 27 dummy pads having the same area as the external connection pads were arranged. The output fluctuation when a voltage of 3 V was applied to the bridge circuit of the acceleration sensor was observed for 10 minutes (600 seconds). In the acceleration sensor according to this example, the output fluctuation after the initial operation (within 30 seconds of energization) was within a range of ± 0.03 mV, and the output fluctuation was very small.

(Example 2)
An acceleration sensor was fabricated with substantially the same configuration as in Example 1 except that the strain detection unit was configured with a piezoresistive element (width: 2 μm) having a one-time folding structure. In the acceleration sensor according to the present embodiment, the output fluctuation during energization is within a range of ± 0.03 mV, and the drift voltage value in the initial operation (within 30 seconds of energization) is 0.03 mV, which is stable immediately after energization. Was possible.

(Example 3)
An acceleration sensor was fabricated with substantially the same configuration as in Example 2 except that the width of the piezoresistive element, which is a strain detector, was set to 3 μm. In the acceleration sensor according to the present embodiment, the output fluctuation during energization is within a range of ± 0.03 mV, and the drift voltage value of the initial operation (within 30 seconds of energization) is less than 0.02 mV. Stable operation was possible.

(Comparative example)
For the conventional acceleration sensor provided with only 11 external connection pads (area: 100 μm × 100 μm) made of Al, the output fluctuation was measured in the same manner as described above, and the output fluctuation after the initial operation was in the range of ± 2.0 mV. Variations appeared in the output, and the output fluctuation was large. Further, the drift voltage value in the initial operation exceeded 2 mV, and the operation was unstable.

  As described above, the embodiments of the present invention are not limited to the above-described embodiments, and can be expanded and modified. The expanded and modified embodiments are also included in the technical scope of the present invention.

1 is an overall perspective view of an acceleration sensor according to the present invention. It is a disassembled perspective view of an acceleration sensor. It is the top view and sectional drawing of an acceleration sensor. It is drawing explaining the detail in a flexible part. It is drawing explaining the distortion | strain detection part which has a folding structure. It is drawing explaining a wiring layout. It is a circuit diagram which shows a detection circuit (bridge circuit). It is drawing explaining another aspect of pad connection. It is drawing explaining the acceleration sensor which has a diaphragm-like flexible part. It is drawing which represents the manufacturing method of the acceleration sensor which concerns on this invention.

Explanation of symbols

1: Acceleration sensor 2: Sensor body 3: Support substrate 110: SOI substrate 120: Silicon film 121: Frame portion 122: Weight joint portion 123: Flexible portion 124: Opening 130: Silicon oxide film 140: Silicon substrate 141: Frame portion 142: weight portion 150: insulating layer 151: contact hole 152: wiring 152 ′: wiring 153: protective layer 160: gap 170: folded portion

R: Strain detector Rx: X-axis direction detection strain detector Rx: Y-axis direction detection strain detector Rz: Z-axis direction detection strain detector P: External connection pad P ′: Dummy pad Vd: Input terminal X ₊ , X ₋ , Y ₊ , Y ₋ , Z ₊ , Z ₋ : output terminals XG, YG, ZG: ground terminals dm: dummy pads

Claims (7)

A weight part;
A frame portion located around the weight portion;
A diaphragm-like or beam-like flexible part that connects the frame part and the weight part;
A strain detector disposed on the flexible part and / or the frame part for detecting a physical quantity in at least one axial direction;
On the flexible part and the frame part, a wiring part that connects the strain detection part so as to constitute a bridge circuit, and
A plurality of external connection pads connected to the strain detection unit via the wiring unit;
A sensor, wherein a dummy pad that is connected to the strain detection unit via the wiring unit and is different from the external connection pad is disposed on the frame unit. The external connection pad has an input pad for inputting a voltage to the bridge circuit, and an output pad for taking out the output voltage of the bridge circuit,
2. The sensor according to claim 1, wherein any one of the dummy pads is connected to the strain detection unit by a wiring unit branched from the wiring unit that connects the input pad and the strain detection unit. 3. The sensor according to claim 2 , wherein any one of the dummy pads is connected to the strain detection unit by a wiring portion branched from the wiring portion that connects the output pad and the strain detection unit.   The sensor according to any one of claims 1 to 3, wherein a sum of areas of the dummy pads is 1.5 times or more of a sum of areas of the external connection pads. 5. The sensor according to claim 1, wherein the dummy pad can be used as a substitute pad for the external connection pad. 6. A method of manufacturing a sensor that detects a physical quantity in at least one axial direction using a semiconductor substrate,
A step of diffusing impurities on one surface of the semiconductor substrate to form a strain detector;
Forming a weight part, a frame part located around the weight part, and a diaphragm-like or beam-like flexible part connecting the frame part and the weight part from the semiconductor substrate;
Forming a wiring part on the flexible part and the frame part for connecting the strain detection part so as to form a bridge circuit;
A plurality of external connection pads connected to the strain detection unit through the wiring unit, and a dummy pad connected to the strain detection unit through the wiring unit and different from the external connection pad Forming on the part. A method for manufacturing a sensor. The manufacturing method according to claim 6, wherein the wiring portion, the external connection pad, and the dummy pad are formed simultaneously.