JP2011038872A - Mems sensor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a MEMS sensor with its detection accuracy enhanced by reducing a displacement amount of a sensor when stress acts on an anchor, through control of the width size T1 of an insulating layer and the width size T2 of a metallic layer, in particular. <P>SOLUTION: This MEMS sensor has: a first member 11; a second member 12; an intermediate member 13 positioned between the first and second members 11 and 12; the anchor 14 formed on the intermediate member 13 as well as the sensor 15 connected to the anchor 14; the insulating layer 16 interposed between the first member 11 and the anchor 14; and the metallic layer 17 interposed between the second member 12 and the anchor 14. A peripheral side surface of the metallic layer 17 is retreated inside a peripheral side surface of the anchor 14. Assuming that the width size of the insulating layer is T1 and the width size of the metallic layer 17 is T2, T1/T2 is in a range of 1.5-3.3. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention includes an intermediate member having a sensor portion and an anchor portion between a first member and a second member, and the first member and the anchor portion are joined via an insulating layer, and the second member and The present invention relates to a MEMS sensor in which the anchor portions are joined via a metal layer.

FIG. 16 is a partially enlarged longitudinal sectional view of a conventional MEMS sensor.
As shown in FIG. 16, the MEMS sensor A is made of, for example, a first member 1 made of silicon, a second member 2 made of silicon, and silicon positioned between the first member 1 and the second member 2. The intermediate member 3 is formed.

  As shown in FIG. 16, the intermediate member 3 includes an anchor portion 4 and a sensor portion 5 connected to the anchor portion 4. The anchor portion 4 has, for example, a quadrangular planar shape. In FIG. 16, the sensor unit 5 is directly connected to the anchor unit 4. However, a thin spring unit or the like may be interposed between the sensor unit 5 and the anchor unit 4. Further, the sensor unit 5 shown in FIG. 16 constitutes one electrode of a capacitive sensor, for example.

  As shown in FIG. 16, the insulating layer 6 is interposed between the first member and the anchor portion 4. For example, the SOI substrate is formed by a laminated structure of the first member 1, the insulating layer 6, and the intermediate member 3 shown in FIG.

  As shown in FIG. 16, an insulating layer 7 is provided on the inner surface 2 a of the second member 2, and a metal layer 8 is interposed between the insulating layer 7 and the anchor portion 4.

  Although not shown, for example, a conductive layer is wired from the metal layer 8 to the inside of the insulating layer 7 so that a detection signal based on a change in capacitance can be output.

  As shown in FIG. 16, the width dimension T <b> 7 of the metal layer 8 is formed sufficiently smaller than the width dimension T <b> 8 of the anchor portion 4. This is because the coefficient of thermal expansion of the metal layer 8 and each member made of silicon are greatly different, and deformation or peeling due to thermal stress is suppressed.

  On the other hand, as shown in FIG. 16, the width dimension T9 of the insulating layer 7 is formed to be equal to or slightly smaller than the width dimension T8 of the anchor portion 4.

  Conventionally, the width dimension T9 of the insulating layer 7 is not particularly controlled, and the width dimension T9 of the insulating layer 7 is sufficiently larger than the width dimension T7 of the metal layer 8.

  For this reason, when stress is applied between the first member 1 and the second member 2, the distribution of stress transmitted to the anchor portion 4 is greatly different from the first member 1 side and the second member 2 side. The anchor portion 4 was easily deformed. As a result, there is a problem in that the sensor unit 5 connected to the anchor unit 4 is bent and the amount of displacement increases and the detection accuracy decreases.

Japanese Patent Laid-Open No. 11-186567 JP 2003-273370 A JP 2000-49357 A

  In the inventions described in Patent Documents 1 to 3, the width dimension of the insulating layer 6 interposed between the first member 1 and the anchor portion 4 and the metal layer 8 interposed between the second member 2 and the anchor portion 4 are described. The relationship with the width dimension is not disclosed.

  Accordingly, the present invention is to solve the above-described conventional problems, and in particular, the displacement amount of the sensor portion when stress is applied to the anchor portion by controlling the width dimension T1 of the insulating layer and the width dimension T2 of the metal layer. It is an object of the present invention to provide a MEMS sensor capable of reducing the size and improving the detection accuracy.

The MEMS sensor in the present invention is
A first member, a second member, an intermediate member positioned between the first member and the second member, an anchor portion formed on the intermediate member, and a sensor portion connected to the anchor portion; An insulating layer interposed between the first member and the anchor portion, and a metal layer interposed between the second member and the anchor portion,
The outer peripheral side surface of the metal layer is receding inward from the outer peripheral side surface of the anchor part,
When the width dimension of the insulating layer is T1 and the width dimension of the metal layer is T2, T1 / T2 is in the range of 1.5 or more and 3.3 or less.

  In the present invention, the bonding strength between the first member and the anchor portion in the insulating layer can be sufficiently maintained, and when stress is applied to the anchor portion from the first member side and the second member side, The amount of displacement can be reduced, and a MEMS sensor with excellent detection accuracy can be obtained.

  In this invention, it is preferable that the outer peripheral side surface of the said insulating layer is formed in the inclined surface from which the width dimension of the said insulating layer becomes small gradually toward thickness direction from T1. The stress applied to the anchor portion can be more effectively dispersed, and the displacement amount of the sensor portion can be further reduced.

  Further, in the present invention, the width T1 of the insulating layer is defined on the bonding side with the first member, and the outer peripheral side surface of the insulating layer is gradually increased from the first member side toward the anchor portion side. It is preferable to incline so that a width dimension may become small.

  Further, in the present invention, when the width dimension of the anchor portion is T3 and the amount of retreat of the outer peripheral side surface of the insulating layer from the outer peripheral side surface of the anchor portion is T4, (T4 / T3) × 100 (%) is It is preferably specified in the range of 10% to 25%. Accordingly, the bonding strength between the first member and the anchor portion in the insulating layer can be sufficiently maintained, and the amount of displacement of the sensor portion when stress is applied to the anchor portion from the first member side and the second member side. Can be reduced more effectively. In particular, as described above, in the form in which the outer peripheral side surface of the insulating layer is inclined so that the width dimension gradually decreases from the first member side toward the anchor portion side, T1 / T2 is set to 3.3. By controlling (T4 / T3) × 100 (%) to 10% or more, the displacement amount of the sensor unit can be reduced more effectively.

  According to the present invention, the bonding strength between the first member and the anchor portion in the insulating layer can be sufficiently maintained, and when the stress is applied to the anchor portion from the first member side and the second member side, the sensor portion Can be reduced, and a MEMS sensor with excellent detection accuracy can be obtained.

The partial expanded longitudinal cross-sectional view which showed typically the MEMS sensor in 1st Embodiment, The partial expanded longitudinal cross-sectional view which showed typically the MEMS sensor in 2nd Embodiment, The partial expanded longitudinal cross-sectional view which showed typically the MEMS sensor in 3rd Embodiment, A plan view of an anchor portion, an insulating layer, and a metal layer in the present embodiment, (A) is a plan view of a MEMS sensor (acceleration sensor) using any one of the structures of FIGS. 1 to 3, and (b) is a MEMS sensor cut from the BB line of (a) and viewed from the arrow direction. A longitudinal sectional view of FIG. 5 is a plan view of a MEMS sensor (acceleration sensor) different from FIG. 5 using any one of the structures in FIGS. FIG. 6 is a perspective view (when movable) of the MEMS sensor shown in FIG. Process drawing (partially enlarged longitudinal sectional view) for explaining a method of manufacturing an insulating layer in which an outer peripheral side surface is formed as an inclined surface in the present embodiment, (A) is a longitudinal cross-sectional view of an experimental sample (Example 1), (b) is a longitudinal cross-sectional view showing a state in which the detection unit is displaced by applying stress from the reference state of (a), A graph showing the relationship between T1 / T2 and displacement L; A graph showing the relationship between (T4 / T3) × 100 (%) and the displacement L; Unlike FIG. 9, the longitudinal cross-sectional view of the experimental sample (Example 2, Example 3) which inclined the outer peripheral side surface of the insulating layer, A graph showing the relationship between T1 / T2 and displacement amount L in Example 1, Example 2 and Example 3, A graph showing the relationship between (T4 / T3) × 100 (%) and displacement amount L in Example 1, Example 2 and Example 3, A TEM photograph showing the shape of the insulating layer interposed between the first member and the anchor part, The partial expanded longitudinal cross-sectional view of the conventional MEMS sensor.

  1 to 3 are partially enlarged cross-sectional views schematically showing the MEMS sensor according to the present embodiment.

  A MEMS sensor 10 according to the first embodiment shown in FIG. 1 includes a first member (support base material) 11, a second member (cap base material) 12, and the first member 11 and the second member 12. It is comprised by the laminated structure with the intermediate member (functional layer) 13 to do.

  The first member 11, the second member 12, and the intermediate member 13 are all formed of silicon. The first member 11 and the second member 12 are formed in a flat plate shape, for example.

  As shown in FIG. 1, the intermediate member 13 includes an anchor portion 14 and a sensor portion 15 connected to the anchor portion 14. As shown in FIG. 1, the sensor unit 15 may be configured to be directly connected to the anchor unit 14, or a spring unit or the like formed with a small thickness is interposed between the sensor unit 15 and the anchor unit 14. Form may be sufficient.

  As shown in FIG. 1, the sensor unit 15 connected to the anchor unit 14 constitutes one electrode in, for example, a capacitance type.

  As shown in FIG. 1, an insulating layer 16 is interposed between the first member 11 and the anchor portion 14. The insulating layer 16 is, for example, silicon oxide. The first member 11, the insulating layer 16, and the intermediate member 13 can constitute an SOI substrate.

  As shown in FIG. 1, an insulating layer 18 is formed on the inner surface of the second member 12, and a metal layer 17 is formed between the insulating layer 18 and the anchor portion 14.

  The metal layer 17 is formed by, for example, eutectic bonding or diffusion bonding of a first metal layer formed on the anchor portion 14 side and a second metal layer formed on the second member 12 side, and a combination of materials. Examples thereof include aluminum-germanium, aluminum-zinc, gold-silicon, gold-indium, gold-germanium, and gold-tin.

  Further, a wiring portion (not shown) is formed in the insulating layer 18 or on the surface from the connection position with the metal layer 17, and is an electrode electrically connected to the wiring portion routed to the outer frame position of the MEMS sensor 10. The pad is exposed. A detection signal can be extracted from this electrode pad.

  As shown in FIG. 1, the insulating layer 16 and the metal layer 17 are not formed at positions facing the sensor unit 15, and are between the sensor unit 15 and the first member 11 and between the sensor unit 15 and the second member 12. There is a space.

  As shown in FIG. 1 and FIG. 4 (plan view of the anchor portion, the insulating layer 16 and the metal layer 17, and the shapes of the insulating layer 16 and the metal layer 17 are indicated by dotted lines), the outer peripheral side surface 16 a of the insulating layer 16, And both the outer peripheral side surfaces 17a of the metal layer 17 are set back inside the outer peripheral side surface 14a of the anchor portion 14.

  As shown in FIGS. 1 and 4, the width dimension of the insulating layer 16 is T1, the width dimension of the metal layer 17 is T2, and the width dimension of the anchor portion 14 is T3. And the retraction amount of the outer peripheral side surface 16a of the said insulating layer 16 seen from the outer peripheral side surface 14a of the anchor part 14 is T4. Here, in FIG. 4, the retraction amount T4 of each outer peripheral side surface 16a is substantially constant from each outer peripheral side surface 14a of the anchor portion 14, but when the retraction amount T4 varies depending on the location, for example, the maximum retraction amount, The reverse amount T4 can be indicated by an average value with the minimum reverse amount.

  As shown in FIG. 4, the planar shape of the anchor portion 14 and the insulating layer 16 is a rectangular shape, and in particular, a square shape is shown in FIG. 4. The planar shape of the metal layer 17 is a circular shape. The planar shapes of the anchor portion 14, the insulating layer 16, and the metal layer 17 are not limited to those shown in FIG.

  Further, the width dimensions T1 to T3 are defined as follows, for example. When formed in a square, the width dimension is defined by the length of one side. In the case of rectangles having different side lengths, the width dimension is defined by the average of the short side and the long side. Alternatively, when formed in a circle, the width dimension is defined by the diameter. When formed in an elliptical shape, the width dimension is defined by the average of the major axis and the minor axis. When formed in a shape other than these, for example, it is defined by an average value of the length dimension in the X1-X2 direction shown in FIG. 4 and the length dimension in the Y1-Y2 direction orthogonal to the X1-X2 direction. The X1-X2 direction and the Y1-Y2 direction follow each side direction if the outer shape of the MEMS sensor 10 is rectangular. Alternatively, the width dimensions T1 to T3 can be defined by a cross-sectional shape that appears when the sensor unit 15 connected to the anchor unit 14 is cut in the extending direction as illustrated in FIG. Alternatively, the width dimensions T1 to T3 can be defined by the average of the major axis and the minor axis.

  In the present embodiment, the ratio (T1 / T2) of the width dimension T1 of the insulating layer 16 to the width dimension T2 of the metal layer 17 is defined within the range of 1.5 or more and 3.3 or less.

  In this embodiment, as shown in FIG. 1, the outer peripheral side surface 17a of the metal layer 17 is retreated from the outer peripheral side surface 14a of the anchor portion 14, and the width dimension T2 of the metal layer 17 is formed with a sufficiently small size. Thereby, even if there is a large difference in thermal expansion between the metal layer 17 and the anchor portion 14 and between the metal layer 17 and the second member 12, distortion and deformation based on thermal stress can be suppressed.

  Furthermore, in the present embodiment, the ratio (T1 / T2) is set to 3.3 or less as described above. Accordingly, the distribution of stress transmitted from the first member 11 side to the anchor portion 14 when stress is applied between the first member 11 and the second member 12, such as thermal strain and bending stress between the members due to the manufacturing process and use environment. And the difference (stress deviation) between the distribution of stress transmitted from the second member 12 side to the anchor portion 14 can be reduced. Therefore, even when stress is applied to the anchor portion 14, the deformation of the anchor portion 14 can be suppressed, so that the displacement amount of the sensor portion 15 connected to the anchor portion 14 can be reduced as compared with the conventional case. Therefore, the MEMS sensor 10 having excellent detection accuracy as compared with the prior art can be obtained. The amount of displacement is defined by the maximum deflection amount of the sensor unit 15 from the reference state when the stress applied to the anchor portion 14 is assumed to be zero when the stress applied to the anchor portion 14 is zero. it can.

  In the present embodiment, the ratio (T1 / T2) is set to 1.5 or more, so that the bonding strength between the first member 11 and the anchor portion 14 bonded through the insulating layer 16 can be sufficiently maintained. I can do it. Further, since the insulating layer 16 is formed by removing unnecessary insulating layers around the insulating layer 16 by etching, for example, the shape and forming position of the insulating layer 16 are the anchor portion 14 and the sensor portion 15 connected to the anchor portion 14. It tends to vary depending on the size, shape, formation location, etching time, etching environment, and the like. Therefore, if the width T1 of the insulating layer 16 is set too small, the entire region of the metal layer 17 is not included in the region of the insulating layer 16 in a plan view shown in FIG. The insulating layer 16 is likely to be in a position shifted from each other. Then, the stress balance applied to the anchor portion 14 from the first member 11 side and the second member 12 side is deteriorated. As shown in the experiment described later, the displacement amount of the sensor unit 15 can be reduced by reducing the ratio (T1 / T2). However, if T1 is too small, the displacement amount of the sensor unit 15 cannot be effectively reduced. . Therefore, in plan view, the lower limit value of T1 / T2 is set to 1.5 in consideration of the manufacturing margin so that the entire region of the metal layer 17 is included in the region of the insulating layer 16.

  Further, by setting the ratio (T1 / T2) to 1.5 or more, eutectic bonding or diffusion bonding can be appropriately performed. That is, as the ratio (T1 / T2) increases, the width dimension difference between the width dimension T1 of the insulating layer 16 and the width dimension T2 of the metal layer 17 increases accordingly. When the pressure is applied when the metal layer 17 is eutectic bonded or diffusion bonded, if the width dimension difference is large, the pressure is sufficiently applied to the metal layer 17 even if the alignment between the members is somewhat shifted. Can be added, and stable bonding can be performed. Therefore, good electrical connection can be obtained, and an electrically stable joint structure can be obtained.

  In the second embodiment shown in FIG. 2 and the third embodiment shown in FIG. 3, the outer peripheral side surface 16a of the insulating layer 16 is inclined. 2 and 3, the ratio (T1 / T2) is defined as 1.5 or more and 3.3 or less, as in FIG.

  In the second embodiment shown in FIG. 2, the width dimension T <b> 1 of the insulating layer 16 is defined on the bonding side with the first member 11. The outer peripheral side surface 16 a of the insulating layer 16 is inclined so that the width dimension of the insulating layer 16 gradually decreases from the first member 11 side toward the anchor portion 14.

  In the third embodiment shown in FIG. 3, the width dimension T <b> 1 of the insulating layer 16 is defined on the joint side with the anchor portion 14. The outer peripheral side surface 16 a of the insulating layer 16 is inclined so that the width dimension of the insulating layer 16 gradually decreases from the anchor portion 14 side toward the first member 11.

  As shown in FIGS. 2 and 3, by forming the outer peripheral side surface 16a of the insulating layer 16 with an inclined surface, the stress applied to the anchor portion 14 from the first member 11 side through the insulating layer 16 is effectively dispersed. It is possible to reduce the amount of displacement of the sensor unit 15 more effectively, and it is known from experiments described later.

  Moreover, it is preferable that the ratio ((T4 / T3) × 100 (%)) of the retraction amount T4 of the insulating layer 16 to the width dimension T3 of the anchor portion 14 is 10% or more and 25% or less. In particular, in the embodiment shown in FIG. 2, the ratio (T1 / T2) is set to 3.3 or lower and the ratio ((T4 / T3) × 100 (%)) is set to 10% or higher. It is known from the experiment described later that the displacement amount of the sensor unit 15 can be reduced more effectively than the embodiment. The upper limit of the ratio ((T4 / T3) × 100 (%)) was set to 25% in consideration of the bonding strength between the first member 11 and the anchor portion 14.

  Specific numerical values will be exemplified. The width dimension T1 of the insulating layer 16 is about 15 to 75 μm. The width dimension T2 of the metal layer 17 is about 10 to 40 μm. The width T3 of the anchor portion 14 is about 30 to 80 μm. Further, the retraction amount T4 is about 1 to 10 μm.

  The structure of this embodiment shown in FIGS. 1 to 3 is applied to the MEMS sensor shown in FIGS.

  FIG. 5A is a plan view of the MEMS sensor (acceleration sensor), and FIG. 5B is a longitudinal sectional view of the MEMS sensor taken along the line BB in FIG. 5A and viewed from the arrow direction. . The plan view of FIG. 5A shows the first member 11 seen through.

  The MEMS sensor shown in FIG. 5 has, for example, a rectangular shape with long sides in the X direction and short sides in the Y direction.

  As shown in FIG. 5A, the intermediate member 13 is formed with a frame layer 21 in the peripheral region, and the inside of the frame layer 21 is a sensor region forming region. In FIG. 5A, the frame body layer 21 is indicated by oblique lines.

  As shown in FIG. 5A, the intermediate member 13 is formed with a first hole 26, a second hole 27, and a third hole 28 that define the outer shape of the sensor portion inside the frame body layer 21. Each hole 26, 27, 28 penetrates the frame layer 21 in the thickness direction.

A frame-shaped insulating layer 16 is formed at the facing portion between the first member 11 and the frame body layer 21.
As shown in FIGS. 5A and 5B, the periphery of the sensor unit is sealed by the metal layer 17 between the frame layer 21 and the insulating layer 18 formed on the opposing surface of the second member 12. Yes.

  As shown in FIG. 5A, the insides of the holes 26, 27, and 28 are sensor portions 36, 37, and 38, respectively.

  5 (a) and 5 (b), a plurality of anchor portions 14 are formed in the holes 26, 27, and 28, and are connected to the anchor portions 14 to be connected to the sensor portions 36, 37, 38 is formed.

  For example, the MEMS sensor shown in FIG. 5 can detect the acceleration in the Z direction in the direction orthogonal to the substrate surface of the first member 11 by the operation of the first movable unit 41 provided in the first sensor unit 36. The movable part 41 is connected to each anchor part 14 and 14 facing in the Y1-Y2 direction via a spring part or a beam part. Further, fixed portions 51 and 53 are formed to extend from the respective anchor portions 14 and 14 facing in the X1-X2 direction. The portion where the movable portion 41 and the fixed portions 51 and 53 are opposed has a comb-like electrode structure, and the capacitance changes when the movable portion 41 moves in the Z direction, so that the movable portion 41 moves. The amount and acceleration can be detected. Further, the acceleration of the first member 11 in the Y direction parallel to the substrate surface of the first member 11 can be detected by the operation of the second movable portion 42 provided in the second sensor portion 37, and provided in the third sensor portion 38. By the operation of the third movable portion 43, the acceleration in the X direction orthogonal to the Z direction and the Y direction can be detected.

  The MEMS sensor shown in FIGS. 6 and 7 has a structure in which the first movable part 60 and the second movable part 61 are supported via the anchor part 14 so as to move in the height direction and in opposite directions. It is.

  6 and 7, only the intermediate member 13 is illustrated, and the first member 11, the second member 12, the insulating layer 16, and the metal layer 17 located above and below the illustration are omitted, but are described with reference to FIGS. 1 to 3. 6 and 7 are joined to the first member 11 and the second member 12 through the insulating layer 16 and the metal layer 17.

  FIG. 8 shows a manufacturing method when the outer peripheral side surface 16a of the insulating layer 16 is formed as an inclined surface as shown in FIGS.

8A, the surface 11a of the first member 11 made of silicon is thermally oxidized to form a first insulating layer (SiO ₂ ) 71 on the surface 11a of the first member 11. On the other hand, the surface 13 a of the intermediate member 13 made of silicon is naturally oxidized to form a second insulating layer (SiOx) 72 on the surface 13 a of the intermediate member 13. As a result, a first insulating layer 71 that is a dense film is formed on the first member 11, and a second insulating layer 72 that is a sparse film is formed on the intermediate member 13.

  Next, in the step of FIG. 8B, the first member 11 and the intermediate member 13 are joined. For example, the first member 11 and the intermediate member 13 can be fusion bonded or room temperature bonded.

  Next, in the process of FIG. 8C, the outer shapes of the first member 11 and the intermediate member 13 are adjusted by an etching or polishing process. As shown in FIG. 8C, the intermediate member 13 is formed with an anchor portion 14, a sensor portion 15 and the like.

  Next, in the step of FIG. 8D, the insulating layer 16 is interposed between the anchor portion 14 and the first member 11 between the first member 11 and the intermediate member 13, and the other region becomes a space. As described above, the first insulating layer 71 and the second insulating layer 72 are removed by isotropic etching using wet etching or dry etching. At this time, the sensor unit 15 is provided with a number of fine holes so that the first insulating layer 71 and the second insulating layer 72 interposed between the sensor unit 15 and the first member 11 are easily etched. The gas can reach the first insulating layer 71 and the second insulating layer 72 between the sensor unit 15 and the first member 11.

  For example, the first insulating layer 71 and the second insulating layer 72 are selectively etched using HF gas. At this time, since the etching rate of the first insulating layer 71 which is a dense film is slower than the etching rate of the second insulating layer 72 which is a sparse film, the first insulating layer 71 and the second insulating layer 72 are equal to each other. As shown in FIG. 8D, the outer peripheral side surface 16a of the insulating layer 16 is formed on an inclined surface. In FIG. 8D, the first insulating layer 71 which is a dense film is provided on the first member 11 side, and the second insulating layer 72 which is a sparse film is provided on the anchor portion 14 side. As shown in d), the outer peripheral side surface 16a is an inclined surface in which the width dimension T1 of the insulating layer 16 gradually decreases from the first member 11 side to the anchor portion 14. Therefore, if the first insulating layer 71 is formed of a sparse film and the second insulating layer 72 is formed of a dense film, the inclined direction of the outer peripheral side surface 16a can be formed opposite to that shown in FIG.

  Note that the MEMS sensor in this embodiment is preferably applied to a physical quantity sensor such as an acceleration sensor, a gyro sensor, or an impact sensor. Also, the sensor unit detection principle is not limited to the capacitance type.

  FIG. 9A is a longitudinal sectional view of an experimental sample of the MEMS sensor in the simulation experiment.

  As shown in FIG. 9A, an intermediate member 83 is interposed between the first member 81 and the second member 82, and the first member 81, the second member 82, and the intermediate member 83 are all formed of silicon. . As shown in FIG. 9A, the intermediate member 83 includes an anchor portion 83a and a displacement detection portion 83b that extends integrally from the anchor portion 83a in one direction.

Further, as shown in FIG. 9A, the anchor portion 83 a and the first member 81 are joined by interposing an insulating layer 84 made of SiO ₂ between the anchor portion 83 a and the first member 81.

Further, as shown in FIG. 9A, an insulating layer 86 made of SiO ₂ is provided on the inner surface of the second member 82, and a metal layer 85 made of Al is interposed between the insulating layer 86 and the anchor portion 83a. The anchor portion 83a and the second member 82 are joined.

Here, the Young's modulus of silicon is 178.00 GPa, the Poisson's ratio is 0.170, the Young's modulus of Al is 70.30 GPa, the Poisson's ratio is 0.350, the Young's modulus of SiO ₂ is 71.00 GPa, and the Poisson's ratio is 0.00. 300.

  The thickness dimension H1 of the first member 81 was 100 μm, the thickness dimension H2 of the second member 82 was 100 μm, and the thickness dimension H3 of the intermediate member 83 was 20 μm. In addition, the thickness dimension H4 of the insulating layer 84 was 1.5 μm, and the thickness dimension H5 of the metal layer 85 was 0.5 μm.

  In the experiment, a plurality of experimental samples obtained by fixing the width dimension T2 (diameter) of the metal layer 85 to 20 μm and changing the width dimension T1 (length of one side) of the insulating layer 84 within a range of 55 to 70 μm. Prepared. In each experimental sample, the center of the metal layer 85 and the center of the insulating layer 84 are aligned in the thickness direction. Then, a pressure of 10 MPa was applied to each experimental sample from the first member 81 side toward the anchor portion 83a and from the second member 82 side toward the anchor portion 83a.

  And the displacement amount L when the displacement detection part 83b displaced as shown in FIG.9 (b) by applying a pressure as mentioned above was measured. In FIG. 9B, the displacement amount of the displacement detector 83b is shown larger than the dimensional ratio obtained from the experiment in order to make it visually easy to understand. The length dimension of the displacement detector 83b extending in one direction was set to 280 μm. Further, the displacement amount L is defined by the maximum deflection amount of the displacement detector 83b when the stress is applied as shown in FIG. 9B from the reference state where the stress is not applied as shown in FIG. 9A.

  FIG. 10 is a graph showing the relationship between T1 / T2 and the displacement L. As shown in FIG. 10, the displacement amount L gradually increases as T1 / T2 increases, but it has been found that the gradient of the displacement amount L begins to become gentle when T1 / T2 approaches 3.3.

  Therefore, in this embodiment, T1 / T2 is set to 3.3 or less. Preferably, T1 / T2 is set to 3.00 or less. As shown in FIG. 10, the smaller T1 / T2, the smaller the displacement amount L, which is preferable. However, as T1 / T2 decreases, the width dimension T1 of the insulating layer 84 decreases, and the bonding strength between the first member 81 and the anchor portion 83a decreases. Further, since the insulating layer 84 is formed by removing unnecessary insulating layers around the insulating layer 84, the shape and position of the insulating layer 84 are the size of the anchor portion 83a and the sensor portion connected to the anchor portion 83a. It tends to vary depending on the sheath shape, formation location, etching time, etching environment, and the like. Therefore, if the width dimension T1 of the insulating layer 84 is too small, the entire region of the metal layer 85 is not included in the region of the insulating layer 84 in plan view, and the metal layer 85 and the insulating layer 84 are displaced from each other. Easy to place. Then, the stress balance applied from the upper and lower surfaces of the anchor portion 83a is deteriorated, and the displacement L cannot be effectively reduced. Therefore, the lower limit value of T1 / T2 was set to 1.5.

  Further, by setting the ratio (T1 / T2) to 1.5 or more, eutectic bonding or diffusion bonding can be appropriately performed. That is, as the ratio (T1 / T2) increases, the width dimension difference between the width dimension T1 of the insulating layer 16 and the width dimension T2 of the metal layer 17 increases accordingly. When the pressure is applied when the metal layer 17 is eutectic bonded or diffusion bonded, if the width dimension difference is large, the pressure is sufficiently applied to the metal layer 17 even if the alignment between the members is somewhat shifted. Can be added, and stable bonding can be performed. Therefore, good electrical connection can be obtained, and an electrically stable joint structure can be obtained.

  Next, the width dimension T3 of the anchor portion 83a is unified to 20 μm, and the receding amount T4 is changed so that the ratio of the receding amount T4 of the outer peripheral side surface 84a of the insulating layer 84 to the width dimension T3 of the anchor portion 83a ((T4 / T3 ) × 100 (%)) and the displacement amount L of the displacement detector 83b were measured. The experimental results are shown in FIG. As shown in FIG. 11, it was found that the displacement L can be effectively reduced when the ratio ((T4 / T3) × 100 (%)) is 10% or more. The ratio ((T4 / T3) × 100 (%)) is more preferably 15% or more.

  Subsequently, as shown in FIGS. 12A and 12B, the outer peripheral side surface 84a of the insulating layer 84 interposed between the first member 81 and the anchor portion 83a was formed on an inclined surface. In FIG. 12A, the outer peripheral side surface 84a is inclined so that the width dimension of the insulating layer 84 gradually increases from the first member 81 side toward the anchor portion 83a. In FIG. The outer peripheral side surface 84a was inclined so that the width dimension of the insulating layer 84 gradually decreased from the side toward the anchor portion 83a. In both FIGS. 12A and 12B, the width dimension T1 of the insulating layer 84, which is compared with the width dimension T2 of the metal layer 85, is defined on the wide side. In the experiment, the inclination dimension (the amount of cutting) T6 of the outer peripheral side surface 84a of the insulating layer 84 was unified to 4 μm. Other dimensions were the same as those in FIG.

  FIG. 13 shows the ratio (T1 / T2) and displacement amount L in Example 1 shown in FIG. 9A, Example 2 shown in FIG. 12A, and Example 3 shown in FIG. It is a graph which shows a relationship.

  As shown in FIG. 13, it was found that the displacement amount L can be effectively reduced when the ratio (T1 / T2) is set to 3.3 or less in any of the examples. In addition, it was found that the displacement amount L can be effectively reduced even in the same ratio (T1 / T2) in Example 2 and Example 3 in which the outer peripheral side surface 84a of the insulating layer 84 is inclined as compared with Example 1. Thus, the amount of displacement L can be reduced by providing an inclination to the insulating layer 84 because the insulating layer 84 is further shaved and the stress applied to the anchor portion 83a by the inclination is more effective. It is thought that this is because it can be dispersed.

  FIG. 14 is a graph showing the relationship between the ratio ((T4 / T3) × 100 (%)) and the displacement L in Example 1, Example 2, and Example 3.

  As shown in FIG. 14, it was found that the displacement amount L can be effectively reduced by setting the ratio ((T4 / T3) × 100 (%)) to 10% or more in any of the examples.

  Further, as shown in FIGS. 13 and 14, in Example 2 and Example 3 in which the outer peripheral side surface 84a of the insulating layer 84 is an inclined surface, the width dimension T1 is defined on the first member 81 side, and the width dimension is determined. It has been found that the displacement amount L can be reduced more effectively in Example 3 including the insulating layer 84 that is gradually reduced from the first member 81 side toward the anchor portion 83a. Therefore, in the embodiment 3, if the ratio (T1 / T2) is set to 3.3 or less and the ratio ((T4 / T3) × 100 (%)) is set to 10% or more, it is more effective. It was found that the displacement amount L can be reduced.

FIG. 15A shows an insulating layer (SiO ₂ ), which is a dense film formed by thermal oxidation (thermal oxidation at about 1000 ° C.) on the inner surface of the first member, and is loosened by natural oxidation on the inner surface of the intermediate member. forming a film in which an insulating layer (SiO _X), a cross-sectional photograph of the remaining insulating layer when removing the unnecessary insulating layer using HF gas. FIG. 15B differs from FIG. 15A in that an insulating layer formed by thermal oxidation is interposed between the first member and the intermediate member without forming a sparse film, and an unnecessary insulating layer is formed using HF gas. It is a cross-sectional photograph of the insulating layer left when removed. FIG. 15C is the reverse of the film configuration of FIG. 15A, that is, an insulating layer that is a sparse film formed by natural oxidation is formed on the inner surface of the first member, and thermal oxidation is performed on the inner surface of the intermediate member. 4 is a cross-sectional photograph of an insulating layer left when an insulating layer which is a dense film is formed and an unnecessary insulating layer is removed using HF gas.

  The outer peripheral side surface of the insulating layer in FIG. 15B is a substantially vertical surface, but in FIGS. 15A and 15C, the outer peripheral side surface of the insulating layer is an inclined surface. As described above, it was found that the outer peripheral side surface of the insulating layer can be formed on the inclined surface easily and appropriately by using the etching rate difference between the sparse film and the dense film.

10 MEMS sensor 11, 81 First member 12, 82 Second member 13, 83 Intermediate member 14, 83a Anchor portion 14a (anchor portion) outer peripheral side surface 15, 36, 37, 38 Sensor portion 16, 18, 84, 86 Insulation Layer 16a (Insulating layer) outer peripheral side surface 17, 85 Metal layer 17a (Metal layer) outer peripheral side surface 21 Frame layer 71 First insulating layer 72 Second insulating layer 83b Displacement detector

Claims (4)

A first member, a second member, an intermediate member positioned between the first member and the second member, an anchor portion formed on the intermediate member, and a sensor portion connected to the anchor portion; An insulating layer interposed between the first member and the anchor portion, and a metal layer interposed between the second member and the anchor portion,
The outer peripheral side surface of the metal layer is receding inward from the outer peripheral side surface of the anchor part,
The MEMS sensor according to claim 1, wherein T1 / T2 is in a range of 1.5 to 3.3, where T1 is a width dimension of the insulating layer and T2 is a width dimension of the metal layer.   The MEMS sensor according to claim 1, wherein the outer peripheral side surface of the insulating layer is formed with an inclined surface in which the width dimension of the insulating layer gradually decreases from T <b> 1 in the thickness direction.   The width dimension T1 of the insulating layer is defined on the bonding side with the first member, and the width dimension of the outer peripheral side surface of the insulating layer gradually decreases from the first member side toward the anchor portion side. The MEMS sensor according to claim 2, which is inclined as follows. (T4 / T3) × 100 (%) is 10% or more and 25%, where T3 is the width dimension of the anchor portion and T4 is the receding amount of the outer peripheral side surface of the insulating layer from the outer peripheral side surface of the anchor portion. The MEMS sensor according to any one of claims 1 to 3, which is defined within the following range.

Images