US20090049911A1 - Integrated micro electro-mechanical system and manufacturing method thereof - Google Patents

Abstract

In the manufacturing technology of an integrated MEMS in which a semiconductor integrated circuit (CMOS or the like) and a micro machine are monolithically integrated on a semiconductor substrate, a technology capable of manufacturing the integrated MEMS without using a special process different from the normal manufacturing technology of a semiconductor integrated circuit is provided. A MEMS structure is formed together with an integrated circuit by using the CMOS integrated circuit process. For example, when forming an acceleration sensor, a structure composed of a movable mass, an elastic beam and a fixed beam is formed by using the CMOS interconnect technology. Thereafter, an interlayer dielectric and the like are etched by using the CMOS process to form a cavity. Then, fine holes used in the etching are sealed with a dielectric.

Description

CROSS REFERENCE TO RELATED APPLICATIONS
• This application is a Continuation of nonprovisional U.S. application Ser. No. 12/216,359 filed Jul. 2, 2008, which is a Continuation of nonprovisional U.S. application Ser. No. 11/208,740 filed Aug. 23, 2005. Priority is claimed based on U.S. application Ser. No. 12/216,359 filed Jul. 2, 2008, which claims the priority of U.S. application Ser. No. 11/208,740 filed Aug. 23, 2005, which claims the priority of Japanese Patent Application JP 2005-050541 filed on Feb. 25, 2005 and Japanese Patent Application JP 2005-226233 filed on Aug. 4, 2005, the contents of which are hereby incorporated by reference into this application.
TECHNICAL FIELD OF THE INVENTION
• The present invention relates to micro electro-mechanical systems (MEMS) and a manufacturing method thereof. More particularly, the present invention relates to a technology effectively applied to an integrated micro electro-mechanical system in which a semiconductor integrated circuit and the MEMS are integrated.
BACKGROUND OF THE INVENTION
• By using the micro-fabrication technology which has realized the improvement in performance and increase in integration density of the semiconductor integrated circuit, the micro electro-mechanical system (hereinafter, referred to as MEMS) technology for forming a mechanical sensor such as a pressure sensor and a acceleration sensor, a miniatured switch, miniature mechanical parts such as a resonator or an oscillator, and a mechanical system has been developed. The MEMS is broadly classified into a bulk MEMS which is formed by processing a silicon substrate itself and a surface MEMS which is formed by repeatedly depositing and patterning thin films on a surface of a silicon substrate. The manufacturing process of the surface MEMS is close to the manufacturing process of semiconductor integrated circuits. The MEMS technologies described above are discussed in pp. 1158 to 1165 of volume 73 (issue 9) of “Applied Physics” (Japan Society of Applied Physics (JSAP), September, 2004).
• The U.S. Pat. No. 6,635,506 discloses the technology for precisely forming a shape of a cavity by using a sacrificial layer when forming the cavity used for the MEMS.
• The U.S. Pat. No. 6,667,245 discloses an example in which a switch is formed as the MEMS. Also, it mentions that it is possible to form the switch at low cost by using the standard technology for forming a multilayer interconnect of an integrated circuit to form the switch.
• The US Patent Application Publication No. 2004/0145056 describes that the MEMS is formed by using a metal layer and a sacrificial layer formed by the standard manufacturing technology of CMOSFET (Complementary Metal Oxide Semiconductor Field Effect Transistor, referred to as CMOS, hereinafter).
• The U.S. Pat. No. 5,596,219 discloses the technology for monolithically forming an integrated circuit and a sensor (actuator). This sensor is formed on a sensor layer made of polysilicon by using the surface micro-machine technology.
SUMMARY OF THE INVENTION
• For example, in the application of MEMS to a sensor, the mechanical deformation of a structure due to the external force or the like is converted into the electrical signal as the piezoresistance change and the capacitance change and then outputted. Further, in general, the above-mentioned output is signal-processed by a semiconductor integrated circuit (LSI: Large Scale Integration, LSI includes CMOS.). Also, in the application of MEMS to an oscillator, the input and output of the oscillators are connected to the RF-IC (Radio-Frequency integrated circuits). As described above, the MEMS is used in combination with the LSI in many cases. In addition, since the operation of the MEMS is not limited to the mechanical operation, the conversion of the different physical values (electrical signal and mechanical deformation in many cases) is required in most of the applications thereof. This conversion mechanism is called a transducer.
• As described above, when the MEMS is used in combination with the signal processing LSI, since each of them is formed on the separate chips, the miniaturization of the entire system is difficult. Since both the MEMS and the LSI are usually formed on a silicon substrate, the direction toward the monolithic integration of them on the same substrate is natural, and it has already been applied in some products.
• For example, an acceleration sensor or a vibration gyroscope using a mass composed of a poly-crystalline silicon film with a thickness of about 2 to 4 μm is integrated with an analog circuit such as a capacitance-voltage converter or an operation amplifier circuit. A sensor mechanical part (arranged on a silicon substrate with a gap) and an analog circuit part are arranged in the different (adjacent) regions on the substrate plane. The sensor mechanical part is entirely covered and sealed with a cover.
• Also, a digital mirror device, in which movable metal films each having a reflection surface are arranged in a matrix form, and the directions of each film are electrostatically controlled to turn on/off the light, thereby realizing a display device, has been produced. The upper part of the device is sealed with a transparent plate which transmits light.
• Furthermore, a technology for forming a RF-MEMS (switch, filter) on a LSI by the so-called copper-damascene interconnect process has been reported. In this technology, both the movable part and the cavity part are formed by the damascene process. In addition, the sealing method after forming the movable part is also described. Further, the method of forming a ME mechanical part and a transducer (signal unit) by using the multilayer interconnect of a LSI has been reported.
• All of the cases described above belong to the category of the surface MEMS. However, the technology for integrating the bulk MEMS with a LSI has been also reported. In the bulk MEMS, since the silicon wafer itself becomes a movable part, the bonding with other wafer (substrate) is required for the sealing and the packaging.
• The first problem to be solved is that the special cavity formation process and the special sealing process are necessary in the conventional MEMS. More specifically, in the manufacture of the MEMS, a special technology is required in addition to the normal manufacturing technology of CMOS. In particular, in the case of an oscillator, the vacuum sealing is necessary to obtain the large vibration value Q, and the air tightness is important for maintaining the properties for a long time. This is true of the case of the integrated MEMS in which the LSI is mounted together with a MEMS. Also in the example in which the multilayer interconnect is used to form the mechanical part, except for the cases of special application, it is preferable to seal the entire thereof. In this case, it is necessary to bond the substrate on which the MEMS is formed and another substrate to be a lid. Also, in the example in which the cavity is fabricated through the damascene process, a special process such as the process of embedding a sacrificial layer into an interlayer dielectric is required. As described above, the special technology in addition to the normal CMOS manufacturing technology is required in the conventional MEMS manufacturing process.
• The second problem is that a film much thicker than a film used in the standard LSI is necessary in order to form a mass with a sufficient mass required in an acceleration sensor and a gyroscope. This is particularly difficult in the case of the integrated MEMS in which a LSI is mounted together with a MEMS for the following reasons. First, it is difficult to form a thick film with a controlled mechanical stress in the standard LSI process. Even if possible, it is difficult to combine it with the fine CMOS process from the viewpoint of the temperature condition of the thermal treatment. Also, in the case where the SOI (Silicon on Insulator) is used, a special process such as the deep trench etching is necessary, and thus, the process is complicated and the cost is increased.
• The third problem is that the manufacturing process of the integrated MEMS in which a LSI is mounted together with a MEMS is complicated, and the chip area is increased.
• An object of the present invention is to provide a technology capable of manufacturing an integrated MEMS without using a special process different from the normal manufacturing technology of a semiconductor integrated circuit, in the manufacturing technology of an integrated MEMS in which a semiconductor integrated circuit (CMOS and others) and a micro machine are monolithically integrated on a semiconductor substrate.
• Also, another object of the present invention is to provide a technology capable of easily manufacturing a mass with a sufficient mass required in an acceleration sensor and a gyroscope at low cost, in the manufacturing technology of an integrated MEMS including an acceleration sensor or a gyroscope.
• Also, still another object of the present invention is to provide a technology capable of simplifying a manufacturing process of an integrated MEMS to reduce the manufacturing cost of a product.
• Further, still another object of the present invention is to provide a technology capable of miniaturizing an integrated MEMS.
• The above and other objects and novel characteristics of the present invention will be apparent from the description of this specification and the accompanying drawings.
• The typical ones of the inventions disclosed in this application will be briefly described as follows.
• An integrated micro electro-mechanical system according to the present invention is the integrated micro electro-mechanical system, in which a micro machine formed by using a manufacturing technology of a semiconductor integrated circuit and a semiconductor integrated circuit are formed on a semiconductor substrate, and the micro machine comprises: (a) a sealed cavity formed by removing a part of an interlayer dielectric formed between the interconnects; and (b) a mechanical structure formed in the cavity, wherein the cavity is formed by using a technology for forming an interconnect of a MOSFET and is sealed by using the technology for forming an interconnect of a MOSFET.
• Also, a manufacturing method of an integrated micro electro-mechanical system according to the present invention is the manufacturing method of an integrated micro electro-mechanical system, in which a micro machine formed by using a manufacturing technology of a semiconductor integrated circuit and a semiconductor integrated circuit are formed on a semiconductor substrate, and the method comprises the steps of: (a) forming a mechanical structure which is a part of the micro machine; (b) forming a layer covering the mechanical structure; (c) forming a cavity in which the mechanical structure is placed; and (d) sealing the cavity, wherein the step (a), the step (b), the step (c) and the step (d) are performed by using a technology for forming an interconnect of a MOSFET.
• The first primary characteristic of the present invention is that a cavity for placing a mechanical structure of a MEMS (micro machine) is formed by using a standard CMOS manufacturing process or a standard interconnect process which is a part of the CMOS manufacturing process (without using a special sealing process). More specifically, first, a movable part also functioning as an electrode which is a part of the mechanical structure of a MEMS is formed in an interlayer dielectric by using the CMOS process (multilayer interconnect process). Then, after forming a (metal) thin film layer with micro holes therein, the interlayer dielectric around the movable part also functioning as an electrode is etched and removed through the micro holes, and then, the micro holes are finally sealed.
• At this time, the structure of the micro machine is placed in the cavity formed by removing a part of the interlayer dielectric formed in the multilayer interconnects below the thin film layer. As the thin film layer, a material with the sufficiently low etching rate to the etching of the interlayer dielectric (for example, upper interconnect layer) is used.
• After the finish of the etching of the interlayer dielectric, the micro holes for etching formed in the thin film layer are sealed by depositing a thin film (CVD dielectric or the like) having relatively isotropic deposition characteristics on the thin film layer. The thin film formation and the etching to remove the interlayer dielectric are performed within the range of the normal CMOS process. The movable part also functioning as an electrode formed in the cavity is formed so as to contain any one of a metal film, a silicon-germanium film, a silicon nitride film, a silicon oxide film, a single crystal silicon film, a polysilicon film, an amorphous silicon film and a polyimide film.
• In addition, the second primary characteristic of the present invention is that, by forming one integral mechanical structure (including a mass and a movable part considered as a mechanically integral structure) using a plurality of LSI layers or interconnect layers, a movable mass with a sufficient mass required in an acceleration sensor and a gyroscope can be formed through the standard interconnect process of the CMOS process. The movable part of the mechanical structure is preferably formed in a cavity and is fixed to an interlayer dielectric surrounding the cavity by a (elastically) deformable LSI material or metal interconnects. The mechanical structure is designed so that the mechanical characteristics thereof are determined based on the dimensions of the structure itself and do not depend on the shape of the cavity. More specifically, the mechanical structure is provided with (1) a fixed part which is fixed to an interlayer dielectric surrounding a cavity and has an enough size substantially considered to be elastically undeformed, (2) a movable part and (3) an elastically deformable part which connects the fixed part and the movable part. By doing so, the dimensional accuracy of the cavity does not influence the mechanical characteristics of the MEMS. Therefore, the high dimensional accuracy of the cavity is not required in comparison to the case where the mechanical characteristics of the structure depend on the shape of the cavity. Usually, the dimensional accuracy of the structure is defined with the accuracy of the interconnect pattern of a LSI. Since this dimensional accuracy is generally much higher than the process accuracy of the bulk MEMS in the conventional technology, the highly accurate mechanical characteristics can be assured.
• Since the structure is formed by using the interconnect layers, the structure itself has not only the mechanical function as a mass but also the electrical function as an electrode and an interconnect. The actuation and the sensing are performed by the electrostatic force and capacitance between the electrically independent electrode fixed to an interlayer dielectric and a movable part also functioning as an electrode. By using the movable part of the integral structure as a mass, for example, an acceleration sensor and a vibration gyroscope (angle rate sensor) are realized. The mechanical connection between the movable part and the surrounding interlayer dielectric (fixed part and elastically deformable part, for example, beam) and the electrical connection (interconnect, actuator capacitor and detecting capacitor) can be made through separate layers constituting the LSI. By sandwiching the movable part between the multilayer interconnects to restrict the movable range of the movable part, it becomes possible to improve the reliability.
• Also, the present invention is characterized in that a MEMS such as a vibration sensor, an acceleration sensor, a gyroscope, a switch and an oscillator is integrated with a LSI, and the mechanical structure of the MEMS is formed from the same layer as the interconnect layer of a LSI (including pad). Alternatively, it is characterized in that a MEMS is stacked and formed on (an area overlapping with) the interconnect of a LSI.
• The effect obtained by the representative one of the inventions disclosed in this application will be briefly described as follows.
• Since a LSI (including CMOS) and a MEMS can be monolithically integrated through the standard CMOS process (LSI process), it is possible to achieve the miniaturization and the cost reduction of the integrated MEMS.
BRIEF DESCRIPTIONS OF THE DRAWINGS
• FIG. 1 is a schematic diagram showing the manufacturing process of an acceleration sensor according to the first embodiment of the present invention;
• FIG. 2 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 1;
• FIG. 3 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 2;
• FIG. 4 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 3;
• FIG. 5 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 4;
• FIG. 6 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 5;
• FIG. 7 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 6;
• FIG. 8 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 7;
• FIG. 9A toFIG. 9C are plan views showing the major layers constituting the acceleration sensor, respectively;
• FIG. 10 is a block diagram showing the circuit configuration of the capacitance detecting circuit;
• FIG. 11 is a schematic diagram showing the cross-sectional structure when the acceleration sensor and the pressure sensor are simultaneously formed;
• FIG. 12 is a schematic diagram showing the cross-sectional structure of the acceleration sensor according to the second embodiment;
• FIG. 13A andFIG. 13B are plan views showing the major layers constituting the acceleration sensor, respectively;
• FIG. 14 is a schematic diagram showing the cross-sectional structure when the acceleration sensor and the pressure sensor are simultaneously formed;
• FIG. 15A toFIG. 15C are plan views schematically showing the configuration and the basic operation of the MEMS switch according to the third embodiment;
• FIG. 16A toFIG. 16D are schematic diagrams showing a part of the manufacturing process of the MEMS switch, respectively;
• FIG. 17A andFIG. 17B are plan views showing the major layers of the MEMS switch, respectively;
• FIG. 18 is a schematic diagram showing the manufacturing process of an acceleration sensor according to the fourth embodiment;
• FIG. 19 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 18;
• FIG. 20 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 19;
• FIG. 21 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 20;
• FIG. 22 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 21;
• FIG. 23 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 22;
• FIG. 24 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 23;
• FIG. 25 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 24;
• FIG. 26A toFIG. 26C are plan views showing the major layers constituting the acceleration sensor, respectively;
• FIG. 27A andFIG. 27B are cross-sectional views showing the acceleration sensor according to the modification examples of the fourth embodiment, respectively;
• FIG. 28 is a schematic diagram showing the manufacturing process of an acceleration sensor according to the modification example of the fourth embodiment;
• FIG. 29 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 28;
• FIG. 30 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 29;
• FIG. 31 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 30;
• FIG. 32 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 31;
• FIG. 33 is a schematic diagram showing the manufacturing process of the acceleration sensor subsequent toFIG. 32;
• FIG. 34 is a schematic diagram showing the manufacturing process of an angle rate sensor (vibration gyroscope) according to the fifth embodiment;
• FIG. 35 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 34;
• FIG. 36 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 35;
• FIG. 37 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 36;
• FIG. 38 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 37;
• FIG. 39 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 38;
• FIG. 40 is a schematic diagram showing the configuration of the angle rate sensor according to the fifth embodiment;
• FIG. 41A andFIG. 41B are plan views showing the operation of the angle rate sensor according to the fifth embodiment, respectively;
• FIG. 42 is a schematic diagram showing the manufacturing process of an angle rate sensor according to the sixth embodiment;
• FIG. 43 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 42;
• FIG. 44 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 43;
• FIG. 45 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 44;
• FIG. 46 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 45;
• FIG. 47 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 46;
• FIG. 48A andFIG. 48B are plan views showing the major layers constituting the angle rate sensor, respectively;
• FIG. 49A andFIG. 49B are plan views showing the operation of the angle rate sensor according to the sixth embodiment, respectively;
• FIG. 50 is a schematic diagram showing the manufacturing process of an angle rate sensor according to the modification example of the sixth embodiment;
• FIG. 51 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 50;
• FIG. 52 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 51;
• FIG. 53 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 52;
• FIG. 54 is a schematic diagram showing the manufacturing process of the angle rate sensor subsequent toFIG. 53;
• FIG. 55 is a diagram showing the configuration of the gas pressure monitoring system for tire seen from the bottom surface of an automobile;
• FIG. 56 is a block diagram of a tire pressure measurement module;
• FIG. 57 is a block diagram of a in-vehicle unit;
• FIG. 58 is a diagram showing the configuration of an anti-skid device for a vehicle seen from the bottom surface of an automobile;
• FIG. 59 is a block diagram of an anti-skid control circuit;
• FIG. 60 is a block diagram showing an air suspension control unit; and
• FIG. 61 is a side view of the vehicle showing the configuration of an air suspension control unit seen from the side of an automobile.
DESCRIPTIONS OF THE PREFERRED EMBODIMENTS
• In the embodiments described below, the invention will be described in a plurality of sections or embodiments when required as a matter of convenience. However, these sections or embodiments are not irrelevant to each other unless otherwise stated, and the one relates to the entire or a part of the other as a modification example, details, or a supplementary explanation thereof.
• Also, in the embodiments described below, when referring to the number of elements (including number of pieces, values, amount, range, and the like), the number of the elements is not limited to a specific number unless otherwise stated or except the case where the number is apparently limited to a specific number in principle. The number larger or smaller than the specified number is also applicable.
• Further, in the embodiments described below, it goes without saying that the components (including element steps) are not always indispensable unless otherwise stated or except the case where the components are apparently indispensable in principle.
• Similarly, in the embodiments described below, when the shape of the components, positional relation thereof and the like are to be mentioned, the substantially approximate and similar shapes and the like are included therein unless otherwise stated or except the case where it can be conceived that they are apparently excluded in principle. This condition is also applicable to the numerical value and the range described above.
• In addition, components having the same function are denoted by the same reference symbols throughout the drawings for describing the embodiments, and the repetitive description thereof is omitted. Note that hatching is used in some cases even in a plan view so as to make the drawings easy to see.
• Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
• In the embodiments described below, after fabricating transistors of a LSI on a silicon (Si) substrate, simultaneously with forming a multilayer interconnect on the transistors, a MEMS is formed in the interlayer dielectric formed between the multilayer interconnect on the same silicon substrate, and then, a cavity is formed and sealed. Alternatively, after forming a MEMS on a silicon substrate, a LSI is fabricated on the same silicon substrate, and then, a cavity is formed and sealed.
First Embodiment
• In the first embodiment, the case where a single axis acceleration (or vibration) sensor is formed as the MEMS will be described.
• FIG. 1 toFIG. 8 are schematic diagrams (cross-sectional views) showing the manufacturing process of the integrated MEMS according to the first embodiment. First, in accordance with the normal manufacturing process of a CMOS integrated circuit,transistors102 for signal processing of the single axis acceleration sensor andcontact holes103 are formed on a silicon substrate (semiconductor substrate)101 (FIG. 1). Next, by using the manufacturing process of the CMOS integrated circuit, a first layer interconnect (M1 layer)104 of thesignal processing transistors102 and anetching stopper film105 used in the etching for forming a cavity (process described later) are formed (FIG. 2).
• Next, through the normal process for manufacturing a CMOS integrated circuit, a multilayer interconnect comprised of a second layer interconnect (M2 layer) and a third layer interconnect (M3 layer) (second layer interconnect and third layer interconnect are not shown.) is formed, and the surface thereof is planarized by the normal chemical mechanical polishing (CMP). Next, after forming aninterlayer dielectric106, viaholes107 are formed in the interlayer dielectric106 (FIG. 3). Then, according to need, a fourth layer interconnect (M4 layer)108 is formed on theinterlayer dielectric106, and a movable mass (movable part)109 of the single axis acceleration sensor, an elastic beam also functioning as interconnect (elastically deformable part)110 and a fixed beam (fixed part)111 are formed on the interlayer dielectric106 (FIG. 4). More specifically, the mechanical structure constituting a part of the single axis acceleration sensor (movable mass, elastic beam, fixed beam) is formed. As described above, this structure is formed from the same layer as the interconnect (for example, fourth layer interconnect108) constituting the semiconductor integrated circuit.
• Further, aninterlayer dielectric112 is formed so as to cover the structure and theinterlayer dielectric112 is planarized according to need by the CMP or the like (FIG. 5). Thereafter, acavity cover film114 having fine (tiny)holes113 for forming a cavity is formed from a fifth layer interconnect (M5 layer) (FIG. 6). Then, theinterlayer dielectric112 and theinterlayer dielectric106 around themovable mass109 are etched and removed through thefine holes113 to form a cavity115 (FIG. 7). At this time, since theetching stopper film105 is formed, the etching does not reach below theetching stopper105. Then, thefine holes113 for forming a cavity are filled up with a dielectric116 to seal the cavity115 (FIG. 8).
• Note that tungsten (W) is used as a material of thefirst layer interconnect104 and thefourth layer interconnect108, aluminum (Al) is used as a material of the second layer interconnect and the third layer interconnect, and tungsten silicide (WSi) is used as a material of the fifth layer interconnect. The materials are not limited to them, but a stacked film of an aluminum film and a titanium nitride (TiN) film can be used as thefirst layer interconnect104 and tungsten can be used as a material of the fifth layer interconnect. The advantage of using the above-described materials for thefirst layer interconnect104 and the fifth layer interconnect is that the etching selectivity (in etching rates) between theinterlayer dielectric106 and theinterlayer dielectric112 can be sufficiently ensured in the etching for forming thecavity115.
• Next, the configuration and operation of the single axis acceleration sensor according to the first embodiment will be described.FIG. 9A toFIG. 9C are schematic diagrams showing the structure pattern in each of the layers constituting the completed single axis acceleration sensor. The plan views of the M1 layer, the M4 layer and the M5 layer are shown inFIGS. 9A,9B and9C, respectively. InFIG. 9A, theetching stopper film105 of the M1 layer functions as a capacitor lower electrode and is connected by thefirst layer interconnect104 to an integrated circuit including thesignal processing transistors102 integrated on the same substrate. InFIG. 9B, themovable mass109 formed in the M4 layer is connected to the fixedbeam111 via theelastic beam110 formed in a spiral shape. Themovable mass109 functions as a capacitor upper electrode and is electrically connected to the integrated circuit including thesignal processing transistors102 by theelastic beam110, the fixedbeam111 and thefourth layer interconnect108. Themovable mass109 formed in thecavity115 is mechanically fixed to theinterlayer dielectric112 via theelastic beam110 and the fixedbeam111. In this configuration, themovable mass109 is moved in the direction vertical to the paper of the drawing by the acceleration in the direction vertical to the paper of the drawing. Therefore, the distance between the capacitor upper electrode composed of themovable mass109 and the capacitor lower electrode composed of theetching stopper film105 is changed and the interelectrode capacitance is changed. By detecting this change in capacitance by the integrated circuit including the signal processing transistors102 (capacitance detecting circuit), the acceleration can be detected as the change in capacitance. More specifically, the single axis acceleration sensor according to the first embodiment can detect the acceleration acting in the direction vertical to the silicon substrate101 (chip).
• As shown inFIG. 9C, thecavity cover film114 is formed in the M5 layer, and thecavity cover layer114 has thefine holes113 used to form thecavity115. Thecavity115 is formed by the etching through the fine holes113. After forming thecavity115, thefine holes113 are filled up.
• As shown inFIG. 9B, the shape of the fixedbeam111 is designed so that a base part thereof is sufficiently thick to prevent the elastic deformation even when the acceleration is applied to themovable mass109. On the other hand, theelastic beam110 positioned at the center of the beam is designed so as to have the smaller width in comparison to that of the fixedbeam111 and a sufficient length because of its spiral shape, and the desired elastic deformation occurs when a predetermined acceleration is applied. Therefore, the mechanical characteristics of the single axis acceleration sensor are determined only by the pattern shape and the thickness of the fixedbeam111, theelastic beam110 and themovable mass109 in the M4 layer, and do not depend on the dimensions and the shape of thecavity115. Since the dimensional accuracy of the fixedbeam111, theelastic beam110 and themovable mass109 is determined by the dimensional accuracy of the M4 layer (dimensional accuracy for forming the interconnect), it is very accurate. Meanwhile, since the dimensions and the shape of thecavity115 are determined by the etching of theinterlayer dielectrics106 and112, the accuracy thereof is not so high. However, it does not influence the mechanical characteristics of the single axis acceleration sensor according to the first embodiment.
• More specifically, the single axis acceleration sensor according to the first embodiment is designed so that the mechanical characteristics are determined by themovable mass109, theelastic beam110 and the fixedbeam111. Therefore, thecavity115 can be formed by the etching technology (etching of interlayer dielectric itself) with lower accuracy than that of the normal manufacturing technology of a CMOS integrated circuit, that is, the case where the mechanical characteristics of the MEMS are determined by the shape of thecavity115.
• On the contrary, in the technology of the U.S. Pat. No. 6,635,506, the mechanical characteristics are influenced by the cavity. Therefore, it is necessary to form the cavity with high accuracy. For its achievement, a sacrificial layer made of a material different from that of the interlayer dielectric is formed in the region of an interlayer dielectric for forming a cavity. As a result, the process for forming the cavity is complicated.
• Meanwhile, in the first embodiment, it is not necessary to form thecavity115 with high accuracy. Therefore, thecavity115 is formed by the etching of theinterlayer dielectrics106 and112 themselves through thefine holes113 without using a sacrificial layer. Also, after forming thecavity115, thefine holes113 are filled up by the dielectric116 to seal thecavity115. The deposition of the dielectric116 which is included in the normal manufacturing technology of a CMOS integrated circuit is used also in this sealing process. More specifically, the process for forming and sealing thecavity115 can be simplified in the first embodiment.
• As described above, in the first embodiment, since thecavity115 can be formed and sealed through the standard CMOS process, the special process for forming and sealing the cavity (packaging process particular to MEMS) which is the major cause of the yield decrease and the manufacturing cost increase in the conventional MEMS manufacturing process becomes unnecessary. Therefore, according to the first embodiment, it is possible to improve the yield, reduce the manufacturing (packaging) cost and improve the reliability. In addition, since the structure of the MEMS (single axis acceleration sensor) can be formed simultaneously with the formation of the interconnect of the LSI, the integration with the LSI can be facilitated.
• Note that the planar shapes of themovable mass109, theelastic beam110 and the fixedbeam111 are not limited to those shown inFIG. 9B. For example, the configuration in which a movable mass located at the center is supported by elastic beams provided at the four corners.
• Next, the capacitance detecting circuit will be described.FIG. 10 is a block diagram showing the circuit configuration of an integrated circuit (capacitance detecting circuit) including thesignal processing transistors102.
• InFIG. 10, the capacitance detected by anacceleration sensor117 is converted into voltage by a C-V (capacitance-voltage)conversion circuit118. Thereafter, the voltage converted by theC-V conversion circuit118 is amplified by anoperation amplifier119 and then digitized by anA-D conversion circuit120. Then, various types of correction such as temperature and amplifier characteristics are performed by amicro processor121 based on the data stored in anon-volatile memory122, and it is outputted as the acceleration from anoutput interface circuit123. Note that the acceleration detecting accuracy can be further improved by simultaneously detecting a suitable fixed reference capacitance or capacitance change between the M5 layer and themovable mass109 to use it as the difference input of the capacitance detecting circuit.
• Next, an application of the single axis acceleration sensor according to the first embodiment will be described. The single axis acceleration sensor according to the first embodiment is mounted together with a pressure sensor for a TPMS (tire pressure monitoring system). The single axis acceleration sensor detects the acceleration based on the displacement of a movable mass due to the centrifugal force from the rotation of the tires and the vibration from the road surface, and determine the operational state of an automobile, that is, whether or not the automobile is running or not. Then, based on the detection result of the single axis acceleration sensor, the frequency of data transmission using RF such as the tire pressure information outputted from the pressure sensor is determined. That is, by providing the single axis acceleration sensor, the frequency of RF transmission of tire pressure information detected by the pressure sensor can be increased when the automobile is running, and the frequency of RF transmission of tire pressure information detected by the pressure sensor can be reduced when the automobile is not running. Consequently, the needless RF transmission can be reduced and the lifetime of battery can be extended.
• The pressure sensor can be formed through the interconnect process similar to the single axis acceleration sensor.FIG. 11 is a cross-sectional view showing the device in which the singleaxis acceleration sensor130 according to the first embodiment and thepressure sensor130 are formed at the same time.
• Alower electrode132 of thepressure sensor131 is formed from the same layer as themovable mass109 of the singleaxis acceleration sensor130. An upper electrode133 (also functioning as diaphragm film) of thepressure sensor131 is formed from the same layer as the cavity cover film114 (sealing film) of the singleaxis acceleration sensor130. The formation of thecavity134 of thepressure sensor131 and the formation of thecavity115 of the singleaxis acceleration sensor130 are performed at the same time throughfine holes135 provided in theupper electrode133 of thepressure sensor131 and thefine holes113 provided in thecavity cover film114 of the singleaxis acceleration sensor130. Similarly, the sealing of theupper electrode133 of thepressure sensor131 and the sealing of thecavity cover film114 of the singleaxis acceleration sensor130 are also performed at the same time. In this manner, thepressure sensor131 can be manufactured in parallel through approximately the same process as that of the singleaxis acceleration sensor130 according to the first embodiment.
• In thepressure sensor131, the pressing force to theupper electrode133 shown inFIG. 11 is changed (position of theupper electrode133 is changed) along with the change in gas pressure around thepressure sensor131. Therefore, the distance between theupper electrode133 and thelower electrode132 is changed and the interelectrode capacitance is changed. Consequently, in thepressure sensor131, the gas pressure can be detected by detecting this interelectrode capacitance.
Second Embodiment
• In this second embodiment, a modification example of the above-described first embodiment will be described. First, in the second embodiment, a 2-axis acceleration sensor which detects the acceleration in two directions within chip (directions orthogonal to each other) will be described. In the acceleration sensor, it is necessary to ensure a predetermined amount of mass of the movable mass. Therefore, the mechanical structure is formed from a pad layer with a relatively large thickness in the interconnect layers.
• FIG. 12 is a schematic diagram showing the cross-sectional structure of the 2-axis acceleration sensor, andFIG. 13A andFIG. 13B are schematic diagrams showing the configuration of the major layers. InFIG. 12, amovable mass202, an elastic beam203 (not shown inFIG. 12) and a fixed capacitor plate (capacitance detecting electrode)204 of the 2-axis acceleration sensor201 are formed from the same metal layer as thepad layer205 of the LSI in which they are monolithically integrated. The 2-axis acceleration sensor201 is formed on the normal LSI206 and acavity207 thereof is formed around themovable mass202 and sealed in the same manner as that of the first embodiment. However, when forming thecavity207, anetching stopper film208 is formed by using an appropriate interconnect layer just below the cavity forming region. Thisetching stopper film208 also functions as an electric shield between the underlying LSI206 (integrated circuit and multilayer interconnect) and the 2-axis acceleration sensor201. As described above, since the mechanical structure of the MEMS (2-axis acceleration sensor) can be formed and stacked on the LSI (interconnect part and device region), the miniaturization of a chip can be realized.
• In the etching for forming thecavity207, it is necessary to ensure the sufficient etching selectivity between theinterlayer dielectric210 to be etched and themovable mass202, theetching stopper film208 and thecavity cover film209 for sealing made of a pad layer material. In this case, thepad layer205 is a stacked film formed by sandwiching an aluminum (Al) film with a thickness of 1500 nm between titanium nitride (TiN) films each having a thickness of 100 nm. By doing so, in the etching for forming thecavity207, the etching selectivity to theinterlayer dielectric210 can be sufficiently ensured. If necessary, it is also possible to form sidewalls made of a titanium nitride film or a silicon nitride (SiN) film in order to prevent the side etching from the side of the aluminum film.
• In order to seal thecavity207 having an area large enough to include the relatively largemovable mass202, it is necessary to ensure the sufficient strength of thecavity cover film209. In this case, a tungsten silicide (WSi) film with a thickness of 1 μm is used. In order to prevent the adhesion and destruction of thecavity cover film209 due to the capillary force of the residual liquid in thecavity207 in the dry process after the etching, the vapor-phase etching using vaporized HF is used to form thecavity207.
• Next, the operation of the 2-axis acceleration sensor according to the second embodiment will be described. As shown inFIG. 13A, themovable mass202 in thecavity207 is fixed to theinterlayer dielectric210 via theelastic beam203 formed from the same layer. Theinterlayer dielectric210 also functions as a fixed beam considered to be elastically undeformed. By forming theelastic beam203 in a folded shape as shown inFIG. 13A, theelastic beam203 is elastically deformed by the force applied to themovable mass202 and the two dimensional position of themovable mass202 is changed in thecavity207. The amount of change is detected as the change in capacitance between amovable capacitor plate211 formed at a part of the movable mass and a fixedcapacitor plate204 fixed to theinterlayer dielectric210 and protruding in thecavity207. Themovable capacitor plate211 and the fixedcapacitor plate204 for detecting the displacement in the two directions (x and y directions) within a chip are arranged in a comb shape in which they are alternately formed in a lateral direction. A pair of fixedcapacitor plates204 sandwiching the onemovable capacitor plate211 are electrically independent from each other, and the capacitance between the fixedcapacitor plate204 and themovable mass202 is detected separately. For example, when themovable mass202 moves in the x direction, the distance between themovable capacitor plate211 and the fixedcapacitor plate204 arranged above and below is changed. More specifically, in themovable capacitor plate211 and one pair of fixedcapacitor plates204 arranged above and below, the distance between themovable capacitor plate211 and one fixedcapacitor plate204 is increased and the distance between themovable capacitor plate211 and the other fixedcapacitor plate204 is decreased. When the distance is changed, the capacitance is also changed. Therefore, by detecting the change in capacitance, the acceleration in the x direction can be detected. Also, when themovable mass202 moves in the y direction, the distance between themovable capacitor plate211 and the fixedcapacitor plate204 arranged side by side is changed. More specifically, in themovable capacitor plate211 and one pair of fixedcapacitor plates204 arranged side by side, the distance between themovable capacitor plate211 and one fixedcapacitor plate204 is increased and the distance between themovable capacitor plate211 and the other fixedcapacitor plate204 is decreased. Therefore, the acceleration in the y direction can be detected.
• These fixedcapacitor plates204 and the movable mass202 (including the movable capacitor plates211) in the x and y directions are electrically connected independently to the signal processing integrated circuit (LSI) integrated on the same semiconductor substrate. When themovable mass202 moves by the acceleration in an arbitrary direction of the two axes, the distance between the fixedcapacitor plate204 and themovable capacitor plate211 is changed and the interelectrode capacitance is changed. By detecting the change in capacitance by the signal processing integrated circuit (capacitance detecting circuit), the acceleration is detected.
• The shape of the beam is designed so that the base part thereof is sufficiently thick in thecavity207 to prevent the elastic deformation even when the acceleration is applied to the mass (fixed part, fixed beam). On the other hand, the center part of the beam is designed so as to have the smaller width in comparison to that of base part and a sufficient length because of its folded shape, and the desired elastic deformation occurs when a predetermined acceleration is applied (elastically deformable part, elastic beam203). Therefore, the mechanical characteristics are determined only by the planar pattern shape and the thickness of the part of the beam and themovable mass202 exposed in thecavity207, and do not depend on the dimensions and the shape of thecavity207. Since the dimensional accuracy of the fixed beam, theelastic beam203 and themovable mass202 is determined by the dimensional accuracy of the patterns of the interconnect layer and the via layer, it is very accurate. Meanwhile, since the dimensions and the shape of thecavity207 are determined by the etching of theinterlayer dielectric210, the accuracy thereof is not so high. However, it does not influence the mechanical characteristics of the 2-axis acceleration sensor according to the second embodiment.
• FIG. 13B shows thecavity cover film209 formed on thecavity207, andfine holes212 used to form thecavity207 are formed in thecavity cover film209. The fine holes212 are sealed with a dielectric when the formation of thecavity207 is finished.
• Next, an example in which the 2-axis acceleration sensor according to the second embodiment is formed simultaneously with the pressure sensor is shown inFIG. 14.FIG. 14 is a cross-sectional view of a compound sensor in which the 2-axis acceleration sensor201 according to the second embodiment is formed simultaneously with apressure sensor220 similar to that in the first embodiment.
• As shown inFIG. 14, the mechanical structure (movable mass202, fixedcapacitor plate204 and the like) of the 2-axis acceleration sensor201 is formed from the same layer as that of thepad layer205, and aninterconnect221 for connecting an upper electrode and aninterconnect222 for connecting a lower electrode of thepressure sensor220 are formed thereon. Subsequently, after forming aninterlayer dielectric223, an opening for connecting upper electrode and an opening for connecting lower electrode are provided on thepad layer205.
• Next, alower electrode224 of thepressure sensor220 is formed and thelower electrode224 is connected to theinterconnect222 for connecting a lower electrode. Subsequently, a dielectric (oxide film)pattern225 for forming a cavity of thepressure sensor220 is formed. Then, a metal film (for example, tungsten film) to be the upper electrode (diaphragm film)226 of thepressure sensor220 and thecavity cover film209 of the 2-axis acceleration sensor201 is formed on the entire surface of the semiconductor substrate. Thereafter, after forming fine holes in the metal thin film on the cavity forming region of thepressure sensor220 and on the cavity forming region of the 2-axis acceleration sensor201, theinterlayer dielectric223 and thedielectric pattern225 are etched through the fine holes. By doing so, thecavity227 of thepressure sensor220 and thecavity207 of the 2-axis acceleration sensor201 are formed. Subsequently, the fine holes formed in the metal thin film are sealed.
• Thereafter, the metal thin film is patterned to form theupper electrode226 of thepressure sensor220 and thecavity cover film209 of the 2-axis acceleration sensor201. Then, a passivation film made of a silicon nitride film is deposited and an opening is formed on thepressure sensor220 and the predetermined pad (not shown).
• As described above, also in this second embodiment, thepressure sensor220 and the 2-axis acceleration sensor201 can be formed almost at the same time.
• According to the second embodiment, since themovable mass202 of the 2-axis acceleration sensor201 can be formed from the interconnect in the same layer as the pad layer (relatively thick layer)205, the mass of themovable mass202 can be greatly increased in comparison to the first embodiment. Therefore, it is possible to improve the sensitivity of the 2-axis acceleration sensor201. Note that, also in the second embodiment, the formation of the mechanical structure and the formation and sealing of the cavity of the 2-axis acceleration sensor can be performed through the normal CMOS process. Therefore, the advantages similar to those of the first embodiment can be realized.
Third Embodiment
• In this third embodiment, the MEMS switch mounted together with the LSI will be described. In the third embodiment, the MEMS switch and the integrated circuit are monolithically mounted by using the normal manufacturing technology of a CMOS integrated circuit (interconnect process). Since the MEMS switch is formed on the multilayer interconnect of the integrated circuit, the increase of the chip area can be prevented. The MEMS switches are used to switch the circuit blocks and to switch the input/output RF devices and the antennas in accordance with the RF (Radio Frequency) wireless communication system. By doing so, the connection to low-power-consumption wireless devices and antennas with small loss can be realized.
• First, the function and basic operation of aMEMS switch300 according to the third embodiment will be described.FIG. 15A toFIG. 15C are plan views schematically showing the configuration and the basic operation of theMEMS switch300 according to the third embodiment.
• The function of theMEMS switch300 according to the third embodiment is to connect or disconnect the input to or from the output in response to the control signal. TheMEMS switch300 has three states, that is, a connection state, a disconnection state and a transition state. In the connection state, as shown inFIG. 15C, twocontacts301aand301bof a centralmovable part301 are in contact with acontact302aof aninput line302 and acontact303aof anoutput line303. Meanwhile, in the disconnection state, as shown inFIG. 15A, the twocontacts301aand301bof the centralmovable part301 are not in contact with thecontact302aof theinput line302 and thecontact303aof theoutput line303. In these connection state and disconnection state, theinput line302 and theoutput line303 of theMEMS switch300 are electrically connected to the signal lines of each integrated circuit. Also, the transition state shown inFIG. 15B is the state corresponding to the transition from the connection state to the disconnection state or from the disconnection state to the connection state, in which theinput line302 and theoutput line303 of theMEMS switch300 are electrically disconnected from the signal lines of each integrated circuit and are connected to the signal line from a switch control unit.
• Next, the function of each component will be described. Theinput line302 is comprised of a fixed part (fixed beam)306 fixed to aninterlayer dielectric305 formed so as to surround acavity304 and amovable part308 including an elasticallydeformable spring part307 and thecontact302a. A part of themovable part308 constitutes oneelectrode309aof a comb-shapeddisplacement actuator309. On the other hand, theother electrode309bof the comb-shapeddisplacement actuator309 is fixed to theinterlayer dielectric305.
• The configuration of theoutput line303 is almost symmetrical to that of theinput line302. More specifically, theoutput line303 is comprised of a fixed part (fixed beam)310 fixed to theinterlayer dielectric305 formed so as to surround thecavity304 and amovable part312 including an elasticallydeformable spring part311 and thecontact303a. A part of themovable part312 constitutes oneelectrode313aof a comb-shapeddisplacement actuator313. On the other hand, theother electrode313bof the comb-shapeddisplacement actuator313 is fixed to theinterlayer dielectric305.
• In addition, the centralmovable part301 is also comprised of almost the same components, that is, it is comprised of a fixed part (fixed beam)314 fixed to theinterlayer dielectric305 formed so as to surround thecavity304 and amovable part316 including an elasticallydeformable spring part315 and thecontacts301aand301b. More specifically, an interconnect having one end fixed to theinterlayer dielectric305 is electrically and mechanically connected to themovable part316 including thecontacts301aand301bvia the elasticallydeformable spring part315. A part of themovable part316 constitutes oneelectrode317aof a comb-shapeddisplacement actuator317. Meanwhile, theother electrode317bof the comb-shapeddisplacement actuator317 is fixed to theinterlayer dielectric305. Note thatFIG. 15A toFIG. 15C are schematic diagrams, and the planar configuration of the spring part and the actuator is simplified.
• Next, the actual operation of the MEMS switch will be briefly described with using the transition from the disconnection state to the connection state as an example. In the disconnection state shown inFIG. 15A, the three comb-shapeddisplacement actuators309,313 and317 are not actuated, and no force is applied to any of the threespring parts307,311 and315. Therefore, theinput line302 and theoutput line303 are switched from the integrated circuit signal line to the actuator control signal line. Then, the voltage is applied between a pair ofelectrodes309aand309bof the comb-shapeddisplacement actuator309 provided in theinput line302. By doing so, themovable part308 is electrostatically actuated and thecontact302aof theinput line302 is displaced to outside. Similarly, the voltage is applied between a pair ofelectrodes313aand313bof the comb-shapeddisplacement actuator313 provided in theoutput line303. By doing so, themovable part312 is electrostatically actuated and thecontact303aof theoutput line303 is displaced to outside.
• In this state, the centralmovable part301 can move in a longitudinal direction with no interference. Further, by applying the voltage between a pair ofelectrodes317aand317bof the comb-shapeddisplacement actuator317 provided in the centralmovable part301, the centralmovable part301 is electrostatically actuated and is displaced upward (FIG. 15B).
• Subsequently, themovable part308 of theinput line302 and themovable part312 of theoutput line303 are returned to the initial positions. Thereafter, when the actuation of the comb-shapeddisplacement actuator317 of the centralmovable part301 is stopped, the centralmovable part301 is fixed to theinput line302 and theoutput line303 by the force of thespring part315. More specifically, thecontacts301aand301bof the centralmovable part301 are connected to thecontact302aof theinput line302 and thecontact303aof theoutput line303, respectively (FIG. 15C). Thereafter, by switching theinput line302 and theoutput line303 to the integrated circuit signal line, it comes to the connection state.
• Next, the manufacturing process of theMEMS switch300 according to the third embodiment will be briefly described. Similar to the second embodiment, the components of theMEMS switch300 are formed from only one interconnect layer, and the manufacturing process thereof is almost identical to that of the first and second embodiments. More specifically, the transistors and the multilayer interconnect are formed through the normal CMOS integrated circuit process, and the mechanical structure of theMEMS switch300 is formed thereon by using the almost identical method shown in the above-described second embodiment. That is, in the third embodiment, instead of forming the structure of the acceleration sensor, the structure of theMEMS switch300 is formed. Similar to the second embodiment, this structure is formed with using a part of the uppermost layer of the multilayer interconnect. However, it is also possible to form the structure with using an intermediate interconnect layer on the memory region in which the number of interconnects is small.
• Also, when forming a cavity around the mechanical structure, a thin film using an interconnect layer is formed as an etching stopper film just below the structure in the cavity forming region. This thin film functions as an electric shield for the transistors and the multilayer interconnects formed below.
• In the third embodiment, the electrical conduction must be obtained when the structures formed of the interconnects are connected. Therefore, it is necessary to prevent the adhesion of dielectrics on the surface of the structure. Also, it is also necessary to prevent the so-called sticking in which the metal bodies are not separated after being connected to each other. For its achievement, in addition to the sealing process described in the first embodiment (refer toFIG. 8), the following process shown inFIG. 16A toFIG. 16D must be performed.
• FIG. 16A toFIG. 16D are schematic diagrams showing a part of the manufacturing process of theMEMS switch300 according to the third embodiment, andFIG. 17A andFIG. 17B are schematic diagrams showing the configuration of the major layers constituting theMEMS switch300. As shown inFIG. 16A, an interconnect (not shown) and astructure320 of the MEMS switch are formed by using the predetermined interconnect layer. Then, aninterlayer dielectric321 is deposited on thestructure320, and a thin film made of a similar interconnect material is formed on theinterlayer dielectric321. Thereafter, holes are formed in this thin film to form acavity cover film322. The holes include at least two types of holes, that is,fine holes323 having a relatively small diameter (about 0.2 to 0.3 μm) and alarger hole324 having a diameter larger than that of thefine hole323. The plan view of thecavity cover film322 in which thefine holes323 and thelarger hole324 are formed is shown inFIG. 17A. As shown inFIG. 17A, thefine holes323 are arranged on the predetermined cavity forming region and thelarger hole324 is provided at the position below which the structure is not present. Note that the dotted line represents thestructure320 formed below thecavity cover film322 via theinterlayer dielectric321.
• Next, theinterlayer dielectric321 around thestructure320 is etched and removed through thefine holes323 and thelarger hole324 to form thecavity325. The plan view showing the layer in which thestructure320 is formed is shown inFIG. 17B. As shown inFIG. 17B, thecavity325 is formed in theinterlayer dielectric321, and thestructure320 is formed in thecavity325.
• Subsequently, thefine holes323 are filled up with a dielectric326 having isotropic deposition characteristics (FIG. 16B). At this time, the dielectric326 is also adhered to the surface of thestructure320 made of, for example, metal due to the deposition gas getting in thecavity325 through the fine holes323. Therefore, when only thefine holes323 are formed in thecavity cover film322, thecavity325 is sealed while the dielectric326 is adhered to the surface of thestructure320. As a result, the electrical conduction cannot be obtained when connecting the MEMS switch. However, thelarger hole324 is formed in addition to thefine holes323 in the third embodiment. Since thislarger hole324 is not filled up, thecavity325 is not sealed yet.
• Next, the dielectric326 adhered to the surface of thestructure320 in thecavity325 is etched through thelarger hole324, and then, the metal surface after the etching is hydrophobically treated so as to prevent the sticking (FIG. 16C). Thereafter, a dielectric327 is deposited by the anisotropic CVD under the reduced pressure to close thelarger hole324. In this manner, thecavity325 is completely sealed (FIG. 16D).
• As described above, according to the third embodiment, since it is possible to remove the dielectric326 formed on the surface of thestructure320 in the sealing process of thecavity325, the improvement of the reliability of the MEMS switch can be achieved. Note that, also in the third embodiment, the structure of the MEMS switch can be formed and the cavity can be formed and sealed through the CMOS process. Therefore, the advantages similar to those of the first embodiment can be realized.
Fourth Embodiment
• In this fourth embodiment, an example in which an integral movable part composed of a plurality of multilayer interconnects is used will be described. As a problem of the acceleration sensor using the surface MEMS, the increase of the mass of the movable mass is relatively difficult. This is because the thickness of the movable mass is determined by the thickness of the interconnect layer. In the fourth embodiment, the method and the structure capable of increasing the mass of the movable mass will be described. In order to increase the mass of the movable mass, an integral structure composed of a plurality of multilayer interconnects is used as a movable part. It is possible to simultaneously form the movable mass and the multilayer interconnects of a LSI.
• FIG. 18 toFIG. 25 are schematic diagrams for describing the manufacturing process of a 3-axis acceleration sensor according to the fourth embodiment, andFIG. 26A toFIG. 26C are schematic plan views showing the configuration of each layer of the structure constituting the 3-axis acceleration sensor.
• First, through the normal CMOS integrated circuit process,signal processing transistors402 andcontact holes403 of the 3-axis acceleration sensor are formed on a silicon substrate401 (FIG. 18). Next, through the similar CMOS integrated circuit process, a first layer interconnect (M1 layer)404 of the integrated circuit, amovable mass405 of the 3-axis acceleration sensor, an elastic beam406 (not shown) also functioning as an interconnect electrically and mechanically connected to themovable mass405 and alower electrode407 described later are formed from the metal patterns (FIG. 19). A schematic diagram of the M1 layer pattern of the 3-axis acceleration sensor is shown inFIG. 26. InFIG. 26, themovable mass405 is fixed to a fixed beam (interlayer dielectric) via theelastic beam406.
• Each of theelastic beam406 and thelower electrode407 is connected to the predetermined interconnects of thesignal processing transistors402 via thefirst layer interconnect404 and the contact holes. Fine holes408 for removing the interlayer dielectric just below themovable mass405 in the latter process are formed in themovable mass405.
• Thereafter, after aninterlayer dielectric409 is deposited by using the normal CMOS integrated circuit process, first layer viaholes410 of the integrated circuit are formed andopenings411 are formed in theinterlayer dielectric409 at the positions corresponding to themovable mass405 of the 3-axis acceleration sensor (FIG. 20). Next, the first layer viaholes410 and theopenings411 are filled with metal (for example, tungsten in this case) by using the normal CMOS integrated circuit process, and the surface thereof is planarized by the CMP. In this case, theopenings411 are formed on the patterns of the M1 layer other than thelower electrode407. Also, in order to prevent the so-called dishing in the CMP, the slit (unremoved pattern of dielectric) is appropriately inserted in the large-area part of themovable mass405 in addition to thefine hole408 for etching.
• Next, through the CMOS integrated circuit process, a second layer interconnect (M2 layer)412 of the integrated circuit is formed and themovable mass405, a movable capacitor plate and a fixed capacitor plate of the 3-axis acceleration sensor are formed from the metal pattern (FIG. 21). The schematic diagram of the pattern of the M2 layer in a part of the 3-axis acceleration sensor is shown inFIG. 26B. As shown inFIG. 26B,movable capacitor plates412aare formed on themovable mass405, and fixedcapacitor plates412bare formed so as to face to themovable capacitor plates412a. The fixedcapacitor plate412bis fixed to theinterlayer dielectric409.
• Thereafter, after aninterlayer dielectric413 is deposited by using the normal CMOS integrated circuit process again, second layer viaholes414 of the integrated circuit are formed andopenings415 are formed in theinterlayer dielectric413 at the positions corresponding to themovable mass405, themovable capacitor plate412aand the fixedcapacitor plate412bof the 3-axis acceleration sensor (FIG. 22). Next, the second layer viaholes414 and theopenings415 are filled with metal (for example, tungsten) by using the normal CMOS integrated circuit process, and the surface thereof is planarized by the CMP. In this case, the pattern of theopenings415 is almost the same as that of the M2 layer. However, in order to prevent the so-called dishing in the CMP, the slit (unremoved pattern of dielectric) is appropriately inserted in the large-area part of themovable mass405.
• Next, after a third layer interconnect (M3 layer)416 of the integrated circuit is formed by using the normal CMOS integrated circuit process again, themovable mass405, themovable capacitor plate412aand the fixedcapacitor plate412bsimilar to those formed in the M2 layer of the 3-axis acceleration sensor are formed from the metal pattern (FIG. 23). The patterns of the second layer viaholes414 and the patterns of the 3-axis acceleration sensor of the M3 layer are similar to those shown inFIG. 26B.
• Further, aninterlayer dielectric417 is deposited and the surface thereof is planarized by the CMP or the like according to need. Then, acavity cover film419 having afine hole418 for forming the cavity at the position corresponding to the center of themovable mass405 is formed from a fourth layer interconnect (M4 layer) (FIG. 24). A plan view of thecavity cover film419 of the 3-axis acceleration sensor is shown inFIG. 26C. Thefine hole418 is provided above the center of themovable mass405, and themovable mass405 is not present just under thefine hole418.
• Thereafter, the interlayer dielectric around themovable mass405 is etched and removed through thefine hole418 to form thecavity420. The etching in the depth direction is stopped on thesilicon substrate401. Since the etching proceeds isotropically, thecavity420 has a round shape. Finally, thefine hole418 is filled with a dielectric421 to seal the cavity420 (FIG. 25). Since the dielectric421 thick enough to have the strength capable of sealing thelarge cavity420 is used, thefine hole418 needs to have a certain size. Also, the relativelythick dielectric421 for the sealing is formed under the anisotropic deposition condition.
• Next, the operation of the 3-axis acceleration sensor according to the fourth embodiment will be described. As shown inFIG. 26A, in thecavity420, themovable mass405 is fixed to the interlayer dielectric via theelastic beam406 formed from the M1 layer. Theelastic beam406 has a square shape as shown inFIG. 26A, and theelastic beam406 is elastically deformed and themovable mass405 is three dimensionally displaced in thecavity420 when a force is applied to themovable mass405. However, the shape shown inFIG. 26A is a mere example and it can be optimized in various ways, for example, formed into the folded zigzag shape as shown in the second embodiment or the like.
• For example, as shown inFIG. 26B, the displacement of themovable mass405 in thecavity420 in two directions (x and y directions) in the chip surface is detected as the change in capacitance between themovable capacitor plate412aformed from the M2 layer, the second layer viahole414 and the M3 layer in a part of themovable mass405 and the fixedcapacitor plate412b(formed from the same layer as themovable capacitor plate412a) fixed to the interlayer dielectric and protruding to thecavity420. The configuration of themovable capacitor plate412aand the fixedcapacitor plate412band the detection principle are the same as those of the second embodiment.
• The displacement in the direction vertical to the chip surface (z direction) is detected by detecting the change in capacitance between alower electrode407 fixed to the interlayer dielectric of the M1 layer just below a part of themovable mass405 which is formed from the M2 or more layers and themovable mass405.
• The fixedcapacitor plates412b(lower electrode407) for each of the x, y and z directions and themovable mass405 are electrically connected independently to the signal processing integrated circuit on thesame silicon substrate401. When themovable mass405 moves in arbitrary three directions due to the acceleration, the distance between themovable capacitor plate412aand the fixedcapacitor plate412bor between themovable mass405 and thelower electrode407 is changed and the interelectrode capacitance is changed. By detecting the change in capacitance in the signal processing integrated circuit (capacitance detecting circuit), the acceleration is detected.
• Theelastic beam406 is designed so that the base part thereof is sufficiently thick in the cavity to prevent the elastic deformation even when the acceleration is applied to the mass (fixed part). On the other hand, the center part of the beam is designed so as to have the smaller width in comparison to that of the base part and a sufficient length because of its folded shape, and the desired elastic deformation occurs when a predetermined acceleration is applied (elastically deformable part). Therefore, the mechanical characteristics are determined only by the planar pattern shape and the thickness of theelastic beam406 and themovable mass405 exposed in thecavity420, and do not depend on the dimensions and the shape of thecavity420. Since the dimensional accuracy of theelastic beam406 and themovable mass405 is determined by the dimensional accuracy of the patterns of the interconnect layer and the via layer, it is very accurate. Meanwhile, since the dimensions and the shape of thecavity420 are determined by the etching of the interlayer dielectric, the accuracy thereof is not so high. However, it does not influence the mechanical characteristics of the 3-axis acceleration sensor according to the fourth embodiment. Also, in the fourth embodiment, since theelastic beam406 and themovable capacitor plate412aare formed in the different interconnect layers, they can be optimally designed regardless of the restrictions of the planar arrangement.
• Also, a protrusion is formed in a part of the M2 layer pattern of themovable mass405, and this protrusion is placed so as to overlap with a part of the patterns of the M1 layer and the M3 layer protruding from the interlayer dielectric to the cavity420 (not shown). By doing so, when themovable mass405 moves largely up and down, the above-described protrusion of themovable mass405 and the protruding parts of the M1 layer and the M3 layer are collided with each other, and the movable range of themovable mass405 is limited. Also, since the protrusion and the protruding parts are deformed when they are collided, the impact at the collision is reduced. Therefore, the impact resistance and the reliability can be improved in the fourth embodiment.
• Since the movable mass is formed from a plurality of interconnect layers in the fourth embodiment, the mass of the movable mass can be increased and the detection sensitivity of the 3-axis acceleration sensor can be improved.
• Note that theelastic beam406 can be designed in various ways. In the fourth embodiment, theelastic beam406 is formed of only the M1 layer. However, it is also possible to form it from all of the M1 layer, the first layer viahole410, the M2 layer, the second layer viahole414 and the M3 layer or the arbitrary combination thereof. For example, the elastic beam and the movable capacitor plate shown in the second embodiment can be formed by using all of them. Alternatively, threeelastic beams406 as shown inFIG. 27A can be formed from only the interconnect layers of the M1 layer, the M2 layer and the M3 layer. In this manner, the sensitivity (deformability) to the acceleration (force) in the longitudinal direction (z direction) is increased. In addition, as shown inFIG. 27B, it is also possible to form theelastic beam406 so as to have the folded zigzag shape in the longitudinal direction. By doing so, the deformation in the longitudinal direction (z direction) is further facilitated. The shape, dimensions and film thickness of themovable mass405 and theplastic beam406 are designed from the viewpoint of the desired acceleration range to be detected and the impact resistance.
• Furthermore, it is also possible to further increase the mass of the movable mass by additionally forming films made of other materials as themovable mass405. More specifically, the material of themovable mass405 in the fourth embodiment is not limited to the interconnect material of the integrated circuit, but the other inorganic or organic material is also available. However, it is necessary to use the material not removed by the etching of the interlayer dielectric. When an oxide film is used as the interlayer dielectric, various metal films such as a silicon-germanium film (SiGe film), a silicon nitride film (SiN film), a silicon oxide film, a single crystal silicon film, an amorphous silicon film, a polysilicon film and a polyimide film are also available. Further, a film other than oxide films can be used as the interlayer dielectric as described below.
• The manufacturing process of a structure in which a thick film is additionally formed to the movable mass in the 3-axis acceleration sensor according to the fourth embodiment will be shown inFIG. 28 toFIG. 33.
• On the structure formed through the process similar to that shown inFIG. 18 toFIG. 23, a thick resistlayer425 is coated, andopenings426 are formed on themovable mass405 formed from the M3 layer through the normal exposure and development process (FIG. 28). Next, a nickel (Ni)film427 is formed in theopenings426 by the electroless plating (FIG. 29) (After forming the nickel film, the surface is polished if necessary.).
• Thereafter, the thick resistlayer425 is removed to expose thenickel film427 to be a thick movable mass structure on themovable mass405 formed from the M3 layer (FIG. 30). Furthermore, apolyimide film428 is coated so as to cover thenickel film427 and the thermal treatment is performed. Thereafter, a tungsten (W) thin film is formed as acavity cover film429 by the sputtering (FIG. 31).
• After formingfine holes430 for etching through the normal exposure method in the tungsten thin film (cavity cover film429), thepolyimide film428 around thenickel film427 is etched and removed through the fine holes430 (FIG. 32). Then, the interlayer dielectric around the movable structure formed from the underlying interconnect film is etched and removed to form acavity431. Thereafter, thefine holes430 for etching are filled with a dielectric432 to seal the cavity431 (FIG. 33).
• As described above, since it is possible to form themovable mass405 including thenickel film427, it is possible to further increase the mass of themovable mass405. Therefore, the detection sensitivity of the 3-axis acceleration sensor can be further improved.
• Note that, also in the fourth embodiment, the mechanical structure of the acceleration sensor can be formed and the cavity can be formed and sealed through the CMOS process. Therefore, the advantages similar to those of the first embodiment can be realized.
Fifth Embodiment
• In this fifth embodiment, an example in which an integral structure composed of a plurality of multilayer interconnects is used will be described. In the structure according to the fifth embodiment, a movable part composed of multilayer interconnects and covered with a dielectric such as an oxide film is fixed to an interlayer film surrounding the cavity by an elastic beam similarly having the interconnect structure therein and covered with a dielectric such as an oxide film in the cavity formed in the interlayer dielectrics. In the embodiments described above, one movable part is composed of one type of contiguous conductors. Therefore, it has only one electrical function as one electrode or interconnect. However, in this fifth embodiment, since it is possible to introduce a plurality of independent interconnects in the structure, more complicated actuation of the movable part and the signal detection can be realized.
• FIG. 34 toFIG. 39 are schematic diagrams for describing a manufacturing process of an angle rate sensor (vibration gyroscope) according to the fifth embodiment.
• First, through the normal CMOS integrated circuit process,signal processing transistors502 of the vibration gyroscope andcontact holes503 are formed on a silicon substrate501 (FIG. 34). Next, by using the CMOS integrated circuit process, a first layer interconnect (M1 layer)504 of the integrated circuit is formed, and asacrificial layer506 is formed in a region corresponding to the cavity below amovable mass505 of the vibration gyroscope. Next, after depositing aninterlayer dielectric507, first layer viaholes508 of the integrated circuit are formed, andopenings509 are formed in the regions corresponding to the cavity surrounding themovable mass505 of the vibration gyroscope (FIG. 35). Next, by using the normal CMOS integrated circuit process, metal (for example, tungsten in this case) is filled in the first layer viaholes508 and theopenings509, and the surface thereof is planarized by the CMP.
• Subsequently, by using the normal CMOS integrated circuit process, second layer interconnects (M2 layer)510 of the integrated circuit are formed, and asacrificial layer511 is formed in a region corresponding to the cavity surrounding themovable mass505 and beams (not shown) of the vibration gyroscope. Also, aninterconnect512 is formed in themovable mass505 and the beams. Subsequently, after depositing aninterlayer dielectric513, second layer viaholes514 of the integrated circuit are formed, andopenings515 are formed in the region corresponding to the cavity surrounding themovable mass505 of the vibration gyroscope. Simultaneously, a viahole516 for connecting the upper and lower interconnects in the movable mass505 (and in the beams according to need) is formed (FIG. 36).
• Next, by using the normal CMOS integrated circuit process, a third layer interconnect (M3 layer)517 of the integrated circuit is formed, and asacrificial layer518 is formed in the region corresponding to the cavity surrounding themovable mass505 and the beams of the vibration gyroscope. Simultaneously, aninterconnect519 is formed in themovable mass505 and the beams. At this time, a part of theinterconnect519 is not connected to theinterconnect512 and is formed as an independent interconnect. More specifically, a plurality of independent movable interconnects are formed in themovable mass505.
• Similarly, after depositing aninterlayer dielectric520, third layer viaholes521 of the integrated circuit are formed, and anopening522 is formed in the region corresponding to the cavity surrounding themovable mass505 of the vibration gyroscope (FIG. 37). Next, by using the normal CMOS integrated circuit process, a fourth layer interconnect (M4 layer)523 of the integrated circuit is formed, and asacrificial layer524 is formed in the region corresponding to the cavity on themovable mass505 of the vibration gyroscope. Furthermore, after depositing a dielectric, aninterlayer dielectric526 havingfine holes525 for forming cavity is formed (FIG. 38).
• Thereafter,sacrificial layers506,511,518 and524 formed from the interconnect materials are etched and removed through thefine holes525 to form thecavity527. Finally, thefine holes525 for etching are filled with a dielectric528 to seal the cavity (FIG. 39). The interconnects and the sacrificial layers are formed of, for example, an aluminum film. However, a tungsten film is also available.
• An Al etching solution used for the etching of the sacrificial layers has the high selectivity to the interconnects. Therefore, when etching the sacrificial layers, the removal rate of the dielectric (oxide film) on the surfaces of themovable mass505 and the beams formed in thecavity527 is extremely low. In this manner, themovable mass505 in which a plurality of independent interconnects are formed can be formed.
• Next, the configuration of the angle rate sensor (vibration gyroscope) formed through the above-described manufacturing process will be described with reference toFIG. 40.
• Aframe structure531 is formed in thecavity530. Thisframe structure531 is fixed to theinterlayer dielectric533 surrounding thecavity530 viabeams532 with rigidity in a detection axis (y) direction extremely higher than that in an actuation axis (x) direction. More specifically, theframe structure531 is easily vibrated in the actuation (x) direction but hardly vibrated in the detection (y) direction. Inside theframe structure531, themovable mass534 is fixed to theframe structure531 viabeams535 with rigidity in an actuation axis (x) direction extremely higher than that in a detection axis (y) direction.
• A comb-shapedfirst actuation electrode536 fixed to theinterlayer dielectric533 surrounding thecavity530 is connected to the predetermined LSI interconnect. A comb-shapedsecond actuation electrode537 fixed to theframe structure531 is connected to the predetermined LSI interconnect outside thecavity530 via the interconnect in thebeam532. The alternating voltage is applied between thefirst actuation electrode536 and thesecond actuation electrode537 by anoscillator538.
• A comb-shapedfirst detection electrode539 fixed to theframe structure531 and asecond detection electrode540 electrically independent from theelectrode539 are connected to the predetermined LSI interconnect provided outside thecavity530 via the interconnect in thebeam532.
• A comb-shapedthird detection electrode541 fixed to themovable mass534 is connected to the predetermined LSI interconnect provided outside thecavity530 via the interconnect in thebeam535, theframe structure531 and the interconnect in thebeam532. An electrostaticcapacitance detecting circuit542 is connected between thefirst detection electrode539 and thethird detection electrode541, and an electrostaticcapacitance detecting circuit543 is connected between thesecond detection electrode540 and thethird detection electrode541.
• All of the electrodes described above are covered with a dielectric. Also, all of the electrode described above are formed from the stacked films of the M2 layer and the M3 layer. The interconnect connected to thesecond actuation electrode537 is formed from the M3 layer and the other interconnects connected to the other electrodes are all formed from the M2 layer.
• Next, the operation of the vibration gyroscope according to the fifth embodiment will be described with reference toFIG. 41. Hereinafter, the actuation axis and the detection axis are considered as the coordinate system fixed to thecavity530. First, by applying the alternating voltage between thefirst actuation electrode536 and thesecond actuation electrode537, theframe structure531 is vibrated in the actuation axis direction. At this time, since thebeam535 connecting theframe structure531 and themovable mass534 has the high rigidity in the actuation direction, themovable mass534 is also vibrated in the actuation axis direction together with the frame structure531 (FIG. 41A).
• Next, when the rotation around the axis vertical to the actuation axis and the detection axis (axis vertical to the paper ofFIG. 41) occurs, themovable mass534 starts to vibrate in the detection axis direction due to the Coriolis force. At this time, since thebeam532 which fixes theframe structure531 to the surrounding surface of thecavity530 has the high rigidity in the detection axis direction, theframe structure531 is not vibrated in the detection axis direction. Therefore, thethird detection electrode541 moves in the detection axis direction relative to thefirst detection electrode539 and thesecond detection electrode540, and the capacitance between thethird detection electrode541 and thefirst detection electrode539 or the capacitance between thethird detection electrode541 and thesecond detection electrode540 is changed. By detecting the change in capacitance, the Coriolis force is measured, and the angle rate is detected (FIG. 41B).
• In the vibration gyroscope according to the fifth embodiment, since a plurality of independent interconnects can be formed in the structure, it is possible to connect independent circuits to the structure. Therefore, since it is unnecessary to perform the separation of the detected signals in the vibration direction, the extremely highly accurate detection of an angle rate can be realized and the signal processing can be greatly simplified. Note that, also in the fifth embodiment, the structure of the vibration gyroscope can be formed and the cavity can be formed and sealed through the CMOS process. Therefore, the advantages similar to those of the first embodiment can be realized.
Sixth Embodiment
• In this sixth embodiment, an example in which a MEMS structure is formed through the process other than the interconnect process and a cavity is formed and sealed for the structure through the interconnect process will be described. As described in the foregoing embodiments, in the case where an interconnect material, that is, metal is used to form beams and a movable part, when the MEMS structure is applied and used as a vibration part, the vibration value Q is small due to the characteristics of the metal material. In this case, a material capable of obtaining the relatively large vibration value Q such as silicon (Si) or the like is suitable. In the sixth embodiment, an example in which the structure of the vibration gyroscope formed through the SOI (Silicon On Insulator) process is sealed through the interconnect process will be described.
• FIG. 42 toFIG. 47 are schematic diagrams for describing the manufacturing process of the vibration gyroscope according to the sixth embodiment, andFIG. 48 is a schematic plan view showing the planar configuration of each layer of the structure constituting the vibration gyroscope. First, in order to form the vibration part on theSOI substrate601,openings604 extending from the surface of theSOI substrate601 to the embeddeddielectric603 are formed in theSOI layer602 around the patterns to be the vibration parts (mass and beam), and then, theopenings604 are filled with a CVD oxide film (HLD film)605 (FIG. 42).
• Next, through the normal CMOS integrated circuit process,transistors606 for actuation and signal processing of the vibration gyroscope andcontact holes607 are formed on the SOI substrate601 (FIG. 43). At this time, afield oxide film608 is formed on theSOI layer602 in the vibration part forming region and its adjacent region.
• Subsequently, by using the normal CMOS integrated circuit process, a first layer interconnect (M1 layer)609 of the integrated circuit is formed, and adetection electrode610 is formed on the central part of the region for forming the vibration part of the vibration gyroscope (FIG. 44). Thereafter, a second layer interconnect (M2 layer)611 and subsequent multilayer interconnects are formed through the normal CMOS integrated circuit process on the integrated circuit. At this time, only an interlayer dielectric is deposited on the vibration part forming region and its adjacent region. After forming the uppermost interconnect, aninterlayer dielectric612 is further deposited and its surface is planarized by the chemical mechanical polishing (CMP) according to need. Then, acavity cover film614 havingfine holes613 for etching is formed on the interlayer dielectric612 (FIG. 45).
• Thereafter, the interlayer dielectric on thevibration part615, theCVD oxide film605 embedded in theopenings604 and the embeddeddielectric603 of theSOI substrate601 below the vibration part (mass and beam)615 are etched and removed through thefine holes613 to form thecavity616 around the vibration part615 (FIG. 46). The etching in the depth direction is stopped on thesilicon substrate601 below the embeddeddielectric603. Finally, thefine holes613 for etching are filled with a dielectric617 to seal the cavity616 (FIG. 47).
• In an application using the vibration characteristics of the mechanical structure (vibration part) like in the sixth embodiment, the influence of the gas resistance around the structure is not negligible. Therefore, thecavity616 is desirably in a vacuum state. However, since the CVD film having the isotropic deposition characteristics is used to fill thefine holes613, the gas pressure at the time of deposition is left in thecavity616, and vacuum sealing is difficult. On the other hand, since the forming pressure of the anisotropic deposition film used to fill the larger hole shown in the third embodiment is close to a vacuum, it is possible to form a near-vacuum state. Therefore, similar to the third embodiment, at least two sizes of holes, that is, larger holes and fine holes are formed in thecavity cover film614 also in the sixth embodiment. Then, the planar shape of thecavity616 is defined through the fine holes, and the vacuum sealing of the cavity is performed through the larger holes. Note that, in the sixth embodiment, different from the third embodiment, the etching of the dielectric formed on the surface of the structure and the surface hydrophobic treatment are not always necessary before the sealing of the larger holes.
• Next, the configuration of the angle rate sensor (vibration gyroscope) formed through the process described above will be described with reference toFIG. 48.FIG. 48A is a plan view showing the SOI layer, in which the vibration gyroscope includes thevibration part615 having aframe structure620 and amass621,first actuation electrodes622 fixed to the surface of thecavity616 andsecond actuation electrodes623 fixed to theframe structure620. In thecavity616, theframe structure620 formed from the SOI layer is fixed to theinterlayer dielectric625 surrounding thecavity616 viabeams624 with rigidity in an actuation axis (x) direction lower than that in other directions. More specifically, theframe structure620 is easily vibrated in the actuation axis (x) direction but hardly vibrated in the detection axis (direction vertical to the paper ofFIG. 48) direction and in the rotation axis (y) direction.
• Inside theframe structure620, themass621 formed from the SOI layer is fixed to theframe structure620 viabeams626 with rigidity in the actuation axis direction and rotation axis direction sufficiently higher than that in other directions. More specifically, themass621 is easily vibrated in the detection axis (z) direction but hardly vibrated in the other detections.
• Two types of electrodes such as thefirst actuation electrodes622 and thesecond actuation electrodes623 are formed from a diffusion layer formed by the ion implantation and are connected to the integrated circuits for actuation and signal processing via the contact holes and the multilayer interconnects. The shape of these electrodes is defined by the etching when forming theopenings604. By applying alternating voltage between thefirst actuation electrode622 and thesecond actuation electrode623, thevibration part615 is vibrated in the actuation axis direction ofFIG. 48.
• FIG. 48B is a plan view showing thedetection electrode610 formed from the M1 layer. By detecting electrostatic capacitance between thedetection electrode610 and themass621, the displacement of the vibration part in the detection axis direction (direction vertical to the paper ofFIG. 48 or the substrate surface of the chip) is detected. The opposing area between the mass621 and thedetection electrode610 is not changed even when themass622 is vibrated (moved) in the actuation axis direction. Therefore, the electrostatic capacitance between the mass621 and thedetection electrode610 almost depends on only the space (distance) therebetween.
• The shape of thebeam626 is designed so that a base part thereof is sufficiently thick and it is not displaced by the designed range of vibration. Therefore, since the mechanical characteristics are determined only by the planar shape and the thickness of themass621 exposed in thecavity616 and do not depend on the dimensions and the shape of thecavity616, it is very accurate.
• Next, the operation of the angle rate sensor according to the sixth embodiment will be described with reference toFIG. 49. By applying alternating voltage between the first actuation electrode and the second actuation electrode, theframe structure620 is vibrated in the actuation axis direction. At this time, since thebeam626 connecting theframe structure620 and themass621 has high rigidity in the actuation axis direction, themass621 is vibrated in the actuation axis direction together with the frame structure620 (FIG. 49A). Next, when the rotation occurs around the rotation axis, the mass621 starts to vibrate in the detection axis direction due to the Coriolis force. As a result, the electrostatic capacitance between the mass621 and thedetection electrode610 is changed. By detecting the change in capacitance, the angle rate is monitored (FIG. 49B).
• As described above, themass621 is formed from only the SOI layer in the sixth embodiment. However, it is also possible to further stack a contact layer and multilayer interconnects on the SOI layer as themass621 in order to increase the mass of themass621. The process will be described below in brief.
• First, in order to form the vibration part on theSOI substrate601,openings604 extending from the surface of theSOI substrate601 to the embeddeddielectric603 are formed in theSOI layer602 around the patterns to be the vibration parts (mass and beam), and then, theopenings604 are filled with a CVD oxide film (HLD film)605 (FIG. 50).
• Next, through the normal CMOS integrated circuit process,transistors606 for actuation and signal processing of the vibration gyroscope andcontact holes607 are formed on the SOI substrate601 (FIG. 51). At this time,openings630 are also formed on theSOI layer602 in the vibration part forming region and its adjacent region.
• Subsequently, through the normal CMOS integrated circuit process, a first layer interconnect (M1 layer)631 and subsequent multilayer interconnects are formed on the integrated circuit. At this time, the multilayer interconnects632 constituting the vibration part are formed also on theopenings630. Then, after forming theinterlayer dielectric633, a third layer interconnect (M3 layer)634 is formed on the integrated circuit, anddetection electrodes635 are formed on the region for forming the vibration part. In this embodiment, thedetection electrode635 is formed from the same layer as the M3 layer. However, it is also possible to form thedetection electrode635 from an appropriate layer in the multilayer interconnects. Next, after forming aninterlayer dielectric636 on the M3 layer and thedetection electrodes635, acavity cover film638 havingfine holes637 for etching is formed on the interlayer dielectric636 (FIG. 52).
• Thereafter, the interlayer dielectric on thevibration part615, theCVD oxide film605 embedded in theopenings604 and the embeddeddielectric603 of theSOI substrate601 below the vibration part (mass and beam)615 are etched and removed through thefine holes637 to form thecavity639 around the vibration part615 (FIG. 53). The etching in the depth direction is stopped on thesilicon substrate601 below the embeddeddielectric603. Finally, thefine holes637 for etching are filled with a dielectric640 to seal the cavity639 (FIG. 54).
• As described above, since the multilayer interconnects are used to form thevibration part615 in addition to the SOI layer, it is possible to increase the mass of thevibration part615, and the detection sensitivity of the vibration gyroscope can be improved.
• It is also possible to use a thick polysilicon film to form the vibration part instead of the SOI layer. In this case, by using the substrate formed by sequentially stacking an oxide film and a polysilicon film with a predetermined thickness on a silicon substrate instead of the SOI substrate, the sixth embodiment can be applied without modification.
• The patterning of the vibration part composed of an SOI layer or a thick polysilicon film, that is, the definition of the planar shape of the vibration part and its peripheral part by the etching and the embedding of the oxide film (sacrificial film) to the etched part can be performed either before or after forming the transistors of the integrated circuit.
• The first point of the sixth embodiment is that the vibration part formed from a SOI layer or formed by stacking a SOI layer and multilayer interconnects and the detection electrodes formed from the multilayer interconnects are combined to constitute the angle rate sensor. Also, the second point thereof is that the vibration part of the angle rate sensor is placed in the cavity formed and sealed in the interlayer dielectric, which does not define the design characteristics of the angle rate sensor. Therefore, the planar shape and the arrangement described above are mere schematic examples, and the modification and the design optimization can be made appropriately.
• Note that, also in the sixth embodiment, the formation of the mechanical structure of the vibration gyroscope and the formation and sealing of the cavity can be performed through the normal CMOS process. Therefore, the advantages similar to those of the first embodiment can be realized.
• Next, the application examples of the MEMS described in the above-described embodiments will be described in the following seventh to ninth embodiments.
Seventh Embodiment
• The entire configuration of a gas pressure monitoring system for tire using the MEMS will be described with reference toFIG. 55 toFIG. 57.FIG. 55 is a diagram showing the configuration of the gas pressure monitoring system for tire seen from the bottom surface of an automobile, and it is comprised of avehicle701,tires702ato702dprovided on left front, right front, left rear and right rear of the vehicle, tirepressure measurement modules703ato703dprovided to each of thetires702ato702d, and an in-vehicle unit704. Also,FIG. 56 andFIG. 57 are block diagrams of the tire pressure measurement module703 (703ato703d) and the in-vehicle unit704, respectively.
• FIG. 56 is a block diagram of the tirepressure measurement module703, and is comprised of oneIC chip705 and abattery706 with the voltage of VBAT for supplying the power to theIC chip705. TheIC chip705 includes apressure sensor circuit707, atemperature sensor circuit708, anacceleration sensor circuit709, analog-digital (A/D)conversion circuits710,711 and712, a dataprocessing control unit713, amemory circuit714, adata transmission circuit715 and adata receiver circuit716. In this case, thepressure sensor circuit707 and thetemperature sensor circuit708 are the circuits for measuring the gas pressure and the temperature of the tires, respectively. Also, theacceleration sensor circuit709 is the circuit for determining whether the tires are rotated. In thepressure sensor circuit707 and theacceleration sensor circuit709, the MEMS (micro machine) constituting the pressure sensor and the acceleration sensor are formed. More specifically, in theIC chip705, in addition to the integrated circuit, the MEMS to be the pressure sensor and the acceleration sensor are formed. As the pressure sensor formed in thepressure sensor circuit707 and the acceleration sensor formed in theacceleration sensor circuit709 according to the seventh embodiment, for example, the pressure sensor and the acceleration sensor described in the first, second and fourth embodiments are used.
• In the MEMS formed in theIC chip705, as described in the foregoing embodiments, the cavity can be formed and sealed through the normal CMOS process. Therefore, the special process for forming and sealing the cavity (packaging process particular to MEMS) which is the major cause of the yield decrease and the manufacturing cost increase in the conventional MEMS manufacturing process becomes unnecessary. As a result, in the seventh embodiment, it is possible to improve the yield, reduce the manufacturing (packaging) cost and improve the reliability of theIC chip705. In addition, since the structure of the MEMS can be formed simultaneously with the formation of the interconnect of the LSI, the integration with the LSI can be facilitated.
• The analog-digital conversion circuits710,711 and712 are the circuits for converting the analog voltage value outputted from thepressure sensor circuit707, thetemperature sensor circuit708 and theacceleration sensor circuit709 into the digital voltage value.
• The data processing control unit713 (1) inputs the digital voltage value converted in the analog-digital conversion circuits710,711 and712, (2) performs the correction computing to correct the pressure measurement value measured in thepressure sensor circuit707, (3) changes the control state in accordance with the output from theacceleration sensor circuit709, (4) outputs the data to thedata transmission circuit715, (5) receives the data from thedata receiver circuit716, and (6) controls the ON/OFF of the individual power supplies of thepressure sensor circuit707, thetemperature sensor circuit708, theacceleration sensor circuit709, thedata transmission circuit715 and thedata receiver circuit716 in response to the EN signals.
• Thememory circuit714 is the circuit for registering the correction value for correcting the pressure measurement value measured in thepressure sensor circuit707 and the ID of the tirepressure measurement module703. Note that the ID number (for example, 32 bits) is used to confirm the tires of one's own car from those of others and the positions of the tires.
• Thedata transmission circuit715 is the circuit for performing the RF transmission of data such as the measurement value corrected by the computing in the dataprocessing control unit713 to the in-vehicle unit704 shown inFIG. 55. The frequency of the carrier wave used in this RF transmission is UHF band, for example, 315 MHz. In this data transmission, the carrier wave which is ASK modulated or FSK modulated by the transmitted data is transmitted.
• On the other hand, thedata receiver circuit716 is the circuit for performing the RF reception of data such as the control signal from the in-vehicle unit704 and it transmits the data to the dataprocessing control unit713. The frequency of the carrier wave received in thedata receiver circuit716 is LF band, for example, 125 kHz, and the carrier wave which is ASK modulated by the transmitted data is transmitted.
• Antennas717 and718 are connected to thedata transmission circuit715 and thedata receiver circuit716, respectively. Also, as thebattery706 for supplying power to theIC chip705, a coin type lithium battery (voltage 3 V) is used.
• Next,FIG. 57 is a block diagram of the in-vehicle unit704. As shown inFIG. 57, the in-vehicle704 is provided with a dataprocessing control unit719 for performing the data input/output and calculation of the data, adata receiver circuit720, adata transmission circuit721, anantenna722 connected to thedata receiver circuit720, anantenna723 connected to thedata transmission circuit721 and adisplay unit724 for displaying the measurement values and cautions.
• The dataprocessing control unit719 receives the data which is RF transmitted from the tire pressure measurement module703 (703ato703d) shown inFIG. 55 via thedata receiver circuit720, and it displays the pressure measurement values and the caution and the warning of the pressure decrease on thedisplay unit724. Further, the dataprocessing control unit719 transmits the control data to the tire pressure measurement module703 (703ato703d) via thedata transmission circuit721. Note that the power required in the in-vehicle unit704 is supplied from the battery (not shown) mounted in the vehicle.
• According to the seventh embodiment, since the MEMS formed by using the standard CMOS process is used to constitute the tire pressure measurement module703 (703ato703d), it is possible to improve the yield and the reliability of the tire pressure measurement module703 (703ato703d). Therefore, it is possible to improve the reliability of the gas pressure monitoring system for tire.
Eighth Embodiment
• The entire configuration of an anti-skid device for a vehicle using the MEMS will be described with reference toFIG. 58 andFIG. 59. In this case,FIG. 58 is a diagram showing the anti-skid device for a vehicle seen from the bottom surface of an automobile, and it is comprised of avehicle801,tires802ato802dprovided on left front, right front, left rear and right rear of the vehicle, brakingforce control actuators803ato803dwhich control the brake provided to each of thetires802ato802d, and ananti-skid control circuit804 for controlling the brakingforce control actuators803ato803d. Also,FIG. 59 is a block diagram of theanti-skid control circuit804, and theanti-skid control circuit804 has oneIC chip805 and a brakingforce calculation circuit806.
• TheIC chip805 hasacceleration sensor circuits807 and808 for detecting the acceleration in the x and y directions of the coordinate axis shown inFIG. 58 and an anglerate sensor circuit809 for detecting the rotation angle rate around the coordinate axis z. The MEMS for constituting the angle rate sensor and the acceleration sensor are formed in these anglerate sensor circuit809 and theacceleration sensor circuits807 and808. More specifically, in theIC chip805, the MEMS to be the angle rate sensor and the acceleration sensor are formed in addition to the integrated circuit. As the angle rate sensor formed in the anglerate sensor circuit809 and the acceleration sensor formed in theacceleration sensor circuits807 and808 according to the eighth embodiment, for example, the acceleration sensor described in the first and second embodiments and the angle rate sensor described in the fifth and sixth embodiments are used.
• In the MEMS formed in theIC chip805, as described in the foregoing embodiments, the cavity can be formed and sealed through the normal CMOS process. Therefore, the special process for forming and sealing the cavity (packaging process particular to MEMS) which is the major cause of the yield decrease and the manufacturing cost increase in the conventional MEMS manufacturing process becomes unnecessary. As a result, in the eighth embodiment, it is possible to improve the yield, reduce the manufacturing (packaging) cost and improve the reliability of theIC chip805. In addition, since the structure of the MEMS can be formed simultaneously with the formation of the interconnect of the LSI, the integration with the LSI can be facilitated.
• Furthermore, theIC chip805 has analog-digital (A/D)conversion circuits810,811 and812, adata correction circuit813 and amemory circuit814. The analog-digital (A/D)conversion circuits810,811 and812 are the circuits for converting the analog voltage value outputted from the anglerate sensor circuit809 and theacceleration sensor circuits807 and808 into the digital voltage value. Thedata correction circuit813 is the circuit for correcting the difference from the ideal output characteristics of the anglerate sensor circuit809 and theacceleration sensor circuits807 and808, and the coefficient of the correction value is registered in thememory circuit814 in advance.
• The configuration of the anti-skid device according to the eighth embodiment has been described above. Next, the operation thereof will be described below in brief. First, the anglerate sensor circuit809 and theacceleration sensor circuits807 and808 formed on theIC chip805 detect the angle rate and the acceleration applied to the vehicle. In addition, a handling angle (steering angle) is also detected. Furthermore, the vehicle speed and the brake operation amount are also detected from outside. Then, when information such as the angle rate, the acceleration, the handling angle, the vehicle speed and the brake operation amount is inputted to theanti-skid control circuit804, the control signals are outputted from theanti-skid control circuit804 to the braking force control actuator803 (803ato803d) so as to prevent the skid of the vehicle. Then, the braking force of thetires802ato802dis controlled by the braking force control actuator803 (803ato803d). In this manner, it is possible to prevent the skid of the vehicle.
• According to the eighth embodiment, since the MEMS formed by using the standard CMOS process is used to form theIC chip805, it is possible to improve the yield and the reliability of theIC chip805. Therefore, it is possible to improve the reliability of the anti-skid device for vehicle.
Ninth Embodiment
• The entire configuration of an air suspension control unit for vehicle will be described with reference toFIGS. 60 and 61. In this case,FIG. 60 is a block diagram showing the airsuspension control unit901, and it is comprised of oneIC chip902, a spring constant/dampingconstant calculation circuit903 and an actuator for controlling the inner pressure of the air suspension. TheIC chip902 has apressure sensor circuit905, atemperature sensor circuit906, anacceleration sensor circuit907, analog-digital (A/D)conversion circuits908,909 and910, adata correction circuit911 and amemory circuit912. In this case, thepressure sensor circuit905 and thetemperature sensor circuit906 are the circuit for measuring the gas pressure and the temperature in the air suspension, respectively. Also, theacceleration sensor circuit907 is the circuit for detecting the acceleration in the vertical direction from the vehicle so as to detect the movement in the vertical direction of the vehicle mainly due to the bumps on the road. In thepressure sensor circuit905 and theacceleration sensor circuit907, the MEMS constituting the pressure sensor and the acceleration sensor are formed. More specifically, in theIC chip902, in addition to the integrated circuit, the MEMS to be the pressure sensor and the acceleration sensor are formed. As the pressure sensor formed in thepressure sensor circuit905 and the acceleration sensor formed in theacceleration sensor circuit907 according to the ninth embodiment, for example, the pressure sensor and the acceleration sensor described in the first, second and fourth embodiments are used.
• In the MEMS formed in theIC chip902, as described in the foregoing embodiments, the cavity can be formed and sealed through the normal CMOS process. Therefore, the special process for forming and sealing the cavity (packaging process particular to MEMS) which is the major cause of the yield decrease and the manufacturing cost increase in the conventional MEMS manufacturing process becomes unnecessary. As a result, in the ninth embodiment, it is possible to improve the yield, reduce the manufacturing (packaging) cost and improve the reliability of theIC chip902. In addition, since the structure of the MEMS can be formed simultaneously with the formation of the interconnect of the LSI, the integration with the LSI can be facilitated.
• The analog-digital conversion circuits908,909 and910 are the circuits for converting the analog voltage value outputted from thepressure sensor circuit905, thetemperature sensor circuit906 and theacceleration sensor circuit907 into the digital voltage value. Thedata correction circuit911 is the circuit for correcting the difference from the ideal output characteristics of thepressure sensor circuit905 and theacceleration sensor circuit907, and the coefficient of the correction value is registered in thememory circuit912 in advance.
• FIG. 61 is a side view of the vehicle, which shows the configuration of an automobile in which the air suspension is mounted. InFIG. 61, the automobile is provided with avehicle body913,tires914aand914bprovided on right front and right rear of the vehicle (only one side is shown here) andair suspensions915aand915bin which the air suspension control unit901 (901aand901b) is mounted and air springs for suspending thevehicle body913 on thetires914aand914bare used.
• The airsuspension control unit901 has the configuration as described above, and the operation thereof will be described below in brief.
• First, the pressure and the acceleration applied to the vehicle are detected by thepressure sensor circuit905 and theacceleration sensor circuit907 formed on theIC chip902. Then, the vehicle speed and the like are detected from outside. Thereafter, when the airsuspension control unit901 receives information such as the pressure, the acceleration and the vehicle speed, the control signal is outputted to theactuator904 so as to prevent the vibration of the vehicle in the vertical direction. As a result, the spring constant and the damping constant of therespective air suspensions915aand915bare controlled to reduce the vibration of the vehicle in the vertical direction.
• According to the ninth embodiment, since the MEMS formed by using the standard CMOS process is used to form theIC chip902, it is possible to improve the yield and the reliability of theIC chip902. Therefore, it is possible to improve the reliability of the anti-skid device for vehicle.
• In the foregoing, the invention made by the inventors of the present invention has been concretely described based on the embodiments. However, it is needless to say that the present invention is not limited to the foregoing embodiments and various modifications and alterations can be made within the scope of the present invention.
• It is also possible to form the structure of MEMS described in the foregoing embodiments so as to contain any one of a metal film, a silicon-germanium film, a silicon nitride film, a silicon oxide film, a single crystal silicon film, a polysilicon film, an amorphous silicon film and a polyimide film.
• The integrated MEMS according to the present invention can be utilized in, for example, an automobile, a mobile device, an amusement machine, a wireless device, an information appliance, a computer and the like.

Claims (4)

1. An integrated micro electromechanical system, comprising:

a semiconductor substrate;

a plurality of transistors formed on said semiconductor substrate;

an interlayer dielectric formed over said plurality of transistors;

a cavity formed in said interlayer dielectric;

an acceleration sensor formed in said cavity, and having a movable mass, a movable capacitor plate fixed to said movable mass, and a fixed capacitor plate fixed to said interlayer dielectric and opposed to said movable capacitor plate; and

an etching stopper film to which a part of an upper surface has been exposed in bottom of said cavity,

wherein said movable mass, said movable capacitor plate and said fixed capacitor plate are composed of a first interconnect layer,

wherein said etching stopper form is composed of a second interconnect layer between said plurality of transistors and said first interconnect layer, and

wherein said etching stopper film functions as an electric shield between said acceleration sensor and an integrated circuit including said plurality of transistors.

2. An integrated micro electro-mechanical system according toclaim 1, further comprising:

a pad electrically connected to said transistor, composed of said first interconnect layer.

3. An integrated micro electromechanical system according toclaim 1,

wherein said first interconnect layer is thicker than said second interconnect layer.

4. An integrated micro electro-mechanical system according toclaim 2,

wherein said first interconnect layer is thicker than said second interconnect layer.

Images