EP1309835A1

EP1309835A1 - Micromechanical speed sensor and a method for the production thereof

Info

Publication number: EP1309835A1
Application number: EP01980234A
Authority: EP
Inventors: Karin Bauer; Tanjo Gleissner; Konrad Lentner; Stefan Sassen; Josef Schalk; Ralf Voss
Original assignee: EADS Deutschland GmbH
Current assignee: Airbus Defence and Space GmbH
Priority date: 2000-08-18
Filing date: 2001-08-06
Publication date: 2003-05-14
Also published as: US20040011130A1; WO2002014787A1; US6898972B2; DE10040537B4; DE10040537A1; JP2004506883A

Abstract

The invention relates to a micromechanical rotational speed sensor comprising a cardanic structure, capable of oscillation. Said sensor consists of two oscillating elements (4, 5), which are mounted so that they can pivot about two axes (A, B) that are aligned in a perpendicular manner in relation to one another. An excitation unit in the form of an electrode (7) causes the first oscillating element (4) to oscillate about the first rotational axis (A). A readout unit in the form of a readout electrode (8) registers a tilting or oscillation of the second oscillating element (5) about the second rotational axis (B) and uses said movement as a measure for the rotational speed of the sensor. Additional elements of mass (6a, 6b), which are symmetrically aligned, are located on the upper face (2a) and the lower face (2b) of the first oscillating element (4), said element forming a rocker. The sensor is produced from at least three individually machined wafers, which are subsequently combined to form a cover section (1), a central section (2) and a base section (3).

Description

Micromechanical rotation rate sensor and method for its production

The present invention relates to a micromechanical rotation rate sensor according to the preamble of claim 1, and a method for producing a micromechanical rotation rate sensor.

Gimbal-suspended, micromechanical rotation rate sensors have e.g. two oscillating frames with a central inertial mass. The mass is brought to resonant vibrations around an excitation axis by an electrostatic drive. When the sensor rotates about an axis of rotation that is perpendicular to the excitation axis, the Coriolis force acts on the oscillating inertial mass. As a result, an oscillation is periodically excited about a readout axis, which is directed perpendicular to the excitation axis and to the axis of rotation. The amplitude of the oscillation thus generated is a direct measure of the yaw rate to be measured. The reading of the amplitude can e.g. electrostatically.

Such micromechanical rotation rate sensors can e.g. in automotive engineering, in aerospace engineering, and in exploration and production processes. For example, the rotation rate sensors can be used for vehicle stabilization, for driving dynamics control and for navigation systems or also within systems for autonomous driving. There are other possible uses in missile navigation and stabilization. In the field of space travel, platforms of this type can be stabilized and their position regulated. Navigation systems, e.g. GPS / INS are supplemented by such rotation rate sensors, especially in the field of avionics. When exploring raw materials, e.g. Drill heads can be controlled with rotation rate sensors. In modern production technologies, rotation rate sensors are used to control robots.

US Pat. No. 4,598,585 describes a yaw rate sensor with a gimbal structure in which a frame is mounted such that it can vibrate about a y-axis. Inside the frame is an element that is vibratable about an x-axis on the Frame is attached. An inertial mass is arranged on the inner element. Drive elements serve to set the frame in vibration around the y-axis. The deflection of the inner element due to the Coriolis force is measured capacitively.

However, the known micromechanical rotation rate sensors have the disadvantage that the measurement accuracy is often inadequate. Very large cross-sensitivities usually also occur. In addition, there is often a high sensitivity to vibrations. In addition, the known yaw rate sensors are usually associated with high manufacturing costs.

The object of the invention is therefore to create a micromechanical rotation rate sensor and to provide a method for its production which has a high sensor sensitivity and low cross sensitivity. In addition, the rotation rate sensor should be robust, have low sensitivity to mechanical vibrations and be inexpensive to manufacture.

This object is achieved by the micromechanical yaw rate sensor according to claim 1 and the method for producing a micromechanical yaw rate sensor according to claim 22. Further advantageous features, details and aspects of the invention are evident from the dependent claims, the description and the drawings.

The micromechanical rotation rate sensor according to the invention comprises a first oscillating element, which is pivotably mounted about a first axis, a second

Vibrating element, which is pivotally mounted about a second axis, which is directed perpendicular to the first axis, an excitation unit to set the first vibrating element in vibration about the first axis and a readout unit for detecting vibrations of the second vibrating element about the second axis, wherein at least two additional mass elements are attached to the first oscillating element, which are aligned symmetrically to a plane which is defined by the first and second axes. The symmetrically aligned, additional mass elements result in a significantly higher sensor resolution and sensitivity. The additional masses or additional mass elements can be made extremely large. This results in a wide outsourcing of the center of gravity symmetrically to the axis of rotation, which causes an extreme increase in sensor sensitivity. The symmetrical structure reduces the cross-sensitivity to yaw rates outside the sensitivity axis of the sensor and reduces the sensitivity to acceleration acting on the sensor. The sensor is inexpensive to manufacture and can be designed to be extremely robust.

The common center of gravity of the two mass elements advantageously lies at the intersection of the first and second axes. This results in maximum symmetry.

The additional mass elements are preferably manufactured separately from the first and / or second oscillating element, in particular the shape, size or material of the mass elements being specifically selected in order to determine the parameters of the sensor. Through the free choice of shape, size and material of the additional masses, the mass distribution, the total mass and the distribution of the moments of inertia of the sensor can be specifically selected. This results in additional configuration options for optimizing the sensor with regard to resolution, cross-sensitivity, shock sensitivity, reduced influence of manufacturing tolerances, or reduced sensitivity to vibrations.

In particular, the free choice of materials for the additional mass elements allows additional masses to be formed with special physical properties that are particularly suitable depending on the requirements of the rotation rate sensor. Furthermore, the sensor element can be trimmed by a special choice of the additional masses, without changes or effects having to be made to the other structures or to the etched-out gimbal structure. The additional masses can be manufactured inexpensively with high precision. Balls are particularly preferably used as mass elements, which can be produced inexpensively with a very low geometry tolerance of, for example, 0.1%. The use of balls results in a very high reproducibility of the mass distribution of the rotation rate sensor. Cuboids, cones, pyramids or truncated pyramids and cylinders can also be used as mass elements, which can also be produced very inexpensively and with a low geometry tolerance. It is particularly favorable to arrange the cones or pyramids with their tips aligned with one another. As a result, the focal points of the individual mass elements are outsourced as far as possible or as far apart as possible.

In particular, the additional mass elements e.g. magnetic properties. This causes a mutual attraction of the additional mass elements so that they align themselves completely symmetrically. Further advantages are the resulting adhesion to the substrate, the possibility of self-calibration, and the

Possibility of magnetic or electromagnetic excitation to vibrate.

The additional mass elements are preferably made of a material that has a higher density than the material of the first and / or the second oscillating element. This leads to a more favorable distribution of the moments of inertia. Here, e.g. Metals, in particular steel, are used as material for the additional mass elements, whereas where for the rest of the sensor structure or for the first and second vibrating elements, e.g. Silicon is used. There is therefore a free choice of material for the additional masses, since the material of the additional masses is not compatible with the processing steps e.g. must be for a silicon wafer from which the oscillating elements or the oscillatable structure is advantageously produced. This can result in an extreme increase in sensor sensitivity in a very cost-effective manner.

For example, the first oscillating element is a rocker and the second oscillating element is a frame, the rocker and the frame forming a cardanic, oscillatable structure, which is fastened in a holding structure. The rotation rate sensor is advantageously produced from at least three joined wafers, which are preferably processed individually. The rotation rate sensor has, for example, a bottom wafer, a middle part wafer and a lid wafer. This results in a reduced complexity in the manufacturing process, as well as the possibility to test the individual components. Furthermore, the yield is increased, which results in reduced costs for the sensors. Furthermore, pit and electrode structures that are inside the sensor after being assembled can be freely designed. In addition, there is a greater choice for electrode material and for the material of the additional masses, since the material has to be compatible with fewer manufacturing steps. The use of identical top and bottom wafers is possible. This enables a structure that is completely symmetrical to the central plane.

The first and the second oscillating element are preferably formed in the central part wafer. In particular, the middle part wafer can be processed on the top and bottom. This ensures symmetry with the center plane, since the masses or additional mass elements can be attached symmetrically. The temperature drift of the sensor properties is reduced by the symmetry.

The bottom and / or the lid wafer is advantageously made of alkali-containing glass wafers, such as, for example, borofloat or pyrex glass, of which, for example, at least one wafer is provided with an electrode structure. Scattering and crosstalk capacities are thereby reduced, since the electrode structure is located on insulating material. In particular, for example, the thermal expansion coefficient is adapted to the silicon of the middle part wafer, which is why the thermal stresses during production can be kept low and which results in a reduced temperature sensitivity of the sensor during operation. The use of alkali-containing glass wafers also enables a reliable connection to the middle part wafer made of silicon by means of an anodic bonding process. The fact that the wafers or the middle part wafers are connected to the bottom wafer and the lid wafer, for example by anodic bonding, results in a reliable connection which requires a maximum temperature of 450 ° C. for the production. This maximum temperature is low enough so that suitably chosen metallizations are not changed, ie there is no oxidation and no formation of alloys. The anodic bonding allows the wafers to be well aligned with one another since no liquid phase occurs during the bonding process. The adjustment tolerance of the rotation rate sensor is therefore usually less than a few μm.

Advantageously, there is a gap distance between the middle part wafer and the bottom wafer or between the middle part wafer and the lid wafer, which is small in relation to the lateral electrode extension. This distance between adjoining wafers is used for electrostatic excitation and / or for capacitive readout of the actuator and / or sensor oscillation of the oscillating elements. The relationship between the gap distance and the lateral electrode extension is e.g. less than 1:20, preferably less than 1:50, and particularly preferably less than 1: 100 or even 1: 1000. This results in very large capacitance values, which in turn produce high electrical signals for the sensors or large electrostatic forces for the actuators enable.

The gap distance for the actuator structure, which enables the excitation vibration of the first vibration element, is preferably greater than the gap distance for the sensor structure, which enables the read-out vibration of the second vibration element. As a result, the actuator oscillation can take place with a very high mechanical amplitude. In addition, the damping of the vibration is lower with a larger gap distance (squeezed film damping), which leads to a higher mechanical amplitude with resonant excitation. On the other hand, the small gap distance in the sensor structure results in a large capacitance and thus a high electrical output signal.

The wafer from which the mechanical structure or the first and second oscillating element is etched is advantageously made from single-crystal silicon. The vibratable structure or gimbal structure of the sensor is made, for example, of a full wafer etched, ie manufactured in bulk technology. The structure capable of oscillation comprises, for example, the first and second oscillation elements and is preferably structured out of the middle part wafer. The use of monocrystalline silicon results in a very low material damping and, furthermore, negligible signs of fatigue and aging. The production in silicon technology leads to low

Manufacturing tolerances at low costs. In addition, silicon has a high mechanical strength with low density, which results in a robust and resilient mechanical structure.

Advantageously, the first and / or the second oscillating element is non-rectangular, i.e. the structure capable of vibrating has a non-right-angled geometry or a symmetrical convex free form. The vibrating elements can e.g. be round or have edges that adjoin each other at an angle of more than 90 °. For example, the vibrating elements can be 8-cornered.

In particular, taking into account large additional inertial masses that cause the sensitivity to be increased, e.g. an enlargement of the capacitance areas with a higher bending stiffness, and thus higher natural frequencies of the frame or the outer vibratable structure.

This achieves a high degree of rigidity for the clamping of the hinge of the frame or the torsion suspension, which cannot be achieved with a rectangular structure. The torsion frequency is essentially determined by the torsion or rotating band as the suspension itself. The rotary band can thus be shortened considerably and a Z-mode of the sensor that can be set almost independently of the torsion frequency and is directed perpendicular to the wafer plane can be achieved.

Due to the geometry described above and the special arrangement of the additional masses, a particularly high sensitivity with a small design can be achieved for a given area of the vibratable structure.

Furthermore, the natural frequency spectrum of the mechanical structure cannot be determined by the Right-angled geometry can be designed more cheaply. Non-right-angled geometries can be found in which the torsional natural frequencies of the rocker or the inner vibratable structure and the frame or the outer vibratable structure are the lowest eigenmodes of the structure and all other modes come to lie at significantly higher frequencies. This enables the required frequency spacing between the mechanical interference spectrum, for example in a harsh environment, and the operating and eigenmodes of the sensor to be guaranteed.

Advantageously, the frequency of the actuator vibration caused by the excitation unit and / or the frequency of the sensor vibration caused by the

Coriolis force is generated, the lowest eigenmodes of the vibratable structure, which is formed by the first and the second vibrating element. This results in a high robustness of the mechanical structure against shock loads and mechanical vibrations.

By using mechanical damping elements, for example mechanical low-pass filters, when building the sensor element, it is possible to separate the rotation rate signal from higher-frequency interference signals. The rotation rate signal has a bandwidth of 0 to 100 Hz, for example. Low-frequency interference signals, the bandwidth of which is comparable to the bandwidth of the rotation rate, cannot or only very strongly suppresses the sensor behavior due to the position of the natural frequencies of the sensor structure. The natural frequencies of the sensor structure, i.e. the actuator and sensor vibration, at approx. 10 kHz, while all other eigenmodes are above it.

In particular, the area ratio between the second vibrating element and the first vibrating element is greater than 5: 1, preferably greater than 10: 1. This area ratio between the frame, which forms the outer sensor structure or the second vibrating element, and the rocker, which forms the inner actuator structure or the first vibrating element, results in a further increase in the electrical capacity

Sensor signal with an optimal mechanical design of the sensor structure or the position of the eigenmodes. Furthermore, there is an extensive decoupling of the Natural frequencies of rocker and frame. This area ratio and the associated ratio of the moments of inertia make it possible to determine the natural frequency of the rocker essentially as a function of the inertial mass or the mass elements, and to determine the natural frequency of the frame essentially as a function of the frame geometry. An almost independent frequency adjustment can thus be achieved, ie the sensor can be trimmed in a simple manner and with high accuracy.

The micromechanical rotation rate sensor preferably has a metallization to form an electrode or electrode structure which is covered with a dielectric layer. This results in passivation, so that the metallization is protected against corrosion. Leakage currents between the insulated electrodes are significantly reduced. Since the metallization in particular only sits on fixed, stationary parts of the sensor, there are hardly any restrictions with regard to the type and the method of passivation.

The micromechanical rotation rate sensor advantageously comprises one or more electrodes, which are surrounded by a closed conductor track. For example, the conductor track can be contacted specifically. The corresponding supply lines can also be surrounded by the closed conductor track. This measure reduces the electrical crosstalk between the electrodes for the sensors and / or for the actuators. Since the metallized electrodes e.g. only in the lid and / or bottom wafer and not in this case on the structured middle part, the Guardel electrodes are easy to contact and have less constraints in terms of their geometry and properties than if they had to be attached to the middle part.

The micromechanical rotation rate sensor preferably has an ohmic pressure contact for connecting the middle part wafer to the bottom wafer or to a bond pad of the bottom wafer. The lid wafer can also be contacted in this way. There is preferably no metallization on the middle part wafer itself. In particular, the entire structure of the middle part wafer has an electrical potential. This makes it possible for the middle part wafer to be electrically connected via standardized wire bond pads which, for example, have a size of 100 μm × 100 μm. As a result, the entire connection pads can be located on one level and arranged side by side. This considerably reduces the effort involved in electrically contacting the sensor element with the associated electronics for actuators and sensors.

Due to the lack of metallization on the middle part wafer, the manufacturing effort of the middle part wafer is significantly reduced. Furthermore, the structure capable of oscillation shows only very low material damping and has no mechanical tension. This also contributes to a reduced temperature dependence of the sensor properties.

The sensor interior can be hermetically sealed, e.g. buried conductor tracks for contacting the electrodes in the sensor interior. The connection between the bond pads and the electrode surfaces in the interior of the sensor by means of buried conductor tracks means that the interior of the sensor can be hermetically sealed and can therefore neither become dirty nor corrode or be changed by moisture or other environmental influences.

The method according to the invention for producing a micromechanical yaw rate sensor comprises the steps: providing at least three wafers; Structuring the individual wafers, a cardanic, oscillatable structure being formed in one of the wafers; Forming an excitation unit to excite a first vibration of the structure; Forming a readout unit for detecting a second vibration of the structure, which is perpendicular to the first vibration; and joining the wafers, wherein the wafer with the oscillatable structure is connected to a further wafer on both sides. This method makes it possible to arrange extremely large symmetrical additional masses on the oscillatable middle section and thereby achieve a significantly higher sensor resolution. The further advantages that result from the production of the rotation rate sensor from at least three individually processed wafers, which are finally joined, have already been described above in connection with the rotation rate sensor according to the invention.

In particular, additional mass elements can be attached to the vibratable structure symmetrically to the axis of the first and / or the second vibration. Furthermore, the wafer with the vibratable structure can be processed on its top and bottom.

Advantageously, the vibratable structure or gimbal structure of the sensor is etched from a single, full wafer, and the suspension of the mechanical or vibratable structure, which forms the middle part of the sensor, is produced in a single etching step. This achieves a high degree of manufacturing accuracy of the geometric structure or structure capable of vibrating, since e.g. there is no need to adjust several masks to each other. It is also possible to carry out the structuring of the silicon with flanks perpendicular to the wafer plane. The middle part itself is therefore symmetrical with high accuracy top-bottom. This excludes an essential source for the quadrature error. There is also a free design of the lateral geometry, e.g. when using the anisotropic plasma etching technique.

Advantageously, a metallization structure is applied to the bottom wafer of the three wafers by means of thin-film technology, which forms, for example, capacitor areas, supply lines and connection pads. This reduces the manufacturing effort, since the complete metallization is only on one wafer. Thin-film technology enables the production of small structures with reproducible thicknesses, which are necessary for a reproducible gap distance. For example, the conductor track width is 10 μm, the conductor track and electrode thickness is 140 nm and the gap distance is, for example, 1.5 μm. The electrical connection of the sensor element to the actuator and sensor electronics takes place, for example, via standardized wire bond pads. Their size is, for example, 100μm x 100μm. The rotation rate sensor or the first oscillation element can be excited in a variety of ways, for example electrostatically, piezoelectrically, magnetostrictively or also magnetically or using additional magnetic masses. In this case, the rotation rate sensor is provided with electrostatic, piezoelectric, magnetostrictive or also magnetic elements or additional magnetic masses.

In addition, a control device can be provided which has electronics for regulating and / or forcing the excitation oscillation. The electronics can be designed by appropriate circuits so that the first oscillating element oscillates in its iron frequency. However, it can also be designed in such a way that the oscillation of the first oscillating element is forced at a specific frequency, which does not have to be the natural frequency. This has the particular advantage that the readout electrodes for measuring the vibration of the second vibrating element do not impair the function of the sensor or falsify the measurement results, even if they are arranged very close to the electrodes for exciting the first vibrating element and / or a free vibration of the would affect the first.

By choosing a suitable electrode shape of the excitation electrodes, the influence of the readout electrodes for measuring the vibration of the second vibrating element is further minimized. For example, the individual electrodes of a pair of excitation electrodes can be divided and controlled separately by the electronics in order to switch off or compensate for the above-mentioned influence.

The readout method can also be implemented in several known ways and in particular e.g. capacitively or optically. In this case, the rotation rate sensor is provided with capacitances or optical elements for reading out the vibration of the second vibration element generated by the Coriolis force.

Continuous self-test and self-calibration functions are possible during operation. In the following, the rotation rate sensor according to the invention is described by way of example with reference to the figures. Show it:

1 shows a section through a micromechanical rotation rate sensor according to a preferred embodiment of the invention;

FIG. 2 shows a plan view of the middle part of the rotation rate sensor according to the preferred embodiment;

3 shows a plan view of an electrode structure of the rotation rate sensor; and

Fig. 4 is a section showing the edge region of the sensor with the connection between the middle part and the bottom part.

FIG. 1 shows a micromechanical rotation rate sensor 10, which is formed from a cover part 1, a middle part 2 and a bottom part 3. These parts are individually processed wafers that were finally assembled. The middle part 2 forms an oscillatable structure with a first oscillating element 4 and with a second oscillating element 5. The first oscillating element 4 forms a rocker and the second oscillating element 5 forms a frame in which the rocker can be pivoted about a first

Axis of rotation A is mounted. The frame or the second oscillating element 5 is mounted within the sensor so as to be pivotable about a second axis of rotation B, which runs perpendicular to the first axis of rotation A in the wafer plane. There is an additional mass element 6a, 6b on the top 2a and on the bottom 2b of the middle part 2 and the first oscillating element 4, respectively. The additional mass elements 6a, 6b are arranged symmetrically to the central part 2 or to the central part plane, which is formed by the axes of rotation A and B. In the preferred embodiment shown, the additional mass elements 6a and 6b are arranged symmetrically to the axis of rotation A and symmetrically to the axis of rotation B, ie there is symmetry to the two axes of rotation A and B and to the intersection of the two axes of rotation A and B. Electrodes 7 together with a controller (not shown) form an excitation unit in order to set the first vibrating element 4 or the rocker in vibrations about the axis of rotation A. Further electrodes 8 together with electronics (not shown) form a readout unit in order to detect the vibrations of the second vibrating element 5 or the frame about the axis of rotation B. The vibrations are excited and read out electrostatically or capacitively. For this purpose, the middle part 2 has an electrical potential.

In the cover part 1 and in the base part 3, recesses 1a and 3a are provided, into each of which a mass element 6a and 6b protrudes. There is between the

Mass elements 6a, 6b and the respectively adjacent structure of the cover part 1 or bottom part 3 a clearance or distance that allows the mass element 6a, 6b connected to the rocker to swing back and forth within the recess 1a or 3a.

There is a gap distance d between the cover part 1 and the middle part 2, and there is also a gap distance el, e2 between the middle part 2 and the bottom part 3. The gap distances d, e1, e2 serve for electrostatic excitation or for the capacitive readout of the actuator and sensor vibration of the oscillatable structure of the middle part 2. The gap distances d, e 1, e2 are in relation to the lateral extension of the electrode structure or to the lateral extension of the Electrodes 8, which are used to read the vibration of the frame or second vibrating element 5, are very small. The ratio of the gap distance d and the gap distance e2 to the lateral extension of the electrode 8 is approximately 1: 100 or less. This results in very high capacitance values for the sensors or large electrostatic forces for the actuators.

The gap distance e 1 between the base part 3 and the first vibrating element 4, which enables the first vibrating element 4 to be tilted or pivoted about the axis of rotation A, is greater than the gap distance e2 between the bottom part 3 and the second vibrating element or frame 5, the the tilting around the axis of rotation B enables. This enables a large mechanical amplitude for resonant excitation, while on the other hand a high electrical output signal is achieved due to the small gap distance in the reading and the associated large capacity.

The lid part 1 has at its edge a projection 11 through which it is firmly connected to the edge 21 of the middle part. By the projection 1 1 or the thereby formed, raised compared to the central region of the lid part, an interior 9 is formed within the sensor 10, the two vibrating elements 4, 5 and the vibratable structure enough space to perform the excitation or. Readout vibration offers. The bottom part 3 also has on its surface a protruding connection area or area 31 which is used to connect the

Bottom part 3 serves to the middle part 2 and thereby offers space for the vibrations.

The base part 3, which can be a base wafer or part of a base wafer, has a greater lateral extent than the other wafer parts or wafers which form the cover part 1 and the middle part 2. In other words, the bottom part 3 has an edge region which extends beyond the edge of the middle part 2 or the cover part 1. Contact surfaces 32 in the form of connection pads are provided on the surface of the base part 3 or of the wafer in the edge region, which serve to contact the metallizations or electrodes 7, 8 in the interior 9 of the sensor. The contact surfaces 32 are connected to the electrodes 7, 8 via conductor tracks 33, the contact surfaces 32, the conductor tracks 33 and the electrodes 7, 8 being formed in one plane on the surface of the base part 3 or lower wafer. The conductor tracks 33 are buried conductor tracks, i.e. they are integrated or incorporated into the wafer. This results in a hermetic or vacuum-tight closure of the interior 9.

FIG. 2 shows a top view of the middle part 2, which can be part of a wafer or also an entire wafer, of the rotation rate sensor according to the preferred embodiment. The inner vibrating element 4 or the rocker is connected to the outer vibrating element 5 or frame by means of two opposite, oscillatory or torsional suspensions 41. The oscillating suspension 41 allows the rocker to tilt or swing about the axis A, which extends through the two suspensions 41. The frame or the outer vibrating element 5 is u

connected by oscillatory or torsional suspensions 42 to the remaining part of the wafer or middle part 2, which forms a holding structure 21 which is fixed between the lid part 1 and the bottom part 3 of the wafer. The vibration-capable suspension 42 of the frame or second vibration element 5 on the holding structure 21 allows the frame to tilt or swing about the axis B, which extends through the two suspensions 42 and is oriented perpendicular to the axis of rotation A of the first vibration element 4.

The upper additional mass element 6a is fastened symmetrically in the center of the vibratable structure formed from the rocker and frame and at the same time in

Center of the rocker or the first vibrating element 4 arranged. The second additional, identically designed mass element 6b is arranged directly below (see FIG. 1).

The vibrating elements 4, 5 have edges 4a, 5a, which are not aligned at right angles to one another, but form an angle α which is greater than 90 ° C. In other words, the structure capable of oscillation, which consists of the two oscillation elements 4 and 5, has a non-right-angled geometry, by means of which an enlargement of the capacitance areas is achieved with a higher flexural rigidity and thus higher natural frequencies of the frame. The suspension 42 or the hinge of the frame can also be greatly shortened. Overall, the natural frequency spectrum of the mechanical structure can be made more economical due to the non-rectangular geometry. The other advantages of the non-rectangular geometry have already been described above.

FIG. 3 shows a top view of a metallization formed on the base part 3. The metallization forms the two flat electrodes 8, which are used for capacitive reading of the vibration of the frame or second vibrating element 5, which is generated due to the Coriolis force when the sensor rotates about a sensitive axis directed perpendicular to the axes of rotation of the vibrating elements 4 and 5 , The electrodes 7 are not shown here, but are configured similarly.

The entire structure of the middle part wafer or middle part 2 has an electrical potential which corresponds to the metallizations or electrodes 7, 8 on the bottom part 2 opposite. As a result, it is not necessary to also apply metallizations to the vibratable structure, which are opposite the electrodes 7, 8 for excitation and for reading (see FIG. 1).

Each electrode 7, 8 is completely surrounded by a ring electrode 12, which encloses both the electrode 7, 8 and the conductor track 33 leading to the outside and the external contact surface 32. The ring electrode 12 can be contacted separately via its own contact surface 34 located outside the sensor interior. The ring electrode 12, which forms a conductor track, reduces the electrical crosstalk between the electrodes 7, 8 for sensors and actuators.

Figure 4 shows the connection between the middle part 2 and the bottom part 3 in the edge region of the sensor 10 in an enlarged view. The middle part wafer or middle part 2 is connected via an ohmic pressure contact 35 to a bond or connection pad of the bottom wafer 3, which is designed in the form of a contact surface 36 on the bottom part 3. The entire connection pads are on one level and are arranged side by side. As a result, both the electrodes 7, 8 and the central part wafer 2 can be electrically contacted in a simple manner and with significantly reduced effort.

In the preferred embodiment, the middle part wafer 2 consists of single crystal

Silicon, while the bottom part 3 and the lid part 2 e.g. from alkali-containing glass wafers, e.g. Borofloat or Pyrex glass. At least one of the wafers is provided with an electrode structure. Of course, other materials are also possible for the sensor parts, the choice of material depending on the respective requirements.

In the present case, the mass elements 6a, 6b are steel balls, which are each supported in a bulge on the top and bottom of the rocker or the first oscillating element 4. The steel balls face each other exactly, so that a high geometry is guaranteed. In a special embodiment, magnetic steel balls are used which align themselves with one another. In addition to steel balls, other shapes and materials are of course also used to design the Mass elements 6a, 6b possible, the sensor parameters being able to be set by suitable selection.

In addition to an electrostatic excitation unit, a wide variety of excitation methods are possible, such as piezoelectric, magnetostrictive, or magnetic excitation methods. The readout method can also be carried out in other known ways, wherein optical readout methods are also possible in addition to capacitive readings.

An example of the manufacture of the sensor 10 is described below.

To produce the sensor, three wafers are processed individually, which form the cover part 1, the middle part 2 and the bottom part 3. The wafers for the cover part 1 and for the base part 3 are structured in such a way that there are recesses in their center for the mass elements 6a, 6b. are, with enough scope to carry out the vibrations. A central area of the respective wafer surface is also lowered relative to the edge area, so that in this area of the cover and bottom part there is a gap distance to the central part 2, which enables the oscillatable structure of the central part to oscillate.

The middle part wafer is processed on the top and bottom so that the symmetry to the middle plane is guaranteed. The vibratable structure of the sensor, consisting of frame and rocker, is etched from a full wafer, the suspensions 41, 42 (see FIG. 2) being produced in one etching step.

Now the mass elements 6a, 6b are attached to the top and bottom of the wafer, which is provided as the middle part 2, for example by gluing or magnetically.

On the wafer, which is provided for the base part 3, a metallization structure is applied by means of thin-film technology, which forms the electrodes or capacitor surfaces as well as the leads and connection pads. The metallization for passivation is covered with a dielectric layer. After the three wafers or sensor parts have been prefabricated separately, they are joined together and firmly connected to one another. Suitable precautions are taken to ensure that there is a vacuum in the interior 9 of the sensor.

The symmetrical sensor effectively avoids sources of error, such as temperature drift in particular, and results in improved measurement results. The symmetrical additional masses on the middle part 2 result in a significantly higher sensor resolution. The highly symmetrical mechanical structure of the sensor leads to high long-term stability and low offset drift. This means that the sensor works stably over the long term and delivers more precise measurement results. It can be subjected to mechanical loads without the measurement results being falsified under such loads.

Claims

claims

1. Micromechanical rotation rate sensor, with a first oscillating element (4) which is pivotally mounted about a first axis (A), a second oscillating element (5) which is pivotably mounted about a second axis (B), which is perpendicular to the first axis ( A) is directed, an excitation unit (7) to set the first vibrating element (6) in vibrations about the first axis (A), and a readout unit (8) for detecting vibrations of the second

Vibrating element (5) about the second axis (B), characterized in that at least two additional mass elements (6a, 6b) are attached to the first vibrating element (4) and are aligned symmetrically to a plane through the first and second axes (A, B) is defined.

2. Micromechanical rotation rate sensor according to claim 1, characterized in that the additional mass elements (6a, 6b) are made separately from the first and second vibrating elements (4, 5), the shape, size and / or

Material of the mass elements (6a, 6b) are selected to determine and / or significantly influence sensor parameters.

3. Micromechanical rotation rate sensor according to claim 1 or 2, characterized in that the mass elements (6a, 6b) are axially symmetrical bodies which are attached to the top (2a) and the bottom (2b) of the first oscillating element (4), the mass elements are preferably spheres, cuboids, cylinders, cones, pyramids or prisms.

4. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that the first vibrating element (4) is a rocker and the second vibrating element (5) is a frame, the rocker and the frame being a Form a cardanic, oscillatable structure, which is fastened in a holding structure (21).

5. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that the additional mass elements (6a, ob) are magnetic.

6. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that the additional mass elements (6a, 6b) are made of a material which has a higher density than the material of the first and second oscillating elements (4, 5).

7. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that it is made of at least three joined wafers, which form a bottom part (3), a middle part (2) and a cover part (1).

8. Micromechanical rotation rate sensor according to claim 7, characterized in that the base part (3) and / or the cover part (1) are made of alkali-containing glass wafers, at least one of the wafers being provided with an electrode structure (7, 8).

9. Micromechanical rotation rate sensor according to claim 7 or 8, characterized in that the first and the second oscillating element (4, 5) are formed in the central part (2), the top and bottom of which is machined.

10. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that a gap distance (d, e1, e2) between adjacent wafer parts or wafers, which enables deflection of the first and / or second oscillating element (4, 5), is small compared to the lateral one Expansion of excitation and / or readout electrodes (7, 8), the ratio between the gap distance and lateral electrode expansion being less than 1:10, preferably less than 1:50, particularly preferably less than 1: 100 or 1: 1000.

1 1. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that a gap distance (e1) for an excitation oscillation of the first oscillating element (4) is greater than a gap distance (e2) for one

Read vibration of the second vibrating element (5).

12. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that it is made of single-crystal silicon, the first and the second oscillating element (4, 5) being made from a single wafer.

13. ^' Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that the first and / or second oscillating element (4, 5) is non-rectangular.

14. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that the first and / or second oscillating element (4, 5) has edges (4a, 5a) which adjoin one another at an angle of more than 90 degrees.

15. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that the frequency of an actuator oscillation and the frequency of the sensor oscillation are the lowest eigenmodes of the first and / or second oscillation element (4, 5).

16. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that the area ratio between the second vibrating element (5) and the first vibrating element (4) is greater than 5: 1, preferably greater than 10: 1.

17. Micromechanical rotation rate sensor according to one of the preceding claims, characterized by a metallization to form an electrode (7, 8) which is covered with a dielectric layer.

18. Micromechanical rotation rate sensor according to one of the preceding claims, characterized by one or more electrodes (7, 8) which is surrounded by a closed conductor track (12).

19. Micromechanical rotation rate sensor according to one of the preceding claims, characterized by an ohmic pressure contact (35) for connecting a

Wafers, which forms a middle part (2), to a further wafer, which forms a bottom or cover part (1, 3).

20. Micromechanical rotation rate sensor according to one of the preceding claims, characterized by a sensor interior (9) which is hermetically sealed, wherein buried conductor tracks (33) are provided for contacting electrodes (7, 8) in the sensor interior (9).

21. Micromechanical rotation rate sensor according to one of the preceding claims, characterized in that the common center of gravity of the mass elements

(6a, 6b) lie at the intersection of the first axis (A) and the second axis (B).

22. A method for producing a micromechanical rotation rate sensor, characterized by the steps: providing at least 3 wafers (1, 2, 3);

Structuring of the individual wafers (1, 2, 3), a cardanic, oscillatable structure (4, 5) being formed in one of the wafers (2);

Forming an excitation unit (7) for exciting a first vibration of the

Structure; Forming a readout unit (8) for detecting a second vibration of the

Structure perpendicular to the first vibration; and joining the wafers (1, 2, 3), wherein ^'the wafer (2) capable of vibrating with the Structure (4, 5) is connected on both sides to a further wafer (1, 3).

23. The method according to claim 22, characterized in that on the vibratable structure (4, 5) additional mass elements (6a, 6b) are attached symmetrically to the axis (A, B) of the first and / or second vibration.

24. The method according to claim 22 or 23, characterized in that the wafer (2) with the vibratable structure (4, 5) is processed on its top and bottom.

25. The method according to any one of claims 22 to 24, characterized in that the oscillatable structure (4, 5) is etched from a single, full wafer (2), with suspensions (41, 42) of the oscillatable structure (4, 5) can be produced in a single etching step.