Patent Application
20200156930
The present application claims the benefit under 35 U.S.C. § 119 of German Patent Application No. 102018219546.3 filed on Nov. 15, 2018, which is expressly incorporated herein by reference in its entirety.
The present invention relates to a micromechanical component. The present invention further relates to a method for producing a micromechanical component.
Micromechanical components, e.g., inertial sensors for measuring acceleration and rate of rotation, are manufactured in mass production for various applications in the automotive and consumer sectors. Rocker structures are preferably used for capacitive acceleration sensors having a detection direction perpendicular to the wafer plane (i.e., in the z-direction). The sensor principle of these rockers is based on a spring-mass system in which in the simplest case a movable seismic mass having two counter electrodes fixed on a substrate forms two plate capacitors. The seismic mass is connected to the base via at least one, for reasons of symmetry usually two torsion springs. If the mass structures on the two sides of the torsion spring are of different size, then a z-acceleration action will induce the mass structure to rotate relative to the torsion spring as axis of rotation. The distance of the electrodes on the side having the greater mass therefore becomes smaller and greater on the other side. The change in the capacitance is a measure for the acting acceleration. Such acceleration sensors are described, for example, in European Patent Application Nos. from EP 0 244 581 A1 and EP 0 773 443 A1.
Various methods have been proposed for compensating for the influence of surface potentials on acceleration sensors, e.g., in German Patent No. DE 103 50536 B3, German Patent Application No. DE 10 2006 057 929 A1, and German Patent Application No. DE 10 2008 040 567 A1. All of the proposals described therein have in common that the problem of the offset drifts is to be solved via special measures and provisions on the circuit side and/or by special test methods. Such measures are very laborious, however, and thus result in significant additional costs of the components.
Some years ago, novel z-sensor designs and technologies were proposed, in German Patent Application No. DE 10 2009 000167 A1 for example, in order, among other things, to improve the parasitic effects due to electrical surface potentials without intervention on the circuit side. German Patent Application No. DE 10 2009 000167 A1 describes a substantially improved robustness vis-a-vis surface potentials and their drifts, since the lower side of the movable structure, which is formed by the second functional layer, was electrically symmetrized vis-a-vis the conductor track plane. The mass asymmetry required for the mechanical sensitivity is here achieved via a third functional layer.
Even these greatly improved structures, however, are in turn sensitive to surface potentials if the upper side of the movable seismic mass in the third functional layer 30 is faced by another electrically conductive plane having parasitic capacitances and resulting parasitic forces, as shown in
While in the design of
German Patent Application No. DE 10 2016 207 650 A1 describes a defined electrical partitioning of electrode surfaces on the cap wafer or in the first functional layer in the area of the additional mass in order to minimize the effects of charge drifts.
A further problem with respect to the boundary surfaces of an asymmetrical rocker design are possible radiometric effects, which may occur at in the event of rapid temperature changes. In such temperature changes, the temperatures of the rocker and the substrate are not in thermal equilibrium, but rather there are temperature gradients perpendicular to the substrate layer, it being possible that e.g. the substrate with the bottom electrodes in the first functional layer is somewhat warmer than the rocker structure in the third functional layer. The thermal gradients induce movements of the gas particles in the sensor cavity, the impacts of which with the movable sensor structure may result in measurable parasitic deflections of the rocker and thus result in offset signals. This effect is described in C. Nagel et al., “Radiometric effects in MEMS accelerometers”, IEEE Sensors 2017, Glasgow, Scotland.
The designs of the sensors of
German Patent Application Nos. DE 10 2009 000 345 A1 and DE 10 2010 038 461 A1 describe rotation-rate sensors having trough-shaped or partially concave sensor masses in order on the one hand to produce top electrodes in the third functional layer and on the other hand to allow for masses that have a light construction, which may offer advantages with respect to their mechanical and electromechanical properties.
One disadvantage of such trough-shaped bodies, however, is the fact that in a drive movement excited parallel to the substrate plane (in-plane), no pure in-plane movement results due to the center of mass having shifted somewhat downward and therefore being below the center of the spring, but rather a small parasitic out-of-plane movement component occurs, which, as sketched in
Micromechanically produced hollow structures are fundamentally known from applications of microfluidics, although these hollow structures are not movable MEMS structures. Hollow structures of a CMOS back end formed by metal oxide stacks are described, for example, in U.S. Pat. Nos. 8,183,650 B2 and 8,338,896 B2, and United States Patent Application No. US 2011 049 653 A1. The structures formed from metal oxide stacks have the disadvantage that the typical thicknesses of the individual functional layers are merely in the range of 1 μm or below.
The metal layers furthermore have thermal expansion coefficients and stress values that markedly differ from those of the surrounding oxide layers. Following the exposure of the structures, both the small thicknesses as well as the great differences in the material parameters of metals and oxides can result in great strain and warping and additionally in changes of the mechanical or geometrical properties across temperature or service life. This yields markedly inferior sensing properties in comparison to micromechanical components formed from silicon layers.
It is therefore an object of the present invention to provide an improved micromechanical component, in particular an improved micromechanical inertial sensor.
In accordance with the present invention, the objective may be achieved in accordance with a first aspect by an example micromechanical component having a movable seismic mass developed in a second and third silicon functional layer, a hollow body being developed in the second and third silicon functional layers, which has a cover element developed in a fourth silicon functional layer.
In this manner, a hollow body made of silicon layers is provided in the movable seismic mass, as a result of which the seismic mass exhibits minimized parasitic effects because surfaces of the rocker device are upwardly and downwardly symmetrized, dimensions of the surfaces upward and downward being largely identical. Since in addition the movable mass is developed from silicon functional layers, the micromechanical components according to the present invention have very favorable properties.
According to a second aspect of the present invention, the objective is achieved by a method for manufacturing a micromechanical component, having the steps:
Preferred developments of the example micromechanical component are the described herein.
One advantageous development of the micromechanical component in accordance with the present invention is characterized by the fact that additionally first electrodes are developed in a first silicon functional layer, the seismic mass being capable of functionally interacting with the first electrodes. This advantageously makes it possible capacitively to detect movements of the seismic mass perpendicular to the substrate plane.
Another advantageous development of the micromechanical component in accordance with the present invention is characterized by the fact that additionally second electrodes are developed in the second, third or fourth functional layers. In this manner, additional stationary electrodes are provided, which further improves the sensing behavior of the micromechanical component.
Another advantageous development of the micromechanical component according to the present invention is characterized by the fact that the layer thicknesses of the second, third and fourth silicon functional layers are greater than approx. 1 μm, which advantageously makes it possible to achieve high stiffness, low warping and large capacitance areas.
Another advantageous development of the micromechanical component according to the present invention is characterized by the fact that the layer thicknesses of the third silicon functional layer is greater than 8 μm, which advantageously makes it possible to achieve high stiffness, low warping and large capacitance areas.
Another advantageous development of the micromechanical component according to the present invention is characterized by the fact that layer thicknesses of the second and fourth silicon functional layers are similar in a defined manner. This ensures that a center of mass of the movable mass is well adjusted in relation to the center point of the spring axis, which largely prevents undesired parasitic movements of the movable mass in the z-direction.
Another advantageous development of the micromechanical component according to the present invention is characterized by the fact that layer thicknesses of the second and fourth silicon functional layers differ by maximally 50%, preferably by maximally 25%. This also makes it possible largely to avoid parasitic deflections of the movable mass in the z-direction.
Another advantageous development of the micromechanical component according to the present invention is characterized by the fact that, at least in sections, a ratio of an area coverage between the second and fourth silicon functional layers and the third silicon functional layer is between three and ten, preferably five. This supports an efficient production of the hollow space in the additional hollow mass using conventional surface micromechanical processes.
The present invention is described below in detail with additional features and advantages with reference to several figures. Identical or functionally identical elements bear the same reference symbols. The figures are intended in particular to elucidate the main principles of the present invention and are not necessarily executed true to scale. For the sake of clarity, it may be provided that not all reference symbols are drawn in all figures.
Disclosed method features result analogously from corresponding disclosed device features and vice versa. This means in particular that features, technical advantages and embodiments relating to the micromechanical component analogously result from corresponding embodiments, features and advantages of the method for operating a micromechanical component and vice versa.
Such standard rockers are simply constructed and widely used, but have some technical problems, which hamper applications with very high requirements regarding offset stability. A significant limitation of the offset stability may be brought about by parasitic electrostatic effects, which are explained below.
For the capacitive evaluation, an electrical effective voltage, for example a pulsed electrical square-wave voltage is applied to the movable structure. In the area of the additional mass, electrostatic forces therefore act between the movable structure and the substrate as soon as an electrical potential difference occurs between the movable structure and the substrate. The forces or the resulting torques result in a parasitic deflection of the rocker. To minimize the electrostatic interaction, an additional conductor track surface is therefore usually situated on the substrate in the area of the additional mass, which has the same potential applied to it as the movable structure.
Theoretically, a freedom from forces may be achieved thereby between the additional mass and the substrate. In practice, however, significant surface charges or effective surface potentials may be present on the conductor track surface connected to the substrate and/or on the lower side of the movable structure, which can still result in parasitic forces and thus in electrical offset signals. These effects are particularly critical if they change across temperature or service life of the product since this results in offset drifts that cannot be corrected by the final calibration of the component.
A core idea of the present invention is in particular to create a micromechanical component, in particular an inertial sensor, having an improved offset stability and sensing characteristic.
In the micromechanical component of the present invention, a symmetrization of sensor masses with respect to parasitic forces (e.g., electrostatic and radiometric forces) is provided when two boundary surfaces exist, both below as well as above movable masses. This is achieved while simultaneously maintaining the mass asymmetries.
Furthermore, it is possible to exploit the advantages of light construction masses for rotation-rate sensors without having to accept parasitic movements of trough-shaped oscillating masses.
Furthermore, a surface micromechanical production method is provided for manufacturing hollow masses for movable MEMS structures.
The mentioned advantages are achieved in accordance with the present invention by a formation of hollow masses for movable MEMS structures, which are formed from three silicon functional layers as well as by a corresponding surface micromechanical production method for manufacturing such hollow masses.
For micromechanical z-acceleration sensors, it is thus possible to achieve a symmetrization with respect to parasitic forces or torques (e.g., electrostatic or radiometric forces/torque) on the upper and lower sides of the movable structure.
For rotation-rate sensors, it is possible in this manner to build very light, but at the same time stiff sensor masses, whose z-coordinate of the mass center of the mass is, in contrast to trough-shaped bodies, at the same elevation as the z-coordinate of the mass center of the spring so that in an in-plane-movement no or only extremely weak parasitic z-movements occur.
By using silicon as functional layer material, it is possible to achieve very favorable mechanical properties having a high temperature stability and service life stability.
The thicknesses of the silicon functional layers may preferably be selected to be relatively great, in particular greater than 1 μm. It is thus possible to build hollow masses that are very stiff and that barely tend to twist or warp.
It is furthermore advantageous to design at least one of the silicon functional layers, preferably the third silicon functional layer, to be particularly thick in order to achieve great masses, high stiffness values and large capacitance areas. Particularly advantageous are layer thicknesses for the third silicon functional layer greater than 8 μm, e.g. 10-50 μm.
Advantageously, it is thereby possible to minimize or compensate for radiometric effects with consequences in the form of parasitic deflections of rocker W in the z-direction. Furthermore, this makes it possible to maintain a pronounced mass asymmetry between the left and the right sides of the rocker since the mass on the right rocker side is formed largely (perforation holes are not shown in the figures for the sake of simplicity) from the thick third silicon functional layer 30 and is thus markedly heavier than the left rocker side.
This also ensures that a high mechanical sensitivity of micromechanical component 100 is maintained.
The hollow masses according to the present invention may also be advantageously used in micromechanical components in the form of rotation-rate sensors. In analogy to
It is additionally particularly preferred that the layer thickness of the third silicon functional layer is chosen to be greater than 8 μm, preferably 10-50 μm, while the layer thicknesses of the second and fourth silicon functional layers may be chosen to be markedly smaller. This advantageously makes it possible on the one hand to achieve hollow masses that are flexurally very stiff, to achieve furthermore great mass differences between hollow masses and filled masses, and finally to achieve stiff springs in the third silicon functional layer, the z-coordinate of the spring coinciding with the z-coordinate of the mass center of the hollow mass and parasitic z-movement components being avoided in an in-plane movement.
As manufacturing method for the spring geometries provided here, it is possible to use a surface micromechanical process described in more detail below, in which the four silicon functional layers 10, 20, 30 and 40 are used, which are preferably formed from polysilicon. The process sequence is shown in
In a substep a), a substrate 1 is provided with a first oxide layer 2, the first silicon functional layer 10 and a second oxide layer 3.
In a substep b), the second silicon functional layer 20 is deposited onto second oxide layer 3 and is patterned by fine trenches.
In a substep c), a third oxide layer 4 is deposited, which closes the trenches on top. This is followed by further process steps, which have no visible effect in the area of the shown hollow mass, however, and are therefore not shown in the figures, that is, the opening of third oxide layer 4 through fine slits and a subsequent etching step of the second silicon functional layer 20 (preferably by isotropic SF6 or XeF2 etching) through the fine oxide openings.
In substep d), a further oxide layer 5 is deposited, whereby all fine openings in third oxide layer 4 are closed. The advantage of the method lies in the fact that it is possible to clear out large areas of second silicon functional layer 20 without leaving significant topography on the surface of oxide layer 5, as known for example from DE 10 2011 080 978 A1. Subsequently, fourth oxide layer 5 is patterned together with third oxide layer 4 in order to allow for contacts between second silicon functional layer 20 and third silicon functional layer 30.
In a substep e), third silicon functional layer 30 is deposited and patterned via fine trenches.
In a substep f), a fifth oxide layer 6 is deposited, and small openings are created in fifth oxide layer 6.
In an etching step in substep g), which is preferably developed as isotropic SF6 or XeF2 etching, sacrificial silicon areas are removed in third silicon functional layer 30.
As indicated, in substep h), the openings in fifth oxide layer 6 are closed again by another oxide layer 7.
Subsequently, seventh oxide layer 7 is patterned together with sixth oxide layer 6 in order to provide electrical contacts between third silicon functional layer 30 and fourth silicon functional layer 40.
In substep i), fourth silicon functional layer 40 is deposited and patterned.
As indicated in substep j), all sacrificial oxides 6, 7 are removed by oxide etching, preferably using gaseous HF, and the sensor structure is exposed.
Ultimately, in substeps a) through j) of
The provided method offers the possibility of cleaning out large areas of third silicon functional layer 30 and nevertheless covering it almost completely with the (merely slightly perforated) fourth silicon functional layer 40.
For example, a ratio between the area coverage of second silicon functional layer 20 and fourth silicon functional layer 40 on the one hand and the area coverage of third silicon functional layer 30 on the other hand may be significantly greater than three, a ratio of ten being possible as well. This is achieved by the perforations, created using etching technology, in the mentioned silicon functional layers, which, at least in sections, in second and fourth silicon functional layers 20, 40 make up approx. 10% to approx. 20% and in third silicon functional layer make up approx. 80% to approx. 90% of the entire area coverage.
In a step 200, a movable seismic mass developed in a second and third silicon functional layer 20, 30 is provided.
In a step 210, a hollow body 36 is developed in the second and third silicon functional layers 20, 30, which has a cover element developed in a fourth silicon functional layer 40.
Although the present invention was described above with reference to concrete exemplary embodiments, in particular acceleration and rotation-rate sensors, one skilled in the art is also able to implement specific embodiments that were not disclosed above or that were disclosed above only partially, without deviating from the essence of the invention. It is in particular possible to use the present invention for other micromechanical components such as e.g. resonators, micromirrors or Lorentz magnetometers.
| Number | Date | Country | Kind |
|---|---|---|---|
| 102018219546.3 | Nov 2018 | DE | national |
