DE19750350C1 - Three-dimensional chip acceleration sensor - Google Patents

Abstract

The sensor has individual acceleration sensors for each of the x,y and z directions, each consisting of three layers, the first of which is a common carrier plate (1) for all three. A seismic test mass for the x-direction sensor is at 90 degrees to that for the y-direction sensor. Lateral deflections of the test masses are measured. The z-direction sensor has an asymmetrically suspended test mass (12) whose torsional movement about its suspension axis (10) is measured for acceleration in the z-direction. The carrier plate is of silicon or ceramic and all other parts are metallic structures. An Independent claim is also included for a method of manufacturing the sensor.

Description

The present invention relates to a three-dimensional chip acceleration sensor and a method for its production by means of UV-assisted micro-electroplating the preamble of claims 1 and 9.

With the high precision and the small size, such a Accelerometer can be used wherever a three-dimensional measurement of Accelerations in a "low cost" manner is required.

Acceleration is a general physical parameter. She accompanies us in various forms through daily life. As diverse as the appearance of Accelerations are as varied as their measuring methods. Accelerometers that based on the piezoelectric, piezoresistive or capacitive measuring principles already developed. With regard to a low temperature coefficient, a high one Long-term stability and in particular a monolithic integration potential Microelectronics feature the surface micromechanically manufactured, capacitively operating Acceleration sensors have significant advantages over other principles. Most However, previously known micromechanical acceleration sensors are one-dimensional Sensors formed (WO 92/03740, EP 0 623 825, DE 44 32 837, DE 41 26 100, etc.). Is A three-dimensional detection of the acceleration is required, three individually working sensors with corresponding mechanical auxiliary devices through specific and very complex assembly techniques arranged on a suitable molded body and interconnected, as in the cases DE 91 13 744, WO 94/14076 and WO 96/04561, which limits the achievable mechanical precision and the degree of miniaturization become. Another disadvantage of this solution is the relatively high outer Wiring effort, which even falsifies the measurement signal in some applications becomes.

In the publication Reuber, Claus: "Acceleration measured by chip", electronics 3 (1995) pp. 58 to 62 describe a chip acceleration sensor that is only in one direction is sensitive. This chip acceleration sensor is based on the capacitive measuring principle and consists essentially of a double comb electrode and a seismic mass, which is attached to a resilient arm. To limit the deflection of the seismic Mass is provided a stop.

From JP patents abstracts of Japan 5-264577 (A) is a chip acceleration sensor known for the detection of three-dimensional accelerations. This is four by one Point arranged seismic masses arranged symmetrically along a carrier. The individual distances of the four seismic cells change according to the acceleration Masses to electrodes attached to the carrier, so that three-dimensional Acceleration values can be determined.

At the 8th International Conference on Solid State Sensors and Actuators 1995 ", pp. 566-569 was a sensor element for three-dimensional determination of the Acceleration presented. A piezoceramic was used as the core element, the one below Acceleration is deformed and a charge is generated. A Group of 16 electrodes were used to measure the charge change on the piezoceramic  appropriate. The accelerations of different axes were then made up of 16 signals costly converted. In addition, the temperature drift of the sensor is very large. On the same conference (pp. 554-557) was a capacitive three-dimensional Accelerometer presented. This has a glass-silicon-glass structure and was manufactured using bulk micromechanics. The three disks were structured separately and hermetically sealed at controlled pressure by anodic bonding. The Here, too, acceleration has to be carried out using a complex calculation from the measurement signals corresponding measuring electrodes can be calculated.

DE 37 41 036 describes a directly measuring three-dimensional chip Acceleration sensor described by the piezoresistive principle. On the one hand, that shows piezoresistive principle disadvantages compared to the capacitive measuring principle, on the other hand its manufacture requires many time and costly processes, e.g. B. epitaxy, Ion implantation of the etch stop layer, diffusion of piezoresistors and a multi-stage lithographic structuring. This manufacture is not with the IC technologies compatible. Therefore, a possibility of integrating the sensor element with the Evaluation circuit out.

The object of the invention is to produce three-dimensional chip Realize acceleration sensors using a simple, inexpensive method. The chip acceleration sensor should be designed so that it for a three-dimensional determination of the accelerations is usable, in each Direction of measurement advantageously based on the differential capacitance measuring principle. The sensor should using a cost-effective technology compatible with the IC process in one Planar process can be produced with a few manufacturing steps.

According to the invention, this task is accomplished by a chip acceleration sensor solved the features specified in claim 1. Further training and refinements result from the associated subclaims. Furthermore, the task is performed by a Method solved with the method steps mentioned in claim 9. Beneficial Process variants result from the subclaims.

An essential feature of the invention is that the chip acceleration sensor integrated structure consisting of three differential capacitive acceleration sensors which are two identical, laterally sensitive, but rotated 90 ° against each other  Sensors for the X and Y direction and a vertically sensitive sensor for the Z direction acts.

The chip acceleration sensor is supported by the combination of a UV Micro electroplating process with a sacrificial layer technique from a three-layer system manufactured. The first layer serves as the base layer for the galvanic structure of the Sensor structure. The connecting tracks or the bond pads on the carrier are also structured out of the first layer. The second layer is the so-called sacrificial layer, which serves as an auxiliary layer for building up the sensor structure and then again completely is etched away. The third layer is the actual sensor structure, the electrochemical one within the spaces of a lithographically structured lacquer on the base layer or the victim layer grows up. The desired moving parts of the sensor structure are then by selective wet chemical etching of the underlying sacrificial layer generated.

The great advantage of the invention is that the three sensor elements for Determination of the acceleration in the three spatial axes using planar micro-electroplating are built up integrated with a substrate. This eliminates the need for mechanical Auxiliary devices, the assembly or the wiring work, the three one-dimensional Acceleration sensors for three-dimensional acceleration measurement are necessary. Of the Chip acceleration sensor according to the invention has a high degree of miniaturization and high mechanical precision.

In each measuring direction, the measurement is based on the differential capacitance principle. Thus, the acceleration can be done in any direction using the so-called Force compensation method can be determined precisely and linearly. The seismic Mass is held in zero position by an electronic control, whereby non-linearities of the Spring deflection can be switched off. In addition, the structure is mechanically less loaded, so that a higher long-term stability of the sensor is guaranteed.

The manufacturing process is particularly simple and economical, since it is expensive Temperature processes, dry etching, doping and X-ray lithography as far as possible be avoided.  

In the laterally sensitive acceleration sensors used, the fixed electrodes built up on an insulated substrate by immovable are firmly connected to the carrier in their central region. In comparison, they are So-called fixed electrodes in the prior art actually only at one end anchored to the carrier. The plate height remains along the finger electrodes and thus the Capacitance of the capacitor constant. This enables a high stability of the structure, with what a high measuring accuracy is guaranteed. A folded spiral spring is advantageous used, which allows a greater deflection of the test mass. Additional blocks on the Lag Islands serve as protection against excessive excursions.

The are advantageous for the vertically sensitive acceleration sensor Torsion springs are not connected directly to the anchor, but via a rectangular ring. This prevents tension from being transmitted from the anchor to the torsion spring. At the same time the asymmetrical mass is slotted to provide favorable dynamic properties to achieve the sensor. The width of the slots should be half the slot spacing be.

The height of the sensor structure is determined by the thickness of the photoresist used determined and can range up to about 30 microns with an aspect ratio of about 6. in the By comparison, acceleration sensors structured out of polysilicon have one Structure height of about 2 µm. High structures improve the rigidity of the Sensor structure. The cross sensitivity in another direction is due to the high rigidity the suspension of the seismic masses is very low.

The acceleration sensor according to the invention is made of galvanized metallic Layers built up that allow or allow direct bonding with metal wire (Au, AlSil) which through the so-called flip-chip bonding a permanent connection with the manufactured external circuit. This eliminates the need for additional metallization.

The acceleration sensor according to the invention is in one with the evaluation circuit so-called additive process, easily integrated on the same substrate. The Evaluation circuit can first be produced using any method (CMOS, bipolar or mixed process). The process for manufacturing the sensor continues additively without intervention in the method for producing the evaluation circuit.

An embodiment of the invention is described below with reference to the drawings explained in more detail. Show it:

Fig. 1: an overall view of an inventive chip-acceleration sensor

FIG. 2 shows a schematic structure of a laterally sensitive accelerometer

FIG. 3 shows a cross-section AA 'as shown in FIG. 2

FIG. 4 shows a cross-section BB 'corresponding to Fig. 2

Fig. 5: a schematic structure of the vertically sensitive acceleration sensor

Fig. 6 shows a cross section CC 'in accordance with FIG. 5

FIG. 7 is a view for illustrating the manufacturing process.

Fig. 1 shows an overall view of an inventive chip-acceleration sensor. The chip acceleration sensor of the invention consists of two identical, at 90 ° to each other twisted laterally sensitive sensor elements (for X- and Y-direction of measurement) and a vertically sensitive sensor element (for Z-direction of measurement ). Each sensor element is constructed like a differential capacitor, ie each has a seismic mass suspended on bearing islands via the bending springs as a movable center electrode and two fixed counter electrodes. The bearing islands for the suspension of the seismic mass and the counter electrodes are led through the conductor tracks to bond pads.

Fig. 2 shows the schematic structure of a lateral acceleration-sensitive sensor on a substrate 1. The sensor has a seismic test mass 2 which is laterally suspended at each end via a folded spiral spring 3 on a bearing island 4 parallel to the surface of the carrier 1 (see arrow in FIG. 2). The distance between the test mass 2 and the carrier 1 corresponds to the thickness of the sacrificial layer used in the manufacturing process. The use of a folded spiral spring 3 allows a greater deflection of the test mass 2 than a simple straight suspension due to its small spring constant, which is only half as large as that of a simple suspension. The additional blocks 5 on the bearing islands 4 are provided to protect against excessive deflection of the test mass 2 . The lateral distance between the blocks 5 and the test mass 2 determines the maximum deflection of the test mass 2 and thus the working range of the laterally sensitive acceleration sensor.

Finger structures 6 protrude from the test mass 2 on the right and left, which are assigned to opposite finger plates of the fixed electrodes 7 a and 7 b. Accordingly, a differential capacitor arrangement is formed between the movable test mass 2 and the two rigid electrodes 7 a and 7 b, ie the movable test mass 2 forms a capacitor with the two fixed electrodes 7 a and 7 b, the capacitance of one capacitor increasing if the capacitance of the other capacitor decreases. The deflection of the seismic test mass 2 or the acceleration can be determined by measuring the capacitance between the movable electrodes 6 and the fixed electrodes 7 a and 7 b.

The fixed electrodes 7 a and 7 b are each firmly connected to the carrier 1 with their central region. A corresponding view is shown in FIG. 3. Here, a cross section through the sensor according to FIG. 2 along AA 'is shown. As can be seen in FIG. 3, the electrodes 7 a and 7 b are firmly connected to the carrier 1 at their central point via the base layer 8 , while the electrodes 6 remain laterally freely movable by the seismic test mass 2 . Along the finger electrodes, the plate area and thus the capacitance of the capacitor between fixed electrodes 7 a and 7 b and the seismic test mass 2 remain constant.

This structure has clear advantages over the sensor structures according to WO 92/03740, DE 44 32 837 A1 and those from Siemens or Analog Devices (see magazine "Sensors and Actuators" A 45 (1994) pp. 7-16 and A 48 (1995) pp. 101-108) which the so-called fixed electrodes are anchored to the support at only one end are. In these, on the one hand, the sensor structures are unstable, which are influenced by the Also move acceleration, which leads to reduced measuring accuracy. On the other hand such electrode arrangements bend downward. This has a decrease in Capacitor area and thus the measuring capacity, which increases the sensitivity of the Accelerometer reduced.

As can be seen in Fig. 2, the fingers of the fixed electrodes 7 a and 7 b are arranged on both sides of the movable mass 2 . The electrical contacting of the finger electrodes on both sides of the seismic test mass 2 takes place by means of an interconnect, which is constructed out of the base layer 8 and is guided below the movable test mass 2 , as shown in FIG. 4.

Fig. 5 shows the schematic structure of the vertically sensitive acceleration sensor on a substrate 1. The basic structure of the vertically sensitive sensor is an asymmetrical test mass 12 suspended by two torsion springs 10 on an armature 11 via two counter electrodes 9 a and 9 b. The anchor 11 is firmly connected to the carrier 1 at its central region via the base layer 8 . The torsion springs 10 is connected by a rectangular ring 15 that is in turn mounted via four Verbindungsstänge 16 on the armature 11 with the armature. 11 This structure prevents the transmission of mechanical stresses from armature 11 to torsion springs 10 . Under the asymmetrical test mass 12 are two electrodes 13 a and 13 b, which are used for self-tests.

In order to maintain the high sensitivity of the sensor, the distance between the asymmetrical test mass 12 and counter electrodes 9 a and 9 b must be as small as possible so that the change in the electrode distance caused by the design leads to a large change in capacitance. With a small distance, the problem of overdamping arises. In order to achieve favorable dynamic properties of the sensor, the asymmetrical test mass 12 is provided with slots 14 above the electrodes 9 a, 9 b, 13 a and 13 b. The optimal width of the slots should be half the slot spacing. The slots are used to reduce air damping and give the sensor a wide bandwidth.

FIG. 6 shows a cross section through the sensor according to FIG. 5 along CC ′. As can be seen, the asymmetrical test mass 12 together with the two electrodes 9 a and 9 b represents a differential capacitor. The distance between the asymmetrical test mass 12 and the electrodes 9 a and 9 b corresponds to the thickness of the sacrificial layer used. An acceleration acting in the Z direction results in a resultant force which leads to torsion of the torsion springs 10 due to a deflection of the asymmetrical test mass 12 . As a result, the two capacitances between the asymmetrical test mass 12 and the fixed electrodes 9 a and 9 b change.

A micromechanical chip acceleration sensor described above can be easily produced according to claim 9 by combining a UV-assisted micro-electroplating process with an opera layer technique, the sequence of which is shown schematically in FIG. 7. In Fig. 7 a laterally sensitive accelerometer (left side) and a vertically sensitive accelerometer (right side) are shown.

A 0.8 μm thick photoresist 17 is first applied to a carrier 1 , which is then structured lithographically. Fig. 7.1 shows the carrier 1 with the patterned photoresist 17th The carrier 1 consists of silicon, the surface of which is electrically insulated with an SiO ₂ or Si ₃ N ₄ layer. Glass, Al ₂ O ₃ or other ceramic materials can also be used as supports.

A layer system (50 nm thickness W / Ti and 300 nm thickness Cu) is deposited on the surface of the structured photoresist 17 by means of sputtering. W / Ti serves as an adhesion promoter and Cu as a base layer 8 . In addition to Cu, other conductive materials can also be used for this layer, as long as they have a high etching resistance against etching solutions for titanium, e.g. B. Au, Ag, Pt, etc. By subsequent lift-off, the bond pads, the interconnects, the counter electrodes of the vertically sensitive acceleration sensor or the bearing pads of the laterally sensitive acceleration sensor are structured, as shown in FIG. 7.2.

A titanium layer 18 with a thickness of approx. 3 μm is applied to the surface of the lift-off structured base layer by means of high vacuum evaporation and then structured with a 2 μm thick etching mask, which is produced by a photolithographic process, so that the areas in where the parts of the microstructure that are to be fixed at the end are to be etched up to the copper base. Titanium is selected as the sacrificial layer material because the Ti surface ensures good adhesion to the thick UV varnish used, because it offers the possibility for a homogeneous deposition of the nickel sensor layer and when etched with hydrofluoric acid, it is very selective to all other materials of the Has sensor system. It is possible to deposit stress-free with thicknesses of up to 10 µm using PVD Titan. The thickness of the Ti layer as a sacrificial layer should on the one hand be very large so that the movable structures have a sufficient distance from the substrate, and on the other hand a large distance with too great a thickness leads to a smaller capacitance of the vertically sensitive sensor. A Ti layer thickness of 3 µm was chosen as a compromise to the contrary requirement. Fig. 7.3 shows the patterned titanium layer.

A thick layer of photoresist 19 is then spun on. A photoresist that is particularly suitable for this process is AZ 4562. With a simple coating, the thickness of the photoresist can reach approximately 10 μm. Even thicker layers of paint can be achieved after softbaking the previous paint layer and repeating the spinning process. Due to the technical limitations of UV lithography, a lacquer thickness of more than 30 µm is not recommended. After drying, the thick photoresist is exposed through a mask that contains the layout of the sensor structure. Subsequent removal of the exposed areas in a developer creates a mold insert for the electroforming of the sensor structure in the photoresist ( Fig. 7.4).

Fig. 7.5 shows the galvanic structure of the sensor structure. Nickel, the actual functional material of the acceleration sensor, is electrochemically deposited from a nickel sulfamate electrolyte into the free spaces of the photoresist 19 on the base layer 8 or the sacrificial layer 18 . The composition of the electrolytes used for the electroforming of the sensor structure is:

Ni sulfate: 350 g / l,
Ni chloride: 5 g / l,
H ₃ BO ₃ : 35 g / l and
Ni carbonate: 3 g / l.

So that a homogeneous Ni layer with an internal tension of less than 10 N / mm ² can be produced, the electrolyte solution is adjusted to a pH value of 3.5 using nitric acid and the following deposition parameters are selected: temperature 55 ° C, current density 1 A / dm ² . The deposition rate is about 0.2 µm / min.

The stripping of the photoresist 19 follows the deposition of the sensor structures. This is done by detaching with acetone or, if the hardbake was carried out at temperatures above 120 ° C., in a mixture of H ₂ SO ₄ and H ₂ O ₂ or by ashing in an oxygen plasma.

In the last manufacturing step, the titanium sacrificial layer is selectively etched with a 5% hydrofluoric acid compared to the other metals. After this process, the sensor structure grown on the titanium layer is freely movable, while the other parts of the sensor structure that have grown up on the Cu base remain firmly anchored to the substrate ( FIG. 7.6).

Claims (12)

1. Three-dimensional chip acceleration sensor consisting of individual acceleration sensors for the acceleration directions (x, y, z), each individual acceleration sensor consisting of three layers ( 1 , 8 , 20 ), the first layer of which is designed as a carrier plate ( 1 ), two spiral springs ( 3 , 10 ) and an associated deflectable test mass ( 2 , 12 ), which can be deflected in the acceleration direction due to an acting acceleration to the fixed electrodes ( 7 a, 7 b; 9 a, 9 b), and from the fixed electrodes ( 7 a, 7 b; 9 a, 9 b) with their finger plates and the test compound ( 2 , 12 ) with their finger plates arranged on both sides, a differential capacitor arrangement is formed, characterized in that the carrier plate ( 1 ) is a common carrier for the individual acceleration sensors to which the seismic test masses ( 2 , 12 ) lie in one plane, the seismic test masses ( 2 ) of the acceleration sensor s for the x direction to the seismic test mass ( 2 ) for the y direction is rotated by 90 ° and the lateral deflection of the seismic test masses is measured, and the acceleration sensor for the z direction has an asymmetrically suspended seismic test mass ( 12 ) and the torsional movement of the seismic mass ( 12 ) around the axis formed by the suspension ( 10 ) is measured for the acceleration in the z-direction, the carrier plate ( 1 ) being made of silicon or ceramic and all other parts of the acceleration sensor being metallic structures . 2. Three-dimensional chip acceleration sensor according to claim 1, characterized in that left and right of the finger plates ( 6 ) of the seismic test mass ( 2 ) fixed electrodes ( 7 a, 7 b) are provided, all electrodes ( 7 a) and all electrodes ( 7 b) are interconnected by a first metal layer ( 8 ). 3. Three-dimensional chip acceleration sensor according to claim 2, characterized in that the fixed electrodes ( 7 a, 7 b) are contacted over their entire length by the first metal layer ( 8 ) with the carrier ( 1 ). 4. Three-dimensional chip acceleration sensor according to one of claims 1 to 3, characterized in that the seismic test mass ( 2 ) via a respective at the ends of the test mass ( 2 ) provided folded bied spring ( 3 ) is attached to the bearing island ( 4 ). 5. Three-dimensional chip acceleration sensor according to claim 1, characterized in that the bearing islands ( 4 ) have additional blocks ( 5 ) for limiting the deflection of the seismic test mass ( 2 ). 6. Three-dimensional chip acceleration sensor according to claim 1, characterized in that the seismic mass ( 12 ) of the acceleration sensor for the z-direction different masses, based on the axis formed by the torsion spring ( 10 ). 7. Three-dimensional chip acceleration sensor according to claim 6, characterized in that the torsion spring ( 10 ) is attached to a rectangular ring ( 15 ) which in turn is connected to the bearing island ( 11 ) by means of supports ( 16 ). 8. Three-dimensional chip acceleration sensor according to claim 6 or 7, characterized in that the seismic test mass ( 12 ) has slots ( 14 ), the width of the slots ( 14 ) being half the distance of the slots ( 14 ). 9. A method for producing a three-dimensional chip acceleration sensor according to one of claims 1 to 8, by means of UV-assisted micro-electroplating, characterized in that

• a) a photoresist layer ( 17 ) is spun onto a carrier ( 1 ) and the photoresist layer is subsequently structured photolithographically, the structure forming the negative for the support points and the fixed electrodes,
• b) a first metal layer ( 8 ) is deposited into the structured photoresist layer ( 17 ) via an adhesion promoter and then the remaining photoresist layer ( 17 ) is removed,
• c) a second metal layer ( 18 ) is deposited over the first metal layer ( 8 ) structured in this way and then structured, the later bases of the fixed structures being exposed,
• d) a thick photoresist layer ( 19 ) is spun onto the second structured metal layer ( 18 ) and subsequently structured, the pattern for the subsequent galvanic structure being formed from this structure,
• e) the actual functional material ( 20 ) of the acceleration sensors in the mold insert of the thick photoresist layer ( 19 ) is electroplated on the first metal layer ( 8 ) and the second metal layer ( 18 ), and finally
• f) the thick photoresist layer ( 19 ) is stripped and then the second metal layer ( 18 ) is completely removed, as a result of which the actual functional material becomes movable.

10. The method according to claim 9, characterized in that the photoresist layer ( 17 ) in step a) is spun on by means of coating. 11. The method according to claim 9 or 10, characterized in that the structuring of the first metal layer ( 8 ) in steps a) and b) is carried out by lift-off. 12. The method according to claim 9, 10 or 11, characterized in that the second metal layer ( 18 ) in steps c) to f) is a sacrificial layer and is deposited as a temporary titanium layer.