The present invention relates to a device according to the preamble of Claim 1. The invention also relates to a method of manufacturing a device according to the preamble of claim 7.

Up to this day vibrating angular rate measurement devices gyros of the tuning fork type or made with the aid of other vibrating bodies have been designed and tested in metal, quartz and in silicium, both of the type with an open loop and with a closed loop. Gyros based upon quartz and silicium have, when compared with the metal fork gyro, well known for a long time (Bamaby et al, Aeronautical Engineering Review, p. 31, Nov 1953) the advantage, that they may be miniaturized and can be manufactured relatively cheaply in long series by batch methods. In contrast to rotating gyros, where the Corioli force developes a constant moment when turned, a vibrating gyro, when turned gives an osciallting moment, where the amplitude is proportional to the turning rate and the phase gives the turning direction. All this means, that there is a lot of electronics needed for excitating and detection and inteipreting of the gyro signal from a vibrating gyro when compared with a conventional rotating gyro. As silicium is a very good electronics material there is much gained with a silicium based gyro integrated with electronics for excitating and sensing. To manufacture and excitate a traditional tuning fork in silicium has not been possible to do easily. The advantage with a tuning fork design is, that it can be made to swing in a dynamically balanced mode, that gives a high Q-value, as little vibration energy is emitted to the surrounding. Even if it would be principally possible to etch a tuning fork in silicium with the aid of modern plasma etching techniques, still there is the problem of exciting. In silicon you cannot directly utilize the piezoelectric effect for exciting, as is possible for quartz, as silicon is no piezoelecric material. Of course it is possible to deposit a piezoelectric layer to silicon, but if you have etched the structure vertically down to the substrate layer, this means, that you shall deposit an evenly thick piezoresistive layer to vertical walls, which can be rather problematic. In stead the usual solution to the exciting problem for silicon based gyros has been to utilize electrostatic excitation perpendicular to the substrate surface, by manufacturing conducting plates on each side of a beam or a tilting plate with a weight, alternatively, to "glue" the entire component to a piezoelectric plate.

The gyro designs disclosed up to now, in metal, silicon and quartz have only been able to measure rotation around one axis, and if more axes have been demanded, more monoaxis gyros have been combined to one multiaxis gyro.

CONFIRMATION COPY

The present invention extends the technology for vibrating gyros, designed from a semiconducting substrate to enable manufacturing vibrating gyros in the same substrate for simultaneous measuring of rotation around three dimension axes, perpendicular to each other, based upon the same manufacturing technique by the features that are given in the characterizing portions of the independent claims 1 and 6. Dependent upon the needs, the invention can be integrated into several configurations on the same substrate and together with accelerometers, which enables simultaneous measurement of rotation around several axes, possibly combined with measurement of acceleration. Furthermore, dependent upon the precision demanded, the device can be made eidier with an open or a closed loop. The utilization of well known silicium technology when manufacturing enables mass production at a low cost with a high precision and reliability.

One embodiment of the device (type I) according to the invention for measurement of angular rate comprises two beams, inclined to the normal plane with or without attached inertial masses, dependent from a frame, which via beams (torsion axes) is connected to the substrate, everything etched from a monocrystalline material. An example of such material is (lOO)-oriented ("Miller index"), monocrystalline silicon, which can be doped to create conducting means, where such are needed. The beams may either be fixed to the same side of the frame or to opposite sides of the frame. All beams have the same longitudinal direction and the same angle to the normal plane, which may be positive or negative in the xz-plane in a coordinate systeme, where the y-axis is directed along the beam and the z-axis is perpendicular to the substrate surface. The beams and the inertial masses, if present, are all formed by anisotropic etching of the monocrystalline material.

A second embodiment of the device (type II) according to the invention for measurement of angular rate comprises two opposite beams, inclined to the normal plane (torsion axes) carrying an inertial mass, which via the beams are dependent upon a frame, which in tum is connected to the substrate via two opposite beams (torsion axes), inclined to the normal plane, which beams are turned 90° in relationship to the first pair of beams, everything being etched from a monocrystalline material. An example of such a material is (100)- oriented ("Miller index"), monocrystalline silicon, which can be doped in order to create conductive means where such are needed. The beams in the pairs of beams have the same longitudinal direction and have the same angle to the normal plane, but the two pairs of beams are turned 90° mutually. The beams and inertial masses, if present, are all formed by anisotropic etching of the monocrystalline material.

The connections of frames to the substrate is achieved by flexible beams that is to say torsion axes, which have an angle to the normal plane, well defined by the crystall

structure. By anisotropic etching from opposite surfaces of semiconducting substrate there are created beams, inclined to the normal plane, which beams connect the vibrating bodies, alternatively the frames to the substrate. The width and thickness of the beams can be dimensioned so that a high bend flexibility is achieved along an axis perpendicular to resp inclined beam and a high bend rigidity along all other directions, or alternatively so that the bend rigidity will be similar i all directions, which can be preferred when beams are utilized as torsion axes.

Configurations according to type I and II are well adapted to an embodiment with a closed loop as well as an open loop. Both configurations are well adapted to integration together with an mono- or multiaxes accelerometer, known from the Swedish Patent document SE 9203468-2. Integration of two sets of the device according to the invention of type I or II, one set of the device of typell and one three-axes accelerometer known from the Swedish Patent Document SE9203648-2 in the same substrate enable simultaneous measurement of all six degrees of freedom necessary to completely describe the movement of a free body in the space.

A device in monocrystalline material is available for manufacture by well known manufacturing methods for semiconductors comprising for instance photolithographic patterning and anisotropic etching which give many advantages including exact control of tolerances, possibility to integrate the complete or part of the electronics for signal treatment into one single common substrate with a relatively small size, and give access to a technology, that enables efficient mass production by batch-manufacture. When a zero- balanced device of the type with a closed loop is desired, conductive surfaces with a pattern can be placed on opposite sides of beams and the inertial masses for example by so called "Anodic bonding" or "Silicon Direct Bonding". "Anodic Bonding" enables joining of quartz glass to silicon, oxides, nitrides and metals at a low temperature (usually 300- 400°) by application of an electrical field over the joint. "Silicon Direct Bonding", which has been known since 1986 ( asky, Applied Physics Letters Vol. 48, p 78,1986) enables joining of e g silicon to silicon, silicon to silicon oxide and silicon oxide to silicon oxide. "Silicon Direct bonding" and "Anodic Bonding" can also be utilized to put mechical stops for the beams so that these are not broken if they are exposed to larger forces than they are intended for, and if the bonding is performed in vacuum, the technics can simultaneously be utilized as a method for vacuum encapsulation of the device.

The invention is described more in detail, reference being made to de enclosed drawings, which show the device according to the invention in two principal embodiments:

Figures 1 - 5 show schematically show the operating principles of the device according to the invention, figure 6 shows, schematically, a plan view and a cross section of one configuration of type I according to the invention, figure 7 shows an alternative embodiment in the same way, of one configuration of type I according to the invention, figure 8 shows, schematically, a plan view, and a cross section of one configuration of type II according to the invention, and figure 9 shows, , schematically, inperspective, one configuration of type II according to the invention with circuits for excitation and detection.

Firstly an explanation of the function of the device according to the invention is given, reference being made to figures 1- 5. The device is of the type gimbal gyro, that is to say type II mentioned above.

Explanation of the function by the aid of Coriolis forces

Usually the function of the gimbal gyro is calculated by the aid of complicated equations of moments. A more inuitive explanation can be made by the aid of Coriolis forces. If a body moves along a straight line in a rotating coordinate system it is subject to a so called Coriolis force. The movement shall be perpendicular to the rotation axis and the resulting Coriolis force then becomes perpendicular to the movement as well as to the rotation axis.

The frame is excited so that it vibrates around the x- axis. If the gyro is rotated, the vibration around the x- axis also will excite a vibration of the mass around the y-axis. This vibration is detected and gives the output signal from the gyro.The design of the mass will determine the sensitivity of the gyro for rotation around the different axes. Reference being made to the limitations set by the manufacturing method with anisotropic etching in silicium an embodiment has to be chosen, that gives a maximal sensitivity.

An ideal gimbal gyro according to figure 1.

Ideally a gimbal gyro shall be extended as far as possible in the z- direction to achieve a large scale factor. Such a gyro, where the mass is represented by two mass points is shown in figure 5. The mass points move in the y- direction due to the vibration around the x- axis. A rotation around the z- axis will create Coriolis forces in the x- direction.The Coriolis force acting upon the upper mass point has reversed direction against the force at the lower mass point, which will give a moment around the y- axis.

A rotation around the x- axis will give rise to Coriolis forces in the z- direction. As these forces are in the same plane the moment around the y- axis will be zero. A rotation around the y- axis will not give rise to any Coriolis forces, as the mass points move parallelly to the y- axis ( no velocity component is perpendicular to the axis of rotation).

Conclusion: The gyro is sensitive to rotation around the z- axis. It is very difficult to manufacture a gyro of this type in silicium when the raw material is this silicium plate.

A flat gimbal gyro according to figure 2.

A simple model of a flat gimbal gyro is shown in figure 6. The mass is represented by four mass points. These move exclusively in the z- direction due to the excited vibration. If the gyro is rotated around the z- axis no Coriolis forces are created as the mass points have no component of velocity that is perpendicular to the rotation. No moment around the y-axis due to Coriolis forces will be generated and thus no vibration around the y- axis will be generated. A "flat" gyro has no sensitivity to rotation around the z- axis. Rotation around the x- axis will give rise to Coriolis forces in the y- direction, but as these forces are parallel to the y- axis no moment around the y- axis is created. Rotation around the y- axis will give rise to Coriolis forces in the x- direction, but as these forces are in the same plane, no moment around the y- axis is created.

Conclusion: A "flat " gyro has no sensitivity and is thus useless. A gyro must have a certain extension in the z- direction to operate.

A thick gimbal gyro according to figure 3.

If the gyro would be manufactured in a relatively thick silicium plate, the mass would have a certain extension in the z- direction. Such a gyro can be represented with eight mass points according to figure 3.

The mass points have a great component of velocity in the z- direction, but also a certain movement in the y- direction. The component of velocity in the z- direction does not give any contribution to the scale factor of the gyro, as we could see in the section above. Rotation around the z- axis gives rise to Coriolis forces in the x- direction, due to the movement in the y- direction of the mass points. The masses at the upper side of the gyro are acted upon by forces with the opposite direction to the forces, that act upon the forces at the lower side. The result will be a moment around the y- axis. Thus the gyro is sensitive to rotation around the z- axis.

Rotation around the x- axis gives rise to Coriolis forces in the z- direction, due to the movement in the y- direction of the mass points. The masses at the upper side of the gyro are acted upon by forces with the opposite direction to the forces, that act upon the forces at the lower side. The forces cancel each other, and the moment is zero.

Rotation around the y- axis does not give rise to any Coriolis forces due to the movement of the mass points in the y- direction. A movement in the z- direction gives Coriolis forces, but their total moment is zero.

Conclusion: A gyro with great extension in the z- direction is sensitive to rotation around the z- axis. The sensitivity is insignificant if the thickness is not great.

A thick unsymmetric gimbal gyro according to figures 4 and 5.

The sensitivity can be increased by designing the mass unsymmetrically, which can be done quite easily by anisotropical etching. The gyro is represented by eight mass points, but the two mass points at the left upper side are made heavier than those at the left lower side. To the right instead the lower side is made heavier that the upper side. The total mass is unchanged. Rotation around die z- axis gives an unchanged total moment.

Rotation around the x-axis gives Coriolis forces in the z- direction, due to the movement of the mass points in the y- direction. The masses at the upper side of the gyro are acted upon by forces in an opposite diection to the forces, that act upon the masses at the lower side. The unsymmetry gives rise to a moment around the y- axis.

Rotation around the y- axis does not give rise to any Coriolis forces due to the movement in the y- direction of the mass points. The movement in the z- direction gives Coriolis forces, but their total moment is zero.

Conclusion: The gyro will be sensitive to rotation around the z- axis as well as around the x- axis. Maximal sensitivity is at an axis between the z-axis and the x- axis.

Figure 6 shows a device according to the type I with two beams 1,2, provided with inertial masses 3,4 at the beam ends. The beams are fixed, at their free ends, to a frame 5, which in turn is carried by two torsional axes 6,7, formed in the same way as the beams. The torsional axes are fixed to the substrate 8,9.

Figure 7 shows an alternative embodiment of type I according to the invention. Here there are two beams 10, 11 extending longitudinally in the frame, fixed in same in both ends, and carrying inertial masses 11, 12 in their middles.

Figure 8 shows a plan view of a configuration of a vibrating gyro according to d e invention of type π, where the frame 13 is excited so that it tilts in the plane around an axis in the longitudinal direction of the beams/torsion axes 14,15 The mass 16 is in turn dependent upon two beams 17,18 which are rotated 90° relative the torsion axes 14,15. When the device is turned around an axis next to perpendicular to the substrate, the Coriolis force will act upon the mass 16 so that it will start swinging around an axis in the longitudinal direction of beams/torsion axes 17,18 as is indicated in the figure.

Figure 9 gives a schematic illustration of how the oscialltion activator and the oscialltion detector may be provided, connected to an electronic unit in a way, known per se

When manufacturing the device in a (100)-oriented silicon substrate the inclined beams are oriented along the /l 10/-directions. Furthermore, when manufacturing the device in a (100) silicon substrate, masking of the silicon surface is firstly made e g by thermic oxidation, which gives a protecting silicon oxide layer on the substrate surfaces. The silicon dioxide is patterned via a photolithographic process on both substrate surfaces in such a way, that the openings formed in the silicium oxide layer are liniated along the /HO/- directions, whereupon this layer is used as a mask at the anisotropic etching. Alternatively silicon nitride, gold, chromium and diverse other materials can be used as mask materials. The openings of the mask material on opposite sides are mutually displaced where a beam is desired in such a way, that a desired beam thickness is achieved after etching. The silicon substrate is then etched through from both sides of the substrate with an anisotropic etch, which selectively etches the (l l l)-planes, that form the beams, much slower than the (100)- and (110)- planes. For example, an etch, based upon ethylenediamine /pyrocathecol (EDP) or upon potassiumhydroxide (KOH) can be used. These etches mentioned are heated to achieve the desired etching effect, KOH to c:a 80° and EDP to c:a 110°.

The manufacture of the device in a semicondor substrate enables, that exciting, detection and electronics for signal treatment can be integrated in the same substrate as the device together with e g filter functions, other computer electronics or servo electronics for a device of the type with a closed loop.

In order to put the device in swinging active exciting is needed, and a number of techniques are available, capacitive, thermic, piezoelectric films etc.The method chosen is dependent upon the embodiment, encapsulation and the demend for precision. With capacitive exciting, the structure is usually excited via an oscillating electrical field between plate capacitors suitably placed. The exciting can be achieved directly between two movable mechanical parts, alternatively between one or several fixed parts and one movable part. Thermic exciting is achived in that the beam/beams is (are) heated locally via a resistance or a source for radiation, e g a pulsed laser so that a tension gradient is achieved in the beam/beams, and if the thermic exciting is created by a frequency, that is close to the natural frequency for the structure, the beam/beams can be made swing at the natural frequency. By depositing one or more layers of piezoeletrical films e g of ZnO to the inclined beams, that are contacted in a suitable way, and then by applying an electrical field over the piezoeletrical film, the beams can be brought to swing.

As is the case for exciting, a number of principles are avialable for detecting of the swinging, capacitive, resonant strain transmitters, piezoelectrical, piezoresistive etc. Dependent upon die embodiment and desired precision the principle is chosen In one embodiment of the type with closed loop, the detection of the oscillation generated by the Coriolis force is suitably made capacitively, as the feed back and possibly present bias is created by an electrical field between the beam/frame/mass and surrounding conductive plates. This gives access to capacitors on each side of the beam/frame/mass, the capacitance of which is varied wid the position of the device and a displacement can be detected with the aid of a capacitance bridge, which in turn gives a signal to a servo systeme, which changes tiie electrical field in such a way, that the beam/frame/mass is (are) returned to the zero position. The detection med resonant strain transmitter can be an interesting detection mediod if the device shall in any case be encapsulated in vacuum. These have a very high Q- value and give very high resolution. Piezoelectrical detection can be an interesting mediod if die exciting is in any case performed piezoelectrically so that this layer is available without any extra process step. A simple method to achieve detection of the movements of the inclined beams is by the aid of die piezoresistive effect, that resistors formed by doping of silicon substrate give. In order to get maximal sensitivity it is suitable to utilize a balanced bridge circuit, compising four resistors located in a suitable way. Furthermore, to get maximal piezoresistive effect, the resistors shall be placed where the strain in the beam, caused by the bend/turning that is desired to be detected, has its maximum. For special applications it is also possible to utilize optical detection of the movement of the beam/frame/mass, e g by utilizing interferece phenomena or by utilizing an arrangemang, where the position of a reflected light beam is dependent upon die position of the beam/frame/mass.

Even if the invention has been described with reference to two principal embodiments, the invention is not limited to diese. The invention is only Umited to what is given in the patent claims.