The present invention relates to a device for measuring pressure according to the preamble of claim 1.

The device is a pressure sensor which is particularly suitable for measuring pressure at high temperatures and in severe environments. The main embodiment can be seen in Fig. 1. The sensor consists of two conducting or semi¬ conducting plates 1 and 2, separated from each other by a thin spacing layer 3, which sealingly encloses a cavity 4 between the plates. The spacing layer is electrically insulating, and the two plates are provided with electrical contacts 5 and 6, which in turn are provided with electrical connections 7 and 8. When the structure of Fig. 1 is exposed to an external pressure, the plates 1 and 2 will be bent until there is galvanic contact between them inside the cavity. If an electrical voltage is applied between the connections 7 and 8, a current will pass through the connected circuit when the external pressure is sufficiently high for contact to occur between the plates. An example is shown in Fig. 2, where the current through a structure manufactured according to Fig. 1 is shown as a function of applied pressure. In this case, there is a sharp increase in current at ca. 1.7 bars when there is galvanic contact between the two plates 1 and 2.

Pressure sensors are preferably manufactured from semi¬ conducting materials and especially from silicon, since silicon technology permits batch manufacturing with a high degree of reproducibility between individuals. There is a plurality of inventions in which piezo-resistive elements have been made by doping in silicon, and then placed on a flexible membrane in order to achieve a pressure-sensitive resistance which can be utilized for measuring the pressure which has caused the bending of the membrane. Such sensors

have the drawback of not being able to operate at temperatures higher than about 150°C.

The structure shown in Fig. 1 can be used as a fibre- optical sensor if the width of the cavity 4 is adapted to the wave length of the light. An application based on this principle has earlier been filed, with No. 9201439-8. In comparison, the present device has the advantage of being very simple, and can therefore be produced at a significantly lower cost.

When the sensor body of Fig. 1 is manufactured using silicon technology, a thermically manufactured silicon dioxide is preferably used as the spacing material 3. Other possibilities are to use Chemical Vapor Deposition, CVD, sputtering, vaporization, or an epitaxial method. In these cases, several materials can be used, and the thickness of the spacing layer can for all the said methods be predicted with a very high accuracy. The manufacturing is achieved by providing one or both plates with a thin layer of the spacing material 3. A pattern is created in one or both layers, for example by means of a photo-lithographic standard method from semi-conductor technology, so that the spacing material is removed from the areas intended for cavities. When the plates are joined together, cavities are formed in those areas where the spacing material has been removed. The depth of the cavities is determined by the thickness of the spacing material, and can thus be predicted with a high degree of accuracy. The method for joining the two layers together is preferably so-called thermal bonding, described in patent application No. 9201439-8. This method provides a very strong and tight joint, suitable for applications in severe environments.

The curve shown in Fig. 2 has been measured in a structure where the plates 1 and 2 are in silicon with a resistivity

of about 10 ohmcm. The lateral form of the cavity is circular with a diameter of 1 mm, and the depth of the cavity is 1.4 μm. The sharp increase in current at the pressure p _k occurs when there is galvanic contact between the plates. The slope of the curve in this section is caused by the contact pressure increasing, whereby the contact resistance decreases between the plates. At pressures higher than p ₀ , there is a slower increase in current. This part of the curve reflects an increase in area of the contact surface between the plates 1 and 2. Fig. 3 illustrates the contact area between the disks 1 and 2 of the sensor structure in Fig. 1. The figure shows a common case, where one of the silicon plates 1 or 2 is significantly thinner than the other or where one of the plates has been attached to a supporting structure so that only one of the plates is bent when the sensor is exposed to an external pressure. When the pressure increases, the contact surface between the plates will increase, thus decreasing the contact resistance, and the current increases in case of a constant applied voltage.

The shape of the output signal is an essential feature of a sensor, and it is often desirable to adapt it in a suitable manner to the application of the component. When the sensor is used as a pressure guard, a relay function with a form according to curve a in Fig. 4 is desirable. When the sensor is used to quantify pressure, a linear function according to curve b-c of Fig. 4 is desirable. The latter curve-shape, which is closely reminiscent of the measured curve of Fig. 2 is transformed into curve-shape a if the contact pressure increases very rapidly with an applied external pressure and if the extension of the contact surface is very small. When quantifying the pressure one wishes to use part c of the curve. Its length can be maximized by decreasing p _k and p ₀ , and making sure that p ₀ =P _k - Such changes in shape of the characteristics can

be obtained by forming the geometry of the sensor in a suitable manner or by using suitably shaped resistive layers in the contact surfaces inside the cavity, as described below.

The magnitude of the pressure p _k , the pressure where galvanic contact is established between the two silicon plates, is determined by the dimensions of the sensor. Thicker plates 1 and 2, a smaller lateral extension of the cavity 4, or a greater depth of the cavity 4 cause a higher p _k . The closing pressure p _k can thus be varied by varying the thickness of one of the plates so that a membrane 10 is formed, as shown in Fig. 5. Using etching technology, very thin membranes can be obtained (Z.Xiao, S.Norrman and O.Engstrδ , Sensors and Actuators, A 41-42, 334, (1994)), for which reason the value of p _k can be varied within a large range.

One way of creating different curve-shapes according to Fig. 4 is by shaping the lateral form of the cavity in a suitable manner. The contact surface which is determined by the pressure, as illustrated in Fig. 3, is determined by the mechanical tension pattern which occurs in the plates

1 and 2, and varies with the lateral shape of the cavity. The component, the characteristics of which are shown in

Fig. 2, has a circular membrane. The lithographic method used to define the shape of the cavity permits easy manufacturing of triangular, square or arbitrary lateral shapes of the cavity, which can be used to influence the shape of the curve of Fig. 4 for pressures higher than p ₀ .

Another way of influencing the shape of the curve of Fig. 4 is to form the topography of the contacting surfaces of the plates 1 and 2 inside the cavity in various manners. For example, one or both surfaces can be provided with a protrusion 9 in the manner shown in Fig. 6, so that the

contact surface decreases, thus increasing the contact pressure. The protrusion can have different shapes with a crowned, pointed or plane contact surface against the opposite plate. A small contact surface causes a high contact pressure, for which reason a very rapid increase of p _k , followed by an essentially constant voltage according to curve-shape a in Fig. 4 is obtained.

Another way of influencing the shape of the curve in Fig. 4 is to, inside the cavity in one or both of the contact surfaces between the plates 1 and 2, shape resistive layers with different geometrical shapes. Fig. 7 shows a pair of examples, 11 and 12, of the geometrical shape of such resistive layers. The point 13 in Fig. 7 shows the initial point of contact when the plates 1 and 2 are brought into galvanic contact due to an external pressure. In case of an increasing pressure, when the contact surface between the plates increases, the increase in contact resistance varies depending on the shape of the geometry of the resistive layer. Shape 11 of Fig. 7 causes a slower decrease of contact resistance than shape 12. Preferably, the resistivity within the areas exemplified by 11 and 12 is smaller than that of the surroundings. The contact resistance will then primarily be determined by the surface within the areas 11 and 12. The areas 11 and 12 with differing resistance can be created by doping, if the plates 1 and 2 are in semi-conducting material, or by a cover of a thin layer of a resistive material, for example doped polychrystaline silicon for an arbitrarily chosen material for the plates 1 and 2.

Another version of pressure sensor within the same basic concept can be obtained if dopings are made in the cavity so that a transistor function is obtained. Fig. 8 shows a structure which is based on the same basic geometry as the version shown in Fig. 5. In Fig. 8, plate 2, which is

assumed to be a semi-conductor has been provided with dopings 18 and 19 of the opposite charge as compared to the charge of the doping of the plate. If for example plate 2 is of the p-type, the areas 18 and 19 are of the n-type. Between the areas 18 and 19 there is left an area of the same kind of charge as that of plate 2. The areas 18 and 19 are metallized with the layers 16 and 17, which are equipped with electrical connections 14 and 15. If, for example, the connection 14 is connected to ground and a positive electrical potential is applied to the connection 15 at the same time that a positive potential is applied to the connection 7, a current is obtained in the surface area between the areas 18 and 19 which can be controlled by the potential applied to the contact 7. The areas 18, 19 and 10 which are connected via 14, 15 and 7 respectively thus constitute source, drain, and gate respectively of a field- effect transistor. The area 20 between source 18 and drain 19, which is decisive for the function of the transistor has been magnified in Fig. 9. The current which passes through area 20 depends on the electrical field which is directed in a perpendicular direction towards the area 20. In case of a constant potential applied to the connection 7 this field will vary with the distance between the plates 1 and 2 which in turn depends on an external pressure. The current between source 18 and drain 19 thus depends on the externally applied pressure and the structure is a "pressure transistor" . Current-voltage characteristics of a field effect transistor are shown schematically in Fig. 10. For transistors of the standard kind, the current- voltage characteristics vary when the gate voltage is varied. In the case of a pressure transistor of the kind described here, the characteristic in case of constant gate voltage can, in a similar manner, be varied by varying the external pressure.

The above-described pressure transistor can be connected to simple transistor circuits in order to increase the sensitivity of a sensor system. Pressure guards with extremely well-defined closing pressures can for example be obtained by letting the transistor be one of the transistor elements in an invertor circuit or in a bi-stabile flip- flop. If the design according to Fig. 8 is manufactured using silicon technology, such circuits can easily be made on one and the same silicon wafer.

A further version of pressure sensor within the same basic concept can be obtained if one of the silicon plates, 1 or 2, for example in the embodiment shown in Fig. 5, is provided with a point inside the cavity as shown in Fig. 10. If an electrical voltage is applied between 7 and 8, this will cause an electrical field constriction around the point 21 so that charge carriers are injected from plate 1 to plate 2 due to field emission. This phenomen is also called Fowler-Nordheim tunneling and is well-known from many areas of physics. A charge carrier which is inside a solid material normally has to have an energy Φ _β in order to be able to exit the material. If there is a sufficiently large electrical surface field, the charge carriers can however exit the material due to tunneling and cause a tunneling current J, which can be expressed as

J = A E ² exp [ - B ( sΦ-V _B B)' ^3/2

where A and B are constants, and E is the electrical field on the surface. Applied to the sensor construction in the present description, the current J occurs between the connections 7 and 8 of Fig. 10. The magnitude of the electrical field is determined by the voltage drop between the plates 1 and 2 inside the cavity, the field

displacement at the point 21, and the distance between the point 21 and plate 2. The distance d between the point 21 and plate 2 as shown in Fig. 12 varies with the pressure, for which reason the electrical field at the point 21 varies with the pressure in case of an applied constant voltage between 7 and 8. From equation (1) it can be seen that the current J is a very sensitive indication of the field E, which in turn depends on , which means that the current between 7 and 8 will become a very sensitive measurement of the applied pressure.

The cavities in the structures which have been described in connection to Figs. 1 to 12 have been assumed to be closed. If parts 1 and 2 are joined together in a vacuum, sensors for measuring absolute pressure are obtained in this way. A sensor for measuring differential pressure relative to a medium with a reference pressure can easily be obtained by creating a cavity which is not closed, in contact with the medium with the reference pressure. An example is shown in Fig. 13 where part 2 has been provided with holes. If the pressure p, in the one area 22 on the one side of the wall 24 is to be measured relative to the pressure p _ref in the other area 23 on the other side of the wall 24, one arranges for the medium in the area 23 to be in contact with the cavity via the holes 25 in part 2.