WO2015119564A1

WO2015119564A1 - Controlling pressure in cavities on substrates

Info

Publication number: WO2015119564A1
Application number: PCT/SE2015/050130
Authority: WO
Inventors: Håkan WESTIN; Peter Wickert
Original assignee: Silex Microsystems Ab
Priority date: 2014-02-07
Filing date: 2015-02-06
Publication date: 2015-08-13
Also published as: EP3102534A4; SE1450135A1; SE538346C2; EP3102534A1

Abstract

The invention relates to a new method of providing different and controlled atmospheres inside cavities on one and the same chip of a MEMS device. With the new method one can refrain from providing separate getter materials, and thereby the manufacturing is simplified. The method comprises providing a first substrate (20) and a second substrate (28) and making at least one depression (22', 22'') in at least one of the substrates(20). A structural component (26) is provided in or on a surface of at least one of the substrates, said structure containing entrapped, absorbed or adsorbed ions, molecules or atoms of a gas. The substrates are bonded together such that a cavity forms and becomes hermetically sealed. The obtained structure is subjected to conditions so as to release the implanted, absorbed or adsorbed gas atoms, ions or molecules from the substrate to provide said controlled atmosphere inside the cavity.

Description

CONTROLLING PRESSURE IN CAVITIES ON SUBSTRATES

The present invention relates to MEMS engineering and in particular to

manufacture of devices having multiple components in closed cavities on the same chip, at least some cavities requiring a different atmosphere.

Background of the Invention

In the continuous strive to reduce size of micro components for use in e.g.

mobile phones and other electronic equipment, the packaging of functions of different types on one and the same chip is an important requirement.

For example it is desirable to provide both accelerometers, gyros, oscillators and other MEMS components such as IR sensor detectors on the same chip in separate sealed cavities. Such components comprise movable members or sensor elements relying on thermal conductivity which each requires a specific atmosphere in order to function properly. Some components require a damping atmosphere in order not cause unwanted self-oscillation, whereas others require vacuum or at least very low pressures in order to function properly. Some devices like gas sensor absorption requires over pressure inside the cavities.

These requirements are contradictory and require special measures in the manufacturing of the devices in question. The state of the art offers a technology of using so called "getters" inside cavities in order to handle and control the atmosphere inside closed cavities.

Getter materials are primarily used for preventing a long term increase of cavity pressure caused by degassing from the wafer material. The getter absorbs released gas molecules and thus the "nominal" pressure can be preserved. The getter technology has been used by e.g. Fraunhofer (US-8,546,928) to provide a solution to the problem of providing cavities with different pressures. In this patent there is disclosed a multiple component which is to be

subsequently individualized by forming components containing active

structures, in addition to a corresponding component which can be used in microsystem technology systems. The multiple component and/ or component comprises a flat substrate and also a flat cap structure which are bound to each other such that they surround at least one first and one second cavity per component, which are sealed against each other and towards the outside.

The first of the two cavities is provided with getter material and due to the getter material has a different internal pressure and / or a different gas composition than the second cavity. Instead of providing getter materials inside cavities, it is also possible to provide a gettering function by transforming the internal surfaces of the cavity so as to create reactive voids.

US-5,840,590 discloses impurity gettering in silicon wafers which is achieved by a new process consisting of helium ion implantation followed by annealing. This treatment creates cavities whose internal surfaces are highly chemically reactive due to the presence of numerous silicon dangling bonds. For two representative transition-metal impurities, copper and nickel, the binding energies at cavities were demonstrated to be larger than the binding energies in precipitates of metal silicide, which constitutes the basis of most current impurity gettering. As a result the residual concentration of such impurities after cavity gettering is smaller by several orders of magnitude than after precipitation gettering. Additionally, cavity gettering is effective regardless of the starting impurity concentration in the wafer, whereas precipitation gettering ceases when the impurity concentration reaches a

characteristic solubility determined by the equilibrium phase diagram of the silicon- metal system. The strong cavity gettering was shown to induce dissolution of metal- silicide particles from the opposite side of a wafer. US-7,564,082 relates i.a. to a semiconductor structure, comprising a gettering region proximate to a device region in a semiconductor wafer. The gettering region includes a precisely-determined arrangement of a plurality of precisely- formed voids through a surface transformation process. Each of the voids has an interior surface that includes dangling bonds such that the plurality of voids getter impurities from the at least one device region. The structure includes a transistor formed using the device region. The transistor includes a gate dielectric over the device region, a gate over the gate dielectric, and a first diffusion region and a second diffusion region formed in the device region. The first and second diffusion regions are separated by a channel region formed in the device region between the gate and the proximity gettering region.

I DE 10 2012 202 183 Al there is disclosed a micro-mechanical structure having a cavity in which there is provided a layer of material for releasing a gas into the cavity.

In US 201 1/0048129 Al there is disclosed an inertial sensor comprising a gas- generating material applied inside a cavity to provide a desired pressure inside said cavity.

Summary of the Invention

The inventors have now devised a new method of providing different and controlled atmospheres inside cavities on one and the same chip of a MEMS device. With the new method one can refrain from providing separate getter materials, and thereby the manufacturing is simplified.

The new method is defined in claim 1. In another aspect there is also provided a MEMS structure/ device having at least two cavities with different pressure prevailing in said cavities. This structure/ device is defined in claim 13. Further scope of applicability of the present invention will become apparent from the detailed description given hereinafter and the accompanying drawings which are given by way of illustration only, and thus not to be considered limiting on the present invention. Brief Descripition of the Drawings

Fig. 1 illustrates a prior art device;

Fig. 2 schematically shows a first step of an embodiment of a process;

Fig. 3 schematically shows a second step of the process; Fig. 4 schematically shows the result of the process of Figs. 1 and 2; Fig. 5 illustrates an alternative embodiment of a process; Fig. 6 shows the result of the process of Fig. 5; Fig. 7 illustrates another embodiment of a process;

Fig. 8 shows the result of the process of Fig. 7; Fig. 9 illustrates a still further embodiment of a process; Fig. 10 shows the result of the process of Fig. 9;

Fig. 11 schematically illustrates another alternative process; Fig. 12 shows the result of the process of Fig. 11 ;

Fig. 13 schematically illustrates another alternative process;

Fig. 14 shows the result of the process of Fig. 13;

Fig. 15a-f shows various possible combinations of the described processes; Fig. 16a-b shows a further embodiment;

Fig. 17 shows a still further embodiment; and

Fig. 18 is a graph demonstrating the effect of the invention using the embodiment in Fig. 16.

Detailed Description of Preferred Embodiments

For the purpose of this application the term "structural component" shall be taken to mean at least a portion of a substrate that has been subjected to a physical and / or chemical change of the surface thereof, either by actually changing the structure of the material in the substrate as such by implanting, absorbing or adsorbing species such as atoms, molecules or ions, or by depositing a material on the substrate, said material also comprising implanted, absorbed, embedded or adsorbed species. Thus, in particular MEMS structures such as gyroscopes, accelerometers etc. are included in this definition.

For the simultaneous housing of a plurality or multiplicity of microsystems containing active structures, for example sensor systems, with different operating pressures and at the wafer or similar level, a corresponding wafer or similar component is provided, containing corresponding active structures, for example, a wafer with one or typically several sensors to be housed. Fig. 1 shows the design of a sensor system according to prior art (US-- 8,546,928) having two sub-modules which are hermetically separated from each other. Micromechanical sensor subsystems 3 and 4 are located on a sensor wafer 1 , the surface of which is essentially flat. Additionally provided is a cover, for example one cap wafer 2, having depressions or recesses corresponding to the sensor areas such that cavities are established when the cover and wafer are bonded together, i.e. solidly connected with sensor wafer 1 through a bonding process. Bonding frame 7 encloses the sensor areas and seals these

hermetically against the outside.

The arrangement of active structures and of the cavities can differ from the arrangement shown in this Figure. As an example, the active structures, for example sensors, may be located in recesses within the sensor wafer with the cap wafer having a flat surface or minimum recesses, depending on the space requirements of the active structures. The active structure may also be located in the cap wafer such that the variants described above would be realized as a mirror image of the shown structure.

The production gas atmosphere, either defined by the bonding process or suitably selected, is encapsulated into the cavities or recesses when both wafers have been connected. The production gas atmosphere consists of at least one gas type A. On activating the first getter material within the first cavity, and depending on the amount of the getter material, this gas type will be at a minimum partially absorbed or will be completely absorbed or will be essentially completely absorbed, and a (partial) vacuum is created within this cavity.

Preferably, the production gas atmosphere will consist of a minimum of two gas types A and B with different reaction characteristics for the first getter material.

Either the cap wafer 2 or the wafer 1 containing the active structure exhibits a recess such that both wafers can be bonded together forming cavities. For each corresponding microsystem component, at least two cavities 5, 6 are provided, however, more cavities can be provided depending on requirements. Preferably within the interior of the cap wafer, in the region of the cavity, a getter material 8 is located such that after completion of the bonding the getter material is only located in the first of the at least two cavities 5, 6. The getter material in the first cavity 5 can absorb molecules of a first gas type A (for example, ¾, O₂, CO₂ or N₂ or any mixture of these gases) . The getter material normally will not react with molecules of a second gas type B (for example, inert gases such as Ar or Ne) . As is known as the state of the art and as discussed above, the getter material will typically be deposited within the cap wafer in an inactive state. The activation of the getter material is typically accomplished through a

temperature-time-process, as is known in the state of the art. After this activation, molecules of type A will have been gettered (absorbed) . After the activation of the getter material, the gases of type A in cavity 5 are absorbed such that the cavity pressure is defined through the remaining molecules of type B. The original cavity pressure defined as the sum (of the partial pressures) of particles of type A and B remains in the second cavities without getter material. The remaining pressure in each of the different subsystems can be determined via the composition of the original gas mixture (A+B) and the amount and type of getter material in at least the first cavities, which must be chosen to be different from the amount and/or type of getter material in the second cavities (inasmuch any getter material is located there at all).

Getter material is brought into the cap of the first cavities in such quantities that after the getter activation, the particles of gas type A have been completely or essentially completely absorbed. Therefore, first cavities 5 will have only (or essentially only) particles of gas type B (or, if the partial pressure of B is zero or almost zero, cavities 5 will have more or less an almost absolute vacuum) , while all gas molecules of type A and, if applicable, type B remain in the gas volume of second cavities 6.

Alternatively, getter material is brought into first cavities 5 in quantities which are not sufficient to completely absorb gas type A, but which will only absorb a fraction of x mol percent (mol-%). After activation of the getter, the gas

atmospheres within the cavities are different such that--due to the reduced amount of getter material and the resulting incomplete gas absorption--the first cavities contain ( 100-x) mol-% of type A plus the total amount of type B, while the gas mixture within the second cavities remains unchanged. This procedure allows adjusting arbitrary pressure ranges. Alternatively, the MEMS component (for example a chip) has two or more cavities 5, 6, with getter material in both cavities or, if there are more than two cavities per component, in at least two of the component cavities. In these cases, the area of the getter material and / or its gas absorption characteristics in both cavities are different such that both or at least two cavities have different end pressures and/or gas compositions after getter activation.

The getter material can be placed into the cavity in any preferred arrangement, for example, as strips or as a surface area or as a structured shape. Preferably, it will be placed onto the cap of the wafer or the like, for example, within its recesses if these exist. Alternatively, the getter material may also be placed on the side of the substrate, for example, to the side of the active structures or even below these structures, if these areas are not needed otherwise.

For components to be produced with more than two cavities, the gas mixture may be composed again of two gas types A and B, and, for example, the first cavities having a getter material in a quantity such that gas type A will be absorbed completely or nearly completely after getter activation, while the second cavities having a getter material which absorbs gas type A within the mixture of (A+B) to a different percentage than the first getter material, and while the third cavities having no getter material at all. Alternatively, the gas mixture may be composed of three or even more gas types A, B, C, . . . . In this case it is advantageous, to place a getter material having one first absorption characteristic with respect to the gas mixture within the first cavity, and another getter material having one second absorption characteristic with respect to the gas mixture within the second cavity. A gas mixture may be composed

exemplary of gas types CO₂, N₂ and Ar. One getter material with one first absorption characteristic may absorb carbon dioxide, but not or only

insignificantly absorb nitrogen. One getter material with one second absorption characteristic may absorb nitrogen and carbon dioxide. A third cavity may remain free of any getter material. As has been indicated already, the production of MEMS components is preferably done as multiple components, for example as wafers, with the wafer or other multiple component device being individualized into single components (for example, chips). Alternatively, the components can obviously also be constructed using a single substrate (for example, a base chip) , suitably chosen to carry the active structure (s), and one cap (for example, a cap chip) which simultaneously covers the at least two cavities and hermetically seals these from each other.

The present invention is able to provide devices having the same functionality as the device according to prior art illustrated in Fig. 1 , but with simpler

manufacturing and at a lower cost. Thus, the invention relates to a new method of providing different and controlled atmospheres inside cavities on one and the same chip of a MEMS device. With the new method one can refrain from providing separate getter materials, and thereby the manufacturing is simplified. The method comprises providing a first substrate and a second substrate. A structural component is provided in or on a surface of at least one of the substrates, said structure containing entrapped, absorbed or adsorbed ions, molecules or atoms of a gas. The substrates are bonded together such that a cavity forms and becomes hermetically sealed. The obtained structure is subjected to conditions so as to release the implanted, absorbed or adsorbed gas atoms, ions or molecules from the substrate to provide said controlled atmosphere inside the cavity.

The pressure in the cavity with the highest pressure is preferably 100 times the pressure in the low pressure cavity, preferably 1000 times higher. The low pressure cavity preferably has a pressure of 1 Pa or less.

The basic idea behind the present invention is to use to advantage the inherent problem of degassing from substrates, which normally is the reason for using getters to eliminate the degassed molecules from a cavity atmosphere. This is achieved by implanting, in a controlled manner, gas molecules (e.g. ¾, O2, CO2 or N2) or atoms (e.g. Ar, Ne, Xe) into a substrate, more precisely in that part of the structure constituting the cavity to be, before the capping process to provide the sealed cavities. Within the inventive concept the implantation of gas species can also be made on structural parts provide inside the cavity. Other gases that are usable can be selected from any gas that is known to degas from materials in vacuum. A non-exhaustive list is (in addition to the ones mentioned above) He, CO, CH₄, H₂0, C₂H₆ and C₃H₈.

There are many possible alternatives to achieve a desired structure, and a few preferred embodiments will be described below.

In its most generic form the method comprises providing a first silicon substrate and a second substrate; making at least one depression in at least one of the substrates; creating a structural component in or on a surface or in a cavity of at least one of the substrates, said structure containing entrapped, absorbed or adsorbed ions, molecules or atoms of a gas; bonding the substrates together such that a cavity forms and becomes hermetically sealed; subjecting the obtained structure to conditions so as to release the implanted, absorbed or adsorbed gas atoms, ions or molecules from the substrate to provide a desired atmosphere inside the cavity.

Thus, in a first embodiment of the present invention, reference being made to Figs. 2-6, a first wafer 20 is pre-prepared by making depressions 22 in its surface, these depressions constituting at least a part of a cavity in which a functional component subsequently is to be provided, see Fig. 2. Such

depression can form the entire cavity if subsequently a lid wafer is bonded to the first wafer, as shown in Fig. 4, but it can also be complementary to a depression in the lid wafer so as to form a cavity having an extension into both wafers. In a subsequent step, see Fig. 3, some depressions 22' are protected with a mask 24, exposing the remaining depressions 22". Now e.g. a PVD or

implantation process (schematically indicated by arrows in Fig. 3) is applied to implant e.g. argon (Ar) into the wafer material in the depressions 22" exposed through the mask 24. The implanted argon is schematically shown by hatching at 26. The implanting process is suitably performed by so called Physical Vapour Deposition or "PVD", commonly indicated as "sputtering". Other variants of PVD include Cathodic Arc Deposition, in which a high-power electric arc discharged at the target (source) material blasts away some into highly ionized vapour to be deposited onto the workpiece; Electron beam physical vapour deposition, in which the material to be deposited is heated to a high vapour pressure by electron bombardment in "high" vacuum and is transported by diffusion to be deposited by condensation on the (cooler) workpiece; Evapourative deposition, in which the material to be deposited is heated to a high vapour pressure by electrically resistive heating in "low" vacuum; Pulsed laser deposition, in which a high-power laser ablates material from the target into a vapour; Sputter deposition, in which a glow plasma discharge (usually localized around the "target" by a magnet) bombards the material sputtering some away as a vapour for subsequent deposition.

Also CVD processes can be used for specific materials, if desired, especially insulators, as well as any other thin film deposition technique well known in the semiconductor industry.

The mask is removed and a lid wafer 28 is bonded, see Fig. 4, suitably by fusion bonding, but other bonding methods can also be used, e.g. eutectic bonding if temperature is an issue (e.g. if metal is present than fusion bonding is not possible).

Normally the bonding is performed under vacuum and as a consequence, in the cavities 22' in which there has been no argon implanted a vacuum will prevail. If then the sealed structure is subjected to mild heating the implanted argon will release in the cavity 22" to provide an atmosphere with a pressure that is different from the pressure in the other cavities.

In an alternative process both depressions are subjected to implanting by PVD but the level of implantation is different. This is illustrated in Fig. 5, wherein after the step shown in Fig. 3, a new mask is provided that covers those depressions 22' in which implantation by sputtering or other means already has been made, and the other depressions 22' are now subjected to a PVD process. However, the degree of implantation, controlled by exposure time and/ or intensity is lower in that second step than in the first step. The lower degree of implantation is schematically illustrated by a less dense hatching 29. After bonding a lid wafer 30 to the first wafer 20 and releasing of argon inside the cavities, the result will be cavities with argon atmospheres with different pressure prevailing inside the cavities, as shown in Fig. 6, i.e. a lower pressure in cavity 22' and a higher pressure in cavity 22".

A still further alternative is to implant argon 72 into a flat first substrate 70 in a defined area corresponding to cavities to be made, by performing PVD, as shown in Fig. 7a. Thereby an appropriate hard mask 74 (or a resist) is provided over the substrate 70. Subsequently a lid substrate 76 is bonded onto the first substrate 70, the lid substrate being provided with depressions 78', 78" corresponding to the depressions 22', 22" in the previously described

embodiment, whereby a structure as shown in Fig. 8a is obtained, which is similar to the structure shown in Fig. 4.

It is also possible to implant the desired gas over the entire surface of a substrate 70' as shown in Fig. 7b, and then bonding a lid 76' having a depression 78"' forming a cavity as shown in Fig. 8b. In Figs. 9 and 10 a still further embodiment is illustrated. Here, a metal 31 of a selected material is deposited onto the bottom of one depression 22' in the substrate 20. The material is selected to have the capability of embedding or retaining ions, atoms or molecules of a gas within its structure. The material can suitably be a metal, but also oxides are usable. Suitably the carrier gas used in the deposition is the gas of choice. In particular argon is useful, since it is a commonly used carrier gas in e.g. sputtering. Thereby the film of the deposited material will inevitably contain gas ions / atoms / molecules trapped in the metal, and since argon is often used as a carrier gas in deposition processes argon will be implanted. If no measures are taken to remove the argon from the metal, once the cavity is appropriately sealed with a lid substrate 32 and the structure is subjected to appropriate conditions, such as mild heating, the argon will degas from the metal and an atmosphere containing argon gas will prevail inside the cavity. Of course, deposition of a film of a suitable material, e.g. metals, insulating materials or semi-conducting materials, can be made in both depressions at different intensities/ exposure times to provide films of different properties, i.e. different argon concentration, so as to provide different pressures in the cavities, once they are sealed, similar to the embodiment shown in Figs. 5 and 6.

Obviously, the film can be deposited on the lid substrate instead, similar to the embodiment of Figs. 7 and 8.

In these embodiments fusion bonding not possible if metal is deposited since it would not withstand the temperatures required without melting.

A further possibility is to provide a compartment 1 10 in which a suitable gas, such as argon or other noble or inert gas, can be provided at an over pressure and wherein a substrate 20 having depressions 22', 22" is placed, see Fig. 1 1. The compartment can be the housing of a standard bonder apparatus. Provided the conditions are appropriate, i.e. high enough pressure and suitably an elevated temperature and exposure for a time long enough, the gas will be adsorbed on or absorbed in the surface to a sufficient degree that when the substrate is provided with a lid and thus the cavities are sealed, and the structure thus obtained is subjected to suitable conditions, the gas will release from the surface and a controlled atmosphere is provided inside the cavities 22', 22", see Fig. 12. The suitable conditions can include heating to an elevated temperature. As an option, the field surfaces 1 12 of the substrate 20 can be masked in order that the adsorption of gas be selectively performed in the depressions 22', 22". A mask 1 14 is indicated in broken lines in Fig. 1 1.

If it is desirable to provide different atmospheres in different cavities, it is possible to let the mask 1 14 cover selected depressions 22' (see Fig. 13) during a first exposure and then remove the mask and continuing the exposure, whereby the concentration of adsorbed gas will be lower in the depression 22' than in the depression 22", see Fig. 14. In Fig. 15a-g several possibilities of combining substrate materials and variations in the exposing of the substrates are shown. Reference numerals are only given in Fig. 15a since all elements are the same in the figures.

Thus, in Fig. 15a a silicon substrate 150 having a depression 152 is exposed to e.g. argon which is implanted, absorbed or adsorbed, and a lid 154 is provided, which could be of any material that can be bonded to the silicon to provide hermetic sealing, e.g. glass or silicon.

In Fig. 15b the substrate can also be of silicon but could also be of e.g. glass. The lid is made of silicon and the argon is implanted, absorbed or adsorbed therein.

In Fig. 15c both substrate and lid are of silicon and both are exposed to argon to provide implanted, absorbed or adsorbed argon.

In Fig. 15d both substrates have been provided with depressions, and only the substrate having a depression is made of silicon and exposed to argon.

Again, in Fig. 15e the lid is made of silicon and has been exposed to argon, and the other substrate can be of some other material, e.g. silicon or glass.

Finally, in Fig. 15f both substrates are made of silicon, both have depressions and both have been exposed to argon. In Fig 16a-b there is shown a structure that is the subject matter of the

EXAMPLE below.

This structure comprises a first substrate 150 built from one wafer 150' having two depressions 151a, 151b on which there has been provided a further wafer 150" comprising two MEMS components 152a, 152b in the form of mechanical resonator structures (such as a gyroscope sensor element) free-hanging over the cavities 151a, 151b, respectively.

On one of the MEMS components 152a the Ar implantation is selectively made (shown by hatching) using suitable masking.

After bonding the structure shown in Fig 16b is obtained. This structure is used in the EXAMPLE. The structures that can be made by the invention are usable in connection with MEMS devices in which there are functional components that require controlled atmospheres to function properly.

In particular the invention is useful if different atmospheres are required in different cavities of a MEMS device and on the same chip. This is in particular useful if components such as gyros and accelerometers, which require different atmospheres, are to be provided in close proximity on one chip.

Including also but not limited to other MEMS devices such as, Resonators and timing devices, RF-switches, IR sensor bolometers or other thermal sensors relying on the heat conductivity of the surrounding atmosphere.

The components in question can be built on either a lid wafer or on a wafer on which a cavity is provided, in which case the component may be provided on the bottom of the cavity.

The embodiments described and shown in the figures are only exemplary of the invention which is limited only by the terms of the claims.

Thus, one could for example envisage the provision of a cavity with argon atmosphere and one with only vacuum.

One other possible variation is to use different gases for different cavities. Thus, it is conceivable to have e.g. argon in one cavity and nitrogen in another.

Also gases with different thermal conductivity properties are required for some MEMS applications. In Fig. 17 a still further embodiment is shown.

Here the cavities in the final structure are not made by forming depression which are closed by bonding the substrates together.

Instead a plurality of areas (two shown) on a surface of a substrate 170' are delimited by a respective closed loop of a suitable bonding material having a finite thickness, preferably in the range 0, 1 - 5 μηι, which thus forms a wall 172 enclosing each area 174a, 174b, and also form a distance member providing the required volume in which a MEMS component (not shown) is moveable. Thus, for the purpose of the application the substrate 170' and the walls 172 together form a first substrate. One of these enclosed areas 174a is subjected to an implanting process to provide e.g. Ar entrapped, adsorbed or absorbed on the surface.

A second substrate 170" (optionally having depressions 174a, 174b, although such are not required), is bonded to the first substrate 170 (comprised of substrate 170' and walls/ distance members 172) whereby a wafer having a plurality of cavities of two kinds, namely one with Ar and one without Ar.

EXAMPLE This example demonstrates the effect according to the invention.

A test production of two wafers using the invention was performed. Both wafers (Wafer 1, Wafer 2) were made in the same way, i.e. using the method according to the invention.

Thus, a first and a second silicon substrate in the form of standard 8" wafers are provided. A plurality of depressions of two different kinds are made in the first substrate by a masking and etching procedure using standard lithography. The first substrate also comprised a MEMS structure in the form of a

mechanical resonator, similar to a gyroscope, arranged to be freely movable over the cavity. This was made for the purpose of this example in order to enable the measurement of internal pressure in the cavities. In practice a Q -value was measured and a pressure corresponding thereto was calculated.

The first substrate was then subjected to implanting with Ar ions, using a shadow mask to implant Ar on selected Si structures thereon, namely on a surface of the gyroscope, such that Ar species became entrapped (implanted) in the surface of the structure on a limited region thereof,

The second substrate was provided with a depression at a location

corresponding to the location on the first substrate where the gyroscope was located in order to allow free movement of the gyroscope.

The two substrates were then placed on each other to form a wafer having the resonator structure freely movable in a cavity. The substrates were bonded together by subjecting them to an elevated temperature (< 400 °C) such that a cavity formed between the substrates and became hermetically sealed, thus forming the test object, Wafer 1.

In the process of bonding the structure, i.e. the substrates forming wafers with two hermetically sealed cavities, wherein one of the cavities had one interior surface comprising implanted Ar species, were subjected to conditions whereby the implanted Ar atoms were released from the substrate to provide said controlled atmosphere inside the cavity.

This procedure was repeated to form a second wafer, Wafer 2. Table 1 (and Fig. 18) show the pressures in the two respective cavities for both wafers, Wafer 1 , "X", and Wafer 2, "+", respectively.

As can be clearly seen the cavity in which Ar atoms were released, shown at B in Fig. 18, from the surface of the gyroscope has a distinctly higher pressure than the other cavity, shown at A in Fig. 18, and as a consequence also a significantly lower Q -value.

The same qualtitative result was obtained for both wafers, which demonstrates the reproducibility of the method.

TABLE 1

Wafer! :

Wafer2:

Die# Ar implanted Estimated Pressure (mTorr) Estimated Pressure (Pa)

Die1 Yes 1 .23E+04 1 .64E+03

Die2 Yes 1 .03E+04 1 .38E+03

Die3 Yes 4.86E+03 6.48E+02

Die4 Yes 4.42E+03 5.90E+02

Die5 Yes 7.79E+03 1 .04E+03

Die6 No 1 .59E+02 2.12E+01

Die7 No 9.81 E+01 1 .31 E+01

Die8 No 3.86E+01 5.15E+00

Die9 No 1 .48E+02 1 .97E+01

Claims

CLAIMS:

1. A method of providing a controlled atmosphere in cavities in silicon based devices, the method comprising providing a first substrate and a second substrate; creating a structural component in or on a surface of at least one of the substrates, said structure containing entrapped, absorbed or adsorbed ions, molecules or atoms of a gas; bonding the substrates together such that a cavity forms and becomes hermetically sealed; subjecting the structure to conditions so as to release the implanted, absorbed or adsorbed gas atoms, ions or molecules from the substrate to provide said controlled atmosphere inside the cavity.

2. The method according to claim 1 , wherein the structural component is created by ion bombardment.

3. The method according to claim 1 , wherein the structural component is created by providing said gas at an overpressure in a confined chamber such that gas is adsorbed on or absorbed in the substrate surface.

4. The method according to claim 1 , wherein the structural component is created by sputtering a material, preferably a metal or an oxide, onto one of the substrates using a carrier gas, whereby the metal or oxide entraps carrier gas atoms or molecules, that subsequently are released by subjecting the final structure to said conditions.

5. The method according to claim 1 , wherein a cavity is formed by making at least one depression in at least one of the substrates and closing said

depression by bonding the substrates together.

6. The method according to claim 1 , wherein the cavity is formed by providing distance elements, suitably in the form of bonding material having a finite thickness, preferably in the range 0, 1 - 5 μπι.

7. The method according to any of claims 6 or 7, wherein the depression is made in the first substrate.

8. The method according to any of claims 6 or 7, wherein the depression is made in the second substrate.

9. The method according to of claims 6 or 7, wherein a depression is made in both substrates.

10. The method according to claim 1 , wherein the structural component is created only on selected parts of the substrate, namely where the cavities are to be provided.

1 1. The method according to claim 1 , wherein the structural component is created only in a depression.

12. The method according to claim 1 , wherein different parts of the substrate are provided with different structural components comprising different amounts of gas so as to create a difference in the concentration of absorbed / adsorbed gas species in said different parts.

13. The method according to claim 12, wherein the different parts are different depressions.

14. The method according to claim 12, wherein the different parts are different surface areas of a substrate.

15. The method according to any of claims 1-7, wherein the structural component covers the entire substrate.

16. The method according to any preceding claim, wherein the gas is selected from H₂, 0₂, N₂,C0₂, CO, He, Ne, Ar, Xe, CH₄, H₂0, C₂H₆ and C₃H₈.

17. A MEMS structure/ device comprising a first and a second substrate bonded together, and a cavity provided between the substrates, each cavity having an atmosphere that is different from the other, wherein no cavity comprises any element having a gettering function.

18. The structure according to claim 17, wherein the atmosphere in one cavity is a gas and in the other vacuum.

19. The structure according to claim 17, wherein the pressure in the cavity with the highest pressure is 100 times the pressure in the low pressure cavity, preferably 1000 times higher.

20. The structure according to claim 17, wherein the low pressure cavity has a pressure of 1 Pa or less.

21. The structure according to claim 17, wherein the atmosphere in one cavity is a first gas at a pressure and in the other the atmosphere is a second gas at a pressure.

22. The structure according to claim 18, wherein the gas in the first cavity and in the second cavity is the same gas but the pressure in the cavities differ.

23. The structure according to claim 18, wherein the pressure in the first cavity differs from the pressure in the second cavity.