METALLIC DEVICE HAVING MOBILE ELEMENT IN A CAVITY OF THE BEOL OF AN INTEGRATED CIRCUIT

Abstract

In order, for example, to improve the ohmic contact between two metal pieces located at a metallization level, these two metal pieces are equipped with two offset vias located at the metallization level and at least partially at the via level immediately above. Each offset via comprises, for example, a nonoxidizable or substantially nonoxidizable compound, such as a barrier layer of Ti/TiN.

Claims

1. A method, comprising: etching a conductive material in a first metallization level to form a first element and a second element configured to be mobile relative to each other; depositing an insulating material over the first element and over the second element in a first via level, the insulating material being further disposed in the first metallization level between the first element and the second element; forming a first opening and a second opening in the insulating material, the first opening and the second opening having a first sidewall disposed in the first via level and contacting an upper surface of the first element and the second element, respectively, the first opening and the second opening further having a second sidewall disposed in the first via level and the first metallization level, the second sidewall in the first metallization level being spaced apart from sidewalls of the first element and the second element; and lining the first opening and the second opening with an electrically conductive material resistant to formation of an electrically insulating compound at surfaces of the electrically conductive material in contact with the insulating material.

2. The method of claim 1, further comprising removing portions of the insulating material in contact with the electrically conductive material to form a housing filled with a fluid.

3. The method of claim 2, wherein the fluid comprises air.

4. The method of claim 1, wherein forming the first opening and the second opening in the insulating material comprises a plasma etching process.

5. The method of claim 1, wherein lining the first opening and the second opening with the electrically conductive material comprises a conformal deposition process.

6. The method of claim 5, wherein the conformal deposition process comprises a chemical vapor deposition process.

7. The method of claim 5, further comprising etching, prior to the lining, surfaces of the first element and the second element exposed by the first opening and the second opening to remove a native oxide formed on the surfaces of the first element and the second element.

8. The method of claim 7, wherein etching exposed surfaces of the first element and the second element comprises a plasma etching process.

9. The method of claim 1, wherein the electrically conductive material comprises an oxidation-resistant material.

10. The method of claim 1, wherein the electrically conductive material comprises at least one of titanium, titanium nitride, tantalum, tantalum nitride, tungsten, ruthenium, or gold.

11. A method, comprising: forming a first element in a first metallization level, the first element being configured to be a mobile element and comprising a first conductive material; depositing a first insulating layer over the first element, the first insulating layer extending partially into a first via level disposed over the first metallization level; etching an opening in the first insulating layer to expose a sidewall of the first element; conformally depositing a second conductive material in the opening to form sidewalls of an appendage, the sidewalls of the appendage physically contacting the sidewall of the first element and extending partially into the first via level; and removing remaining portions of the first insulating layer disposed over the first element and the appendage to expose the first element and the appendage to a fluid, the second conductive material being different from the first conductive material, the second conductive material being resistant to formation of an electrically insulating compound at surfaces in contact with the fluid.

12. The method of claim 11, further comprising filling the opening with a third conductive material to form a center of the appendage, the third conductive material being different from the second conductive material.

13. The method of claim 12, wherein at least one of the first conductive material or the third conductive material comprises copper.

14. The method of claim 11, wherein the fluid comprises air.

15. The method of claim 11, wherein the second conductive material comprises at least one of titanium, titanium nitride, tantalum, tantalum nitride, tungsten, ruthenium, or gold.

16. The method of claim 11, wherein the first element comprises a mobile element in a thermally activatable structure.

17. A method, comprising: forming a first conductive structure in a first metallization layer, the first conductive structure forming a bottom wall of a housing; forming second conductive structures in a first via layer disposed over the first metallization layer, the second conductive structures forming lateral walls of the housing; forming a first conductive element and a second conductive element in a second metallization layer disposed over the first via layer, the first conductive element and the second conductive element disposed within the housing and configured to be mobile relative to each other; and forming a first conductive appendage and a second conductive appendage physically contacting the first conductive element and the second conductive element, respectively, the first conductive appendage and the second conductive appendage having sidewalls comprising an electrically conductive material resistant to formation of an electrically insulating compound at surfaces of the electrically conductive material in contact with fluid in the housing.

18. The method of claim 17, wherein the electrically conductive material is different from a material of the first conductive structure, the second conductive structures, the first conductive element, and the second conductive element.

19. The method of claim 17, wherein forming the first conductive appendage and the second conductive appendage comprises: depositing an insulating layer over the first conductive element and the second conductive element; etching a first opening to expose at least a portion of the first conductive element; etching a second opening to expose at least a portion of the second conductive element; and lining the first opening and the second opening with the electrically conductive material, the electrically conductive material physically contacting the first conductive element and the second conductive element.

20. The method of claim 17, wherein the electrically conductive material comprises an oxidation-resistant material.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Other advantages and characteristics of the invention will become apparent on studying the detailed description of entirely nonlimiting embodiments, and the appended drawings, in which:

[0036] FIGS. 1 to 5 relate to the prior art; and

[0037] FIGS. 6 to 22 relate to various embodiments of the invention.

DETAILED DESCRIPTION OF ILLUSTRATIVE EMBODIMENTS

[0038] Although the invention applies to any device having two pieces mobile relative to one another parallel to the substrate, the invention applies more particularly to integrated electrical devices for mechanical switching or to capacitive devices having a variable, for example thermally variable, capacitive value, for example those described in the aforementioned published patent applications.

[0039] One example of an integrated switching device is described in Patent Application Publication FR 2 984 009 (and counterpart U.S. Pat. No. 8,692,247).

[0040] The main characteristics thereof will be recalled here with reference to FIGS. 1 to 5.

[0041] Referring to FIG. 1, it can be seen that the mechanical switching device or switch CMT is produced within a single metallization level Mi of the interconnection part PITX of the integrated circuit CI, this interconnection part also being commonly referred to by a person skilled in the art by the acronym BEOL.

[0042] This interconnection part PITX may also be referred to as an interconnection region PITX and is located above the substrate SB.

[0043] In the examples described here, the switch CMT is made of metal, more particularly aluminum. This being the case, the metal could be copper, without these two examples implying limitation.

[0044] The switch CMT in this case comprises an assembly or structure ENS1 in the shape of an asymmetric cross. This assembly ENS1 has a first branch BR1A, and a second branch BR1B secured to a beam PTR, also referred to as a central pointer, at two positions EMPA and EMPB respectively located on two opposite faces of the beam PTR. These two locations EMPA and EMPB are separated by a distance d.

[0045] As will be seen in more detail below, the assembly ENS1 is produced by using conventional techniques for producing the metal tracks of the interconnection part PITX, which are used particularly in CMOS technology.

[0046] The left-hand part of FIG. 1 shows the switch CMT, and more particularly the assembly ENS1 encapsulated in an insulating region RIS, whereas the right-hand part of FIG. 1 shows the same assembly after etching of the insulating region so as to expose the branches BR1A and BR1B as well as the beam PTR.

[0047] The assembly ENS1 exposed in this way therefore extends inside a housing LG resulting from the removal of the insulating region RIS, the two branches BR1A and BR1E being secured to the edges BDA and BDB of the housing.

[0048] It has been shown in the article by R. Vayrette et al. entitled: Residual stress estimation in damascene copper interconnects using embedded sensors, Microelectronics Engineering 87 (2010) 412-415, that, after de-encapsulation of an assembly of this type, there is relaxation of stresses which causes a residual longitudinal deformation of the branches, leading to a deviation a of the pointer, here in the clockwise direction.

[0049] Depending on the applications which will be envisaged, and in particular the desired precision, for example in the case of temperature detection, this residual deviation a of the pointer PTR could be taken into account or neglected.

[0050] In the example illustrated in FIG. 2, the switch CMT comprises, in addition to the assembly ENS1, two electrically conductive stops BTA and BTB, for example two metal stops, also produced within the same metallization level as that in which the assembly ENS1 was produced.

[0051] The left-hand part of FIG. 2 represents the assembly ENS1 before de-encapsulation of the insulating region RIS, and the central part of FIG. 2 represents the same assembly ENS1 once de-encapsulated, that is to say exposed from the insulating region RIS. In the example shown here, the residual deviation of the beam PTR (pointer), associated with the relaxation of the stresses during the exposure of the assembly ENS1 from the insulating region RIS which surrounded it, is taken into account.

[0052] In the central part of FIG. 2, the first assembly ENS1 then has a first configuration when it is at a first temperature, for example ambient temperature. In this first configuration, the beam is at a distance from the two stops BTA and BTB and no electrical connection can pass through these stops BTA and BTB and the beam PTR.

[0053] In the right-hand part of FIG. 2, the ENS1 assembly has been thermally activated in order to adopt a second configuration, in which the beam PTR comes in contact with the two stops BTA and BTB.

[0054] This thermal activation is caused by a rise in the temperature of the expansion branches BR1A and BR1B.

[0055] Specifically, because of the rise in temperature, these branches expand and, since the two fastening points of the two branches BR1A and BR1B on the beam PTR are spaced apart in the longitudinal direction of the beam, this expansion causes a nonzero moment of force, resulting in a rotation RTX2 of the beam PTR in the anticlockwise direction and parallel to the substrate, this rotation RTX2 being opposite to the rotation RTX1 (central part of FIG. 2) resulting from the relaxation of stresses on the assembly ENS1.

[0056] Knowing the thermal expansion coefficient of the material forming the expansion branches, the geometry of the branches, in particular their length and their width as well as their thickness, and the spacing d between the two fastening points, the deviation of the pointer PTR resulting from the rotation RTX2 can be simulated easily.

[0057] The stops BTA and BTB can consequently be positioned in the housing in such a way that, above a threshold temperature, the beam PTR comes in contact with these two stops BTA and BTB.

[0058] It then becomes possible to establish an electrical connection passing through the stops BTA, BTB and through the beam PTR.

[0059] Reference will now be made more particularly to the FIGS. 3 to 5, in order to illustrate a way of fabricating an exemplary embodiment of the switch CMT.

[0060] It is assumed in these figures that the assembly ENS1, as well as the stops BTA and BTB, are produced at the metallization level M3 (also referred to as Metal level M3).

[0061] It is then seen (FIG. 3) that the level V2 of vias 2 between the metal level 2 and the metal level M3 and the level V3 of vias 3 between the metal level M3 and the metal level M4 are used in order to form a protective wall for the subsequent oxide etching and in order to make it possible to de-encapsulate the assembly ENS1 and the various stops.

[0062] Furthermore, as illustrated in FIG. 4 both the mobile part of the switch CMT, in this case the beam PTR, and the fixed part, in this case the stop BTA, are produced at the metal level M3. Furthermore, the stop BTA is connected to the metal level M2 by a via V20.

[0063] The switch CMT, and in particular the assembly ENS1, are produced by using conventional steps of fabricating metallization and via levels. More precisely, as illustrated in FIG. 5, after the production of the metal level M2 and of the via level V2, the assembly ENS1, represented here by dashes for the sake of simplicity, is produced in a conventional way by etching the underlying oxide and depositing metal, in this case copper, in the trenches. The assembly is then covered with oxide and the metallization level M4 is then produced.

[0064] After the formation of a conventional nitride layer C1 on the metal level M4, a comb is produced in this metal level M4 so as to form orifices OR.

[0065] Dry isotropic etching is then carried out, followed by wet etching, for example with hydrofluoric acid, so as to remove the insulating region (oxide) encapsulating the assembly ENS1, as well as the various stops, and thereby to produce the housing LG.

[0066] Nonconformal oxide deposition is then carried out in order to form a layer C2 closing the orifices OR.

[0067] Of course, that which has just been described for the metal levels M2, M3, M4 may be generalized to the metal levels M.sub.i1, M.sub.i, M.sub.i+1.

[0068] The conventional method for producing the various upper metallization levels then continues.

[0069] So as to avoid the possible appearance of an oxide on the surface of the two metal pieces mobile relative to one another, for example the beam PTR and the stops BTA and BTB of the prior art device described, provision is generally made to equip two metal parts, mobile relative to one another, with two projecting appendages each comprising at least one electrically conductive compound resistant to the surface formation of an electrically insulating compound in contact with the fluid present in the housing.

[0070] This is illustrated schematically in FIGS. 6 to 20.

[0071] In FIG. 6, the reference 1 denotes an element, fixed or mobile parallel to the substrate of the integrated circuit, of a structure STR housed at least partially in the housing LG of the BEOL part PITX, comprising in this example a branch 10 whose end 100 is intended to be mobile relative to another end of another element.

[0072] It can be seen in this example that the element 1 comprises an appendage 101 secured to the end 100 and projecting from this end loft

[0073] FIG. 6 is a schematic view of the layout type, whereas FIG. 7 schematically illustrates the geometrical configuration of the appendage 101 once produced.

[0074] FIG. 8 is a perspective view of the element 1.

[0075] It can be seen that this appendage 101 comprises an upper part low located at the via level immediately above the metallization level comprising the metal part in movement, and a lower part 1011 extending at least partially, and generally entirely, to the metallization level.

[0076] If it is assumed, for example, that this element 1 is the beam PTR (e.g. metal beam) of the device described with reference to FIGS. 1 to 5, which is produced at the metallization level M3, then the upper part 1010 of the appendage 101 is located at the via level V3 whereas the lower part 1011 of the appendage 101 is located at the metallization level M3.

[0077] The formation of this appendage 101, which is in fact an offset via, will be returned to in more detail below in order to explain the arrangement of these two upper and lower parts 1010 and 1011.

[0078] Specifically, as will be seen in more detail below, this appendage 101 will be produced simultaneously with the vias of the via level V3.

[0079] Thus, even though, as indicated above, there are numerous possibilities for selecting the compound(s) forming this appendage 101, it is particularly advantageous for the appendage 101 to have a composition identical to that of the vias of the integrated circuit.

[0080] Furthermore, by way of example, when the metallization levels are made of aluminum, the vias generally comprise tungsten coated with a barrier layer comprising titanium and titanium nitride.

[0081] This is illustrated more precisely in FIGS. 10 and 12, which are respectively sections along the lines X-X and XII-XII of FIG. 9 and 11.

[0082] FIG. 9 represents two elements 1 and 2 of the structure STR, each having, respectively, the ends 100 and 200 which are mobile relative to one another. Each of the two ends 100 and 200 could be mobile, or alternatively one of the two ends 100 and 200 may be fixed whereas the other is mobile. This is the case, for example, for the beam PTR and the stops BTA and BTB of the device of FIGS. 1 to 5.

[0083] Each end 100 and 200 is respectively equipped with an appendage 101, 201.

[0084] In the configuration illustrated in FIG. 9, the two appendages 101 and 201, facing one another, are spaced apart so as to prevent any electrical connection from passing through the elements 1 and 2.

[0085] It can be seen in FIG. 10 that the central part 10100, 20100 of each appendage 101 and 201 is formed by tungsten coated laterally with the barrier layer 10101, 20101. By way of indication, the barrier layer comprises a 50 angstrom layer of titanium and a 50 angstrom layer of titanium nitride TIN, the titanium nitride layer being in contact with the tungsten.

[0086] In FIG. 11, the two appendages 201 and 101 are in ohmic contact, allowing an electrical connection to pass through the metal elements 1 and 2.

[0087] It can be seen in FIG. 12 that it is in fact the two barrier layers 10101 and 20101 which are in contact, which improves the quality of the ohmic contact and avoids degradation thereof over the course of time, since the barrier layer formed by titanium and titanium nitride will not form on the surface an electrically insulating material in contact with the air present in the cavity and at the usual operating temperatures of the integrated circuit. Furthermore, the use of the appendages 101 and 201 in order to produce the electrical contact makes it possible to increase the contact area in comparison with a conventional configuration (without offset vias), and therefore to reduce the electrical contact resistance.

[0088] Reference will now be made more particularly to FIG. 13 in order to illustrate an embodiment of a method for producing the projecting appendages.

[0089] As explained above, particularly with reference to FIGS. 3 to 5, after having produced the bottom wall of the housing at the metallization level M2, then the via level V2 comprising the corresponding portions of lateral walls of the housing, a layer of aluminum (for example), intended to form the various metal tracks and elements of the device located at the metallization level M3, is deposited on the via level V2 and on the insulating region RIS enclosing the vias of the via level V2.

[0090] Once this deposition of aluminum has been carried out, this layer is etched so as to form the elements 30 and 40 of lateral walls of the housing, the branches 10 and 20 of the two elements 1 and 2 intended to be relatively mobile relative to one another, as well as, outside the housing LG, metal tracks 50 of the integrated circuit.

[0091] Oxide deposition is then carried out again in order to continue to form the insulating region RIS. After chemical-mechanical polishing has been carried out, etching of the vias of the via level V3 is carried out. This etching is conventionally plasma etching and makes it possible to produce, in the insulating region RIS, the orifices 33, 44 and 55 which will be intended to form, after filling with the filling material of the vias, other parts of lateral walls of the housing LG as well as, as regards the orifice 55, a via intended to connect the metal track 50 of the metal level M3 to a metal track located at a higher metallization level.

[0092] This etching also makes it possible to produce two orifices 12 and 22, which are orifices offset in relation to the branches 10 and 20.

[0093] The etching conventionally has a duration long enough to be able to correctly open the orifices 12 and 22 intended to receive the vias over the entire semiconductor wafer assembly, particularly at the ends of the latter.

[0094] For this reason, as regards the orifices 12 and 22, the etching operation will first open the upper parts 120 and 220 of these orifices 12 and 22 then, since these orifices are offset, make it possible to open the lower parts 121 and 221 of these orifices 12 and 22, which are located at least partially, and in general entirely, at the metallization level M3.

[0095] The vias are then filled by conformal deposition of the Ti/TiN barrier layer then by deposition of tungsten, conventionally by chemical vapour deposition (CVD deposition). It should be noted in this regard that this deposition is advantageously preceded by plasma etching (soft sputter etch), which makes it possible to remove the native oxide on the aluminum so that the titanium is in direct contact with the aluminum. This is possible because there is no re-entry of air between the etching chamber and the titanium deposition chamber. The method here is a vacuum method, which advantageously makes it possible to ensure good electrical contact between the barrier layer and the aluminum.

[0096] After chemical-mechanical polishing has been carried out, all the vias of the via level V3 are produced, as well as the appendages 101 and 201 which fill the orifices 12 and 22.

[0097] In this embodiment, the formation of the appendages 101 and 201 will require an additional via level and metallization level in order to close the housing, which represents an albeit reduced extra cost in terms of volume.

[0098] As a variant, it would be possible to avoid forming these additional via and metallization levels by forming, as illustrated schematically in FIG. 14, appendages 201 extending partially or entirely to the metal level M.sub.i (for example the level M3) and extending only partly to the via level V, immediately above (for example the level V3), and by producing the other vias in two steps.

[0099] More precisely, an insulating layer CSS1 thinner than the future insulating region RIS intended to separate the adjacent metallization levels M3 and M4 is first deposited on the metal level M3.

[0100] The appendage 201 secured to the branch 20 is then produced in a manner similar to that described above.

[0101] Since the thickness of the layer CSS1 is less than the distance between metallization levels, the upper part 2010 of the appendage 201 extends only partly to the via level V3, whereas the lower part 2011 of this appendage 201 in this case extends fully to the metallization level M3.

[0102] A complementary insulating layer CSS2 is then redeposited so as to form the insulating region RIS. It is in this region RIS that the other vias V of the via level V3 will then be produced.

[0103] The geometry of the appendages 101 and 201 may be arbitrary, and is not limited to the shape illustrated in FIGS. 6 to 12.

[0104] Thus, as illustrated in FIGS. 15 and 16, the appendage 101 may be located in the vicinity of the end of the element 1 (which may be mobile relative to element 2) and fully overlap this mobile element 1 transversely.

[0105] Likewise, the element 2, which may for example be fixed relative to element 1, is then provided on one of its sides with an appendage 201.

[0106] In FIG. 15, the two appendages 101 and 201 are at a distance from one another, corresponding to an open switch, whereas in FIG. 16 the two appendages 101 and 201 are in ohmic contact, corresponding to a closed switch.

[0107] In FIGS. 17 and 18, which also represent a switch in its open and closed states, the appendage 101 may fully cover the free end of the element 1, which is assumed to be the mobile element relative to element 2, so as to have a part next to the appendage 201 of the element 2.

[0108] The change from a configuration of the open switch type to a configuration of the closed switch type may be due to a thermal activation obtained, for example, by deformation of at least one of the elements, due for example to a rise in temperature of the integrated circuit. This is described for example, in the aforementioned French patent application, or in Patent Application WO 2013/083385 (US2014/266562).

[0109] As a variant, this rise in temperature may be obtained by the Joule effect by flow of a current in at least one of the metal elements.

[0110] As a variant, as illustrated in FIGS. 19 and 20, it is possible for this activation to be obtained by the application of an electrostatic force F generated by activation means known per se. These activation means may, for example, comprise a generator GEN capable of applying a potential difference between the two appendages 102 and 301.

[0111] More precisely, in this case, as illustrated in FIG. 19, the structure STR comprises a first element 1, which is assumed to be a mobile element, equipped at its end with a projecting appendage 101 located next to the projecting appendage 201 of a second element 2, which is assumed to be fixed. The switch is then in an open state.

[0112] Furthermore, a third element 3 is also equipped with an additional projecting appendage 301, which is located next to an additional projecting appendage 102 of the first element 1.

[0113] In view of the projecting nature of the additional appendages, the distance e2 between these appendages is less than the distance e1 which would separate the elements 1 and 3 in the absence of additional appendages 102 and 301.

[0114] As illustrated schematically in FIG. 20, by application of an electrostatic force F between the additional appendages 102 and 301, the appendage 101 comes in contact with the appendage 201 in order to place the switch in its closed state.

[0115] Furthermore, in view of the fact that the distance e2 is less than the distance e1, and the area of the activation capacity is greater because of the presence of the additional appendages, the electrostatic force F necessary in order to place the appendages 101 and 201 in contact is less than the electrostatic force which it would have been necessary to apply between the elements 3 and 1 in the absence of additional appendages 301 and 102. The decrease in the activation voltage, associated with the gain in area, may for example reach a factor of 2.

[0116] The invention does not only apply to integrated mechanical switches, but may also find an application for integrated capacitive devices having a variable, for example thermally variable, capacitive value, such as for example that described in published French Patent Application FR 2 984 010.

[0117] Specifically, as illustrated in FIG. 21, the metal structure STR, comprising the two metal elements 1 and 2 equipped with their respective appendages 101 and 201, may form a capacitor having a first capacitive value C1 when the two appendages 101 and 201 are separated by the distance d1.

[0118] In a second configuration, illustrated in FIG. 22, the element 1 approaches the element 2 so that the distance d2 between the two appendages 101 and 201 is less than the distance d1, which gives the capacitor a second capacitive value C2 different from the capacitive value C1.

[0119] Furthermore, since the appendages 101 and 201 are formed by a compound resistant or refractory to the surface formation of an electrically insulating material, there is no risk of formation on the appendages of such an insulating material, which may form a parasitic dielectric and render incorrect the capacitive values C1 and C2 which have been calculated on the basis of the distances d1 and d2.