Method for producing silicon components

Abstract

The present invention relates to a method for manufacturing a plurality of watch components (1) in silicon, in which a substrate (100) is provided comprising a useful layer (10) of silicon, an intermediate layer (30) of oxide, and a stiffening layer (20), the useful layer (10) is etched to form the watch components (1), the substrate (100) is thermally oxidized with the lower face (2b) of each component (1) remaining in contact with the intermediate layer (30) during the oxidation, then in a single deoxidation step, the oxide previously formed on each component (1) is removed and, for each component, at least the part of the intermediate layer (30) in contact with its lower face (2b).

Description

Technical field of the invention

[0001] The present invention relates to the manufacture of silicon components, in particular watch components such as balance springs, anchors, escape wheels, gear wheels, plates, hands, balance wheels. State of the art

[0002] Silicon watch components are typically manufactured using a substrate - also called a wafer - of the silicon-on-insulator type known by the acronym SOI (silicon on insulator) and comprising a useful layer of silicon in which the components are to be formed, a second layer, typically of silicon, used to stiffen the substrate and, between these two layers, an intermediate oxide layer serving as a stop layer for the etching operations of the useful layer.

[0003] According to a known process, the useful silicon layer is etched to form the components, then a wafer containing at least part of this useful layer bearing the components is "released," that is, detached from the substrate to be subjected to oxidation and deoxidation operations aimed at improving the surface condition of the components or adjusting their dimensions, in particular to adjust their stiffness, before they are finally detached from the wafer.

[0004] In the case of individual components, the components are placed on a support for processing, resulting in laborious movement.

[0005] However, manufacturers are constantly seeking to optimize the manufacturing process of such components, particularly by reducing its duration and therefore the associated costs. Summary of the invention

[0006] One aim of the present invention is therefore to propose a method for manufacturing watch components whose duration and cost are optimized.

[0007] According to the invention, this goal is achieved through the object of claim 1. More specific aspects of the present invention are described in the dependent claims as well as in the description.

[0008] More specifically, one objective of the invention is achieved through a method for manufacturing a plurality of silicon watch components, wherein, in this order: a) a substrate is provided comprising the superposition, in a transverse direction, of at least one useful silicon layer, an intermediate oxide layer and a stiffening layer, the intermediate layer being interposed between the useful layer and the stiffening layer, b) the useful layer is etched to form the watch components, c) the substrate is thermally oxidized, the lower face of each component remaining in contact with the intermediate layer during the oxidation, then d) in a continuous deoxidation step, oxide formed in step c) is removed from each component and, for each component, at least the part of the intermediate layer in contact with its lower face.

[0009] Steps a) to d) are carried out in this order, either directly one after the other, or by inserting one or more intermediate steps between at least two successive steps.

[0010] In this application, and with reference to the substrate's usable layer or to components made within that usable layer, a lower face shall be defined as a face oriented towards the intermediate layer and the stiffening layer of the substrate, and a top face shall be defined as a face opposite such a lower face in the transverse direction. A lateral face connects such lower and upper faces.

[0011] According to the invention, the substrate, already etched, is oxidized while the intermediate layer remains in place and intact over the components. During step c), the underside of the watch components, still attached to the intermediate layer, is therefore not oxidized.

[0012] On each component, oxidation results at least in the formation of a layer of silicon dioxide on its lateral face(s) (corresponding to the etching edges).

[0013] The top face of a component may be oxidized during step c) but this is not necessarily the case, this face may for example have been coated beforehand (in particular by a nitride mask deposited before etching).

[0014] The fact that the components are not oxidized on their underside (or even on their topside) during step c) is not problematic. Indeed, the oxidation in step c) is advantageously a smoothing oxidation of the watch components or an oxidation for sizing them. In the case of smoothing, the primary aim is to correct surface irregularities related to the etching process—and therefore present on the sides or lateral faces of the components.

[0015] Similarly, the height (i.e., the transverse dimension) of the components generally requires no adjustment since this dimension is determined by the thickness of the substrate's effective layer. However, oxidation of the components' lateral faces can allow for adjustment of their width.

[0016] According to the invention, carrying out at least one oxidation on the substrate after etching the useful layer (hereinafter etched substrate) and before releasing the components (in other words, before removing the intermediate layer over them) has several advantages: In the case of individual components, not connected to a part of the useful layer by a fastener, smoothing and / or sizing can be carried out without moving the components one by one on a support.

[0017] In the case of components attached to part of the active layer by a fastener, the components can be preserved during the thermal oxidation step and are not at risk of deformation and / or detachment. Indeed, unlike a wafer which, when subjected to high temperatures in a thermal oxidation furnace, can tend to flex under its own weight, the substrate here is rigidified by its second layer and does not deform.

[0018] Furthermore, the number of steps in the process is optimized and the handling time is reduced, as well as the quantity of treatment products and energy required.

[0019] Whereas conventional oxidation processes release a wafer carrying the components, typically by deoxidizing the substrate, and then perform an additional oxidation and deoxidation sequence on this wafer, the present process requires only one oxidation step and a single substrate deoxidation step. A deoxidation step can therefore be eliminated for the same result.

[0020] As previously mentioned, the oxidation in step c) can, for example, have a smoothing function or a component sizing function.

[0021] In this application, smoothing oxidation should be understood as oxidation solely aimed at improving the surface finish of components, and in which the size of these components is not taken into account. In particular, the choice of the oxidation time or the amount of silicon to be removed from the components is not determined by their dimensions.

[0022] In the case of dimensional adjustment, on the contrary, oxidation can be carried out according to a quantity of silicon to be removed from the components formed in step b) and determined beforehand.

[0023] By quantity of silicon, we mean in particular a thickness (measured orthogonally to the transverse direction) and / or a height (measured in the transverse direction) of silicon of the components formed in step b).

[0024] Typically, in step b), the components can be formed with dimensions larger than the desired dimensions. For example, in the case of a spiral, the dimensions of the spirals formed in the useful layer during step b) of etching may be larger than the dimensions required to obtain spirals with a certain target stiffness.

[0025] However, the thermal oxidation of silicon leads to the formation of silicon dioxide at the expense of silicon. The silicon front retreats to create a new interface with the silicon dioxide. When, during step d) of deoxidation, the silicon dioxide layer is removed, the interface between the silicon and the silicon dioxide becomes the new outer surface of the component.

[0026] Particularly when used for component sizing, the process may include a preliminary step of measuring the etched components and determining the amount of material to be removed to obtain predetermined component dimensions. In this case, the subsequent oxidation is performed according to the amount of material to be removed thus determined.

[0027] In the case of sizing, the smoothing effect is also obtained on the oxidized faces, as an additional intrinsic effect of oxidation / deoxidation.

[0028] The deoxidation in step d) is carried out in a single continuous step, enabling both the removal of the oxide formed in step c) from each component—preferably all the oxide formed in step c) from each component, and the separation of each component from the intermediate layer portion bonded to its lower face. This step thus performs two functions, usually separated in known prior art processes. By continuous, we mean that the deoxidation in step d) is carried out according to a given method (i.e., one that does not vary) and without interruption.

[0029] The process advantageously includes a step d0), prior to step d), in which the duration of step d) is determined. This determination is made taking into account the amount of oxide formed in step c) and an amount of oxide to be removed from the intermediate layer. It must be at least sufficient so that, at the end of step d), each component is free of all oxide on its upper, lower, and lateral faces.

[0030] Step d) can in particular be carried out by wet etching, for example by vapor phase etching, in particular by hydrofluoric acid vapor etching.

[0031] For example, in a direction orthogonal to the transverse direction, each component etched in step b) can be externally delimited by a lateral face forming a continuous and closed contour. In other words, each component can be etched individually in the usable layer, without being attached to the rest of that layer. In this case, the deoxidation in step d) releases each component individually. That is to say, each component is isolated at the end of step d).

[0032] According to another example, the components etched in step b) can each be connected by a fastener to the rest of the useful layer.

[0033] In this case, the process may include the release of a wafer comprising all or part of the useful layer and carrying said components. In particular, step d) may result in the release of such a wafer.

[0034] By wafer, we mean a part of the substrate in which the components are no longer covered, in the transverse direction, by either the intermediate layer or the stiffening layer of the substrate.

[0035] This wafer is then advantageously made up of all or part of the useful layer of the substrate carrying the watch components.

[0036] According to one example, the process may further include, after step d), at least one step e) of oxidation and then deoxidation of the components, in particular of the wafer bearing said components.

[0037] According to one example, the process may include, after step d) or after step e), at least one step f) of final oxidation of the components, in particular of the wafer bearing said components.

[0038] According to one example, the process may further include, after any one of steps d), e), or f), a step g) of detaching each component from its wafer. Brief description of the drawings

[0039] The features and advantages of the present invention will become apparent in more detail in the following description, with an illustrative and non-limiting example given by reference to the attached figures, in which: THE figures 1a to 1g schematically represent the steps of the process according to one implementation method of the invention. figure 2 indicates a sequence of steps, optional or not, of the process according to the invention. Detailed description

[0040] The process described below is suitable for manufacturing all types of silicon watch components, including escape wheels, balance springs, anchors, gear wheels, plates, hands, balance wheels.

[0041] The first step a) of the process consists of providing a suitable substrate 100 (also commonly called a wafer). As illustrated in the figure 1a , the substrate 100 is advantageously of the silicon-on-insulator (SOI) type, and comprises a useful silicon layer 10 in which the component(s) are to be formed, a stiffening layer 20 here of silicon and an intermediate oxide layer 30, typically of silicon dioxide (SiO2), interposed between the useful layer 10 and the stiffening layer 20 in a transverse direction Z of the substrate 100.

[0042] The thickness e1 of the useful silicon 10 layer depends on the desired thickness for the components to be manufactured. For example, it can be between 50 and 500µm.

[0043] The stiffening layer 20 is generally thicker than the useful layer (e2 > e1), so as to ensure the stiffening of the substrate 100.

[0044] The intermediate layer 30 forms a bonding layer between the useful layer 10 and the stiffening layer 20.

[0045] In a step b) of the process, the useful layer 10 of the substrate 100 is etched, for example by deep reactive ion etching (DRIE), to form the watch components.

[0046] The engraving can be done through a photosensitive resin mask 40. In this case, and as illustrated on the figure 1b , the useful layer 10 is covered with a layer of photosensitive resin 42, then, by photolithography, a plurality of apertures 44 are formed in the resin layer 42 (see figure 1c) corresponding to the contours of the plurality of components to be manufactured, with the exception, where applicable, of an unengraved portion or attachment retained to ensure the bond between the component and the rest of the usable layer. For this purpose, the resin layer 42 is typically exposed using a light source, through a photomask 46 representing the desired pattern, as illustrated in the figure 1b .

[0047] By DRIE etching, the active layer 10 of the substrate is then etched across its entire thickness, under the openings 44 of the mask 40, to form the components 1 (see figure 1d ).

[0048] For each component, we define a top face 2a, a bottom face 2b opposite the top face 2a in the transverse direction Z and a side face 2c, with the bottom face 2b of each component in contact with the intermediate layer 30.

[0049] If necessary, the lateral face can define a discontinuous contour, interrupted by a fastener, in a plane orthogonal to the transverse direction Z.

[0050] As illustrated on the figure 1e , the resin layer 42 can then be removed, in particular by oxygen plasma cleaning or by chemical cleaning, typically in a bath or under the effect of a spray containing a product to remove the resin.

[0051] Alternatively, in an unshown embodiment, the process could include, prior to step b) of etching, a step a') of producing a layer of silicon oxide or nitride, typically silicon nitride, on the unetched substrate. Advantageously, this layer is a silicon oxide layer obtained by thermal oxidation of the substrate. In this case, in step b), this oxide or nitride layer is etched through the resin openings to form an additional mask, through which the desired layer will be etched.

[0052] This variant may be of particular interest when the thickness to be engraved (e1) is significant, as the photosensitive resin, "consumed" during DRIE engraving, may not be sufficient in this case.

[0053] In step c) of the process, the etched substrate is then thermally oxidized, usually in a thermal oxidation furnace heated to a temperature between 900 and 1200°

[0054] As the oxidation reaction progresses, the free silicon S at the top and side faces of components 1 is consumed and transformed into silicon dioxide D, gradually pushing back the interface between the silicon and the new layer of silicon dioxide.

[0055] Free silicon (S) refers to silicon that is not covered with silicon oxide. For example, if a nitride mask has been retained on the top of the substrate, the silicon is only free on the lateral faces of the components.

[0056] Retreating the silicon interface eliminates initial surface defects, including irregularities ("scallops") and surface micro-openings related to etching, also resulting in improved mechanical resistance of components 1.

[0057] Conversely, the lower face of each component 1, covered by the intermediate layer 30, does not undergo oxidation. The silicon interface in contact with this intermediate layer 30 therefore remains unchanged and, in particular, is not shifted in the transverse Z direction.

[0058] In a particular variant, when the oxidation function is to adjust the dimension of components 1, step c) can be controlled according to a predetermined quantity of silicon to be removed.

[0059] For example, the process may include a preliminary step of measuring the etched components and determining the amount of silicon to be removed to obtain the desired component dimensions. In this case, the oxidation time is adjusted accordingly.

[0060] According to the invention, step c) of oxidation is followed by a deoxidation step d), illustrated by the figure 1gThis process, by itself, allows the components to be freed from all the oxide present around their periphery. Thus, deoxidation removes both the oxide formed after the etching in step b) on the lateral faces 2c and upper faces 2a of components 1, and that forming part of the intermediate layer 30 in contact with their lower faces 2b. If necessary, it can also remove the oxide or nitride mask formed before the etching in step b) (step a'). Advantageously, at the end of step d), each silicon component formed in step b) has its lower face, upper face, and lateral face(s) free of any coating.

[0061] The deoxidation, carried out in a single step d) continuous, completes the smoothing or sizing of the components 1 by removing the oxide D formed during the previous oxidation, and frees the components by removing more particularly the oxide on the lower face 2b of the components 1.

[0062] Deoxidation can advantageously be achieved by wet etching, in particular by vapor phase etching, especially by hydrofluoric acid vapor etching.

[0063] If the components have been etched without a tether, the components 1 are then individualized and freed from the rest of the substrate 100 in step d).

[0064] If the components 1 are attached to the functional layer 10 of the substrate 100, step d) releases a wafer carrying the watch components 1. This wafer typically consists of the functional layer 10 of the substrate, which carries the watch components, or a portion of this functional layer 10 delimited during step b) of the etching process by a peripheral groove. Releasing such a wafer allows for additional processing steps (e.g., oxidation) of the components on the wafer, or at the very least facilitates their handling during such additional processing steps.

[0065] Once the components or wafer have been released, the process may still include one or more additional sequence(s) e) of oxidation / deoxidation, aimed at further smoothing or dimensioning the components 1 and / or at least one final oxidation step f) of the components or wafer which may be used for thermocompensation and / or mechanical strengthening of the components 1.

[0066] Finally, in the case of components etched with a fastener and attached to a wafer, the process can end with a step g) of detaching the components 1.

Claims

1. A method for manufacturing a plurality of watch components (1) in silicon, wherein, in this order: a) a substrate (100) is provided comprising the superposition, in a transverse direction, of at least one useful layer (10) of silicon, an intermediate oxide layer (30) and a stiffening layer (20), the intermediate layer (30) being interposed between the useful layer (10) and the stiffening layer (20), b) the useful layer (10) is etched to form the watch components (1), c) the substrate (100) is thermally oxidized, the lower face (2b) of each component (1) remaining in contact with the intermediate layer (30) during the oxidation, then d) in a continuous deoxidation step, oxide formed in step c) is removed from each component and, for each component (1), at least the part of the intermediate layer (30) in contact with its lower face (2b).

2. Manufacturing process according to claim 1, wherein the oxidation of step c) is a smoothing oxidation of the watch components (1).

3. Manufacturing process according to claim 1, wherein the oxidation of step c) is a sizing oxidation of watch components (1).

4. Manufacturing process according to claim 3, wherein the oxidation of step c) is carried out as a function of a quantity of silicon to be removed from the components formed in step b) and determined beforehand.

5. A manufacturing method according to any one of claims 1 to 4, wherein step d) is carried out by wet etching, for example by vapor etching, in particular by hydrofluoric acid vapor etching.

6. A manufacturing method according to any one of claims 1 to 5, including the release of a wafer comprising all or part of the useful layer (10) of the substrate (100) and carrying said components (1).

7. A method according to any one of claims 1 to 5, wherein, in a direction orthogonal to the transverse direction, each component etched in step b) is externally delimited by a lateral face forming a continuous and closed contour, and the deoxidation of step d) releases each component individually.

Images