CHEMICAL BONDING METHOD, PACKAGE-TYPE ELECTRONIC COMPONENT, AND HYBRID BONDING METHOD FOR ELECTRONIC DEVICE

Abstract

Substrates that are bonding targets are bonded in ambient atmosphere via bonding films, including oxides, formed on bonding faces of the substrates. The bonding films, which are metal or semiconductor thin films formed by vacuum film deposition and at least the surfaces of which are oxidized, are formed into the respective smooth faces of two substrates having the smooth faces that serve as the bonding faces. The bonding films are exposed to a space that contains moisture, and the two substrates are overlapped in the ambient atmosphere such that the surfaces of the bonding films are made to be hydrophilic and the surfaces of the bonding films contact one another. Through this, a chemical bond is generated at the bonded interface, and thereby the two substrates are bonded together in the ambient atmosphere. The bonding strength γ can be improved by heating the bonded substrates at a temperature.

Claims

1. A chemical bonding method comprising: a step of forming a bonding film, which is a thin film of a metal or semiconductor formed by vacuum deposition and oxidized at least on its surface, on each smooth surface of two base materials having smooth surfaces; a step of exposing the surfaces of the bonding films formed on the two base materials to a space having moisture to hydrophilize the surfaces of the bonding films; and a step of bonding the two base materials by overlaying them with each other so that the surfaces of the bonding films in a hydrophilic state contact each other.

2. A chemical bonding method comprising: a step of forming a bonding film, which is a thin film of a metal or semiconductor formed by vacuum deposition and oxidized at least on its surface, on one of the smooth surfaces of two base materials having smooth surfaces; a step of exposing the surface of the bonding film to a space having moisture to hydrophilize the surface of the bonding film; and a step of bonding the two base materials by overlaying them with each other so that the surface of the bonding film in a hydrophilic state contacts the oxide thin film of a metal or semiconductor formed on the smooth surface of the other base material.

3. A chemical bonding method comprising: a step of forming a bonding film, which is a thin film of a metal or semiconductor formed by vacuum deposition and oxidized at least on its surface, on one of the smooth surfaces of two base materials having smooth surfaces; a step of exposing the surface of the bonding film to a space having moisture to hydrophilize the surface of the bonding film; and a step of bonding the two base materials by overlaying them so that the surface of the bonding film in a hydrophilic state contacts the smooth surface of the other base material, which comprises a hydrophilized or activated metal or semiconductor, or an oxide thereof.

4. The chemical bonding method according to claim 1, which comprises a step of further heating the base materials after the bonding.

5. The chemical bonding method according to claim 4, wherein the heating is performed at a temperature of 400° C. or lower.

6. The chemical bonding method according to claim 1, wherein the overlaying of the two base materials is performed in air.

7. The chemical bonding method according to claim 1, wherein the bonding films are the oxide thin films formed by vacuum deposition.

8. The chemical bonding method according to claim 1, wherein the bonding films are formed by oxidizing at least the surface of the thin films of a metal or semiconductor formed by vacuum deposition.

9. The chemical bonding method according to claim 1, wherein the bonding films are formed as films with many defects.

10. The chemical bonding method according to claim 1, wherein the hydrophilization of the surfaces of the bonding films is performed by taking out the base materials having the bonding films from a vacuum vessel to air.

11. The chemical bonding method according to claim 1, wherein the hydrophilization of the surfaces of the bonding films is performed by introducing moisture into the vacuum vessel in which the vacuum deposition has been performed.

12. The chemical bonding method according to claim 1, wherein the bonding films are formed to have a surface roughness with an arithmetic mean height Sa of 0.5 nm or less.

13. A packaged electronic component with an electronic component sealed inside a hollow package formed by bonding a package body to a lid, a junction between the package body and the lid comprising an intermediate layer comprising a first thin film of a metal or semiconductor oxidized at least on its surface formed on a smooth surface of the package body and a second thin film of a metal or semiconductor oxidized at least on its surface formed on a smooth surface of the lid, an interface between the first and second thin films of the intermediate layer being bonded by chemical bonding, and at least a part of the intermediate layer being formed by an oxide thin film with many defects, and a hollow space of the package being sealed with a gas at atmospheric pressure.

14. A packaged electronic component with an electronic component sealed inside a hollow package formed by bonding a package body to a lid, a junction between the package body and the lid comprising an intermediate layer comprising a thin film of a metal or semiconductor oxidized at least on its surface formed on a smooth surface of either the package body or the lid, an interface between the intermediate layer and a smooth surface of the other of the package body or the lid, which comprises a metal, semiconductor, or oxide thereof, being bonded by chemical bonding, and at least a part of the intermediate layer being formed by an oxide thin film with many defects, and a hollow space in the package being sealed with a gas at atmospheric pressure.

15. The packaged electronic component according to claim 13, wherein the gas at atmospheric pressure is an inert gas.

16. A hybrid bonding method for electronic devices, comprising forming a bonding surface that is at least partially smooth and has an electrode portion and an insulating portion on each of the two electronic devices to be bonded, and bonding the two bonding surfaces by aligning the electrode portions and the insulating portions with each other, the bonding between the two bonding surface comprising: a step of forming a bonding film on each of the two bonding surfaces, which is a thin film of a metal or semiconductor formed by vacuum deposition and oxidized at least on its surface, a step of exposing the surfaces of the bonding films formed on the two bonding surfaces to a space having moisture to hydrophilize the surfaces of the bonding films, and a step of overlaying the two bonding surfaces with each other and bonding them together so that the surfaces of the bonding films in a hydrophilic state contacting each other with the electrode portions and the insulating portions of the two bonding surfaces aligned with each other, the bonding through the bonding films providing continuity between the electrode portions of the two bonding surfaces and making the insulating portions electrically insulated.

17. A hybrid bonding method for electronic devices, comprising forming a bonding surface that is at least partially smooth and has an electrode portion and an insulating portion on each of the two electronic devices to be bonded, and bonding the two bonding surfaces by aligning the electrode portions and the insulating portions with each other, the bonding between the two bonding surface comprising: a step of forming a bonding film on one of the bonding surfaces, which is a thin film of a metal or semiconductor formed by vacuum deposition and oxidized at least on its surface, a step of exposing the surface of the bonding film to a space having moisture to hydrophilize the surface of the bonding film, a step of bonding the two bonding surfaces by overlaying them so that the surface of the bonding film in a hydrophilic state contacts the oxide thin film of a metal or semiconductor formed on the other bonding surface with the electrode portions and the insulating portions of the two bonding surfaces aligned with each other, the bonding through the bonding film providing continuity between the electrode portions of the two bonding surfaces and making the insulating portions electrically insulated.

18. A hybrid bonding method for electronic devices, comprising forming a bonding surface that is at least partially smooth and has an electrode portion and an insulating portion on each of the two electronic devices to be bonded, and bonding the two bonding surfaces by aligning the electrode portions and the insulating portions with each other, the bonding between the two bonding surface comprising: a step of forming a bonding film, which is a thin film of a metal or semiconductor formed by vacuum deposition and oxidized at least on its surface, on one of the bonding surfaces after forming the electrode portions and insulating portions, a step of exposing the surface of the bonding film to a space having moisture to hydrophilize the surface of the bonding film, a step of bonding the two bonding surfaces by overlaying them with each other so that the surface of the bonding film in a hydrophilic state contacts the other hydrophilized or activated bonding surface with the electrode portions and the insulating portions of the two bonding surfaces aligned with each other, the bonding through the bonding film providing continuity between the electrode portions of the two bonding surfaces and making the insulating portions electrically insulated.

19. The hybrid bonding method for electronic devices according to claim 16, wherein the bonding films are formed by oxidizing at least the surfaces of the thin films of a metal or semiconductor formed by vacuum deposition.

20. The hybrid bonding method for electronic devices according to claim 16, wherein the bonding by overlaying the bonding surfaces is followed by a step of heating at a temperature of 300° C. or lower.

21. The hybrid bonding method for electronic devices according to claim 16, wherein a thickness of the bonding films is from 0.3 nm to 5 nm.

22. The chemical bonding method according to claim 2, which comprises a step of further heating the base materials after the bonding.

23. The chemical bonding method according to claim 22, wherein the heating is performed at a temperature of 400° C. or lower.

24. The chemical bonding method according to claim 3, which comprises a step of further heating the base materials after the bonding.

25. The chemical bonding method according to claim 24, wherein the heating is performed at a temperature of 400° C. or lower.

26. The chemical bonding method according to claim 2, wherein the overlaying of the two base materials is performed in air.

27. The chemical bonding method according to claim 3, wherein the overlaying of the two base materials is performed in air.

28. The chemical bonding method according to claim 2, wherein the bonding films are the oxide thin films formed by vacuum deposition.

29. The chemical bonding method according to claim 3, wherein the bonding films are the oxide thin films formed by vacuum deposition.

30. The hybrid bonding method for electronic devices according to claim 2, wherein the bonding films are formed by oxidizing at least the surfaces of the thin films of a metal or semiconductor formed by vacuum deposition.

31. The hybrid bonding method for electronic devices according to claim 3, wherein the bonding films are formed by oxidizing at least the surfaces of the thin films of a metal or semiconductor formed by vacuum deposition.

32. The chemical bonding method according to claim 2, wherein the bonding films are formed as films with many defects.

33. The chemical bonding method according to claim 3, wherein the bonding films are formed as films with many defects.

34. The chemical bonding method according to claim 2, wherein the hydrophilization of the surfaces of the bonding films is performed by taking out the base materials having the bonding films from a vacuum vessel to air.

35. The chemical bonding method according to claim 3, wherein the hydrophilization of the surfaces of the bonding films is performed by taking out the base materials having the bonding films from a vacuum vessel to air.

36. The chemical bonding method according to claim 2, wherein the hydrophilization of the surfaces of the bonding films is performed by introducing moisture into the vacuum vessel in which the vacuum deposition has been performed.

37. The chemical bonding method according to claim 3, wherein the hydrophilization of the surfaces of the bonding films is performed by introducing moisture into the vacuum vessel in which the vacuum deposition has been performed.

38. The chemical bonding method according to claim 2, wherein the bonding films are formed to have a surface roughness with an arithmetic mean height Sa of 0.5 nm or less.

39. The chemical bonding method according to claim 3, wherein the bonding films are formed to have a surface roughness with an arithmetic mean height Sa of 0.5 nm or less.

40. The packaged electronic component according to claim 14, wherein the gas at atmospheric pressure is an inert gas.

41. The hybrid bonding method for electronic devices according to claim 17, wherein the bonding films are formed by oxidizing at least the surfaces of the thin films of a metal or semiconductor formed by vacuum deposition.

42. The hybrid bonding method for electronic devices according to claim 18, wherein the bonding films are formed by oxidizing at least the surfaces of the thin films of a metal or semiconductor formed by vacuum deposition.

43. The hybrid bonding method for electronic devices according to claim 17, wherein the bonding by overlaying the bonding surfaces is followed by a step of heating at a temperature of 300° C. or lower.

44. The hybrid bonding method for electronic devices according to claim 18, wherein the bonding by overlaying the bonding surfaces is followed by a step of heating at a temperature of 300° C. or lower.

45. The hybrid bonding method for electronic devices according to claim 17, wherein a thickness of the bonding films is from 0.3 nm to 5 nm.

46. The hybrid bonding method for electronic devices according to claim 18, wherein a thickness of the bonding films is from 0.3 nm to 5 nm.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0134] FIG. 1 is a cross-sectional electron micrograph (TEM) of Si substrates bonded with Y.sub.2O.sub.3 thin films as bonding films.

[0135] FIG. 2 illustrates the correlation between heat treatment temperature and bonding strength γ for each film thickness (2 nm and 5 nm) in the bonding test results of quartz substrates with TiO.sub.2 thin films as bonding films.

[0136] FIG. 3 illustrates the correlation between film thickness and bonding strength γ at different heating temperatures (unheated, 100° C., 200° C., and 300° C.) in the bonding test results of quartz substrates with ITO thin films as bonding films.

[0137] FIG. 4 illustrates the correlation between film thickness and bonding strength γ for quartz substrates bonded with ITO thin films as bonding films in air and then heated at 300° C. (Example) and for quartz substrates bonded in a vacuum vessel and then heated at 300° C. (Comparative Example), respectively.

[0138] FIG. 5 illustrates the correlation between heat treatment temperature and bonding strength γ for each film thickness (1 nm, 2 nm, and 5 nm) in the bonding test results of quartz substrates with SiO.sub.2 thin films as bonding films.

[0139] FIG. 6 illustrates the correlation between film thickness and bonding strength γ at different heating temperatures (unheated, 100° C., 200° C., and 300° C.) in the bonding test results of quartz substrates bonded with WO.sub.3 thin films as bonding films.

[0140] FIG. 7 illustrates the correlation between the waiting time in air and bonding strength γ in the bonding test results of quartz substrates bonded with ITO thin films as bonding films.

[0141] FIG. 8 illustrates the “blade method” used to measure the bonding strength (free energy at the bonding interface) γ (J/m.sup.2).

[0142] FIG. 9 illustrates hybrid bonding using the chemical bonding method of the present invention.

[0143] FIG. 10A illustrates the preparation process of the wafer A used in Experimental Example 4 (with an electrode film formed).

[0144] FIG. 10B illustrates the preparation process of the wafer A used in Experimental Example 4 (with an Au protective film formed on the electrode film).

[0145] FIG. 10C illustrates the preparation process of the wafer A used in Experimental Example 4 (with an Au protective film and an insulating protective film formed on the electrode film).

[0146] FIG. 11A illustrates the preparation process of the wafer B used in Experimental Example 4 (with an electrode film formed).

[0147] FIG. 11B illustrates the preparation process of the wafer B used in Experimental Example 4 (with an Au protective film formed on the electrode film).

[0148] FIG. 11C illustrates the preparation process of the wafer B used in Experimental Example 4 (with an Au protective film and an insulating protective film formed on the electrode film).

[0149] FIG. 12 illustrates the overlaying of the wafer A with the wafer B used in Experimental Example 4.

[0150] FIG. 13 illustrates the measurement circuit used in Experimental Example 4.

[0151] FIG. 14 illustrates the correlation between temperature change and electrical resistance change when Au electrode portions were bonded by a Ti bonding film (0.5 nm) (Example).

[0152] FIG. 15 illustrates the correlation between temperature change and electrical resistance change when Au electrode portions were bonded directly without a bonding film (Comparative Example).

[0153] FIG. 16 illustrates the correlation between temperature change and electrical resistance change when Cu electrode portions were bonded by a Ti bonding film (0.3 nm) (Example).

[0154] FIG. 17 illustrates the correlation between temperature change and electrical resistance change when Cu electrode portions were bonded by a Mn bonding film (0.3 nm) (Example).

[0155] FIG. 18 illustrates the correlation between temperature change and electrical resistance change when Cu electrode portions were bonded directly without a bonding film (Comparative Example).

[0156] FIG. 19 illustrates the correlation between temperature change and electrical resistance change when Cu electrode portions were bonded by a Ti bonding film (0.5 nm) (Example).

[0157] FIG. 20 illustrates the correlation between temperature change and electrical resistance change when Cu electrode portions were bonded by a Mn bonding film (0.5 nm) (Example).

[0158] FIG. 21 illustrates hybrid bonding (related art).

EMBODIMENTS FOR CARRYING OUT THE INVENTION

[0159] The bonding of base materials using the chemical bonding method of the present invention is described below.

[0160] [Overview of Bonding Method]

[0161] The bonding of base materials by the chemical bonding method of the present invention is performed using a thin film of a metal or semiconductor, which is formed by vacuum deposition such as sputtering or ion plating and has been oxidized at least on its surface, including bonding the base materials by causing chemical bonds at the bonding interface by:

[0162] i) exposing both of the bonding films formed on the smooth surfaces of the two base materials to be bonded to a space containing moisture to hydrophilize them, and then overlaying them with each other, or

[0163] ii) exposing the bonding film formed on the smooth surface of one of the base materials to be bonded to a space containing moisture to hydrophilize it, and then overlaying it in air with an oxide thin film of a metal or semiconductor formed on the smooth surface of the other base material, or

[0164] iii) exposing the bonding film formed on the smooth surface of one of the base materials to be bonded to a space containing moisture to hydrophilize it, and then overlaying it in air with a smooth surface of the other base material made of a hydrophilic or activated metal, semiconductor, or oxide thereof.

[0165] [Base Material (Material to be Bonded)]

(1) Material

[0166] The base material to be bonded by the chemical bonding method of the present invention may be any material on which the bonding film, which will be described below in detail, can be formed by vacuum deposition in a high vacuum atmosphere using a vacuum vessel with an achieved degree of vacuum of from 1×10.sup.−3 to 1×10.sup.−8 Pa, preferably from 1×10.sup.−4 to 1×10.sup.−8 Pa, as an example. In addition to various pure metals and alloys, semiconductors such as Si wafers and SiO.sub.2 base materials, glass, ceramics, resins, oxides, etc. that can be vacuum deposited using the above method may be used as the base material (material to be bonded) in the present invention.

[0167] The two base materials may be a combination of different materials, such as metal and ceramics, as well as a combination of the same material, and even such a combination of different materials can be bonded suitably according to the bonding method of the present invention.

(2) State and Other Properties of Bonding Surface

[0168] The shape of the base material is not particularly limited, and may be any shape from a flat plate to various complex three-dimensional shapes, depending on the application and purpose. However, the part to be bonded with the other base material (bonding surface) must have a smooth surface formed with a predetermined accuracy.

[0169] A plurality of base materials may be bonded to one base material by providing a plurality of smooth surfaces to be bonded on one base material.

[0170] The smooth surface is formed to surface roughness that enables the surface roughness of the formed bonding film or oxide thin film to be 0.5 nm or less in arithmetic mean height Sa (ISO 4287) when the bonding film or oxide thin film described below is formed on this smooth surface. When the smooth surface is directly overlaid with the bonding film after hydrophilization or surface activation, the smooth surface of the base material itself is formed with an arithmetic mean height Sa of 0.5 nm or less.

[0171] The structure that can be bonded to the base material is not particularly limited and various structures can be bonded to the base material, including single crystal, polycrystalline, amorphous, and glassy structures. When the bonding film described below is formed on only one of the two base materials and the other base material is bonded without forming a bonding film, the bonding surface of the other base material without the bonding film must be formed with an oxide thin film so that chemical bonding readily occurs, or subjected to hydrophilization, or activated by removing the oxidized and contaminated layers from the base material surface by dry etching to facilitate chemical bonding.

[0172] [Bonding Film]

(1) General Materials

[0173] The bonding film used for bonding must be oxidized at least on its surface, and can be formed, for example, as a thin oxide thin film that is entirely oxidized to the interior.

[0174] The material of this bonding film is not limited as long as it forms an oxide that is stable in a vacuum and in air, and various metals, semiconductors, and their oxides may be used as the material of the bonding film.

(2) Surface Roughness of Bonding Film

[0175] In order to improve the bonding strength, the bonding interface between the bonding films, between the bonding film and the oxide thin film, or between the bonding film and the smooth surface of the other base material needs to be bonded to a greater extent.

[0176] However, if the surface of the bonding film is uneven, only the contact areas between convex portions are bonded in a point-contact state, resulting in a narrow junction and low bonding strength even if bonding is possible.

[0177] Therefore, it is preferable that the surface of the bonding film can be brought into contact over the entire area of the film surface at the atomic level during bonding so that increased bonding strength can be obtained.

[0178] Such atomic-level contact can be achieved by making the surface roughness (arithmetic mean height Sa) of the bonding film as large as the unit cell when the oxide formed on at least the surface of the bonding film is crystalline.

[0179] Table 1 below shows the crystal structures and lattice constants of typical oxides.

[0180] As is clear from Table 1, the lattice constants of the representative oxides listed below range from 0.3 to 0.5 nm. In order to make the surface roughness of the bonding film as large as the unit cell of the oxide that constitutes at least the surface of the bonding film, the arithmetic mean height (Sa) should be equal to or less than 0.5 nm, which is the upper limit of the numerical range of the above lattice constant, preferably sufficiently smaller than 0.5 nm, and even more preferably 0.3 nm or less, which is the lower limit of the numerical range of the above lattice constant, thereby making it possible to contact the bonding interface at the atomic level via bonding with the OH groups of the adsorbed water molecules.

TABLE-US-00001 TABLE 1 Crystal structures and lattice constants of typical oxides Composition TiO.sub.2 ZrO.sub.2 ZnO MgO Crystal structure Tetragonal Tetragonal Hexagonal Tetragonal crystal (rutile crystal crystal (6 mm) crystal type) Lattice constant a = 0.459 a = 0.515 a = (b=) 0.325 a = 0.421 (nm) c = 0.296 c = 0.521

(3) Deposition Method

[0181] The method for forming a bonding film is not particularly limited as long as it is vacuum deposition capable of forming a thin film of metal, semiconductor, or oxide thereof on a smooth surface of a base material in a vacuum, and the bonding film may be formed by various known methods.

[0182] Such a bonding film formed by vacuum deposition has many structural defects inside the film due to the rapid cooling (quenching) of high-temperature gas phase and liquid phase atoms as they reach the smooth surface of the base material during deposition, which facilitates the generation of bonds with the OH groups of water molecules and thus chemical bonds at the bonding interface.

[0183] In particular, the film can incorporate a large amount of oxygen defects and supersaturated oxygen, and sputtering, which is easy to control these elements, and vapor deposition in combination with oxygen plasma (oxygen radicals) can be suitably used for the formation of the bonding film in the present invention.

[0184] As the bonding film, a thin film of a metal or semiconductor that is oxidized at least on its surface can be used for bonding. For example, in applications where transparency and insulation are required in the junction, the bonding film is formed as an oxide thin film that is entirely oxidized.

[0185] When the bonding film as such an oxide thin film is formed by sputtering or evaporation with oxygen plasma (oxygen radicals), an oxide target may be sputtered, or an oxide solid may be evaporated and deposited, in which the starting material of the deposition is an oxide. Alternatively, the film may be formed by deposition by a method such as reactive sputtering, in which an oxide formed by reacting an oxide-forming element and oxygen in a vacuum vessel is deposited on the smooth surface of the base material.

[0186] In this case, the bonding performance may be enhanced by controlling oxygen defects and supersaturated oxygen to increase the defects inside the film to increase the bonding density of water molecules with OH groups and the atomic mobility on the film surface, or by depositing only a few atomic layers of the surface layer of the bonding film under conditions that result in such a defect-rich condition.

[0187] Furthermore, after forming a metal or semiconductor thin film by sputtering or vapor deposition, the metal or semiconductor thin film formed by vacuum deposition is subsequently oxidized to form an oxide thin film by introducing oxygen into the vacuum vessel or by taking out the base material on which the metal or semiconductor thin film is formed to air, and the oxide thin film may be used as the aforementioned bonding film.

[0188] The surface of the oxide thin film formed in this manner can form many defects as oxide.

[0189] For applications where properties such as transparency and insulation are not required of the bonding film, the bonding film should be oxidized at least on its surface. As described above, a thin film of a metal or semiconductor formed by vacuum deposition may be exposed to a space containing oxygen under conditions such that only its surface portion is oxidized to obtain the aforementioned bonding film.

[0190] In general, the surface roughness of thin films formed by vacuum deposition increases as the film thickness increases. Therefore, if it is necessary to form a relatively thick bonding film, the energy treatment sputtering (ETS) method, which performs sputtering deposition and ion etching simultaneously, may be used to obtain a bonding film with the aforementioned surface roughness (arithmetic mean height Sa of 0.5 nm or less), or the bias sputtering method, in which sputtering is performed while a bias is applied to the substrate, may be used for deposition.

[0191] The ETS method and the bias sputtering method can form thick bonding films while maintaining a small surface roughness.

[0192] The ETS method and bias sputtering method also have significant industrial advantages, such as the ability to form thick oxide thin films with small surface roughness even when the surface roughness of the base material is relatively large, and the elimination of the need for high-precision polishing of the base material surface.

(4) Degree of Vacuum

[0193] Impurity gases such as oxygen, water, and carbon remaining in the vacuum vessel are incorporated into the bonding film to be formed, and degrade the physical properties of the bonding film.

[0194] In addition, when impurity gases such as oxygen and carbon in the vacuum vessel are adsorbed on the surface of the formed bonding film, they stabilize the chemical state of the surface and inhibit the chemical bonding of the bonding film at the bonding interface.

[0195] Therefore, the achieved degree of vacuum of the vacuum vessel must be higher than 10.sup.−3 Pa, which is equal to or less than one hundredth of 10.sup.−1 Pa, where the mean free path of residual gas is equal to the size of the vacuum vessel.

[0196] In order to suppress gas adsorption on the surface of the bonding film, the degree is more preferably better than 10.sup.−4 Pa, which is equivalent to 1 Langmuir.

[0197] It is even better and more ideal to perform thin film formation and bonding in an ultra-high vacuum environment of 10.sup.−6 Pa or lower, while maintaining the purity of the oxygen and other additive gases.

(5) Thickness of Bonding Film to be Formed

[0198] In order to obtain the physical properties of a bonding film, the thickness of the film must be at least equal to or greater than the lattice constant (from 0.3 to 0.5 nm from Table 1 above) when the oxide constituting the bonding film to be formed is crystalline, and the lower limit is 0.3 nm, preferably 0.5 nm.

[0199] On the other hand, when insulating properties are required for a bonding film, a thick thin film may be required from the viewpoint of breakdown voltage. When optical properties are required for a bonding film, a thin film with a certain thickness may be required from the viewpoint of wavelength.

[0200] However, in general deposition methods, increasing the thickness increases the surface roughness, which degrades the bonding performance.

[0201] In this regard, according to the ETS method or bias sputtering method described above, it is also possible to form bonding films with small surface roughness while increasing the thickness. However, a very long deposition time is required to deposit a bonding film of 5 μm or more, which makes it difficult to form industrially. Therefore, the upper limit of the thickness of the bonding film is 5 μm, preferably 1 μm.

[0202] Therefore, the thickness of the bonding film is preferably from 0.3 nm to 5 μm, and more preferably from 0.5 nm to 1 μm.

[0203] [Hydrophilization Treatment]

[0204] The bonding film formed on the smooth surface of the base material as described above is exposed to a space containing moisture before the base material is overlaid with another base material, causing moisture to be adsorbed to the base material surface to hydrophilize the surface.

[0205] Such hydrophilization may be performed by taking out the base material with a bonding film formed by vacuum deposition from a vacuum vessel to air, whereby moisture in air is adsorbed to the surface of the bonding film.

[0206] The hydrophilization of the bonding film may be performed in a vacuum vessel by introducing moisture or air containing moisture into the vacuum vessel in which the bonding film is deposited.

[0207] As described above, in a configuration where a thin (unoxidized) film of a metal or semiconductor is deposited by vacuum deposition, and this thin film of a metal or semiconductor is exposed to an oxygen-containing space after deposition for subsequent oxidation, the base material with a thin film of a metal or semiconductor formed by vacuum deposition may be taken out to air to simultaneously perform oxidation and hydrophilization.

[0208] [Overlaying of Base Materials]

[0209] The two base materials are overlaid with each other so that the surfaces of the bonding films, which are in a hydrophilic state, come into contact with each other, thereby causing a chemical bond at the interface of the bonding films.

[0210] This overlaying of the base materials must be done so that the “bonding films in a hydrophilic state” are in contact.

[0211] When the bonding film is hydrophilized by taking it out to air, the surface of the bonding film eventually becomes hydrophobic due to surface contamination by organic molecules and chemical stabilization over time. However, bonding must be performed when the film is in a hydrophilic state, before it becomes hydrophobic.

[0212] Although the time for which such hydrophilicity is maintained depends on the material of the bonding film to be formed, hydrophilicity is maintained for a relatively long time when hydrophilicity is maintained by taking out to air (thus, bonding is possible for a relatively long time).

[0213] For example, in a bonding test conducted by forming a 5 nm ITO thin film as a bonding film on each of two quartz substrates, the bonding strength hardly changed within 2 hours after the substrates were taken out to air, and gradually decreased after 2 hours. However, even after 24 hours, the bonding strength remained high compared to the bonding strength immediately after taking out, about 70% without heating and about 90% after heating to 300° C. After 165 hours (1 week), the bonding strength remained at about 40% of the bonding strength without heating and 70% after heating to 300° C.

[0214] Thus, as the exposure time increases, surface contamination by organic molecules progresses over time and the surface becomes chemically stabilized and hydrophobic, resulting in a decrease in bonding strength. However, since hydrophilization by taking out to air is maintained for a relatively long period of time, and thus the bonding is maintained for a relatively long time, the time between taking out and bonding should be controlled according to the required bonding strength.

[0215] In the chemical bonding method of the present invention, the hydrophilization of the bonding films and the overlaying of the base materials may be performed in air. When the hydrophilization of the bonding films is performed by introducing moisture into the vacuum vessel as described above, the base materials with the hydrophilized bonding films may be bonded by taking them out from the vacuum chamber and overlaying them with each other in an inert gas atmosphere.

[0216] By this configuration, for example, when an electronic device package is bonded (sealed) by the method of the present invention, an inert gas may be sealed in the package together with the electronic device to protect the electronic device from degradation due to oxidation or other causes.

[0217] [Hydrophilization/Activation of Other Base Material] (when Bonding Film is Formed on Only One Side)

[0218] In the chemical bonding method of the present invention, a bonding film is formed only on the smooth surface of one base material to be bonded, and an oxide thin film is formed on the smooth surface of the other base material by a known method. Alternatively, bonding may be performed by overlaying the smooth surface of one base material with a bonding film on a surface that has been hydrophilized or activated by a known method to facilitate chemical bonding.

[0219] The aforementioned oxide thin film formed on the other base material need not be of the same material as the bonding film formed on the smooth surface of one base material, and these films may be of different materials.

[0220] In the bonding method, activation of the smooth surface of the other base material may be performed in the same vacuum as the formation of the bonding film by removing oxidized or contaminated layers from the smooth surface of the other base material by dry etching or other process.

[0221] The material of the other base material may be a metal, a semiconductor such as Si, or even an oxide of these materials, as long as the base material can be hydrophilized or activated to facilitate chemical bonding, and the material is not particularly limited.

[0222] By using this bonding method, where the bonding film is formed only on the smooth surface of one base material, the bonding film formed by the oxide thin film can be used for electrical insulation between the base materials to be bonded and for adjusting the optical characteristics between the base materials.

[0223] [Heating after Bonding]

[0224] The base materials bonded in the above manner may be further heated by a known method after bonding to increase the bonding strength γ.

[0225] The heating temperature is not particularly limited, but in the case of bonding a substrate with electronic devices, etc., the heating temperature is preferably be 400° C. or lower, for example, about 300° C. to prevent damage to the electronic devices, etc. to significantly improve the bonding strength compared to when the substrate is not heated.

[0226] The bonding strength can be improved in either case, whether the heating temperature is increased in steps up to the target temperature or the heating temperature is increased at once up to the target temperature.

[0227] However, in the experimental example described below (Y.sub.2O.sub.3 bonding film) in which the heating temperature was 300° C., compared with heating to 300° C. by increasing the heating temperature in steps of 100° C., heating to 300° C. at once resulted in a 30% greater increase in the bonding strength γ. Therefore, heating to the target temperature at once after bonding is preferable for improving the bonding strength γ.

[0228] [Application to Hybrid Bonding]

[0229] The chemical bonding method of the present invention described above can be applied to the “hybrid bonding” described above, which is used for three-dimensional integration of multiple electronic devices.

[0230] The method of forming a bonding surface with an electrode portion and an insulating portion on the electronic device to be bonded may be performed as follows as described in FIG. 21: an insulating material (oxide) with a recess that serves as an electrode portion is laminated onto the device, a barrier metal and an electrode material (for example, metal such as Au or Cu) are laminated onto the insulating material, followed by CMP polishing or other processes.

[0231] The bonding surface of the electronic device to be bonded in the hybrid bonding of the present invention should be at least partially smooth. As long as the bonding surface has such a smooth surface, the entire bonding surface may be formed as a smooth surface as illustrated in FIG. 9, or the surface of the electrode portion may be depressed relative to the surface of the insulating portion, as in the case of conventional hybrid bonding described with reference to FIG. 21, or, conversely, the insulating portion may be depressed relative to the surface of the electrode portion.

[0232] On each of the surfaces of the bonding surfaces, as shown in FIG. 9, a thin film of a metal or semiconductor to be a bonding film is formed by vacuum deposition, the thin film of the metal or semiconductor is exposed to air and oxidized to form a bonding film, which is an oxide film with many defects, and this bonding film is hydrophilized by moisture in air.

[0233] Then, after precisely aligning the electrode portions and insulating portions of the two bonding surfaces with each other, the bonding surfaces are overlaid with each other so that the hydrophilic bonding films (oxide films) overlap each other, thereby bonding the bonding surfaces through the bonding films is performed.

[0234] After bonding, the bonded object is heated to a predetermined temperature (for example, 100 to 300° C.) as necessary. By bonding through the bonding films in this manner, the insulating portion becomes electrically insulating without heating after bonding or after heat treatment at a low temperature, and the electrode portion (for example, Cu or Au) becomes conductive without heating after bonding or after heat treatment at a low temperature.

[0235] Conductivity may be obtained as follows: the surface of each electrode portion is formed into a dish shape with a depressed center, and the two bonding surfaces are overlaid with each other so that the insulating portions contact each other, and then heated at a predetermined temperature (for example, 100 to 300° C.) to cause the metals (such as Cu) of the electrode portions to expand thermally and make contact, thereby bringing the bonding films formed on the electrode portion surfaces into mutual contact to bond the electrode portions to each other to obtain conductivity. Furthermore, the height of one or both electrode portions may be formed higher than that of the insulating portion, and the electrode portions may be overlaid with each other so that they contact each other for bonding. That is, the structure and size of the opposing electrode portions and insulating portions are not limited as long as electrical continuity is obtained for the electrode portions to be bonded and electrical insulation is maintained for the insulating portions after bonding.

EXAMPLES

[0236] The following is a description of the results of bonding tests using the chemical bonding method of the present invention.

Experimental Example 1

Example of Bonding Using Oxide Thin Film Deposited by Sputtering Using Oxide Target as Bonding Film

(1) Experimental Method

[0237] Various oxide thin films were formed on smooth surfaces of base materials by sputtering using oxide targets, and these oxide thin films were used as bonding films for bonding.

[0238] The films were deposited by RF magnetron sputtering or bias sputtering under the conditions in Table 2 below.

[0239] The test results, unless otherwise noted, are the results of bonding films deposited by the RF magnetron sputtering method.

TABLE-US-00002 TABLE 2 Sputtering conditions Conditions RF magnetron sputtering Bias sputtering Achieved degree 1 × 10.sup.−6 Pa or less (10.sup.−7 1 × 10.sup.−5 Pa or less (10.sup.−6 of vacuum Pa level) Pa level) Sputtering gas Pure Ar (no oxygen Pure Ar and pure Ar + added) oxygen

[0240] Quartz substrates [2 inches in diameter, surface roughness Sa: 0.12 to 0.14 nm (however, only the quartz substrate used in the bias sputtering had a diameter of 4 inches and surface roughness Sa of 0.13 nm)] and Si substrates (2 inches in diameter, surface roughness Sa: 0.15 nm) were used as bonding targets (base materials). Oxide thin films were formed on these substrates by the sputtering method to form the bonding films.

[0241] The bonding films deposited by either method were vented with nitrogen gas in the load lock chamber of the sputtering equipment and then taken out to air at a humidity of 50% (room temperature: 20° C.) to hydrophilize them, and the two substrates were bonded by overlaying them with each other without pressure so that the bonding films contact each other.

[0242] Bonding was performed immediately after taking out to air.

[0243] After bonding, the free energy γ (J/m.sup.2) at the bonding interface was measured as the bonding strength by the blade method for the unheated sample and the sample heated at 100° C., 200° C., and 300° C. (and 400° C. in some test examples) for 5 minutes.

[0244] Here, the blade method evaluates the bonding strength (surface free energy at the bonding interface) γ based on the peeling length L from the tip of the blade when the blade is inserted into the bonding interface of the two base materials, as indicated in FIG. 8, and the bonding strength γ is expressed by the following equation [M. P. Maszara. G. Goetz. A. Cavigila and J. B. McKitterick: J. Appl. Phys. 64 (1988) 4943]:

γ=3/8×Et.sup.3y.sup.2/L.sup.4

[0245] wherein E is the Young's modulus of the wafer, t is the thickness of the wafer, and y is ½ the thickness of the blade.

(2) Experimental Results

(2-1) Bonding with Y.SUB.2.O.SUB.3 .Thin Films as Bonding Films

[0246] Table 3 shows the measured bonding strength of Y.sub.2O.sub.3 thin films of 2 to 10 nm thickness on each of two quartz substrates as a bonding film.

TABLE-US-00003 TABLE 3 Bonding strength of wafers with Y.sub.2O.sub.3 thin films as bonding films Bonding film thickness Bonding strength γ (J/m.sup.2) (nm/one side) Unheated 100° C. 200° C. 300° C. 2 0.15 0.44 0.62 0.82 5 0.13 0.56 0.73 0.89 10 0.12 0.52 0.72 0.90

[0247] In the entire range of film thickness from 2 to 10 nm, the arithmetic mean height Sa of the Y.sub.2O.sub.3 thin film surfaces was about 0.14 nm or less, which was about the same as that of the quartz substrate surfaces.

[0248] The above results indicate that the bonding strength γ (J/m.sup.2) without heating was low, ranging from 0.12 to 0.15 J/m.sup.2, but confirmed that the bonding could be achieved in air even without heating.

[0249] The bonding strength was not increased by pressure.

[0250] On the other hand, the change in bonding strength γ was measured as the heating temperature of the substrate after bonding was increased in steps of 100° C. The bonding strength increased as the heating temperature increased, reaching 0.82 to 0.90 J/m.sup.2 after heating to 300° C.

[0251] The test results shown in Table 3 above show the heating temperature increased in steps of 100° C., whereas Table 4 below shows the bonding strength γ when the substrates after bonding were heated by increasing the temperature to 300° C. at once.

TABLE-US-00004 TABLE 4 Difference in bonding strength γ based on differences in control conditions of heating temperature (Y.sub.2O.sub.3 film) Bonding strength γ(J/m.sup.2) Heating temperature control conditions Film thickness 5 nm/one side Stepwise temperature increase in steps of 0.89 (quartz substrate) 100° C. up to 300° C. Temperature increased to 300° C. at once 1.23 (quartz substrate) 1.78 (Si substrate)

[0252] The above results confirmed that the increase in bonding strength was greater when the temperature was increased to 300° C. at once, compared to when the heating temperature was increased in steps of 100° C.

[0253] In all samples, bubble generation due to heat treatment was not observed at the bonding interface, indicating that the bonding by the chemical bonding method of the present invention can be used for bonding components that require transparency in the junction and for bonding electronic components (insulating portions) that require homogeneity in the bonding state.

[0254] FIG. 1 shows a cross-sectional TEM image of a bonded sample of Si substrates (γ=1.78 J/m.sup.2) that were bonded with Y.sub.2O.sub.3 thin films (5 nm) as bonding films and then heated to 300° C. at once.

[0255] As shown in FIG. 1, a layer of low density exists at the bonding interface, but the bonding interface was observed to have disappeared in some places.

[0256] The Y.sub.2O.sub.3 thin films contained microcrystals in the interior, and the heat treatment produced atomic rearrangement areas where the bonding interface disappeared.

(2-2) Bonding Using ZrO.SUB.2 .Thin Films as Bonding Films

[0257] Table 5 shows the results of measuring the bonding strength of ZrO.sub.2 thin films of 2 to 10 nm thickness, which were formed as bonding films on the two quartz substrates.

TABLE-US-00005 TABLE 5 Bonding strength of substrates with ZrO.sub.3 thin films as bonding films Bonding film thickness Bonding strength γ (J/m.sup.2) (nm/one side) Unheated 100° C. 200° C. 300° C. 2 0.14 0.16 0.22 0.37 5 0.15 0.16 0.20 0.57 10 0.17 0.18 0.18 0.40 20 0.10 0.10 0.10 0.10

[0258] For film thicknesses of 2 to 10 nm, the bonding strength γ without heating ranged from 0.14 to 0.17 J/m.sup.2, and the change in the bonding strength γ was measured as the heating temperature was increased in steps of 100° C. As a result, the bonding strength γ showed a tendency to increase as the heating temperature increased, and the bonding strength increased to 0.37 to 0.57 J/m.sup.2 when the heating temperature was increased to 300° C.

[0259] On the other hand, for a film thickness of 20 nm (one side), the bonding strength γ without heating was 0.10 J/m.sup.2, which was significantly lower than that for a film thickness of 2 to 10 nm.

[0260] For a film thickness of 20 nm (one side), no increase in the bonding strength γ was observed as the heating temperature was increased.

[0261] These results are presumably due to the fact that the arithmetic mean height Sa of the bonding film surfaces was 0.15 nm or less in the film thickness range of 2 to 10 nm, whereas the arithmetic mean height Sa of the bonding film surfaces increased to 0.22 nm for the film thickness of 20 nm.

[0262] Although the bonding strength γ was lower, it was confirmed that the substrates themselves could be bonded even when 20 nm thick ZrO.sub.2 thin films were used as bonding films.

[0263] In all samples, bubble generation due to heat treatment was not observed at the bonding interface, indicating that the bonding by the chemical bonding method of the present invention can be used for bonding components that require transparency in the junction and for bonding electronic components (insulating portions) that require homogeneity in the bonding state.

(2-3) Bonding Using TiO.SUB.2 .Thin Films as Bonding Films

[0264] FIG. 2 shows the results of measuring the bonding strength of two quartz substrates bonded by 2 nm or 5 nm thick TiO.sub.2 films formed as bonding films on the two substrates.

[0265] For both the 2 nm and 5 nm thick films, the bonding strength γ without heating was 0.1 J/m.sup.2, but reached 1 J/m.sup.2 by heating to 200° C., and reached a plateau by further heating (up to 400° C.) without any significant change in γ.

[0266] However, it was confirmed that the substrates could be bonded under all conditions, and that heat treatment after bonding was effective in improving the bonding strength γ.

[0267] In all samples, bubble generation due to heat treatment was not confirmed at the bonding interface, indicating that the bonding by the chemical bonding method of the present invention can be used for bonding components that require transparency in the junction and for bonding electronic components (insulating portions) that require homogeneity in the bonding state.

(2-4) Bonding Using ITO Thin Films as Bonding Films

[0268] FIG. 3 shows the results of measuring the bonding strength of two quartz substrates bonded by 2 nm, 5 nm, 10 nm, or 20 nm thick ITO thin films formed as bonding films on the two substrates.

[0269] In the entire range of film thickness from 2 to 20 nm, the arithmetic mean height Sa of the ITO thin film surfaces was about 0.15 nm or less, all of which were comparable to the arithmetic mean height Sa of the quartz substrate surface.

[0270] The bonding strength γ without heating was 0.2 J/m.sup.2, but an increase in the bonding strength γ with increasing heating temperature was observed for all film thicknesses with increasing heating temperature.

[0271] After heating to 300° C., the bonding strength γ reached 1.8 J/m.sup.2 for a film thickness of 5 nm, and relatively high bonding strengths of 0.8 to 1.2 J/m.sup.2 were obtained for other film thicknesses.

[0272] FIG. 4 shows the bonding strength of a sample bonded in air with 5 nm thick ITO thin films as bonding films by the method of the present invention (Example) and the results of a sample bonded in a vacuum without taking out the quartz substrates, which also have 5 nm thick ITO thin films, from the vacuum vessel (Comparative Example).

[0273] As is clear from FIG. 4, in the case of the example using a 5 nm thick ITO thin film as a bonding film, the bonding strength γ of 1.7 J/m.sup.2 obtained by heating to 300° C. after bonding in air was comparable to the bonding strength γ of the comparative example obtained when the sample bonded in a vacuum was heated to 300° C., indicating that bonding in air can be as strong as bonding in a vacuum, depending on the conditions.

[0274] In all samples, bubble generation due to heat treatment was not confirmed at the bonding interface, indicating that the bonding by the chemical bonding method of the present invention can be used for bonding components that require transparency in the junction and for bonding electronic components (insulating portions) that require homogeneity in the bonding state.

(2-5) Bonding Using SiO.SUB.2 .Thin Films as Bonding Films

[0275] FIG. 5 shows the results of measuring the bonding strength γ of two quartz substrates bonded by 1 nm, 2 nm, or 5 nm thick SiO.sub.2 films formed as bonding films on the two substrates.

[0276] The bonding strength γ of 1 nm and 2 nm film thicknesses was 0.17 J/m.sup.2 when the films were not heated, but increased with increasing heating temperature, almost saturating at heating above 200° C. The bonding strength γ after heating to 300° C. was 0.87 J/m.sup.2 (film thickness 2 nm).

[0277] In the SiO.sub.2 thin film, the arithmetic mean height Sa of the surface increases rapidly with increasing the film thickness, and the bonding strength γ is lower for bonding with 5 nm thick SiO.sub.2 thin films compared to the film thicknesses of 1 nm and 2 nm.

[0278] However, it was confirmed that bonding was possible in all cases where SiO.sub.2 thin films of any thickness were used as bonding films.

[0279] In bonding using SiO.sub.2 thin films, it was confirmed that the bonding method of the present invention, in which the substrates with SiO.sub.2 thin films are taken out to air, exhibited higher bonding strength γ than that in which the substrates were bonded in a vacuum without taking out them from a vacuum vessel.

[0280] Since crystalline SiO.sub.2 has a very stable diamond-type covalent bonding structure, it is presumed that amorphous SiO.sub.2 has a similarly stable covalent bonding structure. Therefore, it is considered difficult to make a strong direct bond between SiO.sub.2 thin films when bonding in a vacuum.

[0281] On the other hand, in the atmospheric bonding, hydrogen bonds between OH groups are generated by the moisture adsorbed to the surfaces of the SiO.sub.2 films when they are taken out to air, and heat treatment from that state facilitates the formation of Si—Si and Si—O—Si bonds between the SiO.sub.2 thin films, which is considered to have contributed to the increase in bonding strength γ.

[0282] In the bonding with SiO.sub.2 thin films, bubble generation due to heat treatment was not observed at the bonding interface in any of the samples.

(2-6) Bonding with WO.SUB.3 .Thin Films as Bonding Films

[0283] FIG. 6 shows the results of measuring the bonding strength of two quartz substrates bonded by 2 nm, 5 nm, 10 nm, 20 nm, or 50 nm thick WO.sub.3 thin films formed as bonding films on the two substrates.

[0284] In the entire range of film thickness from 2 to 50 nm, the arithmetic mean height Sa of the WO.sub.3 thin film surfaces were constant at about 0.12 nm and was maintained at the same level as that of the quartz substrate surfaces.

[0285] The bonding strength γ was about 0.2 J/m.sup.2 for both film thicknesses without heating, and about 0.9 J/m.sup.2 at maximum after heating to 300° C. (film thicknesses of 5 and 10 nm), indicating that the dependence of bonding strength γ on film thickness is relatively small.

[0286] This is considered to be because the arithmetic mean height Sa of the surfaces of the WO.sub.3 thin films remains constant even as the film thickness increases, as described above.

[0287] The plot in FIG. 6, labeled “Si substrate (5 nm/300° C.),” shows the bonding strength γ of a sample of Si substrates with 5 nm thick WO.sub.3 thin films bonded in air and then heated to 300° C. at once. In this case, the bonding strength γ increased to 1.36 J/m.sup.2.

[0288] The results confirm that higher bonding strength γ was obtained when Si substrates were bonded and heated to 300° C. at once than when quartz substrates were bonded.

[0289] In addition, generation of bubbles due to heat treatment was not observed at the bonding interface in any of the samples bonded using WO.sub.3 thin films as bonding films.

(2-7) Comparison of Bonding Strength Between Various Bonding Films

[0290] Table 6 below shows the comparative results of bonding strength γ when quartz substrates were bonded with the aforementioned thin films of Y.sub.2O.sub.3, ZrO.sub.2, TiO.sub.2, ITO, SiO.sub.2, and WO.sub.3 as bonding films.

[0291] In Table 6, among the results of the bonding tests using each of the above bonding films, only the results for the 2 nm and 5 nm thick films, which have an arithmetic mean height Sa of 0.16 nm or less on the surface, are shown (however, only the 5 nm thick SiO.sub.2 film had a slightly larger arithmetic mean height Sa of 0.22 nm).

[0292] The bonding strength γ after heating to 300° C. represents the bonding strength γ after heating to 300° C. when the temperature was increased in steps of 100° C. The bonding strength γ in parentheses is the value when the temperature was increased to 300° C. at once.

TABLE-US-00006 TABLE 6 Comparison of bonding strength γ between various bonding films Bonding strength γ (J/m.sup.2) After heating to 300° C. [Values in parentheses are Bonding film Film thickness Unheated for temperature rise at once] Y.sub.2O.sub.3 2 nm 0.15 0.82 5 nm 0.13 0.89 (1.23) ZrO.sub.2 2 nm 0.14 0.37 5 nm 0.15 0.57 TiO.sub.2 2 nm 0.12 1.31 5 nm 0.12 0.83 ITO 2 nm 0.16 1.22 5 nm 0.16 1.82 SiO.sub.2 2 nm 0.17 0.87 5 nm 0.03 0.40 WO.sub.3 2 nm 0.26 0.67 5 nm 0.28 0.89

[0293] The above results confirmed that bonding was possible with any of the bonding films, although there were differences in bonding strength due to differences in the constituent elements of the bonding films.

[0294] In addition, it was confirmed that the bonding strength γ increased with heating after bonding in all the bonding films, confirming that heating after bonding was effective in increasing the bonding strength γ.

[0295] In addition, in all of the bonding films, heating to 300° C. significantly improved the bonding strength γ compared to the unheated case, and many of the bonded films showed high bonding strength γ exceeding 1 J/m.sup.2.

[0296] Furthermore, from the measured bonding strength of the Y.sub.2O.sub.3 thin films (5 nm), it can be inferred that further improvement in the bonding strength γ can be achieved by changing the heating conditions for the other bonding films so that they are heated to the target temperature at once.

[0297] In the bonding cases where TiO.sub.2 and SiO.sub.2 thin films were used as bonding films, the heating temperature was further increased to 400° C. and the change in bonding strength γ was measured, but no further improvement in the bonding strength γ was observed. When heating after bonding was performed, it was confirmed that even heating at 400° C. or lower, for example, 300° C., which is unlikely to cause damage to electronic devices, was effective in improving the bonding strength γ.

(2-8) Presence or Absence of Bonding Film, Change in Film Thickness, and Bonding Strength γ

[0298] Table 7 below shows the bonding strength γ of two quartz (SiO.sub.2) substrates polished to an arithmetic mean height Sa of 0.12 nm and bonded (optical contact) in air, the bonding strength γ of two quartz substrates, each with a SiO.sub.2 film (2 nm thick) with an arithmetic mean height Sa of 0.12 nm, bonded in air, and the bonding strength of two quartz substrates, each with a 200 nm thick SiO.sub.2 film formed by bias sputtering, bonded in air.

[0299] The results after heat treatment to 300° C. were measured by increasing the heating temperature in steps of 100° C. after bonding and measuring the bonding strength γ when the temperature reached 300° C.

TABLE-US-00007 TABLE 7 Relationship between the presence of bonding film, changes in film thickness, and bonding strength γ Quartz substrate SiO.sub.2 film SiO.sub.2 film (No film) (2 nm thick) (200 nm thick) Unheated 0.07 J/m.sup.2 0.17 J/m.sup.2 Pure Ar: 0.10 J/m.sup.2 Ar + O.sub.2: 0.14 J/m.sup.2 After heating 0.25 J/m.sup.2 0.87 J/m.sup.2 Pure Ar: 0.42 J/m.sup.2 or more to 300° C. (bubble generated) Ar + O.sub.2: Too many bubbles to analyze

[0300] The constituent element of quartz is SiO.sub.2, which is common to the SiO.sub.2 thin film formed as the bonding film. The above results show that when the arithmetic mean height Sa is the same, the bonding of SiO.sub.2 thin films deposited by vacuum deposition as bonding films achieve a significantly higher bonding strength γ in both the results without heating and after heating to 300° C., compared to direct bonding between quartz substrates.

[0301] The reason for this large difference in bonding strength despite the common constituent elements is that the SiO.sub.2 thin film formed by vacuum deposition has more defects than the quartz substrate, and it is assumed that the presence of these defects is one of the reasons for the increase in the bonding strength.

[0302] The results in Table 7 above confirmed that when using thin films formed by a deposition method that can form a thick film without increasing the surface roughness, such as the bias sputtering method, even relatively thick films of 200 nm could be bonded together with higher bonding strength than when bonding quartz substrates (without film) by optical contact.

[0303] These results indicate that, according to the bonding method of the present invention, bonding can be performed by increasing the film thickness as long as the surface roughness of the bonding film used for bonding can be maintained within a predetermined range.

(2-9) Exposure Time of Oxide Film to Air and Bonding Strength γ

[0304] FIG. 7 shows the results of measuring the bonding strength by forming 5 nm thick ITO thin films as bonding films on two quartz substrates and then taking out the substrates to air, varying the waiting time before bonding. Table 8 shows the magnitude (%) of the bonding strength γ after a given waiting time relative to the bonding strength (100%) when bonded immediately after taking out.

[0305] The arithmetic mean height Sa of the 5 nm thick ITO thin film surfaces was about 0.15 nm or less, which was similar to that of the quartz substrate surfaces.

[0306] The bonding strength γ showed almost no decrease when the waiting time was 2 hours or less.

[0307] Although the bonding strength γ gradually decreased after 2 hours, it was confirmed that the bonding strength γ after 24 hours was also 91% (after heating to 300° C.) of that after bonding immediately after taking out to air, which was 74% (unheated) of that after bonding immediately after taking out to air.

[0308] Furthermore, after 165 hours (1 week), the bonding strength without heating was 40% of that immediately after taking out and 70% after heating to 300° C., but the bonding was still possible.

[0309] Thus, the hydrophilization of the bonding film caused by taking out to air is maintained for a long time (thus, bonding is possible for a long time).

[0310] However, as the surface contamination by organic molecules progresses over time, the surface is chemically stabilized and gradually turns hydrophobic, resulting in a decrease in bonding strength. Therefore, it is necessary to control the time from taking out to bonding within the range where the required bonding strength can be obtained.

[0311] In all samples, bubble generation due to heat treatment was not confirmed at the bonding interface, indicating that the bonding by the chemical bonding method of the present invention can be used for bonding components that require transparency in the junction and for bonding electronic components (insulating portions) that require homogeneity in the bonding state.

TABLE-US-00008 TABLE 8 Ratio of bonding strength γ to bonding strength γ when bonded immediately after taking out (%) Waiting time (hours) 2 24 165 Ratio of bonding Unheated 99 74 39 strength (%) After heating to 99 91 73 300° C.

Experimental Example 2

Example of Bonding Using Metal Thin Films Oxidized in Air as Bonding Films

(1) Purpose of Test

[0312] In Experimental Example 1, in all of the bonding cases, a thin oxide thin film was formed directly on the substrate surface by sputtering using an oxide target, which was used as a bonding film.

[0313] On the other hand, in the present experimental example (Experimental Example 2), metal thin films (unoxidized) were formed on substrates by vacuum deposition, and the substrates with the metal thin films were taken out to air and oxidized afterward to obtain oxide thin films (in some samples, only the surface portion was oxidized) as bonding films, confirming that chemical bonding could be performed with this type of bonding films.

(2) Experimental Method

[0314] Various metal thin films were formed on the smooth surfaces of base materials (quartz or Si substrates) by DC magnetron sputtering using a metal target (pure Ar gas was used). These base materials with the metal thin films were taken out to air at 50% humidity (room temperature of 20° C.) and the oxide thin films obtained by oxidizing the metal thin films at room temperature with air were formed as bonding films, while the films were hydrophilized by moisture in air for bonding.

[0315] Bonding was performed immediately (within 5 minutes) after taking out the base materials from the vacuum vessel.

[0316] Three types of deposition equipment were used for deposition as shown in Table 9 below.

TABLE-US-00009 TABLE 9 Deposition equipment Achieved degree of vacuum Vent gas *1 Equipment 1 1 × 10.sup.−7 Pa or less (10.sup.−8 Pa level) N.sub.2 Equipment 2 1 × 10.sup.−5 Pa or less (10.sup.−6 Pa level) N.sub.2 Equipment 3 1 × 10.sup.−4 Pa or less (10.sup.−5 Pa level) Air *1 Gas to be introduced when returning the vacuum chamber to atmospheric pressure after deposition

[0317] Other conditions, such as the base material (substrate) used, heat treatment conditions after bonding, and bonding strength measurement method, are the same as in the case of “the bonding example using an oxide thin film, which has been deposited by sputtering using an oxide target, as a bonding film”, which was described earlier as Test Example 1.

(3) Experimental Results

(3-1) Bonding Results with Ti Thin Films Oxidized to Form Bonding Films

[0318] 0.3 nm to 1.0 nm thick Ti thin films were formed on two quartz substrates by DC magnetron sputtering, then taken out to air and bonded within 5 minutes.

[0319] The bonding strength γ of the quartz substrates after bonding was measured in an unheated state, after heating to 200° C., and after heating to 300° C., respectively. The measurement results are shown in Table 10.

[0320] The substrates were heated by increasing the heating temperature in steps of 100° C. using a hot plate.

TABLE-US-00010 TABLE 10 Bonding strength γ of bonding in which Ti thin film is oxidized in air to form bonding film Film thick- Vent Bonding strength γ (J/m.sup.2) Substrate Deposition ness (nm) gas Unheated 200° C. 300° C. Sa (nm) equipment 0.3 Air 0.031 0.71 0.94 0.13 Equipment 3 0.5 Air 0.025 0.86 0.95 0.13 Equipment 3 0.3 N.sub.2 0.20 0.58 1.02 Equipment 2 0.6 N.sub.2 0.14 1.02 1.25 Equipment 2 1.0 N.sub.2 0.16 1.25 Broken Equipment 2

[0321] Visual observation of the thin films before bonding showed that both the 0.3 nm thick and 0.5 nm thick Ti thin films were transparent, indicating that they had been oxidized by being taken out to air.

[0322] Although the transparency was advanced even in the 1 nm thick Ti thin film, but the metallic color still remained, suggesting that oxidation occurred only on the surface of the 1 nm thick Ti thin film.

[0323] Although the bonding strength γ in an unheated state after bonding showed a greater value in the case where the vacuum chamber was vented with nitrogen (N.sub.2) gas, it was confirmed that bonding was possible in both samples.

[0324] It was confirmed that heat treatment after bonding increased the bonding strength γ for all the samples.

[0325] The bonding strength γ after heating was higher for the films with greater thickness, and the bonding strength γ after heating to 300° C. was as high as about 1 J/m.sup.2 for the 0.3 nm thick and 0.5 nm thick films, and was so high bonding strength γ for the 1.0 nm thick film that the blade could not be inserted and the quartz substrates broke when the blade was forced to be inserted.

(3-2) Bonding Results Using Oxidized Zr Thin Film as Bonding Film

[0326] 0.3 nm thick or 0.5 nm thick Zr thin films were formed on two quartz substrates by DC magnetron sputtering, and then taken out to air for bonding. The bonding results are shown in Table 11.

[0327] In all cases, bonding was performed within 5 minutes after taking out to air.

[0328] The bonding strength γ after bonding was measured in an unheated state, after heating to 200° C., and after heating to 300° C., respectively. The substrate was heated by increasing the heating temperature in steps of 100° C. using a hot plate.

TABLE-US-00011 TABLE 11 Bonding strength γ of bonding using Zr thin film oxidized in air as bonding film Film thick- Vent Bonding strength γ (J/m.sup.2) Substrate deposition ness (nm) gas Unheated 200° C. 300° C. Sa (nm) equipment 0.3 N.sub.2 0.024 0.48 0.74 0.13 Equipment 1 0.5 N.sub.2 0.028 0.52 0.74 0.13 Equipment 1

[0329] Visual observation of the thin films before bonding showed that both the 0.3 nm and 0.5 nm thick Zr thin films were transparent, indicating that they had been oxidized by exposure to air.

[0330] After bonding, the bonding strength γ without heating was relatively low (about 0.02 to 0.03 J/m.sup.2), but it was confirmed that bonding was possible.

[0331] The bonding strength γ of both samples increased by heat treatment after bonding, and the bonding strength γ after heating to 300° C. was 0.74 J/m.sup.2 for both the 0.3 nm and 0.5 nm film thicknesses, which was a significant increase compared to the unheated state.

Experimental Example 3

Example of Bonding Wafers with Different Bonding Surface Materials

(1) Purpose of Test

[0332] In Experimental Examples 1 and 2 described above, an oxide thin film or a metal film oxidized only on its surface was formed as a bonding film on each surface of the wafers to be bonded, and it was confirmed that chemical bonding was possible with such a bonding film. In Experimental Example 3, it was confirmed that chemical bonding could be achieved when bonding wafers with bonding surfaces made of different materials, as well as when bonding wafers with a bonding film on only one of the bonding surfaces.

(2) Experimental Method

[0333] A 5 nm thick ITO thin film was used as an oxide thin film to be a bonding film. The method of thin film formation is the same as in Experimental Example 1. In some experiments, a 0.5 nm thick Ti thin film was formed by a DC magnetron sputtering method and taken out to air to oxidize the surface. The thin film formation method is the same as in Experimental Example 2, and the equipment used is Equipment 1 in Table 9.

[0334] Other conditions, such as the base material (substrate) used, heat treatment conditions after bonding, and bonding strength measurement method, are the same as in the case of the bonding example using an oxide thin film deposited by the sputtering method with an oxide target as a bonding film, which was described earlier as Experimental Example 1.

(3) Experimental Results

(3-1) Bonding Results of ITO Thin Film and Oxidized Ti Thin Film

[0335] Table 12 shows the results of bonding a 5 nm thick ITO thin film on one substrate and a 0.5 nm thick Ti thin film on the other substrate, which were formed as bonding films, taken out to air almost simultaneously, and bonded within 5 minutes. In the table, the results of bonding 5 nm thick ITO thin films are also shown for comparison, which are the same as those shown in Experimental Example 1.

[0336] The bonding strength γ of 0.35 (J/m.sup.2) was obtained when an ITO film and a surface oxidized Ti film were bonded even in an unheated state, which is greater than that of bonding ITO thin films to each other. The bonding strength γ after heating to 300° C. was 1.23 J/m.sup.2, which was lower than that of ITO thin films bonded to each other, but higher than 1 J/m.sup.2.

[0337] This indicates that, even when bonding oxide films of different types formed by different methods, high bonding strength can be obtained by taking out the thin films to air and hydrophilizing the surfaces after the formation of the thin films, and a higher bonding strength can be obtained by promoting chemical bonding at the bonding interface through heat treatment.

[0338] In this experiment, one ITO thin film and the other Ti thin film were formed respectively, and taken out to air almost simultaneously and bonded within 5 minutes. However, as shown in Experimental Example 1, since a large bonding strength can be obtained when the thin films are bonded within a certain time after being taken out to air, the timing of taking out the formed thin films to air may be different, and the waiting time between taking them out to air and bonding them may be different for the two thin films to be bonded.

[0339] In all samples, bubble generation due to heat treatment was not confirmed at the bonding interface, indicating that the bonding by the chemical bonding method of the present invention can be used for bonding components that require transparency in the junction and for bonding electronic components (insulating portions) that require homogeneity in the bonding state.

TABLE-US-00012 TABLE 12 Bonding of dissimilar thin films and bonding strength γ Bonding Bonding surface 1 surface 2 Bonding strength γ (J/m.sup.2) (Substrate: (Substrate: Un- quartz) quartz) heated 100° C. 200° C. 300° C. ITO film After forming Ti 0.35 0.48 0.68 1.23 (5 nm) (0.5 nm), the surface was exposed to air for oxidation ITO film ITO film 0.20 0.40 0.67 1.68 (5 nm) (5 nm)

(3-2) Bonding Result of Forming ITO Thin Film on One Side

[0340] Table 13 shows the results of bonding an ITO thin film of 5 nm thickness on one substrate and no film on the other substrate. The results show that the ITO thin film was bonded within 5 minutes after it was taken out from the vacuum vessel. The table also shows the results of bonding 5 nm thick ITO thin films, and the results of bonding two quartz substrates in air without forming thin films (optical contact), which are the same as those shown in Experimental Example 1.

[0341] A bonding strength γ of 0.26 (J/m.sup.2) was obtained when a 5 nm thick ITO thin film was formed on one substrate in an unheated state, and 0.41 (J/m.sup.2) was obtained after heating to 300° C. Compared to direct bonding of quartz substrates (optical contact), the bonding with an ITO thin film interposed on one of the substrates as a bonding film achieved a significantly higher bonding strength γ in both results in an unheated state and after heating to 300° C.

[0342] In an unheated state, the bonding strength γ of the bonding with a 5 nm thick ITO thin film on one substrate was slightly higher than that of the bonding with ITO thin films on both substrates. The reason for this may be as follows: the surface roughness of the quartz substrate without ITO thin film is slightly lower than that of the ITO film, and hydrogen bonding can easily occur at the contact interface even if only one of the ITO film surfaces is in a hydrophilic state. However, the bonding strength γ after heating to 300° C. was higher when ITO films were formed on both sides. The reason for this may be as follows: due to the large number of defects on the surface of the ITO film, bonding of atoms at the contact interface is more likely to occur during heat treatment when ITO films are formed on both sides.

[0343] Thus, even in the bonding of quartz substrates with ITO film on only one of the substrates, a higher bonding strength was obtained than in the direct bonding of quartz substrates (optical contact). In addition, as is clear from the aforementioned bonding results of oxidized ITO and Ti thin films bonding is possible regardless of the type of thin film, as long as both surfaces to be bonded are hydrophilic. From this, it is clear that even higher bonding strength can be obtained when the surfaces of quartz substrates without thin films are activated to promote hydrophilization.

TABLE-US-00013 TABLE 13 Formation of bonding film on one side and bonding strength γ Bonding Bonding surface 1 surface 2 Bonding strength γ (J/m.sup.2) (Substrate: (Substrate: Un- quartz) quartz) heated 100° C. 200° C. 300° C. ITO film No film 0.26 0.48 0.41 0.41 (5 nm) (Quartz substrate) ITO film ITO film 0.20 0.40 0.67 1.68 (5 nm) (5 nm) No film No film 0.07 0.07 0.18 0.25 (Quartz (Quartz substrate) substrate)

Experimental Example 4

[0344] Test to Confirm Conductivity of Electrode Portions Bonded by the Chemical Bonding Method of the Present Invention

(1) Purpose of Test

[0345] The purpose is to confirm that continuity can be obtained between electrode portions bonded by the chemical bonding method of the present invention, assuming the application of the method to hybrid bonding.

(2) Experimental Method

(2-1) Making of Wafer A

[0346] A copper (Cu) or gold (Au) electrode film (20 nm thick) was formed over a Ti base film (2 nm thick) on the gray-colored rectangle (20 mm×6 mm) in FIG. 10A of a 2-inch diameter quartz wafer.

[0347] A protective gold (Au) film (30 nm thick) was formed over a Ti base film (2 nm thick) in the central area (12 mm×6 mm) of the electrode film shown with hatching in FIG. 10B.

[0348] Then, a Y.sub.2O.sub.3 insulating protective film (10 nm thick) was formed in the area coinciding with the area where the gold (Au) protective film was formed (the cross-hatched area in FIG. 10C) to make a wafer A.

[0349] The areas of wafer A that are colored gray in FIG. 10C (the areas where the electrode film is exposed) are designated as “electrode portions A”.

(2-2) Making of Wafer B

[0350] A copper (Cu) or gold (Au) electrode film (20 nm thick) was formed over a Ti base film (2 nm thick) on the gray-colored portions in FIG. 11A of a 2-inch diameter quartz wafer.

[0351] A gold (Au) protective film (30 nm thick) was formed over the Ti base film (2 nm thick) in the areas of this electrode thin film, which are hatched in FIG. 11B.

[0352] Then, a Y.sub.2O.sub.3 insulating protective film (10 nm thick) was formed on the cross-hatched portions of the gold (Au) protective film in FIG. 11C to form a wafer B.

[0353] The areas of wafer C that are colored gray in FIG. 11C (the areas where the electrode thin film is exposed) are respectively designated as “electrode portions B”.

(2-3) Bonding Method

[0354] After a Ti or Mn thin film was formed on the entire surfaces of wafer A and wafer B as a bonding film by vacuum deposition, both the wafer A and wafer B were taken out to air to oxidize and hydrophilize the bonding films.

[0355] Then, the wafers A and B were overlaid with each other and bonded so that the electrode portions A on the wafer A and the electrode portions B on the wafer B overlap each other as shown in FIG. 12.

(2-4) Conductivity Evaluation

[0356] Using the wafers A and B bonded by the above method, the measurement circuit shown in FIG. 13 was formed.

[0357] A current was applied to this measurement circuit, and the conduction state and electrical resistance of the junction between the electrode portions A and electrode portions B, which were bonded by the chemical bonding method of the present invention, were evaluated.

(3) Experimental Results

(3-1) Bonding Between Au Electrode Portions

(3-1-1) Example

[0358] For each of the wafer A and B, a gold (Au) thin film (20 nm thick) was formed as an electrode film and a thin Ti film (0.5 nm thick) was formed as a bonding film. The bonding films were oxidized and hydrophilized by exposure to air, and then the wafers A and B were bonded at room temperature to prepare a sample.

[0359] Using the sample, the measurement circuit shown in FIG. 13 was formed. The sample was heated to 200° C. and cooled to room temperature, and the change in electrical resistance with respect to temperature change was continuously measured.

[0360] For heating, the rectangular 78 mm×58 mm area shown by the single-dotted line in FIG. 13 was clamped from above and below with a force of 25 kgf using metal with a flat surface. At this time, the electrode portions A on the wafer A and the electrode portions B on the wafer B overlapped in two areas, and a 5 mm×5 mm Teflon sheet (1.5 mm thick), slightly larger than these areas, was placed in each area and put between them. This applies a pressure of about 5 MPa to the areas where the two electrode portions overlap. In this state, heating was performed at a rate of 0.88° C./sec by gradually increasing the temperature of the electric heaters built into each of the two pieces of metal, and the temperature was held at 200° C. for 5 minutes before being allowed to cool naturally to room temperature.

(3-1-2) Comparative Example

[0361] As a comparative example, a gold (Au) electrode film, an Au protective film, and an insulating protective film were formed on each of the wafer A and wafer B by vacuum deposition, and the electrode portions A of the wafer A and the electrode portions B of the wafer B were directly bonded to each other in air to make a sample without forming bonding films on either the wafer A or wafer B, without forming bonding films.

[0362] Since Au does not oxidize in air at room temperature, atomic rearrangement occurs at the interface of Au electrodes when Au electrode films formed by vacuum deposition are overlaid with each other. Therefore, Au electrodes can be directly bonded to each other without forming a bonding film.

[0363] Using a sample of wafers bonded in this manner, the measurement circuit shown in FIG. 13 was formed. This sample was heated to 200° C. in the same manner as in the example, then cooled to room temperature and the change in electrical resistance with respect to temperature change was continuously measured.

(3-1-3) Measurement Results

[0364] FIGS. 14 and 15 show the measurement results of the example bonded through Ti bonding films (0.5 nm thick) and directly bonded without a bonding film, respectively, and Table 14 shows the electrical resistance of the samples before and after heating.

TABLE-US-00014 TABLE 14 Electrical resistance of junction of Au electrodes Electrical resistance (Ω) Before After heating Bonding film heating (200° C.) Example Ti (0.5 nm) 8.2 7.9 Comparative None (direct bonding of Au) 8.0 7.7 Example

(3-1-4) Consideration

[0365] From the above results, there was no significant difference in the measured electrical resistance between the bonding through Ti bonding films and direct bonding of Au electrodes to each other.

[0366] In the sample of the comparative example, where Au electrodes were bonded to each other without intervening bonding films, the interface resistance of the bonding surfaces was approximately zero. Therefore, the above measurement results indicate that the interface resistance between the Au electrodes is approximately zero even in the sample of the example bonded with an intervening Ti bonding film (0.5 nm).

[0367] It is inferred that these results are due to the fact that the Ti bonding film was as thin as 0.5 nm, and the presence of the bonding film did not suppress the rearrangement of atoms at the Au/Au interface, resulting in atomic rearrangement beyond the bonding interface.

(3-2) Bonding of Cu Electrodes (Bonding Using 0.3 nm Thick Bonding Films)

(3-2-1) Example

[0368] A thin film of copper (Cu) (20 nm thick) as an electrode film and a bonding film were formed on each of the wafer A and wafer B. The bonding films were oxidized and hydrophilized by exposure to air, thereby bonding the wafer A and wafer B at room temperature to make a sample.

[0369] Using the sample, the measurement circuit shown in FIG. 13 was formed. The sample was heated to 200° C., cooled to room temperature, and the change in electrical resistance with respect to temperature change was continuously measured.

[0370] A Ti thin film (0.3 nm thick) or Mn thin film (0.3 nm thick) was formed as a bonding film on each of the surfaces of the wafer A and wafer B, and the wafers were bonded to make samples.

(3-2-2) Comparative Example

[0371] As a comparative example, a copper (Cu) electrode film, an Au protective film, and an insulating protective film were formed on each of the wafer A and wafer B by vacuum deposition, and the electrode portions A of the wafer A and the electrode portions B of the wafer B were directly bonded in air to make a sample without forming a bonding film on either the wafer A or wafer B.

[0372] By bonding in air, a Cu oxide film with a thickness of several nm is formed on the surface of the Cu electrode, resulting in a very low bonding strength in an unheated state, but by raising the temperature to 200° C. after bonding, the Cu oxide film disappears and a strong bonding can be obtained.

[0373] The sample was used to form a measurement circuit similar to that shown in FIG. 13. The sample was heated to 200° C., cooled to room temperature, and the change in electrical resistance with respect to temperature change was continuously measured.

(3-2-3) Measurement Results

[0374] FIG. 16 shows the measurement results of the example bonded through a Ti bonding film (0.3 nm), FIG. 17 shows the measurement results of the example bonded through a Mn bonding film (0.3 nm), and FIG. 18 shows the measurement results of the comparative example bonded directly without a bonding film. Table 15 shows the electrical resistance of each sample before and after heating.

TABLE-US-00015 TABLE 15 Electrical resistance of junction at Cu electrode (bonding film thickness 0.3 nm) Electrical resistance (Ω) Before After heating Bonding film heating (200° C.) Example Mn (0.3 nm) 8.2 7.3 Ti (0.3 nm) 8.2 7.6 Comparative None (direct bonding of Cu) 8.5 7.3 Example

(3-2-4) Consideration

[0375] From the above results, in the state before heating, the resistance of the sample of the example bonded through the Mn and Ti bonding films was lower than that of the sample of the comparative example in which Cu electrodes were bonded directly.

[0376] After heating to 200° C., the resistance values of all the samples of the example, in which bonding was performed through Mn or Ti bonding films, were almost the same as those in the comparative example, in which the Cu electrodes were directly bonded to each other, confirming that the inclusion of the bonding films do not increase the interface resistance at the junction of the Cu electrodes.

(3-3) Bonding of Cu Electrode (Bonding Using a 0.5 nm-Thick Bonding Film)

(3-3-1) Example

[0377] A thin film of copper (Cu) (20 nm thick) was formed as an electrode thin film on each of wafer A and wafer B, and a bonding film was formed on each of them. The bonding films were oxidized and hydrophilized by exposure to air, and the wafer A and wafer B were bonded at room temperature to make a sample.

[0378] Using the sample, the measurement circuit illustrated in FIG. 13 was formed. The sample was heated to 200° C., cooled to room temperature, and then reheated to 300° C. The change in electrical resistance with respect to temperature change was continuously measured.

[0379] A Ti thin film (0.5 nm thick) or Mn thin film (0.5 nm thick) as a bonding film was formed on each of the surfaces of wafer A and B.

(3-3-2) Comparative Example

[0380] As a comparative example, a copper (Cu) electrode thin film, an Au protective film, and an insulating protective film were formed on each of the wafer A and wafer B by vacuum deposition, and the electrode portions A of the wafer A and the electrode portions B of the wafer B were directly bonded in air to make a sample without forming a bonding film on either the wafer A or wafer B.

[0381] Using the sample, a measurement circuit similar to that illustrated in FIG. 13 was formed. The sample was heated to 200° C., and the change in electrical resistance with respect to temperature change was continuously measured.

(3-3-3) Measurement Results

[0382] FIG. 19 shows the measurement results of the example bonded through a Ti bonding film (0.5 nm thick), FIG. 20 shows the measurement results of the example bonded through a Mn bonding film (0.5 nm thick), FIG. 18 shows the measurement results of the comparative example (no bonding film), and Table 16 shows the electrical resistance of each sample before and after heating.

TABLE-US-00016 TABLE 16 Electrical resistance of junction at Cu electrodes (bonding film thickness 0.5 nm) Electrical resistance (Ω) Before After heating Bonding film heating 200° C. 300° C. Example Mn (0.5 nm) 10.2 [+0.17] 7.4 [+0.01] 7.1 [−0.06] Ti (0.5 nm) 11.1 [+0.26] 8.4 [+0.11] 7.8 [+0.01] Comparative None (direct 8.5 7.3 7.7 Example bonding of Cu) The values in [ ] indicate the interfacial resistance ΔR (Ω cm.sup.2) of the junction obtained by comparison with the comparative example (no bonding film).

(3-3-4) Consideration

[0383] In the sample of the example bonded through a bonding film, the interface resistance in the bonding region was larger than that of the sample in which the Cu electrodes were directly bonded without the bonding film in an unheated state.

[0384] However, in both of the samples of the example bonded using either Mn or Ti bonding films (0.5 nm), the interfacial resistance was markedly decreased by heating to 200° C., and further decreased by heating to 300° C.

[0385] In particular, when the bonding film was Mn (0.5 nm), the resistance dropped to the same level as that in Comparison Example where the Cu electrodes were directly bonded to each other after heating to 200° C. After heating to 300° C., it was confirmed that the resistance was lower than that of the comparative example where Cu electrodes were directly bonded to each other.

[0386] Even when the bonding film is Ti (0.5 nm), it was confirmed that heating up to 300° C. lowered the resistance to the same level as that of the comparative example where the Cu electrodes were directly bonded to each other.

[0387] Therefore, it was confirmed that when the chemical bonding method of the present invention was used for hybrid bonding, good conductivity of the electrode portions was ensured by heating at a relatively low temperature of 300° C. or lower as necessary.

Experimental Example 5

[0388] Evaluation Test of Bonding Strength of Electrode Portion

(1) Purpose of Experiment

[0389] The purpose is to confirm that the required bonding strength can be obtained at the electrode portions when the chemical bonding method of the present invention is applied to hybrid bonding.

(2) Experimental Method

[0390] An electrode film made of Au or Cu (both 20 nm thick) was formed as the electrode portion through a Ti base film (2 nm thick) on each quartz substrate (2 inch diameter, surface roughness Sa: 0.10 to 0.13 nm) by sputtering using “Equipment 3” in Table 9, and an additional Ti or Mn bonding film was also formed on the electrode film.

[0391] The quartz substrates were oxidized by venting with air in the load lock chamber of the sputtering equipment, and then taken out to air at 50% humidity (room temperature of 20° C.) to hydrophilize the bonding films, and the two substrates were bonded by overlaying them so that the bonding films contact each other.

[0392] A sample in which the electrode films were directly bonded to each other without the formation of a bonding film was also prepared for comparison.

[0393] After bonding, the bonding strength (free energy at the bonding interface) γ (J/m.sup.2) was measured by the blade method without heating, after heating to 200° C., and after heating to 300° C.

(3) Experimental Results

[0394] The measurement results of the bonding strength γ (J/m.sup.2) are shown in Table 17.

TABLE-US-00017 TABLE 17 Bonding strength of electrode membrane Bonding film Film thickness Electrode on one Bonding strength γ (J/m.sup.2) Quartz Vent Thin film Material side (nm) Unheated 200° C. 300° C. Sa (nm) gas Remarks Au (20 nm) None Broken Broken Broken 0.12 Air Equipment 3 Ti 0.3 0.22 0.29 1.56 0.12 Air Equipment 3 Mn 0.3 0.15 0.31 1.39 0.13 Air Equipment 3 Ti 0.5 0.23 0.23 1.74 0.12 Air Equipment 3 Mn 0.5 0.15 0.39 1.39 0.13 Air Equipment 3 Cu (20 nm) None 0.39 Broken Broken 0.12 Air Equipment 3 Mn 0.3 0.22 0.37 4.66 0.13 Air Equipment 3 Ti 0.5 0.12 0.61 6.92 0.10 Air Equipment 3

(4) Discussion

[0395] It was confirmed that a certain level of bonding strength was obtained, although the bonding strength was lower when the electrode films (electrode portions) were bonded through bonding films compared to when they were bonded directly to each other without bonding films (“None” in Table 17).

[0396] In particular, heating after bonding has been confirmed to significantly improve bonding strength γ (J/m.sup.2), even when bonded through bonding films.

[0397] Therefore, it was confirmed that the chemical bonding method of the present invention, when applied to hybrid bonding, can be used to bond electrode portions with the required bonding strength in an unheated state or by heating at 300° C. or lower as necessary.

Experimental Example 6

[0398] Evaluation Test of Insulation of Insulating Portion

(1) Purpose of Experiment

[0399] The purpose is to confirm that the chemical bonding method of the present invention can be applied to hybrid bonding to obtain insulation in the insulating portion.

(2) Experimental Method

[0400] Ti or Mn bonding films were formed on quartz substrates (2 inches in diameter, surface roughness Sa: 0.12 to 0.13 nm), which are the insulating materials (insulating portions), by sputtering method using the equipment listed in Table 9.

[0401] The quartz substrates were vented in the load lock chamber of the sputtering equipment, and then taken out to air at 50% humidity (room temperature of 20° C.) to hydrophilize the bonding film, and the two substrates were bonded by overlaying them so that the bonding films contact each other.

[0402] After bonding, the sheet resistance of the bonding films was measured by the eddy current method in an unheated state, after heating to 200° C. and 300° C., respectively, by induction.

(3) Experimental Results

[0403] Table 18 shows the sheet resistance of the cases where oxidized Ti or Mn thin films were used as bonding films.

TABLE-US-00018 TABLE 18 Sheet resistance between insulating portions by bonding films of oxidized Ti or Mn thin films Film thickness Bonding on one Vent Sheet resistance (kΩ/sq) Quartz film side (nm) gas Unheated 200° C. 300° C. Sa (nm) Remarks Mn 0.3 Air Measurement Measurement Measurement 0.13 Equipment limit or more limit or more limit or more 3 Mn 0.5 Air Measurement Measurement Measurement 0.13 Equipment limit or more limit or more limit or more 3 Ti 0.3 Air Measurement Measurement Measurement 0.13 Equipment limit or more limit or more limit or more 3 Ti 0.5 Air Measurement Measurement Measurement 0.13 Equipment limit or more limit or more limit or more 3

(4) Discussion

[0404] Visual observation of the thin films before bonding showed that the 0.3 nm and 0.5 nm thick Mn thin films and the 0.3 nm and 0.5 nm thick Ti thin films were both transparent, indicating that they had been oxidized by being exposed to air.

[0405] After bonding, both in an unheated state and after heat treatment, the electrical resistance was so high that it could not be measured by the eddy current method. The measurement limit of the eddy current method used in the experiment is 10 kΩ/sq, and very high electrical resistance exceeding this limit was obtained without heating.

[0406] As a result of the above, it was confirmed that high insulation properties were obtained on the bonding surface of the bonding film.

[0407] Therefore, even when the chemical bonding method of the present invention was applied to hybrid bonding, there was no short circuit between adjacent electrode portions that were separated by an insulating portion due to the presence of the bonding film. In this respect, it was confirmed that the chemical bonding method of the present invention can be used for hybrid bonding.

Experimental Example 7

[0408] Evaluation Test of Bonding Strength of Insulating Portion

(1) Purpose of Experiment

[0409] The purpose is to confirm that the required bonding strength can be obtained in the insulating portions when the chemical bonding method of the present invention is applied to hybrid bonding.

(2) Experimental Method

[0410] Ti or Mn bonding films were formed on quartz substrates (2 inches in diameter, surface roughness Sa: 0.12 to 0.13 nm), which are the insulating materials (insulating portions), by sputtering method using the equipment listed in Table 9.

[0411] The quartz substrates were vented in the load lock chamber of the sputtering equipment, and then taken out to air at 50% humidity (room temperature of 20° C.) to hydrophilize the bonding film, and the two substrates were bonded by overlaying them so that the bonding films contact each other.

[0412] After bonding, the bonding strength (free energy at the bonding interface) γ (J/m.sup.2) was measured by the blade method without heating, after heating to 200° C., and after heating to 300° C.

(3) Experimental Results

[0413] Table 10 shows the bonding strength when oxidized Ti thin films were used as bonding films.

[0414] Table 19 below shows the bonding strength when oxidized Mn thin films were used as bonding films.

TABLE-US-00019 TABLE 19 Bonding strength between insulating portions through oxidized Mn thin films as bonding films Film thickness Bonding on one Vent Bonding strength γ (J/m.sup.2) Quartz film side (nm) gas Unheated 200° C. 300° C. Sa (nm) Remarks Mn 0.3 Air 0.33 0.71 0.94 0.13 Equipment 3 Mn 0.5 Air 0.31 0.85 1.27 0.12 Equipment 3

(4) Discussion

[0415] Visual observation of the thin films before bonding confirmed that both the 0.3 nm thick and 0.5 nm thick Mn thin films were transparent, indicating that they had been oxidized by being taken to air.

[0416] After bonding, the bonding strength γ without heating showed relatively low values, but it was confirmed that bonding was possible in all samples.

[0417] It was confirmed that heat treatment after bonding increased the bonding strength γ for all the samples.

[0418] Therefore, it was confirmed that the chemical bonding method of the present invention, when applied to hybrid bonding, can bond insulating portions with the required bonding strength without heating or, as necessary, with heating at 300° C. or lower.

Experimental Example 8

[0419] Evaluation Test of Bonding Strength of Insulating Portions and Electrode Portions (2)

(1) Purpose of Experiment

[0420] The objective is to confirm that the required bonding strength can be obtained when the chemical bonding method of the present invention is applied to hybrid bonding, even when the bonding films are brought into contact with each other while being heated to a predetermined temperature.

(2) Experimental Method

[0421] The sputtering method using the equipment in Table 9 was used.

[0422] In the experiment on the bonding strength of the insulating portions, a 0.5 nm thick Ti bonding film was formed on quartz substrates (2 inches in diameter, surface roughness Sa: 0.12 to 0.13 nm), which are insulating materials (insulating portions).

[0423] In the experiment on the bonding strength of the electrode portions, a 20 nm thick Au electrode film was formed on a quartz substrate (diameter 2 inches, surface roughness Sa: 0.12 to 0.13 nm) through a Ti base film (2 nm thick), and a 0.5 nm thick Ti bonding film was further formed on the electrode film.

[0424] The quartz substrate was vented in the load lock chamber of the sputtering equipment, and then taken out to air at 50% humidity (room temperature 20° C.) to hydrophilize the bonding film.

[0425] The two quartz substrates were then heated to 200° C. and 300° C., respectively, and bonded by overlaying them so that the bonding films contact each other.

[0426] The bonding strength (free energy at the bonding interface) γ (J/m.sup.2) of the samples bonded by overlaying them in such a heated state was measured by the blade method.

(3) Experimental Results

[0427] The bonding strengths obtained from the experiments are shown in Table 20 below.

TABLE-US-00020 TABLE 20 Bonding strength between insulating portions and between electrode portions when bonding films obtained by oxidizing Ti thin films are overlaid with each other after heating Film thickness Bonding strength γ Bonding on one Vent (J/m.sup.2) Quartz film side (nm) gas 200° C. 300° C. Sa (nm) Remarks Insulating Ti 0.5 N.sub.2 0.64 0.70 0.13 Equipment 1 portion Electrode Ti 0.5 N.sub.2 0.53 0.11 0.13 Equipment 1 portion

(4) Discussion

[0428] In the bonding experiments between insulating portions, visual observation of the thin films before bonding showed that the Ti films had become transparent and had been oxidized by being taken out to air. When the two substrates were heated to 200° C. and 300° C. and bonded by overlaying them so that the bonding films were in contact with each other, good bonding strength was obtained (however, it was slightly lower than the result of bonding them at room temperature in Table 10, followed by heating to the respective temperatures).

[0429] In the bonding experiment between electrode portions, two substrates heated to 200° C. were bonded by overlaying them with each other so that the bonding films were in contact with each other. The resultant bonding strength was higher than that obtained by bonding them at room temperature in Table 17, followed by heating to 200° C. When overlaying was performed after heating to 300° C., the bonding strength decreased, but it was confirmed that the bonding was possible.

[0430] In the bonding experiments between electrode portions, the reason for the lower bonding strength when the electrode portions were overlaid with each other after heating to 300° C. was that the Ti base film (2 nm thick) under the 20 nm thick Au film used as the electrode film diffused due to the heating to 300° C. and the structure of the thin film changed. Therefore, the bonding strength of the electrode portions overlaid with each other after heating to 300° C. is expected to further increase by changing the base film to a material other than Ti.

[0431] In hybrid bonding, the surfaces of the electrode portions may be brought into contact with each other by expanding the metal of the electrode portions through heating, as described with reference to FIG. 21. When the chemical bonding method of the present invention is applied to such hybrid bonding, the bonding films formed on the surfaces of the electrode portions are overlaid with each other under heating.

[0432] The present test example simulates the overlaying of bonding films under such heating conditions. As a result of the test, it was confirmed that bonding was possible even when the bonding films were overlaid with each other at elevated temperatures, and the bonding strength γ increased as the temperature was increased for both samples.

[0433] Therefore, it was confirmed that the chemical bonding method of the present invention, when applied to hybrid bonding in which contact is made between electrode portions or between insulating portions under heating, as in the bonding method described with reference to FIG. 21, can bond insulating portions and electrode portions with the required bonding strength, even when the bonding films are overlaid with each other after heating to 200° C. and 300° C.

[0434] The results in Tables 10 and 17 confirm that the insulating portions and the electrode portions can be bonded to each other with the necessary bonding strength, even when they are bonded without heating.

[0435] These results confirm that the chemical bonding method of the present invention is applicable to all hybrid bonding processes in which overlaying is performed without heating or after heating to 300° C. or lower, and that the bonding strength between insulating portions and between electrode portions can be maintained at the required strength under any temperature conditions.