CHEMICAL BONDING METHOD AND JOINED STRUCTURE
20220199569 · 2022-06-23
Assignee
Inventors
- Takehito SHIMATSU (Sendai-shi, JP)
- Miyuki UOMOTO (Sendai-shi, JP)
- Kazuo MIYAMOTO (Saitama-shi, JP)
- Yoshikazu MIYAMOTO (Saitama-shi, JP)
- Nobuhiko KATOH (Kawasaki-shi, JP)
- Takayuki MORIWAKI (Kawasaki-shi, JP)
- Takayuki SAITOH (Kawasaki-shi, JP)
Cpc classification
H01L2224/29188
ELECTRICITY
H01L21/02
ELECTRICITY
H01L2224/29186
ELECTRICITY
H01L2224/83896
ELECTRICITY
B23K20/00
PERFORMING OPERATIONS; TRANSPORTING
H01L21/2007
ELECTRICITY
International classification
Abstract
A bonded structure includes a first substrate; a second substrate placed opposite to the first substrate; an intermediate layer provided between the first substrate and the second substrate and including a first oxide thin film layered on the first substrate and a second oxide thin film layered on the second substrate; either or both of the first oxide thin film and the second oxide thin film of the intermediate layer being formed of oxide thin films having increased defects; and an interface between the first oxide thin film and the second oxide thin film=being bonded by chemical bonding, and the interface comprising a low-density portion whose density is lower than that of the two oxide thin films.
Claims
1. A bonded structure comprising: a first substrate; a second substrate placed opposite to the first substrate; an intermediate layer provided between the first substrate and the second substrate and including a first oxide thin film layered on the first substrate and a second oxide thin film layered on the second substrate; at least one of the first oxide thin film and the second oxide thin film of the intermediate layer being formed of oxide thin films having increased defects; an interface between the first oxide thin film and the second oxide thin film=being bonded by chemical bonding, and the interface comprising a low-density portion whose density is lower than that of the two oxide thin films.
2. A bonded structure comprising: a first substrate; a second substrate placed opposite to the first substrate; an intermediate layer provided between the first substrate and the second substrate and including an oxide thin film having increased defects layered on the first substrate; an interface between the oxide thin film of the intermediate layer and the second substrate being bonded by chemical bonding, and the oxide thin film at the bonded portion having a low-density portion whose density is lower than that of the oxide thin film.
3. The bonded structure according to claim 1, wherein the interface between the first oxide thin film and the second oxide thin film of the intermediate layer is bonded by chemical bonding with atomic diffusion.
4. The bonded structure according to claim 2, wherein the interface between the oxide thin film of the intermediate layer and the second substrate is bonded by chemical bonding with atomic diffusion.
5. The bonded structure according to claim 1, wherein a material constituting the oxide thin film of the intermediate layer is different from a material constituting the first substrate or the second substrate.
6. The bonded structure according to claim 2, wherein a material constituting the oxide thin film of the intermediate layer is different from a material constituting the first substrate or the second substrate.
7. The bonded structure according to claim 3, wherein a material constituting the oxide thin film of the intermediate layer is different from a material constituting the first substrate or the second substrate.
8. The bonded structure according to claim 4, wherein a material constituting the oxide thin film of the intermediate layer is different from a material constituting the first substrate or the second substrate.
9. The bonded structure according to claim 1, wherein both of the first oxide thin film and the second oxide thin film of the intermediate layer being formed of oxide thin films having increased defects.
10. The bonded structure according to claim 3, wherein both of the first oxide thin film and the second oxide thin film of the intermediate layer being formed of oxide thin films having increased defects.
11. The bonded structure according to claim 5, wherein both of the first oxide thin film and the second oxide thin film of the intermediate layer being formed of oxide thin films having increased defects.
12. The bonded structure according to claim 7, wherein both of the first oxide thin film and the second oxide thin film of the intermediate layer being formed of oxide thin films having increased defects.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0071]
[0072]
[0073]
[0074]
[0075]
[0076]
[0077]
[0078]
[0079]
[0080]
[0081]
[0082]
[0083]
[0084]
[0085]
[0086]
[0087]
[0088]
[0089]
[0090]
[0091]
[0092]
[0093]
[0094]
[0095]
EMBODIMENTS FOR CARRYING OUT THE INVENTION
[0096] The chemical bonding method of the present invention is described below.
[0097] [Overview of Bonding Method]
[0098] The chemical bonding method of the present invention performs chemical bonding using an amorphous oxide thin film formed by vacuum film deposition such as sputtering or ion plating in a vacuum vessel, including superimposing the amorphous oxide thin films formed on smooth surfaces of two substrates to be bonded, or
[0099] superimposing the amorphous oxide thin film (first oxide thin film) formed on the smooth surface of one of the substrates to be bonded on the smooth surface with an oxide thin film (second oxide thin film) formed on the other substrate, furthermore,
[0100] superimposing the amorphous oxide thin film formed on the smooth surface of one of the substrates to be bonded on the smooth surface of the other substrate having an activated smooth surface to generate chemical bonding, preferably chemical bonding with atomic diffusion at the bonding interface to bond the two substrates.
[0101] [Substrate (Material to be Bonded)]
[0102] (1) Material
[0103] The substrate to be bonded by the chemical bonding method of the present invention may be any material on which the above-described amorphous oxide thin film can be formed by sputtering, ion plating, etc., in a high vacuum atmosphere using a vacuum vessel with an ultimate 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 substrates, glass, ceramics, resins, oxides, etc. that can be vacuum deposited using the above method may be used as the substrate (material to be bonded) in the present invention.
[0104] The substrate can be bonded not only between the same materials, such as metal to metal, but also between different materials, such as metal and ceramics.
[0105] (2) State and Other Properties of Bonding Surface
[0106] The shape of the substrate 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 substrate (bonding surface) must have a smooth surface formed with a predetermined accuracy.
[0107] This smooth surface, which is bonded to another substrate, may be provided on one substrate to bond multiple substrates to one substrate.
[0108] This smooth surface is formed to surface roughness that enables the surface roughness of the formed amorphous oxide thin film to be 0.5 nm or less in arithmetic mean height Sa (ISO 4287) when the amorphous oxide thin film described below is formed on this smooth surface. When the smooth surface is surface-activated and the above-described amorphous oxide thin film is superimposed on it, the smooth surface of the substrate itself is formed to an arithmetic mean height Sa of 0.5 nm or less.
[0109] It is preferable that the gas adsorption layer, natural oxidation layer, or other altered layers are removed from the smooth surface of the substrate before the amorphous oxide thin film is formed. For example, the above-described altered layer can be removed by a known wet process such as washing with chemicals, and after the removal of the above-described altered layer, a substrate that has been hydrogen-terminated to prevent gas adsorption again can be suitably used.
[0110] The removal of the altered layer is not limited to the wet process described above, but can also be performed by a dry process, and the altered layer such as the gas adsorption layer and natural oxidation layer can be removed by reverse sputtering or other means through bombarding with rare gas ions in a vacuum vessel.
[0111] In particular, when the altered layer is removed by a dry process as described above, in order to prevent gas adsorption and oxidation on the surface of the substrate until the amorphous oxide thin film is formed after the removal of the altered layer, it is preferable to perform the removal of the altered layer in the same vacuum as that used to form the amorphous oxide thin film, and to form the amorphous oxide thin film following the removal of the altered layer, and it is more preferable to remove the altered layer using an ultra-pure inert gas to prevent the re-formation of an oxidized layer or the like after the removal of the altered layer.
[0112] The structure that can be bonded to the substrate is not particularly limited and various structures can be bonded to the substrate, including single crystal, polycrystalline, amorphous, and glassy structures. When the amorphous oxide thin film described below is formed on only one of the two substrates and the other substrate is bonded without forming an amorphous oxide thin film, the bonding surface of the other substrate without the thin film must be either formed with an oxide thin film so that chemical bonding occurs, or activated by introducing the substrate, whose surface has been hydrophilically treated outside the vacuum vessel, into the vacuum vessel, or by removing the oxidized and contaminated layers on the substrate surface by dry etching in the same vacuum as that used to form the amorphous oxide thin film, thereby facilitating the occurrence of chemical bonding.
[0113] [Amorphous Oxide Thin Film]
[0114] (1) General Materials
[0115] The amorphous oxide thin film used for bonding may be made of any oxide that is stable in a vacuum and in air, and amorphous oxide thin films formed of various oxides may be used.
[0116] (2) Material Selection Based on Electronegativity or Ionicity
[0117] Chemical bonding in oxide thin films is a state of coexistence of covalent and ionic bonding. However, in an amorphous oxide thin film with high covalent bonding properties, surface atoms stabilize their energy states by covalently bonding with each other in a two-dimensional manner, making it difficult for chemical bonding to occur at the contact interface when in contact with other amorphous oxide thin film or the surface of the substrate.
[0118] Therefore, the more an amorphous oxide thin film with high ionic bonding properties is used for bonding, the higher the bonding performance (bonding strength) becomes. In general, the electronegativity of the elements other than oxygen (oxide-forming elements) that form the oxide of the amorphous oxide thin film is smaller than the electronegativity of oxygen (3.44), and the larger the difference between the two, the greater the ionic connectivity.
[0119] When the electronegativity of the above-described oxide-forming elements is A and the electronegativity of oxygen is B, the degree of ionic bonding between them (referred to as “ionicity” in the present invention) is given by the following equation:
Ionicity (%)=[1−exp{−0.25(B−A).sup.2}]×100
[0120] In the chemical bonding of the present invention, it is preferable to bond by forming an amorphous oxide thin film in which the difference (B−A) between the electronegativity B of oxygen and the electronegativity A of the oxide-forming element is 1.4 or more, or the ionicity is 40% or more.
[0121] In particular, when high bonding strength is required, bonding is preferably performed by forming an amorphous oxide thin film in which the difference (B−A) between the electronegativity B of oxygen and the electronegativity A of the oxide-forming element is 1.67 or more, or having ionicity of 50% or more.
[0122] Examples of the oxide-forming element with a difference (B−A) of 1.67 or more from the electronegativity B of oxygen, or with an ionicity of 50% or more include Be, Mg, Al, Sc, Ti, V, Cr, Mn, Zn, Y, Zr, Nb, Hf, Ta or the like.
[0123] Examples of the oxide-forming element with a difference (B−A) of 1.4 or more from the electronegativity B of oxygen, or with an ionicity of 40% or more include, in addition to the oxide-forming elements listed above, Fe, Co, Ni, Cu, Ag, Ge, Ga, In, Sn, B, Si or the like.
[0124] In addition, oxides containing alkali metals, alkaline earth metals, and lanthanides, which have extremely high ionic bonding properties, further increase the ionicity, so even better bonding performance can be expected. Examples of the element include Li, Na, K, Ca, Rh, Sr, Cs, Ba, La, Ce, Pr, Nd, Yb or the like.
[0125] The amorphous oxide thin film may be a composition-modulated film in which the composition of the forming elements is changed in the thickness or in-plane direction, and in particular, may be a film in which only a few atomic layers of the thin film surface are compositionally modulated to a composition with high ionic bonding.
[0126] Alternatively, it may be a multilayer structure with an amorphous oxide thin film with high ionic bonding properties on the surface.
[0127] Furthermore, since the spatial atomic positions of amorphous materials are not as clear as those of crystalline materials, chemical bonding can be achieved between amorphous oxide thin films of different materials if chemical bonding can be achieved at the contact interface. Therefore, in a configuration in which an amorphous oxide thin film is formed on each of the smooth surfaces of one substrate and the other substrate for bonding, the amorphous oxide thin film formed on the smooth surface of one substrate (first oxide thin film) and the amorphous oxide thin film formed on the smooth surface of the other substrate (second oxide thin film) may be formed by oxides composed of different oxide-forming elements.
[0128] In a configuration where an amorphous oxide thin film is formed on one substrate only, and the amorphous oxide thin film formed on one substrate is superimposed on the smooth surface of the activated other substrate for bonding, the material of the substrate may be an oxide or a semiconductor such as Si, as long as the substrate can be activated to make the surface easy to chemically bond, and the material is not particularly limited.
[0129] As described above, since the surface diffusion coefficient of oxide atoms is very small, the bonding interface of the bonded amorphous oxide thin films can have areas of lower density than the density of the amorphous oxide thin films used for bonding (low-density portions), however there are no application problems because bonding can be achieved even with the occurrence of such low-density portions.
[0130] (3) Selection Based on the Physical Properties of the Oxide (Optical, Electromechanical, Etc.)
[0131] In the above description, from the viewpoint of bonding performance, the amorphous oxide thin film was selected on the basis of its electronegativity or ionicity. The preferred oxide-forming elements may be selected in place of, or in combination with, the above-described selection based on electronegativity or ionicity, taking into account the engineering application aspects (for example, refractive index and electromechanical coefficient).
[0132] For example, an amorphous oxide thin film with an appropriate optical refractive index, transmittance, and others is selected for bonding between substrates of optical components that transmit light, and for bonding of electronic devices that apply radio waves, ultrasonic waves, or the like is selected, and an amorphous oxide thin film with an appropriate density, electromechanical coefficient, and others is selected for bonding of electronic devices using radio waves, ultrasonic waves, or the like.
[0133] (4) Surface Roughness of Amorphous Oxide Thin Film
[0134] In order to achieve strong bonding, the bonding interface between the amorphous oxide thin films (first and second oxide thin films) and the amorphous oxide thin film and the smooth surface of the other substrate must be bonded over a wider area.
[0135] However, if the surface of the amorphous oxide thin film is uneven, only the contact area between the convex portions will be bonded in a point contact state, resulting in a narrow bonding area and low bonding strength even if bonding is possible.
[0136] Furthermore, in oxides, the binding energy of oxygen to metal and semi-metal elements, which are the oxide-forming elements, is large, so the transferred atoms are immediately trapped by nearby atoms of different species. Therefore, the travel distance of the atoms is very short, and if only a small gap is generated at the bonding interface, the atoms are trapped by the above-described different atoms on the surface of the same thin film in this area, making it difficult for chemical bonding to occur at the bonding interface with the different atoms of the other oxide film (or substrate).
[0137] Amorphous oxide thin films, because of their amorphous structure, differ from thin films with a crystalline structure in that their atoms exist randomly. However, like stable crystalline oxides, their composition based on stoichiometry is often stable even in an amorphous structure, and it is still difficult for atoms to migrate.
[0138] Therefore, the surface of the amorphous oxide thin film is preferably able to make contact with the entire area of the film surface at the atomic level during bonding so that atoms can move over a wide area of the bonding interface and bond with sufficient strength.
[0139] Such atomic-level contact can be achieved by making the surface roughness (arithmetic mean height Sa) of the amorphous oxide thin film as large as that of a unit cell when the oxide constituting the amorphous oxide thin film is crystalline.
[0140] Table 2 below shows the crystal structures and lattice constants of typical oxides.
[0141] As is clear from Table 2, the lattice constants of the typical oxides listed below are from 0.3 to 0.5 nm. In order to make the surface roughness of the amorphous oxide thin film as large as the unit cell of the oxide, the surface roughness should be 0.5 nm or less, which is the upper limit of the numerical range of the 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 the contact at the bonding interface at the atomic level.
TABLE-US-00002 TABLE 2 Crystal structures and lattice constants of typical oxides Composition TiO.sub.2 ZrO.sub.2 ZnO MgO Crystal structure Tetragonal Tetragonal Hexagonal Tetragonal (rutile type) (6 mm) Lattice constant a = 0.459 a = 0.515 a = a = 0.421 (nm) c = 0.296 ( b = )0.325 c = 0.521
[0142] (5) Film Formation Method
[0143] The method for forming an amorphous oxide thin film is not particularly limited as long as it is a vacuum film forming method capable of forming an oxide thin film having an amorphous structure on a smooth surface of a substrate in vacuum, and the amorphous oxide thin film may be formed by various known methods.
[0144] The amorphous oxide thin film formed by such a vacuum film forming method has many structural defects inside the film due to the rapid cooling (quenching) of high-temperature gas-phase and liquid-phase atoms that reach the smooth surface of the substrate during film formation, which makes it easy for atoms to move, and therefore easy for chemical bonding to occur at the bonding interface.
[0145] In particular, the thin film can incorporate a large amount of oxygen deficiency 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 amorphous oxide thin film in the present invention.
[0146] When forming an amorphous oxide thin film by sputtering or vapor deposition in combination with oxygen plasma (oxygen radicals), the starting material for film formation itself may be an oxide by, for example, sputtering an oxide target or vapor depositing an oxide solid. Alternatively, an amorphous oxide thin film may be formed on a smooth surface of a substrate by, for example, depositing the oxide produced by reacting an oxide-forming element with oxygen in a vacuum vessel, or reactive sputtering.
[0147] It is also possible to increase the number of defects inside a film by controlling oxygen deficiency and supersaturated oxygen to increase the atomic mobility and thereby improve the bonding performance, and it is also possible to form a film under conditions where only a few atomic layers of the surface layer of the amorphous oxide thin film are in such a defect-rich state.
[0148] In general, the surface roughness of amorphous oxide thin films increases as the thickness increases. Therefore, when it is necessary to form a relatively thick amorphous oxide thin film, the film may be formed using the energy treatment sputtering (ETS) method, in which sputtering deposition and ion etching are performed simultaneously, to obtain an amorphous oxide thin film with the above-described surface roughness (arithmetic mean height Sa). This ETS method enables the formation of thick amorphous oxide thin films while maintaining small surface roughness.
[0149] The ETS method also has significant industrial advantages, such as the ability to form thick oxide thin films with small surface roughness even when the surface roughness of the substrate is relatively large, and the elimination of the need for high-precision polishing of the substrate surface.
[0150] (6) Degree of Vacuum
[0151] Impurity gases such as oxygen, water, and carbon remaining in the vacuum vessel are incorporated into the amorphous oxide thin film to be formed, and degrade the physical properties of the oxide thin film.
[0152] In addition, when impurity gases such as oxygen, water, and carbon in the vacuum vessel are adsorbed on the surface of the formed amorphous oxide thin film, they stabilize the chemical state of the surface and inhibit the chemical bonding of the amorphous oxide thin film at the bonding interface.
[0153] Therefore, the ultimate vacuum of the vacuum vessel must be better 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.
[0154] In order to suppress gas adsorption on the surface of the amorphous oxide thin film, ultimate vacuum is more preferably better than 10.sup.−4 Pa, which is equivalent to 1 Langmuir.
[0155] It is even better and more ideal to perform thin film deposition 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.
[0156] (7) Thickness of Amorphous Oxide Thin Film to be Formed
[0157] In order to obtain the physical properties of an amorphous oxide thin 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 2 above) when the oxide constituting the amorphous oxide thin film to be formed is crystalline, and the lower limit is 0.3 nm, preferably 0.5 nm.
[0158] On the other hand, when insulating properties are required for an amorphous oxide thin film, a thick thin film may be required from the viewpoint of breakdown voltage. When optical properties are required for an amorphous oxide thin film, a thin film with a certain thickness may be required from the viewpoint of wavelength. However, in general film formation methods, increasing the thickness increases the surface roughness, which degrades the bonding performance.
[0159] In this regard, according to the ETS method described above, it is also possible to form amorphous oxide thin films with small surface roughness while increasing the thickness. However, a very long deposition time is required to deposit an amorphous oxide thin film of 5 μm or more, which makes it difficult to form industrially. Therefore, the upper limit of the thickness of the amorphous oxide thin film is 5 μm, preferably 1 μm.
[0160] Therefore, the thickness of the amorphous oxide thin film is preferably from 0.3 nm to 5 μm, and more preferably from 0.5 nm to 1 μm.
[0161] (8) Others
[0162] In the chemical bonding method of the present invention, bonding can also be performed by forming an amorphous oxide thin film only on the smooth surface of one substrate to be bonded, activating the surface of the smooth surface of the other substrate to make it easy to chemically bond, and superimposing the smooth surface of the one substrate on which the amorphous oxide thin film is formed.
[0163] In such a bonding method, the activation of the smooth surface of the other substrate may be performed by introducing the substrate whose smooth surface has been hydrophilically treated outside the vacuum vessel into the vacuum vessel, or by removing the oxidized or contaminated layer on the smooth surface of the other substrate by dry etching or other means in the same vacuum as that used to form the amorphous oxide thin film.
[0164] The material of the other substrate may be an oxide or a semiconductor such as Si, as long as the substrate can be activated to make the surface easy to chemically bond, and the material is not particularly limited.
[0165] Thus, by using a bonding method in which the amorphous oxide thin film is formed only on the smooth surface of one of the substrates, the amorphous oxide thin film can also be used for electrical insulation between the substrates to be bonded and for adjusting the optical properties between the substrates.
EXAMPLES
[0166] The bonding test using the chemical bonding method of the present invention is described below.
(1) Test Example 1 (Bonding Using TiO.SUB.2 .Amorphous Thin Film)
[0167] (1-1) Test Outline
[0168] As an amorphous oxide thin film, TiO.sub.2 thin film with an amorphous structure (hereinafter referred to as “TiO.sub.2 thin film”) was formed on the smooth surface of the substrate, and the change in surface roughness of the TiO.sub.2 film formed with respect to the change in thickness was confirmed.
[0169] In addition, the following three substrates (all 2 inches in diameter) were bonded using TiO.sub.2 thin film to check the bonding state and measure the bonding strength (surface free energy at the bonding interface) γ.
[0170] Substrate to be Bonded
[0171] Substrate combination 1: Quartz substrate-quartz substrate
[0172] Substrate combination 2: Sapphire substrate-sapphire substrate
[0173] Substrate combination 3: Si wafer-Si wafer
[0174] The electronegativity of titanium (Ti), an oxide-forming element in TiO.sub.2 thin film, is 1.54, the difference between the electronegativity of oxygen (O) (3.44) and that of Ti (1.54) is 1.9, and the ionicity of Ti is 59.4%.
[0175] (1-2) Bonding Method
[0176] Two substrates were set in a vacuum vessel with an ultimate vacuum equal to or less than 1×10.sup.−6 Pa, and a TiO.sub.2 thin film was formed on the smooth surface of each of the two substrates by RF magnetron sputtering.
[0177] Following the formation of the TiO.sub.2 thin film, the TiO.sub.2 thin films formed on the smooth surfaces of each of the two substrates were superimposed on each other in the same vacuum as that used to form the TiO.sub.2 thin film, and the substrates were bonded by pressurizing them at a pressure of about 1 MPa for 10 seconds without heating.
[0178] After bonding, the samples unheated or heat-treated in air at 100° C., 200° C., and 300° C. for 5 minutes were prepared.
[0179] Among the above-described substrates, for the bonding of the quartz substrate-quartz substrate in the substrate 1, the thickness of the TiO.sub.2 thin film formed on the bonding surfaces of both substrates was varied to 2 nm, 5 nm, 10 nm, and 20 nm per side, and the bonded samples were prepared using the TiO.sub.2 thin films of different thicknesses.
[0180] For the bonding of the sapphire substrate-sapphire substrate in the substrate 2 and the Si wafer-Si wafer in the substrate 3, the bonding was performed with a thickness of 5 nm per side.
[0181] (1-3) Measurement Method
[0182] (1-3-1) Measurement of Surface Roughness
[0183] The change in the surface roughness of the TiO.sub.2 thin film (before bonding) with respect to the change in the thickness of the TiO.sub.2 thin film was measured.
[0184] The arithmetic mean height Sa (ISO 4287) was measured as the surface roughness, and the measurement was performed on a 2 μm square area by atomic force microscope (AFM).
[0185] (1-3-2) Measurement of Bonding Strength (Surface Free Energy at Bonding Interface) γ
[0186] The magnitude of the bonding strength (surface free energy of the bonding interface) γ of the substrates bonded under each of the above bonding conditions was measured using the “blade method”.
[0187] 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 substrates, as indicated in
γ=⅜×Et.sup.3γ.sup.2/L.sup.4
[0188] where E is the Young's modulus of the wafer, t is the thickness of the wafer, and γ is ½ the thickness of the blade.
[0189] (1-4) Test Results
[0190] (1-4-1) Surface Roughness
[0191]
[0192] The arithmetic mean height Sa was the smallest at 0.18 nm for a thickness of 2 nm, and was 0.23 nm for a thickness of 20 nm, although it increased slightly as the thickness increased.
[0193] Thus, the surface roughness of the TiO.sub.2 thin film used in this example is 0.5 nm or less in all cases, which is sufficiently smaller than the lattice constant of TiO.sub.2 (a=0.459, c=0.296: see Table 2).
[0194] (1-4-2) Bonding Strength (Surface Free Energy of Bonding Interface) γ
[0195]
[0196] The bonding strength (surface free energy at the bonding interface) γ after bonding was from 1.0 to 0.62 J/m.sup.2 even for the unheated samples, and increased as the heat treatment temperature increased. After the heat treatment at 300° C., the bonding strength γ of the samples bonded using the TiO.sub.2 thin films of any thickness exceeded 2 J/m.sup.2, with the highest bonding strength γ reaching 2.9 J/m.sup.2 (thickness: 5 nm, heat treatment temperature: 300° C.).
[0197] It was also confirmed that the bonding of the sapphire substrate-sapphire substrate and the Si wafer-Si wafer bonded using the 5 nm thick TiO.sub.2 thin film showed almost the same bonding strength γ as that of the quartz substrate-quartz substrate bonding.
[0198] In the Si wafer-Si wafer bonding, the bonding strength γ was so large that it could not be evaluated by the blade method after heat treatment at 300° C. (the blade could not enter the bonding interface, and if it was inserted forcibly, the Si wafer would break), and it was confirmed that the bonding strength γ was higher than the breaking strength of the Si wafer.
[0199] Therefore, it was confirmed that bonding using the TiO.sub.2 thin film can be performed with industrially usable strength regardless of the substrate material, thickness, presence or absence of heat treatment after bonding, and heat treatment temperature.
[0200] (1-4-3) Bonding State
[0201]
[0202] The layer that appears white between the Si wafer and the TiO.sub.2 thin film is a natural oxide layer of Si that exists on the Si wafer surface. The bonding interface between the TiO.sub.2 films (first and second oxide thin films) is bonded without gaps.
[0203] It is confirmed that there is a slight bright area at the bonding interface between the TiO.sub.2 thin films, and that at the bonding interface, the density of the TiO.sub.2 thin films is slightly reduced (low-density portion) near the bonding interface.
(2) Test Example 2 (Bonding Using 8 Mol % Y.SUB.2.O.SUB.3.—ZrO.SUB.2 .Amorphous Thin Film)
[0204] (2-1) Test Outline
[0205] As an amorphous oxide thin film, Y.sub.2O.sub.3—ZrO.sub.2 thin film with amorphous structure containing 8 mol % Y.sub.2O.sub.3 (hereinafter referred to as “Y.sub.2O.sub.3—ZrO.sub.2 thin film”) was formed, and the change in surface roughness with respect to the change in thickness was measured.
[0206] The bonding of two quartz substrates with a diameter of 2 inches was performed using Y.sub.2O.sub.3—ZrO.sub.2 thin film to confirm the bonding state and to measure the bonding strength.
[0207] Here, Y.sub.2O.sub.3—ZrO.sub.2 is called “stabilized zirconia”. A small amount of yttria (Y.sub.2O.sub.3) was added as a stabilizer because pure zirconia (ZrO.sub.2) is difficult to sinter due to the large volume change caused by the phase transition associated with high temperature changes and cracks in the material during cooling.
[0208] The electronegativity of zirconium (Zr), an oxide-forming element in Y.sub.2O.sub.3—ZrO.sub.2 thin film, is 1.33, the difference between the electronegativity of oxygen (O) (3.44) and that of Zr (1.33) is 2.11, and the ionicity of Zr is 67.1%.
[0209] The electronegativity of yttrium (Y) is 1.22, the difference between the electronegativity of oxygen (O) (3.44) and that of Y (1.22) is 2.22, and the ionicity of Y is 70.8%.
[0210] (2-2) Bonding Method
[0211] Two quartz substrates were set in a vacuum vessel with an ultimate vacuum equal to or less than 1×10.sup.−6 Pa, and Y.sub.2O.sub.3—ZrO.sub.2 thin films were formed on the bonding surfaces of each of the two quartz substrates by RF magnetron sputtering.
[0212] Following the formation of the Y.sub.2O.sub.3—ZrO.sub.2 thin film, the Y.sub.2O.sub.3—ZrO.sub.2 thin films formed on the bonding surfaces of the two quartz substrates were superimposed on each other in the same vacuum as that used to form the Y.sub.2O.sub.3—ZrO.sub.2 thin film, and the bonding was performed by pressurizing the quartz substrates at a pressure of about 1 MPa for 10 seconds without heating them.
[0213] The thickness of the Y.sub.2O.sub.3—ZrO.sub.2 thin film on the bonding surface of the quartz substrate was varied to 2 nm, 5 nm, 10 nm, and 20 nm per side, and the surface roughness (arithmetic mean height Sa) of Y.sub.2O.sub.3—ZrO.sub.2 thin films at each thickness before bonding was measured by atomic force microscopy (AFM) in the same manner as for the above-described TiO.sub.2 thin film, and bonding was performed using Y.sub.2O.sub.3—ZrO.sub.2 thin films formed at each thickness.
[0214] In addition, quartz substrates bonded with each of the above thicknesses were heat-treated in air for 5 minutes without heating or at 100° C., 200° C., and 300° C., respectively, and the bonding strength (surface free energy of the bonding interface) γ of each was measured by the “blade method” described above.
[0215] (2-3) Test Results
[0216] (2-3-1) Surface Roughness Sa and Bonding Strength γ
[0217] The measurements of the surface roughness Sa of the Y.sub.2O.sub.3—ZrO.sub.2 thin film and the bonding strength (surface free energy of the bonding interface) γ of the quartz substrates bonded under each condition are shown in Table 3 below.
TABLE-US-00003 TABLE 3 Surface roughness and bonding strength of Y.sub.2O.sub.3—ZrO.sub.2 thin film Thickness (nm) 2 5 10 20 Surface roughness Sa (nm) 0.17 0.19 0.24 0.27 Bonding Unheated Unmeasurable 1.43 0.45 0.21 strength γ 100° C. Unmeasurable 1.51 0.53 0.24 (J/m.sup.2) 200° C. Unmeasurable 1.81 0.63 0.33 300° C. Unmeasurable Unmeasurable 0.76 0.38
[0218] The surface roughness Sa increased gradually as the thickness increased, however it was still 0.27 nm even at a thickness of 20 nm, and the maximum value was sufficiently small compared to 0.5 nm, and also sufficiently small compared to the lattice constant of ZrO.sub.2 (a=0.515 nm: see Table 2), the main component.
[0219] In the measurement results of the bonding strength (surface free energy of the bonding interface) γ, “Unmeasurable” in Table 3 indicates that the bonding strength was so strong that it was impossible to insert a blade into the bonding interface (the quartz substrate would have broken if a blade was forcibly inserted).
[0220] For the quartz substrates bonded using a 2 nm thick Y.sub.2O.sub.3—ZrO.sub.2 thin film, it has been confirmed that a large bonding strength exceeding the breaking strength of the quartz substrate, which cannot be measured by the blade method, has already been obtained in the unheated state.
[0221] For the quartz substrates bonded using 5 nm thick Y.sub.2O.sub.3—ZrO.sub.2 thin films, a bonding strength γ of 1.43 J/m.sup.2 was obtained immediately after bonding. The bonding strength increased with increasing the heat treatment temperature, and after heat treatment at 300° C., the bonding was so strong that it could not be evaluated by the blade method.
[0222] The bonding strength γ decreased with increasing the thickness of the Y.sub.2O.sub.3—ZrO.sub.2 thin film used for bonding, and at a thickness of 20 nm, the bonding strength γ immediately after bonding was about 0.21 J/m.sup.2, and even after heat treatment at 300° C., it remained at about 0.38 J/m.sup.2.
[0223] Thus, the decrease in the bonding strength γ with increasing thickness is mainly due to the increase in surface roughness of the Y.sub.2O.sub.3—ZrO.sub.2 thin film with increasing thickness. However, industrially usable bonding strength γ was obtained even at the maximum thickness, and it was confirmed that strong bonding could be achieved for all samples.
[0224] When the thicknesses were set to 2 nm and 5 nm, the bonding using the Y.sub.2O.sub.3—ZrO.sub.2 thin film achieved a significantly larger bonding strength γ than the bonding using the TiO.sub.2 thin film (Example 1).
[0225] On the other hand, when the thickness was set to 10 nm and 20 nm, the bonding strength of the bonding using the TiO.sub.2 thin film was higher than that of the bonding using Y.sub.2O.sub.3—ZrO.sub.2 thin film.
[0226] These results can be attributed to the higher bonding strength obtained in the Y.sub.2O.sub.3—ZrO.sub.2 thin film with lower electronegativity (higher ionicity) as a result of the surface roughness Sa being almost identical between the Y.sub.2O.sub.3—ZrO.sub.2 and TiO.sub.2 thin films when the thickness was 2 nm and 5 nm.
[0227] On the other hand, when the thicknesses was 10 nm and 20 nm, the surface roughness Sa of the Y.sub.2O.sub.3—ZrO.sub.2 thin film was larger than that of the TiO.sub.2 thin film, which likely resulted in the larger bonding strength γ for the TiO.sub.2 thin film.
[0228] Therefore, it was confirmed that the smaller the electronegativity (higher ionicity) of the oxide-forming elements and the smaller the surface roughness Sa of the amorphous oxide thin film, the stronger the bonding could be.
[0229] (2-3-2) Bonding State
[0230]
[0231]
[0232] In both samples, the layer that appears white between the Si wafer and the Y.sub.2O.sub.3—ZrO.sub.2 thin film is Si oxide formed on the substrate surface.
[0233] In both samples, the bonding interface between Y.sub.2O.sub.3—ZrO.sub.2 thin films is bonded without gaps.
[0234] It is confirmed that there is a slight bright area at the bonding interface between the Y.sub.2O.sub.3—ZrO.sub.2 thin films, and that at the bonding interface, there is a low-density portion where the density of the Y.sub.2O.sub.3—ZrO.sub.2 thin films is slightly reduced near the bonding interface.
[0235] In the sample without heat treatment (
[0236] These microcrystals were formed immediately after the film was formed, however even with the presence of such a small amount of microcrystals, bonding was possible without any problems. Therefore, it was confirmed that the amorphous oxide thin film to be formed does not necessarily have to have a completely amorphous structure, however even if it contains a small amount of crystalline material in the amorphous structure, there is no problem in bonding.
[0237] In the bonded sample using the Y.sub.2O.sub.3—ZrO.sub.2 thin film, the lattice image of microcrystals in the amorphous material is continuously observed beyond the bonding interface. From this, it is confirmed that atomic diffusion occurs at the bonding interface and crystal lattice rearrangement occurs in this bonding method, and therefore, the above bonding is accompanied by atomic diffusion at the bonding interface, and the occurrence of such atomic diffusion is considered to contribute to a high-strength bonding with no gaps at the bonding interface.
[0238] Since the above-described TiO.sub.2 thin film has a completely amorphous structure and there are no crystal grains, the occurrence of atomic diffusion cannot be confirmed by TEM images. However, since the bonding interface is similarly bonded without gaps and high strength bonding is obtained, it is considered that atomic diffusion at the bonding interface occurred to a small extent in the bonding using the TiO.sub.2 thin film described above and in the bonding using other amorphous oxide thin films described below.
[0239] Furthermore, in the sample heat-treated at 300° C. after bonding (
(3) Test Example 3 (Bonding Using Y.SUB.2.O.SUB.3 .Amorphous Thin Film)
[0240] (3-1) Test Outline
[0241] As an amorphous oxide thin film, an amorphous Y.sub.2O.sub.3 thin film (hereinafter referred to as “Y.sub.2O.sub.3 thin film”) was formed, and the change in surface roughness with respect to the change in thickness was measured.
[0242] The bonding of two quartz substrates with a diameter of 2 inches was performed using Y.sub.2O.sub.3 thin film to confirm the bonding state and to measure the bonding strength.
[0243] The electronegativity of yttrium (Y), which is an oxide-forming element in the Y.sub.2O.sub.3 thin film, is 1.22, the smallest among the materials used in this application, and the difference between the electronegativity of oxygen (O) (3.44) and that of Y (1.22) is 2.22, and the ionicity of Y is 70.8%.
[0244] (3-2) Bonding Method
[0245] The Y.sub.2O.sub.3 thin films formed on the bonding surfaces of two quartz substrates were superimposed on each other in the same manner as in the case of bonding using Y.sub.2O.sub.3—ZrO.sub.2 thin films (Test Example 2) described above, and the bonding was performed by pressurizing the quartz substrates at a pressure of about 1 MPa for 10 seconds without heating them.
[0246] The thickness of the Y.sub.2O.sub.3 thin film on the bonding surface of the quartz substrate was varied to 2 nm, 5 nm, 10 nm, and 20 nm per side, and the surface roughness (arithmetic mean height Sa) of Y.sub.2O.sub.3 thin films at each thickness before bonding was measured by atomic force microscopy (AFM), and bonding was performed using Y.sub.2O.sub.3 thin films formed at each thickness.
[0247] In addition, quartz substrates bonded with each of the above thicknesses were heat-treated in air for 5 minutes without heating or at 100° C., 200° C., and 300° C., respectively, and the bonding strength (surface free energy of the bonding interface) γ of each was measured by the “blade method” described above.
(3-3) Test Results
[0248] (3-3-1) Surface Roughness Sa and Bonding Strength γ
[0249] The measurements of the surface roughness Sa of the Y.sub.2O.sub.3 thin film and the bonding strength (surface free energy of the bonding interface) γ of the quartz substrates bonded under each condition are shown in Table 4 below.
TABLE-US-00004 TABLE 4 Surface roughness and bonding strength of Y.sub.2O.sub.3 thin film Thickness (nm) 2 5 10 20 Surface roughness Sa (nm) 0.15 0.15 0.15 0.20 Bonding Unheated Unmeasurable Unmeasurable 1.8 0.024 strength γ 100° C. Unmeasurable Unmeasurable 1.9 0.39 (J/m.sup.2) 200° C. Unmeasurable Unmeasurable 2.0 0.65 300° C. Unmeasurable Unmeasurable 2.3 0.95
[0250] The surface roughness Sa increased gradually with increasing the thickness, however it was still 0.20 nm even at a thickness of 20 nm, and the maximum value was small enough to be 0.5 nm.
[0251] In the results of the measurement of the bonding strength (surface free energy of the bonding interface) γ, for the thicknesses of 2 nm and 5 nm, a large bonding strength exceeding the breaking strength of quartz was already obtained immediately after bonding (without heating).
[0252] Even for a thickness of 10 nm, a bonding strength of 1.8 J/m.sup.2 was obtained immediately after bonding (unheated), and the bonding strength γ increased with increasing the heat treatment temperature, reaching 2.3 J/m.sup.2 after heat treatment at 300° C.
[0253] For a thickness of 20 nm, the bonding strength γ immediately after bonding (unheated) was about 0.024 J/m.sup.2, however increased to 0.95 J/m.sup.2 after heat treatment at 300° C. The difference in γ with thickness is mainly due to the difference in surface roughness.
[0254] Thus, the bonding performance of the Y.sub.2O.sub.3 thin film was superior to that of the other materials.
[0255] (3-3-2) Bonding State
[0256]
[0257]
[0258] The layer that appears white between the Si wafer and the Y.sub.2O.sub.3 thin film is a layer of Si oxide formed on the substrate surface.
[0259] The bonding interface of the Y.sub.2O.sub.3 thin film disappeared, indicating that it has excellent bonding performance.
[0260] The Y.sub.2O.sub.3 thin film was observed to have very short range lattice fringes in some places, indicating that it was an amorphous layer containing microcrystals, and it was confirmed that the amorphous oxide thin film to be formed does not necessarily have to have a completely amorphous structure, and even if it contains a small amount of crystalline material in the amorphous structure, there is no problem in bonding.
(4) Test Example 4 (Bonding Using Nb.SUB.2.O.SUB.5 .Amorphous Thin Film)
[0261] (4-1) Test Outline
[0262] As an amorphous oxide thin film, Nb.sub.2O.sub.5 thin film with an amorphous structure (hereinafter referred to as “Nb.sub.2O.sub.5 thin film”) was formed, and the change in surface roughness with respect to the change in thickness was measured.
[0263] The bonding of two quartz substrates with a diameter of 2 inches was performed using the Nb.sub.2O.sub.5 thin film to confirm the bonding state and to measure the bonding strength.
[0264] The electronegativity of niobium (Nb), an oxide-forming element in the Nb.sub.2O.sub.5 thin film, is 1.6, the difference between the electronegativity of oxygen (O) (3.44) and that of Nb (1.6) is 1.84, and the ionicity of Nb is 57.1%.
[0265] (4-2) Bonding Method
[0266] The Nb.sub.2O.sub.5 thin films formed on the bonding surfaces of the two quartz substrates were superimposed on each other in the same manner as in the case of bonding using Y.sub.2O.sub.3—ZrO.sub.2 thin films (Test Example 2) described above, and the bonding was performed by pressurizing the quartz substrates at a pressure of about 1 MPa for 10 seconds without heating them.
[0267] The thickness of the Nb.sub.2O.sub.5 thin film formed on the bonding surface of the quartz substrate was varied to 2 nm, 5 nm, 10 nm, 20 nm, 30 nm, 50 nm, 75 nm, and 100 nm per side, the surface roughness (arithmetic mean height Sa) of the Nb.sub.2O.sub.5 thin film at each thickness before bonding was measured by atomic force microscopy (AFM), and bonding was performed using the Nb.sub.2O.sub.5 thin films formed at each thickness.
[0268] In addition, quartz substrates bonded with each of the above thicknesses were heat-treated in air for 5 minutes without heating or at 100° C., 200° C., and 300° C., respectively, and the bonding strength (surface free energy of the bonding interface) γ of each was measured by the “blade method” described above.
[0269] (4-3) Test Results
[0270] (4-3-1) Surface Roughness Sa and Bonding Strength γ
[0271] The measurements of the surface roughness Sa of the Nb.sub.2O.sub.5 thin film and the bonding strength (surface free energy of the bonding interface) γ of the quartz substrates bonded under each condition are shown in Table 5 below.
TABLE-US-00005 TABLE 5 Surface roughness and bonding strength of Nb.sub.2O.sub.5 thin film Thickness (nm) 2 5 10 20 30 50 75 100 Surface roughness Sa 0.17 0.17 0.18 0.18 0.17 0.17 0.17 0.17 (nm) Bonding Unheated 0.54 0.48 0.41 0.40 0.39 0.41 0.34 0.34 strength γ 100° C. 0.77 0.69 0.57 0.56 0.50 0.48 0.46 0.47 (J/m.sup.2) 200° C. 0.93 1.00 0.83 0.80 0.79 0.82 0.65 0.72 300° C. 1.13 1.37 1.23 1.20 1.18 1.22 0.96 1.07
[0272]
[0273] The surface roughness Sa was about 0.17 nm with almost no change in the thickness from 2 nm to 100 nm, and the maximum value was sufficiently small at 0.5 nm.
[0274] The results of the measurement of the bonding strength γ showed that the values ranged from 0.34 to 0.54 J/m.sup.2 immediately after bonding (unheated), however the values increased with the increase in the heat treatment temperature and reached values exceeding about 1 J/m.sup.2 after heat treatment at 300° C., with the highest bonding strength γ being 1.37 J/m.sup.2 (thickness: 5 nm). The change in the bonding strength γ with thickness is small, because the surface roughness Sa does not change much with respect to thickness.
[0275] Even when Si wafers were bonded using Nb.sub.2O.sub.5 thin film with a thickness of 5 nm, the same level of bonding strength γ as when the quartz substrate was bonded was obtained.
[0276] (4-3-2) Bonding State
[0277]
[0278] The layer that appears white between the Si wafer and the Nb.sub.2O.sub.5 thin film is a layer of Si oxide formed on the substrate surface.
[0279] There are slightly brighter areas at the bonding interface compared to the interior of the Nb.sub.2O.sub.5 film, indicating that the density of the bonding interface is slightly lower than that of the interior of the thin film, however the bonding interface of the Nb.sub.2O.sub.5 thin film is bonded without gaps.
(5) Test Example 5 (Bonding Using Al.SUB.2.O.SUB.3 .Amorphous Thin Film)
[0280] (5-1) Test Outline
[0281] As an amorphous oxide thin film, an Al.sub.2O.sub.3 thin film with an amorphous structure (hereinafter referred to as “Al.sub.2O.sub.3 thin film”) was formed, and the change in surface roughness with respect to the change in thickness was measured.
[0282] The bonding of two quartz substrates with a diameter of 2 inches was performed using the Al.sub.2O.sub.3 thin film to confirm the bonding state and to measure the bonding strength.
[0283] The electronegativity of aluminum (Al), an oxide-forming element in the Al.sub.2O.sub.3 thin film, is 1.61, the difference between the electronegativity of oxygen (O) (3.44) and that of Al (1.61) is 1.83, and the ionicity of Al is 56.7%.
[0284] (5-2) Bonding Method
[0285] The Al.sub.2O.sub.3 thin films formed on the bonding surfaces of the two quartz substrates were superimposed on each other in the same manner as in the case of bonding using Y.sub.2O.sub.3—ZrO.sub.2 thin films (Test Example 2) described above, and the bonding was performed by pressurizing the quartz substrates at a pressure of about 1 MPa for 10 seconds without heating them.
[0286] The thickness of the Al.sub.2O.sub.3 thin film on the bonding surface of the quartz substrate was varied to 1 nm, 2 nm, 5 nm, and 10 nm per side, and the surface roughness (arithmetic mean height Sa) of the Al.sub.2O.sub.3 thin films at each thickness before bonding was measured by atomic force microscopy (AFM), and bonding was performed using the Al.sub.2O.sub.3 thin films formed at each thickness.
[0287] In addition, quartz substrates bonded with each of the above thicknesses were prepared and heat-treated in air without heating or at 100° C., 200° C., and 300° C. for 5 minutes, and the bonding strength (surface free energy of the bonding interface) γ of each was measured by the “blade method” described above.
[0288] (5-3) Test Results
[0289] (5-3-1) Surface Roughness Sa and Bonding Strength γ
[0290] The measurements of the surface roughness Sa of the Al.sub.2O.sub.3 thin film and the bonding strength (surface free energy of the bonding interface) γ of the quartz substrates bonded under each condition are shown in Table 6 below.
TABLE-US-00006 TABLE 6 Surface roughness and bonding strength of Al.sub.2O.sub.3 thin film Thickness (nm) 1 2 5 10 Surface roughness Sa (nm) 0.15 0.15 0.15 (0.15) 0.17 Bonding Unheated 0.50 0.47 0.43 (0.67) 0.34 strength γ 100° C. 0.59 0.66 0.56 0.47 (J/m.sup.2) 200° C. 0.86 0.97 0.80 0.66 300° C. 1.05 1.48 1.19 (2.45) 1.07 * In Table 6, the values in parentheses are the measurements using Al.sub.2O.sub.3 thin film formed on Si wafer.
[0291]
[0292] The surface roughness Sa increased slightly with increasing the thickness, however it was still 0.17 nm at a thickness of 10 nm, and the maximum value was sufficiently small at 0.5 nm.
[0293] The results of the measurement of the bonding strength γ showed that the values ranged from 0.34 to 0.50 J/m.sup.2 immediately after bonding (unheated), however the values increased with the increase in the heat treatment temperature and reached values exceeding about 1 J/m.sup.2 after heat treatment at 300° C., with the highest bonding strength γ being 1.48 J/m.sup.2 (thickness: 2 nm).
[0294] As shown in the results for a thickness of 5 nm, the bonding strength γ was higher for Si wafers bonded than for quartz substrates bonded both without heat treatment and heated at 300° C. In the case of Si wafer bonding, the strength reached 2.45 J/m.sup.2 after heat treatment at 300° C.
[0295] (5-3-2) Bonding State
[0296]
[0297] The layer that appears white between the Si wafer and the Al.sub.2O.sub.3 thin film is a layer of Si oxide formed on the substrate surface.
[0298] The bonding interface of the Al.sub.2O.sub.3 thin film appears brighter overall than the interior of the thin film, indicating that the density of the bonding interface is slightly lower than that of the interior of the thin film, however the bonding interface of the Al.sub.2O.sub.3 thin film is bonded without gaps.
(6) Test Example 6 (Bonding Using 9.7 wt % SnO.SUB.2.—In.SUB.2.O.SUB.3 .Amorphous Thin Film)
[0299] (6-1) Test Outline
[0300] Amorphous oxide thin films of 9.7 wt % SnO.sub.2—In.sub.2O.sub.3 (hereinafter abbreviated as “ITO” with the initial letters of Indium Tin Oxide) with amorphous structure were formed and the change in surface roughness with respect to the change in the thickness of the formed films was measured.
[0301] The bonding of two quartz substrates with a diameter of 2 inches was performed using the ITO thin film to confirm the bonding state and to measure the bonding strength.
[0302] The electronegativity of indium (In) and tin (Sn), which are the oxide-forming elements of the ITO thin film, are 1.78 and 1.96, respectively, and from the composition ratio of SnO.sub.2 and In.sub.2O.sub.3, the electronegativity of the oxide-forming elements of the ITO thin film can be considered to be 1.81. The difference between the electronegativity of oxygen (O) (3.44) and that of the oxide-forming elements of the ITO thin film (1.81) is 1.63, and the ionicity of the oxide-forming elements is 48.5%.
[0303] The electronegativity of indium (In) is 1.78, the difference between the electronegativity of oxygen (O) (3.44) and that of In (1.78) is 1.66, and the ionicity of In is 49.8%.
[0304] (6-2) Bonding Method
[0305] The ITO thin films formed on the bonding surfaces of the two quartz substrates were superimposed on each other in the same manner as in the case of bonding using Y.sub.2O.sub.3—ZrO.sub.2 thin films (Test Example 2) described above, and the bonding was performed by pressurizing the quartz substrates at a pressure of about 1 MPa for 10 seconds without heating them.
[0306] The thickness of the ITO thin film on the bonding surface of the quartz substrate was varied to 2 nm, 5 nm, and 10 nm per side, and the surface roughness (arithmetic mean height Sa) of the ITO thin films at each thickness before bonding was measured by atomic force microscopy (AFM).
[0307] In addition, quartz substrates bonded with the ITO thin films of 1 nm, 2 nm, 5 nm, 10 nm, and 20 nm in thickness per side, respectively, were prepared and heat-treated in air without heating or at 100° C., 200° C., and 300° C. for 5 minutes each, and the bonding strength (surface free energy of the bonding interface) γ of each was measured by the “blade method” described above.
[0308] (6-3) Test Results
[0309] (6-3-1) Surface Roughness Sa and Bonding Strength γ
[0310] The results of measuring the surface roughness Sa of the ITO thin film and the bonding strength (surface free energy of the bonding interface) γ of the quartz substrates bonded under each condition are shown in Table 7 below.
TABLE-US-00007 TABLE 7 Surface roughness and bonding strength of ITO thin film Thickness (nm) 1 2 5 10 20 Surface roughness Sa (nm) 0.15 0.15 0.16 Bonding Unheated 0.68 0.74 0.52 0.47 0.43 strength γ 100° C. 0.82 0.90 0.74 0.62 0.62 (J/m.sup.2) 200° C. 1.00 1.11 1.00 0.74 1.11 300° C. 1.23 1.38 1.73 1.11 1.11
[0311]
[0312] The surface roughness Sa increased slightly with increasing the thickness, however it was still 0.16 nm at a thickness of 10 nm, and the maximum value was sufficiently small at 0.5 nm.
[0313] The results of the measurement of the bonding strength γ showed that the values ranged from 0.43 to 0.74 J/m.sup.2 immediately after bonding (unheated), however the values increased with the increase in the heat treatment temperature and reached values exceeding 1 J/m.sup.2 after heat treatment at 300° C., with the highest bonding strength γ being 1.73 J/m.sup.2 (thickness: 5 nm).
[0314] As a result of bonding sapphire substrates and Si wafers using the ITO thin film with a thickness of 5 nm (one side), it was confirmed that the bonding strength γ was approximately the same as when quartz substrates were bonded together.
[0315] (6-3-2) Bonding State
[0316]
[0317] The layer that appears white between the Si wafer and the ITO thin film is a layer of Si oxide formed on the substrate surface.
[0318] The bonding interface of the ITO thin film appears brighter overall than the interior of the ITO thin film, indicating that the density of the bonding interface is slightly lower than that of the interior of the thin film, however the bonding interface of the ITO thin film is bonded without gaps.
[0319] The ITO thin film showed very short range lattice fringes in some places and was an amorphous layer containing microcrystals.
(7) Test Example 7 (Bonding Using Ga.SUB.2.O.SUB.3 .Amorphous Thin Film)
[0320] (7-1) Test Outline
[0321] As an amorphous oxide thin film, the Ga.sub.2O.sub.3 thin film with an amorphous structure (hereinafter referred to as “Ga.sub.2O.sub.3 thin film”) was formed, and the change in surface roughness with respect to the change in thickness was measured.
[0322] The bonding of two quartz substrates with a diameter of 2 inches was performed using the Ga.sub.2O.sub.3 thin film to confirm the bonding state and to measure the bonding strength.
[0323] The electronegativity of gallium (Ga), an oxide-forming element in the Ga.sub.2O.sub.3 thin film, is 1.81, the difference between the electronegativity of oxygen (O) (3.44) and that of Ga (1.81) is 1.63, and the ionicity of Ga is 48.5%.
[0324] (7-2) Bonding Method
[0325] The Ga.sub.2O.sub.3 thin films formed on the bonding surfaces of the two quartz substrates were superimposed on each other in the same manner as in the case of bonding using Y.sub.2O.sub.3—ZrO.sub.2 thin films (Test Example 2) described above, and the bonding was performed by pressurizing the quartz substrates at a pressure of about 1 MPa for 10 seconds without heating them.
[0326] The thickness of the Ga.sub.2O.sub.3 thin film on the bonding surface of the quartz substrate was varied to 1 nm, 2 nm, and 5 nm per side, and the surface roughness (arithmetic mean height Sa) of the Ga.sub.2O.sub.3 thin films at each thickness before bonding was measured by atomic force microscopy (AFM), and bonding was performed using the Ga.sub.2O.sub.3 thin films formed at each thickness.
[0327] In addition, quartz substrates bonded using the Ga.sub.2O.sub.3 thin films of each of the above-described thicknesses were heat-treated in air for 5 minutes without heating or at 100° C., 200° C., and 300° C., respectively, and the bonding strength (surface free energy of the bonding interface) γ of each was measured using the “blade method” described above.
[0328] (7-3) Test Results
[0329] (7-3-1) Surface Roughness Sa and Bonding Strength γ
[0330] The results of measuring the surface roughness Sa of the Ga.sub.2O.sub.3 thin film and the bonding strength (surface free energy of the bonding interface) γ of the quartz substrates bonded under each condition are shown in Table 8 below.
TABLE-US-00008 TABLE 8 Surface roughness and bonding strength of Ga.sub.2O.sub.3 thin film Thickness (nm) 1 2 5 Surface roughness Sa (nm) 0.16 0.16 0.18 Bonding Unheated 0.82 0.83 0.18 (0.40) strength γ 100° C. 1.00 1.23 0.37 (J/m.sup.2) 200° C. 1.53 1.53 1.27 300° C. 2.04 1.93 2.22 (2.77) * In Table 8, the values in parentheses are the measurements using Ga.sub.2O.sub.3 thin film formed on Si wafer.
[0331]
[0332] The surface roughness Sa increased slightly with increasing the thickness, however it was 0.18 nm at a thickness of 5 nm, which was sufficiently small compared to 0.5 nm.
[0333] The results of the measurement of the bonding strength γ showed that the values ranged from 0.82 to 0.18 J/m.sup.2 immediately after bonding (unheated), however with increasing the heat treatment temperature, the bonding strength γ increased, reaching about 2 J/m.sup.2 after the heat treatment at 300° C., with the highest bonding strength γ being 2.22 J/m.sup.2 (thickness: 5 nm).
[0334] The results for the thickness of 5 nm show that the bonding strength γ between the Si wafers was higher than the bonding strength γ between the quartz substrates, both for the film without heat treatment and for the film heated to 300° C. In the case of the Si wafer bonding, the bonding strength reached 2.77 J/m.sup.2 after heat treatment at 300° C.
[0335] (7-3-2) Bonding State
[0336]
[0337] The layer that appears white between the Si wafer and the Ga.sub.2O.sub.3 thin film is a layer of Si oxide formed on the substrate surface.
[0338] The bonding interface of the Ga.sub.2O.sub.3 thin film appears brighter overall than the interior of the thin film, indicating that the density of the bonding interface is slightly lower than that of the interior of the thin film, however the bonding interface of the Ga.sub.2O.sub.3 thin film is bonded without gaps.
(8) Test Example 8 (Bonding Using GeO.SUB.2 .Amorphous Thin Film)
[0339] (8-1) Test Outline
[0340] As an amorphous oxide thin film, a GeO.sub.2 thin film with an amorphous structure (hereinafter referred to as “GeO.sub.2 thin film”) was formed. The bonding of two quartz substrates with a diameter of 2 inches was performed, the bonding state was confirmed, and the change in bonding strength γ with respect to the change in the thickness of the formed film was measured by the “blade method” described above.
[0341] The electronegativity of germanium (Ge), an oxide-forming element in the GeO.sub.2 thin film, is 2.01, the difference between the electronegativity of oxygen (O) (3.44) and that of Ge (2.01) is 1.43, and the ionicity of Ge is 40%.
[0342] (8-2) Bonding Method
[0343] The GeO.sub.2 thin films formed on the bonding surfaces of the two quartz substrates were superimposed on each other in the same manner as in the case of bonding using Y.sub.2O.sub.3—ZrO.sub.2 thin films (Test Example 2) described above, and the bonding was performed by pressurizing the quartz substrates at a pressure of about 1 MPa for 10 seconds without heating them.
[0344] The thickness of the GeO.sub.2 thin film on the bonding surface of the quartz substrate was varied to 1 nm, 2 nm, 3 nm, and 5 nm per side, and the surface roughness (arithmetic mean height Sa) of the GeO.sub.2 thin films at each thickness before bonding was measured by atomic force microscopy (AFM), and bonding was performed using the GeO.sub.2 thin films formed at each thickness.
[0345] In addition, quartz substrates bonded using the GeO.sub.2 thin films of each of the above-described thicknesses were heat-treated in air for 5 minutes without heating or at 100° C., 200° C., and 300° C., respectively, and the bonding strength (surface free energy of the bonding interface) γ of each was measured using the “blade method” described above.
[0346] The surface roughness Sa of the GeO.sub.2 thin film cannot be measured because the surface reacts with moisture in the air and agglomerates, so that the surface roughness Sa was not measured in this test example.
[0347] (8-3) Test Results
[0348] (8-3-1) Surface Roughness Sa and Bonding Strength γ
[0349] The results of the measurement of the bonding strength (surface free energy of the bonding interface) γ of the quartz substrate with respect to the change in the thickness of the GeO.sub.2 thin film are shown in Table 9 below.
TABLE-US-00009 TABLE 9 Surface roughness and bonding strength of GaO.sub.2 thin film Thickness (nm) 1 2 3 5 Bonding Unheated 0.08 0.13 0.14 0.09 strength γ 100° C. 0.57 0.57 0.58 0.35 (J/m.sup.2) 200° C. 0.99 0.99 1.02 0.81 300° C. 1.09 1.21 1.16 (1.47) 0.90 * In Table 9, the values in parentheses are the measurements using GaO.sub.2 thin film formed on Si wafer.
[0350]
[0351] Immediately after bonding (unheated), the bonding strength γ ranged from 0.08 to 0.14 J/m.sup.2, and increased with increasing the heat treatment temperature. After heat treatment at 300° C., the bonding strength exceeded 1 J/m.sup.2 in the range of from 1 to 3 nm thickness, however even the highest bonding strength γ was only about 1.21 J/m.sup.2 (2 nm thickness).
[0352] However, the quartz substrates were bonded to each other regardless of the thickness.
[0353] As a result of bonding Si wafers using the GeO.sub.2 thin film with a thickness of 3 nm (one side), the bonding strength was 1.47 J/m.sup.2 after heat treatment at 300° C., which was slightly higher than that of bonding quartz substrates, however Si wafers were still bonded in this configuration.
[0354] Germanium, the oxide-forming element of the GeO.sub.2 thin film, has the highest electronegativity of 2.01 (the difference from the electronegativity of oxygen is the smallest at 1.43) and the lowest ionicity of 40.0% among the oxide-forming elements of amorphous oxide thin films used in the experiments. Therefore, from the examples of successful bonding using the GeO.sub.2 thin film, it is confirmed that bonding can be performed using an amorphous oxide thin film containing an oxide-forming element with a difference of about 1.4 in electronegativity from oxygen and ionicity of about 40%.
[0355] (8-3-2) Bonding State
[0356]
[0357] The layer that appears white between the Si wafer and the GeO.sub.2 thin film is a layer of Si oxide formed on the substrate surface.
[0358] In the sample after heating at 300° C., the bonding interface between the GeO.sub.2 thin films disappeared, indicating that strong bonding was obtained. This bonding state was also consistent with the fact that the samples bonded using the GeO.sub.2 thin film showed a significant increase in bonding strength of 9.3 times when heated to 300° C. compared to the unheated samples. This is because the melting point of the GeO.sub.2 thin film is 1115° C., which is lower (compared to other oxides), and the effect of the heat treatment at 300° C. to promote atomic diffusion at the bonding interface is relatively greater than that of other oxide thin films.
(9) Test Example 9 (Comparison of Types of Amorphous Oxide Thin Films and Bonding Strength)
[0359] (9-1) Test Outline
[0360] As an amorphous oxide thin film, in addition to the amorphous oxide thin films used in Test Examples 1 to 8, SiO.sub.2 thin films with an amorphous structure (hereinafter referred to as “SiO.sub.2 thin films”) were also formed, and the surface roughness Sa of each film was measured and compared in the same manner as in Test Examples 1 and 2.
[0361] The bonding strength γ of quartz substrate-quartz substrates (2 inches in diameter, neither of which was heated during bonding) bonded using each thin film was measured and compared using the “blade method” described above, both when unheated and after bonding and heat treatment at 300° C. for 5 minutes in air.
[0362] (9-2) Test Results
[0363] (9-2-1) Measurement Results
[0364] The surface roughness Sa of each thin film and the bonding strength γ of the quartz substrate-quartz substrate bonded using each thin film are shown in Table 10 below.
TABLE-US-00010 TABLE 10 Surface roughness and bonding strength of various amorphous oxide thin films Material Test Electronegativity Surface Bonding strength γ (J/m.sup.2) Example (difference Ionicity Thickness roughness Heated at No.) from oxygen) (%) (nm) Sa (nm) Unheated 300° C. Y.sub.2O.sub.3 1.22 (2.22) 70.8 5 0.15 Unmeasurable Unmeasurable (Test Example 3) Y.sub.2O.sub.3—ZrO.sub.2 1.32* (2.12*) 66.8* 5 0.19 1.43 Unmeasurable (Test Example 2) TiO.sub.2 (Test 1.54 (1/9) 59.4 5 0.18 0.77 2.9 Example 1) Nb.sub.2O.sub.5 1.60 (1.84) 57.1 5 0.17 0.48 1.37 (Test Example 4) Al.sub.2O.sub.3 1.61 (1.83) 56.7 2 0.15 0.47 1.48 (Test Example 5) ITO (Test 1.81* (1.63)* 48.5* 5 0.15 0.52 1.73 Example 6) Ga.sub.2O.sub.3 1.81 (1.63) 48.5 2 0.16 0.83 1/93 (Test Example 7) SiO.sub.2 1.9 (1.54) 44.7 2 0.15 0.045 0.045 GeO.sub.2 2.01 (1.43) 40.0 2 — 0.13 1.21 (Test Example 8) *The value for Y.sub.2O.sub.3—ZrO.sub.2 was calculated as the weighted average of ZrO.sub.2 and Y.sub.2O.sub.3, and the value for ITO was calculated as the weighted average of SnO.sub.2 and In.sub.2O.sub.3.
[0365] The surface roughness Sa of the amorphous oxide thin films of both materials was small, less than 0.2 nm.
[0366] It has been confirmed that the smaller the electronegativity of the oxide-forming elements in the amorphous oxide thin film, the greater the bonding strength γ, both for the unheated and heated samples.
[0367] Comparing the unheated samples after bonding, it can be seen that the bonding strength γ tends to increase with decreasing the electronegativity of the oxide-forming elements in the amorphous oxide thin film (with increasing the difference between the electronegativity of the oxide-forming elements and that of oxygen).
[0368] In particular, for the Y.sub.2O.sub.3 thin film with the smallest electronegativity (the difference with the electronegativity of oxygen is the largest), the bonding strength γ is so large that it cannot be evaluated by the blade method even in the unheated state (exceeding the fracture strength of the quartz substrate).
[0369] With the exception of the SiO.sub.2 thin film, the bonding strength γ of all the samples increased with heat treatment, and in the case of the sample using the Y.sub.2O.sub.3—ZrO.sub.2 thin film, heat treatment at 300° C. increased the bonding strength to a level that could not be evaluated by the blade method.
[0370] The bonding strength γ of the samples bonded using the SiO.sub.2 thin film was lower than that of the samples bonded using other amorphous oxide thin films, however the bonding was still possible.
[0371] (9-2-2) Relationship Between Electronegativity, Ionicity, and Bonding Strength γ
[0372]
[0373]
[0374]
[0375] As can be seen from
[0376] Of these, bonding using Y.sub.2O.sub.3 thin film, which has the lowest electronegativity (the largest difference from the electronegativity of oxygen), has excellent bonding performance, as the bonding interface disappears even immediately after bonding (unheated), as described with reference to
[0377] On the other hand, although the bonding strength γ decreased with increasing the electronegativity (the difference in electronegativity with oxygen decreased), the GeO.sub.2 films with an electronegativity of 2.01 (the difference in electronegativity with oxygen is about 1.43), which corresponds to an ionic crystallinity of 40%, could be bonded either unheated or heated to 300° C. after bonding. This indicates that bonding is possible as long as the electronegativity is 2 or less (the difference in electronegativity with oxygen is about 1.4 or more).
[0378]
[0379] For materials with large electronegativity, such as GeO.sub.2 and ITO, heating at 300° C. significantly increased the bonding strength by 9.3 times for GeO.sub.2 and 3.3 times for ITO compared to the unheated condition. Especially for GeO.sub.2, it has been confirmed that the bonding was accompanied by atomic diffusion to the extent that the bonding interface disappeared, as described with reference to
[0380] The reason for the increase in the bonding strength γ of the material with large electronegativity by the heat treatment at 300° C. is thought to be that the melting point tends to be lower for the material with large electronegativity.
[0381] This is because the lower the melting point (i.e., the higher the electronegativity), the greater the effect of promoting atomic diffusion at the bonding interface, and the greater the rate of increase in bonding strength, even when the same heat treatment at 300° C. is applied.
[0382] As an example, the melting points of GeO.sub.2 and ITO, for which the rate of increase in the bonding strength γ by heat treatment was large, were 1115° C. for GeO.sub.2 and about 900° C. for ITO, respectively, which is less than half of the melting point of Y.sub.2O.sub.3 (2425° C.), which has the smallest electronegativity among the examples.
[0383]
[0384]
[0385]
[0386] As indicated in
[0387] Of these, the bonding using Y.sub.2O.sub.3 thin film, which exhibited the maximum ionicity of 70.8%, has excellent bonding performance, as the bonding interface disappears even immediately after bonding (unheated), as explained in
[0388] On the other hand, although the bonding strength γ decreases with decreasing the ionicity, it was confirmed from the bonding example using the GeO.sub.2 thin film (Test Example 8) that bonding is possible up to an ionicity of about 40%.
[0389]
[0390] For materials with small ionicity, such as GeO.sub.2 and ITO, heating at 300° C. resulted in a significant increase in the bonding strength γ compared to the unheated case. This significant increase in the bonding strength γ is thought to be due to the fact that the melting point tends to be lower for materials with smaller ionic properties, and even when the same heat treatment at 300° C. is applied, the lower the melting point (i.e., the smaller the ionic properties), the greater the effect of promoting atomic diffusion at the bonding interface, and the greater the rate of increase in the bonding strength.
[0391] The bonding strength γ of the bonding using Ga.sub.2O.sub.3 thin film was about 14 times higher when unheated and about 50 times higher when heat-treated at 300° C. than that of the bonding using the SiO.sub.2 thin film, and about 6.4 times higher when unheated and 1.6 times higher when heat-treated at 300° C. than that of the bonding using the GeO.sub.2 thin film. These results indicate that a significant increase in the bonding strength γ is obtained when the electronegativity is lower than the electronegativity of Ga, 1.81 (the difference in electronegativity with oxygen is higher than 1.63), or when the ionicity is higher than 48.5% (≈; 50%).