BONDED SUBSTRATE AND A MANUFACTURING METHOD THEREOF, AND A SURFACE ACOUSTIC WAVE DEVICE USING THE SAID BONDED SUBSTRATE

20180294793 ยท 2018-10-11

Assignee

Inventors

Cpc classification

International classification

Abstract

[Object] An object of the present invention is to provide a bonded substrate which is excellent in temperature characteristics and suppresses unnecessary response due to reflection of an elastic wave at a bonding interface.

[Means to Solve the Problems] The present invention is unique in that a bonded substrate is constructed by bonding a LiTaO.sub.3 substrate and a base plate wherein a Li concentration at a base plate-bonding face of the LiTaO.sub.3 substrate is higher than that at a LiTaO.sub.3 substrate-side end face of the bonded substrate, that the difference between the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate and the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate is 0.1 mol % or greater, that the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate satisfies an equation Li/(Li+Ta)100=(50+) mol %, where is in the range of 1.2<<0.5, that the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate satisfies an equation Li/(Li+Ta)100=(48.5+) mol %, where is in the range of 0.5<<0.5, and that the thickness measured from the base plate-bonding face of the LiTaO.sub.3 substrate to the LiTaO.sub.3 substrate-side end face of the finished bonded substrate becomes greater than 5 times but less than 20 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave.

Claims

1. A bonded substrate comprising: a LiTaO.sub.3 substrate; and a base plate bonded to said LiTaO.sub.3 substrate, wherein a Li concentration at a base plate-bonding face of said LiTaO.sub.3 substrate is higher than a Li concentration at a LiTaO.sub.3 substrate-side end surface of said bonded substrate.

2. A bonded substrate as claimed in claim 1, wherein a difference between the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate and the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate is 0.1 mol % or greater.

3. A bonded substrate as claimed in claim 1, wherein the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate satisfies an equation Li/(Li+Ta)100=(50+) mol %, where is in the range of 1.2<<0.5.

4. A bonded substrate as claimed in claim 1, wherein the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate satisfies an equation Li/(Li+Ta)100=(48.5+) mol %, where is in the range of 0.5<<0.5.

5. A bonded substrate as claimed in claim 1, wherein the thickness of the LiTaO.sub.3 substrate in the bonded substrate is greater than 5 times but less than 20 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave.

6. A bonded substrate as claimed in claim 1, wherein an area where Li concentration satisfies said equation for the base plate-bonding face of the LiTaO.sub.3 substrate expands from said bonding face toward the LiTaO.sub.3 substrate-side end face of the bonded substrate by a thickness of 0.1 through 4 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave.

7. A bonded substrate as claimed in claim 1, wherein said bonded substrate possesses an area where Li concentration decreases as it is measured starting from the bonding face of the LiTaO.sub.3 substrate toward the LiTaO.sub.3 substrate-side end face of the bonded substrate.

8. A bonded substrate as claimed in claim 7, wherein said area where Li concentration thus decreases is formed through a thickness of 1 through 5 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave.

9. A bonded substrate as claimed in claim 1, wherein an area where Li concentration satisfies said equation for the LiTaO.sub.3 substrate-side end face of the bonded substrate expands from said LiTaO.sub.3 substrate-side end face of the bonded substrate toward the base plate-bonding face of the LiTaO.sub.3 substrate by a thickness of 1 through 20 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave.

10. A bonded substrate as claimed in claim 1, wherein a crystal orientation of said LiTaO.sub.3 substrate is 36 rotated Y-cut through 49 rotated Y-cut.

11. A bonded substrate as claimed in claim 1, wherein said base plate is made of a material selected from Si, SiC, Spinel, and sapphire.

12. A bonded substrate comprising: a LiNbO.sub.3 substrates and a base plate bonded to said LiNbO.sub.3 substrate, wherein a Li concentration at a base plate-bonding face of said LiNbO.sub.3 substrate is higher than a Li concentration at a LiNbO.sub.3 substrate-side end surface of said bonded substrate.

13. A surface acoustic wave device comprising the bonded substrate as defined in claim 1.

14. A method for manufacturing a bonded substrate comprising: bonding a base plate to a LiTaO.sub.3 substrate in which the Li concentration is greater at the substrate surface than inside the substrate; and removing a surface layer from the LiTaO.sub.3 substrate-side end face of the bonded substrate which lies opposite the bonding face of the LiTaO.sub.3 substrate in a manner such that the Li concentration at the bonding face of the LiTaO.sub.3 substrate becomes greater than that at the LiTaO.sub.3 substrate-side end face of the bonded substrate.

15. A surface acoustic wave device comprising the bonded substrate as defined in claim 12.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0023] FIG. 1 A diagram showing a model of Li content profiles of LiTaO.sub.3 substrates respectively constituting bonded substrates of Example 1 and Example 2.

[0024] FIG. 2 A diagram showing input impedance waveforms of resonators formed respectively on a bonded substrate of Example 1 and a bonded substrate of Comparative Example 2.

[0025] FIG. 3 A diagram showing a result of measuring the values Q of resonators formed respectively on a bonded substrate of Example 1 and a bonded substrate of Comparative Example 2.

[0026] FIG. 4 A diagram showing input impedance waveforms of resonators formed respectively on a bonded substrate of Example 2 and a bonded substrate of Comparative Example 1.

[0027] FIG. 5 A diagram showing a result of measuring the values Q of resonators formed respectively on a bonded substrate of Example 2 and a normal LiTaO.sub.3 substrate of Comparative Example 1.

EMBODIMENTS FOR CARRYING OUT THE INVENTION

[0028] Hereinafter, embodiments of the present invention will be described in detail, but the present invention is not limited thereto.

[0029] The bonded substrate of the present invention is formed by bonding a LiTaO.sub.3 substrate on a base plate. A known method may be used for bonding the LiTaO.sub.3 substrate on the base plate, but considering the base plate's influence to bend the substrate, it is preferable to use the room temperature bonding method. Further, intervening layers made of materials such as oSiO.sub.2, SiO.sub.20.5, a-Si, p-Si, a-SiC, Al.sub.2O.sub.3 and the like may be present between the LiTaO.sub.3 substrate and the base plate.

[0030] The bonded substrate of the present invention is characteristic in that the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate is higher than the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate. Therefore, there is a difference in Li concentration in the thickness direction of the LiTaO.sub.3 substrate bonded to the base substrate.

[0031] In the case where an intervening layer is present between the LiTaO.sub.3 substrate and the base plate, the base plate-bonding face of the LiTaO.sub.3 substrate shall be deemed to mean the face of the LiTaO.sub.3 substrate in contact with the intervening layer.

[0032] Such a LiTaO.sub.3 substrate can be obtained by diffusing Li from the surface of the LiTaO.sub.3 substrate by a known method such as vapor phase diffusion method, for example. Since the Li concentration of the thus obtained substrate at the surface thereof would become large and the Li concentration inside the substrate would be smaller, if the LiTaO.sub.3 substrate is machined by grinding or polishing or the like to expose the inside of the LiTaO.sub.3 substrate and the non-machined surface is made to be the base plate-bonding face, a bonded substrate of the present invention is obtained.

[0033] Further, the LiTaO.sub.3 substrate constituting the bonded substrate of the present invention may itself be a bonded substrate composed of a plurality of LiTaO.sub.3 substrates. It is also possible to manufacture a bonded substrate of the present invention by preparing a LiTaO.sub.3 substrate having a small Li concentration and a LiTaO.sub.3 substrate having a high Li concentration and bonding these to a base plate. For example, on one hand, as a LiTaO.sub.3 substrate with a small Li concentration, a LiTaO.sub.3 substrate of a congruent melt composition manufactured by a usual Czochralski method may be used, and on the other, as a LiTaO.sub.3 substrate having a large Li concentration, a LiTaO.sub.3 substrate of a stoichiometric composition manufactured by a double crucible method may be used.

[0034] Further, the LiTaO.sub.3 substrate constituting the bonded substrate of the present invention may be doped with a metal element such as Fe if necessary, or it may be subjected to a reduction treatment for suppressing the pyroelectricity.

[0035] In the present invention, the difference between the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate and the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate is preferably 0.1 mol % or greater, or more preferably 0.25 mol % or greater, or still more preferably 0.5 mol % or greater. The upper limit of this difference is the difference between the stoichiometric composition (500.5 mol %) and the congruent melt composition (48.50.5 mol %), preferably 2.5 mol % or less, more preferably 2.0 mol % or less, still more preferably 1.5 mol % or less.

[0036] In the present invention, the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate is Li/(Li+Ta)100=(50+) mol %, where is in the range of 1.2<<0.5, preferably in the range of 1.0<<0.5, more preferably in the range of 0.8<<0.5, or still more preferably in the range of 0.8<0.

[0037] From the viewpoints of easiness in manufacturing and manufacturing cost, the LiTaO.sub.3 substrate constituting the bonded substrate of the present invention is preferably manufactured by first preparing a LiTaO.sub.3 substrate having a congruent melt composition through an ordinary Czochralski method and then applying it to a Li diffusion treatment by means of a gas phase method. When the LiTaO.sub.3 substrate is produced in this way, the Li concentration at the substrate surface becomes Li/(Li+Ta)100=(50+) mol % where is in the range of 1.2<<0.5, which is close to the stoichiometric composition.

[0038] The longer the time of applying the Li diffusion treatment is, the more likely warpage and cracking would occur in the substrate, so that it is preferable that the area wherein the Li concentration satisfies Li/(Li+Ta)100=(50+) mol % where is in the range of 1.2<<0.5 extends from the bonding face of the LiTaO.sub.3 substrate toward the LiTaO.sub.3 substrate-side end face of the bonded substrate by a limited distance of about 0.1 through 4 times the wavelength of the surface acoustic wave or the leaky surface acoustic wave. By doing so, it is possible to suppress warping and cracking of the substrate and to increase the value Q of the bonded substrate.

[0039] Further, in the present invention, it is preferable to have an area in which the Li concentration decreases as the measurement moves from the bonding face of the LiTaO.sub.3 substrate toward the LiTaO.sub.3 substrate-side end face of the bonded substrate. Such a bonded substrate can be manufactured by using, as the LiTaO.sub.3 substrate of a bonded substrate, a LiTaO.sub.3 substrate having been subjected to a Li diffusion treatment by a gas phase method. In this case, if a bonded substrate in which the Li concentration profile has a steep variation is used, although it is possible to increase the Value Q of the resonator, there is a possibility that a response noise is generated due to a steep Li concentration variation. For this reason, it is preferable that the area in which the Li concentration decreases from the bonding face of the LiTaO.sub.3 substrate toward the LiTaO.sub.3 substrate-side end face of the bonded substrate spans over 1 to 5 times the wavelength of the surface acoustic wave or the leaky surface acoustic wave. In this way, it is possible not only to manufacture the device easily but also to suppress the response noise caused by the steep Li concentration change.

[0040] In the bonded substrate of the present invention, it is preferable that, on one hand, the Li concentration at the bonding face of the LiTaO.sub.3 substrate satisfies Li/(Li+Ta)100=(50+) mol % where is in the range of 1.2<<0.5, and, on the other hand, the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate satisfies Li/(Li+Ta)100=(48.5+) mol % where is in the range of 0.5<<0.5, and also that the LiTaO.sub.3 substrate is of a congruent melt composition.

[0041] As described above, from the viewpoints of easiness in manufacturing and manufacturing cost, the bonded substrate of the present invention is preferably manufactured based on a LiTaO.sub.3 substrate having a congruent melt composition obtained through an ordinary Czochralski method and having been subjected to a Li diffusion treatment by means of a gas phase method. This way, the Li concentration inside the LiTaO.sub.3 substrate becomes Li/(Li+Ta)100=(48.5+) mol % (wherein 0.5<<0.5), which is close to the congruent melt composition; and after bonding this LiTaO.sub.3 substrate to a base plate, the LiTaO.sub.3 substrate is machined by grinding or polishing or the like to expose the inside of the LiTaO.sub.3 substrate, whereby it is possible to obtain a bonded substrate wherein the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate is Li/(Li+Ta)100=(48.5+) mol % where is in the range of 0.5<<0.5.

[0042] In addition, the area in which the Li concentration satisfies Li/(Li+Ta)100=(48.5+) mol % where is in the range of 0.5<<0.5 can be arbitrarily determined. However, if this area is narrowed relative to the thickness of the LiTaO.sub.3 substrate constituting the bonded substrate, it becomes necessary to lengthen the Li diffusion treatment time, with a consequence that warping and cracking are more likely to occur in the substrate; so this area of the said Li concentration is preferably substantially thick. It is preferable that said area is formed to extend from the LiTaO.sub.3 substrate-side end face of the bonded substrate toward the base plate-bonding face of the LiTaO.sub.3 substrate over about 1 to 20 times the wavelength of the surface acoustic wave or the leaky surface acoustic wave.

[0043] The thickness of the LiTaO.sub.3 substrate in the bonded substrate of the present invention is preferably greater than 5 times but less than 20 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave. In this way, it is possible to suppress an unnecessary response caused by the reflection of the acoustic wave at the bonding interface.

[0044] In the present invention, the wavelength of the surface acoustic wave or the leaky surface acoustic wave are, respectively, the wavelength of the surface acoustic wave or the leaky surface acoustic wave when the bonded substrate is used as the surface acoustic wave device, and they are determined by the frequency of the electric signal input to the bonded substrate (surface acoustic wave device) and the velocity of the surface wave (leaky wave). The velocity of the surface wave varies depending on the material and is about 4000 m/s for LiTaO.sub.3. Therefore, in the case of manufacturing a surface acoustic wave device for 1 GHz from a composite substrate using LiTaO.sub.3 as the piezoelectric single crystal substrate, the wavelength of the surface acoustic wave becomes about 4 m. In the case of manufacturing a surface acoustic wave device for 2 GHz, the wavelength of the surface acoustic wave is about 2 m, and in the case of manufacturing a surface acoustic wave device for 800 MHz, the wavelength of the surface acoustic wave is about 5 m.

[0045] Although the crystal orientation of the LiTaO.sub.3 substrate constituting the present invention can be arbitrarily selected, it is preferable, from the viewpoint of characteristics, that it is 36 rotated Y-cut through 49 rotated Y-cut. Also, the material to make the plate used as the base plate is not particularly limited, but it is preferably selected from Si, SiC, spinel and sapphire.

[0046] Even if a LiNbO.sub.3 substrate is used in place of the LiTaO.sub.3 substrate constituting the present invention, the value Q can be improved similarly as in the case of the bonded substrate using the LiTaO.sub.3 substrate.

[0047] Accordingly, the present invention is also characteristic in that, in a bonded substrate composed of a LiNbO.sub.3 substrate and a base plate bonded together, the Li concentration at a base plate-bonding face of the LiNbO.sub.3 substrate is greater than the Li concentration at a LiNbO.sub.3 substrate-side end face of the bonded substrate.

[0048] In addition, the difference between the Li concentration at the base plate-bonding face of the LiNbO.sub.3 substrate and the Li concentration at the LiNbO.sub.3 substrate-side end face of the bonded substrate is preferably 0.1 mol % or greater.

[0049] It is preferable that the Li concentration at the base plate-bonding face of the LiNbO.sub.3 substrate satisfies Li/(Li+Nb)100=(50+) mol % where is in the range of 1.2<<0.5, and that the Li concentration at the LiNbO.sub.3 substrate-side end face of the bonded substrate satisfies Li/(Li+Nb)100=(48.5+) mol % where is in the range of 0.5<<0.5.

[0050] Furthermore, the thickness of the LiNbO.sub.3 substrate in the bonded substrate is preferably greater than 5 times but less than 20 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave.

[0051] It is also preferable that an area in which the Li concentration satisfies the above formula meant for the base plate-bonding face of the LiNbO.sub.3 substrate of the present invention extends from this bonding face of the LiNbO.sub.3 substrate toward the LiNbO.sub.3 substrate-side end face of the bonded substrate by a distance of 0.1 through 4 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave.

[0052] It is furthermore preferable that the bonded substrate of the present invention has an area extending from the bonding face of the LiNbO.sub.3 substrate toward the LiNbO.sub.3 substrate-side end face of the bonded substrate in which (area) the Li concentration decreases as the depth from the bonding face of the LiNbO.sub.3 substrate increases; and it is further preferable that this area, in which the Li concentration thus decreases, is formed to have a thickness of 1 through 5 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave. In addition to this, the range of the Li concentration in the LiNbO.sub.3 substrate-side end face of the bonded substrate of the present invention is preferably formed over a span of 1 through 20 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave from the LiNbO.sub.3 substrate-side end face of the bonded substrate toward the bonding face of the LiNbO.sub.3 substrate.

[0053] It is preferable that the crystal orientation of the LiNbO.sub.3 substrate of the present invention is rotated from 0 Y through 30 Y cut or 1285 Y cut, and also that the base plate used in the present invention is made of a material selected from Si, SiC, Spinel, and sapphire.

[0054] It is desirable to fabricate a surface acoustic wave device using one of these bonded substrates of the present invention, because a resonator having a large value of Q can be constructed.

[0055] The bonded substrate according to the present invention can be manufactured in the following manner: a base plate and a LiTaO.sub.3 substrate in which the Li concentration is greater at the surface of the substrate than inside the substrate are bonded together; a surface layer at the LiTaO.sub.3 substrate-side end face of the bonded substrate which lies opposite the bonding face of the LiTaO.sub.3 substrate is removed in a manner such that the Li concentration at the bonding face of the LiTaO.sub.3 substrate becomes greater than that at the LiTaO.sub.3 substrate-side end face of the bonded substrate. On this occasion, it is also possible to bond a base plate on each one of the faces of a LiTaO.sub.3 substrate in which the Li concentration is greater at the surface than inside the substrate, and then divide the LiTaO.sub.3 substrate in half across thickness-wise middle plane, to thereby obtain bonded substrates. In this way, two bonded substrates can be obtained from one LiTaO.sub.3 substrate, which is also desirable from the viewpoint of cost.

[0056] Such a manufacturing method can similarly be applied to the case of a LiNbO.sub.3 substrate.

[0057] The Li concentration of the LiTaO.sub.3 substrate or the LiNbO.sub.3 substrate constituting the bonded substrate of the present invention may be measured by a known method, for example, by Raman spectroscopy. In the case of LiTaO.sub.3 single crystal, it is known that there exists a roughly linear relationship between the half width of the Raman shift peak and the Li concentration (Li/(Li+Ta) value) (See 2012 IEEE International Ultrasonics Symposium Proceedings, Page(s): 1252-1255, Applied Physics A 56, 311-315 (1993)).

[0058] Therefore, by using a formula expressing such a relationship, it is possible to evaluate the composition at an arbitrary position in an oxide single crystal substrate.

[0059] A formula expressing a relationship between the half width of the Raman shift peak and the Li concentration can be obtained by measuring the Raman half width of a number of samples whose Li concentrations are different and whose compositions are known; if the conditions of the Raman measurement are identical, it is also possible to use a formula that has been already publicized such as in a literature.

[0060] For example, in the case of lithium tantalate single crystal, the following Formula (1) may be used (See IEEE International Ultrasonics Symposium Proceedings, Page(s): 1252-1255).


Li/(Li+Ta)=(53.150.5 FWHM1)/100 (1)

wherein, FWHM1 is the full width at half maximum of the Raman shift peak around 600 cm.sup.1.

[0061] Please refer to the literature for details of measurement conditions.

[0062] The value Q of the SAW resonator formed on the bonded substrate of the present invention and that of the SAW resonator of a comparative example were obtained through the following Formula (2) described in p. 861 of the publicized literature 2010 IEEE International Ultrasonics Symposium Proceedings Page(s): 861-863. This equation appears as formula (1) in the said literature, but since it overlaps with the above-described formula number, it is given a formula number (2) in this specification.


Q(f)=*(f)*||/(1||.sup.2)

wherein, is the angular frequency, (f) is the group delay time, and is the reflection coefficient measured by a network analyzer.

EXAMPLES

[0063] Hereinafter, examples of the present invention and comparative examples will be described more specifically.

Example 1

[0064] In Example 1, first, a singly polarized 4-inch diameter LiTaO.sub.3 single crystal ingot having a roughly congruent composition and having a Li:Ta ratio of 48.4:51.6 was sliced to obtain a number of 370-m-thick 42 rotated Y-cut LiTaO.sub.3 substrates. Thereafter, in view of a circumstance, the surface roughness of each sliced wafer was adjusted to 0.15 m in arithmetic average roughness Ra value by a lapping procedure, and the after-finish thickness was set to 350 m.

[0065] After the front and back faces of each resulting substrate were finished into a quasi-mirror surface having a Ra value of 0.01 m by planar polishing, the substrate was buried in a powder containing Li, Ta and O with Li.sub.3TaO.sub.4 as a main component. On this occasion, the powder in which Li.sub.3TaO.sub.4 was a main component was prepared by mixing Li.sub.2CO.sub.3 powder and Ta.sub.2O.sub.5 powder at a molar ratio of 7:3, followed by baking at 1300 C. for 12 hours. Then, this powder containing Li.sub.3TaO.sub.4 as a main component was laid in a small container, and a plurality of said sliced wafers were buried in this Li.sub.3TaO.sub.4 powder.

[0066] Then, this small container was set in an electric furnace, which was subsequently filled with an N.sub.2 atmosphere and heated at 900 C. for 20 hours, to thereby cause Li to diffuse into the sliced wafer from the surface toward the middle thereof. Thereafter, during the temperature lowering stage of this diffusion treatment, the thus treated slice substrate was subjected to an anneal treatment at 800 C. for 12 hours, and during the subsequent stage where the wafer was allowed to cool further from 770 C. to 500, an electric field of 4000 V/m was applied in roughly +Z axis direction, and thereafter a treatment was conducted to cause the temperature to fall to the room temperature. Incidentally, after the application of the electric field, the furnace atmosphere may be changed to be the atmosphere.

[0067] Also after this treatment, a rougher side face of each wafer was subjected to a sandblasting to finish it to an Ra value of about 0.15 m, and the other quasi-mirror side face thereof was polished 3 m deep and thus a plurality of LiTaO.sub.3 single crystal substrates were obtained.

[0068] Next, the thus obtained LiTaO.sub.3 substrate and a 230 m-thick Si substrate were bonded together by a room temperature bonding method described in Takagi H. et al, Room-temperature wafer bonding using argonbeam activation From ProceedingsElectrochemical Society (2001), 99-35 (Semiconductor Wafer Bonding: Science, Technology, and Applications V), 265-274. whereby a number of bonded substrates were obtained.

[0069] Specifically, a cleaned substrate was set in a high vacuum chamber; a high-speed atomic beam of argon whose ion beam has been neutralized is irradiated on the surface of the substrate to activate it (activation treatment); thereafter, the LiTaO.sub.3 single crystal substrate was bonded to the Si base plate. The LiTaO.sub.3 substrate-side end face of the thus obtained bonded substrate was subjected to grinding and polishing in a manner such that what remained of the LiTaO.sub.3 substrate on the bonding interface of this bonded substrate came to have a thickness of 28 m, and thus a bonded substrate consisting of a rotated Y-cut LiTaO.sub.3 substrate diffused with Li and the Si base plate was prepared.

[0070] Next, with respect of one of the thus prepared bonded substrates, or more particularly with respect to the middle portion of the bonded substrate, a laser Raman spectrometer (LabRam HR series manufactured by HORIBA Scientific Inc., Ar ion laser, spot size 1 m, room temperature) was used to measure the half-value width FWHM1 of the Raman shift peak around 600 cm.sup.1, which is an indicator of the Li diffusion amount, the measurement proceeding in a depth-wise direction from the LiTaO.sub.3-side surface, and the Li concentration was calculated using the above-described Formula (1); and as the result a Li concentration profile as shown in FIG. 1 was obtained.

[0071] According to the result shown in FIG. 1, regarding this bonded substrate, at the base plate-bonding face of the LiTaO.sub.3 substrate the Li concentration was 49.6 mol % and the value was 0.4 and at the LiTaO.sub.3 substrate-side end face of the bonded substrate the Li concentration was 48.4 mol % and the value was 0.1, so that it was confirmed that the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate was higher than the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate. Further, the difference between the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate and the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate was 1.2 mol %.

[0072] In addition, with respect to the area ranging between 0 m and about 0.5 m in depth as measured from the bonding interface toward the LiTaO.sub.3 substrate-side end face of the bonded substrate, the results indicated Li/(Li+Ta)=49.6 mol % and a presence of pseudo stoichiometry composition. It was confirmed that the bonded substrate had in the vicinity of its surface layer a transition layer having a thickness of about 20 m, in which the Li concentration decreased as the measurement moved toward the LiTaO.sub.3 substrate-side end face of the bonded substrate; it was also confirmed that there was an area ranging from the LiTaO.sub.3 substrate-side end face of the bonded substrate to a depth of about 8 m wherein the Li concentration indicated that Li/(Li+Ta)=48.4 mol % and an existence of a roughly congruent composition.

[0073] Further, when the warpage of this bonded substrate was measured by the interference method using laser light, the value obtained was as large as 200 m, while cracks and chippings were not observed.

[0074] Next, on the LiTaO.sub.3 substrate-side end face of the bonded substrate as obtained in this manner, an Al film having a thickness of 0.2 m was formed by means of a sputtering method, and at the same time, by means of a photolithography using a g-line as the light source, a SAW resonator was formed. On this occasion, the Al electrode was shaped by RIE (reactive ion etching), and a mixed gas of BCl.sub.3, Cl.sub.2, CF.sub.4, and N.sub.2 was used as the gas for this RIE.

[0075] At this time, the thickness of the LiTaO.sub.3 substrate in the bonded substrate was 28 m, which is seven times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave, for the latter were roughly 4 m.

[0076] Then, the thus fabricated evaluation purpose SAW resonator was measured by an RF prober for its property and the measurement result was as shown in FIG. 2. According to the result shown in FIG. 2, in the case of Example 1, it was confirmed that a resonance waveform which is on the whole good was obtained.

[0077] Next, the values of S11 and the group delay time as measured by the above RF prober were input to a recording medium from a network analyzer, and the value Q was obtained by means of the above equation (2), and the result is shown in FIG. 3. According to the result shown in FIG. 3, the maximum value of Q was 1280.

Example 2

[0078] In Example 2, first, a singly polarized 4-inch diameter LiTaO.sub.3 single crystal ingot having a roughly congruent composition and having a Li:Ta ratio of 48.75:51.25 was sliced to obtain a number of 370-m-thick 42 rotated Y-cut LiTaO.sub.3 substrates. Thereafter, in view of a circumstance, the surface roughness of each sliced wafer was adjusted to 0.15 m in arithmetic average roughness Ra value by a lapping procedure, and the after-finish thickness was set to 350 m.

[0079] After the front and back faces of each resulting substrate were finished into a quasi-mirror surface having a Ra value of 0.01 m by planar polishing, the substrate was buried in a powder containing Li, Ta and O with Li.sub.3TaO.sub.4 as a main component. On this occasion, the powder in which Li.sub.3TaO.sub.4 was a main component was prepared by mixing Li.sub.2CO.sub.3 powder and Ta.sub.2O.sub.5 powder at a molar ratio of 7:3, followed by baking at 1300 C. for 12 hours. Then, this powder containing Li.sub.3TaO.sub.4 as a main component was laid in a small container, and a plurality of said sliced wafers were buried in this Li.sub.3TaO.sub.4 powder.

[0080] Then, this small container was set in an electric furnace, which was subsequently filled with an N.sub.2 atmosphere and heated at 900 C. for 10 hours, to thereby cause Li to diffuse into the sliced wafer from the surface toward the middle thereof. Thereafter, during the temperature lowering stage of this diffusion treatment, the thus treated slice substrate was subjected to an anneal treatment at 800 C. for 12 hours, and during the subsequent stage where the wafer was allowed to cool further from 770 C. to 500, an electric field of 4000 V/m was applied in roughly +Z axis direction, and thereafter a treatment was conducted to cause the temperature to fall to the room temperature. Incidentally, after the application of the electric field, the furnace atmosphere may be changed to be the atmosphere.

[0081] After this treatment, a bonded substrate was prepared in the same manner as in Example 1, and the LiTaO.sub.3 substrate-side end face of the thus obtained bonded substrate was subjected to grinding and polishing in a manner such that what remained of the LiTaO.sub.3 substrate on the bonding interface of this bonded substrate came to have a thickness of 40 m, and thus a bonded substrate consisting of a rotated Y-cut LiTaO.sub.3 substrate diffused with Li and the Si base plate was prepared.

[0082] Next, with respect of one of the thus prepared bonded substrates, or more particularly with respect to the middle portion of this bonded substrate, a laser Raman spectrometer (LabRam HR series manufactured by HORIBA Scientific Inc., Ar ion laser, spot size 1 m, room temperature) was used to measure the half-value width FWHM1 of the Raman shift peak around 600 cm.sup.1, which is an indicator of the Li diffusion amount, the measurement proceeding in a depth-wise direction from the LiTaO.sub.3-side surface, and the Li concentration was calculated using the above-described Formula (1); and as the result a Li concentration profile as shown in FIG. 1 was obtained.

[0083] According to the result shown in FIG. 1, regarding this bonded substrate, at the base plate-bonding face of the LiTaO.sub.3 substrate the Li concentration was 49.4 mol % and the value was 0.6 and at the LiTaO.sub.3 substrate-side end face of the bonded substrate the Li concentration was 48.75 mol % and the value was 0.25, so that it was confirmed that the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate was higher than the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate. Further, the difference between the Li concentration at the base plate-bonding face of the LiTaO.sub.3 substrate and the Li concentration at the LiTaO.sub.3 substrate-side end face of the bonded substrate was 0.65 mol %.

[0084] In addition, with respect to the area ranging between 0 m and about 2 m in depth as measured from the bonding interface toward the LiTaO.sub.3 substrate-side end face of the bonded substrate, the results indicated Li/(Li+Ta)=49.4 mol % and a presence of pseudo stoichiometry composition. It was confirmed that the bonded substrate had in the vicinity of its surface layer a transition layer having a thickness of about 8 m, in which the Li concentration decreased as the measurement moved toward the LiTaO.sub.3 substrate-side end face of the bonded substrate; it was also confirmed that there was an area ranging from the LiTaO.sub.3 substrate-side end face of the bonded substrate to a depth of about 30 m wherein the Li concentration indicated that Li/(Li+Ta)=48.75 mol % and an existence of a roughly congruent composition.

[0085] Further, when the warpage of this bonded substrate was measured by the interference method using laser light, the value obtained was as small as 40 m, while cracks and chippings were not observed.

[0086] Next, using this bonded substrate, a SAW resonator whose one-wave length was about 4 m was fabricated, and the thus fabricated evaluation purpose SAW resonator was measured by an RF prober for its property and the measurement result was as shown in FIG. 4. According to the result shown in FIG. 4, in the case of Example 2 too, it was confirmed that a resonance waveform which is on the whole good was obtained.

[0087] At this time, the thickness of the LiTaO.sub.3 substrate in the bonded substrate was 40 m, which is 10 times the wavelength of the surface acoustic wave or that of the leaky surface acoustic wave, for the latter were roughly 4 m.

[0088] Next, the values of S11 and the group delay time as measured by the above RF prober were input to the recording medium from the network analyzer, and the Value Q was obtained by means of the above equation (2), and the result is shown in FIG. 5. According to the result shown in FIG. 5, the maximum value of Q was 1380.

Comparative Example 1

[0089] In Comparative Example 1, a singly polarized 4-inch diameter LiTaO.sub.3 single crystal ingot having a roughly congruent composition and having a Li:Ta ratio of 48.4:51.6 was processed and lapped similarly as in Example 1, and a wafer having an after-finish thickness of 350 m was made.

[0090] Next, after this lapping, the both faces of the wafer were finished by planar polishing into mirror surfaces having a Ra value of 0.0001 m, and thus a plurality of LiTaO.sub.3 single crystal substrates having a roughly congruent composition were fabricated.

[0091] Then, similarly as in Example 1, with respect of one of the thus prepared LiTaO.sub.3 single crystal substrate, or more particularly with respect to the middle portion of this substrate, a laser Raman spectrometer was used to measure the half-value width FWHM1 of the Raman shift peak around 600 cm.sup.1, which is an indicator of the Li diffusion amount, the measurement proceeding in a depth-wise direction from the surface, and the Li concentration was calculated using the above-described Formula (1); and as the result it was found that the Li concentration of this LiTaO.sub.3 substrate remained roughly constant at 48.4 mol % in the depth-wise direction and showed a roughly congruent composition.

[0092] Next, the warping of this substrate was measured by the interference method using a laser beam, and the value obtained was as small as 40 m, and cracks and chipping were not observed.

[0093] Next, on the surface of the substrate as obtained in this manner, an Al film having a thickness of 0.2 m was formed by means of a sputtering method, and at the same time, by means of a photolithography using a g-line as the light source, a SAW resonator whose one-wave length was about 4 m formed. On this occasion, the Al electrode was shaped by RIE (reactive ion etching), and a mixed gas of BCl.sub.3, Cl.sub.2, CF.sub.4, and N.sub.2 was used as the gas for this RIE.

[0094] Then, the thus fabricated evaluation purpose SAW resonator was measured by an RF prober for its property and the measurement result was as shown in FIG. 4. According to the result shown in FIG. 4, in the case of Comparative Example 1, it was confirmed that a resonance waveform which is on the whole good was obtained.

[0095] Next, the values of S11 and the group delay time as measured by the above RF prober were input to the recording medium from a network analyzer, and the value Q was obtained by means of the above equation (2), and the result is shown in FIG. 5. According to the result shown in FIG. 5, the maximum value of Q was 900, and this was lower than the values Q in the cases of Example 1 and Example 2.

Comparative Example 2

[0096] In Comparative Example 2, a singly polarized 4-inch diameter LiTaO.sub.3 single crystal ingot having a roughly congruent composition and having a Li:Ta ratio of 48.4:51.6 was processed similarly as in Example 1, and a wafer having an after-finish thickness of 350 m was made. Next, similarly as in Comparative Example 1, the both faces of this wafer were finished by planar polishing into mirror surfaces having a Ra value of 0.0001 m, and thus a plurality of LiTaO.sub.3 single crystal substrates having a roughly congruent composition were fabricated. Thereafter, similarly as in Example 1, this LiTaO.sub.3 substrate was bonded to a 230-m-thick Si base plate; then the LiTaO.sub.3 substrate-side end face of the thus obtained bonded substrate was subjected to grinding and polishing in a manner such that what remained of the LiTaO.sub.3 substrate on the bonding interface of this bonded substrate came to have a thickness of 28 m, and thus a bonded substrate consisting of a rotated Y-cut LiTaO.sub.3 substrate and the Si base plate was prepared.

[0097] Then, similarly as in Example 1, with respect of one of the thus prepared bonded substrate, or more particularly with respect to the middle portion of this bonded substrate, a laser Raman spectrometer was used to measure the half-value width FWHM1 of the Raman shift peak around 600 cm.sup.1, which is an indicator of the Li diffusion amount, the measurement proceeding in a depth-wise direction from the LiTaO.sub.3 substrate-side end face, and the Li concentration was calculated using the above-described Formula (1); and as the result it was found that the Li concentration of this LiTaO.sub.3 substrate constituting the bonded substrate remained roughly constant at 48.4 mol % in the depth-wise direction and showed a roughly congruent composition so that the Li concentration at the interface between the LiTaO.sub.3 substrate and the base plate was roughly equal to the Li concentration at the LiTaO.sub.3 substrate-side end face.

[0098] Next, the warping of this bonded substrate was measured by an interference method using a laser beam, and the value obtained was as small as 40 m, and cracks and chipping were not observed.

[0099] Next, using this bonded substrate, a SAW resonator whose one-wave length was about 4 m was fabricated in the similar manner as in Example 1, and the thus fabricated evaluation purpose SAW resonator was measured by an RF prober for its property and the measurement result was as shown in FIG. 2. According to the result shown in FIG. 2, in the case of Comparative Example 2, it was confirmed that a resonance waveform which is on the whole good was obtained.

[0100] Next, the values of S11 and the group delay time as measured by the above RF prober were input to the recording medium from a network analyzer, and the value Q was obtained by means of the above equation (2), and the result is shown in FIG. 3. According to the result shown in FIG. 3, the maximum value of Q was 1020, and this was lower than the values Q in the cases of Example 1 and Example 2.

[0101] As is confirmed from the comparison made above between Examples 1 and 2, on one hand, and Comparatives Examples 1 and 2, on the other hand, if a bonding substrate of the present invention is used it is possible to obtain the effects arising from value Q which is greater than is obtained in the case of a conventional LiTaO.sub.3 substrate, and from the fact that there is scarce variation in Li concentration. Also, it was confirmed that when the bonding substrate of the present invention has an a which satisfies 1.2<0.5, especially, and a which satisfies 0.5<<0.5, it is possible to obtain a bonded substrate wherein the value Q is further increased and the warping is reduced, and thus it is possible to obtain a high value of Q favorable for surface acoustic wave device.