JP2015227277A - Method of manufacturing piezoelectric oxide single crystal substrate - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method of manufacturing a piezoelectric oxide single crystal substrate which is good in productivity, low-cost, and free of problems of cracking and warpage, and has excellent temperature characteristics and no pyroelectricity.SOLUTION: There is provided a method of manufacturing a piezoelectric oxide single crystal substrate that includes: a process of cutting an oxide single crystal substrate out by slicing an oxide single crystal ingot comprising a practically congruent composition having been subjected to single polarization processing; a process of burying the substrate in powder including an Li compound; a vapor-phase processing process of performing reforming so that the substrate has a profile having a higher Li toward a surface and decreasing in Li concentration inward by diffusing Li into the substrate from the surface by heating at 850-1,000°C and for 10-50 hours in an atmosphere including no oxygen; and a process of polishing the substrate. The Li concentration profile from the surface to a center part of the substrate can be confirmed from a distribution of half-value widths of Raman shift peaks of almost 600 cm-1 as indexes for Li diffusion amounts measured in a depth direction from the surface.

Description

  The present invention relates to a method for manufacturing a piezoelectric oxide single crystal substrate used for a surface acoustic wave element or the like.

  Recent mobile phone communication systems support a plurality of communication standards, and each communication standard has progressed to a form composed of a plurality of frequency bands. As a part for frequency adjustment / selection of such a mobile phone, for example, a surface acoustic wave (“SAW”) device in which a comb electrode for exciting a surface acoustic wave is formed on a piezoelectric substrate is used. ing.

This surface acoustic wave device is required to have a small size, a small insertion loss, and a performance that prevents unnecessary waves from passing through. Therefore, lithium tantalate: LiTaO ₃ (hereinafter referred to as “LT”) may be used as the material. ) And lithium niobate: LiNbO ₃ (hereinafter, sometimes referred to as “LN”) are used. In particular, many of the 4th generation mobile phone communication standards have narrower transmission / reception frequency band intervals or wider band widths, but on the other hand, the characteristics of surface acoustic wave device materials change with temperature, resulting in frequency changes. Since the selection range is shifted, there is a problem that the filter and the duplexer function are hindered. Therefore, a material for a surface acoustic wave device having a small characteristic variation with respect to temperature and a wide band is desired.

  In the surface acoustic wave device manufacturing process, there are a plurality of steps of applying a temperature of 100 to 300 ° C. to the material. If the surface acoustic wave device material has pyroelectricity, the material is charged to exceed 1 KV. Occurs and discharge occurs. This discharge is not preferable because it decreases the manufacturing yield of the surface acoustic wave device. Further, even if the charge of the surface acoustic wave device material is weak pyroelectricity that attenuates with time, it is not preferable because noise is generated in the electrode of the surface acoustic wave device due to temperature change.

  On the other hand, Patent Document 1 discloses a breakdown mode in which a stoichiometric composition LT obtained mainly by a vapor phase method using copper as an electrode material is destroyed at the moment when high power is input to an IDT electrode. It is described as preferable because it is less likely to occur. Patent Document 2 also has a detailed description regarding the stoichiometric composition LT obtained by the gas phase method. Also, in Patent Document 5 and Non-Patent Document 2, when an LT in which an LT composition is uniformly transformed into Li-rich in the thickness direction by a vapor phase equilibrium method is used for a surface acoustic wave device, its frequency-temperature characteristics Has been reported to be preferable.

However, it has been found that the methods described in these patent documents do not always give favorable results. In particular, according to the method described in Patent Document 5, it takes a long time of 60 hours at a high temperature of about 1300 ° C. to process the wafer in the gas phase, so that the manufacturing temperature is high, the warpage of the wafer is large, and cracks are generated. Since the rate is high, productivity is poor, and there is a problem that the surface acoustic wave device material is expensive. Moreover, since the vapor pressure of Li ₂ O is low and the degree of modification of the sample to be modified varies depending on the distance from the Li source, the variation in characteristics is also large, and considerable improvement is necessary for industrialization.

  In addition, Patent Document 5 describes a manufacturing condition of a plate thickness of 0.5 mmt and a processing temperature of 1200 ° C. to 1350 ° C., but this is an old manufacturing method as it is, and a substrate thickness required for a surface acoustic wave device. It is much thicker. Although it may be possible to thin this substrate after the vapor phase treatment and finish it to the desired thickness, since Li is diffused and distorted, the rate of cracking during processing increases and the processing cost increases. It is obvious that half of the thickness of 0.5 mmt is reduced to 0.25 mmt, and the cost is high even when considering the material cost.

  Further, the lithium tantalate single crystal substrate for a surface acoustic wave element described in Patent Document 5 was found to have weak pyroelectricity in the process of investigation by the present inventors, and this pyroelectricity is removed. Therefore, as an example, when pyroelectric removal was performed by the method described in Patent Document 6, it was impossible to completely remove pyroelectricity.

Then, Patent Document 3, and then proton-exchanged LiNbO ₃ or LiTaO _3, a manufacturing method of attaching a refractive index distribution in the surface layer, such as LiNbO ₃ and LiTaO ₃ are described. However, if proton exchange is performed, the piezoelectricity of LiNbO ₃ , LiTaO _3, or the like is impaired, and thus there is a problem that it cannot be used as a material for a surface acoustic wave device.

In Non-Patent Document 1, 38.5 ° rotated Y-cut LiTaO ₃ (hereinafter referred to as “stoichiometric composition LT or SLT”) having a constant ratio composition prepared by a pulling method using a double crucible is: It is described that the electromechanical coupling coefficient is as high as 20% compared to a molten composition LiTaO ₃ (hereinafter, sometimes referred to as “congruent composition LT or CLT”) having the same composition ratio prepared by a normal pulling method. Has been. However, when the LT described in Non-Patent Document 1 is used, the SLT pulling speed is an order of magnitude smaller than that of the normal pulling method and the cost is high. There is a problem that it is difficult to use for surface wave device applications.

"Practical development of optomedia crystals that support IT," Science and Technology Promotion Coordination Results Report, 2002, Kenji Kitamura Bartasyte, A. et. al, “Reduction of temperature coefficient of frequency in LiTaO3 single crystals for surface acoustic wave applications” Applications of Ferroelectrics held jointly with 2012 European Conference on the Applications of Polar Dielectrics and 2012 International Symp Piezoresponse Force Microscopy and Nanoscale Phenomena in Polar Materials (ISAF / ECAPD / PFM), 2012 Intl Symp, 2012, Page (s): 1-3

JP2011-135245A US Pat. No. 6,652,644 (B1) JP 2003-207671 A JP2013-66032A WO2013 / 135886 (A1) Patent No. 4220997

  Accordingly, the present invention has been made in view of the above circumstances, and its purpose is to provide a piezoelectric oxide unit having good productivity, low cost, no problem of cracking and warping, and good temperature characteristics. It is to provide a method for manufacturing a crystal substrate.

  The inventors conducted extensive studies to achieve the above object, and performed a Li diffusion vapor phase treatment on a substrate having a substantially congruent composition. If it is modified to a crystal structure showing a profile in which the concentration decreases, the warp is small for surface acoustic wave devices, even without modifying to a uniform Li concentration crystal structure near the center in the thickness direction, The present inventors have found that a piezoelectric oxide single crystal substrate having no cracks and scratches and good temperature characteristics can be obtained, and the present invention has been achieved.

  That is, the present invention includes a step of slicing an oxide single crystal ingot having a substantially congruent composition that has undergone a single polarization treatment to cut out the oxide single crystal substrate, and a substrate in a powder containing a Li compound. In the atmosphere that does not contain oxygen in an embedding step and 850 ° C. to 1000 ° C., Li is diffused from the surface of the substrate to the inside by heating for 20 to 50 hours. The method includes a gas phase treatment step for modifying so as to show a profile in which the Li concentration decreases, and a step for polishing the substrate.

  Moreover, it is preferable to perform the gaseous-phase process process of this invention in the atmosphere which does not contain nitrogen atmosphere, inert gas atmosphere, or vacuum oxygen, for example. In that case, it is preferable to perform modification by Li diffusion from the surface to a depth of 10 μm to 50 μm in the thickness direction, and at least a part of the modified range has a pseudo stoichiometric composition. It is preferable. Further, it is preferable that the substrate is annealed at the Curie temperature or higher after the vapor phase treatment step.

  Furthermore, the thickness of the substrate of the present invention is preferably 200 μm or more and 400 μm or less, and the average particle size of the powder containing the Li compound is preferably 0.1 μm or more and 100 μm or less.

  According to the present invention, a piezoelectric oxide single crystal substrate having good temperature characteristics can be manufactured at a low temperature and in a short time as compared with the conventional method, so that the warpage is small and there is no crack or scratch. The substrate can be manufactured with high productivity and low cost.

FIG. 1 shows the Raman profile of the lithium tantalate single crystal substrate obtained in Example 1. FIG. 2 shows an insertion loss waveform of the series resonant SAW resonator formed with the input / output terminals formed on the lithium tantalate single crystal substrate obtained in the first embodiment. FIG. 3 shows an insertion loss waveform of the parallel resonant SAW resonator formed with the input / output terminals formed on the lithium tantalate single crystal substrate obtained in the first embodiment. FIG. 4 shows the temperature dependence of the antiresonance frequency of the SAW resonator formed on the lithium tantalate single crystal substrate obtained in Example 1. FIG. 5 shows the temperature dependence of the resonance frequency of the SAW resonator formed on the lithium tantalate single crystal substrate obtained in Example 1. FIG. 6 shows the Raman profile of the lithium tantalate single crystal substrate obtained in Example 2. FIG. 7 shows an insertion loss waveform of the series resonant SAW resonator formed with the input / output terminals formed on the lithium tantalate single crystal substrate obtained in the second embodiment. FIG. 8 shows an insertion loss waveform of the parallel resonant SAW resonator formed with the input / output terminals formed on the lithium tantalate single crystal substrate obtained in the second embodiment. FIG. 9 shows the temperature dependence of the antiresonance frequency of the SAW resonator formed on the lithium tantalate single crystal substrate obtained in Example 2. FIG. 10 shows the temperature dependence of the resonance frequency of the SAW resonator formed on the lithium tantalate single crystal substrate obtained in Example 2.

  Hereinafter, although an embodiment of the present invention is described in detail, the present invention is not limited to this.

  In the present invention, a single substrate (hereinafter, referred to as “wafer” or “small piece”) is obtained by slicing an oxide single crystal ingot having a substantially congruent composition subjected to a single polarization process. Cut and embed this in a powder containing Li compound, and in an atmosphere that does not contain oxygen, in order to diffuse Li from the substrate surface to the inside, it is heated for 10 hours to 50 hours at 850 ° C to 1000 ° C. The main characteristic is that the surface treatment is performed so as to show a profile in which the Li concentration is higher toward the surface and the Li concentration decreases toward the inside. Examples of the material of the piezoelectric oxide single crystal substrate manufactured according to the present invention include lithium compounds such as lithium tantalate, lithium niobate, and lithium tetraborate.

  Here, the “pseudo stoichiometric composition” in the present invention differs depending on the material. For example, in a lithium tantalate single crystal substrate, Li / (Li + Ta) represents a composition of 0.490 to 0.510. In the lithium niobate single crystal substrate, Li / (Li + Nb) has a composition of 0.490 to 0.510. For other materials, a “pseudo stoichiometric composition” may be defined based on common technical knowledge.

The material of the Li compound contained in the powder is not particularly limited, but is preferably a compound containing the same element as the oxide single crystal substrate material subjected to the modification treatment step. For example, when the material of the oxide single crystal substrate is LiTaO ₃ , Li ₃ TaO ₄ is preferably the main component of the powder material, and when it is LiNbO ₃ , Li ₃ NbO ₄ is the main component of the powder material. It is preferable to use as a component. The powder may be composed of a single compound, but may be a mixture of a plurality of compounds.

  The vapor phase treatment of Li diffusion according to the present invention is a technique in which a defect portion in a crystal structure is filled and a ratio of tantalum to lithium is close to a stoichiometric ratio of 1: 1. Since the diffusion of Li is promoted in a nitrogen atmosphere, an inert gas atmosphere, or a vacuum atmosphere that does not contain oxygen than the atmosphere in which oxygen exists, for example, by reducing the processing time, This is preferable because the warpage can be reduced.

  The reason for this is not necessarily clear, but for example, by setting the reaction atmosphere to a nitrogen atmosphere, oxygen deficient sites other than Li deficient sites are generated in the crystal structure. It is thought that diffusion will progress. In addition to the fact that Li fills the original Li deficiency site, it is thought that some interaction with this oxygen deficiency site works, and as a result, an unprecedented new crystal structure may be formed.

  Further, the Li diffusion gas phase treatment of the present invention is preferably performed at 850 ° C. to 1000 ° C. for about 10 hours to 50 hours. When the heating temperature is too high, the warpage of the substrate increases and the productivity also deteriorates. Further, when the processing time is short, the SAW response characteristic is deteriorated as in the case of Comparative Example 2, while when the processing time is too long, the warpage is increased and the productivity is deteriorated.

  Furthermore, in the gas phase treatment process of the present invention, it is preferable to perform modification by Li diffusion to a depth of about 10 μm to 50 μm in the thickness direction from the surface. If the depth of modification is less than 10 μm, the SAW response characteristics will deteriorate, and if it is modified to a depth of more than 50 μm, the treatment temperature will be high and the time will be longer, which will cause problems of warpage and cracking. Because there is no.

  In the gas phase treatment step of the present invention, it is preferable to modify so that at least a part of the modified range has a pseudo stoichiometric composition. In particular, it is preferable to modify so that the composition of the surface of the substrate becomes a pseudo stoichiometric composition.

  Here, the “pseudo stoichiometric composition” in the present invention differs depending on the material. For example, in a lithium tantalate single crystal substrate, Li / (Li + Ta) represents a composition of 0.490 to 0.510. In the lithium niobate single crystal substrate, Li / (Li + Nb) has a composition of 0.490 to 0.510. For other materials, a “pseudo stoichiometric composition” may be defined based on common technical knowledge.

  As a method for evaluating the composition of the piezoelectric oxide single crystal substrate, a known method such as Curie temperature measurement may be used, but nondestructive local composition can be evaluated by using Raman spectroscopy. It is.

  For lithium tantalate single crystals and lithium niobate single crystals, it is known that an approximately linear relationship can be obtained between the half-value 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)). Therefore, by using an expression representing such a relationship, the composition at an arbitrary position of the oxide single crystal substrate can be evaluated.

The relational expression between the half-width of the Raman shift peak and the Li concentration is obtained by measuring the Raman half-width for several samples with known compositions and different Li concentrations, but the Raman measurement conditions are the same. If there is, a relational expression already known in literature may be used. For example, for lithium tantalate single crystal, the following formula (1) may be used (see 2012 IEEE International Ultrasonics Symposium Proceedings, Page (s): 1252-1255), and for lithium niobate single crystal, the following formula (2 ) Or (3) may be used (see Applied Physics A 56, 311-315 (1993)).
Li / (Li + Ta) = (53.15-0.5FWHM ₁ ) / 100 (1)
Li / (Li + Nb) = (53.03-0.4739FWHM ₂ ) / 100 (2)
Li / (Li + Nb) = (53.29−0.1837FWHM ₃ ) / 100 (3)

Here, FWHM ₁ is the half width of the Raman shift peak near 600 cm ⁻¹ , FWHM ₂ is the half width of the Raman shift peak near 153 cm ⁻¹ , and FWHM ₃ is the Raman shift near 876 cm ^−1. The half width of the peak. Refer to each document for details of the measurement conditions.

  The thickness of the substrate of the present invention is preferably 200 μm or more and 400 μm or less. This is because if the substrate is too thick, the processing temperature must be set high and the processing time becomes long, so that it is difficult to obtain a substrate with little warpage and no cracks or scratches.

  The manufacturing method of the present invention can produce an LT substrate with good temperature characteristics without pyroelectricity, but even if the pyroelectricity is somewhat inferior, there are no problems such as warping and cracking, and SAW waveform characteristics. Can be sufficiently used for a surface acoustic wave device.

  The average particle size of the powder containing the Li compound used in the present invention is preferably from 0.1 μm to 100 μm, and more preferably from 1 μm to 50 μm. When the average particle size of the powder is larger than 100 μm, the contact between the powder and the oxide single crystal substrate becomes non-uniform in the gas phase treatment process, so the dispersion of Li diffusion in the substrate surface increases. This is not preferable because the characteristic variation of the manufactured piezoelectric oxide single crystal substrate becomes remarkable.

  Here, the “average particle diameter” of the powder of the present invention is a particle diameter distribution measured by a laser diffraction / scattering method, and an average diameter (volume average diameter) weighted by the volume of each particle diameter is “ Average particle diameter ". A smaller distribution width or variation in the particle size distribution is preferable because the contact between the powder and the oxide single crystal substrate becomes uniform.

  Next, the present invention will be specifically described based on examples.

<Example 1>
In Example 1, first, a 4-inch diameter lithium tantalate single crystal ingot having a ratio of Li: Ta of approximately congruent composition subjected to a single polarization treatment and a ratio of 48.5: 51.5 was sliced, and Z-cut and A 38.5 ° rotated Y-cut lithium tantalate substrate was cut to a thickness of 300 μm. Thereafter, if necessary, the surface roughness of each slice wafer was adjusted to a Ra value of 0.15 μm by a lapping process, and the finished thickness was 250 μm. Next, a substrate whose one side surface was finished to a quasi-mirror surface with an Ra value of 0.01 μm by plane polishing was embedded in a powder composed of Li, Ta, and O containing Li ₃ TaO ₄ as main components. In this case, as a powder mainly composed of Li ₃ TaO ₄ , Li ₂ CO ₃ : Ta ₂ O ₅ powder was mixed at a molar ratio of 7: 3 and baked at 1300 ° C. for 12 hours. .

When 0.1 g of this powder was dispersed in ethanol and the particle size distribution was measured (Microtrack MT3300II manufactured by Nikkiso Co., Ltd.), the obtained average particle size was 78 μm and the maximum particle size was 262 μm. . Then, such a powder mainly composed of Li ₃ TaO ₄ was spread in a small container, and a plurality of sliced wafers were embedded in the Li ₃ TaO ₄ powder.

Next, this small container is set in an electric furnace, and the inside of the furnace is made into an N ₂ atmosphere and heated at 950 ° C. for 36 hours to diffuse Li from the surface of the slice wafer to the center part, thereby forming a roughly congruent composition. To a pseudo-stoichiometric composition. Thereafter, the slice substrate subjected to this treatment was subjected to an annealing treatment at 750 ° C., which is equal to or higher than the Curie temperature, in the atmosphere for 12 hours. Further, the rough surface side was finished by sandblasting to an Ra value of about 0.15 μm, and the rough mirror side was polished by 3 μm to obtain a plurality of lithium tantalate single crystal substrates. At this time, this substrate was exposed to a temperature above the Curie point, but this substrate was not subjected to a single polarization treatment.

Using one of the substrates manufactured in this way, a laser Raman spectrometer (LabRam HR series manufactured by HORIBA Scientific, Ar ion laser, spot size 1 μm, room temperature) is arbitrarily separated from the outer peripheral side of this substrate by 1 cm or more. When the half width of the Raman shift peak near 600 cm ^−1, which is an index of the amount of Li diffusion, was measured from the surface in the depth direction, the result of the Raman profile shown in FIG. 1 was obtained.

  According to the results of the Raman profile shown in FIG. 1, the value of the half width of Raman decreases as the substrate is closer to the substrate surface in the depth direction of the substrate in the range of about 5 μm to about 50 μm. A profile in which the value of the Raman half width increases is shown.

  From the above results, the substrate manufactured in Example 1 has a higher Li concentration as it is closer to the substrate surface and a lower Li concentration as it is closer to the center of the substrate in the depth direction of the substrate. It was confirmed that a concentration profile was exhibited.

Moreover, since the Raman half-width of the substrate surface is about 6.1 cm ⁻¹ , the following formula (1) is used, the composition of the substrate surface is approximately Li / (Li + Ta) = 0.501, and the pseudo stoichiometry It turned out to be a composition.
Li / (Li + Ta) = (53.15-0.5FWHM ₁ ) / 100 (1)

  Furthermore, when the warpage of the 4-inch substrate subjected to Li diffusion was measured by an interference method using a laser beam, the value was as small as 50 μm, and cracks and cracks were not observed. Furthermore, when these lithium tantalate single crystal substrates were heated with a hot plate and the surface potential was measured, the voltage was 0 V. Therefore, even if the substrate of Example 1 was subjected to heat treatment, It was confirmed that there was no pyroelectricity.

Next, for the small piece cut out from the Z cut and 38.5 ° Y cut, vertical vibration in the thickness direction was given to the main surface and back surface of each piezo d33 / d15 meter (model ZJ-3BN) made by the Institute of Chinese Academy of Voices. As a result of observing the induced voltage waveform, the voltage was 0 V and the piezoelectric constant d33 was also zero. Similarly, by using the d15 unit of the same apparatus and observing the voltage waveform induced by applying vibration in the shear direction to the respective main surface and back surface, a waveform indicating piezoelectricity was obtained. When this small piece was hit with the tip of a synchroscope probe and the piezoelectric response was observed, a waveform showing the piezoelectric response was obtained.
Therefore, since the substrate of Example 1 has piezoelectricity, it was confirmed that it could be used as a surface acoustic wave element.

  Next, the piezoelectric response was observed by hitting the tip of the synchroscope with the tip of the synchroscope about the small piece removed from the surface layer of one side by hand wrapping from the surface layer of each small piece. . Similarly, on the opposite surface of each small piece, a small piece removed by 50 μm thickness from the surface layer by hand wrapping was observed by hitting with the tip of the synchroscope probe, and no piezoelectric response was observed. Further, when this small piece was observed by the d33 / d15 meter to observe the voltage waveform induced by applying vertical vibration in the thickness direction to the main surface and the back surface, the voltage was 0 V and the piezoelectric constant d33 was also zero. . Similarly, a voltage waveform induced by applying vibration in the shear direction to the main surface and the back surface using the d15 unit of the same apparatus did not show piezoelectricity, and the voltage was 0V.

  From the above results, the substrate of Example 1 shows piezoelectricity by modifying to a pseudo stoichiometric composition over a depth of 50 μm from the surface layer portion, but shows piezoelectricity at a portion deeper than 50 μm. From this, it was confirmed that the multi-domain structure has a polarization direction that is not aligned in one direction near the center in the depth direction.

  Next, the surface of the 4 inch 38.5 ° Y-cut lithium tantalate single crystal substrate, which has been polished by performing Li diffusion and annealing, was sputtered to form a 0.05 μm thick Al film. did. Thereafter, a resist was applied to the substrate subjected to this treatment, the SAW resonator and the ladder filter pattern were exposed and developed by an aligner, and patterning for SAW characteristic evaluation was performed by RIE. One wavelength of the patterned SAW electrode was 4.8 μm.

  In this SAW resonator, a series resonance type resonator and a parallel resonance type resonator provided with input / output terminals were formed, and the SAW waveform characteristics were confirmed by an RF prober. The results shown are obtained. 2 and 3 show the SAW waveform characteristics at that time. For comparison, the SAW waveform characteristics of a 38.5 ° Y-cut lithium tantalate single crystal substrate not subjected to Li diffusion treatment are also shown. Show.

  Therefore, from this result, it was confirmed that the substrate of Example 1 also showed good SAW waveform characteristics for a surface acoustic wave device.

Then, the SAW of the sound velocity v _r in away ± 20 mm from the center and the center of the substrate of Example 1 crystal X-axis direction, the insertion loss diagram of a two-port resonator similar parallel resonance type as FIG. 3 Since the frequency at the dip position is the resonance frequency f _a (MHz) and the wavelength is 4.8 μm, it was calculated by the following formula (4).
vr = fa × 4.8 (m / s) (4)

As a result, the variation of the SAW sound velocity v _r of the three points away ± 20 mm from the center and the center of the substrate to the crystal X-axis direction was less 2m / s. For comparison, the sound velocity variation at the same three points on the 38.5 ° Y-cut lithium tantalate single crystal substrate not subjected to Li diffusion treatment was 1 m / s or less.

  Further, when the temperature of the stage was changed from about 16 ° C. to 70 ° C. and the temperature coefficients of the antiresonance frequency and the resonance frequency were confirmed, the results shown in FIGS. 4 and 5 were obtained. According to this result, the temperature coefficient of the anti-resonance frequency of Example 1 is -19 ppm / ° C from FIG. 4, and the temperature coefficient of the resonance frequency is -15 ppm / ° C from FIG. The frequency temperature coefficient was -17 ppm / ° C. For comparison, the temperature coefficient of the 38.5 ° Y-cut lithium tantalate single crystal substrate not subjected to Li diffusion treatment is that the anti-resonance frequency temperature coefficient is -42ppm / ° C, and the temperature coefficient of the resonance frequency is- Since it was 32 ppm / ° C, the average frequency temperature coefficient was -37 ppm / ° C.

  Therefore, since the substrate of Example 1 has a smaller average frequency temperature coefficient and less characteristic variation with respect to the temperature than the substrate not subjected to the Li diffusion treatment, it is confirmed that the temperature characteristics are good. I was able to.

<Example 2>
In Example 2, as in Example 1, from a 4 inch diameter lithium tantalate single crystal ingot having a ratio of Li: Ta of approximately congruent composition subjected to a single polarization treatment in a ratio of 48.5: 51.5, This was sliced, and a Z-cut and 42 ° rotated Y-cut lithium tantalate substrate was cut to a thickness of 300 μm. Thereafter, if necessary, the surface roughness of each slice wafer is adjusted to 0.15 μm in Ra value by a lapping process, Li diffusion treatment is performed under the same conditions as in Example 1, and a pseudo stoichiometric is obtained from the approximate congruent composition. The meter composition was changed.

Next, in Example 2, at different annealing conditions as in Example 1, namely the annealed for 12 hours at the Curie temperature or higher of 1000 ° C. under N ₂ slices wafer. Thereafter, finishing and polishing similar to those of Example 1 were performed to obtain a plurality of lithium tantalate single crystal substrates. In this case, this substrate was exposed to a temperature above the Curie point, but this substrate was not subjected to a single polarization treatment.

For one of the substrates manufactured in this manner, the half-value width of the Raman shift peak near 600 cm ^−1, which is an index of the amount of Li diffusion, was measured from the surface of the substrate in the depth direction. The Raman shown in FIG. Profile results were obtained. According to the results of FIG. 6, the value of the Raman half-value width decreases as the substrate is closer to the substrate surface in the range of about 50 μm in the depth direction from the substrate surface, and the value of the Raman half-value width is closer to the center of the substrate. Showed an increasing profile.

  From the above results, the substrate manufactured in Example 2 has a concentration in which the Li concentration is higher as it is closer to the substrate surface in the depth direction from the substrate surface to about 50 μm, and the Li concentration is decreased as it is closer to the center of the substrate. It was confirmed that the profile was shown.

Further, since the Raman half-value width of the substrate surface is about 6.9 cm ⁻¹ , the composition of the substrate surface is approximately Li / (Li + Ta) when calculated using the above formula (1) in the same manner as in Example 1. = 0.497, indicating a pseudo stoichiometric composition.

  Furthermore, when the warpage of a 4-inch substrate with Li diffusion was measured by an interference method using laser light, the value was as small as 50 μm, and cracks and cracks were not observed. Furthermore, when these lithium tantalate single crystal substrates were heated with a hot plate and the surface potential was measured, the voltage was 0 V. Therefore, even if this was subjected to heat treatment, the surface was not pyroelectric. I was able to confirm.

  Next, in Example 2, a small piece is cut out from the Z cut and the 42 ° Y cut, and the voltage waveform induced by applying vertical vibration in the thickness direction to the main surface and the back surface of the small piece is induced as in Example 1. When observed, the voltage was 0 V and the piezoelectric constant d33 was also zero. Similarly, when a voltage waveform induced by applying vibration in the shear direction to the main surface and the back surface was observed, a waveform indicating piezoelectricity was obtained. Moreover, when this small piece was hit and observed with the probe tip of the synchroscope, a waveform showing a piezoelectric response was also obtained.

  Therefore, from this result, it was confirmed that the substrate of Example 2 also has piezoelectricity and can be used for a surface acoustic wave device.

  Next, in the same manner as in Example 1, when a small piece having a thickness of 50 μm removed from the surface of one side by hand wrapping was observed by hitting with the probe tip of the synchroscope, a smaller voltage was observed in the piezoelectric response. . Similarly, on the opposite surface, when a small piece removed by 50 μm from the surface layer by hand lap was hit with the tip of the synchroscope probe and observed, no piezoelectric response was observed. Similarly to Example 1, when the voltage waveform induced by applying vertical vibration in the thickness direction to each main surface and the back surface was observed, the voltage was 0 V, the piezoelectric constant d33 was also zero, and the main surface The voltage waveform induced by applying vibration in the shear direction to the back surface did not show piezoelectricity, and the voltage was 0V.

  Therefore, from the above results, the substrate of Example 2 also shows piezoelectricity by modifying it to a pseudo stoichiometric composition over a depth of 50 μm from the surface layer portion. Thus, it was confirmed that the multi-domain structure has a polarization direction that is not aligned in one direction near the center in the depth direction.

Next, also in Example 2, the same process as in Example 1 was performed on the surface of the 4-inch 42 ° Y-cut lithium tantalate single crystal substrate that had been subjected to the Li diffusion process and the annealing process to finish the polishing process. When the SAW characteristics were confirmed, the results shown in FIGS. 7 and 8 were obtained. FIGS. 7 and 8 show the SAW waveform characteristics at that time. For comparison, the SAW waveform characteristics of a 42 ° Y-cut lithium tantalate single crystal substrate not subjected to Li diffusion treatment are also shown. Show.
Therefore, from this result, it was confirmed that the substrate of Example 2 also showed good SAW waveform characteristics for a surface acoustic wave device.

  Further, as in Example 1, the antiresonance frequency and the temperature coefficient of the resonance frequency were confirmed, and the results shown in FIGS. 9 and 10 were obtained. According to this result, the temperature coefficient of the anti-resonance frequency of Example 2 is −19 ppm / ° C. from FIG. 9, and the temperature coefficient of the resonance frequency is −21 ppm / ° C. (FIG. 10) from FIG. Thus, the average temperature coefficient was −20 ppm / ° C. For comparison, the temperature coefficient of the 42 ° Y-cut lithium tantalate single crystal substrate not subjected to Li diffusion treatment is that the anti-resonance frequency temperature coefficient is -42ppm / ° C and the resonance frequency temperature coefficient is -32ppm. Since it was / ° C. (FIG. 10), the average frequency temperature coefficient was −37 ppm / ° C.

  Therefore, it is confirmed that the substrate of Example 2 has good temperature characteristics because the average frequency temperature coefficient is small and the characteristic variation with respect to temperature is small compared to the substrate not subjected to the Li diffusion treatment. I was able to.

<Example 3>
In Example 3, as in Example 1, a lapping process and a surface polishing process were performed in the same manner as in Example 1, using a lithium tantalate substrate having a Z-cut and 38.5 ° Y-cut 300 μm-thick congruent composition. At the same time, Li diffusion treatment, annealing treatment and finish polishing were performed under the same conditions as in Example 1 to obtain a plurality of lithium tantalate single crystal substrates for surface acoustic wave elements.

  Next, in Example 3, in a state where a plurality of substrates are stacked, a single polarization process that is not performed in Examples 1 and 2 is performed in the + Z direction of the substrate at a temperature of 750 ° C. that is equal to or higher than the Curie point. This was done by applying an electric field. When the warpage of the 4-inch substrate subjected to the single polarization treatment was measured by an interference method using a laser beam, the value was as small as 50 μm, and cracks and cracks were not observed. When this substrate was heated with a hot plate and the surface potential was observed, the voltage was 2 kV.

  From the above results, the substrate of Example 3 had a small warp and no cracks or cracks on the surface, as in Example 1, but when subjected to heat treatment, it showed strong pyroelectricity. It was to show. This strong pyroelectricity is caused by the single polarization treatment, and the temperature characteristics are slightly inferior to those of Example 1 and Example 2, but it is confirmed that it is better than normal LT. I was able to.

  In addition, when a small piece was cut out from the Z cut and 38.5 ° Y cut and the piezoelectric waveform was observed, the result showing the piezoelectricity was obtained in the same manner as in Example 1, so that it can be used for a surface acoustic wave device. I was able to confirm.

  Further, the same process as in Example 1 was performed on the surface of a 4 inch 38.5 ° Y-cut lithium tantalate single crystal substrate that had been subjected to a Li diffusion process and an annealing process under the same conditions as in Example 1 and finished the polishing process. When the anti-resonance frequency and the temperature coefficient of the resonance frequency were confirmed, the temperature coefficient of the anti-resonance frequency was -32ppm / ° C, and the temperature coefficient of the resonance frequency was -29ppm / ° C. The coefficient was -31 ppm / ° C.

  Therefore, the substrate of Example 3 also has a slightly lower average frequency temperature coefficient than the substrate for comparison shown in FIGS. 4 and 5 (the substrate not subjected to Li diffusion treatment), and has characteristics with respect to temperature. Since the fluctuation was small, it was confirmed that the temperature characteristics were slightly good.

<Example 4>
In Example 4, as in Example 1, using a lithium tantalate substrate with a roughly congruent composition having a thickness of 300 μm with a Z cut and a 38.5 ° rotated Y cut, lapping and surface polishing were performed in the same manner as in Example 1. At the same time, Li diffusion treatment, annealing treatment and finish polishing were performed under the same conditions as in Example 1 to obtain a plurality of lithium tantalate single crystal substrates for surface acoustic wave elements. However, in the Li diffusion treatment step of Example 4, Li ₂ CO ₃ : Ta ₂ O ₅ powder was mixed at a molar ratio of 7: 3 as a powder mainly composed of Li ₃ TaO ₄ , and the temperature was 1300 ° C. Baked for 12 hours and passed through a sieve having an opening of 45 μm. The average particle size of this powder measured by the same method as in Example 1 was 25 μm, and the maximum particle size was 45 μm. Further, this substrate was not subjected to a single polarization treatment.

For one of the substrates manufactured in this way, the half-value width of the Raman shift peak near 600 cm ^−1, which is an index of the amount of Li diffusion, was measured from the surface of the substrate in the depth direction. 1 shows almost the same Raman profile as that in the depth direction from the substrate surface to the depth of about 50 μm. The closer to the substrate surface, the lower the Raman half-width value, and the closer to the substrate center, the higher the Raman half-width value. It was confirmed that the profile to be displayed was shown.

  From the above results, the substrate manufactured in Example 4 has a concentration in which the Li concentration is higher as it is closer to the substrate surface in the range of about 50 μm in the depth direction from the substrate surface, and the Li concentration is decreased as it is closer to the center of the substrate. It was confirmed that the profile was shown.

Next, the SAW sound velocity v _r at the center portion of the substrate of Example 4 and the location ± 20 mm away from the center in the crystal X-axis direction is shown in the insertion loss diagram of the parallel resonance type two-port resonator similar to FIG. Since the frequency at the dip position is the resonance frequency f _a (MHz) and the wavelength is 4.8 μm, it is calculated by the above equation (4).

As a result, the variation of the SAW sound velocity v _r of the three points away ± 20 mm from the center and the center of the substrate to the crystal X-axis direction was less 1 m / s. Therefore, in Example 4, it was found to be smaller than the variation of the SAW of the sound velocity v _r Example 1 (2m / s or less).

<Example 5>
In Example 5, first, a 4 inch diameter lithium niobate single crystal ingot having a ratio of Li: Nb of approximately congruent composition subjected to a single polarization treatment in a ratio of 48.5: 51.5 was sliced, and Z-cut and A 41 ° rotated Y-cut lithium niobate substrate was cut to a thickness of 300 μm. Thereafter, the surface roughness of each slice wafer was adjusted to 0.15 μm in Ra value by a lapping process as necessary, and the finished thickness was set to 250 μm.

Next, a substrate whose one surface was finished to a quasi-mirror surface with a Ra value of 0.01 μm by flat polishing was embedded in a powder composed of Li, Nb, and O containing Li ₃ NbO ₄ as main components. In this case, as a powder mainly composed of Li ₃ NbO ₄ , Li ₂ CO ₃ : Nb ₂ O ₅ powder was mixed at a molar ratio of 7: 3 and baked at 1000 ° C. for 12 hours. . Then, such a powder mainly composed of Li ₃ NbO ₄ was spread in a small container, and a plurality of sliced wafers were embedded in the Li ₃ NbO ₄ powder.

After that, this small container is set in an electric furnace, and the inside of the furnace is made into an N ₂ atmosphere and heated at 900 ° C. for 36 hours to diffuse Li from the surface of the slice wafer to the center part, and from a roughly congruent composition. The pseudo-stoichiometric composition was changed. In addition, the slice substrate subjected to this treatment was subjected to an annealing treatment under N ₂ at 750 ° C. that is equal to or higher than the Curie temperature for 12 hours.

  Further, the rough side was finished by sandblasting to an Ra value of about 0.15 μm, and the rough mirror side was polished by 3 μm to obtain a plurality of lithium niobate single crystal substrates. At this time, this substrate was exposed to a temperature above the Curie point, but this substrate was not subjected to a single polarization treatment.

For one of the substrates manufactured in this manner, the half-value width of the Raman shift peak near 876 cm ^−1, which is an index of the amount of Li diffusion, is measured from the surface of the substrate in the depth direction. In the range of about 5 μm to about 60 μm in the depth direction, the profile of the Raman half-value width decreases as it approaches the substrate surface, and the Raman half-width value increases as it approaches the center of the substrate. The Raman half width on the substrate surface was 17.8 cm ⁻¹ , and the Raman half width at a position of 62 μm or less in the depth direction was 23.0 cm ⁻¹ .

  From the above results, Example 5 shows a concentration profile in which the Li concentration is higher as it is closer to the substrate surface and the Li concentration is lower as it is closer to the center of the substrate in the range of about 5 μm to about 60 μm in the depth direction of the substrate. I was able to confirm.

In addition, since the Raman half-value width of the substrate surface is about 17.8 cm ⁻¹ , when calculated using the following formula (3), the composition in the range is approximately Li / (Li + Nb) = 0.500, It was found that the composition was ichiometric.
Li / (Li + Nb) = (53.29−0.1837FWHM ₃ ) / 100 (3)

  Next, when the warpage of the 4-inch substrate subjected to Li diffusion was measured by an interference method using a laser beam, the value was as small as 50 μm, and cracks and cracks were not observed. Moreover, when these lithium tantalate single crystal substrates were heated with a hot plate and the surface potential was measured, the voltage was 0 V. Therefore, even if the substrate of Example 5 was subjected to heat treatment, It was confirmed that there was no pyroelectricity.

In addition, a small piece cut from the Z cut and 41 ° Y cut was subjected to vertical vibration in the thickness direction on the main surface and back surface of each piezo d33 / d15 meter (model ZJ-3BN) manufactured by the Institute of Chinese Vocal Music. When the induced voltage waveform was observed, the voltage was 0 V and the piezoelectric constant d33 was also zero. Similarly, by using the d15 unit of the same apparatus and observing the voltage waveform induced by applying vibration in the shear direction to the respective main surface and back surface, a waveform indicating piezoelectricity was obtained. When this small piece was hit with the tip of a synchroscope probe and the piezoelectric response was observed, a waveform showing the piezoelectric response was obtained.
Therefore, since the substrate of Example 5 has piezoelectricity, it was confirmed that it could be used as a surface acoustic wave element.

  Next, the piezoelectric response was observed by hitting the tip of the synchroscope with the tip of the synchroscope with respect to the small piece that was removed from the surface layer of one side of the small piece by hand wrapping, and a smaller voltage was observed for the piezoelectric response. . Similarly, on the opposite surface of each small piece, a small piece removed by 60 μm thickness from the surface layer by hand wrapping was observed by striking with the tip of the synchroscope probe, and no piezoelectric response was observed.

  Further, when this small piece was observed by the d33 / d15 meter to observe the voltage waveform induced by applying vertical vibration in the thickness direction to the main surface and the back surface, the voltage was 0 V and the piezoelectric constant d33 was also zero. . Similarly, a voltage waveform induced by applying vibration in the shear direction to the main surface and the back surface using the d15 unit of the same apparatus did not show piezoelectricity, and the voltage was 0V.

  Therefore, from this result, the substrate of Example 5 is modified to a pseudo stoichiometric composition over the depth of 60 μm from the surface layer portion to exhibit piezoelectricity, but at a portion deeper than 60 μm, the piezoelectricity is exhibited. Since it is not shown, it was confirmed that the multi-domain structure where the polarization direction is not aligned in one direction near the center in the depth direction.

  Next, the surface of the 4-inch 41 ° Y-cut lithium niobate single crystal substrate, which has been subjected to the Li diffusion treatment and annealing treatment, was sputtered to form a 0.05 μm thick Al film. did. Thereafter, a resist was applied to the substrate subjected to this treatment, the SAW resonator and the ladder filter pattern were exposed and developed by an aligner, and patterning for SAW characteristic evaluation was performed by RIE. One wavelength of the patterned SAW electrode was 4.8 μm.

  In this SAW resonator, a series resonance type resonator and a parallel resonance type resonator provided with input / output terminals were formed, and when the SAW waveform characteristics were confirmed by an RF prober, the substrate of Example 5 was also obtained. It was confirmed that the SAW waveform characteristic for a surface acoustic wave device was good.

  Further, when the temperature coefficient of the anti-resonance frequency and the resonance frequency was confirmed by changing the temperature of the stage from about 16 ° C. to 70 ° C., the temperature coefficient of the anti-resonance frequency of Example 5 was −34 ppm / ° C. Since the temperature coefficient of the resonance frequency is −50 ppm / ° C., the average frequency temperature coefficient was −42 ppm / ° C. For comparison, the temperature coefficient of the 41 ° Y-cut lithium niobate single crystal substrate without Li diffusion treatment is -56ppm / ° C at the anti-resonance frequency, and- Since it was 72 ppm / ° C., the average frequency temperature coefficient was −64 ppm / ° C.

  Therefore, it is confirmed that the substrate of Example 5 has good temperature characteristics because the average frequency temperature coefficient is small and the characteristic variation with respect to temperature is small compared to the substrate not subjected to Li diffusion treatment. I was able to.

Comparative example

<Comparative Example 1>
Next, Comparative Example 1 to be compared with the present invention will be described. In Comparative Example 1, gas phase treatment is performed on a finished substrate having a thickness of 550 μm over 60 hours at a high temperature of 1250 ° C. described in Patent Document 5. This is an example.

  In Comparative Example 1, a 600 μm thick lithium tantalate substrate having a roughly congruent composition than that of Example 1 was cut out from the same Z-cut and 38.5 ° Y-cut single crystal ingot as in Example 1, and this was performed. Similar to Example 1, the surface roughness was adjusted to Ra value of 0.15 μm by a lapping process as necessary, and the finished thickness was 550 μm. Also, the surface on one side was finished to a quasi-mirror surface with an Ra value of 0.01 μm by flat polishing.

  Next, in a state where the atmosphere in the furnace is an air atmosphere, a Li-phase gas diffusion treatment is performed at 1250 ° C. for 60 hours at a higher temperature for a longer time than in Example 1, and Li is transferred from the substrate surface to the center. Was diffused to change from a roughly congruent composition to a pseudo stoichiometric composition. Thereafter, the Li diffusion-treated substrate was subjected to the same annealing treatment and finish polishing as in Example 1 to obtain a plurality of lithium tantalate single crystal substrates for surface acoustic wave elements. At this time, this substrate was exposed to a temperature equal to or higher than the Curie point, but this substrate was not subjected to a single polarization treatment.

When the Li concentration profile in the depth direction from the surface was measured for one of the substrates thus manufactured, the Raman half-value width indicating the Li concentration was 6.0 cm ⁻¹ , and this value was from the surface layer to the inside. It showed a uniform Li concentration profile.

  In addition, this 4-inch substrate is greatly deformed and cannot be measured by the laser beam interference method that measures the warp. When the amount of warp is measured by a laser displacement sensor, the warp is as large as 1500 μm. A streak-like scratch was observed. Furthermore, when this substrate was heated with a hot plate and the surface potential was observed, the voltage was 1 kV.

  Next, a small piece was cut out from the Z cut and the 38.5 ° Y cut, and the voltage waveform induced by applying vertical vibration in the thickness direction to the main surface and the back surface was observed in the same manner as in Example 1. The voltage waveform was as follows. The voltage waveform induced by applying vibration in the shear direction to the main surface and the back surface was also a waveform indicating piezoelectricity. In addition, when the Z-cut and 38.5 ° Y-cut pieces were struck with the tip of the synchroscope probe, a piezoelectric response waveform was obtained.

  A similar piezoelectric response waveform was also obtained when a small piece of 50 μm thickness was removed from the surface of one side by hand wrapping with the tip of the synchroscope probe. Similarly, on the opposite side of the small piece, a similar piezoelectric response waveform was obtained when the small piece with the surface layer removed by a thickness of 50 μm was struck with the tip of the synchroscope probe.

  From the above results, in the method of Comparative Example 1, since the Li diffusion treatment was performed at a higher temperature for a longer time than in Example 1, the diffusion of Li ions progressed from the surface layer to the center deeper than 50 μm. It was confirmed that the crystal structure had a pseudo stoichiometric composition showing a uniform Li concentration profile across the center in the thickness direction of the substrate.

  Next, the surface of the 4 inch 38.5 ° Y-cut lithium tantalate single crystal substrate, which has been subjected to the Li diffusion treatment and annealing treatment, is subjected to the same treatment as in Example 1 to confirm the SAW waveform characteristics. However, patterning was not possible due to the large warpage of the substrate.

  Therefore, from this result, as shown in Comparative Example 1, when the Li diffusion treatment is performed at a high temperature of 1250 ° C. for 60 hours, a pseudo stoichiometry showing a uniform Li concentration profile from the surface layer to the inside is obtained. Although a crystal structure with a metric composition was obtained, on the other hand, it was confirmed that the manufacturing conditions were high temperature for a long time, so that the warp was large and streak-like scratches were generated on the substrate surface.

<Comparative Example 2>
Next, Comparative Example 2 will be described. In Comparative Example 2, the Li diffusion treatment is performed under the condition of 36 hours at 950 ° C. and 5 hours at 950 ° C. as compared with the case of Example 1. This is an example of changing the time extremely short.

In Comparative Example 2, as in Example 1, first, a 4-inch diameter lithium tantalate single crystal ingot having a ratio of Li: Ta of approximately congruent composition that was subjected to a single polarization treatment was 48.5: 51.5. Was sliced, and a Z-cut and 38.5 ° rotated Y-cut lithium tantalate substrate was cut to a thickness of 300 μm. Thereafter, a plurality of quasi-mirror substrates subjected to the same polishing treatment as in Example 1 were embedded in a small container in which powder mainly composed of Li ₃ TaO ₄ was spread, and 5 ° C. at 950 ° C. in an N ₂ atmosphere. A Li diffusion treatment for a time was applied to change the composition from a roughly congruent composition to a pseudo stoichiometric composition from the surface to the inside.

  Then, after that, the substrate was not subjected to annealing treatment, and the finishing and polishing processes similar to those in Example 1 were performed to obtain a plurality of lithium tantalate single crystal substrates. At this time, this substrate was not subjected to a single polarization treatment.

For one of the substrates thus manufactured, the half-value width of the Raman shift peak near 600 cm ^−1, which is an index of the amount of Li diffusion, was measured from the surface of the substrate in the depth direction. Was 6.5 cm ⁻¹ and the Raman half width from the surface of the sample to a depth of 5 μm was 9 cm ⁻¹ .

  From the above results, in Comparative Example 2, the diffusion of Li progresses at a depth of 5 μm from the surface, and the surface is modified to show a profile in which the Li concentration is higher toward the surface and the Li concentration decreases toward the inside. I was able to confirm that.

  Next, for the small piece cut out from the Z-cut and 38.5 ° Y-cut, vertical vibration in the thickness direction was applied to the main surface and back surface of each of the main surface and back surface using a piezo d33 / d15 meter (model ZJ-3BN) manufactured by the Institute of Chinese Academy of Science When the induced voltage waveform was observed, the voltage was 0 V and the piezoelectric constant d33 was also zero. Similarly, by using the d15 unit of the same apparatus and observing the voltage waveform induced by applying vibration in the shear direction to the respective main surface and back surface, a waveform indicating piezoelectricity was obtained. When this small piece was hit with the tip of a synchroscope probe and the piezoelectric response was observed, a waveform showing a weak piezoelectric response was obtained.

  Therefore, from this result, it was confirmed that the substrate of Comparative Example 2 was modified to a depth of 5 μm from the surface and had weak piezoelectricity, but because the degree of modification was shallow, this substrate was surface acoustic wave. Whether or not it can be used as an element was confirmed by performing the following measurement.

  That is, when the piezoelectric response was observed by tapping the tip of the synchroscope with the tip of the small piece, which was removed from the surface layer on one side of each small piece by hand wrapping, the piezoelectric response showed a smaller voltage than in the above case. Similarly, on the opposite surface of each small piece, when the small piece removed by a thickness of 5 μm from the surface layer by hand wrapping was observed by hitting with the tip of the synchroscope probe, no piezoelectric response was observed. Further, when this small piece was observed by the d33 / d15 meter to observe the voltage waveform induced by applying vertical vibration in the thickness direction to the main surface and the back surface, the voltage was 0 V and the piezoelectric constant d33 was also zero. . Similarly, a voltage waveform induced by applying vibration in the shear direction to the main surface and the back surface using the d15 unit of the same apparatus did not show piezoelectricity, and the voltage was 0V.

  Therefore, from this result, the substrate of Comparative Example 2 has piezoelectricity at a depth of 5 μm, but does not exhibit piezoelectricity at a portion deeper than 5 μm because it is a multi-domain structure without modification. Was confirmed.

  Therefore, after this surface was sputtered to form a 0.05 μm thick Al film, a resist was applied, the SAW resonator and the ladder filter pattern were exposed and developed with an aligner, and SAW was performed with RIE. Patterning for characteristic evaluation was performed. One wavelength of the patterned SAW electrode was 4.8 μm. In this SAW resonator, a series resonance type resonator with input / output terminals and a parallel resonance type resonator are formed, and when the SAW waveform characteristics are confirmed by an RF prober, the SAW response waveform collapses. It was found that it cannot be used as a SAW device.

Therefore, from this result, as in Comparative Example 2, if the Li diffusion treatment is as short as 5 hours, the modification by Li diffusion does not proceed sufficiently, so that the modification depth is as shallow as 5 μm. It was confirmed that the SAW response waveform collapsed and could not be used for surface acoustic wave devices.

Claims (7)

  A step of slicing an oxide single crystal ingot having a substantially congruent composition subjected to a single polarization treatment to cut out the oxide single crystal substrate, a step of embedding the substrate in a powder containing a Li compound, and oxygen In an atmosphere that does not contain, Li is diffused from the surface of the substrate to the inside by heating at 850 ° C. to 1000 ° C. for 10 hours to 50 hours, and the Li concentration is higher toward the surface, and the Li concentration decreases as it goes to the inside. A method for producing a piezoelectric oxide single crystal substrate, comprising: a gas phase treatment step for modifying the substrate so as to exhibit a profile; and a step for polishing the substrate.   2. The method for manufacturing a piezoelectric oxide single crystal substrate according to claim 1, wherein the atmosphere containing no oxygen is any one of a nitrogen atmosphere, an inert gas atmosphere, and a vacuum.   3. The piezoelectric oxide single crystal substrate according to claim 1, wherein in the gas phase treatment step, modification by Li diffusion is performed from the surface to a depth in a thickness direction of 10 μm to 50 μm. Production method.   4. The piezoelectric oxidation according to claim 1, wherein in the gas phase treatment step, the composition is modified so that at least a part of the modified range has a pseudo stoichiometric composition. 5. Of manufacturing a single crystal substrate.   5. The method according to claim 1, further comprising a step of annealing the substrate at a temperature equal to or higher than a Curie temperature and a step of sandblasting the rough surface side of the substrate after the vapor phase treatment step. A method for producing a piezoelectric oxide single crystal substrate according to claim 1.   6. The method of manufacturing a piezoelectric oxide single crystal substrate according to claim 1, wherein the substrate has a thickness of 200 μm or more and 400 μm or less. The method for producing a piezoelectric oxide single crystal substrate according to any one of claims 1 to 6, wherein an average particle size of the powder containing the Li compound is 0.1 µm or more and 100 µm or less.

Images