JP4803157B2 - Load sensor and manufacturing method thereof - Google Patents

Description

  The present invention relates to a smart airbag that is provided under a seat of a vehicle, measures the weight of an occupant, and adjusts the operation method of the airbag according to the weight, a load sensor that measures various loads, and a method of manufacturing the same.

  Conventionally, a load sensor in which a strain sensitive resistor is formed on a metal substrate via a glass layer has been proposed in Patent Document 1 and the like.

  FIG. 17 is a cross-sectional view for explaining the structure of such a conventional load sensor, and is an example of a load sensor having a single-layer wiring. In FIG. 17, the glass enamel 2 is formed on the metal elastic body 1, and the wiring 3a and the strain sensitive resistor 4 are formed thereon. However, when crystalline glass is used for the glass enamel 2 in particular, it is known that when the crystallinity of the precipitated crystals is increased, the surface becomes rough.

  In Patent Document 2, in order to reduce the surface roughness of such crystallized glass, a metal alkoxide layer is formed on the crystallized glass layer to reduce the surface roughness. And it proposes forming a wiring pattern and a strain sensitive resistor on it.

However, since the metal alkoxide layer has lower strength and film quality than general glass, it is difficult to use it in fields where high reliability and strength are required. In addition, since the metal alkoxide layer itself easily interdiffuses with the ruthenium oxide (RuO ₂ ) -based strain sensitive resistor, the gauge factor of the strain sensitive resistor (resistance value change with strain) and TCR (temperature change rate of resistance value). There is a problem that is easily affected.

Here, the problem and subject of the crystallized glass which has been proposed for the load sensor will be described in more detail. Table 1 shows the composition of crystallized glass for wiring single layer proposed in Patent Document 2 and Patent Document 3. MgO is 16～50Wt (wt)%, BaO is 0~50wt%, CaO is 0~20wt%, La ₂ O ₃ is 0~40wt%, B ₂ O ₃ is 5~34wt%, SiO ₂ is 7~30wt %, MO ₂ (M is at least one of Zr, Ti and Sn) 0 to ₅ wt% and P ₂ O _{5 0} to ₅ wt%.

  In addition to these, Patent Document 4, Patent Document 5, Patent Document 6, and Patent Document 7 propose similar crystallized glass compositions.

  However, when a strain sensitive resistor is formed directly on the crystallized glass, there is a problem that characteristics are likely to vary due to the surface roughness. In order to reduce the surface roughness of crystallized glass, for example, Patent Document 2 proposes forming a metal alkoxide layer on the crystallized glass layer and reducing the surface roughness. And it proposes forming a wiring pattern and a strain sensitive resistor on it.

However, since the metal alkoxide layer has lower strength and film quality than general glass, it is difficult to use it in fields where high reliability and strength are required. Also, since the metal alkoxide layer itself is easily interdiffused with the RuO ₂ -based strain sensitive resistor, the gauge factor (resistance value change with respect to strain) and TCR (temperature change rate of resistance value) of the strain sensitive resistor are easily affected. The problem arises. In addition, since these materials are not originally intended for multilayering, unexpected problems are likely to occur when the multilayering is performed.

  Patent Document 8 proposes forming an interface layer as a method for suppressing the influence of mutual diffusion between the glass layer and the resistor.

  FIG. 18 illustrates a case where an interface layer is formed. On the metal elastic body 1, a glass layer 5, a wiring 3 b and a strain sensitive resistor 4 are formed via an interface layer 6.

  Patent Document 8 proposes an interface layer 6 containing particulate alumina suspended in a glass matrix and particulate zinc oxide.

  Table 2 shows the composition of the interface layer 8 proposed in Patent Document 8. Alumina comprises 15 wt% to 35 wt%, zinc oxide 3 wt% to 6 wt%, and glass frit mixture 34 wt% to 53 wt%.

  Table 3 shows the composition of the glass frit mixture. Lead oxide is 50 wt% to 74 wt% (17 wt% to 39 wt% for the interface layer), Boron oxide is 10 wt% to 25 wt% (3 wt% to 13 wt%), and Silica is 8 wt% to 26 wt% (same as above). 2 wt% to 14 wt%).

  Table 4 is a summary of the results of Tables 2 and 3. The composition of the interface layer 8 disclosed in Patent Document 8 is 15 to 41.4 wt% for alumina, 3 to 6 wt% for zinc oxide, 17 to 39.2 wt% for lead oxide, and 2.7 to 13 wt.% For boron oxide. It can be recalculated as 8 wt%, silica 0-6.4 wt%, titania etc. 0-1.6 wt%. Patent Document 8 proposes a higher gauge factor by including both alumina and zinc oxide in the glass.

  However, although it is possible to stabilize the characteristics of the resistor in the device disclosed in Patent Document 8, it is not considered in the problem of multi-layered wiring. Therefore, when multi-layered wiring is described with reference to FIG. 20 described later. This causes problems.

  Amorphous glass has been proposed for matching with such strain sensitive resistors. In Patent Document 9, it is proposed to use lead borosilicate glass instead of crystallized glass for the production of a single-layer wiring load sensor. Patent Document 9 proposes that an insulating layer made of lead borosilicate glass having a film thickness of 30 μm or less per layer is formed into at least three multilayer insulating layers by repeating printing, baking, printing, and baking. . However, in the case of lead borosilicate glass, as described in Patent Document 9, it is known that cracking tends to occur inside when fired at a thickness of about 40 μm.

  Patent Document 10 proposes a load sensor having a single wiring layer. However, in the structure of a conventional load sensor having such a single-layer wiring, since there is no degree of freedom in wiring routing design, it is difficult to mount semiconductors and chip components on the load sensor substrate at high density.

  Furthermore, as proposed in Patent Document 11 and the like, it is desired to mount the load sensor on a passenger car in a state where various components are mounted on the substrate of the load sensor. However, since the metal elastic body is directly fixed to the chassis of the passenger car, noise from the passenger car may be easily picked up. In order to overcome such problems, for example, Patent Document 12 proposes to multilayer the wiring in order to improve the EMI (electromagnetic field interference) characteristics of the on-vehicle load sensor.

  FIG. 19 is a cross-sectional view showing an example of a load sensor in which wirings are multilayered. In FIG. 19, a plurality of lower glasses 7a, 7b and 7c are formed on the metal elastic body 1, and the upper electrodes 8a, 8b and 8c are further formed thereon so that the internal electrode 9a further covers the internal electrodes 9a. A plurality of layers 8c are formed. The wiring 10a and the strain sensitive resistor 12 are formed on the upper glass 8c, and these are protected by the protective layers 11a, 11b and 11c. In addition to EMI countermeasures, the load sensor wiring can be reduced in size and functionality (for example, mounting semiconductor chips and various chip components on the load sensor substrate at high density). It is indispensable, and conventional load sensors with a single wiring layer cannot fully meet these requirements. When wiring is multi-layered in this way, as proposed in Patent Document 12, it is conceivable to repeat multi-layer printing by repeating printing and firing one by one, but the manufacturing cost becomes a problem. Therefore, simultaneous firing of a plurality of layers is desired.

  FIG. 20 is a diagram for explaining a problem in the case where load sensors are multi-layered. The same reference numerals are used for the same parts as in FIG. 19, and reference numeral 11d is a protective layer. It is a figure explaining a mode that the crack 13 generate | occur | produces in an interlayer and a layer especially. According to experiments by the inventors, it has been found that such cracks 13 are likely to occur when different glasses are multilayered.

  For example, when glasses having different coefficients of thermal expansion or amorphous glass and crystalline glass are combined in a sandwich shape, cracks are likely to occur between the layers. It was also found that such problems are likely to occur when a plurality of glass layers are fired simultaneously. As described above, when the glass layers are multi-layered, problems are unlikely to occur when firing one by one individually, but the manufacturing cost is greatly increased.

On the other hand, even if it is attempted to reduce costs by firing multiple layers at once, various problems such as ink leveling during multilayer printing and thermal expansion matching during multilayer firing have occurred in the past. However, the glass materials and construction methods for load sensors could not overcome these problems.
Japanese Patent Laid-Open No. 5-45238 JP-A-6-137805 Japanese Patent No. 2979757 JP-A-8-178768 JP-A-8-304200 JP-A-8-145814 Japanese Patent Laid-Open No. 9-8325 Japanese Patent No. 3010166 JP-A-9-243472 JP 11-326090 A JP 11-351952 A JP 2003-97997 A

  Conventionally, the load sensor that has been proposed has a single layer of wiring, so that it has not been able to sufficiently cope with the high performance required by the market. Therefore, it is desired that the load sensor wiring is multi-layered. However, with the conventional materials, it is difficult to increase the number of layers of the wiring itself, and as the number of layers further increases, the manufacturing cost has been a problem.

  The present invention provides a load sensor that can easily realize multi-layered wiring of a load sensor, and can further reduce the manufacturing cost thereof, and a manufacturing method thereof.

  In addition, by using composite glass as a glass material in contact with at least the metal substrate, it is possible to prevent cracks during refiring and improve mechanical strength. Further, by forming amorphous glass at least under the strain-sensitive resistor, matching with the strain-sensitive resistor, which has been a problem with conventional crystallized glass, is made possible. Even for metal elastic bodies with different thermal expansion coefficients, by adding ceramic powder added to crystallized glass and making it composite glass, crystallized glass and metal substrate, or crystallized glass and amorphous glass The matching can be improved by simultaneous firing with. Thereby, the application range of the product can be expanded, and the leveling property of the printed coating film at the time of multilayer printing by screen printing can be improved by making various pastes curable. For this reason, the pinhole of a coating film can be reduced, the yield of a product can be increased and the firing cost can be reduced, and an object is to provide a high-performance, low-cost load sensor at low cost.

  In order to achieve the above object, an example of the present invention has the following configuration.

The present invention relates to a metal elastic body containing 1 wt% or more of aluminum, 30 to 55 wt% of MgO having an average particle size of 0.1 μm or more and less than 10 μm formed on the metal elastic body, and 5 of B ₂ O _3. Crystalline glass having a range of ˜30 wt%, SiO _{2 of 10} to 25 wt%, BaO of 5 to 25 wt%, alumina of 1 to 30 wt%, CaO of 0 to 6 wt%, ZrO ₂ and SnO ₂ of 0 to 5 wt% The crystalline glass in a weight ratio obtained by simultaneously firing powder and ceramic powder that is one or more of alumina, magnesia, zirconia, silica, calcium oxide, titanium oxide, and spinel having an average particle size of 0.05 μm or more and less than 5 μm powder: the ceramic powder = 95: 5-60: and composite crystalline glass layer is 40, SiO ₂ is formed on the composite crystalline glass layer 40-8 An amorphous glass layer of 0 wt%, CaO 5-15 wt%, PbO 3-15 wt%, alumina 1-10 wt%, ZrO ₂ 2-10 wt%, and formed on the amorphous glass layer A load sensor comprising a wiring pattern and a strain sensitive resistor formed so as to be connected to a part of the wiring pattern, wherein the composite crystalline glass layer and the amorphous glass layer are: It is a load sensor characterized by being fired simultaneously on the metal elastic body , and by using a composite crystalline glass, it is possible to prevent cracking at the time of refiring and improve mechanical strength. Plan.

  The present invention directly prints on a metal elastic body a composite glass paste in which a glass powder having an average particle size of 0.1 μm or more and less than 10 μm and a ceramic powder having an average particle size of 0.05 μm or more and less than 5 μm are dispersed in a resin solution. After that, a composite glass layer is formed by drying and firing in an oxidizing atmosphere of 800 ° C. or higher and lower than 900 ° C., and a wiring pattern and a part of the wiring pattern are connected to the composite glass layer. This is a method of manufacturing a load sensor that forms a strain sensitive resistor. Addition of ceramic powder added to crystallized glass to a metal elastic body with a different thermal expansion coefficient results in composite vitrification. Can improve matching in co-firing of crystallized glass and metal substrate, or crystallized glass and amorphous glass, and manufacture of load sensor It enhanced the quality, function and effect for lowering the cost of manufacturing the load sensor can be obtained.

  The load sensor and its manufacturing method according to the present invention optimizes the glass material used for manufacturing the load sensor, enables multi-layer firing, and further expands the selection range of the metal elastic body. Useful for air bags and the like.

(Embodiment 1)
A load sensor having a multilayered wiring will be described with reference to FIGS. FIG. 1 is a cross-sectional view of a load sensor having a single wiring layer. FIG. 2 is a cross-sectional view of a load sensor having a multilayered wiring. In FIG. 1, reference numeral 1 denotes a metal elastic body, on which a crystallized glass 14a is formed, and a wiring 16a and a strain sensitive resistor 17 are formed thereon via an amorphous glass 15a. A protective layer 18a is formed so as to embed the strain resistor 17 therein.

  In FIG. 2, reference numeral 1 denotes a metal elastic body, on which a crystallized glass 14b is formed with an internal electrode 19 built therein, and a wiring 16b, a strain sensitive film, and the like are further formed on the glass via an amorphous glass 15b. Each of the resistors 17 is formed, and a protective layer 18b is formed so as to embed at least the strain sensitive resistor 17 therein.

  2 is different from FIG. 1 in that the internal electrode 19 is built in the crystallized glass.

  As shown in FIG. 2, in the present invention, the internal electrode 19 can be embedded in the crystallized glass 14b in accordance with the design items and specifications, and the EMI (electromagnetic characteristics or noise resistance) of the load sensor can be improved. Furthermore, it is possible to prevent the glass layer from redissolving and generating cracks when embedding the internal electrode or firing the wiring or the strain sensitive resistor.

  Further, as shown in FIG. 1, by interposing the amorphous glass 15a between the crystallized glass 14a and the strain sensitive resistor 17, the interaction between the crystallized glass 14a and the strain sensitive resistor 17 can be prevented. .

  FIG. 3 is a cross-sectional view for explaining how the load sensor of the present invention is manufactured. First, as preparation, referring to JP-A-6-137805, a predetermined glass raw material was mixed, melted in a platinum crucible, rapidly cooled, and pulverized with a ball mill to prepare a predetermined crystallized glass powder. Next, a resin (ethyl cellulose) dissolved in a solvent (α terpineol) was added to the crystallized glass powder, and the mixture was kneaded using a ceramic three roll to obtain a crystallized glass paste.

  Next, as shown in FIG. 3A, the crystallized glass paste 20a was printed on the metal elastic body 1 in a predetermined pattern. And this was dried and baked at 850 degreeC, and as shown in FIG.3 (B), the 1st layer crystallized glass 21a was formed on the metal elastic body 1. As shown in FIG. Next, as shown in FIG. 3C, the crystallized glass paste 20b was printed in a predetermined shape on the crystallized glass 21a. Then, this was dried and fired at 850 ° C. to form crystallized glasses 21 a and 21 b on the metal elastic body 1 as shown in FIG. Thus, by repeating printing and baking of the crystallized glass layer, it is possible to prevent dust and bubbles in the paste or pinholes that are likely to occur during printing.

  Next, as shown in FIG. 3E, the electrode paste 22 is printed in a predetermined pattern on the crystallized glasses 21a and 21b. And this was dried and baked, and the internal electrode 23 was formed on the crystallized glass 21a, 21b as shown in FIG. 3 (F). Thereafter, as shown in FIG. 3 (G), the crystallized glass paste 20c is printed in a predetermined pattern so as to cover the internal electrode 23, dried, and further baked at 850 ° C. to obtain FIG. 3 (H). Crystallized glass 21c was formed as shown.

  Next, as shown in FIG. 3 (I), an amorphous glass paste 24 is printed in a predetermined pattern on the crystallized glass 21c, and this is dried and fired. As shown in FIG. A glass 25 was formed. Next, a wiring paste (not shown) was printed on the amorphous glass 25 in a predetermined shape, dried, and baked at 850 ° C. to form the wiring 26 as shown in FIG. Next, a resistor paste (not shown) was printed in a predetermined shape between the wirings 26 and baked at 850 ° C. to form a strain sensitive resistor 27 as shown in FIG. Then, a protective paste (not shown) was printed on the strain sensitive resistor 27, and this was repeatedly dried and fired a plurality of times to create a load sensor as shown in FIG.

  The characteristics of the load sensor created in this way were evaluated. As a result, all of the strain sensitivity, temperature characteristics, reliability, noise resistance (EMI), and electrostatic resistance (ESD) were satisfied, and the standard values for in-vehicle use were met. I was satisfied.

  In addition, there are two glass layers between the metal elastic body 1 and the internal electrode 23, and between the wiring 26 and the strain sensitive resistor 27 and the internal electrode 23. There was no need to have 3 layers each.

For comparison, an attempt was made to multilayer the load sensor wiring by using only amorphous glass paste as Conventional Example 1. First, using lead borosilicate (B ₂ O ₃ —SiO _w —PbO) glass paste proposed in Japanese Patent Laid-Open Nos. 9-243472 and 2003-97997, FIG. 1 and FIG. The structure shown was created.

  The (printing-drying-firing) process shown in FIGS. 3A to 3D was repeated three times to form three amorphous glass layers on the metal elastic body 1. Next, as shown in FIGS. 3E to 3F, an internal electrode 23 is formed, and the (printing-drying-firing) process shown in FIGS. Three amorphous glass layers were formed on the electrode. Then, wirings and strain sensitive resistors were formed by the steps of FIGS. 3K and 3L, and finally three protective layers were formed.

  When multi-layered using only such amorphous glass, the characteristics of the load sensor were evaluated, and there was no problem with strain sensitivity, temperature characteristics, reliability, noise resistance (EMI), and electrostatic resistance (ESD). We were able to satisfy the standard value for use. In addition, there was no insulation failure between the metal elastic body 1 and the internal electrode 23 and between the wiring 26 and the strain sensitive resistor 27 and the internal electrode 23 because there were three glass layers each. However, compared with Embodiment 1, the number of glass layers, the number of printing, and the number of firings are large, resulting in high costs.

  Therefore, in an attempt, the same process as in FIG. 1 was performed, and the glass layers between the metal elastic body 1 and the internal electrode 23 and between the wiring 26 and the strain sensitive resistor 27 and the internal electrode 23 were each made into two layers. . However, when such amorphous glass was used and each of the glass layers was composed of two layers, poor insulation occurred frequently and could not be put into practical use. As another attempt, when the thickness of the glass layer was increased to about 40 μm, cracks sometimes occurred inside. As described above, in the case of the conventional amorphous glass, a problem remains when the number of layers is inevitably reduced.

  For comparison, as Conventional Example 2, an attempt was made to multilayer the load sensor wiring using only the crystallized glass paste. As Conventional Example 2, the case where crystalline glass is used will be described.

  The inventors created a crystallized glass paste as a conventional crystallized glass based on the crystallized glass composition proposed in JP-A-6-137805. First, referring to JP-A-6-137805, a predetermined glass raw material was mixed, melted in a platinum crucible, rapidly cooled, and pulverized by a ball mill to prepare a predetermined crystallized glass powder.

  Next, a resin (ethyl cellulose) dissolved in a solvent (α terpineol) is added to the crystallized glass powder, and the mixture is kneaded using a ceramic three roll to form a paste. did. And the process of FIG. 3 (A)-(H) was performed. In the steps of FIGS. 3I to 3J, crystallized glass paste was used instead of amorphous glass. And the protective layer was finally formed through the process of FIG.3 (K)-(L). Thus, the sample of Conventional Example 2 was prepared. When the characteristics of Conventional Example 2 were measured, not only a predetermined sensitivity as a load sensor could not be obtained, but also the resistance value varied greatly and was not practical as a weight sensor. This is probably because the glass layer formed under the strain sensitive resistor was crystalline glass.

  Therefore, a metal alkoxide layer was formed on the crystallized glass layer with reference to JP-A-6-1378805. However, because the strength and film quality are weak, the specifications required for an on-vehicle load sensor could not be fully satisfied. As described above, when crystallized glass is used, even if there are two glass layers between the metal elastic body 1 and the internal electrode 23 and between the wiring 26 and the strain sensitive resistor 27 and the internal electrode 23, high yield is achieved. However, it was found that matching between the strain sensitive resistor and the resistor is difficult.

  For comparison, the composition proposed in Japanese Patent No. 3010166 (US Pat. No. 5,898,359) was examined as Conventional Example 3. First, a commercially available glass paste was printed on a metal elastic body having a thickness of 2 mm and baked at 850 ° C., which was equivalent to the crystalline glass 14a of FIG. On top of this, lead-alumina-borosilicate glass is used as a glass matrix, 15 to 35% of particulate alumina is added thereto, and 3 to 6% of particulate zinc oxide is added to the interface layer, which corresponds to amorphous glass 15a. Formed in the position. With respect to Conventional Example 3 thus prepared, the characteristics as a load sensor were measured, but the expected characteristics were not obtained. This seems to be because the resistor material proposed in Japanese Patent No. 3010166 was not used.

  Further, when the internal electrode is formed in the structure of the conventional example 3, the problem as shown in FIG. 20 may occur. That is, it has been found that the material of Conventional Example 3 has a limit to the multilayer wiring.

  The thickness of the crystallized glass is desirably 10 to 200 μm. When the thickness of the crystallized glass is less than 10 μm, an insulation failure may occur. Moreover, since the material cost of glass will increase when the thickness of crystallized glass exceeds 200 micrometers, it must admit the high cost.

  The thickness of the amorphous glass is preferably 5 to 100 μm. This is because when the thickness of the amorphous glass is less than 5 μm, the strain resistor is in contact with the crystallized glass through pinholes or the like of the amorphous glass, resulting in a disadvantage that the resistance value changes.

  The protective layer covering at least the strain sensitive resistor preferably has a thickness of 200 μm or less. When the thickness of the protective layer exceeds 200 μm, when the protective layer is formed in a predetermined pattern, the end of the pattern is sagged or smudged, and a land for mounting a semiconductor or a chip component (a component formed in the protective layer) A connection hole for mounting, in which the wiring is exposed, may be crushed.

  The overcoat material formed on the strain resistor is preferably a glass layer containing 60 wt% or more of PbO. In the case of a glass paste containing 60 wt% or more of PbO, the firing temperature is about 400 to 700 ° C., which is lower than 150 ° C. by 850 ° C., which is a general firing temperature for strain resistors. Strain resistors are less affected. It is also possible to use a resin for the overcoat. When using resin for the overcoat, silicon resin or epoxy resin is used, and if necessary, glass or ceramic powder is added as a filler to about 5 to 50 wt% to increase the mechanical strength and load. When attaching the sensor to the car, it becomes difficult to be damaged by a tool such as a spanner. When the amount of filler added is less than 5%, the effect of increasing the strength of the resin coating film is not sufficient.

  In addition, when the filler is added in an amount of more than 50 wt%, there is a disadvantage that the overcoat becomes distorted. In the case of an overcoat using a resin, the thickness is desirably 10 μm or more and 200 μm or less. Further, it is desirable that the protective layer covering at least the strain sensitive resistor has a thickness of 200 μm or less. When the thickness of the protective layer exceeds 200 μm, when the protective layer is formed in a predetermined pattern, the end of the pattern is sagged or smudged, and a land for mounting a semiconductor or a chip component (a component formed in the protective layer) This is because a connection hole for mounting, in which the wiring is exposed, is crushed.

  It is also effective to make the overcoat into a plurality of layers. Furthermore, it is also possible to form a second overcoat with a thickness of 10 μm or more and 200 μm or less with a material made of glass or ceramic, and with a material mainly composed of resin on the overcoat. is there. Thus, the reliability of the strain sensitive resistor and the wiring can be improved by multilayering the overcoat using glass or resin. Moreover, a glass material containing PbO or a commercially available resin material can be used for such a member.

(Embodiment 2) Composition of amorphous glass In Embodiment 2, an optimization example of amorphous glass that can cope with multilayer wiring and co-firing with crystalline glass will be described.

First, as shown in composition numbers 1 to 5 in Table 5, SiO ₂ was reduced, and the composition was changed for CaO, ZrO ₂ , Al ₂ O ₃ , and B ₂ O ₃ , but good results were not obtained. Therefore, as shown in composition numbers 6 to 10 in Table 5, when the amount of SiO ₂ is 45 to 65 wt% and the composition is changed for CaO, ZrO ₂ , PbO, Al ₂ O ₃ , and B ₂ O ₃ , good results are obtained. was gotten. On the other hand, as shown in composition numbers 11 to 14 in Table 5, when the amount of SiO ₂ was set to 70 wt% or more, good results were not obtained.

Therefore, as shown in Table 6, the minimum ratio of SiO ₂ is 40 wt% (corresponding to composition numbers 15 to 25) and the maximum ratio is 80 wt% (corresponding to composition numbers 26 to 36), and the ratio of each composition is changed. I tried.

  As a result of further experiments on glass compositions other than this, the best results were obtained with the composition ratios shown in Table 7. Further, in such experiments, it was found that the proportion of lead oxide contained in the amorphous glass diffuses with the strain sensitive resistor and affects the resistance value and the like. This is because the resistor material used by the inventors contained lead oxide, but if a strain sensitive resistor that does not use lead oxide was developed in the future, PbO could be reduced to zero from Table 7. Seem.

These results, as the amorphous glass, as shown in Table 7, SiO ₂ is 40 to 80 wt%, alumina 110 wt.%, CaO is 5 to 15 wt%, PbO is 3 to 15%, ZrO ₂ Was found to be optimal. Further, it was confirmed that by adding an appropriate amount of B ₂ O ₃ , the sinterability of the glass and the matching with the electrode material (adhesion strength, prevention of cracks during simultaneous firing, etc.) can be improved.

  FIG. 4 is a cross-sectional view showing a state in which the amorphous glass thus prepared and a conventional crystallized glass are combined and fired simultaneously. In FIG. 4A, 28 is a crystallized glass paste and 29 is an amorphous glass paste. First, the conventional crystallized glass paste 28 and the amorphous glass paste 29 of the second embodiment are multilayered as shown in FIG. 4 (A), and this is meshed at 850 ° C. (total firing time 1 hour). When co-fired in a furnace, as shown in FIG. 4B, crystalline glass 30a and amorphous glass 31a were obtained, and no abnormality occurred in the internal structure.

For comparison, as a conventional amorphous glass, lead borosilicate (B ₂ O ₃ —SiO ₂ —PbO) glass proposed in Japanese Patent Laid-Open Nos. 9-243472 and 2003-97997 is disclosed. A conventional crystallized glass was simultaneously fired. First, as shown in FIG. 4 (A), a commercially available crystallized glass 29 is printed on the metal elastic body 1 and proposed on Japanese Patent Laid-Open Nos. 9-243472 and 2003-97997. A lead borosilicate system (B ₂ O ₃ —SiO _w —PbO) was printed, fired at 850 ° C. at the same time, and the cross section was observed. The result shown in FIG. 4C was obtained.

  In FIG. 4C, 13 is a crack. As shown in FIG. 4C, the crack 13 was easily generated at the interface between the amorphous glass 31b and the crystalline glass 30b or inside the amorphous glass 31b. Thus, it has been found that it is difficult to fire the amorphous glass simultaneously in the case of the conventional amorphous glass.

  On the other hand, a load sensor was created by combining the amorphous glass proposed in the present invention with a commercially available crystallized glass, and the yield strength was measured using a self-made automobile impact impact device. I found out. In particular, it has also been found that the use of crystalline glass for adhesion to the metal elastic body improves the proof stress of the load sensor itself compared to the case of using only amorphous glass. This seems to be because crystallized glass has higher strength than amorphous glass, and even if cracks occur, it is difficult to propagate like amorphous glass.

  The working temperature or softening point of the amorphous glass is desirably 500 ° C. or higher. When the temperature is less than 500 ° C., when the strain sensitive resistor is baked at a high temperature such as 850 ° C., it may react with the strain sensitive resistor to change the characteristics. The working temperature of the amorphous glass is desirably 700 to 900 ° C. In the case of an amorphous glass having an operating temperature of less than 700 ° C., when the strain sensitive resistor is baked at a high temperature such as 850 ° C., it may react with the strain sensitive resistor and change its characteristics. In the case of an amorphous glass having a working temperature exceeding 900 ° C., it is necessary to use an expensive material excellent in heat resistance for the metal elastic body.

(Embodiment 3) Optimization of crystalline glass In Embodiment 3, even when combined with a commercially available amorphous glass (for example, lead borosilicate B ₂ O ₃ —SiO ₂ —PbO), simultaneous firing is possible. Optimized for crystallized glass.

Table 8 shows examples of optimization of the crystalline glass composition. Table 8 shows the ratio of MgO. As a result, the proportion of MgO has a correlation with crystallinity, and it was found that 30 to 55 wt% is good. Therefore, as shown in Table 9, Table 10, and Table 11, optimization was performed for SiO ₂ , B ₂ O ₃ , BaO, ZrO ₂ , Al ₂ O _{3 and the} like.

As a result, it was found that the addition of SiO ₂ or B ₂ O ₃ improves the sinterability, and the addition of a certain amount of alumina improves the characteristics. The addition amount of CaO was desired to be 0 to 6 wt%. If the added amount of CaO increases, the surface becomes rough, which may cause defects in the screen printing process of other pastes. Although not shown in the above table, when SnO ₂ was also examined, a good result was obtained when the addition amount was 0 to 5 wt%.

As described above, 30～55Wt% is MgO as crystalline glass, alumina 1-30 wt%, SiO ₂ is 10 to 25 wt%, BaO is 5~25wt%, B ₂ O ₃ is 5-30 wt%, CaO Is preferably 0 to 6 wt%, and ZrO ₂ and SnO ₂ are preferably 0 to 5 wt%. As other elements, ZrO ₂ and SnO ₂ can contain 0.1 to 5 wt%.

  PbO and ZnO are not practically harmful to the starting material, powder, or the amount that is mixed in as a small amount of impurities in the course of inking, but ideally not included. It should be noted that if it diffuses somewhat in the firing step, there is no practical problem, but if such elements are contained in the glass by 0.1 wt% or more, the reliability is adversely affected. For example, PbO and ZnO are water-soluble. Therefore, it can be predicted that reliability will be reduced.

  Using the optimized crystallized glass in this way, a load sensor was prepared, and when the yield strength was measured using a self-made automobile impact impact device, a practically sufficient yield strength was obtained. However, when a strain sensitive resistor is directly formed on the crystallized glass, the characteristics of the resistor (temperature characteristics, stability of resistance value, etc.) may be affected.

  In that case, as shown in FIGS. 1 and 2, the electrical characteristics as a load sensor can be stabilized by forming amorphous glass on such crystallized glass. The reason why the crystallized glass of the present invention has higher proof strength than a load sensor made using only ordinary amorphous glass is that the grains constituting the crystallized glass are formed even if cracks occur inside. It seems to stop at the world.

  As described above, in the case of the crystallized glass of Embodiment 3, matching with amorphous glass is high, and co-firing with commercially available amorphous glass is possible, and the cost of the load sensor can be reduced. Met.

Further, the crystallized glass thus obtained is made into a paste, and as a conventional amorphous glass, lead borosilicate (B ₂ O ₃ —) proposed in Japanese Patent Laid-Open Nos. 9-243472 and 2003-97997 is disclosed. (SiO _w —PbO) glass was selected and a co-firing experiment was conducted. As shown in FIG. 4, printed in a predetermined shape on a metal elastic body, and a commercially available amorphous glass paste was printed thereon and fired at the same time. As shown in FIG. No abnormality occurred in the glassy and crystalline glasses.

  For comparison, the crystallized glass proposed in Table 1 was selected as the crystallized glass of the conventional example, and a co-firing experiment as shown in FIG. 4 was similarly conducted. First, as shown in FIG. 4 (A), a conventional crystallized glass paste is printed on a metal substrate, a commercially available amorphous glass paste is printed thereon, and these are simultaneously fired at 850 ° C. As shown in FIG. 4C, it was found that cracks occurred and it was difficult to fire the crystalline glass and the amorphous glass at the same time.

Even if the crack 13 did not occur, the performance as a highly durable load sensor for in-vehicle use was not practical with the conventional amorphous glass. This was thought to be due to the fact that this glass contained a certain amount or more of alumina and a lower proportion of BaO and B ₂ O ₃ than conventional crystallized glass (for example, Table 1). It has also been found that the crystallized glass proposed in the third embodiment has a characteristic that it is easily sintered at the same time as the ceramic powder and made into a composite glass, as will be described later.

(Embodiment 4) Print lamination (curing)
In the fourth embodiment, a case of performing batch firing to reduce the firing cost of a multi-layer printed glass layer will be described. As described in Embodiment 1, when a glass layer is multi-layered, if printing, drying, and firing are repeated for each layer, the number of firings increases as the number of layers increases, leading to an increase in product cost. . However, when the glass is multi-layered, printing and drying are repeated, and the glass layer is multi-layered. The problem that it is easy to occur arises.

  In the fourth embodiment, a method for solving such a problem by adopting a curable paste as the crystallized glass paste will be described.

  First, for comparison, pinholes and thickness unevenness that are likely to occur when a crystalline glass paste is multilayered will be described with reference to FIG. 5 as a conventional example. FIG. 5 is a cross-sectional view illustrating a state in which a crystalline glass paste is multilayer printed on a metal elastic body. As shown in FIG. 5A, the crystallized glass paste 32a printed on the metal elastic body 1 is leveled (surface smoothed) after a predetermined time. This is because the solvent component contained in the crystallized glass paste 32 a does not penetrate into the metal elastic body 1. In FIG. 5B, reference numeral 33a denotes a crystalline glass dry film, which is obtained by volatilizing and drying the solvent after the crystalline glass paste 32a is leveled. FIG. 5C shows a state in which the crystallized glass paste 32b is printed on the crystalline glass dry film 33a.

  The arrow 34 shows a state in which the solvent component contained in the crystalline glass paste 32b is infiltrated into the crystalline glass dry film 33a serving as a base. As a result of the solvent component infiltrated into the base, the solvent component contained in the crystalline glass paste 32b decreases, the viscosity of the crystalline glass paste 32b increases, the paste fluidity decreases, and the surface has irregularities 35 and pinholes. (Not shown) occurs. As a result, as shown in FIG. 5D, irregularities 35 remain on the surface of the crystalline glass dry film 33b. Such irregularities 35 do not disappear even when fired. This is because crystalline glass has less fluidity during firing than amorphous glass.

  As a result, as shown in FIG. 5E, when the internal electrode 36a is formed on such a structure, an insulation failure is likely to occur between the internal electrode 36a and the metal elastic body 1.

  Therefore, the inventors repeated various experiments in order to solve these problems. As a result, it has been found that such a problem is mainly caused by the solvent in the paste soaking into the base as shown in FIG. Therefore, a curable glass paste was developed.

  FIG. 6 shows a state of multilayer printing using a curable glass paste developed by the inventors. 6 (A) to 6 (E) are cross-sectional views for explaining how glass layers are formed in multiple layers on a metal substrate by printing using the curable glass paste of the present invention. FIG. 6A shows a state in which a curable glass paste 37a is printed on the metal elastic body 1 by screen printing or the like. In this state, since the solvent in the curable glass paste 37a does not penetrate into the metal elastic body 1, the metal elastic body 1 is immediately leveled.

  Thereafter, the curable glass paste 37a is cured or insolubilized by heat, light, electron beam or the like to form a cured glass paste 38a shown in FIG. 6B. The surface of the cured glass paste 38a is sufficiently leveled, and the solvent is hardly infiltrated into the cured glass paste.

  Next, as shown in FIG. 5C, a curable glass paste 37b is printed in order to form a second glass layer. In FIG. 6C, since the cured glass paste 38a has been cured, the solvent of the curable glass paste 37b printed thereon hardly penetrates even if it soaks as indicated by the arrow 34. Therefore, the curable glass paste 37b is leveled immediately without generating the unevenness 35, the thickness unevenness, and the pinhole as shown in FIG.

  By curing the leveled curable glass paste 37b, as shown in FIG. 6D, the surface of the cured glass paste 38b forming the second glass layer also has irregularities, uneven thickness, and pinholes. A smooth surface with no unevenness.

  As a result, as shown in FIG. 6E, even if the internal electrode 36b is formed, insulation failure is unlikely to occur between the internal electrode 36b and the metal elastic body 1. In this way, when printing is multilayered using a curable glass paste, it is possible to suppress the occurrence of defects such as irregularities, thickness unevenness, pinholes, etc. in the glass coating film before firing. Increased and reduced manufacturing costs.

  On the other hand, in the case of the conventional glass paste, since the solvent component in the paste soaks into the base, the viscosity suddenly increases immediately after printing and no leveling occurs, resulting in large unevenness on the surface. In the case of screen printing, this unevenness is more likely to occur as the mesh of the screen plate becomes rougher.

  Therefore, even if the thickness of the glass layer is increased and the mesh of the screen plate is increased, unevenness is more likely to occur even if the film thickness is increased, which may reduce the product yield. Therefore, in the past, multi-layer printing was performed using finer meshes (more unevenness is less likely to occur), but the thinner the finer the mesh, the thinner the film thickness. It was necessary to increase the number of man-hours.

  FIG. 7 explains the state of such leveling. In FIG. 7, the X-axis is the time (unit: minutes) after printing, and corresponds to FIGS. 5C and 5D and FIGS. 6C and 6D. The Y axis in FIG. 7 is obtained by measuring the surface irregularities (see C and D in FIG. 5) of the paste surface from immediately after printing until leveling with a non-contact laser surface roughness meter. In FIG. 7, a (corresponding to x mark) is a curable glass paste prepared using crystallized glass powder, and it can be seen that the unevenness is reduced immediately after printing and leveling occurs. Further, in FIG. 7, c (corresponding to a circle) is a curable glass paste prepared using amorphous glass powder, and it can be seen that unevenness is reduced immediately after printing and leveling is performed. On the other hand, b (corresponding to the mark ●) in FIG. 7 is a case of a conventional solvent-dried glass paste, and its unevenness is difficult to reduce after printing.

  In the fourth embodiment, the crystallized glass paste is curable. However, the amorphous glass paste, the internal electric paste, and the wiring paste are also made into such a curable paste, thereby improving the leveling property when multilayer printing is performed. This can greatly improve the cost of the load sensor and increase the yield. Further, in the preparation of such a curable paste, it is possible to use conventionally proposed amorphous glass powder or crystallized glass, and the amorphous glass or crystalline glass proposed in the present invention, or described later. It is easy to apply to composite glass.

(Embodiment 5) Contents of curable insulating paste Next, members used for the curable paste will be described in detail. First, the average particle size of 0.1 to 10 μm is optimal as the glass powder used in the curable glass paste. Glass powder having an average particle size of less than 0.1 μm is expensive. Further, the glass powder having an average particle size exceeding 10 μm may contain a large glass powder of 20 to 30 μm. The glass powder contained in the curable glass paste is preferably in the range of 60 to 80 wt%. When it is less than 60 wt%, when cured after baking, the resulting glass becomes porous (a gap is generated inside), or the film thickness shrinkage during firing is too large, and multiple layers are fired simultaneously. Cracks may occur. Moreover, when glass powder contains 80 wt% or more in the glass paste, the flow component (resin or solvent) in the paste is reduced, and the fluidity of the paste is lowered, which may make it difficult to level. Further, the resin content in the curable glass paste is preferably 1 to 20 wt%. When the resin is less than 1 wt%, the cured coating film has low strength and may be peeled off or removed when multilayer printing is performed. When the resin content exceeds 20 wt%, there is no problem in curing the paste, but when the cured paste is baked, the amount of decomposition of the resin component becomes large, for example, a large amount of smoke is generated in a baking furnace. At the same time, the coating film may swell or crack. The amount of solvent in the paste is preferably 10 to 40 wt%. When the amount of the solvent is less than 10 wt%, the viscosity of the ink is increased, which may affect continuous printability and leveling properties. When the solvent amount is 40 wt% or more, the drying time of the ink becomes long. The curable paste can be made solvent-free, but it is desirable that the solvent is contained in an amount of 10 wt% or more. In the absence of any solvent, a considerable amount of resin must be added to reduce the paste viscosity to that required for screen printing, and this resin may cause cracks during firing. Because. The solvent added to the curable ink preferably has a boiling point of 150 ° C. or higher. A solvent having a boiling point of less than 150 ° C. may be too fast to dry on the printing press, and may affect the continuous printability and the on-press stability of the ink.

  In order to create such a curable paste, a curable resin is selected as the binder component (resin component) of the ink. As the curable resin, light (UV), electron beam (EB), and thermosetting resin are commercially available. In the case of the present invention, since the paste is fired, it is necessary to select a curable resin with a low ash content. Moreover, even if some ash (for example, about 1%) remains, since such ash is often an oxide, in the case of the present invention, it becomes a part of the glass component and often does not affect the quality. .

  The resin component contained in the curable glass paste is desirably 1 to 20 wt%. When the resin component is less than 1 wt%, the physical strength of the cured coating film (particularly before firing) may be low. When the film strength after curing is low in this way, when printing the next glass paste etc. on it by screen printing etc., the coating film may be peeled off or scratched to reduce the product yield There is.

  In particular, when 5 to 100 wt% of the resin in the curable insulating paste is a curable resin, the effect of preventing the penetration of the paste solvent during multilayer printing can be effectively exhibited. In this way, by adding a non-curable resin in the paste in advance, the production of the insulating paste can be facilitated, and the insulating paste is cured on the screen printing machine and cannot be removed later. . For example, by setting 50% of the resin component in the cured insulating paste to thermosetting xylene resin and the remaining 50% to ethyl cellulose resin soluble in the solvent, the ink should be cured on the screen plate. However (or even when it is cured), it can be easily wiped off with a solvent by taking advantage of the solvent solubility of ethyl cellulose. Therefore, since no foreign matter remains on the screen plate, stable printing (or unattended printing) can be performed for a long time. As the curable resin, a resin curable by one or more of heat, light, and electron beam irradiation can be used.

  In view of productivity, a one-component curable resin is more desirable. In the two-component curable resin, curing of the resin starts by adding a reaction initiator to the base resin. However, when an insulating paste is made of such a two-component resin, it is necessary to add and mix a reaction initiator immediately before use, which complicates the process.

  On the other hand, the one-component curable resin is easy to use as the curable resin of the present invention because it is cured by the above method without adding a reaction initiator. In addition, these resins dissolve well in high-boiling solvents widely used for screen printing such as α-terpineol, butyl carbitol, isophorone, etc., so that the viscosity can be easily adjusted. Even when the screen plate is washed after printing, it can be easily cleaned with butyl acetate or the like. As such a curable resin, an epoxy resin and a phenol resin are easy to use. For example, a phenol resin or an epoxy resin can be used. If a two-component curable resin is used for such a purpose, it is complicated to make a paste, and the pot life of the paste after adding a reaction initiator is shortened. In the case of a one-component curable resin, it is not necessary to add a reaction initiator and it is easy to make a paste. For example, a one-component resin typified by xylenephenol formaldehyde resin is easily soluble in solvents such as α-terpineol, isophorone, and butyl acetate, so that it can be easily converted into an ink and easily washed after screen printing.

  When manufacturing glass paste and electrode paste using such a curable resin, it is desirable that the organic solvent added to the paste has a boiling point of 150 ° C. or higher. Since organic solvents having a boiling point of less than 150 ° C. tend to spontaneously evaporate during the printing process, the viscosity of the paste may change during the working process, which may cause pinholes on the printing surface or become difficult to level.

  The resin component in the paste need not be all thermosetting resin. What is necessary is just to make 5-100 wt% in the resinous part contained in a paste into curable resin. By adding a non-curable resin such as ethyl cellulose resin, butyral resin, and acrylic resin, which are widely known as binder resins, even if the ink is cured or gelled on the screen plate, ethyl cellulose It can be easily wiped by taking advantage of solvent solubility.

  The organic solvent contained in the curable paste preferably has a boiling point of 150 ° C. or higher. By including an organic solvent in the curable paste in this way, viscosity adjustment and solid content adjustment of the curable paste are facilitated, so that production quality can be stabilized throughout the year. In the case of an organic solvent having a boiling point of less than 150 ° C., since it spontaneously evaporates during the printing process, the paste viscosity may change during the work process, and the printed coating may vary in thickness. Moreover, it is desirable that the organic solvent contains 10 to 40 wt% of the solvent in the curable paste. When the solvent in the curable paste is less than 10 wt%, the viscosity and solid content adjustment range by the solvent becomes narrow. When the solvent is contained in an amount of 40 wt% or more, the change in the film thickness due to drying of the ink becomes large.

  The viscosity of the curable paste is preferably 10 to 10000 poise from the viewpoint of printability and leveling (the rotation speed in the case of a rotary viscometer is in the range of 10 to 100 rpm, or in the range of 1 to 100 / s at the shear rate). . Such a curable paste preferably has a thixotropy (rheological phenomenon in which the viscosity changes at a shear rate) to some extent. For this reason, the viscosity ratio of the paste (for example, the rotational speed of 10 rpm and 50 rpm, or the shift speed of 1 / s and 5 / s) is 1.0 or more, preferably 1.2 or more. If the viscosity ratio exceeds 5, the thixotropy becomes too high and the time until leveling becomes long, which is not practical.

  The curable glass paste thus prepared has no difference in the glass itself after sintering (after baking at 800 to 900 ° C.) as compared with a solvent-diluted glass paste using a conventional resin such as ethyl cellulose. A clear difference in the firing profile is observed. Therefore, the inventors made a curable glass paste using a phenol resin and a conventional glass paste using an ethyl cellulose resin using a previously proposed crystallized glass powder, and this was added to a predetermined amount of platinum. The sample was set in a container, and thermal analysis (called TG / DTA) from room temperature to 900 ° C. was performed. As a result, the features of each other appeared especially in the DTA graph (not shown). In the case of the conventional type, an endothermic peak in which the solvent component is evaporated at 150 to 200 ° C. was observed.

  Thereafter, an exothermic peak in which the ethylcellulose resin was thermally decomposed around 300 ° C. was observed, and thereafter, there was no particular change, and an exothermic peak presumably due to glass crystallization was observed at around 800 ° C.

  On the other hand, in the case of the curable type, after an endothermic peak in which the solvent component evaporates at 150 to 200 ° C. is observed, a broad peak that seems to be generated by the resin curing at about 300 to 450 ° C. is observed. At 500 to 550 ° C., an exothermic peak was observed at which the cured resin was considered to be thermally decomposed, and thereafter, there was no particular change, and an exothermic peak presumably due to glass crystallization was observed at around 800 ° C.

  From the results of the thermal analysis described above, it can be seen that in the case of a curable resin, the resin is cured, whereby the heat resistance is increased and the thermal decomposition temperature of the paste is increased. Therefore, by using the glass paste used here, as shown in FIG. 10A (note that FIG. 10 will be described in a later embodiment), the inside in which no sintering inhibitor is added. When an electrode was embedded and co-fired with this internal electrode, in the case of a curable glass paste, as shown in FIG. 10B, there was a particular problem (in spite of the internal electrode to which no sintering inhibitor was added). In some cases, multiple layers could be fired simultaneously.

  On the other hand, in the case of the conventional glass paste, various cracks occurred as shown in FIG. From the above experimental results and thermal analysis results, it was found that the use of a curable resin has the effect of preventing cracks when the internal electrode is fired and contracted (considered from around 500 ° C.). . In other words, in the conventional glass paste, the glass powder has a low binding force (because the resin component is not thermally decomposed) at the temperature at which the shrinkage of the internal electrodes begins. It is considered that this is a cause of cracks.

  On the other hand, in the case of curable glass paste, the binding power between the glass powders is strong at the temperature at which the internal electrodes begin to shrink (the resin component whose heat resistance has been increased by being cured still remains). It seems that the glass layer can be prevented from cracking due to shrinkage.

  Thus, it can be seen that by using a curable resin, it is possible to suppress sintering to some extent even with a conventional glass composition. By using such sintering suppression technology using a curable resin, not only enables simultaneous firing of glass and metal, but also between different types of glass (for example, glasses with different softening points and transition points, Even in the case of crystalline glass and amorphous glass), an effect of facilitating simultaneous firing in a multilayered state can be obtained. In this way, by making the resin curable, increasing the heat-resistant temperature of the resin and preventing cracking during sintering, since such resin components are not burned out and remain in the glass and electrodes after firing, It does not affect the characteristics. However, since a resin (organic material) is used, only a sintering suppression effect (or crack prevention by simultaneous firing) up to about 600 ° C. can be obtained, but this also has a considerable crack prevention effect.

  In addition, crack prevention and simultaneous firing of multiple layers can be further facilitated by using in combination with crack prevention technology using these organic substances and the addition of a sintering inhibitor to the electrodes described later (crack prevention technology using inorganic substances). Needless to say.

  Moreover, as for the thickness of the glass paste after hardening or insolubilization, 12-300 micrometers is desirable. If it is less than 12 μm, pinholes may occur during printing. Moreover, when the thickness is less than 12 μm, the thickness of the glass layer after firing is further reduced to less than 10 μm, which may cause insulation failure. When the thickness of the glass paste exceeds 300 μm, even if multilayer printing is performed while curing, the number of printed layers increases, resulting in an increase in cost. On the other hand, when the thickness exceeds 300 μm, the thickness of the glass layer after firing exceeds 200 μm, the thickness and weight of the load sensor increase, and the material cost of the glass also increases.

(Embodiment 6)
Next, as a sixth embodiment, a method for producing a curable crystallized glass paste will be described.

  First, as the crystallized glass, the materials shown in Tables 8 to 11 were weighed in a platinum crucible and melted with high heat. Thereafter, the dissolved glass component was dropped into water and rapidly cooled. Then, this was pulverized for a long time using a ball mill. Then, it classified and prepared glass powder with a particle size of 10 micrometers or less. Next, a curable resin was selected as the ink vehicle, and a curable phenol resin dissolved in a solvent was used. A predetermined amount of glass powder was added thereto, and the mixture was kneaded with a ceramic three roll. A water cooling device was connected to the ceramic roll to prevent resin curing due to roll heat generation during roll kneading. In addition, when one lot is a small amount of about 1 kg, there was no problem even if kneading with three rolls as it was after premixing by hand using a tool such as a spatula in a suitable container (some aggregates) ) Can be detected with the naked eye on the roll, and measures can be taken by increasing the number of rolls as necessary). However, if one lot is 10 kg or more, premixing by hand is too physically intensive, so premixing using a machine such as a planetary mixer or kneader can be followed by kneading with three rolls. desirable.

  The glass paste thus prepared was added with a solvent as desired to adjust the viscosity, and then filtered to complete. In addition, the filtration efficiency can be raised by using a commercially available screen type filter having an opening of 10 to 20 μm and performing pressure filtration as desired.

  Using the curable glass paste thus prepared, as shown in FIGS. 6 (A) to 6 (E), printing and curing were repeated for multilayer printing, but no particular problem could be found. Although this was baked, no abnormality was found. For comparison, an ink was similarly prepared using a conventional ink in which ethyl cellulose was dissolved in α-terpineol instead of a curable phenolic resin as a vehicle. When multilayer printing was performed using this, results as shown in FIGS. 5A to 5E were obtained. Moreover, when this was baked, it was easy to produce the insulation defect between a metal elastic body and an internal electrode. Then, when the cross section of the sample in which the insulation failure occurred was observed, as shown in FIG. 5 (E), the glass layer had a large thickness unevenness, and further, cavities were generated in the sintered glass.

  The curing temperature is preferably 100 ° C. or more and 10 seconds or more, and less than 400 ° C. and less than 30 minutes. When the curing temperature is less than 100 ° C., the ink viscosity may increase when the ink is kneaded or the ink is allowed to stand naturally. When the curing temperature is 400 ° C. or higher, a part of the resin begins to decompose, so it may be difficult to improve the printability by adding such a resin to the curable resin. Further, the curing time is desirably 10 seconds or more and 30 minutes or less. Particularly, a resin that starts to cure within 10 seconds at 100 ° C. may change the viscosity at the time of ink kneading or shorten the pot life.

  If the curing temperature is set to 400 ° C. or more and 30 minutes or more, a part of the added resin may start to decompose, and when the next glass layer or the like is printed thereon, the leveling property is lowered. Can occur.

  Moreover, light can be used for hardening (or insolubilization). For example, it can be cured by using UV (ultraviolet rays) and IR (far infrared rays including far red). Such photocurable resins are commercially available from various manufacturers for screen printing, for example, as an overcoat material for posters. In addition, such photocurable resins are commercially available that can be diluted with a high boiling point solvent having a boiling point of 150 ° C. or higher, such as BCA (butyl carbitol acetate) and α-terpineol. Such a thing can also be used for this invention.

  Here, even if the setting of the curing device is set to 400 ° C., in the case of a load sensor, the metal elastic body has a large heat capacity, and thus the temperature may not increase. In such a case, it may be heated directly using a hot plate or the like.

  In addition, as crystallized glass used for crystallized glass, those other than those shown in Tables 8 to 11 can be used. In particular, in the case of the load sensor proposed in the present invention, since a metal elastic body is used, the thermal expansion coefficient is about 90 × 1E-7 / ° C. (9 ppm / ° C.) to 150 × 1E-7 / ° C. (15 ppm / ° C.). A commercially available crystallized glass can be selected.

  In this case, the average particle size of the glass powder is preferably less than 5 μm. When the average particle size is 5 μm or more, large lumps of 20 to 30 μm may remain, and these may affect the printability. Moreover, as a paste viscosity, the range of 10-10000 poise is suitable. When the viscosity is less than 10 poise, the viscosity is too low, it is difficult to obtain a high film thickness, and there is a possibility that the printed pattern may be blurred or sag. Moreover, when a viscosity exceeds 10,000 poise, a viscosity is too high and a pinhole may generate | occur | produce in a coating film. In the case of such a paste, the viscosity may vary greatly depending on the shear rate (or rotational speed) of the viscometer. Therefore, it is desirable to measure the paste viscosity in a range where the shear rate is 1 to 100 / s (where 1 / s is a unit of the shear rate).

(Embodiment 7)
Next, as Embodiment 7, an amorphous glass paste curable paste will be described.

  First, as the amorphous glass, the materials shown in Tables 5 to 7 were weighed in a platinum crucible and melted with high heat. Thereafter, the dissolved glass component was dropped into water and rapidly cooled. Then, this was pulverized for a long time using a ball mill. Thereafter, glass powder having a particle size of less than 10 μm was collected by classification. Next, as a vehicle for ink, a curable resin was selected, and a curable phenol resin dissolved in a solvent was used. A predetermined amount of glass powder was added thereto, and kneaded by a ceramic three roll. A water cooling device was connected to the ceramic roll to prevent resin curing due to roll heat generation during roll kneading. In addition, when one lot was a small amount of about 1 kg, there was no problem even if kneading with three rolls as it was after premixing by hand using a jig such as a spatula in an appropriate container. Even if some agglomerates remained, they could be detected with the naked eye on the roll, and countermeasures could be taken by increasing the number of rolls as necessary.

  However, if one lot is 10 kg or more, premixing by hand is too physically intensive, so premixing using a machine such as a planetary mixer or kneader can be followed by kneading with three rolls. desirable.

  A solvent was added to the glass paste thus prepared as necessary, the viscosity was adjusted, and then filtered to complete. In addition, the filtration efficiency can be raised by using a commercially available screen type filter having an opening of 10 to 20 μm and performing pressure filtration as desired.

  Using the curable glass paste thus prepared, as shown in FIGS. 5A to 5E, printing and curing were repeated and multilayer printing was performed, but no particular problem could be found. Although this was baked, no abnormality was found.

  For comparison, an ink was similarly prepared using a conventional ink in which ethyl cellulose was dissolved in α-terpineol instead of a curable phenolic resin as a vehicle. When multilayer printing was performed using this, results as shown in FIGS. 5A to 5E were obtained. Moreover, when this was baked, it was easy to produce the insulation defect between a metal elastic body and an internal electrode. Therefore, when a cross section of the sample in which the insulation failure occurred was observed, as shown in FIG. 5E, the glass layer had a large thickness unevenness, and further, there were many cavities in the sintered glass.

  The curing temperature is preferably 100 ° C. or more and 10 seconds or more, and less than 400 ° C. and less than 30 minutes. When the curing temperature is less than 100 ° C., the ink viscosity may increase when the ink is kneaded or the ink is allowed to stand naturally. When the curing temperature is 400 ° C. or higher, a part of the resin begins to decompose, so it may be difficult to improve the printability by adding such a resin to the curable resin. The curing time is preferably 10 seconds or more and 30 minutes or less. Particularly, a resin that begins to harden at 100 ° C. within 10 seconds may change viscosity during ink kneading or shorten the pot life.

  In addition, if it is necessary at 400 ° C. or higher for 30 minutes or more, a part of the added resin may start to decompose, and when the next glass layer or the like is printed thereon, the leveling property may be lowered. .

  Note that materials other than those shown in Tables 5 to 7 can be used for the amorphous glass. In particular, in the case of the load sensor proposed in the present invention, since a metal elastic body is used, a commercially available amorphous glass having a thermal expansion coefficient of about 90 × 1E-7 / ° C. to about 150 × 1E-7 / ° C. A material having good matching with the strain sensitive resistor may be selected. The average particle size of the glass powder is desirably less than 5 μm. When the average particle size is 5 μm or more, large lumps of 20 to 30 μm may remain, and these may affect the printability. Moreover, as a paste viscosity, the range of 10-10000 poise is suitable. When the viscosity is less than 10 poise, the viscosity is too low, it is difficult to obtain a high film thickness, and there is a possibility that the printed pattern may be blurred or sag. Moreover, when a viscosity exceeds 10,000 poise, a viscosity is too high and a pinhole may generate | occur | produce in a coating film. In the case of such a paste, the viscosity may vary greatly depending on the shear rate (or rotational speed) of the viscometer. Therefore, it is desirable to measure the paste viscosity in a range where the shear rate is 1 to 100 / s.

(Embodiment 8)
Next, a method for producing a curable electrode paste will be described as an eighth embodiment.

  First, a conductive powder mainly composed of silver was selected as an internal electric paste, and a curable phenol resin previously dissolved in a solvent was used and kneaded with a metal three roll. A water-cooling device was connected to the three rolls to prevent resin curing due to roll heat generation during roll kneading. In addition, when one lot was a small amount of about 1 kg, there was no problem even if kneading with three rolls as it was after premixing by hand using a jig such as a spatula in an appropriate container. Even if some agglomerates remained, they could be detected with the naked eye on the roll, and countermeasures could be taken by increasing the number of rolls as necessary.

  However, if one lot is 10 kg or more, premixing by hand is too physically intensive, so premixing using a machine such as a planetary mixer or kneader can be followed by kneading with three rolls. desirable. A solvent was added to the electrode paste thus prepared as desired, the viscosity was adjusted, and then filtered to complete. In addition, filtration efficiency can be raised by using a screen-type filter with a commercially available opening of 10 to 20 μm and filtering under pressure as necessary.

  Using the electrode paste thus prepared, as shown in FIGS. 5A to 5E, printing and curing were repeated for multilayer printing, but no particular problem occurred. Although this was baked, no abnormality was found. For comparison, an ink was similarly prepared using a conventional ink in which ethylcellulose was dissolved in α-terpineol instead of a curable phenolic resin as a vehicle. When multilayer printing was performed using this, results as shown in FIGS. 5A to 5E were obtained. Moreover, when this was baked, many unevenness | corrugations resulting from the printing nonuniformity generate | occur | produced in the glass material printed on the electrode paste.

  Insulation failure was likely to occur between the metal elastic body and the internal electrode. Therefore, when a cross section of the sample in which the insulation failure occurred was observed, as shown in FIG. 5E, the glass layer had a large thickness unevenness, and further, there were many cavities in the sintered glass.

  The curing temperature is preferably 100 ° C. or more and 10 seconds or more, and less than 400 ° C. and less than 30 minutes. When the curing temperature is less than 100 ° C., the ink viscosity may increase when the ink is kneaded or the ink is allowed to stand naturally. When the curing temperature is 400 ° C. or higher, a part of the resin begins to decompose, so it may be difficult to improve the printability by adding such a resin to the curable resin. The curing time is preferably 10 seconds or more and 30 minutes or less.

  In particular, when a resin that begins to harden within 10 seconds at 100 ° C. is employed, the viscosity may change during ink kneading and the pot life may be shortened. Also, if the curing temperature is set to 400 ° C. or more and 30 minutes or more, a part of the added resin may start to decompose, and when the next glass layer or the like is printed thereon, the leveling property is lowered. There is.

  In the present invention, the curable paste can be combined with a glass paste as well as an electrode paste.

  Next, the curable resin used for the curable electrode paste will be described. First, as the curable resin, a one-component curable resin is desirable. General phenol resins and epoxy resins can be used for such applications. When an insulating paste is made of a two-component resin, it is necessary to add and mix a reaction initiator immediately before use. Therefore, in a general manufacturing process, the process becomes complicated, and an error in weighing the reaction initiator tends to occur.

  On the other hand, a one-pack type curable resin typified by xylene resin and the like is easy to use as the curable resin of the present invention because it cures only by heating without adding a reaction initiator. The curable electrode paste preferably contains 10 to 80 wt% of metal powder. When the metal powder is 10 wt% or less, necessary conductivity may not be obtained.

  Moreover, when there are more metal powders than 80 wt%, sufficient leveling (smoothing) may not be obtained after printing. The resin component contained in the curable electrode paste is preferably 1 to 20 wt% or less. When the resin component is less than 1 wt%, the physical strength of the unfired insulating layer (after curing) formed by curing may be low. If the strength of the unsintered insulating layer (after curing) is low, the electrode pattern may be cracked, scratched, or dented into a mesh when the next insulating paste is printed on it by screen printing. , May reduce product yield.

  Further, the organic solvent contained in the curable electrode paste preferably has a boiling point of 150 ° C. or higher. By including the organic solvent in the cured insulating paste in this way, the viscosity of the paste and the solid content of the paste can be easily adjusted, so that the production quality can be stabilized over a long period of time. In the case of an organic solvent having a boiling point of less than 150 ° C., it spontaneously evaporates during the printing process, so the paste viscosity changes during the work process, and the thickness variation of the unfired insulating layer (after curing) may occur.

  Moreover, it is desirable that such an organic solvent contains 10 to 40 wt% of the solvent in the paste. When the solvent in the paste is less than 10 wt%, the viscosity and solid content adjustment range by the solvent becomes narrow. When the solvent is contained in an amount of 40 wt% or more, the change in the film thickness due to drying of the ink increases, and in some cases, the inside of the electrode may become porous only after drying (or curing). The viscosity of the insulating paste is preferably 10 poise or more and 10,000 poise or less. If it is less than 10 poise, the ink viscosity is too low to make screen printing difficult.

  If the ink viscosity exceeds 10,000 poise, the viscosity is too high, and it is difficult to perform sufficient leveling no matter how hard the base is cured. In addition, by making 5 to 100 wt% of the resin in the insulating paste into a one-component curable resin, the insulating paste can be easily manufactured, and the insulating paste is cured on the screen printer and cannot be removed later. There is nothing.

  For example, by setting 50% of the resin component in the cured insulating paste to thermosetting xylene resin and the remaining 50% to ethyl cellulose resin soluble in the solvent, the ink should be cured on the screen plate. However (or even when it is cured), it can be easily wiped off with a solvent if the solvent solubility of ethyl cellulose is utilized. Therefore, since no foreign matter remains on the screen plate, stable printing for a long time or unattended printing becomes possible.

  The curing temperature is preferably 100 ° C. or more and 10 seconds or more. When a resin that is easily cured, for example, a resin that is cured at less than 100 ° C. for less than 10 seconds, is used for the paste, the pot life of the paste may be shortened even during the printing process or in a cool and dark place.

  The metal powder preferably has an average particle size of 0.1 to 10 μm, and the paste preferably contains 60 to 80 wt% or more of such metal powder mainly composed of silver. In this way, it is possible to reduce the cost of electronic components by using inexpensive and low-resistance silver for the electrodes. In addition, by adding palladium or platinum or as an alloy as desired, solder wettability, migration prevention, and the like can be achieved, and the reliability of the electronic component can be improved.

  Note that a plurality of unfired insulating layers can be fired at the same time as they are formed on the substrate together with the electrode layers as necessary. In the first to third embodiments, the quality improvement during multilayer printing has been described. However, it is also possible to simultaneously fire a plurality of unfired insulating layers thus cured and formed on the metal substrate. By simultaneously firing a plurality of unfired insulating layers together with the unfired electrode layers in this way, the number of firing furnaces can be reduced, and the manufacturing cost can be greatly reduced.

  Further, the firing temperature is desirably 600 to 950 ° C. (desirably 600 to 900 ° C.). When the temperature is lower than 600 ° C., the electrical conductivity and the adhesive force with the substrate may be insufficient. In addition, when a firing temperature higher than 950 ° C. is required, it is not possible to use an inexpensive silver-based conductor powder as a wiring material, and a metal substrate itself is expensive and special (high proof stress, oxidation resistance, heat resistance). It is necessary to use it. The firing is preferably performed in an oxidizing atmosphere using air containing general oxygen instead of a reducing atmosphere using nitrogen gas, hydrogen gas, or the like. Thus, by baking in an oxidizing atmosphere, the curable resin can hardly remain as residual carbon in the insulator, so that the reliability of the insulating layer can be increased.

  In the case of a curable electrode paste, metal powder having a particle size of 0.1 to 10 μm is 60 to 80 wt%, glass powder or ceramic powder is 1 to 10 wt%, resin is 1 to 20 wt%, and boiling point is 150 ° C. or higher. It is desirable that the solvent contains 10 to 40 wt%. The viscosity is preferably 10 to 10000 poise, and 5 to 100% of the resin is preferably a curable resin.

  When the average particle diameter of the metal powder is less than 0.1 μm, the shrinkage start temperature during sintering is lowered, and it may be difficult to perform simultaneous firing. On the other hand, if it exceeds 10 μm, thickness unevenness is likely to occur, and formation of a fine pattern becomes difficult. Moreover, when the addition amount of the metal powder is less than 60 wt%, the resistance value may increase. On the other hand, when the content is 80 wt% or more, the ink viscosity is easily changed due to a decrease in organic components such as a solvent, and handling is difficult. Glass powder or ceramic powder may not be powder but may be an organic compound (so-called metal resinate). However, when the amount of addition is less than 1 wt%, the effect of addition is reduced and adhesion to the substrate is reduced. May decrease.

  Moreover, when the addition amount exceeds 10 wt%, the resistance value may be increased. Moreover, when resin is less than 1 wt%, the coating-film intensity | strength after hardening of electrode paste falls, and it becomes easy to take or it becomes easy to get damaged. On the other hand, when the resin content exceeds 20 wt%, the coating film strength increases, but the amount of resin decomposition during firing increases, and the resin may swell or crack during firing.

  The boiling point of the solvent is preferably 150 ° C. or higher. When the solvent is 150 ° C. or lower, the paste dries quickly on the printing press, which affects continuous printability. When the solvent is less than 10 wt%, the paste is easily dried and the ink viscosity is easily changed.

  If it exceeds 40 wt%, it takes too much time to dry the ink. When the viscosity is 10 poises or less, the pattern tends to bleed during printing, and it may be difficult to form a highly accurate pattern.

  On the other hand, if the viscosity exceeds 10,000 poise, it is difficult to print a fine pattern because the viscosity is too high. Moreover, when the ratio of curable resin in resin is less than 5 wt%, it may not harden | cure. Further, even if 100% is a curable resin, there is no problem. In this case, the curable resin is preferably a solvent-soluble type. By using a solvent-soluble curable resin, the solid content (other than volatile components) in the ink can be increased, the coating film can be thickened, and the equipment and plate after printing can be washed with an organic solvent. Is good. The paste viscosity can also be adjusted to an optimum value by the amount of solvent.

  The thickness of the internal electrode thus formed is preferably less than 50 μm. When the thickness exceeds 50 μm, the electrode cost increases and the thickness becomes a step, which may affect the formation of the glass layer, wiring, resistor, etc. formed thereon.

(Embodiment 9) Internal power common material Embodiment 9 solves the problem which is easy to occur at the time of simultaneous baking of glass and an internal electrode by optimizing the composition of an internal electrode. First, silver should be 80 wt% or more as the internal electrode, and further, 1 to 20 wt% of the crystalline glass material and alumina material formed above and below the internal electrode should be added as inorganic sintering inhibitors. Is desirable.

  For example, in FIG. 2, it is desirable that a glass material or a composite glass material constituting the crystallized glass 14b sandwiching the internal electrode 19 above and below is added to the internal electrode 19 in advance in the internal electrode paste. . By doing so, the bonding when the internal electrode paste is fired with the crystalline glass can be made stronger, and the thermal expansion coefficient of the internal electrode 19 can be made closer to the crystallized glass 14b. Even so, the stress concentration can be suppressed and cracks are unlikely to occur.

  Next, the case where an internal electrode and a plurality of layers of glass are fired simultaneously will be described with reference to FIG. By firing the plurality of layers in this way, the manufacturing cost can be reduced. In FIG. 8A, crystalline glasses 39 a and 39 b are formed (fired) on the metal elastic body 1. An internal electrode paste 40a is printed in a predetermined pattern on the crystalline glass 39h, and further a crystalline glass paste 41a is printed in a predetermined pattern thereon. Thus, when performing multilayer printing, the leveling property can be improved by making the paste curable, and thickness variations and pinholes of the printed coating film can be reduced.

  In Embodiment 9, simultaneous firing of the internal electrode paste and the crystallized glass paste covering the internal electrode paste will be described.

  In FIG. 8B, the crystalline glass 39c and the internal electrode 42a are obtained by firing the crystalline glass paste 41a and the internal electrode paste 40a at the same time. As shown in FIG. 8 (B), by adding a sintering inhibitor to the internal electrode paste 40a in advance, the internal electrode paste and the crystallized glass paste covering the internal electrode paste are particularly fired. Does not occur.

  On the other hand, as shown in FIG. 8C, when no sintering inhibitor is added to the internal electrode paste 40a, cracks 43 are likely to occur in the periphery of the internal electrode 42b and in the crystallized glass 39c surrounding the periphery. . In FIG. 8C, the crystallized glasses 39a and 39b are pre-fired and crystallized (the crystallization raises the melting point to 1000 ° C. or higher), so that the internal electrode paste 40a is heated to about 850 ° C. The cracks 43 are unlikely to occur even when fired at.

  In addition, as a sintering inhibitor, ceramic powder or metal resinate (organic metal) can be used.

  FIG. 9 is a cross-sectional view for explaining a state in which the plurality of layers are baked together with the crystallized glass serving as a base for the internal electrode. In FIG. 9A, a plurality of crystallized glass pastes 41b and 41c are printed and laminated on the metal elastic body 1, and an internal electrode paste 40b is further printed in a predetermined shape. By adding a sintering suppressing material to the internal electrode paste 40b in advance, as described in FIG. 9B in its cross section, there is a problem with the internal electrode 42b and the crystallized glass 39c and 39d as the base. Does not occur. On the other hand, as shown in FIG. 9C, when no sintering inhibitor is added to the internal electrode paste 40b, cracks 43 are likely to occur around the internal electrode 42b and the crystallized glass 39c, 39d below. Become. The cause of the occurrence of such cracks is the difference in firing shrinkage between metal and glass, and these cracks 43 may also occur at the interface between crystallized glasses and the interface with internal electrodes.

  FIG. 10 is a cross-sectional view showing a state in which a plurality of layers are baked at a time with the upper and lower sides of the internal electrode sandwiched between crystallized glass. In FIG. 10A, a plurality of crystallized glass pastes 41d and 41e are printed on the metal elastic body 1, and the internal electrode paste 40c is further printed in a predetermined shape, and further covered with the crystallized glass paste 41f. Yes. By adding a sintering inhibitor to the internal electrode paste 40c in advance, the internal electrode paste 40c and the crystallized glass pastes 41e and 41f sandwiching the upper and lower sides as shown in FIG. Even if co-firing, no particular problem occurs.

  On the other hand, as shown in FIG. 10C, when the sintering inhibitor is not added to the internal electrode paste 40c, cracks 43 are formed around the internal electrode 42c, the crystallized glass 39f in which it is embedded, and the interface thereof. Likely to happen. These cracks 43 may also occur at the interface between the crystallized glasses and the internal electrode.

  In addition, when adding ceramic powder or crystallized glass as a sintering inhibitor to such an electrode paste, if these are added in powder form, they remain as aggregates after sintering, which may reduce reliability. Therefore, when adding ceramic powder, it is desirable to disperse these in a paste form in advance. It should be noted that by using a glass paste or electrode material as a curable paste, it is possible to prevent the occurrence of cracks during simultaneous firing of a plurality of layers as shown in FIGS.

  This is because by curing the paste, the firing shrinkage start temperature of the cured paste is shifted to a higher temperature side by about 100 to 200 ° C. than when a general resin such as ethyl cellulose is used. Furthermore, a synergistic effect can be expected by adding such a firing inhibitor to the curable paste, and accordingly, the amount of addition of the sintering inhibitor can be reduced, and the conductivity of the electrode can be increased. Since resin is much cheaper, the cost of the electrode material can be reduced.

  Further, by using such a technique, as shown in FIG. 2, the crystallized glass 14b is simultaneously fired while the internal electrode 19 is incorporated, or the internal electrode 19, the crystallized glass 14b, and the amorphous glass 15b are simultaneously fired. Needless to say, the internal electrode 19, the crystallized glass 14b, the amorphous glass 15b, and the wiring 16b can be fired simultaneously. Further, as described above, the simultaneous firing is further facilitated by using the crack prevention technology using the curable resin and the crack prevention technology using the sintering inhibitor described here.

As the internal electrodes, while mainly of silver, as desired, i.e., in accordance with the design instructions and specifications, it is desirable that the Bi ₂ O ₃ is added 1 to 15 wt%. When the added amount of Bi ₂ O ₃ is less than 1 wt%, the adhesive strength with glass may be low. On the other hand, if it exceeds 15 wt%, the resistance value exceeds a predetermined range, or the influence of diffusion of bismuth into the glass layer can be considered. Further, it is desirable to contain 0.5 to 10 wt% of SiO ₂ or CuO. When these are less than 0.5 wt%, the adhesiveness of the glass may decrease. On the other hand, if it exceeds 10 wt%, the resistance value may be increased or the influence of diffusion into glass may be considered. In addition, such SiO ₂ and CuO can be expected to suppress the sintering of the electrode.

(Embodiment 10)
In the tenth embodiment, a wiring material connected to the strain sensitive resistor will be described. The cost of the product can be reduced by soldering various semiconductors and chip parts using wiring connected to the strain sensitive resistor. Such component mounting is not performed by the manufacturer but may be performed by the user. Compared with a hybrid IC using a general alumina substrate or the like, the load sensor has a large heat capacity by using a metal substrate.

  Therefore, even if the solder reflow is performed on the load sensor in advance, it may be difficult to attach the solder compared to a general circuit board. Therefore, it takes time for soldering, and the wiring of the load sensor may be eroded. This solder erosion means that the electrode material that forms the wiring dissolves in the solder, and its thickness decreases or disappears. It is conceivable that the adhesive strength and tensile strength are reduced.

  For such a problem specific to the load sensor, it is desirable to use silver palladium as an electrode material used for wiring. In particular, when the palladium content is 5% or more, preferably 10% or more, the occurrence of solder erosion can be suppressed even under various soldering conditions.

  In addition to solder erosion, the wettability of the wiring material to the solder is also an important factor for soldering to the wiring. In order to improve the wettability with respect to such solder, it is desirable to add bismuth oxide to the wiring material. The addition amount of bismuth oxide is about 1 to 20 wt%, desirably 5 wt% or more and less than 20 wt%. When the addition amount of bismuth oxide is less than 1 wt%, the addition effect may not be obtained.

On the other hand, if it exceeds 20 wt%, the added bismuth oxide may be deposited around the wiring pattern after firing to lower the characteristics or affect the electrical characteristics of the strain sensitive resistor. Further, by adding 0.5 to 5 wt% of SiO ₂ or CuO, the adhesion with the base can be improved.

  Moreover, although it is desirable to use silver palladium as a wiring material for the component mounting portion (or soldered portion), it is not always necessary to use silver palladium for connection to the strain sensitive resistor, and the palladium content is low (for example, 1 wt%), an inexpensive electrode material mainly composed of silver can be used. In this way, it is possible to print the material having a high palladium content separately from the solder mounting portion and the inexpensive wiring material from the other portions.

  As proposed in the present invention, in particular, by setting the amorphous glass paste to a curable type, the leveling property of the wiring pattern printed thereon can be improved, so that the film thickness uniformity is improved.

  In addition, an excellent effect can be obtained particularly when printing is performed by combining different wiring inks. Also, the more uniform the wiring thickness after printing (before firing), the more uniform the thickness of the wiring after firing, so the printability of the strain sensitive resistor printed thereon can also be improved. Component mountability (solderability) in the process can also be stabilized.

(Embodiment 11) Explanation of composite glass (crystalline + ceramic)
Next, as an eleventh embodiment, a method for manufacturing a load sensor using various types of metal elastic bodies will be described. As the use of load sensors expands, there are various demands on the load sensor from the market, and it is necessary to use metal elastic bodies of various shapes and materials according to such needs and applications. For example, there are a difference in appearance dimensions, a difference in substrate thickness, a difference in processing method, a difference in thickness and material, and the presence or absence of a mounting hole. It is necessary to provide a load sensor with multi-layered wiring to the market while using such a member.

  However, the conventionally proposed glass materials have a fixed coefficient of thermal expansion, so that there remains a problem with such measures.

  FIG. 11 explains the warpage of the metal elastic body that becomes such a substrate. FIG. 11A shows an example in which a glass layer 44 is formed on the metal elastic body 1. Reference numeral 45 denotes a warp of the metal elastic body 1. FIG. 11 (B) shows the relationship between the thickness of the metal plate and its warpage. As shown in FIG. 11B, the amount of warpage decreases as the thickness of the metal elastic body 1 increases. However, if the thickness exceeds a certain value, the glass layer 44 may be broken or peeled off. In general, the load sensor optimizes and builds each glass material for a specified metal substrate (material, thickness, shape, etc.), so there is no problem as shown in FIG. . However, even when the same metal elastic body is used, if the thickness is changed to 0.2 mm, 1 mm, 2 mm, and 5 mm, the same glass material cannot be used. This affects not only the reason shown in FIG. 11B but also the difference in processing method due to the change in the thickness of the substrate, the residual amount of processing distortion, and the like. Even when the same metal elastic body is used, punching with a thickness of 1 mm and punching with a thickness of 5 mm are greatly different in processing method and procedure, so that a material having a single thermal expansion coefficient cannot cope with it. Therefore, it is necessary to develop a myriad of glass materials, which causes an increase in cost.

  Therefore, in the eleventh embodiment, a glass material used for a load sensor is made into a composite, and the thermal expansion coefficient thereof is finely adjusted to meet such applications.

  FIG. 12 is an enlarged schematic cross-sectional view of a composite glass. In FIG. 12, the composite powder 48 is formed by co-firing with the ceramic powder 47 dispersed inside the crystallized glass 46.

  Thus, the composite glass formed by co-firing ceramic powder dispersed in the glass is finely adjusted for the type and amount of ceramic powder 47 added thereto, and the composite glass The thermal expansion coefficient can be finely adjusted.

  Conventionally, it has been necessary to develop glass materials in accordance with metal elastic bodies. By compositing the glass material in this way, even one type of glass material can be used for various metal elastic bodies, and the cost of the load sensor can be reduced. In addition, if such a method is used, it is possible to form a composite using amorphous glass in addition to crystallized glass.

  First, as an eleventh embodiment, a load sensor using a composite glass layer, which is a feature of the present invention, will be described. In the eleventh embodiment, composite glass is used for the crystalline glass 14a of FIG. 1 and the crystalline glass 14b of FIG.

  FIG. 13 is a diagram for explaining how the thermal expansion coefficient of the composite glass is changed. FIG. 13A shows the composite glass 48 formed on the metal elastic body 1. FIG. 13B shows the measured thermal expansion coefficient of the composite glass. In FIG. 13B, the X axis is the amount of filler added to the composite glass layer, and the Y axis is the thermal expansion coefficient of the finished composite glass. From FIG. 13B, it can be seen that the thermal expansion coefficient of the composite glass prepared by adding the filler A decreases as the filler addition amount increases. Moreover, in the case of the composite glass to which the filler B is added, it can be seen that the thermal expansion coefficient increases as the filler addition amount increases.

  From FIG. 13B, it can be seen that when the filler addition amount is less than 5 wt%, the change in the thermal expansion coefficient is small, and the filler addition effect may be small. Further, as shown in FIG. 13B, when the filler addition amount exceeds 40 wt%, the composite glass becomes difficult to sinter and its mechanical strength may be lowered. In this way, the variable range of the thermal expansion coefficient can be expanded by adjusting the type and addition amount of the filler, so even a single glass material can be applied to a metal elastic body having various thermal expansion coefficients. .

  When the coefficient of thermal expansion of the crystallized glass 46 in FIG. 12 is 8.5 ppm / ° C., the ceramic powder 47 made of alumina having a coefficient of thermal expansion of 7 ppm / ° C. is added to this, so that the thermal expansion of the finished composite glass is achieved. The coefficient can be lowered. Further, by adding zirconia having a thermal expansion coefficient of 9.5 ppm / ° C. as the ceramic powder 47 of FIG. 12, the thermal expansion coefficient of the finished composite glass can be increased.

  Even when the crystallized glass 46 itself has a large thermal expansion coefficient of 10 or 13 ppm / ° C., the thermal expansion coefficient can be adjusted by compositing in this way.

  In particular, by selecting a crystallized glass 46 having a thermal expansion coefficient smaller than that of the metal elastic body 1, even in a state where a filler (corresponding to filler B) having a higher thermal expansion coefficient than the crystallized glass 46 is added. Compressive stress is generated in the composite glass 48.

  As described above, the thermal expansion coefficient of the crystallized glass 46 itself is unique. However, as proposed in the present invention, ceramic powder is added as a filler to make a composite, and the thermal expansion coefficient is reduced from several% to 10%. It can be increased or decreased by several percent.

  In addition, as for the addition rate of the filler added to crystallized glass, 5 wt% or more and less than 40 wt% are desirable. That is, the ratio between the crystallized glass and the ceramic powder constituting the composite glass is preferably between crystallized glass: ceramic powder = 95: 5 to 60:40. The reason for this is that, as will be described with reference to FIG. 13B, when the amount of filler added is less than 5 wt%, the resulting composite glass 48 has little change in the coefficient of thermal expansion, and the crystallization characteristic of the composite glass 48 is achieved. Advantages of glass 46 (for example, the melting temperature increases after crystallization and the materials are less likely to diffuse each other) and ceramic powders 47 (stable and less likely to diffuse each other, thermal expansion coefficients differ depending on the materials) This is because it may not be possible to obtain a synergistic effect (which is difficult to be remelted by forming the composite glass 48, hardly diffuses after crystallization, and the thermal expansion coefficient can be increased or decreased as necessary).

  Moreover, when the addition amount of a filler exceeds 40 wt%, as shown to FIG. 13 (B), it becomes difficult to sinter composite glass and the intensity | strength may fall. This is because the amount of glass necessary to wet the surface of the ceramic powder 47 is insufficient. In some cases, fine holes or the like may occur in the finished composite glass 48, or the insulation resistance may be lowered. Thus, the amount of filler added is desirably 5 wt% or more and 40 wt% or less of the composite glass, and the thermal expansion coefficient can be increased or decreased within this range. Moreover, when adjusting the thermal expansion coefficient beyond it, it can respond by selecting the crystallized glass 46 which has a different thermal expansion coefficient, and adding various fillers similarly to this.

The ceramic powder 47 added to the composite glass 48 of the present invention includes alumina (Al ₂ O ₃ , thermal expansion coefficient of 6.5 to 8.0 ppm / ° C.), MgO (thermal expansion coefficient of 13 ppm / ° C.), forsterite. (2MgO · SiO ₂ , thermal expansion coefficient is 8 to 11 ppm / ° C. depending on composition ratio), zirconia (thermal expansion coefficient is 10.4 ppm / ° C.), titanium oxide, calcium oxide, magnesium oxide, spinel (MgO · Al ₂ O ₃ ) Etc. are desirable. In addition, the average particle diameter is desirably about 0.01 to 5 μm. If it is less than 0.01 μm, the cost becomes high.

  If the average particle size is larger than 5 μm, simultaneous firing with glass may be difficult. In addition, for commercially available crystallized glass other than the crystallized glass proposed in the present embodiment, by adding ceramic powder as a filler in this way, the thermal expansion coefficient of the finished glass composite material can be finely adjusted. It is possible to optimize the design for any product.

  The ceramic powder preferably has a thermal expansion coefficient of 6 ppm / ° C. or more and less than 15 ppm / ° C. The ceramic powder 47 of less than 6 ppm / ° C. or 15 ppm / ° C. or more may be difficult to match with the metal elastic body 1 even if it is composite vitrified.

  In particular, since the ceramic powder in the composite glass hardly reacts with the glass, it often remains as it is after firing as shown in FIG. Therefore, the calculation of the thermal expansion coefficient of the composite glass can be easily obtained by proportional calculation between the thermal expansion coefficient of the crystallized glass and the thermal expansion coefficient of the ceramic powder added thereto.

Although FIG. 12 is simplified, when a cross section of an actual composite glass is analyzed by SEM (scanning electron microscope) or XMA (X-ray microanalysis), grain boundaries formed by crystallized glass and The diffused ceramic powder is visible. In addition, when these elements are elementally analyzed, a plurality of elements (for example, MgO, SiO ₂ , Al ₂ O _3, etc.) constituting the crystallized glass are mixed and detected from the crystallized glass. Intrinsic elements from the powder (when alumina is used in the ceramic powder, the elements of aluminum and oxygen are specifically detected with high strength.

  In addition, when zirconia oxide is used in the ceramic powder, the elements of zirconia and oxygen are specifically detected with high intensity but can be detected with high concentration, so these elements were originally included in the crystallized glass, or Discrimination is easy for those added as fillers to constitute composite glass.

  The total thickness of the composite glass 48 is preferably at least 10 μm and not more than 200 μm. When the thickness is less than 10 μm, the glass paste is easily affected by a pinhole or the like generated when the glass paste is screen-printed. On the other hand, when the thickness exceeds 200 μm, the cost of the glass increases.

  Moreover, although the manufacturing method of composite glass is baking the glass paste which consists of crystal glass powder and ceramic powder, the composite glass which consists of amorphous glass and ceramic powder can also be created using this method. In this case, a glass paste obtained by blending both amorphous glass powder and ceramic powder at a predetermined ratio (desirably, by weight ratio between amorphous glass powder: ceramic powder = 95: 5 to 60:40), What is necessary is just to bake between 800-900 degreeC. By baking between 800-900 degreeC, it can disperse | distribute in an amorphous glass, without leaving damage to ceramic powder. In this case, some ceramic powder may be dissolved in amorphous glass, but most ceramic powder is dispersed as a kind of filler in amorphous glass. Therefore, if a predetermined analysis method (XMA or the like) is used, it can be determined whether it is amorphous glass or ceramic powder dispersed therein.

  Similarly, the amorphous composite glass prepared by blending the amorphous glass powder with the crystalline glass preferably has a weight ratio of amorphous glass: crystalline glass = 95: 5 to 50:50. . When the ratio of the amorphous glass exceeds 95 wt%, it may not be necessary to make a composite. When the proportion of the amorphous glass is less than 50 wt%, it should be called a composite glass mainly composed of crystalline glass rather than a composite glass mainly composed of amorphous.

The crystalline composite glass prepared by blending amorphous glass with crystalline glass is preferably in a weight ratio of crystalline glass: amorphous glass = 95: 5 to 50:50. When the ratio of the crystalline glass exceeds 95 wt%, it may not be necessary to make a composite. When the ratio of the crystalline glass is less than 50 wt%, it should be called a composite glass mainly composed of amorphous glass rather than a composite glass mainly composed of crystalline. The crystallinity of the glass in the composite glass can be analyzed using XRD (X-ray diffraction) or the like. For example, in the case of amorphous SiO ₂ , a broad rise called halo is observed. In crystalline SiO ₂ (eg, quartz), a plurality of sharp diffraction lines are observed. Moreover, the substance can be easily identified by analyzing the diffraction pattern peak position (referred to as d value) and peak intensity ratio in combination with commercially available substance analysis software.

  When a glass material is used for the overcoat, it is desirable that the lead oxide content in the overcoat glass is 60 wt% or more and less than 95 wt%. When the amount of lead oxide is less than 60 wt%, the firing temperature becomes high, so the strain sensitive resistor is reheated, and the resistance value and temperature characteristics may be shifted.

  Moreover, when the ratio of lead oxide is 95 wt% or more, water resistance may be lowered. Moreover, as for the thickness of overcoat glass, 10 micrometers or more and less than 200 micrometers are desirable. If the thickness is less than 10 μm, pinholes due to dust or bubbles may occur. When the overcoat glass has a thickness of 200 μm or more, the overcoat glass may flow and spread too much during firing, which may cause pattern defects. The overcoat glass is preferably two or more layers. By forming the overcoat glass to have two or more layers, even if pinholes due to dust or bubbles occur in one layer, it is difficult to cause defects.

  In addition, when using resin for an overcoat layer, it is desirable to contain 5 to 50 wt% of ceramic powder in a resin layer. By adding ceramic powder to the resin layer, the physical strength can be increased, so that the wear resistance is increased and the resin layer is hardly damaged. When the content of the ceramic powder is less than 5 wt%, the additive effect may not be obtained. When the content of the ceramic powder is 50 wt% or more, the fluidity and fillability of the overcoat resin layer may be lowered. The thickness of the resin overcoat layer is preferably 10 μm or more and less than 200 μm. If the thickness is less than 10 μm, pinholes due to dust or bubbles may occur. In addition, when the thickness of the overcoat resin layer is 200 μm or more, the overcoat resin may sag and flow at the time of curing so that the pattern defect may occur. The overcoat resin is preferably two or more layers. By using two or more layers of the overcoat resin, even if pinholes due to dust or bubbles occur in one layer of the overcoat, it is difficult to cause defects.

  Moreover, when making overcoat into multiple layers, although multiple layers may be formed by glass or resin, you may combine these. In this case, a first overcoat layer mainly composed of glass or ceramic having a thickness of 10 μm or more and less than 200 μm is disposed on the side closer to the strain sensitive resistor, and a second overcoat layer mainly composed of resin is formed thereon. It is desirable to form it with a thickness of 10 μm or more and less than 200 μm.

  In this case, it is desirable to form an overcoat mainly composed of a resin on an overcoat mainly composed of glass because of the heat treatment temperature. The overcoat material mainly composed of resin can be selected from the product numbers used for protecting various electrical components and mounting semiconductor bare chips. If the thickness of such an overcoat is less than 10 μm, there is a concern about pinholes, and if it is 200 μm or more, it takes a lot of time to form and the cost is high.

  Further, the thickness of the wiring pattern is preferably 3 μm or more and less than 50 μm. If the thickness is less than 3 μm, the electrode forming the wiring pattern may become too thin, and a portion where local conduction cannot be obtained may occur, resulting in an increase in wiring resistance or a capacitance formed between the substrate and the first electrode. Ingredients may be lower than design values. If the thickness exceeds 50 μm, the cost may increase.

  Further, it is desirable that the wiring pattern contains 3 wt% or more and less than 20 wt% of a glass component or a ceramic member. By adding in advance the glass component or ceramic member that forms the composite glass in the wiring pattern, the thermal expansion coefficient can be adjusted to the glass layer, and the adhesive strength with the glass layer can be further increased. Generation of delamination and the like with the layer can be prevented. Here, when the addition amount is less than 3 wt%, such an effect may be small. Further, when the added amount exceeds 20 wt%, the wiring resistance is increased. In particular, since the thermal expansion coefficient is two to three times larger than that of a ceramic substrate such as an alumina substrate like a load sensor, a slight difference in thermal expansion coefficient tends to cause such defects as delamination.

  In addition, when making a composite glass layer into multiple layers, as for the thickness per layer, 5 micrometers or more and less than 70 micrometers are desirable. If it is less than 5 μm, necessary insulation may not be obtained. On the other hand, when the thickness exceeds 70 μm, high-precision printing becomes difficult due to ink smearing or sagging during printing.

  Note that the glass layer, the composite glass layer, the electrode, and the like are preferably baked in an oxidizing atmosphere of 800 to 900 ° C. If it is less than 800 ° C., sufficient sintering strength may not be obtained. Further, when firing at a high temperature exceeding 900 ° C., the firing cost increases, the strength of the metal elastic body may be reduced, and it may be necessary to use an expensive metal substrate with higher heat resistance.

(Embodiment 12)
Next, as a twelfth embodiment, a method for producing a composite glass paste will be described. FIG. 14 explains that pinholes may occur in the composite glass. In FIG. 14, 49 is a ceramic powder aggregate and 50 is a pinhole. As shown in FIG. 14, when the ceramic powder 47 is uniformly dispersed inside the crystallized glass 46, such pinholes are not generated, but the ceramic powder 47 is not sufficiently dispersed, and the ceramic powder aggregate 49 When this occurs, a pinhole 50 may occur as shown in FIG. This is because the ceramic powder does not sinter at the firing temperature of the crystallized glass. In particular, if the ceramic powder remains in the composite glass in the form of aggregates, even if there is a very small probability, the pin It becomes a cause of generation of the hole 50.

  In order to prevent such pinholes 50, a countermeasure can be taken by making the dispersion of the ceramic material in the composite glass uniform.

  Next, the manufacturing method of a composite glass paste is demonstrated using FIG. 15, FIG. FIGS. 15 and 16 are diagrams for explaining a method for producing a composite glass paste. In Embodiment 12, ceramic powder used for composite glass is dispersed in a solvent in advance, and glass powder is added thereto. Kneading. Thus, the glass powder and the ceramic powder can be separately dispersed under optimum conditions for each, thereby providing a composite glass paste that does not contain an aggregate of ceramic powder.

  In FIG. 15A, reference numerals 51a to 51d are predetermined raw materials, specifically, ceramic powder, a solvent, a dispersant, a small amount of resin, and the like. Next, these predetermined raw materials are dispersed by beads 53 in a dispersing device 52 as shown in FIG. As the dispersing device 52, it is desirable to use a stirring type using beads such as a rotating ball mill, a stirring ball mill, or a Dymar mill made by Shinmaru Enterprise, or a vibration type or rotary type dispersing device.

  Ceramic powder can be mixed with a solvent or a small amount of dispersant (or a small amount of resin, depending on the type of resin. Adding a resin will improve the wettability of the ceramic powder. It is difficult to disperse with beads, and when adding a resin, the amount is limited to a small amount), and by dispersing with such a disperser, it can be dispersed without agglomerates.

  In this dispersion, the viscosity of the slurry made of ceramic powder, solvent, dispersant or a small amount of resin is preferably 1 cP (centipoise) or more and 10 P (poise) or less. In the case of a slurry having a low viscosity of 1 cP or less, even if the ceramic powder is unraveled in the disperser, the viscosity is too low and it may be removed when it is taken out and filtered. When the slurry viscosity exceeds 10 poise, the disperser using beads cannot exhibit the dispersing ability, and it may be difficult to remove the beads and the slurry. When using beads, the size of the beads is preferably 10 cm or less. The smaller the bead size is, the higher the dispersion effect is with respect to finer ceramic powders. However, since the energy of collision between the beads is small, it is desirable to use a dispersion device that uses a power of more than a certain horsepower. In the case of such a disperser, especially in the case of a rotary type, depending on the performance of the lip seal (a kind of a seal mechanism of the disperser), it may be difficult to use with beads having a diameter of 0.3 mmnφ or less, so care must be taken. Further, as the bead material, commercially available alumina or zirconia (with yttria) is desirable. Further, by selecting a bead material having the same component as the ceramic powder, even when the beads are polished and mixed into the slurry as impurities, the influence can be prevented.

  The slurry thus prepared is filtered as shown in FIG. In FIG. 15C, 55a is a filtering device, and 56a is a filter. As shown by an arrow 54 in FIG. 15C, the slurry is poured into a filter 56a set in a filtration device 55a, filtered, and then collected in a container 57a.

  In addition, as the filter 56a, a commercially available net having an opening of 10 to 20 μm can be used. However, if a filter obtained by processing a fiber into a pincushion shape, such as a depth type (volume filtration type), is used, a large amount The slurry can be filtered with minimal pressure loss (i.e., in a condition where the filter is not clogged). Further, such filtration may be performed by the weight of the slurry itself, but working efficiency can be increased by using an air pressure (pressure air), a diaphragm pump, or the like. The slurry thus poured as indicated by the arrow 54 in FIG. 15C is filtered and collected in the container 57 a as indicated by the arrow 54.

  Next, with reference to FIG. 16, a state in which a composite glass paste is prepared using ceramic powder dispersed without aggregates will be described. In FIG. 16A, the slurry in which the predetermined materials 51e to 51h and the ceramic powder are dispersed is the one recovered in the container 57a of FIG. 15C, crystallized glass powder, resin, dispersant, and the like. Here, a curable resin may be used as the resin. A part of the resin can also be used as a dispersant. In addition, if necessary, a colorant or the like is added, so that the thickness of each layer can be easily distinguished by color or element, and product management is facilitated.

  In FIG. 16 (B), 58 is a kneader, and a predetermined stirring jig 59 is built in. As such a kneader 58, a planetary mixer, a kneader, an automatic mortar, or the like can be used. Thus, by setting the stirring jig 59 in the kneader 58, even a highly viscous material can be stably kneaded. In particular, in the case of the load sensor proposed in the present invention, the cost can be reduced by using screen printing for printing the composite glass paste.

  However, when the composite glass paste is screen-printed, if the viscosity is too low, the pattern may bleed after printing or a predetermined thickness may not be obtained. Therefore, in the case of kneading such a composite glass, it is often the case that the viscosity is very high (for example, it does not flow under its own weight in a viscosity state of tens of thousands of poises). It is desirable.

  As shown in FIG. 16B, a predetermined amount of each of the predetermined materials 51e to 51h is put into a kneader 58, and the built-in stirring jig 59 is kneaded by rotating as indicated by an arrow 54. . Here, it is desirable to devise the addition order of the materials so that they do not react with each other. For example, when glass powder and resin, slurry and glass powder are kneaded in a large amount at once, an aggregate may be formed. In order to prevent such aggregate formation (for example, it may be called a solvent shock etc.), it can prevent by adding little by little or confirming the combination of things which are hard to aggregate beforehand.

  And as shown in FIG.16 (C), the ceramic powder and glass powder in a slurry can be disperse | distributed uniformly by knead | mixing using kneading machines, such as 3 rolls 60, and a composite glass paste can be created.

  Finally, as shown in FIG. 16D, dust and aggregates in the composite glass paste can be removed by filtering using the filtering device 55b and the filter 56b.

  Next, it explains in detail. First, as an example, a composite glass paste was prepared as shown in FIGS. Only ceramic powder was dispersed and slurried in advance, filtered, and then mixed with crystallized glass powder to prepare a composite glass paste as shown in FIG. First, in FIG. 15A, the predetermined material 51a is ceramic powder. Specifically, an inexpensive alumina powder (several hundred yen / kg) containing a large number of aggregates was used, although it was commercially available with a particle size of 0.4 μm. Next, to this alumina powder, a solvent (such as butyl carbitol acetate) and a dispersant in a small amount (the details of the oil absorption amount are described in JIS-K5101) and a dispersant slightly exceeding the oil absorption amount are added. A low-viscosity alumina slurry was prepared and dispersed for a predetermined time using a commercially available bead mill. The beads were made of alumina and had a diameter of 2 mm. The alumina thus dispersed was evaluated using a grindometer (details are described in JIS-K5600), but aggregates of 3 μm or more were not detected. Further, the particle size distribution of this alumina slurry was measured using a particle size distribution meter, but it was found that even the primary particles were solved.

  In addition, since a predetermined amount of dispersant is added to the alumina slurry, the concentration is high (alumina content is 60 wt% or more, desirably 80 wt% or more), but the viscosity is 10 poise (from the slip rate 1 / s). The dispersion was stable after 24 hours despite being less than (in the range of 100 / s). When this alumina slurry is stored for a long period of time, it can be prevented from settling and re-aggregation of the ceramic powder if it is properly rotated while being set on a rotating stand or the like. The alumina slurry thus prepared could be filtered without any problem with a filter having an opening of 10 μm as shown in FIG. Next, as shown in FIG. 16 (A), the alumina slurry is weighed with other binders and the like for glass powder, premixed as shown in FIG. 16 (B), and finally, FIG. 16 (C). Roll kneading as shown in FIG. The composite glass paste thus prepared could be filtered with a 20 μm filter without any problem. Thereafter, the viscosity is adjusted and completed. The viscosity is preferably 10 poise or more and 10,000 poise or less. When the viscosity is outside this range, printing may be difficult and difficult.

  The load sensor shown in FIG. 15 was prepared using the composite glass paste thus prepared, and there was no particular problem. When the cross section was analyzed by SEM or XMA, as shown in FIG. 12, alumina uniformly dispersed in the crystallized glass 46 was detected.

  Thus, when ceramic powder is dispersed without agglomerates, man-hours are inevitably required, so that it is desired to disperse efficiently at a lower cost. In such a case, by making the ceramic slurry high in concentration, the dispersion efficiency can be increased and the manufacturing cost can be reduced. In particular, it is important to disperse the ceramic slurry at a high concentration (for example, a solvent amount from 10% increase to 50% increase in oil absorption). When the solvent addition amount is less than the oil absorption amount, the viscosity of the ceramic slurry is too high and it is difficult to disperse the beads. In addition, when the oil absorption is increased by 100% (that is, when the oil absorption measured according to JIS is 50 g, the solvent should be 100 g), the viscosity of the ceramic slurry is sufficiently low, but the amount of solvent is too large, In the ink composition, the amount of solvent may increase excessively, and it may be necessary to reduce the solvent in the middle.

(Embodiment 13)
In the thirteenth embodiment, the metal elastic body used in the present invention will be described in more detail.

  In the present invention, crystallized glass is fired on a metal elastic body, so that the heat resistance is naturally excellent, but further proof stress is required. For such applications, an austenitic alloy, an age-hardening alloy (particularly a Ni-based alloy), a Co-based alloy (for example, Inconel), or a ferritic or austenitic heat-resistant steel may be used as needed. For example, by using a ferrite heat-resistant alloy, the strain sensor can be easily processed with high accuracy, and residual stress, processing strain, and the like after the processing can be reduced. Further, if aluminum is contained inside the metal elastic body as necessary, the aluminum added during the heat treatment forms an aluminum oxide film on the metal elastic body, and heat resistance and oxidation resistance can be improved. Furthermore, by adding alumina to the crystalline glass, there is an effect of greatly improving the adhesive strength (peeling strength) in the heat treatment with the metal elastic body containing aluminum.

  As such a member, SUS304, 316, 404, 430, S17, and 444 are suitable as AISI (American Steel Institute) type 300 and 400 series alloys.

  Moreover, even if it is another metal elastic body, it can also be used by heat-treating the surface. Such alloys have a fixed coefficient of thermal expansion in the range of 90 × 1E-7 to 140 × 1E-7 / ° C. depending on their composition. However, when processed into a predetermined shape like a load sensor, strain remains in the metal elastic body during the processing, and when glass paste is printed and baked on this, it deforms beyond the original thermal expansion coefficient. I often do.

  For example, the metal substrate is greatly warped more than the amount of warpage of the substrate calculated from the difference between the thermal expansion coefficients of the glass and the metal substrate. For example, even when the same SUS material is used, even when the thickness is 0.5 mm, when the thickness is 2 mm, and when the thickness is 5 mm, the amount of warpage is processed even when processed into a predetermined substrate shape by pressing (punching). Will change. In particular, even if the same metal material is used, if the thickness changes, the processing method naturally changes, and this warping amount is the same metal composition, but the processing pressure, method, procedure, Changes due to differences in molds. Further, as a countermeasure against such warp, the warp can be corrected after processing, but even if the warp is corrected, it is difficult to make the residual stress and the deformation amount during the heat treatment zero.

  Therefore, when a load sensor is formed on the surface of a metal substrate that is processed into such a predetermined shape and remains stressed, it is necessary to match the thermal expansion coefficient of the glass according to such warpage. In that case, as proposed in the present invention, by making a crystallized glass or an amorphous glass into a composite glass, the thermal expansion coefficient can be finely adjusted while using the same material. However, it can be easily handled on the glass side.

  In addition, as for the thickness of the metal elastic body used for this invention, 1 mm or more is desirable. As shown in FIG. 11, when the thickness of the metal elastic body is less than 1 mm, the metal elastic body side bends with a weak force, so even if the difference in thermal expansion coefficient between the glass and the metal elastic body is slightly shifted. The glass layer is difficult to break or peel off. On the other hand, in the case of the present invention, even when the thickness of the metal elastic body is 1 mm or more (or when it is difficult to deform such as 5 mm or more), the difference in thermal expansion coefficient between the glass and the metal elastic body can be adjusted.

  Thus, there is no upper limit on the thickness of the metal elastic body used in the present invention. However, practically, if the thickness of the metal elastic body exceeds 1 mm, the heat treatment apparatus at the time of forming the glass layer becomes expensive, and the heat unevenness at the time of heat treatment also increases, so the thickness of the metal elastic body should be 1 mm or less. desirable. However, there is no particular problem as long as the uniformity of the heat treatment is possible (preferably within a set temperature ± 5 to 10 ° C.).

  In addition, the heat resistance can be improved by adding aluminum to a metal elastic body. This is because the added aluminum diffuses on the surface during firing of glass or the like and forms an alumina layer. At this time, by adding alumina to all of the crystallized glass layer and the amorphous glass layer, the alumina becomes a common element of each other layer, and the adhesion and matching between the layers can be improved. Durability can be improved. The amount of aluminum (or alumina) added is preferably 1 wt% or more. If it is less than 1 wt%, the effect of addition may not be obtained. Similarly, by adding alumina as a common element to the crystallized glass layer, the amorphous glass layer, and the strain sensitive resistor, matching between the three layers can be improved. In addition, since the matching between the strain sensitive resistor and the amorphous glass layer can be improved, the electrical characteristics of the load sensor are stabilized.

  In addition, a part of the wiring is connected to the strain sensitive resistor to form a predetermined bridge circuit (Wheatstone bridge or the like), or a plurality of through holes formed in the crystalline glass layer and the amorphous glass layer Are electrically connected to each other via internal electrodes. As described above, by using the internal electrodes for the multilayer wiring, it is possible to cope with high-density mounting of chip parts and semiconductor parts, so that the load sensor can be reduced in size and cost.

Further, it is desirable that the wiring partially connected to the strain sensitive resistor has at least Ag of 60 wt% or more and less than 90 wt% and Bi ₂ O ₃ of 5 wt% or more and less than 30 wt%. When silver is less than 60 wt%, the wiring resistance becomes high, which may cause an error in the bridge circuit. Moreover, when silver is 90 wt% or more, even if it tries to solder, silver becomes easy to be eaten and the tensile strength after soldering may fall. If Bi ₂ O ₃ is less than 5 wt%, the adhesive strength to the substrate may be reduced. If it exceeds 30 wt%, excess Bi ₂ O ₃ may ooze out (or ooze out) around the wiring pattern. is there.

(Embodiment 14)
In the fourteenth embodiment, a method for forming a composite glass paste on a metal elastic body will be described. Although it is possible to directly form a composite glass paste on a metal elastic body using screen printing, such a paste is printed in a predetermined pattern on another supporting film (for example, a PET film). This can be transferred onto a metal elastic body and fired in an oxidizing atmosphere at 800 to 900 ° C. together with the metal elastic body.

  Further, such a composite paste is continuously formed into a sheet shape on a support film using a coating machine (forming device such as a ceramic green sheet), and after punching it into a predetermined shape with a mold, a metal elastic body It can also be transferred onto and fired in an oxidizing atmosphere of 800 to 900 ° C. together with the metal elastic body. As described above, by using the metal to be printed as a film from the metal elastic body, continuous printing becomes possible, and the printing cost can be reduced. Further, by forming a sheet, it is possible to prevent pinholes in the coating film, which has been a problem in printing, so that the product yield can be increased. For such effects, reference can be made to JP-A-2003-69192.

The strain sensitive resistor preferably contains at least 10 to 30 wt% PbO, 10 to 60 wt% SiO _2, and 5 to 30 wt% ruthenium compound. In the case of such a strain sensitive resistor, when PbO is less than 10 wt%, GF (corresponding to a change in a resistance value with respect to a gauge factor and strain, which is desirable with higher sensitivity) may be low. On the other hand, if PbO exceeds 30 wt%, the resistance value becomes too high, and even when a bridge circuit is assembled, it may be difficult to amplify with a semiconductor chip.

When SiO ₂ is 10 wt% or less, GF may be low. On the other hand, if it exceeds 60 wt%, the resistance value may be too high to match the input of the semiconductor chip to be used. When the ruthenium compound is 5 wt% or less, the resistance value may become too high or the GF may become small. In addition, when the content is 30 wt% or more, the resistor paste price becomes high, the resistance value becomes too low, making it difficult to match the input specifications of the semiconductor chip, or increasing the power consumption of the entire load sensor There is. Here, examples of the ruthenium compound include RuO, PbRuO compounds (Pb ₂ Ru _{2 O6.5,} etc.), PbBiRuO (Pb ₃ Bi ₂ Ru ₄ O ₁₄ , Pb ₄ Bi ₂ Ru ₃ O ₁₉ , PbBi ₂ Ru ₃ O _10, etc. ). Such a ruthenium compound is an example of a ruthenium compound optimized in the GF increase of the resistor, and is actually an example analyzed by XD (X-ray diffraction method) with a commercially available database (JCPDS). Other pyrochlore type (RuPbO type) can also be used for the application of the present invention.

  The thermal expansion coefficient of the crystallized glass is preferably 8 ppm / ° C. or more and less than 14 ppm / ° C. or the difference in thermal expansion coefficient from the metal elastic body is preferably less than 3 ppm / ° C. It may not be obtained. The crystallinity of the crystallized glass is desirably 20% or more. When the crystallinity of the crystallized glass is less than 20%, the characteristics as crystallized glass may not be obtained. The crystallization rate can be evaluated by X-ray diffraction.

  Note that it is desirable that all baking at 400 ° C. or higher is performed in an oxidizing atmosphere. This is because the organic components remaining in the glass or electrode paste are completely decomposed. When the firing atmosphere is not an oxidizing atmosphere, the organic components are not sufficiently decomposed and remain in the glass, which may reduce the reliability. Moreover, when the amount of air input to the firing furnace is insufficient, a reducing atmosphere is created, and particularly lead oxide or the like is reduced, which may deteriorate the characteristics.

  The load sensor and the manufacturing method thereof according to the present invention optimizes the glass material used for manufacturing the load sensor, enables multi-layer firing, and further expands the selection range of the metal elastic body. This is useful for smart air bags and the like.

Cross section of load sensor with single wiring layer Cross section of load sensor with multi-layered wiring Sectional drawing explaining how to manufacture a load sensor Sectional drawing which shows a mode that the amorphous glass and crystallized glass of this invention are co-fired Sectional drawing explaining a multilayer printing of crystalline glass paste on a metal elastic body Diagram showing multilayer printing using curable glass paste Diagram explaining how leveling works Sectional drawing which shows the case of firing multiple layers of internal electrodes and glass simultaneously Sectional drawing explaining a mode that these multiple layers are baked collectively with the crystallized glass which is an internal electrode as a base Sectional drawing explaining how these multiple layers are fired at once with the upper and lower sides of the internal electrodes sandwiched between crystallized glass The figure explaining the curvature of the metal elastic body used as a substrate Expanded schematic cross section of composite glass Diagram explaining how to change the coefficient of thermal expansion of composite glass Diagram explaining the problem of composite glass The figure explaining the manufacturing method of composite glass paste The figure explaining the manufacturing method of composite glass paste Sectional drawing explaining the structure of the conventional load sensor Explanatory drawing when the interface layer is formed Diagram showing an example of a load sensor with multiple layers of wiring Diagram explaining the problem when multiple layers are used

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Metal elastic body 2 Glass enamel 3a, 3b Wiring 4 Strain sensitive resistor 5 Glass layer 6 Interface layer 7a, 7b, 7c Lower glass 8a, 8b, 8c Upper glass 9a Internal electrode 10a Wiring 11a, 11b, 11c Protective layer 12 Sensing Strain resistor 13 Crack 14a, 14b Crystallized glass 15a, 15b Amorphous glass 16a, 16b Wiring 17 Strain sensitive resistor 18a, 18b Protective layer 19 Internal electrode 20a, 20b, 20c Crystallized glass paste 21a, 21b, 21c Crystal Crystallized glass 22 Electrode paste 23 Internal electrode 24 Amorphous glass paste 25 Amorphous glass 26 Wiring 27 Strain sensitive resistor 28 Crystallized glass paste 29 Amorphous glass paste 30a, 30b Crystalline glass 31a, 31b Amorphous glass 32a, 32b crystalline glass paper G 33a, 33b Dry crystalline film 34 Arrow 35 Uneven 36a, 36b Internal electrode 37a, 37b Cured glass paste 38a, 38b Cured glass paste 39a-39g Crystalline glass 40a, 40b, 40c Internal electrode paste 41a-41f Crystal Glass paste 42a, 42b, 42c Internal electrode 43 Crack 44 Glass layer 45 Warp 46 Crystallized glass 47 Ceramic powder 48 Composite glass 49 Ceramic powder aggregate 50 Pinhole 51a-51h Predetermined raw material 52 Dispersing device 53 Beads 54 Arrows 55a, 55b Filtration device 56a, 56b Filter
57a, 57b Container 58 Kneader 59 Stirring jig 60 Three rolls

Claims (3)

A metal elastic body containing 1 wt% or more of aluminum, ₃₀ to 55 wt% of MgO having an average particle diameter of 0.1 μm or more and less than 10 μm formed on the metal elastic body, 5 to 30 wt% of B ₂ O ₃ , SiO ₂ is 10 to 25 wt%, BaO is 5-25 wt%, alumina 1-30 wt%, CaO is 0~6wt%, ZrO _2, SnO ₂ had an average particle a crystalline glass powder is in the range of 0-5 wt% The crystalline glass powder by the weight ratio: ceramic, which is fired simultaneously with ceramic powder that is one or more of alumina, magnesia, zirconia, silica, calcium oxide, titanium oxide, and spinel having a diameter of 0.05 μm or more and less than 5 μm: the ceramic powder = 95: 5-60: and composite crystalline glass layer is 40, SiO ₂ is formed on the composite crystalline glass layer 40 to 80 wt%, An amorphous glass layer containing ₅ to 15 wt% CaO, 3 to 15 wt% PbO, 1 to 10 wt% alumina, and ₂ to 10 wt% ZrO ₂ , and a wiring pattern formed on the amorphous glass layer A load sensor composed of a strain sensitive resistor formed so as to be connected to a part of the wiring pattern,
The load sensor, wherein the composite crystalline glass layer and the amorphous glass layer are co-fired on the metal elastic body . The load sensor according to claim 1, wherein the strain sensitive resistor includes at least 10 to 30 wt% of PbO, 10 to 60 wt% of SiO2, and ₅ to 30 wt% of a ruthenium compound. MgO having an average particle size of 0.1 μm or more and less than 10 μm is 30 to 55 wt%, B ₂ O ₃ is 5 to ₃₀ wt%, SiO ₂ is 10 to 25 wt%, BaO is 5 to 25 wt%, alumina is 1 to 30 wt%, CaO is 0-6 wt%, ZrO ₂ , SnO ₂ is a range of 0-5 wt% crystalline glass powder and alumina having an average particle size of 0.05 μm or more and less than 5 μm , magnesia, zirconia, silica, calcium oxide, titanium oxide, A metal containing 1 wt% or more of aluminum in a composite glass paste in which ceramic powder which is one or more of spinels is dispersed in a resin solution in a weight ratio of the above glass powder: ceramic powder = 95: 5 to 60:40 A process for forming a composite glass layer by printing directly on an elastic body, followed by drying and baking in an oxidizing atmosphere of 800 ° C. or higher and lower than 900 ° C. And,
Forming an amorphous glass layer on the composite glass layer;
Forming a wiring pattern on the amorphous glass layer;
Forming a strain sensitive resistor so as to be connected to a part of the wiring pattern;
A load sensor manufacturing method comprising :
The method of manufacturing a load sensor, wherein the composite glass layer and the amorphous glass layer are simultaneously fired on the metal elastic body .

Images