To develop a technique for joining dense SiC by reaction-sintering (self-bonding), first an optimum condition was searched for preparing bulk body of SiC by reaction-sintering. Mixture of SiC and carbon powder was pressed at room temperature and siliconized at 1450°C in vacuum for 30min. The sintered body had a bulk density of 3.06g/cm³ and 4-point bending strength of 53kg/mm² (520MN/m²). Since about 10wt% of free Si contained in the reaction-sintered SiC evaporates above 1400°C, especially in vacuum, it is to be noted that the sintered body is obtainable by siliconizing from 1400°C to 1450°C, lower than the conventional technique. A pair of sintered bodies was also joined under the same condition. Furthermore, the powder mixed with polyfurfuryl alcohol was found to be a good joining agent. The resulting joining strength was 34kg/mm² in average. Microscopic examination of the joint showed that it is important to set the joining couple firmly during the siliconizing.

View full abstract

By cooling a number of B₂O₃-Li₂O melts containing 96-20mol% B₂O₃ from the liquidus to room temperature, glass-forming regions have been determined under various cooling rates; Q₁=1.2×10-3, Q2=9.1×10^-3, Q₃=7.9×10^-2, Q₄≅3×10^-1, Q₅≅2, Q₆≅7, Q₇≅50, Q₈2×10³ and Q₉≅10⁵K/s. The glass-forming region, which is restricted in a narrow range of high content of B₂O₃≥86mol% at Q₁, extends continuously with increasing rate of cooling, and glass can be obtained in a range of B₂O₃≥25mol% at the cooling rate of Q₉. An examination of the liquid viscosities in and out of the glass-forming regions in the system revealed that (1) increasing rate of cooling Q decreases the critical liquidus viscosity ηL of the melt and accordingly (2) the critical cooling rate Q for glass formation is dependent on the liQuidus viscosity ηL of the melt. In order to understand the slope of the log Q-logη_L or log Q-logη_L diagram in the system, the composition dependence of the fusion entropy ΔS_f was computed from the equation proposed for glass formation. It was found that η_L and 1/Q change monotonously with composition while Eη/T_L or ΔS_f goes through a maximum or a minimum.

View full abstract

Experiments have been carried out to understand the effects of several factors on the adherence of enamels. Microstructure of the steel-enamel interface of the fired enamel was also investigated. It is confirmed that enamelling steel (extra low carbon rimmed steel) containing 0.023% copper is suitable for the base of the direct-on enamel and also provides good adherence to the regular ground coat enamels. It is shown that the pretreatment is effective in most cases. Concentration profiles of iron across the steel-enamel interface show simple decreasing curves in all cases. There is no correlation between such a profile and the adherence index of the specimen. An “interlocking” structure was observed in the steel-enamel interface. During the firing process, the thickness of this structure increased probably by the action of local electric cells on the steel surface whose cathodes are composed of “smut” (pickling residue) and precipitates from nickel flashing or adherence-promoting agents. As the thick structure develops, the adherence index increased. In conclusion, the mechanical joint through the “interlocking” is the main contributor to the adherence of enamel within an area of practical conditions. Although several chemical reactions take place in the enamelling process, a chemical force mediated by the melting of the iron oxide is universal, but it may be less important.

View full abstract

The mixture of coarse alumina and fine zinc oxide, with different mole fraction of ZnO, X_Z, was heated at 1000°-1200°C for different time. The rate of ZnAl₂O₄ formation was determined, and also the microstructure was observed. The rate of ZnAl₂O₄ formation increased with the decrease of the composition X_Z in the range of X_Z<0.5, but its dependence on the composition was not observed in the range of X_Z>0.5. The microstructure showed the presence of the agglomerates of sintered ZnO, which resulted in the incomplete reaction in the range of X_Z<0.5, and consequently, the alumina particles were not able to react with ZnO because of incomplete mixing. The ZnAl₂O₄ layer grew around alumina particles at the rate of formation in the parabolic relation with time. The diffusion constant controlling the process was determined from the product layer thickness as 3.59×10^-15 to 3.16×10^-13cm²/s at 1000°-1200°C. The temperature dependence of the diffusion constant gave the activation energy of 85.4kcal/mol. The application of Jander's model to the powder mixture with different composition was discussed, and the introduction of three factors to correct the deviation of the practical condition from the Jander's ideal assumption was considered. The first is the factor (1-X_Z)/X_Z to correct the difference between the rate of reaction defined as the ratio of the amount of the product to the amount of the minor component in the mixture and the Jander's rate of reaction defined as the ratio of reacted particles among the main particles. The second is the factor (X'_A/X_A) which represents the ratio of the alumina particles take part in the formation of ZnAl₂O₄ to all alumina particles in the mixture. The third is the factor (X'_Z/X_Z) which represents the ratio of the zinc oxide particles which take part in the formation of ZnAl₂O₄ to all zinc oxide particles in the mixture. The corrected Jander's equation incorporating these factors is expressed as
[1-{1-(X_Z/1-X_Z)(X_A/X'_A)α}^1/3]²=k⋅t (1) or α={1-(1-√k⋅t)³}(X'_A/X_A)(1-X_Z/X_Z) (2)
which reduces to
α_max=(X'_Z/X_Z) (2')
The application of this Eq. (2) shows that in the range X_Z>0.5, the Jander's model is correct, but in the range X_Z<0.5, the correction factors are effective. The Eq. (2) was able to explain the increase of a with decreasing X_Z as a result of the introduction of the factor (1-X_Z)/X_Z and also the incomplete reaction in α-t relation in X_Z<0.5 as a result of the introduction of the factor (X'_Z/X_Z) in Eq. (2).

View full abstract

Relations among microstructure, bending strength and thermal expansion behavior of the aluminum titanate ceramic were investigated. The ceramics were prepared from three kinds of synthesized powders having several grain sizes and various amounts of free corundum. Under crossed nicols field of polarization microscope, the fired bodies showed domain free structures. Grain size of the titanate of specimens fired at 1500°C for 4h varied from 2.3 to 7.2μm depending on the starting powders. Many cracks were observed on their grain boundaries of bodies which had grain size larger than 3μm, whereas only a few cracks were observed in the body with grain size of 2.3μm. In the specimens fired at 1300°C for 2-8h, their grain sizes were less than 1μm, and few cracks were observed on their grain boundaries. Bending strength of the fired bodies increased markedly as their grain sizes decreased below 2.5μm. For bodies fired at 1300°C, their strength increased with increase of firing time and consequently increase of bulk density, and the highest value of 860kgf/cm² was obtained. Although the specimens fired at 1300°C exhibited high bending strengths, the bodies still retained a low thermal expansion coefficient, i.e. 4×10^-6deg-1 from R. T. to 1000°C. Presence of free corundum in the aluminum titanate powders markedly reduced growth rate of the titanate grains.

View full abstract

Two different hydrates of magnesium aluminum sulfate, a freeze-dried salt, MgAl₂(SO₄)₄⋅8-9H₂O, and pickeringite, MgAl₂(SO₄)₄⋅22H₂O which was obtained from hydration of the freeze-dried salt in air, were heated and the decomposition processes to spinel were compared. Both decomposed to form crystalline magnesium and aluminum sulfate anhydrides during dehydration and pickeringite produced well-crystallized sulfate anhydrides since it melted during dehydration. After decomposition of the aluminum sulfate anhydride to γ-alumina, magnesium sulfate anhydride decomposed and reacted with γ-alumina to form a spinel. The freeze-dried salt reacted completely to form a spinel at 1000°C; however the pickeringite left some unreacted magnesia as periclase at the same temperature. It was inferred that well crystallized sulfate anhydrides formed from the pickeringite tended to decompose at higher temperatures and did not react completely to form spinel.

View full abstract