Dependence of mechanical properties of electroporcelain on quartz particle size subjected to cyclic tests

The aim of the paper is the reveal effect of the quartz particle size on the mechanical properties of the electroporcelain subjected to cyclic test. This paper describes the influence of quartz on the mechanical properties of electroporcelain. The amount of quartz was controlled during sample preparation. Eight groups of samples were manufactured by extrusion with the addition of 0, 1, 2, and 4 wt.% of quartz and two different quartz grain sizes of 200 μm and 64 μm. Samples were fired at the temperature of 1320 °C, which is a common temperature for this type of ceramics. Once the nominal value of mechanical strength for each group was determined, the samples were cyclically tested using 50% of the nominal mechanical strength. The samples were tested by a 3-point bending machine for obtaining mechanical strength. The obtained values exhibited, that with an increasing amount of quartz, the mechanical strength decreased. On the other hand, the amount of quartz did not have a significant impact on cyclic tests. The influence of the amount of quartz on the value of mechanical strength after 100 and 500 cycles was negligible. However, the quartz grain size had a significant effect on the mechanical strength. The smaller grain size of impurities led to a lower degradation effect.


Introduction
The estimated remaining lifetime of installed insulators has been always studied, but in the last few years, more and more researchers have been studying this question, mostly in close collaboration with operating energy companies.Especially the KEPCO Research Institute and Sungkyunkwan University in South Korea have been actively publishing studies on different insulator aging mechanisms since 2019 [1][2][3].Insulator aging and replacement strategies have as well been studied by an IEEE working group [4].
The aforementioned studies have defined four main aging mechanisms, a) expansion of the cement, b) corrosion of the metal parts, c) mechanical and electrical stresses on the insulators core, and d) drying out of bitumen.Resistance against mechanical and electrical stresses is related to the microstructure of the ceramic insulators.
Today a huge majority of ceramic high voltage insulators are made of grades of C-120 Alumina Porcelain and C-130 High Strength Alumina Porcelain as specified by the IEC 60672 [5].Cristobalite and quartz porcelain are as well used as insulator materials in low-voltage distribution applications below 36 kV.Curiously the IEC 60672 specifies the grades only by the mechanical strength as new after production.Aging, fatigue resistance, and microstructure are not even mentioned.This is one reason why these insulator lifetime studies are so important nowadays.
The required mechanical strength of C-120 and C-130 can be achieved by various mixtures of raw materials, which then lead to different microstructures and different lifetime expectations between the manufacturers.The importance of the microstructure and its impact on the lifetime has been known for a long time, and several papers were published by Johannes Lieberman [6,7].Lieberman defines the ideal microstructure for the C-130, which should contain ≥ 40 % corundum, ≤ 15 % mullite, and residual quartz content of < 1%.The requirement of maximum quartz content is not included in the IEC 60672, but many manufacturers have this requirement in their technical specifications.
A. Rawat and R. S. Gorur [8] tested 30 years old insulators and they were able to establish clear relations between the lost mechanical and dielectric strength and microstructure.The results confirm the theory of Lieberman's ideal microstructure at the Power Frequency Puncture Test: the failed samples were having quartz crystals size > 50 μm and high general quartz content.
Keekeun Kim et al. [9] have demonstrated the role of corundum content as a major element in increasing resistance to aging.They were able to establish a model between the corundum content and a resistance to aging and predicted the tensile strength of a 43-year-old cap & pin insulator.
The quartz in the C-130 (high-strength porcelain) should be entirely replaced by aluminum oxide (alumina, Al2O3).The alumina is partly coming from the fired clay and chamotte and the rest is added as purified alumina.Impurified bauxite can replace pure alumina, but the bauxite contains a lot of impurities, mainly iron and titan oxides, which are giving a brownish or grey color to the porcelain otherwise white.Further, the negative impact of impurities on the mechanical strength must then be compensated by extra alumina oxide content in the fired product.As the manufacturers try to minimize the use of alumina, which is the most expensive mineral used in the receipt, there is an increased risk of residual quartz.
The residual quartz might be originated from the raw material mixture itself, just by the simple fact that alumina content is not high enough.Even when the alumina content is optimized there might appear residual quartz when the particle size is too big.The particles do not have time to melt into the glass matrix when they are too big.The particle size can be high when the milling time is short or has been shortened for cost savings.
Another reason for big particles could be intentional addition of coarse grain fractions to the body.The coarser grains fraction reduces shrinkage during drying, which can prevent sample damage.This is important for casting or extrusion.The isostatic process, or dry process, does not need a long drying step and the particle size is smaller than in the plastic process, which makes it easier to control the residual quartz content.
Finally, the residual quartz might come from an inapt firing curve.The curve might be too short for melting the quartz particles.And on the other hand, the firing temperature should be as low as possible so the time cannot be compensated for by a higher temperature.Then the cooling should be fast above 1000 °C and slower at the quartz (β → α) transformation zone.All these parameters can have an impact on the presence of residual quartz and its grain size.
According to Liebermann [6], reduction of quartz in the receipt, formation of early eutectic melt phases, increasing the alumina content to above 60%, low firing temperature, and fast cooling above 1000 °C reduce the residual quartz.G. Fassbinder [10] explains that the great disadvantage of the quartz is the (β → α) transition at 573 °C, which is causing about 0.7 vol.% shrinkage during the cooling [11].At this temperature, the solidified glass matrice resulting from the feldspar cannot follow the shrinkage of the quartz grains.The effect is high radial tension on the quartz grains and a tangential pressure on the surrounding glassy phase [12].The material can support this if the grains are small enough, but the large grains will not withstand the tension and gets microcracks [13], which is then reducing the porcelain strength.
The existence of microcracks is known to reduce the resistance to aging.Especially when the insulators are submitted to cyclic stresses, both tension and bending, like on the overhead lines, on circuit-breakers, or high insulator columns exposed to strong wind loads.During cyclic stresses, the microcracks start to propagate and the insulators are slowly losing their strength.In applications where the insulators are only under permanent compression, there is no significant crack propagation.
Based on these facts, the investigation of the influence of residual quartz content on mechanical properties and aging is required.The aim of this paper is to determine the effect of the amount of residual quartz and fraction on the modulus of rupture of electroporcelain mixture grade C-130 subjected to cyclic tests.

Materials and methods
The idea of this research was to assess the impact of defects caused by quartz grains on the mechanical properties of electroporcelain.For this reason, the porcelain mixture was prepared in PPC Insulators at Cab (Slovakia) according to standard industrial C-130 basic receipt, from feldspar (26 wt.%), chamotte (20 wt.%), clay (20 wt.%), and alumina (34 wt.%).The raw materials were wet milled to an average 5 µm particle size.After milling the slurry passed through fine sieves to remove coarse dirt and strong magnets to eliminate all potential iron impurities.After that, the granulate was made by spray drying process where the slurry was sprayed directly into hot air, and instantly dry granulate droped to the transport belt to the silos.
The quartz particles (64 µm and 200 μm) were added into the base mixture separately using wet milling in the following amounts: 1 wt.%, 2 wt.%, and 4 wt.%.As a reference sample mixture with addition of 0 % of quartz was used.The sample body was prepared by a laboratory extruder with Ø10 mm and length of ~150 mm.After drying the bars were manually glazed by dipping.After glaze drying, the samples were fired at a temperature of 1320 °C which was verified by pyrometric rings.
The modulus of rupture (R) was determined by the 3-point bending machine LabTest 2.025-2 according to IEC 60672-2 [14].At first, the nominal load was obtained by cracking 10 samples.Based on this value, each group of samples was cycled with 50% of the nominal value.In the frame of the present research 100 and 500 cycles were done with a strain rate of 20 N/s and an outer support span of 100 mm.After the breakage, the test bars were inspected to be sure that the breakage was not caused by an external defect in the glazing.

Results and discussion
The mineralogical analyses were executed by X-ray diffraction (XRD).The results are shown in table 1 and table 2. The main mineralogical phases are corundum, mullite, and amorphous (glassy) phase.The input materials naturally contain some quartz impurities, which is the reason for 0.86% and 0.89% of their content in the samples without added quartz.The test bars with 0 wt.% added quartz contained an average of ~0.88 % of quartz.During the comparison of fractions of 64 μm and 200 μm, it was evident that the samples made from 64 μm fraction after thermal treatment contained less quartz.One possible explanation is that the finest fraction dissolves better.However, the difference in quartz content between fractions except for 0% added quartz is ~ 50%.Another reason for these results is the accuracy of measurement.A. Ali et al, [15] state that, the detection limit of XRD is about 4 wt.% of the sample.
Figure 1 describes the dependence of added quartz amount on the modulus of rupture.At a glance, it can be seen that added quartz negatively affects mechanical properties.Figure 1 shows the nominal value of R depending on the amount of added quartz and grain size.The finer quartz grains caused a less intense decrease in R in comparison with 200 μm quartz.One percent of 200μm quartz led to a ~ 30 % decrease in the modulus of rupture.However, 64 μm grains resulted in a decrease in R by only about 5 %.Another decreases occur with the increase in the added quartz content.The decrease in the value of R was not linear.The biggest drop was observed after adding 1% of quartz.An additional increase in quartz content did not lead to a similarly significant effect as in the aforementioned case.However, with increasing quartz content, the modulus of rupture decreased.After obtaining the nominal value of R for each group of samples, cyclic tests were performed.During the cyclic testing, the samples were loaded by 50% of the maximal load from the nominal value of R.After 100 cycles, the samples were broken to acquire R.The results of the 100 cycles test are presented in figure 2. From these obtained results, it can be concluded that the decrease of R is strongly dependent on the grain size.While R decreased by ~30 % after 1 % of additional quartz with a fraction of 200 μm, the smaller grain size, 64 μm, caused a decrease at the same amount of additional quartz content only by ~5 % after the 100 cycles.A similar effect was observed after 500 cycles (see figure 3).This phenomenon can be described by the different thermal expansions of quartz grain and the amorphous phase during cooling.These differences cause the formation of microcracks which weaken the structure of the sample.If the size of crystals is below 20 μm, according to G. Fassbinder [10] quartz grains are harmless.A closer look at the data reveals, that in most cases, the error bars are bigger for 200 μm quartz fraction.A possible explanation can be that the bigger grains are unevenly distributed which causes a disproportion in the amount of defects in the examined sample.This can cause a value dispersion.The IEC 60672-3 [5] stated that the minimal value of flexural strength for this material is 160 MPa of glazed samples.Samples after 100 and 500 cycles showed almost the same result as the nominal.The differences are smaller than the measurement uncertainty.

Conclusions
The modulus of rupture of electroporcelain samples grade C-130 with additional quartz content subjected to cycle test was studied.It was found that: • additional quartz negatively affects the modulus of rupture, however, the grain size determines by how much.1% of additional quartz decreased the modulus of rupture by 30% if 200 μm quartz grains are added but only ~5 % if the quartz grain size was 64 μm, • the quartz content of 2% and 4% has a small effect comparing to 1% of added quartz, • the difference in thermal expansion of the amorphous phase and quartz grain size of 5 μm are small enough to weak the samples, • 100 and 500 cycles of 50% nominal load not shown any significant influence of cycling on the modulus of rupture in this range,

Figure 1 .
Figure 1.Nominal value of R of samples.

Table 1 .
The mineralogy analysis after the firing process (200 μm quartz) in wt.%

Table 2 .
The mineralogy analysis after the firing process (64 μm quartz) in wt.%