Available online at www.sciencedirect.com
CERAMICS INTERNATIONAL
Ceramics International ] (]]]]) ]]]–]]] www.elsevier.com/locate/ceramint
The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories A. Mertken, C.G. Aneziris Institute of Ceramic, Glass and Construction Materials, TU Bergakademie Freiberg, Agricolastraße 17, 09596 Freiberg, Germany Received 4 August 2014; received in revised form 15 September 2014; accepted 16 September 2014
Abstract Modern steel casting plants require functional refractory components such as monobloc stoppers, submerged entry nozzles, and ladle shrouds. The functional flow control components in steel casting plants are often made of carbon bonded alumina refractories containing approximately 30% residual carbon after coking. This study investigated the mechanical and thermo-mechanical behavior of different functionalized Al2O3–C materials with additions based on alumina nanosheets, carbon nanotubes, semiconductive silicon, and a carbon content reduced to 20%. Furthermore, the curing temperature and the mixing order of the raw materials were altered. By optimizing the curing and mixing conditions of the samples that included all the additives, high residual strengths (with absolute values of up to 14.51 MPa before thermal shock, after the first thermal shock, 12.11 MPa, and after the fifth thermal shock, 13.87 MPa, and a relative values of up to 19.81% and 4.29% respectively) could be recorded after thermal shock treatment. & 2014 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Keywords: B. Nanomaterial; C. Semiconductor; D. Graphite; D. Alumina; E. Refractory
1. Introduction With high throughputs and extreme demands on plant materials, modern steel casting plants require high-quality, functional refractory components, such as monobloc stoppers, submerged entry nozzles, and ladle shrouds. High corrosion resistance against both steel and slag melts as well as high thermal shock performance and oxidation resistance at elevated temperatures are typical properties of these functional components. These properties are achieved through specific combinations of materials and specially designed microstructures [1]. One class of refractory materials that fulfills these requirements is the carbon-containing and carbon bonded refractories. The combination of carbon and oxides reduces the material's wettability by slags and molten metals, and also results in a n Corresponding author. Tel.: þ49 3731 393927, þ49 3731 392505; fax: þ 49 3731 392419. E-mail addresses:
[email protected] (A. Mertke),
[email protected] (C.G. Aneziris).
higher thermal shock resistance due to the reduced thermal expansion of the composite and its improved thermal conductivity [3]. The functional flow-control components in steel casting plants are often made of carbon bonded alumina refractories containing approximately 30% residual carbon after coking [2]. One disadvantage due to the flow of molten steel through Al2O3–C refractories is the so-called ‘carbon pick-up’ during steel casting. Deep decarburized steels have a tendency to dissolve the carbon of refractory linings, which has a negative impact on their properties [4–8]. As already mentioned carbon bonded alumina refractories contain 30% residual carbon after coking. With regard to the ‘carbon pick-up’ the carbon content should be reduced. However, the properties of carbon bonded alumina refractories should be maintained. Previous work has demonstrated that admixture of semiconductive silicon has a positive effect on the properties of refractories. This could be attributed to the property of semiconductive silicon to transfer electrons to the macro-molecules of the carbon-based binding phase [19]. With
http://dx.doi.org/10.1016/j.ceramint.2014.09.090 0272-8842/& 2014 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
2
this transfer a higher amount of carbon in the samples after coking has been registered. With this higher carbon content inside the binding phase, resulting in a lower porosity the mechanical properties such as the cold modulus of rupture and thermal shock resistance could be improved [19,20]. Another approach is the use of additives on the nanometer scale [21]. Tamura et al. [22] and Aneziris et al. [3] proposed the use of nanoscale material additions in refractory applications as early as 2003. In the past, the use of nanoscale materials in carbon bonded refractories was hindered by their extremely high cost of production, as well as difficulties in achieving homogenous distributions in the mixtures. In recent years, the production of nanoscale materials has become increasingly cost effective. Furthermore, a range of developments in process engineering has facilitated the production of greater volumes of nanoscale materials [23]. Further investigations have also shown that the addition of minor amounts of nanoscale materials can improve the performance of the refractories [23–36]. Roungos and Aneziris [1] explored the addition of alumina sheets in combination with carbon nanotubes in Al2O3–C materials. Moreover, it has been demonstrated that mechanical and thermo-mechanical properties are improved by the formation of the Al3CON phase and amorphous whiskers with platelet structures [14,37,38]. This approach has been also investigated for the MgO–C system by Li et al. [39] and Zhu et al. [40]. Another disadvantage of carbon bonded alumina refractories is the oxidation of carbon, resulting in the loss of positive properties. To improve the oxidation resistance of carbon bonded refractories, antioxidants are incorporated in the mixtures. In general, fine metallic powders (Si, Al) in the range of up to 150 mm are used as antioxidants, as well as carbides (SiC, B4C, Al4SiC4, Al8B4C7) and boron-containing oxides (fluxes) in the range of up to 100 mm [9–16]. With the addition of Si, for example, the CO (g) of the surrounding atmosphere forms SiO (g). During this reaction, the CO (g) is reduced to C (s) [13]. In a subsequent step, the formed SiO reacts again with the CO (g) to form deposits as thin protective layers of SiO2 (s) on the graphite surface [14,17]. This reaction causes a volume expansion, which can close the pores of the refractories [9,18]. If boron is used, this effect may take place due to the formation of a glass phase
[11,13,16]. Furthermore, a combination of different antioxidants may be used to further improve the performance of the refractories [10,16]. In the present work, the mechanical and thermo-mechanical behavior was investigated for different functionalized Al2O3–C materials with additions based on alumina nanosheets, carbon nanotubes, and semiconductive silicon. For this purpose, model compositions with and without alumina grains were investigated. The aim of this contribution was to investigate compositions with excellent mechanical and thermomechanical behavior. Another aim was to increase the carbon binder yield after coking in further investigations. A further purpose was to determine the influence of the curing temperature and the mixing order of the raw materials on the properties of carbon bonded alumina refractories. 2. Experimental For the model experiments without alumina grains, the following raw materials were used: natural graphite with very fine grains (99.50 wt%o40 mm, d50=8.50–11.00 mm) and with a carbon content of 90–96 wt%; flaked, coarse-grade graphite (95.00 wt%471 mm, d50=140 mm) with 87–98 wt% carbon content, both produced by Graphit Kropfmühl AG, Hauzenberg, Germany; and fine metallic silicon powder (Elkem, Oslo, Norway) of high purity (99.50 wt%o150 mm, d50=17.30 mm). In addition to these raw materials, different nanoscale materials and a silicon semiconductor as listed in Table 1 were used. In all experiments, a liquid phase (PF 7280 FL 01) and a powder phenolic resin (0235 DP), both from Momentive Specialty Chemicals, Iserlohn, Germany, were added as binders. Furthermore, hexamethylenetetramine (Momentive Specialty Chemicals, Iserlohn, Germany) was used as a curing agent. In the first step, the model compositions without alumina grains were produced to investigate interactions between the binder, the graphite, and the additives. Therefore, the graphite and the novolak powder resin were premixed in a paddle mixer (ToniMix, Toni Technik Baustoffprüfsysteme GmbH, Model 6209) for 3 min. Next, the nanoparticles and the liquid resin were mixed for 3 min. Partial agglomeration of the raw materials occurred during mixing due to the absence of coarse grains. In Table 2, the compositions of the different model mixtures are listed.
Table 1 Special additives (nanoscale materials and semiconductor). Additives
Producer
Abbreviation
Purity (wt%) Average particle size
Carbon nanotubes (C)
Timesnano (China)
TN
Z 95.00
Alumina nanosheets (α-Al2O3) Silicon
Sawyer (USA)
AS
Silchem, (Germany)
Si (special) resp. Si1
Outer Inner diameter diameter 450 nm 5 15 nm 95.0099.80 10 250 nm –
o64 mm
Specific surface area (m2/g) Length 440.00 10 20 mm
Specific resistance Doping (Ohm cm) –
–
940
–
–
–
1.201.30
Phosphorous
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
After mixing, discs with a diameter of 15 mm and a thickness of 8 mm were pressed with a hand press (Holzmann-Maschinen, Dornpresse DOP 2000) at 30 MPa. In Table 3, the compositions of the mixtures with alumina grains are presented. For the experiments with the alumina grains, white fused alumina (99.46 wt% Al2O3, 0.45 wt% Na2O, d50 ¼ 76.40 mm) with a maximum grain size of 0.2 mm (Treibacher Industrie AG, Althofen, Austria) and coarse-grade tabular alumina (99.50 wt% Al2O3, max. 0.4 wt % Na2O, d50 ¼ 0.35 mm) with a maximum grain size of 0.60 mm (Almatis GmbH, Ludwigshafen, Germany) were used. The production method was used for all Al2O3–C samples and is comparable with standard industrial practice, being close to the mixing range described in [7]. At first, all basic materials were mixed together in an Eirich compulsory mixer (Maschinenfabrik Gustav Eirich GmbH & Co KG, Hardheim, Germany) at room temperature for 1 min. After mixing the graphite, alumina, metallic silicon and resin, a premix of the nanoparticles with the doped Si was added. The two nanoscale additives and the doped Si were blended in a separate vessel in advance. After mixing for 5 min, the mixtures were uniaxially pressed to bar-shaped samples (25 25 150 mm3) at 100 MPa using a RUCKS press (RUCKS Maschinenbau GmbH, Glauchau, Germany).
3
Two different curing conditions were investigated in this research. The pressed samples with and without alumina grains were cured in a dryer (Model TR 120 from Nabertherm GmbH, Lilienthal, Germany) with curing condition I. They were heated from ambient conditions at a heating rate of 1 K/min up to 200 1C, and the curing temperature was maintained for 2 h. In Section 3.3, the samples were heated under curing condition II in steps (in 20 min up to 85 1C, and held at this temperature for 90 min; in 30 min up to 120 1C, and held at this temperature for 120 min, in 150 min up to 180 1C) up to 180 1C, and held at that temperature for 90 min. Afterwards, all samples were coked for 5 h at a temperature of 1000 1C under a reducing atmosphere (in a retort filled with coke breeze) in a furnace (Model LH 15/14; Nabertherm GmbH, Lilienthal, Germany). This temperature was chosen based on the work of Roungos and Aneziris [1] and Stein [19]. The heating rate was 3 K/min. Subsequently, the samples with and without alumina grains were investigated with the aid of an environmental scanning electron microscope (ESEM; Philips Type XL30). During the evaluation of the SEM analysis, some special structures were detected. To evaluate the consistency of these structures, electron backscatter diffraction (EBSD) analyses were carried out with the aid of a Philips XL30 equipped with an EBSD
Table 2 Model mixtures without alumina grains. Raw materials
Fine graphite Coarse graphite Novolac liquid Novolac powder Metallic silicon Hexa. Carbon nanotubes Alumina nanosheets Si (special)
Compositions (wt.-%) I Reference
II R20-Si-TNAS
III R20-Si
IV R20-TNAS
30.68 30.68 6.14 12.27 18.41 1.84
30.30 30.30 6.06 12.12 16.67 1.82 0.91 0.30 1.52
30.67 30.67 6.13 12.27 16.87 1.84
30.77 30.77 6.15 12.31 16.92 1.85 0.92 0.31
1.53
Table 3 Compositions of the Al2O3–C samples. Raw materials
Fused alumina Tabular alumina Fine graphite Coarse graphite Novolac liquid Novolac powder Metallic silicon Hexa. Carbon nanotubes Alumina nanosheets Si (special)
Compositions (wt.-%) Ia Reference
IIa R20-Si-TNAS
IIIa R20-Si
IVa R20-TNAS
29.10 38.90 10.00 10.00 2.00 4.00 6.00 0.60
29.10 38.90 10.00 10.00 2.00 4.00 5.50 0.60 0.30 0.10 0.50
29.10 38.90 10.00 10.00 2.00 4.00 5.50 0.60
29.10 38.90 10.00 10.00 2.00 4.00 6.00 0.60 0.30 0.10
0.50
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
4
system TSL from Edax/Ametek. Measurements of the bulk density and the extent of the open porosity of the coked samples were carried out according to DIN EN 993-1 by using toluol as an intrusion liquid. Furthermore, the mechanical properties of the samples were determined by measuring their cold modulus of rupture (CMOR) according to DIN EN 993-6 (TIRA Test 28100, TIRA GmbH, Schalkau, Germany). Based on the measurements of the CMOR, the static Young's modulus was calculated. The thermal shock resistance was determined with compressed air according to DIN EN 993-11. Here, the CMOR was measured after the first and, for some samples, after the fifth thermal shock (with 5 samples for each mixture). The pore size distribution was identified by means of a mercury porosimeter (Pascal 140/440, Thermo Electron Corporation, Karlsruhe, Germany). These measurements were realized on the samples after coking, as well as after the first thermal shock.
graphite and the silicon particles were homogeneously distributed inside the samples. Whiskers were formed through the addition of semiconductive silicon, Fig. 1(b). It was not possible to detect the doped silicon after the coking process. An energy-dispersive X-ray spectroscopy microanalysis (EDAX) revealed that these whiskers consisted of silicon, carbon and oxygen, indicating that these phases were most likely SiC whiskers. However, a special feature of normal SiC whiskers is that they exhibit acicular growth, while the observed whiskers tended to form clouds without elongated shapes. With the addition of carbon nanotubes and alumina sheets, acicular SiC whiskers were formed, as shown in Fig. 1(c). Finally, by adding the doped Si and the nanoparticles, the SiC whiskers grew. In this case, the growth of whisker inside the pores of the samples could also influence the strength of the products.
3. Results and discussion
Table 4 lists the physical as well as mechanical and thermomechanical properties of coked samples with alumina grains based on curing condition I. The highest open porosity was observed for the samples of the mixture IIa, while the lowest was noted for mixture IIIa. With respect to the reference composition, the mixtures containing nanoscale material exhibited higher degrees of open porosity. However, the addition of doped Si alone led to the lowest values. This leads to the ability of phosphor doped silicon to emit electrons. Due to this ability it
3.1. Samples without alumina grains cured according to curing condition I In Fig. 1(a), the microstructure of the reference sample without alumina grains is presented, whereby the light white spots denote the silicon particles and the flat gray parts are the amorphous carbon (position I) and the crystalline graphite (position II). The
3.2. Samples with alumina grains based on curing conditions I
Fig. 1. ESEM micrographs of the fracture surfaces of samples without alumina grains: reference (mixture I, BSE picture), magnification 5000 (a); semiconductive silicon (mixture III), magnification 5000 (b); nanoparticles (mixture IV), magnification 5000 (c); semiconductive silicon and nanoparticles (mixture II), magnification 5000 (d). Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
5
Table 4 Properties of the samples with the alumina grains. Compositions
Ia (R20) IIa (R20-Si-TNAS) IIIa (R20-Si) IVa (R20-TNAS)
Properties Open porosity (%)
Bulk density (g/cm3)
Average pore radius (mm)
CMOR
CMOR1TS
Strength loss (1 TS) (%)
20.747 0.28 22.057 0.13 19.347 0.22 21.937 0.84
2.4970.00 2.4670.01 2.5170.00 2.4270.04
0.95 0.97 0.97 1.21
6.607 0.37 6.147 0.17 6.807 0.20 6.637 0.34
5.8470.41 6.2670.33 6.8570.28 6.6370.49
11.51 2.06 0.71 0.05
Fig. 2. Pore size distributions of coked samples with alumina grains based on curing condition I.
Fig. 3. SEM micrographs of fracture surfaces of coked samples with alumina grains: reference, magnification 5000 (a); Reference, magnification 10000 (b); R20-TNAS, magnification 5000 (c); R20-TNAS, magnification 10000 (d). Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
6
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
influences the carbon yield as well as the graphitization of the binder carbon. This resulted in decrease of the open porosity. Another result of this influence is the highest CMOR in comparison to the other compositions. In the future the carbon yield of the samples has to be detected to investigate the influence of the semiconductor on the carbon yield. The CMOR values of all compositions were low and in the range of 6 MPa. Normally, the CMOR values of compositions with 30 wt% residual carbon after coking achieve the 8 MPa level. After experiencing thermal shock, all functionalized compositions exhibited very good performance. In particular, the combination of nanoscale additions with the doped Si led to an increase of approximately 2%, whereby the achieved CMOR value of 6 MPa was, again, not sufficient when compared to state of the art materials. This could be attributed to curing condition I and the mixing procedure, which led to high pore sizes. Fig. 2 presents the pore size distribution of the coked samples with alumina grains based on curing condition I. Based on this curing condition, the addition of nanoparticles (R20-TNAS) led to the highest content (approximately 35 vol% in comparison to 16– 18 vol% for all other compositions) of pores in the range of 1000–2000 nm. In Fig. 3(a) and (b), the microstructure of the reference composition is presented. Carbon flakes (position I) as well as the glassy carbon matrix (position II) can be identified. Individual whiskers were formed in some spaces. The small number of
formed whiskers was very long and thin, and an accumulation of whiskers could be discerned. With the addition of nanoparticles, the formation of whiskers increased, Fig. 3(c) and (d). In particular, whiskers formed in the spaces between some grains and on the tops of graphite grains. In some spaces, the whiskers did not grow as often nor did they grow as long as they did in other parts of the sample, which would indicate that the formation of whiskers was hindered. The formation of longer whiskers was only observed at the edges of the particle. The whisker growth in these areas and at this coking temperature (1000 1C) was probably related to a vapor–solid (VS) and/or a vapor–liquid– solid (VLS) reaction [41–45]. In the images of sample IIIa (R20-Si), only a formation of agglomerated whiskers/fibers in the carbon bonded matrix was identified (Fig. 4). Based on EDX analysis, small amounts of Fe were also registered inside these agglomerates. Fe impurities were detected in the used resin. Individual alumina grains were observed on top of the agglomerates. There was no formation of an individual whisker in the entire sample. Furthermore, and in comparison to the images of the samples without alumina grains, no networks of short whiskers could be detected. With the addition of alumina grains, the silicon and carbon particles are further dispersed, and thus the formation of a uniform network of whiskers is prevented. Another difference between these samples was the content of the binder. Without alumina grains there is a higher content of
Fig. 4. SEM micrographs of fracture surfaces of coked samples with alumina grains: samples with semiconductive silicon R20-Si, magnification 3000 (a) and R20-Si, magnification 10000 (b).
Fig. 5. SEM micrographs of fracture surfaces of coked samples with alumina grains: samples with semiconductor and nanoparticles R20-Si-TNAS, magnification 5000 (a) and R20-Si-TNAS, magnification 10000 (b).
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
binder resulting on blowing up of the binder. This blowing up leads to increased porosity values. By combining nanoparticles and doped-Si, whiskers grew and formed networks all over the sample, especially at places were a high concentration of carbon was measured with the aid of EDX (Fig. 5). Whiskers also formed in some of the pores. However, no tight networks could be detected on the sample. To identify the possible reasons for the thermal shock behavior of mixture IIa after the first thermal shock, some SEM images were generated (Fig. 6).
7
During the thermal shock, whiskers of the mixture IIa (R20Si-TNAS) grew in bigger areas (Fig. 6(a)) and inside some cracks (Fig. 6(b)). They were also covered by amorphous carbon and grew between the grains. Furthermore, the formation of SiC whiskers with a length of around 1 mm and a thickness of around 300 nm (Fig. 6(c)) influenced the thermal shock behavior. These whiskers probably grew during the heating process before the thermal shock, and prevented crack propagation. The balls on top of the whisker surfaces contain Si and O. This formation is also investigated by Li [46]. 3.3. Change of the curing temperature (curing conditions II) and the mixing order
Fig. 6. SEM micrographs of fracture surfaces of coked samples with alumina grains after the first thermal shock: agglomeration of SiC whiskers 10000 (a); formation of whiskers inside of cracks 5000 (b); and SiC whiskers of about 1 mm 20000 (c).
In comparison to the results of previous research by Roungos and Aneziris [1] and Stein and Aneziris [20], the CMOR values achieved were very low. The main differences between these contributions are the curing temperature used and the mixing order of the raw materials. In order to overcome this discrepancy in this part of this contribution, both the curing temperature and the mixing order were adjusted according to [1,20]. The mixing order of the latter [20] had two main differences from the mixing order adapted in Section 3.2. The first was the addition of the liquid resin, it was added after the two alumina grain types and premixed for approx. three minutes. The other difference compared to the mixing order of Stein and Aneziris [20] is the addition of the nanoparticles. They were added after the liquid resin. With this order, improved distribution of the nanoparticles in the batch could be achieved. In the last step, the fines were introduced into the mix and the complete batch was mixed for an additional 5 min. Moreover, a curing temperature of 180 1C was selected in further trials, as already proposed by Roungos and Aneziris [1]. In a final approach, the influence of the new mixing procedure as well as the new curing temperature was investigated. With respect to Section 3.3, the nanoparticle and silicon semiconductor materials were added in the same mass as described in Table 3. To detect any possible effect of the curing temperature, some of the samples were cured at 180 1C and some at 200 1C. Like the samples in Section 3.2, these samples also had the same mixing order as the samples in Section 3.2. The samples with the new mixing order were marked with the number 5 behind the composition name. The composition of the reference material R20 is related to a commercially available mix with a reduced graphite content of 20 wt%. One significant difference for the samples could be observed in the median pore diameter (Table 5). The median pore diameter of the samples of Section 3.2 with a curing temperature of 200 1C was higher than the median pore diameter of the samples with a curing temperature of 180 1C. By changing not only the curing temperature to 180 1C, but also the mixing order, the open porosity values and the median pore diameter decreased, and the pore sizes were shifted to lower pore classes (Fig. 7). The values of the CMOR before and after thermal shock are presented for the new mixing and curing approaches, and are listed in Table 6. In Section 3.2, the CMOR of all compositions achieved values of around 6 MPa. The CMOR of the
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
8
reference material (R20) only increased with the change of the mixing order, to 10.31 MPa. This result was compatible with the decrease in the levels of open porosity. Equivalent results were observed for the mixtures with the addition of doped Si (R20-Si). The samples with the lowest open porosity achieve the highest CMOR values. This result could be attributed to the change in the mixing order, which led to the more homogenous growth of whiskers. However, in comparison with the reference composition, the CMOR was already increased by the change in the curing temperature, while the highest CMOR
Table 5 Physical properties of samples with different additives, mixing order and curing temperatures. Compositions
R20_180 1C R20_5_180 1C R20-Si_180 1C R20-Si_5_180 1C R20-TNAS_180 1C R20-Si-TNAS_5_200 1C R20-Si-TNAS_5_180 1C
Physical properties Open porosity (%)
Bulk density (kg/m3)
Median pore diameter d50 (mm)
21.2470.31 20.0871.06 20.9670.35 19.0671.00 20.9170.86 19.8870.36 18.9670.79
2.487 0.01 2.507 0.03 2.477 0.01 2.527 0.02 2.477 0.03 2.507 0.01 2.527 0.04
0.77 0.50 0.79 0.51 0.82 0.43 0.63
Fig. 7. Pore size distribution of samples cured at 200 1C and 180 1C.
of 13.20 MPa was achieved with the new combined curing/ mixing conditions. Furthermore, in the case of the nanoscale additions, the new curing approach led to higher CMOR values in comparison to the old curing temperature, whereby the R20Si-TNAS_5_180 1C sample with additions and changes in both the curing and mixing orders led to the highest CMOR values after coking (14.51 MPa) as well as after thermal shock (12.11 MPa). Comparing the changes in relative strength loss, the best results were achieved with the samples based on the R20-Si-TNAS_5_200 1C composition and processing procedure. The CMOR were also measured after the fifth thermal shock, which was only made using the samples with the highest CMOR before thermal shock and the lowest strength loss after thermal shock. Here the samples with the addition of the semiconductor (R20-Si_5_180 1C) again experienced a decrease in CMOR. This could have been caused by the formation of cristobalite, which was already shown in Section 3.2. After the fifth thermal shock of the samples with the carbon nanotubes, the alumina sheets and the semiconductive silicon, and with a curing temperature of 200 1C, there was also a decrease in the CMOR. Comparing the samples with the same additives and the same mixing order, the influence of the thermal shock resistance of the curing temperature could be observed. After the fifth thermal shock of the samples with the carbon nanotubes, alumina nanosheets, semiconductive silicon, and the curing temperature of 180 1C, there was an increase of the CMOR. This suggests that the formation of phases increased during the thermal shock, and in particular during the heating-up process. This could have been a formation of new SiC whiskers or the formation of the Al3CON phases, as mentioned by Roungos et al. [37]. Therefore, with the combination of all of the changes of the reference material, the thermal shock resistance could be increased. The SEM micrographs of the two samples with nanoparticles and doped Si based on new curing/mixing combinations before as well as after the thermal shock are presented in Fig. 8. With a curing temperature of 180 1C, there was a huge formation of whiskers above the entire surface of the sample not exposed to thermal shock (R20-Si-TNAS_5_180 1C). The whiskers grew in agglomerations between the individual grains and at the edges of pores and cracks. There were also whiskers that grew between the individual agglomerations (Fig. 8(a)). In (Fig. 8(b)), the homogenous distribution of the whiskers can be seen. The growth of whiskers occurred in particular at the
Table 6 Mechanical properties before and after thermal shock with different additives, mixing orders, and curing temperatures. Compositions R20_180 1C R20_5_180 1C R20-Si_180 1C R20-Si_5_180 1C R20-TNAS_180 1C R20-Si-TNAS_5_200 1C R20-Si-TNAS_5_180 1C
Properties CMOR (MPa)
CMOR1TS (MPa)
Strength loss (1TS) (%)
CMOR5TS (MPa)
Strength loss (5TS) (%)
6.6970.52 10.3170.15 9.7070.40 13.2071.13 10.4170.16 10.5670.91 14.5170.26
6.1870.18 6.9570.36 – 9.6671.53 9.9370.54 10.3770.46 12.1170.81
8.19 48.32 – 36.56 4.87 1.83 19.81
– – – 8.01 7 0.38 – 8.2070.76 13.8970.96
– – – 39.30 – 22.35 4.29
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
9
Fig. 8. SEM micrographs of fracture surfaces of coked samples with different curing temperatures and mixing order with and without thermal shock. Agglomeration of SiC whiskers (sample R20-Si-TNAS_5_180 1C) 5000 (a); growth of SiC whiskers on the surface of the sample (R20-Si-TNAS_5_180 1C) 1000x (b); SiC whiskers of approx. 5 mm (sample R20-Si-TNAS_5_180 1C_1TS) 5000x (c); growth of SiC whiskers on the surface of the sample (R20-SiTNAS_5_180 1C_1TS) 1000 (d); entangled SiC whiskers (sample R20-Si-TNAS_5_200 1C) 5000 (e); accumulation of whiskers on the surface of the sample (R20-Si-TNAS_5_200 1C) 1000 (f); agglomeration of whiskers (sample R20-Si-TNAS_5_2001C_1TS) 5000x (g); and points of agglomeration of whiskers on the surface of the sample (R20-Si-TNAS_5_200 1C_1TS) 1000 (h).
10
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
edges of graphite grains. The whiskers exhibited a length of up to 5 mm and a diameter of several nm. After a thermal shock, it seemed that the density of the whisker network decreased, while the length of the whiskers increased up to 7 mm (Fig. 8 (c)). The number of whisker agglomerations on the surface seems to decrease after thermal shock (Fig. 8(d)). The reason for the decrease in the number of whiskers could be the oxidation of reactive whiskers during the thermal shock and the formation of SiO2. This could not be discerned during the SEM observations because of the amorphous structure of SiO2. Through this reaction, the level of oxidation of graphite could be reduced, and a better thermal shock performance as achieved with the new mixing/curing conditions was possible. Many factors influenced the thermal shock performance of the refractory. Furthermore, the formation of longer whiskers was helpful in forming links with other whiskers. With the change of the curing temperature, the whiskers became thinner and longer (Fig. 8(e)), with some of the whiskers forming hooks that facilitated their entanglement with adjacent whiskers. While there was a high degree of whisker formation on the surface (Fig. 8(f)), surface coverage of the sample decreased after the first thermal shock (Fig. 8(h)). The whiskers formed in areas with low density were shorter and had no distinct linkage (Fig. 8(g)). The formation of whiskers depended on many parameters, such as the coking temperature, the concentration of the involved species, the partial pressure of CO (g), and the porous structure [46–48]. The influence of the curing temperature on whisker formation could not be discerned. In comparison to the results of Section 3.1, whisker formation increased. With the changing of the mixing order, an increase in the level of whisker growth was possible, as investigated by other authors [49]. The change of the curing temperature increased the formation of the resite lattice during polymerization. This in turn led to increased and homogenous formation of a glassy carbon lattice during pyrolysis. This, among other factors, increased the thermal shock resistance, which was further influenced by the mixing order. With a homogenous distribution, whisker formation increased and the distribution of the antioxidants was improved. Therefore, carbon oxidation decreased during the thermal shock. To determine the phases exactly, an EBSD analysis was made of the samples R20-Si-TNAS_5_180 1C and R20-SiTNAS_5_200 1C. It revealed that the formed phases were cubic SiC or β-SiC (Fig. 9). This phase formation depended on the surrounding atmosphere during coking, the pressure of the surrounding gas, and the level of impurities [50]. There also seemed to be a difference between the thicknesses of the whiskers formed. With the curing temperature of 200 1C, thinner whiskers could be formed. With the curing temperature of 180 1C, the whiskers were thicker and formed a closer network. This must be confirmed by further research and measurements in the future, but there was a slight tendency for the curing temperature to influence whisker thickness during formation. The formation of Al3CON phases could not be discerned. The quality of the EBSD was not sufficient for automated phase identification, one reason being the undefined
Fig. 9. SEM micrograph of whisker formation on sample R20-Si-TNAS_5_180 1 C (a) and the EBSD from the point marked (b), with identification as β-SiC(c); SEM micrograph of a whisker formation on sample R20-Si-TNAS_5_200 1C (d) and the EBSD from the point marked (e), with identification as β-SiC(f).
geometry. It is also not possible to determine the Al3CON phases with an XRD analysis due to the insufficient degree of their deposition. Furthermore, Roungos et al. [37] describe the
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
difficulty in determining this phase. It is possible that there may have been a reaction between the nanoparticles that resulted in the formation of Al3CON phases, but their content was too low to confirm this. 4. Conclusion In this contribution, the influence of different additives such as nanoparticles and semiconductors on the properties of Al2O3–C refractories as a function of curing temperature and mixing conditions was investigated. By optimizing the curing and mixing conditions in samples with both additives, high residual strengths (with absolute values of up to 14.51 MPa before thermal shock, after the first thermal shock, 12.11 MPa, and after the fifth thermal shock, 13.87 MPa, and a relative values of up to -19.81% and -4.29% respectively) were registered after thermal shock treatment. They were also the only samples that exhibited increases in CMOR after the fifth thermal shock. Therefore, the curing temperature as well as the mixing order had a positive influence on the thermo-mechanical properties of the refractory materials. With the aid of EBSD, cubic SiC formation was identified for both favored compositions. In the future, the growth and the interlocking of SiC whiskers in the favored compositions will be investigated in more detail, as will be their impact on creep resistance. Acknowledgment The authors would like to thank the "DFG (Deutsche Forschungsgemeinschaft), Germany" for funding this work within the Priority Program 1418 (Grant number AN 322/ 16-2). The authors would also like to thank Dr. G. Schmidt for the SEM examinations, and Dr. H. Berek for the EBSD examinations. The authors would like to thank A. McDonnell for proof reading the article. References [1] V. Roungos, C.G. Aneziris, Improved thermal shock performance of Al2O3–C refractories due to nanoscaled additives, Ceram. Int. 38 (2012) 919–927. [2] R. Amavis, Refractories for the steel industry, Elsevier Science Publishers Ltd., England, 1989, p. 47–58 (S). [3] C.G. Aneziris, D. Borzov, J. Ulbricht, Magnesia-carbon bricks: a highduty refractory material, Interceram. Refract. Manual (2003) 22–27. [4] A. Baulig, Die Vermeidung der Wiederaufkohlung bei der Herstellung von ULC-Stählen, EUR (Luxembourg),2001. [5] G. Kunz, Ladle refractories for clean steel production, RHI Bull. 2 (2010) 30–40. [6] W. Pluschkell, C. Knoche, R. Berger, Wiederaufkohlung von Stahlschmelzen niedrigen Kohlenstoffgehaltes durch Stranggießpulver, Stahl Eisen 112 (2) (1992) 81–84. [7] R. Scheel, Wiederaufkohlung von kohlenstoffarmen Stahlschmelzen durch Feuerfeststoffe sowie Giessschlacken, Abschlussbericht. Technische Forschung Stahl, Kommission der Europäischen Gemeinschaften, 14485, 1993. [8] B. Xu, et al., Study on the oxidation kinetics of in situ β-sialon bonded Al2O3–C refractories, Adv. Mater. Res. (2014) 1021–1025.
11
[9] M.N. Khezrabadi, J. Javadpour, H.R. Rezaie, R. Naghizadeh, The effect of additives on the properties and microstructures of Al2O3–C refractoriesk, J. Mater. Sci. 41 (2006) 3027–3032. [10] S. Zhang, A. Yamaguchi, A comparison of Al Si and Al4SiC4 added to Al2O3–C refractories, in: Proceedings of the Unified International Technical Conference, USA, 1997, pp. 861–869. [11] H. Sunayama, M. Kawahara, T. Mitsuo, K. Sumitomo, The effect of B4C addition on the oxidation resistance of Al2O3–C and Al2O3–SiC–C refractories, in: Proceedings of the Unified International Technical Conference, USA, 1997, pp. 841–849. [12] W. Vieira Jr., B. Rand, The nature of the bond in silicon-containing alumina–carbon refractory composites – Part I, in: Proceedings of the Unified International Technical Conference, USA, 1997, pp. 831–840. [13] A. Yamaguchi, Self-repairing function in the carbon-containing refractory, Int. J. Appl. Ceram. Technol. (2007) 490–495. [14] S.W. Zhang, Next generation carbon-containing refractory composites, Adv. Sci. Technol. 45 (2006) 2246–2253. [15] E. Luhrsen, A. Ott, Immersion nozzles for metal melts, United States Patent 5.171.495, 15 December 1992. [16] T. Wang, A. Yamaguchi, Antioxidation behavior and effect of Al8B4C7 added to carbon-containing refractories, J. Ceram. Soc. Jpn. 108 (9) (2000) 818–822. [17] A. Schwechheimer, Glasuren zum Oxidationsschutz von kohlenstoffhaltigen feuerfesten Bauteilen für den Stahlstrangguss, Dissertation, Clausthal-Zellerfeld (2000). [18] A. Yamaguchi, S. Zhang, J. Yu, Effect of refractory oxides on the oxidation of graphite and amorphous carbon, J. Am. Ceram. Soc. 79 (9) (1996) 2509–2511. [19] V. Stein, Contribution to the characteristic improvement of carbon bonded doloma refractories by addition of functional ceramic materials, Freiberger Forschungshefte, vol. 905, Techn. Univ. Bergakademie Freiberg, Freiberg, Sachs, 2011. [20] V. Stein, C.G. Aneziris, Low-carbon carbon bonded alumina refractories for functional components in steel technology, J. Ceram. Sci. Technol. 5 (2) (2014) 115–124. [21] M.J. Pitkethly, Nanomaterials – the driving force, Mater. Today 7 (2004) 20–29. [22] S. Tamura, T. Ochiai, S. Takanaga, T. Kanai, H. Nakamura, Nano-Tech refractories – 1: The development of the nano structural matrix, in: Proceedings of the Unified International Technical Conference on Refractories, Japan,2003, pp. 517–520. [23] C.G. Aneziris, S. Jin, Y. Li, V. Roungos, Interactions of carbon nanotubes in Al2O3–C refractories for sliding gate applications, in: summary Booklet (Abstracts) of the Unified International Technical, Brazil,2009, p. 8. [24] C.G. Aneziris, U. Klippel, W. Schärfl, V. Stein, Y. Li, Functional refractory material design for advanced thermal shock performance due to titania additions, Int. J. Appl. Ceram. Technol. 4 (6) (2007) 481–489. [25] H. Hattanda, T. Yotabun, T. Tsuda, T. Ochiai, S. Tamura, Nano-Tech. refractories – 7: application of nano-structured matrix to SN plates, in: Proceedings of the Unified International Technical Conference on Refractories, Germany,2007 pp. 204–207. [26] K. Haren, K. Morikawa, J. Yoshitomi, K. Asano, Improvement of thermal spalling resistance of alumina–graphite materials by nano-technology, in: Proceedings of the Unified International Technical Conference on Refractories, Germany,2007 pp. 358–361. [27] S. Takanaga, Y. Fujiwara, M. Hatta, T. Ochiai, S. Tamura, Nano-Tech. refractories – 3: development of MgO-rimmed MgO–C brick, in: Proceedings of the Unified International Technical Conference on Refractories, USA,2005, pp. 148–151. [28] T. Matsui, K. Goto, Y. Yamada, N. Taki, Characteristics and applications of nano-tech. magnesia carbon bricks, in: Proceedings of the Unified International Technical Conference on Refractories, USA,2005, pp. 176–179. [29] S. Takanaga, T. Ochiai, S. Tamura, T. Kanai, H. Nakamura, Nano-Tech Refractories – 2: the application of the nano structural matrix to MgO–C bricks, in: Proceedings of the Unified International Technical Conference on Refractories, Japan, 2003, pp. 521–524.
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090
12
A. Mertke, C.G. Aneziris / Ceramics International ] (]]]]) ]]]–]]]
[30] Y. Shiratani, T. Yotabun, K. Chihara, T. Ochiai, S. Tamura, Nano-Tech Refractories – 4: the application of the nano structural matrix to SN plates, in: Proceedings of the Unified International Technical Conference on refractories, USA, 2005, pp. 575–578. [31] A. Sen, B. Prasad, J. Sahu, N. Sahoo, J. Tiwari, Effect of nano-oxides and anti-oxidants on corrosion and erosion behavior of submerged nozzle for longer sequence casting of steel, in: Summary Booklet (Abstracts) of the Unified International Technical, Brazil, 2009, p. 8. [32] M. Braulio, M. Rigaud, A. Buhr, C. Parr, V. Pandolfelli, Spinelcontaining alumina-based refractory castables, Ceram. Int. 37 (2011) 1705–1724. [33] S. Dutta, P. Das, A. Das, S. Mukhopadhyay, Significant improvement of refractoriness of Al2O3–C castables containing calcium aluminate nanocoatings on graphite, Ceram. Int. 40 (2014) 4407–4414. [34] T. Lipinski, E. Drygalska, C. Tontrup, The influence of additions of nanostructured Al2O3-powder on the high temperature strength of high alumina refractories, in: Summary Booklet (Abstracts) of the Unified International Technical, Brazil, 2009, p. 12. [35] V. Arasu, S. Adak, A. Chattopadhyay, C. Kamath, Influence of nanoadditives on thermo-mechanical properties of alumina castables, in: Summary Booklet (Abstracts) of the Unified International Technical, Brazil, 2009, p. 24. [36] Q. Wang, Y. Li, M. Luo, S. Sang, T. Zhu, L. Zhao, Strengthening mechanism of graphene oxide nanosheets for Al2O3–C refractories, Ceram. Int. 40 (2014) 163–172. [37] V. Roungos, C.G. Aneziris, H. Berek, Novel Al2O3–C refractories with less residual carbon due to nanoscaled additives for continuous steel casting applications, Adv. Eng. Mater. (2012). [38] T. Zhu, Y. Li, S. Sang, S. Jin, Y. Li, L. Zhao, et al., Effect of nanocarbon sources on microstructure and mechanical properties of MgO–C refractories, Ceram. Int. 40 (3) (2013) 4333–4340.
[39] Y. Li, M. Luo, S. Jin, S. Sang, L. Zhao, Microstructural evolution of multi-walled carbon nanotubes in the presence of mixture of silicon and silica powders at high temperatures, Ceram. Int. 38 (2012) 4105–4110. [40] T. Zhu, Y. Li, M. Luo, S. Sang, Q. Wang, L. Zhao, et al., Microstructure and mechanical properties of MgO–C refractories containing graphite oxide nanosheets (GONs), Ceram. Int. 39 (2013) 3017–3025. [41] H. Peng, X. Zhou, H. Lai, N. Wang, S. Lee, Microstructure observations of silicon carbide nanorods, J. Mater. Res. 9 (2000) 2020–2026. [42] G.G. McMahon, J. Carpenter, T. Malis, On the growth mechanism of silicon carbide whiskers, J. Mater. Sci. (1991) 5655–5663. [43] R.S. Wagner, W. Ellis, The vapor- liquid-solid mechanism of crystal growth and its application to silicon, Trans. Met. Soc. AIME 233 (1965) 1053–1064. [44] P.C. Silva, J. Figueiredo, Production of SiC and Si3N4 whiskers in Cþ SiO2 solid mixtures, Mater. Chem. Phys. 72 (2001) 326–331. [45] G. Yang, R. Wu, Y. Pan, J. Chen, R. Zhai, L. Wu, et al., Direct observation of the growth process of silicon carbide nanowhiskers by vapor–solid process, Physica E 39 (2007) 171–174. [46] Y. Li, C.G. Aneziris, X. Yi, S. Jin, N. Li, Formation of dumbbell-shaped β-SiC whiskers in Al2O3–ZrO2–C composite refractories, Interceram. Refract. Manual (2005) 20–23. [47] Q. Li, C. Wang, S. Tie, Synthesis of bamboo-like SiC whiskers from waste silica fume, Cryst. Res. Technol. 49 (2014) 290–297. [48] R. Wu, G. Yang, Y. Pan, J. Chen, R. Zhai, L. Wu, et al., Prism-shaped SiC nanowhiskers, J. Alloys Compd. 453 (2008) 241–246. [49] Y. Li, X. Chen, Y. Li, S. Sang, L. Zhao, Effect of multiwalled carbon nanotubes on the thermal conductivity and porosity characteristics of blast furnace carbon refractories, Metall. Mater. Trans. A 41 (2010) 2383–2388. [50] H. Salmang, H. Scholze, R. Telle, Keramik, 7th ed., Springer, Berlin, New York, 2007.
Please cite this article as: A. Mertke, C.G. Aneziris, The influence of nanoparticles and functional metallic additions on the thermal shock resistance of carbon bonded alumina refractories, Ceramics International (2014), http://dx.doi.org/10.1016/j.ceramint.2014.09.090