Elsevier

Ceramics International

Volume 44, Issue 11, 1 August 2018, Pages 12055-12064
Ceramics International

Thermal shock resistance of refractory composites with Zirconia and Silicon-Carbide inclusions and alumina binder

https://doi.org/10.1016/j.ceramint.2018.03.217Get rights and content

Abstract

With the goal of designing a castable refractory for an aerospace application with optimum resistance to thermal shock, three different particle-reinforced ceramic composites are designed and compared. Different volume fractions of Silicon Carbide (SiC) particles, Zirconia (ZrO2) bubbles, and Zirconia solid particles dispersed in an alumina (Al2O3) matrix are used in the fabrication of these castables. Destructive and nondestructive testing procedures are implemented to evaluate their thermomechanical properties, both before and after a custom designed thermal shock cycle. After demonstrating the applicability of thermal shock indices, the variation of these indices due to thermal shock is measured experimentally and utilized as a design tool. Multiple micro-scale damage mechanisms were observed, all of which are various forms of structural deformation.

Introduction

Refractories can be made of various compositions as well as through different production processes, which in turn would lead to a diverse range of properties. Consequently, refractories can be classified in various ways based on each of their major properties [1]. Some important properties of refractories include: resistance to high temperatures and thermal shock, chemical inertness, resistance to mechanical load, resistance to corrosion, erosion and impact [2], [3]. From another perspective, refractories can also be classified based on how they are shaped and implemented in the structure. Some refractories come in pre-formed shapes, whereas some others can be shaped in situ and without joints so as to form an integral component [4], [5]. This latter group of refractories is called monolithic refractories [4]. One subgroup of monolithic refractories is castables [5]. Castables are composed of refractory grains, which are dispersed in a binding matrix. The castable, depending on the amount of this binding matrix material, can be a self-flowing castable or a vibration castable [6]. After addition of a suitable liquid (e.g. water), the solution is poured into the target location to form the desired refractory shape or structure after solidification due to chemical reaction [4].

Among different properties of interest for castables, thermal shock resistance is arguably the critical one, as castables are frequently subjected to severe thermal loads. In this work we examine the thermal shock resistance of castable refractories made of ZrO2 and SiC inclusions and hydratable alumina binders. In the remainder of this section, a brief summary of the current state of knowledge in thermal shock resistance of castables is presented, which provides the context for our work described in the subsequent sections.

The presence of a temperature gradient can give rise to thermal stresses in solid materials [7]. However, if this temperature gradient is applied suddenly, it can lead to “thermal shock” [7]. This temperature gradient can be the result of sudden heating, which leads to hot shock, or it can be caused by sudden cooling, which causes cold shock [8]. Thermal shock resistance can be defined as the ability of the material to withstand different forms of failure that may take place during rapid cooling or heating [7], [9]. Thermal shock resistance is not an intrinsic property of a material and it is strongly related to the size [10] and shape [11] of the material as well as duration and the method by which thermal gradient is applied [8], [9], [12]. Nonetheless, some of the properties of the solid that can affect its thermal shock resistance include coefficient of thermal expansion (CTE) [7], [8], [13], [14], thermal conductivity [7], [8], tensile strength [7], [8], modulus of elasticity [7], [8], toughness [8], thermal diffusivity [7] and Poisson’s ratio [7], [15]. For metals, thermal stresses may cause small plastic deformations; whereas in contrast, due to linear elasticity of ceramic materials, large stresses can be generated in response to thermal shock [16].

In order to analytically predict the resistance of homogeneous ceramics to thermal shock, multiple parameters have been proposed [7], [17]. Although as previously emphasized thermal shock resistance depends upon various factors, including the shape of the sample, the aforementioned thermal shock parameters are defined in terms of intrinsic properties of the material only. One such thermal shock parameter is R, which is defined as follows [7]:R=σf(ν)Eαwhere E, α, and ν represent elastic modulus, CTE, and Poisson’s ratio, respectively. Also, parameter σ can either represent tensile strength (σt) in the case of cold shock, or represent crushing strength (σc) in the case of the hot shock. The definition of the function f(ν) depends on the state of stress. The value of f(ν) can be equal to 1 or (1ν) or (12ν) for uniaxial, biaxial and triaxial stress, respectively. The parameter R describes the maximum allowable temperature gradient that the material can withstand without crack initiation, and it is used for hard thermal shocks where the Biot number is relatively high. Biot number can be defined as [8]:β=bhλwhere b is a thickness, h is the coefficient of surface heat transfer and λ is the thermal conductivity. In a similar fashion, the criteria for mild thermal shock, R, is defined as follows [7]:R=λσf(ν)Eα

The R parameter describes the maximum allowable heat flux, and it is used in cases where the Biot number is relatively low. Together, R and R parameters can provide us with a useful description about crack initiation in a ceramic material due to cold or hot thermal shock, both of which can take place either by hard shock or by mild shock regimes. From a different perspective, another category of thermal shock parameters are developed to describe the resistance of the ceramic to crack propagation. One such parameter is R defined as follows [7], [17]:R=Eσ2f(ν)

which from the perspective of strain energy, describes the resistance to spalling. As can be inferred from their corresponding equations, resistance to crack initiation and crack propagation are two qualities of the ceramics that are competing with one another, and depending on the application, one may be preferred to the other [17]. In general, at least two thermal shock parameters (or merit indices) need to be implemented to characterize the response of material to various conditions of thermal shock, and no single one of these parameters would be adequate independently [7]. Fig. 1 qualitatively describes how each pair of the aforementioned indices can evaluate certain aspects of thermal shock response of the material under examination, collectively.

Notice that these parameters, in conjunction with many other thermal shock indices are essentially established to portray the thermal shock resistance of homogeneous ceramics; thus, their validity for describing composite materials such as refractories can be considered as ad-hoc [18]. Also, they are not exploited for characterizing the thermal shock cycle (the temperature profile) applied to the ceramic materials. Moreover, for ceramic composites the crack initiation and crack propagation steps can be more intertwined; thus, implementing both groups of parameters simultaneously for characterizing the thermal shock response seems to be a reasonable approach, as has been done by other researchers [19], [20], [21]. With this background, in the present work we start by hypothesizing applicability of the thermal shock parameters for describing thermal shock resistance of our composites, and after proving the hypothesis, implement them as comparative design tools.

In the existing body of experimental research on thermal shock resistance of refractory materials, different quenching techniques have been designed for simulating cold shock. The specimen may be placed in contact with a cold metal rod, or immersed in a quenching medium. For the latter technique, different quenching media have been implemented, such as room-temperature water, boiling water, room temperature air, different types of oils and alcohols, and preheated salt [15], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32]. Hot shock is also simulated experimentally by means of a flame [33], [34], [35], [36], contact with molten metal bath [37] and many other methods [28], but most frequently by using different types of furnaces in the lab [37], [38], [39], [40].

Damage assessment of thermal shock can be performed by means of destructive as well as non-destructive testing procedures. Measuring changes in elastic modulus as a result of thermal shock, using either ultrasonic method or resonance method has been one of the most common nondestructive practices for characterizing thermal shock damage [37], [38], [39], [40], [41], [42]. Among destructive methods of characterizing thermal shock damage, wedge splitting, three-point-bending, and compression tests are the most common procedures for measuring strength, modulus of rupture (MOR), work of fracture and other properties [12], [21], [23], [24], [25], [31]. In addition to the previously mentioned destructive and nondestructive tests, Scanning Electron Microscopy (SEM), Energy Dispersive Spectroscopy (EDS), as well as other imaging techniques are frequently practiced to provide insight about microstructure, changes and failures due to thermal shock, and crack deflection mechanisms [19], [20], [21], [22], [23], [24], [36], [43], [44], or even implemented as an independent tool for characterizing thermal shock damage by inspecting morphology of cracks, density of crack, etc. [31], [45], [46], [47].

Thermal shock resistance of refractories may be improved from the perspective of crack initiation, as well as crack propagation [44], [48]. To prevent crack initiation, properties of the material need to be optimized to obtain higher thermal shock crack initiation indices [44], [48]. As an instance, increasing strength and decreasing elastic modulus would be helpful according to Eq. (1). Ye et al. [49] reported improvement in the strength of the hydratable alumina-bonded castable through the addition of magnesite aggregates, and Ceylantekin et al. [19] improved cold shock resistance of MgO-spinel by the addition of ZrO2 particles. On the other hand, to prevent crack propagation in the material, a pre-existing high concentration of short cracks may be one solution [44], [48]. Miyaji et al. [21] reported an escalation in fracture energy of alumina-based refractory castables due to addition of eutectic aggregates which had higher capacities in crack deflection. Martinović et al. [41] investigated the influence of heat treatment on different properties including crushing, flexural and tensile strength of refractories. Soboyejo et al. [50] improved cold shock resistance by addition of Na2O which enhanced bridging mechanisms in an aluminosilicate refractory. Other researchers designed multilayered refractory ceramics with crack deflection mechanism at their interfaces, and in comparison to monolithic ceramics, observed improvement in resistance to crack penetration caused by thermal shock [36], [43], [44]. From a different perspective, self-healing refractories may also unravel a new repertoire of solutions in the future [3], [51].

In order to design our castables however, we have turned our attention to the microstructure and properties of the constituents, so as to optimize the response of the refractory composite to both damage initiation and damage propagation. Through a number of trials, three compositions were identified for further investigation in this research.

Porosity is one of the factors that seems to be always present in the structure of castables to various extents, and researchers have long been investigating the role of porosity and its impact on various properties including thermal shock response. In these studies, porosity has been artificially created using different techniques, including the addition of various types of fugitive inclusions such as rice starch, corn starch, and potato starch in different procedures as pore-forming agent [27], [52], [53], sintering of the ceramic powder [27], or gelation of the slurry [54]. Different experiments are then performed on such porous samples before and after thermal shock so as to characterize their properties. Yuan et al. [27] implemented fugitive inclusions as well as sintering method, and showed that the value of elastic modulus and thermal conductivity is independent of the method of porosity generation, and only depends on the volume fraction of porosity. Although not mentioned in their publication, this perhaps means that in their experiments they did not detect any influence from the shape of the porosity on the properties as well. However, they reported that the fracture properties were dependent upon the porosity generation procedure. Jin et al. [55] studied the thermal shock behavior of porous ZrB2-SiC ceramics, and discovered that both porosity volume fraction and pore size can influence MOR, fracture toughness, and thermal shock damage resistance in different ways. Saho et al. [30] established that fracture toughness, elastic modulus, and MOR will decrease with increase in porosity. Jin et al. [55] concluded that most porous samples displayed higher thermal shock residual strength than the dense samples, and findings of Chong and Yansheng [56] agree with this. Conversely, Shao et al. [30] did not find the introduction of porosity to be a reliable technique for enhancing the thermal shock response in their specific ceramic, due to the large reduction in mechanical properties. While it is not the focus of the current investigation, it seems that there exists plenty of room for further exploration of different aspects of the presence of porosity in castables. In general, immersion in water seems to be a regular practice in the literature for measuring the volume fraction of porosity [19], [20], [41], [42]. Consequently, we measure porosity in our refractories using the immersion method, and assimilate it to our calculations.

In many refractory composites alumina is used as the matrix material, since it benefits from reliable refractory properties such as high degree of hardness, melting point, and good electrical and thermal insulation properties [47]. Specifically, Low Cement Castables (LCC) which contain lower contents of calcium and finer grains of alumina, benefit from superior thermal shock resistance, as opposed to Conventional Castables (CC) [6], [47]. The binder used in our castable refractory composites is a low cement alumina as well. Moreover, SiC and ZrO2 particulates are used as reinforcements with varying volume fractions. Thermal conductivities, CTEs, and elastic moduli of SiC and ZrO2 are significantly different. As a result, thermomechanical properties of the resulting castables can be tailored by appropriate choices of the volume fractions of the particulate inclusions.

Motivated by the potential use of castables in aerospace applications, we started with the hypothesis that thermal shock parameters developed for monolithic ceramics can also adequately characterize the thermal shock response of our composite ceramics. If this is the case, then the thermal shock resistance of castables with SiC particulate inclusions should be superior to that of castables with ZrO2 inclusions. To test this hypothesis, we manufactured three refractory composites containing different volume fractions of ZrO2 and SiC inclusions, dispersed in an alumina matrix. These refractories were designated as “Castable Zirconia” (CZ), “Castable Silicon-Carbide” (CS), and “Castable Silicon-Carbide Zirconia” (CSZ). Next, by measuring their properties and thus thermal shock indices, the validity of our hypothesis was examined. Experimental results showed an enhancement in the room temperature thermal shock indices due to addition of SiC, and thus, validated our hypothesis. This allowed us to take this idea one step further by inspecting variations in damage initiation and propagation indices due to thermal shock, and utilizing the results to identify the most suitable composition. This was accomplished by applying an identical custom designed thermal shock cycle to the samples of different composite types, and measuring their properties before and after the exposure to the thermal shock cycle. For this purpose, elastic modulus, crushing strength, MOR, and CTE of the samples were measured using the ultrasonic method, compression tests, three-point-bending tests, and dilatometry, respectively. Additionally, via SEM studies, multiple modes of failure were detected in micro-scale, all in the form of structural deformation. To the best knowledge of the authors, aside from preliminary results of our research [57], refractory composites similar to the ones in this work have never been compared, and experimental investigation of variation in the thermal shock indices due to thermal shock and implementation of those results as a design tool have also not been attempted. In the succeeding sections, an overview of our experimental procedures is presented, which is followed by the analysis of the results obtained from the experiments and conclusions drawn based on the analysis.

Section snippets

Materials and methods

Aluminum molds were manufactured to cast cylindrical as well as bar shaped samples. Mold release was used prior to casting samples. Batches were prepared using a 6 quart Kitchen Aid mixer to blend ingredients (dry and wet) for several minutes before casting. By implementing this procedure, three different compositions of samples were prepared with proportions that are detailed in Appendix A. In brief, the first composition named CZ, was mainly composed of calcia-stabilized Zirconia both in the

Thermomechanical properties

For all of the cylinders, the volume and mass of the samples were measured pre-shock, as well as about one day after thermal shock for the samples that were exposed to it. As a result, the average density of CZ, CS and CSZ were found before and after thermal shock. These values are tabulated in Table 1. Additionally, a permanent mass loss was detected in all of the refractory compositions with any of the two different shapes (cylinders and bars), as a result of the thermal shock cycle. The

Conclusions

In the present work, we have started with the hypothesis of applicability of thermal shock indices for designing refractory composites. On this basis, three different compositions of refractory castables were designed, characterized, and compared in terms of their room temperature mechanical properties as well as resistance to a specific thermal shock cycle. These comparisons were performed using destructive and nondestructive testing procedures, including compression, three-point-bending,

Acknowledgments

This work was supported in part by an STTR Grant (NNX14CK10C) from NASA, United States.

References (63)

  • S. Marenovic et al.

    Thermal shock damage characterization of refractory composites

    Ceram. Int.

    (2008)
  • S. Martinović et al.

    Influence of sintering temperature on thermal shock behavior of low cement high alumina refractory concrete

    Compos. Part B Eng.

    (2014)
  • N.C. Acikbas et al.

    Thermal shock behaviour of α:β-sialon-tin composites

    Ceram. Int.

    (2014)
  • P.K. Panda et al.

    Thermal shock and thermal fatigue study of ceramic materials on a newly developed ascending thermal shock test equipment

    Sci. Technol. Adv. Mater.

    (2002)
  • P.K. Panda et al.

    Thermal shock and thermal fatigue study of alumina

    J. Eur. Ceram. Soc.

    (2002)
  • D. Jakobsen

    Thermal shock behaviour of laminated multilayer refractories for steel casting applications reinforced by residual stresses

    Ceram. Int.

    (2016)
  • F. Damhof et al.

    Experimental analysis of the evolution of thermal shock damage using transit time measurement of ultrasonic waves

    J. Eur. Ceram. Soc.

    (2009)
  • M. Hamidouche et al.

    Thermal shock behaviour of mullite ceramic

    Ceram. Int.

    (2003)
  • R. Grasset-Bourdel et al.

    Influence of thermal damage occurrence at microstructural scale on the thermomechanical behaviour of magnesia-spinel refractories

    J. Eur. Ceram. Soc.

    (2012)
  • A.P. Luz et al.

    Thermal shock damage evaluation of refractory castables via hot elastic modulus measurements

    Ceram. Int.

    (2013)
  • C.P. Jiang

    A study of the mechanism of formation and numerical simulations of crack patterns in ceramics subjected to thermal shock

    Acta Mater.

    (2012)
  • M. Collin et al.

    The morphology of thermal cracks in brittle materials

    J. Eur. Ceram. Soc.

    (2002)
  • S. Martinovic et al.

    Implementation of image analysis on thermal shock and cavitation resistance testing of refractory concrete

    J. Eur. Ceram. Soc.

    (2010)
  • G. Ye et al.

    Hydration of hydratable alumina in the presence of various forms of MgO

    Ceram. Int.

    (2006)
  • M. Samadzadeh et al.

    A review on self-healing coatings based on micro/nanocapsules

    Prog. Org. Coat.

    (2010)
  • W. Pabst et al.

    Elasticity of porous ceramics - A critical study of modulus-porosity relations

    J. Eur. Ceram. Soc.

    (2006)
  • W. Pabst et al.

    Isothermal and adiabatic Young’s moduli of alumina and zirconia ceramics at elevated temperatures

    J. Eur. Ceram. Soc.

    (2013)
  • T. Shimizu et al.

    Thermal conductivity of high porosity alumina refractory bricks made by a slurry gelation and foaming method

    J. Eur. Ceram. Soc.

    (2013)
  • X. Jin et al.

    Thermal shock behavior of porous ZrB2-SiC ceramics

    Mater. Sci. Eng. A

    (2013)
  • T. Hiratoko et al.

    Thermal properties of Ca-doped stabilized zirconia under high pressure and high temperature

    Ceram. Int.

    (2014)
  • S.K. Durrani et al.

    Synthesis and characterization of low density calcia stabilized zirconia ceramic for high temperature furnace application

    Mater. Chem. Phys.

    (2006)
  • Cited by (13)

    • Role of fatigue in damage development of refractories under thermal shock loads of different intensity

      2020, Ceramics International
      Citation Excerpt :

      The length-to-width ratio is 5–6 and 4 for smaller and larger cross-sections, respectively. Cubical and cylindrical samples are also used [17,23,25,29]. For those the side length varies between 36 mm and 50 mm.

    • Micromechanical modeling of thermo-mechanical properties of high volume fraction particle-reinforced refractory composites using 3D Finite Element analysis

      2020, Ceramics International
      Citation Excerpt :

      Afterwards, the postprocessing techniques for extracting final results are explained and finally, analytical schemes that are used for comparison with FE results are briefly discussed. Additionally, from our previous studies we have the VF of all constituents in each composite type [2], repeated here in Table 2. For all FE simulations, the input for VF of each constituent is the corresponding value from this table.

    View all citing articles on Scopus
    View full text