Elsevier

Acta Materialia

Volume 48, Issue 8, 11 May 2000, Pages 1655-1665
Acta Materialia

Analysis and prediction of thermal shock in brittle materials

https://doi.org/10.1016/S1359-6454(00)00011-2Get rights and content

Abstract

The indentation quench method has been studied through both experiments and modeling. Practical results are obtained for alumina, whisker reinforced alumina, cermet and high speed steel. The crack growth vs temperature difference curves show no crack growth at very low ΔTs, stable crack growth at medium ΔTs and unstable crack growth above a certain ΔTTU) and these regimes are explained in the analysis. The stable and unstable crack growth are governed by the combination of residual and thermal stress. An expression has been derived for the prediction of thermal shock resistance and it is shown that the fracture toughness is of great importance. The presence of residual stress results in the greater sensitivity of the indentation-quench method compared to other approaches, and also makes it possible to define specific values of ΔT adapted to specific applications. The method can be used to explore susceptibility to thermal fracture in a range of brittle materials on condition that it is possible to insert an indentation precrack.

Introduction

Materials that are used in high temperature applications are often exposed to rapid temperature changes which cause thermal stresses and risks for thermal shock damage. Examples are as varied as energy conversion systems, electronic devices and cutting tools. The quenching-strength test is a common laboratory approach to determine the thermal shock resistance of ceramics 1, 2. Heated samples are quenched into a coolant, generally water, and the remaining strength after quenching is evaluated by bending. The remaining strength is plotted as a function of the temperature difference, ΔT, over which the sample is quenched and the critical temperature difference, ΔTC, is defined as the point at which the material shows a drastic drop in remaining strength. The drop in strength is caused by a limited growth of at least one of the most harmful defects. The harmfulness of a defect depends on both size and location, and both of these are statistically distributed. A consequence of this is that a great number of specimens is needed for the quenching-strength test. Statistical effects can be reduced by introducing precracks with known size and location, and for this purpose indentation techniques have been used during the last 20 years 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14. These investigations have the combination of indentation and rapid cooling in common. Although the main approach is similar, the investigators have used different technical solutions; e.g. the quenching media have been as varied as air 4, 5, 10, helium [6], liquid nitrogen 8, 13, and water 3, 7, 9, 11, 12, 14. Among these, water will cause the most rapid cooling and is consequently an appropriate choice for the testing of materials with high thermal shock resistance. The investigators have also used different approaches to evaluate the indentation–rapid cooling tests. In early investigations 4, 5, 6, the thermal shock resistance was evaluated by observing if crack growth had occurred or not and the lowest temperature difference giving crack growth was denoted ΔTC. During the last decade, most investigators have noticed that the crack growth is stable at low thermal load and the amount of crack growth has been measured 8, 9, 11, 12, 13, 14. Only a few investigators have treated the transition from stable to unstable growth. It has been suggested that the sudden drop in the retained strength at a certain temperature difference is connected to a transition from stable to unstable crack growth [7].

An indentation–quench test to evaluate the thermal shock resistance of brittle materials is currently being explored in detail 11, 14, 15. This test treats localized cracks with known crack geometry, which makes modeling possible. As the artificial cracks are made in the center of the sample surface, edge effects are avoided. This paper explores the indentation–quench test through both experiment and modeling and provides the basis to interpret experimental results for a range of quenching conditions and materials. It is further shown that the indentation quench method is applicable, not only to obviously brittle materials like ceramics, but even to those cermets and steels that fail by cracking. The results emphasize the importance of toughness in defining the resistance to thermal failure.

Section snippets

Materials and quenching experiments

Four different materials were investigated: (1) high-purity densely sintered fine-grained alumina with an average grain size <5 μm (Procera Sandvik), (2) alumina reinforced with 30 vol.% of silicon carbide whiskers (Sandvik Coromant), (3) titanium-based cermet, grade CT530 (Sandvik Coromant), and (4) high-alloy high speed steel, grade ASP2060 (Erasteel). The samples were in the form of plates (1) 13 mm diameter×4 mm, , 13 mm square×4 mm, and (4) 15 mm diameter×4 mm. The high speed steel grade

Crack growth at the surface

The mean percentage crack length increase with respect to the as-indented crack length has been calculated from the growth of the individual cracks along the surface. Figure 1 shows the crack growth as a function of the temperature difference, ΔT, for the four materials.

The pattern of crack growth is similar for all materials and can be divided into three regimes:

  • Regime A. At very low ΔT no significant crack growth can be detected.

  • Regime B. In a medium ΔT interval the crack growth is stable.

Analysis of thermal propagation of indentation cracks

The three regimes in Fig. 1 and the changes in shape of the propagating crack need to be explained. We will use stress intensity calculations to predict the conditions under which cracks will grow 10, 19 and it will be shown that the results can be explained by considering the effects of thermal and residual stress for different quenching conditions. Alumina has been used as a model material in the analysis, but the results are applicable to all brittle materials.

The total stress intensity of

Estimation of the surface heat transfer coefficient

We can use the results from the previous section to obtain information on the surface heat transfer coefficient from a knowledge of the thermal stress. The use of fracture mechanics in the analysis implies parity between the thermal stress at fracture and the mechanical strength. This kind of parity has been the starting-point for many investigations 4, 5. It is reasonable to anticipate that the thermal stress at the onset of unstable crack growth is equal to the mechanical strength of the

Prediction of thermal shock resistance

A method to predict the thermal shock resistance has been desired for a long time. The ΔTU is a promising parameter to predict, since the relation between residual and thermal stress intensity can be derived at this point. From Section 4.4. we know that ΔTU occurs when the derivative of the total stress intensity equals the derivative of the toughness. According to equation (1) the total stress intensity is the sum of the thermal and residual stress intensities. We then have:d(Kthermal,surf)dc+d

Modeling

The analysis is valuable for the understanding and interpreting of results from thermal shock resistance measurements using the indentation–quench method. The model confirms that the crack growth is governed by a combination of thermal and residual stress. The presence of residual stress is mainly advantageous for the method. First it makes the method more sensitive and second it results in a regime with stable crack growth, which makes it easier to define specific values of ΔT, such as ΔTU or Δ

Conclusions

The indentation–quench technique has been studied through both experiments and modeling. The quench medium was 30°C water. Results obtained for four different materials show that the crack growth vs temperature difference curve can be divided into three regimes—no crack growth at very low ΔTs, stable crack growth at medium ΔTs and unstable crack growth above a certain ΔTTU). The absence of crack growth at low ΔTs is mainly due to a very low surface heat transfer coefficient on sample

Acknowledgements

The authors would like to thank Kristin Breder for helpful discussions. This work has been performed within the Center Inorganic Interfacial Engineering, supported by the Swedish National Board for Industrial and Technical Development (NUTEK) and the following industrial partners: Erasteel Kloster AB, Ericsson Cables AB, Höganäs AB, Kanthal AB, OFCON AB, Sandvik AB, Seco Tools AB and Uniroc AB.

References (43)

  • F. Osterstock

    Mater. Sci. Engng

    (1993)
  • F. Osterstock et al.

    J. Euro. Cer. Soc.

    (1996)
  • T. Fett et al.

    Engng Fracture Mech.

    (1990)
  • R.F. Cook et al.

    Acta metall.

    (1988)
  • S.J. Bennison et al.

    Acta metall.

    (1989)
  • Y. Kim et al.

    Mater. Sci. Engng

    (1991)
  • G. Vekinis et al.

    Acta metall. mater.

    (1990)
  • D.P.H. Hasselman

    J. Am. Ceram. Soc.

    (1970)
  • D. Lewis
  • D.P.H. Hasselman et al.

    Am. Ceram. Soc. Bull.

    (1978)
  • K.T. Faber et al.

    J. Am. Ceram. Soc.

    (1981)
  • D.B. Marshall et al.
  • D. Johnson-Walls et al.

    J. Am. Ceram. Soc.

    (1985)
  • Gong, J. H., Guan, Z. D. and Jiang, D. C., in Fracture Mechanics of Ceramics, Vol. 10, ed. R. C. Bradt. Plenum Press,...
  • S.R. Choi et al.

    J. Am. Ceram. Soc.

    (1994)
  • M. Saadaoui et al.

    J. Mater. Sci. Lett.

    (1996)
  • T. Andersson et al.

    J. Am. Ceram. Soc.

    (1996)
  • Legrendre, F. and Osterstock, F., J. Mater. Sci. Lett., 1997, 16,...
  • Andersson, T. and Rowcliffe, D. J., J. Euro. Ceram. Soc., in...
  • Collin, I. K. M. and Rowcliffe, D. R., Ceramic Science and Enginineering Proceedings, Vol. 20.3 ed. E. Ustundag and G....
  • R.G. Munro

    J. Am. Ceram. Soc.

    (1997)
  • Cited by (173)

    View all citing articles on Scopus
    View full text