Mechanical Properties and Performance of Engineering Ceramics and Composites VIII, Volume 34, Issue 2

Dileep Singh, Jonathan Salem, Dileep Singh, Jonathan Salem

• Publisher: Wiley-American Ceramic Society
• Year: 2013
• Print ISBN: 9781118807477
• eBook ISBN: 9781118807453
• Edition: 1
• Topic: Technology & Engineering
• Subtopic: Materials Science
• Index: Technology & Engineering

Ceramic Engineering and Science Proceedings Volume 34, Issue 2 - Mechanical Properties and Performance of Engineering Ceramics and Composites VIII A collection of 21 papers from The American Ceramic Society's 37th International Conference on Advanced Ceramics and Composites, held in Daytona Beach, Florida, January 27-February 1, 2013. This issue includes papers presented in Symposium 1 -
Mechanical Behavior and Performance of Ceramics and Composites.

Processing and Properties of Fibers and Ceramics

THE EFFECT OF THE ADDITION OF CERIA STABILISED ZIRCONIA ON THE CREEP OF MULLITE.

D. Glymond*, M. Vick^°+, M.-J. Pan^°, F. Giuliani*^,+, L.J. Vandeperre*

*Centre of Advanced Structural Ceramics & Department of Materials, Imperial College London, South Kensington Campus, London SW7 2AZ, UK

⁺ Department of Mechanical Engineering, Imperial College London, South Kensington Campus, London SW7 2AZ, UK

^° Naval Research Laboratory, Washington, DC 20375

ABSTRACT

Mullite is considered a promising candidate for ceramic recuperators in turbo propelled engines, due to due to its low thermal conductivity, adequate thermal shock resistance, low cost, low density, thermodynamic stability, and reasonable strength at high temperatures. Unfortunately, the limited fracture toughness of mullite (~1.8–2.8 MPa m^1/2) is considered too low. Improving the fracture toughness to 4.7 MPa m^1/2 is possible by the addition of ceria stabilised zirconia (CSZ). However, the addition of CSZ to mullite may also affect other properties. In this paper the effect of the addition of CSZ on the creep resistance is described by comparing a mullite-zirconia composite made with a commercially available mullite powder against the creep behaviour of mullite made from the same mullite powder. The stress exponent is close to 1 and the activation energies for creep were similar at 426±38 kJ mol⁻¹ and 452±15 kJ mol⁻¹. However, the resistance to creep as expressed by a given strain rate for a given stress is 250 °C lower in the 20 vol% zirconia-mullite composite compared to the baseline.

INTRODUCTION

Mullite is an important material for high temperature structural applications due to its low coefficient of thermal expansion, good creep resistance and strength at these temperatures.^1–2 The main drawback of mullite as a structural material is its fracture toughness of only 1.8–2.8 MPa m^{1/2 3–4}.

There are a variety of methods to increase the toughness of a ceramic such as crack deflection⁵, self-reinforcement by elongated grains to create crack bridging,⁶ and second phase particles.⁷ One method, which has been shown to toughen mullite to a large extent, is the addition of zirconia. Bender et al. measured a toughness of 4.7 MPa m^1/2 when adding 18 vol% of ceria-stablised zirconia (CSZ).⁸ With toughness values in this range a number of structural applications within turbine engines are within reach, and thus the effect of this addition on other properties needs to be quantified to determine overall feasibility.

The high creep resistance of mullite is well known^9–13, and any effect on this property will have a large effect on the performance of the material.^{1, 2, 14} Therefore in this paper the effect of the addition of CSZ on the resistance to compressive creep is investigated.

EXPERIMENTAL

A commercial mullite powder (KM 101, KCM corporation, Nagoya, Japan) with a nominal composition of 28 wt% SiO₂ and 71.9 wt% Al₂O₃ and less than 0.1 wt% other oxides was used. This powder is very close to the silica rich end of the single phase mullite region represented by the nominal composition 3 Al₂O₃ − 2 SiO₂. The Ce-stabilised zirconia contained 12 mol% CeO₂ (Grade CSZ-12) and was obtained from Iwatani America, USA.

Composites containing 20 vol% CSZ were prepared by ball milling the powder mixture in acetone for 5 hours. Poly-ethylene glycol (PEG 400, Alfa Asaer, UK) was added as a binder. After milling, the slurries were dried in an oven at 70 °C overnight followed by uni-axial compaction at 30 MPa before sintering in air at 1500 °C for 5 hours. Samples of the mullite powder on its own were produced by vacuum hot pressing the powder at 1650 °C for 2 hours under 25 MPa pressure.

The density of the sintered pellets was determined using Archimedes’ principle. The theoretical densities were calculated using the rule of mixtures using 3.16 Mg m⁻³ for mullite and 6.23 Mg m⁻³ for Ce-stabilised zirconia¹⁵. The microstructure was characterised by observing fracture surfaces in a scanning electron microscope (JEOL-5610LV, Jeol, Japan).

All samples for creep tests were produced in a 2:1 aspect ratio. For the reference material consisting of mullite alone, samples with dimensions of 5 mm × 5 mm × 10 mm were machined out of the hot pressed billet, while for the mullite-zirconia composites, cylindrical samples measuring 12 mm high by 6 mm diameter were produced by sintering directly. All samples were polished plane parallel for compressive creep testing in a custom built testing rig consisting of a MRF Vacuum furnace mounted in a 100 kN universal test frame with graphite push rods and molybdenum heat shields and elements. The samples were heated under vacuum to the test temperature at 50 °C min⁻¹, soaked at temperature for 20 minutes under a pinch load of 150 N and then tested under load for 1 hour unless failure occurred earlier. After testing the CSZ-mullite composites at temperatures (~1100–1300 °C) and stresses (10–100 MPa) relevant to some of the gas turbine applications, initial tests indicated that the creep rate in the hot pressed mullite was much lower. Therefore, it was decided to find the temperature difference needed to obtain the same creep strain rate for the same stress for both types of mullite, and therefore subsequent testing on the hot pressed mullite was carried out at higher temperatures (1400–1550 °C).

RESULTS

Table 1 shows the density of the samples tested. Both sets of samples were near full density after processing. Figure 1 shows micrographs of the fracture surface of the two materials. The mullite + CSZ sample shows bright equiaxed zirconia grains embedded in a matrix of gray, elongated, mullite grains and evidence of intergranular fracture, whereas the hot pressed mullite shows mainly transgranular failure.

Figure 1. (a) Secondary electron micrograph of the hot pressed mullite, (b) back scattered electron micrograph of the mullite + 20 vol% CSZ composites. Grey areas are mullite, brighter areas are CSZ.

Table 1. Density of the different materials

Material	Density (%)
Hot pressed mullite	98.8 ± 0.2
Mullite + 20 vol% CSZ	99.9 ± 0.2

Figure 2 shows examples of typical traces of strain versus time. The initial creep rate tends to be high (primary creep), followed by a stable, slower, steady state creep rate (secondary creep) until the material starts to fail and the creep rate enhances again as can be seen for example in the trace for 52.8 MPa in Figure 2. Hence, to calculate the strain rate, the initial higher strain rate is ignored and the slope of the linear section is determined. These steady state strain rates have been plotted against stress for the mullite-zirconia composite in Figure 3. Over the limited range of stresses tested, the creep rate is related to the stress in the usual manner through:

Figure 2. Strain against time for mullite with 20 vol% CSZ at 1300 °C at 3 stress levels.

Figure 3. Strain rate versus applied stress the mullite + 20 vol% CSZ composite

(1)

where

is the strain rate, A is a pre-exponential constant, σ is the stress and n the stress exponent. The fact that most lines are parallel, indicates that the stress exponent varies little. As shown in Table 2, this is indeed the case with the stress exponent close to 1 for most temperatures. It is worth pointing out that it is clear from Figure 3 that even at 1300 and 1350 °C the stress exponent would have been closer to 1 had the data point at 52.8 MPa not been included in the analysis.

Figure 4 shows the same strain rate against stress plot for the hot pressed mullite. Again parallel lines were found and as illustrated in Table 2, the stress exponent is again nearly constant and close to 1. Moreover, the results have been arranged in Table 2 to illustrate that for approximately the same creep rate at the same stress, the temperature has to be 250 °C higher for the hot pressed mullite than for the composite containing CSZ.

Figure 4. Strain rate versus stress for the hot pressed mullite.

Table 2. Strain rate and stress exponent for ...

Table of contents

• Cover
• Half Title page
• Title page
• Copyright page
• Preface
• Introduction
• Characterization and Modeling of Ceramic Matrix Composites
• Processing and Properties of Fibers and Ceramics
• Author Index