Mechanical Properties and Performance of Engineering Ceramics and Composites IX, Volume 35, Issue 2

Dileep Singh, Jonathan Salem, Dileep Singh, Jonathan Salem

• Publisher: Wiley-American Ceramic Society
• Year: 2014
• Print ISBN: 9781119031185
• eBook ISBN: 9781119031314
• Edition: 1
• Topic: Technology & Engineering
• Subtopic: Materials Science
• Index: Technology & Engineering

The Ceramic Engineering and Science Proceeding has been published by The American Ceramic Society since 1980. This series contains a collection of papers dealing with issues in both traditional ceramics (i.e., glass, whitewares, refractories, and porcelain enamel) and advanced ceramics. Topics covered in the area of advanced ceramic include bioceramics, nanomaterials, composites, solid oxide fuel cells, mechanical properties and structural design, advanced ceramic coatings, ceramic armor, porous ceramics, and more.

Creep, Fatigue, and Damage Characterization

ANISOTROPIC CREEP BEHAVIOR OF A UNIDIRECTIONAL ALL-OXIDE CMC

Katia Artzt, Stefan Hackemann, Ferdinand Flucht and Marion Bartsch

German Aerospace Center (DLR) Cologne, Germany

ABSTRACT

This paper is intended to give an overview of the creep behavior of unidirectional porous all-oxide CMC including experimental results as well as numerical simulations. The creep behavior was investigated by means of creep tests proving a tension- compression asymmetry. For compression creep, stress and temperature dependencies were determined and described by power-law creep equations. Activation energy was similar for all fiber orientations whereas the stress exponent differed with respect to the loading direction. Thus, for modeling the creep behavior of CMC lamina anisotropic creep must be taken into account. Within commercial finite element software implemented anisotropic creep laws are rare. The anisotropic approach according to Hill was tested for compression creep with regard to the applicability and limitations for CMCs. Since the Hill-approach did not capture the experimental results sufficiently, compression creep was further investigated by simulations on a microscopic scale via unit cells. Thereby the effect of different creep parameters for fibers and matrix and the resulting lamina’s deformation rate could be investigated. It became apparent that additionally compaction of the porous matrix has to be included in the numerical description.

INTRODUCTION

Al-oxide CMCs based on alumina fibers and matrices are favorable materials for use at high temperature in oxidizing atmospheres. One possibility for CMCs is the application as combustion liner. As a major drawback oxide fibers show the lowest creep resistance of all ceramic fibers¹ due to their predominant ionic bonds. Therefore, the knowledge of the creep behavior becomes important for construction and dimensioning of components undergoing creep deformation over the life time.

In this study, WHIPOX^™ material (‘Wound Highly Porous Oxide CMC’) is investigated². In the first processing step fiber bundles are heat treated to diminish organics, infiltrated with alumina slurry and deposited on a rotating mandrel. Due to the winding process various geometries of components can be achieved and the deposition angle of the fibers can be changed producing laminas with different fiber orientations (e.g. ±45°). The green body is dried and sintered whereby the special microstructure of WHIPOX^™ evolves. The weak porous matrix (Figure 1) leads to the desired high damage tolerance due to crack deflection at fibers. The matrix porosity is 40-85 Vol.% and the composite has an overall porosity of 20-50 Vol.%. These materials are based on alumina fibers (Nextel^™ 610) or aluminosilicate fibers (Nextel^™ 720) and an aluminosilicate or pure alumina matrix. The aluminosilicate type shows higher creep resistance³, but the alumina variant reveals better mechanical properties and higher heat conductivity. The investigated material consists of 3000 Denier Nextel^™ 610 alumina fibers embedded in a pure alumina matrix. The material is sintered at 1573 K for 1h dwell time. To gain an overview of the creep characteristics of fibers and matrix, creep experiments with a quasi-unidirectional composite (±2°) were conducted.

Figure 1. Microstructure of WHIPOX^™ (SEM micrograph).

EXPERIMENTAL RESULTS

Tension and compression creep experiments were conducted on WHIPOX^™ with quasi-unidirectional fiber architecture (±2°). An angle of ±2° was chosen to maintain a better handling in wet state before drying and sintering. Stress, temperature and fiber orientation were varied within the experiments. The definition for the fiber orientation with regard to the load axis used can be obtained from Figure 2.

Figure 2. Definition of the samples’ fiber orientation by the angle α between fiber and load axis. In this case, a compression load is indicated by the arrows.

Figure 3 and Figure 4 illustrate some examples of strain rate versus strain − curves for tension (Figure 3) and compression creep tests (Figure 4).

Figure 3. Tension creep rates for different stresses and fiber orientations (Left: 0°. Middle: 45°. Right: 90°.).

Figure 4. Strain rates vs. the absolute value of strain or time in compression creep. Left: different fiber orientations at 1473 K and −30 MPa. Right: Stress variation within one experiment (1448K, 0° fiber orientation) for determination of the stress exponent.

The tension experiments were conducted at stresses between 2 and 60 MPa depending on the fiber orientation due to the different strength of fibers and porous matrix (Figure 3). For tensile experiments of the 90° specimens stresses of 2-4 MPa were chosen. Damage emerged at low strains, indicated by the increase of strain rate, and fracture occurred at small strains of 0.4-1%. In contrast to the 90° orientation, the 0° samples exhibit a longer almost steady state creep regime (Figure 3, left). Experiments were terminated after 6 to 12% strain without fracture. The 45° samples show deformation characteristics of both the 0° and the 90° samples with a steady state creep regime and intermediate fracture strains between 2 to 3% (Figure 3, middle).

An example for orientation dependent compression creep at −30 MPa and 1473 K is illustrated in Figure 4 (left). The 0° specimen deforms with the lowest strain rate. With increasing angle between load and fiber axes, the absolute value of the strain rate rises. For all fiber orientations the negative creep rate is maximum at the beginning and decreases over the course of time. Only short-time (6-8 h) experiments were performed.

In case of compression tests, temperature and stress dependencies were determined according to the Norton power-law creep equation

(1)

with the creep rate

, the constant A, the stress σ, the stress exponent n, the creep activation energy Q, the universal gas constant R and the temperature T. Equation (1) characterizes the steady state creep. If experimental results do not show secondary steady state creep, the minimum creep rate is chosen for creep evaluation. For compression creep no lowest or steady state creep rate could be observed. Therefore, the stress exponent (Table I) was determined by stress variation within a single experiment (e.g. Figure 4 right). For determining the activation energy experiments at different temperatures but at similar creep strains were evaluated.

For all fiber directions activation energies were found in the magnitude of about 700 kJ/mol. This is consistent with the results of an earlier creep study of WHIPOX^{TM 4}. Contrary to the activation energy the stress exponent is not constant but varies from about 3 for fiber alignment parallel to the load (0°) to 1.9 (90°) (Table I). This indicates different stress exponents for the single components fiber and matrix. Fibers presumably dominate creep in 0° orientation with a stress exponent of 3. A similar value of 2.9 to 3.4 assigned to interface-reaction-controlled creep for fiber bundles and single fibers in tension experiments was already reported elsewhere ^4-10. Considering the 90° orientation the stress exponent is about 1.9 indicating a larger influence of the matrix. Because the exact portion of matrix and fiber creep deformation in 90° has not been in...

Table of contents

• Cover
• Half Title page
• Title page
• Copyright page
• Preface
• Introduction
• Creep, Fatigue, and Damage Characterization
• Processing and Properties of Carbides
• Processing and Properties of Non-Carbides
• Oxidation and Healing
• Delamination, Chipping, and Wear
• Index