Strength degradation mechanisms in NiAl alloy coated ... · Strength degradation mechanisms in NiAl alloy coated sapphire fibers Von der Fakultät für Georessourcen und Materialtechnik

Strength degradation mechanisms in NiAl alloy coated sapphire fibers

Von der Fakultät für Georessourcen und Materialtechnik der Rheinisch-Westfälischen Technischen Hochschule Aachen

zur Erlangung des akademischen Grades eines

Doktors der Naturwissenschaften

genehmigte Dissertation

vorgelegt von Master of Science (Physik)

David Edward Hajas

aus Jönköping, Schweden

Berichter: Univ.-Prof. Jochen M. Schneider, Ph.D. Univ.-Prof. Dr.-Ing. Christoph Leyens

Tag der mündlichen Prüfung: 01. Februar 2008

Diese Dissertation ist auf den Internetseiten der Hochschulbibliothek online verfügbar

Shaker VerlagAachen 2008

Materials Chemistry Dissertation

No.: 08 (2008)

David E. Hajas

Strength degradation mechanismsin NiAl alloy coated sapphire fibers

Bibliographic information published by the Deutsche NationalbibliothekThe Deutsche Nationalbibliothek lists this publication in the DeutscheNationalbibliografie; detailed bibliographic data are available in the Internet athttp://dnb.d-nb.de.

Zugl.: Aachen, Techn. Hochsch., Diss., 2008

Copyright Shaker Verlag 2008All rights reserved. No part of this publication may be reproduced, stored in aretrieval system, or transmitted, in any form or by any means, electronic,mechanical, photocopying, recording or otherwise, without the prior permissionof the publishers.

Printed in Germany.

ISBN 978-3-8322-7471-9ISSN 1861-0595

Shaker Verlag GmbH • P.O. BOX 101818 • D-52018 AachenPhone: 0049/2407/9596-0 • Telefax: 0049/2407/9596-9Internet: www.shaker.de • e-mail: [email protected]

iii

Abstract

The efficiency of a gas turbine can be increased by increasing the

combustion zone temperature. For this new materials are needed. NiAl

strengthened with single crystal �-Al2O3 (sapphire) fibers is considered as

a load bearing component in the combustion zone turbine blades.

However, due to strength degradation of the fibers during composite

fabrication, sufficient strengthening of NiAl can not be achieved. The goal

of this thesis is to identify strength degradation relevant mechanisms in

order to minimize the strength degradation during the production of

Intermetallc Matrix Composites (IMC’s) in the future. NiAl and IP75

(Ni45Al45Cr7,5Ta2,5) were considered as matrix materials. The presence of

an interlayer on the fiber strength was studied; hexagonal BN (h-BN) for a

weak interface, Y and Hf for enhanced interface strength by compound

formation. The strength of coated fibers was evaluated by tensile testing

and compared to the strength of uncoated fibers. The effect of temperature

during diffusion bonding was studied. None of the investigated coupling

concepts yields the improvement in tensile strength of NiAl or IP75

necessary for high temperature structural applications due to process

related fiber strength degradation.

iv

By systematic examination of fracture and fiber surfaces as well as

chemical analysis, the following strength degradation mechanisms have

been identified: Twinning of the rhombohedral plane (r-plane) of the

sapphire crystal as a consequence of the evolving thermal stress during

cooling. Chemical reactions between fiber and matrix material and/or

impurities and surface diffusion of Al2O3 into irregularities in the adjacent

matrix material like cracks in the h-BN interlayer both lead to altering of the

fiber surface morphology. In combination with the evolving thermal stress

during cooling, fracture mirror formation at the surface flaws takes place,

degrading the fiber strength.

In this work the fiber strength degradation mechanisms active during

individual process steps are identified. The significance of thermal stress

induced fiber damage is emphasized and it is concluded that the

elimination or a significant reduction thereof is the largest challenge which

has to be addressed to explore the strengthening potential of sapphire

fibers for IMC´s in the future.

Based on the results presented here, a strategy for obtaining high

strength IMC´s in the future is compiled. It is suggested that the strength

degradation may be avoided or minimized by: lowering the thermal stress

by increasing the fiber volume fraction and/or by using a ductile interlayer.

v

Zusammenfassung

Der Wirkungsgrad einer Gasturbine kann durch eine Erhöhung der

Brennkammertemperatur gesteigert werden. Dazu werden neuartige

Hochtemperaturwerkstoffe benötigt. Langfaserverstärktes NiAl mit

einkristallinen �-Al2O3-Fasern (Saphir) soll die lasttragende Komponente in

der Brennkammerturbinenschaufel darstellen. Aufgrund von

Festigkeitserniedrigungen der Saphirfaser kann jedoch keine ausreichende

Verstärkung des NiAl erreicht werden. Das Ziel dieser Arbeit ist es, die

relevanten Festigkeitserniedrigungsmechanismen zu identifizieren, um

zukünftig die Festigkeitserniedrigung während der Herstellung des

intermetallischen Komposites zu minimieren. Als Matrixwerkstoffe werden

NiAl und IP75 (Ni45Al45Cr7,5Ta2,5) verwendet. Der Einfluss einer

Zwischenschicht auf die Faserfestigkeit wird untersucht: hier werden zum

einen hexagonales BN (h-BN) für eine schwache Anbindung und zum

anderen Y und Hf für eine starke Anbindung durch Compoundbildung

eingesetzt. Die Festigkeit von beschichteten Fasern wird mittels Zugtests

ermittelt und mit der Festigkeit von unbeschichteten Fasern verglichen. Der

Temperatureinfluss während des Diffusionsschweißens wird untersucht.

Aufgrund von prozessbedingten Faserfestigkeitserniedrigungen zeigt

keines der untersuchten Bindungskonzepte einer Verbesserung der für

Hochtemperaturanwendungen benötigten Festigkeit.

vi

Mittels systematischer Untersuchungen der Bruch- und

Faseroberflächen sowie chemischer Analysen können die folgenden

Festigkeitserniedrigungsmechanismen identifiziert werden:

1. Zwillingsbildung auf rhomboedrischen Kristallebenen (r-Ebene) des

Saphirkristalls als Folge von thermischen Spannungen, die während des

Abkühlens entstehen. 2. Chemische Reaktionen zwischen Faser- und

Matrixwerkstoff und/oder Verunreinigungen und 3. Oberflächendiffusion

von Al2O3 aufgrund Unebenheiten im Matrixwerkstoff und Rissen in der

h-BN-Schicht. Die Mechanismen 2 und 3 führen zu einer Veränderung der

Faseroberflächenmorphologie. In Kombination mit den thermischen

Spannungen während des Abkühlens kommt es an den

Oberflächenunebenheiten zur Bildung sogenannter „Fracture mirrors“, die

zu einer Festigkeitserniedrigung führen.

In dieser Arbeit werden die Festigkeitserniedrigungsmechanismen

während der einzelnen Prozessschritte der Kompositherstellung

identifiziert. Die Bedeutung der durch thermische Spannungen induzierten

Faserfestigkeitserniedrigung wird herausgestellt. Die Vermeidung bzw.

eine deutliche Verringerung der Festigkeitserniedrigung stellt eine große

Herausforderung dar. Ohne die Lösung dieses Problems kann das

Verstärkungspotential von Saphirfasern in intermetallischen Kompositen in

zukünftig nicht genutzt werden.

vii

Auf Grundlage der hier vorgelegten Ergebnisse wird eine Strategie

zur Herstellung hochfester intermetallischer Komposite zusammengestellt.

Vorschläge zur Vermeidung bzw. zur deutlichen Verringerung der

Festigkeitserniedrigung sind: Erniedrigung der thermischen Spannungen

durch eine Erhöhung des Faservolumenanteils und/oder durch das Nutzen

einer duktilen Zwischenschicht.

viii

Preface

The work presented in this thesis is a part of the project funded by

the Deutsche Forschungsgemeinschaft (DFG) within the collaborative

research centre (SFB) 561 “Thermally highly loaded, porous and cooled

multi-layer systems for combined-cycle power plants”. It was exclusively

carried out at the RWTH Aachen University in the Materials Chemistry

group.

ix

Publications

Papers contributing to the thesis

Paper I

Effect of a BN Interlayer on the Tensile Strength of NiAl Coated

Sapphire Fibers

D.E. Hajas, S. Kyrsta, J.M. Schneider

Scripta Materialia 55 (2006) 219-222

Paper II

Strength Degradation Mechanisms in h-BN/NiAl Coated Sapphire

Fibers with a Reactive Hf or Y Interlayer

D.E. Hajas, S. Kyrsta, S. Richter, J. Mayer, J.M. Schneider

Materials Science and Engineering A 491(2008) 207-213

Paper III

Tensile Strength of IP75 Coated Sapphire Fibers

D.E. Hajas, T. Gebhardt, H. Chen, J.M. Schneider

Submitted for publication

(September 2007)

x

Paper IV

Microstructure and Mechanical Properties of Continuous Al2O3 Fiber

Reinforced Ni45Al45Cr7.5Ta2.5 Alloy (IP57) Matrix Composites

Y. Zhong, D. Hajas, W. Hu, H. Chen, G. Gottstein

Philosophical Magazine 87 (2007) 1019-1032

Paper V

Long Fibre Reinforced NiAl: High Temperature Material for Turbine

Blades

M. Rosefort, C. Dahmen, A. Bührig-Polaczek, W. Hu, H. Chen, Y. Zhong,

G. Gottstein, D.E. Hajas, J.M. Schneider

Advanced Engineering Materials 8 (2006) 730-735

Paper VI

A New Method to Examine Interfacial Reactions of a Multilayered

System NiAl-Hf-h-BN on a Sapphire Fiber

S. Richter, S. Kyrsta, D.E. Hajas, J.M. Schneider, J. Mayer

Microchimica Acta 155 (2006) 257-262

xi

Publications related to the topics of the thesis

Paper VII

Thermal and Chemical Stability of Cr2AlC Containing �-Al2O3

Reinforced NiAl Composites

M. Scholz, D.E. Hajas, B. Hallstedt, D. Music, J.M. Schneider

In manuscript

Paper VIII

Effect of a V2AlC Interlayer on the Tensile Strength of NiAl Coated

Sapphire Fibers

T. Gebhardt, D.E. Hajas, M. Scholz, B. Cappi, R. Telle, J.M. Schneider

Submitted for publication

(January 2008)

xii

Other publications

Paper IX

Oxidation of Cr2AlC in the Temperature Range 1220 to 1410°C

D.E. Hajas, M. to Baben, J.M. Schneider

In manuscript

Paper X

Integrated Approach for the Development of Advanced, Coated Gas

Turbine Blades

R. Herzog, N. Wanke, I. Steinbach, B. Hallstedt, C. Walter, J. Müller,

D.E. Hajas, E. Münstermann, J.M. Schneider, R. Nickel, D. Parkot,

K. Bobzin, E. Lugscheider, P. Bednarz, O. Trunova, L. Singheiser

Advanced Engineering Materials 8 (2006) 391-418

Paper XI

The Sandfish´s Skin: Morphology, Chemistry and Reconstruction

W. Baumgartner, F. Saxe, A. Weth, D.E. Hajas, D. Sigumonrong,

J. Emmerlich, M. Singheiser, W. Böhme, J.M. Schneider

Journal of Bionic Engineering 4 (2007) 1-9

xiii

Paper XII

Oxidation Resistance of Al2O3 Infiltrated Open Porous Iron Foams

D.E. Hajas, Y. Jiang, S. Angel*, S.L. Schulze*, W. Bleck*, and

J.M. Schneider

In manuscript

Paper XIII

Chemical Vapor Infiltration of �-Al2O3 into Open Porous Iron Foam

with TiN Interlayer

Y. Jiang, D.E. Hajas, S. Angel S.L. Schulze, W. Bleck, and J.M. Schneider

In manuscript

xiv

Acknowledgements

I sincerely thank my Professor Jochen M. Schneider who gave me

the chance to do this work, believed in me, helped me to develop, and

pushed me in the right direction when I got a bit lost. Thanks Jochen, you

are an excellent supervisor and Professor.

Many other people have supported me in my work. First of all I

would like to thank all my colleagues for helping me with measurements

and scientific discussions: The workshop wise guys, Hr. Horbach, Hr.

Kaiser and the workshop staff for the unlimited help to build experimental

setups, solving problems in the lab, valuable discussions, and just for

having a really nice time whenever being in the workshop,

Dr. Münstermann, for proof reading and being really helpful whenever it

was needed, Bengt, for the support whenever thermodynamics was an

issue, Dr. Chen, for the stress calculations, Stepan, for the lunch breaks

filled with interesting political discussions, Fr. Hauck and Fr. Strauch for the

administrative work. Thanks to all my students, Thomas, Moritz, Carolin,

Mirjam, Yan, Jamie, Wang and Huang for helping me with experiments and

measurements.

I thank all my friends that have indirectly helped me with my work.

Arif O., Jiri D., Timur P., Sebastian W., Halil G., Raphael M., Florian R.,

xv

Gyurika F., Gábor M., “Gyula & Kati “, “Balazs & Flavi” , Egér, Zsazsa,

Gyula K., Evi S. and Sara F. You have made my lunch breaks, vacations,

and my spare time to events that I was always looking forward to and

enjoying.

I want to thank my parents, my brother and my sister for your

support and for believing in me. Last but not least I want to thank my

wonderful wife, Kati, and my son Philip for making me so happy.

Financial support granted by the Deutsche Forschungsgemeinschaft

(DFG) within the Collaborative Research Centre (SFB) 561: “Thermally

highly loaded, porous and cooled multi-layer systems for combined cycle

power plants” is gratefully acknowledged.

xvi

xvii

Enjoy the present, because you don’t know what tomorrow brings

To my family

xviii

Contents

xix

1. Introduction………………………………………………………..-1-

2. Review……………………………………………………….……... -7-

2.1 Materials ……………………………………………..………... -7-

2.1.1 NiAl……………………………………………………….. -7-

2.1.2 IP75………………………………………………………. -11-

2.2 Fiber reinforcement…………………………………………... -13-

2.3 Coupling concepts…………………………………………….-13-

2.3.1 Hexagonal BN interlayer……………………………….. -14-

2.3.2 Y and Hf interlayer……………………………………… -15-

3. Methods of research………………………………………....... -23-

3.1 Composite fabrication……………………………………….. -23-

3.2. Interlayer deposition…………………………………………. -25-

3.2.1 h-BN interlayer………………………………………….. -25-


3.3 Matrix material deposition…………………………………... -27-

3.3.1 Deposition of the NiAl matrix…………………………… -30-

3.3.2 Deposition of the IP75 matrix…………………………... -31-

3.4 Thermal stress simulations………………………………... -32-

3.5 Tensile testing………………………………………………… -33-

3.6 Annealing………………………………………………………..-34-

3.7 Etching………………………………………………………….. -35-

3.8 Characterization………………………………………………. -35-

3.8.1 Morphology and fracture surface analysis…………… -35-

3.8.2 Chemical composition………………………………….. -36-

3.8.3 Structure…………………………………………………. -38-

xx

3.9 Statistical methods…………………………………………… -39-

3.9.1 Weibull modulus………………………………………… -40-

3.9.2 Characteristic life……………………………………….. -42-

3.9.3 Confidence interval……………………………………... -43-

4. Results and discussion…………………………………. -49-

4.1 Characterization of the matrix- and interlayer materials. -49-

4.1.1 NiAl matrix……………………………………………….. -50-

4.1.2 IP75 matrix………………………………………………. -52-

4.1.3 Morphology and microstructure……………………….. -54-


4.2 Fiber strength…………………………………………………..-63-

4.2.1 Tensile strength of NiAl coated sapphire fibers with

and without a h-BN interlayer…………………………..-63-

4.2.2 Tensile strength of IP75 coated sapphire fibers with

and without a h-BN interlayer…………………………. -71-

4.2.3 Tensile strength of h-BN/NiAl coated sapphire fibers

with a reactive Hf or Y interlayer………...................... -89-

5. Conclusions………………………………………………………. -103-

6. Future work……………………………………………………… -107-

xxi

xxii

1. Introduction

- 1 -

1. Introduction

One way to increase the efficiency of gas turbines is to increase the

combustion zone temperature. Therefore, new high temperature

compatible materials are being developed. Of special interest are the

combustion zone turbine blades. NiAl has a number of attractive properties

for high temperature applications, such as high melting point, good thermal

conductivity [1], low density, and excellent oxidation resistance [2, 3].

However, low creep resistance at high temperature as well as poor fracture

toughness at room temperature and low thermal shock resistance limits its

use for turbine blade applications [2]. One way to address the low creep

resistance and strength is by strengthening NiAl with ceramic fibers

forming an Intermetallic Matrix Composite (IMC). Single crystal �-Al2O3

(sapphire) fiber has here been selected due to its high strength at room

and elevated temperatures (2.1-3,4 GPa at RT and 1.0-1.7 GPa at

1100°C[4]) as well as chemical stability [5].

It is well known in the literature that the fiber matrix coupling, i.e. the

physical and chemical interaction between fiber and the surrounding matrix

material, plays an essential role for the resulting physical properties of the

composite [6]. Due to a low fracture toughness of NiAl below the brittle to

1. Introduction

- 2 -

ductile transition temperature (approximately 400°C) and due to super

ductility above approximately 600°C, the requirements of the fiber-matrix

coupling for IMC´s vary with temperature. At low temperatures NiAl is

brittle, hence weak coupling is beneficial. This promotes fiber pull out and

crack deflection, resulting in an enhancement of the macroscopic

composite toughness [7]. However, at high temperatures NiAl is super

ductile, hence strong coupling is needed for efficient load transfer from the

matrix to the fibers to enhance the composite strength and creep

resistance [6, 8]. By the introduction of an interlayer the fiber matrix

coupling is altered as well as the macroscopic properties of the composite.

In this work several coupling concepts have been developed and

investigated which will be described in detail.

Sapphire fibers have been incorporated into a wide variety of matrix

materials. However, it has been reported that the fiber strength is degraded

during composite fabrication [9-12]. In this work the tensile strength of

coated sapphire fibers with and without an interlayer is compared with the

strength of uncoated fibers. The fabrication process of IMC´s has been

simulated by annealing coated and uncoated fibers at the diffusion bonding

temperature. The goal of this thesis is to identify strength degradation

relevant mechanisms active between �-Al2O3 fibers and the surrounding

1. Introduction

- 3 -

matrix or interlayer in order to minimize the strength degradation during

production of IMC’s in the future.

1. Introduction

- 4 -

List of references [1] Intermetallic compounds, Vol. 2, D. B. Miracle and R. Darolia,

(John Wiley & Sons Ltd, 1994).

[2] Overview No.104 - The Physical and Mechanical Properties of

NiAl, D. B. Miracle, Acta Metallurgica Et Materialia 41 (1993) 649.

[3] Ductility in intermetallic compounds, A. M. Russell, Advanced

Engineering Materials 5 (2003) 629.

[4] Data sheet for c-axis oriented sapphire fibers (www.photran.com).

[5] Alumina as a Ceramic Material, W. H. Gitzen, The American

Ceramic Society Special Publications, Vol. 4, 1970.

[6] Interfaces in Metal-Matrix and Intermetallic-Matrix Composites,

R. J. Arsenault, Composites 25 (1994) 540.

[7] The Role of Coating Compliance and Fiber-Matrix Interfacial

Topography on Debonding in Ceramic Composites,

R. J. Kerans, Scripta Metallurgica Et Materialia 32 (1995) 505.

[8] An Introduction to Composite Materials, Second Edition, D. Hull

and T. W. Clyne, (Cambridge University Press, 1981).

[9] Processing and Mechanical-Properties of Al2O3 Fiber-

Reinforced NiAl Composites, R. R. Bowman, A. K. Misra, and

S. M. Arnold, Metallurgical and Materials Transactions A-Physical

Metallurgy and Materials Science 26 (1995) 615.

[10] Effects of Composite Processing on the Strength of Sapphire

Fiber-Reinforced Composites, J. B. Davis, J. Yang, and A. G.

Evans, Acta Metallurgica Et Materialia 43 (1995) 259.

1. Introduction

- 5 -

[11] Effect of composite fabrication on the strength of single crystal

Al2O3 fibers in two Fe-base alloy composites, S. L. Draper and

B. J. M. Aikin, Materials Science and Engineering A-Structural

Materials Properties Microstructure and Processing 266 (1999) 18.

[12] Al2O3 Fiber Strength Degradation in Metal and Intermetallic

Matrix Composites, S. L. Draper and I. E. Locci, Journal of

Materials Research 9 (1994) 1397.

1. Introduction

- 6 -

2. Review

- 7 -

2. Review

In this chapter the relevant properties of the two matrix materials

NiAl and IP75 and two coupling concepts investigated in this work are

presented and reviewed. The first coupling concept is based on weakening

the fiber matrix bond in order to enhance the composite toughness and

lower thermal stresses. In the second coupling concept additionally a

reactive interlayer is introduced in order to strengthen the fiber matrix bond

during compound formation.

2.1. Materials

2.1.1. NiAl

As already discussed, NiAl possesses an attractive combination of

properties for use as turbine blades material in gas turbines or aircraft

engines [1]. It has a high congruent melting temperature (1638 °C) [2]

which is approximately 300 °C higher than conventional Ni-based super

alloys [3]. Its density (5.9 g/cm3) is approximately two thirds of Ni-based

super alloys and hence result in a lower self induced stress in rotating

turbine blades [2]. Furthermore, NiAl has a high thermal conductivity

(70 - 80 W/mK) which is 3-8 times higher than Ni-based super alloys. This

2. Review

- 8 -

results in a significant reduction of thermal gradients and hence lower

thermal stress and an improved thermal fatigue resistance [3]. NiAl

possesses an excellent oxidation resistance due to the formation of

continuous protective �-Al2O3 scale above 1050 °C [4]. The parabolic rate

constant is two orders of magnitude lower than for typical Ni-based super

alloys [5]. However, there are two principal challenges of NiAl for use as

turbine blade material [3]: 1) Poor toughness and damage tolerance at

T>400°C. According to von Mises criterion, five independent slip systems

are required for a poly-crystalline material to yield [6]. In NiAl only three

independent slip systems are available, and hence the material can not

undergo extensive plastic deformation at room temperature and fractures

before substantial plastic flow occurs [1]. 2) Low strength and creep

resistance at elevated temperature. The latter will be addressed in this

work by reinforcing the NiAl with ceramic fibers. This will be discussed in

detail in chapter 2.2.

NiAl is an intermetallic compound and possesses an ordered B2

(CsCl-type) cubic crystal structure [1]. The structure consists of two

interpenetrating primitive cubic cells, as shown in figure 1.

2. Review

- 9 -

Figure 1. CsCl crystal structure of NiAl with two interpenetrating Ni and Al

primitive cells.

The Al atoms occupy the cube corners of one sublattice, and the Ni atoms

occupy the cube corners of the other sublattice. For body centered cubic

(b.c.c.) structures such as NiAl, diffusion-controlled deformation

mechanisms begin to dominate above a homologous temperature of

0.5 Tm, which results in the low strength and creep resistance of NiAl.

Above the ductile to brittle transition temperature which is approximately

400 °C [3], polycrystalline NiAl becomes very ductile, and plastic

elongations of over 40 % are observed at 600 °C [3]. Polycrystalline NiAl

has a tensile strength of 170 MPa at room temperature and 70 MPa at

800°C [2]. A typical Ni-based super alloy exhibits a tensile yield strength of

1100 MPa at RT and up to 770 MPa at 1000°C [7].

2. Review

- 10 -

The B2 structure is stable for large deviations from stoichiometry,

from 45 to 69 at.% Ni, as shown in the Ni-Al phase diagram (figure 2) [8].

The Coefficient of Thermal Expansion (CTE) of NiAl at 1000°C is 15.1 K-1

and does not vary significantly with composition [9].

Figure 2. The Ni-Al phase diagram shows a large stability range (45 to

69 at.% Ni) for the B2 NiAl phase [8].

2. Review

- 11 -

2.1.2. IP75

NiAl with Cr and Ta alloying referred to as IP75 [10] is also

considered as matrix material in this work. The attempts to address the

challenges posed by harnessing NiAl for efficient energy conversion have

lead to the development of NiAl alloys. Zeumer and Sauthoff [11] reported

that by adding Ta to NiAl a significantly higher strength and creep

resistance can be obtained at high temperatures due to the formation of

the ternary Laves phase NiAlTa with C14 structure besides the NiAl B2

phase. However, the brittleness at room temperature and low fracture

toughness of the NiAlTa is comparable with poly crystalline NiAl. By

addition of Cr, a third Cr rich disordered phase with A2 structure is formed

and the alloy shows an improved plastic deformability and fatigue

resistance [12, 10]. It was established that the most promising composition

with respect to strength, creep resistance, brittleness, corrosion resistance,

and thermal shock resistance was 45 at.% Ni, 45 at.% Al, 7.5 at.% Cr, and

2.5 at.% Ta [10]. However, despite of improved ductility by the addition of

Cr the fracture toughness of IP75 at RT remains low, 6-10 MPa(m1/2) [13]

(table 1). It is comparable to that of polycrystalline NiAl, 4-6 MPa(m1/2) [2]

(table 1) and is well below the typical values for Ni based super alloys

which exceed 50 MPa(m1/2) at RT [7] (table 1). Its tensile strength at RT,

2. Review

- 12 -

however, is considerably higher than for NiAl: 440 MPa at RT and 200-300

MPa at 800 °C (table 1), but still too low for applications as structural

material in turbine blades. The previously discussed data on the

mechanical properties of NiAl, IP75, and superalloy CMSX-4 are

summarized in table 1.

NiAl IP75 Ni-based superalloy

(CMSX-4)

Tensile strength [MPa]

RT (ultimate tensile strength)

800°C (yield strength)

170 [9]

70 [9]

440 [13]

200-300 [13]

900 [14]

1100 [14]

CTE [K-1] 15.1 [9] 13.5 [15] -

Fracture toughness [MPa(m1/2)]

RT

4-6 [2]

6-10 [13]

�50 [7]

Table 1. Tensile strength, coefficient of thermal expansion (CTE), and

fracture toughness values of NiAl, IP75, and super alloy CMSX-4 at room

temperature (RT) and 800°C.

2. Review

- 13 -

2.2. Fiber reinforcement

Experimental results have consistently shown that the strength of

sapphire fibers incorporated in metallic or intermetallic matrices is reduced

during the composite fabrication process [16-22].

The main reason for this identified in the literature is a chemical

reaction between Al2O3 and matrix and/or binders and/or impurities in the

matrix material [17, 19-21]. However, thermal stresses that arise due to

CTE mismatch between fiber and matrix e.g. NiAl CTE=15.1�10-6 K-1 [9]

and �-Al2O3 fibers (c-direction) CTE=6.9�10-6 K-1 [23] (Note that these

values vary with temperature), and morphological changes of the fiber

surface due to surface diffusion of Al2O3 into matrix grain boundaries

during high temperature treatment [16-18, 20], and hot pressing induced

fiber fracture [18, 20] may also degrade the fiber strength.

2.3. Coupling concepts

Three different fiber-matrix coupling concepts are evaluated in this

work as presented in figure 3, namely: NiAl deposited directly onto the

fiber without any interlayer (figure 3a), NiAl deposited onto the fiber with a

2. Review

- 14 -

h-BN interlayer (figure 3b) and NiAl with a h-BN interlayer and an

additional reactive Y or Hf interlayer (figure 3c).

Figure 3. Schematic illustration of the different coupling concepts

investigated in this work: (a) NiAl alloy deposited onto the Al2O3 fiber

without interlayer, (b) with a h-BN interlayer and (c) with an Y or Hf

interlayer between the NiAl alloy and the h-BN interlayer.

2.3.1. Hexagonal BN interlayer

Strong coupling between the matrix and the fiber may result in a low

composite toughness at room temperature due to the lack of energy

dissipative mechanisms upon loading [24]. By introducing a weak bonding

between fiber and matrix, for example due to an interlayer (figure 3b),

other energy dissipative mechanisms like fiber pullout and crack deflection

may be enabled upon loading, increasing the fracture energy and hence

the composite toughness [25, 26]. Hexagonal BN (h-BN) is widely used as

2. Review

- 15 -

interlayer material in brittle composites (e.g. SiC-fiber/BN/SiC-matrix

composites [27]), which due to Van der Waals bonding between the basal

planes in the hexagonal lattice [28] may deflect cracks.

Hu et al. [29] performed structural analysis of the h-BN interlayer

used here with transmission electron microscopy (TEM). It was observed

that as deposited films exhibit a turbostratic structure which transforms into

a hexagonal structure during annealing with the hexagonal basal planes

aligned along the fiber c-axis, which is beneficial for achieving low interface

shear strength due to van der Waals bonding [29]. Furthermore, it was

established that the deposited h-BN does not form reaction products with

�-Al2O3 fiber at the composite fabricating temperature which is important in

order to maintain a high fiber strength [30, 17-21]. Reichert et al. [31]

performed chemical analysis of the h-BN interlayer and established that it

contains approximately 4 at.% O and C dissolved in the h-BN.

2.3.2. Y and Hf interlayer

Reichert and co-workers proposed to introduce refractory metal

interlayers, e.g. Y or Hf, between the h-BN interlayer (figure 3c) and the

NiAl matrix to enhance the fiber matrix coupling [32]. It was envisioned that

during composite fabrication (diffusion bonding) the refractory metal reacts

2. Review

- 16 -

with the fiber improving the fiber-matrix coupling by compound formation.

The reaction path in the Al2O3/h-BN/Y system in the temperature range

1100-1300°C has been investigated earlier [32]. The effect of the presence

of the matrix material NiAl is not discussed in reference 32. Reactions

between Y and Al2O3 take place forming several different Y oxides, of

which Y3Al5O12 is the most stable one. The reaction products obtained

during annealing of the h-BN/Hf/NiAl system for 20 h at 1350 °C were

reported to be: AlN, Al2O3, HfB2 and HfN [33].

2. Review

- 17 -

List of references [1] Ductility in intermetallic compounds, A. M. Russell, Advanced

Engineering Materials 5 (2003) 629.

[2] Overview No.104 - The Physical and Mechanical Properties of

NiAl, D. B. Miracle, Acta Metallurgica Et Materialia 41 (1993) 649.

[3] Intermetallic compounds, Vol. 2, D. R. Miracle DB, (John Wiley &

Sons Ltd, 1994).

[4] The Oxidation Behavior of NiAl. 1) Phase-Transformations in

the Alumina Scale during Oxidation of NiAl and NiAl-Cr Alloys,

M. W. Brumm and H. J. Grabke, Corrosion Science 33 (1992) 1677.

[5] Effect of 0.1 at.-Percent Zirconium on the Cyclic Oxidation

Resistance of Beta-NiAl, C. A. Barrett, Oxidation of Metals 30

(1988) 361.

[6] Independent Slip Systems in Crystals, G. W. Groves and A. Kelly,

Philosophical Magazine 8 (1963) 877.

[7] Intermetallic Matrix Composites III, C. H. Ward and A. S.

Culbertson, edited by R. R. B. J.A. Graves, J.J. Lewandowski

(Materials Research Society, San Francisco, 1994), Vol. 350, p. 3.

[8] Binary alloy phase diagrams, Vol. 1, T. B. Massalski, (American

Society for Metals, , Ohio 44073, 1986).

[9] Intermetallic compounds, Vol. 2, D. B. Miracle and R. Darolia,

(John Wiley & Sons Ltd, 1994).

[10] Deformation behaviour of intermetallic NiAl-Ta alloys with

strengthening Laves phase for high-temperature applications, III) Effects of alloying with Cr, B. Zeumer and G. Sauthoff,

Intermetallics 6 (1998) 451.

2. Review

- 18 -

[11] Intermetallic NiAl-Ta alloys with strengthening Laves phase for

high temperature applications, I) Basic properties, B. Zeumer

and G. Sauthoff, Intermetallics 5 (1997) 563.


strengthening Laves phase for high-temperature application, 2) Effects of alloying with Nb and other elements, B. Zeumer and

G. Sauthoff, Intermetallics 5 (1997) 641.


strengthening Laves phase for high-temperature applications - IV. Effects of processing, B. Zeumer, W. Sanders, and


[14] Tensile Behavior of a New Single-Crystal Nickel-Based

Superalloy (CMSX-4) at Room and Elevated-Temperatures,

A. Sengupta, S. K. Putatunda, L. Bartosiewicz, J. Hangas,

P. J. Nailos, M. Peputapeck, and F. E. Alberts, Journal of Materials

Engineering and Performance 3 (1994) 664.

[15] Multiphase intermetallic alloys for structural applications,


[16] Processing and Mechanical-Properties of Al2O3 Fiber-

Reinforced NiAl Composites, R. R. Bowman, A. K. Misra, and




Fiber-Reinforced Composites, J. B. Davis, J. Yang, and

A. G. Evans, Acta Metallurgica Et Materialia 43 (1995) 259.

2. Review

- 19 -

[18] Effect of composite fabrication on the strength of single crystal

Al2O3 fibers in two Fe-base alloy composites, S. L. Draper and

B. J. M. Aikin, Materials Science and Engineering A-Structural

Materials Properties Microstructure and Processing 266 (1999) 18.

[19] Tensile Behavior of Al 2O3/FeAl+B and Al2O3/FeCrAlY

Composites, S. L. Draper, B. J. M. Aikin, and J. I. Eldridge,

Metallurgical and Materials Transactions a-Physical Metallurgy and

Materials Science 26 (1995) 2719.

[20] Al2O3 Fiber Strength Degradation in Metal and Intermetallic

Matrix Composites, S. L. Draper and I. E. Locci, Journal of


[21] Influence of Interfacial Reactions on the Fiber-Matrix Interfacial

Shear-Strength in Sapphire Fiber-Reinforced NiAl(Yb) Composites, S. N. Tewari, R. Asthana, R. Tiwari, R. R. Bowman,

and J. Smith, Metallurgical and Materials Transactions A-Physical


[22] Niobium Coating Effects on the Tensile-Strength of Sapphire

Fibers, E. R. Trumbauer, J. R. Hellmann, and D. A. Koss,

Composites 25 (1994) 625.

[23] Data sheet for c-axis oriented sapphire fibers, www.photran.com.

[24] Influences of Interfacial Bonding Strength and Scatter of Fiber

Strength on Tensile Behavior of Unidirectional Metal Matrix Composites, S. Ochiai and K. Osamura, Journal of Materials

Science 23 (1988) 886.

[25] Interface engineering in oxide fibre/oxide matrix composites,

K. K. Chawla, C. Coffin, and Z. R. Xu, International Materials

Reviews 45 (2000) 165.

2. Review

- 20 -

[26] Interface design for oxidation-resistant ceramic composites,

R. J. Kerans, R. S. Hay, T. A. Parthasarathy, and M. K. Cinibulk,

Journal of the American Ceramic Society 85 (2002) 2599.

[27] Effect of a boron nitride interphase that debonds between the

interphase and the matrix in SiC/SiC composites,

G. N. Morscher, H. M. Yun, J. A. DiCarlo, and L. Thomas-Ogbuji,


[28] Third-order elastic constants and pressure derivatives of the

second-order elastic constants of hexagonal boron nitride,

V. M. Mathew, C. S. Menon, and K. P. Jayachandran, Journal of


[29] Interface structure, chemistry and properties of NiAl

composites fabricated from matrix-coated single-crystalline Al2O3 fibres (sapphire) with and without an hBN interlayer,

W. Hu, T. Weirich, B. Hallstedt, H. Chen, Y. Zhong, and G. Gottstein,

Acta Materialia 54 (2006) 2473.

[30] Strength degradation of sapphire fibers during pressure casting

of a sapphire-reinforced Ni-base superalloy, R. Asthana,

S. N. Tewari, and S. L. Draper, Metallurgical and Materials

Transactions a-Physical Metallurgy and Materials Science 29 (1998)

1527.

[31] Influence of BN fiber coatings on the interfacial structure of

sapphire fiber reinforced NiAl composites, K. Reichert, K. Wen,

R. Cremer, W. Hu, D. Neuschütz, and G. Gottstein, Applied Surface

Science 179 (2001) 150.

2. Review

- 21 -

[32] Reaction paths in the system Al2O3-hBN-Y, K. Reichert,

O. Oreshina, R. Cremer, and D. Neuschütz, Applied Surface

Science 179 (2001) 138.

[33] A new method to examine interfacial reactions of a multilayered

system NiAl-Hf-hBN on a sapphire fibre, S. Richter, S. Kyrsta,

J. M. Schneider, D. E. Hajas, and J. Mayer, Microchimica Acta 155

(2006) 257.

2. Review

- 22 -

3. Methods of research

- 23 -


In this chapter the deposition processes for interlayer and matrix

materials are described and the characterization methods used. A

description of the tensile strength measurement and the evaluation of

the measured data by statistical methods follows.

3.1. Composite fabrication

There are several methods known in the literature to incorporate

fibers into a matrix material. Examples of these are the fiber foil fiber

technique [1, 2], the powder metallurgical approach [3], and the matrix

coated fiber technique [4]. The latter was chosen in this work due to its

advantages over the other techniques, namely the homogeneous

distribution of fibers in the matrix, the precise control of fiber volume

fraction in the IMC by control of the matrix coating thickness. Most

importantly, deposition processes such as magnetron sputtering allow

the introduction of an interlayer which can be used for tailoring the fiber

matrix coupling.

The composite fabrication process is schematically shown in

figure 1. The fabrication process includes three main steps: i) interlayer

deposition, ii) matrix deposition and iii) diffusion bonding.


- 24 -

Figure 1. Schematic drawing of the composite fabrication process,

including three main steps: i) interlayer deposition, ii) matrix deposition

and iii) diffusion bonding.

For the composite fabrication a bundle of coated fibers are

consolidated by diffusion bonding under pressure at 40 MPa and

1300°C. However, within this work only the deposition processes i) and

ii) and the characterization of single fibers are discussed. The diffusion

bonding iii) and characterization of the diffusion bonded composite is

outside the scope of this work. The reader is refered to references [5-9]

where the composite consolidation process and the characterization has

been described in detail.


- 25 -

Single crystal c-axis oriented �-Al2O3 fibers (Saphikon) with an

average diameter of approximately 130 �m were cleaned in an

ultrasonic bath, first in acetone and subsequently in methanol. Two

fiber lengths were used in this work, 75 mm and 110 mm.The interlayer

materials Y and Hf and the matrix materials IP75 and NiAl were

deposted by magnetron sputtering, whereas the h-BN interlayer was

deposited by Chemical Vapor Deposition (CVD).

3.2. Interlayer deposition

3.2.1. h-BN interlayer

The h-BN coating was deposited in a vertical hot wall reactor

(figure 2) with a � 44 mm reaction tube and 220 mm long isothermal

zone at 930°C and at atmospheric pressure.

Figure 2. Experimental setup for chemical vapor deposition of a h-BN

interlayer onto Al2O3 fibers.


- 26 -

A gas mixture of H2, NH3, and C3H12B3N3 “Trimethylborazine” (TMB),

was used for the deposition. The flow of H2 and NH3 was 97% and 3%,

respectively, of the total gas flow. TMB was added by letting 10 % of the

H2 flow through a bubble chamber containing TMB at 40°C. 108 pieces

of fibers were coated in one deposition experiment, suspended in the

isothermal zone. A coating thickness of approximately 1 �m was

realized after 30 minutes of deposition. Deposition mechanisms and

kinetics for CVD of h-BN are described in detail in references [10] and

[11].


The fibers were placed in a rotatable holder and Y or Hf was

deposited onto the h-BN coated fibers. The experimental setup for

sputtering of Y or Hf is displayed in figure 3.


- 27 -

Figure 3. Experimental setup for sputtering Y or Hf interlayer onto Al2O3

fibers.

Both Y and Hf were sputtered from circular (90 mm diameter) elemental

targets. 96 pieces of 75 mm long Al2O3 fibers were coated in one

deposition experiment. An approximately 500 nm thick Y (or Hf) layer

was deposited with a power density of 5,5 W/cm2. The base pressure

before sputtering was approximately 6x10-6 mbar and the deposition

was carried out in a 6x10-3 mbar Ar atmosphere. Exposure of the coated

fibers to air caused the coatings to spall off the fibers. This was

prevented by deploying a shutter which allowed deposition of NiAl

subsequently to the Y or Hf deposition without removing the fibers from

the vacuum.

3.3. Matrix material deposition

For the deposition of matrix materials (NiAl and IP75), two

different experimental setups were used. For the NiAl deposition ,setup


- 28 -

1 was exclusively used (figure 4a), which was later scaled up for the

deposition of IP75 matrix (figure 4b), here called setup 2. For both

setups a shutter was deployed in order to prevent deposition before

stable processing conditions had been attained. The base pressure was

approximately 6.0x10-6 mbar. Prior to deposition, the fibers were

cleaned by Ar (99.999% purity) RF-plasma etching at a pressure of

6.0�10-3 mbar, a power of 250 W, and a DC-Bias of -350 V for

5 minutes. Both the NiAl and the IP75 were sputtered by means of

DC-magnetron sputtering with a power density of 11,4 W/cm2 at each

cathode under Ar atmosphere at a pressure of 6.0�10-3 mbar. With

setup 1, 96 fibers could be coated in one deposition experiment,

whereas with the scaled up “setup 2” 720 fibers could be coated in one

deposition experiment.


- 29 -

(a) (b)

Figure 4. (a) The experimental setup for NiAl deposition (setup 1). With

this setup 96 fibers can be coated in one deposition experiment. (b) The

experimental setup for IP75 deposition (setup 2). The fiber holders

rotate according to the planetary gear principle. With this setup 720

fibers can be coated in one deposition experiment.

The sapphire fibers were placed in holders with a tweezer before

deposition (figure 5). In setup 1, one holder was used, whereas in

“setup 2” 15 individual holders each containing 48 fibers were used.


- 30 -

Figure 5. Fiber holder for the old setup containing 96 fibers, before

deposition.

3.3.1. Deposition of the NiAl matrix

For the deposition of NiAl two opposite magnetron cathodes

(CemeCon CK288, figure 6), as shown in figure 5a, were used. 14

pieces of 2 mm thick (10x40) mm2 nickel plates (99.0% purity) were

clamped onto the aluminum sputtering targets (88 x 200) mm2

(99.5% purity) to allow for sputtering of stoichiometric NiAl with the

composition of approximately 50 at.% Ni and 50 at.% (figure 6). Sputter

deposition and characterization of NiAl have been reported by Reichert

et al. [12].


- 31 -

Figure 6. Sputter target for depsotion of NiAl consists of Ni plates

clamped onto an Al sputter target.

An IMC fiber volume fraction of 30 or 50% was realized after 165

and 74 minutes of sputtering, respectively. The deposition rate was 327

nm/min.

3.3.2. Deposition of the IP75 matrix

As described earlier IP75 was deposited with setup 2. Cast IP75

plates were used as target material. 8 plates with a total area of

(88x200) mm2 were clamped with Cu clamps onto each cathode

(figure 7).


- 32 -

Figure 7. Cast IP75 plates held in place by Cu clamps were used as

sputter target.

A coating thickness of 54 �m IP75 corresponding to a fiber

volume fraction of approximately 30% was realized after 600 minutes of

deposition with a power density of 11.4 W/cm2 at each cathode under

6x10-3 mbar Argon atmosphere. The substrate temperature was

measured with a shielded thermocouple. The highest temperature was

approximately 1100 K in front of the cathodes.

3.4. Thermal stress simulations

The thermal residual stresses on the fiber exerted by the IP75

coating for a single coated fiber with and without a h-BN interlayer were

simulated by using ABAQUS, a commercial Finite Element Methode

FEM) code [13]. The two-dimensional axisymmetric model for coated

fiber discriminates two zones: (i) the elastic fiber and (ii) the

elastic/plastic matrix that surrounds the fiber. A stress free temperature

Tsf of 850 K [14] for NiAl alloy (IP75) was assumed. Due to delamination


- 33 -

of the h-BN basal planes [7], the interface in the NiAl composite with h-

BN interlayer was considered to be a purely frictional interface with a

friction coefficient of �f = 0.21. Details on the model description can be

found in reference [15].

3.5. Tensile testing

The fiber tensile strength at room temperature and ambient

humidity was measured by using an in-house built pull test machine

(figure 8a). The two ends of the fibers were glued to metal plates

containing grooves which were attached to a holder (figure 8b). This

assures that the fibers are pulled perpendicular to the fiber axis and

provide an easy way to insert and attach the fiber ends to the

crossheads of the pull test machine (figure 8a).


- 34 -

(a) (b)

Figure 8. (a) Pull test machine used for the tensile testing of single fibers.

(b) Schematic of insertion of a fiber onto the crossheads for tensile testing.

Displacement controlled conditions with a crosshead speed of

5.0 �m/s and a gauge length of 50 mm were used for all tests. The

sapphire fiber diameter used for the calculation of the ultimate strength

was measured by scanning electron microscopy. However, the fibers

with NiAl matrix were analysed using a light microscope. During tensile

testing the fibers were immersed in grease in order to prevent the fiber

from breaking into several pieces. The fibers were then cleaned in an

ultra sonic bath with acetone and methanol before further analysis.

3.6. Annealing

The annealing treatment of the coated fibers was carried out

under similar conditions as the diffusion bonding of the composite [16]:

1300 or 1350°C under flowing Ar (99.995% purity) with 17.5 vol.% H2

(99.999% purity) for 1 h. The heating and cooling rate was set to

6.7°C/min and 2.5°C/min, respectively.


- 35 -

3.7. Etching

Two different solutions were used to remove the coating from the

fibers: 1) 33 vol.% nitric acid (HNO3), 17 vol.% hydrochloric acid (HCl)

and 50 vol.% deionized water or 2) 23 vol.% nitric acid (HNO3), 54 vol.%

hydrochloric acid (HCl) and 23 vol.% deionized water. The coatings

were removed more efficiently in the latter solution. Tensile strengths of

as received and etched fibers were compared, and no difference in

strength was established. Hence the etching solutions do not affect the

tensile stenght of the fibers.

3.8. Characterization

3.8.1. Morphology and fracture surface analysis

Fracture surface morphologies were evaluated and fracture

initiation sites were identified by imaging fracture and fiber surfaces

with a JEOL JSM-6480 Scanning Electron Microscope (SEM). The

fractured fibers were placed in a sample holder in the microscope

according to the schematic drawing (figure 9).


- 36 -

Figure 9. Setup for imaging of fracture surface of single fibers with

SEM.

With the current setup the fracture- and the fiber-surface could be

observed 360° around the fiber axis. By examination of the fracture

surface, the origin of the fracture initiation could be assessed. With the

coating removed by etching possible strength degrading flaws on the

fiber surface were detected.

3.8.2. Chemical composition

For qualitative chemical analysis Energy Dispersive X-ray

spectroscopy (EDX) with an EDAX Genesis2000 detector was

employed. The chemical composition of all deposited films was

quantitatively determined either by the standardless RZAF method or


- 37 -

using a compound standard, where the composition of the standard was

determined by Wavelength Dispersive X-ray spectroscopy (WDX). The

constituents of reaction products, percipitates etc. detected on fiber

surfaces were only qualitatively determined. The reason for this is the

limitation of EDX analysis method to quantitatively determine the

composition of a material that is not homogeneous and microscopically

smooth.

Focused Ion Beam (FIB) was used to expose the reaction zone in

the h-BN/Y/NiAl and h-BN/Hf/NiAl specimens by milling a bevel in the

matrix and interlayers with a Ga+ ion beam according to the principle

demonstrated in figure 10.


- 38 -

Figure 10. Setup for milling a bevel in the matrix and interlayer material

with Focused Ion Beam (FIB) [17].

The reaction zone was then mapped by Electron Probe Micro Analysis

(EPMA) in order to determine the spatial distribution of the elements.

The method has been described in detail in reference [17].

3.8.3. Structure

X-ray diffraction (XRD) with a General Area Detection Diffraction

System (GADDS) equipped for micro diffraction was deployed for

determination of the structure of the deposited films and reaction

products. The focusing geometry shown in figure 12 allows spatially

resolved structural analysis with a lateral resolution of approximately

0.5 mm.


- 39 -

Figure 12 illustrates the focusing geometry of the General Area

Detection Diffraction System (GADDS) equipped for micro diffraction,

which allows spatially resolved structural analysis, with a lateral

resolution of approximately 0.5 mm.

3.9. Statistical methods

Statistics help to extract representative information and draw

conlcusions from a large amount of measured data. In order to ascertain

significant changes in tensile strength between groups of fibers that

have experienced different treatments, e.g. deposition-, etching-,

heat-treatment etc., the strength data were statistically evaluated by

calculating significant statistical parameters. Since the composite is

constituted of a population of fibers, not only the absolute strength

values of the fibers should be considered, but also the statistical spread

of these.


- 40 -

3.9.1. Weibull Modulus

Ceramic fibers are brittle and a fracture is initiated at the surface

flaw that generates the largest stress intensity according to the Griffith

theory [18]. These strength limiting flaws can be morphologic or internal

inhomogeneities like chemical reaction sites or pores. The flaws are a

member of some determinable distribution [18]. It has been proven that

the Weibull distribution [19] is usedfull for describing the ultimate

strength of ceramic fibers. In its simplest form it can be described by the

two parameter relation:

��

�

� �

��

��

mc

cP0

exp1)(��

� ,

were P is the probability of failure, m is the Weibull modulus, �c is the

uniaxial tensile stress to rupture, and �0 is a scale parameter with the

same dimension as �c. From the equation above a linear relation exists,

when ln(ln(1/1-P)) is plotted against ln (�C):

� � � �0lnln)(1

1lnln ��

mmP c

c

��

�

� �

��

��

.

From this equation it can be seen that m is the slope of a straigh line

and m�ln(�0) is the y intercept. P(�c) is a cumulative probability function,


- 41 -

also called estimator and has a value between zero and one. Several

forms of estimators exist, but for small sample sizes the estimator has

proved to have its most appropriate form:

� �N

Pi5,0i �

� ,

where i is the ith result of the ranked critical fracture stresses (�c,,i) [20,

21]. Based on these equations and the strength data, a Weibull diagram

can be plotted. The Weibull modulus is then calculated from the slope of

a linear fit through the critical fracture stress (�c) vs. ln(ln(1/1-P)) data.

There are several methods of fitting a straight line, for example the

Least Square Method (LSM). However, it has been reported that it may

lead to large errors in estimating the slope [22]. With the LSM the best

fit is obtained for a slope where the sum of vertical square distances

(y-distances) is as small as possible [23]. Large errors occur when all

except one or a few points conform nicely to a straght line. Then the

slope does not produce the best fit to the data points. Instead it is

recommended to use the Maximum Likelihood Method (MLM) [22] by

Cohen [24]. Simulation results by Masson et al. [22] (figure 13a and b)

demonstrate the difference in the slope of the linear regression between

the two methods when one extreme value exists. It is seen that a large

estimation error occurs using the LSM (figure 13b). Hence, the Weibull


- 42 -

modulus m was here calculated with the MLM as described by Dodson

[25].

(a) (b)

Figure 13. (a) Linear fit with the Maximum Likelihood Method compared to

the (b) Least Square Method with one value largely deviating from the linear

regression [22].

3.9.2. Characteristic life

For Weibull distributed data the strength at a given failure

probability and not its mean value should be considered [22]. Hence,

the characteristic life, �0, was considered here which is the strength at a

given failure probability of 0,632, meaning that 63,2% of the population

fails at this strength regardless of the value of m. This value is also

known as the scale parameter for the Weibull distribution [25]. �0

corresponds to the interflexion point of the cumulative Weibull

distribution function which has an S like shape (figure 14). At this point


- 43 -

the slope is the largest and hence the least estimation error is obtained

[22].

Figure 14. Cumulative Weibull distribution for two different Weibull

moduli, m, demonstrating the interflexion point at a failure probabaility of

0.632.

3.9.3. Confidence Interval

Confidence interval is the interval where with a certain probablity

the true mean value of �0 exists. A confidence of 95% was chosen,

meaning that with a 95% certainty the true mean value of �0 exists

within the interval. The interval aproaches exactness as the sample size

increases. Hence, when comparing two mean values, a larger number

of experimental data will decrease the estimation error and hence lead

to a better estimate. The interval was calculated with the MLM, as

described by Dodson [25].


- 44 -

List of references [1] Fabrication of metal-matrix composites by hot-triaxial-

compaction of foil/fiber/foil lay-ups, P. D. Nicolaou and

H. R. Piehler, Scripta Materialia 39 (1998) 1077.

[2] Influence of BN fiber coatings on the interfacial structure of sapphire fiber reinforced NiAl composites, K. Reichert,

K. Wen, R. Cremer, W. Hu, D. Neuschütz, and G. Gottstein,

Applied Surface Science 179 (2001) 150.

[3] Intermetallic Fe-Al layers obtained by the powder cloth

method, F. Binczyk, S. T. Skrzypek, and A. Gierek, Powder

Technology 94 (1997) 259.

[4] Fiber Coating Concepts for Brittle-Matrix Composites,

J. B. Davis, J. P. A. Lofvander, A. G. Evans, E. Bischoff, and

M. L. Emiliani, Journal of the American Ceramic Society 76

(1993) 1249.

[5] Thermal residual stress (TRS) analysis in continuous Al2O3

fiber reinforced NiAl composite, H. Chen, W. Hu, A. Atiser, Y.

Zhong, and G. Gottstein, International journal of materials

research 97 (2006) 10.

[6] High-temperature creep behavior and hotpressing

consolidation of NiAl, H. Chen, W. P. Hu, Y. L. Zhong, and

G. Gottstein, Zeitschrift Fur Metallkunde 95 (2004) 1074.



W. Hu, T. Weirich, B. Hallstedt, H. Chen, Y. Zhong, and G.

Gottstein, Acta Materialia 54 (2006) 2473.


- 45 -

[8] Long fibre reinforced NiAl: High temperature material for

turbine blades, M. Rosefort, C. Dahmen, A. Buhrig-Polaczek,

W. P. Hu, H. Chen, Y. L. Zhong, G. Gottstein, D. E. Hajas, and J.

M. Schneider, Advanced Engineering Materials 8 (2006) 730.

[9] Microstructure and mechanical properties of continuous

Al2O3 fibre reinforced Ni45Al45Cr7.5Ta2.5 alloy (IP75) matrix composites, Y. Zhong, D. E. Hajas, W. Hu, H. Chen, and G.

Gottstein, Philosophical Magazine A 87 (2007) 1019

[10] Constitution and deposition mechanisms of hexagonal boron nitride formed by CVD from trimethylborazine

A. Jörg, E. Zimmermann, M. Schierling, R. Cremer, and

D. Neuschütz, in CVD XIV and Euro CVD-11, The

Electrochemical Society (Pennington, USA, 1997).

[11] Kinetics of Chemical-Vapor-Deposition of Boron-Nitride from

a Gas-Mixture of Trimethylborazine, Ammonia, and Hydrogen at 900 to 1050-Degrees-C and 1 Bar Total Pressure, A. Jörg,

D. Neuschutz, and E. Zimmermann, Journal De Physique Iv 5

(1995) 167.

[12] Sputter deposition and film characterization of NiAl on

sapphire fibres, K. Reichert, C. Martinez, S. Kyrsta, R. Cremer,

and D. Neuschütz, Vacuum 71 (2003) 241.

[13] ABAQUS User’s Manual, Version 6.3, Vol. K. Hibbit, Sorensen.,

(Providence, 2002).

[14] Processing and Mechanical-Properties of Al2O3 Fiber-Reinforced NiAl Composites, R. R. Bowman, A. K. Misra, and




- 46 -

[15] Finite-element analysis of the hot-pressing consolidation of

continuous Al2O3 ribers-reinforced NiAl composites, H. Chen,

W. Hu, and G. Gottstein, Zeitschrift Fur Metallkunde 96 (2005)

710.

[16] Untersuchungen zu chemischer Stabilität,

Grenzflächenstruktur und Debonding-Verhalten der von kontinuierlichen Al2O3-Fasern verstärkten NiAl-

Verbundwerkstoffe W. Hu, K. Wen, H. Chen, S. Kyrsta, and G. Gottstein,

Verbundwerkstoffe und Werkstoffverbunde, (Wien, 2003), 153.

[17] A new method to examine interfacial reactions of a

multilayered system NiAl-Hf-hBN on a sapphire fibre,

S. Richter, S. Kyrsta, J. M. Schneider, D. E. Hajas, and J. Mayer,

Microchimica Acta 155 (2006) 257.

[18] Fracture of Brittle Solids-Second Edition, B. Lawn,

(Cambridge University Press, 1975).

[19] A Statistical Distribution Function of Wide Applicability,

W. Weibull, Journal of Applied Mechanics-Transactions of the

Asme 18 (1951) 293.

[20] On the Estimation of the Weibull Modulus, B. Bergman,

Journal of Materials Science Letters 3 (1984) 689.

[21] Experimental Probability Estimators for Weibull Plots,

J. D. Sullivan and P. H. Lauzon, Journal of Materials Science

Letters 5 (1986) 1245.

[22] Some Guidelines for a Consistent Use of the Weibull

Statistics with Ceramic Fibers, J. J. Masson and E. Bourgain,

International Journal of Fracture 55 (1992) 303.

[23] Statistical Techniques in Business & Economics, J. N. Brown,

Journal of the American Statistical Association 74 (1979) 728


- 47 -

[24] Multi-Censored Sampling in 3 Parameter Weibull

Distribution, A. C. Cohen, Technometrics 17 (1975) 347.

[25] Weibull Analysis, B. Dodson, (ASQ Quality Press, Wisconsin,

1994).


- 48 -

4. Results and discussion

- 49 -


The goal of this work is to identify strength degradation relevant

mechanisms active during composite fabrication in sapphire fiber

reinforced NiAl or IP75 with or without a h-BN or Y (or Hf) + h-BN

interlayer. Chemical composition, structure, and morphology of the

deposited matrix- and interlayer-materials before and after annealing as

well as results from the tensile testing and the subsequent characterization

are presented in section 4.1 and 4.2. The effect of different interlayers and

matrix materials as well as processing conditions and heat treatment on

the fiber strength is discussed and strength degradation mechanisms are

identified.

4.1. Characterization of the matrix- and interlayer

materials

In the following section, the chemical composition, the structure, and

the morphology of the deposited matrix and interlayer materials before and

after annealing are presented. These data are important for process

development and for determining the effect of heat treatment on the

structure and chemistry of the interlayer and matrix materials.


- 50 -

4.1.1. NiAl matrix

The composition of the as deposited NiAl was approximately Ni0.55Al0.45

and hence within the stability range of the B2 NiAl phase between 45 to 69

at.% Ni [1]. The X-ray diffractogram of the sputtered NiAl indicates phase

pure NiAl (diffractogram i in figure 1). After annealing in Ar + H2

atmosphere, traces of �-Al2O3 were detected (diffractogram ii in figure 1).

All sputtered films contained 4-5 at.% oxygen which may be incorporated

in the NiAl during the sputtering process. �-Al2O3 has been observed in the

NiAl by transmission electron microscope (TEM), which may have been

formed due to a reaction between Al and the O incorporated in the NiAl

matrix during deposition [2].


- 51 -

Figure 1. i) X-ray diffraction data indicate the presence of pure NiAl before

annealing. ii) After annealing the formation of �-Al2O3 was detected.

Sputtered NiAl is growing with a columnar structure (figure 2a)

which transforms into dense NiAl after annealing (figure 2b). A thin Al2O3

scale is formed on the surface of the NiAl during annealing (figure 2b).

This observation indicates that the �-Al2O3 observed by XRD

(diffractogram ii in figure 1) may also be formed during the annealing

treatment due to reaction between Al and residual oxygen in the annealing

atmosphere.


- 52 -

(a) (b)

Figure 2. Cross sectional scanning electron microscope images of NiAl on

�-Al2O3 fibers: (a) Columnar growth morphology of as deposited NiAl. (b)

After annealing a thin Al2O3 scale is formed on the NiAl surface.

4.1.2. IP75 Matrix

The chemical composition of the cast IP75 target was: Ni: 45.2 at.%,

Al: 45.5 at.%, Cr: 6.6 at.% and Ta: 2.7 at.% (figure 3a), which is close to

the composition of IP75 defined by Sauthoff et. al. [3], namely: Ni:

45.0 at.%, Al: 45.0 at.%, Cr: 7.5 at.% and Ta: 2.5 at.%. The deposited films

were Ni depleted (Ni: 38.9 at.%) (figure 3a coating (a)), and hence

according to the Ni-Al phase diagram, Al3Ni2 and NiAl phases may be

present (figure 3b, coating (a)). The Ni content was increased to 43.8 at.%

(figure 3a and b, coating (b)) by sputtering additional Ni by attaching a Ni

plate on the sputtering target.


- 53 -

(a) (b)

Figure 3. (a) Chemical composition of the IP75 target and the sputtered

IP75 films. (b) Ni-Al phase diagram [1].

The resulting films consisted of phase pure NiAl as shown in the X-ray

diffractogram (figure 4). However, during annealing hexagonal- and cubic-

Cr2Ta laves phases are formed. These results indicate that Cr and Ta are

dissolved in NiAl after deposition. �-Al2O3 was also detected after

annealing and may be formed by reaction between Al and the dissolved

oxygen and/or oxygen from the annealing atmosphere as discussed

earlier.


- 54 -

Figure 4. (i) Diffractogram of sputtered and (ii) annealed IP75 coating

indicating the formation of cubic and hexagonal Cr2Ta as well as �-Al2O3

during annealing.

4.1.3. Morphology and microstructure

Similar to sputtered NiAl, IP75 grows with columnar morphology

(figure 5a and b, where a is the cross section of the sputtered IP75 and b

is the surface morphology). The cast microstructure of IP75 (figure 5c)

does not develop during annealing. Instead of this precipitates are formed


- 55 -

along the columnar grain boundaries (figure 5d) which are developed

during deposition. The chemical composition data suggest that the

precipitates contain Cr and Ta. It is reasonable to assume that the

precipitates constitute the Cr2Ta Laves phase detected with XRD after

annealing.


- 56 -

(a) (b)

(c) (d)

Figure 5. (a) Columnar structure growth of IP75. (b) Surface morphology

of sputtered IP75. (c) Microstructure of cast IP75. (d) Microstructure of

sputtered and annealed IP75.


A micrograph of the FIB cut of the as deposited h-BN/Y/NiAl layer is

displayed in figure 6a, while the for 2h at 1350°C annealed sample is


- 57 -

shown in figure 6b. Before annealing (figure 6a) the h-BN, Y and NiAl

layers are distinguishable. Chemical analysis (not shown) indicates the

presence of O in the Y interlayer, which may be due to oxygen

incorporation from residual gas [4]. After annealing, the formation of pores

was observed in the NiAl as well as in the Y interlayer (figure 6b), which

may be due to volume changes associated with chemical reactions or

depletion of material due to diffusion processes. In addition, Al was

detected in the Y-interlayer and precipitates formed between the Y and

NiAl layers. Chemical analysis (not shown) indicates the presence of Al

and N in the precipitates.


- 58 -

(a) (b)

Figure 6. (a) Scanning electron microscope (SEM) micrograph of an as

deposited h-BN/Y/NiAl coating cut by focused ion beam (FIB). Chemical

analysis indicates the presence of O in the Y interlayer after

deposition. (b) After 2 h of annealing additionally Al was detected in the Y

interlayer and Al and N containing precipitates were detected between the

Y and NiAl layers.

The X-ray diffractogram of the as deposited h-BN/Y/NiAl matrix

indicates the presence of Y2O3 and AlYO3, however, no metallic Y was

detected (figure 7).


- 59 -

Figure 7. X-ray diffractogram of the h-BN/Y/NiAl matrix separated from the

fiber, before, after 2, and after 20 h of annealing. The diffractogram

indicates that AlN, �-Al2O3, Y2O3, and several AlxYyOz phases are formed

during annealing.

The formation of oxides in the interlayer is consistent with the chemical

analysis discussed earlier. After annealing for 2 h, Y2O3, AlYO3, Al3Y5O12,

AlN, and �-Al2O3 was formed. After 20 h additionally YAl and Al2Y4O9 was

formed. These results support the notion of Al diffusion into the Y interlayer

during annealing. �-Al2O3 and AlN have been identified as reaction

products in the Al2O3/h-BN/NiAl system earlier, which are formed due to


- 60 -

reaction between Al and O incorporated in the NiAl matrix, and between h-

BN and NiAl, respectively [2].

SEM investigation of cross sections of annealed sapphire fibers with

a h-BN/Hf/NiAl coating showed extensive pore formation in the NiAl matrix

along the columnar grain growth boundaries (figure 8a and b) as well as

on the h-BN/Hf, Hf/NiAl interface (figure 8c). After 20 h of annealing time

the fraction of pores in the matrix and on the interface increased. Within

the pore walls and along the NiAl grain boundaries precipitates were

detected (figure 8b and c) which may be formed by diffusion processes

and/or chemical reactions.


- 61 -

(a) (b)

(c)

Figure 8. (a) Fracture cross section of the h-BN/Hf/NiAl coated fiber with

pores in the NiAl matrix visible. (b) Magnification of a matrix pore with

precipitates. (c) Magnification of an interface pore with precipitates.

The X-ray diffractogram of the h-BN/Hf coated Al2O3 fibers indicates

the presence of metallic Hf after deposition (figure 9).


- 62 -

Figure 9. X-ray diffractogram of the h-BN/Hf/NiAl matrix separated from

the fiber, before, after 2, and after 20 h of annealing. The diffractogram

indicates that �-Al2O3 and HfB2, AlN and HfN are formed during annealing.

After annealing HfB2 and HfN were identified as the only reaction

products besides AlN and �-Al2O3 discussed earlier. Hence, the

precipitates detected by SEM along the h-BN/Hf interface and the NiAl

grain boundaries may be HfB2 and HfN. No constitutional changes were

observed after annealing for 20 h.


- 63 -

4.2. Fiber strength

In the following section the effect of different interlayers and matrix

materials as well as processing conditions and heat treatment on the fiber

strength is discussed in order to identify the strength degradation relevant

mechanisms active in the composites.

4.2.1. Tensile strength of NiAl coated sapphire fibers with and

without a h-BN interlayer

The characteristic strength values measured for as-received �-Al2O3

fibers as well as NiAl coated fibers with and without a h-BN interlayer

before and after heat treatment are displayed in figure 10.


- 64 -

Figure 10. Characteristic strength of as-received and coated sapphire

fibers grouped according to treatment. � Al2O3, � Al2O3/h-BN,

� Al2O3/NiAl, � Al2O3/h-BN/NiAl.

The as received fibers (�) have a characteristic strength of

�0 = 2.42 � [0.08, 0.07] GPa, where the first number in the brackets is the

upper and the second number the lower confidence interval. This is

consistent with the manufacturer’s specifications of the fiber strength being

2.1-3.4 GPa at RT [5]. No significant change in �0 was detected after

annealing and/or etching of the as-received fibers. Also no indications of

reactions or morphological changes could be observed on the fiber

surface.


- 65 -

The results from the pull tests of the as deposited h-BN coated

fibers (�) show that the CVD process does not affect the fiber strength.

Chemical reactions between h-BN and �-Al2O3 are not expected at a CVD

coating temperature of 930°C. However, morphologic changes were

detected on the fiber surface at the crack initiation sites (figure 11a).

Chemical analyses indicate that the surface flaws contain traces of Si, Mg

and Ca. SEM images taken on the fractured surfaces support the notion

that these flaws may serve as crack initiation sites (figure 11b).

(a) (b)

Figure 11. (a) Typical morphologic artifact on the fiber surface around the

fiber fracture after the CVD of h-BN. (b) Typical fiber fracture surface

indicate fracture initiation at the surface flaw.


- 66 -

Results from the chemical analysis and the SEM investigation

indicate that chemical reactions between the contaminants and the fiber

give rise to morphologic changes. Possible sources of the contamination

include the CVD reactor tube, the precursor gases and the cleaning

agents.

Similar morphologic changes in the vicinity of the crack initiation

sites, with the same contaminants, were observed on h-BN coated fibers

which were subsequently coated with NiAl (�). However, with the NiAl

coating on top of the h-BN interlayer the fiber strength decreased

significantly to �0 = 1.56 � [0.09, 0.10] GPa which corresponds to

approximately 64% of the initial sapphire fiber strength. The decrease in

fiber strength may be due to thermal stress that arises upon cooling after

the NiAl coating process. The altered morphology may act as stress

concentration site. The compressive stress may induce micro cracks

around the surface flaws thus decreasing the strength of the fiber. When

the NiAl was removed by subsequent etching, the strength of the fibers

increased to �0 = 1.96 � [0.13, 0.14] GPa which indicate that thermal

stresses are released, but the fibers do not retain their initial strength which

is consistent with the notion discussed above.

After annealing of the h-BN coated fibers 84 % of their initial strength

was retained, �0 = 2.04 � [0.18, 0.16], whereas the strength of h-BN/NiAl


- 67 -

coated fibers decreased to 1.45 � [0.16, 0.14] GPa, which corresponds to

approximately 60 % of the initial strength. On the h-BN/NiAl coated fibers

after annealing additionally ridges were observed on the fiber surface

meandering along the fiber surface (figure 12).

Figure 12. Ridge formation on the fiber surface of a BN/NiAl coated fiber

after annealing.

No evidence for contamination was detected on the ridges. Different

thermal expansion coefficients of �-Al2O3, BN, and NiAl may cause the BN

interlayer to crack during the annealing process. The ridges can then be

formed due to surface diffusion of Al2O3 into the cracks in the BN. SEM

analysis of the fractured surface indicate that the fracture initiates at the

ridges and at sites where chemical reactions have taken place. Thermal


- 68 -

stresses from the NiAl in combination with the altered surface morphology

may be responsible for the decrease in fiber strength observed here.

Fibers coated with NiAl (�) did not exhibit any significant change in

critical strength neither with coating nor with the coating etched away. No

morphological changes or chemical reactions could be identified. However,

during the annealing process all the coated fibers broke into pieces.

Fractured fibers large enough for pull testing cracked upon handling with

tweezers. Hence, the strength was assumed to be zero, �0 = 0. Fiber

pieces examined by optical microscopy after etching showed severe

cracking of the fiber (figure 13a). Surface cracks could be identified by

SEM, which also revealed grain boundary like imprints on the fiber surface

(figure 13b).


- 69 -

(a) (b)

Figure 13. (a) Annealed NiAl/Al2O3 with the coating etched away captured

with an optical microscope indicates severe cracking within the fiber.

(b) NiAl grain boundary imprints in the sapphire surface and cavities with

traces of Si.

The imprints have been observed earlier for the NiAl/Al2O3 system [6]

and it is thought that surface diffusion processes are responsible for the

morphological changes [7]. Cavities were also identified on the fiber

surface (figure 13b). Chemical analysis indicates traces of Si. The cavities

may arise due to chemical reactions between Si and Al2O3 during the

annealing process. Si may originate from the Ni or the Al target material

which according to the manufacturer’s specification contains < 0.15 at.%

and < 0.30 at.% Si, respectively.


- 70 -

Simulation results show that thermal stresses may induce fracture as

the fiber is cooled down from 1300°C to RT [8]. It has been reported that

the compressive strength of �-Al2O3 decreases below the tensile strength

in the c-axis direction at temperatures of >400°C [9]. At temperatures

between approximately 520°C and 410°C the compressive stress asserted

by the NiAl coating is larger than the compressive strength of the sapphire

fiber in the c-axis direction which may cause the observed fiber fracture [8].

The extensive fiber fracturing observed here has to the author’s knowledge

not been observed earlier and may be explained by the presence of larger

thermal stresses which may be caused by the larger coating thickness

used in this work compared to previously reported works [6] [7]. The

thicknesses reported range from 1-3 �m for a NiAl matrix [6] to 15 �m for a

Mo matrix [7], compared to the approximately 54 �m used in this work.


- 71 -

4.2.2. Tensile strength of IP75 coated sapphire fibers with and

without a h-BN interlayer

The characteristic strength of IP75 coated fibers with and without a

h-BN interlayer before and after heat treatment are displayed in figure 14.

Figure 14. Characteristic strength of as-received and coated sapphire

fibers grouped according to treatment. � Al2O3, � Al2O3,/h-BN

� Al2O3/IP75 � Al2O3/h-BN/IP75.

The characteristic strength of the as received sapphire fibers was

2.50 � 0,17 GPa. A typical fracture surface of an as received fiber is

displayed in figure 15.


- 72 -

Figure 15. Fracture surface of an as received fiber.

No large facets, steps or preferential crack direction can be

discerned. No significant change in �0 is detected after annealing and/or

etching of the as-received fibers. Similarly, no strength degradation of

fibers coated with only h-BN was measured before or after annealing.

Significant strength degradation was observed for the IP75 coated

fibers without a h-BN interlayer after the PVD process. Approximately 23%

(570 MPa) of the initial strength was retained for coated fibers and 35%

(880 MPa) with the coating etched away which indicates deposition

process induced damage of the fibers. No morphological changes were

detected on the fiber surface at the vicinity of the fracture surface within the

resolution of the SEM, and hence the strength degradation cannot be

explained by altered fiber surface morphology. The small but significant

increase in strength of etched fibers may be explained by the existence of


- 73 -

cracks in the IP75 coating, acting as stress concentrators on the fiber

during tensile testing. Such cracks in the IP75 may arise during cooling

from deposition temperature and/or during tensile testing.

Two distinct fracture surface morphologies (figured 16a and b) were

observed compared to the as received fibers (figure 15).

(a) (b)

Figure 16. (a) Fracture surface (type 1) of IP75 coated sapphire fiber

without BN interlayer after coating. A facetted fracture surface of weak

fibers may result from crack propagation along weakened twinned

interfaces of the Al2O3 crystal. (b) Fracture surface (type 2) of IP75 coated

sapphire fiber without h-BN interlayer after coating was mirror smooth,

referred to as fracture mirror in the literature [10].

The first one (type 1) was that faceting occurred (figure 16a) at an

angle of approximately 54° between the fracture plane and the fiber axis.


- 74 -

The three rhombohedral planes (r-planes) of the sapphire crystal lattice

have an angle of 57,6° to the c-axis. A deviation of up to 8° has been

reported in the literature for c-axis oriented sapphire fibers [11]. Hence, it is

reasonable to assume that the fracture planes are the r-planes of the

sapphire crystal lattice. The observed facetted fracture surface, commonly

referred to as ”coarse hackle”, was also reported by Firstone and Heuer

[12] on fractured c-axis oriented sapphire fibers, who also identified the

facets as the r-planes of the sapphire crystal lattice.

Type 2 fracture surface morphologies were observed on fibers which

separated during etching before tensile testing (figure 16b). The latter

indicates that they fractured during the deposition process, but were held

together by the coating. These fracture surfaces did not follow

crystallographic planes as the type 1 fracture, but were microscopically

smooth (figure 16b) and are therefore here referred to as fracture mirrors.

The fracture mirror is the fracture surface morphology of pre cracks

present in the fibers before tensile testing. Mirror morphologies are

typically observed during tensile loading as a consequence of slow crack

growth from a flaw [10]. Sherman and Be´ery [13] observed that fracture

mirrors in sapphire are generated by low velocity crack growth with low

kinetic energy in the direction with the maximum tensile component [14]


- 75 -

[13]. Hence, the existence of mirrors indicates that the fibers were

subjected to tensile stress before tensile testing.

After deposition bending of the fibers was observed (bending radius

approximately 40 cm). Bending is caused by an inhomogeneous coating

thickness distribution around the fibers which may induce tensile stress of

approximately 65 MPa. This stress is comparably low, and may not be

sufficient to induce the observed pre cracks. However, tensile stresses

may be generated locally by dislocation pile ups on the fiber matrix

interface. Koehler [15] showed that large tensile stresses are produced

ahead of an array of dislocations which pile up at a barrier (i.e. at the fiber

matrix interface here). Later, Stroh [16] showed that the stress

concentration around the piled up group of dislocations at the barrier can

be sufficient to initiate and grow a crack in a brittle material [16]. The stress

necessary for dislocation movement may be generated during cooling

which causes the tensile stresses locally on the fibers by the CTE

mismatch (Al2O3 (c-direction at 100°C) [5]: 6.9x10-6 K-1 and NiAl [17]:

15.1x10-6 K-1, where the latter is the main phase in IP75) and a

temperature difference during cooling. Thermal stress induced dislocation

movement may cause in turn localized tensile stress acting on fiber surface

segments. The notion of dislocation pile up induced fracture initiation on

the interface is supported by TEM investigations on NiAl/sapphire fiber


- 76 -

composites [18-20]. It was observed that high dislocation densities are

created in the NiAl due to the evolving thermal stresses in the NiAl matrix

during thermal cycling [18-20]. These build up considerable

inhomogeneous elastic strain fields on the fiber matrix interface which

according to reference [19] resulted in complete fiber fracture.

A simple experimental setup was constructed to test our notion on

fracture mirror formation during cooling. The experimental setup consists of

a single fiber suspended vertically by two alligator clips. The force acting

along the fiber c-axis in this configuration is less than 0.1 N (� 8 MPa) and

is induced by the weight of the clip. A current was applied through the IP75

coating to heat up the fiber. The current was increased until the fiber was

glowing. The temperature, measured with a pyrometer, was 980±20°C.

During heating and holding the fiber at approximately 980°C, no fracture is

observed. However, as the fiber is cooled, the fiber fractured. The fracture

surfaces of these fibers were mirror smooth. One example is shown in

figure 17. This supports the here presented notion that the fracture mirror

formation can take place during cooling from the deposition/annealing

temperature.


- 77 -

Figure 17. Fracture surface morphology of an IP75 coated sapphire fiber

without a BN interlayer that fractured during cooling after heating to 980°C.

Fracture surfaces exhibiting a combination of fracture mirror and

coarse hackle were also observed on fibers with degraded strength (figure

18).

Figure 18. Fracture surface (type 1 and 2) of IP75 coated sapphire fiber

without BN interlayer after tensile testing and subsequent etching.


- 78 -

Generally the larger the observed fracture mirror was the lower fiber

strength was measured. The strength of ceramic fibers is limited by stress

concentrating flaws such as surface cracks [14]. The size of the flaw

required to account for the observed strength degradation was evaluated.

The strength of a fiber containing a semi-elliptical surface crack was

calculated as a function of the crack length b (figure 19). Based on the

here observed fracture surface geometry (see for example figure 18), the

ratio between the axes a and b (see inset in figure 19) of an elliptic crack

was close to one. The critical fiber strength, �c, was calculated by

��c=KIC/F�(�b) [21] [22]. Stress intensity factors, F, for this fracture surface

geometry were taken from reference [22] and the fracture toughness of

sapphire was taken from reference [23] (KIC=2.4 MPa�m).

The measured tensile strength of fibers with well defined fracture

mirrors vs. crack length as determined from SEM images is shown in

figure 19. For comparison the calculated tensile strength vs. crack length

data based on reference [22] is also given. Very good agreement is

obtained. This is consistent with our previously introduced notion that the

presence of fracture mirrors is responsible for the strength degradation

observed here.


- 79 -

Figure 19. Critical strength vs. crack length. A comparison between

calculations based on the fracture geometry shown in figure 18 and the

tensile test data.

In contrast to the fibers coated with only IP75, no strength

degradation was measured for fibers with a h-BN interlayer after the PVD

process. For these fibers no pronounced type 1 or 2 fracture surfaces were

discerned.

The simulation results (figure 20) show the compressive axial

thermal stress, �zz, that develops in a sapphire fiber deposited with IP75

with and without a h-BN interlayer along its c-axis at different

temperatures.


- 80 -

Figure 20. FEM simulation results show the c-axis compressive stress and

resolved shear stress on the rhombohedral planes in the sapphire lattice

exerted on the fiber by the IP75 coating with and without a h-BN interlayer.

The twinning stress of sapphire after Scott and Orr. [24] and Schmid and

Harris [9] is displayed for comparison. The data from Scott and Orr indicate

that twin formation in fibers without a h-BN interlayer may take place during

cooling.


- 81 -

Sputtered IP75 coatings develop a columnar structure [25] which

considerably differs from the cast IP75 microstructure [26]. It is reasonable

to assume that lower thermal residual stresses develop due to the under-

dense microstructure. The simulation results may therefore be considered

as an upper boundary of thermal stresses that may develop during cooling.

For comparison, the compressive strength of sapphire along the c-axis is

given, which was established experimentally by Scott and Orr [24]. They

reported that twin formation on the r-planes of the sapphire crystal limits its

c-axis compressive strength. Hence, the resolved shear stress on the

r-planes is also displayed in figure 20, which is equal to the c-axis

compressive stress multiplied by the Schmid factor of 0.452 [9]. The

measured deviation of the crystal c-axis angle to the fiber axis discussed

earlier implies up to 5 % difference in resolved shear stress on the

r-planes, and therefore the resolved shear stress on specific r-planes may

be even higher than the values displayed here, implying a lower

compressive strength of the fibers. When two twins intersect, atoms at the

intersection are shared by two differently oriented twins, producing large

stress concentrations at the intersection which may lead to fracture. This

has been identified as the reason why the compressive strength is equal to

the twinning stress [27, 9, 24]. In the work of Scott and Orr [24], sapphire

rods (9 mm long and 3 mm across) were tested in compression at


- 82 -

temperatures ranging from 623 to 1373 K. It was observed that the twin

formation stress of samples with ground surfaces exhibited strong

temperature dependence, whereas polished samples exhibited an

athermal behavior (figure 20), indicating that the twin formation stress and

hence the compressive strength depends on the surface conditions, which

has also been observed earlier by Bertolot and Scott [27]. It is assumed

that a twin forms at a pre existing twin nucleus or local stress concentration

sites originating from surface damage induced during cutting, polishing or

grinding the specimen [27, 24]. Schmid and Harris [9] measured a strong

temperature dependence of the twin formation with polished samples,

similar to the ground samples by Scott and Orr. The data from Schmid and

Harris are also plotted in figure 20 for comparison. They did not attribute

the twin formation stress to surface finish. However, as displayed in

figure 20, the temperature region where Scott and Orr established a

dependence on surface finish (between 773 and 900 K), contains only one

data point in the work of Schmid and Harris [9]. The resolved shear stress

on the r-planes for twin formation by Scott and Orr [24], was as low as

12.6 � 6 MPa for polished samples above 773 K (figure 20). Based on

these literature values, the compressive stress on fibers without a h-BN

interlayer is exceeded by more than 100% during cooling from deposition

temperature (figure 20) and twinning has to be expected. The interface


- 83 -

between the twinned and untwinned crystal is weaker than the untwinned

crystal itself [28, 29], leading to a decrease in tensile strength. This is

consistent with the lower strength value measured for these fibers. The

fracture in a twinned sapphire crystal tends to follow the weak interface

[28, 29] which is consistent with the coarse hackle fracture surface

morphology observed here. According to the discussion above fibers with a

combination of type 1 and 2 fracture surfaces may contain r-plane twins as

well as fracture mirrors before testing.

According to the simulation results, the h-BN interlayer decreases

the developing thermal stresses below the twin formation stress during

cooling (figure 20), which is consistent with the observed fracture surface

morphologies as well as tensile strength results discussed earlier.

During annealing, IP75 coated fibers without a h-BN interlayer

fractured and tensile testing was therefore not possible. This has also been

observed on fibers coated with NiAl and may be due to thermally induced

stress causing fracture of the sapphire [30]. It is likely that the same

mechanism is at least in part responsible for the strength degradation of

these annealed fibers as for the “as deposited fibers” discussed earlier.

This is supported by the fracture surface analysis, where large smooth

fracture surfaces, i.e. mirrors, were identified, similar to the one displayed

in figures 16b and 17. The annealed and etched fiber surfaces also


- 84 -

contained cavities and pits, as well as acid resistant reaction products

adhering to the fiber surface (figure 21). Chemical analysis indicates that

these precipitates contain Ta and O, which may have been formed due to

chemical reactions between fiber and matrix. These may act as additional

stress concentration sites during cooling.

Figure 21. Fibers with IP75 coating without a h-BN interlayer exhibited a

rough surface with Ta rich precipitates after annealing. The surface

morphology indicates chemical reactions between fiber and matrix.

After annealing IP75 coated fibers with a h-BN, 13% (0.34 GPa) of

the strength was retained (figure 14). With the coating etched away, the

strength rose to 0,75 GPa, corresponding to 30% of the initial strength.

This indicates that the fibers are permanently damaged after the annealing

process. Generally much larger fracture mirrors were observed after


- 85 -

annealing than on the IP75 coated fibers after deposition discussed earlier

(figure 22a-b, where the frame in fig 22a marks the crack initiation site,

displayed in 22b). This is consistent with the comparably low fiber strength

measured.

An altered fiber surface morphology was observed at the vicinity of

the fracture surfaces where the cracks were initiated (figure 22b). Also

ridges (figure 22c) and grooves (figure 22d) were observed meandering

on the fiber surface, which may contribute to strength degradation.


- 86 -

(a) (b)

(c) (d)

Figure 22. (a) After annealing, the strength of the IP75 coated fibers with

h-BN interlayer were strongly degraded. The fracture surfaces generally

exhibited large fracture mirrors. (b) Magnification of the section marked in

figure 22a, shows an altered fiber surface morphology, which may arise

due to chemical reaction. Traces of Cu, Si and Cr were detected by

chemical analysis. Fracture surfaces analysis indicates that the crack


- 87 -

initiates at these sites. (c) Ridges and (d) grooves with traces of mainly Si

and Cr were observed on the fiber surface after etching of the annealed

IP75/BN coated fibers. These may form due to chemical reactions with the

fiber through cracks in the h-BN interlayer.

Traces of mainly Si, but also Cr and Cu, were detected by chemical

analysis on the ridges and in the groves, indicating that the morphological

changes may be due to chemical reactions on the fiber surface during

annealing. The meandering pattern indicates that these reactions may take

place through cracks in the h-BN interlayer, which may form during the

annealing process as a consequence of the thermal stress due to

differences in CTE between the constituents of the composite and/or due

to thermally induced pore formation [31]. The deposited h-BN interlayers

have 89% of the theoretical density [32]. Densification induced stress at

the composite fabrication temperature of 1300°C due to crystallization of

the turbostratic structure [2] may also lead to the observed cracking of the

BN interlayer. Cu most likely originates from the Cu clamps used to fasten

the IP75 sputtering target. Si may be introduced in the IP75 during the

casting process [33] and forms SiO with �-Al2O3 already at T >925°C [34].

Al2SiO5 is identified by X-ray diffraction in the cast IP75 alloy [26], and

hence Si is present in the alloy. Cr which is an alloying element in IP75 has


- 88 -

been observed to react with �-Al2O3 at 1250°C [35] (compared to 1300°C

used here). Hence the roughened surface morphology may be caused by

chemical reactions between �-Al2O3 fiber and Si impurities and/or Cr at the

annealing temperatures of 1300°C. The altered fiber surface morphology

may act as a stress concentration site, leading to fiber strength

degradation. The fracture mirror formation in fibers with a BN interlayer

after annealing may happen at locally tensile strained surface segments

due to dislocation pile up [18, 15, 16, 19, 20] and or inhomogeneous NiAl

coating thickness along the fiber as discussed above. Additionally mirrors

may also form at stress concentration sites on the surface during loading,

similarly to a notched specimen. This was observed in the work of

Sherman [13], where the introduction of a notch in the sapphire specimen

resulted in the formation of a large fracture mirror during loading. Stresses

sufficient for propagation of a fracture mirror may also arise at chemical

reaction sites in the fiber due to volume changes associated with the

reaction.


- 89 -

4.2.3. Tensile strength of h-BN/NiAl coated sapphire fibers with

a reactive Hf or Y interlayer

The tensile test results of Al2O3/h-BN/Y/NiAl (�) and

Al2O3/h-BN/Hf/NiAl (�) before, after 2, and after 20 hours annealing at

1350°C are displayed in figure 23.

Figure 23. Tensile strength of uncoated Al2O3 (�) Al2O3/h-BN/Y/NiAl

(�) and Al2O3/h-BN/Hf/NiAl (�) before, after 2, and after 20 h annealing at

1350°C.

The strength of uncoated fibers before and after annealing for 2 h

(�) is 2.42�[0.08, 0.07] GPa and 2.24�[0.09, 0.08] GPa, respectively. The


- 90 -

tensile strength of the as deposited fibers is reduced to

1.96�[0.26, 0.23] GPa for fibers with an Y interlayer and to

1.75�[0.21, 0.19] GPa for fibers with a Hf interlayer. After annealing, fibers

containing Y or Hf interlayers display a significant decrease in tensile

strength, to 1.13�[0.17, 0.14] GPa for the Y interlayer and to

1.17�[0.28, 0.24] GPa for the Hf interlayer. Uncoated fibers retain their

initial strength after annealing. Annealing for 20 h decreased the tensile

strength further to 0.83�[0.24, 0.19] GPa for Y interlayer and

0.96�[0.16, 0.14] GPa for Hf interlayer. This corresponds to 34 and 40% of

the as received strength.

Grooves meandering on the fiber surface were detected by SEM

after annealing the Al2O3/h-BN/Y/NiAl composite (figure 24a-c), where the

frame in figure 24a marks a section of the groove imaged in figure 24b.

Three regions, 1-3, are distinguishable in the micrograph of the groove

shown in figure 24b. Region 1: Al2O3 fiber surface with no morphological

changes, indicating that no detectable chemical interactions between fiber

and coating takes place in this region during annealing. Region 2: This

region exhibits a roughened morphology compared to region 1. Al and O

with traces of Y were detected by EDX, which is an indication of chemical

reactions between fiber and Y and/or diffusion of Y. Region 3: A deep

groove with large concentration of Y but also the presence of N and B was


- 91 -

detected. The altered surface morphology and the large concentration of Y

indicate chemical reactions between fiber and Y. On a BSE image (figure

24c), the Y enrichment in the groove is visible. Except for the large

particles in region 3, also smaller Y enriched particles in region 2 are

distinguishable.

(a) (b) (c)

Figure 24. (a) Grooves meandering on the fiber surface of h-BN/Y/NiAl

coated fibers were observed after annealing. (b) Magnification of a groove

indicates 3 distinguishable regions (regions 1-3). (c) Backscattered

electron (BSE) image indicates the presence of Y in the groove (region 3)

as well as in region 2.

Cross sections of the fiber-coating system were investigated after

annealing. A BSE image at the location of a crack in the h-BN interlayer is

displayed in figure 25.


- 92 -

Figure 25. Backscattered electron (BSE) image of the h-BN/Y/NiAl coating

on the Al2O3 fiber after annealing at the position of a crack in the h-BN

interlayer. The image suggests that chemical reactions take place through

cracks in the h-BN interlayer.

Here, the fiber and the different layers are visible, as well as a pore

in the NiAl matrix. The chemical analysis indicates that the Y interlayer

contains O, Al, and N, which is consistent with the chemical analysis on the

FIB cross sections discussed earlier. The Al content and the local growth

of the Y interlayer around the crack in combination with the pore formation

may be due to Al diffusion from the NiAl into the Y interlayer. The crack in

the h-BN interlayer supplies a surface diffusion path from the Y rich layer to

the Al2O3, and since Y3Al5O12 is more stable than �-Al2O3 [36], the groove

may form by chemical reaction between the fiber and Y.


- 93 -

As discussed earlier, cracks in the h-BN interlayer may be created

during cooling after deposition or after the annealing treatment, due to the

development of thermal stresses arising from extensive differences in the

CTE´s between the Al2O3 fiber (CTE=6�10-6 K-1) [5], h-BN

(CTE=1.1-8.6�10-6 K-1) [37] and NiAl (CTE=15�10-6 K-1) [17].

The fracture surfaces of fibers containing Y or Hf interlayer were

examined with SEM and EDX-Electron Probe Micro Analysis (EPMA), in

order to detect the fracture initiation sites. On fibers with a Hf interlayer a

groove in the h-BN interlayer with traces of Hf was detected at the position

of fracture initiation (figure 26a-b), which may be a result of reaction

between Hf and h-BN. This is consistent with the formation of HfB2 and

HfN as observed by XRD. At the corresponding location on the fiber

surface of the etched fiber, a ridge was detected (figure 26c).


- 94 -

(a) (b)

(c)

Figure 26. (a) Cross section of the fracture surface of a h-BN/Hf/NiAl

coated fiber after annealing. (b) A groove in the h-BN interlayer with the

presence of Hf indicates chemical reactions between Hf and h-BN. (c) On

the corresponding position on the fiber surface a ridge was detected, which

may be caused by surface diffusion of Al2O3 into the groove in the h-BN

interlayer.


- 95 -

As discussed earlier the ridges may be a result of surface diffusion

of the Al2O3 fiber during annealing into cracks in the h-BN interlayer. The

sole presence of Al and O in the ridge strengthens this notion. Ridge

formation on coated sapphire fibers due to surface diffusion of Al2O3 during

heat treatment into irregularities in the adjacent matrix, such as grain

boundaries, have been observed earlier [7].

Fracture surface analysis generally indicates that fracture is initiated

where the fiber surface morphology was altered. Grooves caused by

chemical reactions as discussed earlier are identified as fracture initiation

sites for Y coated fibers whereas ridges and cavities may serve as fracture

initiation sites for Hf coated fibers. Hf was detected in the cavities which

may indicate that they are formed due to localized reactions between Al2O3

fiber and Hf.

One example for the formation of a cavity is shown in the insert of

figure 27, which may have acted as the fracture initiation site for the

fracture surface shown in figure 27.


- 96 -

Figure 27. A fracture mirror on the fracture surface of a h-BN/Hf/NiAl

coated fiber may have formed due to stress concentration on a surface

flaw shown in the inserted figure.

At the fracture initiation sites smooth fracture surfaces topographies, so

called fracture mirrors, were observed, which may be a result of crack

growth initiated at surface flaws [21], such as a grooves (figure 26a-c), or

cavities (figure 26), or ridges (figure 26c) as described above. During

cooling, thermal stresses arise due to CTE mismatch between fiber and

matrix and flaws may act as a stress concentration site where crack growth

can occur. Hence, the observed fracture mirror topography suggests that

the fiber was fractured before testing. It is reasonable to assume that the

presence of fracture mirrors has an adverse effect on the fiber strength.


- 97 -

No significant difference in size or morphology of the surface

morphological features was noticed between fibers annealed for 2 and

20 h, which is consistent with the similar strength values measured.

According to the results presented here, the largest individual

challenge for maintaining high fiber strength is eliminating the thermal

stress induced fiber damage. We have shown that despite of a substantial

lowering of thermal stresses by weakening the fiber matrix bond with a

h-BN interlayer, the fiber strength is degraded due to thermal stress

induced fracture mirror formation at surface flaws. Possibly the strength

degradation related fiber damage can be prevented by lowering the

thermal stresses further e.g. by increasing the fiber volume fraction in the

IMC [8] and/or by modifying the fracture properties of the reinforcing fiber

and by avoiding chemical reactions on the fiber surface by the choice of

matrix and interlayer material.


- 98 -

List of references

[1] Intermetallic compounds, Vol. 2, D. B. Miracle, (John Wiley &

Sons Ltd, 1994).



W. Hu, T. Weirich, B. Hallstedt, H. Chen, Y. Zhong, and G. Gottstein,

Acta Materialia 54 (2006) 2473.


strengthening Laves phase for high-temperature applications III. Effects of alloying with Cr, B. Zeumer and G. Sauthoff,

Intermetallics 6 (1998) 451.

[4] Hydrogen uptake in alumina thin films synthesized from an

aluminum plasma stream in an oxygen ambient, J. M. Schneider,

A. Anders, B. Hjorvarsson, I. Petrov, K. Macak, U. Helmersson, and

J. E. Sundgren, Applied Physics Letters 74 (1999) 200.

[5] Data sheet for c-axis oriented sapphire fibers (www.photran.com)

[6] Al2O3 Fiber Strength Degradation in Metal and Intermetallic Matrix Composites, S. L. Draper and I. E. Locci, Journal of



Fiber-Reinforced Composites, J. B. Davis, J. Yang, and

A. G. Evans, Acta Metallurgica Et Materialia 43 (1995) 259.

[8] Thermal residual stress (TRS) analysis in continuous Al2O3 fiber

reinforced NiAl composite, H. Chen, W. Hu, A. Atiser, Y. Zhong,

and G. Gottstein, International journal of materials research 97

(2006) 10.


- 99 -

[9] Effects of crystal orientation and temperature on the strength of

sapphire, F. Schmid and D. C. Harris, Journal of the American

Ceramic Society 81 (1998) 885.

[10] Observing and interpreting fracture surface topography, D. Hull,

(Cambridge University Press, Cambridge, 1999).

[11] Slow Crack-Growth in Sapphire Fibers at 800-Degrees to 1500-Degrees-C, S. A. Newcomb and R. E. Tressler, Journal of the

American Ceramic Society 76 (1993) 2505.

[12] Creep Deformation of 0 Degree Sapphire, R. F. Firestone and

A. H. Heuer, Journal of the American Ceramic Society 59 (1976) 24.

[13] Fracture mechanisms of sapphire under bending, D. Sherman

and I. Be'ery, Journal of Materials Science 35 (2000) 1283.

[14] Fracture of Brittle Solids-Second Edition, B. Lawn, (Cambridge

University Press, 1975).

[15] The Production of Large Tensile Stresses by Dislocations, J. S.

Koehler, Physical Review 85 (1952) 480.

[16] The Formation of Cracks as a Result of Plastic Flow, A. N. Stroh,

Proceedings of the Royal Society of London Series A-Mathematical

and Physical Sciences 223 (1954) 404.

[17] Overview No.104 - The Physical and Mechanical-Properties of NiAl, D. B. Miracle, Acta Metallurgica Et Materialia 41 (1993) 649.

[18] Interfaces in Metal-Matrix and Intermetallic-Matrix Composites,

R. J. Arsenault, Composites 25 (1994) 540.

[19] Dislocations in continuous filament reinforced W/NiAl and Al2O3/NiAl composites, L. Wang, K. Xu, R. R. Bowman, and

R. J. Arsenault, Metallurgical and Materials Transactions A-Physical



- 100 -

[20] Interfaces in Continuous Filament-Reinforced Al2O3/NiAl

Composites, L. Wang, K. Xu, R. R. Bowman, and R. J. Arsenault,

Metallurgical and Materials Transactions A-Physical Metallurgy and


[21] Effect of flaw shape on strength of ceramics, G. K. Bansal,


[22] Stress intensity factors handbook, Vol. 2, Y. Murakami,

Pergamon press (1986) 659.

[23] Fracture Toughness of Quartz and Sapphire Single Crystals at Room Temperature, M. Iwasa, T. Ueno, and R. C. Bradt, Journal of

the Society of Materials Science, Japan 30 (1981) 1001.

[24] Rhombohedral Twinning in Alumina, W. D. Scott and K. K. Orr,


[25] Long fibre reinforced NiAl: High temperature material for

turbine blades, M. Rosefort, C. Dahmen, A. Buhrig-Polaczek,

W. P. Hu, H. Chen, Y. L. Zhong, G. Gottstein, D. E. Hajas, and

J. M. Schneider, Advanced Engineering Materials 8 (2006) 730.

[26] Microstructure and mechanical properties of continuous Al2O3

fibre reinforced Ni45Al45Cr7.5Ta2.5 alloy (IP75) matrix composites,

Y. Zhong, D. E. Hajas, W. Hu, H. Chen, and G. Gottstein,

Philosophical Magazine A 87 (2007) 1019

[27] Compressive Creep of Al2O3 Single Crystals, R. l. Bertolot and

W. D. Scott, Journal of the American Ceramic Society 54 (1971)

286.

[28] Mechanical Behavior of Sapphire, H. Conrad, Journal of the

American Ceramic Society 48 (1965) 195.


- 101 -

[29] Fracture and Twinning in Sapphire (Alpha-Al203) Crystals,

E. Stofel and H. Conrad, Transactions of the Metallurgical Society of

AIME 227 (1963) 1053.

[30] Effect of a BN interlayer on the tensile strength of NiAl coated

sapphire fibers, D. E. Hajas, S. Kyrsta, and J. M. Schneider,

Scripta Materialia 55 (2006) 219.

[31] Effect of a boron nitride interphase that debonds between the

interphase and the matrix in SiC/SiC composites, G. N.

Morscher, H. M. Yun, J. A. DiCarlo, and L. Thomas-Ogbuji, Journal

of the American Ceramic Society 87 (2004) 104.

[32] Constitution and deposition mechanisms of hexagonal boron nitride formed by CVD from trimethylborazine

A. Jörg, E. Zimmermann, M. Schierling, R. Cremer, and

D. Neuschütz, in CVD XIV and Euro CVD-11, The Electrochemical

Society (Pennington, USA, 1997).

[33] Production-scale processing of a new intermetallic NiAl-Ta-Cr

alloy for high-temperature application Part I. Production of

master alloy remelt ingots and investment casting of combustor liner model panels, M. Palm, J. Preuhs, and G. Sauthoff, Journal of

Materials Processing Technology 136 (2003) 105.

[34] Silicon-on-Sapphire Epitaxy by Vacuum Sublimation - Leed-

Auger Studies and Electronic Properties of Films, C. C. Chang,

Journal of Vacuum Science & Technology 8 (1971) 500.

[35] Effect of Cr on Interfacial Interaction and Wetting Mechanisms in Ni-Alloy Alumina Systems, P. Kritsalis, V. Merlin, L. Coudurier,

and N. Eustathopoulos, Acta Metallurgica Et Materialia 40 (1992)

1167.


- 102 -

[36] Reaction paths in the system Al2O3-hBN-Y, K. Reichert,

O. Oreshina, R. Cremer, and D. Neuschütz, Applied Surface

Science 179 (2001) 138.

[37] Hexagonal Boron Nitride: Fabrication, Properties and

Applications, A. Lipp, K. A. Schwetz, and K. Hunold, Journal of the

European Ceramic Society 5 (1989) 3.

5. Conclusions

- 103 -

5. Conclusions

One way to increase the efficiency of a gas turbine is by increasing

the combustion zone temperature. For this new materials are needed that

can withstand the increase in temperature without degradation of their

mechanical properties. Sapphire fiber reinforced NiAl is of special interest

for the combustion zone turbine blades. However, experimental results

have shown that the strength of the fibers is reduced during composite

fabrication. The goal of this thesis was to identify strength degradation

relevant mechanisms active between �-Al2O3 fibers and the surrounding

matrix or interlayer, in order to minimize the strength degradation during

production of IMC’s in the future. These data are important for formulating

future strategies for the production of IMC´s.

Strength degradation mechanisms of NiAl or IP75 coated fibers with

and without an Y (or Hf) and/or h-BN interlayer have been assessed.

Fracture- and fiber- surfaces after tensile testing were evaluated by

microscopy and chemical analysis. The effect of temperature during

diffusion bonding as well as the presence of an interlayer on the tensile

strength has been studied.

5. Conclusions

- 104 -

Strength degradation was measured for IP75 coated fibers without a

h-BN interlayer after deposition, indicating the introduction of deposition

process related damage. This notion is supported by the observation of

fracture mirrors and coarse hackle fracture surface morphologies. The

coarse hackles indicate twinning of the rhombohedral Al2O3 crystal planes,

which was predicted by simulation results as a consequence of the

evolving stress during cooling after deposition. Surface flaws in

combination with thermal stress on the fiber matrix interface may cause

local stress concentrations on the fiber, which may be responsible for the

fracture mirror formation.

Tensile testing of annealed NiAl or IP75 coated fibers without a h-BN

interlayer was not possible due to fracturing of the fibers. Large fracture

mirrors were observed, which indicates that thermal stress induced fracture

mirror formation during cooling from composite fabrication temperature is

the main strength degradation mechanism after annealing, when no h-BN

interlayer is present.

IP75 or NiAl coated fibers containing a h-BN/Y (or Hf) or a h-BN

interlayer degraded their strength after annealing. Fracture mirrors with

altered fiber surface morphology at the vicinity of the fracture surface

caused by Al2O3 surface diffusion into cracks or irregularities in the

5. Conclusions

- 105 -

interlayer material as well as chemical reactions between fiber and matrix

or interlayer elements (Cr, Y or Hf) or impurities (Si) were observed. The

fracture mirror formation may be a consequence of the developing thermal

stresses at the sites with altered surface morphology causing fiber strength

degradation.

Based on these observations the following strength degradation

relevant mechanisms for coated sapphire fibers with or without an

interlayer are proposed: Twinning of the rhombohedral plane (r-plane) of

the sapphire crystal as a consequence of the evolving thermal stress

during cooling. Chemical reactions between fiber and matrix material

and/or impurities and surface diffusion of Al2O3 into irregularities in the

adjacent matrix material like cracks in the h-BN interlayer which both lead

to altering of the fiber surface morphology. In combination with the evolving

thermal stress during cooling, fracture mirror formation at the surface flaws

takes place, degrading the fiber strength.

In this work the fiber strength degradation mechanisms active during

individual process steps are identified. The significance of thermal stress

induced fiber damage is emphasized and it is concluded that the

elimination or a significant reduction thereof is the largest challenge which

5. Conclusions

- 106 -

has to be addressed to explore the strengthening potential of sapphire

fibers for IMC´s in the future.

Based on the results presented here, a strategy for obtaining high

strength IMC´s in the future is compiled. It is suggested that the strength

degradation may be avoided or minimized by: lowering the thermal stress,

by increasing the fiber volume fraction, and/or by using a ductile interlayer.

6. Future work

- 107 -

6. Future work

In this work it has been concluded that a significant reduction or

elimination of thermal stress induced fiber damage is the largest challenge

which has to be addressed in order to make use of the strengthening

potential of sapphire fibers for IMC´s in the future. It is suggested that the

strength degradation may be avoided or minimized by: i) lowering the thermal

stresses, ii) modifying the fracture and twinning properties of the

reinforcing fiber and/or iii) inhibiting the altering of the fiber surface

morphology due to chemical reactions or diffusion into irregularities in the

surrounding matrix or interlayer material.

Thermal stresses can be lowered by increasing the fiber volume

fraction in the IMC [1] or by introducing a damage tolerant interlayer

referred to as compliant interlayer in the literature [2], which

accommodates stress by yielding. Such a ductile interlayer also inhibits

crack formation on the interface during thermal cycling and impedes crack

propagation from the matrix into the fiber [2]. An example for such an

interlayer may be V2AlC [3].

The twin formation stress in sapphire can be altered by the addition

of dopants. It has been observed that the twinning stress of sapphire is

increased 3 fold by addition of Ti to the Al2O3 [4]. This is explained by

6. Future work

- 108 -

creation of stacking faults on the twin planes, which makes twin

propagation more difficult [4]. The effect of addition of other elements on

the fracture and twinning properties has not been investigated and may be

aim of future investigations. However, collaboration with the fiber

manufacturer is needed. The fiber properties may also be altered by

replacing the sapphire fiber with other another fiber material. Several

commercial ceramic fibers exist, like SiC fibers (e.g. SCS6 by Specialty

materials inc.) as well as poly-crystalline Al2O3 fibers (e.g. Nextel series by

3M) which both have been developed for high temperature applications [2].

The SCS6 fiber is of special interest due to its excellent high temperature

strength, > 3 GPa, at 1000°C, which is approximately three times higher

than the strength of the sapphire fibers used here.

SiC fibers have been incorporated in several matrix materials such

as: NiAl [5], FeAl [6] and TiAl [7] and TiAlNb [8] [9]. However, from the

literature it can be learned that chemical reactions between fiber and

matrix has limited its use. In order to utilize SiC as reinforcing fibers, these

chemical reactions have to be inhibited. This could be achieved by

introducing a chemically stable interlayer between fiber and matrix which

should also be ductile as discussed earlier. Ductile interlayers like Mo and

Nb have been investigated. However, chemical stability is not achieved for

these interlayers [2].

6. Future work

- 109 -

By reviewing the reaction products reported in the literature, it is

possible to identify reaction products which possess properties that may be

suited as interlayer. For instance it has been reported in the literature that

Ti3SiC2 forms as a reaction product between a TiAl and SiC, and is hence

chemically stable with SiC. This is also supported by the quaternary

Ti-Si-Al-C phase diagram, where a “tie line” exists between SiC and

Ti3SiC2 up to >1700°C [10]. Ti3SiC2 belongs to a group of materials called

MAX phases and which are known to possess some ductility at room

temperature [11]. The chemical stability in combination with ductility makes

this phase interesting as interlayer material. Ti3SiC2 can be deposited by

magnetron sputtering [12]. Initial growth of a TiCx interlayer was observed

on both MgO and Al2O3 substrates. However, growth characteristics of

Ti3SiC2 on SiC substrate have still not been investigated and can be the

scope of future investigation. Also the chemical stability between NiAl and

Ti3SiC2 remains to be investigated.

Deposition of the above discussed interlayers and chemical stability

investigations should be the primary scope of future work, followed by

assessment of strength degradation mechanisms using the methodology

presented in this thesis.

6. Future work

- 110 -

List of references

[1] Thermal residual stress (TRS) analysis in continuous Al2O3 fiber

reinforced NiAl composite, H. Chen, W. Hu, A. Atiser, Y. Zhong,

and G. Gottstein, International journal of materials research 97

(2006) 10.

[2] Intermetallic Matrix Composites III, Vol. 350, A. K. Misra, edited

by R. R. Bowman. J.A. Graves, J.J. Lewandowski (Materials

Research Society, San Francisco, 1994) 73.

[3] Tensile strength of NiAl coated sapphire fibers with a V2AlC

interlayer, G. Gebhardt, D. E. Hajas, B. Cappi, R. Telle, and

J. M. Schneider, Submitted for publication (January 2008).

[4] The effect of titanium doping on the high-temperature

rhombohedral twinning of sapphire, E. Savrun, W. D. Scott, and

D. C. Harris, Journal of Materials Science 36 (2001) 2295.

[5] Deposition, Characterization, Thermochemical Compatibility

and Failure Analysis of Multilayer Coated Silicon-Carbide Fiber-Reinforced Metal-Matrix Composites, R. S. Rastogi and

K. Pourrezaei, Journal of Materials Processing Technology 43

(1994) 89.

[6] Compatibility of Fe-40a1 with Various Fibers, S. L. Draper,

D. J. Gaydosh, M. V. Nathal, and A. K. Misra, Journal of Materials

Research 5 (1990) 1976.

[7] Reaction Zone Microstructure in a Ti3Al + Nb/SiC Composite,

S. F. Baumann, P. K. Brindley, and S. D. Smith, Metallurgical

Transactions a-Physical Metallurgy and Materials Science 21 (1990)

1559.

6. Future work

- 111 -

[8] Interface Control and Design for Sic Fiber-Reinforced Titanium

Aluminide Composites, H. P. Chiu, S. M. Jeng, and J. M. Yang,

Journal of Materials Research 8 (1993) 2040.

[9] Effect of Fiber Coating on the Mechanical-Behavior of SiC

Fiber-Reinforced Titanium Aluminide Composites, S. M. Jeng,

J. M. Yang, and J. A. Graves, Journal of Materials Research 8

(1993) 905.

[10] Phase equilibria at 1000 degrees C in the Al-C-Si-Ti quaternary

system: an experimental approach, J. C. Viala, N. Peillon,

F. Bosselet, and J. Bouix, Materials Science and Engineering A-

Structural Materials Properties Microstructure and Processing 229

(1997) 95.

[11] The M(N+1)AX(N) phases: A new class of solids; Thermodynamically stable nanolaminates, M. W. Barsoum,

Progress in Solid State Chemistry 28 (2000) 201.

[12] Growth of Ti3SiC2 thin films by elemental target magnetron sputtering, J. Emmerlich, H. Hogberg, S. Sasvari, P. O. A. Persson,

L. Hultman, J. P. Palmquist, U. Jansson, J. M. Molina-Aldareguia,

and Z. Czigany, Journal of Applied Physics 96 (2004) 4817.

Curriculum Vitae

David Hajas

Passstrasse 99 52070 Aachen E-Mail: [email protected] Persönliche Daten geboren am 11.06.1976 in Jönköping, Schweden Nationatilität: schwedisch Familienstand: verheiratet, ein Kind Berufstätigkeit 06/2006 – heute Leiter der Arbeitsgruppe für Werkstoffcharakterisierung,

Lehrstuhl für Werkstoffchemie der RWTH Aachen, Deutschland 10/2002 – heute Wissenschaftlicher Mitarbeiter, Lehrstuhl für Werkstoffchemie der

RWTH Aachen, Deutschland 01/2001 – 12/2001 Laborbetreuung, Institut für Physik, „Linköpings Universitet“,

Linköping, Schweden 06/1998 – 08/1998 Lehrer, „Erik Dahlbergsgymnasiet“, Jönköping, Schweden

- Unterricht in Mathemathik und Physik (Sommerschule) 01/1998 – 12/1998 Vorsitzender der Studienrichtungsvertretung (SNOrdf), „Linköpings Universitet“, Linköping, Schweden

08/1995 – 06/1996 Lehrhilfe in Mathemathik, Physik, Chemie und Biologie, „Per

Brahegymnasiet“ und „Junedahlskolan“, Jönköping, Schweden

Ausbildung 10/2002 – 01/2008 Promotionsstudium, Lehrstuhl für Werkstoffchemie der RWTH

Aachen, Deutschland - Faserschädigungsmechanismen in NiAl/Al2O3-

Verbundwerkstoffen - Abschluss: Doktor der Naturwissenschaften

09/1996 – 03/2002 Studium, ”Linköpings Universitet”, Linköping, Schweden

- Physik mit Schwerpunkt Festkörperphysik - Abschluss: Master of Science

09/1999 – 06/2000 „Universidad de Alicante“, Alicante, Spanien - Sprachstudium

09/1992 – 05/1995 „Erikdahlbergsgymnasiet“, Jönköping, Schweden

- Abitur mit Schwerpunkt Chemie Sprachen Schwedisch (Muttersprache) Ungarisch (Muttersprache) Englisch (Verhandlungssicher) Deutsch (Verhandlungssicher)

Spanisch (Fließend)

Documents

Strength degradation mechanisms in NiAl alloy coated ... · Strength degradation mechanisms in NiAl alloy coated sapphire fibers Von der Fakultät für Georessourcen und Materialtechnik