All-Oxide Ceramic Matrix Composites Thermal Stability during Tribological Interactions with Superalloys

(1)

All-Oxide Ceramic Matrix Composites

Thermal Stability during Tribological Interactions with Superalloys

Daniel Vazquez Calnacasco

Materials Engineering, master's level (120 credits)

2021

Luleå University of Technology

(2)

– i –

To my parents, Carmen and Héctor.

“We make our world significant by the courage of our questions and the depth of our answers.” – Carl Sagan

(3)

– ii –

Preface

This project was performed between September 2019 and May 2021 as part of the Advanced Materials Science and Engineering (AMASE) Master Program, coordinated by the European School of Materials (EUSMAT) through an Erasmus+ scholarship.

The work focused on the interactions between a ceramic matrix composite and a superalloy when subjected to tribological testing and was carried out under the supervision of professors Marta-Lena Antti and Farid Akhtar at the Division of Engineering Materials of Luleå University of Technology (Sweden) in collaboration with GKN Aerospace Engine Systems, Sweden.

The composites studied in this work are often referred to in the literature with different terminologies involving the acronym “CMC” for Ceramic Matrix Composites, preceded by a suffix, such as in: i) “Oxide” or “All-Oxide” CMC (OCMC), ii) Oxide-Oxide CMC (Ox-Ox or Ox/Ox CMC), iii) Continuous-Fiber Ceramic Composites (CFCC), iv) Long Continuous-Fiber Composites (LFC) v) Ceramic Continuous-Fiber-Matrix Composites (CFMC) and vi) Fiber Reinforced Ceramics (FRC & FRCMC). In this work the term OCMC is preferred.

(4)

– iii –

Abstract

The challenges faced in today’s industry require materials capable of working in chemically aggressive environments at elevated temperature, which has fueled the development of oxidation resistant materials. All-Oxide Ceramic Matrix Composites (OCMC) are a promising material family due to their inherent chemical stability, moderate mechanical properties, and low weight. However, limited information exists regarding their behavior when in contact with other high-temperature materials such as superalloys. In this work three sets of tribological tests were performed: two at room temperature and one at elevated temperature (650 °C). The tests were performed in a pin-on-disk configuration testing Inconel 718 (IN-718) pins against disks made with an aluminosilicate geopolymeric matrix composite reinforced with alumina fibers (N610/GP). Two different loads were tested (85 and 425 kPa) to characterize the damage on both materials.

Results showed that the pins experienced ~ 100 % wear increase when high temperature was involved, while their microstructure was not noticeably affected near the contact surface. After high temperature testing the OCMC exhibited mass losses two orders of magnitude higher than the pins and a sintering effect under its wear track, that led to brittle behavior. The debris generated consists of alumina and suggests a possible crystallization of the originally amorphous matrix which may destabilize the system.

(5)

– iv –

Acknowledgements

“If I have seen further, it is by standing upon the shoulders of giants” – Isaac Newton

I am deeply grateful with my supervisor Professor Marta-Lena Antti for her support during these complicated years, thank you for being so accessible, understanding, and empathic; for always taking the time to listen to my ideas and inspire me with your passion, while sharing a cup of tea. I would also like to thank you for introducing me into this amazing field that I had so longed for, I enjoyed every step and I hope these new steps take me in a new direction. I always admired the clarity of your mind, and the precision of your ideas, I hope one day I will also master those skills.

A sincere thank you to Professor Farid Akhtar for his patience and guidance, for all the practical problems he solved during the development of the project and for his sharp ideas and interesting perspectives which shaped the evolution of this work.

I would also like to thank the team of scientist and engineers that shared a little bit of their expertise (and a big bit of their good mood) with us every time we met: Géraldine, Per, Bengt, Guillaume, thank you for all the nice and fruitful remarks and ideas; for making a space for us in your (probably very busy) schedules, and for your availability and commitment.

A huge acknowledgment goes to the team at the European School of Materials (EUSMAT) and to Professor Flavio Soldera who are always keeping an eye on us students. To the coordinators of the program in each of my home universities Professors David Horwat and Lennart Wallström, thank you for your help along the way. And an enormous (and maybe uncommon) thank you to the European Union for the financial support without which this dream could have never been possible.

I have been lucky enough to develop this project in a great working environment, and I would like to express my gratitude to those who made my everyday life a really nice experience: special thanks to Lars and Erik for all the pedagogic crosswords and quirky conversations; and for sharing with me your vast practical and theoretical knowledge; to Leonardo Pelcastre, thank you for the profilo-therapy sessions and for all the good advice. To my generation friends Gabriel and Mercedes, for being there with me when things got hard, I hold you in a very special place in my heart. To my new friends, Khalifa, Arai, Kim, Nacho, Myrna, Anish, Adriá, Rahma, Yago, Bilal, Mariana, thank you for the (corona-friendly) fikas and dinners, for the laughs and conversations, I could have not asked for better company.

(6)

– v –

Abbreviations

/A Alumina matrix

/AS Aluminosilicate matrix

/GP Geopolymeric matrix

/M Mullite matrix

/M-A Mullite-alumina matrix

3YSZ 3 mol% yttria stabilized zirconia

Al2O3 Alumina

AlPO4 Berlinite

B4C Boron carbide

CTE Coefficient of thermal expansion CVD Chemical vapor deposition

E Young’s modulus

EDS Energy-dispersive x-ray spectroscopy

IN718 Inconel 718

KIC Critical stress intensity factor

lc Critical fiber length

LTU Luleå tekniska universitet , Luleå technical university MgAl2O4 Magnesium aluminum spinel

MgO Magnesia n Number of mols N610 Nextel™ 610 fiber N720 Nextel™ 720 fiber NH Non-homogeneous Ø Diameter

OCMC All-oxide ceramic matrix composite(s)

OM Optical microscopy

Q Quadrants for traceability during characterization

RT Room temperature

SiC Silicon carbide

SiO2 Silica Tm Melting temperature ZrO2 Zirconia γ Surface tension Γ Fracture energy ε Strain εT Emissivity ν Poisson’s ratio σ Stress

σUTS Ultimate tensile strength

(9)

– 1 –

Chapter 1

1. Introduction

The technological trend in the aerospace sector points towards the use of higher engine operating temperatures and pressure ratios to improve the propulsion system’s efficiency while decreasing noise and emission generation [1], [2]. Current research efforts focus on the development of novel light-weight materials for combustion environments capable of working at higher loads and temperatures than current superalloys. The components manufactured with these new materials would present less stringent maintenance and cooling requirements. This has fueled the growth of the Ceramic Matrix Composites (CMCs) market, whose value is expected to increase from $ 4,857.6 million USD in 2018 to 11,516.1 million USD by 2026 [2]–[4].

(10)

– 2 –

Figure 1.1.- Global CMC market share per composite type and expected growth [12].

The rapid degradation of components by oxidation has led to the assessment of other materials for this application, oxide ceramics being particularly appealing due to their inherent resistance to oxidation at temperatures above 1000 °C [14] as well as other advantages such as: i) high melting point, ii) low density, iii) chemical stability and iv) good mechanical properties in terms of hardness, tensile strength, elastic modulus, and creep resistance. However, oxides present a mixture of covalent and ionic interatomic bonds that results in a brittle nature, which in component design translates as low reliability, discarding them as viable manufacturing materials [15].

The brittleness of oxide ceramics is overcome by the introduction of other phases acting as reinforcements either in the form of fibers, whiskers, or particles, which allows them to exhibit: i) ductile-like behavior; ii) increased fracture toughness, iii) high resistance to thermal shock [5]–[10] and iv) improved dynamic load capacity [16]. This combination of phases receives the name of composite, the introduced phase is called “reinforcement” and the reinforced material receives the name of “matrix”. When the chemical composition of both matrix and reinforcement is based on oxides, the composite receives the name of All-oxide Ceramic Matrix Composite (OCMC).

All-oxide ceramic matrix composites (OCMC) have become the second most required material in 2018 slightly above C/C designs (Figure 1.1) according to a study published in 2019 by Allied Market Research, being a material family capable of working in high-temperature and chemically aggressive environments; however, their use is restricted to low-load applications due to their limited mechanical properties in comparison with non-oxide designs. The current OCMC field is mainly focused on their mechanical performance, in particular on the improvement of the inherently weak interlaminar strength and creep resistance, the latter being strongly influenced by the reinforcing fiber and its integrity, which is currently damaged during the manufacturing process due to the high temperatures involved [4].

(11)

– 3 –

No studies have been reported in the literature regarding i) the thermal behavior of this matrix design, in particular its emissivity which is the main heat transfer mechanism used to avoid component overheating due to the low thermal conductivity of oxide ceramics [19], [20]; and ii) its tribological performance which is an important aspect to consider since those components manufactured for hot aerospace applications will most likely be in contact with other high-temperature materials in their working environment. This work aims at filling the knowledge gap by subjecting a new OCMC design manufactured with a geopolymeric matrix and reinforced with alumina fibers, to a tribological test against Inconel 718 (IN-718).

There are three main chapters in this document: i) this first one provides an overview of the work, its objectives, relevance and scope; ii) the second chapter is a review of the relevant aspects of the state of the art of all-oxide ceramic matrix composites, aimed at any professional interested in the topic, with useful references and basic descriptions for those new to the field; iii) the third section is a scientific report that details the practical knowledge generated during the development of the experiments as well as a description of the data acquisition methodology, analysis of the results, discussion and conclusions.

1.1. Framework and Objectives

The scope of this work is focused on the behavior of a fiber reinforced geopolymeric aluminosilicate matrix OCMC (further referred to as “N610/GP”) when used as a disk in pin-on-disk tests, subjected to sliding contact against pins manufactured with the superalloy Inconel 718 (IN-718). The test schedule contemplates two testing temperatures: 25 and 650 °C, with the application of two different loads at each temperature i.e. 85 and 425 kPa. The study also contemplates the measurement of the thermal emissivity of the N610/GP composite in the testing range (25 °C – 650 °C).

The objective of this work is to answer the following research questions:

i. Which are the wear mechanisms, and the extent and nature of the degradation experienced by the materials after a tribological test at room temperature (RT)?

ii. How does high temperature affect both materials and their response to the test? iii. What is the nature of the debris generated?

iv. How does the thermal emissivity of a geopolymeric matrix OCMC change with temperature?

Since the OCMC material is already defined for this work, there are several aspects of composite manufacture and analysis that will not be covered in depth in this work, such as i) other reinforcing geometries besides the N610 fibers (particles, whiskers, sheets); ii) other possible chemical compositions for matrix or reinforcement (organic compounds, carbides, nitrides); iii) the study will also be limited to the tribological interaction between the materials studied, and although a brief scope of the tensile performance of the composite is provided, no testing is involved in the experimental part regarding tensile, flexural or compressive properties.

1.2. Methodology and Limitations.

(12)

– 4 –

was designed to accurately evaluate the effect of different parameters on the tribological interactions between the geopolymer matrix composite and IN-718, however it cannot fully represent the behavior of the material in working conditions, which can be further assessed through “component”, “bench” and “field” tests.

A pin-on-disk test was selected over the reciprocating-pin or reciprocating-cylinder wear tests due to its similarity with the contact geometry in working conditions. The N610/GP sample was only tested in the disk position due to the sample geometry provided by the manufacturer (sheet form). The testing parameters, namely temperatures and loads, were defined in collaboration with GKN to simulate as accurately as possible the working conditions of the component during its lifetime. After testing, the degradation of the materials was studied directly in the contact surfaces, and the extent of the damage was measured in their cross-sections using the following characterization techniques: Scanning Electron Microscopy (SEM), Optical Microscopy (OM), Energy Dispersive Spectrometry (EDS) and microhardness. A study of the debris generated was also performed using X-ray diffraction (XRD) and EDS. The emissivity test (thermal properties) was performed through an external service provider, and limited information was provided regarding the equipment or methodology used.

1.3. Advantages and Benefits of OCMC

The advantages of OCMCs are determined depending on the material to which they are compared. Against superalloys, their main advantages are in terms of weight reduction and higher operating temperatures: since i) lighter rotating components already imply a reduction of the engine weight, but also have the beneficial side effect of ii) subjecting the shaft to lower stresses, allowing thinner shaft designs and further reducing weight; furthermore, the high-temperature capabilities of these materials would allow other benefits such as iii) allowing engines to work at their ideal combustion temperatures and to iv) produce thrust with the air volume currently diverted for cooling, improving the overall efficiency of the system; and v) simplifying component design and manufacturing by eliminating the need for the complex cooling channels present in today’s designs due to the current material limitations; These improvements would lead to vi) lower fuel consumption and emission generation. When the comparison is made against non-oxide designs, OCMCs also exhibit some advantages such as i) superior fracture toughness ii) better long term durability in oxidizing environments under modest loads up to 1200 °C (after which the degradation of the fibers occurs) [21]; iii) lower thermal conductivity which also makes them interesting as thermal protection systems for spacecraft and hypersonic vehicles [22]; iv) similar price, for instance a kilogram of N720/AS is in the range of 4.99 x104_{– 5.96 x10}4 _{SEK, while a kilogram of SiC/SiC composite lies between 2.93 x10}4_{– 5.18 x10}4_[23];

v) low probability of catastrophic failure originated at pores or notches [24], [25] vi) easy oxide fiber fabric manufacture due to their lower stiffness and strength [26].

The advantages mentioned above come at the expense of i) reduced mechanical performance (σUTS,

(13)

– 5 –

1.4. Commercial Providers

OCMCs have been developed for decades and are currently commercialized by a few companies and institutions summarized along with their products in Table 1.1. General Electrics (GE) and the University of California, Santa Barbara (UCSB) have also produced CMCs in the past but they are not currently active [4]. The School of Aerospace Engineering of the Tsinghua University in China, and the Solid-State Physics Institute of the Russian Academy of Sciences are also performing research on the topic. The latter has proposed a new method to manufacture and weave oxide fibers to reduce fabric cost with promising results [28].

Table 1.1.- OCMC manufacturing companies and institutions and their commercial designs.

Company | Institution Developed products

Airbus

Group Innovations branch (Germany) UMOX

Alliance Techsystems (ATK) - Composite

Optics Inc. (COI) (United States) COI-610/AS, COI-720/AS, COI-720/A. Axiom Materials

(United States) [13]

AX-7820-610

(Axiom N610/A and N720/AS designs) Cytec Ltd.

(UK) [25] N720/A

Fraunhofer Institute for Ceramic

Technologies and Systems (Germany) Custom designed oxide CMCs with no brand General Electrics

(United States) Ox-Ox CMC, GE-610/GEN-IV

Institute of Materials Research, High-temperature and Functional Coatings

– German Aerospace Center [29]–[32] WHIPOX, UMOX, OXIPOL Pyromeral Systems

(France) [33], [34]

PyroXide and the non-oxide designs PyroSic and PyroKarb

University of California, Santa Barbara

(United States) UCSB-610/M, UCSB-720/M

Walter E.C. Pritzkow Spezialkeramik (Germany) [21], [28]

Keramikblech designs with an A + 3YSZ matrix and N610 fiber reinforcement (FW12, FW30) WPX Faserkeramik GmbH

(Germany) [11] WHIPOX, WPX-610/A

1.5. Industrial Applications

(14)

– 6 –

Another example, taken from a sinter metal production plant, concerns a lift gate used to divide two furnace chambers, one with an oxidizing atmosphere at 600 °C and another one with a N2/H2

atmosphere between 1100 and 1280 °C. Metallic materials and carbon/carbon composites were discarded due to the stringent working environment however the use of an OCMC component allowed more than 29 200 h of un-interrupted production.

A) B) C) D)

Figure 1.2.- Comparison of a metallic flame tube A) before and B) after 1 000 h operation, and an OCMCs flame tube C) before and D) after 20 000 h operation [21].

Several tests have been performed in engine combustion environments. In 2003 a set of all-alumina CMCs (fiber and matrix) protected with a Friable Graded Insulation (FGI) was field tested in the combustion liners of a Centaur 50S gas turbine generator, logging 12 582 h (other sources claim 25 000 h and 109 starts [4]) and showing minor degradation under operating conditions [35], [36]. In 2014 Boeing finished ground- and flight tests of OCMCs for its CLEEN program. The tests showed better thermal performance than Inconel and 20 % less weight than titanium while operating continuously at 816 °C [1]. This company also finished the sub-element and sub-component testing of a nozzle assembly using a N610/AS center-body (Figure 1.3) with the ground testing and prototype flight being expected between 2015 and 2016.

Other Boeing projects involving OCMCs are related to the design of the exhaust system of a Rolls-Royce Trent 1000 engine working in collaboration with ATK-COIC and Albany Engineered Composites. The company has also been working on the use of OCMCs to shield thermal protection systems (constituted of lightweight ceramic foam tiles) against wear, impact and high-temperature conditions [4].

(15)

– 7 –

Another company working with OCMCs is GE Aviation, which in 2011 manufactured a divergent exhaust seal for the F414 engine using an OCMC internally called “Ox-Ox”. In 2014 this company used Ox-Ox to replace a graphite/epoxy composite as manufacturing material for the four piece panel that encloses the engine core and the nine piece mixer (Figure 1.4), as well as other components such as the centerbody and core cowls of its Passport-20 engine which powers Bombardier 7 000 and 8 000 ultra-long range business jets, saving around 20 kg in the mixer alone [1].

Figure 1.4.- GE Aviation mixer and centerbody manufactured with OCMC [37].

Other components manufactured with porous OCMCs are nozzle flaps for fighter aircraft, helicopter exhaust ducts (Figure 1.5), seal/casing shrouds and stationary vanes in turbine engines [4]. Currently the replacement of titanium parts in subsonic jet engines with OCMCs is being studied with the aim of improving component durability without switching to heavier alloys such as Inconel [4].

Figure 1.5.- OCMC helicopter exhaust system [4].

(16)

– 8 –

Figure 1.6.- NASA Glenn Research Center and Rolls Royce Liberty Works OCMC subscale mixer nozzle assembly [4].

(17)

– 9 –

Chapter 2

2. Background

This chapter is divided in three parts, the first one provides an overview of the state of the art regarding all-oxide ceramic matrix composites (OCMCs) focusing on the geopolymeric matrix technology as an innovation. The second part allows the comparison of geopolymers with conventional matrices in terms of manufacturing technologies and mechanical performance. In its third part, this chapter contains the theoretical background required to understand the thermal and tribological parameters controlled in the experimental part of this work, allowing the interpretation of the data gathered and the results obtained.

As an introduction, OCMCs is a material family originally developed to manufacture high-temperature components using oxide ceramics, a practice not viable in their monolithic form due to their brittle behavior. A problem overcome with the introduction of a reinforcing phase of the same nature to improve their mechanical properties, however the initial approach still resulted in low σUTS due to the

high diffusion rates of oxides [38]. This led to the introduction of a third phase, which was initially conceived as a fiber coating whose aim was only to prevent the chemical interactions between fibers and matrix, as well as the physical contact between fibers, which is also detrimental for their performance. However the coating proved to be critical for the mechanical behavior of the material and received the name of “interphase” [38].

(18)

– 10 –

• Improvement of mechanical performance: which is normally evaluated in terms of i) tensile/flexural strengths ii) elastic moduli, iii) fracture toughness, iv) interfacial frictional stresses (thermal and radial), and v) creep behavior [39].

• Thermal behavior: i) thermal conductivity and emissivity, ii) thermal stability during long exposures to high-temperatures and iii) thermal shock resistance [39].

• Practical feasibility: i) raw material availability, ii) complexity of manufacturing process, iii) cost of the manufacturing equipment.

Toughness and pseudo-plasticity were achieved with the introduction of the interphase, which forces cracks to follow constantly changing planes and directions during growth, avoiding rapid and uncontrolled failure. This phenomenon known as “crack deflection” causes different failure mechanisms such as bridging and pull-out which allow OCMCs to exhibit non-brittle failure (Figure 2.1 left) and will be further addressed in section 2.5.1 while discussing the mechanical behavior of OCMCs. The energy required for failure is also related to the mechanical properties of each component, and the size, distribution, shape, and number of defects produced during manufacturing and handling.

In the following chapters, the characteristics and influence of each component is addressed as well as their relevant properties and performances, first separately in chapters 2.1 (matrix), 2.2 (fibers and fabric), 2.3 (interphases) and 2.4 (coatings); and then when acting as a whole (chapter 2.5).

(19)

– 11 –

2.1. OCMC Matrices – Oxide Matrices

The matrix has two main contributions to the performance of a CMC: i) it transfers the load to the reinforcement and ii) protects it from the atmosphere. Matrices can be divided depending on their chemical composition or morphology. The latter classification divides them in dense, porous and hybrid matrices.

Regarding the chemical composition, from the 24 existing high melting temperature oxides (> 1700 °C), mullite and alumina were found to be the most viable due to their low thermal conductivity, relatively low thermal expansion coefficient, moderate fracture strength and chemical stability [39]. However the low fracture toughness, sinterability and flexural strength of mullite and alumina were found to limit the optimization of the composite, which led to the study of the effect of small additions of other ceramics such as 3 mol% Yttria Stabilized Zirconia (3YSZ, normally in the range of ~15 %) [28].

Table 2.1 is a summary of the properties of those matrices which could be compared with the aluminosilicate matrix used in the OCMC studied in this work, however the information available for geopolymeric matrices in the literature is limited.

Table 2.1.- Room temperature (RT) properties of relevant candidates for OCMC matrices [14], [38], [40].

Oxide E [GPa] σUTS [MPa] Poisson ratio ν KIC [MPa*m1/2_] Fracture Energy Γ [Jm-2_] Lattice Energy JL [kJ/cm3_] Al2O3 410 600 0.23 2 - 6 44.7 624 ZrO2 140 < 400 0.32 3.6 - 555 3YSZ 205 45 [10] - - 93 - Al6Si2O8 145 - - - 39 191

Initially, dense matrix CMCs were developed resembling as much as possible to monolithic ceramics both in properties and manufacturing techniques. It was soon understood that the absence of a crack deflecting mechanism between fibers and matrices resulted in the same brittle behavior exhibited by their monolithic counterparts. This brittleness led to the development of porous matrices which was favored by the existence of a wide range of fibers but a lack of available interphases [41], this matrix design allows crack deflection without the need of optimized interphases, simplifying the manufacturing process [42].

(20)

– 12 –

strength and vi) inferior wear resistance [41] vii) propensity to sintering embrittlement after long exposures to high temperatures [4].

Porous ceramic matrices normally contain between 35 - 50 % porosity only in the matrix (without considering the reinforcement) and 25 - 40 % in total [4]. They are normally composed of two dissimilar phases: i) a continuous 3D network which dictates the long term stability of the matrix against sintering, and ii) a less refractory ceramic or glass binder which controls the mechanical integrity of the matrix [41].

The microstructure of a common porous matrix composite is shown in Figure 2.2, where the pores constitute a considerable volume of the matrix. The pore fraction can be measured with different methods such as open porosity and mercury intrusion, more information regarding these techniques can be found elsewhere [6], [8], [11], [13], [24].

Figure 2.2.- Typical microstructure of a porous matrix oxide-oxide CMC [4].

The internal voids of porous matrices are normally classified as: shrinkage cracks, microcracks and pores (which can be further sub-divided into nano- (<200 nm) and micropores (>200 nm) [24]), however the relative amount of each defect type is rarely known, and affects the mechanical performance of the composite differently.

The chemical composition is another characteristic that defines the behavior of the matrix. Aluminosilicate matrices are normally constituted of alumina and held by a silica binder which may be a highly porous filler or a continuous silica film. Aluminosilicate matrices are designed to improve the long-term thermal stability of the composite by inhibiting shrinkage through the slow sintering of mullite which is strengthened by the quick sintering of fine alumina particles. Examples of commercial versions of this design are the COI-610/AS and COI-720/AS produced by ATK-COIC.

All-alumina matrices are the most researched items and include designs such as the: i) COI-720/A produced with colloidal alumina mixing simultaneously two different particle diameters; and ii) WHIPOX (Wound Highly Porous Oxide Matrix), a composite which exhibits porosities between 60 - 80 % [42] after being sintered at high temperatures ~ 1300 °C which damages the reinforcing fibres (N610 or N720) [4].

(21)

– 13 –

2.1.1. Geopolymeric Matrices1

The matrix of the OCMC studied in this work is a geopolymer (or “inorganic polymer”), a term that refers to solid ceramic-like X-ray amorphous polymeric aluminosilicate materials produced through the reaction of a geological mineral with an alkali silicate solution, which present good thermal stability even over 1000 °C [43]. Their manufacturing process involves low processing temperatures similar to those used for thermosetting resins [18] which is an interesting feature considering that current OCMC manufacturing methods involve high-temperature steps that damage the reinforcing fibers. By avoiding fiber damage, geopolymer matrices promise composites with higher mechanical performance and more reliable behavior.

Geopolymer matrix composites are normally reinforced with particles, platelets or long fibers, which include carbon (reported to provide tensile strengths exceeding 500 MPa), basalt and glass fibers, as well as SiC, Al2O3, Mullite and Boron fibers. Another benefit of the low processing temperature is that

it allows the reinforcement of ceramic-like matrices with organic fibers such as PVA, aramid and cellulose based (flax) as well as protein-based (wool) fibers, a feature not possible before.

The synthesis of a geopolymeric matrix requires i) a reactive solid source of sufficiently reactive silica and alumina, and ii) an alkaline solution containing alkali metal hydroxides, silicates and sometimes aluminates. This results in a wide variety of possible raw materials and therefore of final mechanical properties, more details regarding the manufacturing process of OCMCs with this matrix is provided in section 2.5.1.

The mechanical performance of the geopolymer matrix depends on the nature of the cations (typically alkali metal ions) present in the microstructure, since they balance the negative charge associated with tetrahedral aluminate units that combine with silicate units to form a three-dimensional structure. In the reaction process first proposed by Davidovits [43], the matrix is formed by the condensation of tetrahedrally coordinated aluminosilicate units called polysialates (sialate being an abbreviation of silicon-oxo-aluminate [44]) consisting of tetrahedral SiO2 and AlO2 units linked by shared oxygen

atoms, with alkali cations (either Na+_{or K}+_{) that compensate the charge generated by the presence}

of the tetrahedral aluminate units.

The fundamental polysialate oligomer units are: i) polysialate (PS, Si:Al = 1), ii) polysialate siloxo (PSS, Si:Al = 2) and iii) polysialate disiloxo (PSDS, Si:Al = 3) and their structure is shown in Figure 2.3, however the sialate nomenclature cannot describe the compositional range of geopolymers, since compounds with Si:Al ratios over 3 have been reported.

A) B) C)

Figure 2.3.-Polysialate units constituting a geopolymer structure A) Polysialate, B) Polysialate siloxo and C) Polysialate disiloxo [18].

Solid-state magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectroscopy has been used to provide evidence of tetrahedral silicate and aluminate units (Figure 2.4), as well as information regarding the location and hydration states of the monovalent charge-balancing cations.

(22)

– 14 –

Figure 2.4.- Geopolymer structure based on Nuclear Magnetic Resonance (NMR) data [18]. Besides geopolymeric matrices there are other developments in matrix technology for OCMCs. New studies (performed in non-oxide but transferable to all-oxide systems) suggest that the addition of non-oxide particles (such as SiC [45] and B4C,) might improve the failure management of CMCs.

These systems are called self-healing and introduce compounds whose objective is to react with oxygen as it diffuses through a crack. The compounds form a liquid oxide (SiO2 and B2O3 respectively)

with higher volume than the original particle, that fills the crack and crystallizes on its walls. The crystallization process bonds the crack walls due to its highly exothermic nature. This results in autonomic recovery of strength for the composite. In the conventional studied cases the self-healing mechanism prevents atmospheric oxygen from interacting with the SiC or C fibers [46] however its effect as a crack blunting mechanism might be interesting for oxide systems.

In terms of structure and morphology, new nanoceramic oxide matrices characterized by dimensions less than 100 nm in either grain size or fiber diameter are being produced from Al2O3 and

ZrO2.Nanoceramic matrix composites exhibit better physical and mechanical properties than

conventional composites (such as creep resistance [47]). The better performance is attributed to i) the proportion of atoms at grain boundaries and surfaces, and to ii) the change in behavior of physical phenomena (generation of dislocations, ferromagnetism and quantum confinement) that occurs under a characteristic scale length [48].

(23)

– 15 –

2.2. OCMC Reinforcements – Oxide Fibers and Fabric

The reinforcement is a phase with at least one of its dimensions smaller than 500 μm, and sometimes of the order of a micron which enhances the mechanical properties of the matrix, being conventionally hard, strong and stiff although there are exceptions [49].

Almost any material can be used as a reinforcement, however only those capable of withstanding high-temperatures are viable for CMC applications. The first characteristic used to classify a reinforcement is its morphology, the nomenclature is shared with the polymer and metal matrix fields and defines:

• “Whiskers” as cylindrically shaped, short (typically 0.6 μm in diameter and 10 – 80 μm in length [23]) single crystalline bodies with a diameter of less than one micron [50], [51] • “Particles” as small (10-100 μm) spherical, cubical or platelet shaped bodies (or any regular

or irregular geometry) with approximately equal dimensions in all directions, either completely solid or hollow

• “Fibers” as cylindrical bodies whose long axis is several orders of magnitude greater than the cross-sectional axis, characterized by their “aspect ratio” which is the ratio between the former and latter dimensions [51].

In terms of chemical composition, fibers may be: i) organic polymeric with degradation temperatures inferior to 500 °C; ii) carbon fibers which degrade over 450 °C in oxidizing atmospheres (2800 °C in non-oxidizing); iii) glass fibers with melting/softening points under 700 °C; and iv) polycrystalline or v) amorphous ceramic fibers with different degradation temperatures. These fibers summarize all non-metallic fibers with the exception of those manufactured via solidification of glass melts based on silicate systems (glass fibers) or minerals such as basalt (mineral fibers) [52], further in this project only oxide fibers will be discussed but information regarding non-oxide fiber properties can be found elsewhere [49], [52]–[55].

The production of ceramic fibers started in the 1960s with B and SiC fibers, however the first oxide (alumina) fibers appeared in the 70s and were rapidly adopted as reinforcements for metals. Oxide fibers may consist of a polycrystalline or single crystal structures, and may be largely classified in: i) single-phase alpha alumina fibers (Fiber FP, Almax and N610), ii) Zirconia reinforced alumina fibers (PRD-166), iii) Alumina-silica fibers (Saffil, Altex, N312, N440, N480 and N720) and iv) other large-diameter fibers [56]. For more details regarding the physical and mechanical properties of available oxide fibers please refer to references [6], [27], [32], [39], [52], [56]–[68].

Polycrystalline long fibers are normally produced through either a direct or an indirect process. Direct processes spin inorganic precursors (salt solutions, sols or precursor melts, also called dopes) into “green fibers” and apply a high-temperature heat treatment to eliminate the non-ceramic components. On the other hand, indirect processes apply pre-ceramic precursors on non-ceramic fibers either by soaking or deposition, both followed by pyrolysis, more information regarding each process can be found elsewhere [52].

The mechanical properties of a fiber are determined by its features at different scales: i) molecular: the chemical structure of the fiber determines its thermal and chemical stability, the ionic and covalent bonds show the highest energy content mostly when oriented three-dimensionally; ii) supramolecular: fiber properties depend on its molecular orientation and crystallinity; and iii) macroscopic: their properties are affected by their cross-section, diameter homogeneity, porosities, structural flaws and surface properties (roughness and surface energy) which affect their adhesion to the matrix and wetting behavior [52].

(24)

– 16 –

depends on the mechanical and physical properties of the matrix and the fiber [69] (more information regarding this parameter and other characteristics of the fiber/matrix interface can be found elsewhere [69][51]), although according to current standards used in material selection software, short fibers are considered under 5 mm long [23].

When considering which reinforcement to use, the selection depends on the OCMC manufacturing process, short fibers and particles are easily added to matrix precursors which are then shaped (by extrusion, molding, or casting) and sintered, providing homogeneous random fiber/particle distribution and therefore quasi-isotropic mechanical properties.

However, whiskers and particles do not provide some of the crack deflection mechanisms that long fibers do, such as crack bridging and pull-out, which allow long fiber reinforced composite to exhibit higher impact strength, tensile modulus and tensile strength, as well as lower elongation at failure [70]. Another important aspect to consider is that whisker/particle reinforced composites normally do not benefit from the improvements offered by available interphases, only recent studies are considering this approach [71].

The OCMC studied in this work is reinforced with N610, a single-phase α-alumina fiber, therefore this design will be further explored.

2.2.1. Single Phase Alpha-Alumina Fibers (N610)

Most commercially available fibers are based on α-alumina, the most stable allotropic phase of this compound, which is formed after heating over 1000 °C (Figure 2.5). Sometimes small additions of silica are made to control grain growth, but this lowers the E and creep strength, therefore it is not common.

Figure 2.5.-Structural transformations of alumina [56].

The production of α-alumina fibers with controlled grain growth and porosity is obtained by dry spinning and pyrolyzing a slurry containing α-alumina particles with controlled granulometry in an aqueous solution of aluminum salts.

(25)

– 17 –

Mitsui Mining and 3M followed in the early 1990s with the development of Almax and Nextel 610 respectively.

Nextel 610 (N610) is a polycrystalline fiber (Figure 2.6) with a grain size of 0.1 μm, produced by a sol-gel technique that presents the highest strength of its design. It contains a small amount of silica (0.35 wt%) for grain growth control, and α-Fe2O3 (0.67 wt%) used in production to nucleate the

α-alumina grains. It offers twice the σUTS of FP but has lower resistance to creep, exhibiting higher creep

strain rates at 900 °C due to its finer grain size and grain boundary chemistry. Its failure mechanism occurs through the coalescence of cavities into large cracks, whose growth occurs initially in an intergranular manner, changing quickly to an inter and transgranular combination, and returning to intergranular at catastrophic failure [66].

Figure 2.6.- Scanning electron microscopy of the cross sectional section of N610 [56]. N610 loses strength after being exposed to high-temperatures without a matrix [65] and the lower performance is attributed to grain growth and fiber-fiber sintering. To have a reference of this grain growth, Table 2.2 provides the grain size of this fiber after being exposed for 1 h at 900 °C and 1400 °C showing an increase of ~ 430 nm [32], [65].

The effect of different matrices on the high-temperature grain growth (1300 °C for 5 h) of N610 fibers has also been studied in WHIPOX (an alumina matrix) and Keramikblech type FW12 (an alumina-3YSZ matrix) [72]. It was found that when the temperature remained under 1200 °C grain size was only slightly increased, but over 1300 °C the effects become critical depending on exposure time and manufacturing routes. WHIPOX fibers exhibited smaller grains in the fiber core.

During manufacturing some fibers go through a process called “sizing”, which consists of the application of a polyvinyl alcohol (PVA) coating to protect them from interaction with the atmosphere until used, this sizing must be removed before the composite manufacturing. N610 fibers have also been used to test the effects of desizing on tensile strength using different atmospheres[67], concluding that temperature is more critical than gas flow (from 100 to 200 standard cubic centimeter per minute).

General information can be found elsewhere regarding other designs [14], [52], [53], [56], and particular information about specific fibers is also available such as for FP [73]; Nitivy ALF [52]; Saffil [63]; YAG and magnesia/chromia doped alumina fibers [74]; Spinel [59]; Zircon [4]; YAG-Al2O3

(26)

– 18 –

Table 2.2.- Grain growth of oxide fibers aged at different temperatures. Values reported in nanometers.

Source and comments <1000 °C 1100 °C 1200 °C 1300 °C 1400 °C [32] 1 h ageing N610 fiber, only alumina grains N720 fiber Alumina grains Mullite grains 900°C 90 40 280 100 40 280 110 40 280 170 80 290 520 150 310 [65] 1 h ageing N610 900°C 89 96 - 175 - [66] 1 min exposure N610 - 69±7 69±8 75±8 91±11 [72] 5 h exposure WHIPOX™ FW12™ - - - 2 146 155 - [68] N610 N720 1000 °C 70 45 130 ~60 >160 ~135 - _- 2.2.2. Fabric Properties

Although fiber properties are important, the study of a single fiber does not represent the behavior of a composite since commonly fibers are arranged in a certain configuration to obtain a usable component. There are three relevant fabric related properties that must be considered when studying CMCs: i) “Denier” (den) which is an old and common concept used to give an idea of the mass of the fiber per unit length, taken from the textile industry where one gram of silk fiber was 9 000 meters long. This measure unit has been slowly changed for the “dtex” defined in the International System of Units as the mass in grams of a fiber 10 000 meters long; ii) “Weave”: which refers to the patterns in which fibers are braided such as 2x2 twill (TW), 4 harness-satin (HS), 5HS, 8HS and semi-unidirectional (SU), each with different morphologies; and iii) “Fiber volume” (Vf) which can be

calculated knowing the weight (Wf) and density (ρf) of the fibers as well as the total volume of the

composite (Vc) using Eq. 2.1 [13]:

𝑉𝑓=

𝑊𝑓

𝜌𝑓∗ 𝑉𝑐

∗ 100 Eq. 2.1

Other important concepts in fabric weaves are “warp” and “fill” (also called “weft” or “woof”) which refer to the fiber orientation in the loom (weaving machine):

• “warps” are those fabrics positioned parallel to the manufacturing direction of the fabric (which also coincide to the longest dimension of the finished fabric in Figure 2.7 A, being the vertical fibers in (Figure 2.7 B).

(27)

– 19 –

• “fills” which are perpendicular (horizontal in Figure 2.7 B), from these it must be noted that a simple thread of the fill crossing the warp is called “pick” [28].

A) B) C)

Figure 2.7.- Fabric characteristics A) Fill and warp fiber directions [68], B) 8 harness satin (HS) [76] and C) 2x2 twill [77] fiber/bundle weaves.

Due to the bending suffered by the fiber during weaving, the warp is generally more damaged than the weft [78] and the fabrics have different mechanical behavior depending on the loading direction, for example the EF-11 fabric (N720 fibers) has a tensile strength of 32 kg/cm in the warp and 30 kg/cm in fill directions, lower than the DF-11 fabric with 46 kg/cm in both directions [13].

Current practices use fibers in an 8HS weave which present good draping capability for curved components [13], however it has been proven that different weaves have no effect in fiber bundle mechanical tests [28]. The composite studied in this work uses a 2x2 twill weave pattern (Figure 2.7 C). It has been shown that tight fabrics and rovings can be easily processed with current techniques and have similar mechanical properties as standard fabrics [28].

Higher denier fabrics have lower apparent strength (i.e. maximum force withstood in a tensile test divided by the initial area of the bundle) but the strength of the composites manufactured with similar fibers but different fabric patterns were only different in through thickness flexural tests, where the strength of the material depends on the undulations of the outer layer fabrics [28].

(28)

– 20 –

Table 2.3.- Fill and warp N610 fiber tensile properties [28].

Denier Apparent Strength [MPa]

Characteristic

Strength [MPa] Weibull modulus

1500 (10k Fill) 3 _{1574 ± 147} ₃₃₃₁ _4.1 1500 (20k Fill) 1725 ± 196 3190 4.8 1500-8HS Fill 1578 ± 159 3153 4.7 1500-8HS Warp 1990 ± 171 3427 3.4 3000-5HS Fill 1300 ± 60 2905 4.0 3000-5HS Warp 1455 ± 128 3049 4.3 4500-Twill Fill 1280 ± 129 2543 3.4 4500-Twill Warp 1444 ± 262 3333 3.3 4500-4HS Fill 1240 ± 178 3216 3.2 4500-4HS Warp 1188 ± 8 3193 2.5 4500-5HS Fill 1437 ± 131 3322 3.8 4500-5HS Warp 1368 ± 88 3247 3.1 4500-8HS Fill 1130 ± 25 2699 2.9 4500-8HS Warp 1432 ± 36 3136 3.4 10 000 (10k Warp) 1070 ± 104 2373 2.5 20 000 (20k Warp) 1003 ± 69 2152 2.6 2.3. OCMC Interphases

Interphases were initially conceived as fiber coatings whose purpose was to protect fibers from environmental and thermal degradation when a crack generated in the matrix. A secondary objective being to prevent fiber-to-fiber contact which also affects the fiber performance. However, it was soon understood that their influence on the mechanical properties of the composite went further, being directly related to its toughness and pseudo-plasticity. Figure 2.8 shows the difference in fracture behavior between a brittle composite without an interphase (A) and a tough composite with a monazite interphase (B).

Although the benefits of an interphase have been well documented, current commercial OCMCs do not include them in their designs due to the additional manufacturing steps required for their implementation, which elevates costs and production complexity without a considerable improvement in mechanical properties when compared to those offered by porous matrix OCMCs. However, to fully optimize the mechanical properties dominated by the matrix such as through thickness and interlaminar strengths, the matrix must be dense and act in collaboration with an optimized interphase.

(29)

– 21 –

A) B)

Figure 2.8.- Fracture surfaces of a A) N610/A with no interphase and B) N610/A with a monazite interphase where deflection within the coating can be appreciated [4].

The characteristics required from a good interphase include crack bridging, crack blunting and crack deflection mechanisms as well as chemical and thermal stability in oxidizing/corrosive (and slightly reducing) environments. Interphases must also be chemically and morphologically compatible with both matrix and fiber while providing debonding and frictional sliding which favor fiber pull-out performance [79]. Different interphase philosophies exist:

• Low toughness interphases. • Easy cleavage structures. • Fugitive interphases. • Fugitive/oxide interphases. • Porous interphases. • Refractory metal coatings.

The optimal interphase characteristics such as interphase thickness, frictional stresses and de-bond length, depend on the combination of composite constituents as well as on the type of interphase designed. Their characterization considers the uniformity and continuity of the interphase, its thickness, porosity and chemical composition (determined through x-ray diffraction or TEM) [4]. The composite studied in this work does not include an interphase in its design, therefore this topic will not be further addressed, however Monazite (LaPO4) is nowadays a very promising candidate for

this application, more information can be found elsewhere for this interphase [4], [22], [41], [80]–[82] and others [4], [29], [41], [67], [79], [83]–[87].

2.4. OCMC Insulating Coatings

Insulating coatings are among the most promising technologies for performance improvement of OCMCs, defined as a layers that protect the composite from aggressive environments. They are normally divided in Environmental Barrier Coatings (EBC) and Thermal Barrier Coatings (TBC). Coating application has been proven effective in the protection of all-alumina composites in the combustion liners of Centaur 50S gas turbines (Friable Graded Insulation FGI [36]), and burner rig tests (YAlO3 [88]). Studies have suggested that YAG FGIs may increase Temperature Rotor Inlet

(30)

– 22 –

Although these coatings have been developed, no tests have been performed on samples to quantify their impact in the mechanical properties of OCMCs after or during high-temperature exposure, considering that the studied composite does not possess an insulating coating, more information about this topic will not be provided, however it can be found elsewhere [20], [89]

2.5. OCMC – Overview

The mechanical performance of an all-oxide ceramic matrix composite (OCMC) will depend on the characteristics of its components, in particular the volume fraction of the reinforcement, as well as its shape and size, although the orientation of the matrix grains also plays an important role. These parameters will be strongly influenced by the manufacturing process and the composite design.

2.5.1. Conventional OCMC Manufacturing Process

Conventional OCMC manufacturing in general is a challenging field due to the similarity in chemical composition between fibers and matrices, which results in microstructural and morphological changes occurring at similar temperatures and pressures for both components [4]. As it has been already discussed, this normally forces to find a balance between fiber damage and matrix consolidation due to the high process temperatures involved.

In general, any OCMC manufacturing process (summarized in Figure 2.9) considers the synthesis of the reinforcing fibers as the first step, which are then woven into fabric and coated with the interphase (if the composite design requires it) to be then infiltrated with the matrix precursor to create a pre-impregnated fabric called prepreg. Manufacturing then proceeds with the stacking of several prepregs to achieve the shape and dimensions of the final component. The process finishes with several firing steps that consolidate the composite.

(31)

– 23 –

2.5.1.1. Infiltration

After resin preparation, the manufacturing process may follow one of three routes: i) pultrusion, ii) preform infiltration, which was commonly assisted by vacuum and pressure in the early 2000’s [91]; or iii) prepregs [4], [21], being the latter the most commonly used nowadays due to its simplicity. Infiltration is the insertion of the matrix precursor into a fabric body to obtain a “green body”, which is a component close in dimensions to the finished part, but with lower mechanical properties. Full infiltration of the inter-fiber space with the precursor is unachievable with current technology, forcing designers to control the void presence in the CMC to ensure small homogeneously distributed porosities instead of their agglomeration which lead to cracking [4].

In the prepreg route fiber layers are impregnated with the resin (binder) using squeegees and rollers to ensure complete wetting of the fibers. The prepregs are then stacked to form the final component geometry. These prepregs can be stored without any loss of strength, although their elastic modulus is negatively affected [18].

During stacking it is recommended to remove excess water to improve the adhesion between layers, and to perform short drying steps between every couple of layers. The objective is to avoid drying the fully stacked laminate since this could have a negative influence in the final mechanical properties of the composite, mostly when complex 3D architectures are manufactured due to shrinking.

The control of the solution/slurry used to fabricate the CMC is critical since it affects its homogeneity and final product performance. In the conventional route, the oxide powder is dispersed in a carrier such as water, along with other additives (binders, dispersants, plasticizers) and its particle size is limited by the fiber diameter [4]. The dispersed powder is then infiltrated into the composite through one of the following techniques:

Prepreg / lamination [4], [21], [92], which is a common method used due its advantages, being simple,

inexpensive and fast and allowing the production of near-net shape components. This method has the downside of providing the lowest fiber volume content of all infiltration processes resulting in the lowest strengths. In the prepreg process a fabric is submerged in the slurry to obtain prepregs, which are then stacked onto a tooling or mold and pre-consolidated using vacuum bagging and warm temperatures (< 150 °C to eliminate water and remaining solvents). This manufacturing technique is normally followed by pressure-less sintering which will be further discussed in the following chapter.

Pressure infiltration / mold pressing / flexible injection [4], [21], [90], [92]: is a method used for

high-quality components that offers high and well-defined fiber volume contents and allows near-net shape manufacturing, however it is limited to parts with rather low complexity. In pressure infiltration, layers of fabric are stacked into a mold in the desired orientation, then slurry is poured into the mold and a combination of pressure and sometimes vacuum is applied forcing the fabric to act as a filter retaining the matrix particles from the slurry [26]. In the filtering stage the fabric must be constrained to prevent its movement during the application of pressure, which may cause regions with a thicker matrix between layers and therefore lower fiber volume fractions. The flexible injection method uses a flexible membrane to apply pressure to promote transverse impregnation.

Vibration assisted vacuum infiltration is similar to pressure infiltration, except that no pressure is

(32)

– 24 –

Freeze forming / freeze drying / freeze gelation is a method in which ceramic powder is dispersed in

a sublimable solvent (water or camphene), along with binders and dispersants to form a slurry which is infiltrated into the fiber tow. The tows are then stacked and warm-molded to form the green body, which is then cooled under the freezing point of the solvent so that the solidification front pushes the ceramic particles. Then the composite is either freeze-dried under vacuum (water), or sublimated in air at RT (camphene), preventing the build-up of capillary pressure associated with the drying phenomenon. The result is a porous matrix where the pores replicate the frozen solvent structure. The green body is then sintered, retaining the pore structure. This technique has been investigated for the formation of N610/SiO2-Mullite composites with interesting results [4].

Electrophoretic Deposition EPD infiltrates colloidal sol into a conductive coated fabric using an electric

field, the process is fast in the beginning but slows down as the thickness of the infiltrated layer increases. This process is easily applicable to a single fabric but thicker parts are not viable at least with current technology, this limitation restricts it to the manufacture of prepregs [4], [92].

2.5.1.2. Consolidation and Post-Processing

Consolidation is the step through which the matrix acquires its final mechanical properties, this process conventionally involves a high-temperature step (over 1000 °C) whose purpose is to eliminate residual non-ceramic components. In the conventional practice, warm- and hot-pressing are the favorite consolidation processes for OCMCs [91] although the colloidal route has also been used to produce porous matrices [94].

More information can be found in the literature regarding the most common conventional consolidation processes: i) Pressureless sintering (1000 - 1200 °C [4], [92]), ii) Hot pressing (HP) /

Hot Isostatic Pressing (HIP) / Field Assisted Sintering (FAST/SPS) (> 1000 °C) [4], [92] iii) Chemical Vapor Deposition (CVD) / Chemical Vapor Infiltration (CVI) [4], [92] iv) Directed Metal Oxidation (DIMOX) / Melt Infiltration / Reaction Bonding [4], [92] v) Microwave Sintering [95] vi) Strong Magnetic Field Alignment [96] and vii) Spark Plasma Sintering [27], [96], [97].

It is important to emphasize that these conventional manufacturing processes require consolidation temperatures normally exceeding 1000 °C, which, as was discussed in section 2.2.1, compromises the performance of the fibers.

(33)

– 25 –

A) B) C)

D) E) F)

Figure 2.10.- Effects of different cutting techniques on an N610/ 85 % Alumina 15 % 3YSZ porous matrix composite [10]:

Laser A) mesoscopic structure and re-melted B) fiber ends and C) matrix particles. Water jet D) mesoscopic structure E) fiber ends with delamination F) matrix part.

2.5.1.3. Environmental Impact

An important consideration to make when developing a new material or technology is the effect that it will have on the environment i.e., the impact of the i) raw material extraction, ii) waste generation and energy consumption during production and iii) disposal of end-of-life components.

An interesting tool for this is the Life Cycle Assessment (LCA) which considers the resources (mostly water and electricity) and emissions (greenhouse gases) involved in i) raw material extraction, ii) component design/production, iii) packaging/distribution, iv) use/maintenance and v) disposal; comparing them with the benefits of the technology’s implementation in the same terms (resources and emissions). Three different impact approaches can be taken: i) “cradle to grave”, which considers from the mining of raw materials up to disposal of the manufactured component; ii) “cradle to gate”, which measures the impact of extraction and manufacture finishing when the product is ready to be sold; and iii) “cradle to cradle”, which applies for components which can be almost completely recycled such as PET and aluminum bottles. The International Organization for Standardization (ISO) has released a series of requirements and guidelines for life cycle assessment (LCA) in the ISO 14044 standard, which identifies the following steps: i) Goal & Scope definition ii) Inventory analysis: data collection iii) Impact assessment and iv) Interpretation [98].

Impact assessment relates the production waste and emissions with one of the following categories: i) Global warming ii) Ozone layer depletion iii) Acidification iv) Eutrophication v) Photochemical oxidation formation, vi) Ecotoxicity (water, land or air) vii) Toxicity for humans, viii) Fossil depletion ix) Respiratory effects x) Ionizing radiation xi) Carcinogenic and xii) Non-carcinogenic effects xiii) Mineral extraction and ixx) non-renewable resources.

(34)

– 26 –

Raw Material Extraction and OCMC Manufacturing

Considering the OCMC manufacturing process as: i) raw material extraction, ii) milling, iii) precursor slurry preparation (for either fiber coating or matrix), iv) fabric infiltration, v) consolidation and vi) final machining; the impact of the process can be summarized in terms of electricity (required for milling, vacuum pumps, vibrating/pressing equipment, and electrical furnaces), water, ceramic scrap and CO2

emissions, the latter being fairly low since the furnaces used for consolidation are normally electric. Alumina can be reviewed to provide an example of the extraction impact of raw material. This oxide is normally extracted from bauxite through the Bayer process which produces around 0.05 ton of CO2

for every ton of aluminum manufactured [99], calculated from the direct emissions of stationary combustion equipment (CO2, CH4, and N2O) as proposed in [100].

The preparation of alumina powder for the next processing step (grain size adjustment and homogenization) requires an average water consumption of 20 L4_{(assuming a square meter, 12.7}

mm alumina tile), 60 % of which is used during milling, however water consumption in ceramic facilities is low being normally recirculated. The energy consumption is 32 kWh, from which 90 % is consumed by the spray drying and firing steps of the process, which produce the alumina granulates to be further prepared for fiber infiltration [101], [102].

An environmental handbook published in Germany [102] provides information and restrictions related to environmental impact for the manufacture of technical ceramics in the following categories:

• Air: samples taken must not exceed 5 mg/Nm3_{of fluorine, 10 mg/Nm}3 _{of total dust, 50 mg/Nm}3

of sulphur dioxide and 0.1 mg/Nm3 _{of Cd/Tl/Hg per element.}

• Dust: fine quartz dust (<5 μm) is restricted to maximum 0.15 mg/Nm3_{, while for non-specific}

dust is 4 mg/Nm3_.

• Noise: any equipment exceeding 85 dB must be installed out of permanent workplaces. The facility must not exceed 50-60 dB during the day and 35-45 dB at night. Residential areas shall be at least 500 m from the facility.

• Water: a sample mixed during 2 h shall not present more than 100 ml of filterable solids, 80 mg/l of chemical oxygen demand, 0.5 mg/l of lead and 0.07 mg/l of cadmium while the total suspended solids of a random sample shall not exceed 0.5 mg/l.

• Soils: most of the residues of ceramic manufacturing can be reused as raw material for production, there are almost no negative consequences for soil.

The material selection software GRANTA EduPack 2020 reports that in order to produce 1 kg of N720/AS from the alumina ore (mining, milling and manufacturing), between 31 700 and 35 000 MJ of energy are required, producing between 2370 – 2620 kg of CO2 and requiring between 180 – 199

liters of water [23]. The same source reports that a facility specialized in manufacturing of a N720/AS composite through molding uses for each kilogram of composite between 10 – 22 MJ of energy, and produces 1.5 – 1.65 kg of CO2 requiring 22.3 – 33.5 liters of water, these values considered for

manufacturing only [23], which implies that a company has registered water, gas and energy consumption, as well as emission generation per production area for several months to provide an accurate historical behavior [103].

4_{For this calculation, the raw material preparation process is considered similar to that required for}

(35)

– 27 –

Benefits of Technology Implementation

For the assessment of impact in society, economy and environment, an interesting framework proposed by the UN in 2016 (Agenda 2030), provides a series of goals whose aim is to achieve a better and sustainable life quality [104]. The agenda considers the following goals:

• No poverty. • Zero hunger.

• Good health and well-being . • Quality education.

• Gender equality.

• Clean water and sanitation. • Affordable and clean energy. • Decent work and economic growth. • Industry innovation and infrastructure.

• Reduced inequalities.

• Sustainable cities and communities. • Responsible consumption and

production. • Climate action. • Life below water. • Life on land.

• Peace justice and strong institutions. • Partnership for the goals.

The main environmental benefits of OCMC implementation are related to affordable and clean energy, development of sustainable cities and communities, and climate action through the reduction of emission levels in gas turbine plants for power generation, and the improvement of commercial aircraft propulsion. In both applications, the availability of materials capable of withstanding high-temperatures allows engines to work at optimum fuel-burn efficiency reducing emissions and allowing a better performance, as has been demonstrated by the experiment performed on a Centaur 50S gas turbine engine, which reduced NOx levels from 25 to 15 ppmv and CO from 50 to 10 ppmv [35].

Waste Disposal and Recycling

There are six options for handling materials at the end of the product’s life: i) Reuse, when the product’s life can be extended, ii) Re-engineer, which incorporates the waste into a new product, iii) Recycle, when it can be reprocessed into a new piece of the same product grade, iv) Downcycle, when it is reprocessed into a lower grade material v) Combustion, when the calorific content of the waste can be recovered and vi) Landfill, when it is disposed. The latter being the less accepted [23]. OCMCs do not present any complications related to radiation, toxicity flammability or cancerogenic nature as compared to waste from other sources such as nuclear, and they do not find the same recycling challenges as polymer and metal matrix composites since the chemical composition and mechanical behavior of both reinforcement and matrix is very similar or at least compatible. However, these materials must be downcycled through milling, and reintroduced to the manufacturing flow as an aggregate or filler replacement, which requires a considerable amount of energy and specialized equipment, hindering their rapid reusability. For this reasons only a 0.1 % of this material family is normally recycled [23]. If the energy requirement is overcome, the end-of-life parts can be used to manufacture other ceramic based components, mostly refractory due to the nature of the OCMCs components.

All-Oxide Ceramic Matrix Composites Thermal Stability during Tribological Interactions with Superalloys