Project Topics
Get Complete Project Material File(s) Now! »
SiC, SiC fibers and SiCf/SiCm composites
SiC physico-chemical properties are interesting for nuclear applications. How-ever, its use as structural material is limited by its brittle nature, typical of ceramic materials. SiC-based composites, SiCf/SiCm, allow to overcome bulk SiC mechani-cal limitations while retaining its properties, hence placing themselves as one of the few realistic candidates for high temperature and nuclear structural applications. The achievement of high temperature resistant composites is indebted to the de-velopment of near-stoichiometric and highly crystalline SiC fibers. In this chapter, SiC fundamental properties as well as the properties of the successive generations of SiC fibers and their influence on the behavior of the composites are described.
SiC
Silicon carbide is a non-oxide covalent ceramic compound with chemical formula SiC. Figure 2.1 shows an assessed equilibrium phase diagram of the Si-C system. The diagram shows a peritectic transformation at 2545 ± 40 ◦C and an eutectic transfor-mation at 1404 ± 5 ◦C that involve the intermediate compound SiC.
The process to obtain SiC followed by Acheson was rather simple. By combining silica sand and petroleum coke into a high temperature electric furnace, he was able to obtain the reaction SiO2 + 3C → SiC + 2CO.24 Since then, more refined manu-facturing methods such as sintering, direct conversion, gas phase reaction and polymer pyrolysis, appeared to cover the incrementing needs of SiC with high crystallinity levels for different purposes.26 For instance, chemical vapor deposition (CVD) is one of the most familiar gas phase reaction methods for the synthesis of high crystallinity, stoi-chiometric and high purity β-SiC whereas polymer pyrolysis is often used in production of continuous SiC fibers. The CVD process can produce SiC in solid form by epitaxial growth from a gas phase reactant at relatively low temperatures, between 900 ◦C and 1100 ◦C, without the use of sintering aids. Either methyltrichlorosilane (CH3SiCl3) or ethyltriclorosilane (C2H5SiCl3) are commonly used as gas reactants and hydrogen as carrier gas. Depending on the reactant.
The methane produced by using ethyltriclorosilane as reactant easily decomposes in free C with hydrogen generation producing undesirable C-rich phases or C layers in CVD-SiC.
The basic structural unit of this compound is a C atom surrounded by four Si atoms forming a covalent-bonded tetrahedron. One of the characteristics of SiC is its remarkable number of different stable stoichiometric solid crystalline phases. Each of these phases, known as polytypes, can be achieved by altering the stacking sequence of the basic SiC tetrahedron. Figure 2.2 shows the fundamental tetrahedron as well as different stacking sequences. Among the more than 200 reported polytypes,27 the most common are those which present cubic, hexagonal and rhombohedral structures with repetition sequences of 3, 4, 6 and 15 Si-C pairs.28 Following the Ramsdell notation,29 SiC polytypes are denoted as nX-SiC , where n designs the repetition period of the stacking sequence and X the crystallographic structure, these structures are commonly known as 3C-, 4H-, 6H-, 15R-SiC. The 3C-SiC polytype, also known as β-SiC, results to be the only stacking sequence that has cubic symmetry. The remaining polytypes, which show hexagonal or rhombohedral symmetry, are known as α-SiC. The occurrence and stability of different SiC polytypes primarily depend on the temperature. For instance, the 3C-SiC polytype is a metastable phase and will transform into α-SiC at very high temperatures, above 1900–2000 ◦C.30 In addition, the presence of impurities and the deviation from strict Si:C stoichimetry play a role in polytype stability.
Each of the polytypes exhibits unique physico-chemical properties which make SiC an attractive material for industrial applications ranging from heat engines and elec-tronic devices to nuclear systems. Table 2.1 shows some characteristics of the most common SiC polytypes. Despite all SiC polytypes have almost the same Si–C bond length of 1.89 ˚A and a bilayer height along the c-axis of 2.52 ˚A, each of them have different lattice constants as well as different electronic band structures—hence the optical and electronic properties—due to their different crystalline structures. All the polytypes have indirect band structures with band gaps increasing monotonically with the polytype hexagonality (h). For instance, typical band gap values for 3C (h=0), 6H (h=0.33) and 4H (h=0.5) yield 2.4, 3 and 3.2 eV respectively.
SIC, SIC FIBERS AND SICF/SICM COMPOSITES
On the other hand, different polytypes have similar elastic modulus and fracture toughness due to the similarity of the Si–C bonds. Regarding the elastic modulus at RT, ERT, SiC single crystals and CVD-SiC exhibit high values ranging from 390 to 690 GPa.30 For the latter, grain size is considered to have a negligible effect on ERT whereas it decreases with increasing porosity or impurity concentration.12 In addition, the elastic modulus slightly decreases with temperature with typical values of 0.9×ERT at 2000 ◦C. Despite the small impact of the temperature on SiC mechanical properties, it is remarkable the low fracture toughness of SiC, which typical values are between 2.4 and 5.1 MPa m1/2 and are at least ten times lower than for steel alloys.12 Such a brittle behavior makes bulk SiC not suitable for its use as structural material.
Thermal creep is also a relevant mechanical property for structural nuclear appli-cations. Primary and steady state creep strains have been reported in the literature for CVD-SiC. The former occurs immediately upon loading and tends to saturate with time. In severe loading conditions, primary creep strain in CVD-SiC can reach as high as ∼1% at 1923 K depending on the quality of the material. Steady state or secondary creep rates for CVD-SiC have been only measured above ∼1650 K as they are typically too small to be measured under this temperature.12 CVD-SiC creep rates for tem-perature ranging from 1655 to 1743 K are reported to be proportional to the applied flexural strength with an activation energy of 6.6±0.9 eV.12 Also, the underlying creep mechanisms are related to grain boundary (GB) diffusion of C atoms and dislocation glide below 1923 K. Over this temperature, creep seems controlled by a climb-controlled glide mechanism.
SIC, SIC FIBERS AND SICF/SICM COMPOSITES
SiC based composites
Reinforcing SiC with continuous SiC fibers allows overcoming the brittle failure of bulk SiC inadequate for structural applications.33 Indeed, the composite obtained is lightweight, damage tolerant, tough, and strong and exhibit a much greater resistance to high temperature environments than metals or other conventional engineering ma-terials.10 The chemical vapor infiltration (CVI) method can produce high crystalline SiCf/SiCm composites with excellent properties. The composite is made by the den-sification of a coated SiC fiber preform by infiltration of gaseous reactants, commonly methyltrichlorosilane (CH3SiCl3) and hydrogen.33 The resulting high purity SiC ma-trix provides good irradiation resistance to the composite, necessary for nuclear appli-cations.11 Despite the quality of the achieved SiC matrix, CVI produced composites have a residual porosity of 10–15%. The porosity level not only affects the composite properties, such as thermal conductivity and mechanical behavior, but also its leak-tightness.34 For nuclear structural applications, total leak-tightness is mandatory to contain gaseous fission products so different solutions to prevent SiCf/SiCm leakage have been proposed. For instance, covering the inner surface of the composite with a layer of high density monolithic SiC7 or with a metallic liner35 would allow fission gas retention.
The mechanical behavior of CMCs displays features that differentiate them from the other composites, such as polymer matrix composites or metal matrix composites, and from homogeneous (monolithic) materials. These features are due to heteroge-neous and multiscale composite microstructure and the respective properties of the constituents—fibers, matrix and fiber-matrix interphase.33 SiCf/SiCm pseudo-ductile behavior is achieved by micro-crack formation and propagation across the SiC matrix and crack deflection in the fiber-matrix interphase. The latter consist of a thin film of a compliant material, normally a single layer of pyrocarbon (PyC) or successive layers of PyC/SiC, with low shear stress deposited on the fiber surface prior to the matrix infiltration. In addition to deflect the matrix cracks, the interphase acts as a mechanical fuse and grants a good load transfer between the fiber and the matrix.34,36 For composites with negligible matrix creep, subcritical crack growth is controlled by the time dependent elongation of crack-bridging fibers.16 Finally, the ultimate failu of the composite generally occurs after crack saturation of the matrix and fiber over-load. This phenomenon is highly influenced by stochastic features as SiC fibers are brittle ceramics sensitive to present randomly distributed flaws which may act as stress concentrators.
SiC fibers1
In the early 80’ s, first generation SiC fibers developed by Prof. Yajima39 caught the interest from the aerospace and aeroengine industries. The availability of small diameter ceramic fibers meant the possibility of producing structural ceramic materials capable to work at temperatures where metallic alloys could not operate. Nevertheless, the characteristics of these fibers were not the expected in terms of elastic modulus and creep resistance hence triggering out an intensive R&D process that eventually led to the so-called third generation of SiC fibers whose properties approach to those of high-purity SiC.
◦C for 14 h obtaining a precursor which, although difficult to spin, could be converted to ceramic fibers.38 Finally, as it can be observed in Figure 2.3, it was found that the cross-linking stage that turned the PCS infusible was crucial in the obtention of high quality SiC fibers.
SiC fibers
First generation SiC based fibers
First generation fibers were fabricated by making infusible the PCS by cross-linking in air between 145 ◦C and 200 ◦C. After cross-linking, PCS fibers were pyrolyzed to induce the evaporation of low weight components and the reaction of C with Si. The resultant fiber was composed by nano-sized β-SiC surrounded by a disordered Si−O−C phase due to the O introduced during the cross-linking stage.
The two first SiC fibers available were Nicalon, commercialized by Nippon Carbon in 1982, and Tyranno, commercialized by Ube industries in 1987. The difference between these fibers was found in the fabrication process. The PCS precursor of Tyranno fibers was doped with titanium (polytitanocarbosilane, PTCS) as sintering aid.
Both fibers had a very similar chemical composition; however, despite the denom-ination of SiC fibers, the poor stoichiometry and the presence of free C resulted in poor mechanical properties at high temperatures when compared to bulk SiC. First generation SiC fibers had a low elastic modulus and crept above 1000 ◦C. Also they exhibited a poor chemical stability. Superficial silica layers appeared when heated in air above 1200 ◦C and CO formation induced the apparition of pores which caused the decohesion of the SiO2 − SiC interphase. Table 2.2 gathers the main characteristics of first generation fibers.
SIC, SIC FIBERS AND SICF/SICM COMPOSITES
Second generation SiC fibers
First generation SiC fibers poor characteristics were generally caused by their poor stoichiometry and the presence of the Si − O − C amorphous phase.
On the one hand, the O content of Nicalon fibers was introduced during their cross-linking stage. Therefore, a new O-free cross-linking method using e− irradiation was used to break the Si−CH3, Si−H, C −H chemical bonds allowing the formation of Si and C bonds. After the irradiation stage, the remaining free radicals trapped withing the fiber were removed by submitting the cross-linked fibers to a short heat treatment at 327 ◦C. The resultant SiC fibers, called Hi-Nicalon and produced by Nippon Carbon, presented an O content as low as the 0.5 wt.%.
On the other hand, the successor of the Tyranno, the Tyranno LOX-E, still exhibited high O content as it was not only introduced during the cross-linking stage but also due to the presence of T i on the polymeric precursor. T i addition to the PCS precursor was via the alkoxide T i(OR)4, with R = CnH2n+1, hence introducing O despite the O-free cross-linking stage. In order to reduce the O content, the metallic additive to the PCS precursor shifted from T i to Zr resulting in two types of fibers, Tyranno ZE and Tyranno ZM.
O reduction resulted in improved physical properties for Hi-Nicalon fibers. For instance, the elastic modulus was increased by a 35% with respect Nicalon fibers and thermal creep below 1000 ◦C was negligible. However, the improvement of Tyranno fibers was not so significant though their reduced O content. Table 2.3 gives the main characteristics of second generation SiC fibers.
Third generation SiC fibers
Low-O SiC fibers showed in general improved properties with respect first generation fibers. However, second generation SiC fibers still had a poor stoichiometry due to the excess of free C, causing a poor stability at high temperatures (cf. Table 2.3). As their properties remained insufficient to fulfill the requirements for their devised applications, efforts were directed towards achieving near-stoichiometric fibers which implied the reduction of the C excess present in Hi-Nicalon fibers. In order to do so, the pyrolysis of PCS fibers was carried out under a H2 atmosphere.
In the case of Tyranno fibers, the path followed to improve their stoichiometry was to change the chemical composition of the polymeric precursor. The metallic sintering aid shifted from Zr and T i to Al compounds. Also, the polymeric precursor, polyaluminocarbosilane (PACS), was cured by a two-stage pyrolysis process. First stage allowed oxide phases to decompose and CO to outgas above 1200 ◦C. In the second stage, the Al content aided sintering the SiC grains between 1600 ◦C and 1800 ◦C. The resulting SiC fibers, called Tyranno SA 3 (TSA3), contained up to 2 wt.% of Al.38 In addition, these fibers presented higher thermal stability and better properties than previous Tyranno fibers (LoxM, LoxE, ZMI and ZE).42 Table 2.4 gathers the main characteristics of third generation SiC fibers.
As it is shown in Figure 2.4, these fibers exhibit unique microstructures as a result of their fabrication process. Both fibers consist of highly faulted β-SiC grains surrounded with free turbostratic C, i.e. unmatched graphene layers, and very low oxygen con-tent.43,44 As it is shown in Figure 2.5, the diametrical free C distribution in HNS fibers is rather constant whereas TSA3 fibers present a decreasing free C concentration from the core to the surface. The heterogeneous C concentration of the latter is a result of the reactions occurring during the two-step thermal treatment and the presence of the metallic sintering aid.
The main microstructural differences between these fibers is found in their grain size with mean values of 20 nm for the HNS fibers and 200 nm for the TSA3 fibers. The smaller grain size of the former implies the smoother surface roughness of HNS fibers shown in Figure 2.6 which in turn produce composites with superior mechanical properties due to lower shear stresses between fiber and matrix.
SiC-based composites are referred frequently as the most promising structural materials for nuclear applications. Whereas there is an extensive characterization of the irradiation effects on single and polycrystalline SiC among the literature, including comprehensive bibliographic reviews,11,12 the characterization of the ir-radiation effects in SiC fibers still requires further investigation. In this chapter, a bibliographic review of the irradiation effects in bulk SiC and in the different generations of SiC fibers is presented.
Irradiation damage creation
Before detailing the irradiation effects in SiC, it is necessary to recall the different processes that produce the displacement of the target lattice atoms and how to quantify irradiation damage.1
When an energetic incident particle elastically interacts with a lattice atom there is a kinetic energy exchange between them. If this transmitted energy is higher than the threshold displacement energy of the knocked lattice atom, Ed, it will be ejected from its equilibrium site leading to a Frenkel pair formation: a vacancy and an interstitial atom. Also, as schematized in Figure 3.1, if the kinetic energy transfer is high enough, the displaced atom may have enough kinetic energy to displace not only one but many 1A detailed description of the displacement process and a full derivation of the different models of irradiation damage quantification can be found in Ref.
ION AND NEUTRON IRRADIATION EFFECTS IN SIC AND SIC FIBERS
atoms of the lattice which, in turn, will cause other displacement processes. Displace-ment cascades depend on the energy given to the primary knock-on atom (PKA) by the incident particle and the displacement cascade process will continue until the ki-netic energies of the displaced atoms are below the Ed. As a result of the displacement cascade, after thermal recombination of some of the displaced lattice atoms, a wide va-riety of stable lattice defects, ranging from isolated Frenkel pairs to large defect clusters (cf. Figure 3.1), can induce changes in the microstructure and properties of the host material.
Normally, the changes observed in a material exposed to irradiation will depend on the amount of stable defects which, in turn, will vary according to the total number of displaced atoms or dose. The most extended model to calculate the total number of displaced atoms by an indicent particle is the model proposed by Norget, Robinson and Torrens,47 (NRT) based on the initial work by Kinchin and Pease.48 The NRT model (Eq. 3.1) gives the total relative number of atomic displacements, νNRT , produced by a PKA with kinetic energy EP KA:
Irradiation damage creation
atomic displacements by elastic collisions and 0.8 is the displacement efficiency. There-fore, given an incident particle flux, the rate of atomic displacements, Rd, will be proportional to the target atom density, N, and the integration of the NRT damage function over the incident particle energy spectrum taking into account the interaction probability, thus:
One of the advantages of this model is that it is independent of the incident par-ticle and thus allows the comparison of the damage obtained with different types of irradiation such as neutrons, ion or electron irradiation. On the other hand, dpaN RT unit does not take into account defect recombination thus overestimating the number of total displaced atoms. Indeed, when a displaced target atom fills the vacancy left by another displaced it has no effect on the lattice configuration of the material.
With the NRT model, the problem to quantify the damage generation in a target material is reduced to the quantification of the PKA production by an incident flux of particles. It is at this moment when the nature of the incident of particle influences the damage process. In contrast to the straightforward neutron-nucleus elastic inter-actions due to their null electric charge, the interaction of charged particles with the target lattice cannot be simplified to a hard-sphere collision and needs to be otherwise described.
The energy loss of an incident charged particle is also due to Coulomb interactions between the incident particle and the target electron cloud. Therefore, the energy loss process of a charged particle within a material can be considered as two separated processes according to the type of interaction. The different processes are known as the electronic stopping (Se) and the nuclear stopping (Sn) powers, which refer respectively to the inelastic and the elastic processes that causes the gradual energy loss on the particle. The former can be described as inelastic collisions that result in small energy losses in which the target electrons are excited or ejected from their shells and dissipate their energy through thermal vibration of the target. The latter can be described as ion-atom and atom-atom collisions due to repulsive Coulomb potentials.
Figure 3.2 shows the morphology of the damage produced by different charged particles with an energy of 1 MeV as well as the damage profile of different particles with different energies as compared to the neutron flat energy deposition and damage profile. Also, areas with dominant electronic and nuclear energy regimes are highlighted in red and blue respectively.
Table of contents :
1 Introduction
2 SiC, SiC fibers and SiCf/SiCm composites
2.1 SiC
2.2 SiC based composites
2.3 SiC fibers
2.3.1 First generation SiC based fibers
2.3.2 Second generation SiC fibers
2.3.3 Third generation SiC fibers
3 Ion and neutron irradiation effects in SiC and SiC fibers
3.1 Irradiation damage creation
3.1.1 Monte Carlo simulations for irradiation damage estimation
3.2 Irradiation effects in SiC single crystals and CVD-SiC
3.2.1 Irradiation induced amorphization
3.2.2 Irradiation induced swelling
3.2.3 Degradation of physical properties
3.2.4 Thermal annealing of irradiation effects
3.3 Irradiation effects in SiC fibers
3.3.1 Irradiation induced surface degradation
3.3.2 Irradiation induced amorphization, densification and swelling
3.3.3 Irradiation induced degradation of physical properties
3.3.4 Influence of the fiber in SiCf/SiCm irradiation stability
4 Materials & Methods
4.1 Materials
4.2 Ion-irradiation facilities
4.2.1 JANNUS
4.2.2 GANIL
4.3 In situ tensile test device: MiniMecaSiC
4.4 Characterization techniques
4.4.1 Micro-Raman spectroscopy
4.4.2 Transmission electron microscopy
4.4.2.1 TEM thin foils preparation
4.4.3 Environmental scanning electron microscopy
5 Characterization of the ion-amorphization threshold conditions of third generation SiC fibers
5.1 Introduction
5.2 Ion-irradiation conditions
5.3 Results
5.3.1 Microstructural characterization of as-received materials
5.3.2 Ion-amorphization kinetics at RT
5.3.3 Ion-amorphization as a function of the irradiation temperature
5.4 Discussion
5.5 Conclusions
6 Characterization of the effects of thermal annealing on ion-amorphized 6H-SiC and third generation SiC fibers
6.1 Introduction
6.2 Experimental conditions
6.2.1 Materials
6.2.2 In situ E-SEM
6.2.3 In situ TEM
6.3 Results
6.3.1 Thermal annealing induced cracking
6.3.2 Thermal annealing induced recrystallization
6.3.2.1 Ion-amorphized 6H-SiC single crystal
6.3.2.2 Ion-amorphized HNS fiber
6.3.2.3 Ion-amorphized TSA3 fiber
6.4 Discussion
6.5 Conclusions
7 In situ characterization of ion-irradiation creep of third generation Tyranno SA3 SiC fibers
7.1 Introduction
7.2 Experimental conditions
7.2.1 Fiber selection and preparation
7.2.2 Ion-irradiation conditions
7.2.3 Ion-flux-induced temperature rise estimation
7.3 Results
7.3.1 Thermal creep
7.3.2 Influence of the irradiation temperature on in situ tensile tests
7.3.3 Characterization of irradiation creep at high irradiation temperatures
7.4 Post-mortem characterization
7.5 Discussion
7.6 Conclusions
8 Summary Conclusions & Future Work
