正在加载图片...
J, Vicens et al./Aerospace Science and Technology 7(2003)135-146 nearby pure crystalline carbon layer as well as the planar some carbon units embedded in an amorphous oxycarbide face between these reaction phases gave rise to new intergranular phase [211 statements. These observations were performed in SIC/CAS As proposed by Le Strat et al. and Le Strat [25, 26 the by Bonney and Cooper [5]. They proposed a solid state reaction begins by oxygen diffusion from the glass to the displacement reaction being rate-limited by the diffusion fibre and inside the fibre through the silicon oxycarbide of oxygen through the silica layer. Other models including phase where it exchanges with silicon or carbon. At the a passive oxidation reaction of Sic by Co has also been same time, silicon and carbon elements diffuse through the proposed [3, 30]. Benson et al. 3] supposed two steps in concentration gradient from the fibre to the glass. Oxygen carbon layer formation, depending on the partial pressure of which migrates in the oxycarbide phase comes in contact CO. The first stage is the formation of an initial carbon layer with SiC, forming a new oxycarbide phase. The carbon and by the reaction of Sic and Oz corresponding to reaction(1). silicon resulting from the oxidation of SiC and SiO.Cy and In the second stage, the carbon layer formation is controlled the pre-existing carbon in the fibres migrate towards the by the two following chemical reactions fibre surface. Silicon is incorporated in the glass forming 2Cs+O2(g)→2COg) SiOz. The carbon crystallizes at the glass interface and the carbon layer grows at the fibre surface. The oxygen SiCs+2 CO(g)- 3Cs SiO(s) ( amount available for diffusion decreases progressively in at the matrix/fibre interface and at the carba e spectively the glass. This slows down the reaction rate. A phase reorganization then occurs leading to the formation of the faces.In this model the carbon layer thickness depends on two sublayers. In the transition layer, the oxycarbide phase the oxygen diffusion rate inside the matrix and on that of tends to an equilibrium composition by rejecting some CO(g) in the carbon layer elements in excess. This induces the formation of an oxygen A phenomenological model describing the kinetics of rich silicon oxycarbide phase(45 at %), the growing of the carbon interfaces formation at SiC-glass interfaces of SiC nanocrystals and the carbon unit formation SIC/LAS and SiC/CAS composites has been presented by Qi et al. [38]. The model shows how variables(process pa- rameters, addition of oxides such as As203, melting history 5. ic/Bn dual-coated Nicalon-fibre-reinforced which can modify the oxygen activity in the glass)influence glass-ceramic matrix composites the carbon layer thickness and its stability Recently a model has been published by Hahnel et al. [18]. 5.1. The concept of the dual Sic/BN interphase They proposed a multiple-step reaction to describe their ob- ervations of the complex phases arrangement(carbon, car- This part describes recent works performed by the team bon and silica, silica) between the fibre and the matrix(Sic of J.J. Brennan at the United Technologies Research Cen- Nicalon/Duran composite). The thermodynamic approach ter, East Harford, Connecticut, USA, on a new Sic/bn dual ields to a four stage model of layer formation including the coated Sic Nicalon fibre reinforced BMAS matrix com- active oxidation of sic posites. Mechanical behavior and associated microstructure An interesting work performed in the SiC/Pyrex compos- changes were very carefully found out [6,, 40-43, 491 [25, 26] described a mechanism to explain the two transi- The embrittlement of the glass-ceramic matrix which tions and carbon layers formation very often observed at the appears in oxidizing environment can be either the con- fibre/glass interface in many composites [17, 22, 35]. This has sequence of oxygen diffusion to the fibre/matrix interface not been clearly explained in previous models. In this model via matrix microcracks reacting with the carbon layer or a three points were emphasized. A first point concerns with the"pipeline diffusion"from cut ends of fibres exposed to air at formation of the TL oxygen-rich interphase located between the composite surface. To maintain an oxidative stability at the CL and the fibre as illustrated before. Recent results sup- high temperatures, one approach consists of using fibre coat- ported the existence of Sio, Cy phases instead of SiO2 in ings applied to the fibres prior to composite processing Such TL. These oxycarbides are the same phases as those in the coatings should have two key functions: a mechanical fuse SiC Nicalon fibres but with higher contents compared with to allow crack deflection and load transfer from matrix to fi- SiC Nicalon fibre. A second point of this description is the bre and an improvement of the oxidation resistance. Studies resence of a certain amount of dissolved oxygen atoms in were carried out to replace the carbon-rich zone with a sim the glass with a fast diffusion rate(10-8 cm-s-)and ilar weakly bonded coating which also improves oxidative of non-bridging oxygen atoms with diffusion coefficients stability. CVD coatings such as BN and Sic or BN were de- between 10-14-10-12 cm s-I. It was also suggested that posited on SiC Nicalon fibres, on SiN-C HPZ fibre(Dow the diffusing species are atomic oxygen instead of O2(g) or Corning Corp )[44] and extended to composite with Hi of Co(g). From these considerations, diffusion lengths( Dt) Nicalon fibre [1]. These coated fibres were incorporated into were calculated using the hot pressing duration of the com- LAS, BMAS [8, 40] and BsAs [l]. The BN coatings(100- posites. A third point concerns the particular microstruc- 300 nm thick) were either amorphous or partly turbostratic ture of the Nicalon fibre consisting of SiC nanocrystals with as it is usually the case for low temperature BN-CVD. TheJ. Vicens et al. / Aerospace Science and Technology 7 (2003) 135–146 143 the nearby pure crystalline carbon layer as well as the planar interface between these reaction phases gave rise to new statements. These observations were performed in SiC/CAS by Bonney and Cooper [5]. They proposed a solid state displacement reaction being rate-limited by the diffusion of oxygen through the silica layer. Other models including a passive oxidation reaction of SiC by CO has also been proposed [3,30]. Benson et al. [3] supposed two steps in the carbon layer formation, depending on the partial pressure of CO. The first stage is the formation of an initial carbon layer by the reaction of SiC and O2 corresponding to reaction (1). In the second stage, the carbon layer formation is controlled by the two following chemical reactions: 2Cs + O2(g) → 2 CO(g) (2) SiCs + 2 CO(g) → 3Cs + SiO2(s) (3) Reactions (2) and (3) are supposed to occur respectively at the matrix/fibre interface and at the carbon/fibre inter￾faces. In this model the carbon layer thickness depends on the oxygen diffusion rate inside the matrix and on that of CO(g) in the carbon layer. A phenomenological model describing the kinetics of carbon interfaces formation at SiC–glass interfaces of SiC/LAS and SiC/CAS composites has been presented by Qi et al. [38]. The model shows how variables (process pa￾rameters, addition of oxides such as As2O3, melting history which can modify the oxygen activity in the glass) influence the carbon layer thickness and its stability. Recently a model has been published by Hähnel et al. [18]. They proposed a multiple-step reaction to describe their ob￾servations of the complex phases arrangement (carbon, car￾bon and silica, silica) between the fibre and the matrix (SiC Nicalon/Duran composite). The thermodynamic approach yields to a four stage model of layer formation including the active oxidation of SiC. An interesting work performed in the SiC/Pyrex compos￾ite [25,26] described a mechanism to explain the two transi￾tions and carbon layers formation very often observed at the fibre/glass interface in many composites [17,22,35]. This has not been clearly explained in previous models. In this model three points were emphasized. A first point concerns with the formation of the TL oxygen-rich interphase located between the CL and the fibre as illustrated before. Recent results sup￾ported the existence of SiOxCy phases instead of SiO2 in TL. These oxycarbides are the same phases as those in the SiC Nicalon fibres but with higher contents compared with SiC Nicalon fibre. A second point of this description is the presence of a certain amount of dissolved oxygen atoms in the glass with a fast diffusion rate (∼10−8 cm−1·s−1) and of non-bridging oxygen atoms with diffusion coefficients between 10−14–10−12 cm·s−1. It was also suggested that the diffusing species are atomic oxygen instead of O2(g) or of CO(g). From these considerations, diffusion lengths (Dt) were calculated using the hot pressing duration of the com￾posites. A third point concerns the particular microstruc￾ture of the Nicalon fibre consisting of SiC nanocrystals with some carbon units embedded in an amorphous oxycarbide intergranular phase [21]. As proposed by Le Strat et al. and Le Strat [25,26], the reaction begins by oxygen diffusion from the glass to the fibre and inside the fibre through the silicon oxycarbide phase where it exchanges with silicon or carbon. At the same time, silicon and carbon elements diffuse through the concentration gradient from the fibre to the glass. Oxygen which migrates in the oxycarbide phase comes in contact with SiC, forming a new oxycarbide phase. The carbon and silicon resulting from the oxidation of SiC and SiOxCy and the pre-existing carbon in the fibres migrate towards the fibre surface. Silicon is incorporated in the glass forming SiO2. The carbon crystallizes at the glass interface and the carbon layer grows at the fibre surface. The oxygen amount available for diffusion decreases progressively in the glass. This slows down the reaction rate. A phase reorganization then occurs leading to the formation of the two sublayers. In the transition layer, the oxycarbide phase tends to an equilibrium composition by rejecting some elements in excess. This induces the formation of an oxygen￾rich silicon oxycarbide phase (45 at %), the growing of the SiC nanocrystals and the carbon unit formation. 5. SiC/BN dual-coated Nicalon-fibre-reinforced glass–ceramic matrix composites 5.1. The concept of the dual SiC/BN interphase This part describes recent works performed by the team of J.J. Brennan at the United Technologies Research Cen￾ter, East Harford, Connecticut, USA, on a new SiC/BN dual coated SiC Nicalon fibre reinforced BMAS matrix com￾posites. Mechanical behavior and associated microstructure changes were very carefully found out [6,8,40–43,49]. The embrittlement of the glass–ceramic matrix which appears in oxidizing environment can be either the con￾sequence of oxygen diffusion to the fibre/matrix interface via matrix microcracks reacting with the carbon layer or a “pipeline diffusion” from cut ends of fibres exposed to air at the composite surface. To maintain an oxidative stability at high temperatures, one approach consists of using fibre coat￾ings applied to the fibres prior to composite processing. Such coatings should have two key functions: a mechanical fuse to allow crack deflection and load transfer from matrix to fi- bre and an improvement of the oxidation resistance. Studies were carried out to replace the carbon-rich zone with a sim￾ilar weakly bonded coating which also improves oxidative stability. CVD coatings such as BN and SiC or BN were de￾posited on SiC Nicalon fibres, on SiN–C HPZ fibre (Dow Corning Corp.) [44] and extended to composite with Hi Nicalon fibre [1]. These coated fibres were incorporated into LAS, BMAS [8,40] and BSAS [1]. The BN coatings (100– 300 nm thick) were either amorphous or partly turbostratic as it is usually the case for low temperature BN-CVD. The
<<向上翻页向下翻页>>
©2008-现在 cucdc.com 高等教育资讯网 版权所有