Directed melt oxidation and nitridation of aluminium alloys: a comparison B S.S. Daniel and V.s. R Murthy Department of Materials and Metallurgical Engineering, Indian Institute of Technology, Kanpur 208016, India Received 4 September 1995; accepted 21 September 1995 Directed melt oxidation and nitridation are recently developed in-situ methods for the formation of metal-ceramic composites. Both processes are based on liquid-gas reaction, wherein a nolten alloy is reacted with a gaseous species, i.e. oxygen or nitrogen, to form A2O,/Al or AIN/A composite microstructures Composite formation is controlled by the alloy composition partial between these two processes, there are differences in the microstructural development. Furthe pressure of gaseous species, processing tomperature and time. Although there are similarities the growth rates of these composites are accelerated using filler materials which provide sites for secondary nucleation and in turn compete with the primary growth of the ceramic phase. the mechanical properties of Al,O Al and AIN/Al are comparable, but AIN/Al exhibits higher thermal conductivity and sensitivity to moisture. Finally, the applications and limitations of these composites are presented Keywords: oxidation; nitridation; in-situ composite Introduction reactions, there are differences between the two. First To meet the demand of recent high-technology applica- in the former method the reacting gas is injected into tions, a large number of metal-matrix composites the molten pool and the reaction time is relatively short (MMCs)were developed adopting suitable processing to control the process effectively. In the latter the reac- methodologies-3. These conventional processing tion is a continuous process over an extended period of methods, however, have some limitations, i.e. residual time. Second, by modifying both alloy and gas compo- microporosity, uneven distribution of reinforcing mate- sition, a large number of reinforcing constituents can be rials(inhomogeneity), non-wetting of reinforcements, formed On the other hand, in direct melt reactions a matrix-rcinforccment interfacc microstructure control singular reinforcing phase is possible. Third, by blowing and its cleanliness. etc. To overcome these barriers. gases, fine(10-20 um) and isolated particles are formed several innovative processing methods have been devel- and, beyond a certain percentage, higher-volume frac oped, wherein the reinforcements are formed by in-sitt tions of dis ssible, whereas, in the reactions. These processes are broadly classified based DIMOX/PRIMEX processes the ceramic phases that on their reactant phases, i.e. liquid-gas, liquid-solid and re formed are large in percentage terms and moreover solid-solid reactions. In liquid-gas reaction processing the metal and ceramic phases are interconnected there one method of forming a reinforcing phase is by injec- is also a greater fiexibility in the volume fraction of rein tion of a reactive gas into a liquid alloy reservoir. forcing phase, i.e. both metal and ceramic matrix Depending on the alloy and gas composition, one or composites are feasible by controlling process variables more phases are formed+6. Another technique which en recently developed is again based on The basic principle liquid-gas reaction, but uses a different approach. In In directed melt composite reactions a molten alloy is this process(developed by Lanxide Corporation, USA) reacted with a gaseous species (sometimes a gaseous the liquid alloy is treated in a static/flowing gas en- mixture)to develop a metal-ceramic composite AIN/Al composites are formed as the reaction product. nitriding, ALO,Al or AIN/Al composites are obtain.or vironment. Depending on the atmosphere, AlOyAl or Depending on the atmosphere, whether oxidizing or hese processes are now commonly referred to as respectively(Figure 1). Under suitable conditions a directed melt oxidation(DIMOX), and direct melt reaction product initially forms on the surface of th Although both processes are based on liquid-gas uous wich y and the product grows outward by contin- nitridation(PRIMEX) respectively molten all kingof liquid alloys through microchannels that are present within the reaction product Thus the Correspondence to v.S.R. Murth final composite contains an interconnected and inter. penetrating network of metal and ceramic. Further 261-306995/03015507 Materials Desian Volume 16 Number 3 1995 155
Directed melt oxidation and nitri~ation of a~uminium alloys: a comparison B. S. S. Daniet and V. S. R. gushy Department of Materials and Metalfurgical Engineering, Indian Institute of Tec~ffology, Ka~~ur ZU8U76, vodka Received 4 September 7995; accepted 21 September 1995 Directed melt oxidation and nitridation are recently developed in-situ methods for the formation of metal-ceramic composites. Both processes are based on liquid-gas reaction, wherein a molten alloy is reacted with a gaseous species, i.e. oxygen or nitrogen, to form AI&/AI or AINfAl composite microstructures. Composite formation is controlled by the alloy composition, partial pressure of gaseous species, processing temperature and time. Although there are similarities between these two processes, there are differences in the microstructural development. Further, the growth rates of these composites are accelerated using filler materials which provide sites for secondary nucleation and in turn compete with the primary growth of the ceramic phase. The mechanical properties of AI,OdAI and AIN/AI are comparable, but AIN/AI exhibits higher thermal conductivity and sensitivity to moisture. Finally, the applications and limitations of these composites are presented. Keywords; oxidation; nitridation; in-situ composite Introduction To meet the demand of recent high~technology applications, a large number of metal-matrix composites (MMCs) were developed adopting suitable processing methodologies’“. These conventional processing methods, however, have some limitations, i.e. residual microporosity, uneven distribution of reinforcing materials (inhomogeneity), non-wetting of reinforcements, matrix-reinfor~ment interface mi~rost~cture control and its cleanliness, etc. To overcome these barriers, several innovative processing methods have been developed, wherein the reinforcements are formed by in-situ reactions. These processes are broadly classified based on their reactant phases, i.e. liquid-gas, liquid-solid and solid-solid reactions4. In liquid-gas reaction processing, one method of forming a reinforcing phase is by injection of a reactive gas into a liquid alloy reservoir. Depending on the alloy and gas composition, one or more phases are formed4”. Another technique which has been recently developed is again based on liquid-gas reaction, but uses a different approach. In this process (developed by Lanxide Corporation, USA) the liquid alloy is treated in a static/flowing gas environment. Depending on the atmosphere, AI,O,/Al or AIN/AI composites are formed as the reaction product. These processes are now commonly referred to as directed melt oxidation (DIMOX)7,S and direct melt nitridation (PRIMEX)‘,“, respectively. Although both processes are based on liquid-gas Correspondence to V. S. R. Murthy 026%3069/95/030155-07 reactions, there are differences between the two. First, in the former method the reacting gas is injected into the molten pool and the reaction time is relatively short to control the process effectively. In the latter the reaction is a continuous process over an extended period of time. Second, by modifying both alloy and gas composition, a large number of reinforcing constituents can be formed. On the other hand, in direct melt reactions a singular reinforming phase is possible. Third, by blowing gases, fine (lo-20 pm) and isolated particles are formed and, beyond a certain percentage, higher-volume fractions of dispersoids are not possible, whereas, in the DIM~~RIMEX processes the ceramic phases that are formed are large in percentage terms and, moreover, the metal and ceramic phases are interconnected. There is also a greater flexibility in the volume fraction of reinforcing phase, i.e. both metal and ceramic matrix composites are feasible by controlling process variables. The basic principle In directed melt composite reactions a molten alloy is reacted with a gaseous species (sometimes a gaseous mixture) to develop a metal-ceramic composite. Depending on the atmosphere, whether oxidizing or nitriding, AI,OJAl or AlN/Al composites are obtained respectively (~~g~~~ I). Under suitable conditions, a reaction product initially forms on the surface of the molten alloy and the product grows outward by continuous ‘wicking’ of liquid alloys through microchannels that are present within the reaction product. Thus the final composite contains an interconnected and interpenetrating network of metal and ceramic. Further, Materials St Desian Volume 76 Number 3 1995 155
Directed melt oxidation and nitridation in aluminium alloys: B S S. Daniel and V.S. R Murthy Vapor phase Oxidising Nitriding Atmosphere Metal-Ceramic Alloy AlO Crucible Vapor phase B Metal-c。rami Composit Molten Alloy A20 Figure 1 Schematic of outward growth of ceramic/ metal reaction product(A)into free space and( B)through filler material. The corresponding microstructures in oxidizing and nitriding atmospheres are show these in-situ reactions can be utilized to obtain 'multi Additionally, these elements also improve the wettability phase composites by placing filler materials in the path of liquid alloys during wicking. In addition to the volatile of the growth direction(Figure 1), lo, The objective of species, elements such as Si, Ge and Sn are also added to using filler materials is mainly to increase growth rates, in control the reaction kinetics". In oxidation experiments addition to modifying the properties. a wide range of there is a choice of adding these elements either in materials are produced using different filler materials. elemental form (in the liquid alloy) or as surface Some examples are SiC/AlO Al, Al,O AlO,Al, dopants, Addition of surface oxides is believed to TiB,AIN/AL, SiNAIN/AL, etc,. All composit reduce the incubation period obtained by gas-metal reactions exhibit not only good In contrast, in nitridation experiments both II and IIB mechanical properties(KIC, stiffness, wear and corrosion elements are added to the alloy. In this process, group II resistance)but also improved electrical and thermal elements not only improve the wettability but also act as (shock resistance) properties compared to their ceramic a gettering agent in maintaining the oxygen partial pres- counterparts. The potential advantages of this method sure below a critical level. In the dimOx process are low processing cost, simplicity, near-net shape and AlOy Al composite growth was reported in binary alloy lexibility in filler material selection like Al-Mg, Al-Zn and Al-Li alloys where there are few limitations in nitridation For instance, AIN/Al composite growth was not seen in Al-Mg or Al-Si Process variables binary alloys up to 1450oC when a 4oC/min heating rate When pure aluminium is oxidized a thin oxide film is Al-Si-Mg exhibited accelerated weight gain beyond formed on the surface preventing further oxidation of the 1200oC19. In contrast, Scholz and Greil2D,21 reported metal. The oxidation behaviour of pure aluminium is conversion of Al to AIN in Al-2.3Li and Al-2.5Mg drastically changed with small amounts of volatile alloys. Additionally, in nitridation, small quantities of elements such as Mg, Zn, Li and Na-3. Initially, these iron were found to aid the infiltration whereas in oxida elements diffuse at a rapid rate to form a porous oxide tion, iron additions give rise to undesirable intermetallic on the surface. However, at a later stage, the surface which in turn affect the mechanical properties". In both oxide dissociates under a concentration gradient and processes, nickel addition was found to be beneficial helps to maintain the Mg level in the liquid reservoir because it refines the composite structure 156 Materials design Volume 16 Number 3 1995
Vapor Phase A Metal-C~amic Composite A1203 Crucible Vapor Phase B Filler Material Metal-Ceramic Composite v ‘AI203 Crucible Directed melt oxidation and nitridation in ~lu~iniu~ alloys: B. S. S. t)anje/ and V. S, R. Mu&y Nitriding Atmosphere Figure 1 Schematic of outward growth of ceramicl metal reaction product (A) into free space and (B) through filler material. The co~esponding microstructures in oxidizing and nitriding atmospheres are shown these in-situ reactions can be utilized to obtain ‘multiphase’ composites by placing filler materials in the path of the growth direction (F&z 1)‘,‘*. The objective of using filler materials is mainly to increase growth rates, in addition to modifying the properties. A wide range of materials are produced using different filler materials. Some examples are SiC/AI,O,/Al, Al,O,/Al,O,/Al, TiB~AlN/~, Si~N~AlN/Al, etc7,‘*. All composites obtained by gas-metal reactions exhibit not only good mechanical properties (&, stiffness, wear and corrosion resistance) but also improved electrical and thermal (shock resistance) properties compared to their ceramic counte~a~, The potential advantages of this method are low processing cost, simplicity, near-net shape and ~exibility in filler material selection. Process variables A flay selection When pure al~inium is oxidized a thin oxide film is formed on the surface preventing further oxidation of the metal. The oxidation behaviour of pure aluminium is drastically changed with small amounts of volatile elements such as Mg, Zn, Li and Na”-‘3. initially, these elements diffuse at a rapid rate to form a porous oxide on the surface. However, at a later stage, the surface oxide dissociates under a concentration gradient and helps to maintain the Mg level in the liquid reservoir’4. Additionally, these elements also improve the wettability of liquid alloys during wicking. In addition to the volatile species, elements such as Si, Ge and Sn are also added to control the reaction kineticP. In oxidation experiments there is a choice of adding these elements either in elemental form (in the liquid alloy) or as surface dopants”,“. Addition of surface oxides is believed to reduce the in~bation period. In contrast, in nitridation experiments both II and IIB elements are added to the alloy. In this process, group II elements not only improve the wettability but also act as a gettering agent in maintaining the oxygen partial pressure below a critical level. In the DIMOX process, Al,03/Al composite growth was reported in binary alloys like Al-M& Al-Zn and Al-Li alloys@ where there are few limitations in nitridation. For instance, AlN/Al composite growth was not seen in Al-Mg or Al-Si binary alloys up to 1450°C when a 4Wmin heating rate was employed. However, ternary alloys such as Al-Si-Mg exhibited acc&erated weight gain beyond 1200°C’9. In contrast, Scholz and Grei120,2’ reported conversion of Al to AlN in Al-2.3Li and Al-2.5Mg alloys. Additionally, in nitridation, small quantities of iron were found to aid the in~ltration” whereas in oxidation, iron additions give rise to undesirable intermetallics, which in turn affect the mechanical properties2’. In both processes, nickel addition was found to be beneficial because it refines the composite structurez3. 156 Materials & Design Volume 16 Number 3 1995
Directed melt oxidation and nitridation in aluminium alloys: B. S.s. Daniel and V. s R Murthy equilibrium thermodynamics, it is established that an oxygen partial pressure of -10 Pa is required for AIN formation. This is never achieved, but fortunately in the presence of local gettering agents such as Mg, Li or Na introduced as alloying additives, kinetics overrules Addition of H, to N,(forming gas) hel oxygen activity low and further improves the kinetics of the process,. Better wetting of the preform surface by the liquid metal is anticipated in a nitrogen atmosphere Composite growth Nucleation and growth mechanisms during high temperature oxidation of liquid aluminium alloys have been studied, 4: 20.27. In the early stages, when the alloy is held at above the liquidous temperature a duplex oxide(MgO MgAl O4) layer forms on the surface These oxide layers are porous and contain intercon nected microdiscontinuities. The liquid metal from the underneath reservoir wicks through these channels and emerges as small nodules on the surface. Many such Aloy Reservo nodules coalesce to form cauliflower, type of colony on the surface. Finally, several such colonies form planar oxidation front. At each stage, nodules are covered with a duplex layer and magnesium (or any Figure 2 Cross-sectional volatile species) in the alloy reservoir depletes with time When the sium content in the alloy drops belot a threshold value (i.e. 0.3 wt% Al2O3)columnar crystals Temperature and time nucleate and grow to several tens of microns uniter In addition to chemistry of the melt, control of the ruptedly (Figure 2). ALO, develops mainly as a with tion kinetics is established by varying temperature growth direction maintained parallel to the c-axis of time to arrive at the desired microstructure. At temper Al,o, " 1. These Al,O, crystals grow in an interconnected atures below 950@C selective conversion of magnesium manner, but appear isolated in two-dimensional to MgO-MgAl2O, takes place. On the other hand, sections. Finally, when there is a bulk growth temperatures above [400c lead to formation of a ALOyAl composite, an oxygen gradient builds porous material. Hence, process temperatures are across the layers and Mgo and MgA1 O in the layers confined between 1000oC and 1300%C .In contrast. underneath dissociate to give Mg", Al*and O-ions conversion of Al to aIN can be achieved at relatively Mg*ions diffuse to the surface and maintain the non lower temperatures(2750C)when compared to oxida- protective Mgo layer, whereas AP*and 02-contribute tion. Further, in nitridation, temperature has a more to growth of Al,03 3, 4.28 pronounced effect on structure; at lower temperatures, The number of steps that are involved in aIN/Al predominantly metal-matrix composites are obtained, composite growth are smaller compared to oxidation whereas at higher temperatures, metal-dispersed Here no complex surface oxides are formed. However, eramic matrix are generated. Such flexibility is not the basic processes such as nodule formation by wicking observed in oxide systems, where the final microstruc- and the conversion of these nodules to composite struc ture is ceramic dominant. To some extent, temperature ture ture are seen(Figure 3). The specific volume ratio of so dependent on the the nitride layer(VAIN/VA=1.26)is greater than unity composition of the alloy and gaseous species and is expected to form a protective nitride surface layer. Magnesium present in the alloy is reported to Atmosp have a catalytic effect on nitride formation such that Both reaction processes are sensitive to partial pressure helps to transfer the surface reaction into a volume of oxygen.4. Nagelbergreported a Po4 dependency reaction". Based on the kinetics of composite forma- for the oxidation of Al-10Si-3Mg alloy at a weight gain tion, Aluminium to aIn conversion is broadly classified of 0. 2 g/cm, whereas no significant influcnce of oxygen into four reaction domains. They are: (1)passivating partial pressure was noticed in a complex alloy". Thus surface nitridation(observed in many binary alloys), oxidation reaction kinetics appears to be largely depen-(2) n-controlled volume nitridation, ( 3)volume dent on alloy composition. On the other hand, for nitric involving outward growth of AIN-AI ture-free nitrogenous atmosphere is essential. From AIN-Al microstructures in domains(3)andya% effective conversion of Al to AIN, an oxygen and mois- and (4) break-away nitridation. The morphologies Materials Design Volume 16 Number 3 1995 157
Figure 2 Cross-sectional microstructure depicting various phases of directed melt oxidation. Inset shows thr~-dimensional view of A&O, crystals Temperature and time In addition to chemistry of the melt, control of the reaction kinetics is established by varying temperature and time to arrive at the desired microst~cture. At temperatures below 950°C selective conversion of magnesium to Mg~MgAl~~~ takes place. On the other hand, temperatures above 1400°C lead to formation of a porous material. Hence, process temperatures are confined between 1000°C and 1300°C. In contrast, conversion of Ai to AlN can be achieved at relatively lower temperatures (2750°C) when compared to oxidation”. Further, in nitridation, temperature has a more pronounced effect on structure; at Iower temperatures, predominantly metal-matrix composites are obtained, whereas at higher temperatures, metal-dispersed ceramic matrix are generated. Such flexibility is not observed in oxide systems, where the final microstructure is ceramic dominant. To some extent, temperature and time combinations are also dependent on the composition of the alloy and gaseous species. Atmosphere Both reaction processes are sensitive to partial pressure of oxygen14~24. Nagelberg” reported a pg; dependency for the oxidation of Al-IOSi-3Mg alloy at a weight gain of 0.2 g/cm*, whereas no significant influence of oxygen partial pressure was noticed in a complex alloy”. Thus oxidation reaction kinetics appears to be largely dependent on alloy composition. On the other hand, for effective conversion of Al to AIN, an oxygen and moisture-free nitrogenous atmosphere is essential. From equilibrium the~odynamics, it is es~blished that an oxygen partial pressure of -lOmBPa is required for AIN fo~ation, This is never achieved, but fortunately in the presence of local gettering agents such as Mg, Li or Na introduced as alloying additives, kinetics overrules. Addition of H2 to N2 (forming gas) helps to keep the oxygen activity low and further improves the kinetics of the process’. Better wetting of the preform surface by the liquid metal is anticipated in a nitrogen atmosphere2? Composite growth Nucleation and growth mechanisms during high temperature oxidation of liquid aluminium alloys have been studied*3,~~26,“. In the early stages, when the alloy is held at above the liquidous temperature, a duplex oxide (MgO + MgA1,03 layer forms on the surface. These oxide layers are porous and contain interconnected microdiscontinuities. The liquid metal from the underneath reservoir wicks through these channels and emerges as small nodules on the surface. Many such nodules coalesce to form a ‘cauliflower’ type of colony on the surface. Finally, several such colonies form a planar oxidation front. At each stage, nodules are covered with a duplex layer and magnesium (or any volatile species) in the alloy reservoir depletes with time. When the magnesium content in the alloy drops below a threshold value (i.e. 0.3 wt% Al,OJ columnar crystals nucleate and grow to several tens of microns uninterruptedly (Figure 2). A&O, develops mainly as OL with growth direction maintained parallel to the c-axis of A1203”“. These A&O, crystals grow in an interconnected manner, but appear isolated in two-dimensional sections. Finally, when there is a bulk growth of AI,O,/Al composite, an oxygen gradient builds up across the layers and MgO and MgA1204 in the layers underneath dissociate to give Mg2’, A13+ and 0” ions. Mg” ions diffuse to the surface and maintain the nonprotective MgO layer, whereas AP’ and 0”” contribute to growth of Al20313,24? The number of steps that are involved in AlN/Al composite growth are smaller compared to oxidation. Here no complex surface oxides are formed. However, the basic processes such as nodule formation by wicking and the conversion of these nodules to composite structure are seen (Figure 3)29. The specific volume ratio of the nitride layer ( VxlN/VA, = 1.26) is greater than unity and is expected to form a protective nitride surface layer. Magnesium present in the alloy is reported to have a catalytic effect on nitride formation such that it helps to transfer the surface reaction into a volume reaction2’. Based on the kinetics of composite formation, Aluminium to AlN conversion is broadly classified into four reaction domains. They are: (1) passivating surface nitridation (observed in many binary alloys), (2) diffusion-controlled volume nitridation, (3) volume nitridation involving outward growth of AlN-Al and (4) break-away nitridation. The morphologies of AIN-Al mi~rostructures in domains (3) and (4) are Materials & Design Volume 16 Number 3 1995 157
Directed melt oxidation and nitridation in aluminium alloys: B S S. Daniel and V.S. R Murthy A B AIN/AI Allo 100 C D 20 um 10m Figure 3 SEM micrographs showing various domains of nitride formation. (A)Surface nitridation and globule formation in the early stages of nitridation.(B)AIN/Al composite microstructure. (C)Deep etched sample showing morphology of interconnected AIN. (D) Ain crystal growth Table 1 Properties of Al2O, and AIN AIN Corundum Wurtzite 4.76;c1299a3.l1;c498 N】 RIDAIJON 140-170 TCE(25500C)(×10◆C Dielectric constant at 1 MHz Flexural strength(MPa) 5 Fracture toughness(MPa vm)2-3 Elastic modulus(GPa) 320-370 300-310 (15 N scale) 95.5 94.5 computed for different alloys during oxidation varies from 89 to 400 kJ/mol 1, 4 263. A typical growth for the Figure 4 Schematic diagram showing weight gain rates during DiMOX process can be up to 5-8 mm/day, but it can be further improved to 30 mm/day when preforms are shown in Figure 3. The AIN formed in the composite used32. The activation energy for AIN formation is 100 has a wurtzite structure with the growth direction kJ/mol and the growth rates are at least three orders oriented along the [0001] direction". The liquid metal higher compared to oxidation(Table 1) exhibits good wettability with aiN, and finds it easy to wick through the AIN crystals during growth. In oxida- Composites using filler materials tion, wicking is due to capillary action aided by magne- In the direct melt infiltration of composites, the growth in an oxide of reaction product(Al, O, /Al or AIN/AD)is restricted or system is expected to be slower channelled within the cavities of a filler material. The Both in nitridation and oxidation, different stages in filler material can be loose powder (or fibres )or owth are distinctly visible in themogravimetric analy- sintered porous ceramic body, which is usually in the sis (TGa)(Figure 4). The activation energy values shape of the final product. During infiltration, limited 158 Materials Design Volume 16 Number 3 1995
Directed melt oxidation and nitfidation in a/u~i~iu~ a/fop: B. S. S. Daniel and V. S. R. hubby Figure 3 SEM micrographs showing various domains of nitride formation. (A) Surface nitridation and globule formation in the early stages nitrida .tion. (B) AlN/Al composite microstructure. (C) Deep etched sample showing morphology of inter~nnected AIN. (D) AIN crystal grov during break-away nitridatiin MgA$O~ Formation Tim4 - Figure 4 Schematic diagram showing weight gain rates during directed melt oxidation and nitr~dation shown in Figure 3. The AlN formed in the composite has a wurtzite structure with the growth direction oriented along the [OOOI] direction”. The liquid metal exhibits good wettability with A1N3’, and finds it easy to wick through the AlN crystals during growth. In oxidation, wicking is due to capillary action aided by magnesium-induced wetting. Hence, liquid rise in an oxide system is expected to be slower. Both in nitridation and oxidation, different stages in growth are distinctly visible in themogravimetric analysis (TGA) (Figure 4). The activation energy values of vth Table 1 Properties of A&O, and AIN Property AR AIN Crystal structure Corundum Wurtzite Lattice parameter (A) a 4.76; c 12.99 a 3.11; c4.98 Theoretical density &m/cm’) 3.98 3.26 Thermal conductivity (Wm.‘K-l) 26 14cL170 TCE (25-SOO’C) (x lO?‘C) 7.1 4.19 Dielectric constant at 1 MHz 9.5 Resistivity at RT (&rn) >10J4 ;Oli Flexural strength (MPa) 400 280-320 Fracture toughness (MPa t’;;) 2-3 Elastic modulus (GPa) 320-370 300-3 10 Rockwell hardness (I.5 N scale) 95.5 94.5 computed for different alloys during oxidation varies from 89 to 400 kJ/mo111~t4~~~31. A typical growth for the DIMOX process can be up to 5-8 mm/day, but it can be further improved to 30 mm/day when preforms are used32. The activation energy for AIN formation is -100 kJ/moI and the growth rates are at least three orders higher compared to oxidation (Table 1)9,20. Composites using filler materials In the direct melt in~ltration of composites, the growth of reaction product (Al,O,/Al or AlN/AI) is restricted or channelled within the cavities of a filler material. The filler material can be loose powder (or fibres) or a sintered porous ceramic body, which is usually in the shape of the final product. During infiltration, limited 158 Materials & Design Volume 16 Number 3 1995
Directed melt oxidation and nitridation in aluminium alloys: B. S.S. Daniel and V.S. R Murthy oxygen-induced reactions. To reduce the degradation and imt bility between the matrix and come essentia Nicalon SiC fibre preforms are coated to prevent surface oxidation and minimize the interfacial reactions33 During infiltration, filler materials act as secon nucleation sites for the reaction product(AlO, or AIN and refine the composite microstructure (Figure 5), while maintaining a growth direction similar to that of primary growth 22. Secondary nucleation seen in oxide systems is attributed to the reduction of oxide coatings Sic that are inherently present on the filler material (e.g. 5 ur Sio2 or SiC) or that which are formed(ZnO, Mgo vapour deposits)during infiltration. For instance, silica B present on SiC is reduced by the advancing aluminium 3SO2+4A→3Si+2Al2O3 and Al,O, tends to nucleate on the surface of Sic par- ticles. The reaction by-product, silicon, occupies portion of channel space and in extreme cases growth elements(e.g. Zn)that are added to the liquid melt form th front. The deposited Zno is 5 um reduced by the advancing aluminium to give rise to AlO, nucleation Figure 5 Secondary nucleation of (A)AL, 0, on SiC particulate and 3Zn0+ 2Al-Al2O3+ 3Zn (B)AIN oI (Nicalon) fibi In AlO, filler materials, newly formed oxide reactions between the preform and liquid alloys are epitaxial growth, whereas on Sic particles,no nd to enhance the lographic matching was evident". In nitridation wettability. In nitridation, adverse reactions are rather secondary nucleation is verge. limited, however, in an oxidative atmosphere the rein- nous, but a clear mechanism relating to their surface forcements can lose their basic properties due to chemistry is not known Table 2 Mechanical properties of direct melt oxidized composites,> Property Al,O Al composite SiCPAlOyA Al,OAl AlO/Al Uniaxial 2D, 12HSW Youngs modulus(GPa) Shear modulus(GPa) Flexural strength 720±150 540± (3-point bend test) (WM 20)(WM 22)(max 880) (max 620) (MPa vm (max 29) (max 16) 414-2100 0.1370.312 Vickers hardness(GPa 141-15.0 at700°C l18-39.6 Coeff. of thermal expansion 9.3-11.0 Dielectric constant o C-axis of Al O, ⊥ to c-axis of Al2O3 4 awith variation in residual metal content Weibull modulu Materials Design Volume 16 Number 3 1995 159
Directed melt oxidation and nitridatjo~ in aiuminium alloys: f3. S. S. Daniel and V. S. R. Murthy Figure 5 Secondary nucleation of (A) A&O, on Sic particulate and (B) AlN on Sic (Nicalon) fibres reactions between the preform and liquid alloys are inevitable and these reactions tend to enhance the wettability. In nitridation, adverse reactions are rather limited, however, in an oxidative atmosphere the reinforcements can lose their basic properties due to oxygen-induced reactions. To reduce the degradation and improve the compatibility between the matrix and the fibre, coatings become essential. For instance, Nicalon Sic fibre preforms are coated to prevent surface oxidation and minimize the interfacial reactions33. During infiltration, filler materials act as secondary nucleation sites for the reaction product (A&O, or AlN) and refine the composite microst~cture (r;igure 5), while maintaining a growth direction similar to that of primary growth iLz2 . Secondary nucleation seen in oxide systems is attributed to the reduction of oxide coatings that are inherently present on the filler material (e.g. Si02 or Sic) or that which are formed (ZnO, MgO vapour deposits) during infiltration. For instance, silica present on Sic is reduced by the advancing aluminium via the displacement reaction 3SQ + 4Al-+ 3Si + 2Al,O, (1) and Al,O, tends to nucleate on the surface of Sic particle.?. The reaction by-product, silicon, occupies a portion of channel space and in extreme cases growth is inhibited due to a ‘choking’ effect3’. Further, the solute elements (e.g. Zn) that are added to the liquid melt form a vapour and are deposited on the filler material surface ahead of the growth front. The deposited ZnO is reduced by the advancing aluminium to give rise to A&O, nucleation23. 3Zn0 f 2Al+ Al& + 3Zn (2) In A&O, filler materials, newly formed oxide exhibits epitaxial growth, whereas on Sic particles, no crystallographic matching was evident34. In nitridation, secondary nucleation is expected to be simply heterogenous, but a clear mechanism relating to their surface chemistry is not known. Table 2 Mechanical properties of direct melt oxidized composites’“*3~‘,4~ Property Al,O,/AI composite without preforma Young’s modulus (GPa) 88-304 Shear modulus @Pa) 41-123 Flexural strength 45-525 (4-point) (MPa) (3-point bend test) Fracture toughness 2.9-9.5 Wa ~‘3 Compressive strength 414-2100 @@a) Poisson’s ratio 0.137-0.312 Vickers hardness (GPa) 1.41-15.0 at 700°C - Thermal conductivity 1 t s-39.6 (Wm-’ K-l) Coeff. of thermal exapansion 9.3-I 1 .o (X 10*/K) Dielectric constant [j to c-axis of A&O, 8.0 I to c-axis of A&O3 6.4 With variation in residual metal content bWeibull modulus 403d Al,OrtAl 301 312 (WMb 20) 5.9 - 8.3 4.9 36 9.0 SiCd SiC~Al~O~lAl Al~O,IAI Uniaxial 2D, I2HSW 324 - 334 720 k 150 540 f 60 (WM 22) (max 880) (max 620) 6.9 27 k3 15 + 1 (max 29) (max 16) - - 6.4 3.7 82 7.5 Materials & Design Volume 16 Number 3 1995 159
Directed melt oxidation and nitridation in aluminium alloys: B S. S. Daniel and V. S R. Murthy Table 3 Mechanical properties composites,37. Materi Flexural Tensile Young's Shear Poissons persian (MPa) (MPavm)(MPa) (GPa) (×10)(wmK)(kgmm2) AIN/Al 66-40095-11.1 AIN/AIN/Al TiB2/AIN/Al 13.5 9.5 42-457165-18065-720.260.54 2865064 ALo/AIN/AI275+50°8+1 523+5410.74+2.1 92-94 "Naturally aged with 0. 4 um reaction layer with 1.5 um reaction layer As the liquid wicks through the preform, occasionally Applications and limitations the reaction front is blocked and results in macroporos- Keeping the advantages(simplicity, flexibility and cost ity, whereas local solidification shrinkage of metal in the of the process in mind, a large number of patents have composite gives rise to microporosity. These porosities already been filed on these materials. Landini et are claimed to increase with temperature, magnesium made valve inserts using SiC and Al O, particulate rein- content in the alloy and reinforcing particle size forced Al2O Al composites. These components were Porosities can be controlled to a limited extent either by successfully tested on a diesel engine. Similarly, compo- sothermal holding between the oxidation temperature nents made from SiC particulate -reinforced Al20-Al and the liquidus of the alloy or by altering the alloy composites(AInax products, USA) have been used for composition hydrocyclone cone liners and apexes and the perform- Finally, there were attempts to model the complex ance was superior compared to conventionally used infiltration process 5-37 However, so far there are no parts". SiC(Nicalon) fibre reinforced AL,O,/Al com- realistic models to estimate the growth rates in any of posites are seen as future materials for gas turbine hese system applications". The other probable applications of oxide composites include wear parts for pumps, jet mill noZZ les,automotive parts, etc On the other hand Mechanical properties AIN/Al composites(with their superior thermal conduc- Like many composite materials, the properties of tivity) are likely to find a place in the electronics pack- The major limitations of these processes are the type and volume fraction of the reinforcing phase. The properties of reinforcing phase are listed in 1. It takes a long time to develop oxide-nitride rein Tables 2 and 3. Although the mechanical properties of forced composites and hence the process cannot be AIN and Al,O, are almost similar, ain has a greater used for producing very large components thermal conductivity (ten times higher)and lower 2. Residual porosities(both micro and macro)are thermal expansion coefficient as compared to Al 03 limitation in achieving best mechanical perform AIN exhibits excellent thermal shock resistance and ance. due to these inherent defects these materials stable up to -1000oC in an oxidizing atmosp re unlikely to find applications in critical areas Additionally, ain has excellent corrosion resis 3. ain base composites are sensitive to moisture and good wettability with aluminium. The only limita attack and hence applications are restricted to tion of ain is that it reacts with H,0 in a hydrolysis selected areas such as the electronics industry. If the reaction to form various hydroxide phases, liberating AIN evolved in the process can be made water monia and heat. As a result, mechanical properties resistant like the recently developed AIn powders deteriorate to a large extent (chemically altered composition) AIN/Al compos- The mechanical properties of oxide and nitride base ites can be used successfully in various applications omposites are summarized in Tables 2 and 3. Average flexural strength(280-400 MPa)and fracture toughness values(3-9 MPa vm) of both composites are compa- The recently developed direced melt reaction processes rable even when filler materials are used.,. However, (oxidation and nitridation) are flexible and attractive in- properties are supcrior when fibre reinforcements are situ methods for developing metal-ceramic composites utilized.Further,these properties of fibre reinforced Composite formation in aluminium alloys is strongly composites are dependent on chemistry, strength of the dependent on composition, processing temperature, fibre and its interfacial interaction time and partial pressure of the gaseous species. For 160 Materials Design Volume 16 Number 3 1995
Directed melt oxidation and nitr~dation in ~iu~ini~~ alloys: B. S. S. Daniel and V. S. R. bulky Table 3 Mechanical properties of AlNlAl compositef~2s~37~39 Material Flexural Fracture Tensile Young’s Shear Poisson’s Thermal Thermal strength toughness strength modulus modulus Knoop ratio NW expansion (MPat/m) (MPa) conductivity hardness KW (X lodK-‘) (W m-‘K-‘) (kg mm*) Sintered AIN 28CL3.50 3.3 1200 AlNiAi 366-400 9.5-11.1 186240 223-i 124 AINJAINIAI 340 5.9 240 8-11 -100 TiBdAIN/AI 540 13.5 240 Al~O~~AlN/Al 340 9.5 242-451 165-180 65-72 0.26-&54 286506 Al,O,~AlN/Al 275450’ 8Flb 523+554’ 10.74%!.1’ SiCjAIN/Af 9.2-9.4 “Naturally aged bWith 0.4 pm reaction layer With 1.5 nm reaction layer As the liquid wicks through the preform, occasionally the reaction front is blocked and results in macroporosity, whereas local solidification shrinkage of metal in the composite gives rise to microporosity. These porosities are claimed to increase with temperature, magnesium content in the alloy and reinforcing particle size34. Porosities can be controlled to a limited extent either by isothermal holding between the oxidation temperature and the liquidus of the alloy or by altering the alloy composition22. Finally, there were attempts to model the complex infiltration process 35-37 . However, so far there are no realistic models to estimate the growth rates in any of these systems. Mechanical properties Like many composite materials, the properties of directed melt in-situ reacted composites are dependent on the type and volume fraction of the reinforcing phase. The properties of reinforcing phase are listed in Tables 2 and 3. Although the mechanical properties of AIN and Al,O, are almost similar, AlN has a greater thermal conductivity {ten times higher) and lower thermal expansion coefficient as compared to AI,O,. AIN exhibits excellent thermal shock resistance and is stable up to -1000°C in an oxidizing atmosphere. Additionally, AIN has excellent corrosion resistance and good wettability with aluminium. The only limitation of AlN is that it reacts with Hz0 in a hydrolysis reaction to form various hydroxide phases, liberating ammonia and heat, As a result, mechanical properties deteriorate to a large extent3’. The mechanical properties of oxide and nitride base composites are summarized in Tables 2 and 3. Average flexural strength (280-400 MPa) and fracture toughness values (3-9 MPa y’m) of both composites are comparable even when filler materials are used37s39@. However, properties are superior when fibre reinforcements are utilized. Further, these properties of fibre reinforced composites are dependent on chemistry, strength of the fibre and its interfacial interaction4’. 160 Materials & Design Volume 16 Number 3 1995 Applications and limitations Keeping the advantages (simplicity, flexibility and cost) of the process in mind, a large number of patents have already been filed on these materials. Landini et ~1.~’ made valve inserts using SiC and A&O3 particulate reinforced Al,O,/Al composites. These components were successfully tested on a diesel engine. Similarly, components made from SIC particulate-reinforced A120j-Al composites (Alnax products, USA) have been used for hydrocyclone cone liners and apexes and the performance was superior compared to conventionally used parts4’. SIC (Nicalon) fibre reinforced Al,03/A1 composites are seen as future materials for gas turbine applications43. The other probable applications of oxide composites include wear parts for pumps, jet mill nozzles, automotive parts, etc3’. On the other hand, AlN/Al composites (with their superior thermal conductivity) are likely to find a place in the electronics packaging industry. The major limitations of these processes are: 1. It takes a long time to develop oxide-nitride reinforced composites and hence the process cannot be used for producing very large components. 2. Residual porosities (both micro and macro) are a limitation in achieving best mechanical performance. Due to these inherent defects, these materials are unlikely to find applications in critical areas. 3. AlN base composites are sensitive to moisture attack and hence applications are restricted to selected areas such as the electronics industry. If the AlN evolved in the process can be made waterresistant like the recently developed AlN powders (chemically altered composition)38 AlNiAl composites can be used successfully in various applications. Summary The recently developed direced melt reaction processes (oxidation and nitridation) are flexible and attractive inS&X methods for developing metal-ceramic composites. Composite formation in aluminium alloys is strongly dependent on composition, processing temperature, time and partial pressure of the gaseous species. For
Directed melt oxidation and nitridation in aluminium alloys: B. S. S. Daniel and V.S. R. Murthy self-sustainment of the process, addition of highly diffu kinetics of aluminium alloys nitridation. In Proceedings of sive elements(along with silicon)is essential. Compared Ceramic and Metal Matrix Composites, Ed. Mostarhaci Pergamon Press, New York, 1989, 302-311 to oxidation, nitridation is more sensitive to reaction 20 Scholz. H. and Greil. p nitridation reactions of molten AH-Mg temperature and partial pressure of the reactant gas Si)alloys, J. Mat. Sci. 1991. 26, 669-677 Further, in nitridation, a wide range of metal and 21 Scholz, H. and Greil, P. Synthesis of high purity aIN by nitrida- eramic combinations(MMC and CMC) is possible. 22 Manor, E. et al. Microstructure evolution of SiC/ALO/A! alloy While extensive studies have been carried out on t omposite formation via the DiMOX process(with and 1993,76,1771787 23 Breval. E. ef al. Structure of aluminum nitridealuminum and without filler materials), only limited research has been aluminium oxide/aluminum composites produced by the directed undertaken on nitridation. Although Al,/Al and oxidation of aluminum. J. Am. Ceram. Soc. 1993. 76, 1865 AIN/Al composites exhibit similar mechanical proper- 24 Antolin, S et al. Formation of Al; O metal composites ties, the moisture sensitivity of AIn restricts the appli directed oxidation of molten aluminum-magnesium-Si alloys: Part 1, Microstructural development. J. Am. Cerum cation of AIN/Al composite material 992.75.147-154 25 Aghajanian, M.K. et al. Fabrication of metal matrix composites by a pressureless infiltration technique. J. Mat. Sci. 1991. 26 Acknowledgements 447-454 The authors are grateful to the board of research in Nagelberg, A.S. ef al. Formation of Al,O metal com the directed oxidation of molten aluminum-magnesium Nuclear Sciences(BRNS)for the research grant no alloys: Part Il, Growth kinetics. J. Am. Ceram. So 34/4/93-G 455462 27 Murthy. V.S.R. and Rao, B S Microstructural development in the directed melt oxidized(DIMOX)Al-Mg-Si alloys. J Mater References Sct.1995、30.3091-309 I Clyne, T.W. and Withers, P J. An Introduction to Metal Matrix 8 Schmaizried. H. and Laqua, w. Multicomponent oxides in Everell, R. K, and Arsenault,R. J. Me(us uva composites: 29 Kagawa, Y et al. Directed nitridation of liquid aluminium alloy 3 K. K. Chawla, Composite materials: Science and Engineering Growth process and modeling. Ceram. Eng. Sci. Proc. 1993. 14 inger-Verlag, New York. 1987 4 Koczak, M.J. and Premkumar, M.K. Emerging technologies for 30 Rhee. S. K. Wetting of ceramics by liquid aluminium. J. Am u production of MMCs. J Metals 1993, 45 5 Premkumar, M. K and Chu, M. G. Synthesis of TiC particulates 31 Dhandapani, s. P. et ul. Growth and microstructure of Sic preforms. Acta Metall. Mater 1994, 42, 649-656 6 Koczak M ]. and Kumar, K S. US Patent No. 4808372(March 32 Claussen, N and Urquhart, A W Directed metal oxidation In Concise Encyclopedia of Advanced Ce ramic Materials. Ed. Brook 7 Newkirk, M.S. et al Preparation of L dixide d radia Press, Oxford, 1992, pp. 124-129 composites: Matrix formation by the ation of 33 Fareed, A.S. et al. Development of BN/SiC duplex fibre coatings molten metals. Ceram Eng Sci Proc 1987.8.87 8 Newkirk, M.S.et al. Formation of Lanxide ceramic composite 81-89 34 Muralidhar, H. R. et al. Growth of Al, 0 /Al composites from 9 Creber. D. K. er ul owth by nitridation of Al-Zn alloys, Personal communication aluminium alloys, Ceran Sci Proc.1988,9,975-982 35 Martins, G. P. et al. Modeling of infiltration kinetics for liquid process for the fabri- metal processing of composites. Metall. Trans. 1988, 19B, 95-101 cation of metal matrix composites. SAMPE Quarterly 1989, 20 36 Hilg, W.B. Melt infiltration approach to ceramic matrix mposites. J. Am. Ceram. Soc. 1988, 71, C96-C99 1 Nageiherg, A. s Growth kinetics of AL,O, /metal ites 37 Toy, C and Scott, w. D Ceramic-metal composite produced by from a complex aluminium alloy. Solid State Ionics 1989, 32/33, melt infiltration. J. Am, Ceram. Soc. 1990. 73 97-101 783-788 38 Groat, E. A and Mroz, T. J. Aquous slip casting of stabilized 12 Breval, E. et ul. Mici sTructure of AlOyAl composites made by AIN powders. J. Mat Sci. 1994, 73, 75-78 directed oxidation of aluminium. J. Am. Ceram. Soc. 1990, 73 69 Lewis Ill, D. In-situ reinforcement of metal-matrix composites 610-2614 In Metal Matrix Composites: Processing and interfaces, Eds 3 Salas, O et al Nucleation and growth of Al, O, metal composite and Arsenault. R. J. Academic Press. London by oxidation of aluminium alloys. J. Mat. Res. 1991, 6 1991,pp.12l-150 1964-198 40 Andersson, C A and Aghajanian, M K. The fracture toughen 14 Vlach, K. C. et al, a thermogravimetric study of the oxidative ing mechanism of ceramic composites containing adherent ductile of AlO Al alloy composites. J. Mat. Res. 1991, 6. metal phases. Ceram. Eng. Sci. Proc. 1988,9, 621-626 +1 Landini, D. J. and Lesher, H. D. Ceramic erg, A.S. Observation on the role of mg inserts. In Proceedings of Third international Symposium on directed oxidation of Al-Mg-Si alloys. J. Mat. Res. 1992, 7 Ceramic Materials and Components for Engines. Ed. tenney. V J,, Las Vagas, 1988, pp. 1536-1551 16 Xiao, P and Derby, B. Alumin 42 Weinstein, J and Rossing, B. Application of a new ceramic/metal by the directed oxidation of using magnesia as a Irface dopant.JAm. Ceram. 77.1761-1770 nents. In High Performance Composites for the 1990's, Eds Da 17 Xiao, P and Derby, B. Alumina/aluminum composites formed S.K., Ballard, C. P. and Marikar. f, The Minerals. Metals and by the directed oxidation of aluminum hydroxide Materials Society, UK, 1990 as a surface dopant. J. Am. Ceram Soc. 7-1776 43 Schrioky, G. H et al. Ceramic composites for gas turbine engines Engineering 1992. 3, 247-259 Aeroengine Congress and Exposition, Toronto, Canada.4-8 19 LeHuy. H. and Dallarie, S. Effects of Si and Mg dopants on the June 1989 Materials Design Volume 16 Number 3 1995 161
Dhcfed melt oxidation and n&id&ion in aluminium alloys: B. S. S. Daniel and V. S. R. Murthy self-sustainment of the process, addition of highly diffusive elements (along with silicon) is essential. Compared to oxidation, nitridation is more sensitive to reaction temperature and partial pressure of the reactant gas. Further, in nitridation~ a wide range of metal and ceramic combinations (MMC and CMC) is possible. While extensive studies have been carried out on the composite formation via the DIMOX process (with and without filler materials), only limited research has been undertaken on n~tridation. Although Al,OJAJ and AlN/Al composites exhibit similar mechanical properties, the moisture sensitivity of AIN restricts the application of AlNiAl composite material. Acknowledgements The authors are grateful to the Board of Research in Nuclear Sciences (BRNS) for the research grant no. 34/4/93-c. References 1 2 3 4 5 6 7 8 9 10 II 12 13 I4 15 16 I7 18 19 Clyne, T. W. and Withers, P. J. An Introdl~~tfun to Metuf Matrix C~~?~u.s~fe.s, Cambridge Universjty Press, Cambridge, 1993. Everett, R. K. and Arsenault, R. J. Metul Mutrix Composites: Pracessing und Inte$uces, Academic Press, New York, 199 I. K. K. Chawla, Cumposite muteriufs: Science und Engineering, Springer-Verlag, New York. 1987 Koczak, M. J. and Premkumar, M. K. Emerging technologies for the in-situ production of MMC’s. J. Metufs 1993, 45, 44-48 Premkumar. M. K. and Chu. M. G. Svnthesis of Tic narticulates and their segregation during solidification in in-situ processed AI-Tic composites. Metaff. Truns. 1993, 24A, 2358-2362. Koczak. M. J. and Kumar, K. S. US Patent No. 4808372 (March 1989) Newkirk, M. S. et ul. Preparation of Lanxide ceramic matrix composites: Matrix formation by the directed oxidation of molten metals. Cerum. ,!%g. Sci. Proc. 1987, 8, 879-885 Newkirk, M. S. et ul. Formation of Lanxide ceramic composite materials. J: Mat. Res. 1986, 1, 81-89 Creber, D. K. et ai. AIN composite growth by nitridation of aluminium alloys. Ceram. Eng. Sci. Proc. 1988, 9, 975-982 Aghajanian, M. K. et al. A new infiltration process for the fabrication of metal matrix composites. SAMPE Quarterly 1989, 20, 43-46 Nagelberg, A. S. Growth kinetics of Al,O,/metal composites from a complex aluminium alloy. S&i& State IO&S 1989, 32133, 783-788 Breval, E. er uI. Microstructure of Al,C$‘AI composites made by directed oxidation of aiuminium. J. Am. Ceram. Sot. 1990. 73. ~ 2610-2614 Salas, 0. et ul. Nucleation and growth of Al,O,imetal composites by oxidation of aluminium alloys. J. h&t. Res. 1991. 6, 1964-1981 Vlach, K. C. rr rrl., A thermogravimetric study of the oxidative growth of Al,O,IAl alloy composites. .I. &fur. Res. 1931, 6, 1982-1995 Nagelberg, A. S. Observation on the role of Mg and Si in the directed oxidation of Al-Mg-Si alloys. J. Mut. Res. 1992, 7, 265--268 Xiao, I’. and Derby. B. Alumin~aluminum composites formed by the directed oxidation of aluminum using magnesia as a surface dopant. J. &n. Cerum. Sot. 1994,77, 1761-1770 Xiao, P. and Derby, B. Alumina/aluminum composites formed by the directed oxidation of aluminum using sodium hydroxide as a surface dopaut. J. Am. Cerum. SOE. 1994, 77, 1771-1776 Jayaram, V. Ceramic composites by melt oxidation, Materials Engineering 1992,3,247-259 LeHuy. H. and Daltarie, S. Effects of Si and Mg dopants on the 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 3.5 36 37 38 39 40 41 42 43 kinetics of aluminium alloys nitridation. In Proceedings of Ceramic and Metaf Matrix Composites, Ed. Mostarhaci. H., Pergamon Press, New York, 1989, pp. 302-311 Scholz, H. and Greil, P. Nitridation reactions of molten Al-(Mg, Si) alloys, J. Mat. Sci. 1991. 26, 669-677 Scholz, H. and Greil, P. Synthesis of high purity AIN by nitridation of Li-doped Al-melt. J. Euro. Cerum. Sot. 1990,6, 237-242 Manor, E. et ul. Microstructure evolution of SiC/Al,O,IAI alloy composites produced by melt oxidation. J. Am. Cerum. Sot. 1993.76, 1777-1787 Breval, E. er til. Structure of aluminum nitride/aluminum and aluminium oxide/aluminum composites produced by the directed oxidation of alumjnnm. 3. Am. Cerum. SK 1993. 76. 1865-1868 Antolin, S. et al. Formation of AlzO,/metal composites by the directed oxidation of molten aluminum.-magnesium-silicon alloys: Part I, Microstructural development. J’. Am. C.&m. Sot. f992.75. 147-154 Aghajanian, M. K. et ul. Fabrication of metal matrix composites by a pressureless infiltration technique. J. Mut. Sci. 1991. 26, 447-454 Nagelberg, A. S. et ul. Formation of AI,OJmetal composites by the directed oxidation of molten aluminum-magnesium-silicon alloys: Part II, Growth kinetics. J. .4m. Cerum. Sot. 1992, 75. 455-462 Murthy. V. S. R. and Rao, B. S. Microstructural development in the directed melt oxidized (DIMOX) Al-Mg-Si alloys. J. Mater. sci. 1995.30. 3091-3097 Schmaizried. H. and Laqua, W. Multicomponent oxides in oxygen potential gradients. ~~iduff~n 01 ~etfffs 1981, IS, 339-323 Kagawa, Y. ef ul Directed nitridation of liquid aluminium alloy: Growth process and modeling. Ceram. Eng. Sci. Proc. 1993, 14. 776-78 I Rhee. S. K. Wetting of ceramics by liquid aluminium. J. &I. Ctrrm. SW. 1970, 53, 386-389 Dhandapani, S. P. et ul. Growth and microstructure of A&O,-Sic-Si(A1) composites prepared by reactive infiltration of SiC preforms. Acfu ~~~ffl~. Muter. 1994, 42, 649656 Claussen, N. and Urquhart, A. W. Directed metal oxidation. In Concise Erzcyclo~diff crf Advanced Ceramic Muteri~ls. Ed. Brook, R. J., Pergamon Press, Oxford, 1992, pp. 124-129 Fareed, A. S. et al. Development of BN/SiC duplex fibre coatings for fibre reinforced alumina matrix composites fabricated by direct melt oxidation. Cerum. Eng. Sci. Proc. 1993, 14, 794-801 Mural~dhar, H. R. et ul. Growth of AI,OJAl composites from Al-Zn alloys, Personal communication Martins, G. P. et ul. Modeling of infiltration kinetics for liquid metai processing of composites. ~e~u~~. Truns. 1988, 193,95-l0f Hillig, W. B. Melt infiltration approach to ceramic matrix composites. J. Am. Cerum. Sot. 1988. 71, C96-C99 Toy, C. and Scott, W. D. Ceramic-metal composite produced by melt infiltration. J. Am. Cerum. Sot. 1990, 73, 97-101 Groat, E. A. and Mroz, 1. J. Aquous slip casting of stabilized AlN powders. J. &fut. Sci. 1994,73, 75-78 Lewis 111, D. In-situ reinforcement of metal-matrix composites. In Metal Matrix Composites: Processing and Interfrrces. Eds Everett, R. K. and Arsenault. R. J. Academic Press, London, 1991, pp. 121-150 Andersson, C. A. and Aghajanian, M. K. The fracture toughening mechanism of ceramic composites containing adherent ductile metal phases. Cerum. Eng. Sci. Proc. 1988,9, 621-626 Landini, D. J. and Lesher, H. D. Ceramic composite valve seat inserts. In Proceedings of Third ~~tern~tio~ai Symp~sfum on Ceramic Muteriafs and Components for Engines, Ed, Tennery, V. J., Las Vagas, 1988, pp. 1536-1551 Weinstein, J. and Rossing, 3. Application of a new cerami~metal composite technology to form net shape wear resistant components. In High Per~orm~~~e Comp~.~ites for the 199&s, Eds Das, S. K., Ballard, C. P. and Mmrikar, F., The Minerals, Metals and Materials Society, UK, 1990, pp. 339-360 Schrioky, G. H. et ul. Ceramic composites for gas turbine engines via a new process, Paper presented at Gas Turbine and Aeroengine Congress and Exposition, Toronto, Canada. 4-8 June 1989 Materials & Design Volume I6 Number 3 1995 161