Review article: RE-SiAlONa¤b- composites. Formation, thermal stability, phase relationships, reaction densification

Artigo revisão: compósitos TR a¤b SiAlON. Formação, estabilidade térmica, relações de fases, densificação por reação

Review article: RE-SiAlON a¤b - composites. Formation, thermal stability, phase relationships, reaction densification

(Artigo revisão: Compósitos TR a¤b SiAlON. Formação, estabilidade térmica, relações de fases, densificação por reação)

V. A. Izhevskiy¹ , L. A. Gênova, J. C. Bressiani

Instituto de Pesquisas Energéticas e Nucleares, CNEN/SP

C.P. 11049, Pinheiros, 05422-970, S. Paulo, SP, Brazil.

Abstract

In view of the considerable progress that has been made over the last several years on the fundamental understanding of phase relationships, microstructural design, and tailoring of properties for specific applications of rare-earth doped SiAlONs, a clear review of current understanding of the basic regularities lying behind the processes that take place during sintering of SiAlONs is timely. Alternative secondary phase development, mechanism and full reversibility of the a«b transformation in relation with the phase assemblage evolution are elucidated. Reaction sintering of multicomponent SiAlONs is considered with regard of wetting behavior of silicate liquid phases formed on heating. Regularities of SiAlON’s behavior and stability are tentatively explained in terms of RE-element ionic radii and acid/base reaction principle.

Keywords: SiAlON, phase transformation, structure evolution, N-melilite, reaction sintering

Resumo

Em vista do considerável progresso obtido nos últimos anos quanto à relação entre fases, o design microestrutural e a otimização de propriedades para aplicações específicas de SiAlONs dopados com elementos de terras-raras, se faz necessário uma clara revisão dos mecanismos envolvidos nos processos que ocorrem durante a sinterização dos SiAlONs. O desenvolvimento de segundas fases alternativas, os mecanismos e a completa reversibilidade da transformação a«b, em relação ao conjunto das fases são elucidados. Sinterização por reação de SiAlONs multi-componentes é analisada em relação à molhabilidade da fase líquida a base de silicato formada no aquecimento. Propõe-se uma explicação para a regularidade e estabilidade do comportamento dos SiAlONs em termos do raio iônico dos elementos de terras-raras e do princípio de reação ácido/base.

Palavras-chave: SiAlON, transformação de fases, evolução estrutural, N-melilita, sinterização por reação

INTRODUCTION

SiAlONs is a general name for a large family of the so-called ceramic alloys based on silicon nitride. Initially, they were discovered in the early 70-ies [1-3] and were developed actively since. The peak of activities in SiAlON research belongs to the 80-ies in which time fully dense polycrystalline bodies were prepared by pressure-less sintering techniques. Profound understanding of basic regularities of the interrelation between the starting powder's properties, processing parameters and the properties of consolidated materials was achieved. The majority of the SiAlON phases were investigated and characterized. The latter applies to b-SiAlON, a-SiAlON, AlN polytypoid phases, O’-SiAlON, as well as to a variety of crystalline secondary phases, mostly silicates, alumosilicates, and oxynitrides. Nitrogen-rich glasses were also thoroughly investigated. Although the initial goal of creating a single-phase silicon nitride based ceramics without any intergranular amorphous phases due to the transient liquid phase sintering was not achieved, SiAlON ceramics became one of the commercially produced high-tech ceramic materials.

There are two SiAlON phases that are of interest as engineering ceramics, b -SiAlON and a-SiAlON, which are solid solutions based on b - and a -Si₃N₄ structural modifications, respectively. b -SiAlON is formed by simultaneous equivalent substitution of Al-O for Si-N and has most commonly been described by the formula Si_6-zAl_zO_zN_8-z. In this formula z can be varied continuously from zero to about 4.2 [4-6]. The homogeneity region of b -SiAlON extends along the Si₃N₄-(Al₂O₃-AlN) line. The unit cell contains two Si₃N₄ units. The a -SiAlON phase region has a fairly limited two-dimensional extension in the plane of the Me-Si-Al-O-N phase diagram, the b - SiAlON thus falling on one border of the plane mentioned. Of special interest are the Me-Si-Al-O-N systems where the a -SiAlON phase is stabilized by the added Me ion, where Me is Li, Mg, Ca, Y, and the rare-earth metals except La, Ce, Pr, and Eu [7-9]. The a -SiAlON unit cell content, comprising four Si₃N₄ units, can be given in the general formula Me_mSi_12-(pm+n)Al_(pm+n)O_nN_16-n for a metal ion Me^p+. Two substitution mechanisms may act; the first is similar to that of b -SiAlON with n (Si+N) being replaced by n (Al+O). The second mechanism is further replacement of pmSi⁴⁺ by pmAl³⁺, the electron balance being retained by incorporation of mMe^p+ into the a -phase structure.

a - SiAlON was most extensively studied in the Y-Si-Al-O-N system, where the two-dimensional phase field extension can be expressed as Y_xSi_3-(3x+n)Al_(3x+n)O_nN_4-n, with 0.08 < x < 0.17 and 0.13 < n < 0.31 [8, 10-13]. The lower solubility limit is located at x » 0.08 and 0.20 for cations with the radius of 0.095 to 0.1 nm, and significantly higher, x = 0.25, for small cations such as ytterbium with a radius of » 0,085 nm [14]. It was also shown that by changing the overall composition in the Y-SiAlON system, (a + b ) -SiAlON ceramics can be prepared with a varying a-SiAlON:b - SiAlON phase ratio [15, 16]. a -SiAlON always appears as equiaxed grains in the microstructure of the material, while b-SiAlON forms elongated grains with an aspect ratio of 4 to 10. Hardness increases markedly with increasing a -SiAlON phase content, whereas the fracture toughness decreases. High a -SiAlON content also improves oxidation resistance of the material and the thermal shock resistance. It is also evident from its chemical composition that a-SiAlON offers possibility of reducing the amount of residual glassy grain-boundary phase by incorporating constituents of sintering aids into the crystal structure, which otherwise will be present as substantial amounts of residual glassy phase and deteriorate the high temperature performance of the material. Thus, a -SiAlON enables the preparation of the practically glass-free SiAlON composites, which may become interesting candidates for high-temperature structural and engineering applications.

The possibility of varying the a -SiAlON:b -SiAlON phase ratio by slightly changing the overall composition was shown to open many possibilities to prepare Y-SiAlON ceramics with desired properties. This is of the utmost importance in connection with tailoring the properties for specific applications. Other ways of changing properties of SiAlON ceramics by means of replacing Y₂O₃ by other RE-oxide additives were also investigated [17]. Precisely this direction of research lead to discovery of the full reversibility of the a « b transformation for certain phase assemblies under conditions of post-sintering heat treatment [18-20]. It has been observed that the a -SiAlON phase is less stable at low temperatures and decomposes into rare-earth-rich intergranular phase(s) and b -SiAlON with a remarkably elongated crystal morphology. The extent of the transformation is more pronounced with increased heat-treatment temperature. The mechanism of the fully reversible a « b transformation was investigated by the same authors [21]. This transformation provides an excellent possibility for optimizing phase content and microstructure without further additions of oxides and nitrides merely by heat treatment at appropriately chosen temperatures. In this way premeditated values of hardness, strength and toughness can be achieved from a single starting composition, which opens new perspectives for their successful application.

The objective of the present review, therefore, is to provide a comprehensive overall view of the new developments in SiAlONs that will facilitate advances of this subject, and allow areas for further study to be more clearly identified. The review will cover such important issues as thermal stability of different RE-doped (a -b )-SiAlON composites, phase relationships in this systems, alternative secondary phases for SiAlON ceramics, and mechanisms of reversible a «b transformation.

Occurrence of nitrogen melilite and nitrogen woehlerite solid solutions containing aluminum

Nitrogen melilite and woehlerite phases have been recently closely re-evaluated as the candidates for alternative secondary phases in

Si₃N₄ based ceramics after the Al-containing solid solutions of this phases were discovered and investigated [22-25]. Nitrogen-containing (or

N-) melilites and nitrogen-containing (or

N-) woehlerites, the latter also known as

J-phase, were discovered relatively early in the course of silicon nitride development and at the time were considered as undesirable secondary phases due to their poor oxidation resistance at temperature between 900 and 1100

⁰C [26]. Under such conditions they oxidize rapidly to crystobalite and

y-RE₂Si₂O₇ with a ~ 30% increase in specific volume, and this results in catastrophic cracking and total failure of material. N-melilites have the general formula

R₂Si₃O₃N₄ (where

R= yttrium or rare earth metal) and occur as an intermediate phase during sintering of a number of

Si₃N₄-based ceramics. In particular,

N-melilite is located in the phase compatibility regions of a -(a - b) -

SiAlONs, which in theory can produce a good microstructure of SiAlONs N-melilite as the only secondary phase.

N-melilite

(M(R)), has a tetragonal structure that is built up of an infinite linkage of corner-sharing

Si (O,N)₄ units in tetrahedral sheets perpendicular to the [001] direction. Sandwiched between these sheets are

Y³⁺ or rare earth

Ln³⁺ ions which hold the silicon oxynitride layers together [27]. The structure of

N-woehlerite, or

J-phase, is monoclinic and contains either

Si₂O₅N₂ or

Si₂O₆N ditetrahedral units containing

N in the bridging position and arranged lengthwise along the

a-axis. These units are linked by

RE polyhedra. Both phases are very stable refractory compounds and occur in almost all rare-earth systems [27]. With a very high melting point (~ 1900

⁰C for yttrium

N-melilite [28]) and a composition close to a nitrogen-rich limit of the liquid region, these phases fulfill some of the ideal requirements for a grain boundary phase in

SiAlON ceramics. However, the most important distinction of melilite-woehlerite type of secondary phases is that they do not have a eutectic with the matrix

Si₃N₄ while the traditional oxide-base secondary phases have one unconditionally. This becomes absolutely obvious if one considers the phase diagram of

R-Si-(Al/O)-N systems. A liquid region exists at high temperature, separating the Si

₃N

₄ corner and the plane of additive oxides and their compounds. In contrast, there is no liquid region between

M(R)/J(R) and

Si₃N₄. Although some liquid does form during initial heating because of the ternary eutectic of

SiO₂,

Al₂O

₃, and

R₂O₃, the liquid should be consumed largely by

Si₃N₄ to form

M(R) or

J(R). Thus, sintering of the latter composition is relatively difficult at the intermediate temperature for lack of liquid, but it improves drastically once the melting point of these phases is exceeded. Indeed, the lack of liquid at the intermediate temperature could be advantageous because it reduces the possibility of melt evaporation and massive powder relocation. Additionally, both melilite and woehlerite type phases crystallize readily on cooling directly from the liquid, which leads to quicker and much more complete crystallization as compared to any devitrification process. The problem of catastrophic oxidation at intermediate temperatures is solved by formation of melilite and woehlerite based solid solutions containing aluminum. The general formulas of these solid solutions are

R₂Si_3-xAl_xO_3+xN_4-x and

R₄Si_2-xAl_xO_7+xN_2-x, respectively. The solid solutions will be further referred to as

M'(R) and

J'(R). In these phases,

Si-N bonds are progressively being replaced by

Al-O bonds as the

Al solubility in the structure increases and this increase in oxygen content improves the oxidation stability.

Until now the regularities of solid solubility of aluminum in M(R) and J(R) were studied for the majority of RE elements (Nd, Sm, Gd, Dy, Yb, and Y) that are used in SiAlON production. It was unequivocally shown that the solid solubility range of M’(R) phases in respect of Al are determined by ionic radius of the RE element. With decreasing ionic radius the solubility limits in melilite solid solution decrease. For elements with large ionic radii, i. e., R= Nd and Sm, up to one Si can be replaced by Al. Consequently, yttrium melilite has the lowest observed aluminum solubility (x= 0.6) in rare earth elements melilite solid solutions [29]. The N-melilite phase is not stabilized by Yb and La [30-32], and therefore no study on these M’(R) was conducted. It was, however, shown that Yb forms J’(R) solid solution that can serve as an alternative secondary phase in this system. On the other hand, La forms only the M’ phase with the highest substitution level (x= 1) while the M’ (La) homogeneity range does not exist in La-Si-Al-O-N system. Here it must be specifically emphasized that phase pure M(R) or M’(R) were never formed neither

in the form of a secondary phase, nor if synthesized separately according to stoichiometry. Appreciable amounts of

J’(R),

REAG or other phases were present depending on the

RE oxide used, the amount of

J’(R) increasing with the decrease of the

RE ionic radius. The results of the

M’(R) solubility range determination systems are summarized in

Table 1 [33].

The data on the correlation between the x value in M’(R) solid solution formula and the lattice parameters of M’(R) were determined for the majority of RE stabilizing elements [22, 29, 33]. Linear increase of the lattice parameters with increasing of the x values was observed for all systems investigated up to a certain x value for respective M’(R) phases. However, these x values do not necessarily represent true solubility limits as all samples contained additional phases, as it was mentioned earlier, some of them rare earth rich. It can, however, be noted that the c/a ratios are almost constant with increasing x-value for each M’(R) phase which indicates that the structural changes introduced when Si-N pairs are replaced by Al-O ones are isotropic in all M’(R) phases studied.

Since the nominal x value, i. e., the one determined by the powder mixture composition, was not relevant due to the multiphase nature of the synthesized materials, SEM-EDS studies have been used to estimate solubility limits of Al in various M’(R) phases. The maximum Al content, as determined in [29], were x=0.98, 0.9, 0.78, 0.71, and 0.57 for R= Nd, Sm, Gd, Dy, and Y, respectively. Thus, in M’ (Nd) and M’(Sm) systems one Si-N pair per formula unit can be replaced by one Al-O pair in accordance with previous findings [22, 34, 35] while the true solubility limits in the M’ (Gd) and M’ (Dy) systems seem to be somewhat less (0.75) and even smaller in the M’ (Y) system (0.60). These findings also confirm that the limit of Al solubility decreases with the ionic radii of the RE element [36]. It was also shown, that the substitution of Si-N by Al-O in the M(R) structure produces a lattice expansion [29]. The fact that the maximum value of lattice expansion was found in the M’ (Nd) system (1.66%) and the minimum one in the M’ (Y) system (0.70%) further supports the conclusion that the investigated systems exhibit different solid solubility ranges.

A tentative attempt to rationalize the interrelation between the ionic radii of RE element and the solubility limits of different M’(R) phases was made in [33] and is based on the simple considerations regarding the structure of melilite-type phases. The projection of the yttrium melilite on (100) may be schematically represented as shown in Fig. 1. With non-bridging anions taking up positions on the apex of the tetrahedra each silicon oxynitride layer of the N-melilite structure possesses an overall negative charge that has to be balanced by the RE cations. These cations separate the silicon oxynitride layers by a finite distance. In systems with small cation sizes such as yttrium or dysprosium, the negatively charged layers are allowed to come relatively closer together and possibly create a repulsive force which can destabilize the melilite structure. A structurally unstable M’(R) phase may result in a low Al accommodation capability. However, the exact mechanism involved in this process is currently unknown.

M'(R) phases is their behavior at high temperatures primarily in relation with densification of respective compositions.

Al-free melilites tend to become unstable at temperatures exceeding 1800

⁰C. Melting of melilites is thought to be impossible because the edge of the liquid region in the four component

R-Si-O-N plane does not reach as far as the melilite composition (i. e., 67 eq.% nitrogen). Instead, the melilite phases are believed to decompose with significant loss of nitrogen, resulting in the formation of

J phase. For

M'(R), stability at high temperature is determined by two factors. With increasing aluminum content, the limiting composition of

M'(R) gradually moves towards the liquid region, and at the same time the liquid region can accommodate a higher nitrogen concentration with increasing temperature, approaching the terminal composition of

M'(R). Current preparative evidence suggests that the liquid region eventually reaches the

M'(R) composition, and

M'(R) melts [22]. This generates large amounts of liquid which, in combination with

Si,

N, and other components, promotes the precipitation of a -

SiAlON phase and accelerates the final densification. This effect has been widely observed in several

SiAlON systems [34, 37].

It is clear that the melting temperature of M’(R) varies in different systems according to the densifying cation, but a low melting temperature for M’(R) is a key factor in the preparation of dense

a - and (a + b)

- SiAlON ceramics. According to [22] the

M’(Sm) melts at about cooling 1700

⁰C and on cooling at normal furnace cooling rate

M’(Sm) reprecipitates, suggesting that any liquid with

M’(R) composition does not form a glass. Although it has not yet been confirmed that dissolution of aluminum in

M(R) actually lowers the melting temperature of these phases, but the closeness of the

M’(R) composition to the liquid-forming region certainly dominates the behavior of

M’(R) phases at high temperatures. As it will be discussed further, the high temperature behavior of

M’(R) phases plays an important role in the reversible a « b transformation in

SiAlON ceramics. Another important factor in designing and processing of various silicon nitride based materials,

(a

- b)

- SiAlONs in particular, are the phase relations of

M'(R) with the neighboring phases in different

M-Si-Al-O-N systems. The most thorough and systematic research of this subject was accomplished in [38, 39]. This research covered the majority of

RE-metal oxides (

RE=

La,

Gd,

Dy,

Er,

Y,

Nd,

Sm, and

Yb) that are used as sintering aids in complex

SiAlON systems. Firstly, it was shown that the nominal compositions of

N-melilite do not form a single-phase material for neither of the RE elements investigated. Moreover, it was unequivocally shown that

M(R) is not formed with

La, which, depending on the synthesis conditions, forms a mixture of

K-phase (

La₂O_3× Si₂N₂O) and

1:2 phase (

La₂O_3× 2Si₃N₄) in this or that proportion. The results of the attempts of

M(R) synthesis with different RE elements are listed in

Table 2.

Si₂N₂O, which enters the compositions of

J (

2R₂O_3× Si₂N₂O) and

K (

R₂O_3× Si₂N₂O) phases, formed by different

RE elements as "impurity" or additional phases (see

Table 2), and light rare earth oxides as more basic than heavy rare earth oxides. The various minor phases formed in the presence of melilite then can be viewed as the product of the following reactions:

1:2 phase La₂O_3× 2Si₃N₄ (strong base / strong acid)

K(R) phase La₂O_3× Si₂N₂O (strong base / weak acid)

J(R) phase 2Y₂O_3× Si₂N₂O (weak base / weak acid)

For example, as the basicity of

R₂O₃ decreases with decreasing ionic radius, it prefers to react with a weak acid to form J phase rather than

K or

1:2 phase. It is also interesting to note that the molar ratio of base to acid in

K(R) and

J(R) increases from 1:1 to 2:1; this reflects the need for a larger amount of the weaker base to react with the same acid.

Interesting results and trends were discovered for M’(R) phases when an attempt was made to synthesize the nominal composition

R₂Si₂AlO₄N₃.

M’(R) solid solution was found in all cases except ytterbium (

Table 3). Overall, the systematic trend of phase distribution discovered was very similar with the one for

M(R), with the minor phases favoring

K-phase, in the case of lanthanum, and

J-phase, in the case of heavy rare-earth elements. However, some differences were also discovered.

First, in the case of La, melilite solid solution was formed. This is in contrast with Al-free composition, which did not form La-melilite. The lattice parameters of M’ (La) were larger than all of

the other

M(R) and

M’(R) found so far (

a = 7.855

_Å and

c = 5.120

_Å ). The composition of

M’ (La) was found to be

La_1.98Si_1.82Al_0.96O_4.21N_3.08, i. e., essentially an

x = 1 compound in the

R₂Si_3-xAl_xO_3+xN_4-x series. Similar compositional investigations revealed that the x value decreases monotonically with the atomic number of the

RE-element. This can be interpreted as a size effect. From the interrelation between the ionic radii of

RE elements and both

x value and (cubic root of) unit cell volume,

(a

²c)

^1/3, from

M(R) and

M’(R) (

Fig. 2) it is clear that the unit cell has expanded in

M’(R) relative to

M(R). This is accompanied by the substitution of

Si-O bond (bond length 1.62

_Å) and

Si-N bond (1.74

_Å) by

Al-O bond (1.75

_Å ). Apparently, in the case of lanthanum, which has a very large radius, only after maximal substitution,

x = 1, is it possible to incorporate lanthanum into the melilite structure. In case of ytterbium, which has a very small radius in contrast, it is apparently stable only in the limit of smallest volume realizable at

x = 0.

The lattice parameters of melilite and its solid solution obtained for different rare-earth elements at different x values can be correlated to each other, according to [39], by the following formula:

where r is the ionic radius (in angstroms) of the rare-earth ion using Ahrens scale. These formulas were obtained by a regression analysis of the experimental data on XRD-determined lattice parameters of M(R) and M’(R) formed by different RE-elements and can be used for identifying the composition of

R₂Si_3-xAl_xO_3+xN_4-x once the identity of the rare-earth element is known. As to the two other phases,

K and

J, that were present in the synthesized samples the mechanism of formation and the trends of stability were also formulated in [39]. Formation of

K phase was attributed to the reaction of

R₂O₃ with the surface-oxidized

Si₃N₄, forming

Si₂N₂O. In the case of the aluminum-containing compositions, the oxidized nitride was supposed to be incorporated into

M’(R) solid solution thus preventing the formation of the

K phase. For

J phase, formation of

J’(R) solid solutions was confirmed for a number of

RE elements. Moreover, it was shown that the stability of both

J(R) and

J’(R) increases with the decrease of ionic radii of

RE elements. The

x value for

J’(R) was shown to increase with the ionic radius of the rare-earth element, which is similar to the trend found for

M’(R). Recognizing the systematic trends in

M’(R),

J’(R), a’, and b’ compositions, which vary with different rare-earth elements, a series of tentative diagrams were suggested in [39] to delineate the phase relationships between melilite and neighboring phases. The rare-earth aluminoxynitride,

R₂AlO₃N, denoted as 1:1, was included into consideration, based on the findings of [41]. Four prototypical diagrams, presented in

Fig. 3-6, illustrate these relationships:

(1)

For lanthanum, there is no a

’ solid solution; the

x = 1 compound of

M’(R) forms. In addition,

La₂Si₆O₃N₈ (1:2 phase) exists as a neighboring phase (see

Fig. 3);

(2)

For neodymium and samarium, there is an extensive solid solution of both a

’ and

M’(R), but only

J phase, not

J’(R) solid solution, is compatible with

M’(R). In addition,

R₂AlO₃N forms and is compatible with

M’(R) and

J (see

Fig. 4). The compatibility of

RAlO₃ (

R =

Nd,

Sm) with

M’(R) were not resolved. According to [39],

M’(R) is compatible with

RAlO₃, while according to [42]

SmAlO₃ may be just a product that forms on cooling at some specific conditions and, in fact, is not compatible with

M’(R). In [39] only coexistence of

M’(R),

R₂AlO₃N and

J phase at 1550

⁰Cwas found.

(3)

For gadolinium, dysprosium, europium, and yttrium, there was found an extensive solid solution of a

’, but the solid solution range of

M’(R) was shown to be compatible with

M’(R), while the compound

R₂AlO₃N failed to form in this case (see

Fig. 5).

(4)

Ytterbium was shown to enter a

’ solid solution and

J’(R) solid solution. However, only

M(R), not

M’(R), was proven to form in this case (see

Fig. 6).

Subsolidus phase relationships in

R

₂O

₃ – Si

₃N

₄ – AlN – Al

₂O

₃ systems As it was already mentioned before, the importance of using rare earth oxides for the densification of silicon nitride ceramics was recognized in recent years. Not only are they very effective along with alumina for densification, either singly or in combination with yttria, but they can also be accommodated in the a -

Si₃N₄ lattice forming a-

SiAlON (a-

Si₃N₄ solid solution with the composition

Ln_xSi_12-(m+n)Al_m+nOnN_16-n , also designated as a’), thus providing the opportunity for decreasing the transient liquid phase content after sintering and hence reducing the amount of residual grain boundary glass. Therefore, phase relationships in

Ln-Si-Al-O-N systems are of particular interest. The most well investigated system is

Y-Si-Al-O-N system. Most of the early work in this system has been restricted to the region bounded by

Si₃N₄ -b-

SiAlON- Al₂O₃ -

SiO₂and

Y₂O₃, which does not include the solid solution a’

- SiAlON.

Huang at al. reported the a-

SiAlON formation in the system

Si₃N₄-AlN-Y₂O₃[43] and

Si₃N₄ -

AlN- rare earth oxide [14]. The systems

Si₃N₄ -

AlN- Y₂O₃and

Si₃N₄-AlN:3AlN-Al₂O₃ : AlN have also been studied by the same authors [44]. Later, with the emergence of a’

- SiAlON, information on the nitrogen-rich part of the

Si, Al, Y/N, O became necessary and the investigations in this system were completed [45]. Thirty-nine- compatibility tetrahedra had been established in the region

Si₃N₄-

AlN- Al₂O_3- Y₂O₃ (

Table 4). The subsolidus phase relationships in the region

Si₃N₄-AlN-YN-Y₂O₃ were also determined. Only one compound,

2YN:Si₃N₄ , was confirmed in the binary system

Si₃N₄-YN contrary to the early results of Thompson [46] who reported the existence of three compounds:

6YN:Si₃N_4, 2YN:Si₃N₄ , and

YN:Si₃N₄. This contradiction could be explained by the purity of the

YN used as a starting material, and in that case the X-ray diffraction lines of

2YN: Si₃N₄ and

YN: Si₃N₄.reported by Thompson may be regarded as a mixture of melilite (

Y₂Si₃O₃N₄),

J phase (

2Y₂O₃: Si₂N₂O), and other oxygen-containing phases. The solubility limits of the a-

SiAlON on the

Si₃N₄-YN:3AlN join were determined to range from m= 1.3 to m= 2.4 in the formula Y

_m/3Si

_12-mAl

_mN

₁₆. No quinary compounds were found. Seven compatibility tetrahedra were established in the region

Si₃N₄-AlN-YN-Y₂O₃ (

Table 4).

Sixty eight compatibility tetrahedra were established in the system Si,Al,Y/N,O : thirty-nine in the region bounded by Si₃N₄-SiO₂-AlN- Al₂O₃-Y₂O₃ , seven in the region bounded by Si₃N₄-AlN- YN- Y₂O₃ , and twenty-two in region Si₃N₄-b₆₀-Al₂O₃-SiO₂ and Y₂O₃. All this data are graphically presented in Fig. 7-11.

Subsolidus phase relationship in Ln₂O₃- Si₃N₄- AlN-Al₂O₃ systems until now were thoroughly investigated for most of RE elements, although not in such a detailed and profound way as the Y- doped one. The determination of the whole system, represents a very large amount of work and most initial studies have restricted to the sub-systems Ln-Si-O-N, Ln-Al-O-N (Ln = Ce, Pr, Nd and Sm) and some planes involving a - and b - SiAlON [14, 34]. It is known that elements in the lanthanide (Ln) series are similar to yttrium in compound formation and phase relationships. Previous research on Ln-Si-O-N subsystems [47], indicates that the phase relationships are nearly the same as in the Y-Si-O-N system, but in Ln-Al-O-N systems [48] the phase relationships were shown to vary with atomic number of the rare earth element. The phase relationships in the systems with high Z-value of the rare earth element are similar to those in Y-Al-O-N [49]. In the systems Ln-Al-O-N (where Ln = Ce® Sm) to nitrogen containing compounds exist – Ln₂AlO₃N and Ln₁₂Al₁₂O₁₈N₂(magneto-plumbite, MP compound) – and no garnet phase occurs; instead the 1:1 aluminate phase becomes stable [48]. The Ln₂AlO₃N and MP compounds do not occur in those systems containing rare earth oxides with cations smaller than that of Gd [14, 48]. Therefore, in the five-component systems with the low Z-value rare earth elements, the phase relationships become slightly different from those in Y-Si-Al-O-N. In the oxygen-rich region where two five-component phases [50, 51] – U-phase ( Ln₃Si_3-xAl_3+xO_12+xN_2-x ) and W-phase ( Ln₄Si₉ Al₅O₁₀N) – exist, the former is stable for rare earth cations between La and Dy the latter stable only in La, Ce and Nd systems. None of the phases above occur in the Y system. The stability of N-melilite (Ln₂Si_3-xAl_xO_3+xN_4-x) in different are earth systems was discussed in the previous section of this paper.

Most closely and most recently the subsolidus phase relationships were studied in Nd and Sm [52], and Dy [53] containing Ln₂O₃-Si₃N₄-AlN- Al₂O₃ systems. The attention paid to the former two elements is caused by the efforts to substitute Y by some cheaper RE elements in Si₃N₄-based ceramics without hindering the properties. Dy containing system was chosen because dysprosium is one of the central elements in the rare earth series: in this region, phase relationships in Ln-Si-Al-O-N systems are changing from those observed for the low-Z elements (La, Nd, Sm) to those observed for high-Z elements (Er, Yb... and Y). It was therefore interesting to find out whether these changes occur simultaneously or whether there is overlap in this region between the different phase relationships observed for the high-Z and low-Z rare earths.

According to [52], Sm has the same behavior as Nd in regard of phase formation, but it was noted that melting temperatures were lower in the samarium system and consequently, at a given temperature, the size of the liquid-forming region is larger. For both Sm and Nd systems 44 compatibility tetrahedra were established (Table 5). Some specific features of the systems investigated as compared to the Y-Si-Al-O-N system were discovered. Unlike the Y-containing system where J-phase or woehlerite (Y₄Si₂O₇N₂) forms a continuous solid solution with YAM (Y₄Al₂O₉) neither in Sm- nor in Nd- containing system the formation of such solid solutions was observed, although the J-phase itself was readily formed. The N-melilite solid solution M’(R) was found to be compatible with b-SiAlON phase only in the range b₀-b₁₀ . Moreover, according to the XRD data Nd- and Sm- melilite (Ln₂O_3- Si₃N₄) seems to have priority ever b-Si₃N₄ in accommodating Al and O (thereby forming M’(R) phase) and therefore the tie lines between melilite and b-SiAlON do not run parallel. The tie line between b-Si₃N₄ and SmAlO₃ was proven not to exist on account of the results of heat treatment of a number of relevant compositions at 1350 and 1550 ⁰C that resulted in formation of M’ and apatite (Sm₁₀(SiO₄)₆N₂) only. Graphically these results are presented in Fig. 12-16.

In the Dy-Si-Al-O-N system the region defined by the four end-members Si₃N₄, AlN, Al₂O₃ and Dy₂O₃ , that includes the phases a -SiAlON, b -SiAlON and AlN polytypoids, and covers the range of compositions used for the design and processing of commercial multi-phase sialon ceramics was investigated [53]. The binary tie-lines and 42 compatibility tetrahedra were established based on experimental results and on the more extensive knowledge of phase relationships in the Y-Si-Al-O-N and Sm(Nd)-Si-Al-O-N systems (Table 6).

The DyAG (Dy₃Al₅O₁₂) is stable phase and is compatible with all the AlN polytypoid phases, b-SiAlON (from b₁₀ to b₆₀) and a-SiAlON (oxygen-rich) forming twelve Al–polytypoids- containing compatibility tetrahedra (Fig. 17) and one a’-b’-containing compatibility tetrahedron, a’-b₁₀-DyAG-12H (Fig. 18) unlike the Sm(Nd)-containing systems where the YAG does not occur. The melilite solid solution M’ is compatible with both b - and a - SiAlON phases, and with DyAG forming two compatibility tetrahedra: a’- b₀-b₁₀-M’ and a’- b₁₀-M’- DyAG (Fig. 19). Moreover, in the Dy-containing system the J-phase (Dy₄Si₂O₇N₂) forms a continuous solid solution with Dy₄Al₂O₉ unlike the Sm(Nd)-containing ones and similar with Y and Yb [25, 49, 54]. It was also shown that the compound DyAlO₃ is compatible only with M’ (melilite solid solution) and Dy-rich terminals of AlN polytypoids (from AlN to 21R), and therefor the tie line between LnAlO₃ and b’ that occurs in the Nd(Sm)-Si-Al-O-N systems does not exist in the Dy-Si-Al-O-N system. All the results are summarized in Fig. 20.

Subsolidus phase relationships for other RE-elements, although not studied in such detail as the ones presented above, could be easily derived by analogy if the RE-ionic radius is taken into account appropriately.

Thermal stability of RE-a-SiAlONs and the reversibility of a « b SiAlON transformation

The oxide densification additives used for sintering of SiAlON ceramics are generally insoluble in the b-SiAlON structure and even though most of them are a-SiAlON formers, some additive remains as a residual M-Si-Al-O-N intergranular gassy phase, which degrades the mechanical and chemical properties of the material above the glass-softening temperature, which normally is about 900-1100 ⁰C.

Post-sintering heat treatment is one of the accepted methods the main objective of which includes eliminating or minimizing the grain boundary glass in nitrogen ceramics by converting it into refractory crystalline phase(s) thus improving the materials performance at elevated temperatures. This is usually achieved via a conventional glass-ceramic process during which the glass devitrifies at a temperature above its Tg point but without being melted [51, 55]. There has been much work in understanding of the grain boundary crystallization, but very little attention has been paid to the effects of the post-sintering heat treatment on the stability of SiAlON phases, while the characterization of the finished product has always focused on the account of the residual glass and the type, amount and microstructure of the resulting crystalline oxynitride phase. However, in the course of the study of densification and post sintering heat treatment of (a-b)-SiAlON compositions containing selected rare earth oxide additives (Yb₂O₃, Dy₂O₃, Sm₂O₃ either singly or in combination with Y₂O₃) Mandal et al. [18] reported that the transformation from a’ to b’ readily occurred in some rare earth a’ + b’ composite materials when heat treated between 1000 and 1600 ⁰C and that extent of the transformation was more pronounced with increasing temperatures.

a’ and b’ are two different crystallographic modifications of SiAlON structures [56]. Although both phases are basically built up of corner sharing (Si, Al)(O, N)₄ tetrahedra, there is a distinct difference in the atomic arrangement between the two phases [57]. The transformation between a’ and b’ phases has a reconstructive nature, which involves the breaking of chemical bounds and substantial atomic diffusion, and hence requires significant amounts of thermal energy. As a result of the strong covalent nature of the bounding associated with both the a’ and b’ structures, the atomic diffusivity of the species making up the lattices is inherently low. It is, therefore, generally assumed that the a’ « b’ phase transformation requires a liquid phase to assist the necessary atomic diffusion by analogy with the transformation from a-Si₃N₄ to b-Si₃N₄, which did not occur without the presence of a liquid phase in which a-Si₃N₄ grains could dissolve above ~1400 ⁰C [58]. However, the actual role of liquid phase in a’ to b’ transformation has not been fully understood.

According to the observations of [18], the fast cooling (cooling rate of ~ 450 ⁰C/min in the interval 1775-900 ⁰C) of the RE-SiAlON samples performed in order to preserve the intergranular phase in a fully amorphous condition led to a marked increase of the a’-phase content. Thus the original samples designated b-SiAlON being reheated to the sintering temperature and rapidly cooled contained some a’-SiAlON, and the original samples designated (a-b)-SiAlON were almost pure a-SiAlON. The effect was more marked in the case of samples densified with higher atomic number rare earth oxides. Subsequent annealing of these samples at various temperatures in the range 1000-1800 ⁰C yielded the a’ to b’ transformation accompanied by crystallization of various secondary phases. Both the extent of the a’® b’ transformation and the resulting secondary phase assembly were obviously dependent on the RE oxide used, specifically the ionic radius of the RE element. The degree of the a’® b’ transformation was again more pronounced for the higher atomic number rare earth oxide sintering additives.

The dependence of the phase assemblages after heat treatment on the rare earth oxide used is presented in Fig. 21. Therefore, it was concluded that the a-SiAlON phase is only stable at high temperatures (typically in excess of 1550 ⁰C) and that it transforms to b-SiAlON plus other Ln-rich grain boundary phases at lower temperatures.

Further experiments showed that the amount of grain boundary liquid phase in the sample had marked effects on the transformation, and the introduction of an additional amount of glass into a previously stable a’ composition could significantly destabilize the a’ phase and promote the a’ ® b’ transformation [21]. After comparing the results in different rare earth SiAlON systems, it was further suggested by the same authors that the residual grain boundary liquid phase was one of the most important factors influencing the transformation. The rate of the transformation was stated to be dependent on the amount and viscosity of the intergranular liquid phase present during the heat treatment.

Shen et al. [59] on the other hand, proposed two possible transformation routs, one involving a liquid phase, i.e. a’+ liquid ®b’+M’ and other being a direct decomposition of the a’ phase to form b’ and the aluminum-containing melilite (Sm₂Si_3-xAl_xO_3+xN_4-x , M’(R)) phases in the Sm (a- b)- SiAlON system, without the involvement of a liquid phase. However, a more in-depth investigation of the thermal decomposition of samarium and ytterbium stabilized a-SiAlON phases revealed that some self-generated liquid phase is involved in the process of a’® b’ transformation [60]. Unique microstructural features resulting from the a- to b-SiAlON transformation have been observed in Sm and Yb (a- b)-SiAlON samples. The newly formed b’-grains, namely the ones formed as a result of a’ ® b’ transformation, contain a high density of dislocations and a large amount of ultra-fine (~40 nm) spherical inclusions which are often associated with dislocations. The conclusion that such microstructure is a unique feature associated with the a’ to b’ transformation was based on careful structural analyses by several methods. It was also shown that the transformed grains grow as a result of the epitaxial heterogeneous nucleation on the pre-existing b’-phase.

According to [61], the a’ ® b’ transformation occurs in two stages. The first stage involves relatively high amounts of intergranular glassy phase, which provides the liquid phase necessary for transformation since the post-sintering heat treatment is conducted at the temperatures exceeding the melting point of the glass. At this stage the transformation occurs according to the mechanism described in [59]. After the initial period of heat treatment, the volume percentage of the residual grain boundary glass is reduced dramatically; hence its role in the transformation would be severely restricted. Moreover, the M’ phase formed during the first stage is very stable at the annealing temperatures that have been applied (1450 ⁰C). Once it has formed it would not re-melt to become a grain boundary liquid phase and to be involved in a’ ® b’ transformation at this temperature [41, 59]. The rate of the a’ ® b’ transformation at the second stage, i.e. after devitrification of the intergranular glass was accomplished, was observed to be somewhat lower than at the first stage but the process continued at a nearly constant rate, suggesting an additional mechanism(s) to be operative until the system is closed to equilibrium. The activation energy of a reconstructive phase transformation will be significantly reduced if the atomic diffusion is assisted by a liquid phase and/or it there exist nuclei to reduce the interfacial energy. The occurrence of the self-generated liquid in transformed b’ grains leads to a logical assumption that this liquid phase plays a major role in promoting atomic diffusion and facilitating a’® b’transformation after crystallization of the grain boundary liquid. Because it involves a liquid phase emerging from the transforming a’ phase itself, the transformation from a’ to b’ can be proceed without involving much grain boundary liquid and a steady rate of the transformation may be expected.

The appearance of the RE-rich spheroids in the b’ phase results in thermodynamic instability, hence these nano-sized inclusions would eventually diffuse out of the transformed b’ grains through the easy path of dislocations. It is therefore, suggested in [61] that the amount and viscosity of the self-generated liquid within the transforming a’ grains is a determining factor in controlling the rate of the a’® b ’ transformation when most at the grain boundary glass has been crystallized.

Without much involvement of grain boundary glass, the a’ ® b’ transformation at the second stage may be regarded as a decomposition process of a-SiAlON phases. This suggests that, below a certain temperature, some of the a’ compositions have a higher free energy than b’ phases, and the transformation from a’ to b’ would occur if it could be thermally activated. Because of the compositional difference between a’ and b’ phases, the decomposition of an a’ phase must produce other phases(s) in addition to a b’ composition. If in the [59] it was suggested that a’ decomposes with the formation of a’ and M’(R), the more recent investigations in the Sm-Si-Al-O-N system [61] where the EDS X-ray mapping technique was applied for quantification of AlN polytypoid phases showed that another, more complicated, reaction path seems to be more plausible for description of Sm a’–SiAlON phases decomposition, namely .

Further attempts to reach even more complete understanding of the mechanism of this interesting transformation together with the evaluation of different rare-earth-doped duplex (a–b)-SiAlONs for long term high temperature stability are presented in [62, 63]. Both aspects are of special interest since many applications for (a–b)-SiAlON ceramics are in the temperature range where the a’ ® b’ transformation will continue with consequent continuous change in properties. The research accomplished in [62] was mostly aimed at elucidation of the effect of starting composition, type of rare earth sintering additive and amount of liquid phase on the a « b transformation during post sintering heat treatment. Therefore, five different (a–b)-SiAlON and a–SiAlON compositions have been prepared using neodymium, samarium, dysprosium and ytterbium. The reasons for preparation of these compositions were as follows:

- first group was formulated as the original starting composition used in [18] and has been shown to readily undergo a « b SiAlON transformation;

- the second group was formulated to produce the phase composition » 30% a-SiAlON and 70% b-SiAlON ;

- the third was also aimed for the two-phase region but was more a–SiAlON rich (» 50% a–SiAlON and 50% b-SiAlON); this composition was useful for comparison with the first two compositions to understand the effect of liquid phase, since the amount of liquid is less and the viscosity is higher than in the other two;

- the fourth group was formulated for an a–SiAlON with m = 1 and n = 1.7 (oxygen-rich corner of the a–SiAlON phase region) and was designed to determine whether all different cation size rare earth oxide sintering additives give pure a–SiAlON, since the borders of a–sialon phase field have only been established precisely in the yttrium system;

- the fifth group was designed as an a–SiAlON composition with m = 1.5 and n = 1.5, representing the maximum nitrogen content a–sialon composition which can be prepared without LnN.

A combination of extensive X-ray and microstructural observations on sintered and heat treated samples showed that the a–SiAlON composition in mixed (a–b)-SiAlON materials is always located at the edge of the a–SiAlON phase region and is less stable with respect to a–b-SiAlON transformation than in single phase a–SiAlON phase region. For (a–b)-SiAlON composites the transformation was shown to proceed more readily with the increase of the amount and the decrease of the viscosity of liquid phase formed during heat treatment. For the a–SiAlON starting compositions prepared both within or the edge of the a–SiAlON phase region, the regularities of the ease of phase transformation are of a more complex nature. A combination of thee factors influences the ease of transformation: the cation size of the sintering additive, the presence of b-SiAlON grains and also the amount and viscosity of liquid phase present during heat treatment. The role of the liquid phase in the phase transformation for all compositions was elucidated in the same way: an extra amount (10% and 20%) of glass powders of overall composition Ln: Si: Al = 1:1:1 to the samples. Such additions led to unequivocal intensification of the a’® b , transformation in all samples with the exception of the ones from the fifth group, that were doped with small ionic size of sintering additives (Dy₂O₃ and Yb₂O₃). For these materials the presence of b-SiAlON grains was necessary for triggering the transformation. Moreover, the increase of the number of such grains (10% b-SiAlON with z = 0.8 was added to one of the mixtures) enhanced the transformation substantially, while such grains might possibly act as a nucleation site for a ® b SiAlON transformation. Only then the amount and viscosity of liquid became important factors for the overall reaction a–SiAlON + b-SiAlON + (21R) + liquid ® b-SiAlON + (21R) + grain-boundary phase(s). b-SiAlON nucleation sites provided more favorable kinetic conditions for transformation in case of large ionic size sintering additives, but the process itself proceeded without any additional ones. The overall results are summarized in Table 7, where increasing number of "+" marks indicate the more dominant mechanism influencing a « b SiAlON transformation.

Somewhat similar research was carried out for duplex a–b-SiAlON stabilized with Nd, Yb and Y. The former two RE elements were chosen because they represent the size end members of rare earth cations known to stabilize the a’–phase, while Y served as a reference. The long-term high-temperature stability of the different duplex (a–b)-SiAlON ceramics was of special interest. Phase and microstructural development and the volume fraction and chemistry of residual intergranular phases were characterized and related to mechanical properties. It was shown that on sintering the duplex (a–b)-SiAlON starting powder compositions result in equilibrium phase compositions consisting of a’, b’ and a residual glassy phase. The analytical XRD investigations showed that a’ and b’ substitution levels and the equilibrium composition and volume fraction of the glass are dependent upon the starting composition as well as the a’ stabilizing cation. However, the b’/(a’+b’) weight ratio was shown to depend mostly on the RE stability element used. The long-term heat treatment experiments at 1450 ⁰C clearly indicate that Y and Yb give a stable a–SiAlON phase, while Nd gives a less stable one that during prolonged heat treatment at 1450 ⁰C decomposes into crystalline phases: melilite, b-SiAlON, and 21R polytypoid. These observations are fully consistent with the results of [62]. Overall, the above described experimental results on a’-phase stability and a « b SiAlON transformations strongly indicate that for duplex

(a–b)-SiAlON composites designed for high-temperature applications RE elements with smaller ionic radii present an optimal choice.

Reaction densification of a - and (a - b )- SiAlONs

In the previous parts of the present review the importance of proper choice of initial composition, of the type of RE-oxide used as a sintering additive and of some processing parameters for optimal phase assemblage formation, and hence of the desired properties of SiAlONs were discussed. However, an even more important issue is to achieve the complete densification of SiAlON materials, preferably by means of pressure-less sintering. In order to optimize this process basic understanding of the densification behavior of SiAlONs is necessary.

Densification of SiAlONs occurs by reaction liquid-phase sintering or "transient liquid sintering", the latter concept specifically applied to SiAlONs. Intensive investigations have been done on the reaction sintering of SiAlON in the metal oxide-Si₃N₄-AlN-SiO₂-Al₂O₃system [64-75]. Some of the earliest work was carried out in the MgO-Si₃N₄-AlN-SiO₂-Al₂O₃ system [64-67]. It was found that a transient phase, which formed during the hot pressing of a -Si₃N₄ with MgO, reacted with a -Si₃N₄ above 1400 ⁰C to form b-Si₃N₄ and a vitreous phase [64]. This reaction was found to expedite the sintering kinetics and the reaction pathway of a mixture of MgO-Si₃N₄-AlN-SiO₂-Al₂O₃ depended on the amount of MgO added [64-66]. The addition of Y₂O₃, instead MgO to aSi₃N₄-AlN-SiO₂ mixture resulted in a material with diphasic b’-SiAlON-Y₃Al₅O₁₂ (YAG) microstructure with improved mechanical properties [67]. Later the attempts were made to obtain dense b-SiAlON ceramics in the Si₃N₄-AlN-SiO₂-Al₂O₃ plane [68-70]. In these studies, it was shown that the sintering kinetics depended on the composition of starting powder mixtures, even though the final composition was the same. In these and other reaction sintering and hot-pressing studies, the densification mechanisms were identified to be solution-reprecipitation [64-67], fast particle rearrangement [73], Cobble creep [73], or grain boundary sliding [72, 74].

Recently, it was shown by Hwang and Chen that reaction hot pressing of a’- and a’+ b’-Y-SiAlON took place in three stages [75]. Further, it was found out that the wetting properties of the Y₂O₃- Al₂O₃-SiO₂ eutectic melt controlled the densification behavior of the powder compact. The first stage was identified with the formation of the ternary oxide eutectic and YAG. The shrinkage in this stage was due to the redistribution of liquid in the powder compact and slight improvement in the packing efficiencies. The second stage was identified with the wetting of AlN particle and formation of b₆₀. The preferential wetting of AlN by the ternary oxide melt caused localization of the liquid, leading to a delaying effect on the second shrinkage step. The third stage occurred with the dissolution of Si₃N₄ and formation of the final phase. Massive particle rearrangement was found to be the dominant densification mechanism. It also caused Al enrichment in the liquid, leading to formation of b ₆₀ as a transient phase. Thus the physical/chemical characteristics of the liquid, in particular its wetting behavior, and the kinetic pathway of the intermediate reactions are clearly important factors in reaction densification of multicomponent systems.

A survey of the literature indicates that until recent time there was almost no knowledge of the wetting characteristics of various ternary oxide eutectic melts on nitrides. Most studies on reaction sintering have been restricted to the documenting of phase evolution and densities. Only limited information may be gleaned from the reports of transient/intermediate phases of these studies. The reaction sequence were studied in Ca-, Y- and Nd-a-SiAlON, and, according to [76, 77], gehlenite (Ca₂Al₂SiO₇) was found to be transient phase in Ca-a-SiAlON, whereas N-melilite (R₂Si₂O₃N₄) formed during the densification of Y- and Nd-a-SiAlON. N-melilite was also reported to be a transient phase in Nd- and Sm-doped a- SiAlON [79]. The same phase was identified as a transient one during the pressureless sintering of lanthanide (Y and Nd to Ln, except Pm) – modified a -SiAlON [80]. K-phase (RSiO₂N) and b-Si₃N₄phase along with melilite were determined in the case of Nd-a - SiAlON in [31]. Furthermore, in the same work J-phase (R₄Si₂O₂N) and b-Si₃N₄ were reported to occur along with melilite in the case of Sm- and Gd-a-SiAlON, and J and b -Si₃N₄in case of Dy-, Y-, Er- and Yb-a-SiAlON. The reaction sequence in hot-isostatic-pressed and pressureless-sintered, rare-earth-doped (a - b )- SiAlONs have been studied in [6, 80-83]. Formation of 12H and La-silicon-oxynitride as the transient phase in the case of Y- or Y/La-doped samples were reported. Generally it has not been known what controls the formation and selection of various phases. The first indication that a general understanding of reaction sequence during sintering of SiAlON materials may, after all, be based on simple physical chemical considerations governing surface activities of nitrides with reactive liquids was given in [75]. There it was shown that in Y-SiAlON system, Y₂O₃-Al₂O₃-SiO₂ melt preferentially wets AlN, leading to the formation of b₆₀-SiAlON as a transient phase, which delays densification in the early stages due to liquid trapping.

A more systematic research of the wetting behavior of the ternary eutectic melt M₂O₃-Al₂O₃-SiO₂ and its interrelation with phase development during reaction hot pressing of a-Si₃N₄, AlN, Al₂O₃, and M₂O₃ powder mixtures (M = Li, Mg, Ca, Y, Nd, Sm, Gd, Dy, Er, and Yb) was accomplished in [84], and in [85] these results were tied in with densification. The evolution of phases on heating was studied by means of heating up the mixtures to some intermediate temperatures, cooling down, and subsequent XRD characterization. The wetting experiments were to study the trends in the wetting behavior of the ternary oxide melts on the nitrides. For this propose ternary eutectic compositions were prepared and pressed into pellets, which were placed on substrates prepared from Si₃N₄ and AlN by hot pressing without additives. These "sandwiches" were then heated to different temperatures, in a nitrogen atmosphere, where the pellets fused. On cooling, the contact angles were measured from photomicrographs. Since the wetting characteristics are sensitive to the composition (including the oxygen/nitrogen content in the melt) and the temperature, the true equilibrium was not achieved during the experiments [84]. Therefore, these experiments can not be considered as yielding the actual values of the wetting angle. Lather, an insight into preferential wetting of the nitride powders by the ternary oxide melt in a multicomponent system was provided.

It was shown that the wetting behavior of the ternary oxide melt towards Si₃N₄ and AlN was strikingly different, depending on the metal oxide used. Such a preference in reaction sequence was explained in [84] in terms of acid-base chemistry. According to the Pearson’s principle, hard acids preferentially react with soft bases [40]. Empirically, polarizability, electronegativity, and charge to size ratio [40] can be used as a measure of hardness of as acid or a base [86]. Recently, a model of pH₀ (the point of zero charge that corresponds to the pH at which the immersed solid has a zero net surface charge) values calculation for ionic metal oxides based on the ionizational potential of the metal ion was proposed [87]. If the oxides used in [84] are put in order of decreasing basicity according to this model, and if the acidity of the nitrides is assumed to be determined by the acidity of their native oxide layers (in the case Si₃N₄ will be more acidic than AlN), than both the experimental wetting data and the sequence of the intermediate reaction products in relation to their Si/Al ratio will be in total accordance with the above mention acid-base principle. Therefore, the harder acid, Si₃N₄, reacts with hard bases of oxide melts of Li, Ca, Mg, and lighter rare earth metals up to Gd, whereas the soft acid AlN reacts with the soft bases of the heavier rare-earth metals down to Dy. In the order of decreasing Si/Al ratio and additive basicity, typical intermediate phases are O’ (SiN₂O₂), M-cordierite (Mg₂Al₄Si₅O₁₈), M’ (R₂Si₂AlO₄N₃), and b₆₀-SiAlON (Si_0.8Al_4.2O_4.2N_3.8).

The temperatures for the various reactions, which can be used to understand the shrinkage behavior, were also identified in [84] based on the observed wetting/reaction behavior. The reactions identified during the reaction densification of a-SiAlON are:

(i) the eutectic formation (at temperature T₁),

(ii) wetting of a nitride powder and intermediate phase precipitation (at temperature T₂),

(iii) secondary wetting of the other nitride powder (at temperature T₃),

(iv) dissolution of the intermediate phase (at temperature T₄), and

(v) precipitation of the final phase, a -SiAlON (at temperature T₅)

The schematic representation of expected shrinkage curves when the eutectics melt preferentially wets Si₃N₄ and AlN are shown in Fig. 22 (a) and (b), respectively. In these curves, the beginning of various shrinkage steps are identified with some of the above characteristic temperatures. Wetting of AlN is shown to lead to small or no immediate shrinkage in the powder compact because of the low AlN content of the typical SiAlON compact [75]. As it is shown in Fig. 22 (b), though, a subsequent shrinkage step occurs when the majority Si₃N₄ is wetted. Variation of these shrinkage curves may also be possible. For example, it is known that the initial precipitation of the intermediate phase usually starts with partial wetting of the first nitride, i.e., T₂ could be significantly lower than the temperature when complete wetting is achieved. Also, the dissolution of intermediate phase may or may not be higher than T₅. Indeed T₄ may even be lower than T₃, as in case of Li and Ca, where wetting and dissolution of the second nitride (AlN) occurs after the intermediate phase (O’) is dissolved and partial formation of a-SiAlON is already achieved [85]. Experimental data on the characteristic temperatures of shrinkage determined in [85] for a number of systems are presented in Table 8. Experimental shrinkage curves obtained by means of hot pressing with a constant treating rate [84] were in good agreement with the ones, predicted by the sequence of wetting reactions.

Reaction densification of M-SiAlON occurs in three stages. The first stage is associated with the formation of the SiO₂-Al₂O₃-M₂O₃ ternary eutectic (at temperature T₁). The second stage is associated with the preferential wetting of majority nitride powder, Si₃N₄. This occurs at T₂ when Si₃N₄ is wetted first. If AlN is wetted first, then no shrinkage step is observed at T₂ and the densification is delayed until T₃, when Si₃N₄ is wetted finally.

The third stage involves the dissolution/melting of the intermediate phase (at temperature T₄) if T₄ is low; otherwise a distinct third stage is not seen while a gradual densification continues following the second stage. In the extreme case of very high T₄, this may result in poor density as in case of Mg, where cordierite precipitates as an intermediate phase.

The formation temperature of the final a-SiAlON phase (T₅) is not crucial in most cases, unless a sudden secondary precipitation of a-SiAlON at this later stage (T’₅) drains the liquid and retards densification. Such material must be densified at temperatures below T’₅. To achieve full densification at relatively low temperatures, a low wetting temperature for the second nitride (T₃) and low dissolution/melting temperature of the intermediate phase (T₄) are required. The dominant process for densification in SiAlON ceramics was shown to be massive particle rearrangement although some dissolution facilities full densification by providing the necessary liquid [85].

More generally, the amount of liquid is controlled by the phase diagram, dissolution/melting of intermediate phase, distribution of liquid and secondary precipitation of a-SiAlON. The effect of viscosity of liquid as opposite to the most of expectations, is relatively insignificant. The different characteristic temperatures that vary widely among the SiAlON systems obscure its effect in most cases. For example, according to [85], although in heavier rare-earth SiAlONs in isothermal hot pressing the densification time increased slightly in the order of Dy, Er, and Yb, most probably due to the influence of viscosity, the densification kinetics in constant heating rate do not reflect the same trend, because of non-monotonic variation of characteristic temperatures.

CONCLUSIONS

a¤b- SiAlON composites represent a family of ceramic materials promising from the standpoint of tailoring for specific applications due to the flexibility of their phase composition and microstructure. The latter could be easily modified not only by means of initial composition alterations but rather by means of post-sintering heat treatment. This possibility is provided by the fully reversible a « b SiAlON transformation that occurs in the presence of some secondary phases, specifically Al-containing nitrogen melilite solid solutions, M'(R). The mechanism of this transformation is similar to the one occurring in silicon nitride, however in the case of SiAlONs the transformation is not only fully reversible but also may occur in the absence of substantial amounts of intergranular liquid phase since internally-generated liquid phase forms at certain temperatures and under certain conditions. The amount of the latter, although rather small, is sufficient for promoting the reconstructive a¢ « b¢ phase transformation.

Stability of phases (a¢, M’(R), J’(R), etc.) and of phase assemblages is determined by the rare-earth ionic radius. Therefore, definite properties of a¤b-SiAlON composites may be achieved by using certain RE-oxides as sintering additives.

Sintering behavior of SiAlONs, which can be described as reaction densification in the presence of liquid phase, can be rationalized in terms of simple physical chemical considerations, i. e. of the acid/base reaction principle. Shrinkage and formation of the transient phases during sintering as well as the final phase and microstructure of SiAlON ceramics depends on the preferential wetting of the nitrides by the RE-containing silicate melt.

Overall, the recent findings indicate that a good basis for further SiAlON development has been achieved due to diverse in-depth research of a variety of RE-Si-Al-O-N systems.

ACKNOWLEDGMENTS

The financial support of the research activities of Dr. V. A. Izhevskyi by CNPq during his stay at IPEN as an invited scientist is highly appreciated.

REFERENCES

[1] Y. Oyama, O. Kamigaito, "Solid solubility of some oxides in Si₃N₄", Jpn. J. Appl. Phys. 10, (1971) 1637-42 .

[2] K. H. Jack, "Solid solubility of alumina in silicon nitride", Trans. J. Brit. Ceram. Soc. 72 (1973) 376-78.

[3] K. H. Jack, W. I. Wilson, "Ceramics based on the Si-Al-O-N and relates systems", Nature Phys. Sci. (London), 238, (1972) 28-29 .

[4] L. J. Gauckler, H. L. Luckas, G. Petzow, "Contribution to the phase diagram Si

58

[5] K. H. Jack, "The significance of structure and phase equilibria in the development of silicon nitride and sialon ceramics", Sci. Ceram. 11, (1981) 125-42.

[6] T. Ekström, P.-O. Käll, M. Nygren, "P.-O. Olssen. Dense single-phase

24

[7] Z. K. Huang, W.-Y. Sun, D.-S. Yan, "Phase relations of the

₃₄4

[8] S. Hampshire, H. K. Park, D. P. Thompson, K. H. Jack, "

274

[9] D. P. Thompson, "The crystal chemistry of nitrogen ceramics", Mater. Sci. Forum 47, (1989) 21-42.

[10] M. Mitomo, H. Tanaka, K. Muramatsu, N. Li, Y. Fujii, "The strength of a-Sialon ceramics", J. Mater. Sci., 15 (1980) 2661-61

[11] M. Mitomo, F. Izumi, Y. Bando, Y. Sakikawa, "Characterization of a-Sialon ceramics", in: Ceramic Components for Engines. Edited by S. Somiya et al. KTK Science Publishers, Tokyo, Japan, (1984) 377-86.

[12] K. Ishizawa, N. Ayuzawa, A. Shiranita, M. Takai, N. Uchida, M. Mitomo, "Some properties of a-sialon ceramics", in: Ceramic Components for Engines. Edited by W. Bunk and H. Hausner, German Ceramic Society, Bad Honnet, Germany, (1986) 511-18.

[13] D. Stutz, P. Greil, G. Petzow, "Two-dimensional solid solution forming of Y-containing a- Si₃N₄", J. Mater. Sci., Lett., 5 (1986) 335-36.

[14] Z. K. Huang, T. Y. Tien, D. S. Yen, "Subsolidus phase relationship in the Si₃N₄-AlN-Rare Earth oxide systems", J. Am. Ceram. Soc., 69 (1986) C-241-C-242.

[15] M. Mitomo, "In situ microstructure control in silicon nitride based ceramics", in: Advanced Ceramics II. Edited by S. Somiya. Elsevier, Barking, Essex. U.K., (1988) 147-61.

[16] T. Ekström., "Effect of composition, phase content and microstructure on the performance of yttrium sialon ceramics", Mater. Sci. Eng., A109 (1989) 341-49.

[17] T. Ekström, P.-D. Käll, M. Nygren, P. O. Olsson, "Mixed a- and b- (Si-Al-O-N) materials with yttria and neodymia additions", Mater. Sci. Eng., A105/106, (1988) 161-65.

[18] H. Mandal, D. P. Thompson, T. Ekström, "Reversible a « b SiAlON transformation in heat-treated sialon ceramics", J. Eur. Ceram. Soc., 12 (1993) 421-429.

[19] H. Mandal, D. P. Thompson, T. Ekström, "Optimization of sialon ceramics by heat treatment", in: Third Euro - Ceramics, Vol. 3, ed. P. Duran and J. F. Fernandez. Faenza Editrice Iberica S.L., Spain, (1993) 385-390.

[20] H. Mandal, D. P. Thompson, Y. W. Sun, T. Ekström, "Mechanical property control of rare earth oxide densified a- b ", in: 5^th International Symposium on Ceramic materials and Components for Engines, ed. D.S. Yan, Y.R. Fu and S.Y. Shi. World Scientific Publishers. Singapore, (1995) 441-446.

[21] H. Mandal, D. P. Thompson, "Mechanism of a - b SiAlON transformation", in: Fourth Euro – Ceramics, ed. G. Galus Gruppo Editoriale. Faenza Editrice S.p.A , Faenza, Italy, (1995) 273-280.

[22] Y.-B. Cheng, D. P. Thompson, "Aluminum – containing nitrogen melilite phases", J. Am. Ceram. Soc. 77, (1994) 143-48.

[23] H. Camuscu, H. Mandal, D. P. Thompson, "Optimized high-temperature sialon ceramics containing melilite as the grain boundary phase", in: 21^st Century Ceramics, British Ceramic Proceedings, ed. D. P. Thompson , 55, (1996)239-248.

[24] H. Mandal, Y.-B. Cheng, D. P. Thompson, " a- SiAlON ceramics with a crystalline melilite grain-boundary phase", in: Proceedings of the 5^th International Symposium on Ceramic materials and Components for Engines, 1994 Shanghai, ed. D.S. Yan, Y.R. Fu and S.Y. Shi. World Scientific Publishers. Singapore, (1995) 202-205.

[25] V. A. Ijevskii, U. Kolitsch, H. J. Seifert, I. Wiedmann, F. Aldinger, " Aluminum- containing Ytterbium Nitrogen Woehlerite solid solutions. Synthesis, Structure and some properties", J. Eur. Ceram. Soc. 18 (1998) 543-52.

[26] K. H. Jack, "Fabrication of dense nitrogen ceramics", in: Processing of Crystalline Ceramics. Ed. By H. Palmour III, R. F. Davis and T. M. Hane. Plenum Press, New York, (1978) 561-78.

[27] R. Marchand, A. Jayaweera, P. Verdier, J. Lang, "Preparation and characterization of new oxynitrides in the lanthanide-sialon-oxide-nitride system", Acad. Sci. Paris, C. R. 283, (1976) 675-77.

[28] K. H. Jack, "Silicon Nitride Sialons and related ceramics", in Ceramics and Civilization, vol. III, High-Technology ceramics. American ceramic Society, Columbus, OH, (1986) 259-88.

[29] P. L. Wang, H.Y. Tu, W. Y. Sun, D. S. Yan, M. Nygren, T. Ekström, "Study on the solubility of Al in the melilite systems R₂Si_3-x Al_xO_3+xN_4-x with R= Nd, Sm, Gd, Dy and Y", J. Eur. Ceram. Soc., 15 (1995) 689-695.

[30] R. K. Harris, M. J. Leach, D. P. Thompson, "Nitrogen-15 and oxygen-17 NUR spectroscopy of silicates and nitrogen ceramics", Chem. Mater.,4 (1992) 260-267.

[31] P. L. Wang, W. Y. Sun, D. S. Yan, "Formation and densification of R-a’-SiAlONs (R= Nd, Sm, Gd, Dy, Er and Yb)", in: Silicon Nitride Ceramics: Scientific and Technological Advances, vol. 287, Eds. I.-W. Chen, P. F. Becher, M. Mitomo, G. Petzow and T.-S. Yen. Materials Research Society, (1993) 387-92.

[32] M. J. Hoffman, G. Petzow, "Microstructural design of Si₃N₄ based ceramics", Ibid., 3-14.

[33] K. S. Chee, Y.-B. Cheng, M. E. Smith, "The solubility of aluminum in Rare Earth Nitrogen Melilite Phases", J. Eur. Ceram. Soc., 15 (1995) 1213-1220.

[34] S. Slasor, K. Liddel, D. P. Thompson, "The role of Nd₂O₃ as an additive in the formation of a’and b’ sialons", Br. Ceram. Proc., 37 (1086) 51-64.

[35] P.-O. Käll, T. Ekström, "Sialon ceramics made with mixtures of Y₂O₃-Nd₂O₃ as sintering aids", J. Eur. Ceram. Soc., 6 (1990) 119-127.

[36] B. K. Vainstein, V. M. Fridkin, V.L. Indenbom, "Crystallochemical radu systems", In Modern Crystallography II, Structure of Crystals, Ed. By B. K Veinstein, A. A Chernov and L. A Shuvalov, Springer, Berlin, Heidelberg, (1982) 78-81.

[37] G. Z. Cao, R. Metselar, "a- Sialon Ceramics: A Review", Chem. Mater., 3 (1991) 242-53.

[38] Z. K. Huang, S. Y. Liu, A. Rozenflanz, I. W. Chen, "Sialon composites containing rare-earth melilite and neighboring phases", J. Am. Ceram. Soc., 79 (1996) 2081-90.

[39] Z. K. Huang, I. W. Chen, "Rare-earth melilite solid solution and its phase relations with neighboring phases", Ibid. 79 (1996) 2091-97.

[40] J. E. Huheey, "Inorganic Chemistry", Harper and Row. New York, (1983).

[41] Z. K. Huang, D.-S. Yan, T. S. Tien, "Compound formation and melting behavior in the AB compound and rare earth oxide systems", J. Solid State Chem., 85 (1990) 51-55.

[42] I. W. Chen, D. P. Thompson, "Preparation and grain boundary devitrification of samarium a-SiAlON ceramics", J. Eur. Ceram. Soc., 14 (1994) 13-21.

[43] Z. K. Huang, P. Greil, G. Petzow, "Formation of a-Si₃N₄ solid solution in the system Si₃N₄-AlN-Y₂O₃", J. Am. Ceram. Soc., 66 (1983) C-96-C-97.

[44] W. Y. Sun, T. S. Tien, T. S. Yen, "Solid solubility limit of single-phase region of a-SiAlON solid solutions in the system Si, Al, Y/N, O", J. Am. Ceram. Soc., 74 (1991) 2547-50.

[45] W. Y. Sun, T. S. Tien, T. S. Yen, "Subsolidus phase relationships in part of the system Si, Al, Y/N, O: the system Si₃N₄-AlN-YN-Al₂O₃-Y₂O₃", J. Am. Ceram. Soc., 74 (1991) 2753-58.

[46] D. P. Thompson, "Phase relationship in Y-Si-Al-O-N ceramics", in: Proceedings of the 21^st. University Conference on Ceramic Science, Tailoring Multiphase and Composite Ceramics, edited by R. E. Tressler et al. Plenum Press, New York, (1986) 79-91.

[47] M. Mitomo, F. Izumi, S. Horiuchi, Y. Matsui, "Phase relationships in the system Si₃N₄-SiO₂-La₂O₃", J. Mater. Sci., 17 (1982) 2359-66.

[48] W. Y. Sun, T. S. Yen, T. S. Tien, "Subsolidus phase relationships in the system RE-Al-O-N (Where R= rare earth elements)", J. Solid State Chem., 95 (1991) 424-29.

[49] W. Y. Sun, Z. K. Huang, T. S. Tien, T. S. Yen, "Phase relationships in the system Y-Al-O-N", Mater. Sci. Lett., 11 (1991) 67-69.

[50] H. Mandal, D. P. Thompson, T. Ekström, "Heat treatment of Ln-Si-Al-O-N glasses", in: Proc. 7^th. Irish Mater. Forum Conf. IMF7, Eds. M. Buggy and S. Hampshire. Trans. Tech. Publications, Switzerland, (1992) 187-203.

[51] D. P. Thompson, "New grain-boundary phases for nitrogen ceramics", Mater. Res. Soc., Symp. Proc., 287 (1993) 79-92.

[52] W. Y. Sun, D. S. Yan, L. Gao, H. Mandal, D. P. Thompson, "Subsolidus phase relationships in the systems Ln₂O₃-Si₃N₄-AlN-Al₂O₃ (Ln = Nd, Sm)", J. Eur. Ceram. Soc., 15 (1995) 349-355.

[53] W. Y. Sun, D. S. Yan, L. Gao, H. Mandal, D. P. Thompson, "Subsolidus phase relationships in the systems Dy₂O₃-Si₃N₄-AlN-Al₂O₃", Ibid., 16 (1996) 1277-1282.

[54] U. Kolitsch, V. A. Ijevskii, H. J. Seifert, I. Wiedmann, F. Aldinger, "Formation and general characterization of a previously unknown ytterbium silicate (A-type Yb₂SiO₅)", J. Mater. Sci. 32 (1997) 6135-39.

[55] M. H. Lewis, "Crystallization of grain boundary phases in silicon nitride and SiAlON ceramics", in: Silicon Nitride 93, Key Engineering Materials, Eds. M. J. Hoffmann, P. F. Becker and G. Petzow, Trans. Tech. Pub., 89-91 (1993) 333-38.

[56] K. H. Jack, "Review: SiAlONs and related nitrogen ceramics", J. Mater. Sci. 11 (1976) 1135-58.

[57] S. Hampshire, "Nitride Ceramics", in : Materials Science and technology, Structure and properties of Ceramics, ed. M. Swain, VCH, Weinheim, 11 (1994) 121-71.

[58] S. Hampshire, K. H. Jack, "The kinetics of densification and phase transformation of nitrogen ceramics", in: Special Ceramics 7, Proc. Brit. Ceram. Soc., Eds. DW Taylor and P. Popper, 31 (1981) 37-49.

[59] Z. Shen, T. Ekström, M. Nygren, "Temperature stability of samarium-doped a- SiAlON ceramics", J. Eur. Ceram. Soc., 16 (1996) 43-54.

[60] R. Zhao, Y. B. Cheng, "Microstructural feature of the a- to b-SiAlON phase transformation", J. Eur. Ceram. Soc. 16 (1996) 529-34.

[61] R. Zhao, Y. B. Cheng, "Decomposition of Sm a-SiAlON phases during post-sintering heat treatment" J. Eur. Ceram. Soc., 16 (1996) 1001-08.

[62] H. Camuscu, D. P. Thompson, H. Mandal, "Effect of starting composition, type or rare earth sintering additive and amount of liquid phase on a « b SiAlON transformation", J. Eur. Ceram. Soc., 17 (1997) 599-613.

[63] L. K. L. Falk, Z. J. Shen, T. Ekström, "Microstructural stability of duplex a - b SiAlON ceramics", J. Eur. Ceram. Soc., 17 (1997) 1099-1112.

[64] P. Drew, M. H. Lewis, "The microstructure of silicon nitride ceramics during hot-pressing transformations", J. Mater. Sci. 9 (1974) 261-69.

[65] M. H. Lewis, B. D. Powell, P. Drew, R. J. Lumby, B. North, A. J. Taylor, "The formation of single-phase Si-Al-O-N ceramics", J. Mater. Sci., 12 (1977) 61-74.

[66] M. H. Lewis, A. R. Bhatti, R. J. Lumby, B. North, "The microstructure of sintered Si-Al-O-N ceramics", J. Mater. Sci., 15 (1980) 103-13.

[67] M. H. Lewis, R. J. Lumby, "Nitrogen ceramics: liquid phase sintering" Powder Metall., 26 (1993) 73-81.

[68] G. K. Layden, "Process development for pressureless sintering of SiAlON ceramic components", Report No. R175-91072-4. United Technologies Research Center, (1976).

[69] S. Boskovich, L. J. Gauckler, G. Petzow, T. Y. Tien, "Reaction sintering forming b-Si₃N₄ solid solution in the system Si, Al / N, O", I: Sintering of Si₃N₄-AlN mixtures. Powder Metall. Int., 9 (1977) 185-89.

[70] S. Boskovich, L. J. Gauckler, G. Petzow, T. Y. Tien, "Reaction sintering forming b-Si₃N₄ solid solution in the system Si, Al / N, O", II: Sintering of Si₃N₄-SiO₂-AlN mixtures. Powder Metall. Int., 10 (1978) 180-85.

[71] S. Boskovich, L. J. Gauckler, G. Petzow, T. Y. Tien, "Reaction sintering forming b-Si₃N₄ solid solution in the system Si, Al / N, O", III: Sintering of Si₃N₄-AlN-Al₂O₃ mixtures. Powder Metall. Int., 11 (1979) 169-71.

[72] M. N. Rahman, F. L. Riley, R. J. Brook, "Mechanisms of densification during reaction hot-pressing in the system Si-Al-O-N", J. Am. Ceram. Soc., 63 (1980) 648-53.

[73] M. Kuwabara, M. Benn, F. L. Riley, "The reaction hot pressing of compositions in the system Al-Si-O-N corresponding to b ’- SiAlON", J. Mater. Sci., 15 (1980) 1407-16.

[74] M. Havier, P. L. Hansen, "Hot-pressing and a-b’phase transformation of compositions corresponding to b ’- SiAlON", J. Mater. Sci., 25 (1990) 992-96.

[75] S. L. Hwang, I. W. Chen, "Reaction hot pressing of a and b ’- SiAlON ceramics", J. Am. Ceram. Soc., 77 (1994) 165-71.

[76] W. Y. Sun, P. A. Walls, D. P. Thompson, "Reaction sequences in preparation of Sialon ceramics", in: Non-Oxide Technical and Engineering Ceramics Proc. Inter. Conf., Limerick, Ireland, 1985, Ed. S. Hampshire. Elsevier Applied Science, London, England, (1986) 105-17.

[77] S. Slasor, D. P. Thompson, "Preparation and characterization of yttrium-a’- SiAlONs", Ibid., 223-30.

[78] S. Hampshire, K. P. J. O’Reilly, M. Leigh, M. Redington, "Formation of a’- SiAlONs with neodymium and samarium modifying cations", in: High Tech Ceramics. Edited by P. Vincezini. Elsevier Science, Amsterdam, Netherlands, (1987) 933-40.

[79] K. P. J. O’Reilly, M. Redington, S. Hampshire, M. Leigh", Parameters affecting pressureless sintering of a’- SiAlONs with lanthanide modifying cations", in: Silicon Nitride Ceramics Scientific and Technological Advances, Proc. Mat. Res. Soc. Symp., 287 (Boston, MA, Nov. 30- Dec. 3, 1992). Ed. I. W. Chen, P. F. Pedier, M. Mitomo, G. Petzow, and T. S. Yen. . Materials Research Society, Pittsburgh, DA, (1993) 393-98.

[80] T. Ekström, "SiAlON ceramics sintered with yttria and rare earth oxides", Ibid., 121-32.

[81] T. Ekström, P. O. Olssen, " b ’- SiAlON ceramics prepared at 1770 ^oC by hot-isostatic pressing", J. Am. Ceram. Soc., 72 (1989) 1722-24.

[82] T. Ekström, P. O. Olssen, "Pressureless sintering of SiAlON ceramics with mixed Y₂O₃-La₂O₃ additions", J. Mater. Sci., Lett., 6 (1989) 119-27.

[83] T. Ekström, P. O. Olssen, "HP- sintered b- and a-b- SiAlONs densified with Y₂O₃ and La₂O₃ additions", J. Mater. Sci., 25 (1990) 1824-32.

[84] M. Menon, I. W. Chen, "Reaction densification of a’- SiAlON. I: Wetting behavior and acid-base reactions", J. Am. Ceram. Soc., 78 (1995) 545-52.

[85] M. Menon, I. W. Chen, "Reaction densification of a’- SiAlON. II: Densification behavior", Ibid., 553-59.

[86] R. G. Pearson, "Absolute electronegativity and hardness: application to inorganic chemistry", Inorg. Chem., 27 (1988) 734-40.

[87] A. Carre, F. Roger and C. Varinot. Study of acid/base properties of oxide, oxide glass and glass-ceramic surfaces. J. Colloid Interface Sci. 154, (1990) 174-83.

(Rec. 12/98, Ac. 02/99)

(Publicação financiada pela FAPESP)

¹

Invited scientist, on leave from the Institute for Problems of Material Science, National Academy of Sciences of Ukraine, Kiev, Ukraine.

Publication Dates

• Publication in this collection
  01 June 2000
• Date of issue
  Feb 1999

History

• Received
  Dec 1998
• Accepted
  Feb 1999