Materials Science Research

The emergence of new techniques to produce uniform-size, spherical particles offers opportunities to quantitatively study sintering kinetics. The ability to control the average particle size and to form uniformly packed green microstructures enhances the applicability of many conventional experimental studies, such as isothermal and constant heating rate dilatometry and surface area behavior.

Monodisperse, spherical TiO₂ powders, synthesized by controlled hydrolysis of titanium alkoxides, were employed as a model experimental system. Densification kinetics were determined by isothermal and constant heating rate dilatometry using heat treated (crystallized) and amorphous powders. Microstructural evolution was followed by SEM observations for isothermal conditions of 1060, 1100 and 1160°C and was correlated to the sintering process.

A rigorous description of the geometric evolution of microstructure during sintering is presented, based upon two alternate spacefilling constructs: cells associated with the grains in the system, and bipyramids associated with the cell faces. Relations are derived between the rate of densification and the rate of annihilation of vacancies and also appropriate average concentration gradients in the solid phase at its surface. An efficiency factor is defined which describes the conversion of the volume of vacancies annihilated to the global volume shrinkage of the system. These relationships are free from simplifying geometric assumptions, and have potential application to all stages of powder processing.

The sintering of a row of spherical particles is computer simulated by a method which eliminates geometric approximations. Results for neck growth, and the rate law exponent for evaporation-condensation, surface diffusion, volume diffusion and grain boundary diffusion have been simulated and compared with those obtained from integrated sintering equations by others. The sintering involving five major mechanisms simultaneously is also simulated. The result shows that significant errors could occur if geometrical approximations are involved. It also shows that the rate law exponent is dependent on variables like heating rate, sintering temperature and atmosphere.

The development of the topological model of sintering is reviewed. It is pointed out that the decay of the topological network of channels and junction pores (i.e., sintering) is directly comparable to other network decay processes, such as that of grain growth in a polycrystalline aggregate, wherein surface tension both drives and directs the geometric changes that occur. Through this analogy some unfamiliar, though important, features of sintering have been deduced.

Herring’s scaling law⁽¹⁾ considers the particle size dependence of microstructural change, and notably of sintering during the processing of power compacts. On the basis that the driving force, transport path length, transport area and the volume to be transported are proportional to R^-1, R, R², and R³, respectively, where R is the particle size, the times for equivalent geometric change among particles of different sizes can be formulated as: where subscripts 1 and 2 represent two different powders of initial particle size R₁,₀ ^and R_2,0, respectively, and m is an integer. When particle size ratio is maintained throughout the sintering process, the integer m corresponds to a certain transport mechanism (i.e., m = 1: viscous flow; m = 2: evaporation and condensation; m = 3: volume diffusion; m = 4: surface diffusion or grain boundary diffusion).

Pore coarsening, frequently observed during sintering, is usually regarded as a consequence of pore migration with the grain boundaries. Experimental observations on grain growth and pore coalescence during heat treatment of copper, reported in the present work, however support the conclusion that the major mechanism of pore coarsening is Ostwald ripening by a grain boundary diffusion process rather than grain boundary controlled pore migration.

The primary characteristic of a ceramic compact prior to sintering is its green density. The green density obviously determines the amount of shrinkage required to densify a ceramic, but its influence upon the densification rate or the microstructure of the ceramic is not well understood. A full description of microstructure development must include the role of green density. Green density effects are also important in understanding processing-related defects. Green density inhomogeneities on a microscopic scale have recently figured in models of the development of strength-limiting flaws in ceramics [1,2].

The fundamental sintering behavior of iron is examined to determine possible success in enhancing the rate or improving the properties. Within this scope, both liquid phase and activated sintering treatments are considered. Various additions are discussed in terms of diffusivity and phase diagram data. Small concentrations of additions like B, C, N and P are identified as candidate additives for enhanced sintering when combined with various refractory or transition metals.

There is a significant difference in the important processing steps between metals and ceramics. Most of the microstructure development and control which are used to vary properties in metallic systems occur during post-consolidation thermal and mechanical treatments such as recrystallization and precipitation. Only recently are similar techniques being applied to ceramics. For example, a great deal of effort is underway in trying to increase the work of fracture of ceramics by transformational toughening¹ and microcracking². Nevertheless, if the fracture strength of such materials is determined by large random processing flaws, even though the average strength is increased by toughening, the sample to sample variation will still make the systems designer reluctant to utilize the material. Therefore, an improved understanding of ceramic microstucture development is the most critical area which will lead to new, improved, and more widespread use of high technology ceramics.

ZnO compacts were sintered at 650 to 750°C. in dry air, dry argon, argon with up to 50% oxygen, and air-water vapor mixtures. The progress of sintering was measured by mercury porosimetry from which linear shrinkages, volume shrinkages, surface areas, and pore diameters were derived. In the water free atmospheres, initially about 80% loss in surface area and small pore growth occurred, followed by more rapid shrinkage than predicted with relation to surface area reduction. No relation was found with oxygen pressure. The presence of water vapor in air caused pore growth at 650°C. related linearly with shrinkage but without relation to the water vapor pressure.

During sintering, particles and pores change size and shape to decrease the interfacial energy of the system. The pores shrink and the grains or particles grow. Since surface energy is the driving force for both processes, their comparative rates determine the densification rate and microstructure evolution. Since the goal of sintering is to obtain a desired microstructure, it is critical to understand the relationship between the two processes.

Microstructure control during sintering of ceramics is necessary in order to tailor the microstructure to obtain the properties desired in the final material. Such control can only be achieved through a thorough understanding of the relative contributions of the mass transport mechanisms of bulk, grain boundary, and surface diffusion, and vapor transport. Unfortunately, there is a paucity of diffusion data in the literature making it difficult to compare experimental results to sintering models^1,2. On the other hand, vapor phase transport coefficients exist or can easily be calculated permitting comparison between experiment and models if vapor transport is the dominant transport mechanism³.

The reactive sintering of diamond with titanium was studied. The experiments were carried out at pressures of 4.3 and 7.0 GPa in a “TOROID” type pressure chamber. Under these conditions it was possible to investigate reactive sintering at temperatures up to 1973°K without graphitization of diamond.

An x-ray diffraction single profile analysis method is presented which allows the calculation of the particle size distribution function (PSD). This method is verified by using it to analyze simulated diffraction profiles from single, duplex, triplex, and normal distributions of particle sizes. This method is employed to study the sintering of a nickel catalyst.

A method is developed to unfold x-ray diffraction patterns of metal to obtain the pattern from particles residing on a crystalline support. This method is employed to produce the x-ray pattern of cobalt containing particles in a 9.5% Co-ZSM-5 catalyst. The resulting pattern indicates the metallic cobalt particles contain a high density of basal plane faults.

The paper is mainly concerned with wetting and spreading in catalyst-substrate systems and with the roles played by the interactions among catalyst, atmosphere and substrate in the above phenomena. The first part of the paper analyzes the thermodynamics of spreading, using three different approaches. The first of these considers a thick film on a substrate, the second one improves on the first by considering the film thickness to be smaller than the range of the interaction forces between a catalyst atom and the substrate across the film, and the third one starts from a given loading of the catalyst and tries to obtain information, for given interaction forces, on the organization of the catalyst atoms on the substrate. The analysis indicates when a two-dimensional phase of single atoms of catalyst dispersed over the surface of the substrate exists alone, and when it is in equilibrium with a large crystallite. The Ostwald ripening mechanism of sintering occurs only when a two-dimensional surface phase can exist, while migration and coalescence of the crystallites is the mechanism of sintering. when such a phase does not exist. Conditions are identified under which the crystallites spread over the surface of the substrate to extended planar shapes with a steep variation of angle near the leading edge. This happens when the interactions with the substrate are moderately strong and is a consequence of the fact that the wetting angle at the leading edge is larger than the thermodynamic macroscopic wetting angle which is reached at a distance of a few nanometers from the leading edge. This means that total spreading can occur at some distance from the leading edge, while the atoms near the leading edge still have a finite wetting angle with the substrate. When the above interactions are strong, the crystallite will disintegrate and disperse over the surface of the substrate. The second part of the paper contains an explanation of the higher mobility of the catalyst atoms on the surface of the crystallites above the so called Tammann temperature. The third part presents a selection of experimental results obtained in this laboratory which provide evidence for the extension of the crystallites during heating in O₂ followed by contraction during subsequent heating in H₂, as well as on their splitting and change in shape.

The average metal crystallite size in supported metal catalysts usually increases during use. This increase in average metal crystallite size, i.e. decrease in metal surface area, is one of the causes of catalyst deactivation. For some supported metal catalysts the average metal crystallite size in deactivated catalysts can be decreased by appropriate regeneration procedures. The processes by which increases in metal crystallite sizes occur is referred to as sintering, and processes which result in decreases in metal crystallite size are called redispersion.

In this paper, data obtained with conventional supported metal catalysts, i.e. metals supported on high surface area porous carriers, are examined with the aim of determining the mechanisms of sintering and redispersion. The systems examined will be restricted to platinum metals on carriers which do not react with the surrounding atmosphere, e.g. metals supported on carbon will not be included in the discussion. The data strongly supports the hypothesis that redispersion and sintering for these non-reacting systems occurs via transport of atomic or molecular species, and not by migration of entire metal or metal oxide particles.

A number of highly sinterable ceramic powders have been rapidly sintered by insertion of small samples into preheated furnaces, rapid heating of small specimens in low thermal mass furnaces, and passing tube-shaped samples through short hot zone furnaces. More recently, aluminum oxide has been sintered by passing through gas plasmas. Heating rates in the neighborhood of 100 ºK/s and densification rates >1%/s have been achieved in the plasma sintering. Sintering models and computer simulation shed some light on the effect of rapid heating on the various sintering mechanisms and the interplay among the sintering mechanism.

Gamma and alpha alumina powders having an average particle size of 9.6nm and 20nm respectively were sintered isothermally and at constant heating rates. With increase in heating rate, the gamma-alpha phase change occurred at higher temperatures, and densification of pellets originally made of gamma powder was enhanced. Pellets made of alpha powder were not sensitive to change in heating rates. Initial densification was interpreted in terms of Coble creep and grain boundary diffusion of oxygen. In gamma pellets from 1400°C to 1800°C this mechanism was aided by dislocation climb, which reduces the activation energy by approximately 0.57 of the actual value of 441 kJ/mol calculated from shrinkage of alpha pellets at 1200°C to 1800°C. The dislocation climb was caused by transformation of gamma to alpha phase, so gamma pellets densified more (90% theoretical) than alpha (70% theoretical) at 1800°C.

The use of highly dynamic shock waves to alter ceramic and metal powders to enhance sinterability has been described in the literature for a variety of materials (1–4). Recently, a group of target materials has been intensively investigated, with the use of alumina as one of these focus materials here at North Carolina State University (5–8).

The densification of SrTiO₃ using LiF as a sintering aid was investigated. In the temperature range 700 to 900°C, the sintering kinetics of ½ to 2w% LiF doped SrTiO₃ appears to be controlled by liquid phase sintering. Densification was found to be a function of cation stoichiometry and temperature as well as LiF content. For example, 2w% LiF additions with 4 hour sintering at 910°C, specimens of composition Sr_1.03TiO₃ achieved a density of 96% TD, whereas, for the same treatment of those of compositions Sr₉₈ TiO₃ achieved only 75% TD. The interaction between LiF and SrTiO₃ is very complex and both X-ray diffraction and weight loss measurements suggest that it might involve the formation of Li and F containing compounds.

The usefulness of traditional models to establish the sintering mechanisms in complex systems typical of industrial practices is questionable since the systems usually consist of many particles which are nonuniform in size and shape. In such systems many mechanisms of mass transport can be active. Other complications such as grain growth and particle rearrangement which are even observed in studies involving uniform spherical particles¹ make interpretation of shrinkage during sintering even more difficult.

The sintering of combustion-synthesized titanium carbide was investigated relative to that of commercially available powders. Depending on the preparation parameters, the synthesized powders had variable shrinkages at the same sintering conditions. Moreover, the commercial powder experienced a higher densification than any of the powders prepared in this study. These observations are interpreted in terms of the influence of free carbon on the sintering process. Approximately calculated amounts of free carbon show an inverse relationship with the measured shrinkages. Experimental verification for this is provided by sintering data on carbon-doped commercial powders of titanium carbide.

The activation energies for sintering were determined as 390 ± 26 and 458 ± 13 kJ.mol^-1 for the commercial and combustion-synthesized powders, respectively.

Cr and Mn are very difficult to sinter into dense compacts under normal sintering conditions presumably because of their ease of oxidation. In addition, the relatively high vapor pressures of these two metals, limits the use of high sintering temperatures. In this study, it was attempted to obtain dense compacts of Cr and Mn by utilizing activated sintering through the addition of Group VIII transition metals, which has been well established for the sintering of a number of refractory metals [1] – [15]. Ni and Pd were selected as the most promising activators [14] for the sintering enhancement of Cr and Mn powder compacts. The effect of the quantity of additive and of the particle size of the sample powders on the sintering densification has been investigated as a function of sintering temperature.

The densification by the reactive phase calsintering of a compact formed from an oxide precursor is presented. The essential features of the process are the development of a reactive, high surface area matrix upon calcination of the precursor and the presence of a reactive liquid phase during sintering. The densification process is described for Fe₂O₃ doped dolomite in terms of the elimination of intra-aggregate ana interaggregate porosities via liquid phase sintering. It is demonstrated that high densities (≥95%) and rapid sintering kinetics can be obtained with coarse (>10μm) precursor powders.

From our previous investigations ^1–4, we have come to the conclusion that the consolidation of powder may have to be considered as a unique process. According to our results⁵ we can comprehend materials consolidation, i.e. sintering as a unique process in which the relative parameter of the system may be changed. These parameters, define the change in materials activity during the sintering and pressing processes⁵ and are presented schematically by the block diagram in Fig. 1.