Synthesis of Ceramic Reinforcements in Metallic Matrices during Spark Plasma Sintering: Consideration of Reactant/Matrix Mutual Chemistry

Abstract

Metal–ceramic composites are obtained via ex-situ or in-situ routes. The in-situ route implies the synthesis of reinforcement in the presence of a matrix and is often regarded as providing more flexibility to the microstructure design of composites than the ex-situ route. Spark plasma sintering (SPS) is an advanced sintering method that allows fast consolidation of various powder materials up to full or nearly full density. In reactive SPS, the synthesis and consolidation are combined in a single processing step, which corresponds to the in-situ route. In this article, we discuss the peculiarities of synthesis of ceramic reinforcements in metallic matrices during SPS with a particular consideration of reactant/matrix mutual chemistry. The formation of carbide reinforcements in Cu, Al, and Ni matrices is given attention with examples elaborated in the authors’ own research. Factors determining the suitability of reactive SPS for manufacturing of composites from a matrix/reactants system and features of the structural evolution of the reaction mixture during sintering are discussed.

1. Introduction

Composite materials containing ceramic compounds can be manufactured via ex- situ [1,2] or in-situ routes [3,4,5]. The ex-situ route is based on adding a previously synthesized compound to the matrix. Alternatively, the reaction can be conducted in situ in the presence of a matrix (which is in the solid or liquid state). The in-situ approach often provides more flexibility in terms of the microstructure tailoring and design of composites, as reactants of different chemical nature and particle sizes can be used. This increases the number of variable parameters during the processing of a composite. The in- situ approach is suitable for synthesizing particles of specific morphologies [6,7].

The authors of this article have studied the formation features, microstructure, and properties of in-situ metal–ceramic composites formed by spark plasma sintering (SPS) [8,9,10,11]. Those studies dealt with the synthesis of carbide reinforcements in the presence of a metal matrix. Below, the peculiarities of the synthesis of ceramic reinforcements in metallic matrices during SPS are shown with consideration of reactant/matrix mutual chemistry. This article was conceived as a discussion of different systems to formulate practical recommendations pertaining to reactive SPS of metal–ceramic composites. Factors determining the suitability of reactive SPS for manufacturing of a composite from a matrix/reactants system are discussed. Based on the authors’ previous research, examples of the in-situ composite structures formed via reactive SPS are presented.

2. Advantages and Limitations of Spark Plasma Sintering for the Production of In-Situ Metal–Ceramic Composites

SPS allows fast consolidation of various powder materials up to full or nearly full density [12]. The attention of researchers to reactive SPS is due to the possibilities to synthesize and consolidate the material within a single processing step [13]. Fast heating of the tooling/sample assembly enabled by the use of pulsed electric current makes SPS especially feasible for the formation of materials with fine sizes of structural elements, usually in the range from several nanometers to several micrometers.

SPS tooling that is commonly used (Figure 1) was designed for solid-state sintering, which allows maintaining the inner structure of the powder particles while avoiding extensive crystallite/grain growth. As the powder mixtures are processed in a rigid die, it is necessary to avoid the formation of large quantities of liquid, as this complicates the process from a technological viewpoint. The removal of the sample from the die becomes a serious issue when the liquid spreads between the punches and the die walls and solidifies there upon cooling. Chemical interaction of the liquid with the tooling material leads to the product contamination by carbon and a reduction in the tooling service life. Therefore, the inability of the SPS devices to process samples, whose formation is associated with the formation of large quantities of liquid, is an inherent limitation and needs to be taken into account in the industrial practice.

As the formation of ceramic reinforcements occurs via exothermic reactions, the heating method and heating rate of the mixtures are important for the outcome of the synthesis [14,15]. In the SPS processing of exothermic mixtures, the reaction heat can influence the microstructure formation of the product. On the one hand, the reaction heat is dissipated through the walls of the die and punches. On the other hand, the temperature can rise locally in the areas of inter-particle contacts, as demonstrated in [8,9], owing to the effect of electric current. The heat generation and dissipation processes need to be considered when different processing options of materials by SPS are attempted [16].

3. Considerations of Reactant/Matrix Chemistry and Structure of the Reaction Mixture

The key points of the discussion presented below are summarized in

Table 1. Characteristics of A-B and A-B-C systems (reaction mixtures of elemental powders), examples of metal–ceramic composites to be produced from these mixtures, and factors influencing the synthesis completeness (formation of the desired reinforcing phase).

No.	Binary A-B/Ternary System A-B-C	Characteristics of the Reaction Mixtures A-B and A-B-C	Example of Reaction Mixture and Metal–Ceramic Composite to Be Formed	Factors Influencing the Synthesis Completeness
I	A-B	AmBn compounds can form	Ti–C (Ti–TiC composite)	- initial particle size of B - sintering time - sintering temperature
II	A-B-C	Both B and C are insoluble in solid A	Cu–W–C (Cu–WC composite)	- B/C initial contact area - sintering time - sintering temperature
III	A-B-C	Both B and C are insoluble in solid A, AmBn and AxCy compounds can form	Al–Ti–C (Al–TiC composite)	- B/C initial contact area - sintering time - sintering temperature
IV	A-B-C	B is soluble in solid A, C is insoluble in solid A	Cu–Ti–C (Cu–TiC composite)	- B/C initial contact area - generation and “activation” of diffusion paths for B in A (when diffusion of B through A is a rate-limiting step) - sintering time - sintering temperature
V	A-B-C	Both B and C are soluble in solid A	Ni–W–C (Ni–WC composite)	- B/C initial contact area - generation and “activation” of diffusion paths for B or C through A (when diffusion of B or C through A is a rate-limiting step) - sintering time - sintering temperature

. The matrix of a composite is designated as A. Let us first consider binary reaction mixtures A-B (case I). In the case of compounds containing the element forming a metal matrix, the distribution of the other (non-metallic) component in the reaction mixture determines the structure of the composite [17].

Mechanical milling of powder mixtures can assist the SPS processing of metal–ceramic composites by facilitating mixing of the reactants, shortening the diffusion distances, generating new crystallite/grain boundaries, and refining the structure of the final product [18]. For example, in the binary Ti–C high-energy milled powder mixture of the stoichiometric composition, TiC with crystallites about 100 nm in size was obtained after heating the mixture up to 900–1000 °C and holding it for 30 min at the maximum temperature [19]. When TiC is synthesized in a mixture not subjected to high-energy milling, the size of TiC crystals is of the order of several micrometers [20].

The formation of a transient liquid (contact melting) [21] is possible in powder mixtures not subjected to pre-alloying (alloying before sintering). This liquid phase disappears as it reacts with the solid phase, unless the composition corresponds to the eutectic alloy. The formation of a liquid phase facilitates the occurrence of chemical reactions. If the system is fully alloyed prior to sintering, the liquid phase forms at a temperature predicted by the corresponding phase diagram for the mixture composition.

In a ternary A-B-C mixture, in which B and C are to form a compound, the transformation to occur is more complex, as B and C reactants have to combine in the presence of a matrix acting as a separating medium or a barrier. There are four situations depending on the chemical properties of the components (cases II, III, IV and V in Table 1). Case II describes a situation, in which neither B nor C can form solid solutions with A (matrix). For the reaction between B and C to occur, a direct contact of the particles is required, while the matrix particles act as a diffusion barrier. An example of such systems is Cu–W–C with WC/W₂C carbide reinforcements to form within a Cu matrix [22].

If the goal is to synthesize TiC in an Al matrix (as described by case III), the traditional SPS under pressure does not appear to present the best solution. In our experiments, a Ti–C–3Al mixture was mechanically milled and processed by SPS. Ti and Al form thermodynamically stable compounds. The TiAl₃ phase formed after SPS at 600 °C together with certain amounts of Al₄C₃. Holding the sample for 30 min at this temperature did not result in the formation of the target phase (TiC). It was shown that a temperature exceeding the melting point of aluminum is necessary to induce the interaction of TiAl₃ with carbon [23]. When Ti–C–Al mixtures containing 10–50 wt.% of Al are ignited, the formation of TiC–Al composites occurs in the mode of self-propagating high-temperature synthesis [24]. In our studies, we have found that it is possible to synthesize TiC in an Al-rich reaction mixture (Ti–C–3Al atomic composition, 57 wt.% of Al) via a two-step processing: SPS at 550 °C (not allowing extensive interaction of Ti and Al to occur) followed by fast heating of the compact up to 800 °C (Figure 2a,b). The product of annealing was a porous two-phase (bulk) composite, Al–TiC (Figure 2c). The second processing step was pressureless sintering in vacuum. After the experiment was completed, an Al residue was found on the walls of the container holding the sample.

In case IV, B forms A(B) solid solutions, while C does not dissolve in A, as in the Cu–Ti–C system. The synthesis of TiC in the presence of copper has been described in [25]. Figure 3a shows the microstructure of a Ti–C–3Cu agglomerate formed after 5 min of high-energy milling. In the powder agglomerates obtained by milling, dark-gray stripes correspond to titanium embedded in the Cu–Cu(Ti) matrix; the latter is light gray. The structural changes caused by diffusion are indicated by red arrows in Figure 3b; gray areas now surround the dark-gray stripes. Upon heating, titanium diffused into copper and, where possible, reacted with carbon after a certain temperature was reached. The microstructure of the fully reacted TiC–Cu composite obtained by SPS at 900 °C from the Ti–C–3Cu mixture milled for 5 min is presented in Figure 3c. As the milling time of the mixture increases, the shape of the thermal explosion thermogram changes, as seen in Figure 4a [25]. The partial formation of the reaction product at the milling stage reduced both the ignition and the maximum temperatures developed in the system upon thermal explosion. The formation of large quantities of liquid (molten copper) was avoided when milling (for 5 min and longer) prior to SPS was used (Figure 4b). It should be noted that melting at the inter-particle contacts caused by electric current passage is local and does not cause any technological complications.

When both B and C form solid solutions with A (case V), they can diffuse within the matrix that separates them. The matrix, therefore, is not a diffusion barrier. An example is a Ni(W) alloy obtained by mechanical alloying and sintered in contact with graphite foil by SPS [10,11]. The subsurface layers of compacts sintered from the non-milled and mechanically milled Ni–W mixtures had different microstructures. In the case of non-milled mixture, carbon reacted with the very surface layers of the material only. In the compact sintered from the milled mixture, particles of WC were found at a distance of about 100 μm from the interface with the graphite foil (a layer of the WC–Ni composite formed). The formation of WC particles within a 100-μm layer was due to diffusion of carbon into the Ni(W) alloy facilitated by the developed network of grain boundaries in the mechanically milled alloy.

4. Practical Recommendations for Implementing the Synthesis of Metal–Ceramic Composites via SPS and Directions for the Future Research

The following practical recommendations can be formulated when a metal–ceramic composite is to be obtained via reactive SPS. First, a possibility of forming a reinforcing phase of the desired composition needs to be considered. For that, the elemental composition of the reaction mixtures should be properly selected. Second, a question of the suitability of the system for being processed by reactive SPS needs to be raised. Is it possible to conduct the synthesis of a reinforcing phase while keeping the matrix in the solid state? Will there be sufficient time for the reaction diffusion to take place and what temperatures are required for that? Third, methods to improve the distribution of the components of the reaction mixture and those to generate/“activate” the diffusion paths in the matrix (generation and structural modification of grain boundaries) need to be applied.

Reaction mixtures to be processed into metal–ceramic composites by SPS should meet the following requirements (by chemistry and structural state):

(1)

in order to fully use the potential of SPS for the production of fine-grained materials, the synthesis reaction should be fast enough (should be completed within minutes) (if the reaction requires prolonged holding for its completeness, grain growth of the matrix can ensue);

(2)

upon the reaction, only a limited amount of liquid should form (extensive formation of liquid should be avoided).

To fulfill these requirements, after selecting a proper chemistry of the mixture, mechanical milling can be used to improve the distribution of the reactants and reduce the synthesis temperatures.

In the future research, the grain structure of the inter-particle (inter-agglomerate) contact areas that experienced overheating (and, possibly, melting) during SPS, and compositional differences between those and the particle (agglomerate) volume, would be of particular interest. It is also necessary to determine the specific characteristics of metal–ceramic composites obtained by SPS in relation to materials produced by other methods.