Nitrogen and silicon diffusion in the Si3N4 crystal lattice

Get Complete Project Material File(s) Now! »

Chapter 2: Review of production of silicon nitride bonded silicon carbide (SNBSC) by reaction bonding.

Silicon nitride bonded silicon carbide (SNBSC) refractory bricks are produced commercially by a reaction bonding technique with silicon nitride (RBSN) as the binder phase. Unlike pure silicon nitride products, silicon carbide grains and an organic binder are added to silicon powder to form a compact. The mixture is shaped into a ª greenº brick by vibration and pressing (Figure 2-1).
After shaping, the green bricks are stacked onto trolleys, inserted into a controlled atmosphere furnace (Figure 2-2) and heated. Initially the bricks are heated under normal atmospheric conditions with temperature up to 400°C, in order to burn out the binder. Then the temperature is raised to1100-1450°C and the materials are held at this temperature in a nitrogen atmosphere while an exothermic nitridation reaction -bonding occurs (Equation 2-1)[25].
Equation 2-1 3Si (s) + 2N2 (g) = Si3N4 (s) H(1100°C) = -723.3 kJ/Mole
The heat generated due to the exothermic nature of the nitridation reaction should be considered when calculating the total heat balance of the nitridation process. Failing to include this additional heat in the overall heat balance, could lead to increased temperature above the melting temperature of silicon (1410°C) in the ª greenº bricks. This will cause premature local melting of the silicon powder, reduction of the surface area exposed to nitrogen, and complete nitridation will be impossible, leaving unreacted silicon in the final product. Temperatures above the 1500°C, and the presence of molten liquid silicon, will make the conversion from to Si3N4 feasible resulting in Si3N4 rich bonding phase.
SNBSC refractories are usually fabricated from ª commercial gradeº silicon powder with a typical mean particle size varying between 10-25 µm and a green bulk density of 1500-1600 kg/ m3 [25]. An unusual and beneficial feature of the RBSN fabrication route is the fact that the overall dimensions of the ª greenº compact do not change during ª firingº , despite a 22% increase in volume on conversion of Si to Si3N4. A close tolerance (approximately 0.1%) on the dimensions of a finished component can readily be maintained, thereby avoiding expensive machining after firing. The reason for this phenomenon lies in the fact that the only reduction in void space in the compacted powder is the consequence of the molar volume expansion on conversion of silicon to silicon nitride. There is no sintering in the normal sense and the volume expansion is accommodated entirely within the pore structure of the nitriding powder. The density of Si and Si3N4 are 2.32 and 3.19 mg/m3 respectively and their molar volumes are 0.363 cm3 /mol for the Si and 0.440 cm3 /mol for the Si3N4 [21].
The nitridation of a compact involves the following two steps:
1. Movement of nitrogen into the continuously changing pore system of the compact.
2. Chemical reaction between silicon and nitrogen.
The nitridation process can proceed via two different mechanisms which will then determine the type of product produced[21, 25]:
• The dominant process is a vapour phase reaction which involves vaporization of silicon and a vapour phase reaction with nitrogen. The vapour pressure of Si at 1350°C is 10-7 atmospheres and it implies a silicon evaporation rate of 10-6kg/m2sec1 [25]. The supply of silicon vapours and reaction with nitrogen according to Equation 2-2 is in accordance with the observed reaction rates.
Equation 2-2 3Si (l) + 2N2 (g) = Si3N4 (s)
This process is a typical chemical vapour deposit (CVD) reaction, which lead to formation of Si3N4. The reason Si3N4 phase dominates in products from the vapour phase reaction could be related to the lack of freedom of the species to order themselves into the energetically favoured Si3N4 structure at the growth site. The independent mobility of the reactants in this process is restricted when precipitation occurs in a host solid.
• A vapour liquid solid (VLS) reaction process which involves solution of nitrogen in liquid silicon alloys that are formed in the presence of contaminations such as Al, or Fe (commercial silicon powders can contain 0.9wt% Fe). The impurities will reduce the formation temperatures of liquid phase to below 1207°C creating the condition to formation of Si3N4.
• A minor process, occurring in a presence of high-purity silicon, is the adsorption/reaction, in which nucleation of silicon nitride on a solid silicon surface leads to formation of a coherent layer of silicon nitride (mostly Si3N4). The formation of this layer is rapid when PN2 = 1 atmosphere. This nitridation process follows several stages which control the kinetics of the reaction:
1. The first stage involves the formation of Si3N4 nuclei on the Si surface, followed by growth of these nuclei. This is thought to be initiated by the reaction between chemisorbed nitrogen and silicon, the latter arriving at the reaction site by a combination of surface diffusion and an evaporation/condensation process (in this stage, formation of Si3N4 could also proceed according to the CVD process mentioned above) (Figure 2-3). During this stage the kinetics are linear with time.
A few parameters influence the nitridation process:
1. Silicon nitridation is a temperature sensitive and strongly exothermic reaction (Equation 2-1). Thus overheating could lead to premature melting of the silicon and reduction of its surface area. In contrast, if the temperature is kept below 1400°C, the nitridation reaction of conventional silicon powder become very slowly and ultimately effectively stops; completion of the nitridation thus becomes very difficult.
2. Nitrogen and silicon diffusion in the Si3N4 crystal lattice is too slow to give significant rates of nitride formation.
3. The nitridation reaction bonding process is extremely sensitive to impurities in the silicon powder and/or in the furnace environment. Reproducibility of product quality is hard to ensure.
During the nitridation process the temperature in the core of the brick can easily rise above the melting point of Si due to the heat generated from the exothermic reaction (Equation 2-1) and the lack of heat dissipation from the brick core to the exterior surface. This can happen even at heating temperatures of 1100°C and it will determine the properties of the finished refractory. Thus in this work, a computer model that estimates the temperature profile in the brick cross section during the nitridation reaction has been developed using the following assumptions:
Ambient (furnace) temperature: 1100°C, Thermal conductivity at 1100°C= 20 W/(m*K), Si conversion rate: 4%/hour [25], Nitrogen velocity at the brick surface 2m/sec.
As seen in Figures 2-6 and 2-7, the temperature on the outer surface of the brick could reach 1442 ëC and 1452 ëC in the core due to the heat generated by the exothermic reaction and the poor heat dissipation from the core of the brick to its exterior surface. Temperature overshoot is generally controlled in practice by controlling the rate of nitrogen supply [25]. However, it is expected that during the nitridation process the heat dissipation from bricks which are surrounded in the furnace by other bricks may give rise to highly inhomogeneous surface cooling. Hence the temperatures in the core of the brick could get significantly higher than predicted by the model.
Temperatures above the silicon melting point (1410°C) would lead to premature local melting of the silicon powder, particularly in a poorly mixed green block, and will form of an amorphous glassy phase thus reduce the surface area exposed to nitrogen. Excessive heat also influences to Si3N4 conversion, that could occur in temperatures as low as 1450°C [26] and will result in formation of Si3N4 especially in the hotter parts of the brick.
To overcome the overheating problem, the nitrogen supply can be constrained to allow initial nitridation to build up an adequate restraining, skeletal microstructure of bridging Si3N4 [21], which will then limits coalescence of the molten silicon particles. Only then is the temperature increased in order to accelerate and complete the nitridation reaction. This is usually achieved using multi step time-temperature furnace control, with a program which consists of a slow ramp to 1300°C followed by a holding time until the reaction rate has virtually ceased and than followed by a final period at 1450°C to complete the conversion of the silicon to nitride. The reason for the step increase in the nitridation rate with increasing temperature has not been satisfactory explained [21].

Chapter 1: Introduction to aluminium reduction cell structure and properties
1.1. Introduction to aluminium reduction cell structure and operations
1.1.1. Aluminium reduction cell structure
1.1.2. Cell heat balance
1.1.3. The reduction cell sidewalls
1.1.4. Summary
1.2. Review of sidewall refractory materials for aluminium smelting.
1.2.1. Carbon based sidewall materials
1.2.1.1. Types and properties
1.2.1.2. Wear mechanisms of carbon sidewall
1.2.2. Alternative materials for sidewall
1.2.2.1. Types and properties of SiC based materials:
1.2.2.2 Wear mechanism of SiC-based refractories
Chapter 2: Review of production of silicon nitride bonded silicon carbide (SNBSC) by reaction bonding.
2.1. Nitrogen and silicon diffusion in the Si3N4 crystal lattice
2.2. The effects of contamination in the nitridation system
2.3. / Si3N4 ratio and conversion of to Si3N4
2.4. Summary
Chapter 3: Previous studies of the oxidation and corrosion resistance of Si3N4 bonded SiC
3.1. Introduction
3.2. Laboratory scale corrosion test methods
3.2.1. Oxidation tests in air and CO2/O2 atmospheres
3.3. Corrosion mechanisms
3.4. The contribution of different parameters on the corrosion rate
3.5. Conclusions
Chapter 4: Microstructural analysis
4.1. Crystal structures of and Si3N4
4.2. Quantitative analysis of and Si3N4 using X-ray diffraction
4.3. Solid state NMR analysis
4.3.1. Introduction
4.3.2. Experimental method
4.4. Raman spectroscopy
4.5. X-ray photoelectron spectroscopy (XPS)
4.6. SNBSC material morphology
4.6.1. SEM analysis
4.7. Summary
Chapter 5: Experimental methodology for corrosion testing
5.1. Quantifying volume loss using the ISO 5017 standard method
5.2. Lab scale corrosion testing with electrolysis
5.2.1. “Half immersed” configuration
5.2.2. Electrolysis corrosion experiments using a rotating anode
5.2.3. Gas phase corrosion test configuration
5.3. CO2 sparging experiment
5.4. Summary
Chapter 6: Corrosion of SNBSC refractories in an aluminium reduction cell environment
6.1. Industrial scale trial
6.2. Laboratory scale corrosion experiments
6.2.1. Effect of porosity in SNBSC bricks
6.2.1.1. The effect of porosity on corrosion
6.2.2. Summary
6.2.3. Effect of SiC/Si3N4 ratio on the corrosion rate of SNBSC bricks
6.2.4. Effect of / Si3N4 ratio on the corrosion rate of SNBSC bricks
6.2.4.1. Corrosion rate as function of Si3N4 content
6.2.5. The effect of free silicon on the corrosion rate of SNBSC bricks
6.2.6. The effect of environment -bath phase vs. gas phase attack
6.3. Alteration of SNBSC samples in the gas phase, during electrolysis tests
6.3.1. Thermal treatment of SNBSC samples
6.3.2. Morphology changes
6.3.3. Electrolysis corrosion test
6.4. Summary
Chapter 7: Mechanisms of degradation of SNBSC refractories
7.1. Reactivity of sidewall materials to reduction cell atmosphere
7.2. Hypothesis – vs. Si3N4 and the influence of morphology on corrosion rate
7.3. Discussion of the laboratory scale results
7.3.1. Proposed mechanism of degradation
7.4. Summary of mechanism of degradation of SNBSC
Chapter 8: Summary and conclusions
8.1. Conclusions from this study
8.2. Suggestions for future work.
References
GET THE COMPLETE PROJECT