Why Particles Stick Together

The Problem
Powdered titanium dioxide (or titania) is used widely as a pigment and in other applications. Titania exists naturally as one of three crystalline phases (or polymorphs): rutile, anatase and brookite. Made by the ‘chloride’ process, it is critically important to achieve a specific mean particle size and particle size distribution to optimise the product’s light scattering properties. The process involves a high temperature (1500-2100K) flame reactor in which gaseous titanium tetrachloride combines with gaseous oxygen to produce titanium dioxide and chlorine gas. Practical studies have empirically linked temperature and the use of additives to particle size effects and phase stability. However, these studies are limited by their inability to unravel the competing effects of the three sub-processes, nucleation, growth and coagulation.

The Challenge
A detailed understanding of the sub-processes was required if the complex reactor environment was to be modelled successfully. Using computer simulation techniques it is possible to isolate the effects of the three sub-processes. The current poor insight into the atomic processes underlying the bulk behaviour is a limitation in other possible simulation approaches such as computational fluid dynamics (CFD). The possibility to develop such parameters using molecular dynamics (MD) modelling of the sub-processes was a major challenge for this investigation aimed at understanding the fundamentals linking optimal reaction conditions to powder phase, particle size and size distribution.

The Solution
The Daresbury MD code DL-Poly was used to study the thermodynamic, structural and transport properties of micro-clusters of rutile titania at a range of temperatures typical of the chloride process - the first time that this had ever been attempted. Simulations were performed on 1245 atom clusters between 1000 and 3000K. Rutile is shown to be the stable phase and the coagulation process was shown to be long timescale (ns) and markedly influenced by surface ion diffusion. Atoms at or near the surface exhibit high diffusion rates which is important in cementing particles together. The temperature dependence of coagulation observed experimentally is not found in the simulation for pure rutile nanoclusters. This suggests strongly an involvement of chlorine adsorption and desorption on the titania clusters leading to their observed ‘stickiness’ in the manufacturing process.

The Benefits
The originality of the simulation on ceramic nanoclusters has potential benefits in modelling the more general process of powder sintering by which powders are fused to create a solid mass at high T and P.
The customer has gained a unique insight into the fundamental chemistry of a complex process which indicates a vital area for further study into the thermochemistry of chlorine adsorption and desorption on titania clusters.