Agglomeration of Titanium Dioxide (TiO2) particles

1 June 2021

Mihai Gherghe, MSc and Sandra J. Veen, PhD

Agglomeration of primary particles is a well known phenomenon occurring in particle suspensions. When designing a new formulation, it can often make the difference between a good quality product and a bad quality one. Depending on the properties of the particles and the continuous phase, the formation of agglomerates can lead to poor dispersion of particles and ultimately, sedimentation or creaming. Some well known options to increase suspension stability include adding a dispersant, using a different particle size distribution, choosing a different particle shape or opting for a surface treated particle. To determine the most feasible prevention strategy, however, the cause of the clustering of primary particles needs to be understood. 

A well known example of particle agglomeration is the case of TiO2 in water. Titania is a widely used material in multiple industries including paints and coatings, cosmetics and paper industry. Regardless of the purpose, a high dispersion efficiency in the wet state is needed in order to achieve the best formulation properties. Due to its wide applicability, we would like to use titania as an example to demonstrate how RheoCube can help in gaining more insight in agglomeration processes in particle suspensions.

Why simulations?

Typical techniques used to study agglomerate structure and dimensions include scanning/transmission electron microscopy (SEM/TEM) and light and X-ray scattering. Those techniques require extensive time for set-up and analysis, specialized operators and a substantial investment in the actual devices. There are also limitations to their applicability. Conventional SEM/TEM only offer, at best, a semi-three-dimensional perspective on agglomerates [1]. Moreover, the need for dried samples introduces the uncertainty that the images viewed may not represent the structures in solution. Simple light scattering measurements of suspensions are often limited to lower particle concentrations, and hindered by sedimentation and ongoing aggregation. There are very advanced experimental measurement methods that can give more information (e.g. cryo-SEM/TEM, tomography, near field scattering). But access to these highly specialized set-ups is scarce and obtaining good results is very time and labor consuming, if it succeeds at all. 

In applications such as the assembly of colloidal particles, simulations can offer fast and reliable insights and can aid in understanding aggregation phenomena. RheoCube offers the possibility to design your system according to your needs. It also offers the option to observe and analyse the agglomeration process as it naturally occurs in the sample. There is no need to adjust or prepare the system for measurement. With the titania example we will demonstrate what can be learned about agglomeration from using simulations in general and RheoCube in particular.

Titania as an example

There are many factors that influence the extent of agglomeration in a particle suspension. Examples include e.g. particle size (distribution), particle shape, surface area, electrostatic interactions. For the titania example, we used RheoCube to study the effect of particle shape on the extent of agglomeration, by comparing two systems: one containing perfectly spherical particles and one containing slightly dented, or rough particles. For more information on RheoCube’s settings for creating particles of different shapes, check out this blogpost

The systems we worked with consist of 20 volume percent TiO2 particles of two different shapes, suspended in water. The diameter of the particles is 400 nm. To encourage particle contact and thus faster agglomeration a shear rate of 1000 s-1 was applied to the suspensions. The agglomeration tendency is partly due to strong attraction between TiO2 particles. Particle adhesion is also encouraged by the relatively poor compatibility between TiO2 and water (bulk phase). In RheoCube, particle-particle and particle-fluid compatibility are based on HSP theory.

Rough Particle

Rough Particles

The two videos and accompanying images here (above and below) show a full 3D visualization of the two different TiO2 systems as they are being sheared.

Smooth Particle
Smooth Particles

The rough particles seem to lock into a more fixed position within an agglomerate once they come in contact. In the lower part of the rough particles simulation box, an agglomerate can be observed rotating along the direction of the shear flow, suggesting that interparticle attractions are strong enough to withstand agglomerate breakup due to shear.

The details of agglomeration

Visualization alone can offer a great deal of information but quantifying agglomeration becomes a necessity when comparing different cases with apparently similar results. RheoCube’s data analysis capabilities offer a multitude of tools for quantitative analysis. For this study we have chosen to analyse agglomeration by comparing the number of clusters formed under shear and their size at an initial and final time. 

TiO2 agglomerates of spherical particles (left) and rough particles (right), formed after 2.25 ms, visualized and analyzed via the Cluster Analysis widget. The inset photos are visualizations of agglomerates colored by cluster groups. The bar charts depict the number of clusters formed and their volume. Green bars represent clusters at the beginning of the simulation (t0) while purple bars represent clusters after 2.25 ms.

As a general rule, the decrease in number of clusters indicates a more extensive agglomeration of particles. The total number of clusters composed of spheres as primary particles remains the same: 33 clusters found at the beginning and the end of the simulation. The changes are more visible when looking at cluster volumes. By the end of the simulation, the largest cluster loses 2 primary particles down to 7 particles. At the same time, the number of two- and three-particle clusters increases. Primary particles stick together to form small agglomerates, while already formed larger clusters break into smaller ones. This suggests that the attractive energy between clustered particles is lower than the energy resulting from shearing forces.

On the other hand, cluster break-up is not observed in the case of rough particles. After 2.25 ms, the total number of clusters decreases by ~60% and a more significant drop in the number of low volume clusters can be observed. Also, the largest cluster volume is ~5 times higher than in the case of spherical particles. Due to the increase in particle surface area resulting from the addition of roughness [2], the contact interaction energy between two particles touching is increasing as well. As a result, rough particles are more susceptible to forming larger and more stable clusters than spherical particles.  


Agglomeration of particles is a complex phenomenon, dependent on multiple aspects. One of those determining factors is the roughness of the primary particle. A slight degree of denting in the TiO2 particle surface can lead to larger agglomerates that are harder to break, compared to spherical, smoothed surfaces. Such conclusions can help researchers in designing smarter and more cost effective formulations, without the need of specialized analysis techniques which are often limited in result interpretation or hardly accessible. While RheoCube’s complex models can efficiently account for complex phenomena such as agglomeration, it also offers a multitude of intuitive options for data interpretation and analysis.


[1] Koeylue, U., Xing, Y., & Rosner, D. E. (1995). Fractal morphology analysis of combustion-generated aggregates using angular light scattering and electron microscope images. Langmuir, 11(12), 4848-4854.

[2] Tanner, R. I., & Dai, S. (2016). Particle roughness and rheology in non colloidal suspensions. Journal of Rheology, 60(4), 809-818.