Cluster–cluster aggregation controlled by the number of intercluster connections: kinetics of aggregation and cluster mass frequency

https://doi.org/10.1016/j.jcis.2004.06.010Get rights and content

Abstract

The aggregation of colloids in the presence of hydrodynamic forces was investigated, employing a numerical model that took into account the masses of the individual clusters and the number of intercluster connections established when two clusters stuck together. The number of possible connections was determined by analyzing all the possible nonoverlapping configurations of stuck clusters. This operation was done for a couple of clusters of various masses, taking into account the assembly of clusters of even and uneven masses. The formulation of the constraints established a certain hierarchy in the sticking on a basis compatible with the irregular fracture model of Horwatt and co-workers. As a result, the permanent sticking of large clusters required the formation of a large number of connections, whereas that of small clusters might be realized even with a small number of connections. Thus, the aggregation started with the features of the standard reaction-limited process and this cluster growth became progressively inhibited as a result of the prevailing effects of the connection constraints. The cluster-mass frequency showed the emergence at least of a second population whose bell-shaped mass distribution was superimposed on the monotonically decreasing distribution resulting from the reaction-limited aggregation process. The results of the numerical study were confronted with those previously obtained in the aggregation of hydrated polystyrene latex particles dispersed in 1 M sodium chloride solution. The two striking features—the aggregate growth kinetics and the mass distribution function—were common to the computer-generated clusters and the latex aggregates.

Introduction

Despite the fact that aggregation (and fragmentation) of colloidal particles has been the subject of extended and diverse experimental, numerical, and theoretical investigations [1], [2], [3], [4], the aggregation process controlled by the internal cohesion of the agglomerate has only been invoked to interpret recent experimental results [5]. This special aspect and its effect on the kinetics of aggregation and on the cluster-mass-frequency function were taken into account in a numerical investigation. In our opinion, looking at the control of the internal cohesion of newly forming clusters may address various experimental situations, and especially the two following cases.

The first case corresponds to aggregation in the presence of hydrodynamic forces, which is very important from a practical standpoint. As pointed out by Mason [6] and Gregory [7], two particles approaching each other in a shear field form a doublet that can be either transient or permanent. In the former case, the doublets rotate through a certain angle before fluid forces separate the particles. This image may be adapted to two colliding clusters. One may imagine that the breakup issue is relatively reduced when the encounter between the two clusters immediately induces a great number of intercluster connections. Conversely, when the two clusters are only connected by a single bond, the sticking certainly will be transient [8]. Regarding the given conformation of the stuck clusters, the knowledge of the number of intercluster connections is thus of major interest for the determination of the evolution of the sticking toward rupture or permanent consolidation.

The second case is relative to systems of marginal instability such as hairy or hydrated colloids and colloidal materials for which intercolloid attraction forces are greatly reduced due to the small value of their Hamaker constant, such as low-density biological materials [5], [9], [10]. One may imagine that permanent aggregation requires the formation of multiple interparticle connections between the aggregates.

Obviously, this supplementary constraint may complete the already existing conditions imposed to interpret the slow aggregation affecting colloids suspended in aqueous media of small ionic strength (Fuchs) [11]. Similar conditions were expressed in the algorithm of the standard reaction-limited aggregation process [12], [13], [14], [15], [16], W(i,j)=W11(ij)σ, where W(i,j) expresses the sticking probability of colliding clusters of mass i and j and W11 the probability of doublet formation. The exponent σ constitutes the usual variable in the numerical investigation. Below or above the critical σc, the mass frequency function—the number concentration c(n) of colloids composed of n particles as a function of the aggregate mass n—displays a bell-shaped or a continuously decreasing curve, respectively. Only results relative to W11=0.01 and σ=0.4 greater than σc are presented here, insofar as the objective of the investigation is to determine to what extent the connection constraint modifies the standard reaction-limited aggregation process.

In order to fix a realistic constraint on the fulfillment of the permanent sticking of two clusters, the constraint was elaborated according to the irregular fracture model of Horwatt et al. [17]. This model assumed that the force binding the fragment to the parent cluster depended on the number of bonds that must be severed in the cluster breakup. Therefore, this model is briefly presented first and the expression of the constraints relative to the number of intercluster connection and the relative sizes of the cluster and the fragment are given in the second section.

In the third section, we detailed the algorithm of the complete aggregation process and the results relative to the kinetics of aggregation and the mass-frequency function of the aggregates. Finally, we present some experimental results of the aggregation of hydrated polystyrene latex [5], which display modes of cluster growth and mass frequency curves similar to those provided by the present numerical simulation.

Section snippets

Theoretical

Investigation of the aggregate breakup in the presence of hydrodynamic forces requires modeling of both the hydrodynamic forces FH acting on the aggregate and the forces FC responsible for the aggregate cohesion [17]. First, the hydrodynamic forces and the mechanical characteristics of the aggregates (considered as nonpermeable spheres) are briefly developed according to the presentation of Horwatt et al., and second, the model of the aggregate cohesion is developed according to the model

Numerical methods

First of all, clusters of various masses were synthesized employing the algorithms of the standard diffusion- and reaction-limited aggregation processes [20], [21] in order to diversify the fractal dimension. In the second step, clusters were agglomerated two by two and the number v of connections between adjacent particles was determined for each agglomerate conformation. Although off-lattice models are more realistic concerning the short-range structure of 3D systems, the cubic lattice model

The connection frequency and the number v of intercluster connections

Fig. 3 shows schematic 2D configurations for agglomeration of two clusters of masses 34 and 41. Conformation (a) is secured by seven intercluster connections and thus resists fragmentation when the threshold value A(v,i,j) of permanent sticking was set to 10 in the simulation run (Eq. (6)). Conversely, sticking with conformations (b) or (c) is transient since the corresponding A(v,i,j) values are equal to 15.5 and 31, respectively.

Fig. 4, Fig. 5 correspond to agglomeration of clusters generated

Conclusion

The present numerical study had the objective to lay the foundations of a new aggregation process taking into account the connection constraints that may exert the major role when hydrodynamic forces additionally restrict the establishment of permanent sticking. The investigation was developed on 2D and 3D systems, but without other indications, all presented results concerned 3D systems that are of major interest in most areas. In the conclusion of Ref. [5], we noticed that the outlined

Acknowledgements

This work was carried out with the support of the Centre National de la Recherche Scientifique (CNRS, France). S. Stoll is warmly acknowledged for helpful discussions. The Université Louis Pasteur (Strasbourg) is acknowledged for giving access to the computers of the CURRI. The reviewers are warmly acknowledged for improving the presentation of the results.

References (39)

  • J. Widmaier et al.

    J. Colloid Interface Sci.

    (1998)
  • S.G. Mason

    J. Colloid Interface Sci.

    (1977)
  • S.W. Horwatt et al.

    Powder Technol.

    (1992)
  • Y. Tatek et al.

    Powder Technol.

    (2001)
  • Y. Tatek et al.

    Powder Technol.

    (2001)
  • D.N. Sutherland

    J. Colloid Interface Sci.

    (1967)
  • E. Pefferkorn et al.

    Colloids Surf. A

    (1998)
  • S. Stoll et al.

    J. Colloid Interface Sci.

    (1992)
  • S. Stoll et al.

    J. Colloid Interface Sci.

    (1992)
  • R.G. Horn et al.

    Chem. Phys. Lett.

    (1989)
  • L.A. Spielman

    J. Colloid Interface Sci.

    (1970)
  • E.P. Honig et al.

    J. Colloid Interface Sci.

    (1971)
  • A.E. Gonzalez et al.

    Physica A

    (2002)
  • A.E. Gonzalez et al.

    J. Colloid Interface Sci.

    (2002)
  • R. Jullien et al.

    Aggregation and Fractal Aggregates

    (1987)
  • T. Vicsek

    Fractal Growth Phenomena

    (1989)
  • J. Gregory

    Crit. Rev. Environ. Control

    (1989)
  • Cited by (6)

    • A novel multiscale model for the simulation of polymer flash nano-precipitation

      2017, Chemical Engineering Science
      Citation Excerpt :

      The main novelty of this work lies on the use of a multiscale approach, capable of describing FNP at three scales: molecular-scale, nanoparticle-scale and macro-scale. It represents therefore an improvement with respect to other similar works (Tatek and Pefferkorn, 2004; Wang and Larson, 2015; Yuan et al., 2010) and it is consistent with other research efforts conducted in this area (Cheng et al., 2010; Li et al., 2006; Rajagopalan, 2001; Yan and Xie, 2013). The three scales communicate among each other, as suggested by other authors (Capretto et al., 2011; Rajagopalan, 2001), and are modeled respectively with MD, Smoluchowski PBE and CFD, without resorting to other mesoscale techniques (Li et al., 2006; Yan and Xie, 2013).

    • Modelling of aggregate restructuring in a weakly turbulent flow

      2012, Colloids and Surfaces A: Physicochemical and Engineering Aspects
    • Chapter 17 The Cohesion of Fractal Agglomerates: An Elementary Numerical Model

      2007, Handbook of Powder Technology
      Citation Excerpt :

      In this numerical simulation of the RLA process with connection constraints expressed by equation (23), the part of the usual RLA process is determined by applying equation (2) with P11=0.01 and σ=0.4. This allows one to determine to what extent the connection constraint modifies the kinetics of the aggregate growth and the aggregate mass frequency resulting from the application of the standard algorithm of the RLA process [126]. In the usual RLA process, the permanent sticking of the cluster is realized if the sticking probability Pij (equation (2)) is greater than the randomly selected number.

    • Destabilization of polystyrene latex particles induced by adsorption of polyvinylamine: Mass, size and structure characteristics of the growing aggregates

      2006, Journal of Colloid and Interface Science
      Citation Excerpt :

      The aggregation process further developed with features of the diffusion-limited aggregation process when the interfacial polymer layer was mainly composed of trains. When irreversible and reversible aggregation processes concomitantly contribute to the aggregate growth, several aggregate populations emerge in the system and grow at different rates [12,13]. First of all, the paper provides a theoretical sight of the aggregation process based on relationships derived from numerical simulation studies.

    • New developments in carbon black dispersion

      2005, KGK Kautschuk Gummi Kunststoffe
    1

    Present address: Département de physique, Université d'Ottawa, 150 Louis Pasteur, Ottawa, Ontario K1N 6N5, Canada.

    View full text