Fouling of microfiltration membranes by bidisperse particle solutions

Highlights

• Nanoparticle tracking enables the visualization of membrane fouling in filtration processes.

• Internal membrane fouling consists of three stages: fouling site formation, growth, and coalescence.

• The presence of small fraction of large particles significantly increases the retention of small particles.

Abstract

Fouling of filtration membranes is a key process that degrades performance and reduces membrane lifetime. Nevertheless, obtaining direct information about fouling mechanisms is still a significant challenge, especially fouling that occurs within the internal pore spaces. Here, we combined highly multiplexed single-particle tracking methods with alternating laser excitation, which allowed the direct visualization of the transport of nanoparticles within microfiltration membranes, where the advective motion and retention of 200 nm and 40 nm diameter particles was simultaneously imaged within filtration membranes with 650 nm nominal pore size, as membranes became increasingly fouled as a function of total particle throughput. We found that internal membrane fouling consisted of three stages, fouling site formation, fouling site growth, and fouling site coalescence. Larger particles had a greater chance to be retained initially, nucleating the fouling sites. These sites tended to capture other particles passing through the membrane, causing the fouling sites to grow in size. Eventually, isolated fouling sites grew large enough to coalescence with neighboring sites, blocking a large region of membrane pores, leading to significant fouling. Importantly, the presence of small numbers of large particles significantly increased the retention of small particles and also accelerated the rate of particle retention. This mechanistic information about internal membrane fouling will assist in the design and optimization of filtration processes to reduce membrane fouling and expand the fundamental understanding of complex mass transport of polydisperse particle distributions.

Introduction

Microfiltration is a pressure-driven separation process for removing micron-sized particles and microorganisms from process liquids. Microfiltration has been widely used in industrial applications, including pharmaceutical sterilization [[1], [2], [3]], water treatment [[4], [5], [6]], and beverage and dairy processing [7,8]. Fouling is a crucial problem during separations using microfiltration membranes, and can be categorized as external fouling or internal fouling. External fouling is caused by cake formation, which occurs on a membranes’ upstream surface due to the accumulation of large particles that cannot enter the membrane pores, while internal fouling is caused by pore plugging and/or pore narrowing, which occurs within the porous structure of the membrane due to physical deposition and/or adsorption of small particles that are able to enter the membrane pores [9].

Microfiltration membrane fouling is conventionally studied by measuring permeate concentration and flux rate as a function of time [[10], [11], [12], [13], [14]]. Interpretations drawn from such experiments are based on indirect model-dependent analysis of ensemble-averaged information about particle transport and membrane fouling. While such interpretations can have practical utility, they provide limited insights into the complex transport behavior of particles within the highly heterogeneous and interconnected membrane environment. Supplementing these measurements, scanning electron microscope (SEM) images taken before and after fouling can serve to verify proposed mechanisms [15,16]. For example, Tracey and Davis [10] proposed protein fouling mechanisms in microfiltration membranes by studying the resistance to flow versus time, and these experiments were complemented by cross-sectional SEM images of the fouled membranes [15], which showed that fouling mechanisms changed from internal fouling dominated to external fouling with increasing foulant size. This type of endpoint characterization provides important qualitative insights into fouling mechanisms but does not allow visualization of the entire fouling process, which is required for a thorough understanding of fouling formation and fouling reduction. To this end, various types of advanced visualization methods have been developed to monitor membrane fouling in real-time. For example, Trinh et al. [17] used three-dimensional optical coherence tomography to characterize membrane fouling by oil emulsions. Also, Ben Hassan et al. [18] observed the formation of bidisperse cakes using confocal laser scanning microscopy, which provided unique insights into the cake formation process on membrane surfaces in the presence of a bisdisperse suspension.

Microfluidics has provided a powerful approach to probe particle motion inside porous media in situ, and significantly advanced the fundamental understanding of particle transport mechanisms in porous environments [[19], [20], [21], [22], [23], [24], [25], [26], [27], [28]]. Pioneering work used highly-ordered interconnected microchannels as model porous media to understand physico-chemical and hydrodynamic effects on particle deposition [20,21]. Wyss et al. [19] proposed general clogging mechanisms by observing clogging in a well-designed microfluidic device comprising an array of parallel narrow channels. Dersoir et al. [25] further investigated the influence of geometric pore structure on the clog morphology by using microchannels with various geometric features. Although this work provided important insights into the fundamentals of particle transport in porous media, direct visualization of particle transport, and related internal fouling in commercial heterogeneous filtration membranes is still limited, especially when foulant particle size is smaller than 200 nm, which is the approximate resolution limit of a typical optical microscope.

Foulants in real-world scenarios have a large variety of particle sizes, and particle size can affect the performance of the microfiltration significantly [[29], [30], [31], [32]]. For example, Chang et al. [29] found that permeate quality was better in the presence of multiple sized particles because a cake layer deposited on the membrane surface with various sized spheres was more likely to have a smaller porosity compared to one with single-sized spheres. In addition, in situ observation of cake formation by a bisdisperse suspension indicated that the presence of small particles can solidify cake structures by migrating and filling pore voids of the cake formed by large particles [18]. Although size effects in the microfiltration process have been extensively studied, the effects of polydisperse particle distributions, especially when the particles are smaller than the pore size, are not well understood.

Here, we combined highly multiplexed single-particle tracking methods with alternating-laser excitation (ALEX) imaging to simultaneously visualize nanoparticle motion deep in the porous microfiltration membrane and elucidate fouling processes associated with the presence of both large and small particles. Single-particle tracking is a powerful tool to investigate dynamic processes in complex and confined environments, including heterogeneous porous polymer membranes and biological materials [[33], [34], [35], [36], [37], [38], [39], [40], [41], [42]]. In these single-particle tracking experiments, particle motion was continuously imaged using fluorescence microscopy, particle positions were localized and linked into trajectories using a custom tracking algorithm. To track two particle populations, we employed alternating-laser excitation, which enables imaging of multiple fluorescent signals simultaneously by rapid switching between excitation wavelengths [[43], [44], [45]]. By using nanoparticles of two different sizes with contrasting excitation spectra, we were able to visualize bidisperse particle motion in the membrane simultaneously. These direct observations of spatio-temporal single-particle trajectories provided explicit information about the formation and development of fouling sites in the membranes, and elucidated the correlation of various sized particles in confined environments. The findings presented here provide important insights into internal fouling mechanisms along with valuable reference data for the rational design of microfiltration processes to reduce membrane fouling.