Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Capillary filtering of particles during dip coating

Capillary filtering of particles during dip coating Capillary ltering of particles during dip coating 1,  2, 3 2 Alban Sauret, Adrien Gans, B en edicte Colnet, Guillaume 2 4, 5 1 Saingier, Martin Z. Bazant, and Emilie Dressaire Department of Mechanical Engineering, University of California, Santa Barbara, CA 93106, USA Surface du Verre et Interfaces, UMR 125, CNRS/Saint-Gobain, 93303 Aubervilliers, France FAST, CNRS, Univ. Paris-Sud, Univ. Paris-Saclay, 91405 Orsay, France Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Department of Mathematics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA (Dated: 10 May 2019) An object withdrawn from a liquid bath is coated with a thin layer of liquid. Along with the liquid, impurities such as particles present in the bath can be transferred to the withdrawn substrate. Entrained particles locally modify the thickness of the lm, hence altering the quality and properties of the coating. In this study, we show that it is possible to entrain the liquid alone and avoid contamination of the substrate, at suciently low withdrawal velocity in diluted suspensions. Using a model system consisting of a plate exiting a liquid bath, we observe that particles can remain trapped in the meniscus which exerts a resistive capillary force to the entrainment. We characterize di erent entrainment regimes as the withdrawal velocity increases: from a pure liquid lm, to a liquid lm containing clusters of particles and eventually individual particles. This capillary ltration is an e ective barrier against the contamination of substrates withdrawn from a polluted bath and nd application against bio-contamination. asauret@ucsb.edu arXiv:2011.13796v1 [cond-mat.soft] 27 Nov 2020 2 I. INTRODUCTION A solid body dipped in a liquid bath emerges covered with a thin layer of liquid. This familiar phenomenon occurs, for instance, when withdrawing a probe from a biological sample, but it is also essential in many industrial settings [1]. Dip coating is commonly used to cover surfaces with a liquid layer of constant thickness, from several to hundreds of micrometers, depending on the conditions [2]. The thickness of the entrained lm depends on the competition between the capillary pressure gradient and the viscous force at the liquid/solid interface through several factors such as the withdrawal speed, the physical properties of the uid and surface [1, 3{5]. Recent advances in materials science combine modi ed surfaces and complex uids to develop new coatings, with interesting physical and optical properties. For a perfectly wetting Newtonian uid, the rst solution describing the thickness of the liquid lm was proposed by Landau and Levich [3] and then Deryagin [4]. Most of the research on liquid lm entrainment has since focused on the great diversity of homogenous uids and substrates and their in uence on the lm properties [6]. Indeed, the solutions to the hydrodynamic problem and the lm features depend on the nature of the uid, through bulk properties and boundary conditions. As non-Newtonian uids are commonly used in industrial applications, studies have also investigated the in uence of the rheological properties on the lm thickness [7, 8]. Surfactants modify the interfacial properties of the uid and the result of the dip-coating process. Experiments show that surfactants increase the thickness of the entrained liquid layer by modifying the ow in the liquid bath [9{13]. Much less emphasis has been placed on complex or heterogeneous uids. In particular, a liquid bath can contain impurities in the form of rigid or soft particles, which may be entrained in the coating. For instance, microorganisms, sediments, and fragments of plastics are dispersed in natural water bodies and can easily contaminate solid surfaces. This transfer of particles is useful for biological processes and environmental measurements, but can also contribute to the dissemination of pathogens and pollutants. It is thus essential to be able to predict the conditions under which a particle dispersed in a liquid phase will be entrained in the liquid lm and contaminate the substrate. Previous work on dip coating of suspensions has evidenced the entrainment of colloidal particles by a liquid lm [14{16]. After withdrawal, the colloidal particles form complex patterns on the solid surface. The geometry of the patterns depends on the particle concentration, withdrawal speed, and evaporation of the liquid. Despite these recent advances, the transfer of dilute particles whose size is comparable to the lm thickness continues to challenge the continuum description that applies to coated lms of colloidal suspensions. Over the past few years, studies have explored the in uence of particles on interfacial dynamics, when the length scales of a liquid thread or lm and the particle diameter are comparable. Because the particles deform the air-liquid interface, they strongly modify the liquid lm geometry and instabilities [17{25]. In the dip-coating con guration, only a few studies have investigated the fate of individual non-Brownian particles in the meniscus and their entrainment in the lm [26, 27]. An experimental study by Kao et al. focused on the self-assembly of a monolayer of entrained particles in the liquid lm [28] and bubbles [29]. Recently, Colosqui et al. considered 2D numerical simulations of a small number of dilute particles and showed the existence of two regimes of entrainment that result in the formation of clusters of particles, either within the meniscus or in the liquid lm [30]. They also suggested that no particles are entrained in the liquid lm below a critical withdrawal velocity, corresponding to a coating lm thinner than the particle diameter. However, no systematic validation of this prediction has been performed. In addition, the validity of the entrainment condition provided by Colosqui et al. remains to be validated experimentally. Numerically, the cost of the 2D (and 3D) numerical simulations is too high to explore the validity of the entrainment condition. In addition, most applications involve 3D particles (spheres, aggregate, etc.) whose shape can in uence the threshold for particle entrainement and the formation of clusters. There is therefore a need of systematic experiments to determine the conditions under which a particle is entrained in a liquid lm during a dip coating process. 3 In this paper, we show experimentally that capillary e ects can prevent suciently large particles from being entrained in the liquid lm during the dip coating process. We also demonstrate that particles up to six times larger than the liquid lm can contaminate the withdrawn surface and we show the relevance of our ndings for microorganisms. In section II, we rst present our experimental approach: a glass plate is withdrawn from a liquid bath containing dispersed particles. In section III, we highlight the existence of three regimes for increasing withdrawal velocity: one where the coating lm is free of particles, one in which the lm contains individual isolated particles and one in which the entrained particles are assembled into clusters. We rationalize these observations in section IV and provide a discussion of the scaling law that describes the entrainment. Finally, in section V, we demonstrate that the capillary ltration mechanism is an original and e ective way to prevent the contamination of a substrate dipped in a polluted liquid, containing particles or microorganisms. II. EXPERIMENTAL METHODS To study the entrainment of non-Brownian particles in a coated lm, we withdraw a glass plate from a bath of viscous dilute suspension at constant velocity U . The plate entrains a liquid lm whose thickness in the absence of particles obeys the classical Landau-Levich law [1, 3, 31]. The experiments are performed with density-matched polystryrene particles dispersed in silicone oil. We use high density silicone oil, AP 100, AR 200 or AP 1000 (from Sigma Aldrich). Because of possible small temperature variations during the experiments, we measure the evolution of the o o viscosity and the density of the silicone oils in the temperature range T 2 [16 C; 26 C] with an Anton Paar MCR 501 rheometer. The dynamic viscosity for the silicone oil AP 100 is captured by the expression  = 0:1315 0:0051 (T 20) (in Pa.s), for the AR 200:  = 0:2425 0:0077 (T 20) 100 200 and for the AP 1000:  = 1:4227 0:0669 (T 20), where T is the temperature in degree Celsius. The density can be calculated with the following expressions, for the silicone oil AP 100: = 1:0626 0:0008 (T 20) (in g:cm ), for the AR 200:  = 1:046 0:0008 (T 20) and for 100 200 the AP 1000:  = 1:0871 0:0:0008 (T 20), where T is the temperature in degree Celsius. The o 1 silicone oils have an interfacial tension at 20 C of = 21  1 mN:m . The polystyrene particles of radius a = [20; 40; 70; 125; 250] m and average density of  = 1055 kg m (Dynoseeds TS - Microbeads) are dispersed in the liquid. Between two experiments the suspension was thoroughly mixed and over the timescale of an experiment, typically between a few seconds to a few tens of minutes, the suspension can be assumed to be neutrally buoyant. Indeed, the settling velocity of a particle in a quiescent uid is due to a density mismatch between the uid and the particle, and is given by [32] 2   g a v = ; (1) where  =   is the density di erence between the particle and the uid,  is the dynamic p f viscosity, g is the gravitational acceleration, and a is the radius of the particle. We compare the gravity-induced displacement of the particles with the distance traveled by the plate during an experiment. We found that this gravity-induced displacement is negligible for all suspensions and entrainment velocities. Therefore, the suspension can be considered as neutrally buoyant over the timescale of the experiment. The volume fraction of particles in suspension  is de ned as the ratio of the volume of particles to the total volume of uid. We work with dilute suspensions, i.e.,  < 5 %. We can thus neglect the collective e ects associated with dense mixtures, including their non-Newtonian rheology. In particular, we assume that the presence of particles does not signi cantly modify the viscosity of the dilute suspension. This approximation holds for many situations, including samples with biological contaminants, typically at very low concentrations. Once homogeneous, the suspension is transferred to a container (110 mm long  40 mm wide 150 mm high) whose vertical motion is actuated by a stepper motor (Zaber), with a speed up 4 (a) (b) (c) (d) FIG. 1. (a) Schematic of the dip coating experimental setup. Experimental visualizations of the three entrainment regimes during the withdrawal of the substrate from a suspension (2% of PS particles of radius a = 125 m in AP 100). (b) liquid only, (c) entrainment of clusters and (d) entrainment of individual particles. The time increases from left to right. Scale bars are 500 m to 2:4 cm s with a 2 % precision. The solid substrate, a glass plate of thickness l = 2 mm, width w = 100 mm and height h = 120 mm is initially dipped in the suspension and remains stationary during the experiment to prevent vibrations from disturbing the coated lm. Prior to the experiments, the glass plates were thoroughly cleaned with Isopropanol (IPA) and dried with compressed air. A new container was used for each set of experiments. As the container moves downward, the plate emerges from the bath at a constant velocity U and is coated with a liquid lm over the 6.5-cm vertical course of the motor. A schematic of the experimental setup is shown in Fig. 1(a). The dip coating dynamics, including the lm thickness and composition are recorded with a digital camera (Nikon D7100) and a 200-mm macro lens focused above the meniscus. To detect the presence of the smallest particles (a = 20 m), we also use a long working distance microscope lens (10X EO M Plan Apo Long Working Distance In nity Corrected from Edmund Optics). The thickness of the liquid lm without particles is determined using a gravimetry method. The volume of suspension removed from the container during a withdrawal can be obtained from the di erence of mass measured by a scale m and the uid density , with V = m=. The main issue with this method is the presence of a lower edge e ect [26]. Therefore, we perform two withdrawal experiments with di erent dipping lengths l > l . Subtracting the masses obtained with the two 2 1 dipping lengths allows de ning and subtracting the volume of uid that remains attached to the bottom of the plate after the withdrawal and we obtain the average uid thickness h. In addition to the gravimetry methods used to measure the thickness of the liquid lm in the pure silicone oil, we perform measurements of the lm thickness using a light absorption method, similarly to Vernay et al [33]. When the light passes through the coated liquid lm, its absorption is proportional to the thickness of the lm following the Beer-Lambert law of absorption. We rst perform a careful calibration through samples of controlled thickness and determine that for the Sudan Red and the silicone oil used (AP 100), the intensity response of the Sudan Red matches the Beer-Lambert law of absorption h = log (I=I ) =(" c). 0 5 (a) (b) FIG. 2. (a) Evolution of the surface density of particles entrained ' (hollow blue rectangles) and individual particles entrained ' (blue lled circles) when increasing the withdrawal velocity U for a suspension s;in of polystyrene particles dispersed in silicone oil AP 100 ( = 5 % and a = 70 m). The vertical dashed line and dash-dotted line show the threshold velocities for clusters entrainment U and individual particle cl entrainement U , respectively. The insets illustrates the composition of the coated lm in the di erent in regimes. Scale bars are 500 m. (b) Threshold velocity U for the entrainment of individual particles when in varying the radius of the particles a and the liquid phase: silicone oil AP 100 (blue squares), AR 200 (red circles), AP 1000 (green diamonds). The experiments are performed only for dilute suspensions,  < 0:5%. III. RESULTS A. Phenomenology We begin by examining the entrainment of particles in the liquid lm for di erent withdrawal velocities. We observe three entrainment regimes when increasing the withdrawal velocity U [Fig. 1(b)]. At low withdrawal velocity, no particles are entrained in the lm. The particles are trapped in the meniscus, which e ectively acts as a capillary lter. At large withdrawal velocity, isolated particles are entrained in the liquid lm and contaminate the substrate. The thickness of the liquid lm increases locally around the particles, which modi es the properties and stability of the liquid coating. At intermediate velocity, clusters or aggregates of particles are entrained in the liquid lm. The assembly of the particles, i.e., the formation of the clusters takes place in the meniscus. Indeed, the meniscus traps individual particles. When the local density of particles is large enough, clusters form, cross the meniscus and get entrained in the lm. The withdrawal velocity at which clusters are rst observed depends on the probability of the particles to assemble in the meniscus and thus on the volume fraction of the suspension and the withdrawal length. Based on these qualitative observation, the capillary ltration, which is responsible for trapping particles in the meniscus, occurs at low withdrawal velocity where it prevents contamination of the thin liquid lm and contributes to the formation of clusters at intermediate withdrawal velocity. This mechanism is not relevant at high withdrawal velocity. B. Entrainement threshold To better understand the physical mechanisms involved in capillary ltration and determine the conditions under which particles are not entrained in the liquid lm, we carry out systematic ex- periments varying the withdrawal speed and the properties of the suspension. Here, we consider a 6 individual no particles particles (a) (b) FIG. 3. (a) Rescaled thickness of the liquid lm h=` without particles for varying capillary number Ca and silicone oil AP 100 (green sybmols), AR 200 (blue symbols), AP 1000 (green symbols). The dotted line 2=3 is the theoretical prediction h=` = 0:94 Ca . (b) Thickness of the liquid lm when varying the capillary number and withdrawing the glass plate from a suspension of polystyrene particles in silicone oil (AP 100) with  = 0:28 % and a = 70 m particles. The vertical lines indicate the three di erent regimes and the insets show typical lm compositions on the plate. dilute suspension ( < 5%). The experiment consists in increasing the withdrawal velocity while keeping other parameters xed and recording the number of particles entrained per unit surface area of lm [Fig. 2(a)]. We record the total number of particles per unit area ' and the total number of individual isolated particles ' , i.e., particles that are not in a cluster. At low withdrawal s;in velocity, we observe that no particles are entrained in the liquid lm whose thickness remains well predicted by the Landau-Levich law. Increasing the withdrawal velocity leads to a rst transition at U above which only clusters are entrained in the liquid lm, and no individual particles are cl observed. Finally, increasing the withdrawal velocity further, beyond U , leads to the entrainment in of individual particles. We rst focus on the situation when individual particles are entrained in the liquid lm, and we shall discuss later the entrainment of clusters. To focus on the individual particle regime, we consider very diluted suspensions ( < 0:5%). Indeed, at low particle concentration, the probability to assemble clusters in the meniscus is low. The eciency of the ltration process drops only slightly when clusters start being entrained, but then increase suddenly as isolated particles are entrained. Using particles of di erent radii, we measure the threshold velocity U at which individual par- in ticles are entrained in the liquid lm, and we show that the larger the particle is, the greater the threshold velocity is [Fig. 2(b)]. The capillary ltration is more ecient for large particles as they are trapped in the meniscus for a larger range of withdrawal velocity. The properties of the uid also determine the eciency of the capillary ltration. Using di erent oils, we vary the viscosity of the uid. Increasing the viscosity of the uid decreases the threshold velocity U for all particle in sizes [Fig. 2(b)]. The less viscous the liquid is, the more ecient the capillary ltration. For dilute suspensions, the entrainment of particles depends on the thickness of the coated lm. The thickness of liquid deposited on a plate withdrawn from a bath of Newtonian uid scales with a power-law of the capillary number Ca =  U= in the limit of small capillary numbers, typically Ca < 10 [31]. The thickness of the liquid layer h is uniform along the plate and follows the 2=3 Landau-Levich-Deryaguin law (LLD law): h = 0:94 ` Ca , where ` = =( g) is the capillary c c length of the uid. We performed dip coating experiments with silicone oils, without particles, and measured the lm thickness by classical gravimetry methods [12, 26, 31]. Our measurements are reported in Fig. 3(a) and show an excellent agreement with the classical LLD law in the range clusters 7 FIG. 4. (a) Schematic of the forces acting on a particle in the liquid meniscus. (b) Threshold capillary number Ca for individual particle entrainement in the liquid lm. The symbols are the experimental in results for di erent silicone oils (blue squares: AP 100, red circles: AR 200, green diamonds: AP 1000) and 3=4 the dotted line is the scaling Ca = 0:24 Bo . in 4 2 Ca 2 [10 ; 2  10 ] for all silicone oils. For dilute suspensions, we will, therefore, use the LLD model to describe the thin lm and the forces acting on the particles. Experimentally, we observe that the presence of a small amount of particles does not modify signi cantly the thickness of the coating lm suciently far from the particles [Fig. 3(b)]. Indeed, experiments performed with a diluted suspension ( = 0:28 % of 2 a = 140 m particles in silicone oil AP 100) show that we recover the classic Landau-Levich law in all regimes, far from the particles, as reported in Fig. 3(b). IV. DISCUSSION A. Threshold of particles entrainement The condition of ltration depends on the forces acting on a particle in the meniscus, the capillary and viscous forces, as represented in Fig. 4(a). The viscous force is responsible for the entrainment of the particles in the liquid lm whereas the capillary force which opposes the deformation of the air/liquid interface prevents the particles from entering the liquid lm, hence leading to capillary ltration. Colosqui et al. suggested that a 2D particle is entrained when the thickness of the liquid lm at the stagnation point, h , is larger than the particle size. During the dip coating of a plate, the thickness at the stagnation point is given by h =` = 3 (h=` ) (h=` ) =Ca [3, 34]. For a particle to be entrained in the liquid lm, the viscous drag c c force on the particle needs to be larger than the resistive capillary force. An upper bound for the entrainment threshold is given by the condition h > 2 a. Thus, the scaling for particle entrainment, 2=3 with h ' 3 h = 3 e = 2:82 ` Ca , becomes 3=4 Ca = 0:59 Bo ; (2) where Bo = (a=` ) is the Bond number. This condition is derived based on geometrical criteria and is the same as the one suggested by Colosqui et al. [30]. The geometrical condition is equivalent to assuming that the particle is entrained in the liquid lm when the capillary force vanishes. For capillary numbers greater than Ca , the capillary force is not sucient to lter out individual particles, i.e., to prevent them from being entrained in the liquid lm. 8 FIG. 5. Evolution of the surface density of (a) the total particles entrained ' and (b) only the individual particles entrained ' when varying the capillary number Ca and for di erent volume fractions  = 1% s;in (green diamonds),  = 2% (blue circles) and  = 5% (red squares). In (a) the colored dashed lines show the transition between the entrainment of liquid alone and the entrainment of particles in clusters. In (b) the vertical line shows the transition between the cluster regime and the individual particle regimes. (c) Mean size of the clusters entrained in the liquid lm. The insets show di erent morphologies of clusters. (d) Diagram (; Ca) reporting the existence of the three regimes: liquid alone, clusters and individual particles, obtained for  == 33 mm and various volume fraction . The black lines are guides for the eye. Our experiments reveal that the scaling law captures the trend, as a 3=4 power law of the Bond number, but the prefactor is not correct. In Fig. 4(b), we report the threshold value of the capillary number as a function of the Bond number for the experiments originally presented in Fig. 2(b). 3=4 All the data points collapse on the line corresponding to the scaling law Ca / Bo , for Bo in varying over three orders of magnitude. This power law is consistent with the model. The prefactor measured experimentally is equal to 0.24, and di ers from the value obtained in the model. This di erence likely comes from the condition that the capillary force has to vanish is too restrictive. This indicates that the force balance between the capillary e ects and viscous force needs to be re ned to predict accurately the condition of entrainment. As a quantitative model for the capillary force in the regime of large deformation of the meniscus is still missing, the experiments presented in this paper are the only method to quantitatively determine the entrainment threshold. B. Entrainement of clusters In the previous section, we reported on the entrainment of individual particles in very dilute suspensions ( < 0:5%). Depending on the particle volume fraction, and the length over which the plate is withdrawn from the suspension, , we also observe an intermediate regime, in which clusters of particles are entrained. The clusters assemble in the meniscus and cross into the lm for 3=4 Ca < 0:24 Bo , before the entrainment of individual particles. The formation of clusters in the meniscus e ectively decreases the eciency of the ltration process. To determine the conditions of cluster formation and entrainment in the liquid lm, we conduct withdrawal experiments with suspensions of 140 m particles in the AP 100 silicone oil. For di erent volume fractions of particles in suspension 0:3% <  < 5%, we measure the surface fraction of particles entrained in the liquid lm ' over a constant withdrawal distance of the plate  = 33 mm. We also measure the surface fraction of individual particles ' [Fig. 5(a)-(b)]. We de ne two entrainment thresholds, one for s;in the clusters and one for individual particles using ' and ' = ' + ' , where ' is the s;in s s;cl s;in s;cl surface fraction of particles in clusters. The results indicate that the threshold capillary number at which individual particles are entrained on the plate does not depend on the volume fraction  of the suspension. For the suspensions used here, Ca ' 2:2  10 for all particle densities [Fig. 5(b)]. This result is consistent with the isol model presented in the previous paragraph: Ca only depends on Bo, which is independent of . In contrast, the threshold capillary number above which clusters are entrained in the liquid lm 9 decreases as the volume fraction of the suspension  increases [Fig. 5(a)]. In this regime, the individual particles that are not able to enter the liquid lm are joined by other particles being advected along streamlines toward the meniscus, thus forming clusters driven by capillary forces in the vicinity of the stagnation point [35{38]. The formation of clusters both distorts the interface locally and increases the viscous drag on the cluster. Above a certain velocity, when the probability of forming a cluster of a given size is sucient, clusters are able to enter the lm. The eciency of the capillary ltration is therefore dependent upon the particle concentration. Indeed, for a given volume of suspension entrained, the number of particles trapped in the meniscus increases with the volume fraction of the suspension  and the withdrawal length , making this process time dependent. A greater number of particles can then assemble into clusters in the vicinity of the stagnation point during the withdrawal of the plate from the bath of suspension. It is important to note that the number of particles that can assemble in the meniscus not only depends on the volume fraction  but also on the withdrawal length of the plate . We report experiments performed with a constant withdrawal length  = 33 mm, but increasing the value of  leads to more particles collected in the vicinity of the stagnation point and decreases the threshold capillary number for clusters entrainment. The trapping of particles in the meniscus not only increases the probability for particles to be in a cluster but it can also determine the size and shape of those clusters. Increasing the length also allows for more time for entrained particles to aggregate within the lm, due to capillary forces. We measure the average size of the clusters entrained on the substrate as a function of the capillary number Ca for di erent values of the volume fraction of the suspension . The results are reported in Fig. 5(c) and show that, on average, the larger the clusters are, the lower the capillary number at which they are entrained is. Large clusters are entrained far below the threshold value obtained for individual particles Ca . Indeed, clusters composed of 4 to 5 particles are entrained at capillary in numbers that are equal to about half the threshold value to entrain isolated particles. The threshold capillary numbers obtained for clusters of di erent sizes are independent of the concentration of the suspension. Only the probability to assemble a cluster of a ven size in the meniscus varies with . The quantitative prediction of the entrainment of a cluster, depending on its size and shape, requires an analytical expression for the capillary force. Qualitatively, the clusters are composed of a single layer of particles and the presence of a larger number of particles in a cluster increases its surface area and the viscous lubrication force whereas the capillary force depends on the width of the cluster. This allows the cluster to overcome the capillary ltering mechanism induced by the meniscus at a smaller lm thickness than individual particles. We also note that the amount of particles entrained 3=4 on the plate for Ca < Ca is much smaller than when the threshold Ca ' 0:24 Bo is reached, as in the majority of particles remain isolated in the meniscus [Fig. 5(a)]. Finally, we summarize our ndings in a phase diagram of the three regimes as a function of the capillary number Ca and the volume fraction  in Fig. 5(d). The threshold for individual particle entrainment, which is the main entrainment mechanism for very dilute suspensions, is independent of the volume fraction . However, for more concentrated suspensions, capillary ltration is limited by the formation of clusters which depends on  and on the withdrawal length . Therefore, a substrate withdrawn from a liquid bath containing particles is more likely to be contaminated at lower withdrawal velocities if the volume fraction of the suspension, or the amount of contaminants, is substantial. C. Application to biological contamination In rst approximation, the capillary ltration mechanism described here can be extended to bio- logical organisms. The threshold capillary number or velocity is useful information to prevent the contamination of an object removed from a liquid bath containing dilute microorganisms. This ltration mechanism can be valuable in biological applications in which a probe is dipped in water containing bacteria or microalgae. Biological species are usually very diluted in water, so we only consider the threshold obtained 10 Phytoplankton Parasite Bacteria Choanoflagellate Larvae (1b) (2b) Contamination No contamination (2a) (1a) -2 0 1 2 3 10 10 10 10 FIG. 6. Velocity threshold leading to the contamination of a substrate withdrawn from contaminated water (white region). In the grey region, the substrate is expected to be withdrawn from the polluted water without contamination thanks to the capillary ltering mechanism. The black corresponds to Eq. 3. The arrow indicates the range of size of di erent biological organisms [39{44]. The lled symbols and the hollow symbols indicates contaminated and non-contaminated substrates, respectively. The blue circles are experiments with Synechocystis sp. PCC6803 and the red symbols are experiments with Chlamydomonas reinhardtii. The insets photographs are obtained from withdrawal experiments from suspensions of Chlamydomonas reinhardtii (2 a  10 m) [(1a)-(1b)] and Synechocystis sp. PCC6803 (2 a  3 m) [(2a)-(2b)]. At low withdrawal velocity, the micro-organisms are not entrained: for (1a), U = 0:36 mm:s and for (2a) U = 0:48 mm:s . At larger withdrawal velocity, the micro-organisms contaminate the substrate: for (1b), U = 1 1 2:18 mm:s and for (2b) U = 1:03 mm:s . The scale bars are 200m. for the entrainment of individual particles. Assuming the shape of biological organisms is spherical, 3=4 we can use the relation obtained previously, Ca ' 0:24 Bo in terms of dimensional quantities: in 3=4 3=4 1=4 3=2 U = 0:24 a : (3) Water has a surface tension = 72 mN=m, a density  = 1000 kg=m and a dynamic viscosity = 10 Pa:s. In Fig. 6 we plot the threshold velocities at which biological contaminants are expected to be entrained in the liquid lm on the at plate based on physical values found in the literature [39{ 46]. For microorganisms such as bacteria and parasites, the substrate needs to be pulled out of 2 1 1 the polluted water at velocities smaller 10 10 (mm:s ) to prevent contamination. For larger microorganisms such as phytoplankton and larvae, the threshold speed increases, of the order of a few tens of centimeters per second. These threshold values, under which capillary ltration prevents contamination, can provide guidelines to operate probes and mixers and for rinsing processes in biological and industrial applications. To obtain a better estimate of the threshold velocity at which biological particles will contaminate a substrate, the shape, the deformability and the motility of the particles would need to be taken into account. However, the values derived for hard particles already give a good estimate of the threshold velocities for microorganisms. Indeed, we performed experiments with dilute suspensions of two di erent micro-organisms in water at usual volume fraction  < 0:1%: Chlamydomonas reinhardtii and Synechocystis sp. PCC6803. An illustration of the results is reported in the inset of Fig. 6 and shows that below the expected threshold, the substrate is withdrawn from the polluted 11 water with no microorganism. The thresholds obtained with these two micro-organisms are in fair agreement with the theoretical prediction. The small discrepancies may be due to the shape, size distribution, and mechanical properties of the microorganisms. At velocity smaller than the expected threshold, the substrate is not contaminated by microorganisms. V. CONCLUSION During the withdrawal of a at substrate from a liquid bath containing particles or biological micro-organisms, the meniscus can e ectively act as a capillary lter that could be sometimes more e ective than a passive lter as used in micro uidic devices that clog over time, reducing their eciency [47{50]. Below a given lm thickness on the plate, the meniscus generates a strong capillary force at the stagnation point and prevents the particles from being entrained in the liquid lm and contaminating the substrate. Assuming that the thickness at the stagnation point is smaller than a fraction of the particles size, i.e. that the capillary number is small enough, the particles are trapped 3=4 in the liquid lm for Capillary number smaller than a threshold value Ca = 0:24 Bo . For larger volume fraction, typically larger than a few percents, some trapped particles assemble into clusters that can be entrained in the liquid lm. The capillary ltering mechanism described here should ,apply to a wide range of quasi-spherical particles, including biological micro-organisms. Besides, the threshold dependence on the particle size could allow particle sorting by size using the same ltration mechanism. ACKNOWLEDGEMENTS We thank H. A. Stone, C. Colosqui, and G. M. Homsy for important discussions and P. Brunet for providing the Chlamydomonas reinhardtii and the Synechocystis sp. PCC6803. This work was supported by the French ANR (project ProLiFic ANR-16-CE30-0009) and partial support from a CNRS PICS grant n 07242. [1] E. Rio and F. Boulogne, \Withdrawing a solid from a bath: How much liquid is coated?" Advances in Colloid and Interface Science 247, 100{114 (2017). [2] K. J. Ruschak, \Coating ows," Annual Review of Fluid Mechanics 17, 65{89 (1985). [3] L. Landau and B. Levich, \Dragging of a liquid by a moving plate," Acta Physicochim. (USSR) 17, 42{54 (1942). [4] B. Deryagin and A. Titievskaya, \Experimental study of liquid lm thickness left on a solid wall after receeding meniscus," Dokl. Akad. Nauk USSR 50, 307{310 (1945). [5] D. Qu er e, \Fluid coating on a ber," Annual Review of Fluid Mechanics 31, 347{384 (1999). [6] H. C. Mayer and R. Krechetnikov, \Landau-levich ow visualization: Revealing the ow topology responsible for the lm thickening phenomena," Physics of Fluids 24, 052103 (2012). [7] M. Maillard, J. Boujlel, and P. Coussot, \Solid-solid transition in landau-levich ow with soft-jammed systems," Physical Review Letters 112, 068304 (2014). [8] M. Maillard, J. Boujlel, and P. Coussot, \Flow characteristics around a plate withdrawn from a bath of yield stress uid," Journal of Non-Newtonian Fluid Mechanics 220, 33{43 (2015). [9] C.-W. Park, \E ects of insoluble surfactants on dip coating," Journal of Colloid and Interface Science 146, 382{394 (1991). [10] A. Q. Shen, B. Gleason, G. H. McKinley, and H. A. Stone, \Fiber coating with surfactant solutions," Physics of Fluids 14, 4055{4068 (2002). [11] R. Krechetnikov and G. M. Homsy, \Experimental study of substrate roughness and surfactant e ects on the landau-levich law," Physics of Fluids 17, 102108 (2005). [12] R. Krechetnikov and G. M. Homsy, \Surfactant e ects in the landau{levich problem," Journal of Fluid Mechanics 559, 429{450 (2006). 12 [13] J. Delacotte, L. Montel, F. Restagno, B. Scheid, B. Dollet, H. A. Stone, D. Langevin, and E. Rio, \Plate coating: in uence of concentrated surfactants on the lm thickness," Langmuir 28, 3821{3830 (2012). [14] M. Ghosh, F. Fan, and K. J. Stebe, \Spontaneous pattern formation by dip coating of colloidal sus- pensions on homogeneous surfaces," Langmuir 23, 2180{2183 (2007). [15] S. Watanabe, K. Inukai, S. Mizuta, and M. T. Miyahara, \Mechanism for stripe pattern formation on hydrophilic surfaces by using convective self-assembly," Langmuir 25, 7287{7295 (2009). [16] D. D. Brewer, T. Shibuta, L. Francis, S. Kumar, and M. Tsapatsis, \Coating process regimes in particulate lm production by forced-convection-assisted drag-out," Langmuir 27, 11660{11670 (2011). [17] R. J. Furbank and J. F. Morris, \An experimental study of particle e ects on drop formation," Physics of Fluids 16, 1777{1790 (2004). [18] M. Buchanan, D. Molenaar, S. de Villiers, and R. M. L. Evans, \Pattern formation in draining thin lm suspensions," Langmuir 23, 3732{3736 (2007). [19] C. Bonnoit, T. Bertrand, E. Cl ement, and A. Lindner, \Accelerated drop detachment in granular suspensions," Physics of Fluids 24, 043304 (2012). [20] M. Z Miskin and H. M Jaeger, \Droplet formation and scaling in dense suspensions," Proceedings of the National Academy of Sciences (2012). [21] L. A. Lubbers, Q. Xu, S. Wilken, W. W. Zhang, and H. M. Jaeger, \Dense suspension splat: Monolayer spreading and hole formation after impact," Physical Review Letters 113, 044502 (2014). [22] W. Mathues, C. McIlroy, O. G. Harlen, and C. Clasen, \Capillary breakup of suspensions near pinch- o ," Physics of Fluids 27, 093301 (2015). [23] J. Kim, F. Xu, and S. Lee, \Formation and destabilization of the particle band on the uid- uid interface," Physical Review Letters 118, 074501 (2017). [24] Y. E. Yu, S. Khodaparast, and H. A. Stone, \Armoring con ned bubbles in the ow of colloidal suspensions," Soft Matter 13, 2857{2865 (2017). [25] Y. E. Yu, S. Khodaparast, and H. A. Stone, \Separation of particles by size from a suspension using the motion of a con ned bubble," Applied Physics Letters 112, 181604 (2018). [26] M. Ouriemi and G. M. Homsy, \Experimental study of the e ect of surface-absorbed hydrophobic particles on the landau-levich law," Physics of Fluids 25, 082111 (2013). [27] A. Gans, E. Dressaire, B. Colnet, G. Saingier, M. Z. Bazant, and A. Sauret, \Dip-coating of suspen- sions," Soft matter 15, 252{261 (2019). [28] J. C. T. Kao and A. E. Hosoi, \Spinodal decomposition in particle-laden landau-levich ow," Physics of Fluids 24, 041701 (2012). [29] J. C. T. Kao, A. L. Blakemore, and A. E. Hosoi, \Pulling bubbles from a bath," Physics of Fluids 22, 061705 (2010). [30] C. E. Colosqui, J. F. Morris, and H. A. Stone, \Hydrodynamically driven colloidal assembly in dip coating," Physical Review Letters 110, 188302 (2013). [31] M. Maleki, M. Reyssat, F. Restagno, D. Qu er e, and C. Clanet, \Landau{levich menisci," Journal of Colloid and Interface Science 354, 359{363 (2011). [32] H. Lamb, Hydrodynamics (Cambridge university press, 1993). [33] C. Vernay, L. Ramos, and C. Ligoure, \Free radially expanding liquid sheet in air: time-and space- resolved measurement of the thickness eld," Journal of Fluid Mechanics 764, 428{444 (2015). [34] R Krechetnikov, \On application of lubrication approximations to nonunidirectional coating ows with clean and surfactant interfaces," Physics of Fluids 22, 092102 (2010). [35] P. A Kralchevsky and K. Nagayama, \Capillary interactions between particles bound to interfaces, liquid lms and biomembranes," Advances in Colloid and Interface Science 85, 145{192 (2000). [36] D. Vella and L Mahadevan, \The cheerios e ect," American Journal of Physics 73, 817{825 (2005). [37] K. J. Stebe, E. Lewandowski, and M. Ghosh, \Oriented assembly of metamaterials," Science 325, 159{160 (2009). [38] M. Cavallaro, L. Botto, E. P Lewandowski, M. Wang, and K. J Stebe, \Curvature-driven capillary migration and assembly of rod-like particles," Proceedings of the National Academy of Sciences 108, 20923{20928 (2011). [39] J. Dervaux, M. C. Resta, and P. Brunet, \Light-controlled ows in active uids," Nature Physics 13, 306 (2017). [40] J. S. Guasto, R. Rusconi, and R. Stocker, \Fluid mechanics of planktonic microorganisms," Annual Review of Fluid Mechanics 44, 373{400 (2012). [41] G. L. Mino, ~ M. A. R. Koehl, N. King, and R. Stocker, \Finding patches in a heterogeneous aquatic environment: ph-taxis by the dispersal stage of choano agellates," Limnology and Oceanography Letters 13 2, 37{46 (2017). [42] R. Rusconi, J. S. Guasto, and R. Stocker, \Bacterial transport suppressed by uid shear," Nature Physics 10, 212 (2014). [43] R. Stocker, \Marine microbes see a sea of gradients," Science 338, 628{633 (2012). [44] G. Vesey, P. Hutton, A. Champion, N. Ashbolt, K. L. Williams, A. Warton, and D. Veal, \Application of ow cytometric methods for the routine detection of cryptosporidium and giardia in water," Cytometry: The Journal of the International Society for Analytical Cytology 16, 1{6 (1994). [45] S. Jana, A. Eddins, C. Spoon, and S. Jung, \Somersault of paramecium in extremely con ned environ- ments," Scienti c reports 5, 13148 (2015). [46] W. M. Durham and R. Stocker, \Thin phytoplankton layers: characteristics, mechanisms, and conse- quences," Annual Review of Marine Science 4, 177{207 (2012). [47] A. Lenshof and T. Laurell, \Continuous separation of cells and particles in micro uidic systems," Chem- ical Society Reviews 39, 1203 (2010). [48] A. Sauret, E. C. Barney, A. Perro, E. Villermaux, H. A. Stone, and E. Dressaire, \Clogging by sieving in microchannels: Application to the detection of contaminants in colloidal suspensions," Applied Physics Letters 105, 074101 (2014). [49] E. Dressaire and A. Sauret, \Clogging of micro uidic systems," Soft Matter 13, 37{48 (2017). [50] A. Sauret, K. Somszor, E. Villermaux, and E. Dressaire, \Growth of clogs in parallel microchannels," Physical Review Fluids 3, 104301 (2018). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Physics arXiv (Cornell University)

Capillary filtering of particles during dip coating

Loading next page...
 
/lp/arxiv-cornell-university/capillary-filtering-of-particles-during-dip-coating-xTcYLR3A6G
ISSN
2469-990X
eISSN
ARCH-3341
DOI
10.1103/PhysRevFluids.4.054303
Publisher site
See Article on Publisher Site

Abstract

Capillary ltering of particles during dip coating 1,  2, 3 2 Alban Sauret, Adrien Gans, B en edicte Colnet, Guillaume 2 4, 5 1 Saingier, Martin Z. Bazant, and Emilie Dressaire Department of Mechanical Engineering, University of California, Santa Barbara, CA 93106, USA Surface du Verre et Interfaces, UMR 125, CNRS/Saint-Gobain, 93303 Aubervilliers, France FAST, CNRS, Univ. Paris-Sud, Univ. Paris-Saclay, 91405 Orsay, France Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA Department of Mathematics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA (Dated: 10 May 2019) An object withdrawn from a liquid bath is coated with a thin layer of liquid. Along with the liquid, impurities such as particles present in the bath can be transferred to the withdrawn substrate. Entrained particles locally modify the thickness of the lm, hence altering the quality and properties of the coating. In this study, we show that it is possible to entrain the liquid alone and avoid contamination of the substrate, at suciently low withdrawal velocity in diluted suspensions. Using a model system consisting of a plate exiting a liquid bath, we observe that particles can remain trapped in the meniscus which exerts a resistive capillary force to the entrainment. We characterize di erent entrainment regimes as the withdrawal velocity increases: from a pure liquid lm, to a liquid lm containing clusters of particles and eventually individual particles. This capillary ltration is an e ective barrier against the contamination of substrates withdrawn from a polluted bath and nd application against bio-contamination. asauret@ucsb.edu arXiv:2011.13796v1 [cond-mat.soft] 27 Nov 2020 2 I. INTRODUCTION A solid body dipped in a liquid bath emerges covered with a thin layer of liquid. This familiar phenomenon occurs, for instance, when withdrawing a probe from a biological sample, but it is also essential in many industrial settings [1]. Dip coating is commonly used to cover surfaces with a liquid layer of constant thickness, from several to hundreds of micrometers, depending on the conditions [2]. The thickness of the entrained lm depends on the competition between the capillary pressure gradient and the viscous force at the liquid/solid interface through several factors such as the withdrawal speed, the physical properties of the uid and surface [1, 3{5]. Recent advances in materials science combine modi ed surfaces and complex uids to develop new coatings, with interesting physical and optical properties. For a perfectly wetting Newtonian uid, the rst solution describing the thickness of the liquid lm was proposed by Landau and Levich [3] and then Deryagin [4]. Most of the research on liquid lm entrainment has since focused on the great diversity of homogenous uids and substrates and their in uence on the lm properties [6]. Indeed, the solutions to the hydrodynamic problem and the lm features depend on the nature of the uid, through bulk properties and boundary conditions. As non-Newtonian uids are commonly used in industrial applications, studies have also investigated the in uence of the rheological properties on the lm thickness [7, 8]. Surfactants modify the interfacial properties of the uid and the result of the dip-coating process. Experiments show that surfactants increase the thickness of the entrained liquid layer by modifying the ow in the liquid bath [9{13]. Much less emphasis has been placed on complex or heterogeneous uids. In particular, a liquid bath can contain impurities in the form of rigid or soft particles, which may be entrained in the coating. For instance, microorganisms, sediments, and fragments of plastics are dispersed in natural water bodies and can easily contaminate solid surfaces. This transfer of particles is useful for biological processes and environmental measurements, but can also contribute to the dissemination of pathogens and pollutants. It is thus essential to be able to predict the conditions under which a particle dispersed in a liquid phase will be entrained in the liquid lm and contaminate the substrate. Previous work on dip coating of suspensions has evidenced the entrainment of colloidal particles by a liquid lm [14{16]. After withdrawal, the colloidal particles form complex patterns on the solid surface. The geometry of the patterns depends on the particle concentration, withdrawal speed, and evaporation of the liquid. Despite these recent advances, the transfer of dilute particles whose size is comparable to the lm thickness continues to challenge the continuum description that applies to coated lms of colloidal suspensions. Over the past few years, studies have explored the in uence of particles on interfacial dynamics, when the length scales of a liquid thread or lm and the particle diameter are comparable. Because the particles deform the air-liquid interface, they strongly modify the liquid lm geometry and instabilities [17{25]. In the dip-coating con guration, only a few studies have investigated the fate of individual non-Brownian particles in the meniscus and their entrainment in the lm [26, 27]. An experimental study by Kao et al. focused on the self-assembly of a monolayer of entrained particles in the liquid lm [28] and bubbles [29]. Recently, Colosqui et al. considered 2D numerical simulations of a small number of dilute particles and showed the existence of two regimes of entrainment that result in the formation of clusters of particles, either within the meniscus or in the liquid lm [30]. They also suggested that no particles are entrained in the liquid lm below a critical withdrawal velocity, corresponding to a coating lm thinner than the particle diameter. However, no systematic validation of this prediction has been performed. In addition, the validity of the entrainment condition provided by Colosqui et al. remains to be validated experimentally. Numerically, the cost of the 2D (and 3D) numerical simulations is too high to explore the validity of the entrainment condition. In addition, most applications involve 3D particles (spheres, aggregate, etc.) whose shape can in uence the threshold for particle entrainement and the formation of clusters. There is therefore a need of systematic experiments to determine the conditions under which a particle is entrained in a liquid lm during a dip coating process. 3 In this paper, we show experimentally that capillary e ects can prevent suciently large particles from being entrained in the liquid lm during the dip coating process. We also demonstrate that particles up to six times larger than the liquid lm can contaminate the withdrawn surface and we show the relevance of our ndings for microorganisms. In section II, we rst present our experimental approach: a glass plate is withdrawn from a liquid bath containing dispersed particles. In section III, we highlight the existence of three regimes for increasing withdrawal velocity: one where the coating lm is free of particles, one in which the lm contains individual isolated particles and one in which the entrained particles are assembled into clusters. We rationalize these observations in section IV and provide a discussion of the scaling law that describes the entrainment. Finally, in section V, we demonstrate that the capillary ltration mechanism is an original and e ective way to prevent the contamination of a substrate dipped in a polluted liquid, containing particles or microorganisms. II. EXPERIMENTAL METHODS To study the entrainment of non-Brownian particles in a coated lm, we withdraw a glass plate from a bath of viscous dilute suspension at constant velocity U . The plate entrains a liquid lm whose thickness in the absence of particles obeys the classical Landau-Levich law [1, 3, 31]. The experiments are performed with density-matched polystryrene particles dispersed in silicone oil. We use high density silicone oil, AP 100, AR 200 or AP 1000 (from Sigma Aldrich). Because of possible small temperature variations during the experiments, we measure the evolution of the o o viscosity and the density of the silicone oils in the temperature range T 2 [16 C; 26 C] with an Anton Paar MCR 501 rheometer. The dynamic viscosity for the silicone oil AP 100 is captured by the expression  = 0:1315 0:0051 (T 20) (in Pa.s), for the AR 200:  = 0:2425 0:0077 (T 20) 100 200 and for the AP 1000:  = 1:4227 0:0669 (T 20), where T is the temperature in degree Celsius. The density can be calculated with the following expressions, for the silicone oil AP 100: = 1:0626 0:0008 (T 20) (in g:cm ), for the AR 200:  = 1:046 0:0008 (T 20) and for 100 200 the AP 1000:  = 1:0871 0:0:0008 (T 20), where T is the temperature in degree Celsius. The o 1 silicone oils have an interfacial tension at 20 C of = 21  1 mN:m . The polystyrene particles of radius a = [20; 40; 70; 125; 250] m and average density of  = 1055 kg m (Dynoseeds TS - Microbeads) are dispersed in the liquid. Between two experiments the suspension was thoroughly mixed and over the timescale of an experiment, typically between a few seconds to a few tens of minutes, the suspension can be assumed to be neutrally buoyant. Indeed, the settling velocity of a particle in a quiescent uid is due to a density mismatch between the uid and the particle, and is given by [32] 2   g a v = ; (1) where  =   is the density di erence between the particle and the uid,  is the dynamic p f viscosity, g is the gravitational acceleration, and a is the radius of the particle. We compare the gravity-induced displacement of the particles with the distance traveled by the plate during an experiment. We found that this gravity-induced displacement is negligible for all suspensions and entrainment velocities. Therefore, the suspension can be considered as neutrally buoyant over the timescale of the experiment. The volume fraction of particles in suspension  is de ned as the ratio of the volume of particles to the total volume of uid. We work with dilute suspensions, i.e.,  < 5 %. We can thus neglect the collective e ects associated with dense mixtures, including their non-Newtonian rheology. In particular, we assume that the presence of particles does not signi cantly modify the viscosity of the dilute suspension. This approximation holds for many situations, including samples with biological contaminants, typically at very low concentrations. Once homogeneous, the suspension is transferred to a container (110 mm long  40 mm wide 150 mm high) whose vertical motion is actuated by a stepper motor (Zaber), with a speed up 4 (a) (b) (c) (d) FIG. 1. (a) Schematic of the dip coating experimental setup. Experimental visualizations of the three entrainment regimes during the withdrawal of the substrate from a suspension (2% of PS particles of radius a = 125 m in AP 100). (b) liquid only, (c) entrainment of clusters and (d) entrainment of individual particles. The time increases from left to right. Scale bars are 500 m to 2:4 cm s with a 2 % precision. The solid substrate, a glass plate of thickness l = 2 mm, width w = 100 mm and height h = 120 mm is initially dipped in the suspension and remains stationary during the experiment to prevent vibrations from disturbing the coated lm. Prior to the experiments, the glass plates were thoroughly cleaned with Isopropanol (IPA) and dried with compressed air. A new container was used for each set of experiments. As the container moves downward, the plate emerges from the bath at a constant velocity U and is coated with a liquid lm over the 6.5-cm vertical course of the motor. A schematic of the experimental setup is shown in Fig. 1(a). The dip coating dynamics, including the lm thickness and composition are recorded with a digital camera (Nikon D7100) and a 200-mm macro lens focused above the meniscus. To detect the presence of the smallest particles (a = 20 m), we also use a long working distance microscope lens (10X EO M Plan Apo Long Working Distance In nity Corrected from Edmund Optics). The thickness of the liquid lm without particles is determined using a gravimetry method. The volume of suspension removed from the container during a withdrawal can be obtained from the di erence of mass measured by a scale m and the uid density , with V = m=. The main issue with this method is the presence of a lower edge e ect [26]. Therefore, we perform two withdrawal experiments with di erent dipping lengths l > l . Subtracting the masses obtained with the two 2 1 dipping lengths allows de ning and subtracting the volume of uid that remains attached to the bottom of the plate after the withdrawal and we obtain the average uid thickness h. In addition to the gravimetry methods used to measure the thickness of the liquid lm in the pure silicone oil, we perform measurements of the lm thickness using a light absorption method, similarly to Vernay et al [33]. When the light passes through the coated liquid lm, its absorption is proportional to the thickness of the lm following the Beer-Lambert law of absorption. We rst perform a careful calibration through samples of controlled thickness and determine that for the Sudan Red and the silicone oil used (AP 100), the intensity response of the Sudan Red matches the Beer-Lambert law of absorption h = log (I=I ) =(" c). 0 5 (a) (b) FIG. 2. (a) Evolution of the surface density of particles entrained ' (hollow blue rectangles) and individual particles entrained ' (blue lled circles) when increasing the withdrawal velocity U for a suspension s;in of polystyrene particles dispersed in silicone oil AP 100 ( = 5 % and a = 70 m). The vertical dashed line and dash-dotted line show the threshold velocities for clusters entrainment U and individual particle cl entrainement U , respectively. The insets illustrates the composition of the coated lm in the di erent in regimes. Scale bars are 500 m. (b) Threshold velocity U for the entrainment of individual particles when in varying the radius of the particles a and the liquid phase: silicone oil AP 100 (blue squares), AR 200 (red circles), AP 1000 (green diamonds). The experiments are performed only for dilute suspensions,  < 0:5%. III. RESULTS A. Phenomenology We begin by examining the entrainment of particles in the liquid lm for di erent withdrawal velocities. We observe three entrainment regimes when increasing the withdrawal velocity U [Fig. 1(b)]. At low withdrawal velocity, no particles are entrained in the lm. The particles are trapped in the meniscus, which e ectively acts as a capillary lter. At large withdrawal velocity, isolated particles are entrained in the liquid lm and contaminate the substrate. The thickness of the liquid lm increases locally around the particles, which modi es the properties and stability of the liquid coating. At intermediate velocity, clusters or aggregates of particles are entrained in the liquid lm. The assembly of the particles, i.e., the formation of the clusters takes place in the meniscus. Indeed, the meniscus traps individual particles. When the local density of particles is large enough, clusters form, cross the meniscus and get entrained in the lm. The withdrawal velocity at which clusters are rst observed depends on the probability of the particles to assemble in the meniscus and thus on the volume fraction of the suspension and the withdrawal length. Based on these qualitative observation, the capillary ltration, which is responsible for trapping particles in the meniscus, occurs at low withdrawal velocity where it prevents contamination of the thin liquid lm and contributes to the formation of clusters at intermediate withdrawal velocity. This mechanism is not relevant at high withdrawal velocity. B. Entrainement threshold To better understand the physical mechanisms involved in capillary ltration and determine the conditions under which particles are not entrained in the liquid lm, we carry out systematic ex- periments varying the withdrawal speed and the properties of the suspension. Here, we consider a 6 individual no particles particles (a) (b) FIG. 3. (a) Rescaled thickness of the liquid lm h=` without particles for varying capillary number Ca and silicone oil AP 100 (green sybmols), AR 200 (blue symbols), AP 1000 (green symbols). The dotted line 2=3 is the theoretical prediction h=` = 0:94 Ca . (b) Thickness of the liquid lm when varying the capillary number and withdrawing the glass plate from a suspension of polystyrene particles in silicone oil (AP 100) with  = 0:28 % and a = 70 m particles. The vertical lines indicate the three di erent regimes and the insets show typical lm compositions on the plate. dilute suspension ( < 5%). The experiment consists in increasing the withdrawal velocity while keeping other parameters xed and recording the number of particles entrained per unit surface area of lm [Fig. 2(a)]. We record the total number of particles per unit area ' and the total number of individual isolated particles ' , i.e., particles that are not in a cluster. At low withdrawal s;in velocity, we observe that no particles are entrained in the liquid lm whose thickness remains well predicted by the Landau-Levich law. Increasing the withdrawal velocity leads to a rst transition at U above which only clusters are entrained in the liquid lm, and no individual particles are cl observed. Finally, increasing the withdrawal velocity further, beyond U , leads to the entrainment in of individual particles. We rst focus on the situation when individual particles are entrained in the liquid lm, and we shall discuss later the entrainment of clusters. To focus on the individual particle regime, we consider very diluted suspensions ( < 0:5%). Indeed, at low particle concentration, the probability to assemble clusters in the meniscus is low. The eciency of the ltration process drops only slightly when clusters start being entrained, but then increase suddenly as isolated particles are entrained. Using particles of di erent radii, we measure the threshold velocity U at which individual par- in ticles are entrained in the liquid lm, and we show that the larger the particle is, the greater the threshold velocity is [Fig. 2(b)]. The capillary ltration is more ecient for large particles as they are trapped in the meniscus for a larger range of withdrawal velocity. The properties of the uid also determine the eciency of the capillary ltration. Using di erent oils, we vary the viscosity of the uid. Increasing the viscosity of the uid decreases the threshold velocity U for all particle in sizes [Fig. 2(b)]. The less viscous the liquid is, the more ecient the capillary ltration. For dilute suspensions, the entrainment of particles depends on the thickness of the coated lm. The thickness of liquid deposited on a plate withdrawn from a bath of Newtonian uid scales with a power-law of the capillary number Ca =  U= in the limit of small capillary numbers, typically Ca < 10 [31]. The thickness of the liquid layer h is uniform along the plate and follows the 2=3 Landau-Levich-Deryaguin law (LLD law): h = 0:94 ` Ca , where ` = =( g) is the capillary c c length of the uid. We performed dip coating experiments with silicone oils, without particles, and measured the lm thickness by classical gravimetry methods [12, 26, 31]. Our measurements are reported in Fig. 3(a) and show an excellent agreement with the classical LLD law in the range clusters 7 FIG. 4. (a) Schematic of the forces acting on a particle in the liquid meniscus. (b) Threshold capillary number Ca for individual particle entrainement in the liquid lm. The symbols are the experimental in results for di erent silicone oils (blue squares: AP 100, red circles: AR 200, green diamonds: AP 1000) and 3=4 the dotted line is the scaling Ca = 0:24 Bo . in 4 2 Ca 2 [10 ; 2  10 ] for all silicone oils. For dilute suspensions, we will, therefore, use the LLD model to describe the thin lm and the forces acting on the particles. Experimentally, we observe that the presence of a small amount of particles does not modify signi cantly the thickness of the coating lm suciently far from the particles [Fig. 3(b)]. Indeed, experiments performed with a diluted suspension ( = 0:28 % of 2 a = 140 m particles in silicone oil AP 100) show that we recover the classic Landau-Levich law in all regimes, far from the particles, as reported in Fig. 3(b). IV. DISCUSSION A. Threshold of particles entrainement The condition of ltration depends on the forces acting on a particle in the meniscus, the capillary and viscous forces, as represented in Fig. 4(a). The viscous force is responsible for the entrainment of the particles in the liquid lm whereas the capillary force which opposes the deformation of the air/liquid interface prevents the particles from entering the liquid lm, hence leading to capillary ltration. Colosqui et al. suggested that a 2D particle is entrained when the thickness of the liquid lm at the stagnation point, h , is larger than the particle size. During the dip coating of a plate, the thickness at the stagnation point is given by h =` = 3 (h=` ) (h=` ) =Ca [3, 34]. For a particle to be entrained in the liquid lm, the viscous drag c c force on the particle needs to be larger than the resistive capillary force. An upper bound for the entrainment threshold is given by the condition h > 2 a. Thus, the scaling for particle entrainment, 2=3 with h ' 3 h = 3 e = 2:82 ` Ca , becomes 3=4 Ca = 0:59 Bo ; (2) where Bo = (a=` ) is the Bond number. This condition is derived based on geometrical criteria and is the same as the one suggested by Colosqui et al. [30]. The geometrical condition is equivalent to assuming that the particle is entrained in the liquid lm when the capillary force vanishes. For capillary numbers greater than Ca , the capillary force is not sucient to lter out individual particles, i.e., to prevent them from being entrained in the liquid lm. 8 FIG. 5. Evolution of the surface density of (a) the total particles entrained ' and (b) only the individual particles entrained ' when varying the capillary number Ca and for di erent volume fractions  = 1% s;in (green diamonds),  = 2% (blue circles) and  = 5% (red squares). In (a) the colored dashed lines show the transition between the entrainment of liquid alone and the entrainment of particles in clusters. In (b) the vertical line shows the transition between the cluster regime and the individual particle regimes. (c) Mean size of the clusters entrained in the liquid lm. The insets show di erent morphologies of clusters. (d) Diagram (; Ca) reporting the existence of the three regimes: liquid alone, clusters and individual particles, obtained for  == 33 mm and various volume fraction . The black lines are guides for the eye. Our experiments reveal that the scaling law captures the trend, as a 3=4 power law of the Bond number, but the prefactor is not correct. In Fig. 4(b), we report the threshold value of the capillary number as a function of the Bond number for the experiments originally presented in Fig. 2(b). 3=4 All the data points collapse on the line corresponding to the scaling law Ca / Bo , for Bo in varying over three orders of magnitude. This power law is consistent with the model. The prefactor measured experimentally is equal to 0.24, and di ers from the value obtained in the model. This di erence likely comes from the condition that the capillary force has to vanish is too restrictive. This indicates that the force balance between the capillary e ects and viscous force needs to be re ned to predict accurately the condition of entrainment. As a quantitative model for the capillary force in the regime of large deformation of the meniscus is still missing, the experiments presented in this paper are the only method to quantitatively determine the entrainment threshold. B. Entrainement of clusters In the previous section, we reported on the entrainment of individual particles in very dilute suspensions ( < 0:5%). Depending on the particle volume fraction, and the length over which the plate is withdrawn from the suspension, , we also observe an intermediate regime, in which clusters of particles are entrained. The clusters assemble in the meniscus and cross into the lm for 3=4 Ca < 0:24 Bo , before the entrainment of individual particles. The formation of clusters in the meniscus e ectively decreases the eciency of the ltration process. To determine the conditions of cluster formation and entrainment in the liquid lm, we conduct withdrawal experiments with suspensions of 140 m particles in the AP 100 silicone oil. For di erent volume fractions of particles in suspension 0:3% <  < 5%, we measure the surface fraction of particles entrained in the liquid lm ' over a constant withdrawal distance of the plate  = 33 mm. We also measure the surface fraction of individual particles ' [Fig. 5(a)-(b)]. We de ne two entrainment thresholds, one for s;in the clusters and one for individual particles using ' and ' = ' + ' , where ' is the s;in s s;cl s;in s;cl surface fraction of particles in clusters. The results indicate that the threshold capillary number at which individual particles are entrained on the plate does not depend on the volume fraction  of the suspension. For the suspensions used here, Ca ' 2:2  10 for all particle densities [Fig. 5(b)]. This result is consistent with the isol model presented in the previous paragraph: Ca only depends on Bo, which is independent of . In contrast, the threshold capillary number above which clusters are entrained in the liquid lm 9 decreases as the volume fraction of the suspension  increases [Fig. 5(a)]. In this regime, the individual particles that are not able to enter the liquid lm are joined by other particles being advected along streamlines toward the meniscus, thus forming clusters driven by capillary forces in the vicinity of the stagnation point [35{38]. The formation of clusters both distorts the interface locally and increases the viscous drag on the cluster. Above a certain velocity, when the probability of forming a cluster of a given size is sucient, clusters are able to enter the lm. The eciency of the capillary ltration is therefore dependent upon the particle concentration. Indeed, for a given volume of suspension entrained, the number of particles trapped in the meniscus increases with the volume fraction of the suspension  and the withdrawal length , making this process time dependent. A greater number of particles can then assemble into clusters in the vicinity of the stagnation point during the withdrawal of the plate from the bath of suspension. It is important to note that the number of particles that can assemble in the meniscus not only depends on the volume fraction  but also on the withdrawal length of the plate . We report experiments performed with a constant withdrawal length  = 33 mm, but increasing the value of  leads to more particles collected in the vicinity of the stagnation point and decreases the threshold capillary number for clusters entrainment. The trapping of particles in the meniscus not only increases the probability for particles to be in a cluster but it can also determine the size and shape of those clusters. Increasing the length also allows for more time for entrained particles to aggregate within the lm, due to capillary forces. We measure the average size of the clusters entrained on the substrate as a function of the capillary number Ca for di erent values of the volume fraction of the suspension . The results are reported in Fig. 5(c) and show that, on average, the larger the clusters are, the lower the capillary number at which they are entrained is. Large clusters are entrained far below the threshold value obtained for individual particles Ca . Indeed, clusters composed of 4 to 5 particles are entrained at capillary in numbers that are equal to about half the threshold value to entrain isolated particles. The threshold capillary numbers obtained for clusters of di erent sizes are independent of the concentration of the suspension. Only the probability to assemble a cluster of a ven size in the meniscus varies with . The quantitative prediction of the entrainment of a cluster, depending on its size and shape, requires an analytical expression for the capillary force. Qualitatively, the clusters are composed of a single layer of particles and the presence of a larger number of particles in a cluster increases its surface area and the viscous lubrication force whereas the capillary force depends on the width of the cluster. This allows the cluster to overcome the capillary ltering mechanism induced by the meniscus at a smaller lm thickness than individual particles. We also note that the amount of particles entrained 3=4 on the plate for Ca < Ca is much smaller than when the threshold Ca ' 0:24 Bo is reached, as in the majority of particles remain isolated in the meniscus [Fig. 5(a)]. Finally, we summarize our ndings in a phase diagram of the three regimes as a function of the capillary number Ca and the volume fraction  in Fig. 5(d). The threshold for individual particle entrainment, which is the main entrainment mechanism for very dilute suspensions, is independent of the volume fraction . However, for more concentrated suspensions, capillary ltration is limited by the formation of clusters which depends on  and on the withdrawal length . Therefore, a substrate withdrawn from a liquid bath containing particles is more likely to be contaminated at lower withdrawal velocities if the volume fraction of the suspension, or the amount of contaminants, is substantial. C. Application to biological contamination In rst approximation, the capillary ltration mechanism described here can be extended to bio- logical organisms. The threshold capillary number or velocity is useful information to prevent the contamination of an object removed from a liquid bath containing dilute microorganisms. This ltration mechanism can be valuable in biological applications in which a probe is dipped in water containing bacteria or microalgae. Biological species are usually very diluted in water, so we only consider the threshold obtained 10 Phytoplankton Parasite Bacteria Choanoflagellate Larvae (1b) (2b) Contamination No contamination (2a) (1a) -2 0 1 2 3 10 10 10 10 FIG. 6. Velocity threshold leading to the contamination of a substrate withdrawn from contaminated water (white region). In the grey region, the substrate is expected to be withdrawn from the polluted water without contamination thanks to the capillary ltering mechanism. The black corresponds to Eq. 3. The arrow indicates the range of size of di erent biological organisms [39{44]. The lled symbols and the hollow symbols indicates contaminated and non-contaminated substrates, respectively. The blue circles are experiments with Synechocystis sp. PCC6803 and the red symbols are experiments with Chlamydomonas reinhardtii. The insets photographs are obtained from withdrawal experiments from suspensions of Chlamydomonas reinhardtii (2 a  10 m) [(1a)-(1b)] and Synechocystis sp. PCC6803 (2 a  3 m) [(2a)-(2b)]. At low withdrawal velocity, the micro-organisms are not entrained: for (1a), U = 0:36 mm:s and for (2a) U = 0:48 mm:s . At larger withdrawal velocity, the micro-organisms contaminate the substrate: for (1b), U = 1 1 2:18 mm:s and for (2b) U = 1:03 mm:s . The scale bars are 200m. for the entrainment of individual particles. Assuming the shape of biological organisms is spherical, 3=4 we can use the relation obtained previously, Ca ' 0:24 Bo in terms of dimensional quantities: in 3=4 3=4 1=4 3=2 U = 0:24 a : (3) Water has a surface tension = 72 mN=m, a density  = 1000 kg=m and a dynamic viscosity = 10 Pa:s. In Fig. 6 we plot the threshold velocities at which biological contaminants are expected to be entrained in the liquid lm on the at plate based on physical values found in the literature [39{ 46]. For microorganisms such as bacteria and parasites, the substrate needs to be pulled out of 2 1 1 the polluted water at velocities smaller 10 10 (mm:s ) to prevent contamination. For larger microorganisms such as phytoplankton and larvae, the threshold speed increases, of the order of a few tens of centimeters per second. These threshold values, under which capillary ltration prevents contamination, can provide guidelines to operate probes and mixers and for rinsing processes in biological and industrial applications. To obtain a better estimate of the threshold velocity at which biological particles will contaminate a substrate, the shape, the deformability and the motility of the particles would need to be taken into account. However, the values derived for hard particles already give a good estimate of the threshold velocities for microorganisms. Indeed, we performed experiments with dilute suspensions of two di erent micro-organisms in water at usual volume fraction  < 0:1%: Chlamydomonas reinhardtii and Synechocystis sp. PCC6803. An illustration of the results is reported in the inset of Fig. 6 and shows that below the expected threshold, the substrate is withdrawn from the polluted 11 water with no microorganism. The thresholds obtained with these two micro-organisms are in fair agreement with the theoretical prediction. The small discrepancies may be due to the shape, size distribution, and mechanical properties of the microorganisms. At velocity smaller than the expected threshold, the substrate is not contaminated by microorganisms. V. CONCLUSION During the withdrawal of a at substrate from a liquid bath containing particles or biological micro-organisms, the meniscus can e ectively act as a capillary lter that could be sometimes more e ective than a passive lter as used in micro uidic devices that clog over time, reducing their eciency [47{50]. Below a given lm thickness on the plate, the meniscus generates a strong capillary force at the stagnation point and prevents the particles from being entrained in the liquid lm and contaminating the substrate. Assuming that the thickness at the stagnation point is smaller than a fraction of the particles size, i.e. that the capillary number is small enough, the particles are trapped 3=4 in the liquid lm for Capillary number smaller than a threshold value Ca = 0:24 Bo . For larger volume fraction, typically larger than a few percents, some trapped particles assemble into clusters that can be entrained in the liquid lm. The capillary ltering mechanism described here should ,apply to a wide range of quasi-spherical particles, including biological micro-organisms. Besides, the threshold dependence on the particle size could allow particle sorting by size using the same ltration mechanism. ACKNOWLEDGEMENTS We thank H. A. Stone, C. Colosqui, and G. M. Homsy for important discussions and P. Brunet for providing the Chlamydomonas reinhardtii and the Synechocystis sp. PCC6803. This work was supported by the French ANR (project ProLiFic ANR-16-CE30-0009) and partial support from a CNRS PICS grant n 07242. [1] E. Rio and F. Boulogne, \Withdrawing a solid from a bath: How much liquid is coated?" Advances in Colloid and Interface Science 247, 100{114 (2017). [2] K. J. Ruschak, \Coating ows," Annual Review of Fluid Mechanics 17, 65{89 (1985). [3] L. Landau and B. Levich, \Dragging of a liquid by a moving plate," Acta Physicochim. (USSR) 17, 42{54 (1942). [4] B. Deryagin and A. Titievskaya, \Experimental study of liquid lm thickness left on a solid wall after receeding meniscus," Dokl. Akad. Nauk USSR 50, 307{310 (1945). [5] D. Qu er e, \Fluid coating on a ber," Annual Review of Fluid Mechanics 31, 347{384 (1999). [6] H. C. Mayer and R. Krechetnikov, \Landau-levich ow visualization: Revealing the ow topology responsible for the lm thickening phenomena," Physics of Fluids 24, 052103 (2012). [7] M. Maillard, J. Boujlel, and P. Coussot, \Solid-solid transition in landau-levich ow with soft-jammed systems," Physical Review Letters 112, 068304 (2014). [8] M. Maillard, J. Boujlel, and P. Coussot, \Flow characteristics around a plate withdrawn from a bath of yield stress uid," Journal of Non-Newtonian Fluid Mechanics 220, 33{43 (2015). [9] C.-W. Park, \E ects of insoluble surfactants on dip coating," Journal of Colloid and Interface Science 146, 382{394 (1991). [10] A. Q. Shen, B. Gleason, G. H. McKinley, and H. A. Stone, \Fiber coating with surfactant solutions," Physics of Fluids 14, 4055{4068 (2002). [11] R. Krechetnikov and G. M. Homsy, \Experimental study of substrate roughness and surfactant e ects on the landau-levich law," Physics of Fluids 17, 102108 (2005). [12] R. Krechetnikov and G. M. Homsy, \Surfactant e ects in the landau{levich problem," Journal of Fluid Mechanics 559, 429{450 (2006). 12 [13] J. Delacotte, L. Montel, F. Restagno, B. Scheid, B. Dollet, H. A. Stone, D. Langevin, and E. Rio, \Plate coating: in uence of concentrated surfactants on the lm thickness," Langmuir 28, 3821{3830 (2012). [14] M. Ghosh, F. Fan, and K. J. Stebe, \Spontaneous pattern formation by dip coating of colloidal sus- pensions on homogeneous surfaces," Langmuir 23, 2180{2183 (2007). [15] S. Watanabe, K. Inukai, S. Mizuta, and M. T. Miyahara, \Mechanism for stripe pattern formation on hydrophilic surfaces by using convective self-assembly," Langmuir 25, 7287{7295 (2009). [16] D. D. Brewer, T. Shibuta, L. Francis, S. Kumar, and M. Tsapatsis, \Coating process regimes in particulate lm production by forced-convection-assisted drag-out," Langmuir 27, 11660{11670 (2011). [17] R. J. Furbank and J. F. Morris, \An experimental study of particle e ects on drop formation," Physics of Fluids 16, 1777{1790 (2004). [18] M. Buchanan, D. Molenaar, S. de Villiers, and R. M. L. Evans, \Pattern formation in draining thin lm suspensions," Langmuir 23, 3732{3736 (2007). [19] C. Bonnoit, T. Bertrand, E. Cl ement, and A. Lindner, \Accelerated drop detachment in granular suspensions," Physics of Fluids 24, 043304 (2012). [20] M. Z Miskin and H. M Jaeger, \Droplet formation and scaling in dense suspensions," Proceedings of the National Academy of Sciences (2012). [21] L. A. Lubbers, Q. Xu, S. Wilken, W. W. Zhang, and H. M. Jaeger, \Dense suspension splat: Monolayer spreading and hole formation after impact," Physical Review Letters 113, 044502 (2014). [22] W. Mathues, C. McIlroy, O. G. Harlen, and C. Clasen, \Capillary breakup of suspensions near pinch- o ," Physics of Fluids 27, 093301 (2015). [23] J. Kim, F. Xu, and S. Lee, \Formation and destabilization of the particle band on the uid- uid interface," Physical Review Letters 118, 074501 (2017). [24] Y. E. Yu, S. Khodaparast, and H. A. Stone, \Armoring con ned bubbles in the ow of colloidal suspensions," Soft Matter 13, 2857{2865 (2017). [25] Y. E. Yu, S. Khodaparast, and H. A. Stone, \Separation of particles by size from a suspension using the motion of a con ned bubble," Applied Physics Letters 112, 181604 (2018). [26] M. Ouriemi and G. M. Homsy, \Experimental study of the e ect of surface-absorbed hydrophobic particles on the landau-levich law," Physics of Fluids 25, 082111 (2013). [27] A. Gans, E. Dressaire, B. Colnet, G. Saingier, M. Z. Bazant, and A. Sauret, \Dip-coating of suspen- sions," Soft matter 15, 252{261 (2019). [28] J. C. T. Kao and A. E. Hosoi, \Spinodal decomposition in particle-laden landau-levich ow," Physics of Fluids 24, 041701 (2012). [29] J. C. T. Kao, A. L. Blakemore, and A. E. Hosoi, \Pulling bubbles from a bath," Physics of Fluids 22, 061705 (2010). [30] C. E. Colosqui, J. F. Morris, and H. A. Stone, \Hydrodynamically driven colloidal assembly in dip coating," Physical Review Letters 110, 188302 (2013). [31] M. Maleki, M. Reyssat, F. Restagno, D. Qu er e, and C. Clanet, \Landau{levich menisci," Journal of Colloid and Interface Science 354, 359{363 (2011). [32] H. Lamb, Hydrodynamics (Cambridge university press, 1993). [33] C. Vernay, L. Ramos, and C. Ligoure, \Free radially expanding liquid sheet in air: time-and space- resolved measurement of the thickness eld," Journal of Fluid Mechanics 764, 428{444 (2015). [34] R Krechetnikov, \On application of lubrication approximations to nonunidirectional coating ows with clean and surfactant interfaces," Physics of Fluids 22, 092102 (2010). [35] P. A Kralchevsky and K. Nagayama, \Capillary interactions between particles bound to interfaces, liquid lms and biomembranes," Advances in Colloid and Interface Science 85, 145{192 (2000). [36] D. Vella and L Mahadevan, \The cheerios e ect," American Journal of Physics 73, 817{825 (2005). [37] K. J. Stebe, E. Lewandowski, and M. Ghosh, \Oriented assembly of metamaterials," Science 325, 159{160 (2009). [38] M. Cavallaro, L. Botto, E. P Lewandowski, M. Wang, and K. J Stebe, \Curvature-driven capillary migration and assembly of rod-like particles," Proceedings of the National Academy of Sciences 108, 20923{20928 (2011). [39] J. Dervaux, M. C. Resta, and P. Brunet, \Light-controlled ows in active uids," Nature Physics 13, 306 (2017). [40] J. S. Guasto, R. Rusconi, and R. Stocker, \Fluid mechanics of planktonic microorganisms," Annual Review of Fluid Mechanics 44, 373{400 (2012). [41] G. L. Mino, ~ M. A. R. Koehl, N. King, and R. Stocker, \Finding patches in a heterogeneous aquatic environment: ph-taxis by the dispersal stage of choano agellates," Limnology and Oceanography Letters 13 2, 37{46 (2017). [42] R. Rusconi, J. S. Guasto, and R. Stocker, \Bacterial transport suppressed by uid shear," Nature Physics 10, 212 (2014). [43] R. Stocker, \Marine microbes see a sea of gradients," Science 338, 628{633 (2012). [44] G. Vesey, P. Hutton, A. Champion, N. Ashbolt, K. L. Williams, A. Warton, and D. Veal, \Application of ow cytometric methods for the routine detection of cryptosporidium and giardia in water," Cytometry: The Journal of the International Society for Analytical Cytology 16, 1{6 (1994). [45] S. Jana, A. Eddins, C. Spoon, and S. Jung, \Somersault of paramecium in extremely con ned environ- ments," Scienti c reports 5, 13148 (2015). [46] W. M. Durham and R. Stocker, \Thin phytoplankton layers: characteristics, mechanisms, and conse- quences," Annual Review of Marine Science 4, 177{207 (2012). [47] A. Lenshof and T. Laurell, \Continuous separation of cells and particles in micro uidic systems," Chem- ical Society Reviews 39, 1203 (2010). [48] A. Sauret, E. C. Barney, A. Perro, E. Villermaux, H. A. Stone, and E. Dressaire, \Clogging by sieving in microchannels: Application to the detection of contaminants in colloidal suspensions," Applied Physics Letters 105, 074101 (2014). [49] E. Dressaire and A. Sauret, \Clogging of micro uidic systems," Soft Matter 13, 37{48 (2017). [50] A. Sauret, K. Somszor, E. Villermaux, and E. Dressaire, \Growth of clogs in parallel microchannels," Physical Review Fluids 3, 104301 (2018).

Journal

PhysicsarXiv (Cornell University)

Published: Nov 27, 2020

There are no references for this article.