Microfluid Nanofluid (2012) 12:761–770 DOI 10.1007/s10404-011-0907-1

RESEARCH PAPER

Transient flow of microcapsules through convergent–divergent microchannels E. Leclerc • H. Kinoshita • T. Fujii D. Barthe`s-Biesel

•

Received: 27 July 2011 / Accepted: 24 October 2011 / Published online: 3 December 2011 Ó Springer-Verlag 2011

Abstract The study deals with a microfluidic method to investigate the transient behavior of microcapsules in flow. The technique consists of investigating ovalbumin microcapsules passing through a convergent–divergent microchannel made of PolyDiMethylSiloxane. We work with three types of square microchannel with, respectively, cross section values of h 9 h = 30 9 30, 50 9 50 and 70 9 70 lm. The microchannels length is L = 3h. We analyze the kinetics of deformation of the microcapsules in the microchannels for velocity ranging from 2 to 5 cm/s and for microcapsule size ratio d/h ranging from 0.9 to 2.5. The relaxation process at the pore outlet is modeled using an exponential relaxation law. We show that that the relaxation time at the divergent outlet depends on the microcapsule size ratio d/h. Thanks to the analytical expression of the relaxation, we extract a shear modulus of the membrane equal to 0.04 N/m. This value is consistent with the value of 0.07 N/m that we found using the steady state analysis performed in cylindrical glass capillaries. Thus, it is interesting to notice that the microcapsule behavior based on a simple analytical model can be successfully described despite the complex flow situation consisting of deformable microcapsule in confined square microchannels. Keywords Microcapsules  Convergent divergent microchannels  Transient flows E. Leclerc (&)  D. Barthe`s-Biesel UMR CNRS 6600, Biome´canique et Bioinge´nierie, Universite´ de Technologie de Compie`gne, Compie`gne, France e-mail: [email protected] H. Kinoshita  T. Fujii Institute of Industrial Science, The University of Tokyo, Tokyo, Japan

1 Introduction Bioartificial micro-capsules consist of an internal liquid medium protected by a semi-permeable membrane. Usually, this membrane has a thickness which is small compared to the radius of curvature to enhance the transmembrane exchanges. Such micro capsules are used to protect active molecules and to release them at specific locations (Gombotz and Wee 1998; Chan 2005). This release is achieved by rupture of the capsule (under mechanical stress or chemical attack) or by diffusion. Capsules are largely used in industry such as cosmetic (protection, slow release of active substances), pharmaceutical (vectorization and/or protection of molecules, slow release of active reagents), food (fermentation, flavor conservation) and biomedical (production of antibodies or others bio molecules of therapeutic interest) industry (Fairhurst and Loxley 1954; Orivea et al. 2008; Chambina et al. 2007). These various applications imply a good control of the membrane mechanical properties among others (permeability, biocompatibility, etc.). A few methods have been proposed to assess the mechanical properties of capsule membranes. They are based on the measurement of the capsule deformation under a well-defined stress. Large millimeter-sized artificial capsules can be manipulated fairly easily and are thus amenable to evaluation through squeezing between two plates (Feng and Yang 1973; Lardner and Pujara 1980; Carin et al. 2003; Risso and Carin 2004; Rachik et al. 2006) or deformation in a shear flow (Chang and Olbricht 1993; Walter et al. 2001). The manipulation of small micron size capsules and cells is much more difficult as it requires the use of micro techniques such as partial aspiration of the membrane in a micropipette under a given pressure (Hochmut 2000) or poking the capsule with an

123

762

AFM probe (Fery and Weinkamer 2007). These techniques are difficult to use in a routine way. There is thus a need for a method to measure the mechanical properties of micrometer-sized capsules that can be implemented fairly easily. We have designed such a method where a micro-capsule suspension is flowed in a small tube with diameter of the same order as the capsule size. The capsules, which are initially spherical, are deformed into slug or parachute shapes depending on the size ratio and the flow strength. A comparison of the particle deformed profile and velocity with the predictions of a numerical model of a flowing capsule allows us to determine the shear modulus of the membrane (Lefebvre et al. 2008; Chu et al. 2010). These measurements are made for a steady deformed state of the capsule and necessitate long enough tubes. One drawback is then that the pressure drop in the tube can become quite large. This limits the range of flow rates that can be used. An alternative is to flow the capsules through short pores and to analyze their transient behavior. A convenient way to produce versatile pore geometries is to use the microfluidic technique. We thus propose to study to the behavior of a microcapsule flowing through square cross-section microchannels with different transversal dimensions. This particular geometry creates transient flow situations in the convergent section where the flow is accelerated and in the divergent section where the flow is decelerated. We will show in particular how the relaxation process at the exit of the pore can be analyzed in terms of the capsule elastic properties. In the first section, we describe the experimental set-up and the capsule preparation. Then, we give results showing the motion and deformation of capsules through pore of different dimensions. The relaxation process at the exit of a pore is followed in time. In the last section, we discuss the experimental results and analyze the relaxation with an analytical model designed for an initially spherical capsule relaxing in an unbounded domain.

2 Materials and methods 2.1 Micro-channel design and fabrication The microchannels are fabricated by conventional PDMS replica moulding and then bonded on a thin PDMS layer spin coated on a thin slide glass. Each pore has a straight central part with a square crosssection h 9 h and length L = 3h. The symmetrical inlet and outlet sections are comprised between cylinders of height h and radius RC = 2.5h, as shown in Fig. 1. We have designed three pores corresponding to cross-sections

123

Microfluid Nanofluid (2012) 12:761–770

Fig. 1 Pore design: the central part has a square section with side h and length L

30 9 30 lm (S1 series); 50 9 50 lm (S2 series) and 70 9 70 lm (S3 series). 2.2 Microcapsules suspension preparation Microcapsules with a a cross-linked ovalbumin membrane are prepared by means of a method developed earlier (Andry et al. 1996). A 10 % (w/v) ovalbumin (Sigma) solution in pH 6.8 phosphate buffer (6 mL) is emulsified (Heidolph RGL 500 stirring motor, Prolabo-VWR, France) for 5 min at room temperature in 30 mL of cyclohexane containing 2 % (w/v) sorbitan trioleate, using a stirring rate of 1,550 rpm. Then, 40 mL of a 2.5 % (w/v) solution of terephthaloyl chloride in a chloroform:cyclohexane (1:4, v/v) mixture is added to the emulsion and stirring is continued for 5 min. The reaction is ended by dilution with 40 mL cyclohexane. The resulting microcapsules are separated, and washed successively with cyclohexane, water containing 2% (w/v) polysorbate, and finally thrice with pure water. The mean diameter of the population is 68 ± 17 lm, as determined by a laser diffraction technique (Coulter Particle Sizer, type LS 230, BeckmanCoulter, France). The microcapsules are stored in distilled water. A 60 lL volume of capsule sediment is mixed with 2 mL of 100% glycerin. The resulting suspending medium correspond to a solution of 97 % glycerin. After the suspension is prepared, the experiments are done within the next hour to minimize osmotic pressure effects between the capsules and the glycerin. Experiments are performed at a room temperature of 20°C. 2.3 Experimental procedure The fluids are perfused with a syringe pump with flow rates ranging from 1 to 10 lL/min. The visualizations are done with a high speed CMOS phantom camera (V7-1) on a conventional optical inverted microscope and connected to a computer. The exposition time is set to 90 ls with images

Microfluid Nanofluid (2012) 12:761–770

763

recorded at 5,000 (relaxation analysis) to 10,000 frames per second (flow in the pore analysis). The profile contours are analyzed with the ImageJ image processing software. The length scale is determined from the known pore width. Because of varying optical contrast of the pictures, the extraction of the outline of the capsule cannot be automated and has to be performed manually by placing points along the contour (between 30 and 50, depending on the complexity of the shape). An example of typical video images is shown in Fig. 2. We use the microchannel width as internal etalon. Due to optical effects, the square channel walls appear as thick lines on the images. The channel width W is then measured between the tops or the bottoms of the two lines. The typical dimension of the channel width is of order 17 pixels for 50 lm in the relaxation process analysis (5,000 fps). In the motion analysis inside the pore, working at 10,000 fps, the channels width is of order of 25 pixels for 50 lm. Then, we estimate the error on the length scale of the capsules L0 to be of order of 2 pixels leading to 6 lm and 4 lm of error respectively. This results to estimate an error of order of 8% The size ratio d/h is defined as the ratio between the diameter d of the circular microcapsule profile before the pore inlet and the height h of the pore (Fig. 1). Following Risso

et al. (2006), we characterize the deformation of the microcapsules by means of the maximum length of the deformed micro-capsules in the axial Lx and lateral Ly directions. The lengths Lx, Ly are measured as a function of time (as given by the recording software) or of the position X/h of the capsule front along the pore axis (X/h = 0 at the beginning of the square section part of the pore). Due to the video camera set up, the measurements inside the pore and in the pore outlet (for the relaxation phenomenon) are extracted from independent images sequences. Finally, the velocity V of the microcapsule is obtained by dividing the distance covered by the front of the capsule over several frames by the corresponding time (as given by the recording software).

Fig. 2 Flow of a microcapsule along pore S2 (V = 1.7 cm/s, d/h = 1.5). a The microcapsule enters the constriction, b the microcapsule flows in the central part of the constriction, c the

microcapsule exits from the pore, d the microcapsule has relaxed back to a spherical shape. Images (a–c) come from the same sequence, d comes from a second sequence

3 Results 3.1 Microcapsules flow through a S2 pore 3.1.1 Flow phases As an example, we show typical observations of a capsule flowing through a S2 pore with velocity V = 1.7 cm/s

123

764

(Fig. 2). The microcapsule has a circular shape with d/h = 1.8 at the pore inlet (Fig. 2a). At the pore entrance, the microcapsule begins to take a slug shape: this is characterized by a longitudinal elongation and a transversal compression (Fig. 2a). Inside the pore, the capsule has a slug shape with a convex rear (Fig. 2b). It should be noted that even for larger velocities (up to 5 cm/s), a concave rear curvature shape is never observed inside the pore. At the exit, the pressure increases, the capsule thus expands laterally and the rear becomes flat or sligthly concave (Fig. 2c). After flowing out of the pore, the capsule recovers its original circular shape after a relaxation phase. We show the values of Lx/h and of Ly/h extracted from the images sequences as a function of the nose position X/h for five size ratios (d/h = 0.9; 1; 1.1; 1.3 and 1.4) and for V = 2.9 cm/s in Fig. 3a, b. The effect of velocity is presented in Fig. 3c, d. Those graphs show that the capsule takes a slug shape inside the pore because for X/h = 2, the width Ly/h = 1 is that of the pore and the length Lx/h is smaller than 2. The slug length increases with the size ratio (Fig. 3b) or with the velocity (Fig. 3d). Altogether, the overall capsule elongation in the pore is Lx/d B 1.18. At the outlet, the transversal expansion Ly/h is larger than the initial value d/h.

Microfluid Nanofluid (2012) 12:761–770

all capsules, while Lx0/h takes different values depending on the size ratio and the velocity as shown in Fig. 3. The end of the relaxation is defined by Lx = Ly = d when the microcapsule has recovered its initial circular shape. We find that for d/h = 1.5, the relaxation of Ly is independent of the velocity as shown in Fig. 4a. The relaxation ends when the front of the capsule reaches the position (X - L)/h&5. A small overshoot of Ly during the relaxation occurs for large size ratio (d/h = 1.8) as shown in Fig. 4c. The overshoot amplitude increases with the velocity of the microcapsule. However, the velocity does not influence significantly the relaxation distance of Ly when d/h = 1.8 (Fig. 4c). When Ly is plotted as a function of time rather than distance, we find again that the relaxation process does not depend significantly on velocity for the range of values tested. The axial length Lx first decreases due to the lateral expansion and the formation of the parachute, and then increases back to its initial value Lx = d. The amplitude of this phenomenon increases with velocity as shown in Figs. 4b, d. Altogether, the end of relaxation is independent of the velocity of the microcapsules but is size ratio dependent. 3.3 Effect of the cross section on the microcapsule deformation

3.2 Microcapsule relaxation at the exit of a S2 pore 3.3.1 Capsule flow through a S1 pore The microcapsules relaxation is measured at the exit of a S2 pore for capsules such that d/h [ 1. The deformed profile lengths Lx0, Ly0 at the beginning of relaxation, correspond those measured for X = L. Thus, Ly0/h = 1 for Fig. 3 Evolution of Ly/h and Lx/h along the S2 pore; a, b for various size ratio at V = 2.9 cm/s; c, d for two velocities at d/h = 1.1; X/h is the position of the front of the microcapsule. Arrows show the pore inlet and outlet

123

The advantage of the S1 pores with the small cross section is that they lead to larger size ratios (up to 2.5) while using the same microcapsules suspension. The different stages of

Microfluid Nanofluid (2012) 12:761–770

765

Fig. 4 Evolution of Lx/h and Ly/h during the relaxation from the pore exit; a, b for d/h = 1.5; c, d for d/h = 1.8; Arrows show the pore outlet

Fig. 5 Microcapsule with an aspect ratio of d/h = 2 flowing at V = 2 cm/s inside a S1 pore

deformation of a microcapsule (d/h = 2, V = 2cm/s) flowing through a S1 pore are shown in Fig. 5. We find that the behavior of the microcapsules with size ratio ranging between d/h = 1.4 and 1.75 is qualitatively similar in the S1 and S2 pores: after the elongation in the inlet, a steady slug shape is reached in the center of the pores (Fig. 6a, b). However, for d/h = 2 the slug shape is barely reached as shown in Fig. 6b (square data): when the capsule tip is at the exit, its back has just entered the pore since the capsule length is Lx/h = 3. The overall capsule elongation in a S1 pore is at most Lx/d B 1.5. It is clear that the confinement in the S1 pore increases the capsule deformation as compared to the S2 pore. For example, for d/h = 1.4 and

V = 2.9 cm/s in the S2 pore, the maximum elongation is Lx/d = 1.14, whereas for d/h = 1.5 and V = 2cm/s in the S1 pore, the elongation is Lx/d = 1.26 (Fig. 6). In the S1 pores, we never observed the parachute shape of the microcapsules. At the exit, a transversal expansion of the microcapsule occurs leading to a concave rear as shown in Fig. 5d, e. This leads to a radial expansion and to an axial contraction as shown in Fig. 6a, b. The full relaxation out of a S1 pore could not be analyzed because the initial capsule shape was not recovered within our observation window. Furthermore large capsules with aspect ratio d/h = 2.5 are damaged when they exit from the pore as shown in Fig. 7.

123

766

Microfluid Nanofluid (2012) 12:761–770

Fig. 6 a–d Effect of the size ratio and velocity on Lx and Ly inside a S1 pore. Arrows show the pore inlet and outlet

For d/h \ 1, the relaxation curves of the microcapsules are different from the ones observed in the S1 and S2 pores. The behavior of Lx/d is qualitatively the same as in pore S2 : a small decrease is followed by a return to the initial shape. However, since the capsule is smaller than the pore, it has room to expand laterally during the exit process, so that there is an overshoot of Ly/d over the equilibrium value of 1 which is due to the adverse pressure gradient in the outlet section (Fig. 9). This overshoot is similar to the one describe experimentally and numerically in cylindrical pipes (Lefebvre et al. 2008; Risso et al. 2006).

4 Analysis of results and discussion Fig. 7 Damage of a d/h = 2.5 microcapsule flowing through a S1 pore

3.3.2 Capsule flow through a S3 pore In those pores, the size ratio of most of the microcapsules ranges between 0.8 and 1.0 leading to weak wall effects and to a spherical shape at the pore inlet when d/h B 1. No significant deformation can be measured for velocities below 4 cm/s and d/h \ 1. Capsules with a size ratio larger than 1 are very few and seem to be more fragile than the smaller ones as they are often damaged by the passage through the pore. The different stages of deformation are shown in Fig. 8 for a capsule (d/h = 0.9, V = 4 cm/s) flowing through a S3 pore. The capsule takes a slug shape inside the pore with no tendency towards the formation of a parachute. The concave rear occurs at the exit for the highest flow velocity (V C 6 cm/s) or for d/h [ 1.

123

4.1 Mechanics of capsule flow in a pore The Reynolds number of the external flow is given by Re = V h/m, where m % 10-3 m2/s is the kinematic viscosity of glycerin. For the S2 pore and a velocity V = 3 cm/s, we find Re = 1.5 9 10-3. The flow of the capsule through the pore is thus inertialess and governed by a balance between the deforming viscous and pressure forces exerted by the external liquid and the shape restoring elastic stress in the membrane. The viscosity of the internal liquid being much smaller than the external one, the viscous shear stress in the internal liquid may be neglected and the main hydrodynamic internal force is a uniform pressure. The flow process is qualitatively similar to the one where a spherical capsule (radius a) flows through a cylindrical pore (radius R) with hyperbolic entrance and exit sections. This situation has been modeled numerically

Microfluid Nanofluid (2012) 12:761–770

767

Fig. 8 Flow of a microcapsule through pore S3 flowing at 4 cm/s for d/h = 0.9. a The microcapsule before the constriction, b the microcapsule enters the constriction, c the microcapsule flows in the central part of the constriction, d the microcapsule exits from the constriction

Fig. 9 Effect of the size ratio and velocity on Ly relaxation in the S3 pores

(Que´guiner and Barthe`s-Biesel 1997; Lefebvre and Barthe`s-Biesel 2007). During the entrance phase, the upstream pressure and flow stream push the capsule into the tube. The membrane elongates at the front as the capsule enters the pore. When the whole capsule is inside the pore, the elastic tension must balance the viscous pressure drop in the external liquid. Since the external pressure is larger at the rear than at the front, the membrane curvature is larger at the front than at the rear, thus leading to slug or parachute shapes of the capsule as can be noted in Figs. 2b, 5c, 8c. As the front of the capsule enters the exit section, it becomes free to expand laterally since the exit section opens up quickly. The capsule is then subjected to the elastic forces that tend to return the profile to the initial

circular shape and to a large pressure difference between the back and the top. This pressure difference leads to flat or slightly concave back profiles as shown in Figs. 2c, 5d, 8d. When the whole capsule is out of the pore, the outer flow has almost stopped, the outer pressure is uniform and the inertialess relaxation process back to the initial shape is driven by the elastic energy stored in the membrane. The numerical models (Que´guiner and Barthe`s-Biesel 1997; Lefebvre and Barthe`s-Biesel 2007) of the flow of a capsule through cylindrical pores show that the capsule deformed profile becomes steady only after its front has reached a finite position in the tube. This entrance length depends on the size ratio and on the flow strength. It is of order 3.5–5 pore radii for small capsules (a/R \ 1) and increases to 5.5–8 radii for large capsules (a/R = 1.2 - 1.4). This means that the capsule shapes we observed were probably not quite steady except those in the S3 pore (Fig. 8c). The absence of parachute in the S3 pore may be attributed to a small pre-inflation due to osmotic effects during storage (Lefebvre and Barthe`s-Biesel 2007, Lefebvre et al. 2008). The models also find that there is a critical flow strength past which no steady state can be obtained and the capsule undergoes continuous elongation. This occurs when the viscous stress becomes too large to be balanced by the membrane elastic forces. This critical flow strength deceases when the ratio a/R increases. It could be a similar phenomenon which leads to capsule damage in the S1 section pore (Fig. 7) where the size ratio is large. 4.2 Condition for slug formation Capsules such that d/h [ 1 can take a slug shape only if they are small enough to fit in the whole pore. Assuming

123

768

Microfluid Nanofluid (2012) 12:761–770

that the upstream shape of the capsule in the inlet is roughly an ellipsoid with diameters d and h, the capsule volume is p A ¼ hd 2 6 Assuming then that the capsule becomes a slug with a cylindrical central part of radius h/2 closed by two spherical caps, its volume can be evaluated as p p A ¼ ðLc  hÞh2 þ h3 4 6 where Lc is the total length of the slug. Equating these two expressions, we find the relation between the capsule initial diameter and the slug length    1=2 d 3 Lc ¼ 1 þ1 h 2 h The condition for the capsule to fill the pore is Lc = L. The maximum capsule diameter follows readily d=h  2 for

L=h ¼ 3

Indeed, in Fig. 5 we show a capsule with size ratio d/h = 2 that fills the S1 pore. The assumption of a cylindrical middle part for the slug probably overestimates the length of the slug as the capsule may use some of the corner space of a square section pore. If we assume a square central section for the slug the maximum values become d/h = 2.2. Of course, the membrane of large capsules (d/h [ 1) is deformed and thus stressed before entering the pore. This leads to an internal pressure which is larger than the pressure in the spherical capsule. 4.3 Transient behavior The transient entrance or exit of the capsule into the pore is very difficult to analyze in a quantitative fashion as there is no model of the flow of a closely fitting spherical capsule into a square section pore. It is interesting to analyze the relaxation process in the pore outlet. The best parameter to consider is the length Ly(t) when it evolves monotonously from Ly = h to Ly = d for d/h [ 1 (Figs. 4, 9). Since the relaxation is inertialess, it is well approximated by an exponential law   Ly ðtÞ Ly ð0Þ ¼1 1 ð1Þ expt=sc d d where sc is the capsule relaxation time. The relaxation begins at time t = 0 when the nose exits from the pore which implies Ly(0) = h. We recast the exponential law as

123

Fig. 10 Relaxation function Y as a function of time for two size ratios and for a S2 pore

 3 2 L ð0Þ 1  1  yd 5 ¼ t=sc Y ¼ ln4 Ly ðtÞ=d

ð2Þ

and plot Y as a function of time for the relaxation curves in Figs. 4 and 9 where the time t corresponding to each position X is recorded by the camera. As shown in Fig. 10, we find that the exponential law fits the data well and leads to relaxation times sc = 2.3 9 10-3s for d/h = 1.5 (R2 = 0.998) and sc = 2.4 9 10-3s for d/h = 1.8 (R2 = 0.907). In principle, we could also use the time evolution of the axial length Lx to obtain the capsule relaxation time. But Lx first decreases as the capsule is compressed. This implies that the exponential relaxation would start from the minimum value of Lx. This minimum is difficult to measure with accuracy and furthermore, the relaxation being shorter, is difficult to assess. As there is no available model of the relaxation of a capsule constrained between two walls, we are left with the model of the relaxation of an initially spherical capsule in an unbounded fluid medium at rest, which has been solved by Barthe`s-Biesel and Rallison (1981). They considered the case where the capsule was slightly deformed into an ellipsoid and left to recover its initial shape under the influence of the elastic forces in the membrane. They find that the relaxation time is given by s¼

3ð19k þ 16Þð2k þ 3Þ la pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 3G 5ð19k þ 24Þ  5377k þ 14256k þ 9792 s

ð3Þ

where l is the suspending fluid viscosity, Gs is the shear elastic modulus of the capsule membrane, a is the capsule radius, and k is the ratio of the internal fluid viscosity to the

Microfluid Nanofluid (2012) 12:761–770

external one. In the present study the external fluid viscosity of glycerin is roughly 1 Pas whereas the internal viscosity of the albumin solution is 0.03 Pas, leading to a ratio k & 0. The largest relaxation time is then la s ¼ 2:29 ð4Þ Gs We now equate the value of the relaxation time given by Eq. (4) to the experimental one, taking a = d/2 and find Gs = 0.045 N/m for d/h = 1.8 and Gs = 0.036 N/m for d/h = 1.5. These values are somewhat lower than the one Gs = 0.07 N/m obtained from the analysis of the steady shape of the same capsules flowing in a cylindrical tube (Lefebvre et al. 2008). This can be attributed to the fact that the measured relaxation time is larger than the theoretical one. This increase in relaxation time may be attributed to viscous friction forces near the walls (recall that there are no walls in the model). What is needed now is a model of the flow of capsules in narrow channels to which the experimental results can be compared. Altogether the order of magnitude of the value we obtain for the elastic shear modulus of the capsule membrane is certainly correct if slightly underestimated. It is interesting to note that it has been obtained with a simple analytical model that is easier to use than a complex numerical code that solves for fluid structure interactions in a confined domain.

5 Conclusion In conclusion, we have investigated the transient behavior of microcapsule flowing in convergent divergent micro-channels. The results show that the capsules are able to deform to squeeze through the narrow channels and still recover their initial shape provided the deformation remains moderate with an elongation ratio less than 1.2. Larger elongations seem to either lead to a plastic deformation of the capsule or to breakup. The analysis of the exit from the pore allows us to measure a characteristic relaxation time when the capsule returns to its initial shape. The full analysis of the relaxation process necessitates a complete fluid-structure mechanical model that takes into account the confinement due to the channel wall. This is a very complex model which is not available presently. However, using a simple model of the relaxation of an initially spherical capsule in an unbounded fluid domain, we were able to infer a value of the capsule membrane shear elastic modulus with the proper order of magnitude. This is encouraging because this very simple relaxation model could be used to analyze relaxation experiments of capsules flowing through narrow channels and thus discriminate between different capsule populations even if the exact value of the shear modulus is a little off.

769 Acknowledgments This work was supported by the Conseil Re´gional de Picardie (projects lFIEC and MODCAP). The experimental work was performed in collaboration with Pr. Teruo Fujii research group (French-Japanese SAKURA grant). We thank Florence Edwards Levy from University of Reims who provided the microcapsules used in this study. We also thank Takuji Okamoto for the microchips fabrication and Pr. Marie Oshima (IIS, University of Tokyo) who provided the high speed phantom camera.

References Andry MC, Edwards-Le´vy F, Le´vy MC (1996) Free amino group content of serum albumin microcapsules.iii. a study at low ph values. Int J Pharm 128:197 Barthe`s-Biesel D, Rallison JM (1981) The time-dependent deformation of a capsule freely suspended in a linear shear flow. J Fluid Mech 113:251–267 Carin M, Barthe`s-Biesel D, Edwards-Le´vy F, Postel C, Andrei D (2003) Compression of biocompatible liquid filled hsa-alginate capsules: determination of the membrane mechanical properties. Biotechnol Bioeng 82:207 Chambina O, Voilleya A, Gharsallaoui A, Roudauta G, Saurela R (2007) Applications of spray-drying in microencapsulation of food ingredients: an overview. Food Res Int 40:1107–1121 Chang KS, Olbricht WL (1993) Experimental studies of the deformation and breakup of a synthetic capsule in steady and unsteady simple shear flow. J Fluid Mech 250:609–633 Chang T (2005) Therapeutic applications of polymeric artificial cells. Nat Rev 4:221 Chu TX, Salsac A-V, Leclerc E, Barthe`s-Biesel D, Wurtz H, Edwards-Le´vy F (2010) Comparison between measurements of elasticity and free amino group content of ovalbumin microcapsule membranes: discrimination of the cross-linking degree. J Colloid Interface Sci 355:81–88 Fairhurst D, Loxley A (1954) Micro- and nanoencapsulation of waterand oil-soluble actives for cosmetic and pharmaceutical applications. Particle Sciences Inc., Bethlehem Fery A, Weinkamer R (2007) Mechanical properties of micro- and nanocapsules: single capsule measurements. Polymer 48:7221–7235 Feng WW, Yang WH (1973) On the contact problem of an inflated spherical nonlinear membrane. J Appl Mech 40:209–214 Gombotz WR, Wee S (1998) Protein release from alginate matrices. Adv Drug Deliv Rev 31:267 Hochmuth RM (2000) Micropipette aspiration of living cells (review). J Biomech 33:15–22 Lardner TJ, Pujara P (1980) Compression of spherical cells. Mech Today 5:161–176 Lefebvre Y, Barthe`s-Biesel D (2007) Motion of a capsule in a cylindrical tube: effect of membrane pre-stress. J Fluid Mech 589:157–181 Lefebvre Y, Leclerc E, Barthe`s-Biesel DD, Walter J, Edwards-Le´vy F (2008) Flow of artificial microcapsules in microfluidic channels: a method for determining the elastic properties of the membrane. Phys Fluids 20:123102–123112 Orivea G, Hernndeza R, Castroa M, Murua A, Portero A, Pedraz JL (2008) Cell microencapsulation technology: towards clinical application. J Controlled Release 138:76–83 Que´guiner C, Barthe`s-Biesel D (1997) Axisymmetric motion of capsules through cylindrical channels. J Fluid Mech 348:349– 376 Rachik M, D. Barthe`s-Biesel D, Carin M, Edwards-Le´vy F (2006) Identification of the elastic properties of an artificial capsule membrane with the compression test : effect of thickness. J Colloid Interface Sci 301:217

123

770 Risso F, Carin M (2004) Compression of a capsule: mechanical laws of membranes with negligible bending stiffness. Phys Rev E 69:061601–061608 Risso F, Colle-Paillot F, Zagzoule M (2006) Experimental investigation of a bioartificial capsule flowing in a narrow tube. J Fluid Mech 547:149–173

123

Microfluid Nanofluid (2012) 12:761–770 Walter A, Rehage H, Leonhard H (2001) Shear induced deformation of microcapsules: shape oscillations and membrane folding. Colloids Surf A: Physicochem Eng Aspects 183–185:123–132