Controlling molecular transport in minimal emulsions

Abstract

Emulsions are metastable dispersions in which molecular transport is a major mechanism driving the system towards its state of minimal energy. Determining the underlying mechanisms of molecular transport between droplets is challenging due to the complexity of a typical emulsion system. Here we introduce the concept of 'minimal emulsions', which are controlled emulsions produced using microfluidic tools, simplifying an emulsion down to its minimal set of relevant parameters. We use these minimal emulsions to unravel the fundamentals of transport of small organic molecules in water-in-fluorinated-oil emulsions, a system of great interest for biotechnological applications. Our results are of practical relevance to guarantee a sustainable compartmentalization of compounds in droplet microreactors and to design new strategies for the dynamic control of droplet compositions.

ARTICLE

Received 24 Dec 2014 | Accepted 8 Dec 2015 | Published 22 Jan 2016

Emulsions are metastable dispersions in which molecular transport is a major mechanism driving the system towards its state of minimal energy. Determining the underlying mechanisms of molecular transport between droplets is challenging due to the complexity of a typical emulsion system. Here we introduce the concept of minimal emulsions, which are controlled emulsions produced using microuidic tools, simplifying an emulsion down to its minimal set of relevant parameters. We use these minimal emulsions to unravel the fundamentals of transport of small organic molecules in water-in-uorinated-oil emulsions, a system of great interest for biotechnological applications. Our results are of practical relevance to guarantee a sustainable compartmentalization of compounds in droplet micro-reactors and to design new strategies for the dynamic control of droplet compositions.

DOI: 10.1038/ncomms10392 OPEN

Controlling molecular transport in minimal emulsions

Philipp Gruner1, Birte Riechers1,2, Benot Semin1,3, Jiseok Lim1,4, Abigail Johnston1, Kathleen Short1 & Jean-Christophe Baret1,2

1 Max-Planck-Institute for Dynamics and Self-Organization, Droplets, Membranes and Interfaces, Am Fassberg 17, DE-37077 Goettingen, Germany. 2 CNRS, Univ. Bordeaux, CRPP, UPR 8641, 115 Avenue Schweitzer, 33600 Pessac, France. 3 Laboratoire de Physique Statistique, Ecole Normale Suprieure, UPMC Univ Paris 06, Universit Paris Diderot, CNRS, 24 rue Lhomond, 75005 Paris, France. 4 School of Mechanical Engineering, Yeungnam University, Gyeongsan 712-749, Republic of Korea. Correspondence and requests for materials should be addressed to J.-C.B. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10392

An emulsion is the dispersion of one uid into another, stabilized by surfactant molecules1,2. Emulsions have a wide range of technological applications, including the use

in food products, paints, cosmetics, chemical synthesis and drug delivery3. In recent years, droplet-based microuidics has been proposed as a means for miniaturization and automatization of biochemical assays. The billions of microcompartments contained in an emulsion provide an environment ideal for the parallelization of assays49. This concept was shown to be very powerful for applications relying on high-throughput parallelized measurements such as drug screening10,11, biomarker analysis1214, cell screening1517 or directed evolution of enzymes18,19. These emulsions are unconventional in the sense that individual droplets have a unique and time-varying composition, depending on their initial loading and the biochemical processes occuring within them. From a physics perspective, emulsions are systems intrinsically out of thermodynamic equilibrium1,2. They are kinetically stabilized in a metastable state by the use of surfactant molecules. The role of the surfactant molecules is to increase the height of the energy barrier between the local energy minima of the system and its global minimum; this minimum is reached with a simple system where both phases are separated by an interface of minimal energy and in which the chemical potentials of all species is homogeneous. In addition to the classical ageing processes of occulation, coalescence, gravitational separation and Ostwald ripening, solute transport drives these emulsions towards equilibrium. The heterogeneity in droplet composition leads to an increased number of local minima in the energy landscape and hence the relaxation to the global minimum can follow a complex path, resulting in complex dynamics.

We consider the kinetics of equilibration of concentration differences between droplets containing solutes which are poorly soluble in the continuous phase. We use uorinated oils as the continous phase and aqueous droplets as the dispersed phase. This system is of particular interest for biochemical applications20,21. Although the solubility of organic molecules in uorinated oils is normally very low22,23, surfactant molecules mediate their solubility, as they do for organic systems24, through their amphiphilic character20. Thus, there is a nite solubility of encapsulated compounds in the continuous phase, which can lead to cross-talk between droplets. In droplet-based microuidic experiments, the droplets are designed to be independent micro-reactors, with the assumption that encapsulated compounds are not leaking into the continuous phase. This assumption is not valid in the presence of cross-talk between droplets2529. Accessing quantitative information on mass transport in emulsions is experimentally challenging as the microenvironments of droplets must be precisely controlled to allow for quantitative analysis. Microuidics is a powerful approach in the production of calibrated emulsions, providing access to a wide range of time scales and experimental conditions21,3032. We have previously demonstrated how to measure chemical equilibration in three-dimensional systems33. Macromolecular additives, solubilized in a fraction of droplets, were shown to concentrate organic solutes in these droplets33. This result suggests that the local chemical equilibrium between phases is reached faster than the equilibration of concentrations in each phase by diffusion. Surfactants acting as an energy barrier to phase partioning at the time scale of the whole transport process is therefore excluded.

To further reduce the complexity of the emulsion system, we produce, order and immobilize droplets in a controlled fashion7,34. We introduce the concept of minimal emulsions made of an assembly of xed monodisperse droplets, with controlled centre-to-centre distances. The microenvironment of each droplet is precisely controlled, to a level unreachable in bulk emulsication. As a result, we access fundamental information on

the rate-determining step of transport. We demonstrate that the transport processes follow a universal law based on Fickian diffusion, described using simple thermodynamic arguments. We further use our understanding of the process to effectively control chemical transport between microreactors. We demonstrate the simple control and programming of chemicals in emulsions for targeted delivery into droplets at rates compatible with the typical time scales of biochemical assays. To the best of our knowledge, this result is the rst demonstration of the full control of molecular transport between emulsion droplets. Our approach provides the tools to actively control the composition of droplets, for example in the continuous exchange of buffer or media conditions, without the need to individually manipulate droplets35.

ResultsMinimal emulsions. Emulsions are prepared by dispersing one uid into another. Shearing liquidliquid mixtures in bulk leads to polydisperse emulsions while liquids co-own in microuidic channels form highly monodisperse emulsion droplets36. The monodispersitycrucial for many biochemical applicationsis a key feature for exploring emulsion science21,26,27,31,37. We use these microuidic techniques to control the order and spacing of the emulsion droplets reducing the number of degrees of freedom. In these minimal emulsions, parameters such as the volume fraction of the dispersed phase are precisely dened. We prepare one-dimensional arrays of droplets with alternating composition in three steps. First, we produce a train of monodisperse droplets with alternating composition, then the train of droplet is actuated on-demand towards the storage and analysis zone, and nally the droplets are immobilized by halting the ow of the continuous phase. Alternating droplets are produced with two opposing T-junctions (Supplementary Fig. 1 and Supplementary Note 1). When one of the aqueous uids is advancing into the nozzle, the other stream is blocked resulting in a reliable production of droplets with alternating properties3841. To implement a function for on-demand switching of droplets, we design a pressure controlled hydrodynamic switch. The uid ow is divided at a Y-junction into two microuidic channels of different width. In the absence of any additional applied pressure, droplets preferentially ow towards the wider microuidic channel owing to a lower hydrodynamic resistance7. Upon increase of the pressure level in the wider microuidic channel, the pressure balance is inverted and droplets are directed towards the narrower channel (Supplementary Fig. 1). For the immobilization of droplets we use anchors and rails for guiding and storing droplets34,42,43 (Fig. 1a,b) and create regular arrays of droplets with controllable centre-to-centre distances dc (Fig. 1c,d). Our system is thus reduced to the equivalent of a single membrane with periodic boundary conditions, given that all membranes between the droplets are equivalent and that edge effects are negligible33.

We use uorophores as model molecules. We will focus on three molecules, uorescein, resorun and rhodamine 6G which are all exchanged but exhibit time scales of exchange well separated, from minutes for rhodamine 6G to days for uorescein (Fig. 2c,d). Resorun is known to display exchange with a time scale of hours and will be used as a model dye for convenience33. Two droplet populations of identical size are produced, where uorophores are only present in one population (resorun sodium salt, 100 mM). The surfactant is a peruoropolyether-polyethylenoxide block copolymer (PFPEPEGPFPE) stabilizing our emulsions against coalescence. The intensity of the emitted uorescent light is proportional to the concentration of the uorophore in the relevant concentration range between0.1 and 100 mM. Hence, the concentration of resorun sodium

2 NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10392 ARTICLE

a c

Flow direction

b d

hc

hr

I

y

wr

wt

x d

dc

r

wc

Figure 1 | Microuidics for the control of minimal emulsions. (a) Sketch of the trapping channel based on rails and anchors34,42,43. (b) The

geometry of the channels determines the spacing distance between droplets (hc 20 mm, hr 15 mm, wr 20 mm, wt 80 mm, wc 100 mm,

l varies between 36 and 6 mm). (c) Time sequence (100 ms interval) of droplets buffering each other from one local minimum in surface energy to the next (scale bar, 100 mm). (d) Sketch of the membrane model used to describe the transport process in the array based on the transport through a single membrane (see Supplementary Fig. 2 and Supplementary Note 2).

a b c

t=0 t =25 h t =50 h

t =5 min t =10 min t =13 min

d

e

f

Figure 2 | Molecular transport for two uorophores. Fluorescein (ac) is exchanged at a time scale of order several days in a water-in-uorinated oil emulsion stabilized by a PFPEPEGPFPE surfactant (in HFE-7500). At production, the bright droplets contain Fluorescein at 100 mM in PBS water and the dark droplet PBS only. In contrast, Rhodamine 6G (df) is exchanged over minutes (scale bar, 100 mm). At production, the bright droplets contain Rhodamine 6G at 100 mM in millipore water and the dark droplet millipore water only. Resorun is an intermediate case with a time scale of exchange of the order of 1 h (ref. 33).

salt in individual droplets is determined by the uorescence intensity. Droplets are stored in our arrays and uorescence images are recorded every 5 min until the uorophore concentrations between the two populations are equilibrated (Fig. 3ac). The dynamics of transport of uorophores from the initially lled towards the initially empty droplets are measured by analysing time sequences of uorescence images.

The concentration difference between the two droplet populations decays exponentially (Fig. 3d,e). The sum of intensity

over all droplets stays constant during the experiment (Fig. 3d), indicating that photobleaching is negligible and that the concentration of the dye in the continuous phase is negligible compared with the concentration in the droplet. This latter point is expected from the low solubility of organic molecules in uorinated oils. This result experimentally conrms that the alternating one-dimensional array is an eigenmode of the discrete diffusion equation considering that the transport occurs as a diffusive process through membranes arranged at the node of a one-dimensional lattice33. The time scale of the exchange in the array is therefore directly linked to the permeability of one single membrane. In addition, we recover that the time scale of the transport is inversely proportional to the concentration of surfactant (Fig. 3f; ref. 33). Next, we perform measurements of the relaxation to equilibrium for various droplet spacing, from E1 mm (touching droplets) to 30 mm (Fig. 3g). In all cases, the relaxation is exponential and the time scale t decreases as the spacing increases. Quantitatively, the time scale of exchange is related to the membrane permeability P using t V/SP33, where

V is the droplet volume and S its surface area. Figure 4 shows the permeability as a function of the dimensionless spacing between the droplets d dc/2r (where dc is the centre-to-centre distance

between the droplets of radius r), for various concentrations of surfactant. For a value d 1, undeformed spherical

droplets would touch leading to innite permeabilities in the diffusion-limited model. We observe experimentally a strong increase of the permeability as the droplets approach towards contact (Fig. 4), but the limit d 1 is not reached as the droplet

deform.

Permeation model. In principle, the process of molecular exchange is predicted by thermodynamics: the chemical potential of the solute must equilibrate over the whole system through molecule transport. The process of molecular exchange involves four steps44: in the rst step, solute molecules diffuse towards the wateroil interface; in the second step, the molecules partition between the water and the oil phase in the vicinity of the interface; third, the molecules diffuse through the oil phase towards a neighbouring droplet; and nally, as a fourth step, the molecules partition again, between the oil and the initially empty droplet until the equilibrium is reached. To date, the rate-determining step of molecular transport in water-in-oil emulsions has not been explicitly identied. However, recent numerical simulations have suggested that the transport of molecules across the phase boundary is the rate-limiting step45. In our experiments, the kinetics of the equilibration across the interface is constant and only the diffusive time scale in the oil varies. Qualitatively, since we observe a decay of the transport rate, the process must be diffusion limited. Quantitatively, we analyse the equilibration kinetics in our geometry for a purely diffusive process (that is, the rst, second and fourth steps are signicantly faster than the third). To understand the inuence of droplet spacing on the time scale of transport, we use a permeation model (Supplementary Fig. 2 and Supplementary Note 2). For two reservoirs separated by an innite membrane of thickness l, the permeability P of the membrane is given by P KD/l, where D is the diffu

sion coefcient in the membrane and K ceqoil=ceqaq is the

partitioning coefcient of the molecule between the membrane and the reservoirs. In our geometry, the membrane is not described by the single parameter l but includes two parameters, the spacing between the centres of the droplets dc and their radius r. For simplicity, we approximate the trapped droplets as cylinders. From the geometry of the problem, we consider only the contribution of the gradient in the direction of the

NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10392

a b c

Time

d e f g

100

100

101

102

100

100

Fluorophore concentration (M)

Experimental data

Exp(t/)

9101

80

7101

5101

7101

60

c/c 0

c/c 0

3101

c/c 0

40

5101

2101

dc

20

1 wt% surf

2 wt% surf

5 wt% surf

101

3101

0 0 5 10 15 20 25 30 Time (h)

0 5 10 15 20 25 30 Time (h)

0 2 4 6 8 10 Time (h)

0 2 4 6 8 10 Time (h)

Figure 3 | Fabrication and control of minimal emulsions. (a) Sketch of the device for droplet production and storage as alternating row (see Supplementary Fig. 1 and Supplementary Note 1). (b) On one single chip, several packing distance are tested under the same condition. (c) White-light micrograph of an immobilized row of droplets (scale bar, 100 mm) and time evolution of one row of droplets (0, 1.5, 3, 6, 12 and 24 h after immobilization, for a surfactant concentration C 1 wt% and a distance dc 4 mm.) in uorescence. (d) Relaxation to equilibrium of the average uorophore concentrations

of initially lled (lled circles) and empty droplets (empty circles) (e) The concentration difference between the two droplet populations decays exponentially. (f) Variation of the kinetics as a function of surfactant (surf) concentration (for dc 15 mm). (g) Relaxation dynamics of ve one-dimensional

droplet microarrays for edge-to-edge distance dc 2r between the droplets of 30 mm (black), 15 mm (red), 10 mm (blue), 4 mm (pink) and 1 mm (olive)

(for C 1 wt%).

1 1.1 1.2 1.3 1.4 dc/2r

We compare the analytical expression equation 1 to our experimental results in Fig. 4. Qualitatively, the shape of the model curve ts the experimental data over the whole range of parameters tested. For quantitative comparison, we t the experiments with a single t parameter D, the diffusion coefcient of the dye in the oil. The value of K for the various surfactant concentrations (1, 2 and 5 wt%) is obtained independenty from the shake-ask method via spectroscopy (Fig. 4). The values of K are systematically smaller than 0.1, conrming the low solubility of resorun in the oil phase. In addition, we found, to a rst approximation, a linear variation of K with the surfactant concentration. Our results show that surfactant molecules mediate the solubility of organic molecules in uorinated oils and play the key role for the transport of organic molecules in uorinated emulsions. The values of the diffusion coefcient D obtained as a tting parameter are in all cases D 7.90.3 10 9 cm2 s 1.

Considering its size, the freely diffusing uorophore is expected to diffuse with a diffusion coefcient on the order of 10 5 to 10 6 cm2 s 1. We therefore conclude that the diffusing object is much bigger than a single uorophore molecule. With the

StokesEinstein equation, we estimate an average diameter of B200 nm for the diffusing object. This size correlates with the size of surfactant assemblies obtained via dynamic light scattering (B200 nm, See Supplementary Fig. 3 and Supplementary Note 3).

We therefore conclude that assemblies of surfactant molecules, such as reverse micelles or vesicles, provide a nanoscopic environment for the solubilization of organic molecules in uorinated oils and mediate the transport of small organic molecules. Our expression for the permeability shows a divergence close to droplet contact with a scaling PB(d 1) 1/2 (Supplementary

Fig. 2 and Supplementary Note 2). Physically, we expect two mechanisms to limit the divergence. First, our hard sphere model

a

108

b

0.09

1 wt% surfactant 2 wt% surfactant 5 wt% surfactant

0.06

K

Permeability (ms1 )

0.03

0 0 1 2 3 4 5 Surfactant conc (wt%)

109

D=7.6109 cm2s1 K=0.082

D=7.8109 cm2s1 K=0.033

D=8.2109 cm2s1

K=0.016

Figure 4 | Oil Permeability P. (a) The permeability is a function of the centre-to-centre distance between the droplets dc normalized by the droplet diameter 2r. Experimental data are tted using equation (1) (see Supplementary Fig. 2 and Supplementary Note 2) to obtain the diffusion coefcient of the dye in the oil-surfactant mixture using the values of K determined independently (b). We obtain D 7.6

10 9 cm2 s 1 (5% surfactant), D 7.8 10 9 cm2 s 1 (2% surfactant),

D 8.2 10 9 cm2 s 1 (1% surfactant). The value indicates that transport

is mediated by surfactant aggregates affect the partitioning coefcient of the organic molecule between both the phases (error bars arestandard deviation over three experiments); conc, concentration.

dropletdroplet centres expressed in a dimensionless form using d dc/2r (Supplementary Note 2):

P

KDr

!

: 1

p

4

d

p d2 1

arctan

r

d 1

d 1

4 NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10392 ARTICLE

a b

0 s

100

Fluorophore concentration

c(t) (M)

80

30 s

PDMS

PDMS

Oil+ Surfactant + Dye

60

60 s

40

90 s

20

120 s

0 0 0.5 1 1.5 2 2.5 3 3.5 Time t (min)

c

d

Time

Oil + Surf.

+ Fluorescein

Oil + Surf.

+ Resorufin

Oil + Surf.

+ Fluorescein

Oil + Surf.

+ Resorufin

Figure 5 | Flow-induced targeted delivery. (a) A droplet trapped in a microchannel is exposed to an oil doped with uorescent dye (see inset). (b) The partitioning to the droplet leads to an increase of uorescence (scale bar, 40 mm). The rate of molecule uptake by the droplet is given by the delivery rate from the oil (Qoil 0.2 ml min 1). The principle for a

controlled uptake of multiple dyes is shown in c. (c) Droplets of 100 mm in diameter are trapped in a microuidic chamber. We selectively inject uorescent dyes (uorescein and resorun) from the oil and surfactant mixture into the droplet. The dye injected depends on the position of the droplet in the chamber. (d) Experimental realization: uorescent micrograph of the droplets showing the dye uptake occuring at a time scale of the order of 1 min. Here Qoil 0.2 ml min 1.

will break down close to contact. Indeed, forcing the oil lm between the droplets is made at the expense of the increase of pressure in the lm. Droplets will ultimately deform when the capillary pressure g/r is of the order of the lm pressure leading to a non-zero spacing. Second, the partitioning between the oil and the water phases might become the limiting step. However, we will show that this transfer occurs at a much shorter time scale and is not the limiting step in most of the cases.

Flow-induced targeted delivery. To measure the rate of transport accross the oilwater interface, we use a second microuidic approach. By doping the uorous continuous phase with resorun, making use of the surfactant-induced solubility, we induce the uptake of resorun by a droplet trapped in a microchannel (Fig. 5). The time scale of uptake is measured as tB45 s, and is several orders of magnitude smaller than the rates measured in emulsions. Interestingly, it is directly given by the ux of dye carried by the oil. This result indicates that even for short time scales, there is no signicant effect of the partitioning kinetics to slow down the transport across the interface. From this time scale, we can construct a permeability P V=StB10 6 m s 1

(for a droplet size of 50 mm in diameter). This permeability is at least two orders of magnitude larger than the permeabilities obtained in Fig. 4, nally showing that the transport in emulsions is not inuenced by the kinetics of partitioning. The transport process is entirely limited by the uorophore transfer through the continuous phase, from spaced droplets to almost touching droplets, in ow or in quiescent conditions. There is no signicant inuence of a kinetic barrier to the partitioning at the

interface, in contradiction to recent theoretical models45. We further use our approach to demonstrate that droplets trapped in a wide channel are selectively fed by chemicals, using a proper design of microchannels (Fig. 5c,d). In this case, solutions of oil and surfactant doped with resorun or uorescein are injected towards immobilized droplets. The laminar ow provides a control of the delivery towards specic droplets. Here, we obtain a direct control of the droplet composition through the control of the continuous phase.

Emulsion-based targeted delivery. As the partitioning coefcient between the dispersed and the continous phase is the important control parameter for mass transport in emulsions, acting on partitioning coefcients is a key concept to control transport rates. We have previously shown that BSA (bovine serum albumin) added to the aqueous phase slows down the rate of transport through partitioning26,33. In other systems, sugar additives have been used to slow down the transport29. The solubility of organic molecules in an aqueous medium is also altered by the addition of salt46, commonly known as the salting out effect. By increasing the salt concentration, more water molecules are involved in solvation, decreasing the number of water molecules available to hydrate the uorophores. As a result, the solubility in the aqueous phase decreases with increasing salt concentration.

We use this principle to design a scenario where we obtain the transfer of uorophores from one droplet population to the other (Fig. 6a,b). While one population initially contains uorophores and sodium chloride (125 mg), the other population does not contain any additives. Consequently, the partition coefcient of the uorohpores is different between the two droplet populations (Fig. 6c). We observe that the uorophores are enriched in the droplets free of sodium chloride. We obtain an exponential relaxation for the concentration difference from the initial state to the reversed state (Fig. 6a,b). An equal concentration is obtained transiently in the course of the exchange (at times of the order of 1 h in this case). At a later stage, the nal ratio of concentrations in both types of droplets is given by the ratio of the partitioning coefcients with and without salt. Quantitatively, we again correlate the permeability of the oil with the partitioning coefcient (Fig. 6c). Finally, it should be noted that the nal state of the system is not given by two populations of droplets of the same size. At longer times, the transport of the dispersed phase itself becomes signicant to balance the osmotic pressure difference related to the imbalance of salt concentration30,47,48. Therefore, our method provides means to circulate along various paths of the relaxation to the equilibrium of our dispersion.

It is interesting to compare the transport of organic molecules to the transport of the dispersed phase itself. Ostwald rippening is observed in uorocarbon oils but the kinetics of transport is independent of the surfactant concentration (Supplementary Fig. 4 and Supplementary Note 4). As a consequence, there is no signicant increase of water solubility with increasing surfactant concentration. Therefore, the structure of the supramolecular assemblies is not compatible with water-lled micelles, but is rather compatible with an oil-in-oil vesicle-type assembly as observed for organic block copolymers49 (Fig. 7). This structure is also compatible with the experimental observation that various dyes are exchanged with different time scales26,33.

Surfactant-based chemical extraction. To gain additional insights into the surfactant-mediated solubility, we performed a nal set of experiments. Here, we address the opposite case, namely the extraction of compounds from droplets. The principle derived from our previous experiment is to increase the partitioning towards the oil phase to extract the molecules. Emulsions

NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10392

a

b c

1

1.5108

0.5

Time

c/c 0

Permeability (ms1 )

1.2108

108

8109

5109

3109

0

0.5

Experimental data Model fit 2exp(t/[afii9848])1

1 0 1 2 3 4 5Time (h)

0.2 0.3 0.5 0.8 K

Figure 6 | Emulsion-based targeted delivery. (a) White-light picture of an immobilized row of droplets and time sequence of uorescence signals (0, 0.5, 1, 1.5, 2, 2.5, 3 h after immobilization, scale bar, 100 mm). (b) Relaxation of the concentration difference towards the opposite composition. A complete inversion of the concentration is obtained according to an exponential relaxation for droplets initially containing 100 mM resorun sodium salt and10 mg ml 1 sodium chloride (lled circles, see inset) and millipore water droplets (empty circles, see inset). (c) Dependency of permeability P on the partitioning coefcient K: the transport rate is driven by the largest partitioning coefcient in the system. Inset: dependency of partitioning coefcient K on the salt concentration in the aqueous phase cNaCl (error bars are standard deviation over three experiments).

Aqueouscore Fluorinated

core

Vesicular structure

a b

Micellarstructure Oil Oil

~200 nm

Figure 7 | The possible structures of the surfactant supramolecular assemblies. (a) The swolen micelle structure with an aqueous core, (b) the vesicle structure where the surfactant is either extended or forms bilayers. The vesicle structure is compatible with our experimental results. With the micelle structure, the exchange rate of water should depend on surfactant concentration and all dyes should have the same exchange rates (see Supplementary Fig. 4 and Supplementary Note 4).

were formed with 0.5 wt% of PFPEPEGPFPE surfactant comprising an additive, a carboxylic acid surfactant (Krytox-FSH) at low (o1%) and high (30%) mass fractions. To visualize the extraction of compounds from droplets, the organic uorophore rhodamine 6G was added to the aqueous phase. The uorophore was chosen for two major reasons: rst, it is uorescent in both phases, and second, it is a water soluble molecule, that does not partition into the pure uorous phase. Interestingly, we observe that the efciency of the uorophore extraction strongly depends on the surfactant type. In the presence of the sole PFPEPEG PFPE surfactant (0.5 wt%), the uorophore molecules are retained in the droplet (Fig. 8a,b). In contrast, the addition of the carboxylic acid surfactant to the PFPEPEGPFPE surfactant with a surfactant fraction of 30% results in an immediated and complete extraction of the uorophores into the continuous phase (Fig. 8c,d). In microuidics, the process of extraction is completed within less than 1 s (Fig. 8c,d). This experiment shows a direct control on the process of extraction by additives solubilized in the uorous phase and conrms that the time scale for partitioning is several orders of magnitude shorter than the time scale of the diffusive process.

Unravelling relevant interactions. In addition, the use of rhodamine 6G sheds light onto the mechanism of partitioning. To quantitatively analyse the extraction of the uorophores into

the uorous phase, aqueous solutions of rhodamine 6G (100 mM) are exposed to a uorous phase (HFE7500) containing various concentrations of the carboxylic acid surfactant (Fig. 8e,f). In the absence of surfactant, no signicant extraction of molecules is observed. With increasing concentration of the carboxylic acid surfactant, the amount of organic molecules being extracted into the uorous phase is increasing up to full extraction (495%). The absorbance in the uorous phase is proportional to the concentration of surfactant up to about 100 mM. Above this 1:1 molar ratio, no further increase in absorbance, but a bath-ochromic shift is observed. The acid acts as a molecular receptor for rhodamine 6G resulting in the efcient extraction of the organic solutes into the uorous phase.

To validate the stoichiometry of the complex formed in the continuous phase from the two solute species present in the system, a continuous variations experiment is conducted (Fig. 8g,h). The total molar concentration of the acidic surfactant and the uorophore is constant, but their mole fractions are varied. The maximum of absorption in the uorous phase is found at a 1:1 molar ratio indicating that a complex with the corresponding 1:1 stoichiometry is formed.

Specic noncovalent interactions are known to increase the solubility of organic molecules in uorous liquids50 through hydrogen bonding or ion pairing. Fluorosurfactants with a carboxylic acid head group form strong hydrogen bonds with organic molecules51,52. Fluorous carboxylic acids act as molecular receptor for organic molecules signicantly increasing their solubility in uorous liquids53. Nitrogen hydrogen bond are also efcient acceptors51, as shown for pyridines51,52. In summary, noncovalent interactions signicantly improve the extraction of organic molecules into a uorous phase and the efciency is strongly dependent on the compatibility of substrate and receptor.

This result obtained with a seemingly different system compared with the PFPEPEGPFPE molecule used above is important to explain the interactions within the supramolecular assemblies of the block copolymer. The amide group binding the uorinated chains to the PEG group is a good candidate to be at the origin of the interactions between the surfactant assemblies and the solutes. We want to point out that although hydrogen bonding is a good candidate to explain the dyesurfactant interactions, obtaining a direct a priori correlation between structure and exchange rate remains challenging.

With respect to applications of droplet-based microuidics, it becomes clear that impurities of carboxylic acids can have a

6 NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10392 ARTICLE

a

c

e g

1

0

0 0.2 0.4 0.6 0.8 1

101

100

101

102

103

12 34 56 78 910

Absorbance

at 500 nm

1

0.1

Absorbance

Absorbance

1

0.8

0.6

0.4

0.2

0

11 12 13 14 15 16 17 18 19

Absorbance at 500 nm

0.01 0.1 1

Surfactant concentration (mM)

Mole fraction krytox

b d

400 450 500

500 600 600 650

700

550

Wavelength (nm) Wavelength (nm)

f h

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Figure 8 | Surfactant-mediated extraction of solutes. (ad) Fluorescence images of microuidic droplet production and resulting emulsions in the presence of 0.5% PFPEPEGPFPE surfactant and additionally low (a,bo1%) or high (c,d 30%) mass fractions of carboxylic acid uorosurfactants

(scale bar, 300 mm). The aqueous and uorous phase ow rate are 1 and 4 ml min 1, respectively. (eh) Macroscopic partitioning experiments.(e) Absorption spectra of the uorous phase after equilibration shown for various concentrations of the additive. The inset shows the absorption at 500 nm. (f) Corresponding images of the partitioning experiments. Aqueous solutions of rhodamine 6G (initial concentration: 100 mM) are exposed to the uorous phase with additive concentrations from 1 to 9 as: 0, 3.75, 7.51, 18.8, 37.5, 93.8, 188, 375 and 751 mM. Solution 10 contains only the

PFPEPEGPFPE surfactant at a concentration of 877 mM (0.5 wt%). (g) Absorption spectra of the uorous phase for various mole fractions of uorophore and surfactant. Inset: absorbance at 500 nm as a function of the mole fraction of the additive (Job plot). (h) Images of the corresponding partitioning experiments. In samples 1119, the fraction of the additive increases as: 0, 0.125, 0.25, 0.375, 0.5, 0.625, 0.75, 0.875, 1.0.

tremendous effect on the performance of the compartmentalization system. Hence, characterizing the amount of carboxylic acid remaining after synthesis is crucial. Various methods such as NMR and IR measurements may be applied. However, these methods may not be sufciently sensitive to trace impurities in a concentration range relevant for typical assays applied for biotechnological purposes (o100 mM). Phase partitioning of a uorescent indicator, such as rhodamine 6G, from a more protic solvent may be a sensitive method for the determination of the concentration of the residual carboxylic acid. The analysis of several surfactant batches by our partitioning method shows how the level of traces of side products in surfactant samples is determined (Supplementary Fig. 5 and Supplementary Notes 5 and 6). Finally, this chemical view of the interactions provides guidelines to design surfactants to guarantee efcient encapsulation or in contrast favour selective extraction of compounds from droplets.

DiscussionOur experiments show that microuidics provides tools to efciently manipulate droplets to prepare, order and store them in a controlled manner; emulsions with a precisely dened microstructure are obtained for quantitative studies of physico-chemical processes at the microscopic level. We show that the uorophore transport in uorinated emulsions, used as a model for organic molecule transport, is, in all our experiments, limited by the diffusive transport through the continuous phase. The dependence of the transport process on the droplet spacing is fully consistent with an analytical model based on the proper description of the permeability of the oil membrane separating the droplets. Increasing the spacing between droplets is an efcient strategy in reducing the exchange of material between droplets. In combination with a decrease of concentration of surfactant, we have shown a decrease in the rate of transport by a factor of about 30. In a bulk emulsion, the equivalent strategy would be to increase the continuous phase volume fraction which is technically challenging. The values of the diffusion coefcient of the uorophores in the continuous phase obtained

experimentally show that the transport of uorophores is mediated by large assemblies of surfactant molecules. Simple additives, such as sodium chloride, BSA33 or sugars29, not only affect the rate of transport but also the distribution of organic molecules among the droplets. We demonstrate how to use this concept for the targeted delivery of compounds, a potential new mechanism for actively feeding droplets from external sources. In practice, special care should be taken when changing buffer conditions in biochemical applications or using additives such as encoding uorophores which might affect interactions and partitioning. In contrast, understanding and controlling this process is essential to deliver molecules from one droplet to the next and might provide new tools for the chemical control of the content of emulsion droplets. Besides straightforward applications in droplet-based microuidic systems, we believe that our approach will be applicable to emulsion-based synthesis where transport of reagents between compartments is crucial. Our system might also provide additional insights to understand how organic molecules can be concentrated in a population of microcompartments, a question relevant for compartmentalization through phase separation in cells54, for prebiotic chemical systems55,56 and for the design of minimal functional micro-compartments55,57,58.

Methods

Chemicals. Resorun sodium salt, uorescein and rhodamine 6G (Sigma-Aldrich) solutions were prepared by dissolution in millipore water or NaCl solutions. Droplets were produced in uorinated oil (HFE-7500, 3M) and stabilized against coalescence by a peruoropolyetherpolyethyleneglycol block-copolymer surfactant (PFPEPEGPFPE, Critical Micellar Concentration B0.03% (weight fraction)33). The surfactant was a kind gift from Dr E. Mayot, prepared from the carboxylic acid Krytox (157-FSH, Dupont) and polyethyleneoxide (Sigma-Aldrich), adapting the synthesis scheme described by Holtze et al.59 and Scanga et al.33,60 The infra-red spectra of the surfactants used here, a detailed protocol for surfactant synthesis and the analysis of the rhodamine 6G partitioning towards the oil phase are provided in Supplementary Fig. 5 and Supplementary Notes 5 and 6. After mixing the surfactant in the uorinated oils, the solutions are stable and stored in closed glass asks and handled at room temperature.

Microuidic device fabrication. Chips were made of Norland Optical Adhesive 81 (NOA81). Briey, a positive SU-8 mould is made by two-layer photolithography to fabricate a negative PDMS stamp by replica moulding. A drop of NOA81 is

NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10392

deposited on the patterned PDMS substrate. For curing, the liquid lm is covered with a blank PDMS substrate and exposed to ultraviolet light (4 min, Polylux PT (Dreve)). For uidic connection, holes are punched into the crosslinked NOA81 layer (Harris Uni-Core Punch, 0.75 mm). Subsequently, the layer is stuck to a glass substrate, which was previously coated with NOA81 glue (500 r.p.m.) and then crosslinked (4 min, Polylux PT (Dreve)), improving bonding. Ports (Upchurch) are glued on top of the punched holes. A second layer of NOA81, made in the same way, is put in contact with the rst layer to provide overlying microuidic chambers that can be lled with water during the experiment. The completed chip is exposed to ultraviolet light for 40 min (Polylux PT (Dreve)). Aquapel (PPG Industries) was used to hydrophobize the channels: the aquapel solution is used as received and simply injected from a gastight syringe into the channels. The channels are then dried with nitrogen.

Device operation. A pressure-driven pump (Fluigent, MFCS-8C) was used to control the ow in the microuidic device. For droplet production, the two aqueous solutions were co-own with the uorinated oil containing a distinct concentration of surfactant (1, 2 or 5 wt%). The pressure levels of the aqueous phases were set to 280 mbar each, the oil phase was set to 260 mbar. For on-demand direction of droplets at the hydrodynamic switch, the pressure level in the wider microuidic channel was controlled. Up to typically 110 mbar, all droplets ow towards the wider microuidic channel. Above E180 mbar, the behaviour is inverted such that all droplets ow through the narrow microuidic channel towards the experimental zone of the chip. They buffer each other from one to the next local minimum in surface energy (Supplementary Fig. 1 and Supplementary Note 1). Lowering the switch pressure to 50 mbar leads to an immediate stop of delivery of droplets towards the experimental zone. However, the droplets are still buffering each other such that no droplets will be immobilized in between the local minima of surface energy. We immobilize droplets in ve independent one-dimensional microarrays, each with varying distances between the droplets (0, 3, 10, 15, 30 mm edgeedge distance). Each microarray consists of 13 to 15 droplets.

Fluorescence measurement and data processing. Images were taken every5 min with a digital camera (Canon, EOS D600). A light emitting diode (CoolLED pE-2, 550 nm) was used for the excitation of uorophores. The recorded intensity in the red channel was found to be proportional with the dye concentration in the relevant range (0100 mM). Images were analysed with homemade scripts using the standard toolbox in Matlab.

Flow-induced target delivery. We used two independent dyes (uorescein and resorun). Solutions of uorescein (alternatively resorun) are prepared as followed: a mixture of oil and surfactant (at 1%(w/w) concentration) is prepared and equilibrated for several days with a uorescein (alternatively resorun) solution at 100 mM in a closed vessel. At the end of the incubation, the oil is recovered and loaded in a syringe. Both solutions are injected in independent microchannels, feeding the microuidic chamber towards empty droplets (that is, droplets containing only buffer). The laminar ow prole of the oil ensures that single droplets are targeted. When the droplet buffer composition is similar to the composition of the solution used during the incubation, the equilibrium concentration of the uorescein (respectively, resorun) is the concentration used initially.

Measurement of partitioning coefcients. The partition coefcients between the aqueous and the uorinated phase are determined by the shake ask method. A total 500 ml of HFE-7500 with a distinct concentration of surfactant (0.55 wt%)

and the same amount of millipore water containing 100 mM of resorun sodium salt and in some cases sodium chloride (025 mg ml 1) are put into contact in a glass vial. Careful pipetting ensures that no emulsication is obtained. The mixture of the two liquids is incubated for at least 72 h. The partition coefcient is calculated by measuring the change in uorescence intensity of the aqueous phase with a microplate reader (Spectra Max Paradigm, Molecular Devices).

References

1. Bibette, J., Leal-Calderon, F. & Poulin, P. Emulsions: basic principles. Rep. Prog. Phys. 62, 6961033 (1999).

2. McClements, D. Nanoemulsions versus microemulsions: terminology, differences, and similarities. Soft Matter 8, 17191729 (2012).

3. Rosen, M. J. Surfactants and Interfacial Phenomena, i-xiii (John Wiley & Sons, Inc., 2004).

4. Tawk, D. S. & Grifths, A. D. Man-made cell-like compartments for molecular evolution. Nat. Biotechnol. 16, 652656 (1998).

5. Taly, V., Kelly, B. T. & Grifths, A. D. Droplets as microreactors for high-throughput biology. ChemBioChem 8, 263272 (2007).

6. Teh, S.-Y., Lin, R., Hung, L.-H. & Lee, A. P. Droplet microuidics. Lab Chip 8, 198220 (2008).

7. Seemann, R., Brinkmann, M., Pfohl, T. & Herminghaus, S. Droplet-based microuidics. Reports on Progress in Physics 75, 016601 (2012).

8. Guo, M. T., Rotem, A., Heyman, J. A. & Weitz, D. A. Droplet microuidics for high-throughput biological assays. Lab Chip 12, 21462155 (2012).

9. Theberge, A. B. et al. Microdroplets in microuidics: an evolving platform for discoveries in chemistry and biology. Angew. Chem. Int. Ed. 49, 58465868 (2010).

10. Baret, J.-C., Beck, Y., Billas-Massobrio, I., Moras, D. & Grifths, A. D. Quantitative cell-based reporter gene assays using droplet-based microuidics. Chem. Biol. 17, 528536 (2010).

11. Miller, O. J. et al. High-resolution dose-response screening using droplet-based microuidics. Proc. Natl Acad. Sci. USA 109, 378383 (2012).

12. Pekin, D. et al. Quantitative and sensitive detection of rare mutations using droplet-based microuidics. Lab Chip 11, 21562166 (2011).

13. Taly, V. et al. Multiplex picodroplet digital PCR to detect KRAS mutations in circulating DNA from the plasma of colorectal cancer patients. Clin. Chem. 59, 17221731 (2013).

14. Didelot, A. et al. Multiplex picoliter-droplet digital PCR for quantitative assessment of DNA integrity in clinical samples. Clin. Chem. 59, 815823 (2013).

15. Baret, J.-C. et al. Fluorescence-activated droplet sorting (FADS): efcient microuidic cell sorting based on enzymatic activity. Lab Chip 9, 18501858 (2009).

16. Debs, B. E., Utharala, R., Balyasnikova, I. V., Grifths, A. D. & Merten, C. A. Functional single-cell hybridoma screening using droplet-based microuidics. Proc. Natl Acad. Sci. USA 109, 1157011575 (2012).

17. Sjostrom, S. L. et al. High-throughput screening for industrial enzyme production hosts by droplet microuidics. Lab Chip 14, 806813 (2014).

18. Agresti, J. J. et al. Ultrahigh-throughput screening in drop-based microuidics for directed evolution. Proc. Natl Acad. Sci. USA 107, 40044009 (2010).

19. Wang, B. L. et al. Microuidic high-throughput culturing of single cells for selection based on extracellular metabolite production or consumption. Nat. Biotechnol. 32, 473478 (2014).

20. Krafft, M. P. Fluorocarbons and uorinated amphiphiles in drug delivery and biomedical research. Adv. Drug Deliv. Rev. 47, 209228 (2001).

21. Baret, J.-C. Surfactants in droplet-based microuidics. Lab Chip 12, 422433 (2012).

22. Scott, R. The solubility of uorocarbons. J. Am. Chem. Soc. 70, 40904093 (1948).

23. Simons, J. H. & Linevsky, M. J. The solubility of organic solids in uorocarbon derivatives. J. Am. Chem. Soc. 74, 47504751 (1952).

24. Kabalnov, A. Can micelles mediate a mass transfer between oil droplets? Langmuir 10, 680684 (1994).

25. Mazutis, L. et al. Multi-step microuidic droplet processing: kinetic analysis of an in vitro translated enzyme. Lab Chip 9, 29022908 (2009).

26. Courtois, F. et al. Controlling the retention of small molecules in emulsion microdroplets for use in cell-based assays. Anal. Chem. 81, 30083016 (2009).

27. Bai, Y. et al. A double droplet trap system for studying mass transport across a droplet-droplet interface. Lab Chip 10, 12811285 (2010).

28. Wu, N. et al. Management of the diffusion of 4-methylumbelliferone across phases in microdroplet-based systems for in vitro protein evolution. Electrophoresis 31, 31213128 (2010).

29. Sandoz, P. A., Chung, A. J., Weaver, W. M. & Carlo, D. D. Sugar additives improve signal delity for implementing two-phase resorun-based enzyme immunoassays. Langmuir 30, 66376643 (2014).

30. Thiam, A., Bremond, N. & Bibette, J. From stability to permeability of adhesive emulsion bilayers. Langmuir 28, 62916298 (2012).

31. Bremond, N. & Bibette, J. Exploring emulsion science with microuidics. Soft Matter 8, 1054910559 (2012).

32. Brosseau, Q., Vrignon, J. & Baret, J.-C. Microuidic dynamic interfacial tensiometry (mDIT). Soft Matter 10, 30663076 (2014).

33. Skhiri, Y. et al. Dynamics of molecular transport by surfactants in emulsions. Soft Matter 8, 1061810627 (2012).

34. Fradet, E. et al. Combining rails and anchors with laser forcing for selective manipulation within 2d droplet arrays. Lab Chip 11, 42284234 (2011).

35. Abate, A. R., Hung, T., Mary, P., Agresti, J. J. & Weitz, D. A. High-throughput injection with microuidics using picoinjectors. Proc. Natl Acad. Sci. USA 107, 1916319166 (2010).

36. Thorsen, T., Roberts, R. W., Arnold, F. H. & Quake, S. R. Dynamic pattern formation in a vesicle-generating microuidic device. Phys. Rev. Lett. 86, 41634166 (2001).

37. Woronoff, G. et al. New generation of amino coumarin methyl sulfonate-based uorogenic substrates for amidase assays in droplet-based microuidic applications. Anal. Chem. 83, 28522857 (2011).

38. Zheng, B., Tice, J. D. & Ismagilov, R. F. Formation of droplets of in microuidic channels alternating composition and applications to indexing of concentrations in droplet-based assays. Anal. Chem. 76, 49774982 (2004).

39. Hung, L. H. et al. Alternating droplet generation and controlled dynamic droplet fusion in microuidic device for Cds nanoparticle synthesis. Lab Chip 6, 174178 (2006).

8 NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10392 ARTICLE

40. Frenz, L., Blouwolff, J., Grifths, A. D. & Baret, J.-C. Microuidic production of droplet pairs. Langmuir 24, 1207312076 (2008).

41. Tang, S. K. Y. et al. A multi-color fast-switching microuidic droplet dye laser. Lab Chip 9, 27672771 (2009).

42. Abbyad, P., Dangla, R., Alexandrou, A. & Baroud, C. N. Rails and anchors: guiding and trapping droplet microreactors in two dimensions. Lab Chip 11, 813821 (2011).

43. Dangla, R., Lee, S. & Baroud, C. N. Trapping microuidic drops in wells of surface energy. Phys. Rev. Lett. 107, 124501 (2011).

44. Zwolinski, B. J., Eyring, H. & Reese, C. E. Diffusion and membrane permeability. J. Phys. Chem. 53, 14261453 (1949).

45. Chen, Y., Gani, A. & Tang, S. Characterization of sensitivity and specity in leaky droplet-based assays. Soft Matter 12, 50935103 (2012).

46. Poulsen, S., Harrington, R. & Drever, J. The solubility of toluene in aqueous salt solutions. Talanta 48, 633641 (1999).

47. Boitard, L. et al. Monitoring single-cell bioenergetics via the coarsening of emulsion droplets. Proc. Natl Acad. Sci. USA 109, 71817186 (2012).

48. Hofmann, T. W., Hanselmann, S., Janiesch, J.-W., Rademacher, A. & Bhm, C.H. J. Applying microdroplets as sensors for label-free detection of chemical reactions. Lab Chip 12, 916922 (2012).49. Mai, Y. & Eisenberg, A. Self assembly of block copolymers. Chem. Soc. Rev. 41,

59695985 (2012).

50. Vincent, J. M. Non-covalent associations in uorous uids. J. Fluorine Chem. 129, 903909 (2008).

51. ONeal, K. L. & Weber, S. G. Extraction of pyridines into uorous solvents based on hydrogen bond complex formation with carboyxlic acid receptors. Anal. Chem. 79, 31173215 (2007).

52. ONeal, K. L. & Weber, S. G. Molecular and ionic hydrogen bond formation in uorous solvents. J. Phys. Chem. B 113, 113149 (2009).

53. Palomo, C., Aizpurua, J. M., Loinaz, M. J., Fernandez-Berridi, M. J. & Irusta, L. Scavenging of uorinated N,N-dialkylureas by hydrogen binding: a novel separation method for uorous synthesis. Org. Lett. 3, 23612364 (2001).

54. Hyman, A. A. & Simons, K. Cell biology. beyond oil and water-phase transitions in cells. Science 337, 10471049 (2012).

55. Oparin, A. I. The origin of life and the origin of enzymes. Adv. Enzymol. Relat. Areas Mol. Biol. 27, 347380 (1965).

56. Fallah-Araghi, A. et al. Enhanced chemical synthesis at soft interfaces: A universal reaction-adsorption mechanism in microcompartments. Phys. Rev. Lett. 112, 028301 (2014).

57. Sokolova, E. et al. Enhanced transcription rates in membrane-free protocells formed by coacervation of cell lysate. Proc. Natl Acad. Sci. USA 110, 1169211697 (2013).

58. Spruijt, E., Sokolova, E. & Huck, W. T. S. Complexity of molecular crowding in cell-free enzymatic reaction networks. Nat. Nanotechnol. 9, 406407 (2014).

59. Holtze, C. et al. Biocompatible surfactants for water-in-uorocarbon emulsions. Lab Chip 8, 16321639 (2008).

60. Scanga, R. et al. Fluorinated amphiphilic block copolymers to stabilize water-in-uorocarbon emulsions. Polymer Preprints 50, 148149 (2009).

Acknowledgements

J.-C.B. acknowledges the nancial support by the SFB-755 Nanoscale Photonic Imaging, the ERC (FP7/2007-2013 /ERC Grant agreement 306385-SofI), the IdEx Bordeaux and the Rgion Aquitaine. B.R. acknowledges the IMPRS for Physics of Biology and Complex Systems for nancing her fellowship. A.J. and K.S. aknowledge the support by the IAESTE. P.G. and B.R. acknowledge additional support from the GGNB doctoral school. Fruitful discussions with Dr B. Hutchison, Dr E. Mayot and Dr V. Taly are warmly acknowledged as well as the experimental support from Dr E. Akoury andDr M. Prior and the technical support of the chemistry facility at the Max Planck Institute for Biophysical Chemistry (Goettingen, Germany) and at the Centre de Recherche Paul Pascal (Pessac, France).

Author contributions

P.G. performed the research, analysed the data and wrote the paper; B.R. performed the research, analysed the data and contributed to the writing. B.S., J.L. and B.R. contributed new reagents or analytical tools; A.J. and K.S. performed the experiments and contributed to the writing, J.-C.B. designed the research, analysed the data and wrote the paper.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Gruner, P. et al. Controlling molecular transport in minimal emulsions. Nat. Commun. 7:10392 doi: 10.1038/ncomms10392 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

NATURE COMMUNICATIONS | 7:10392 | DOI: 10.1038/ncomms10392 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9