Interfacial Crystallization within Liquid Marbles

: We report interfacial crystallization in the droplets of saline solutions placed on superhydrophobic surfaces and liquid marbles filled with the saline. Evaporation of saline droplets deposited on superhydrophobic surface resulted in the formation of cup-shaped millimeter-scaled residues. The formation of the cup-like deposit is reasonably explained within the framework of the theory of the coffee-stain effect, namely, the rate of heterogeneous crystallization along the contact line of the droplet is significantly higher than in the droplet bulk. Crystallization within evaporated saline marbles coated with lycopodium particles depends strongly on the evaporation rate. Rapidly evaporated saline marbles yielded dented shells built of a mixture of colloidal particles and NaCl crystals. We relate the formation of these shells to the interfacial crystallization promoted by hydrophobic particles coating the marbles, accompanied with the upward convection flows supplying the saline to the particles, serving as the centers of interfacial crystallization. Convective flows prevail over the diffusion mass transport for the saline marbles heated from below.

Liquid marbles, introduced in the pioneering investigations of Quéré et al., are non-stick droplets coated with nano-or micron-sized particles, which are usually hydrophobic [11][12][13][14][15]. Liquid marbles are not hermetically coated by the solid particles. The respirability of liquid marbles makes them suitable for the cultivation of microorganisms and cells [16][17][18][19][20]. Liquid marbles may also be used as mini-reactors and bio-reactors [21,22]. Kinetics of the evaporation of liquid marbles was addressed in [23][24][25][26]. Use of liquid marbles for non-traditional computing was reported recently [27][28][29][30]. Liquid marbles may be actuated by UV and IR light and Marangoni flows [31][32][33], and demonstrate obvious potential for micro-fluidics applications [34,35]. Actually, liquid marbles opened new horizons for the researchers studying the liquid state of the matter. We report here the use of liquid marbles for controlled interfacial crystallization.

Results
We carried out comparative experiments in which droplets of aqueous NaCl solutions we aporated under conditions of superhydrophobicity when (i) naked aqueous NaCl solution drople ere evaporated when placed on superhydrophobic surfaces, and when (ii) marbles filled wi ueous NaCl solution and coated with lycopodium powder (a strongly hydrophobic yellow-ta st-like powder, consisting of the dry spores of clubmoss plants) were evaporated when placed o perhydrophobic substrates, as depicted in Figure 1. When naked saline droplets are evaporated on superhydrophobic surfaces, very differe enarios of evaporation are possible, namely, the evaporation may take place under the pinned trip e, and under the sliding triple line, as discussed in [36]. The evaporation may be accompanied b e Cassie-Wenzel wetting transition and may take place within the Cassie wetting regime [36,37 e studied evaporation of the naked saline droplets under ambient conditions (t = 25 °C) and als der the constant temperature of the substrate of t = 70 °C. The characteristic time scales aporation were ≅ 24 for the "slow" and ≅ 4 for the rapid evaporation processes Both of temperature regimes supplied very similar results, namely, evaporation occurred und e slight variation in the apparent contact angle, i.e., 142° 149°, as shown in Figure   hereas the contact line was strongly pinned as demonstrated in Movie S1 and Movie S2. The lo ntact angle hysteresis and low sliding angles inherent for the pristine saline droplets demonstrate at the initial wetting regime was the Cassie-like one. The wetting regime at which high appare ntact angles are accompanied with the pronounced pinning of the triple line is inherent for the s lled "rose petal effect" [37][38][39][40]. However, in our experiments, the situation was much mo mplicated due to the crystallization of NaCl occurring in a course of evaporation of the droplet When naked saline droplets are evaporated on superhydrophobic surfaces, very different scenarios of evaporation are possible, namely, the evaporation may take place under the pinned triple line, and under the sliding triple line, as discussed in [36]. The evaporation may be accompanied by the Cassie-Wenzel wetting transition and may take place within the Cassie wetting regime [36,37]. We studied evaporation of the naked saline droplets under ambient conditions (t = 25 • C) and also under the constant temperature of the substrate of t = 70 • C. The characteristic time scales of evaporation were τ ev 24 min for the "slow" and τ ev 4 min for the rapid evaporation processes.
Both of temperature regimes supplied very similar results, namely, evaporation occurred under the slight variation in the apparent contact angle, i.e., 142 • < θ app < 149 • , as shown in Figure 2, whereas the contact line was strongly pinned as demonstrated in Movies S1 and S2. The low contact angle hysteresis and low sliding angles inherent for the pristine saline droplets demonstrated that the initial wetting regime was the Cassie-like one. The wetting regime at which high apparent contact angles are accompanied with the pronounced pinning of the triple line is inherent for the so-called "rose petal effect" [37][38][39][40]. However, in our experiments, the situation was much more complicated due to the crystallization of NaCl occurring in a course of evaporation of the droplets, and it is reasonable to suggest that high pinning of the triple line arises from the high affinity of water to NaCl crystals, formed in the vicinity of the triple line, as will be demonstrated in detail below.

Discussion
Evaporation of saline droplets placed on a superhydrophobic surface resulted in the formation of the cup-shaped millimeter-scaled residues, as shown in Figures 1a and 3a and Movies S1 and S2, and was observed for both slow and rapid evaporation regimes. It seems that the formation of this cup-like deposit may be explained within the framework of the theory of the coffee-stain effect developed in [41,42]. The authors of [41,42] related the formation of coffee-stain deposits to a couple of main physical reasons, namely: contact line pinning and intensive evaporation from the edge of the drop. When the apparent contact angle is obtuse, we have one more geometric factor strengthening the evaporation rate in the vicinity of the triple line, which is shown schematically in Figure 4. Consider two molecules evaporated from the sites labeled "I" and "II" depicted in Figure 4. A molecule evaporated from the site labeled "I" has a chance to be absorbed by a droplet, whereas a molecule evaporated from site "II" has a chance to be absorbed by the superhydrophobic substrate. This factor strengthens the evaporation rate in the vicinity of the triple line. A pinned contact line, in turn, induces an outward, radial fluid flow when there is evaporation at the edge of the drop; this flow replenishes the liquid that is removed from the edge and gives rise to the formation of the coffeestain cup-like deposits shown Figures 1a and 3a and discussed in detail in [41,42]. Crystallization of NaCl at the circumference of the contact area, in turn, strengthens the effect of pinning. Thus, crystallization works as a positive feedback, enhancing the coffee-stain effect and providing the stability of the apparent contact angle in a course of evaporation, illustrated in Figure 1a. Our observations support recent experimental findings reported in [43], in which the crystallization of CaCl2 salt in a droplet was investigated, and it was revealed that the crystallization rate along the

Discussion
Evaporation of saline droplets placed on a superhydrophobic surface resulted in the formation of the cup-shaped millimeter-scaled residues, as shown in Figures 1a and 3a and Movies S1 and S2, and was observed for both slow and rapid evaporation regimes. It seems that the formation of this cup-like deposit may be explained within the framework of the theory of the coffee-stain effect developed in [41,42]. The authors of [41,42] related the formation of coffee-stain deposits to a couple of main physical reasons, namely: contact line pinning and intensive evaporation from the edge of the drop. When the apparent contact angle is obtuse, we have one more geometric factor strengthening the evaporation rate in the vicinity of the triple line, which is shown schematically in Figure 4. Consider two molecules evaporated from the sites labeled "I" and "II" depicted in Figure 4. A molecule evaporated from the site labeled "I" has a chance to be absorbed by a droplet, whereas a molecule evaporated from site "II" has a chance to be absorbed by the superhydrophobic substrate. This factor strengthens the evaporation rate in the vicinity of the triple line. A pinned contact line, in turn, induces an outward, radial fluid flow when there is evaporation at the edge of the drop; this flow replenishes the liquid that is removed from the edge and gives rise to the formation of the coffee-stain cup-like deposits shown Figures 1a and 3a and discussed in detail in [41,42]. Crystallization of NaCl at the circumference of the contact area, in turn, strengthens the effect of pinning. Thus, crystallization works as a positive feedback, enhancing the coffee-stain effect and providing the stability of the apparent contact angle in a course of evaporation, illustrated in Figure 1a. Our observations support recent experimental findings reported in [43], in which the crystallization of CaCl 2 salt in a droplet was investigated, and it was revealed that the crystallization rate along the contact line of the droplet is many times higher than in the direction of the droplet radius. Obviously, the heterogeneous crystallization prevails on the homogeneous bulk nucleation.

Condens. Matter 2018, 3, x FOR PEER REVIEW 4 of 11
contact line of the droplet is many times higher than in the direction of the droplet radius. Obviously, the heterogeneous crystallization prevails on the homogeneous bulk nucleation.  . Scheme illustrating the formation of the coffee-stain deposits under evaporation of a naked saline droplet evaporated on superhydrophobic surfaces. A molecule evaporated from the state labeled "I" has a chance to be absorbed by a droplet, whereas a molecule evaporated from site "II" has a chance to be absorbed by the superhydrophobic substrate. Red arrows illustrate the internal flows resulting in the formation of the NaCl residue. Now, consider the evaporation of liquid marbles filled with the NaCl solutions. In contrast to the naked saline droplets, the apparent contact angle of the marbles decreased gradually in a course of evaporation from ca. 150° to ca. 115°, as shown in Figure 5 and Movie S3. It should be mentioned contact line of the droplet is many times higher than in the direction of the droplet radius. Obviously, the heterogeneous crystallization prevails on the homogeneous bulk nucleation.  . Scheme illustrating the formation of the coffee-stain deposits under evaporation of a naked saline droplet evaporated on superhydrophobic surfaces. A molecule evaporated from the state labeled "I" has a chance to be absorbed by a droplet, whereas a molecule evaporated from site "II" has a chance to be absorbed by the superhydrophobic substrate. Red arrows illustrate the internal flows resulting in the formation of the NaCl residue. Now, consider the evaporation of liquid marbles filled with the NaCl solutions. In contrast to the naked saline droplets, the apparent contact angle of the marbles decreased gradually in a course of evaporation from ca. 150° to ca. 115°, as shown in Figure 5 and Movie S3. It should be mentioned . Scheme illustrating the formation of the coffee-stain deposits under evaporation of a naked saline droplet evaporated on superhydrophobic surfaces. A molecule evaporated from the state labeled "I" has a chance to be absorbed by a droplet, whereas a molecule evaporated from site "II" has a chance to be absorbed by the superhydrophobic substrate. Red arrows illustrate the internal flows resulting in the formation of the NaCl residue. Now, consider the evaporation of liquid marbles filled with the NaCl solutions. In contrast to the naked saline droplets, the apparent contact angle of the marbles decreased gradually in a course of evaporation from ca. 150 • to ca. 115 • , as shown in Figure 5 and Movie S3. It should be mentioned that the apparent contact angle of a liquid marble is not a "true interfacial angle" due to the fact that it is formed by the micro-rough surface of a marble coated with colloidal particles. that the apparent contact angle of a liquid marble is not a "true interfacial angle" due to the fact that it is formed by the micro-rough surface of a marble coated with colloidal particles. We establish that the eventual shape of the solid residue, built of the mixture of colloidal particles and NaCl crystals, depends strongly on the temperature of evaporation and the concentration of the solution. When the marble was slowly evaporated under ambient conditions, (the characteristic time of the process was ≅ 35 min) the process yielded a single relatively large (ca. 1 mm) NaCl crystal that formed at the bottom of the evaporated saline marble. The situation was rather different for rapidly evaporated liquid marbles. When evaporation was conducted under high temperatures (t = 70 °C), the characteristic time of the process was ≅ 5 min and two different scenarios of crystallization were observed. When the NaCl solutions were unsaturated (14.9% w/w and 8.1% w/w), the crystallization took place mainly at the base of the marble, yielding the flat residues shown in Figure 6. The most interesting and intriguing result was obtained under the rapid ( = 70°C, ≅ 5 min) evaporation of saturated NaCl solutions. In this case, small crystals of NaCl (the characteristic size of crystals was ca. 10-50 µm) grew uniformly over the entire marble shell. In other words, interface crystallization stimulated by the particles coating a marble took place [1][2][3][4][5][6]. The NaCl crystals merged with the colloidal particles, thus forming the uniform hemisphere-like dented shells depicted in Figures 7b and 8d. The SEM images of the shells, representing the mixture of NaCl crystals and colloidal particles, are depicted in the inset (c) of Figure  7. The residues depicted in Figures 7b and 8d are markedly stronger than the residues emerging at room temperature. It should be emphasized that in this case, the interfacial crystallization promoted by the hydrophobic particles coating the marble was observed. The mass transport promoting the interfacial crystallization under slow and rapid evaporation is depicted schematically in Figure 8a,c. We establish that the eventual shape of the solid residue, built of the mixture of colloidal particles and NaCl crystals, depends strongly on the temperature of evaporation and the concentration of the solution. When the marble was slowly evaporated under ambient conditions, (the characteristic time of the process was τ ev 35 min) the process yielded a single relatively large (ca. 1 mm) NaCl crystal that formed at the bottom of the evaporated saline marble.
The situation was rather different for rapidly evaporated liquid marbles. When evaporation was conducted under high temperatures (t = 70 • C), the characteristic time of the process was τ ev 5 min and two different scenarios of crystallization were observed. When the NaCl solutions were unsaturated (14.9% w/w and 8.1% w/w), the crystallization took place mainly at the base of the marble, yielding the flat residues shown in Figure 6. The most interesting and intriguing result was obtained under the rapid (t = 70 • C, τ ev 5 min) evaporation of saturated NaCl solutions. In this case, small crystals of NaCl (the characteristic size of crystals was ca. 10-50 µm) grew uniformly over the entire marble shell. In other words, interface crystallization stimulated by the particles coating a marble took place [1][2][3][4][5][6]. The NaCl crystals merged with the colloidal particles, thus forming the uniform hemisphere-like dented shells depicted in Figures 7b and 8d. The SEM images of the shells, representing the mixture of NaCl crystals and colloidal particles, are depicted in the inset (c) of Figure 7. The residues depicted in Figures 7b and 8d are markedly stronger than the residues emerging at room temperature. It should be emphasized that in this case, the interfacial crystallization promoted by the hydrophobic particles coating the marble was observed. The mass transport promoting the interfacial crystallization under slow and rapid evaporation is depicted schematically in Figure 8a,c.   The upward flow, supplying NaCl to the hydrophobic particles coating the marble, is reasonably attributed to the thermal convection, inevitable when a marble is placed on a hot plate. Consider that   The upward flow, supplying NaCl to the hydrophobic particles coating the marble, is reasonably attributed to the thermal convection, inevitable when a marble is placed on a hot plate. Consider that the change in temperature of the saline in the course of evaporation was registered with the thermal camera as ca. Δ ≅ 25°. In order to attain a qualitative understanding of the process occurring within   The upward flow, supplying NaCl to the hydrophobic particles coating the marble, is reasonably attributed to the thermal convection, inevitable when a marble is placed on a hot plate. Consider that the change in temperature of the saline in the course of evaporation was registered with the thermal camera as ca. Δ ≅ 25°. In order to attain a qualitative understanding of the process occurring within The upward flow, supplying NaCl to the hydrophobic particles coating the marble, is reasonably attributed to the thermal convection, inevitable when a marble is placed on a hot plate. Consider that the change in temperature of the saline in the course of evaporation was registered with the thermal camera as ca. ∆t 25 • . In order to attain a qualitative understanding of the process occurring within the marbles under their evaporation, let us understand the hierarchy of time scales involved in the process. The characteristic time scale of evaporation is τ ev 5 min. The characteristic time scale of the diffusion mass transport within the marbles/saline droplets is supplied by Equation (1): where R is the radius of a marble, and D is the diffusion coefficient of NaCl in water; substitution of R 1.0 × 10 −3 m and D 1.5 × 10 −9 m 2 s (see [44]) yields the estimation τ di f f 2.7 × 10 3 s 45.0 min. The characteristic time scale of thermal equilibration in the marbles heated from below τ eq may be estimated as: where α 0.14 × 10 −6 m 2 s is the thermal diffusivity of saline (see [45]); calculation with Equation (2) yields the estimation τ therm 0.5 min. Thus, the hierarchy of time scales supplied by Equation (3) takes place: We recognize from Equation (3) that the diffusion mass transport is a slow process on the time scales of evaporation and thermal equilibration, occurring in the saline marbles subject to heating. Therefore, it is reasonable to attribute the dramatic influence exerted by temperature on the morphology of the eventual residue to the upward convective flows (shown schematically in Figure 8c), which supply saline to the hydrophobic particles, coating the marble and promoting the interfacial crystallization, and yielding entire shells shown in Figure 7b. Contrastingly, slow evaporation of the liquid marbles yielded flat deposits, as shown in Figure 8b, containing the large (~1 mm) NaCl crystal formed at the base of the marble. This conclusion is supported by the direct calculation of the Peclet number supplied by Equation (4): where u is the characteristic velocity of the convective flows (see Figure 8c). Direct visualization of the convective flows taking place within the heated marbles under t = 70 • C supplied the value of the typical velocity of convective flows estimated as u 0.3 mm s (see Movie S2). Substitution of the aforementioned values of the physical parameters yields the following estimation: Pe 4 × 10 2 1. This means that within the regime of rapid evaporation of the saline marbles, the bulk convective flows prevail over the diffusion mass transport, thus justifying the suggested mechanism of shells' formation.
Our observations are in accordance with those reported in [46], in which crystallization during evaporation of saturated Na 2 SO 4 and NaCl droplets was studied. The authors of [46] reported that on a hydrophobic surface, contrary to what happens for the hydrophilic one, the NaCl crystallization appears not at the liquid-air interface but rather at the solid-liquid interface. In our experiments, this was the case for both naked NaCl droplets evaporated on superhydrophobic surfaces and also for NaCl liquid marbles. Let us quote [46]: 'Sodium chloride crystals are found to form preferentially in contact with a nonpolar area (air or hydrophobic solid). This is an important observation, as hydrophobic treatments have been used frequently in the past to prevent crystallization and damage on buildings. Our observations suggest that this may be counterproductive due to the tendency of NaCl to crystallize on nonpolar surfaces'. Our results support this conclusion; indeed, the sodium chloride crystals prefer to grow on hydrophobic surfaces in contact with NaCl solutions, filling naked droplets placed on superhydrophobic surfaces and saline marbles coated with hydrophobic particles.
The dented shape of the shell emerging from the evaporation of saline marbles, as shown in Figure 8c,d, is noteworthy. This shape resembles the shape of Leidenfrost droplets deposited on hot plates, as discussed in [47]. It is reasonable to attribute the formation of the dent to a gas pocket, formed by the evaporation of water when a marble is heated from below.

Conclusions
We conclude that the saline droplets placed on superhydrophobic surfaces and liquid marbles filled with saline demonstrate interfacial crystallization over the course of water evaporation. Evaporation of naked saline droplets resulted in the formation of cup-shaped millimeter-scaled residues. The formation of the cup-like deposit is reasonably explained within the framework of the theory of the coffee-stain effect [41,42]. Evaporated saline marbles heated from below gave rise to dented shells built of a mixture of colloidal particles and NaCl crystals. Qualitative analysis of the processes taking place under the evaporation of saline marbles supports the assumption that the diffusion mass transport is a slow process on the time scales of evaporation and thermal equilibration of heated marbles. Formation of the entire shells built of the lycopodium and NaCl crystals under rapid evaporation of the heated saline marbles takes place under the high values of the Peclet numbers [48]. We relate the formation of these shells to the interfacial crystallization promoted by the hydrophobic particles coating the marbles, accompanied by the upward convection flows supplying the saline to the hydrophobic particles.
Liquid marbles containing aqueous solution of NaCl were coated by lycopodium powder. Then, 5-20 µL drops containing three different concentrations of aqueous NaCl solution (25.9% w/w; 14.9% w/w and 8.1% w/w) were used to prepare the liquid marbles. Saline droplets were deposited using a precise micro-syringe on the superhydrophobic surface. The superhydrophobic surface was covered with a layer of aforementioned powders. Rolling of droplets yielded the formation of marbles enwrapped with the specified hydrophobic powders and containing the solution.
Evaporation of liquid marbles and naked saline droplets was performed on a superhydrophobic surface. The evaporation process, carried out on the superhydrophobic surface, was studied at ambient conditions (t = 25 • C) and 70 • C, where 70 • C was the temperature of the hot plate. The relative humidity was RH = 44%.
A Nikon 1 v3 camera, digital microscope BW1008-500X and filed emission scanning electron microscope (MAIA3 TESCAN) were used to capture images and movies of the marbles during evaporation. A Rame-Hart 500 goniometer was used to capture the side view of the liquid marbles during evaporation. Temperature distribution was studied using a Therm-App Pro TAS19AV-M25A-HZ LWIR 7.5-14 µm thermal camera. The resolution, accuracy, sensitivity and the frame rate of the camera were 640 × 480 pixels (>300,000 pixels), NEDT (noise equivalent differential temperature) < 0.03 • C and 25 frames/s, respectively.
Supplementary Materials: The following are available online at http://www.mdpi.com/2410-3896/5/4/62/s1, Movie S1 represents evaporation of a 5 µL 25.9% w/w aqueous NaCl drop placed on the superhydrophobic surface (side view) (speed 64X). The temperature of the superhydrophobic surface is t = 25 • C; Movie S2 represents evaporation of a 5 µL 25.9% w/w aqueous NaCl drop placed on a superhydrophobic surface (top view) (speed 4X). The temperature of the superhydrophobic surface is t = 70 • C; Movie S3 represents evaporation of a 10 µL liquid marble made by lycopodium powder and filled with 25.9% w/w aqueous NaCl solution (side view). The temperature of the superhydrophobic surface is t = 70 • C.
Author Contributions: S.S. and E.B. proposed the concept. P.K.R. and I.L. designed and performed the experiments. E.B. wrote the paper. The results were analyzed by S.S., E.B., I.L. and P.K.R. Authorship must be limited to those who have contributed substantially to the work reported. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.