9636
2009, 113, 9636–9639 Published on Web 07/01/2009
Imaging and Estimating the Surface Heterogeneity on a Droplet Containing Cosolvents Xiaohua Fang,† Bingquan Li,† Jun Wu,† Charles Maldarelli,§ Jonathan C. Sokolov,‡ Miriam H. Rafailovich,‡ and Ponisseril Somasundaran*,† Langmuir Center for Colloids and Interfaces, Department of Earth and EnVironmental Engineering, Columbia UniVersity, New York, New York 10027, Department of Materials Science and Engineering, Stony Brook UniVersity, Stony Brook, New York 11794, and Department of Chemical Engineering, City College of CUNY, New York, New York 10031 ReceiVed: May 7, 2009; ReVised Manuscript ReceiVed: June 17, 2009
Cosolvents have numerous applications in many industries as well as scientific research. The shortage in the knowledge of the structures in a cosolvent system is significant. In this work, we display the spatial as well as the kinetic distribution of the cosolvents using droplets as paradigms. When an alcohol/water-containing sessile droplet evaporates on a substrate, it phase segregates into a water-enriched core and a thin alcohol prevailing shell. This is considered to be due to the different escaping rate of solvents out of the liquid-vapor (l-v) interfaces. In between the core and shell phases, there exists a rough and solid-like liquid-liquid (l-l) wall interface as marked by the fluorescent polystyrene spheres and imaged by a confocal microscope. Holes and patches of beads are observed to form on this phase boundary. The water-dispersed beads prefer to partition within the core. The shell prevails in the droplet during most of the drying and shrinks with the l-v boundary. By monitoring the morphological progression of the droplet, the composition of the cosolvent at the liquid-vapor interface is obtained. Cosolvents (mostly alcohol/water) have numerous applications in pharmaceutical, cosmetic, as well as food industries. They are also widely adopted in scientific research such as organic synthesis and micro/nanofabrications.1-5 The general need for cosolvents is to disperse molecules that have poor solubility in another liquid2 and to form/modify functional structures.3 Upon evaporation, the distribution of liquids and the way they volatize significantly affects the movement of solutes and thus the drying pattern.4 The liquids which are in direct contact with the solutes during drying would also control their morphologies and functionalities.5 Compared with their wide applications in chemistry, the structures of the cosolvent system are barely studied.6-8 Empirically, the solution-dissolving power, the Hilderbrand solubility parameter δ, has been expressed as the linear addition of the contribution from each of the cosolvent components. This method assumes the cosolvents to be a uniform system, which is oversimplified and has failed to explain the many practical phenomena encountered during research, that is, the formation of porous materials,9 the complex conformation transformation in macromolecules,10,11 or the best yield of products at certain mixing ratios of cosolvents.12 In general, at least one of the liquids in the cosolvent system is volatile. The assumed experimental loading compositions might change considerably in even milliseconds due to evaporation, leading to poor experimental repeatability. To be able to manipulate the experimental conditions by tuning the cosolvent composition, * To whom correspondence should be addressed. † Columbia University. § City College of CUNY. ‡ Stony Brook University.
10.1021/jp904272a CCC: $40.75
a basic understanding of the cosolvent system is necessary. In this study, we focus on revealing the spatial as well as the kinetic distribution of the cosolvents using droplets as paradigms. The bulk solution behavior is expected to be similar to the evaporating droplet system. A technique is developed to monitor the in situ cosolvent compositions at the liquid-vapor interfaces. We observed the drying of water/ethanol and water/propanol droplets on octadecyltrichlorosilane (OTS, Aldrich) covered hydrophilic silicon surfaces.13 A CAM 200 optical contact angle meter (KSV instruments Ltd., Helsinki, Finland) was used to observe the droplet evaporation. The side views of the droplets were recorded upon loading. The experiment was conducted in ambient air with 40 ( 2% humidity. The droplet size was controlled to be similar (∼2 µL). Fluorescent polystyrene spheres (Duke Scientific) with a diameter of 1 µm were adopted to track the liquid distributions and distinguish between the cosolvents. The water-suspended beads were first mixed with water and then with the other cosolvents. The mixture was ultrasonicated for 24 h before the evaporation study. A Leica TCS SP2 confocal microscope (Leica Camera Group, Germany) was placed under the droplets during drying. Projections from three-dimensional confocal image reconstructions of beads in pure water and beads in water/alcohol mixtures droplet are shown in Figure 1. The stock bead solution is prepared in water. When pure water droplets dry, the beads are distributed uniformly over the entire volume of the droplet (Figure 1a). The loss of water at the l-v interface concentrates the fluorescent beads near the surface, forming a thin skin across the interface. The addition of alcohol, that is, ethanol or propanol, on the other hand, creates a rough 3D boundary for the beads and confines them into the core of the droplet (Figure 2009 American Chemical Society
Letters
J. Phys. Chem. B, Vol. 113, No. 29, 2009 9637
Figure 1. 3D reconstructions of 1 µm polystyrene beads distribution in droplets containing cosolvents; (a) a pure water droplet; (b) a droplet with original propanol/water loading ratio of 1:1; (c) side view of the droplet in (b); (d) side view of a droplet with original ethanol/water loading ratio of 1:1 showing the rough solid-like liquid-liquid interfacial boundary; (e,f) snapshots of a droplet with original ethanol/water loading ratio of 9:1 at (e) the initial stage and (f) the later stage of drying.
1b-d). Holes and patches of beads are observed to form on this phase boundary. The beads in the core area oscillate slowly near an equilibrated position, indicative of the high viscosity of the liquids in the core. The rough boundary denotes the interface between liquids of distinct properties. Liquids with low intermolecular forces tend to partition preferentially on the l-v side. The radii ratio of the core/shell is a function of the cosolvent type. The shell is thicker for propanol/H2O droplets than that for droplets with a similar composition of ethanol/ H2O. The initial mixing ratio also affects the thickness of the shell, although not as significantly. Upon evaporation, more beads appear in the shell area, and the l-l interface shrinks with the droplet. (Figure 1e,f). The l-v interface of the outer shell does not show a pure liquid behavior. Upon being loaded on OTS, droplets with a 1:1 ethanol/water ratio show a starting angle of 56°, as compared with the 26° for pure ethanol and 105° for pure water. Droplets with lower ethanol content show a higher initial contact angle and vice versa, indicating the coexistence of both liquids at the l-v front. This is related to the lower l-v and l-s (OTS) interfacial tension of ethanol. Pure water droplets dry initially at a constant contact base diameter mode with a steep angle recline from 105 to 89° (Figure 2a). They then adopted a constant contact angle mode at ∼90°, until the last stage of drying, where the liquid disappears drastically (Figure 2b). All alcohol-containing droplets have greater starting contact areas and smaller contact angles on OTS than those on pure water droplets. During their initial evaporation, the droplets increase their contact angles and shrink the contact bases at a rate higher than that for water. This indicates that the major loss of the cosolvent at this preliminary stage of drying could be the OTS-liking liquids, that is, ethanol. Similar to water droplets, a stable angle period is observed after the initial anglechanging process until the last stage of drying, the time span of which depends on the ethanol/water ratio. For droplets that are composed primarily of ethanol content, this constant contact angle mode was barely discernible. For a singlecomponent droplet, the drop volume decreasing rate per unit drop height, -(dV/dt)/h or λ, is related to only liquid
Figure 2. The (a) contact angle and (b) contact base diameter changes for droplets evaporating with a series of ethanol/water ratios.
properties. Therefore, λ remains a constant during drying. For droplets containing multiple vaporizing liquids, the escaping rate of the drop components out of the l-v interface would be nonidentical. As a result, it is expected that the liquid composition would not remain a constant with time, leading to a varying λ. The composition of the cosolvent at the liquid-vapor interface can be evaluated
9638
J. Phys. Chem. B, Vol. 113, No. 29, 2009
Letters
from the behavior of the droplet. The volume loss rate of the droplet scales with the drop height h, as (dV/dt) ) -(2πDM/FRT)(P* - P∞)h.14 Here, F, M, and D are the overall drop liquid density, molecular weight, and diffusion coefficient of the molecules in air, R is the universal gas constant, T is temperature, and P* and P∞ are the vapor pressures of the liquid on the drop surface and at infinity distance away from the drop, respectively. On a cosolvent droplet liquid-vapor interface containing xA fraction of liquid A and (1 - xA) fraction of liquid B, the coexistent liquids both contribute to the volume loss rate of the droplet. In general, it is assumed that at the l-v boundary, the liquids and vapors are nondistinguishable. Under such an assumption, ideal mixing conditions were considered for both noninteracting as well as liquids with strong interactions. The liquid loss rate can then be expressed by I ) IA + IB ) SA(t)JA + SB(t)JB ) xAS(t)JA + (1 - xA)S(t)JB, with J ) -D(∂C/∂r) and S ) 2πr2(1 - cos θ). Here, θ is the contact angle of the droplet with the substrate, J represents the evaporative flux, D is the diffusion coefficient of the liquid molecules in air, C is the liquid molecule’s concentration at the l-v interface, r is the drop the radius of curvature, and S(t), SA(t), and SB(t) are the surface areas taken by the droplet, component A, and componenet B, respectively. Combining the above equations leads to an expression of xA as
(
xA ) DBMBPB* +
)[
FBRT dV /h / DBMBPB* - DAMAPA* + 2π dt (FB - FA)RT dV /h (1) 2π dt
]
The Droplet’s shrinking liquid-vapor front contains 100% of the liquid on the liquid side. In this case, PA* and PB* equal to the saturated vapor pressure of the component drop liquid and can be estimated by the intermolecular forces δ of the liquid molecules
ln P* ) -
Vmδ2 + K1 - 1 RT
(2)
where Vm represents the molar volume of the liquid and K1 is a constant for all of the liquid.14,15 DA and DB are the diffusion coefficients of the molecules in air as calculated from the Fuller, Schettler, and Griddings (FSG) method.16 After ideal mixing, the liquid density of the drop F can be expressed by the linear summation of the components, as xAFA + (1 - xA)FB. Combining eqs 1 and 2 and inserting the solubility parameter δ measured using the same method for pure liquid components, we can measure the composition xA on a cosolvent droplet at the liquid-vapor interface by observing the droplet morphology as a function of time during evaporation, the result of which is shown in Figure 3a and b. The starting content of alcohol measured using this method scales with the initial loading composition of the samples. The discrepancy observed here is due to the fast evaporation of ethanol before the recording of the data at the beginning stages of drying. Upon drying, the l-v interfacial mixing ratio (alcohol/water) decreases with time. This is attributed to the more vigorous transportation of alcohol molecules through the l-v interface than water molecules. When the free alcohol molecules in the system escape totally, the mixture of the drops reaches a steady composition stage when
Figure 3. The calculated ethanol/propanol content for droplets evaporating with a series of (a) ethanol/water and (b) propanol/water ratios.
both of its components dry at similar rate, which is very slow. Stable assemblies between alcohol and water molecules are believed to be formed. Depending on the experimental conditions (i.e., OTS, 25 °C, 40% humidity), the drying time at this stage can be varied considerably. For pure liquids, it has been expected that xA will be a straight line that is parallel to the x-axis, which is obeyed well by the water curve shown in both graphs. However, from the shapes of the curves in Figure 3a and b, we detected a little impurity in both the ethanol and propanol used at later stages of drying. In summary, when an alcohol/water-containing sessile droplet evaporates on a substrate, it phase segregates into a water-enriched core and a thin alcohol prevailing shell. This is considered to be due to the different rates of escape of the solvents from the liquid-vapor (l-v) interfaces. In between the core and shell phases, there exists a rough and solid-like liquid-liquid (l-l) wall interface as marked by the fluorescent polystyrene spheres. The water-dispersed beads prefer to partition within the core. The shell prevails in the droplet during most of the drying time and shrinks with the l-v boundary. Outside of the l-l phase boundary is the liquid shell with lower solubility parameters, while most of the water remains in the droplet core. By monitoring the morphological progression of the droplet, the composition of the cosolvent at the liquid-vapor interface can be inferred. The method described here can be used to measure the in situ fraction of liquids at the evaporation front, which can improve our understanding as well as our ability to control many cosolvent-related physical chemical phenomena.
Letters References and Notes (1) Kim, D.; Jeong, S.; Moon, J.; Kang, K. Mol. Cryst. Liq. Cryst. 2006, 459, 45. (2) Sun, Y.; Bai, S.; Gu, L.; Tong, X. D.; Ichikawa, S.; Furusaki, S. Biochem. Eng. J. 1999, 3, 9. (3) Chennamsetty, N.; Bock, H.; Scanu, L. F.; Siperstein, F. R.; Gubbins, K. E. J. Chem. Phys. 2005, 122, 094710. (4) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Nature 1997, 389, 827. (5) Scharnagl, C.; Rief, M.; Friedrich, J. Biochim. Biophys. Acta 2005, 1749, 187. (6) Lee, S. Y.; Chalikian, T. V. J. Phys. Chem. B 2009, 113, 2443. (7) Froba, A. P.; Wasserscheid, P.; Gerhard, D.; Kremer, H.; Leipertz, A. J. Phys. Chem. B 2007, 111, 12817. (8) Smith, P. E.; Mazo, R. A. J. Phys. Chem. B 2008, 112, 7875. (9) Chen, H. M.; He, J. H.; Tang, H. M.; Yan, C. X. Chem. Mater. 2008, 20, 5894.
J. Phys. Chem. B, Vol. 113, No. 29, 2009 9639 (10) Fokkens, M.; Schrader, T.; Klarner, F. G. J. Am. Chem. Soc. 2005, 127, 14415. (11) Gardinier, W. E.; Baker, G. A.; Baker, S. N.; Bright, F. V. Macromolecules 2005, 38, 8574. (12) Gomez-Tagle, P.; Vargas-Zuniga, I.; Taran, O.; Yatsimirsky, A. K. J. Org. Chem. 2006, 71, 9713. (13) Zheng, J. W.; Zhu, Z. H.; Chen, H. F.; Liu, Z. F. Langmuir 2000, 16, 4409. (14) Fang, X.; Li, B.; Petersen, E.; Ji, Y.; Sokolov, J. C.; Rafailovich, H. M. J. Phys. Chem. B 2005, 109, 20554. (15) Fang, X.; Li, B.; Gewaily, D.; Sokolov, J. C.; Rafailovich, H. M. Appl. Phys. Lett. 2005, 87, 094103. (16) Yaws, C. L. Handbook of Transport Property Data; ButterworthHeinemann: Gulf, TX, 1999; Vol. 133, p 666.
JP904272A
