Microflow and Crack Formation Patterns in Drying ... - ACS Publications

Jul 10, 2008 - Massachusetts General Hospital, HarVard Medical School, and Shriners Hospital for Children,. Boston, Massachusetts 02114. ReceiVed ...
0 downloads 0 Views 3MB Size
7688

Langmuir 2008, 24, 7688-7697

Microflow and Crack Formation Patterns in Drying Sessile Droplets of Liposomes Suspended in Trehalose Solutions Dana R. Adams,† Mehmet Toner,*,‡ and Robert Langer† Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, and Center for Engineering in Medicine and Surgical SerVices, Massachusetts General Hospital, HarVard Medical School, and Shriners Hospital for Children, Boston, Massachusetts 02114 ReceiVed December 7, 2007. ReVised Manuscript ReceiVed February 21, 2008 Anhydrobiotic preservation potentially provides a means of long-term storage of mammalian cells in carbohydrate glasses under ambient conditions. During desiccation, sessile droplets of glass-forming carbohydrate solutions exhibit complex phenomena, including fluid flow, droplet deformation, and crack formation, all of which may alter the cell preservation efficacy. Cell-sized liposomes were employed as a model system to explore these phenomena in diffusively dried sessile droplets of trehalose solutions. Two factors were identified that strongly influenced the features of the desiccated droplets: the underlying surface and the liposomes themselves. In particular, the surface altered the droplet shape as well as the microflow pattern and, in turn, the moisture conditions encountered by the liposomes during desiccation. A ring deposit formed when the droplets were dried on polystyrene, as would be expected owing to the capillary flow that generally occurs in pinned droplets. In contrast, when dried on the more hydrophilic glass slide, the resulting droplets were thinner, and the liposomes accumulated near their centers, which was an unexpected result likely owing to the glass-forming nature of trehalose solutions. As might be anticipated given the variations in liposome distribution, the choice of surface also influenced crack formation upon continued drying. In addition to providing a preferential path for drying, such cracks are relevant because they could inflict mechanical damage on cells. The liposomes themselves had an even more profound effect on crack formation; indeed, whereas cracks were found in all droplets containing liposomes, in their absence few of the droplets cracked at all, regardless of the surface type. These complex drying dynamics merit further investigation in the development of anhydrobiotic preservation protocols, particularly with regard to the role therein of surface hydrophobicity and the cells themselves.

1. Introduction Anhydrobiotic preservation, inspired by the variegated organisms that naturally survive desiccation in a state of “suspended animation,”1–3 offers the hope of a means of long-term storage of mammalian cells under ambient conditions via nontoxic4 glassforming carbohydrate preservatives such as trehalose. Biopreservation has many applications in tissue engineering, stem cell research, and cell transplantation; anhydrobiotic preservation would be especially advantageous in situations where cryopreservation is particularly challenging or prohibitively expensive, such as during transport, in developing countries or on the battlefield. Long-term anhydrobiotic preservation requires the transition of the carbohydrate solution to the glassy state upon drying. This vitrification halts or slows molecular mobility,5 thus conferring biomolecular stability and inhibiting metabolic reactions, although carbohydrates may play other roles in biomolecular stabilization * Corresponding authors. E-mail: [email protected]. † Massachusetts Institute of Technology. ‡ Massachusetts General Hospital, Harvard Medical School, and Shriners Hospital for Children. (1) Alpert, P.; Oliver, M. J. Drying Without Dying In Desiccation and SurViVal in Plants: Drying Without Dying; Black, M., Pritchard, H. W., Eds.; CAB International: Wallingford, U.K., 2002; pp 3–43. (2) Crowe, J. H.; Crowe, L. M. Nat. Biotechnol. 2000, 18, 145–146. (3) Hoekstra, F. A. Pollen and Spores: Desiccation Tolerance in Pollen and the Spores of Lower Plants and Fungi In Desiccation and SurViVal in Plants: Drying Without Dying; Black, M., Pritchard, H. W., Eds.; CAB International: Wallingford, U.K., 2002; pp 185–205. (4) Guo, N.; Puhlev, I.; Brown, D. R.; Mansbridge, J.; Levine, F. Nat. Biotechnol. 2000, 18, 168–171. (5) Green, J. L.; Angell, C. A. J. Phys. Chem. 1989, 93, 2880–2882. (6) Tsvetkova, N. M.; Phillips, B. L.; Crowe, L. M.; Crowe, J. H.; Risbud, S. H. Biophys. J. 1998, 75, 2947–2955.

such as forming hydrogen bonds to “replace” the evaporated water.6 Thus, trehalose is an effective preservative owing, at least in part, to its higher glass-transition temperature (113 °C for trehalose dihydrate),7 compared to that of many other sugars.5 Although promising results have been achieved with mammalian cells via freeze-drying,8 convective drying,9,10 and diffusive drying,9 much work remains before the goal of longterm anhydrobiotic storage is achieved. Diffusive drying is of particular interest because this method most closely mimics the conditions experienced by anhydrobiotic organisms in nature and is also convenient and economical to employ: sessile droplets of trehalose-containing cell suspensions may be simply air-dried at low humidity. Surprisingly complex drying kinetics arise even in the isothermal diffusive drying of trehalose solution droplets without cells.7 Beginning with a trehalose concentration comparable to that in naturally anhydrobiotic organisms (e.g., 200 mM11) considerable desiccation (to a moisture content of approximately 9 wt %) is required for vitrification at room temperature7 because the solution glass-transition temperature decreases with the trehalose concentration. The trehalose concentration is typically highly heterogeneous in such a desiccating solution, with a thin glassy skin forming first at the air–liquid interface.12 Recent experiments for trehalose solutions drying in micro(7) Aksan, A.; Toner, M. Langmuir 2004, 20, 5521–5529. (8) Goodrich, R. P.; Sowemimo-Coker, S. O.; Zerez, C. R.; Tanaka, K. R. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 967–971. (9) Chen, T.; Acker, J. P.; Eroglu, A.; Cheley, S.; Bayley, H.; Fowler, A.; Toner, M. Cryobiology 2001, 43, 168–181. (10) McGinnis, L. K.; Zhu, L.; Lawitts, J. A.; Bhowmick, S.; Toner, M.; Biggers, J. D. Biol. Reprod. 2005, 73, 627–633. (11) Tunnacliffe, A.; de Castro, A. G.; Manzanera, M. Cryobiology 2001, 43, 124–132. (12) Aksan, A.; Irimia, D.; He, X.; Toner, M. J. Appl. Phys. 2006, 99, 064703. (13) Pauchard, L.; Allain, C. Europhys. Lett. 2003, 62, 897–903.

10.1021/la703835w CCC: $40.75  2008 American Chemical Society Published on Web 07/10/2008

Microflow and Crack Formation Patterns

channels suggest that, unlike the situation hypothesized for polymeric carbohydrates,13 the formation of a thin glassy skin significantly impedes further evaporation,12 delaying the transition of cells to a state of suspended animation because they would likely be found in the underlying solution. Once the skin has formed, further drying of glass-forming carbohydrate solutions may eventually lead to other complexities, such as buckling of the skin14,15 and crack formation.7 The inhomogeneous moisture conditions potentially experienced by cells within a drying droplet of trehalose solution are further complicated by microflow fields, which may alter the distribution of cells within the droplet. Although this phenomenon has been largely ignored in the anhydrobiotic cell preservation literature, the flow of dispersed particles in drying sessile droplets has been investigated for diverse applications including DNA stretching,16–18 nanoparticle patterning,19,20 and ink-jet printing.21–23 Diffusive drying of sessile droplets generally gives rise to capillary flow resulting in particulate deposition near the contact line if it is pinned,24 with the width of the deposited ring depending on parameters such as the particle volume fraction19,21 and size;20,25 however, little is known of the behavior of particulates in desiccating glass-forming solutions. The purpose of this study is thus to investigate these complex drying phenomena as they relate to anhydrobiotic preservation, employing suspensions of giant liposomes in a trehalose solution as a model cell system. Droplets of the solution were air dried on two surface typesscell culture polystyrene and glasssand the effects of the substrate type on liposome flow, droplet deformation, and crack propagation were assessed. Drying on polystyrene produced a ring deposit of liposomes near the contact line, as would be expected from capillary flow, whereas drying on glass resulted in the accumulation of liposomes at the droplet center, although the droplet shape deformation was similar for both surface types, resulting in a depression at the center. As drying progressed, a netlike hierarchical crack pattern propagated from the edge of the droplets dried on polystyrene and from both the edge and the center for droplets dried on glass. An increase in liposome concentration produced a branchlike pattern for droplets dried on polystyrene, whereas trehalose solution droplets without liposomes rarely exhibited cracks at all.

2. Materials and Methods Giant vesicles were prepared via the film-swelling method,26–28 as detailed in previous work.29 Briefly, a stock solution of 1,2(14) Gorand, Y.; Pauchard, L.; Calligari, G.; Hulin, J. P.; Allain, C. Langmuir 2004, 20, 5138–5140. (15) Pauchard, L.; Allain, C. Phys. ReV. E 2003, 68, 052801. (16) Chopra, M.; Li, L.; Hu, H.; Burns, M. A.; Larson, R. G. J. Rheol. 2003, 47, 1111–1132. (17) Smalyukh, I. I.; Zribi, O. V.; Butler, J. C.; Lavrentovich, O. D.; Wong, G. C. L. Phys. ReV. Lett. 2006, 96, 177801. (18) Jing, J.; Reed, J.; Huang, J.; Hu, X.; Clarke, V.; Edington, J.; Housman, D.; Anantharaman, T. S.; Huff, E. J.; Mishra, B.; Porter, B.; Shenker, A.; Wolfson, E.; Hiort, C.; Kantor, R.; Aston, C.; Schwartz, D. C. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 8046–8051. (19) Maenosono, S.; Dushkin, C. D.; Saita, S.; Yamaguchi, Y. Langmuir 1999, 15, 957–965. (20) Chon, C. H.; Paik, S.; Tipton, J. B., Jr.; Kihm, K. D. Langmuir 2007, 23, 2953–2960. (21) Park, J.; Moon, J. Langmuir 2006, 22, 3506–3513. (22) Kajiya, T.; Nishitani, E.; Yamaue, T.; Doi, M. Phys. ReV. E 2006, 73, 011601. (23) Wang, J.; Evans, J. R. G. Phys. ReV. E 2006, 73, 021501. (24) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Nature 1997, 389, 827–829. (25) Conway, J.; Korns, H.; Fisch, M. R. Langmuir 1997, 13, 426–431. (26) Reeves, J. P.; Dowben, R. M. J. Cell. Physiol. 1969, 73, 49–60. (27) Needham, D.; Evans, E. Biochemistry 1988, 27, 8261–8269. (28) Needham, D.; McIntosh, T. J.; Evans, E. Biochemistry 1988, 27, 4668– 4673. (29) Adams, D. R.; Toner, M.; Langer, R. Langmuir 2007, 23, 13013–13023.

Langmuir, Vol. 24, No. 15, 2008 7689 distearoyl-sn-glycero-3-phosphocholine (DSPC) (Avanti Polar Lipids, Inc., Alabaster, AL) was prepared in chloroform (10 mg/mL), to which was added a small amount of the fluorescent lipid analogue 1,1′-didodecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiIC12(3)) (Invitrogen-Molecular Probes, Eugene, OR) such that the DSPC/DiIC12(3) molar ratio was 1000:1. Next, 50 µL of the solution was spread on a sanded Teflon (Scientific Commodities, Inc., Lake Havasu City, AZ) disk cut to fit in the bottom of a 50 mL glass beaker, which was then placed in a vacuum oven (Thermo Scientific-NAPCO, Waltham, MA) overnight at 710 mmHg vacuum and room temperature. The resulting lipid films were hydrated with 4 mL of 0.22-µm-filtered 200 mM trehalose (Sigma-Aldrich, St. Louis, MO) in ultrapure water heated to 60–65 °C. The covered beakers were then placed in a water bath (at the same temperature) and allowed to sit for 12 h before harvesting the resulting liposome suspension via pipet aspiration. For some experiments, the liposomes were concentrated by centrifugation for 30 min at 325g followed by removal of the supernatant. Some experiments were also performed with a solution from which the liposomes had been removed by syringe filtration (0.22 µm) in order to determine the effect of any free lipid. After mixing the liposome suspension, a micropipet was employed to form 6 or 10 µL droplets on either cell culture polystyrene (Becton, Dickinson and Company, Franklin Lakes, NJ) or glass (Fisherfinest premium microscope slides: Fisher Scientific Research, Pittsburgh, PA); it should be noted that the type of glass has a significant effect on the droplet contact angle and thus the results obtained. Trehalose solution (200 mM) droplets without liposomes were prepared in the same manner, as were those containing trehalose and 50 µM calcein (Sigma-Aldrich, St. Louis, MO). The droplets thus formed were then dried in a dry box (Thermo Fisher Scientific, Waltham, MA) containing calcium sulfate (W. A. Hammond Drierite Co. Ltd., Xenia, OH), which has a relative humidity of approximately 5%.9 The dry box was kept at room temperature and covered with aluminum foil to shield the samples from light. To determine the percentage of droplets cracked in long-term (approximately two month) dry box storage, 40 droplets were analyzed on polystyrene and 50 droplets on glass, all 10 µL in volume. For many of the micrographs, 6 µL droplets were employed because this smaller volume allowed more of the droplet to be imaged, particularly for those droplets on glass since they spread out more, without qualitatively affecting the flow or pattern obtained. Those samples dried under the microscope were desiccated in enclosed chambers containing calcium sulfate to mimic the conditions experienced in the dry box. These were created from 100-mm-diameter × 20-mm-high polystyrene dishes for cell culture (Becton, Dickinson and Company, Franklin Lakes, NJ) into which were glued two parallel Teflon spacers approximately 30 mm apart. In the center, between the spacers, the droplet was placed either on the polystyrene itself or on a glass slide that fit between the spacers. The remaining volume of the cell culture dish, separated by the Teflon spacers, was filled with calcium sulfate. The dish was then covered with the supplied lid, and the edges sealed with Parafilm M (Alcan Packaging, Chicago, IL). An inverted microscope, the Axiovert 200 M (Carl Zeiss, Oberkochen, Germany) equipped with an AxioCam MRm, was utilized to image the resulting droplets. The liposome distribution during drying was obtained from the fluorescence intensity variation within the droplet. The radial symmetry of this distribution allowed for simple characterization via the variation in fluorescence intensity with radial position, which was obtained using ImageJ, the public domain image analysis program available from the National Institutes of Health;30 a plug-in for calculating radial intensity profiles has already been developed.31 This program was employed to integrate the pixel intensity values along concentric circles emanating from the droplet center; the integrals were then normalized by the length of integration. The integration was carried out over a slice of the (30) Rasband, W. S. ImageJ, v.1.35n; U.S. National Institutes of Health: Bethesda, MD, 1997–2006; http://rsb.info.nih.gov/ij/. (31) Carl, P. Radial Profile Extended Plugin for ImageJ, v.1; U.S. National Institutes of Health: Bethesda, MD, 2006 http://rsb.info.nih.gov/ij/plugins/ radial-profile-ext.html.

7690 Langmuir, Vol. 24, No. 15, 2008

Adams et al.

Figure 1. Initial droplet shape with resulting liposome distribution after drying. (A) Side view photograph of initial droplet deposited on polystyrene with corresponding top view (B) phase contrast and (C) fluorescence micrographs after drying, compared with (D) a photograph of the initial droplet on glass with corresponding (E) phase contrast and (F) fluorescence micrographs after drying. All droplets had a volume of 6 µL; the droplets in the micrographs were dried for approximately 45 min. The magnification for the micrographs was 2.5×; the scale bars represent 1000 µm for all images.

droplet rather than its entirety because the droplets were too large to fit in the microscope field of view (Figure 1C,F). For the 6 µL droplets used in the numerical analysis (Figure 2), the droplet formed on polystyrene was integrated over a 154° region. Given that the droplet on glass spread out more, it was integrated over a smaller slice of 110°. A similar procedure was followed for the calcein fluorescence intensity profiles in Figure 4. For both droplet types, this resulted in a liposome distribution profile that was relatively smooth at the edge of the droplet, where the length over which the total intensity is averaged is higher, but jagged at the center where the integration is carried out and normalized over a very small length. Data very near the center of the droplet, for which r/R < 0.1, where r is the radial position and R is the contact radius of the droplet, was thus excluded because it is unlikely to be meaningful. The rest of the data was smoothed via a cubic spline contained within MATLAB 7.0 (The MathWorks, Natick, MA). The smoothing parameter p was chosen such that minimal alteration of the (already smooth) data at the edge of the droplet occurred. Parameter p may range from zero to unity, where zero provides a linear fit to the entire data set and unity gives the (variational) cubic spline interpolant. Thus, a value near unity was chosen. Given that the integrals were averaged over a greater length at the edge of the droplet than toward the center, the degree of smoothing was radially weighted as (r/R)b, where b is a constant. The fluorescence intensity profiles for calcein in Figure 4 were not smoothed. The contact angles of droplets of the liposome suspension in 200 mM trehalose on the two substratesscell culture polystyrene and glassswere assessed by photographing the droplets with a Canon EOS 5D camera equipped with a macrolens. The droplets were back lit to increase the contrast with their surroundings. The contact angle was then determined using ImageJ.30

3. Results and Discussion The liposome preparation procedure resulted in nearly spherical liposomes having a median diameter of 7 µm,29 which is comparable to that for mammalian cells. Droplets of 200 mM

trehalose solution with and without liposomes were deposited on polystyrene and glass owing to their common use in cell culture and preservation; for both surface types, the droplets initially took on a spherical cap shape with a higher contact angle for the more hydrophobic polystyrene surface. As the droplets air dried, flow occurred within the droplet resulting in very different liposome distribution patterns for the two substrates. For polystyrene, a ring deposit formed, which was an expected result from the literature on the drying of aqueous particle suspensions; the previously unobserved accumulation of liposomes at the center of the droplet on glass likely owes to the glass-forming nature of the trehalose solution. The surface type, as well as the liposomes themselves, also had a strong effect on crack formation. In the absence of liposomes, most of the droplets did not crack upon further drying. When liposomes were present, cracks formed in a similar fashion to those observed in a variety of other colloidal systems. 3.1. Liposome Microflow and Droplet Deformation: Effect of Surface Type. For both surfaces, the droplets initially adopted spherical cap shapes (Figure 1A,D), as is expected given their small size for which capillary, rather than gravitational, effects dominate.32 For droplets of the same volume (6 µL), those formed on the more hydrophobic polystyrene had a smaller contact radius (1.91 ( 0.03 mm) than those formed on glass (2.84 ( 0.07 mm). The initial contact angle of the droplet formed on the more hydrophobic polystyrene surface was 53 ( 3° whereas that on glass was 16.7 ( 0.8°. For each substrate, desiccation failed to alter the droplet radius (i.e., the contact line was pinned); however, drying deformed the droplet shape, creating a ridge at the edge surrounding a central depression. Drying of sessile droplets of the liposome suspensions with trehalose on the two different surfaces led to strikingly different (32) de Gennes, P.-G.; Brochard-Wyart, F.; Quéré, D. Capillarity and Wetting Phenomena: Drops, Bubbles, Pearls, WaVes; Springer: New York, 2004; p 291.

Microflow and Crack Formation Patterns

Figure 2. Liposome distribution profiles during drying. The liposome distribution kinetics are shown for (A) polystyrene (black, 12 min; blue, 20 min; red, 44 min) and (B) glass (black, 2 min; blue, 10 min; red, 43 min). For both surfaces, 6 µL droplets were dried under the microscope in enclosed chambers containing desiccant. The dots display the fluorescence intensity data at various radial positions obtained from image analysis, and the lines are the cubic smoothing spline fits to the data. The smoothing parameter, p, for the profiles in A was 0.9999 with a weighting power b of 4; for the profiles in B, p was 0.99999, and b was 4, except for the profile at 10 min, for which b was 7. The early-time (black) spline fit for the polystyrene data is represented by a dashed line very near the droplet edge because the high initial curvature creates an optical artifact in this region (0.86 < r/R < 1), and thus no intensity data was available.

liposome distribution patterns. After drying on polystyrene for approximately 45 min, at which point no more liposome movement was observed, a clear ring deposit consistently formed near the edge of the droplet (Figure 1B,C). In contrast, drying the droplets on glass until the distribution ceased shifting (approximately 25 min) resulted in the deposition of the liposomes near the center of the droplet with a wide region near the edge

Langmuir, Vol. 24, No. 15, 2008 7691

that was essentially liposome-free (Figure 1E,F). For both surface types, the patterns formed were approximately radially symmetric. Ring deposits, such as the one created by the liposomes on polystyrene, form owing to capillary flow and have been observed in many particulate-containing droplets.24,33 In contrast, the accumulation of the liposomes at the center of the droplet on glass was unexpected. Glass in general does not cause this behavior because other suspension types dried upon it have created ring deposits. Indeed, as long as the droplet remains pinned, outward capillary flow and thus a ring deposit are expected regardless of the surface type.33 Nor does the slightly larger droplet radius appear to be the cause of the central accumulation; ring deposits have been observed even for much larger drops (e.g., 15 cm).33 Ring deposits have also been observed for droplets having low contact angles comparable to that for the liposome suspension on glass: an aqueous microsphere suspension droplet having a contact angle of 14° was observed to form a ring deposit.24 For a pinned contact line, failure to form a ring deposit has been previously observed for aqueous solutions only in special chambers, wherein evaporation occurred mainly from the droplet center,33 or for cases where the particles sedimented or adhered rapidly to the surface.34 In such systems, the solute deposits uniformly. The glass-forming nature of the trehalose solution is thus the likely cause of the liposome accumulation at the droplet center. 3.1.a. Kinetics of Liposome Pattern Formation. The liposome flow patterns during drying were observed in order to discover the mechanism responsible for the disparate final distributions on polystyrene and glass. On both surfaces, the liposomes were initially observed to flow toward the droplet edge. For all drying times observed, once the deposited droplet was brought into focus, the liposomes flowed more rapidly on polystyrene than on glass; for example, at 8 min the maximum velocity was approximately 10 µm/s for polystyrene and about 5 µm/s for glass. The observed velocities were highest near the droplet center and decreased with time for both surface types. A ring deposit formed on both surfaces during the early phase of drying but was more pronounced for polystyrene than glass. After its formation, the liposomes at the droplet center continued to flow outward toward the ring, while the ring itself, as well as liposomes outside the ring, began to shift inward toward the droplet center. During the final drying phase, liposomes inside the ring reversed direction and also began to flow toward the droplet center. Though the flow toward the droplet center was slower, on the order of 1 µm/s, it was still conspicuous for the droplet on glass, whereas on polystyrene these flows were discernible only upon close inspection. Thus, while the sequence of microflow patterns was similar for both surfaces, outward flows dominated for the droplets on polystyrene whereas inward flows dominated for those on glass, ultimately creating the very different patterns. The kinetics of the liposome distribution is displayed quantitatively in Figure 2. Initially, the liposomes are distributed evenly throughout the droplet volume. The early profile (black) thus just approximates the droplet height: higher at the center (r/R ) 0) and lower at the edge (r/R ) 1). As the droplet begins to dry, the liposomes flow toward the edge initially for both droplet types, giving a profile with a slight peak at the edge of the droplet (blue); this peak is more distinct in the case of the droplet dried on polystyrene (Figure 2A) than for the droplet dried on glass (Figure 2B). As drying continues, the peak continues to increase in height and width for the droplet dried on polystyrene as more (33) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Phys. ReV. E 2000, 62, 756–765. (34) Sommer, A. P.; Gheorghiu, E.; Cehreli, M.; Mester, A. R.; Whelan, H. T. Cryst. Growth Des. 2006, 6, 492–497.

7692 Langmuir, Vol. 24, No. 15, 2008

Figure 3. Transient surface wrinkle formation in desiccating trehalose solution microdroplets. These phase-contrast micrographs display the formation of transient surface wrinkles in microdroplets on (A) polystyrene and (B) glass. In both cases, 6 µL droplets of 200 mM trehalose with liposomes were employed. The droplet on polystyrene was imaged 73 min after placement in the dry box whereas the droplet on glass was imaged after 41 min because it is thinner and dries more quickly. In both cases, the transient wrinkles formed upon removal of the droplets from the dry box and exposure to ambient humidity. The wrinkles are stationary rather than traveling. They appear to form only where there is a thin glassy surface skin overlying a liquid solution. For long drying times such as those pictured above, the center of the droplet on glass is very thin and likely has little solution underlying the skin; thus the wrinkles do not form there. Wider wrinkles appear to indicate a deeper underlying solution. The wrinkles typically relax within a few minutes; the heat generated by the microscope light may also contribute to the dissipation of the wrinkles, which typically occurs within a few seconds after imaging. The scale bars represent 1000 µm.

liposomes flow from the center to the droplet edge (Figure 2A, red); the maximum of the peak shifts inward slightly as drying continues. In contrast, for the droplet on glass, the inward flow in the late stages of drying is significant such that the shape of the profile inverts (Figure 2B, red) and is lower at the edge than in the early (black) profile, although a small peak remains at the very edge of the droplet. 3.1.b. Proposed Mechanism of Liposome Pattern Formation. As mentioned above, the overall flow of liposomes to the edge of the droplets dried on polystyrene is expected from capillary flow owing to contact line pinning, as has been observed for many other types of dispersed particulates.24,33 Were the contact line not pinned, the droplet contact radius would decrease with

Adams et al.

time owing to evaporation; pinning prevents this, resulting in fluid flow toward the droplet edge to maintain a constant contact radius.24 Furthermore, for contact angles of less than 90°, the evaporative flux is higher at the edge of the droplet than at the center,35 which enhances outward flow. Although the flow of particulates to the droplet edge is explained by an existing theory for evaporating sessile droplets of a pure liquid,24,33 certain features of the liposome outward flow differ qualitatively from those predicted, likely owing at least in part to the multicomponent nature of the solution. For example, the liposomes flow most rapidly near the center rather than near the edge of the droplet. Furthermore, the liposome velocities were lower, for each time observed, in the droplet with a smaller initial contact angle and decreased with time for a given droplet as the contact angle decreased; the pure liquid droplet theory predicts an increase in velocity as drying progresses. Unlike the case of a pure liquid, drying of a trehalose solution alters the transport properties of the remaining liquid. This is particularly significant with regard to the viscosity, which increases sharply with decreasing moisture content.36 Differences in evaporative flux, both between droplets on different surfaces and within a given droplet, may thus cause differences in viscosity and with it flow. The pure liquid droplet theory neglects viscous effects33 and thus could not be expected to capture all of the observed flow features. Furthermore, the theory assumes that the particulates are evenly distributed vertically,33 which is not the case as the liposomes sediment with time. Whereas most of the liposomes flow to the droplet edge on polystyrene creating a ring deposit, a fraction remain in the center of the droplet. Most of these liposomes appear to be from the upper end of the size distribution,29 and their failure to flow outward likely owes simply to more rapid sedimentation. In the absence of trehalose, DSPC liposomes adsorb to polystyrene, forming a uniform deposit upon droplet drying; however, adsorption is insignificant for DSPC liposomes in the presence of trehalose, particularly in the early stages of drying when the solution is still liquid and flowing.29 Sedimentation of a fraction of the larger particles at the droplet center has been observed in other desiccating droplets containing multiple particle sizes.37 Although the mobile liposomes flow outward from the center of the droplet for nearly all of the drying process on polystyrene, there is a slight shift in the peak of the liposome distribution inward from the droplet edge with time that is not observed in simpler systems such as aqueous solutions of microspheres.24 Given that the contact line remains pinned, this deviation likely results from the glass-forming nature of the solution. Glassy skin formation initiates at the edge of the droplet, progressing inward as the droplet dries. The presence of a skin is indicated by the transient surface wrinkles that form upon exposure of the drying droplet to an abrupt increase in humidity. Similar wrinkles form in layered polymer systems;38 in such systems, thermal contraction of the underlying soft layer wrinkles the overlying rigid layer.39 For the trehalose solution droplet, water condensation may slightly expand the overlying skin; owing to the confined (pinned) geometry, this creates wrinkles where underlying liquid is still present. At early drying times, such wrinkles form only at the droplet edge; eventually, the skin covers the entire droplet surface, (35) Hu, H.; Larson, R. G. J. Phys. Chem. B 2002, 106, 1334–1344. (36) He, X.; Fowler, A.; Toner, M. J. Appl. Phys. 2006, 100, 074702. (37) Sommer, A. P.; Ben-Moshe, M.; Magdassi, S. J. Phys. Chem. B 2004, 108, 8–10. (38) Genzer, J.; Groenewold, J. Soft Matter 2006, 2, 310–323. (39) Bowden, N.; Brittain, S.; Evans, A. G.; Hutchinson, J. W.; Whitesides, G. M. Nature 1998, 393, 146–149.

Microflow and Crack Formation Patterns

Langmuir, Vol. 24, No. 15, 2008 7693

Figure 4. Dried trehalose solution microdroplet with hydrophilic dye. Microdroplets (6 µL) initially containing 200 mM trehalose with 50 µM calcein (a fluorescent hydrophilic dye) were dried in the dry box for 8 h. The phase contrast (A), and corresponding fluorescence micrograph (B), as well as the fluorescence intensity profile (C) are shown for the droplet on polystyrene. Also, the phase contrast (D), and corresponding fluorescence micrograph (E), as well as the fluorescence intensity profile (F) are shown for the droplet on glass. The scale bars represent 1000 µm.

and wrinkles form where liquid remains trapped underneath (Figure 3). Physically touching the droplet surface provides another means of verifying the existence of the skin at a given location.40 Recent experiments suggest that despite its thinness the glassy skin presents a significant barrier to evaporation in trehalose solutions.12 If evaporation is hindered at the edge of the droplet, then the liposomes may no longer flow to the very edge but rather to a point some finite distance inward. This point would shift further inward with time as skin formation progresses. The region underneath the thin glassy surface skin would still be liquid, allowing the liposomes to flow back away from the contact line. The formation of a glassy skin starting at the droplet edge could thus account for the observed slight inward shift of the peak of the liposome distribution with time. Eventually, the liposomes sediment or the viscosity increases sufficiently such that no further significant changes in the distribution are observed, leaving a ring deposit near, but not at, the contact line. Another factor that may contribute to the slight inward shift is that, although the glassy skin at the edge of the droplet remains pinned, the still-liquid region underneath it, as well as the liquid region closer to the droplet center, may undergo depinning as the contact angle becomes very small. For pure water, depinning has been observed at a contact angle of 2–4°.35 For low contact angles, evaporation of a depinned droplet results in fluid motion toward the center of the droplet.20,41 The thin glassy skin with underlying liquid creates an unusual situation for which the deposition of particulates at the droplet edge is reversible, to the extent that the particles do not sediment or become trapped in the skin. In the more typical case without a glass-forming solution, particles are essentially irreversibly deposited at the edge as it dries. In later stages, depinning may (40) Aksan, A.; Morris, S. C.; Toner, M. Langmuir 2005, 21, 2847–2854. (41) Petsi, A. J.; Burganos, V. N. Phys. ReV. E 2006, 73, 041201.

cause an inward flow of particles in the still-liquid region closer to the center of the droplet, resulting in a broadening of the ring deposit or in some cases more complex patterns,42 but this does not disturb those particles already deposited at the contact line.20 The reversibility of particle deposition suggests that altering the dynamics of glassy skin formation will lead to changes in the liposome distribution within the dried droplet. Indeed, this was observed to be the case because altering the contact angle, and thus desiccation kinetics, had a significant effect on the liposome distribution, as was observed for the droplet dried on glass. In contrast to the case of the droplet dried on polystyrene, drying on glass resulted in an accumulation of the liposomes at the center of the droplet, giving a largely liposome-free ring at the edge rather than a ring deposit. Although the final liposome distribution for the droplet dried on glass is unusual, the initial flow patterns are those expected from capillary flow. During the initial drying phase, the liposomes were observed to flow radially outward, as with the droplet on polystyrene, albeit at a slower velocity; however, a distinct ring never had time to form at the edge owing to the slower outward flow and larger contact radius, as compared to the polystyrene case. Once the glassy skin begins to form, the liposomes that did move to the edge may then begin to shift inward toward the center via the same mechanism as droplets on polystyrene, except for the few that become trapped in the skin, likely resulting in the slight peak in the distribution observed at the edge of the final profile. Given that the droplet is thinner and likely dries faster, skin formation may progress more quickly toward the droplet center than in the case of the droplet on polystyrene, shifting the point of highest evaporative flux toward the center of the droplet more rapidly. Furthermore, the critical contact angle below which depinning may occur, for the solution under the skin, would be reached sooner for the droplet on glass. (42) Deegan, R. D. Phys. ReV. E 2000, 61, 475–485.

7694 Langmuir, Vol. 24, No. 15, 2008

3.1.c. Droplet Deformation and Trehalose Distribution. Interestingly, the peak of the distribution for the remaining trehalose solution does not necessarily coincide with that for the liposomes. When viewed obliquely, a ridge was visible at the dried droplet edge for glass as well as for polystyrene, indicating the accumulation of trehalose solution near the droplet edge. Calcein, a hydrophilic dye that would be expected to have a distribution similar to that of trehalose when added to the solution at low concentration, likewise formed a ring deposit on both substrates (Figure 4). The center of the dried droplet was thinner for the droplet on the glass slide; its thickness was comparable to the size of the liposomes given that protrusions owing to the liposomes were visible near the center, which was not the case for the droplet dried on polystyrene. Also, transient surface wrinkles did not form in the thin film at the droplet center on glass, suggesting that little underlying liquid remained after skin formation (Figure 3B). The ridge was also narrower for the solution dried on glass. In both cases, the ridge likely forms owing principally to outward capillary flow. Buckling of the glassy skin may also play a role as observed in drying polymeric solutions,22,43 including those composed of carbohydrates such as dextran;14,15 however, the deformation of trehalose droplets is expected to vary from that observed with dextran owing to the greater rigidity of the trehalose glassy skin.40 Instead of deforming, the more rigid trehalose glassy skin may instead crack, as discussed in the next section. For the polystyrene case, the ridge location is not surprising because the liposomes and the trehalose then follow a similar distribution with the peak of the liposome distribution being only slightly closer to the edge than the peak of the trehalose ridge. It is more unexpected for the droplet dried on glass because the trehalose solution ridge exists in the liposomefree region, indicating a segregation effect. During the final stages of drying in non-glass-forming solutions, the contact line may cyclically depin and repin, giving rise to competing flows towards the droplet center and edge, owing to dewetting and capillary effects, respectively.42 Although the glassy skin at the edge of the trehalose solution droplet remains pinned throughout the drying process, the still-liquid solution underlying the skin, as well as the liquid core near the droplet center, may undergo depinning and repinning. Given that the liposomes sediment, vertical variations in these competing flow profiles could lead to the segregation phenomena observed on glass. The capillary flow velocity is lowest near the substrate,44 such that this flow would affect the sedimented liposomes less than the trehalose and calcein molecules. Moreover, depinning may have more of an influence on the liposomes than on small hydrophilic molecules since the former are concentrated at the liquid–substrate interface, which is most affected by dewetting.42 Accumulation of the liposomes at the droplet center may be further enhanced by interparticle capillary forces, which come into play when sufficient evaporation has occurred such that the particles are only partially immersed.45 Though it is unlikely that they are principally responsible for the inward flow of liposomes in the drying droplet, recirculating flows may enhance the inward flow in a manner that favors the liposomes. Although liposomes were not observed to be traveling in opposite directions in different focal planes, it may be that the flow lacked sufficient strength to carry the liposomes upward. One possibility is Marangoni convection. Temperature-gradient-driven flows are typically weak in evaporating aqueous solutions;46 however, the inhomogeneous trehalose concentration along the droplet surface might provide a sufficient surface tension gradient for flow, particularly after (43) Head, D. A. Phys. ReV. E 2006, 74, 021601. (44) Hu, H.; Larson, R. G. Langmuir 2005, 21, 3963–3971.

Adams et al.

the formation of the glassy skin over part of the droplet. Buoyancydriven flows are also possible, although they are typically more significant in thicker solutions and thus it is unclear why they would form in the thinner droplet on glass rather than on polystyrene. Whatever their cause, vertical variations in the microflow profile that favor inward liposome movement do appear to play some role because drying the droplet in pendent mode resulted in less pronounced inward liposome motion; however, at least part of this effect likely owes to sedimentation and trapping of the liposomes in the glassy skin because the inward flow was also slightly less pronounced for pendent droplets on polystyrene. Further investigation is necessary to pinpoint the exact mechanism behind this intriguing segregation phenomenon. 3.1.d. Summary of Liposome Microflow and Droplet Deformation In summary, competing fluid motion effects are present during desiccation of sessile droplets, namely, motion towards the droplet edge, owing to capillary flow, and towards the center, owing to dewetting. Different flows may dominate for different surface types and droplet components. Surface type affects the droplet shape, drying rate and glassy skin propagation, which in turn alter fluid flow velocity. Sedimentation changes the response of liposomes to fluid flows because their velocity profiles vary vertically in the drying droplet. This complex set of effects creates disparities in the distribution of liposomes on different surfaces, as well as segregation of liposomes from small hydrophilic molecules on a given surface. 3.2. Crack Formation: Effect of Surface Type and Liposome Concentration. In addition to affecting the dried droplet shape and liposome distribution, the surface type has a strong effect on crack formation, as do the liposomes themselves. The surface type affects both the likelihood and pattern of cracks for trehalose solution droplets in the absence of liposomes. When liposomes are present, all droplets form cracks, the pattern of which is dependent on the surface employed. 3.2.a. Crack Formation in the Absence of Liposomes. In the absence of liposomes, most of the droplets do not crack at all, even after months of storage in the dry box; this is true for droplets dried on polystyrene (Figure 5A) as well as those dried on glass (Figure 5D). In cases without liposomes where cracks did form (i.e., 35% of the droplets for polystyrene and 8% of the droplets for glass), the surface had a significant effect on the pattern of crack formation. Both patterns had a hierarchical netlike topology, although the cracks were found throughout the droplet for the case of polystyrene (inset, Figure 5A), whereas for droplets dried on glass the netlike cracks were largely confined to the ridge near the droplet edge, with a few approximately parallel cracks in the center (inset, Figure 5D). Droplets formed on polystyrene that cracked also often displayed a deep circular macroscale crack just inside the ridge at the edge of the droplet (inset, Figure 5A). The spacing between cracks has been previously observed to increase approximately linearly with the thickness of the dried solution for many systems.47,48 A similar result was found here: the cracks were closer together for the thinner droplet on the glass slide compared to that on the polystyrene surface, with the exception of the thin film at the center of the droplet on glass where few cracks were found. Previous experiments with a highly concentrated (∼1 M) trehalose solution resulted in random rather than hierarchical cracks,7 likely owing to the difference in drying dynamics. Crack patterns similar to those for the droplet on polystyrene (inset, Figure 5A) were observed for highly concentrated solutions of trehalose with low-molecular-weight (45) Denkov, N. D.; Velev, O. D.; Kralchevsky, P. A.; Ivanov, I. B.; Yoshimura, H.; Nagayama, K. Langmuir 1992, 8, 3183–3190. (46) Hu, H.; Larson, R. G. J. Phys. Chem. B 2006, 110, 7090–7094.

Microflow and Crack Formation Patterns

Langmuir, Vol. 24, No. 15, 2008 7695

Figure 5. Effect of surface type and liposome concentration on crack formation. Trehalose solution droplets were dried on polystyrene (A) without liposomes, (B) with a low concentration of liposomes, and (C) with a high concentration of liposomes. The inset in A shows the unusual case of crack formation in the absence of liposomes. The dark circles in the centers of B and C are bubbles. The higher liposome concentration in C relative to that in B results in a wider and likely higher liposome ring deposit, as has been observed for other particulates upon an increase in concentration.21 Droplets were also dried on glass (D) without liposomes, (E) with a low liposome concentration, and (F) with a high liposome concentration. The inset in D is for the uncommon case of crack formation in the absence of liposomes. The inset in E shows a close-up of cracks emanating from the liposomes in the depression in the center of the droplet; the ridge is to the left of the inset image, and the center of the droplet is to the right. The higher liposome concentration in F relative to that in E results in a larger central liposome accumulation region, likely owing to packing constraints. All droplets initially contained 200 mM trehalose and had a volume of 6 µL, except for the panel A inset and panel C, which both had a volume of 10 µL. Each image was obtained using phase contrast; the fluorescence image displaying the liposomes was then overlaid in red. The magnification for all of the images is 2.5×, except for the inset in E for which it is 10×. The scale bars represent 1000 µm, except for that in the inset of E, which is 50 µm.

dextrans whereas those with higher-molecular-weight dextrans did not crack,7 as was found for most droplets of low trehalose concentration employed in the current study. The surface type was also previously observed to have a profound effect on highly concentrated trehalose solutions: surfaces coated with a dry lubricant eliminated crack formation for all compositions.7 Alterations in the droplet-substrate interaction have the potential to influence the glassy skin mechanical stress profile and, in turn, the likelihood and pattern of cracks. Such interactions may be mediated by the contact angle, with its concomitant effects on dried droplet shape, thickness, and remaining moisture distribution, as well as the adhesion of the droplet to the surface and defects in the substrate itself. The concentration and composition of the glass-forming solution also play a role through their effects on drying dynamics and glassy skin elasticity. 3.2.b. Crack Formation in the Presence of Liposomes. More relevant to cell preservation is the case wherein liposomes are present. When even a low liposome concentration was present, all of the droplets began forming cracks within a couple of hours for both polystyrene (Figure 5B) and glass (Figure 5E); hierarchical netlike crack patterns formed that were similar to those in the absence of liposomes for the few cases where cracks appeared. For the droplets containing low liposome concentrations dried on polystyrene, tiny cracks first formed at the droplet edge and propagated radially inward. Although cracks were observed to emanate from liposomes at the very edge of the droplet, there were additional cracks that did not seem to originate at liposomes; these cracks were approximately evenly spaced along the edge

of the droplet. For low liposome concentrations, the liposomes are unlikely to be present near the skin, except perhaps the liposomes trapped at the very edge of the droplet. Although they are soft, deformable bodies, it is possible that these liposomes create defects in the skin that nucleate crack formation, after which point cracks form at regular intervals as has been reported in the literature.47 The possibility that liposomes may nucleate cracks was further supported by observations with pendent drops, wherein many cracks were observed to emanate from liposomes that had sedimented and become trapped in the skin. For the droplets that did crack in the absence of liposomes, this nucleation role was perhaps played by substrate defects, dust, or small bubbles.49 The cracks do not appear to be caused by free lipid accumulating at the droplet surface because the removal of liposomes via filtration resulted in crack-free droplets. Similarly, the addition of low concentrations of hydrophilic dyes such as calcein did not result in crack formation (Figure 4A,D). After nucleation of the radial cracking, additional cracks began to form approximately perpendicular to the initial ones. Subsequent cracks propagate so as to meet the previous cracks at approximately a 90° angle because formation of a given crack releases stress in only one direction.50 The direction of crack propagation thus depends on the previous cracks. This hierarchical method of propagation eventually leads to a smooth netlike pattern of cracks. Such patterns have been reported in dried thin films (47) Allain, C.; Limat, L. Phys. ReV. Lett. 1995, 74, 2981–2984. (48) Bohn, S.; Platkiewicz, J.; Andreotti, B.; Adda-Bedia, M.; Couder, Y. Phys. ReV. E 2005, 71, 046215.

7696 Langmuir, Vol. 24, No. 15, 2008

Figure 6. Crack topology distribution. Trehalose solution droplets with a low concentration of liposomes cracked in a netlike fashion upon desiccation on both polystyrene and glass surfaces. The number of sides of each domain partitioned off by cracks, as well as the number of first neighbor domains, was counted for 50 domains per droplet. The experiment was performed in triplicate on each surface and the average fraction of domains in a given bin plotted as a histogram; the error bars characterize the variation among droplets. For both surfaces, 6 µL droplets initially composed of 200 mM trehalose with liposomes were desiccated in the dry box for 28 hrs.

of colloids, gels, and ceramic glazes50 that look very similar to those observed for the dilute liposome suspensions. For a low liposome concentration, droplets dried on glass likewise exhibited crack propagation from the droplet edge, where there were a few liposomes. In contrast to the case of droplets dried on polystyrene, those dried on glass also displayed crack propagation from the center of the droplet, where most of the liposomes were located. For the droplets dried on glass there were additional short jagged cracks around some individual liposomes found in the central thin film, particularly those closer to the ridge; the droplet center had very few cracks in the absence of liposomes even when cracks did form at the ridge. Netlike cracks generally do not form in very thin layers51 such as the one at the center of the droplet dried on glass; rather, defects, in this case owing to the liposomes, nucleate jagged cracks in a nearly simultaneous, rather than hierarchical, fashion.52 For the liposomes toward the center but near the ridge, straighter, longer cracks emanated from the liposomes toward the ridge where they became part of the netlike pattern (inset, Figure 5E). For high liposome concentrations, a jagged branchlike crack pattern was observed for the droplets dried on polystyrene (Figure 5C) rather than the smooth netlike pattern found for low liposome concentrations (Figure 5B). At high enough concentrations, the liposomes may be packed so that they are in or near the skin over much of the droplet surface, in contrast to the low-liposomeconcentration case where the liposomes are likely present near the skin only at the very edge of the droplet. This may cause defect formation in the skin near the liposomes, resulting in the jagged cracks. Such patterns have been reported previously in the literature; evaporation-induced surface shrinkage causes these rough fracture patterns in disordered media because of its breaking threshold heterogeneity.53 The droplets dried on glass having high liposome concentrations (Figure 5F) were similar in appearance to those with low liposome (49) Tirumkudulu, M. S.; Russel, W. B. Langmuir 2005, 21, 4938–4948. (50) Bohn, S.; Pauchard, L.; Couder, Y. Phys. ReV. E 2005, 71, 046214.

Adams et al.

concentrations (Figure 5E), displaying a netlike crack pattern in the ridge around the edge of the droplet and small jagged cracks around some of the liposomes in the center, although the width of the netlike patterned region was somewhat narrower. Given that most of the liposomes accumulate in the depression near the center, it is perhaps not surprising that the result is largely unchanged from the low-liposome-concentration case because the ridge area near the droplet edge where the hierarchical cracks form remains largely free of liposomes. A more quantitative description of the crack topology may be found by counting the number of sides of each domain partitioned off by cracks as well as the number of immediately adjacent domains bordering a given domain; such a description is shown for liposome suspensions dried on polystyrene and glass in Figure 6. For a given surface, the topological distribution varied little from droplet to droplet. Theoretically, it can be shown that for hierarchical crack formation the average topological domain should have four sides and six immediate neighbors.54 This appears to be a satisfactory description of the netlike crack patterns observed.

4. Conclusions In summary, the substrate on which cells are dried is an often overlooked consideration in anhydrobiotic preservation that likely affects the conditions experienced by cells during drying as well as their final moisture content, particularly for cells in suspension. Using cell-sized liposomes as a model system, very different distribution results were obtained on two surfaces relevant to cell preservation: a glass slide and a polystyrene dish. Interestingly, the liposomes accumulated in the center of the droplet dried on the glass slide rather than at the droplet contact line as would be expected from capillary flow; this unusual result is likely due at least in part to inhomogeneous glassy skin formation at the droplet surface. Given the likely spatial variation of moisture within the droplet, not only vertically owing to skin formation but also radially owing to droplet deformation and later crack formation, the overall moisture content typically used as an indicator of molecular mobility in anhydrobiotic preservation is possibly misleading. Given that the liposomes accumulate mainly in the thin film at the center of the droplet on glass, the overall moisture content would likely be an overestimate of the moisture surrounding the liposomes, because most of the remaining solution appears to exist in the liposome-free ridge at the droplet edge; in contrast, the overall moisture content would potentially underestimate the moisture conditions for liposomes dried on polystyrene. The distribution patterns may also be relevant to interactions between liposomes (or cells), particularly for the preservation of liposomes that are prone to fusion. Additionally, the various crack-formation patterns on the two surfaces may have either a detrimental or beneficial effect. It is conceivable that, similar to crack formation in ice during cryopreservation,55 cracks in dried trehalose cause mechanical damage to the cells encapsulated therein. On the other hand, cracks present a preferential path for evaporation47 after the formation of the glassy skin and thus speed further drying of the underlying solution. Given their importance in long-term anhydrobiotic storage, these possibilities warrant further investigation. The study of the mechanisms underlying the cell-sized liposome distribution patterns observed here also benefits emerging applications such as ink-jet printing of cell suspension droplets (51) Cafiero, R.; Caldarelli, G.; Gabrielli, A. J. Phys. A 2000, 33, 8013–8028. (52) Bohn, S. Colloids Surf., A 2004, 263, 46–51. (53) Colina, H.; de Arcangelis, L.; Roux, S. Phys. ReV. B 1993, 48, 3666– 3676.

Microflow and Crack Formation Patterns

for tissue engineering,37,56,57 particularly because the addition of glass-forming carbohydrates and subsequent drying could provide a means of preserving these constructs.58 More generally, self-organized patterns, such as those created by particle distribution, wrinkle formation, and crack propagation, have diverse applications in microscale and nanoscale fabrication and (54) Bohn, S.; Douady, S.; Couder, Y. Phys. ReV. Lett. 2005, 94, 054503. (55) Rabin, Y.; Olson, P.; Taylor, M. J.; Steif, P. S.; Julian, T. B.; Wolmark, N. Cryobiology 1997, 34, 394–405. (56) Mironov, V.; Boland, T.; Trusk, T.; Forgacs, G.; Markwald, R. R. Trends Biotechnol. 2003, 21, 157–161. (57) Calvert, P. Science 2007, 318, 208–209. (58) Flickinger, M. C.; Schottel, J. L.; Bond, D. R.; Aksan, A.; Scriven, L. E. Biotechnol. Prog. 2007, 23, 2–17. (59) Ball, P. The Self-Made Tapestry: Pattern Formation in Nature; Oxford University Press : New York, 2001; p 312.

Langmuir, Vol. 24, No. 15, 2008 7697

also showcase the beauty and complexity of nature.59 Multicomponent glass-forming solutions appear to be a particularly rich system for further study of these useful and fascinating phenomena. Acknowledgment. This research was supported in part by grants from the National Institutes of Health (nos. DK46270 and U54-CA119349). Further support was provided by the U.S. Army Research Office through the Institute for Soldier Nanotechnologies at MIT. The information contained herein does not necessarily reflect the position or policy of the U.S. government, and no official endorsement should be inferred. We are grateful to Professor Howard Brenner of MIT for his helpful suggestions. LA703835W