Formation of Salt Crystal Whiskers on Porous Nanoparticle Coatings

Oct 30, 2009 - contact with the coating. Sheetlike crystals formed from the base growth of whiskers that had fallen flat onto the porous coating surfa...
1 downloads 0 Views 3MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Formation of Salt Crystal Whiskers on Porous Nanoparticle Coatings Heng Zhang, Zhen Wu,† and Lorraine F. Francis* Department of Chemical Engineering and Materials Science, University of Minnesota, 151 Amundson Hall, 421 Washington Avenue SE, Minneapolis, Minnesota 55455. †Current address: The Dow Chemical Company, Shanghai, China. Received August 5, 2009. Revised Manuscript Received October 8, 2009 Salt crystal whiskers were grown from aqueous solution on porous nanoparticle silica coatings. Coated substrates were partially immersed in an aqueous potassium chloride solution and then kept in a controlled relative humidity chamber for whisker growth. The salt solution was pulled into the porous coating, reaching a steady level about 1 h after immersion. Crystals with whisker morphologies, typically 2-50 μm in lateral dimension and up to ∼1 cm in length, emerged from the coating surface at a position above the original liquid level. Crystallites pushed upward by attached whiskers indicated a base growth mechanism in which ions are added to the surface of a growing whisker that is in contact with the coating. Sheetlike crystals formed from the base growth of whiskers that had fallen flat onto the porous coating surface. The effects of solution concentration and relative humidity on growth were characterized and used to elaborate the transport phenomena and growth mechanisms. Salt whiskers were also grown on bare substrates immersed in salt solutions containing nanoparticles. In this case, growth occurred below the original contact line on coatings created by convective assembly.

Introduction The earliest accounts of the formation of whiskerlike salt crystals on the surfaces of porous materials saturated with salt solution date back to the late 1800s, as reviewed by Nabbaro and Jackson.1 In one of these early reports,2 aqueous sodium silicate gels containing dissolved salts were prepared and, on drying, irregular masses of fibrous, whiskerlike crystals appeared on their surfaces. Others studied whisker formation on porous materials, such as clay and porcelain, when they were in contact with salt solutions.1,3-6 Most research concerned alkali halide whiskers with diameters on the order of micrometers to tens of micrometers. Work on this topic continued into the early 1970s, a time when there was considerable interest in single-crystal whiskers.7 Since then, only a few reports of salt whisker growth from solution-saturated porous materials have appeared.8-10 This lack of attention is contrasted by a resurgence of interest in vaporliquid-solid crystal growth, a technique originally developed to create single-crystal whiskers7 and recently adapted to the production of nanowires.11 This article explores solution-based whisker growth using porous nanoparticle coatings as a platform for growth and controlled growth conditions. Several mechanisms have been proposed for whisker growth on solution-saturated porous materials. The first step is the same for *Corresponding author. E-mail: [email protected]. Phone: 1-612-6250559. Fax: 1-612-626-7246. (1) Nabarro, F. R. N.; Jackson, P. J. In Growth and Perfection of Crystals: Proceedings of the International Conference on Crystal Growth; Doremus, R. H., Roberts, B. W., Turnbull, D., Eds; John Wiley & Sons: New York, 1958; pp 14-98. (2) Fells, H. A.; Firth, J. B. Proc. R. Soc. London 1926, 112, 468–474. (3) Kober, P. A. J. Am. Chem. Soc. 1917, 39, 944–950. (4) Charsley, P.; Rush, P. E. Philos. Mag. 1958, 3, 508–512. (5) Tarjan, L., Matrai, M., Eds. Laboratory Manual on Crystal Growth; Akademiai Kiado: Budapest, Hungary, 1972. (6) Gyulai, Z. Z. Phys. 1954, 138, 317–321. (7) Levitt, A. P., Ed. Whisker Technology; Wiley-Interscience: New York, 1970. (8) Yellin, N.; Zelingher, N.; Ben-dor, L. J. Mater. Sci. 1986, 21, 504–506. (9) Yellin, N.; Ben-dor, L.; Zelingher, N. J. Mater. Sci. 1986, 21, 2648–2650. (10) Srivastava, R.; Chandra, S. Prog. Cryst. Growth and Charact. Mater. 2002, 44, 133–139. (11) Xia, Y.; Yang, P.; Sun, Y.; Wu, Y.; Mayers, B.; Gates, B.; Yin, Y.; Kim, F.; Yan, H. Adv. Mater. 2003, 15, 353–389.

Langmuir 2010, 26(4), 2847–2856

all: the pore fluid becomes supersaturated during evaporation, and salt crystals appear on the porous material. The whisker morphology has been attributed to selective growth at the extremity (tip growth) or from beneath the crystal (base growth). Sears12,13 proposed the tip growth mechanism. In this mechanism, a thin layer of liquid adheres to the whisker surface and is driven up to the tip by a surface tension gradient arising from a salt concentration gradient; growth occurs at an axial dislocation located at the tip of the whisker. Some studies supported this mechanism6,14 and others did not.4,15,16 Amelinckx17,18 proposed that the base growth of whiskers could be launched from a defect at the edge of the whisker. Shichiri15 considered another mode of base growth; he proposed that 2D nucleation occurs on the bottom surface of the whisker, which is in contact with a liquid layer, and then growth occurs by ion addition there. Zehnder and Arnold19 asserted a similar mechanism in their study of efflorescence. No mechanism for whisker growth has been universally accepted, indicating that the mechanism may depend on the particulars of the material and growth conditions. The crystallization of salts inside and on masonry, concrete, and art objects also has been studied for many years.20-26 In service, building materials may be exposed to a source of water that contains a dissolved salt, and on drying, salt crystallizes on the surface (efflorescence) and inside the material (subflorescence). (12) Sears, G. W. Acta Metall. 1955, 3, 367–369. (13) Sears, G. W. J. Chem. Phys. 1957, 26, 1549–1552. (14) Kato, N. J. Phys. Soc. Jpn. 1955, 10, 1024–1025. (15) Shichiri, T. J. Cryst. Growth 1974, 24-25, 350–353. (16) Shichiri, T.; Kato, N. Acta Metall. 1965, 13, 785–796. (17) Amelinckx, S. Philos. Mag. 1958, 3, 425–428. (18) Amelinckx, S. J. Chem. Phys. 1959, 31, 1687–1688. (19) Zehnder, K.; Arnold, A. J. Cryst. Growth 1989, 97, 513–521. (20) Evans, I. S. Rev. Geomorphol. Dyn. 1970, 19, 155–177. (21) Rodriguez-Navarro, C.; Doehne, E. Earth Surf. Processes Landforms 1999, 24, 191–209. (22) Rodriguez-Navarro, C.; Doehne, E.; Sebastian, E. Cem. Concr. Res. 2000, 30, 1527–1534. (23) Flatt, R. J. J. Cryst. Growth 2002, 242, 435–454. (24) Theoulakis, P.; Moropoulou, A. J. Porous Mater. 1999, 6, 345–358. (25) Steiger, M. J. Cryst. Growth 2005, 282, 455–469. (26) Scherer, G. W. Cem. Concr. Res. 2004, 34, 1613–1624.

Published on Web 10/30/2009

DOI: 10.1021/la902902k

2847

Article

Zhang et al.

The microstructure of efflorescence takes a variety of forms; in some cases, whiskers are found.27,28 Researchers in this field have considered the transport phenomena that govern the position of the liquid meniscus in the porous material. Lewin29 predicted the location of the saltwater meniscus, and presumably the site of salt deposition, by setting equal the rate of liquid flow to the surface and the rate of exit from the surface by evaporation. The rate of liquid flow to the surface is influenced by the pore structure and fluid properties, and the evaporation rate depends mainly on the relative humidity. Experimental studies of efflorescence agree with this general framework; namely, efflorescence is more likely under conditions of high relative humidity30 where this balance allows the meniscus to remain at the surface. Puyate and co-workers31,32 described the transport phenomena in more detail. They developed a 1D model that includes a diffusion-convection equation to determine the ionic composition distribution in the pore liquid. In this research, we develop a method to create salt crystals from solution using nanoporous coatings as a platform for growth and potassium chloride as a representative alkali halide. The effects of humidity and solution concentration are characterized. Electron and optical microscopy results are used to determine the transport phenomena that influence whisker formation and the mechanism of whisker growth. In addition, a new whisker growth mode-growth during convective assembly or free particle growth-is presented, and a new morphology-a salt crystal sheet-is described.

Experimental Section

Figure 1. Experimental setup for whisker growth on porous coatings. Not to scale. Table 1. Conditions Used for 2-Day Whisker Growth Studies

Preparation of Porous Coatings. Coatings were prepared by

dip coating aqueous silica dispersions onto 1  3 in.2 glass slides. Glass slides were cleaned with detergent in an ultrasonic bath, rinsed with ethanol and distilled water, and then dried in an oven at 120 C for 1 h. Coating dispersions were prepared using Cabot PG-022 (Cabot Corporation), an aqueous dispersion of cationically surface-treated silica nanoparticle aggregates (150 nm aggregates of 20 nm individual particles). The stock dispersion was diluted with distilled water to a solids loading of 10 wt % and ultrasonicated for 10 min before coating. Coatings were prepared by dip coating at a rate of 0.2 mm/s. Immediately after deposition, the back and side edges of the substrate were wiped clean and then the glass slide with a coating on one side was dried at 70 C for 1 h. The coating thickness was ∼0.35 μm, as determined by SEM. See Supporting Information. Whisker Growth on Porous Coatings. The experimental setup for whisker growth is shown in Figure 1. A salt solution was prepared by mixing the required amount of potassium chloride (99.6%, Mallinckrodt) with distilled water and ultrasonicating until completely dissolved (∼10 min). The salt solution was added to a 50 mL graduated cylinder outfitted with a small rubber stopper to hold the coated substrate. The graduated cylinder with solution was then placed into a humidity-controlled box (Electrotech Systems, Inc., model 503). The 24 in. W  18 in. D  15 in. H box contains a humidity sensor and a small fan for circulation. Preliminary experiments showed that the position of the graduated cylinder relative to the fan was important to the growth and hence the position was kept constant (25 cm vertical distance and 20 cm horizontal distance from the fan) for all studies reported here. (27) Lewin, S. Z.; Charola, A. E. Scan. Electron Microsc. 1980, 551–558. (28) Charola, A. E.; Lewin, S. Z. Scan. Electron Microsc. 1979, 379–386. (29) Lewin, S. Z. In Conservation of Historic Stone Buildings and Monuments; National Academy Press: Washington, DC: 1982; pp 120-144. (30) Goudie, A. S. Earth Surf. Processes Landforms 1993, 18, 369–369. (31) Puyate, Y. T.; Lawrence, C. J.; Buenfeld, N. R.; McLoughlin, I. M. Phys. Fluids 1998, 10, 566–575. (32) Puyate, Y. T.; Lawrence, C. J. Phys. Fluids 1998, 10, 2114–2116.

2848 DOI: 10.1021/la902902k

case

RH (%)

standard 80 relative humidity (RH) 60 effect 65 70 75 solution concentration effect

80 80 80 80 80

KCl conc (wt %) 2.9 2.9 2.9 2.9 2.9 2.9 1.5 2.9 9.1 23.1

whisker morphology long and straight short with crystal crust short with crystal crust short and curved long and straight or curved long and straight straight straight straight straight with exiaxed crystals

After the desired relative humidity was achieved, the box was briefly opened to put the coated substrate in place. The setup was then left undisturbed for the growth period, typically 2 days. During this time, the level of the salt solution dropped because of evaporation; the concentration of the salt solution in the graduated cylinder remained nearly constant because the liquid volume change was less than ∼2%. At the completion of a growth experiment, the coated substrate was carefully removed and allowed to dry under ambient conditions. The effects of two variables-relative humidity and solution concentration-were explored, as described in Table 1. The standard conditions, chosen after considerable preliminary research, represent conditions that produced mostly solitary straight whiskers. The standard solution concentration, 2.9 wt %, is about 10% of the saturation concentration for KCl in water. Variations from the standard conditions were made to explore the effects of experimental conditions on growth. Studies of Liquid Flow. The initial flow of salt solution into the porous coating was visualized using a digital optical microscope (Hirox MX series). For imaging purposes, the coated substrate was held vertically in a 10 mL beaker and the assembly was placed near the transparent wall of the humidity control box. The microscope’s macrozoom lens (MX-MACROZ VI) was Langmuir 2010, 26(4), 2847–2856

Zhang et al.

Figure 2. SEM micrographs of the whisker growth zone under standard growth conditions: (a) overview and (b) high magnification of a whisker base. mounted horizontally and focused on the coated substrate from outside of the box. After the required relative humidity was reached, salt solution or water was poured into the beaker and imaging began. The relative position of the liquid front in the porous coating was recorded at 15 s intervals. The effect of relative humidity and salt solution concentration on the liquid front movement was characterized. Steady state was achieved in 1 h or less. Whisker Growth Visualization. The growth of whiskers was captured using an optical microscope (Olympus BX60) outfitted with a CCD camera. A growth setup was designed for the geometry of this standard microscope. A coated glass slide was placed on the microscope stage, and the coating surface was covered with 250 μL of KCl solution (2.9 wt %). The solution wicked to the edge of the slide, which was also coated with silica nanoparticles, and whisker growth occurred outward from the coated edge. Images of whiskers emerging from the edge were gathered sequentially. This experiment was carried out under ambient conditions with relative humidity in the range of 65 to 70%. Whisker Growth with Free Particles. Whiskers were also grown on bare glass substrates from salt solutions that contained silica nanoparticles. A 2.9 wt % KCl solution was modified by the addition of a Cabot PG022 dispersion such that the quantity of silica was approximately 0.00025 wt %. Growth was carried out in the same way as described for the coated substrates. The relative humidity was 70%, and the growth time was 2 days. Characterization. After growth was complete, whiskers were characterized by several methods. Coated substrates with whiskers attached were imaged with a digital optical microscope (Hirox MX series). Substrates were also carefully fractured into smaller pieces for examination with field-emission scanning Langmuir 2010, 26(4), 2847–2856

Article

Figure 3. SEM micrographs of an early growth stage (1 day under standard conditions): (a) typical morphologies and (b) silica particles from the porous coating on the top of a whisker. electron microscopy (FESEM). Specimens were coated with 5 nm of platinum before imaging in a Hitachi S-4700 FESEM. Microdiffraction (Bruker-AXS microdiffractometer) was used to verify the portions of coatings that were exposed to salt solution. Single whiskers and sheets grown under standard conditions were detached from the substrate and examined by microdiffraction. One whisker and one sheet were additionally characterized by using a Bruker-AXS X-ray platform diffractometer.

Results Whisker Growth under Standard Conditions. Whiskers were first visible to the naked eye after approximately 7 h. More whiskers appeared as the liquid level in the graduated cylinder dropped slowly. These whiskers were identified as KCl by X-ray diffraction, as described later. After 2 days of growth under standard conditions, a visible zone of whisker growth was found well above the original liquid level. A digital image is in the Supporting Information. Figure 2a shows an SEM image of the whisker growth zone after the 2-day growth period. Most whiskers are straight with rectangular cross sections, lateral dimensions in the range of 2-50 μm, and lengths of up to ∼5 mm. On the basis of a sample of 225 whiskers, the average lateral dimension was 11 μm with a standard deviation of 7.5 μm. Figure 2 also shows crystals with sheetlike morphologies and fallen whiskers, as described more below. In addition, two other whisker morphologies were observed occasionally under standard conditions: whiskers with corrugated surfaces and tapered whiskers that were larger at the base (Supporting Information). Figure 2b shows a whisker base at higher magnification. Here, the surface of the porous coating is visible along with a dendritic DOI: 10.1021/la902902k

2849

Article

Figure 4. SEM micrographs of (a) fallen whiskers that are propped up and remain in contact with the porous coating such that an extension grows and (b) a fallen whisker in contact with a substrate that continues to grow into two sheets connected by the original whisker in the middle (inset).

pattern. The dendritic pattern appears to be a consequence of the crystallization of the pore liquid after the specimen is removed from the growth chamber. Interestingly, this crystallization takes place within or below the porous coating. A layer of exiaxed crystals is around the whisker base. The morphology in an early stage of growth is shown in Figure 3a. A scattering of small exiaxed crystals and tiny whiskerlike structures is apparent on the coating surface, along with a dendritic pattern. Some tiny whiskers are tilted, and small crystals appear on the tip of other larger whiskerlike structures. These small crystals could be seeds for whisker growth. Occasionally, silica particles from the porous coating could be observed on top of the whisker, as shown in Figure 3b. Whiskers removed from the coating after 2 days of standard growth were examined by X-ray diffraction. Ten long, straight whiskers were chosen and examined by X-ray microdiffraction. The results showed that the whiskers are single crystals of KCl with bounding surfaces belonging to the {100} or {110} family of planes; these two sets of planes were indistinguishable by the microdiffraction method. See Supporting Information. One whisker was further investigated with a single-crystal X-ray diffractometer, and the crystal faces were unambiguously identified as the {100} family. Fallen whiskers, detached from the substrate, were also observed (Figure 2a). These whiskers were sometimes propped up by neighboring whiskers. Figure 4a shows a pair of whiskers that fell in this manner but maintained contact with the porous coating, 2850 DOI: 10.1021/la902902k

Zhang et al.

Figure 5. (a) Digital images of the liquid front in the porous coating: (left to right) 1 min, 20 min, and 1 h after immersion in the KCl solution at 80% RH. (b) Effect of relative humidity on the liquid front position of KCl solution (2.9 wt % KCl) in porous coatings and the liquid front position of water in a porous coating at 80% RH.

which allowed further growth. In other cases, an entire whisker fell flat onto the porous coating and then continued to grow as a sheet. In Figure 4b, a single fallen whisker led to the growth of two sheets, which are connected on the top by the whisker. The right sheet appears to have grown faster or started growing sooner than the left sheet. See the Supporting Information for additional examples. Sheets were also examined by X-ray methods, and the faces of the sheets were in the {100} family of planes. Visualization Studies. The first study characterized the movement of water or salt solution into the porous coatings under different conditions. A glass substrate with a porous coating was placed vertically in water or salt solution, and a digital optical microscope was used to track the upward motion of a liquid front into the porous coating. Figure 5a shows an example of the progression of the liquid front under standard conditions. After 1 h, the front reaches the end of the coating. Figure 5b is a summary of the liquid front movement. When water was used, the liquid front rose quickly to about 1 mm above the contact line under the standard humidity condition (80% RH). However, when a salt solution was used under the same conditions, the liquid front rose by >2 cm to the end of the porous coating itself. A hazy zone appeared near the front edge for salt solutions. This hazy zone moved up the coating as the liquid Langmuir 2010, 26(4), 2847–2856

Zhang et al.

Article

Figure 6. Optical micrographs of the growth of whiskers on a porous silica coating with increasing growth time (t1 < t2 < t3 < t4 with a time interval of 5 s). The central whisker has a small crystal attached to its tip, making the growth apparent by extension from the base.

penetrated further and disappeared gradually after the liquid front reached the upper end of the coating. When the relative humidity was lowered, the liquid front height for a given salt solution decreased. X-ray microdiffraction was used to demonstrate the validity of liquid front tracking. The specimen used to track the position of a KCl solution front at 75% RH was examined above and below the liquid front line. KCl crystals were found below but not above the line, confirming that the salt solution had not reached above the front. The second experiment captured whisker growth under nearly standard conditions. Figure 6 shows sequential images of whiskers growing on the edge of a glass slide, as imaged with an optical microscope. As described, the experimental setup was different than that used in the growth experiments; however, the important features were the same. Salt solution imbibes the porous coating; evaporation starts, followed by salt crystallization and whisker formation. In Figure 6, the central whisker in the images increases in length at a rate of about 0.02 mm/s. As growth occurred, some whiskers were observed to sway, change orientation, and sometimes fall over or leave the field of view. The small crystal at the tip of this whisker is pushed further away in the process. This result indicates that the whisker grows from the bottom or base upward. Multiple experiments were carried out and support this conclusion. See Supporting Information for another example. Effect of Relative Humidity. The effect of relative humidity on whisker growth was studied by varying the relative humidity from standard conditions. Whiskers were observed over a range of humidity of 60-83%; however, whiskers grown at the lower end of this range were severely stunted by a nonwhisker salt crust. As the relative humidity increased, the position of the whisker growth zone shifted from closer to the liquid level to further away, consistent with the trend in liquid penetration. The whisker morphology also changed with relative humidity. See Table 1. As the relative humidity dropped from 80% (standard) to 75%, whiskers became more curved and irregular. Langmuir 2010, 26(4), 2847–2856

Figure 7. SEM micrographs of the whisker growth zone under conditions of (a) 65% relative humidity with all other conditions the same as the standard condition and (b) 60% relative humidity with all other conditions the same as the standard condition.

Curved whiskers may be related to a nonuniform supply of salt solution to the whiskers, as suggested by previous researchers.19 The quantity of whiskers appeared to be greater at 75% RH as compared with that at 80%, but whisker dimensions were similar. DOI: 10.1021/la902902k

2851

Article

Zhang et al.

Figure 8. Digital images of specimens grown from solutions of varying KCl concentration, as indicated, with all other conditions the same as the standard condition. The right end was the one immersed in the salt solution.

Lowering the humidity further to 65% led to a change in morphology as shown in Figure 7a. Fine, short whiskers and a continuous layer of nonwhisker crystals covered the whisker growth zone. Figure 7b shows this morphology for whiskers grown at even lower relative humidity (60%). Here, it appears that the crust of exiaxed KCl crystals choked the salt solution supply to the whiskers. Effect of Solution Concentration. Figure 8 shows specimens grown from KCl solutions with varying concentration at 80% RH. See also Table 1. As the concentration increased, whiskers appeared earlier in the 2-day growth period. The quantity of whiskers increased, and the growth zone moved closer to the original liquid level. Regardless of concentration, whiskers were straight and had about the same lateral dimensions. At the highest concentration, exiaxed crystals were incorporated in the growth zone. Whisker Growth on Bare Substrate with Free Particles. This method involved placing a bare glass substrate into KCl solution that contains free silica particles. Approximately 2 h after the glass slide was immersed, the leading edge (the contact line) turned white first and then whiskers quickly appeared in this region. Unlike growth on porous coatings, whiskers appeared right beneath the original liquid level and formed a substantial mass. See the Supporting Information for a digital image. This growth mode required the relative humidity to be below 75%. No whiskers appeared at higher relative humidity. The small concentration of silica particles in the growth solution was essential to the formation of whiskers. Without the silica particles, a nonwhisker crust formed on the glass slide in the course of 2 days. Figure 9 shows SEM images of the whisker growth zone of a specimen formed using free particle growth conditions. The whiskers have a similar morphology to that of whiskers grown on porous coatings. However, more whiskers were formed by this growth mode at the same relative humidity and salt concentration. 2852 DOI: 10.1021/la902902k

Some curved whiskers and sheet morphologies were also found. A coating of silica particles was formed coincident with the whiskers. After drying, this coating was cracked and peeling, as shown in Figure 9b.

Discussion In this section, the transport phenomena that create the conditions for crystallization are discussed along with a whisker formation mechanism. Models developed for wicking and salt degradation of concrete27,29,31,32 and convective assembly33,34 are used to explain the transport issues that combine to produce a whisker growth zone. A whisker formation mechanism is then proposed and compared with mechanisms set forth previously in the literature.12,15,18 In the two subsections, the focus is on the standard growth conditions, with results from the effects of relative humidity and concentration used to support the hypotheses. Transport Phenomena. Three phenomena are important to establishing the conditions for crystallization on the porous coating: capillary-driven flow of the salt solution into the porous coating, evaporation of water from the coating surface, and diffusion of ions in the solution-filled pore space. See Figure 1. At the start of the experiment, salt solution is pulled into the coating by capillary pressure and, simultaneously, evaporation occurs from the coating surface. The effect of evaporation on capillary rise has been reported in the literature.35,36 Here, we (33) Brewer, D. D.; Allen, J.; Miller, M. R.; Santos, J. M. D.; Kumar, S.; Norris, D. J.; Tsapatsis, M.; Scriven, L. E. Langmuir 2008, 24, 13683–13693. (34) Norris, D. J.; Arlinghaus, E. G.; Meng, L; Heiny, R; Scriven, L. E. Adv. Mater. 2004, 16, 1393–1399. (35) Fries, N.; Odic, K.; Conrath, M.; Dreyer, M. J. Colloid Interface Sci. 2008, 321, 118–129. (36) Hall, C.; Hoff, W. D. Proc. R. Soc. London, Ser. A 2007, 463, 1871–1884.

Langmuir 2010, 26(4), 2847–2856

Zhang et al.

Article

tension, γ, of 72 mN/m, a contact angle, θ, of 0, and a pore radius, r, of 50 nm. The density and viscosity of the water are 1000 kg/m3 and 1 mPa s, respectively, and the coating thickness is ∼0.35 μm. The permeability of the porous coating,37 κ = φ3/5(1 φ)2S2], is ∼1.7  10-18 m2 using measured values for the pore fraction (φ = 0.6) and surface area of 184 m2/g, corresponding to S = 4  108 m2 surface/m3 volume. The evaporation flux of water at 80% relative humidity was experimentally determined to be 8  10-5 kg m-2 s-1. Using these approximations, the final liquid front position is ∼0.15 mm, which is in the same range as the observed height (∼1 mm). The front position for salt solutions is much greater than that of water (Figure 5), an effect that cannot be explained by changes in liquid properties, such as the small (∼10%) increases in surface tension38 and viscosity,39 on the parameters in eq 3 or by the effect of salt on the mass balance. The difference may be related to the effect of salt on the evaporation rate. The evaporation rate depends on the relative humidity (RH) in the surrounding atmosphere, as well as the gas flow conditions: JE ¼ kðp - p¥ Þ

Figure 9. SEM of whiskers grown on bare substrates by the free particle method. Growth occurred for 2 days at 70% relative humidity using a 2.9 wt % KCl solution that contained 0.00025 wt % silica particles (150 nm aggregates).

include a brief analysis of the experimental setup in order to identify important parameters and discuss the results. To estimate the final liquid front position, L, the mass per unit time entering the coating by capillary flow via the cross-section (HW, where H is the thickness and W is the width of the coating) is equated with the mass per unit time exiting by evaporation via the coating surface (LW): JC ðHWÞ ¼ JE ðLWÞ

ð1Þ

JE is the evaporation mass flux. JC is the capillary-driven mass flux, which acts primarily up the length of the coating33 and is given by Darcy’s law JC ¼

KFΔP μL

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi KHFΔP L ¼ μJE

Pe ¼

ð3Þ

The parameters in eq 3 can be approximated for the standard condition of 80% RH and assuming pure water as the liquid. The pressure drop, ΔP = 2γ cos θ/r, is ∼3 MPa, using a surface Langmuir 2010, 26(4), 2847–2856

In eq 4, k is the mass-transfer coefficient, and p* and p¥ are the vapor pressure of water in the gas just above and far from the evaporating liquid surface, respectively. For pure water, p* is the saturation vapor pressure of water, psat. p¥ depends on the percent relative humidity, RH, and psat (p¥ = (RH/100)psat). For evaporation from liquid-filled pores with tiny menisci, the curvature lowers p* relative to a flat liquid surface. A calculation with the Kelvin equation40 reveals that this effect is significant only for evaporation from nanosized pores. For example, the saturated vapor pressure over water with a 50 nm radius of curvature is ∼98% of that over a flat surface. In addition, dissolved salt also lowers p*. The equilibrium vapor pressure over a saturated KCl solution at 25 C, for example, is 84.34% of that over pure water.41 The decrease in p* from this effect is significant,36 resulting in a decrease in the vapor-pressure difference (p* - p¥) and hence a lower JE for a given relative humidity. The depression of p* is one factor that lowers JE and hence raises L for salt solutions. The other is local salt crystallization. Because simultaneous capillary-driven flow and evaporation establish a region of liquid-filled pore space in the coating, the concentration of ions within the pore liquid increases. Capillarydriven flow brings ions in the water to the evaporating surface, which creates a higher local concentration, and the potential for diffusion to equalize this gradient. Puyate et al.31,32 define a Peclet number (Pe) to represent the competition between convection (capillary-driven flow) and diffusion

ð2Þ

where ΔP is the pressure drop in the liquid due to capillarity; μ and F are the viscosity and density of the liquid, respectively; and κ is the permeability of the coating. By substitution, L is

ð4Þ

vp, ave l D

ð5Þ

where vp,ave is the average velocity in a pore, l is the length over which capillary-driven flow and diffusion take place, and D is the (37) Schidegger, A. E. The Physics of Flow Through Porous Media; University of Toronto Press: Toronto, Canada, 1957; p 104. (38) Zhang, H.; Han, S. J. Chem. Eng. Data 1996, 41, 516–520. (39) Ali, K.; Shah, A. A.; Bilal, S.; Shah, A. A. Colloids Surf., A 2009, 337, 194– 199. (40) Hiemenz, P. C.; Rajagopalan, R. Principles of Colloid and Surface Chemistry; Marcel Dekker: New York, 1997; pp 261-262. (41) CRC Handbook of Chemistry and Physics, 89th ed.; Lide, D. R., Ed.; CRC Press: Boca Raton, FL, 2008.

DOI: 10.1021/la902902k

2853

Article

Zhang et al.

diffusion coefficient. The pore velocity is related to JC and the volume fraction of pores (φ):

Using approximations, Pe is estimated to be on the order of 25. Hence, capillary-driven flow dominates; the salt concentration rises at the liquid-vapor interface, eventually becoming supersaturated and launching crystallization. The haziness that developed in the coating near the wetting line is evidence of local crystallization where a supersaturated condition is reached. The hazy zone moved together with the liquid front, indicating that the crystals form and then dissolve as the liquid front moves upward. This local crystallization clogs pores, which along with the depressed p* decreases the evaporation rate42 and leads to a higher L for salt solutions as compared with that for water (Figure 5). It is possible that some form of creeping crystallization43 occurs in which the surface is sealed with a salt layer and the liquid climbs in the coating beneath it. If such a layer forms, it does not persist because the haziness is transient and whiskers eventually form in a zone beneath the liquid front position. The lower L for salt solutions imbibed at lower RH appears to be a consequence of increased JE; however, pore clogging may also play a role if the advancing liquid is arrested by more significant local crystallization. The origin and location of the whisker zone is a result of capillary-driven flow, evaporation, and diffusion. Qualitatively, this combination of effects can be understood as follows. The solution entering the coating has the same concentration as the salt solution in the graduated cylinder, but as this liquid travels up the porous coating under the influence of the pressure gradient, evaporation takes place and the salt concentration rises. Thus, the salt concentration in the pore liquid increases with distance away from the liquid pool. At some position up the length of the coating, the supersaturation is high enough and crystal nucleation and growth proceed. This location is the whisker growth zone. Salt is consumed in this zone, so solution that might proceed on to further reaches of the coating is depleted in salt and less likely to drive further growth. Therefore, above and below the zone, crystallization is less likely. This explanation of the growth zone squares with the observations of the effects of solution concentration and relative humidity. As the initial concentration of the salt solution increased, the position of the growth zone shifted to lower levels, closer to the liquid level. In this case, less evaporation is needed to raise the concentration of the solution as it travels up the coating and hence the zone is closer to the liquid level. Decreasing the relative humidity from the standard case had the same effect: a lower whisker growth zone. In this case, the lower humidity led to more rapid evaporation and a faster increase in concentration as the liquid travels up the coating. The whisker growth zone for the free particle case is also a consequence of flow, evaporation, and diffusion. The results show

that when a small quantity of silica particles was added to the salt solution, whiskers grew on bare glass substrates. This situation is more complex than the one discussed above; one hypothesis is presented here. In a manner similar to the convective assembly of colloidal crystals,34 free particles are pulled up to the contact line where the liquid wets the substrate and they deposit there to form a coating as evaporation proceeds and the liquid level drops. Evaporation also increases the salt concentration in the pore fluid, and it apparently does so very effectively, perhaps because of the adjacent supply in the bulk liquid because whiskers begin to form quickly at the initial point of contact. Whiskers continue to grow as the liquid level drops and more particles assemble into a coating. Whisker Growth Mechanism. KCl forms in the rock salt crystal structure, which is cubic (space group Fm3m). Ordinarily, cube-shaped crystals form.44 Whisker morphology is not expected on the basis of crystallography. In this work, visualization studies and features of the whisker zone microstructures provide evidence for whisker growth by ion addition to the bottom face of the crystal, which is in contact with the porous coating. We will call this a base growth mechanism, which is the term used by Shichiri.15 As described above, the pore fluid becomes supersaturated and crystals form. The location of the first crystals is difficult to confirm because once specimens are removed from the growth chamber the residual salt solution in their pores dries and crystals inevitably form. SEM images taken in the early stages show some crystals on the surface of the coating. See Figure 3a. These crystals are believed to be the initial stage of whiskers. Silica particles were found occasionally on the tops of some crystals and whiskers (Figure 3b), indicating that some crystals originated beneath the surface. It seems more likely that crystals form on the surface because growth inside the pore space is confined.45 In addition, Shahidzadeh-Bonn and co-workers46 reported that the interfacial tension is lowered during crystallization, which could enhance the spreading of liquid out of the pores to the surface of the coating. A crystal on the surface can grow uniformly as long as salt solution wets the exterior surfaces of the crystal. This uniform growth, however, can be sustained only if the salt solution contacts all of the surfaces and is uniformly saturated everywhere. At some point, this condition breaks down and the liquid may contact the base of the crystal only from beneath and on the edges of the crystal. The crystal layer at the whisker base (Figure 3b) is consistent with the existence of a liquid layer wetting the crystal (whisker) surface during growth. The wetting liquid layer provides stability to the whisker. When the sample is removed from the solution, the liquid layer at the crystal (whisker) base dries, crystallizing in a ring. At some point in the process, growth from the underside or base of the crystal occurs at a faster rate than elsewhere and a whisker emerges perpendicular to the substrate. Shichiri15 and Zehnder and Arnold19 previously reported this mechanism for whisker growth on bulk materials. Two key pieces of experimental evidence point to the base growth mechanism in the case of whisker growth on porous coatings. First, sequential optical microscope images (Figure 6) show a straight whisker extending without disrupting a small crystal attached near the tip of the whisker. The whisker appears to be pushed upward from beneath. Though gravity works against this action, the effect is expected to

(42) Espinosa Marzal, R. M.; Scherer, G. W. In Proceeedings of the 11th International Congress on Deterioration and Conservation of Stone; Nicolaus Copernicus University Press: Torun, Poland, 2008; Vol. I, pp 81-88. (43) Washburn, E. R. J. Phys. Chem. 1927, 31, 1246–1248.

(44) Lian, L; Tsukamoto, K.; Sunagawa, I. J. Cryst. Growth 1990, 99, 150–155. (45) Scherer, G. W. Cem. Concr. Res. 1999, 29, 1347–1358. (46) Shahidzadeh-Bonn, N.; Rafai, S.; Bonn, D.; Wegdam, G. Langmuir 2008, 24, 8599–8605.

vp, ave ¼

JC =F KΔP ¼ φ φμL

ð6Þ

On the basis of the analysis of Brewer and co-workers,33 l = L and therefore Pe ¼

2854 DOI: 10.1021/la902902k

KΔP φμD

ð7Þ

Langmuir 2010, 26(4), 2847–2856

Zhang et al.

Article

Figure 10. Schematic drawing of whisker and sheet growth mechanism in a time sequence. (t1 < t2 < t3 < t4 < t5 < t6). t1, whisker starts growing; t2, whisker continues to grow; t3, whisker tilts; t4, whisker falls; t5, sheet starts to grow, fed from liquid in the porous coating beneath; and t6, a salt sheet is formed.

be small given the small mass of the whiskers. The upward movement of the whiskers was observed repeatedly. Shichiri15 documented a similar effects for the growth of NaCl whiskers on bulk porcelain. Second, morphologies arising from fallen whiskers can only conceivably be the result of a base growth mechanism. Fallen but propped up whiskers touching the porous coating are pushed up further at their contact points by continued crystal growth supplied from beneath (Figure 4). More dramatic are sheet morphologies that arise when a fallen whisker makes extensive contact with the porous coating and growth continues on the face of the whisker in contact with the coating, forming a sheet. These features were commonly observed. Figure 10 is a schematic drawing of a time sequence of whisker and sheet growth by the mechanism proposed above. Taller and narrower sheets were also observed to fall onto the coating and grow from their contacting faces into 3D objects. Additionally, sheet morphologies formed on specimens created by the free particle method indicate that the same growth mechanism is at work in this case. There are several possible reasons for the preference for growth on the face in contact with the porous coating. First, the supply of salt solution is abundant there as compared with the layer of adhering salt solution on the other faces.45,47 Another possibility is that a dislocation or other defect mechanism is at play. The rough surface of the porous coating leads to the possibility that the crystal surface in contact with the coating is defective and the growth rate is enhanced.48 The literature is mixed on the role of defects and dislocations in whisker growth from solution. Shichiri16,49,50 studied the dissolution behavior of whiskers and noted that they appear to be defect-free, whereas Amelinckx51 documented dislocations in whiskers by a decoration method. We made multiple attempts to locate dislocations in our whiskers, but none were found. Some research on tabular-shaped alkali halides52 suggests that stacking faults may also lead to rapid and preferred directions of crystal growth. More work is needed to (47) Taber, S. Proc. Natl. Acad. Sci. U.S.A. 1916, 2, 659–664. (48) Frank, F. C. Discuss. Faraday Soc. 1949, 5, 48–54. (49) Shichiri, T.; Kinoshita, H.; Kato, N. Int. Conf. Cryst. Growth 1967, 385– 388. (50) Shichiri, T.; Kato, N. J. Cryst. Growth 1968, 3-4, 384–390. (51) Amelinckx, S. J. Appl. Phys. 1958, 29, 1610–1611. (52) Bennema, P.; Bogels, G.; Bollen, D.; Mussig, T.; Meekes, H. Imaging Sci. J. 2001, 49, 1–32.

Langmuir 2010, 26(4), 2847–2856

establish conclusively the role of defects in whiskers grown by the method described here.

Conclusions Salt crystal whiskers and sheets were formed using nanoporous coatings as a platform for growth. Whereas this report documented the growth of KCl, the method can also be adapted to create whiskers and sheets of other materials, such as NaCl and KBr. The steps in the growth process were established by characterizing the effects of solution concentration and relative humidity on growth. The solution is pulled into the porous coating by capillarity; concurrent evaporation leads to building concentration in the pore liquid and a whisker growth zone where the concentration reaches a critical level for the formation of salt crystals. Growth occurs by ion addition to the crystal surface that is in contact with the porous coating. This mechanism indicates the possibility of using seeds on the porous coating surface to control whisker geometry. The results presented here show the potential of this method: thin crystal sheets formed from whiskers as “seeds”. Salt whiskers and sheets with controlled dimensions may find applications related to their properties or their water solubility. For example, they can be added to polymers and then leached to create controlled porosity for biomaterial applications.53,54 Finally, a copious number of whiskers was generated from a salt solution containing nanoparticles in contact with a bare substrate under controlled relative humidity conditions. Here, the step of capillary flow into the coating was eliminated and growth occurred rapidly, apparently by the same mechanism. Acknowledgment. We thank the Petroleum Research Fund for supporting this work through grant 43388-AC and the generous supporters who established the Shell Professorship and the L. E. Scriven Chair. We also thank Wieslaw Suszynski for assistance with the visualization studies, Damien Brewer for insightful conversations about flow through porous media, Karan Jindal for providing information on porous silica coatings, and Victor G. Young for single-crystal X-ray diffraction of (53) Mikos, A.; Thorsen, A. J.; Czerwonka, L. A.; Bao, Y.; Langer, R.; Winslow, D. N.; Vacanti, J. P. Polymer 1994, I, 1068–1077. (54) Hou, Q.; Grijpma, D. W.; Feijen, J. Biomaterials 2003, 24, 1937–1947.

DOI: 10.1021/la902902k

2855

Article

whiskers and sheets. Parts of this work were carried out in the University of Minnesota I.T. Characterization Facility, which receives partial support from NSF through the NNIN program. Supporting Information Available: SEM of a nonporous silica coating, digital images of specimens formed under

2856 DOI: 10.1021/la902902k

Zhang et al.

standard growth conditions on porous coatings and free particle growth, X-ray microdiffraction data, other whisker morphologies, an additional sequence of optical micrographs showing whisker growth as a function of time, and additional images of sheet formation. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2010, 26(4), 2847–2856