Surface Degradation during Separation of Crystals from Solution:Minimizing the Shut-off EffectFieke J. van den Bruele, Kess M. Marks, Bram Harmsen, Alinda L. Alfring, Hannah Sprong,Willem J. P. van Enckevort,* and Elias Vlieg
Radboud University Nijmegen, Institute for Molecules and Materials, Heyendaalseweg 135, 6525 AJ, Nijmegen, The Netherlands
ABSTRACT: Deterioration of crystal surfaces during removalfrom solution prior to observation poses a problem for ex situ(atomic force) microscopy studies. We use potash alum(KAl(SO4)2·12H2O) crystals as a model system and investigatethe variation of the following parameters: size, orientation,lifting rate, humidity, airflow, rinsing, and pollution. A model todescribe the effects occurring during removal is developed andimplications for other systems are presented to minimize thediscrepancy between in situ and ex situ surfaces. Thisdiscrepancy is minimized for potash alum when a crystal islifted slowly from its solution in a vertical position at a humiditybelow room humidity and in the absence of air flow.
1. INTRODUCTIONCrystal growth from solution is an important field of research.A major part of industrial and natural crystallization proceedsfrom solution. The same holds for crystal growth ofbiomolecules, such as proteins. Examples are the crystallizationof simple compounds, like L-ascorbic acid1 and salts,2,3 andmore complex compounds, such as active pharmaceuticalingredients4,5 and proteins.6−8 The circumstances of crystalgrowth can have an influence on properties important forindustry, such as morphology, purity and the growth of theproper polymorph. The surface of a crystal reveals informationabout its growth. The best way to study the surface is in itssolution using microscopy. However, the amount of micro-scopic techniques suited for in situ measurements is limited.Some techniques, such as Scanning Electron Microscopy,cannot be used in solution, whereas others, such as AtomicForce Microscopy9,10 (AFM), Scanning Tunneling Microscopy(STM) and optical microscopy,11,12 have limited possibilities orare quite elaborate when used in solution.Ex situ measurements on the other hand present a problem
as well. During removal of the crystal from the growth solutiona thin film of (super)saturated solution will remain on itssurface, which evaporates relatively fast to give very roughstructures, obscuring the original growth patterns. This effect iscalled the shut-of f ef fect and presents a serious problem wheninvestigating crystal faces for growth features. An exampleshowing the surface degradation of a potash alum crystalsurface by the shut-off effect is presented in Figure 1.To be able to do ex situ measurements, this shut-off effect
needs to be minimized. In literature, various techniques tominimize the difference between in situ and ex situ measure-ments are used. But this issue still remains a point ofdiscussion.10 Essentially, three different approaches to tackle
the problem have been reported in the past. In the first, thecrystal is rinsed and/or pulled through a liquid layer, using adifferent solvent with negligible solubility, for examplehexane.13−15 This gives good results for optical microscopybut does not succeed in creating surfaces suitable for AFM.This technique creates a surface that is rough on a nanometerscale.13 A second approach is rinsing the crystal surface withpure solvent after removal from the growth solution and thendrying it by nitrogen flow.9,10 This method improves the
Received: November 15, 2011Revised: February 28, 2012Published: March 22, 2012
Figure 1. Potash alum crystal removed from solution by hand andsubsequently dried in air. The surface is heavily deteriorated.
surface quality, as it removes artifacts introduced by the shut offeffect. However, as a result of this short dissolution period, it islikely that the patterns are more representative of dissolutionthan of growth. A third approach is removing the adheringsolution after separation from the bulk solution using a jet ofcompressed air or argon. Using this method, Ester et al.10,16
were successful in obtaining the original growth patterns on the{010} faces of potassium hydrogen phthalate (KAP) crystals.As far as we know, there are no general descriptions and/ormethods to minimize the shut-off effect.The shut-off effect is mainly influenced by three parameters.
These are the amount of fluid and solvate remaining on thesurface, the speed of solvent evaporation and dewetting effects.The first determines the amount of material deposited afterseparation, the second influences the tendency for fast 2Dnucleation or kinetic roughening17,18 and the third determinesthe formation of dewetting patterns during evaporation. Theseparameters are influenced by environmental factors. In ourstudy, we investigate the following factors, both from anexperimental and a theoretical point of view: orientation, crystalsize, lifting rate, airflow, relative humidity (RH), rinsing andpollution. As a model compound, we used potassium aluminumsulfate growing from aqueous solution. For comparison of ourex situ surface patterns with in situ AFM observations, we referto Reyhani et al.,19 who showed that in situ the potash alum{111} surfaces have atomically smooth terraces with somewhatragged step edges.We present a descriptive model for the processes during the
removal of a surface from solution and use this to define theimportant parameters for minimization of the shut-off effect forother systems. This leads to a new approach in reducing theshut-off effect.
2. EXPERIMENTAL SECTIONPotash alum (KAl(SO4)2·12H2O, 98% pure from Sigma-Aldrich)crystals were grown from water (ultrapure, 18 MΩ/cm resistance and<3 ppb organic content) solutions by a two-step procedure. First, seedcrystals (approximately 0.5 cm in size) are grown at high super-saturation (7−10% at RT). Then, the seeds are placed in a solution,2% supersaturated at room temperature, which was heated up to 50°C. The crystal first partially dissolves and as the solution cools down,the crystal grows at a slower rate. In this way, crystals of at least 1 cmin size, suitable for further experiments, are obtained.For studying the influence of environmental factors on the shut-off
effect, the setup schematized in Figure 2 is used. First, a specimencrystal is placed with its largest {111} face in a vertical position on aglass filter (G0). This crystal/filter combination is lowered in anaqueous solution (volume 100 mL), saturated at room temperature,but slightly heated (5−10 °C) at the moment the crystal is placed.Then, the solution is allowed to cool down to room temperature andequilibrated for 2−3 h. In this way, artifacts due to the shut off effectduring specimen preparation are removed and a smooth surface isobtained during the equilibration period. To avoid clogging of theglass filter, all experiments were done with saturated solutions.After the growth/equilibration period, the crystal is lifted slowly out
of the solution using the glass filter, which is suspended from a liftingapparatus using thin fishing lines. Controlled pulling rates down to 0.1mm per minute can be realized.To avoid direct contact between the crystal and the hydrophilic
glass filter, the crystal was placed on a small “wad” of aluminum foil.Other holding mechanisms that make no or very small contact withthe crystal surface and have a smooth contact surface will also work.Relative humidity (RH) is controlled and air flow is restricted by
placing the setup in a closed plexiglass box. The RH is monitoredduring the whole experiment. To control the humidity one or severallarge Petri dishes filled with water for high (80−100% RH), potassium
hydroxide pellets for low (0−30% RH) and saturated potash alumsolution for intermediate (60−80% RH) humidity are placed in thebox during the experiments. Air flow during experiments is regulatedby adjusting the size of an opening in the Plexiglas box or by activeblowing with a nitrogen flow.
For rinsing the crystal surface after removal from the solutiondistilled n-heptane (Merck) or distilled chloroform (Fisher) are used.Special care is taken that all solvents are very pure and glasswork wascleaned thoroughly to prevent silicon oil or other pollutions tointerfere with the results. For removal of the crystal through a layer ofinert liquid on top of the growth solution, the same pure heptane wasused.
The surface quality of the large {111} faces is studied using opticaldifferential interference contrast microscopy (DICM) (Leica DM RX)and AFM (Dimension D3100 Digital Instruments) in tapping mode.These faces are the slowest growing ones and therefore are expected toexhibit the clearest step patterns. Observation is always done withintwo hours after the experiment to avoid postremoval effects. Uponrepetitive AFM scanning at room RH (30−50%), the surface does notshow changes.
3. RESULTS AND DISCUSSION
Upon separation of a crystal surface from its solution, severalprocesses take place simultaneously; the model we use todescribe these processes is depicted in Figure 3.
Surface Orientation. Placing some saturated solution ontop of a horizontal {111} face of a potash alum crystal showsthat the liquid contracts to a droplet, regardless of relativehumidity. The contact angle of the droplet is low, about 4 ± 2degrees, corresponding to incomplete wetting. This amplifiesirregular drying patterns on the surface when kept horizontal,because the remaining fluid contracts into droplets, oftenembedding irregularities on the surface. Upon further drying, awealth of dewetting patterns, obscuring the original growthfeatures, are obtained.For a vertical orientation of the surface, on the other hand,
the incomplete wetting causes droplets or a liquid film to“slide” downward due to gravity and surface tension effects.This sliding mechanism was confirmed by a simple experimentin which a potash alum crystal was removed from solution byhand with two {111} surfaces vertical. Careful observationusing a magnifying-glass and keeping the face vertical showed aclosed liquid layer with a sharp boundary on top that slowly
Figure 2. Set-up used for investigating the influence of external factorson the shut-off effect during removal of a crystal from its growthsolution.
receded downward (approximately one to a few mm perminute).In this study, only vertical surfaces are used to examine the
influence of the environmental factors on the shut-off effect.Surface Size. During removal of the crystal, a meniscus will
be present at the contact line between the crystal and itssolution (Figure 3). The height of the meniscus is given by20
= − θ γρ
⎡⎣⎢
⎤⎦⎥h
g(1 sin )
21/2
(1)
with γ the solution surface free energy, ρ the liquid density, gthe gravitational acceleration and θ the liquid-crystal contactangle. For a saturated potash alum solution at 20 °C γ isestimated 0.074 J/m2 using the approach by Marcus21 and ρ =1.05 × 103 kg/m2.22 Using g = 9.8 m/s2 and θ = 4° incombination with the above two values gives a meniscus heightof 3.7 mm, which agrees well with the observed value of 3−4mm. After separation of the crystal from the solution a largepart of the meniscus liquid will remain on the lower part of thecrystal surface leading to an excessive shut-off effect. Therefore,it is essential that the size of the crystal exceeds the meniscusheight. For this reason, we used crystals with sizes larger than10 mm.Lifting Rate. The surface quality of the crystals deteriorates
with lifting speed. Keeping the other conditions constant andvarying the pulling rate we found an optimum value of about 1mm per minute. A further lowering did not improve the results.
This was found by using optical microscopy as well as AFM(Figure 4).
The main reason for the above behavior is that at high liftingrates the liquid has not enough time to flow from the surface. Inorder to estimate this effect, we use the result for completewetting derived by Wilson.23 In that case, the thickness of thefilm is given by the relation:
= μρ − α
− α− α
⎛⎝⎜
⎞⎠⎟
⎧⎨⎩⎫⎬⎭
dUg
Ca
Ca
1
(1 sin )0.9458( )
0.10685cos(1 sin )
1/2
1/21/6
1/2(2)
with μ = viscosity, U = lifting speed and α is the angle betweenthe crystal face and the direction perpendicular to the liquidsurface, which is 90° in our case. Ca is the capillary length,which is defined as
= μγ
CaU
(3)
Using μ = 1.31 × 10−3 kg/m s for the saturated potash alumsolution at 20 °C, the thickness of the adhering layer can becalculated as a function of the lifting speed. The result is givenin Figure 5. For a lifting speed of 1 mm/minute (for which weobtained the best results), the calculated layer thickness isabout 100 nm. Such a layer is expected to flow downwardunder the influence of gravity. Using the relation for thedownward flow rate of a film24
=ρμ
−u zg
sz z( ) ( 0.5 )2(4)
with s the film thickness and z the distance from the solidsurface, we obtain a downward flow rate of the outer layer (z =
Figure 3. Illustration of processes that occur during removal of avertical crystal surface from solution. The thick solid line representsthe lifting rate dependent film pulled down by gravity for a nonfullywetting solution. The dashed line is the resulting humidity-dependentnanofilm thickness. The dotted line represents the film for a fullywetting solution partially pulled down by gravity and evaporating untilthe full surface is covered with an equilibrium film.
Figure 4. (a) Optical DICM image of a potash alum surface removedat 1.5 mm/min at RH < 20%, showing a smooth surface covered withtiny irregularities; (b) AFM image of the same crystal showing aregular pattern of steps, each 0.5 nm high; (c) optical image of apotash alum surface removed at 15 mm/min at RH < 20%, showingsigns of liquid film contraction; (d) AFM image of the same crystalshowing rough deposition of material.
s = 100 nm) equal to 2.4 × 10−3 mm/minute. This value isnegligible compared to the lifting rate of 1 mm/minute. For asaturated solution, a 100 nm solution layer corresponds to agrowth of approximately 16 monomolecular layers of 0.5 nmthickness each after solvent evaporation. This is quitesubstantial and likely gives a dramatic shut-off effect. Theliquid profile for a system with complete wetting is indicated bythe dotted line in Figure 3.Although the downward flowing macroscopic film is quite
thin, bulk hydrodynamics are still applicable. The limit of thisapproach is on the order of 10 (water or hydrated ion)diameters, that is, about 10−20 nm as follows from thetheoretical work by Guo et al.25
For potash alum crystals, the wetting is not complete andtherefore the above model only holds for fast pulling rates. Thepartial wetting causes the liquid film to slowly recededownward, leaving an ultrathin equilibrium water layer (fromhereon called ”nanofilm”). The lifting rate dependent fluid filmis indicated by the solid line and the nanofilm by the dashedline in Figure 3. If the pulling rate is equal or slower than thereceding velocity of the film, only this nanofilm remains on thesurface in contact with the upper end of the meniscus. Thisresults in minimal deposition due to slow solvent evaporation atthe meniscus. Slow lifting (at low RH) gives results that arecomparable to the in situ growth patterns imaged by Reyhani etal. using AFM,19 as shown in Figure 4b.Humidity. The equilibrium thickness of the nanofilm will
depend on the RH. For NaCl (100), the thickness was reportedto range from one water monolayer at low RH to several tens ofmonolayer near the deliquescence RH.26,27 In our case, thethickness of the nanofilm varies over the surface. Very near thecontact line of the liquid meniscus and the solid, the RH is notfar from the deliquescence value (which is 90% RH for potashalum) and at the top of the face the film thickness approachesthe equilibrium value for the applied RH.Using in situ DICM to observe a (111) potash alum surface
in a closed cell with 80−100% RH revealed no surface details.However, after opening, the cell and exposure to lower roomRH revealed rapid formation of a shallow dendrite pattern onthe crystal surface, indicating a rapid evaporation of a solutionnanofilm. Interesting is the observation that even under thishigh RH condition a droplet of saturated solution on the crystalsurface does not spread. This indicates that a nanofilm coversthe surface, but the rest of the solution is gathered in droplets,
preferentially embedding irregularities on the surface. Theproperty of a liquid unable to spread on its own adsorbednanofilm is known as “autophobicity”.28
Macroscopic film and droplet formation can be avoided byslowly lifting the crystal from its mother liquor as explainedbefore. However, the thickness of the remaining nanofilm islargely controlled by the relative humidity. Pulling the crystal atdifferent RH and subsequent transfer from the humiditycontrolled box to room conditions for inspection gives differentdewetting patterns (Figure 6). Lifting at 60% RH results in a
spinodal dewetting pattern, which is characteristic for anultrathin film.29 At 78% RH, we obtained a shallow dendritepattern, indicating a thicker nanofilm. We think that thesedewetting structures result from the sudden decrease in RH toroom values during removal of the crystal from its humiditycontrolled environment. This leads to a decrease in equilibriumnanofilm thickness and subsequent fast growth of the last fewlayers.The best results are obtained for a RH < 20%, where a clean
pattern of steps is obtained (Figure 4b), with minimal shut-offeffect. We think that at this RH the nanofilm is reduced to onemonolayer in a similar way as found for (100) NaCl.26 Duringthe slow lifting, the RH at a given surface location reducesslowly from a value not far from deliquescence near themeniscus contact line to a much lower value. This leads to avery slow, controlled evaporation of the nanofilm to minimalthickness. Transfer to room RH leads to a slight increase innanofilm thickness, resulting in slight etching of the surface,presumably less than a monolayer. This dissolves the minimalamount of material deposited through evaporation occurring atand near the meniscus during lifting.
Airflow. Airflow is avoided in our setup (Figure 2) as thisstrongly increases the evaporation of the adhering liquid andthus the postgrowth deposition rate. Simple tests confirm thatthe presence of slight air flow leads to serious degradation ofthe surface patterns.Following the positive results on KAP crystals reported in
literature,10,16 removal of solution by strong gas flow wasinvestigated as well. In this case we subjected the surface of analum crystal to a jet of compressed nitrogen immediately after
Figure 5. Liquid layer thickness versus pulling rate assuming completewetting.
Figure 6. Optical DICM images of alum surfaces examined at roomhumidity (30−50%) (a) removed at 60% RH showing spinodaldewetting patterns and (b) removed at 78% RH showing dendriticdewetting patterns.
(rapid) separation from its solution. The optical surface isindeed of high quality, although some patterns characteristic offast growth can be recognized, such as relatively steep growthhillocks and droplet tracks. AFM measurements, however, showa kinetically roughened surface (Figure 7) and no step patternscan be distinguished anymore on this surface.
Rinsing. Rinsing the crystal after separation from its growthsolution using a liquid that is immiscible with water and doesnot dissolve existing surface structures is a common method toreduce the shut-off effect prior to optical examination of acrystal surface.13−15 To investigate the usefulness of rinsing forAFM observation, we applied rinsing after lifting a crystal fromsolution (1.5 mm/min) at high (75%) RH, using a flow ofheptane or chloroform for a period of a few seconds. Thisresulted in smooth surfaces suitable for optical DICMmicroscopy (Figure 8a), but the AFM images show that thesurface has recognizable steps partially covered by a thindendritic layer (Figure 8b). Both the results in literature,13−15
where the crystal was quickly removed from the solution andrinsed, and our observation indicate that rinsing is capable ofremoving part of the macroscopic layer and droplets from thesurface but leaves the nanofilm intact.
Assuming a laminar boundary layer during rinsing, a roughestimation of the liquid flow velocity u(z) in this layer as afunction of the distance z from the crystal surface is given by24
=δ
−δ∞ ⎜ ⎟ ⎜ ⎟
⎛⎝
⎞⎠
⎛⎝
⎞⎠
u zu
z z( ) 32
12
3
(5)
with δ the boundary layer thickness and u∞ the bulk flow rate.Neglecting the second term of this equation, it follows that thetime needed for removal of a solution sublayer at height z ≪ δis given by
= ≅ δ∞t
Lu z
Lu z( )
23 (6)
where L is the size of the crystal surface rinsed. Using δ = 10−5
to 10−4 m, u∞ = 10−1 m/s and L = 10−2 m, one obtains for asublayer of height of z = 100 nm and a removal time of 7−70 sfor δ = 10−5 to 10−4 m respectively. In this and earlierinvestigations15 rinsing times of approximately five secondswere used. Due to the retarded flow near the crystal surface,these short rinsing times are incapable of removing the lowestfew tens of nanometers of a liquid layer on top of a crystal. Thisexplains our observations and those reported in theliterature13−15 of good optical surface quality but roughsurfaces found using AFM.In our case, only a nanofilm and no macroscopic layer was
present as a consequence of the slow lifting speed. Here, afterrinsing, the final ex situ surface was still covered with a partiallayer of disordered material, indicating fast precipitation, shownin Figure 8b. This again shows why the rinsing technique is notsuitable for AFM studies.We also investigated the removal of the mother liquid by
pulling the crystal surface slowly through a heptane layer on topof the growth solution. The main part of the solution isremoved in this way but a laminar layer of water still remains onthe potash alum surface. The evaporation of this layer results insimilar structures as found for rinsing with heptane.
Pollution. A somewhat more trivial, but still very importantpoint is the absence of “pollution” in the solution. Pollution bysmall particles as well as crystal defects and rough surfacestructures cause protrusions at the surface, which act as pinningpoints for the receding liquid film. In this way, solution islocally collected, often leading to trapped, isolated droplets.This gives rise to uncontrolled local growth and strongdewetting effects as shown in Figure 9.
4. EVALUATION AND GENERALIZATION
During separation of a crystal from its solution a liquid film of100 nm to 10 μm thickness (depending on the lifting rate)remains on its surface. In the case of complete wetting (contactangle θ = 0) this film flows downward by gravity at anextremely slow pace. As this film stays adhered for a longperiod, solvent evaporation results in a substantial post growthdeposition, which leads to a shut-off effect. This makes thesurface unsuitable for ex-situ microscopic examination.However, if the wetting is not complete (θ > 0), then thedownward flow of the liquid layer is assisted by capillary effects.This leads to a sharp boundary between the film and anequilibrium nanofilm on the surface, which recedes downward.If the lifting rate of the crystal from its solution is slower thanthe retraction rate of the boundary line, then only an ultrathinequilibrium nanofilm is left on its surface.
Figure 7. AFM image of a kinetically roughened potash alum surfaceobtained after rapid removal of the adhering solution layer by using ajet of compressed nitrogen.
Figure 8. (a) Optical image of a growth hillock on an alum surfaceremoved at 75% RH and rinsed with heptane. (b) AFM image of analum surface removed at 75% RH and rinsed with heptane showingdisordered precipitates.
The small contact angle angle of θ = 4 ± 2° indicates that thewetting of the {111} faces of potash alum is not complete, butstill considerable. Therefore, the adhering solution layer recedesslowly. Less wetting makes the liquid film (or droplets) easierto retract, for instance assisted by a jet of compressed air asdemonstrated for (010) KAP.13,16 As the (010) face of thiscrystal is terminated by benzene rings and thus is hydro-phobic,30 we expect that for KAP the wetting of the adheringsolution is less.The thickness of the remaining nanofilm is determined by
the RH and can range from one monolayer at low RH toseveral tens of solution monolayers close to the deliquescencevalue. To obtain a nanofilm as thin as possible, the RH duringthe separation of the crystal should be kept as low as possible.In our case, the best results were obtained by lifting the crystalat RH < 20%. After separation of the crystal at low RH andtransferring it to room conditions with higher RH, one or a fewmonolayers of water are attracted to the surface. This dissolves(a part of) the small, rough structures, formed during theremoval process. Transfer from high RH to room RH, however,leads to rapid partial evaporation of the nanofilm, which resultsin spinodal or dendritic dewetting patterns.Implications for Other Systems. Our specific approach
cannot directly be applied to other systems, but the modeldescription may help to obtain better ex situ surfaces for othercases as well. The following aspects for other systems arediscussed: solvent, wetting properties, RH, concentration andheterogeneous systems.For solvents other than water, the room vapor pressure is
always 0%, which leads to evaporation of the adhering nanofilmupon transfer of the specimen to room conditions. Dependingon the vapor pressure during lifting, this creates dewettingpatterns or causes kinetic roughening on the crystal surface,which can be substantial if the evaporation rate of the liquid ishigh, which is common for many organic solvents. The vaporpressure should therefore be kept as low as possible duringlifting.Surface and interfacial energies of the crystal-solution system
are other factors that govern the formation of dewettingpatterns during and after separation of the crystal from thegrowth system. As stated before, complete wetting leads to adetrimental shut-off effect. Using a liquid that poorly wets thecrystal surface gives good results as the liquid layer retracts
faster during pulling and, generally, the remaining equilibriumfilm is thinner. Using a less viscous solvent is also helpful, as itcreates a thinner film, which also recedes faster during pulling.The solute concentration of the mother liquid is very high in
our model system. This makes it difficult to obtain a cleansurface suitable for ex-situ observation. For many crystal growthsystems the solute concentration is much less, which impliesthat the nanofilm formed after separation from the solutionholds not enough material to create a monolayer. Dependingon the size of a growth unit, concentrations of 10−3 M and lesswill cause less than 0.1 monolayer deposition and therefore willnot extensively disturb or change the original growth patterns.For heterogeneous surface deposition, for example, mono-
layer growth of organic materials on substrates,31 theconcentrations used are often low. In this case, the mainfocus during removal will be on the prevention of rearrange-ment of the material. This can be done by slowly lifting at lowRH, in a similar way as for homogeneous deposition. Quickremoval and subsequent equilibration under 100% vaporpressure is another option,30 if examination under this highvapor pressure is possible. In a number of cases, only amonolayer is desired. If the monolayer is strongly adhering tothe substrate, rinsing using the same solvent as used fordeposition will remove the extra layers.
5. CONCLUSIONSIn this paper we have studied the factors that cause thedegradation of the surface patterns during separation of acrystal from its growth solution, prior to examination bymicroscopy. Potash alum grown from aqueous solution wasused as a model system. The best results were obtained forvertical orientation using a slow lifting rate of 1 mm/minute ata RH < 20%, which is lower than room RH, and the absence ofairflow. In this way, clear patterns of steps, each 0.5 nm high,with atomically smooth terraces in between could be imaged byAFM. A model describing the process during removal andsubsequent drying is proposed.It is not trivial to obtain an immaculate surface for ex situ
microscopic observation for crystals grown from solution.However, the difference between in situ and ex situ surfaceobservation can be minimized to a large extent. Based on ourmodel the following environmental factors are recommended ifapplicable for the system: low lifting rates, poor wetting, a lowsolvent viscosity and a low solvent RH. The use of a solventwith a low saturation concentration also decreases the shut-offeffect.
NotesThe authors declare no competing financial interest.
■ REFERENCES(1) Srinivasan, K.; Devi, K. V. Cryst. Res. Technol. 2010, 45, 946−952.(2) Kowacz, M.; Prieto, M.; Putnis, A. Geochim. Cosmochim. Acta2010, 74, 469−481.(3) Mullin, J. W. Crystallization, Fourth ed.; Butterworth-Heinemann:Oxford, 2001; pp 315−402.(4) Alvarez, A. J.; Myerson, A. S. Cryst. Growth Des. 2010, 10, 2219−2228.(5) Kratochvil, B. Chem. Listy 2010, 104, 823−830.(6) Vekilov, P. G. Nanoscale 2010, 2, 2346−2357.
Figure 9. Optical DICM image of a potash alum surface showingaccumulation of material and dewetting effects around a pinning point.
(7) Verma, S.; Shlichta, P. J. Prog. Cryst. Growth Charact. Mater. 2008,54, 1−120.(8) McPherson, A. Crystallization of biological macromolecules; ColdSpring Harbor Laboratory Press: New York, 1999.(9) Cheng, M.; Li, M. W.; Guo, J. L.; Cao, Y. C. Sci. China: Technol.Sci. 2010, 53, 1554−1561.(10) Yang, S. F.; Su, G. B.; Tang, J. B.; Mao, W. J.; Wu, M.; Li, Z. D.J. Cryst. Growth 1999, 203, 425−433.(11) Tsukamoto, K. J. Cryst. Growth 1983, 61, 199−209.(12) Maruyama, M.; Tsukamoto, K.; Sazaki, G.; Nishimura, Y.;Vekilov, P. G. Cryst. Growth Des. 2009, 9, 127−135.(13) Ester, G. R.; Price, R.; Halfpenny, P. J. J. Cryst. Growth 1997,182, 95−102.(14) van Enckevort, W. J. P.; van der Linden, W. H. J. Cryst. Growth1979, 47, 196−202.(15) van Enckevort, W. J. P.; Bennema, P.; van der Linden, W. H. Z.Physik. Chem. 1981, 124, 171.(16) Price, R.; Ester, G. R.; Halfpenny, P. J. Proc. R. Soc. A 1991, 455,4117−4130.(17) Kotrla, M. Czech. J. Phys. 1992, 42, 449−544.(18) van Veenendaal, E.; van Hoof, P.; van Suchtelen, J.; vanEnckevort, W. J. P.; Bennema, P. Surf. Sci. 1998, 417, 121−138.(19) Reyhani, M. M.; Freij, S.; Parkinson, G. M. J. Cryst. Growth1999, 198, 258−263.(20) Adamson, A. W. Physical Chemistry of Surfaces; John Wiley andSons: New York, 1990; p 390.(21) Marcus, Y. J. Chem. Eng. Data 2010, 55, 3641−3644.(22) Kubota, N.; Shimizu, K.; Itagaki, H. J. Cryst. Growth 1985, 73,359−363.(23) Wilson, S. D. R. J. Eng. Math. 1982, 16, 209−221.(24) Coulson, J. M.; Richardson, J. F. R. Chemical Engineering, 2nded; Pergamon Press: Oxford, 1964; Vol. 1.(25) Guo, Z.; Zhao, T. S.; Shi, Y. Phys. Fluids 2006, 18, 067107.(26) Arsic, J.; Kaminski, D. M.; Radenovic, N.; Poodt, P.;Graswinckel, W. S.; Cuppen, H. M.; Vlieg, E. J. Chem. Phys. 2004,120, 9720−9724.(27) Ewing, G. E. In: Intermolecular Forces and Clusters II; Wales, D.,Ed.; Springer: Berlin, 2005; Structure and Bonding, Vol. 116, pp 1−25.(28) Hare, E. F.; Zisman, W. A. J. Phys. Chem. 1955, 59, 335−340.(29) Herminghaus, S.; Jacobs, K.; Mecke, K.; Bischof, J.; Fery, A.;Ibn-Elhaj, M.; Schlagowski, S. Science 1998, 282, 916−919.(30) van den Bruele, F. J.; Elemans, J.; Rowan, A. E.; van Enckevort,W. J. P.; Vlieg, E. Langmuir 2010, 26, 498−503.(31) Schreiber, F. Prog. Surf. Sci. 2000, 65, 151−256.
