We present a protocol for a facile orientation specific deposition of plate-like mesoporous SBA-15 silica particles onto a surface (mesopores oriented normal to surface). A drop of an aqueous dispersion of particles is placed on the surface and water vaporizes under controlled relative humidity. Three requirements are essential for uniform coverage: particle dispersion should not contain aggregates, a weak attraction between particles and surface is needed, and evaporation rate should be low. Aggregates are removed by stirring/sonication. Weak attraction is realized by introducing cationic groups to the surface. Insight into the mechanisms of the so-called coffee stain effect is also provided.

By depositing mesoporous particles on a surface, one can increase the effective surface area by a factor of a hundred and at the same time achieve a steric molecular selectivity. There is also an increased possibility of introducing a chemical functionality. For a mesoporous 2D hexagonal structure, it is important to control the orientation of the pores relative to the surface. If the mesoporous material is grown directly on the surface, one typically obtains a parallel orientation of the pores.1,2 These pores are not directly accessible from the solution. To obtain a pore orientation perpendicular to the surface, we have pursued the alternative strategy of adsorbing prefabricated, already calcined, particles from the solution. By making the particles in a separate synthesis, one obtains a greater flexibility with respect to chemical modification and control of particle morphology, presence or absence of micropores,3,4 and improved structural integrity from calcination, all at the cost of introducing a deposition step.

There are a number of techniques that can be used to deposit particles on a surface. Using a dispersion of plate-like mesoporous particles of micron size, we have tested several possibilities. The set targets have been (i) only a monolayer of particles, (ii) particles oriented with the pore direction perpendicular to the surface, (iii) uniform distribution of particles, and (iv) high particle density. In exploratory tests, both dip and spin coating gave unsatisfactory results, while the simple method of placing a drop of the suspension on the surface5–7 provided promising results. We are yet to explore other methods, frequently used for depositing zeolites, such as the Langmuir-Blodgett technique and layer-by-layer deposition. However, the drop deposition method requires no additives so it is cheaper and readily scalable. In a scientific context, studying the drying of individual drops is to be preferred relative to the industrially more relevant method of spray coating. When a drop of a suspension dries, particles are often deposited in complex patterns. The most noted example is the so-called “coffee stain” effect.7–9 It takes its name from the familiar dark ring that forms at the periphery of a dried stain of coffee. The ring deposit is due to a transport of fluid from the center of the drop to the periphery. This occurs when the drop dries at a constant radius.

In the present paper, we present a method of depositing mesoporous SBA-15 particles from a drop in a way that suppresses the formation of a “coffee stain” ring as well as other inhomogeneous patterns. This is achieved by controlling the interaction between particles and surface as well as the evaporation rate. The latter is accomplished by letting the drying occur under controlled but varied relative humidity. Apart from providing a promising deposition method, the study also reveals some basic mechanisms behind the “coffee stain” effect.

The synthesis follows a protocol published by Linton and Alfredsson10 Pluronic P104 (0.48 g) was dissolved in 1.6 M HCl (18.75 g) in a Teflon-lined polypropylene bottle with a screw cap. The sealed bottle was placed in a water bath and heated to 55 °C. The solution was stirred continuously at 250 rpm with a magnetic stirrer. Tetramethyl orthosilicate (TMOS) (0.715 ml) was added and the stirring temporarily increased to 500 rpm for 1 min. The sealed bottle was kept at 55 °C for 24 h and then transferred to an oven and kept at 80 °C under static conditions. This was followed by filtration, washing of the material with water, and then drying in ambient atmosphere. The organic polymer was burned off by calcination at 500 °C in air for 6 h. This synthesis protocol produces particles with platelet morphology (height 300 nm and width 800 nm) with the primary mesopores running perpendicular to the short axis.

All surfaces used in this work were purchased from Linköping University (Department of Chemistry, IFM, Linköping University) and SWI (Semiconductor Wafer, Inc., Taiwan). They were prepared from polished silicon wafers (p-type, boron-doped, resistivity of 1-20 Ω cm) and thermally oxidized at approximately 900 °C in an oxidation furnace to yield a silicon oxide layer of roughly 300 Å (i.e., a silica surface). The surface contains approximately 5 silanol groups per square nanometer and can thus be easily modified.11 The surfaces are extremely flat, being routinely used as surfaces for ellipsometry measurements.

Strips of the surfaces with a width of approximately 10 mm were cut and then cleaned in a base mixture of 25% NH4OH (Merck), 30% H2O2 (Honeywell), and H2O at volume ratio of 1:1:5 for 5 min at 80 °C followed by cleaning in an acid mixture of 37% HCl (Merck), 30% H2O2, and H2O at volume ratio of 1:1:5 for 5 min at 80 °C. The cleaned surfaces were rinsed in water and stored in 99% ethanol until further use.

The cationic silica surfaces were prepared by liquid phase silanization of the virgin silica surfaces12 described above. The silica surfaces were dried in a flow of nitrogen gas and then plasma cleaned (Harrick Scientific Corp., model PDC-3XG) for 5 min at 0.04 millibars to remove any organic contaminants, followed by drying in an oven at 150 °C for 30 min to remove any moisture. The surfaces were incubated in anhydrous toluene with 2% 3-aminopropyltriethoxy silane (APTES, Fluka) under a nitrogen atmosphere for 2 h. The modified surfaces were then sonicated first in toluene, then in 1:1 toluene/ethanol mixture and finally in ethanol to remove any unreacted species. Finally, the surfaces were dried at 120 °C and then stored in 99% ethanol until further use.

Dry and calcined SBA-15 powder was mixed with deionized water to a concentration of 0.025 wt. %. The dispersion was stirred with a magnetic stirrer for roughly 24 h and then sonicated for various lengths of time (generally 4 h). It was found that 4 h of sonication resulted in depositions with substantially fewer aggregates. Right before the deposition experiment, the dispersion was shaken manually to avoid any sedimentation of material. The isoelectric point of silica is around pH 2 and the particles will hence have an inherent negative charge in the dispersions.13 See supplementary material for SEM micrographs demonstrating the influence of sonication on the dispersion (Fig. S1)14 and the small-angle x-ray diffractograms of the material before and after the sonication process (Fig. S2).14 

A custom built chamber (see supplementary material for schematic drawing of the chamber (Fig. S3)14) was used to achieve an environment with controlled relative humidity. The chamber consists of a metallic cylinder with inner diameter of 58 mm and height of 70 mm. A plastic lid, with a small opening fitted with a cork, covers the cylinder. A HumiSys humidity generator (HumiSys LF, InstruQuest, Inc., USA) registers humidity and temperature via a probe inside the chamber and maintains the relative humidity by injecting humid air into the chamber.15 

Prior to depositing the drop, the surface (∼1 cm2) was placed inside the chamber and the relative humidity (RH) was adjusted to the desired value (10%-90%). When a stable RH level was reached, the cork in the lid was removed, a small drop of dispersion (5 μl) was carefully placed on the surface and then finally the chamber was sealed using the cork. This procedure took approximately 30 s. The deposition process momentarily alters the RH but the set value is restored within a few minutes. The sample was left until complete evaporation had occurred, as determined by visual inspection. Complete evaporation took between 20 min and several hours depending on the value of the RH.

The deposits of SBA-15 particles on the surfaces were investigated by scanning electron microscopy. Each surface was attached to a sample holder and sputter coated with Au/Pd to reduce charging. The micrographs were recorded with a JEOL JSM-6700 microscope operating at 10 kV.

Small angle x-ray diffraction (SAXD) measurements of the calcined material were performed on a GANESHA SAXS system (SAXSLAB, Denmark) with a configuration yielding a q-range of 0.06-0.26 Å−1. The powder samples were placed in quartz capillaries sealed with wax. See supplementary material for small-angle x-ray diffractograms (Fig. S2).14 

Drops of aqueous dispersions of mesoporous silica SBA-15 were placed on a surface. When the water had evaporated, the particles form a deposit on the surface. The deposition patterns were analyzed by SEM. The dispersion contained 0.025% w/w of particles. The drop volume was 5 μl and the drop formed was studied for both types of surfaces in a small but finite contact angle. The resulting drop radius was between 1.5 and 2 mm depending on the surface substrate. The total number of particles per drop was approximately 1 × 107 and, provided a monolayer is formed, this covers an area of around 5 × 10−6 m2, which corresponds to approximately 60% of the initial area of the drop-surface interface.

Two qualitatively different substrate surfaces were investigated; cationic and the native silicon oxide. Some features were common for the deposits in the two situations. Particles deposit in a monolayer configuration, with the exception of a few cases of small three-dimensional aggregates on the surface. The occurrence of these did not depend on deposition conditions and we conclude that they are due to a small population of such oligomeric aggregates in the original dispersion. Another case is where multilayer adsorption is found in the “coffee stain” rings, where the average particle coverage exceeds the available area. Particles deposit in a clear majority of cases with the flat side facing the underlying surface. In some cases, we observe 2D-aggregates on the surface and these are likely formed on the surface rather than in the bulk.

In spite of the local similarities between the deposits on the two types of surfaces, there were distinct differences in the over-all deposit patterns. Fig. 1 shows SEM micrographs of the deposits onto the virgin silica surface (see supplementary material for enlarged version of Fig. 1 (Fig. S4)14). The main pictures show a low magnification image over the whole deposit area, while the insets show a higher magnification picture of a local patch. For Fig. 1(a) the RH was 10% while for Fig. 1(b) RH = 90%. Both samples show a similar deposition pattern with areas of high particle density separated by areas of low particle density. The images shown in Fig. 1 are representative for all deposition experiments irrespective of the relative humidity during evaporation.

FIG. 1.

Low magnification SEM micrographs of plate-like particles of SBA-15 deposited on a hydrophilic silica surface at (a) 10% RH and (b) 90% RH. The deposition is non-uniform, clearly seen from the higher magnification SEM micrographs shown in the respective insets. Scale bars in the insets are 20 μm and 10 μm for (a) and (b), respectively.

FIG. 1.

Low magnification SEM micrographs of plate-like particles of SBA-15 deposited on a hydrophilic silica surface at (a) 10% RH and (b) 90% RH. The deposition is non-uniform, clearly seen from the higher magnification SEM micrographs shown in the respective insets. Scale bars in the insets are 20 μm and 10 μm for (a) and (b), respectively.

Close modal

For the cationic surface, the deposition pattern depends systematically on the value of the relative humidity in the range 10%–90%. The left column of Fig. 2, images (a)-(e), shows the overall pattern, while the right column, images (f)-(j), illustrates the deposition at the periphery of the original drop (see supplementary material for enlarged version of Fig. 2 (Fig. S5)14). The general feature is that away from the periphery the deposition is relatively uniform, while at the rim a clear “coffee stain” pattern is apparent in the range RH = 10%–70%. The width of the ring decreases with increasing RH, while the radius remains constant within experimental uncertainty. Fig. 3 summarizes how the width of the ring deposit, the diameter of the ring, and the time of evaporation depend on the value of the relative humidity. One can also note that the ring has a wedge shape with more multilayer deposition in the direction towards the center of the drop.16 

FIG. 2.

Left column: Low magnification SEM micrographs of plate-like particles of SBA-15 deposited on a silica surface modified with cationic surface groups at (a) 10% RH, (b) 30% RH, (c) 50% RH, (d) 70% RH, and (e) 90% RH. Right column: Higher magnification SEM micrographs at the point/position of the pinning of the three-phase contact line for samples prepared at (f) 10% RH, (g) 30% RH, (h) 50% RH, (i) 70% RH, and (j) 90% RH.

FIG. 2.

Left column: Low magnification SEM micrographs of plate-like particles of SBA-15 deposited on a silica surface modified with cationic surface groups at (a) 10% RH, (b) 30% RH, (c) 50% RH, (d) 70% RH, and (e) 90% RH. Right column: Higher magnification SEM micrographs at the point/position of the pinning of the three-phase contact line for samples prepared at (f) 10% RH, (g) 30% RH, (h) 50% RH, (i) 70% RH, and (j) 90% RH.

Close modal
FIG. 3.

Graphs showing the variation in the three parameters (a) ring width, (b) ring diameter, and (c) time of evaporation with increasing relative humidity during deposition of plate-like particles of SBA-15 on a silica surface modified with cationic surface groups.

FIG. 3.

Graphs showing the variation in the three parameters (a) ring width, (b) ring diameter, and (c) time of evaporation with increasing relative humidity during deposition of plate-like particles of SBA-15 on a silica surface modified with cationic surface groups.

Close modal

Our results show that drop deposition is a simple method to control deposition of anisotropic particles without requiring additives or further chemical processing. We note that although closepacked monolayers have been produced for materials such as zeolite microcrystals, covalent bonds between organic species added to the microcrystal surface after synthesis are responsible for the extended network formed.17,18 Obtaining uniform closepacked layers of bare anisotropic inorganic particles remains a challenge. Our early experiments on this system, where positively charged surfaces were suspended in the solution of dispersed particles (similar to layer-by-layer deposition conditions) also showed that this method did not result in a well oriented close-packed monolayer. Thus for such particles, particle geometry, flow, and electrostatic interactions are all essential factors to control orientation and distribution.

Depending on deposition conditions, we observe three qualitatively different patterns for the deposit. For the virgin surface, at all relative humidities, particles deposit in an inhomogeneous irregular pattern. As seen in Figs. 1(a) and 1(b) the details of the irregularity are influenced by the relative humidity and the deposition is more homogeneous at 90% RH than at 10% RH. There are traces of a “coffee stain” deposit at the original ring of the drop, but there are also regions with high deposit density at more central locations. For the cationic surface, we observe at low relative humidities, see Figs. 2(a) and 2(f), a distinct “coffee stain” band located at the rim of the originally deposited drop. As the relative humidity is increased, the width of this drop decreases and at 90% RH the coffee stain effect has disappeared altogether.

We discuss these three limiting cases by starting with the observations for the cationic surface at 10% RH. As the water evaporates, the volume of the drop decreases and for a perfectly homogeneous system, the drop shape should be determined by the surface free energies and, to some extent, the effect of gravity. In such a case, the drop radius should decrease with time. The extent of particle deposition during this time would depend on the effective particle–surface interaction, which will be more attractive as the particle concentration increases. The observation of a distinct “coffee stain” band in Fig. 2(f) shows a clear deviation from this “idealized” behavior. As is also clear from visual observations during the drying process, the three phase contact line at the rim of the drop is pinned, so that drying only affects the thickness of the drop, which becomes gradually more uniform. The most likely mechanism for the pinning is an initial rapid adhesion/deposition of particles onto the surface due to electrostatic attraction between the positively charged surface and the negatively charged particles. This initial presumably diffusion controlled deposition should occur uniformly over the surface under the drop. As evaporation leads to a decrease in volume, there is a surface tension driven force on the three phase contact line. Particles on the surface create a restoring force, tending to keep the contact line at a fixed position and with sufficient line density of particles the contact line is pinned. As evaporation continues, there is a net liquid flow from the center to the periphery of the drop. This flow brings a transport of particles towards the rim and the deposition rate is higher at the rim than away from the contact line. One contribution to the outwards flow has a geometrical origin. If gravity effects are neglected, the shape of the drop on a planar surface corresponds to a spherical cap. With the contact line pinned during evaporation, this cap will correspond to a sphere with increasing radius. Thus, the height at the center will decrease more rapidly than the height in the periphery. If one assumes that the evaporation rate per area is constant across the drop, the surface tension will give rise to an outwards flow. This flow will be even more pronounced if the evaporation rate is higher at the periphery than in the center, which is a likely scenario when the rate determining step in the evaporation is due to heat transfer as is normally the case for aqueous systems.19 Thus, for the case shown in Figs. 2(a) and 2(f), one has an initial deposition of particles on the surface, which results in a pinning of the contact line. As the evaporation continues, the pinning is strong enough to keep the contact line at its original position. An outward liquid flow leads to an increased particle deposition close to the contact line resulting in the “coffee stain” ring combined with a rather uniform particle deposition over the remaining surface.

The second case we discuss is the deposition of particles on the virgin silica surfaces shown in Fig. 1. Here, the electrostatic interaction between particle–surface is repulsive since surface and particles are both negatively charged. The particles thus adsorb less readily on the surface. We observe a weak “coffee stain” ring at the original position of the three phase contact line. However, the images show that the contact line can only have been pinned for a short period. After some time, the surface tension force is strong enough to pull the contact line away from the original position. It appears, particularly from Fig. 2(a) that there is a stick slip behavior of the movement of the contact line. From the initial position it tends to retreat to the equilibrium position given by the current volume (and the surface free energy). When the evaporation continues, a tension builds up and there is a new slip stage. However, this process is influenced by local inhomogeneities/defects on the surface leading to a loss of the circular symmetry in the deposition.

Finally, the third case is the cationic surface, where the evaporation occurs at a high relative humidity where no “coffee stain” ring is observed (Fig. 2(j)). The particles are deposited with uniform density over the whole area originally covered by the drop. However, we conclude that the initial particle deposition is large enough to pin the contact line during the whole evaporation period. This conclusion is based on the trend shown in Figs. 2(f)-2(j) on the observed uniform deposition as well as on visual observations during the evaporation process. So what explains the difference between the cases of 10% RH and 90% RH? The time of evaporation changes by a factor of 50 from 5 min to 250 min. Assuming other factors are unchanged, this indicates that the flow rate within the drops differs by a factor on the order of 50. The particle transport driven by liquid flow is opposed by the diffusional flow of the particles. We thus suggest that at 90% RH the particle transport from liquid flow induced by evaporation is fully balanced by the diffusional liquid flow. When this is the case, there is a uniform particle density in the liquid and this leads to a uniform deposition density on the surface.

The effectively uniform particle density on the surface combined with the observation that particles adsorb with a uniform orientation indicates that the surface adsorption occurs under reversible conditions. One important factor is that although there is, for the cationic surface, an electrostatic attraction between surface–particle the average surface charge density decreases as the negatively charged particles adsorb. There is no extra electrolyte added to the aqueous solvent resulting in a long Debye length, so that an adsorbed particle effectively excludes a large area on the surface. Thus, one could expect that an initial fast diffusion-controlled deposition is followed by slower adsorption. Another effect of the long Debye length is that for initially adsorbed particles there can be a solvent “cushion” allowing for a reorientation to the most favorable direction relative to the surface.

We have presented a method for uniform orientation specific surface deposition of plate-like mesoporous silica SBA-15 particles. We have reached a surface coverage of about 10% (this value was obtained from evaluating SEM micrographs and only counting the particles that have a perfect orientation on the surface, i.e., this is a minimum value for the coverage), and we foresee that it is possible to reach higher densities by further optimization. The uniform coverage is obtained by combining an initially attractive particle–surface interaction with having a controlled evaporation rate. It is then possible to maintain internal (near) equilibrium distribution of particles in the solution so that the deposition rate becomes uniform across the surface. We also report deposition experiments with high evaporation rates for which a distinct “coffee stain” ring is observed. When the deposition is made on a virgin, slightly negatively charged, silica surface, there is a repulsive component to the particle–surface interaction having the consequence that the contact line is not pinned. It moves during evaporation in a stick-slip manner that is perturbed by the presence of surface defects. The end result is a clearly non-uniform particle deposition.

In addition to a protocol for controlled particle deposition, the experiments also provide insights into the mechanisms leading the formation of a coffee stain ring. For example, in Figs. 2 and 3 we present unique data on how the presence and the width of the coffee stain ring depend on the evaporation rate. We also show, by comparing Figs. 1, 2, and 3, how the pinning of the contact line is essential for the observation of the ring deposit. This pinning is furthermore controlled by the characteristics of the surface–particle interaction.

We thank Bernard Cabane, ESPCI, ParisTech., for insightful discussions. Financial support by the Swedish Foundation for Strategic Research, the Swedish Research Council through the Linnaeus Grant “Organizing Molecular Matter,” and an individual project grant from the Swedish Research Council (V.A.)

1.
D. A.
Doshi
,
A.
Gibaud
,
V.
Goletto
,
M.
Lu
,
H.
Gerung
,
B.
Ocko
,
S. M.
Han
, and
C. J.
Brinker
,
J. Am. Chem. Soc.
125
,
11646
(
2003
).
2.
D.
Grosso
,
F.
Cagnal
,
G. J. de A. A.
Soler-Illia
,
E. L.
Crepaldi
,
H.
Amenitsch
,
A.
Brunet-Bruneau
,
A.
Bourgeois
, and
C.
Sanchez
,
Adv. Func. Mater
14
,
309
(
2004
).
3.
T.
Kjellman
,
N.
Reichhardt
,
M.
Sakeye
,
J.-H.
Smått
,
M.
Lindén
, and
V.
Alfredsson
,
Chem. Mater.
25
,
1989
(
2013
).
4.
N.
Reichhardt
,
T.
Kjellman
,
M.
Sakeye
,
F.
Paulsen
,
J.-H.
Smått
,
M.
Lindén
, and
V.
Alfredsson
,
Chem. Mater.
23
,
3400
(
2011
).
5.
L.
Malfatti
,
Y.
Tokudome
,
K.
Okada
,
S.
Yagi
,
M.
Takahashi
, and
P.
Innocenzi
,
Microporous Mesoporous Mater.
163
,
356
(
2012
).
6.
R. D.
Deegan
,
Phys. Rev. E
61
,
475
(
2000
).
7.
R. D.
Deegan
,
O.
Bakajin
,
T. F.
Dupont
,
G.
Huber
,
S. R.
Nagel
, and
T. A.
Witten
,
Nature
389
,
827
(
1997
).
8.
R. D.
Deegan
,
O.
Bakajin
,
T. F.
Dupont
,
G.
Huber
,
S. R.
Nagel
, and
T. A.
Witten
,
Phys. Rev. E
62
,
756
(
2000
).
9.
G.
Berteloot
,
A.
Hoang
,
A.
Daerr
,
H. P.
Kavehpour
,
F.
Lequeux
, and
L.
Limat
,
J. Colloid Interface Sci.
370
,
155
(
2012
).
10.
P.
Linton
and
V.
Alfredsson
,
Chem. Mater.
20
,
2878
(
2008
).
11.
D. P.
Chang
,
M.
Jankunec
,
J.
Barauskas
,
F.
Tiberg
, and
T.
Nylander
,
ACS Appl. Mater. Interfaces
4
,
2643
(
2012
).
12.
D. P.
Chang
,
M.
Jankunec
,
J.
Barauskas
,
F.
Tiberg
, and
T.
Nylander
,
Langmuir
28
,
10688
(
2012
).
13.
Colloidal Silica: Fundamentals and Application
, edited by
H. E.
Bergna
and
W. O.
Roberts
(
Taylor & Francis Group
,
Boca Raton
,
2006
), Vol.
131
.
14.
See supplementary material at http://dx.doi.org/10.1063/1.4897282 for SEM micrographs demonstrating the influence of sonication on the dispersion (Fig. S1); the small-angle x-ray diffractograms of the material before and after the sonication process (Fig. S2); schematic drawing of the chamber (Fig. S3); enlarged version of Fig. 1 (Fig. S4); and enlarged version of Fig. 2 (Fig. S5).
15.
T.
Mokhtari
,
Q. D.
Pham
,
C.
Hirst
,
B. M. D.
O’Driscoll
,
T.
Nylander
,
K. J.
Edler
, and
E.
Sparr
,
Langmuir
30
,
9991
(
2014
).
16.
Á. G.
Marín
,
H.
Gelderblom
,
D.
Lohse
, and
J. H.
Snoeijer
,
Phys. Rev. Lett.
107
,
085502
(
2011
).
17.
J. S.
Park
,
Y.-J.
Lee
, and
K. B.
Yoon
,
J. Am. Chem. Soc.
126
,
1934
(
2004
).
18.
J. S.
Park
,
G. S.
Lee
,
Y.-J.
Lee
,
Y. S.
Park
, and
K. B.
Yoon
,
J. Am. Chem. Soc.
124
,
13366
(
2002
).
19.
K.
Sefiane
and
C. A.
Ward
,
Adv. Colloid Interface Sci.
134–135
,
201
(
2007
).

Supplementary Material