pubs.acs.org/Langmuir © 2010 American Chemical Society
Tunable Pyramidal Assemblies of Nanoparticles by Convective/Capillary Deposition on Hydrophilic Patterns Made by AFM Oxidation Lithography B. Viallet, L. Ressier,* L. Czornomaz, and N. Decorde Universit e de Toulouse, INSA-CNRS-UPS, LPCNO, 135 avenue de Rangueil, 31077 Toulouse Cedex 4, France Received February 8, 2010. Revised Manuscript Received March 10, 2010 Close-packed pyramidal assemblies of 100 nm latex nanoparticles were made by convective/capillary deposition on hydrophilic patterns created by oxidation lithography using atomic force microscopy (AFM). We demonstrated that the substrate temperature during convective/capillary assembly is a key experimental parameter in finely tuning the geometry of these pyramids and thus the total number of nanoparticles forming each 3D assembly. The volume and shape of these nanoparticle assemblies are discussed and compared to simulations.
I. Introduction The directed assembly of colloidal nanoparticles onto desired localized areas of substrates is necessary to fabricate reproducible, reliable nanoparticle-based devices.1,2 Among all techniques, the combination of convective/capillary deposition3 and chemical patterning has been identified as an efficient approach for the realization of close-packed nanoparticle assemblies with various geometries.4-9 This method consists of dragging a meniscus of a colloidal suspension onto a surface presenting a wetting contrast obtained by chemical patterning. In this case, colloidal nanoparticles are deposited on the most wettable areas of the patterned surface. We have shown in a previous paper that this wetting contrast can be obtained by combining chemical functionalization and AFM oxidation lithography.9 We demonstrated that convective/capillary deposition onto such patterned surfaces is an efficient way to control the local 3D assembly of any kind of colloid on hydrophilic patterns. The versatility of this original method was demonstrated by applying it to 100 nm colloidal latex nanoparticles and E. coli bacteria. In the present work, we characterize and analyze the crucial role of the substrate temperature during the convective/capillary assembly of latex nanoparticles onto this kind of chemically functionalized surface patterned by AFM oxidation lithography.
consists of functionalizing a silicon substrate with a hydrophobic self-assembled monolayer (SAM) of octadecyltrimethoxysilane (OTMS) by chemical vapor deposition (CVD) (Figure 1a). The experimental conditions used for this silanization, optimized in previous work,10 lead to a 2.3-nm-thick dense monolayer presenting a water contact angle of about 100°. The aim of this hydrophobic functionalization is to prevent nanoparticle deposition outside of the defined patterns. Hydrophilic SiOx patterns are then created by AFM oxidation lithography in contact mode under ambient conditions (20 °C, 60% RH) using a negative dc bias of -10 V (Figure 1b). Under these conditions, previous studies revealed that the hydrophilic silicon oxide patterns protrude from the surrounding intact OTMS monolayer of about 2 nm.9,11 The use of AFM oxidation lithography on a silicon substrate covered by an OTMS selfassembled monolayer is a convenient, fast, and versatile way to fabricate hydrophilic oxide patterns with nanometer precision, surrounded by a hydrophobic background, without requiring any expensive clean-room or vacuum equipment. The last step of the process is the convective/capillary assembly of colloidal nanoparticles onto this chemically patterned substrate (Figure 1c). It consists of dragging a droplet of the colloidal suspension onto the substrate. For the present experiments, we used an aqueous suspension of 100 nm latex nanoparticles
II. Experimental Section Figure 1 gives a schematic description of the nanoparticle assembly method used in this work. The first step of the process *Corresponding author. E-mail:
[email protected].
(1) Shipway, A. N.; Katz, E.; Willner, I. Chem. Phys. Chem. 2000, 1, 1–52. (2) Koh, S. J. Nanoscale Res Lett. 2007, 2, 519–545. (3) Dimitrov, A. S.; Nagayama, K. Langmuir 1996, 12, 1303–1311. (4) Jonas, U.; del Campo, A.; Kruger, C.; Glasser, G.; Boos, D. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5034–5039. (5) Fan, F.; Stebe, K. J. Langmuir 2004, 20, 3062–3067. (6) Fustin, C. H.; Glasser, G.; Spiess, H. W.; Jonas, U. Langmuir 2004, 20, 9114– 9123. (7) Maury, P.; Escalante, M.; Reinhoudt, D. N.; Huskens, J. Adv. Mater. 2005, 17, 2718–2723. (8) Ling, X. Y.; Malaquin, L.; Reinhoudt, D. N.; Wolf, H.; Huskens, J. Langmuir 2007, 23, 9990–9. (9) Ressier, L.; Viallet, B.; Beduer, A.; Fabre, D.; Fabie, L.; Palleau, E.; Dague, E. Langmuir 2008, 24, 13254–13257. (10) Ressier, L.; Martin, C.; Viallet, B.; Grisolia, J.; Peyrade, J. P. J. Vac. Sci. Technol., B 2007, 25, 17–20. (11) Sugimara, H.; Hanji, T.; Hayashi, K.; Takai, O. Ultramicroscopy 2002, 91, 221–226.
Langmuir 2010, 26(7), 4631–4634
Figure 1. Directed assembly of colloidal nanoparticles by combining chemical functionalization, AFM oxidation nanolithography, and convective/capillary deposition: (a) grafting of a hydrophobic OTMS self-assembled monolayer by CVD, (b) fabrication of hydrophilic SiOx patterns by AFM oxidation nanolithography, and (c) directed assembly of nanoparticles by convective/capillary deposition.
Published on Web 03/16/2010
DOI: 10.1021/la1005852
4631
Letter
Figure 2. AFM topography images of a pyramidal assembly of 100 nm latex nanoparticles on a 9 μm 11 μm SiOx pattern for a substrate temperature of 22 °C during convective/capillary deposition: (a) 3D view (scan size, 15 μm; Z scale, 1 μm) and (b) top-view enlarged image of panel a (scan size, 6 μm; Z scale, 1 μm). (volumic concentration of 2%). The meniscus dragging speed was fixed at 10 μm s-1 and the surrounding relative humidity was maintained at 40%. The only parameter that was varied was the substrate temperature.
III. Results and Discussion 1. Experimental Results. Figure 2 presents typical AFM images in tapping mode of the resulting nanoparticle assembly for a substrate temperature of 22 °C during convective/capillary deposition. In agreement with previous work, this Figure shows that the created SiOx pattern is covered with a pyramidal closepacked assembly of latex nanoparticles.9 Real-time observations by optical microscopy during convective/capillary deposition provided evidence of a three-step assembly mechanism.9 In the first step, a droplet of the colloidal suspension is dragged onto the unmodified OTMS monolayer. In this case, no deposition of nanoparticles occurs because the hydrophobic behavior of this SAM leads to a high dynamic contact angle of the colloid droplet dragged on the surface. The second step begins when the meniscus arrives at the hydrophilic silicon oxide patterns. At that point, the contact angle of the dragged droplet decreases. The contact angle on the SiOx patterns is then small enough to enable colloid assembly in front of the meniscus by convective deposition. These first two steps are very similar to those previously reported in the literature in the case of colloid assembly onto silicon substrates, chemically patterned by photolithography and silanization.4,6 The third step is dominated by capillary phenomena. The colloidal suspension droplet is elastically deformed by the oxide patterns until the meniscus line breaks,12 leaving a small amount of colloidal suspension that completely wets the already deposited colloids. The remaining solvent finally evaporates. Figure 3 clearly reveals that the number of nanoparticle layers in the created pyramids (and thus the pyramid volume) strongly depends on the substrate temperature during convective/capillary assembly. The complete set of experimental data (blue dots) reported in Figure 4 quantifies this temperature dependence. Two regimes are actually evidenced on this figure: the first one, for temperatures varying from 15 to 27 °C, corresponds to the exponential-like growth of the pyramid volume with the substrate temperature. The second regime occurs for temperatures higher than 27 °C. In that case, the pyramid volume no longer depends on the substrate temperature. It remains constant, at a value of around 37 μm3 under the present experimental conditions. (12) Joanny, J. F.; de Gennes, P. G. J. Chem. Phys. 1984, 81, 552–562.
4632 DOI: 10.1021/la1005852
Viallet et al.
In both regimes, the heights and the number of layers of the pyramids, experimentally measured by AFM, are consistent with compact arrangements of 100 nm nanoparticles (Table 1). Table 1 also gives the corresponding number of nanoparticles forming the pyramids. For instance, pyramids containing 10 layers are made of 51 000 close-packed latex nanoparticles. 2. Simulations and Discussion. The exponential-like growth of the pyramid volume with the substrate temperature in the first regime was modeled and compared with experimental results. Real-time observations by optical microscopy revealed that there is no significant concentration of nanoparticles toward the meniscus before nanoparticle assembly on the oxide patterns in this regime. We therefore made the hypothesis that the nanoparticle pre-concentration at the three-phase contact line can be neglected. The volume of the nanoparticle pyramids Vpyramid can then be expressed by eq 1 Vpyramid ¼
c Vevap D
ð1Þ
where c is the volumic concentration of nanoparticles in the colloidal suspension (2% in the present case), D is the compactness of the nanoparticle assembly (0.67 for the bottom and top layers of nanoparticles and 0.74 for the other layers, in the present case), and Vevap is the volume of the colloidal suspension evaporated above the oxide pattern. Vevap is given by eq 2, where S(t) is the evaporation surface and J(T) is the flux of evaporation that depends only on the substrate temperature T (the surrounding relative humidity is constant in our experiments): Z Vevap ¼ JðTÞ
SðtÞ dt
ð2Þ
t
We experimentally measured the evaporation flux J(T) by recording real-time videos of the evaporation of a calibrated water droplet at various temperatures. The experimental values of the evaporation flux obtained were well fit by the Hertz-Knud13,14 sen formula. R The t S(t) dt term depends on the pattern geometry, the meniscus dragging speed, and the hydrophilic/hydrophobic contrast between the oxide patterns and the surrounding functionalized substrate. This term is quite difficult to measure experimentally with a good accuracy. We therefore used it as a fit parameter in our model. Very good agreement between experimental and R simulated nanoparticle volumes was obtained for a value of t S(t) dt = 257 μm2 s in all experiments. This optimum fit value is consistent with the experimental value that we roughly estimated by measuring the area protruding in front of the meniscus using real-time videos during nanoparticle assembly. The results of this model are plotted in Figure 4 (red triangles). This very good agreement between experimental data (blue dots) and simulation data (red triangles) confirms the validity of our hypothesis. The main hypothesis was that the nanoparticle concentration near the meniscus on the hydrophobic surface was the same as in the original suspension during the first step of the deposition. It is worth noting that this hypothesis was not valid in previous work reported in the literature regarding the capillary assembly of colloidal nanoparticles on topographical patterns.15 In that case, (13) Barrett, J.; Clement, C. J. Colloid Interface Sci. 1992, 150, 352–364. (14) Ward, C. A.; Fang, G. Phys. Rev. E 1999, 59, 429–440. (15) Malaquin, L.; Kraus, T.; Schmid, H.; Delamarche, E.; Wolf, H. Langmuir 2007, 23, 11513–21.
Langmuir 2010, 26(7), 4631–4634
Viallet et al.
Letter
Figure 3. AFM top-view topography images (scan size, 15 μm; Z scale, 1 μm) and associated horizontal and vertical sections (blue curves) of
the pyramidal assembly of 100 nm latex nanoparticles on 9 μm 11 μm SiOx patterns for various substrate temperatures during convective/ capillary deposition: (a) 20, (b) 24, (c) 25, and (d) 27 °C. The horizontal and vertical sections (red curves) of simulated water droplets on 9 μm 11 μm hydrophilic patterns corresponding to the nanoparticle pyramids obtained at these temperatures are also plotted for comparison.
Figure 4. Evolution of the volume of the nanoparticle pyramids with the substrate temperature during convective/capillary deposition. Comparison of the experiment (blue dots) with the model (red triangles).
indeed, nanoparticle accumulation at the three-phase contact line was found to be essential in triggering the assembly process. In the second regime, the pyramid volume remains constant with the substrate temperature because the nanoparticle assembly cannot be larger than the suspension droplet left when the meniscus broke. The constant value of the pyramid volume indicates that the modification of the droplet contact angle induced by the substrate temperature has a negligible impact on the volume of the suspension droplet left on hydrophilic patterns in the explored temperature range. The volume of the suspension droplet depends only on the pattern geometry and the meniscus Langmuir 2010, 26(7), 4631–4634
dragging speed. It is thus the same in both regimes. However, contrary to the first regime, the remaining suspension droplet in the second regime is completely saturated in nanoparticles because of the high flux of evaporation induced by higher substrate temperatures. Moreover, the shapes of the pyramidal nanoparticle assemblies measured by AFM (blue curves on sections of Figure 3) were compared to the shapes of simulated water droplets (red curves on sections of Figure 3) using Surface Evolver software.16 For these simulations, the water droplets were constrained to a rectangle corresponding to the geometry of the oxide patterns, the contact angle of the droplet with respect to the surface was fixed, and the volume of the droplet was adjusted to match the shape of the pyramids as observed by AFM. The pyramid shapes obtained at temperatures of up to 24 °C (Figure 3a,b) are symmetric and correspond to the computed droplet shapes. Because of the high hydrophilic/hydrophobic contrast between the oxide pattern and the surrounding functionalized substrate, the meniscus broke during the third step of deposition, leaving a droplet of a concentrated suspension of nanoparticles. The remaining liquid evaporated, and the nanoparticles were reorganized to adopt the shape of this droplet. For temperatures higher than 24 °C, the pyramids can be fit by a symmetric droplet in the direction perpendicular (“horizontal section” in Figure 3c,d) to the meniscus motion but not in the direction parallel (“vertical section” in Figure 3c,d). In that case, the concentration of nanoparticles in the remaining suspension droplet was so high that the nanoparticles did not have enough freedom of movement and time to reorganize. The shape of the (16) Brakke, K. A. Exp. Math. 1992, 1, 141–165.
DOI: 10.1021/la1005852
4633
Letter
Viallet et al.
Table 1. Comparison of the Experimental Heights of Nanoparticle Pyramids Measured by AFM and the Theoretical Heights of Close-Packed Nanoparticle Pyramids Containing the Same Number of Layers substrate temperature during deposition (°C) number of layers in the pyramids experimental pyramid heights measured by AFM (nm) theoretical pyramid heights in the case of a compact arrangement (nm) number of particles in the pyramids
nanoparticle assemblies is thus close to the shape of the meniscus at the end of the second step of deposition, just before the meniscus breaks. The slope of the pyramid on the front side of the droplet is lower than the droplet profile. That is consistent with the contact angle hysteresis theory.17 The front slope has a low contact angle corresponding to a receding meniscus, which is lower than the equilibrium contact angle. The back slope is higher and is given by the nanoparticle accumulation. Finally, an accurate observation of Figure 3a reveals that the area of the first layer of nanoparticles is smaller than the area of the oxide pattern. Fine contrast tuning of all AFM images in nanoparticle pyramids leads to the same conclusion. Our simulations showed that the gap between the first layer of nanoparticles and the edge of the oxide patterns corresponds to the domain where the thickness of the liquid droplet was smaller than the nanoparticle diameter.
IV. Conclusions Combining chemical functionalization, AFM oxidation lithography, and convective/capillary deposition is an efficient way to fabricate pyramidal assemblies of close-packed nanoparticles. We demonstrated that the substrate temperature during convective/ (17) Rapha€el, E.; De Gennes, P. G. J. Chem. Phys. 1989, 90, 7577–7584.
4634 DOI: 10.1021/la1005852
20 2 190 182 13 000
22 3 270 263 19 000
24 5 430 427 30 000
25 7 585 590 39 000
27 10 800 835 51 000
capillary deposition is a crucial parameter in accurately adjusting the geometry of the pyramidal assemblies. Experimentally, two regimes were evidenced. In the first regime, for temperatures ranging from 15 to 27 °C under our experimental conditions, the volume of the pyramidal nanoparticle assemblies exponentially increases with the substrate temperature. This evolution can be quantitatively explained by the increase in the evaporation flux with the substrate temperature during convective/capillary assembly. In the second regime, for temperatures higher than 27 °C, the volume of the pyramidal nanoparticle assemblies remains constant with temperature. In that case, the volume of the nanoparticle pyramids is actually limited by the volume of the suspension droplet left on the hydrophilic patterns when the meniscus broke. Concerning the shape of the nanoparticle pyramids, these are symmetric for temperatures of up to 24 °C because nanoparticles have enough freedom to reorganize during solvent evaporation with respect to the shape of the droplet left on the hydrophilic patterns. For higher temperatures, the nanoparticle pyramids are no longer symmetric in the direction parallel to the meniscus motion. In that case, the concentration of nanoparticles in the remaining suspension droplet being too high prevents nanoparticle reorganization. The shape of the nanoparticle pyramids is thus close to the shape of the meniscus just before it breaks.
Langmuir 2010, 26(7), 4631–4634
