Royal Society Publishing

Evaporative lithographic patterning of binary colloidal films

Daniel J. Harris, Jacinta C. Conrad, Jennifer A. Lewis


Evaporative lithography offers a promising new route for patterning a broad array of soft materials. In this approach, a mask is placed above a drying film to create regions of free and hindered evaporation, which drive fluid convection and entrained particles to regions of highest evaporative flux. We show that binary colloidal films exhibit remarkable pattern formation when subjected to a periodic evaporative landscape during drying.

1. Introduction

The ability to pattern colloidal films is of growing importance for novel coatings (Martinez & Lewis 2002; Harris et al. 2007; Harris & Lewis 2008), metallized ceramic layers (Masuda et al. 2003) and even high throughput DNA screening (Jing et al. 1998). Most patterning approaches guide colloidal assembly either through chemical (Aizenberg et al. 2000; Lee et al. 2002; Zheng et al. 2002) or topographical (van Blaaderen et al. 1997; Lin et al. 2000; Yin et al. 2001; Lee et al. 2004) modification of the underlying substrate or by application of an external field (Richetti et al. 1984; Hayward et al. 2000; Ristenpart et al. 2003, 2004; Bhatt et al. 2005; Velev & Bhatt 2006). However, those approaches often require multiple processing steps, while only guiding the deposition of a few particle layers. To overcome such limitations, we recently introduced a new route for patterning colloidal films, known as evaporative lithography (Harris et al. 2007).

Our approach builds on prior observations of fluid flow and particle migration in freely evaporating aqueous (Chiu & Cima 1993; Deegan et al. 1997, 2000; Routh & Russel 1998; Deegan 2000; Shmuylovich et al. 2002; Hu & Larson 2005a; Smalyukh et al. 2006) and organic (Berg et al. 1966; Savino et al. 2002; Hu & Larson 2005b, 2006; Girard et al. 2006; Ristenpart et al. 2007) droplets. In aqueous colloidal droplets, contact line pinning and higher evaporation rates at the edge of the drop lead to the outward flow of fluid and entrained particles, yielding the well-known ‘coffee-ring’ effect (Deegan et al. 1997). By contrast, in freely evaporating, non-aqueous drops, the outward flow of fluid and entrained particles is reversed due to Marangoni stresses (Hu & Larson 2006). Evaporative cooling and inefficient heat transfer through the drop induce a temperature gradient, which in turn leads to an inverse gradient in surface tension across the drop’s surface such that the temperature is lowest and the surface tension is highest at the centre of the drop. Recirculating flows develop as fluid is transported from regions of lowest surface tension to regions of highest surface tension. Unlike aqueous colloidal films, the majority of the particles are deposited in the centre of the drop, rather than at its edge.

By introducing a periodically varying evaporative landscape above unary colloidal films, we have reported direct (Harris et al. 2007) and inverse (Harris & Lewis 2008) pattern formation during drying of aqueous and non-aqueous systems, respectively. In this paper, we investigate pattern formation in binary colloidal mixtures that are suspended in water. During evaporative lithography, both particle populations are entrained within the fluid and migrate to regions of highest evaporative flux. The observed pattern formation can be tuned by varying the mask design, mixture composition and particle size ratio.

Below, we review the fundamental phenomena that give rise to pattern formation during evaporative lithography in §2. In §3, we describe the pattern formation observed in binary colloidal films composed of microsphere–nanoparticle mixtures that are subjected to a periodically varying evaporative landscape during drying. Finally, we provide suggestions for other areas of exploration using this nascent approach in §4.

2. Evaporative lithography: a brief review

Evaporative lithographic patterning of unary colloidal suspensions has been demonstrated for aqueous (Harris et al. 2007) and non-aqueous systems (Harris & Lewis 2008). In prior efforts, a mask composed of a periodic array of holes of diameter dh and pitch P was placed above the drying suspension, as shown in figure 1a. A finite separation distance was maintained between the mask and underlying droplet (or film), as shown in figure 1b, where hg defines the initial gap height between the mask and underlying film. As drying proceeds, the film experiences an evaporative landscape that is modulated by the mask above it.

Figure 1.

Schematic illustration of evaporative lithographic patterning. (a) Top view of a binary colloidal mixture drying beneath a patterned mask that contains a hexagonal array of holes with diameter dh and pitch, or centre-to-centre spacing P, (b) side view of the initial gap height hg between the mask and underlying binary colloidal suspension and (c) magnified side view of resulting binary film produced from an initial suspension that contains a high concentration of colloidal microspheres (depicted as grey spheres) and a dilute concentration of nanoparticles (depicted as green spheres). (Note: the dried film consists of a periodic array of discrete nanoparticle features (shaded green in (c)) embedded within a continuous colloidal network.)

Figure 2.

Calculated evaporation profile of a colloidal film dried under a mask with dh=250 μm, P=5dh and hg≈150 μm. Adapted from Harris et al. (2007).

Through finite-element modelling (FEM), we showed that this evaporative landscape varies periodically in accord with the mask design. The maximum evaporative flux, Embedded Image, occurs under the open regions of the mask, whereas the minimum evaporative flux, Embedded Image, approaches zero under the masked regions of the drying film (see figure 2; Harris et al. 2007). Although fluid evaporation is higher in the open regions, surface tension acts to keep the drying film flat. In aqueous systems, fluid flows towards the evaporating regions to compensate for the fluid loss; concurrently, entrained particles accumulate beneath the open regions, as shown in figure 1c. When the difference in evaporative flux, Embedded Image, is normalized by that observed for a freely evaporating film Jfree, we find that ΔJ/Jfree must exceed unity to induce significant pattern formation. Importantly, evaporation is strongly suppressed under the masked regions when either hg is small or P is large, thereby yielding ΔJ/Jfree>1. In prior work, we identified a critical value of the gap height-to-pitch ratio, Embedded Image, of 0.3, below which pattern formation occurs (Harris et al. 2007).

The initial concentration of colloidal species within the drying drop (or film) also plays an important role in pattern formation. In the dilute limit, colloidal particles segregate into discrete patterned features with a final height (hf) as small as several micrometres and lateral dimensions of the order of the hole diameter, dh. However, as the colloid volume fraction increases, there is a transition from discrete patterned features to continuously patterned films. This transition takes place when the diameter of the patterned features df is equal to the pitch P, which occurs when the initial microsphere volume fraction ϕμ exceeds a critical value Embedded Image. When Embedded Image, even thicker films are formed that possess a patterned surface topography. To determine Embedded Image or the mask design shown in figure 1a, we set df=P and equate the total initial microsphere volume within a hexagonal region around each open feature, Embedded Image, to the final microsphere volume within each patterned hemi-ellipsoidal feature, Embedded Image, where hi is the initial film height (≈100 μm). We assume that ϕf=0.64, which corresponds to the random close-packed volume fraction for monodisperse spheres, and use the experimentally determined value of hf/df≈0.02 (Harris et al. 2007). For unary films dried under a mask with a periodic array of holes (dh=250 μm, P=5dh), we find that Embedded Image. Below, we demonstrate how this transition can be exploited in binary mixtures to preferentially segregate one species in a continuous film of another, as illustrated in figure 1c.

3. Pattern formation in binary colloidal films

To explore evaporative lithographic patterning of binary colloidal mixtures, we first created an aqueous suspension of silica microspheres and sulphonated polystyrene nanoparticles with respective volume fractions of ϕμ=0.3 and ϕnano=10−3 and particle radii of am=0.59 μm and an=10 nm. Figure 3 shows a patterned film that was prepared under a mask with dh=250 μm, P=5dh and hg≈30 μm. These particle volume fractions were purposely chosen such that ϕμ>ϕ* and ϕnanoϕ*, where ϕ*∼0.05 for this mask design, as discussed in §2. Under these conditions, the microspheres quickly consolidate to form a continuous patterned film, while the nanoparticles remain entrained in the fluid and segregate to the regions of highest evaporative flux yielding discretely patterned features. The surface topography of the dried binary films shown in figure 3a is analogous to that observed for unary colloidal films at identical ϕμ. To determine the extent of nanoparticle segregation, we imbibed the patterned binary film with a dimethylsulphoxide (DMSO) : water solution that matched the refractive index of the silica microspheres. The refractive index mismatch between the imbibed solution and the polystyrene nanoparticles allowed direct visualization of their distribution (figure 3b). The magnified view of a nanoparticle-rich feature provided in figure 3c reveals a diameter of approximately 250 μm, which is roughly equivalent to the hole size dh within the mask.

Varying the mask design alters the evaporative landscape above the drying film, and, hence, the resulting patterns are formed. Figure 4af shows a series of fluorescence images acquired for binary films dried under masks with varying dh and P at a constant hg≈30 μm. These films were produced from binary suspensions initially composed of ϕμ=0.3 and ϕnano=10−3 and particle radii of am=0.59 μm and an=10 nm, which were dried beneath masks of dh=250 or 500 μm and P=2, 5 or 10dh. Again, the microspheres form a continuous patterned film, similar to that observed in figure 3a, whose surface topography varies with dh and P. In this set of experiments, fluorescently labelled nanoparticles were used to both circumvent the need to imbibe the DMSO : water solution and enhance the resolution of the imaged features. The fluorescence micrographs clearly show that nanoparticles segregate laterally to yield a periodic array of discrete, nanoparticle-rich features embedded within the continuous microsphere films.

Figure 3.

Optical images of a patterned microsphere–nanoparticle film dried under a mask with dh=250 μm and P=5dh. (a) Only the microspheres are visible in the dried film. (b) After imbibing the film with an index matching DMSO : water solution, the silica film becomes transparent, leaving the polystyrene nanoparticles clearly visible (scale bar: 5 mm). (c) A magnified view of a single nanoparticle feature shows the lateral segregation of the nanoparticles (scale bar: 250 μm).

Figure 4.

Fluorescence images of patterned binary films dried under masks with dh=250 μm (ac) or 500 μm (df) and varying centre-to-centre distance between holes, P, of 2, 5 or 10dh, in which only the fluorescent nanoparticles are shown. Scale bar: 5 mm. Adapted from Harris et al. (2007).

Figure 5.

Fluorescence images of binary colloidal films prepared from suspensions of constant ϕμ=0.3, ϕnano=10−3 and varying aμ/anano, in which only the fluorescent nanoparticles are shown. The films were dried under masks with dh=250 μm, P=5dh and hg≈30 μm. (a) amicro/anano=47, (b) amicro/anano=25, (c) amicro/anano=10.6, (d) amicro/anano=15.8, (e) amicro/anano= 8.3, (famicro/anano=3.5, (g) amicro/anano=4.7, (h) amicro/anano=2.5 and (i) amicro/anano=1.1.

Figure 6.

Plot of the patterned feature diameter as a function of aμ/anano for films assembled from binary suspensions of fixed ϕμ=0.3 and ϕnano=10−3, but varying size ratio, aμ/anano. The films were dried under a mask with dh=250 μm and P=5dh. (Note: inset depicts the interstitial pores formed between close-packed microspheres and surrounding nanoparticles in binary mixtures of high size ratio.)

In the initial stage of drying, silica microspheres and polystyrene nanoparticles are transported to regions of high evaporative flux by particle convection. As drying proceeds, the microspheres consolidate into a close-packed network that contains interstitial pores of radius ap≈0.15 aμ. Since ϕμ∼0.5 ϕf, this transition occurs when roughly 50 per cent of the water has evaporated from the film. Because ϕnanoϕ*<ϕf, the nanoparticles remain entrained in the liquid well beyond this point; moreover, since ananoap, they are able to migrate through the porous microsphere network. As the drying front recedes into the film, capillary tension at the liquid menisci creates a pressure gradient, Pc=−2γlv/ap, that further enhances nanoparticle segregation, where γlv is the liquid–vapour surface tension. Using the measured surface tension of 27 mNm−1, the calculated pressure drop is approximately −6 atm.

To investigate the effects of the particle size ratio on pattern formation, we prepared a series of binary colloidal suspensions of fixed ϕμ=0.3 and ϕnano=10−3, but varying aμ/anano. The films were dried under a mask with dh=250 μm, P=5dh and hg≈30 μm. Figure 5 shows the fluorescence images of the dried films; the diameter of the patterned features increased as aμ/anano decreased. The extent of nanoparticle segregation is quantified by plotting the diameter of the patterned features as a function of aμ/anano (figure 6). When aμ/anano>7, significant nanoparticle segregation occurs and the diameter of the observed patterned features is comparable to dh, as shown in figure 6. At a critical value of aμ/anano∼7, the nanoparticle size is now equivalent to the characteristic pore size, i.e. anano=ap≈0.15aμ. When aμ/anano≤7, there is an abrupt change in feature size as the transport of nanoparticles within this porous network becomes hindered. In actuality, the microsphere network contains a distribution of pore sizes, hence some nanoparticle segregation persists even when anano>ap. Not surprisingly, as aμ/anano decreases even further, there is a concomitant rise in the size of the patterned, nanoparticle-rich features. Ultimately, when aμ/anano≈1, the features are comparable in size to the pitch P, indicating that pattern formation has been effectively suppressed.

4. Conclusions

Evaporative lithographic patterning offers many advantages over other assembly routes, including its simplicity and the ability to control particle deposition without substrate modification. Importantly, it provides a promising new avenue for patterning soft materials, including colloidal, polymeric and biomolecular species. Our investigation, albeit confined to a narrow compositional window, demonstrates that evaporative lithography can be readily extended to binary mixtures. Looking towards the future, we anticipate that a rich array of patterned films can be produced by varying the respective microsphere and nanoparticle concentrations and exploring non-aqueous systems.


This work is based on research supported by the National Science Foundation (grant no. DMR-0652424). The authors gratefully acknowledge the contributions of H. Hu (Proctor and Gamble), who carried out FEM analysis, and W. Wu, who assisted with figure preparation.



View Abstract