View source for Controlling the Stability and Reversibility of Micropillar Assembly by Surface Chemistry

Mariko Matsunaga, Michael Aizenberg and [[Joanna Aizenberg]]

"Controlling the Stability and Reversibility of Micropillar Assembly by Surface Chemistry"

J. Am. Chem. Soc. 133 (14), 5545–5553 (2011)


Entry by [[Meredith Duffy]], AP 225, Fall 2011


'''Keywords:''' [[self-assembly]], [[capillarity]], [[micropillar]], [[adhesion]], [[thiol]]


==Summary==

Self-assembled structures and systems are prevalent in nature, which has resulted in high interest in developing synthetic analogs. While many groups pride themselves on the stability and irreversibility of their self-assembled systems, Matsunaga et al. stress the relevance of dynamic systems whose assembly can be tuned and reversed indefinitely on demand. Here, they investigate the capillarity-driven self-clustering of micropillar arrays, analyzing the force balance that determines cluster stability and modulating the clustering behavior through the use of various solvents and functional groups.

==Methods and Results==

Standard soft lithography methods were used to form polyepoxide pillars of diameter 1.5 μm and height 10 μm in an square lattice of pitch 8 μm. These dimensions were chosen purposefully to limit the maximum cluster size to 2x2 pillars in order to reduce system complexity (whereas [[Self-Organization of a Mesoscale Bristle into Ordered, Hierarchical Helical Assemblies|previous work]] had clusters of up to 10x10 pillars). The array was sputter-coated with a thin gold coating to prevent interactions between the native epoxide surfaces, then modified with various thiols using one of three methods: immersion in an ethanol-based thiol solution, vapor deposition, or soft contact with a thiol-coated PDMS stamp. Surfaces were rinsed in ethanol then air-dried.  Clusters were examined with optical microscopy and SEM.

The mechanism of self-assembly is illustrated in Scheme 1 below. The impetus for clustering is the capillary force of ethanol droplets between pillars. Ethanol rises up the centers of lattice squares, then pulls the tops of the pillars together as it evaporates, resulting in four-pillar clusters. However, the elasticity of the pillars provides a restoring force, which must be overcome by something else in order to maintain stable clustering once all the ethanol has evaporated. The group's previous work showed how adhesive forces between the thiol modifications could fill this role. That said, different thiols have different strengths of adhesion based on the chemical bonds their end groups form (NH2 ~ OH < COOH < SH); the magnitude of this adhesive force A1 compared to that of the elastic force E determines the stability of clusters formed by evaporation. Moreover, whether rinsing with a second solvent can weaken or dissolve these chemical bonds determines the reversibility of stable cluster formation. 

[[Image:madwiki9s1.png]] [[Image:madwiki9p2.png]]

To illustrate the dependence of cluster stability on thiol choice, the authors recorded the thiol-modified arrays during and after the ethanol evaporation and drying steps. Although four-pillar clusters formed during evaporation for all 8 thiols tested (which was expected since the formation of clusters depends only on capillary force exceeding elastic restoring force, a property of material and geometry), Figure 2 reveals that not all of these clusters were stable upon drying due to the different strengths of chemical adhesion. Only 1-5% of pillars remained clustered, and only in pairs, for the thiols in (c), (h) and (i), whereas no clusters at all remained for cysteamine (b). Moderately stable cluster formation can be seen in (e) and (g), whereas the thiols in (d) and (f) proved to have strong adhesive forces resulting in over 90% stable four-pillar clusters. These results were consistent across the three methods of thiol deposition.

It is worth noting that the chemical formulas given for the thiols are abbreviated to emphasize the relevant parts: the length of the carbon chain and the adhesive end group. Indeed, the differences in stability correlated well with the types of bonds expected to dominate. For example, the alkyl end groups of 1-Dodecanethiol (i) can only form van der Waals forces with dissociation energies < 1 kcal/mol and clusters spontaneously disassemble; in contrast, the carboxyl-terminated thiols in (e) and (g) form hydrogen bonds with dissociation energies of 6-8 kcal/mol and consequently form moderately to very stable clusters. Additionally, longer chain length generally created substantially more stable clusters. However, in both of these cases, dithiols (g and h) proved the exception. Though they have the strongest binding forces, they may have been more inclined to bind with their own pillar at both ends than with chains on adjacent pillars, particularly when given the flexibility of a longer carbon chain, leaving their carbon chains partially or mostly responsible for (weak) adhesion between pillars.


Next, the authors demonstrated the ability to repeatedly reverse cluster formation for some thiols by treating the dried stable clusters with new solvents. They determined that water and organic solvents like ethanol, acetone and chloroform easily disassembled the stable clusters formed with carboxyl- or alcohol-terminated thiols (d, e and f). (However, the only moderately stable clusters of the short dithiol chains were irreversible in these solvents, presumably due to their covalent bonds.) Once placed under ethanol, the newly disassembled pillar arrays once more formed stable clusters; this cycle could be repeated many times, even for selectively patterned substrates. This is demonstrated in Figure 5, where pillars were functionalized with thiol in an array of circles by microcontact printing, then subjected to ethanol, dried, subjected to chloroform, and dried again. (For brevity only one cycle of reversible pillar formation is shown below but the original paper shows two to illustrate the repeatability of the process.)

[[Image:madwiki9p5.png|center]]


Interestingly, the rates of cluster dissociation measured for each solvent did not follow the order of polarity of the solvents, despite the polarity of the hydrogen bonds forming the adhesions. In fact, with higher dielectric constant dissembly rate decreased: chloroform was the fastest, while water the slowest. The authors offer the possible explanation that less polar solvents better penetrate the disordered chain regions of the polymer monolayers, disturbing van der Waals forces and thereby destabilizing the geometric integrity of the ordered network of hydrogen bonds at the surface. Conversely, by this theory, the polar solvents must compete against the tightly woven hydrogen bonding networks, consequently taking them longer to disassemble clusters.


==Conclusions==

The authors demonstrate and explain effectively the interactions between the capillarity, elastic, and chemical forces involved in stable and reversible self-clustering of micropillars, and in particular draw attention to some of the previously unexplored roles of chemical adhesion at the mesoscale level (counteracting elasticity). Additionally, with this relatively simple self-assembly system, they have provided a method of quantitatively comparing strengths of most chemical bonds using relatively macroscale observations, since for a set chain length the percentage of pillars that form stable clusters is a direct indication of this bond strength. Overall, this system's potential applications in particle capture-release systems for drug delivery and in generating switchable surface properties (similar to the switchable adhesion of a gecko's foot) are likely to generate interest in self-assembly research and beyond.