Flexible Fabrication of Microarrays of Microwells

Microwell technology, although in its infancy, has already shown enormous potential. Microwells have for example been applied in a number of optoelectronics-related applications where microwells were able to offer fine control over, and manipulation of, crystal nucleation. In the biological arena microwells have been produced that have allowed the selection and immobilization of single cells; while micro-patterned wells that maintain human embryonic stem cells (hESC) undifferentiated for up to three weeks, and limit colony growth have been reported. A number of approaches have been used to generate the required structures. This includes a variety of soft-lithographic techniques, often using a combination of polydimethylsiloxane (PDMS) based materials and photolithography to produce “stamps” which can be coated with various polymers, biochemicals, gels etc., prior to stamping onto the receiving substrate to yield the desired microwell pattern. Photolithographic methods have also been used to transfer images from a mask to a metal surface, with subsequent selective etching to generate the desired microwells. Several templating methods, based on self-assembly, have been used to create patterned substrates with sub-micron features. This includes the use of ordered arrays of colloidal particles, microporous materials, polymers with rod-coil architecture or honeycomb structures, self-organized surfactants, and microphase-separated block copolymers. A very recent approach that has been applied to the process of microwell generation is the deposition of organic solvents onto a solid polymer surface via inkjet printing, however only single wells were generated, while in many cases inkjet printing generated wells with major “irregularities or bumps” within the well itself. Here we report the rapid fabrication of arrays of 3D microwells by contact printing various solvents upon dual-polymer layered coated substrates using a microarray printer. This approach offers access to a highly controllable variety of microwells in an array type format, with high densities of wells, each of which is individually addressable by the printer. The general approach is shown in Scheme 1 in which a contact-based microarrayer equipped with solid pins was used to deposit organic solvent onto polymer coated substrates, thereby causing local dissolution and microwell formation. For our application the microwells were designed to bind cells and thus the slides used were coated with a base layer of chitosan (thickness ∼1.0 lm) and an upper layer of polystyrene (the thickness of the coated PS layer was either 1.2 lm or 2.4 lm as measured by scanning electron microscopy (SEM)). The chitosan also played an important role as an adherent, preventing the detachment of the PS layer from the glass slide when the slide was used in aqueous environments. The process of contact printing and polymer dissolution is clearly solvent dependant and a number of solvents (acetophenone, toluene, and ethyl acetate) were initially selected and analyzed for their potential to generate wells. Microwell formation only occurred with acetophenone, but initially these were non uniform. However, using a mixture of solvents (acetophenone and ethyl acetate) substantially improved the characteristics of the microwells with the best results being observed with a 2:1 ratio of acetophenone/ethyl acetate (see Supporting Information for further details). Using this solvent combination, well dimension and density of printing could be controlled by the make-up of the coating on the substrate, the sizes of the solid pin heads (150 lm or 100 lm), the amount of solvent printed, as well as subtle features such as contact time (the time the pin is in contact with the slide during printing (for this study this was fixed as 100 ms) and the number of “stampings” (1-10) used to generate the wells. The density of the wells could be controlled by altering the pitch (the distance between the centers of two printed drops from 1500–400 lm) and the number of stamps. Using a 150 lm pin and a pitch distance of 1500 lm the density of the wells was 50 per cm, but this could be increased up to 600 wells per cm area with a 400 lm pitch (Fig. 1A). Using smaller pins (100 lm diameter) and a spacing distance of 220 lm the density of the wells could be increased to generate 12800 addressable wells on a standard glass slide (Fig. 1B). Much greater numbers of stamps and/or decreases in pitch distance lead to merging of the wells. Each stamp deposited about 0.2 nL of solvent (small pin) and 0.5 nL (large pin) per stamping and thus well size, depth and morphology could be finely controlled by varying the number of stamps used during the deposition process. This is shown clearly in Figure 2A-C where increasing the number of stamps gave a controlled increase in well size, although this reached a maximum after about 8–10 stamps. C O M M U N IC A TI O N