Advances in Droplet Microfluidics for High-Throughput Experimentation

Abstract

High throughput experimentation (HTE) is increasingly important in both academic and industrial research environments. While the utility in screening large libraries for many applications is promising, the analytical testing remains a major bottleneck to discovering desired samples. New analytical tools are required that can process and screen large datasets with minimal time and waste. This work focuses on the development and advancement of droplet microfluidics to achieve HTE. A limiting factor to droplet microfluidics as a screening technology is the transfer of small molecules between droplet samples. This work first investigates droplet carryover or “crosstalk” using electrospray ionization mass spectrometry (ESI-MS). Using this label-free analysis method, 36 small molecules are screened for extent of crosstalk, and the behavior is modelled using physicochemical properties. Important drivers of crosstalk are elucidated, such as analyte hydrophobicity, surfactant identity, oil composition, surfactant concentration, and flow duration. A new surfactant is tested which is shown to selectively retain small molecules that resemble the surfactant head chemical structure. We next apply fundamental droplet investigations to the development of novel unit operations for the processing of droplet samples. In one tool, dielectrophoresis (DEP) is used to split and trap 400 – 1000 pL microfluidic droplets in flow at rates of 25 droplets/s. The splitting effect is characterized to show the utility of the tool in microfluidic workflows with varying droplet sizes and applications. A second tool, high-dilution reagent addition, is developed to enable temporal control for synthetic and biological screening. Two microfluidic device designs are developed to dilute droplet samples from 2 – 150-fold with less than 1% carryover between droplet samples. One device was shown to be compatible with droplets generated from a multiwell plate, which was applied to testing a photocatalysis reaction with 1% reagent consumption compared to µL-scale experiments. The second device is shown to provide 10-fold dilution at > 40 droplets/s without carryover and a 97% success rate, which is a significant improvement on previous designs. The development of a microfluidic liquid-liquid extraction is shown as an advanced tool for sample processing. A microfluidic chip is designed to achieve slug flow nanoextraction (SFNE). An automated system is built using the device, employing an autosampler to inject samples into the chip, and an in-line UV detector to analyze small molecules partitioning between aqueous and organic phases. Flow modulation and analytical performance is characterized using the automated SFNE system. A mini-screen of octanol-water partition coefficients is performed using the system, leading to a platform that meets industry standards with 10-fold less time and 40-fold less volume than a previous miniaturized method. Lastly, the development a droplet-based analytical method to test and sort cell variants for lysine production is presented. An ESI-MS method is shown to perform reliable detection of lysine from 10-30 nL microfluidic droplets even with a complex cell media background. Individual E. coli cells are encapsulated and grown in microfluidic droplets, and lysine production from a known E. coli variant is measured. Using mass-activated droplet sorting, the above strategies were used to analyze and sort cells with high lysine production, with 5.8-fold improvement of lysine production from a mixed variant population after 2.5 h of sorting. The resulting pool is immediately translatable to validation methods such as sequencing.