ABS 018: Microfluidic Mixing in Droplet-Based Injectors for Time-Resolved Serial Femtosecond Crystallography of NQO1 Enzyme

Angela Zeng ¹ ², Sayante Sen ¹ ³, Diandra Doppler ¹ ³, Mukul Sonker ¹ ³, Alexandra Ros ¹ ³

¹ Center for Applied Structural Discovery, Biodesign Institute, Arizona State University, Tempe, AZ 85281
² School of Engineering, Brown University, Providence, RI 02912
³ School of Molecular Sciences, Arizona State University, Tempe, AZ 85287

The Van Wickle Journal (2026) Volume 2, ABS018

Introduction: Time-resolved serial femtosecond crystallography (TR-SFX) is a structural biology technique that uses femtosecond X-ray free-electron laser pulses to capture molecular motions in proteins. This method can reveal dynamic structural changes that are difficult to resolve using traditional crystallography. However, TR-SFX is highly sample-intensive because protein microcrystals must be continuously delivered into the X-ray beam, and a large fraction of the sample can be wasted between pulses. Droplet-based sample delivery offers a strategy to reduce sample consumption by encapsulating protein crystals in aqueous droplets surrounded by an immiscible oil phase, allowing better synchronization between sample delivery and X-ray pulses.

This project focused on characterizing the Side-Mixer Droplet Generator 200 (SM-DG200), a microfluidic device designed to generate droplets containing human NAD(P)H:quinone oxidoreductase 1 (NQO1) enzyme crystals. NQO1 is a therapeutic target due to its overexpression in many solid tumors. The goal was to identify flow rate and electrical triggering conditions that reliably produce droplets at approximately 120 Hz, matching the repetition rate required for experiments at the Linac Coherent Light Source (LCLS). By optimizing droplet generation parameters, this work supports more efficient sample delivery for future TR-SFX experiments on NQO1.

Methods: The SM-DG200 droplet generator was used to produce aqueous droplets within an immiscible oil phase. Flow rates of the aqueous and oil phases were systematically varied to determine conditions that generated stable droplet frequencies near the 120 Hz target. Electrical stimulation was applied at the Y-junction of the device to induce electrowetting, which altered aqueous phase attachment to the channel wall and improved control over droplet pinch-off behavior. Voltage amplitude and pulse duration were adjusted to evaluate their effects on droplet generation. Droplet production was monitored using microscopy, and frequency values were experimentally measured across different flow rate combinations. Waterfall plots were used to verify droplet timing consistency and identify parameter combinations suitable for synchronization with LCLS beamtime experiments.

Results: Droplet generation frequency increased approximately linearly with total flow rate. Several viable flow rate combinations produced frequencies near the 120 Hz target. The optimal tested condition was an aqueous flow rate of 5.5 µL/min and an oil flow rate of 21.2 µL/min, which generated droplets at approximately 121.36 Hz. Waterfall plots confirmed consistent droplet timing and reliable lock-in behavior using electrical triggering parameters of 1.5 ms duration and 75 V amplitude. These results identify operating conditions suitable for future LCLS experiments.

Discussion: These findings demonstrate that droplet-based microfluidic sample delivery can be tuned to match the frequency requirements of TR-SFX experiments while reducing sample waste. Optimizing flow rate and electrical triggering parameters enables reliable droplet generation and synchronization with X-ray pulses. This approach may improve the efficiency of structural studies on NQO1 and other protein targets by conserving valuable crystal samples. Future work should further validate device performance during beamtime experiments and test droplet stability under protein crystal loading conditions.