Fig. 2: The comprehensive dataset includes SE and DE droplets produced using a wide variety of device geometries, fluid properties, and flow rates.
From: Design automation of microfluidic single and double emulsion droplets with machine learning

a The SE dataset includes 43 polycarbonate devices that varied orifice width, channel depth, outlet width, and continuous and dispersed fluid inlet widths; each device was used over multiple flow rate ratios and capillary numbers to generate SEs with diameters of 25–250 μm at rates of 5–500 Hz. A representative image of SEs for a single experiment is shown. b The DE dataset includes 3 PDMS devices (each with two flow-focusers) and 6 inner, 1 middle, and 4 outer fluids. Devices were tested at different flow rates of inner, middle, and outer fluids (different capillary numbers and flow rate ratios for each junction). The orifice widths of the PDMS devices were 15, 22.5, and 30 μm at the first junction, with an aspect ratio of 1, while the orifice size at the 2nd junction was twice the size of the 1st junction. Several biologically relevant fluids with different viscosities and interfacial tensions were used to generate droplets with diameters of 15 (inner) to 54.2 (outer) μm at rates of 1800–11,800 Hz. A representative image of DEs for a single experiment is shown. Scale bars represent 50 μm. Source Data are provided as a Source Data file.