Figure 1
Robotic fluid delivery components, operation principle, and characterization of switching performance and well-to-well carryover across a 96-well plate. (a) Model of the robotic platform comprised of a multiwell plate holder with x-y motion and a servo linkage with vertical z-motion to raise and lower an inlet tube connected to a microfluidic device mounted on a microscope stage (Supplementary Video S1). (b) Control schematic of the robotic platform, microfluidic device valves, and microscope optical components. Device connections include an inlet (in) tube, backpressure (bp) reservoir, valve, and tube, and an outlet (out) valve and tube. Robotic positions are computer-controlled by Arduino (1) and all valves and microscope illumination are controlled by Arduino (2). (c) Top view of an example microfluidic device with a multiwell plate inlet (in), and backpressure (bp) port, and one waste outlet (out). Valve states (red X, closed) are illustrated below during well flow (left) and well-to-well switching (right). (d) Side view of the raised inlet tubing and magnified views demonstrating proper backpressure (bp) balance: bp too low can introduce an air bubble in the inlet tubing whereas bp too high can cause backward flow and a drop to emerge from the inlet tubing. (e) Timing of valve actuation and servo motion during well-to-well transfer and dynamics of the fluid switch monitored with fluorescein dye. Upper graphs show the time course of relative fluorescence intensity beginning with the switch from a buffer well to fluorescein at t = 0, and a switch back to buffer at t ≃ 90 s. Inset shows a magnified view from 110–130 s (i.e. 20–40 s after the second tube switch). Below, corresponding actuation states of backpressure (bp) and outlet (out) valves and inlet tube z-position (purple arrows) during the short (~2 s) transition time (tΔ). Fluid switching in the microfluidic device is complete after a filling delay (tfill, vertical red lines) dependent on flowrate and inlet tube volume. (f) Multiwell plate map showing 96 alternating buffer and fluorescein dye wells scanned in a “snake” pattern. (g) Heat map shows percent carryover without wash steps from fluorescent intensity in the microfluidic channel, calculated for each buffer well relative to the prior fluorescein well after a fill delay of 30 s. (h) Histogram of well-to-well carryover without wash steps across 48 wells, from data shown in g. (i) Sequential tubing wash steps further reduce well-to-well carryover. Data are shown using a fill delay of 30 s.
