The dynamics of capillary flow in an open-channel system featuring trigger valves

Abstract

Trigger valves are fundamental features in capillary-driven microfluidic systems that stop fluid at an abrupt geometric expansion and release fluid when there is flow in an orthogonal channel connected to the valve. The concept was originally demonstrated in closed-channel capillary circuits. We show here that trigger valves can be successfully implemented in open channels. We also show that a series of open-channel trigger valves can be placed alongside or opposite a main channel resulting in a layered capillary flow. We developed a closed form model for the dynamics of the flow at trigger valves based on the concept of average friction length and successfully validated the model against experiments. For the main channel, we discuss layered flow behavior in the light of the Taylor-Aris dispersion theory and in the channel turns by considering Dean theory of mixing. This work has potential applications in autonomous microfluidics systems for biosensing, at-home or point-of-care sample preparation devices, hydrogel patterning for 3D cell culture and organ-on-a-chip models.

Keywords: Capillary microfluidics; Fluid dynamics; Friction length; Open microfluidics; Stop valves; Trigger valves.

Fig. 1

Considerations and operation of trigger valving in open channels. (a) Open-channel device with one TGV highlighting the intersection of the main channel and side channel and key aspects of TGV devices. (Inset) Drawings of the side channel opening indicating where pinning occurs (upper left panel) and a view of the trigger valve gate with the step from the bottom of the main channel to the bottom of the side channel (lower left panel)—a critical feature preventing flow into the empty main channel. Accompanying Keyence profilometer photo of an example open TGV (right panel). (b) A series of images illustrating TGV release in open channels starting with an empty main and side channel (bi) with the same channel dimensions as the sketch in i and ii. (bii) After liquid is added to the side channel, the fluid pins at the edge of the opening to the main channel where it remains stationary. (biii) When the main channel fluid makes contact with the pinned fluid, the side channel liquid is de-pinned. (biv) Fluids from main channel and side channel flows then stabilize over time.

Fig. 2

Sketch of the device with the three TGVs, placed on the same side (a) or on opposite sides (b), in the notations, a star denotes the main channel. (c) Schematic of the flow at a node (intersection of the TGV and main channel) used for the calculation. The letter Q corresponds to a volumetric flow rate, and P to a pressure.

Fig. 3

Distance between TGVs provides timed control of fluid release. Progression of flow in devices with parallel TGVs spaced approximately 15 mm (a, Device 1) and 30 mm (b, Device 2) apart. Comparison of model (solid line) with experimental fluid front travel distance (diamonds) at the meniscus over time for device 1 (c, left) and 2 (d, left). Experimental data was averaged for three trials (n = 3). Model is presented in segments corresponding to the calculated travel distances prior to the TGV release (orange, z0), between the first and second TGV (blue, z1), between the second the third valves (yellow, z2) and between the third valve and the paper pad (red, z3). Experimental velocities (diamonds) were calculated from the travel distance and fluid velocities of each channel were calculated from the travel distance progression and are compared against the calculated model velocities (dashed line) using a dynamic contact angle model (V0, orange) for the inlet. Model is shown at the first (blue, V1), second (yellow, V2), and third (red, V3) valves due to the increase in velocity upon TGV release. Data is averaged from three (n = 3) trials. Raw data can be found in Figure SI. 5.1.

Fig. 4

Open channel flow experiments show an increase in velocity after each valve opening. Progression of fluid flow through device with shorter distances between TGVs (a). (b) Comparison of the theoretical model (solid line) with experimental fluid front travel distance at the meniscus (diamonds) for devices with three side channels in parallel. Experimental data were averaged for four trials (n = 4). Model is presented in segments corresponding to the calculated travel distances prior to the TGV release (orange, z0), between the first and second TGV (blue, z1), between the second the third valves (yellow, z2) and between the third valve and the paper pad (red, z3). (c) Fluid velocities were calculated from the travel distance and the experimental data (diamonds) compared against the calculated model (dashed line) velocities using a dynamic contact angle model (orange, V0) for the inlet. Model is shown at the first (blue, V1), second (yellow, V2), and third (red, V3) valves due to the increase in velocity upon TGV release. Raw data can be found in Figure SI. 5.3.

Fig. 5

Fluid layers released by the TGVs stabilize in width over time. (ai) Image of the fluid layers in a parallel configuration. Circles indicate where layer thickness measurements were taken for each layer (0.5 mm after each valve). Plots of the measured layer widths over time for the parallel channel configuration after the first (aii), second (aiii), and third (aiv) valves. Data are shown as an average of three replicates (n = 3) with error bars indicating standard deviation. Layer widths are reported at a defined point in the channel (see diagram in SI. 5.7). (b) A side view image of the fluid layering after all three valves were released by undyed (clear) nonanol. Order of colored nonanol used in the side channels for part b were adjusted for visualization of the layers and are not the same as in part a.

Fig. 6

Trigger valves enable on-chip nitrite detection via the Griess reaction. (a) Schematic of device operation using the Griess reaction. Reagents 1 and 2 (R1 and R2) are added to the side reservoirs and are subsequently released by the flow of the sample with interspersed mixing steps via circular paper pads. Pads function as mixers as well as control (C) and test (T) zones. (b) Representative photograph of a device to detect nitrites in bacon samples. (c) Representative images of paper pads from Griess reaction with a 5 mg/L nitrite ion standard solution (positive control), deionized and distilled water (negative control), Lake Union (Seattle, Washington) water sample, rain runoff water sample, Seattle tap water sample, and a bacon sample. A red plus sign denotes a positive test result; A green minus sign denotes a negative test result (pink not visible by eye or spectrophotometer readout yielding a concentration below 1 mg/L). Experiments were done in triplicate (n = 3) with different devices (Figure SI. 5.9). Concentration data for the environmental and meat samples measured using the standard benchtop method can be found in Table SI. 5.1.