Continuous Flow with Reagent Injection on an Inlaid Microfluidic Platform Applied to Nitrite Determination

Abstract

A continuous flow with reagent injection method on a novel inlaid microfluidic platform for nitrite determination has been successfully developed. The significance of the high-frequency monitoring of nutrient fluctuations in marine environments is crucial for understanding our impacts on the ecosystem. Many in-situ systems face limitations in high-frequency data collection and have restricted deployment times due to high reagent consumption. The proposed microfluidic device employs automatic colorimetric absorbance spectrophotometry, using the Griess assay for nitrite determination, with minimal reagent usage. The sensor incorporates 10 solenoid valves, four syringes, two LEDs, four photodiodes, and an inlaid microfluidic technique to facilitate optical measurements of fluid volumes. In this flow system, Taylor-Aris dispersion was simulated for different injection volumes at a constant flow rate, and the results have been experimentally confirmed using red food dye injection into a carrier stream. A series of tests were conducted to determine a suitable injection frequency for the reagent. Following the initial system characterization, seven different standard concentrations ranging from 0.125 to 10 µM nitrite were run through the microfluidic device to acquire a calibration curve. Three different calibrations were performed to optimize plug length, with reagent injection volumes of 4, 20, and 50 µL. A straightforward signal processing method was implemented to mitigate the Schlieren effect caused by differences in refractive indexes between the reagent and standards. The results demonstrate that a sampling frequency of at least 10 samples per hour is achievable using this system. The obtained attenuation coefficients exhibited good agreement with the literature, while the reagent consumption was significantly reduced. The limit of detection for a 20 µL injection volume was determined to be 94 nM from the sample intake, and the limit of quantification was 312 nM. Going forward, the demonstrated system will be packaged in a submersible enclosure to facilitate in-situ colorimetric measurements in marine environments.

Keywords: environmental monitoring; lab on chip; microfluidics; nitrite; ocean sensors.

Figure 1

(A) Concept model for the improved inlaid absorbance cell, with light coming from an external source and directed through the cell to a detector. (B) CAD of the microfluidic chip for continuous flow analysis. The interfacing ports to the syringes are labeled R1, R2, P1 and P2, corresponding to the syringes in panel (C). The engraved microprisms are at the center of the inlaid circular apertures. (C) Fluid diagram for continuous flow on a microfluidic system. Syringes R1 and R2 contain the reagent that is periodically injected into the fluid stream in pre-programmed volumes. Syringes P1 and P2 alternate between pulling fluid through the system from either the standard or sample port and pushing fluid out through the waste ports.

Figure 2

(A) Photograph of the bench-top testing setup for the continuous flow system. Four individually programmed Tecan syringe pumps were used to move fluid through the system, and ten solenoid valves were mounted to the testing apparatus to enable fluid control. A separate custom-made selector manifold was used to switch between fluids for automated calibrations. A 3D-printed alignment piece was used to hold the LEDs and photodiodes in position above the source and detector prisms, respectively. All electronics were controlled by a microcontroller programmed in C. (B) Photograph of the microfluidic chip with circular inlays. Each absorbance cell contained two inlays, one at the detector side and one at the source side. Three asymmetrical alignment holes were used to ensure correct orientation and positioning within the testing apparatus or in situ instrument. (C) Electrical block diagram for the testing apparatus. An STM32 BlackPill microcontroller and a custom PCB were used to automate and coordinate the electronic components of the system. Three solenoid drivers were used to control up to 24 solenoid valves, with only 10 valves used for this work. An external PC was used to create scripts and program the microcontroller to test configurations.

Figure 3

(A) Dispersion of a 100 µL injected dye plug after traveling 1.01 m at 100 µL min⁻¹. The normalized concentration from a simulated plug that has undergone dispersion based on the Taylor–Aris simulation model is overlaid onto the absorbance signal recorded during the dispersion of a 100 µL plug of red dye. (B) Dispersion of a 50 µL injected dye plug after traveling 1.01 m at 100 µL min⁻¹. The simulated signal is overlaid onto the recorded signal during a 50 µL injection. (C) Dispersion of a 10 µL injected dye plug after traveling 1.01 m at 100 µL min⁻¹. The simulated signal is overlaid onto the recorded signal during a 10 µL injection. (D–G) Voltage recordings from multiple 50 µL dye injections into a 100 µL min⁻¹ carrier stream. The four signals represent four different timing intervals between the end and start of each new injection: 30 s, 60 s, 80 s and 100 s. The area marked by red dashed lines shows the disturbance in the flow caused by the opening and closing of valves and the switching of directions of the syringes.

Figure 4

Voltage measured by the photodiode during automated nitrite calibrations with nitrite concentrations listed in µM. The areas shaded in grey represent the signal obtained during four consecutive reagent injections into a carrier stream composed of sequential standard fluids, with nitrite concentrations ranging between 0.125 µM and 10 µM. The areas shaded in white represent the rinsing between standards, where Milli-Q is flushed through the system. Panel (A) contains the filtered voltage from the calibration performed with 50 µL of reagent, while panels (B,C) contain the filtered voltages from the calibrations performed with 20 µL and 4 µL of reagent, respectively. The areas enclosed by dashed red lines correspond to the profiles shown in Figure 5.

Figure 5

(A) Voltage profile for a single 50 µL reagent injection into a 10 µM carrier stream. The visible dual peaks are a result of an incomplete reaction in the center of the plug upon reaching the absorbance cell. (B) Voltage profile for a 50 µL reagent injection into a 0.125 µM carrier stream. The initial increase followed by a decay in voltage is a result of the Schlieren effect. The corrected absorbance profile for a single reagent injection into a 10 µM (C) and 0.125 µM (D) carrier stream. Dual peaks are still visible, but the decay and rise in signal due to the Schlieren effect are removed. (E) Absorbance profile of 4 consecutive 50 µL reagent injections into Milli-Q carrier stream. The rises and decays in the absorbance are the result of the Schlieren effect.

Figure 6

Corrected absorbance signals calculated by subtracting the averaged values of blank absorbance from the sample absorbance. A MATLAB (R2021b) script is used to integrate the absorbance signal produced by each injection, shown as the shaded colored areas between each injection peak and the time axis. Panel (A) contains the absorbance signal from the calibration performed with 50 µL of reagent, while panels (B,C) contain the absorbance signal from the calibrations performed with 20 µL and 4 µL of reagent, respectively.

Figure 7

Panel (A) calibration curves produced using the continuous flow system, with 4 µL, 20 µL and 50 µL reagent injection volumes. The absorbance profile for each of the four consecutive reagent injections for every standard fluid is integrated and averaged to produce a resultant value, which is plotted against the final nitrate concentration. The linearity of the calibration slope for each reagent injection volume is displayed alongside the slope value and dependent variable offset. Panel (B) displays a closer view of the data sets with nitrite concentrations below 1 µM. Panel (C) shows the residual plot of the linear regressions. Measurements were performed at room temperature or 21–23 °C.