Multi-wavelength microflow cytometer using groove-generated sheath flow

Abstract

A microflow cytometer was developed that ensheathed the sample (core) fluid on all sides and interrogated each particle in the sample stream at four different wavelengths. Sheathing was achieved by first sandwiching the core fluid with the sheath fluid laterally via fluid focusing. Chevron-shaped groove features fabricated in the top and bottom of the channel directed sheath fluid from the sides to the top and bottom of the channel, completely surrounding the sample stream. Optical fibers inserted into guide channels provided excitation light from diode lasers at 532 and 635 nm and collected the emission wavelengths. Two emission collection fibers were connected to PMTs through a multimode fiber splitter and optical filters for detection at 635 nm (scatter), 665 nm and 700 nm (microsphere identification) and 565 nm (phycoerythrin tracer). The cytometer was capable of discriminating microspheres with different amounts of the fluorophores used for coding and detecting the presence of a phycoerythrin antibody complex on the surface of the microspheres. Assays for Escherichia coli were compared with a commercial Luminex flow cytometer.

Fig. 1

Chevrons and sheath flow generation. Core fluid was surrounded by sheath fluid by first sandwiching the core stream between the two sheath streams via fluid focusing. The chevrons perform the final sheathing function by moving some of the sheath fluid to above and below the core.

Fig. 2

The PDMS module was 3.17 × 5.71 cm and had connections for optical fibers and fluidics as shown in the figure. Excitation fibers (bottom) were connected to the 635 nm and 532 nm lasers. Two multimode beam dump fibers were inserted across the channel from the excitation fibers (upper left) to capture excess excitation light and reduce scatter. The microsphere identification fiber (upper right) and the phycoerythrin/light scatter fiber (lower left) were each connected to a multimode fiber splitter. Each leg of the splitters was connected to a PMT as shown in the diagram.

Fig. 3

Simulation of chevron sheathing using COMSOL. The core fluid and sheath fluid enter the channel simulation from the left and flow at rates of 10 μL/min and 100 μL/min per channel, respectively. To reduce computation time, only the top half of the channel (above the dotted line) was modeled. (a) top view of channel. (b) side view of section down the middle of the channel showing chevron locations (c) sheathed core fluid as viewed at cross-section “A” (enlarged). For (b) and (c) the top half of the channel was combined with its mirror image to form a full image of the channel.

Fig. 4

Alignment of fibers in the interrogation region. To observe the alignment of the optical fibers with the center of the flow channel, each fiber was connected to a laser while the channel was filled with Cy5 dye solution and photographed. Three images were combined to show the overlap of the output of the single mode 635 nm excitation fiber (bottom), the phycoerythrin excitation fiber (bottom right) and the acceptance angle of the microsphere identification fiber (top right).

Fig. 5

Microsphere identification plot. (a) Shown are the coded microsphere sets 50 (○), 73 (△), 77 (▽), and 81 (□). Each point represents the pulse area normalized by its width for the 665 and 700 nm microsphere identification wavelengths detected for each pulse that was over threshold. (b) Data from (a) normalized using the light scatter signal for each microsphere. Set 77 had phycoerythrin attached to the microspheres.

Fig. 6

Phycoerythrin detection results. Shown is the mean and standard deviation of the phycoerythrin signal for each microsphere set tested. Signal was the area under each pulse normalized by the pulse width.

Fig. 7

Microsphere identification plots for E. coli assays. Shown are the coded microsphere sets 50 (○), 73 (△), 81 (□) and 98 (▽), for the 10⁷ cfu/ml sample on the (a) Microflow cytometer. Bead identification signal was pulse area normalized by scatter. Microsphere set 81 had anti-E.coli attached to the microspheres (b) bead identification from the Luminex for the same microsphere sets.

Fig. 8

Results from E. coli assays. The fluorescence pulse from each microsphere was integrated and normalized by pulse width. Shown are the mean and standard deviation for the fluorescence signals from microsphere sets 50 (black), 73 (medium gray), 81 (dark gray) and 98 (light gray) for each concentration tested. Microsphere set 81 (dark gray) had anti-E.coli attached to the microspheres.

Fig. 9

The E. coli assays were performed on the microflow cytometer and Luminex systems. Shown is the mean and standard deviation of the (a) fluorescence signal integrated and normalized by pulse width from the microflow cytometer and (b) results from the Luminex system using the same samples.