Mixing Performance of a Cost-effective Split-and-Recombine 3D Micromixer Fabricated by Xurographic Method

Abstract

This paper presents experimental and numerical investigations of a novel passive micromixer based on the lamination of fluid layers. Lamination-based mixers benefit from increasing the contact surface between two fluid phases by enhancing molecular diffusion to achieve a faster mixing. Novel three-dimensional split and recombine (SAR) structures are proposed to generate fluid laminations. Numerical simulations were conducted to model the mixer performance. Furthermore, experiments were conducted using dyes to observe fluid laminations and evaluate the proposed mixer's characteristics. Mixing quality was experimentally obtained by means of image-based mixing index (MI) measurement. The multi-layer device was fabricated utilizing the Xurography method, which is a simple and low-cost method to fabricate 3D microfluidic devices. Mixing indexes of 96% and 90% were obtained at Reynolds numbers of 0.1 and 1, respectively. Moreover, the device had an MI value of 67% at a Reynolds number of 10 (flow rate of 116 µL/min for each inlet). The proposed micromixer, with its novel design and fabrication method, is expected to benefit a wide range of lab-on-a-chip applications, due to its high efficiency, low cost, high throughput and ease of fabrication.

Keywords: Reynolds number; diffusion; lamination; microfabrication; microfluidics; micromixer; split and recombine.

Figure 1

An overview of the geometry of the proposed micromixer. (a) Design of the micromixer compromising of three identical mixing units (Type A); 3D view of the mixing unit is illustrated. Furthermore, cross section views of the fluid splitting and recombining zone are illustrated. (b) Other three types of mixing units sequence, which was studied in this paper. Dashed lines are used to represent the rear triangle.

Figure 2

Explosive view of the fabrication method. Three patterned pressure sensitive adhesive (PSA) layers are sandwiched between two cut PMMA substrates shaping the microfluidic device (type A). Three PSA layers built the 3D micromixer geometry, as illustrated in Figure 1. Inlet and outlet holes were suited on the upper substrate and three adhesive layers to allow fluids to enter the micromixer.

Figure 3

Calibration of the green component and equivalent gray value for different sample concentration ratios (yellow to blue dye). Images of five ratios (from pure yellow to pure blue) are illustrated on the chart. A curve was fitted to find a practical correlation, used during the image-based MI assessment.

Figure 4

Generation of sub-streams represented by flow streamlines starting from inlets (blue and red). Streamlines are calculated with regard to velocity vectors obtained from the fluid flow simulation. Total number of 500 streamlines with a uniform density of starting position is plotted for each inlet. Due to the novel geometry of the mixing units, streamlines mix together as magnified views are illustrated for: (a) T-shaped inlet, (b) after passing through first mixing unit, (c) second mixing unit and (d) third mixing unit.

Figure 5

Experimental results of the micromixer performance. (a) Macroscopic image of the proposed device, mixing blue and yellow dyes at Re = 1. The fluid layers mimicked the pattern predicted by the numerical simulations, illustrated in Figure 4. Fluid layers were generated as expected. A slight misalignment was caused during the fabrication step. Based on our experience, the device had robust output against it and performed well. (b) A straight channel was fabricated with the same channel width, height and length. It was tested with the same test condition as (a). Two phases were lightly mixed at the outlet in the absence of 3D mixing elements. (c) Microscopic images of the main channel after three mixing units for three different Reynolds numbers (Re = 0 1, 1 and 10) show that an incomplete mixing was captured at the outlet for high flow rate (Re = 10). In contrast, more efficient mixing was achieved in the case of lower fluid flow rates. Images show the region at the distance of 0.5 mm downstream of mixing units, where MI is calculated.

Figure 6

Numerical simulation of the mixing process (micromixer type A). (a) Comparison of mixing performance of the proposed design and a straight channel with same cross section size and boundary conditions (Re = 0.1). (b) Mas fraction distributions at the inlet and after each mixing unit are illustrated. Location of the planes is shown on the micromixer. (c) Plot of mass fraction along an imaginary line middle of the above-mentioned planes. Schematic position of the plotting line is illustrated.

Figure 7

Variation of the experimental and numerical MI after each mixing unit versus Reynolds number (Re = 0.1, 1, 10 and 100) in logarithmic scale. Maximum MI was achieved at lower Reynolds numbers, due to more mass transport caused by molecular diffusion.

Figure 8

Impact of geometry type of the micromixer (orientation of the sequence of three mixing units) on the mixing performance through MI measurements. Experimental data are presented for a range of Reynolds numbers and four micromixer designs, illustrated in Figure 1.