Spontaneous diffusiophoretic separation in paper-based microfluidic device

Microfluidic paper-based analytical devices (μPADs) for separating particles have been playing a key role for point-ofcare diagnostics in the area of remote settings. While splendid separation methods using μPADs have been explosively developed, they still require external devices inducing external field. In this work, the spontaneous separation method in μPADs was suggested by leveraging convective flow (the imbibition of paper and nanoporous medium) and diffusiophoresis by ion exchange medium. Especially, the paper’s fast imbibition was utilized as driving particles at the first stage, which results in fast overall processing in contrast to the spontaneous separation method of microfluidic chip integrated with only ion exchange medium. Therefore, our novel spontaneous selective preconcentration method based on μPADs would have key potential to be used in portable point-of-care devices in remote settings.


Introduction
Microfluidic paper-based analytical devices (μPADs) have received significant attentions in recent years due to the properties of paper such as low cost, portability, bio-degradability, handy portability and simple fabrication [1][2][3]. These properties make the paper-fluidic device attractive in remote settings and resource limited settings. μPADs have been studied for point-of-care diagnostics in remote settings, in which diverse methods have been developed for separation and preconcentration of samples [4][5][6][7]. Recently, separation and preconcentration of biomolecules has been demonstrated with paper-based ion concentration polarization (ICP) platform. Since conventional ICP preconcentration method has several benefits such as high preconcentration factor and no need of complex buffer exchange, ICP has been extensively studied for the selective preconcentration of biomolecules [8][9][10][11][12]. However, because typical ICP platform required an external electrical field for performing the separation, paper-based ICP separator also needed an external power source and cause the methods to be against to the original concept of μPADs.
Our previous work presented the spontaneous preconcentration (or selective preconcentration) method in microfluidic chip with leveraging convective flow through microchannel over diffusiophoresis induced by imbibition and ion exchange of ion exchange medium, respectively [13][14][15][16][17][18]. Especially, we had demonstrated that the velocity of the convective flow by the imbibition of the medium deviated from original Darcy's law using non-uniformly patterned ion exchange medium, but this work has the limitation of time-consuming and complex process [13]. Therefore, in this work, we would present the spontaneous diffusiophoretic separation of multiple particles using μPADs. We analyzed the results using the mechanism of previous works such as diffusiophoresis and imbibition, while a part of result was found to be contradictory to such mechanisms so that one need further investigations for the exact fundamentals. However, presenting results can be directly employed for practical μPAD based selective preconcentration platform. Lee and Kim Micro and Nano Syst Lett (2020) 8:6 Materials and methods

Device fabrication
Paper-based separation device was fabricated by cellulose paper (Whatman grade 1, Sigma Aldrich, USA) with slide glass, as shown in Fig. 1a. The paper has 180 μm of thickness and a mean pore diameter of 11 μm. Using a commercial wax printer (ColorQube 8570, Xerox, USA), the channel was pattered with hydrophobic wax region (black area in Fig. 1b). The printed channel was then baked at 120 °C for 10 min. Then, Nafion was dropped in a broad region of paper for forming ion exchange medium as shown. The Nafion-patterned paper was dried at 95 °C for 10 min. Finally, to avoid sample evaporation, adhesive tapes were attached on both sides of the paper channel.

Experimental apparatus
The motions of fluorescent particles were captured by an inverted fluorescence microscope (IX53, Olympus, Japan) and the images were analyzed by CellSens program (Olympus, Japan). The snapshots were captured at every 5 min.

Mechanism of spontaneous separation method in μPADs
When ion exchange medium (e.g. Nafion) met water containing non-protonic cation, the protons inside the medium were exchanged with other cations, and the concentration gradient was generated due to the different diffusivity of proton and other cations as shown in Fig. 2a. This mechanism is called diffusiophoresis, which spontaneously induces the motion of charged particles under the electrolyte concentration gradient. In this configuration, the particles are affected by diffusiophoretic velocity (U DP ) and also the velocity of convective flow through microchannel (U μ ) with the opposite direction to U DP . This convection can be caused by the imbibition of ion exchange medium and paper itself.
Both velocities (U DP and U μ ) are proportional to the time −1/2 [14]. Thus, they were plotted as a straight line with a negative slope in log scale graph as shown in Fig. 2b. U DP of a particle depends on diffusiophoretic constant (D DP ) as a function of zeta potential and radius of particle, etc. However, we used paper-based device to enhance the efficiency of separation time unlike our previous work which has the limitation of slow processing (6-12 h) [13], leading to a complicated behavior of U μ in Fig. 2b. When a paper contacts with water, capillary forces created a fast fluid flow into the paper. Thus, U μ induced by imbibition of paper could be plotted as a straight line above any other U DP s, meaning the fastest velocity (dotted line) before the time of finishing imbibition of paper (t p ) as shown in Fig. 2b. Thus, particle 1 and 2 would move to the Nafion by U μ . After t p , the imbibition by paper was stopped and the imbibition of Nafion only affected U μ and, as our previous work suggested, U μ was deflected due to the non-uniformly patterned medium at the critical time of convective flow (t c ) [13]. Then, before the time of separating particles (t s ), the particles would move away from Nafion, but each particle The plot of U DP s of particle 1 and 2 with different diffusiophoretic constant, and U μ induced by the imbibition of paper and Nafion would move as the opposite direction after t s , which implies that two types of particles with different diffusiophoresis constant would be separated in μPADs. Following section will experimentally demonstrate this scenario.

Experimental demonstration of spontaneous separation
As shown in Fig. 3, two particles with different size were spontaneously separated according to the aforementioned mechanism in Fig. 2. Figure 2 showed that the competition between diffusiophoresis and the imbibition of nanoporous medium lead to spontaneous separation in the paper-based platform. In Fig. 3a, U μ induced by paper's fast imbibition drove multiple particles from reservoir to the region near the Nafion (see green region of image at t = 0 min). After 2 h, the particles were almost completely separated and preconcentrated. The separation resolution (R s ) calculated by (peak to peak distance)/ (average width of bands) under Gaussian distribution assumption [19] led to 1.25 which showed reasonable separation. Compared to our previous work that utilized only microfluidic channel-based device spent 6 h to the completion, we can conclude that the separation in paper-based device was performed faster. Furthermore, the separated particles can be observed with naked eye using fluorescent lamp even after 1 days as shown in Fig. 3b. In this image, soaked water in the paper was completely evaporated. Although the paper has random cellulose network, the fast convective flow induced by imbibition of the paper is obvious in the direction to the nanoporous medium, so that the process time might be random, but the separation phenomenon would finally occur.
However, previous literature suggested that D DP of larger particle is usually larger than that of small particle. This meant that 0.04 μm particle should be accumulated near Nafion and 0.2 μm particle should be repelled from Nafion. This is contradictory to our present observation [13] and it may come due to the complex geometry of cellulose and interaction of particles with it. In the random (or natural) pore network, various convective flow such as recirculation can be included in the analysis [20][21][22][23]. Thus, one need further investigations for the exact fundamentals in paper-based device.

Conclusions
In this work, the spontaneous particle separation method by diffusiophoresis in μPAD platform was demonstrated without any external power sources. The separation was leveraged by convective flow induced by imbibition of ion exchange medium over diffusiophoresis. To overcome the slow processing of our previous work [13], the paper network was designed for utilizing the fast imbibition of paper and, thus, the faster process time was achieved as expected. Although this method has still slow process time to apply in practical cases, this method has an advantage of no need for external power source, so that this method can be useful for separating external power-sensitive bio-samples and a time-insensitive lab on a chip application such as environmental monitoring and food monitoring. While separation order was different from previously reported mechanisms so that one Fig. 3 a Time-evolving microscopic images of spontaneous diffusiophoretic separation and b the separated particles captured by regular cellphone camera need further investigations for the exact fundamentals, presented results can be directly employed for practical μPAD based selective preconcentration platform and one can expect the applications of this work for the portable diagnosis devices without any external power sources.