Microfluidic Channel Fabrication Process Utilizing Nd:YVO$_{4}$ Laser-Melting Technique

Ju-Nan Kuo(郭如男)$^{\hyperlink{s}{**}}$, Jian-Liang Wu(吴建良)

doi:10.1088/0256-307X/34/4/044201

⇦Chinese Physics Letters, 2017, Vol. 34, No. 4, Article code 044201⇨ Microfluidic Channel Fabrication Process Utilizing Nd:YVO$_{4}$ Laser-Melting Technique * Ju-Nan Kuo(郭如男)**, Jian-Liang Wu(吴建良) Affiliations Department of Automation Engineering, National Formosa University, Yunlin 632 Received 19 October 2016 ^*Supported by the Ministry of Science and Technology of Taiwan of China under Grant No MOST 104-2221-E-150-056
^**Corresponding author. Email: junan@nfu.edu.tw Citation Text: Kuo J N and Wu J L 2017 Chin. Phys. Lett. 34 044201 Abstract A simple and repeatable method is proposed for fabricating microfluidic channels on polydimethylsiloxane (PDMS) substrates. In the proposed approach, ridge structures with the required microchannel dimensions are formed on the surface of a borosilicate glass substrate by means of a laser-induced melting process. The patterned substrate is then used as a mold to transfer the microchannel structures to a PDMS layer. Finally, the PDMS layer is aligned with a glass cover plate and is sealed using an oxygen plasma treatment process. The proposed patterning technique is a maskless method, and is thus cheaper and more straightforward than conventional lithography techniques. Moreover, unlike direct laser ablation methods, the proposed method requires significantly less input energy, and therefore minimizes thermal effects such as substrate cracking and distortion. The feasibility of the proposed fabrication method is demonstrated by measuring the capillary filling speed of human blood plasma in microfluidic channels with cross-section sizes of $19.5\times2.5$, $17.0\times1.6$, and $7.6\times1.1$ μm$^{2}$ (width$\times$height), respectively, and temperatures of 4$^\circ\!$C, 25$^\circ\!$C and 37$^\circ\!$C. It is shown that the filling speed reduces with a reducing channel cross-section size, a lower operating temperature, and an increased filling length. DOI:10.1088/0256-307X/34/4/044201 PACS:42.62.-b, 47.61.-k, 47.55.nb © 2017 Chinese Physics Society Article Text The large surface-to-volume ratio of microscale devices prompts a strong capillary effect. Hence, capillary flows are exploited in many microfluidic systems. For example, in biochips, lengthy microchannels are often used to deliver different solutions to the reaction or analysis regions by means of capillary forces. Capillary action provides an effective means of simplifying the microfluidic system design since it eliminates the need for external accessories such as syringe pumps or electrical/centrifugal driving devices.[1-3] The literature thus contains many proposals for capillary-driven microfluidic systems. For example, Bouaidat et al.[4] proposed a capillary-driven microfluidic system consisting of two planar parallel surfaces (one entirely hydrophobic and the other mainly hydrophobic but containing hydrophilic pathways) separated by spacers. Zimmermann et al.[5] designed capillary pumps consisting of microstructures with dimensions ranging from 15 to 250 μm for regulating the flow properties of autonomous capillary systems (CSs). Kuo et al.[6] proposed a capillary-driven micromixer with a square-wave microchannel for blood plasma mixing. Microfluidic systems have significant potential for performing rapid analyses and biochemical applications on a single chip (e.g., lab-on-a-chip[7,8] and micro-total-analysis-systems (μTAS)).[9,10] However, advances in microscale patterning techniques are essential to fully exploit the potential of such systems. One of the most important components of microfluidic platforms is the fluidic channel network, which plays a critical role in many applications, including blood assays,[11,12] DNA sequencing,[13] drug delivery,[14] and single molecule detection.[15,16] The literature contains many different lithography techniques for the fabrication of fluidic channels. However, these methods are time consuming and require special equipment. Moreover, lithography techniques often require the use of a mask, which adds to the cost of the fabrication process and limits the maximum attainable spatial resolution due to the optical diffraction limit effect. To address this problem, various researchers have proposed direct laser writing methods for creating fluidic channels by ablating glass substrates[17] or thin metal films.[18] However, direct laser ablation requires the use of a high input energy, and therefore induces various undesirable thermal effects, such as a higher surface roughness, cracking, and substrate distortion. Jariwala et al.[19] proposed a two-photon absorption (TPA) polymerization-assisted ablation technique for the patterning of fluidic channels on a photoresist layer. However, the proposed method requires a post-ablation development process and induces breakdown under high-energy conditions. Processing of materials by an ultrafast laser has evolved significantly over the last decade, which can greatly reduce heat diffusion in the material.[20] However, it is a very expensive technique. Accordingly, this study proposes a simple and repeatable method for fabricating microfluidic channels, in which a Nd:YVO$_{4}$ laser is first used to pattern ridge structures on a glass substrate by means of a partial melting process and the substrate is then used as a mold to replicate microchannels with the corresponding dimensions on a PDMS layer. Notably, the dimensions of the ridge structures depend on the laser power, the scan speed and the spot size. Thus by carefully manipulating these parameters, two-dimensional microfluidic channels with the desired channel width and depth can be easily obtained. The feasibility of the proposed fabrication technique is evaluated by measuring the capillary filling speed of human blood plasma in microchannels with various cross-section sizes ($19.5\times2.5$, $17.0\times1.6$, and $7.6\times1.1$ μm$^{2}$ (width$\times$ height)) and temperatures (4$^\circ\!$C, 25$^\circ\!$C and 37$^\circ\!$C) Figure 1 presents a schematic illustration of the optical setup used to create the ridge structures on the surface of the glass substrate. As is shown, the optical system comprised a 200 mW diode-pumped frequency-doubled Nd:YVO$_{4}$ laser with a central wavelength of 532 nm, a variable neutral density (ND) filter with a power meter (PM130, Thorlabs) to control the optical intensity of the output beam, and a focusing objective lens with three different magnification settings (i.e., 10$\times$, 20$\times$ and 40$\times$) to focus the laser beam to different spot sizes. In performing the patterning process, the laser beam was maintained in a fixed position and the glass substrate was translated using a motorized $X$–$Y$ stage.

Fig. 1. Optical setup used to perform laser processing.

Figure 2 illustrates the main steps in the proposed fabrication process. The optical absorbance of glass in the visible light range is less than 10%. As a result, a large input energy is required to melt and remove the glass surface. As described above, a higher energy input leads to a rougher surface finish, a greater propensity for cracking, and a higher risk of thermal distortion. Consequently, the glass substrate was spin coated with a thin (18 μm) layer of PDMS prior to the patterning process to improve its absorption properties (Fig. 2(a)). To create the microchannels, the Nd:YVO$_{4}$ laser was tightly focused on the surface of the glass substrate (a borosilicate glass slide with a thickness of 1.1 mm) and the substrate was then translated in such a way as to produce a pattern of parallel lines with a fixed offset. In performing the patterning process, the laser was operated in the continuous wave mode and the input energy intensity was restricted to less than 180 mW through an appropriate control of the laser power (100–180 mW), scan speed (0.5–2.5 mm/s) and spot size (0.5–1.3 μm theoretically). As a result, the substrate experienced only a partial melting and evaporation effect in the heated region, giving rise to the formation of resolidified 'ridges' on the glass surface.[21] According to the atomic force microscopy (AFM, DI-3100, Veeco) observations, the average surface roughness (Ra) within the scanned area ($100\times100$ μm$^{2}$) was approximately 280 nm. Note that Lin et al.[22] proposed a very smooth etched surface that could be obtained by iterating dipping and etching procedures with ultrasonic agitation. After laser processing, the PDMS film was removed to reveal the ridge structures on the glass surface (Fig. 2(b)). The patterned glass substrate was then used as a mold to transfer the microchannel design to a PDMS layer (Fig. 2(c)). Finally, the PDMS layer was aligned with a glass cover plate and sealed using an oxygen plasma process (Fig. 2(d)).

Fig. 2. Main steps in proposed fabrication procedure for PDMS microchannels. AFM image shows the ridge structures formed on the surface of the glass substrate.

The practical feasibility of the PDMS microchip was evaluated by performing capillary filling experiments using venous blood samples collected from healthy male donors. Platelet-poor plasma was extracted from the whole blood via centrifugation at 3000 rpm for 10 min. The plasma was then drawn into the microchannel under the effects of capillary action and allowed to flow through the channel until it reached the exit. The change in position of the moving meniscus over time was captured at a rate of 30 frame/s using a charge-coupled device (CCD) camera attached to a microscope. The filling rate of the plasma was investigated under three different temperature conditions (i.e., 4$^\circ\!$C, 25$^\circ\!$C and 37$^\circ\!$C) and three different microchannel cross-section sizes, respectively. The capillary filling speed in a rectangular channel can be described theoretically by the classical Washburn model.[23] According to this model, the position of the moving meniscus $x$ as a function of time $t$ can be determined as[2] $$\begin{align} x=\sqrt {\frac{\gamma \cos (\theta )ht}{3\mu }},~~ \tag {1} \end{align} $$ where $\gamma$ is the surface tension of the liquid in air, $\theta$ is the contact angle of the liquid on the channel walls, $h$ is the channel height, and $\mu$ is the viscosity of the liquid. Figure 3 plots the width and height of the ridges formed on the glass surface for laser powers in the range of 100–180 mW, three different magnification settings of the objective lens, and a constant laser scan speed of 2.5 mm/s. As shown in Fig. 3(a), the ridge width remains approximately constant with an increasing laser power for a magnification setting of 40$\times$. However, for lower magnification settings of 10$\times$ and 20$\times$, respectively, the width increases as the laser power increases. Similarly, for a magnification setting of 10$\times$, the height also increases with an increasing laser power, as shown in Fig. 3(b). Therefore, by manipulating the laser power, scan speed and magnification setting, ridges with various widths and heights can be fabricated. In the present study, microchannels with three different cross-section sizes were prepared, namely $19.5\times2.5$ (type 1), $17.0\times1.6$ (type 2) and $7.6\times1.1$ μm$^{2}$ (type 3) (width/depth).

Fig. 3. (a) Width and (b) height of ridges as a function of the laser power given three different magnification settings (i.e., 10$\times$, 20$\times$ and 40$\times$) and a constant scan speed of 2.5 mm/s.

Fig. 4. Variation of the contact angle of plasma on cleaned flat PDMS surface over time.

PDMS is inherently hydrophobic (contact angle of water approximately 110$^\circ\!$),[24] and hence the capillary flow effect is suppressed.[25] Previous studies have shown that oxygen plasma treatment not only enables the high-strength bonding of PDMS to materials such as glass, but also induces a change in the PDMS surface properties from hydrophobic to hydrophilic.[26] Thus in the present study, the patterned PDMS substrates were cleaned using an oxygen plasma process prior to bonding with the glass cover plate to enhance the capillary flow effect. The advancing contact angle of the plasma on a flat PDMS surface was measured after cleaning using a contact angle analyzer (FTA 188, First Ten Angstroms). As shown in Fig. 4, the contact angle had a value of approximately 16$^{\circ}$. In other words, the surface treatment process yielded a significant improvement in the hydrophilicity of the PDMS surface.

Fig. 5. (a) Variation of meniscus position with filling time, and (b) variation of meniscus velocity with filling length. Note that the results are obtained under room-temperature conditions in both cases.

Figure 5(a) shows the measured position of the capillary meniscus as a function of the square root of the filling time in the three different microchannels under room temperature conditions (i.e., 25$^{\circ}\!$C). It is seen that a near perfect linear relationship exists between the meniscus position and the square root of the filling time for all three channels (i.e., $x^{2}\varpropto t$). Moreover, it is observed that the deeper channels fill more rapidly than the shallower channels ($x^{2}/t \varpropto h$). In other words, the experimental results are consistent with the Washburn model given above. The inset in Fig. 5(a) shows an experimental image of the capillary meniscus in the type-2 microchannel at time $t=1$ s. Owing to the strong capillary force, plasma was introduced into the channel without any external driving force. Figure 5(b) shows the variation of the plasma filling velocity in the three microchannels under room-temperature conditions. From inspection, the average flow velocities are 1.15, 0.97 and 0.73 mm/s in channels of types 1–3, respectively. Note that the average velocity is computed over a total channel length of 9 mm in each case. In other words, the flow rate reduces with a reducing channel cross section. Moreover, for a given cross-section size, the flow velocity reduces with an increasing filling length as a result of the greater viscous drag force. The operating temperature has a critical effect on the viscosity of plasma, and therefore also affects the flow velocity.[27] Accordingly, a series of filling experiments were performed under three different operating temperatures, namely 4$^{\circ}\!$C, 25$^{\circ}\!$C and 37$^{\circ}\!$C, respectively, where the temperatures were specifically chosen to coincide with those used in typical plasma applications. The test temperatures of 4$^{\circ}\!$C and 37$^{\circ}\!$C were obtained by cooling the chip in a fridge and heating the chip in an oven, respectively. Figure 6 shows the experimental results obtained for the flow velocity in the type-1 microchannel under each of the considered temperatures. As is expected, the flow velocity increases with an increasing temperature due to the corresponding reduction in the plasma viscosity. From inspection, the average flow velocities are found to be 0.93, 1.15 and 1.70 mm/s for operating temperatures of 4, 25 and 37$^{\circ}\!$C, respectively. Similar results were obtained for channels of types 2 and 3. Note that the results are omitted here for reasons of brevity. The average flow velocities in the type-2 channel were found to be 0.87, 0.97 and 1.47 mm/s for operating temperatures of 4, 25 and 37$^{\circ}\!$C, respectively. Meanwhile, those for the type-3 channel were found to be 0.52, 0.73 and 0.82 mm/s, respectively.

Fig. 6. Variation of meniscus velocity with filling length given different operating temperatures. Note that the results correspond to the type-1 microchannel.

In summary, this study has presented a simple, repeatable and inexpensive method for fabricating microfluidic channels via the direct laser processing of glass substrates. In contrast to existing laser ablation methods, in which material removal is achieved through a high-temperature melting and evaporation process, the present method utilizes a low-temperature partial-melting and solidification effect to create the microstructures required for microchannel replication. As a result, the thermal input to the glass substrate is significantly reduced, thereby reducing the surface roughness, suppressing crack formation, and preventing thermal distortion. Moreover, unlike conventional lithography processes, the proposed method does not require a mask (and is thus cheaper and quicker) and can be used to produce microchannel networks with any configuration. The practical feasibility of the proposed method has been demonstrated by performing plasma capillary filling experiments. The results have shown that the capillary filling speed reduces with a reducing channel cross section, a reducing operating temperature and an increasing filling length. Overall, the results presented in this study show that the proposed fabrication method has significant potential for the low-cost fabrication of microfluidic devices for a wide variety of biomedical and biochemical applications. The access provided to fabrication equipment by the Common Lab for Micro/Nano Science and Technology of National Formosa University is greatly appreciated. References Design of a Compact Disk-like Microfluidic Platform for Enzyme-Linked Immunosorbent AssayCapillary filling speed of water in nanochannelsYoung 4ever—the use of capillarity for passive flow handling in lab on a chip devicesSurface-directed capillary system; theory, experiments and applicationsCapillary pumps for autonomous capillary systemsDesign optimization of capillary-driven micromixer with square-wave microchannel for blood plasma mixingLab-on-chip technologies: making a microfluidic network and coupling it into a complete microsystem—a reviewMicrofluidic platforms for lab-on-a-chip applicationsMicro Total Analysis Systems. 1. Introduction, Theory, and TechnologyLatest Developments in Micro Total Analysis SystemsSimultaneous detection of C-reactive protein and other cardiac markers in human plasma using micromosaic immunoassays and self-regulating microfluidic networksMicrofluidic chip for blood cell separation and collection based on crossflow filtrationDNA Translocation in Inorganic NanotubesNanoengineered device for drug delivery applicationDNA Fragment Sizing by Single Molecule Detection in Submicrometer-Sized Closed Fluidic ChannelsAn Artificial Nanopore for Molecular SensingRapidly Prototyped Three-Dimensional Nanofluidic Channel Networks in Glass SubstratesDirect Laser Writing on Electrolessly Deposited Thin Metal Films for Applications in Micro- and NanofluidicsMicro-fluidic channel fabrication via two-photon absorption (TPA) polymerization assisted ablationUltrafast laser processing of materials: from science to industryViscosity of transient melt layer on polymer surface under conditions of KrF laser ablationA fast prototyping process for fabrication of microfluidic systems on soda-lime glassThe Dynamics of Capillary FlowCapillary Kinetics of Water in Homogeneous, Hydrophilic Polymeric Micro- to NanochannelsMicromolding in Capillaries: Applications in Materials ScienceRapid Prototyping of Microfluidic Systems in Poly(dimethylsiloxane)Experimental and numerical investigation of capillary flow in SU8 and PDMS microchannels with integrated pillars

[1]	Lai S Y, Wang S N, Luo J, Lee L J, Yang S T and Madou M J 2004 Anal Chem. 76 1832
[2]	Tas N R, Haneveld J, Jansen H V, Elwenspoek M and van den Berg A 2004 Appl. Phys. Lett. 85 3274
[3]	Eijkel J C T and van den Berg A 2006 Lab Chip 6 1405
[4]	Bouaidat S, Hansen O, Bruus H, Berendsen C, BauMadsen N K, Thomsen P, Wolff A and Jonsmann J 2005 Lab Chip 5 827
[5]	Zimmermann M, Schmid H, Hunziker P and Delamarche E 2007 Lab Chip 7 119
[6]	Kuo J N, Liao H S and Li X M 2017 Microsys. Technol. 23 721
[7]	Abgrall P and Gué A M 2007 J. Micromech. Microeng. 17 R15
[8]	Haeberle S and Zengerle R 2007 Lab Chip 7 1094
[9]	Reyes D R, Iossifidis D, Auroux P A and Manz A 2002 Anal. Chem. 74 2623
[10]	Arora A, Simone G, SaliebBeugelaar G B, Kim J T and Manz A 2010 Anal. Chem. 82 4830
[11]	Juncker D, Michel B, Hunziker P and Delamarche E 2004 Biosens Bioelectron 19 1193
[12]	Chen X, Cui D F, Liu C C and Li H 2008 Sens. Actuators B 130 216
[13]	Fan R, Karnik R, Yue M, Li D, Majumdar A and Yang P 2005 Nano Lett. 5 1633
[14]	Sinha P M, Valco G, Sharma S, Liu X and Ferrari M 2004 Nanotechnology 15 S585
[15]	Foquet M, Korlach J, Zipfel W, Webb W W and Craighead H G 2002 Anal. Chem. 74 1415
[16]	Saleh O A and Sohn L L 2003 Nano Lett. 3 37
[17]	Ke K, Jr Hasselbrink E F and Hunt A J 2005 Anal. Chem. 77 5083
[18]	Lorenz R M, Kuyper C L, Allen P B, Lee L P and Chiu D T 2004 Langmuir 20 1833
[19]	Jariwala S, Tan B and Venkatakrishnan K 2009 J. Micromech. Microeng. 19 115023
[20]	Malinauskas M, Zukauskas A, Hasegawa S, Hayasaki Y, Mizeikis V, Buividas R and Juodkazis S 2016 Light-Sci. Appl. 5 e16133
[21]	Weisbuch F, Tokarev V N, Lazare S and Débarre D 2002 Surf. Sci. 186 95
[22]	Lin C H, Lee G B, Lin Y H and Chang G L 2001 J. Micromech. Microeng. 11 726
[23]	Washburn E W 1921 Phys. Rev. 17 273
[24]	Jeong H E, Kim P, Kwak M K, Seo C H and Suh K Y 2007 Small 3 778
[25]	Kim E, Xia Y N and Whitesides G M 1996 J. Am. Chem. Soc. 118 5722
[26]	Duffy D C, Olivier J C M, Schueller J A and Whitesides G M 1998 Anal. Chem. 70 4974
[27]	Saha A A, Mitra S K, Tweedie M, Roy S and McLaughlin J 2009 Microfluid. Nanofluid 7 451