Fabrication of glass-based microfluidic devices with dry film photoresists as pattern transfer masks for wet etching

Abstract

A simple, cheap and rapid method is developed to fabricate glass-based microfluidic devices with dry film photoresists (DFR) as pattern transfer masks for wet etching. In this method, the DFR mask for wet etching can be easily achieved by a one-step lamination, and no expensive facilities and materials are used; therefore, both the difficulty and the cost of fabrication of glass microchips with etched microchannels are reduced greatly compared with those in conventional methods. With the DFR mask, mass-production of glass microchips can be achieved efficiently and controllably in general laboratories. The fabricated glass microfluidic devices feature very flexible design of microchannels, good chemical compatibility and optical properties, easy modification of channel surface wettability, mass producibility and satisfactory reproducibility. We demonstrate the utilities of fabricated glass microchips in the preparation of monodisperse water-in-oil (W/O) and oil-in-water (O/W) emulsions, and the formation of a stable laminar flow interface and concentration gradient in microchannels.

---

Introduction

Microfluidic technologies, which are also called lab-on-a-chip (LoC) systems, have developed to be among the most effective methods for manipulation of fluids in channels with dimensions of tens to hundreds of micrometres.¹ Due to their miniaturization, easy integration and reduced reagent consumption,^2,3 microfluidic technologies have been applied to a broad range of applications varying from biological to chemical fields,^4–8 such as cell biology,^9,10 medical diagnose,¹¹ microreactors,^12–14 chemical synthesis,^15,16 preparation of microparticles.^17–20 For achieving efficient performances in various applications, microfluidic devices are usually expected to be optically transparent, mechanically strong, thermally stable, easily modifiable, flexibly designable, and simply and cheaply available for mass production. However, how to combine these features effectively in one chip simultaneously is still a challenging issue. Partly, that is why LoC systems are still not being applied extensively.^21,22

Currently, most of the microfluidic devices are fabricated of plastic, silicon and glass. Due to the characteristics of flexible designing and precise size controlling of channels based on soft lithography method, polydimethylsiloxane (PDMS) has become one of the most commonly used plastic materials in the microfluidic community.^23,24 Unfortunately, the somewhat complex and costly fabricating processes of PDMS devices as well as the fundamental incompatibilities of PDMS with many solvents have limited the wide applications of PDMS microfluidic devices to some extents.²⁵ Silicon microfluidic devices have good chemical compatibility and excellent thermal stability.^26–28 However, besides expensive and fragile, silicon is opaque to both visible and ultraviolet light so that it can not be used in conventional optical applications.¹ Glass is a cheap material that widely used in LoC systems because of many desirable properties, such as optical transparency, chemical inertness, rigidity, high-temperature resistance, and easy modification.^29,30 Assembly of glass capillaries^16,31–33 as well as patterned cover slips³⁴ on microscope glass slides and wet chemical etching of channels on glass plates^35–41 are usually used for fabrication of microfluidic devices. Among them, glass capillary devices and glass patterned cover slip devices can be easily fabricated in general labs; however, assembly of glass capillaries and assembly of patterned cover slips on microscope glass slides usually require skilled manual operation, which reduces the efficiency of device preparation and the precision of device dimensions. Chemical wet etching method combined with photolithography can make a flexible designing of microchannel with a high precision, and thus shows less dependence on manual operation. To date, chemical wet etching has been verified to be a good method for fabricating glass microfluidic devices. The conventional masks for the chemical wet etching in glass are usually manufactured by deposition of metal layers³⁶ and spin-coating of photoresists^38,39 onto glass substrates. However, the preparation of both metal layers and photoresist layers is an extremely time-consuming and expensive process,⁴⁰ which limits the application of chemical wet etching method in the preparation of glass microfluidic devices to a large degree. For this reason, people continue to seek simple and low cost approaches to prepare wet etching masks for fabrication of glass microfluidic devices. In this way, Coltro and co-workers⁴⁰ developed a method using toner layers as mask for wet chemical etching, in which the device layouts were printed on wax paper and transferred onto glass surfaces by a hot press. Santana and co-workers⁴¹ reported a method for fabricating glass microchannels by wet chemical etching using masks made by xurography in vinyl adhesive films, in which the patterns of microchannels were transferred to the vinyl adhesives using a cutting plotter. Rodriguez and co-workers³⁷ developed an alternative method relies on transferring patterns of microchannels in PDMS onto glass by using PDMS molds, in which the PDMS molds were used to confine the etching solution and then the patterns were etched on the glass; however, this process is still dependent on construction of a silicon masters using photolithography. Therefore, development of a simple, cheap and rapid method for fabricating wet etching masks is still very important and essential for construction of microfluidic devices that are optically transparent, mechanically strong, thermally stable, easily modifiable, flexibly designable, and simply and cheaply available for mass production.

Here we report on a simple, cheap and rapid method for fabrication of glass microfluidic devices with dry film photoresist (DFR) as pattern transfer mask for wet etching. DFR is a cheap and widely used pattern transfer material in printed circuit board (PCB) technologies.⁴² Recently, DFRs have been reported to be useful for applications where sub-micro resolution is not required, such as sealing microfluidic devices,⁴³ constructing microchannels,^44–47 encapsulating nanochannels,⁴⁸ and fabricating electroplating moulds;⁴⁹ however, there has been few report about using DFRs as pattern transfer masks for mass production of glass microfluidic devices. Compared with the masks made of liquid photoresist or metal, DFRs exhibit good chemical resistance in HF based solution and excellent adhesion to glass substrates,^44,48 which are very important for wet etching process. In this work, we use the DFR as the pattern transfer masks for wet etching to fabricate glass microfluidic devices for the first time. The network of microfluidic channels is designed by AutoCAD software flexibly. Microfluidic devices are fabricated by a one-step lamination of DFR onto glass substrates, simplified photolithography operations, wet etching process using HF solution and rapid UV-bonding. Compared with the conventional etching masks that fabricated by deposition of metal layers and spin-coating of photoresists, only a one-step lamination operation is needed in our method to immobilize the DFR onto glass substrate to form an etching mask, which greatly reduces the difficulty and cost for fabrication of glass microfluidic devices. Moreover, the DFR can be easily laminated onto glass substrate in a large area due to the avoidance of complex processes of spin-coating and deposition. Importantly, Simple and cheap mass-production of glass microfluidic devices is potentially available by combining this DFR mask method with the mature manufacturing process of PCB in the electronics industry, while the conventional methods of spinning of liquid photoresists or deposition of metal layers are not competent. The fabricated glass microfluidic devices are featured with very flexible design of microchannel networks, easy spatial patterning of channel surface wettability, and good chemical compatibility and optical properties. We demonstrate their utility in various applications such as generation of monodisperse water-in-oil (W/O) and oil-in-water (O/W) emulsions, parallel laminar flow interface, and concentration gradient in microchannels. We believe that the method proposed in this work provides an efficient way to fabricate the glass microfluidic devices and is significant to the development of microfluidics.

Materials and methods

Fabrication of glass microfluidic devices

The glass devices are fabricated by simplified photolithography and chemical wet etching (Fig. 1). The commercially available soda-lime microscopic glass slides (76.2 mm × 25.4 mm, thickness of 1–1.2 mm) and the glass sheets with bigger dimensions (178 mm × 59 mm, thickness of 2–2.2 mm) are used to fabricate the glass chips (Fig. 1a). Both kinds of glass substrates are washed using an ultrasonic cleaner with acetone, ethanol and deionized (DI) water (18.2 MΩ, Milli-Q, Millipore) sequentially, and then dried in an oven at 60 °C for 2 h. The glass microfluidic devices are fabricated with the following procedures:

Fig. 1 A schematic illustration of the process for fabricating glass microfluidic device by using dry film photoresist (DFR) and soda-lime glass slide. (a) Glass substrate. (b) Laminating the DFR on the glass substrate using an office laminator. (c) UV exposing with a printed photomask on the DFR. (d) Developing and transferring the microchannel patterns onto the DFR. (e) Wet etching and DFR striping for transferring microchannel patterns onto the glass substrate. (f) Bonding the cover glass with drilled holes onto the glass substrate with etched microchannels using UV-curable glue.

(1) Laminating the DFR on the glass substrate using an office laminator (Fig. 1b). The DFR (Etertec HT-115T, Eternal Chemical Co., Ltd) with a line resolution of 50 μm is laminated onto the cleaned glass substrate by using an office PCB laminator (KH320C, Fuzhou Kehai Electronic Co., Ltd) at 95 °C, with the help of a droplet of water between the DFR and the glass to eliminate the air bubbles between the DFR and the glass effectively.

(2) UV exposing with a printed photomask on the DFR (Fig. 1c). The microchannel patterns on photomasks are designed using software AutoCAD and printed on a transparent film using a high-resolution laser printer (25400 dpi). UV exposure with the printed photomask on the DFR is carried out using a low-cost UV exposure box (KVB30D, Kinsten Pty Ltd).

(3) Developing and transferring the microchannel patterns onto the DFR (Fig. 1d). The designed structure is then developed by spraying an aqueous solution of 1.43% (w/v) sodium carbonate onto the exposed area in less than 2 min at 30 °C. The samples are rinsed with DI water to remove the sodium carbonate and dried by a hair drier. After that, the samples are putted on a hotplate at 65 °C for 1–2 h to remove residual water molecules in DFR layer to increase the adhesion of DFR to the glass substrate. All the DFR operations are carried out under yellow light.

(4) Wet etching and DFR striping for transferring microchannel patterns onto the glass substrate (Fig. 1e). After baking the DFR coated glass substrates, the substrates are putted in an etch tank (ET10, Kinsten Pty Ltd) containing buffered oxide etchant (BOE) solution. The solution temperature is controlled by a thermostated water bath, and a K-type thermocouple is used to monitor the liquid temperature in the etch tank. In order to remove the precipitated particles that formed while etching soda-lime glass in a HF-containing solution, 20 vol% concentrated HCl is added in BOE solution (NH₄F (40 wt%) : HF (49 wt%) = 7 : 1) to achieve a smoother etched surface by referring to previously published works.^35,38,50,51 The BOE solutions are diluted with DI water in various ratios to obtain different concentrations. After the etching process, the glass substrates are ultrasonically treated to peel off the DFR layer, and then rinsed with acetone, ethanol and DI water in an ultrasonic bath for 5 min. Finally, the glass substrates are dried with a hair drier. The morphology of the etched channels is observed by a scanning electron microscope (SEM, G2 Pro, Phenom).

(5) Bonding the cover glass with drilled holes onto the glass substrate with etched microchannels using UV-curable glue (Fig. 1f). Prior to sealing the chips, the glass slides are cut into small pieces (25 mm × 25 mm) to fit the dimensions of the etched glass substrates. Then, access holes with diameter of Ø1.1 mm are drilled on the glass slides at the corresponding points using a ultrasonic drilling machine. The glass slides are ultrasonically washed with DI water and acetone to remove the drilling dust. Subsequently, the glass chips are sealed using UV-curable glue according to a previously published method.³⁴ Briefly, the cover glass slide containing access holes and the glass substrate with etched microchannels are aligned together and then an appropriate amount (20–40 μl) of UV-curable adhesive (1018, Ruikang Science and Technology Co., LTD) is introduced slowly into the gaps between the glass plates using a syringe needle. The UV-curable glue enters automatically into the gaps between glass plates due to the capillary force. The position of the UV-curable adhesive solution between the two glass plates is observed in real time using a microscope. In addition, to prevent the UV-glue from entering into the microchannels, a swab is used to remove the redundant UV-glue from the gaps between glass plates. The chips are placed under a UV lamp for 30 s to cure the glue completely when the adhesive solution reaches the edges of the microchannels. Finally, polyethylene (PE) pipes (Ø1.09 mm × 0.38 mm) are connected to the outlet and inlets respectively by epoxy resin (no. 14250, Devcon®).

Performance test of glass microfluidic chips

Leakage test of the chips is carried out by using a simplified pressure test devices according to a previously published method.³⁴ After the pressure test, 1 mol l⁻¹ NaOH, 1 mol l⁻¹ HCl and acetone are respectively injected into the microchannels using syringe pumps (PHD 2000, Harvard Apparatus) at 2000 μl h⁻¹ for 10 min. Then, the microchannels are washed with DI water for several minutes and are filled with methylene blue dye for observation of leakage.

Microfluidic preparation of monodisperse emulsions

To demonstrate the flexibility of modification, the channels are modified spatially by self-assembled monolayer (SAM) method.⁵² We make the channels hydrophobic by injecting 10 vol% chlorotrimethylsilane ((CH₃)₃SiCl (CTMS), Chengdu Kelong Chemical Reagents Co., Ltd) in hexane into the devices at 0.5–1 ml min⁻¹ for 2 min. To make the glass channels much more hydrophilic, 10 vol% HF (Chengdu Kelong Chemical Reagents Co., Ltd) is introduced into the device at 0.5–1 ml min⁻¹ for 2 min. All the modified channels are slightly rinsed with DI water after modification. A hydrophobically modified T-junction device is used to prepare W/O emulsions, while a hydrophilically modified cross-junction device is used to prepare O/W emulsions. Mixtures of benzyl benzoate (BB) (Sinopharm Chemical Reagent Co., Ltd) and soybean oil (SO) (Kerry Oils & Grains Co., Ltd) are used as the oil phases, and 5–10% (w/v) glycerin aqueous solutions are used as the aqueous phases. 2–5% (w/v) polyglycerol polyricinoleate (PGPR 90) (Danisco) and 1% (w/v) Pluronic F-127 (Sigma-Aldrich) are used as lipophilic and hydrophilic surfactants respectively. Syringes and syringe pumps (PHD 2000, Harvard Apparatus) are used for driving fluids into the microchannels. The formation processes of emulsion droplets are recorded by a high-speed camera (Mico3, 1024MC, Phantom). The freshly prepared W/O emulsions containing Eosin-Y dye and the O/W emulsions containing Lumogen® F Red 300 (LR300, BASF) are observed by a fluorescent microscope (DM 4000, Leica) under I3 filter set (BP 450–490) and A filter set (BP 340–380) respectively. All the experiments are carried out at room temperature (22 °C).

Laminar flow operations in microchannels

A Y-junction device and a concentration gradient generator device are used to illustrate the utilities of our chips under laminar flows. Specially, the Y-junction device is used for formation of the parallel laminar flow interface in microchannels that can be used for microfabrication or synthesis based on interface reactions,^53,54 and the concentration gradient generator device is used to generate the molecular gradients in the microchannels that can be used for cell biology studies.⁵⁵ Fluids are infused into the devices using the syringe pumps with a flow rate of 500 μl h⁻¹. Sudan red and methylene blue are used as dyes to prepare solutions with red and blue colours. The laminar flows in microchannels are observed by an optical microscope (BX 61, Olympus).

Results and discussion

Performance of DFR as a pattern transfer mask for wet etching

The mask preparation for wet chemical etching is very important for fabricating the glass microfluidic devices, because it heavily affects the wall roughness and structure dimensions of etched microchannels. In particular, the adhesion of mask to substrate surface is extremely important during the wet etching process. With DFRs as masks for wet etching of glass microchannels in this study, the wet etching processes are satisfactorily controllable.

The survival time of DFR mask and the dimensions of the etched channels in HF-based etching solutions with different concentrations are studied at varied temperatures. When the temperature of the buffered oxide etchant (BOE) solution increases, the average etching rate of the soda-lime glass increases, but the survival time of the DFR mask decreases (Fig. 2a). Different from the case using spin-coated liquid photoresist mask,³⁹ when the temperature of the BOE solution is fixed, the survival time of DFR mask presents a slightly downward trend as the concentration of BOE in etching solution decreases. The main reason is that the DFR is a kind of water-soluble dry film. When the concentration of the BOE solution decreases, the water ratio in the etchant solution increases; therefore, the dissolution of DFR leads to a reduced survival time of the DFR mask. With increasing the concentration of BOE solution, the average etching rate of the soda-lime glass increases at first and then gradually declines. All of the maximum average etching rates at 20, 30 and 40 °C are obtained by etching in the dilute 1/2 BOE solution (Fig. 2a). With a compromise of all the factors, the deepest microchannels with a depth of 102 μm are achieved under the conditions of both 1/1 BOE at 20 °C and 1/2 BOE at 30 °C (Fig. 2b). Due to the lower concentration of BOE and higher average etching rate, the 1/2 BOE at 30 °C is selected as the optimized etching condition. The survival time of DFR mask in 1/2 BOE at 30 °C is 31 min, and the average etching rate of the soda-lime glass substrates under this condition is 3.29 μm min⁻¹ (Fig. 2a).

Fig. 2 Controllable wet etching processes. (a) Effects of the ratio of buffered oxide etchant (BOE) solution and temperature on the survival time of DFR etching mask and the average etch rate of soda-lime glass microchannels. (b) Effects of the ratio of BOE solution and temperature on the maximum etch depth of microchannels. (c) SEM images of cross-section views of microchannels etched in different time periods. (d) Effects of etch time on the width and depth of microchannels.

Straight microchannels with a width of 300 μm are designed to investigate the morphology and the dimensions of etched channels with varied etching time periods, and the results show that smooth etched surfaces and controllable dimensions of microchannels are efficiently achieved (Fig. 2c). The depth of etched microchannels can be flexibly controlled from ∼10 μm to ∼100 μm by controlling the etching time (Fig. 2d). Due to the isotropic etching of glass substrates, the width of the etched channels increases over the etching time (Fig. 2d). To demonstrate the controllability of this method, we etch nine microchannels with the same geometry and size, and the measured widths and depths at different time periods show that satisfactory reproducibility (please see Fig. S1 in the ESI† for details).

Characterization of prepared glass microchips

To demonstrate the flexibility of the method for design and fabrication of glass microfluidic devices, six types of microchannels are designed including Y-junction (Fig. 3a1 and b1), T-junction (Fig. 3a2 and b2), cross-junction (Fig. 3a3 and b3), double flow-focusing channel (Fig. 3a4 and b4), snake-like channel (Fig. 3a5 and b5), and concentration gradient generator channel (Fig. 3a6 and b6). All the channels are etched in 1/2 BOE solutions at 30 °C for 20 min, and the resulted depths are ∼78 μm. With flexible designs, the patterns of microchannels transferred onto DFR after developing processes exhibit good quality (Fig. 3a). As shown in Fig. 3a, the edges of the developed structures are clear and neat, and no residual resist is observed in the channels at all. After wet etching and DFR striping, the patterns of designed channels are transferred onto glass substrates precisely, and the etched channels have smooth surfaces and neat edges (Fig. 3b). The minimum width of microchannel transferred onto DFR in this study is 50 μm, and a microchannel with width of 85 μm is achieved on glass slide after 5 min etching in BOE solution (please see Fig. S2 in the ESI† for details). This minimum feature size is dependent on the line resolution of Etertec HT-115T DFR. The results indicate that the DFRs have good chemical resistance to HF based etchant and good adhesion to glass substrates.

Fig. 3 Perfect transfer of different microchannel patterns from printed photomasks onto DFR and glass substrates. (a1–a6) Patterns of microchannels transferred onto DFR after developing processes. (b1–b6) Photographs of etched microchannels on the glass substrates after wet etching processes, in which the microchannels are filled with methylene blue solution for eye catching. (a1 and b1) Y-junction channel, (a2 and b2) T-junction channel, (a3 and b3) cross-junction channel, (a4 and b4) double flow-focusing channel, (a5 and b5) snake-like channel, and (a6 and b6) concentration gradient generator channel.

A rapid UV-bonding method is used to seal the microchannels. Compared with other glass-to-glass bonding methods, UV-bonding has several advantages, such as low operation temperature, no rigorous requirements for surfaces of glass substrates, simple and fast, and no deformation of the channels during the bonding process.³⁹ The low dependence on environment of this approach makes it applicable in general laboratories. Previously, Chen et al.³⁹ used spinning to apply a thin layer (∼10 μm) of the UV-glue to seal the microchannels. Here, we utilize an easy way of UV-bonding for fabrication of microfluidic devices, and only 2 min is needed to seal one set of microchip. A syringe needle is used to introduce a small volume (20–40 μl) of UV-glue slowly into the gaps between the glass plates, and the adhesive solution spreads uniformly and slowly into the gaps under the action of capillary force (please see Fig. S3 in the ESI† for details). As shown in Fig. S3,† the glue-bonding process is easy to control. The UV-glue fills the gaps between the glass plates after being introduced into the gaps for 80 s (Fig. S3k†), and then it stops at the edges of the microchannels even after 5 min (Fig. S3i†). Much smaller interface between the gaps of the glass plates makes the capillary force between the glass plates is much stronger than that between the microchannels and the glass plate, so that the UV-glue is prevented from entering into the microchannels when the volume of the UV-glue is in an appropriate range. In this work, a large number of glass-bonding experiments show that, the optimal volume of UV-glue used in bonding process is 20–40 μl for 25 mm × 25 mm glass microchips. With this optimal volume of UV-glue, the glass chips are all completely bonded but the etched channels are not blocked, and a high-yield bonding (∼100%) is achieved. When the volume of the UV-glue is respectably 20 μl, 30 μl and 40 μl, the resulted thickness of the UV-glue layer after bonding process is 24 μm, 36 μm and 40 μm (please see Fig. S4 in the ESI† for details). The glass microchips after UV-bonding show satisfactory and reliable durability. No solution leakage or glue clogging is observed in the fabricated microchips (Fig. 4). The pressure tests of the glass chips are carried out according to a previously published method,³⁴ and the results show that the glass chips are sealed satisfactorily without any fluid leakage even when the test pressure is as high as 0.8 MPa (please see Fig. S5 in the ESI† for details). Chemical resistance tests are carried out by subsequently introducing the solution of 1 mol l⁻¹ NaOH, 1 mol l⁻¹ HCl and acetone into the microchips, and the results show that the microchannels bonded by UV-glue retain their original shapes (please see Fig. S6 in the ESI† for details).

Fig. 4 Photographs of glass microchips after bonding with cover glasses using UV-curable glue, in which the microchannels are filled with methylene blue solution. (a) Y-junction device, (b) T-junction device, (c) cross-junction device, (d) double flow-focusing device, (e) snake-like device, and (f) concentration gradient generator device. The inserted pictures are patterns on photomasks that designed by AutoCAD.

Preparation of monodisperse W/O and O/W emulsions

To demonstrate the applications of the glass microchips, a hydrophobically modified T-junction microfluidic device is used to generate W/O emulsions (Fig. 5a), and a hydrophilically modified cross-junction device is used to prepare O/W emulsions (Fig. 5b). The emulsion droplet sizes are highly controllable by adjusting the flow rate ratios of Q_O/Q_I, and the droplet size decreases when Q_O/Q_I value increases (Fig. 5a1–a3 and b1–b3). The prepared emulsions are highly monodisperse (Fig. 5a3 and b3), and the coefficients of variation (CV) of the W/O and O/W emulsions are both less than 2%. The results show that the microchannel surfaces of the glass microchips can be spatially modified easily and stably.

Fig. 5 Demonstrations of different applications of prepared glass microchips. (a) Preparation of monodisperse W/O emulsions in a T-junction device, in which oil phase A = BB : SO (1 : 1, v/v) + 5% (w/v) PGPR and aqueous phase B = DI water + 10% (w/v) glycerol + 1% (w/v) F127 + 0.3% (w/v) Eosine-Y fluorescent dye. (a1) High-speed optical micrographs of the preparation processes of W/O emulsions with different flow rates, in which all the inner aqueous phase flow rates (Q_I) are 300 μl h⁻¹, and the outer oil phase flow rates (Q_O) are 1500 μl h⁻¹, 2500 μl h⁻¹ and 4500 μl h⁻¹ respectively. (a2) Dependence of the droplet diameter (D) on the relative flow rate Q_O/Q_I. (a3) Optical micrographs of W/O emulsion droplets prepared with outer oil phase flow rates of 1500 μl h⁻¹, 2500 μl h⁻¹ and 4500 μl h⁻¹ respectively. (b) Preparation of monodisperse O/W emulsions in a cross-junction device, in which oil phase C = BB + 5% (w/v) PGPR + 0.3% (w/v) LR 300 and aqueous phase D = DI water + 10% (w/v) glycerol + 1% (w/v) F127. (b1) High-speed optical micrographs of the preparation processes of O/W emulsions with different flow rates, in which all the inner aqueous phase flow rates (Q_I) are 300 μl h⁻¹, and the outer water phase flow rates (Q_O) are 2000 μl h⁻¹, 5000 μl h⁻¹ and 12 000 μl h⁻¹ respectively. (b2) Dependence of the droplet diameter (D) on the relative flow rate Q_O/Q_I. (b3) Optical micrographs of O/W emulsion droplets prepared with outer water phase flow rates of 2000 μl h⁻¹, 5000 μl h⁻¹ and 12 000 μl h⁻¹ respectively. (c) Generation of parallel laminar flow interface. (d) Generation of concentration gradient. In (c) and (d), the red solution is dyed with sudan red and the blue one is dyed with methylene blue, and both flow rates of the red and blue fluids are 500 μl h⁻¹. The scale bars in (a and b) are 200 μm.

Laminar flows in microchannels

To further demonstrate the utilities of our glass microchips, a Y-junction device and a concentration gradient generator device are used to respectively generate laminar flow interface (Fig. 5c) and concentration gradient in microchannels (Fig. 5d). As shown in Fig. 5c, a stable laminar flow interface is generated between the red and blue fluids. When we change the flow rates of the two fluids at different ratios, the interfaces are stable all the time. Specially, the formation of the laminar flow interface in microchannels is useful for microfabrication and synthesis based on interface reactions.^53,54 Fig. 5d shows a photograph of a representative microfluidic network used for generating concentration gradients, in which the magnified view of the channel clearly shows a stable colour gradient from blue to red perpendicular to the direction of flow. Such a molecule concentration gradient is useful in cell biology studies.⁵⁵

Mass production of glass microfluidic devices

Mass production of microfluidic devices in a simple and cheap way is the key for microfluidics to go from laboratory scale to industrial scale.^21,22 The DFR can be easily laminated onto glass substrate in a large area due to its unique structure,⁵⁶ so it is a highly potential strategy for mass production of glass microfluidic devices by using DFR as the pattern transfer mask for wet etching. Herein, T-junction devices are designed to demonstrate the utilities of our method for mass preparation of glass microchips. As shown in Fig. 6a, 24 T-junction channels are transferred onto a glass substrate in a single operation (Fig. 6a1 and a2), and then the microchips are bonded after cut into pieces (Fig. 6a3). The whole fabrication process only takes 12 h, which means that an average of 30 min is needed for one chip. To the best of our knowledge, such a short time for fabricating glass microfluidic devices with wet etching method has never been reported before. The mass-produced glass microchip can be used as modular unit to be assembled in parallel for achieving functional amplification, such as preparation of emulsions, microreactors, chemical synthesis, drug screening, etc. Here, assembled glass microchips in parallel are demonstrated to mass-produce emulsions (Fig. 6b). The emulsions prepared with the assembled T-junction glass microchips exhibit good monodispersity (Fig. 6b3 and b4).

Fig. 6 Demonstration of mass production of monodisperse W/O emulsions using assembled glass microchips in parallel. (a) Mass production of glass microchips. (a1) Patterns of 24 microchannels transferred onto the DFR layer on a large glass substrate, (a2) the photographs of 24 etched microchannels on the glass substrates after etching processes, and (a3) a batch of glass microchips after cutting into pieces and bonding. (b) Illustration of assembled microchips for mass production of monodisperse W/O emulsions. (b1) Photograph of distribution of a batch of T-junction glass microchips on a self-made shelf that fabricated by a 3D printer, in which A and B are the entrances of continuous and dispersed fluids, and C is the exit of fluid, (b2) photograph of assembled microfluidic devices in parallel after connected by PE pipes, (b3) high-speed optical micrograph of W/O emulsions nearby the exit C, and (b4) optical micrograph of W/O emulsions prepared with assembled glass microchips in parallel. The scale bars in (b3 and b4) are 200 μm.

With our method using DFR as pattern transfer mask for wet etching, very low cost is needed for fabrication of a glass microchip. Predictably, such a simple and cheap technology opens up a range of possibilities for productively manufacturing glass-based microfluidic devices.

Conclusions

In this work, we have developed a simple, cheap and rapid method for fabrication of glass microfluidic devices using DFR as the pattern transfer mask for wet etching. Compared with the conventional methods with wet etching masks that fabricated by deposition of metal layers and spin-coating of photoresists, our method only needs a one-step lamination to form DFR mask on the glass substrate, which greatly reduces the difficulty and cost for the fabrication of etched glass microchips. With the simple DFR mask operation, the glass microchips can be fabricated easily and controllably. By combining this DFR mask method with the mature manufacturing process of printed circuit board in the electronics industry, it could be very easy to achieve simple and cheap mass-production of glass microfluidic devices with flexibly designed microchannels.

Acknowledgements

The authors gratefully acknowledge support from the National Natural Science Foundation of China (21136006), the Program for Changjiang Scholars and Innovative Research Team in University (IRT1163), State Key Laboratory of Polymer Materials Engineering (sklpme2014-1-01), and the Specialized Research Fund for the Doctoral Program of Higher Education by the Ministry of Education of China (20120181110074).

References

1. G. M. Whitesides, Nature, 2006, 442, 368–373.
2. T. Thorsen, S. J. Maerkl and S. R. Quake, Science, 2002, 298, 580–584.
3. S. Y. Teh, R. Lin, L. H. Hung and A. P. Lee, Lab Chip, 2008, 8, 198–220.
4. S. K. Sia and G. M. Whitesides, Electrophoresis, 2003, 24, 3563–3576.
5. D. N. Breslauer, P. J. Lee and L. P. Lee, Mol. Biosyst., 2006, 2, 97–112.
6. P. Mitchell, Nat. Biotechnol., 2001, 19, 717–721.
7. H. Stone and S. Kim, AIChE J., 2001, 47, 1250–1254.
8. A. Günther and K. F. Jensen, Lab Chip, 2006, 6, 1487–1503.
9. J. El Ali, P. K. Sorger and K. F. Jensen, Nature, 2006, 442, 403–411.
10. N. Ye, J. Qin, W. Shi, X. Liu and B. Lin, Lab Chip, 2007, 7, 1696–1704.
11. P. Yager, T. Edwards, E. Fu, K. Helton, K. Nelson, M. R. Tam and B. H. Weigl, Nature, 2006, 442, 412–418.
12. E. M. Chan, R. A. Mathies and A. P. Alivisatos, Nano Lett., 2003, 3, 199–201.
13. J. Atencia and D. J. Beebe, Nature, 2004, 437, 648–655.
14. L. Frenz, A. El Harrak, M. Pauly, S. Bégin-Colin, A. D. Griffiths and J. C. Baret, Angew. Chem., Int. Ed., 2008, 47, 6817–6820.
15. A. J. Demello, Nature, 2006, 442, 394–402.
16. C. C. Lee, G. Sui, A. Elizarov, C. J. Shu, Y. S. Shin, A. N. Dooley, J. Huang, A. Daridon, P. Wyatt and D. Stout, Science, 2005, 310, 1793–1796.
17. L. Y. Chu, A. S. Utada, R. K. Shah, J. W. Kim and D. A. Weitz, Angew. Chem., Int. Ed., 2007, 46, 8970–8974.
18. R. K. Shah, J. W. Kim, J. J. Agresti, D. A. Weitz and L. Y. Chu, Soft Matter, 2008, 4, 2303–2309.
19. S. H. Kim, A. Abbaspourrad and D. A. Weitz, J. Am. Chem. Soc., 2011, 133, 5516–5524.
20. W. Wang, M. J. Zhang, R. Xie, X. J. Ju, C. Yang, C. L. Mou, D. A. Weitz and L. Y. Chu, Angew. Chem., Int. Ed., 2013, 52, 8084–8087.
21. G. M. Whitesides, Lab Chip, 2011, 11, 191–193.
22. G. M. Whitesides, Lab Chip, 2013, 13, 11–13.
23. J. R. Anderson, D. T. Chiu, H. Wu, O. J. Schueller and G. M. Whitesides, Electrophoresis, 2000, 21, 27–40.
24. D. C. Duffy, J. C. McDonald, O. J. Schueller and G. M. Whitesides, Anal. Chem., 1998, 70, 4974–4984.
25. J. P. Rolland, R. M. Van Dam, D. A. Schorzman, S. R. Quake and J. M. DeSimone, J. Am. Chem. Soc., 2004, 126, 2322–2323.
26. K. F. Jensen, MRS Bull., 2006, 31, 101–107.
27. A. Datta, I. Y. Eom, A. Dhar, P. Kuban, R. Manor, I. Ahmad, S. Gangopadhyay, T. Dallas, M. Holtz and H. Temkin, IEEE Sens. J., 2003, 3, 788–795.
28. R. L. Hartman and K. F. Jensen, Lab Chip, 2009, 9, 2495–2507.
29. C. Iliescu, Inform. Midem., 2006, 36, 204–211.
30. L. Rojas, R. Norarat, M. Napari, H. Kivistö, O. Chienthavorn and H. Whitlow, Nucl. Instrum. Methods Phys. Res., Sect. B, 2013, 306, 296–298.
31. A. Utada, E. Lorenceau, D. Link, P. Kaplan, H. Stone and D. Weitz, Science, 2005, 308, 537–541.
32. W. Wang, R. Xie, X. J. Ju, T. Luo, L. Liu, D. A. Weitz and L. Y. Chu, Lab Chip, 2011, 11, 1587–1592.
33. E. Q. Li, J. M. Zhang and S. T. Thoroddsen, J. Micromech. Microeng., 2014, 24, 015019–015029.
34. N. N. Deng, Z. J. Meng, R. Xie, X. J. Ju, C. L. Mou, W. Wang and L. Y. Chu, Lab Chip, 2011, 11, 3963–3969.
35. M. Stjernström and J. Roeraade, J. Micromech. Microeng., 1998, 8, 33–38.
36. T. McCreedy, Anal. Chim. Acta, 2001, 427, 39–43.
37. I. Rodriguez, P. Spicar-Mihalic, C. L. Kuyper, G. S. Fiorini and D. T. Chiu, Anal. Chim. Acta, 2003, 496, 205–215.
38. C. H. Lin, G. B. Lee, Y. H. Lin and G. L. Chang, J. Micromech. Microeng., 2001, 11, 726–732.
39. Q. Chen, G. Li, Q. H. Jin, J. L. Zhao, Q. S. Ren and Y. S. Xu, J. Microelectromech. Syst., 2007, 16, 1193–1200.
40. W. K. T. Coltro, E. Piccin, J. A. F. da Silva, C. L. do Lago and E. Carrilho, Lab Chip, 2007, 7, 931–934.
41. P. Pessoa de Santana, T. P. Segato, E. Carrilho, R. S. Lima, N. Dossi, M. Y. Kamogawa, A. L. Gobbi, M. H. Piazzeta and E. Piccin, Analyst, 2013, 138, 1660–1664.
42. R. H. Wopschall, Solid State Technol., 1986, 29, 153–157.
43. P. Vulto, T. Huesgen, B. Albrecht and G. Urban, J. Micromech. Microeng., 2009, 19, 077001–077005.
44. P. Vulto, N. Glade, L. Altomare, J. Bablet, L. Del Tin, G. Medoro, I. Chartier, N. Manaresi, M. Tartagni and R. Guerrieri, Lab Chip, 2005, 5, 158–162.
45. P. Leech, N. Wu and Y. Zhu, J. Micromech. Microeng., 2009, 19, 065019–065024.
46. N. B. Palacios-Aguilera, J. Bastemeijer, V. R. S. S. Mokkapati, J. R. Mollinger and A. Bossche, ICQNM 2011, The Fifth International Conference on Quantum, Nano and Micro Technologies, 2011, pp. 71–76.
47. V. R. S. S. Mokkapati, L. Zhang, H. R. Mollinger, J. Bastemeijer and J. Bossche, Quantum, Nano and Micro Technologies, 2009. ICQNM'09. Third International Conference on, IEEE, 2009, pp. 86–89.
48. C. Shen, V. R. S. S. Mokkapati, F. Santagata, A. Bossche and P. M. Sarro, Solid-State Sensors, Actuators and Microsystems Conference (TRANSDUCERS), 2011 16th International, IEEE, 2011, pp. 854–857.
49. H. Lorenz, L. Paratte, R. Luthier, N. F. de Rooij and P. Renaud, Sens. Actuators, A, 1996, 53, 364–368.
50. G. Spierings, J. Mater. Sci., 1993, 28, 6261–6273.
51. C. Iliescu, J. Jing, F. E. Tay, J. Miao and T. Sun, Surf. Coat. Technol., 2005, 198, 314–318.
52. B. Zhao, J. S. Moore and D. J. Beebe, Science, 2001, 291, 1023–1026.
53. P. J. Kenis, R. F. Ismagilov and G. M. Whitesides, Science, 1999, 285, 83–85.
54. X. Luo, D. L. Berlin, J. Betz, G. F. Payne, W. E. Bentley and G. W. Rubloff, Lab Chip, 2010, 10, 59–65.
55. S. K. Dertinger, D. T. Chiu, N. L. Jeon and G. M. Whitesides, Anal. Chem., 2001, 73, 1240–1246.
56. T. Huesgen, G. Lenk, B. Albrecht, P. Vulto, T. Lemke and P. Woias, Sens. Actuators, A, 2010, 162, 137–144.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1 showing the reproducibility of etched microchannels, Fig. S2 showing the minimum pattern transfer size of Etertec HT-115T DFR, Fig. S3 showing the spreading positions of the UV-glue between the glass plates during the glue-introducing process, Fig. S4 showing the resulting thickness of the UV-glue layer after bonding, Fig. S5 showing the pressure test of the UV-bonded glass microchip and Fig. S6 showing the chemical resistance of the UV-bonded microchannels. See DOI: 10.1039/c4ra15907a

---