Low temperature and deformation-free bonding of PMMA microfluidic devices with stable hydrophilicity via oxygen plasma treatment and PVA coating

Abstract

In the present study, a low temperature and deformation-free bonding method to seal hydrophilicity-stable microchannels in polymethyl methacrylate (PMMA) microfluidic devices was proposed. Oxygen plasma pre-/post-treatments and polyvinyl alcohol (PVA) coating were performed on the PMMA microfluidic substrates before thermal bonding. Different surface treatments were characterized using a confocal profilometer, tensiometer, attenuated total reflection Fourier-transform infrared spectrometer (ATR-FTIR), differential scanning calorimeter (DSC), and scanning electron microscope (SEM). The effects of surface modification on bond strength and microchannel integrity were studied. Oxygen plasma treatment prior to PVA coating was found to be significant in improving the coating uniformity and enhancing the adhesion between the PVA layer and the PMMA base. The PMMA microchannel substrate could be bonded to a homogeneous PMMA cover plate at about 70 °C (30 °C lower than the PMMA glass transition temperature, T_g) with the coated PVA layer serving as a medium material which had high wettability, high surface energy and low T_g. With oxygen plasma post-treatment of the coated PVA layer, the bond strength was improved and comparable to that obtained by pure thermal bonding at a temperature near the T_g of PMMA. Due to the low bonding temperature, the microchannel integrity was well retained with negligible deformation. The hydrophilicity stability of different surface treatments was evaluated and compared under both dry and wet storage conditions for a month. The results suggested that the PVA-coated PMMA substrates with oxygen plasma post-treatment present the highest hydrophilicity and lowest hydrophobic recovery. As a demonstration, monodispersed oil-in-water (O/W) droplets with volumes of sub-10 nL were successfully and reliably generated in the hydrophilic microfluidic devices fabricated with the proposed bonding method.

---

1. Introduction

Since the 1990s, microfluidics has become an enabling technology that offers many advantages including device miniaturization, reduced costs, minimized consumption of sample, accelerated analytical speed, and improved sample throughput.^1,2 Over the past two decades, many researchers have been attracted by the easy access of polydimethylsiloxane (PDMS) based microfluidics fabricated by soft lithography.^3,4 However, the demand of commercializing microfluidic devices with a killer application has driven people to use thermoplastics such as polycarbonate (PC), poly(methyl methacrylate) (PMMA), cyclic olefin copolymer (COC) and cyclic olefin polymer (COP).^5,6 Thermoplastics have been recognized as advantageous alternatives because they are relatively inexpensive and amenable to volume production processes (e.g. micro injection molding, micro hot embossing and micro thermoforming).^7,8 However, there are a few drawbacks intrinsic to thermoplastics. Most synthetic thermoplastic surfaces are hydrophobic in nature. This characteristic restricts thermoplastic devices in many applications. For instance, the introduction of aqueous solution in a hydrophobic thermoplastic micro- or nano-channel requires high external pumping pressure and high chip burst strength. Due to the hydrophobicity, biomolecules could irreversibly adsorb onto the hydrophobic thermoplastic surface, leading to biofouling and poor analytical performance.⁹ Moreover, thermoplastics present low surface free energy, limiting the bond strength developed between the mating thermoplastics. At elevated temperatures, it is easier to bond thermoplastic substrates together, but all or part of the previously defined microstructures could be distorted simultaneously.

Therefore, a variety of surface modification techniques have been employed to increase the hydrophilicity and the surface free energy.^10,11 Surface modification with plasma is fast, straightforward and homogeneous. It could tailor the surface chemistry and topography of the thermoplastic surface by oxygenation, degradation and etching.^12,13 However, the hydrophilization is typically not stable, and either a partial or complete hydrophobic recovery is usually observed on the treated thermoplastics.¹⁴ Compared to plasma treatment, UV/ozone irradiation could generate higher surface energy and a thin surface layer with lower T_g, making it possible to bond mating substrates at low temperature with high strength,¹⁵ but the hydrophilicity could also be largely lost upon DI water and solvents.¹⁶ Photografting could permanently functionalize the surface property to be either hydrophilic or hydrophobic.¹⁷ The hydrophilization on the very inert thermoplastic COC by two-step photografting was found to significantly reduce the protein adsorption.^18,19 Other one-step photografting techniques were also investigated to improve the bond strength of COC microfluidic devices.^20,21 However, both polar monomer and photo-initiator should be carefully prepared to be exposed under long time ultraviolet illumination. Moreover, the harsh solvents (e.g. acetone) used in photografting are not applicable to many thermoplastics (e.g. PMMA, PC and PS). In the literatures, many works on surface modification have partially aimed at either a reliable surface functionalization technique for a proof-of-concept microfluidic application or a versatile bonding method for the fabrication of disposable microfluidic chips, leaving the combination of these two sectors as an afterthought. In order to realize the potential of microfluidic devices, we should consider its final function at the earlier design and fabrication stages.⁶

In this study, we propose a reliable bonding method with surface modification aiming at those applications requiring high and stable hydrophilicity. The proposed bonding method comprises two-stage oxygen plasma treatments and one simple polyvinyl alcohol (PVA) coating by thermal immobilization, followed by one low temperature thermal bonding process. PVA is a water-soluble polymer, which has been frequently used in biomedical applications due to its good film-forming ability, good mechanical strength, high hydrophilicity, biocompatibility, good chemical resistance and non-toxicity.^22,23 PVA coating has recently been introduced to microfluidics for surface hydrophilization in electrophoresis chips made from silica capillaries,²⁴ PDMS,²⁵ PMMA²⁶ and COC.¹⁶ However, in these experiments, the PVA coating was applied to the already enclosed channels by external pumping, rinsing and then thermal incubation. Furthermore, those works focused only on the effects of PVA on the improvement of application performance. In this paper, we evaluate the utility of PVA coating and oxygen pre-/post-plasma treatments on both microchannel walls and substrate surfaces as a method for effective bonding at low temperature without deformation, meanwhile for surface functionalization of many hydrophilic applications. The proposed fabrication method was successfully demonstrated by an oil-in-water droplet generation experiment, which is also believed to be applicable to many other potential applications, such as biological analyses with low biomolecular adsorption, and electrophoresis separation with low electroosmotic flow (EOF), etc.

2. Experimental setups and procedures

2.1. Fabrication of microfluidic substrates

The whole fabrication process flow of microfluidic devices is shown schematically in Fig. 1(a–e). A microfluidic pattern was transferred from a plastic transparency onto a layer of AZ9260 positive photoresist through photolithography, followed by a Si wafer through deep reactive-ion etching (DRIE). The etched silicon wafer was used as the primary micromold to make the secondary metallic micromold by micro hot embossing of a Zr-based bulk metallic glass (BMG) plates supplied by Liquidmetal® Technology. After silicon stripping, the BMG plate was cut into a 25 mm × 25 mm micromold and mounted in a 75 mm × 25 mm mold core in a Battenfeld HM25/60 injection molding machine to replicate the microchannels in PMMA SUMIPEX® LG2 (Sumitomo Chemical, Japan). The molding conditions were predetermined as follows: pre-drying (80 °C, 4 hours), melt temperature (270 °C), mold temperature (85 °C), injection pressure (1000 bar), holding pressure (680 bar), holding time (12 s) and cooling time (60 s).²⁷ The fabricated BMG micromold, injection-molded microfluidic substrate and blank cover plate are shown in Fig. 1(f–h). After micro injection molding, the molded parts were cut into 25 mm × 25 mm and 25 mm × 50 mm footprint sizes. The latter type was used in the lap shear test to evaluate the bond strength. Before bonding, several surface treatments were performed on the PMMA substrates, including oxygen plasma pre-treatment, PVA coating and oxygen plasma post-treatment. A custom-made thermal bonding system was then used to bond the microfluidic substrate with the mating PMMA blank cover plate in varied bonding conditions.

Fig. 1 Fabrication process of microfluidic devices: (a) Si micromold fabrication by photolithography and DRIE, (b) BMG micromold fabrication by micro hot embossing, (c) PMMA substrate and cover plate fabrication by micro injection molding, (d) surface treatments through PVA coating and oxygen plasma pre- and post-treatment, and (e) low temperature thermal bonding. Photos of (f) fabricated BMG micromold, (g) injection-molded microfluidic substrate and (h) injection-molded blank cover plate.

2.2. PVA coating formation

The chemical structures of PVA and PMMA are shown in Fig. 2. PVA powder (average M_w 31 000–50 000 g mol⁻¹, 87–89% partially hydrolysed) was purchased from Sigma-Aldrich. A 2% w/v PVA aqueous solution was prepared by dissolving PVA powder into deionized (DI) water with stirring and heating at 85 °C for 24 hours. The PMMA substrates were immersed in the PVA aqueous solution at room temperature (RT, ∼22 °C) for 10 min, then removed from the solution and incubated in an oven at 75 °C for 10 min for thermal immobilization. The above procedure could be repeated for several cycles resulting in a multilayer PVA coating. After the final cycle of PVA coating, the substrates were put in a dry cabinet for 24 hours. The immobilization temperature chosen is 75 °C, which is much lower than the nominal heat deflection temperature of 90 °C supplied by the material manufacturer and T_g of 100 °C measured by DSC at 20 °C min⁻¹, in order to avoid any undesirable microchannel deformation during the thermal incubation. In the literature,²⁸ a similar immobilization procedure with baking at 60 °C for 20 min showed a reliable film coating on PMMA samples. In another reported work, the multi-layer PVA could be dynamically and irreversibly coated onto PMMA substrates even without thermal immobilization.²⁹

Fig. 2 Chemical structures of (a) the base material PMMA and (b) the coating material PVA.

2.3. Characterization

The thickness of coated PVA layer was measured using a confocal profilometer by measuring the boundary steps between coated portion and uncoated portion. The latter was previously covered with a Scotch tape before PVA coating. A JEOL scanning electron microscope was also used to characterize the coated layer thickness on the cross-sectioned PMMA substrates. Depending on the microfluidic application, the number of coating cycles was varied to determine the acceptable coating thickness in order to keep the microchannel dimension integrity to a great extent. The measurements were taken on samples with and without oxygen plasma pre-treatment, and at least four measurements were taken on each sample.

The contact angle measurement was used to calculate the surface energy (γ_sv) according to the Young's eqn (1) (ref. 30) shown below,

γ_sv = γ_sl + γ_lv cos θ + π_e (1)

where γ_sv is the free energy of the solid in contact with vapour, γ_sl is the free energy of the solid covered with liquid, and γ_lv is the free energy of the liquid-vapour interface, θ is the contact angle between solid and liquid, and π_e is the equilibrium pressure of the adsorbed vapour on the solid which could be assumed to be zero. In order to calculate the surface energy γ_sv from the contact angle, the second unknown variable γ_sl in above equation must be determined. According to the Owens–Wendt–Rabel–Kaelble (OWRK) method, also known as the extended Fowkes method, the interfacial energy γ_sl could be calculated based on the solid and the liquid surface energies and the interactions between the phases.³¹ These interactions could be interpreted as the geometric mean of a dispersive component (γ^d) and a polar component (γ^p) of the surface energy, as shown in eqn (2)–(5):

γ_sv = γ^d_sv + γ^p_sv (2)

γ_lv = γ^d_lv + γ^p_lv (3)

γ_sl = γ^d_sl + γ^p_sl (4)

After combining eqn (1)–(5), we could get

At least two liquids with known disperse and polar components of the surface energy are required to determine the solid surface energy. Therefore, the γ^d_sv and γ^p_sv can be determined by measuring the contact angles of these liquids against the solid. In this work, DI water (H₂O) and ethylene glycol (C₂H₆O₂) were used to measure the contact angles.

Static contact angles were measured at RT by the sessile-drop mode using a Theta Attension optical Tensiometer. A monochromatic cold led light source and a smooth lighting integrating sphere were used in this tensiometer to minimize undesirable sample evaporation. A 5 μL droplet of testing liquid was deposited on the sample surface, and after stabilization, a droplet image was captured. The static contact angle was calculated using Young–Laplace drop profile fitting method, and the results were averaged from 10 readings obtained on each sample.

The chemical compounds of the PMMA surface before and after treatments were determined using a Perkin-Elmer GX Fourier-transform infrared spectrometer (FTIR) by attenuated total reflection (ATR). A total of 32 accumulation scans were carried out for each spectrum. The frequencies were reported in wavenumbers (cm⁻¹).

The surface roughness measurements were performed using the confocal profilometer. The measured surface area in each sample was kept constant at 93 μm × 93 μm. The reported R_a value was averaged from at least five topographies captured on each sample.

The coating stability in terms of hydrophilicity (or hydrophobic recovery) was monitored by measuring the static contact angle multiple times over one month after various surface treatments under two different storage conditions, modified from the method used in the literature.¹⁴ The first set of samples was stored in the dry cabinet conditions: RT and relative humidity of 25%. The second set of samples was stored by immersion in the DI water at RT. Before each contact angle measurement, the samples were removed from the DI water and put in the dry cabinet at RT for 24 hours.

Differential scanning calorimetry (DSC) was carried out on a Q200 DSC instrument (TA Instruments, USA) to determine the T_g of PMMA and PVA. PMMA pellets were pre-dried in an oven at 80 °C before the measurement. PVA films with desired weight of about 15 mg were prepared in an aluminium pan through several repetitions of the coating recipe described in the previous section. The heating rate of DSC measurement was 20 °C min⁻¹, and the second heating cycle was recorded. The inflection midpoint of the heat flow shift was determined as the T_g.

2.4. Bonding performance evaluation

Bond strength was evaluated by lap shear test on an Instron testing machine. For each measurement, a pair of bonded substrates with overlapping area of 25 mm × 25 mm was tested at RT. The distance between upper and lower grips was kept constant and the crosshead moving speed of 0.5 mm min⁻¹ was applied for each testing. The bond strength was calculated as the maximum failure load divided by the overlapping area. At least three bonded chips were tested to get the averaged bond strength and standard deviation.

After lap shear test, the topographies of microchannels on the detached substrate were measured using the confocal profilometer. The morphology of detached surfaces and the cross-sectioned microchannel after thermal bonding were characterized using the SEM.

2.5. Oil-in-water (O/W) droplet generation setups

In the fabricated droplet generation chip, the oil droplets were formed in the aqueous phase at the cross junction by focusing the oil stream with aqueous streams coming from two sides. The oil used in the setup was M5904 mineral oil (Sigma Aldrich, Singapore), while the aqueous phase was the mixture of 80.4%w/w G7757 glycerol (Sigma Aldrich, Singapore), 19.1% w/w DI water and 0.5% w/w L4509 sodium dodecyl sulphate (SDS) (Sigma Aldrich, Singapore). Glycerol was added to the aqueous phase to increase the viscosity while SDS was added to prevent the coalescence of droplets. The liquids were driven from syringes, with controlled volumetric flow rate, using KDS200 syringe pumps (KD Scientific, USA). The videos of the droplet formation were captured using HiSpec-4 high speed camera (Fastec Imaging, USA).

3. Results and discussion

3.1. Surface characterization

3.1.1 Coating layer thickness. As to the application of microfluidic devices, the coating layer thickness is important as it could change the microchannel dimension and may affect the functional results. When the coating process was repeated, multi-layer assembly could be formed on the PMMA surface, due to the hydrophobic interactions between main chain groups (–CH₂–CH₂–) and the intermolecular hydrogen bonding of hydroxyl groups between the adsorbed PVA in the dry state and PVA chain in aqueous solution.³² The thickness of coating layer could depend on the solution concentration and the hydrophilicity of the base material.³³ Fig. 3 shows the measured PVA layer thickness through different coating cycles with and without oxygen plasma pre-treatment. It is apparent from the figure that the coating layer on the pristine PMMA surface was thicker with dramatic thickness variance. Due to the hydrophobic property of the pristine PMMA surface, the PVA coating was non-uniform across the whole surface. By contrast, the oxygen plasma treated surface had better wettability to the aqueous PVA solution, resulting in thinner and more uniform coating layer on the substrate. The micrographs in Fig. 3 show the cross-sections of two thermally bonded chips, each with two-cycle PVA coatings. Without oxygen plasma treatment, the two-cycle coating layer (A) was about 3.2 μm, while the coating layer (B) with oxygen plasma pre-treatment was only about 1.2 μm. It also should be noted that the surface activation could improve the surface free energy and enhance the adhesion between the base material and the coating layer, which usually could be achieved by UV/ozone or corona discharge. The adhesion reliability is very important, especially in the end applications, e.g. oil-in-water droplet generation in a microchannel with hydrophilic PVA coating, which will be presented in the latter section of this paper.

Fig. 3 PVA coating layer thickness with respect to varied cycles of coating on pristine and oxygen plasma pre-treated PMMA substrates. Inset SEM micrographs show the cross-sections of bonded PMMA substrates.

3.1.2 Contact angle and surface energy. The surface wettability and surface energy were examined by measuring the contact angles using DI water and ethylene glycol as the testing liquids with their known surface energy components³⁴ listed in Table 1. Fig. 4 shows the average contact angles of these two testing liquids and the calculated surface energies on the PMMA substrate with different treatments. The hydrophobic surface of the pristine PMMA had a static water contact angle of 75.0° ± 0.9°, agreeing quite well with the previous reported values of 75 ± 2° (ref. 35) and 76 ± 4° (ref. 36). The oxygen plasma treatment could significantly reduce the water contact angle to 42.7° ± 2.2°. After PVA coating, the water contact angle decreased further to 20.0° ± 2.4° showing considerably increased surface hydrophilicity. This could be attributed to the hydrophilic nature of PVA, containing several polar functional groups. After the oxygen plasma post-treatment on the coated PVA layer, the water contact angle decreased further to 13.9° ± 1.7°. After each treatment, both the polar component and the total surface energy increased. As the polar component corresponds to the various polar groups (e.g. –OH, C–O, C=O, O–C=O, et al.) on the modified surface, the water contact angle decreased as the polar component of surface energy increased.

Table 1 Surface energy components of DI water (H₂O) and ethylene glycol (C₂H₆O₂) at 20 °C

Liquid	γ^dlv (mJ m^−2)	γ^plv (mJ m^−2)	γlv (mJ m^−2)
DI water	21.8	51.0	72.8
Ethylene glycol	29.3	19.0	48.3

---

Fig. 4 Static contact angles of DI water and ethylene glycol, and calculated surface energies with dispersive and polar components for the pristine and treated surfaces (treatments in an accumulative sequence).

3.1.3 FTIR. ATR-FTIR measurements were carried out to study the differences in modified chemical structure of PMMA with different treatments. Fig. 5(a) shows the ATR-FTIR spectra of PMMA before and after different surface treatments, while Fig. 5(b) shows the subtracted spectra.

Fig. 5 (a) ATR-FTIR spectra and (b) the spectral differences: O—pristine, OP—pristine and OPO–OP, where ‘O’ represents oxygen plasma pre- or post-treatment, and ‘P’ represents PVA coating.

The spectral change after the oxygen plasma treatment was noticeable with two absorption peaks at 1722 cm⁻¹ and 1140 cm⁻¹ showing the presence of C=O and C–O groups, respectively. The increase in these oxygen-containing functional groups revealed the oxygen insertion into the polymer matrix during the plasma treatment. This finding correlates well with the earlier observations that, due to these functional groups the hydrophilicity and surface energy could be increased. This hydrophilic surface with higher surface energy could facilitate the uniform coating as we observed in the thickness measurement. Moreover, the high surface energy could also promote the adhesion between the coating layer and the base substrate which will be discussed in the following section in terms of bond strength measured by lap shear test.

After PVA coating, the subtracted spectrum showed a broad noticeable absorption peak around 3100–3600 cm⁻¹ which could be due to the presence of hydroxyl group (–OH) of PVA molecules. The absorption peaks corresponding to –CH stretching vibrations, CH₂, CH₃ and vibration of CH₂ bond were also observed. The bands corresponding to carbonyl functional groups were due to the residual acetate groups remaining after the hydrolysis of polyvinyl acetate (PVAc). All the above observed peaks in the spectrum clearly revealed the major groups associated with PVA chemical structures, indicating the successful attachment of PVA onto the PMMA substrate.

In the subtracted spectrum between PVA coating and oxygen plasma post-treatment, there was only slight increase in –OH and C–O associating with the insignificant change in both water contact angle and surface energy. In general, there was no noticeable change in the spectrum, indicating that the main chemical composition and chain interaction of PVA was unchanged after oxygen plasma post-treatment.

3.1.4 Surface roughness. It is known that the increase in surface energy especially the polar components could either be attributed to the increased polar groups, or originate with the increased surface roughness.¹⁵ The comparative surface roughness values and typical observations on samples with different treatments are shown in Fig. 6. After micro injection molding, the roughness on the pristine PMMA was about 10 nm, which then increased to 14 nm after oxygen plasma treatment. The change in surface roughness could be resulted from etching or degradation of PMMA during plasma treatment.¹⁵ After PVA coating the surface became smoother with a roughness value of about 7 nm, which indicated that the effect of surface roughness on surface energy was negligible compared to the effect of the increased polar groups as analyzed in the previous section. It is evident from Fig. 6 that the oxygen plasma post-treated surface exhibited a higher surface roughness of 11 nm, which could also be attributed to etching or oxygenation in the oxygen plasma process.³⁷ The higher surface roughness probably could contribute to the improved wettability and surface energy, which might not be solely explained by ATR-FTIR observations.

Fig. 6 (a) Calculated surface roughness values and (b) typical surface morphologies of the pristine and treated surfaces (treatments in an accumulative sequence).

3.2. Thermal bonding performance

3.2.1 Bond strength. The bond strength of PMMA substrates with different treatments were measured and compared, as shown in Table 2. For the untreated PMMA substrates, the bond strength of 0.34 MPa could be obtained at 85 °C, which is in the vicinity of the glass transition range. When thermal bonding was immediately conducted after oxygen plasma pre-treatment, the bond strength at 85 °C could be improved to 0.45 MPa. This increment in bond strength could be attributed to the increase in the surface free energy especially the polar component and the surface roughness, resulted from the oxygen plasma pre-treatment. During thermal bonding process, the increased surface energy could facilitate the interaction between the treated surfaces and enhance the intimate contact.¹⁵

Table 2 Bond strength of thermally bonded PMMA with different treatments (bonding pressure 1 MPa and boning time 15 min)

Treatment on PMMA	Temperature (°C)	Bond strength (MPa)
Untreated	70	Nil
	85	0.34
O2 plasma	70	Nil
	85	0.45
PVA without O2 plasma	70	0.13
O2 plasma + PVA	70	0.18
	85	0.26
O2 plasma + PVA + O2 plasma	70	0.32

---

When a PVA layer was coated onto the PMMA substrates without oxygen plasma pre-treatment, the obtained bond strength was weak at 0.13 MPa. From the SEM micrograph shown in Fig. 7(a), the coated PVA layer was easily detached from the PMMA surface after strength test. This result indicated that de-bonding mainly occurred at the coating interface, due to the inferior adhesion between the coated PVA and the PMMA. Without surface activation, PVA can only be attracted to the PMMA surface by hydrogen bonding,²⁹ because the surface charge could arise from enhanced autolysis of water at hydrophobic surfaces with the preferential adsorption of a hydroxide ion.³⁸ With oxygen plasma pre-treatment, the attachment of PVA layer onto the PMMA surface could be enhanced due to the increased surface free energy, as analyzed in the previous section. Besides, when PMMA was oxidized by oxygen plasma, more hydrogen-bond donors could be supplied at the PMMA surfaces for the spontaneous adsorption of the coating polymer. The enhancement in adhesion between PVA and PMMA could be interpreted from Fig. 7(b). The coated PVA layer still adhered firmly onto the PMMA surface, while the zig-zag failure observed on PVA was due to the shear force during test.

Fig. 7 Morphologies of failure surfaces after lap shear tests of samples (a) without and (b) with oxygen plasma pre-treatment.

During thermal bonding, the PVA films coated on the mating PMMA surfaces served to bond to each other, and the T_g should be taken into account because it correlates with the molecular segmental mobility and determines the processing window of temperature. The DSC curves of PMMA and prepared PVA film are presented in Fig. 8. The T_g of pristine PMMA and PVA was measured to be 100 °C and 71 °C, respectively. The measured T_g of PVA was smaller than the value 85 °C reported in the literature,³³ but agreed well with other reported results between 59 °C and 74 °C.³⁹ When the applied bonding temperature (70 °C) was around the T_g of PVA, the molecular chains could actively move and interdiffuse to each other at the mating interface, and the obtained bond strength was 0.18 MPa. When the temperature was elevated to 85 °C, the bond strength could be improved to 0.26 MPa; however the microchannel was distorted at the same time.

Fig. 8 DSC curves showing the measured T_g of PMMA and PVA, respectively.

It was interesting to note that when the coated PVA were then treated again with the oxygen plasma, the obtained bond strength was noticeably improved to about 0.32 MPa at 70 °C, which was close to the strength obtained at 85 °C for untreated PMMA. The possible reason for this substantial improvement in bond strength could be that, the post-treatment of oxygen plasma served to improve the surface wettability and enabled easier intimate contact between the surfaces during thermal bonding. The roughened surface could also increase the surface contact area and enhance the molecular diffusion across the bonding interface. It was reported that due to the complexity involved in the oxygen plasma process (simultaneous etching and oxygenation), some amorphous and crystalline fractions could be removed by plasma and the T_g of PVA film could decrease.³⁷

3.2.2 Microchannel integrity. For the microchannel dimensional integrity analysis, the PMMA substrates with microchannel design dimension of 200 μm × 200 μm were injection-molded and bonded with PVA coating and oxygen plasma pre- and post-treatment. The thermal bonding experiments were conducted at temperature of 70 °C with bonding pressure of 1 MPa and bonding time of 15 min. The morphology and dimension of microchannel were examined using an SEM and a confocal profilometer. The SEM micrographs shown in Fig. 9(a–d) present the microchannels on the embossed BMG micromold, the injection-molded pristine PMMA substrate, the PVA-coated PMMA substrate with oxygen plasma pre-treatment, and the manually detached PMMA substrate after thermal bonding, respectively. It is apparent from Fig. 9(c) that, all the surfaces became smoother after PVA coating, indicating the successful attachment of PVA on both the substrate surface and the microchannel sidewalls. After thermal bonding, the manually detached substrate presented the fracture of PVA layer as shown in Fig. 9(d). The microchannel profiles on the samples used in the SEM observation were measured as shown in Fig. 9(e–g) for comparison. The microchannels were replicated from the BMG micromold to the PMMA substrate with high fidelity by micro injection molding, and the dimension was negligibly affected by the coated thin PVA layer. It is also evident that the microchannel integrity was well retained without distortion after low temperature thermal bonding. A thermally-bonded droplet generation chip is shown in Fig. 9(h) and the cross-sectioned microchannel is shown in Fig. 9(i). It can be seen from the SEM micrograph that after thermal bonding, the microchannel walls remained vertical and no obvious microchannel collapse was observed. All the above observations could be attributed to the fact that the bonding process was conducted at the temperature much lower than the heat deflection temperature (90 °C) and the T_g (100 °C) of the bulk PMMA.

Fig. 9 (a–d) SEM micrographs of microchannels on (a) BMG micromold, (b) injection-molded pristine PMMA substrate, (c) PVA-coated PMMA substrate, and (d) manually detached PMMA substrate after thermal bonding; (e–g) profile comparison of the microchannel measured on the samples used in the SEM observation; (h) photo of a thermally-bonded droplet generation chip with (i) SEM micrograph showing the cross-sectioned microchannel.

3.3. Hydrophilicity stability

Fig. 10(a) and (b) show the hydrophilicity stability test results of treated PMMA samples exposed to air and DI water, respectively. The contact angle of 75.0° on the pristine PMMA surface is included for reference.

Fig. 10 Static water contact angles of hydrophilicity stability tests measured on the surface-treated PMMA substrates under (a) dry and (b) wet storage conditions.

The oxygen plasma treated PMMA surfaces present evident hydrophobic recovery from 42.7° (±2.2°) to 56.2° (±1.6°) and 56.1° (±2.7°) after one month for the samples exposed to air and DI water, respectively. This recovery could still extend till the contact angle reaches to near the original level of pure PMMA.^14,40 The results showed that, either in dry or in wet storage environment, PMMA with oxygen plasma treatment has very instable hydrophilicity over a period of time. Hydrophobic recovery of the surface could also cause the inconsistency in the wetting properties of microfluidic chips during the end applications.

The aging results of the PVA coated samples exposed to the air showed that the surface hydrophilicity could remain stable with the contact angle around 20° for over one month. The stability correlated fairly well with the reported results in the literature.²⁸ After the oxygen plasma post-treatment, the modified surface underwent a quick hydrophobic recovery in the air from 13.9° (±1.7°) to 16.5° (±1.9°) after one day as shown in the inset figure, and its wettability stabilized with the contact angle still lower than that of the surface without plasma post-treatment. This observation suggested that the oxygen plasma post-treatment could permanently modify the PVA surface chemistry and morphology resulting in a more hydrophilic surface.

After the PVA coated samples were immersed in DI water, the measured contact angle increased from 20.0° (±2.4°) to 30.9° (±1.6°) within 5 days and then stabilized for the rest testing days. The similar trend in contact angle variation was also observed for the coated samples with oxygen plasma post-treatment, which increased from 13.9° (±1.7°) and stabilized at about 22.8° (±1.7°) after 5 days. Due to the plasma post-treatment, the hydrophilicity could be retained at a higher level with the contact angle close to that of PVA exposed to the air. It also should be noted that PVA is a water-soluble synthetic polymer. If the coated PVA layer was dissolved in the DI water, the surface wettability would be dominated by the exposed base PMMA. However there was no full hydrophobicity recovery observed on the DI water immersed samples. It has been reported that the solubility of PVA in water depends on its degree of hydrolysis, and the PVA grade with high degree (87–89% in this case) of hydrolysis has low solubility in water.⁴¹ The presence of large amount of hydroxyl groups in highly hydrolysed PVA could enhance the intra- and inter-molecular hydrogen bonding of adjacent hydroxyl groups and subsequently could greatly reduce the water solubility. The water temperature must be raised to above 70 C° to accelerate the dissolution of PVA,³³ which was highly evident in the aqueous solution preparation. In this study, the increase in contact angle could be attributed to the initial dissolution of PVA into the water due to the residual hydrophobic acetate groups. After reaching the equilibrium state, the hydrophilicity could then stabilize for very long time.

All the above results suggested that the PVA coated PMMA substrate with oxygen plasma post-treatment presented the lowest contact angle and the most stable hydrophilicity, either in dry or wet storage conditions. Therefore, the microfluidic chips fabricated with the proposed modification and bonding method could be stored with highly stable hydrophilicity and used for long shelf-life hydrophilic applications.

3.4. Oil-in-water (O/W) droplet generation

With the proposed fabrication method, several droplet generation chips were fabricated and stored under DI water condition. After one week of immersion in the DI water, they were tested for the oil-in-water droplet generation. Fig. 11 shows the generated droplet volumes at different flow rate ratios between the continuous aqueous phase and the dispersed oil phase. Each data point in the graph of the figure was collected from video capturing of over 100 generated droplets, and analysed by a self-written software using OpenCV image processing library. As expected, within the same flow rate ratio, the droplet volume decreased as the continuous phase flow rate increased. This was due to the increment of viscous stress from the continuous phase streams that made the dispersed stream to break up faster. The insets in the figure show the micrographs of the freshly formed oil droplets at flow rate ratio of 5, with different continuous phase flow rate: 500, 1000, 1500, and 2000 μL h⁻¹, respectively. The results from this experiment showed the bonded device was capable of generating monodispersed oil droplets with controllable sizes and frequencies using different sets of flow rates.

Fig. 11 Effect of continuous aqueous flow rate on droplet volume at different flow rate ratios, Q_c/Q_d = 2.5, 5, and 10; inset micrographs show the monodispersed oil droplets in aqueous solution at flow rate ratio of 5 with different continuous phase flow rates of 500, 1000, 1500, and 2000 μL h⁻¹, respectively.

The adhesion stability of PVA coating onto PMMA surface could also be evaluated from the delamination of PVA layer during microfluidic applications. Control chips were fabricated by thermal bonding of pristine PMMA substrates at 85 °C under 1 MPa without oxygen plasma pre-treatment. The procedure of dynamic coating PVA in the microchannel was adopted from the literatures.^16,26 The 2% w/v PVA solution was injected into the sealed microchannels using a syringe at RT for 10 min. Then the microchannels were emptied by pumping air through the chip for 5 min. Upon drying, the chip was incubated in an oven at 70 °C for 10 min. The filling/evacuation/incubation process was repeated for two times for a better coating. Fig. 12 shows the downstream microchannel on the droplet generation chip. It could be found in Fig. 12(a) that it was quite easy for the coated PVA layer to delaminate from the microchannel sidewalls during the experimental operation. The delaminated PVA layer was washed away by the aqueous solution and accumulated at the downstream microchannel. In contrary as shown in Fig. 12(b), no delamination of PVA was observed, indicating that the oxygen plasma treatment could improve the coating adhesion and the uniformity, ensuring a hydrophilicity-reliable and stable coating during the application. Moreover, due to the low temperature bonding method, the clarity and surface finish of PMMA substrates could be highly preserved after bonding, as comparatively shown in Fig. 12.

Fig. 12 Photos of the downstream microchannel during oil-in-water droplets generation showing (a) the delaminated PVA layer and (b) the improved adhesion without PVA delamination by oxygen plasma pre-treatment.

The PVA coating stability was evaluated by measuring the in vitro solubility of PVA film in the applied aqueous solution (the mixture of glycerol, DI water and SDS), in order to estimate the loss of PVA in the bonded chip over time during droplet generation. The PVA film was prepared in an aluminium pan and post-treated by oxygen plasma to simulate the coating layer in the bonded microchannel. The film was then immersed in the aqueous solution at RT for solubility measurement. No detectable weight loss of PVA film was recorded after 5 days of solution immersion. The result indicated that the coated PVA layer in bonded microchannel could be used for long shelf-life droplet generation because PVA with high degree of hydrolysis has negligible solubility in water and glycerol at RT. It also should be noted that PVA is insoluble in organic solvents, e.g. mineral oil, and to the best of our knowledge there is no reported interaction between PVA and SDS.

4. Conclusions

The low wettability and low surface energy of thermoplastics could affect the ease of bond strength development during thermal bonding and limit the use of thermoplastic microfluidic devices for many applications. In this study, a novel bonding method via PVA coating and oxygen pre- and post-treatments, followed by a low temperature thermal bonding process, was proposed as an alternative fabrication method for obtaining sealed PMMA microchannels with high dimensional integrity and highly stable hydrophilicity.

The oxygen plasma pre-treatment was found to be effective in improving the coating layer uniformity and enhancing the adhesion between the coated PVA layer and the PMMA surface due to the increased polar groups and the roughened surface. Due to the high surface energy and low T_g, the coated PVA layer could serve as a medium material to bond the mating substrates at the temperature lower than the glass transition range of the substrate material. Therefore, the undesirable microchannel distortion coming along with the conventional thermal bonding method could be avoided. The oxygen plasma post-treatment was found to be effective in increasing surface energy of the coated PVA by roughening its surface. At the bonding temperature close to T_g of the PVA layer, the bond strength developed with oxygen plasma post-treatment could be improved and comparable to that obtained on pristine PMMA by pure thermal bonding at a temperature near the T_g of PMMA. Moreover, the oxygen plasma post-treatment could further render the surface to be more hydrophilic and more stable in either dry or wet storage environments. The monodispersed oil-in-water (O/W) droplets were successfully and reliably generated in the hydrophilic microfluidic device fabricated with the proposed bonding method.

Acknowledgements

The authors gratefully acknowledge financial support from the Singapore-MIT Alliance (SMA) program in Manufacturing Systems and Technology (MST) for this work. H. Yu would also want to thank the support of Nanyang Technological University Research Scholarship.

References

1. G. M. Whitesides, The origins and the future of microfluidics, Nature, 2006, 442, 368–3673.
2. E. K. Sackmann, A. L. Fulton and D. J. Beebe, The present and future role of microfluidics in biomedical research, Nature, 2014, 507, 181–189.
3. D. C. Duffy, J. C. McDonald, O. J. A. Schueller and G. M. Whitesides, Rapid prototyping of microfluidic systems in poly(dimethylsiloxane), Anal. Chem., 1998, 70, 4974–4984.
4. M. A. Unger, H. P. Chou, T. Thorsen, A. Scherer and S. R. Quake, Monolithic microfabricated valves and pumps by multilayer soft lithography, Science, 2000, 288, 113–116.
5. H. Becker and L. E. Locascio, Polymer microfluidic devices, Talanta, 2002, 56, 267–287.
6. L. R. Volpatti and A. K. Yetisen, Commercialization of microfluidic devices, Trends Biotechnol., 2014, 32, 347–350.
7. H. Becker and C. Gärtner, Polymer microfabrication technologies for microfluidic systems, Anal. Bioanal. Chem., 2008, 390, 89–111.
8. P. N. Nge, C. I. Rogers and A. T. Woolley, Advances in microfluidic materials, functions, integration, and applications, Chem. Rev., 2013, 113, 2550–2583.
9. C. Zilio, L. Sola, F. Damin, L. Faggioni and M. Chiari, Universal hydrophilic coating of thermoplastic polymers currently used in microfluidics, Biomed. Microdevices, 2014, 16, 107–114.
10. M. Kitsara and J. Ducrée, Integration of functional materials and surface modification for polymeric microfluidic systems, J. Micromech. Microeng., 2013, 23, 033001.
11. C. W. Tsao and D. L. DeVoe, Bonding of thermoplastic polymer microfluidics, Microfluid. Nanofluid., 2009, 6, 1–16.
12. Z. Wu, N. Xanthopoulos, F. Reymond, J. S. Rossier and H. H. Girault, Polymer microchips bonded by O₂-plasma activation, Electrophoresis, 2002, 23, 782–790.
13. S. Roy and C. Y. Yue, Surface modification of COC microfluidic devices: A comparative study of nitrogen plasma treatment and its advantages over argon and oxygen plasma treatments, Plasma Processes Polym., 2011, 8, 432–443.
14. V. Jokinen, P. Suvanto and S. Franssila, Oxygen and nitrogen plasma hydrophilization and hydrophobic recovery of polymers, Biomicrofluidics, 2012, 6, 016501.
15. H. Shinohara, T. Kasahara, S. Shoji and J. Mizuno, Studies on low-temperature direct bonding of VUV/O₃-, VUV- and O₂ plasma-pre-treated poly-methylmethacrylate, J. Micromech. Microeng., 2011, 21, 085028.
16. C. W. Tsao, L. Hromada, J. Liu, P. Kumar and D. L. DeVoe, Low temperature bonding of PMMA and COC microfluidic substrates using UV/ozone surface treatment, Lab Chip, 2007, 7, 499–505.
17. T. Rohr, D. F. Ogletree, F. Svec and J. M. J. Fréchet, Surface functionalization of thermoplastic polymers for the fabrication of microfluidic devices by photoinitiated grafting, Adv. Funct. Mater., 2003, 13, 264–270.
18. C. Li, Y. Yang, H. G. Craighead and K. H. Lee, Isoelectric focusing in cyclic olefin copolymer microfluidic channels coated by polyacrylamide using a UV photografting method, Electrophoresis, 2005, 26, 1800–1806.
19. T. B. Stachowiak, D. A. Mair, T. G. Holden, L. J. Lee, F. Svec and J. M. J. Fréchet, Hydrophilic surface modification of cyclic olefin copolymer microfluidic chips using sequential photografting, J. Sep. Sci., 2007, 30, 1088–1093.
20. R. K. Jena, C. Y. Yue and L. Anand, Improvement of thermal bond strength and surface properties of Cyclic Olefin Copolymer (COC) based microfluidic device using the photo-grafting technique, Sens. Actuators, B, 2011, 157, 518–526.
21. S. Roy, C. Y. Yue, S. S. Venkatraman and L. L. Ma, Low-temperature (below T g) thermal bonding of COC microfluidic devices using UV photografted HEMA-modified substrates: High strength, stable hydrophilic, biocompatible surfaces, J. Mater. Chem., 2011, 21, 15031–15040.
22. E. Chiellini, A. Corti, S. D'Antone and R. Solaro, Biodegradation of poly (vinyl alcohol) based materials, Prog. Polym. Sci., 2003, 28, 963–1014.
23. C. Shao, H. Y. Kim, J. Gong, B. Ding, D. R. Lee and S. J. Park, Fiber mats of poly(vinyl alcohol)/silica composite via electrospinning, Mater. Lett., 2003, 57, 1579–1584.
24. D. Belder, A. Deege, F. Kohler and M. Ludwig, Poly(vinyl alcohol)-coated microfluidic devices for high-performance microchip electrophoresis, Electrophoresis, 2002, 23, 3567–3573.
25. D. Wu, Y. Luo, X. Zhou, Z. Dai and B. Lin, Multilayer poly(vinyl acohol)-adsorbed coating on poly(dimethylsiloxane) microfluidic chips for biopolymer separation, Electrophoresis, 2005, 26, 211–218.
26. J. J. Shah, J. Geist, L. E. Locascio, M. Gaitan, M. V. Rao and W. N. Vreeland, Surface modification of poly(methyl methacrylate) for improved adsorption of wall coating polymers for microchip electrophoresis, Electrophoresis, 2006, 27, 3788–3796.
27. H. Yu, S. B. Tor, N. H. Loh and A. K. Asundi, Effect of injection-molding-induced residual stress on microchannel deformation irregularity during thermal bonding, J. Micromech. Microeng., 2013, 24, 015012.
28. M. Kitsara, C. E. Nwankire, L. Walsh, G. Hughes, M. Somers, D. Kurzbuch, X. Zhang, G. G. Donohoe, R. O'Kennedy and J. Ducrée, Spin coating of hydrophilic polymeric films for enhanced centrifugal flow control by serial siphoning, Microfluid. Nanofluid., 2014, 16, 691–699.
29. Y. Zhang, G. Ping, N. Kaji, M. Tokeshi and Y. Baba, Dynamic modification of poly(methyl methacrylate) chips using poly(vinyl alcohol) for glycosaminoglycan disaccharide isomer separation, Electrophoresis, 2007, 28, 3308–3314.
30. D. K. Owens and R. C. Wendt, Estimation of the surface free energy of polymers, J. Appl. Polym. Sci., 1969, 13, 1741–1747.
31. R. Good and C. van Oss, The Modern Theory of Contact Angles and the Hydrogen Bond Components of Surface Energies, in Modern Approaches to Wettability, ed. M. Schrader and G. Loeb, Springer, 1992, pp. 1–27.
32. T. Serizawa, S. Hashiguchi and M. Akashi, Stepwise assembly of ultrathin poly(vinyl alcohol) films on a gold substrate by repetitive adsorption/drying processes, Langmuir, 1999, 15, 5363–5368.
33. M. Kozlov, M. Quarmyne, W. Chen and T. J. McCarthy, Adsorption of poly(vinyl alcohol) onto hydrophobic substrates. A general approach for hydrophilizing and chemically activating surfaces, Macromolecules, 2003, 36, 6054–6059.
34. D. H. Kaelble and J. Moacanin, A surface energy analysis of bioadhesion, Polymer, 1977, 18, 475–482.
35. D. Raghavan, M. VanLandingham, X. Gu and T. Nguyen, Characterization of heterogeneous regions in polymer systems using tapping mode and force mode atomic force microscopy, Langmuir, 2000, 16, 9448–9459.
36. L. Brown, T. Koerner, J. H. Horton and R. D. Oleschuk, Fabrication and characterization of poly(methylmethacrylate) microfluidic devices bonded using surface modifications and solvents, Lab Chip, 2006, 6, 66–73.
37. D. J. Upadhyay and N. V. Bhat, Pervaporation studies of gaseous plasma treated PVA membrane, J. Membr. Sci., 2004, 239, 255–263.
38. J. K. Beattie, The intrinsic charge on hydrophobic microfluidic substrates, Lab Chip, 2006, 6, 1409–1411.
39. S. N. Cassu and M. I. Felisberti, Poly(vinyl alcohol) and poly(vinyl pyrrolidone) blends: Miscibility, microheterogeneity and free volume change, Polymer, 1997, 38, 3907–3911.
40. K. Tsougeni, N. Vourdas, A. Tserepi, E. Gogolides and C. Cardinaud, Mechanisms of oxygen plasma nanotexturing of organic polymer surfaces: From stable super hydrophilic to super hydrophobic surfaces, Langmuir, 2009, 25, 11748–11759.
41. C. A. Finch, Polyvinyl alcohol: properties and applications, Wiley, New York, 1973, p. 22.

---