Microwave plasma treatment of polymer surface for irreversible sealing of microfluidic devices

Abstract

Microwave plasma was generated in a glass bottle containing 2–3 Torr of oxygen for plasma treatment of a polymer surface. A “kitchen microwave oven” and a dedicated microwave digestion oven were used as the power source. Poly(dimethylsiloxane) (PDMS) slabs treated by a 30 W plasma for 30–60 s sealed irreversibly to form microfluidic devices that can sustain solution flow of an applied pressure of 42 psi without leaking. Experimental set up and conditions for the production of a homogeneous plasma to activate the PDMS surface for irreversible sealing are described in detail. The surface of a microwave plasma-treated PDMS slab was characterized using atomic force microscopy (AFM) and attenuated total reflection–Fourier Transform infrared spectroscopy (ATR-FTIR). The plasma-treated surface bears silica characteristics.

---

Introduction

Poly(dimethylsiloxane) (PDMS) is one of the major types of material used for the fabrication of microfluidic devices.^1,2 PDMS is relatively low cost. The polymer is transparent down to a UV wavelength of 240 nm. PDMS also has a high breakdown voltage and is chemically inert towards most reagents. Importantly, negative replicas can be mass-produced readily by casting and curing prepolymer on a positive relief master of the desired microchannel structure.^3–6 The negative replica is sealed with another flat slab of PDMS to form a microfluidic device. Reversible sealing by electrostatic interaction and irreversible sealing by chemical bond formation are possible. Reversible sealing is convenient and simple but the device can only withstand solution flow of a pressure of ≤5 psi.⁶ In contrast, microfluidic devices prepared using irreversible sealing can sustain a relatively high pressure of 30–50 psi.⁶

The surface of PDMS slabs is often treated in a radio-frequency (RF) plasma cleaner for irreversible sealing.^1,4,6 The PDMS replica and flat slab are exposed to an air or oxygen plasma of gas pressure 0.2–0.6 Torr for 1 or 2 min. The slabs are then brought into conformal contact to form the irreversible seal.⁴ Plasma surface treatment using corona discharge^7–9 and microwave-oven-generated plasma¹⁰ have also been reported. The corona discharge method requires a relatively long treatment time of 10 min. The microwave-oven-generated plasma method is interesting in that a ubiquitous “kitchen microwave oven”, instead of the relatively expensive and single-purpose RF plasma cleaner, is used. The method, however, operates at a relatively low pressure of 10⁻³ Torr and requires a vacuum desiccator as the plasma chamber and a steel electrode to initiate the plasma. In this paper, we report a microwave-oven-generated plasma in a glass bottle containing a few torr of oxygen. The oxygen plasma starts spontaneously upon the application of microwave power. The plasma was used to treat the surface of PDMS slabs for irreversible sealing. The plasma-treated PDMS surface was characterized using attenuated total reflection-Fourier Transform infrared spectroscopy (ATR-FTIR) and atomic force microscopy (AFM).

Experimental

Chemicals and apparatus

Negative PDMS replicas were cast from a classical cross-type silicon positive relief master using Sylgard 184 silicone elastomer and curing agent (Dow Corning Corporation, Midland, MI, USA). The PDMS replicas and flat slabs were cleaned using HPLC grade methanol (Fisher Scientific, Nepean, ON, USA) prior to microwave plasma treatment. Two microwave digestion ovens were used to generate the microwave plasma: a dedicated microwave digestion oven of maximum power 1200 W (Mars 5, CEM Corporation, Matthews, USA) and a microwave digestion oven modified from a domestic microwave oven of maximum power 600 W (MK-I, Xin-Ke Institute of Applied Microwave Technology, Shanghai, China). Both microwave ovens operate at 2450 MHz. A pressure gauge (Baratron 127AA-00100, MKS Instruments, Massachusetts, USA) with associated electronic readout (PDRC1C, MKS Instruments) was used to measure the pressure of the glass plasma chamber during pump down and gas filling. The maximum operating pressure of the pressure gauge was 100.00 Torr. A syringe pump (NE-1000 Multi-Phaser, New Era Pump System, Farmingdale, NY, USA) equipped with a 250 µL microsyringe (Hamilton Company, Reno, NV, USA) was used in the leakage test of the microchips.

Preparation of PDMS slabs

The Sylgard 184 silicone elastomer and curing agent were degassed in a vacuum dessicator at 10–20 Torr for 30 min. Approximately 10 g of the polymer base and curing agent (weight ratio = 10∶1) was mixed thoroughly in a beaker and transferred to a Petri dish. The silicon positive relief master was placed in the Petri dish for negative replica production. The mixture was cured in an oven at 100 °C for one hour.¹¹ The PDMS plate or negative replica (thickness = 1–2 mm) was peeled off from the Petri dish and cut into a desirable size (typically 2 cm × 4 cm).

Microwave plasma generation and plasma treatment

A schematic diagram of the microwave plasma chamber is shown in Fig. 1. A 250 mL rectangular glass bottle of dimensions of approximately 6 cm × 6 cm × 13 cm and a 3 cm circular opening (Schott Duran, Mainz, Germany) was used as the chamber. The bottle was sealed with a rubber stopper (diameter = 4 cm). The original screw cap of the bottle was used to keep the stopper in place. A through-hole of diameter 0.6 cm was bored at the center of the stopper and the screw cap for vacuum feed-through. A PTFE disk (diameter = 2.9 cm, thickness = 0.3 cm) with a 0.6 cm opening at the center was placed over the rubber stopper to shield the rubber stopper from the microwave plasma. The vacuum feed-through is an L-shaped J.Young vacuum valve (POR/3/RA, J.Young Scientific Glassware Ltd., London, UK) with a pair of extended glass arms. One of the arms (length = 13 cm, o.d. = 0.6 cm, i.d. = 0.3 cm) was fed through the rubber stopper and served as a gas inlet and as a support of the quartz rack for the PDMS slabs (see below). The other arm of the J.Young valve was connected to the vacuum pump and oxygen supply via a brass Swagelok connector (6 mm tube) and a tee (Fig. 1). The glass chamber was detached from the gas lines at the Swagelok connector prior to microwave plasma treatment. The detachable quartz rack is a slightly tapered quartz tube (length = 4 cm, i.d. = 0.6 cm, o.d. = 0.9 cm) with two quartz hooks (length = 0.5 cm, diameter = 0.2 cm). Two PDMS slabs can be anchored on the hooks simultaneously for plasma treatment.

Fig. 1 Schematic diagram of the plasma chamber.

Successful sealing of the PDMS replica and flat slab requires meticulous cleaning of the plasma chamber and the PDMS slabs. The chamber and the J.Young connector were rinsed with distilled water and methanol in turn before the experiment. The PDMS replica and PDMS flat slab were placed in a beaker of distilled water and subjected to ultrasonic cleaning for 10 min. The PDMS plates were then rinsed with methanol and fastened on the quartz rack. The chamber was sealed with the rubber stopper and the screw cap and pumped down to 5 Torr. Methanol on the PDMS plates dried rapidly at that pressure. Oxygen (inlet pressure = 5 psi) was then introduced into the chamber by opening the valve to the gas cylinder and closing the valve to the pump for approximately 3 s. The gas pressure in the chamber increased to approximately 100 Torr. The chamber was finally pumped down to 2–3 Torr. The J.Young valve was closed and the chamber was detached from the Swagelok connector. The pressure in the chamber stayed constant for at least 5 min.

The glass chamber was placed upright in the microwave oven (with the J.Young valve at the bottom). Forward power of 150 W is sufficient to produce a plasma inside the chamber. To avoid damage of the microwave power supply by the reflected microwave power, 80 mL of water was placed in the control vessel of the CEM microwave digestion oven to absorb the excessive electromagnetic energy. Similarly, a beaker containing 200 mL of distilled water was placed in the MK-I oven to absorb the reflected microwave power. Depending on the input power and the location of the chamber in the oven, exposure duration of 5 to 60 s was needed to activate the PDMS surface for irreversible sealing. The PDMS plates were placed in conformal contact immediately after plasma treatment to form the irreversible seal.

The input power was selected via the front panel of the microwave ovens. The input power could be set explicitly using the CEM Mars 5 oven. The input power of the Xin-Ke MK-I oven was indicated as low, mid-low, mid, mid-high, and high levels only. The input power of each level was estimated from the duration of microwave irradiation and the temperature rise of the 200 mL water in the beaker. To estimate the power of the microwave plasma, the glass chamber was placed in the oven and irradiated with microwaves to generate a plasma for a fixed period of time. The increase in temperature of the water in the control vessel or the beaker was used to estimate the microwave energy absorbed by the water. The difference in input energy and the energy deposited in the water was used to estimate the power of the microwave plasma.

Safety precautions

Precaution should be taken to avoid explosion and implosion of the glass bottle. The bottle should be pumped to 2 or 3 Torr or higher and filled to atmospheric pressure or lower. The bottle should be filled with oxygen slowly to avoid ejection of the rubber stopper and associated gas lines due to sudden increase in gas pressure. The chamber should not be used if there is any crack or flake on the wall. The glass chamber is hot (up to 100 °C or higher) after microwave irradiation and must be handled with proper heat insulation.

To avoid damage to the microwave oven, a glass of water should be placed in the oven to absorb the reflected microwave power. The temperature of the water should be checked after each experiment and kept below the boiling point.

Leakage test of the PDMS chip

The PDMS chips were tested for leakage by applying a flow of water of 100 µL min⁻¹ in the microchannel using a syringe pump. The stainless steel needle of a microsyringe was replaced with a CE capillary (o.d. = 330 µm, i.d. = 50 µm, length = 20 cm). The syringe and the capillary were filled with distilled water. The free end of the CE capillary was connected to the end of the microchannel by insertion of the capillary into the PDMS slab. Water was pumped through the microchannel (length = 3.5 cm, cross sectional area = 50 × 50 µm²) continuously for 5 min. The channel was observed under an optical microscope at 10× magnification to check for leakage, if any.

Surface characterization using AFM and ATR-FTIR

The plasma-treated PDMS surface was examined using an atomic force microscope (AFM) (Nanoscope IIIa, Digital Instruments, Santa Barbara, California, USA) operated in tapping mode. The AFM tip (n-type silicon) was mounted on a cantilever (Fluoroware Inc., German). The force constant was 45–61 N m⁻¹. The resonance frequency was 270–290 kHz. Sensitivity was 0.06–0.1 V nm⁻¹. The image size was 512 × 512 pixels.

ATR-FTIR spectra of the PDMS chips were acquired using a BioRad FTS165 spectrometer (BioRad Laboratory, Hercules, California, USA) with a single-pass ZnSe ATR cell (PCL-11M, Harrick Scientific Corporation, NY, USA). Spectral resolution was 4 cm⁻¹. Each spectrum was an average of 32 scans.

Results and discussion

Plasma generation and treatment

A reason of uneven heating in a microwave oven is inhomogeneous distribution of microwave energy in the oven.^12,13 The microwave “node” size is approximately 6 cm. The node pattern varies with vertical position in the oven. To attain homogeneous time-integrated microwave irradiation, a turntable is often used to rotate the food or the samples in the oven. In this study, microwave plasma was generated in a glass chamber containing 2–3 Torr of oxygen using a microwave digestion oven (CEM Mars 5) and a domestic microwave oven (Xin-Ke MK-I). If the chamber was placed on the turntable directly, the plasma in both ovens did not fill the whole chamber and the intensity of plasma emission changed with the location of the glass chamber in the oven because of inhomogeneous distribution of microwave energy in the ovens. Spatial inhomogeneity of microwave energy distribution was especially prominent in the Xin-Ke MK-I oven. The microwave plasma dimmed or extinguished at the left corner towards the back of the oven. For both microwave ovens, the emission intensity of the microwave plasma was higher and appeared to be more homogeneous spatially if the plasma chamber was placed near the exit of the waveguide. The plasma also filled the whole chamber and was large enough to cover the entire 4 cm PDMS slab. The exit of the waveguide was located at the top of the ovens. In the following sections, the chamber was elevated to a position close to the exit of the waveguide, with the top of the plasma chamber at approximately 5 cm from the ceiling of the ovens.

The modes of power output of the CEM Mars 5 and Xin-Ke MK-I ovens are different. The Xin-Ke MK-I oven delivers the maximum power (approximately 560 W) during the ON period of the duty cycle. The duty cycle depends on the power setting. For example, “high” power means 100% duty cycle and “mid” power means approximately 50% duty cycle. The duration of one duty cycle is fixed at 30 s. The long duty cycle can cause difficulty in plasma treatment. For example, at the lowest power setting of “low”, the duration of microwave plasma ON was approximately 2 s in each 30 s duty cycle. The duration of plasma OFF is 28 s. Such a long waiting period between power ON renders plasma treatment irreproducible. At the next level of power setting of “mid-low”, the duration of plasma ON is approximately 7 s in each duty cycle, which is already long enough to cause damage to the PDMS surface (see below). The CEM Mars 5 oven is more flexible in power setting. Both the peak power and the duty cycle can be adjusted. Peak power of 300, 600, and 1200 W are available. Duty cycle can be set arbitrarily. For example, if average input power of 180 W is desired, the oven can be set to 30% duty cycle of 600 W or 60% duty cycle of 300 W. In addition, the duration of a duty cycle of the CEM oven is relatively short (1 duty cycle = 0.2 s). The microwave plasma would go on-and-off in both ovens if the duty cycle was less than 100%. The modulation frequency of the emission intensity of the microwave plasma, however, is much higher in the CEM oven.

The power of microwave plasma was estimated from the increase in temperature of the water in the oven after microwave irradiation. The measurement aimed to provide a rough estimation of the order of magnitude of the plasma power. Rigorous thermal measurement strategy was not applied. Table 1 summarizes the approximate plasma power versus input power for the ovens. The duration of irradiation was 1 min. Using the CEM Mars 5 oven with input power of 180 W (power setting = 600 W, duty cycle = 30%), the plasma power was approximately 30 W. Using the Xin-Ke MK-I oven with “mid-low” power (input power approximately 150 W), the estimated microwave power was 50 W. The power of the microwave plasma depends on the configuration (especially the duty cycle) and input power setting of the microwave oven. The magnitude of the plasma power, however, is comparable to the input power of commercial RF plasma cleaners.

Table 1 Estimated power of the microwave plasma and the microwave ovens

Microwave oven	Power setting	Input power/W	Microwave plasma power/W
Xin-Ke MK-I	Mid-low	150	50
	High	560	200
CEM Mars 5	600 W, 30% duty cycle	180	30
	600 W, 80% duty cycle	480	90

---

The power of the microwave plasma increases with the input power (Table 1). For the Xin-Ke MK-I oven, plasma power was approximately 200 W at “high” input power (100% duty cycle). Such high plasma power, however, is not desirable. The PDMS surface becomes opaque within a few seconds of plasma treatment. Sealing of the PDMS slabs also fails. In fact, using the Xin-Ke MK-I oven, the ON period of the microwave power supply varies from 7 to 30 s within one duty cycle for power setting of “mid-low” (one level above the minimum input) to “high” (maximum). Plasma treatment of one duty cycle, no matter the power setting, is already too long and the PDMS slabs will not seal. Plasma treatment of 5 s, however, results in adequate sealing of the PDMS slabs. (The microwave power was interrupted 5 s after plasma ignition). An apparently more gentle plasma is produced using the CEM Mars 5 oven (average input power = 180 W; input power setting = 600 W or 300 W with duty cycle = 30% and 60%, respectively), probably due to the much shorter duration of each duty cycle (0.2 s). Plasma treatment of 30–60 s is adequate for irreversible sealing of the PDMS slabs. The surface of the PDMS slab turns white and becomes opaque upon plasma treatment of 4 min or longer.

In addition to the consideration of input power, oxygen gas pressure in the chamber must be kept at 5 Torr or below for successful PDMS sealing. Although no visible damage was observed on the PDMS surface after plasma treatment at oxygen pressure >5 Torr for 30–60 s, sealing of the PDMS slabs failed.

Surface characterization of plasma-treated PDMS

Fig. 2 shows the surface topology of plasma-treated PDMS slabs. The PDMS surface was treated with microwave plasma for 1 min (input power = 180 W, oxygen gas pressure = 2.6 Torr). Fig. 2a shows typical AFM image of a plasma-treated PDMS surface. The surface is relatively smooth and featureless. The root-mean-square (RMS) roughness is approximately 0.20 nm, of similar magnitude of the roughness of a glassy surface.¹⁴ In contrast, the RMS roughness of untreated PDMS surface is 0.35 nm. In addition, fine cracks were found occasionally on the plasma-treated PDMS surface (Fig. 2b). The cracks (width = 50–100 nm and length ≥5 µm) indicate oxidation of the polymer surface to bear silica-like properties.¹⁵ The cracks may be a result of mechanical stress exerted on the silica-like surface during handling of the PDMS slabs or a result of elasticity difference between the bulk and the oxidized layer.^15,16 Interestingly, the roughness of PDMS surface increases significantly if two PDMS slabs are sealed irreversibly and then torn off by force (Fig. 2c). RMS roughness of the torn surface is larger than 20 nm. Irreversible sealing is a result of Si–O–Si bond formation between two PDMS substrates. Peeling of the PDMS slabs may rupture the PDMS bulk near the interface rather than breaking the relatively strong Si–O–Si bonds. The rough surface also accounts for the inability of re-sealing the PDMS chips once the two irreversibly sealed PDMS slabs are separated.

Fig. 2 AFM images of (a) plasma-treated PDMS surface, (b) plasma-treated PDMS surface with micro-cracks, (c) surface of a PDMS slab that was torn off from an irreversible-sealed microfluidic device.

The ATR-FTIR spectrum of the PDMS surface with 4 min oxygen plasma treatment is shown in Fig. 3. The input power of the CEM Mars 5 microwave oven was 180 W. In agreement with the findings of Hillborg and Gedde,⁷ the general pattern of the IR spectrum of the polymer did not change with the duration of plasma treatment. A broad absorption peak at 3400 cm⁻¹, however, appeared upon plasma treatment (Fig. 3). The peak corresponds to the hydroxyl groups that formed on the PDMS surface during plasma treatment.⁷ The peak height (absorbance) increases with the duration of plasma treatment (Fig. 4). In the last section, the optimal plasma treatment duration was found to be 30–60 s using the CEM oven. The irreversible seal fails with prolonged plasma treatment. Increase in the quantity of oxygen insertion (stronger absorption at 3400 cm⁻¹ in Fig. 4) does not result in a stronger seal between the PDMS slabs. The mechanism of the loss in sealing ability is not known currently, but is probably related to favorable bond formation between neighboring oxygen groups as the population of the oxygen groups on the surface increases.

Fig. 3 ATR-FTIR spectrum of microwave plasma-treated PDMS slab. The arrow indicates the absorption peak at 3400 cm⁻¹ that appeared after oxygen-plasma treatment.

Fig. 4 ATR-FTIR absorption peak height (absorbance) at 3400 cm⁻¹versus the duration of plasma treatment.

Applicable pressure range of the PDMS microchip

Water flow of 100 µL min⁻¹ was pumped through the microchannel (50 µm × 50 µm × 3.5 cm) of two PDMS chips separately. The flow rate is much higher than normal CE flow rate in order to determine the applicable pressure to the chips. The PDMS chips appeared intact with no water leakage over a pumping duration of 5 min. The pressure drop inside the microchannel can be estimated using the following equation,¹⁷
where Δp = pressure drop inside the channel; f = friction coefficient; l = channel length; d_h = hydraulic diameter; ρ = density of water; v = linear velocity of the flow. The calculated pressure drop in the channel is approximately 42 psi. The applicable pressure range of the microwave plasma-treated PDMS microchips is comparable to literature value for PDMS chips that were treated with RF plasma.⁶

Conclusions

Treatment of PDMS surfaces using a microwave-oven-generated plasma for irreversible sealing of PDMS slabs was demonstrated. The plasma chamber is readily constructed in a chemistry laboratory. Economic “kitchen microwave oven” or the commonly available dedicated microwave digestion oven can be used to generate the microwave plasma. The microwave plasma can potentially be applied to other plasma cleaning or treatment processes as a low cost alternative to dedicated plasma cleaners.

Acknowledgements

This work was supported by a grant from the Research Grant Council of the Hong Kong Special Administrative Region, China (Project No. HKU 7009/04P), Seed Funding Programme for Basic Research of HKU (Project No. 200411159031), and NSFC key project (20299035) of National Natural Science Foundation of China.

References

1. J. C. McDonald and G. M. Whitesides, Acc. Chem. Res., 2002, 35, 491–499.
2. J. M. K. Ng, I. Gitlin, A. D. Stroock and G. M. Whitesides, Electrophoresis, 2002, 23, 3461–3473.
3. D. Qin, Y. Xia and G. M. Whitesides, Adv. Mater., 1996, 8, 917–919.
4. D. C. Duffy, J. C. McDonald, O. J. A. Schueller and G. M. Whitesides, Anal. Chem., 1998, 70, 4974–4984.
5. Y. Xia and G. M. Whitesides, Annu. Rev. Mater. Sci., 1998, 28, 153–184.
6. J. C. McDonald, D. C. Duffy, J. R. Anderson, D. T. Chiu, H. Wu, O. J. A. Schueller and G. M. Whitesides, Electrophoresis, 2000, 21, 27–40 .★ A review paper on the fabrication and applications of PDMS microfluidic devices. The fabrication process was described in detail in the first half of the paper, from the first step of soft lithography to the final step of sealing of microfluidic devices.
7. H. Hillborg and U. W. Gedde, Polymer, 1998, 39, 1991–1998.
8. K. W. Ro, K. Lim, H. Kirn and J. H. Hahn, Electrophoresis, 2002, 23, 1129–1137.
9. B. Wang, Z. Abdulali-Kanji, E. Dodwell, J. H. Horton and R. D. Oleschuk, Electrophoresis, 2003, 24, 1442–1450.
10. B. T. Ginn and O. Steinbock, Langmuir, 2003, 19, 8117–8118 .★ This paper was the first to describe microwave-oven-generated plasma for polymer surface treatment.
11. Sylgard 184 silicone elastomer product information, Dow Corning Corporation, Midland, MI, USA.
12. A. Steyn-Ross and A. Riddell, The Physics Teacher, 1990, 28, 474–476 .★ The vertical and lateral distribution of the microwave energy in a microwave oven cavity was shown and discussed.
13. M. Vollmer, Physics Education, 2004, 39, 74–81. ★ The physics of microwave plasma in a microwave oven was discussed in detail. Magnetron, waveguide, and standing waves in the resonant cavity were discussed.
14. P. K. Gupta, D. Inniss, C. R. Kurkjian and Q. Zhong, J. Non-Cryst. Sol., 2000, 262, 200–206.
15. G. Bar, L. Delineau, A. Hafele and M. H. Whangbo, Polymer, 2001, 42, 3527–3632.
16. H. Makamba, J. H. Kim, K. Lim, N. Park and J. H. Hahn, Electrophoresis, 2003, 24, 3607–3619.
17. N.-T. Nguyen and S. T. Wereley, Fundamentals and Applications of Microfluidics, Artech House, Boston, 2002, pp. 36–38.

---