Boiling regimes in uncoated polydimethylsiloxane ... - Springer Link

Heat Mass Transfer (2010) 46:1253–1260 DOI 10.1007/s00231-010-0655-x

ORIGINAL

Boiling regimes in uncoated polydimethylsiloxane microchannels with a fine wire heater Yanyan Lu • Fen Wang • Hao Wang

Received: 24 November 2009 / Accepted: 26 July 2010 / Published online: 5 August 2010 Ó Springer-Verlag 2010

Abstract Polydimethylsiloxane (PDMS) is a type of gas permeable media widely used in microfluidic applications. In this work multiphase patterns and boiling curves in PDMS square microchannels were experimentally investigated. Very fine platinum wires with diameter of 50 lm were embedded through the microchannels and serving as heater. The multiphase patterns were visualized by means of high speed CCD camera with microscope. Curves of temperature versus heat flux on the wire heaters were plotted. Based on the evolution of multiphase patterns, five boiling regimes were classified, that is, single phase, bubble formation, slug formation, slug dominated and dry out. Interestingly, the bubbles were generated from the channel walls rather than the heater surface, and so-called ‘‘droplets-in-bubble’’ phenomenon drew attention in which bunches of microdroplets kept forming, growing, and disappearing within the big bubbles. The boiling curves were plotted and compared to boiling in open space and in glass tubes. The heat transfer in the PDMS microchannels got deteriorated when the bubbles formed.

1 Introduction Gas or vapor bubbles are increasingly employed or encountered in microdevices such as various bubble-powered actuators [1], pumps [2, 3], engines [4], bio-particle actuators [5], and also cooling devices by means of boiling heat transfer [6, 7]. Numerous studies have been carried out

Y. Lu F. Wang H. Wang (&) Department of Energy and Resources Engineering, College of Engineering, Peking University, Beijing 100871, China e-mail: [email protected]

to clarify bubble dynamics in microstructure. Among them, bubbles in microchannels have been paid great attentions [8–12]. Recent years polydimethylsiloxane (PDMS), a type of porous material, are widely used in microchips due to its advantages in visualization and fabrication [13–17]. Most PDMS chips are used at room temperature, however heating conditions with consequent multi-phase flows are still encountered in many cases [7, 16–18]. For example, Choban et al. [16] and Shah et al. [17] used PDMS in fuel cell as membrane, taking advantage of PDMS porosity. In process such as polymerase chain reaction (PCR), heating is applied and bubble formation should be taken care [19– 21]. Due to the increasing demand of heating or cooling, it is necessary to have a comprehensive understanding about the possible boiling process and bubble dynamics in PDMS channels. Related study could also help the understanding of boiling in porous media [6]. Bubble formation in PDMS channels have been investigated [22–24]. Shin et al. [19] expressed that the bubble formation was due to the porosity of PDMS and then coated the channel with Parylene to avoid bubble. Niu et al. [25] rooted the bubble formation to the hydrophobicity of the PDMS chamber surface and used wetting process to avoid the bubble formation. Toriello et al. [26] used a similar wetting process. Liu et al. [21] found that the bubble formation was strongly related to the microfeatures inside the reactors and inside the chip bonding interface, especially the part near the inner corners of the microreactors. Gas permeability of PDMS and the wetting property of PCR sample also have influence on the air bubble formation. Very limited systematical studies, however, were available for bubble dynamics and flow pattern during heating process in PDMS channels especially for those uncoated. Among the very few studies, Huh and Kim [27, 28]

123

1254

Heat Mass Transfer (2010) 46:1253–1260

made an experimental investigation on flow boiling in a coated PDMS rectangular microchannel with serpentine platinum microheater array on a Pyres substrate. They studied the local boiling heat transfer coefficients and single bubble inception, growth, departure, and elongated bubble behavior in coated channels. Bubbly/slug flow and elongated slug/semi-annular flow were observed. As above, it is necessary to conduct systematical studies to explore possible features of phase change in gas-permeable PDMS channels under heating. In the present work, uncoated PDMS-glass microchannels were fabricated and fine wires inside served as heater. The bubble and droplets behaviors during the heating process in the channels were systematically investigated.

2 Experimental setup

(a) Power supply Thermocouple A V

PDMS chip

Data acquisition

Tested channel Microscope

(b) Glass slide Reservoir Heater wire in microchannel

A: Put master into the dish

B: Pour prepolymer and cure

C: Peel PDMS replica from master, punch holes and put the thin Pt wire

D: Oxidize PDMS replica and

The experimental setup consisted of four parts, i.e., test section, visualization system, data acquisition system and power supply as schematically illustrated in Fig. 1. The test section was a PDMS microchannel with a fine platinum wire inside the microchannel. The depth of the channel was around 120 lm and the width and length was 180–220 lm and 7–10 mm, respectively. The heater was a fine platinum wire with diameter of 50 lm and length 10–15 mm. The PDMS microchannel was fabricated by means of standard soft lithography method [29]. The procedure of the fabrication is summarized in Fig. 2. The pattern of channels in the photomasks was replicated in SU8-2050. The microfluidic channels were molded by casting a layer of PDMS prepolymer mixture with a mass ratio of A:B = 10:1 in a disposable culture dish. After peeling the

Heater wire

Master

PDMS

glass slide in plasma and seal

Fig. 2 The fabrication procedure of the microchannels

PDMS replica from the master, punching holes and putting the thin platinum wire inside the channel. Holes were punched at the ends of channels served as reservoirs. Then oxygen plasma (Diener Electronic: Femto) surface treatment was employed to treat the PDMS and the glass slide for 30 s at 75 W for the purposes of bonding them together. Note the plasma treatment could enhance the wettability of the PDMS surfaces, which however, would not last forever since the wettability would practically recover after some time, e.g. 3 days later in [20]. In this paper, all experiments were conducted after the PDMS wettability recovered, and the contact angle of water on the glass slide maintained about 70°, except the special one explained in Sect. 3.2. The visualization system consisted of a CCD (Monochrome Cooled Digital Camera Head DS-Qi1Mc, Nikon, Japan), a high speed CCD camera (X-MOTION, AOS technologies AG), a fluorescence microscope (TI-U INVERTED, Nikon, Japan) and a computer. Data acquisition system consisted of several K-type thermocouples, a data acquisition device for thermocouples (UTL/D08LS1V0N, You-Tuo, China) and two universal meters (METRA Hit 23s; Agilent U1251A). Power is supplied to the platinum wire by a direct current power supply (WYK5030, Hua-Tai, Yang-Zhou, China). 2.1 Temperature measurement

Three holes for thermocouples

Fig. 1 Experimental setup. a A schematic illustration of the experimental system; b PDMS chip with microchannel and heater wire

123

In this work, direct current passed through the platinum wire and caused the temperature increasing. The voltage and current supply to the heater wire were measured by two


1255

universal meters. By measuring the current and voltage of the platinum wire, the heating power as well as the resistance of the wire was obtained. Based on the good correlation between the resistance and temperature of platinum, the temperature of the whole wire in the circuit was calculated. An error analysis showed that the overall uncertainty of the wire temperature measurement was ±1 K. The temperature of the PDMS 2 mm from the channel wall was qualitatively estimated by three thermocouples located are shown in Fig. 3. They were inserted into the PDMS slab. The thermocouples were calibrated prior to the experiments and the final temperature of the PDMS near the channel was the mean temperature of the three measured points. In the following section, ‘‘PDMS temperature’’ is represented by this averaged temperature.

2.2 Conditions and operation procedure In the experiment the pressure on the microchannel kept at one atmosphere by keeping the two reservoirs opens to atmosphere. Deionized water (DI water) used as working fluid. In most experiments the fluid was not degassed. In some experiments degassing was conducted by boiling twice and vacuuming for 30 min, and the results were compared to those without degassing, as shown in Fig. 12a. Before each experiment, working fluid at room temperature was pumped in until the channel and the reservoirs were filled up (the volume capacity of each reservoir was about 70 lL). Power was then supplied to the platinum wire. The evolution of the multiphase patterns was recorded. The voltage imposed on the heater wire was increased step by step. Each time the voltage was increased, the parameters were recorded 2 min after.

(a) PDMS

~2.5 ~2.5

Thermocouples ~2

~0.2

3 Experimental results 3.1 Direct observation of boiling regimes

~10 Glass slide

(b) ~3

Fig. 3 Thermocouple locations (mm) for PDMS temperature measurement: a top view; b front view

A series of photographs for boiling behaviors were caught by CCD camera. Based on the observations, from low to high heat flux the process was divided into five regimes. That is, single phase regime when the channel was fulfilled with only liquid, bubble formation regime when bubbles appeared, slug formation regime when bubbles were big enough to occupy the channel section, slug dominated

Fig. 4 Sequence of images show the varying of droplets in a big bubble (2,000 frames/s, wire heat flux 0.8 9 105 W/m2, 47°C): a droplets were inside the big parent bubble; b droplets merged and left a blank area; c tiny droplets appeared and new cycle started; d droplets kept growing and merging

123

1256


regime when the channel was occupied by huge slugs, and dry out regime when there was no obvious liquid phase found in channel. Details are introduced as following. 3.1.1 Bubble formation regime and droplets-in-bubble phenomena When the input power was about 0.5 9 105 W/m2, wire temperature 40°C, and PDMS temperature 30°C, bubbles were observed in the channel, as shown in Fig. 4a. Distinct features of the bubbles were noticed: (a)

The bubbles rooted from the PDMS channel walls instead of the heater wire, even though the heater wire must have higher temperature. The bubbles were in

Heater wire

Droplets on walls

Evaporation

Bubble surface Vapor/gas

Condensation

Liquid bridge PDMS side wall

Fig. 5 A hypothetical scheme to explain droplets-in-bubble phenomenon

Fig. 6 The draining out process of a droplet in bubble (heater wire heat flux 0.5 9 105 W/m2, 41°C). a A big droplet within the parent bubble was going to contact the bubble boundary; b, c the droplet was drained out; d tiny droplets appeared in blank area

123

big sizes (at least tens of micrometers). Most of them were on the side wall, and touched the top or bottom wall at the same time. No bubble detachment was found during experiments. (b) The wall temperature required for the bubble formation was low. The wire temperature was around 40°C and PDMS temperature about 30°C. Therefore the temperature of the channel wall must be between these two temperatures, which was far below the saturation temperature 100°C. Obviously the bubbles should content mixture of vapor and also noncondensable gases. The gases could origin from the porosity of PDMS, or from the wedge between PDMS and glass slide [21]. Comparing experiments as shown in Fig. 12a indicated that the degassing or not of the working fluid did not alter the bubble formations, which should be because the channels already stored enough gases. (c) Even more interestingly, there were many microdroplets growing in the established big bubbles. An example is illustrated in Fig. 4. A big bubble (named ‘‘parent bubble’’ in Fig. 4a) formed on the channel wall. Its inside was not blank but bunch of microdroplets were found as marked. These droplets had different sizes, bunches of them were tiny with diameter \10 lm, a few of them could reach tens of micrometers. One of important behaviors of the droplets was that they were not stable but varying quickly. They were repeating a cycle of forming, growing, merging and vanishing as illustrated in Fig. 4a–d. In the blank area where they vanished, the


1257

Fig. 7 Bubble slug and droplets inside (heater wire heat flux 1.0 9 105 W/m2, 50°C) with different wall wettability: a hydrophobic (contact angle on the glass and on PDMS was about 70° and 103°, respectively); b hydrophilic (contact angle on the glass and on PDMS was about 25°)

evaporation on the hot side and condensation on the cold sides. The condensation produced droplets on channel walls. Small droplets grew into big ones. When one was big enough to have a contact with the bubble boundary, it connected to the liquid surrounding the bubble, and formed a liquid bridge as marked in Fig. 5. The big droplet was in this way drained out in a flash. In our experiments pieces of experimental evidence were captured. One of them was shown in Fig. 6. It is seen within a time period of \0.001 s the big droplet was drained out and vanished. Fig. 8 Slug dominated 1.3 9 105 W/m2, 78°C)

in

channel

(heater

wire

heat

flux

newly born tiny droplets appeared and a new cycle started. A hypothetical scheme is made in Fig. 5 trying to understand the mechanisms behind the droplets-in-bubble phenomenon. That is, for a big bubble on the channel wall, its one side was close to the heater wire which had high temperature, while the other sides were on the channel walls which were relatively cold. Therefore, a non-equilibrium state was established which leaded to

3.1.2 Slug formation When the wire heat flux was increased to around 0.6 9 105 to 1.2 9 105 W/m2, wire temperature 50–70°C and PDMS temperature about 34–38°C, the bubbles generated near the sidewall were promoted to expand or coalescent with each other, and formed slugs as shown in Fig. 7. The characteristics of this regime is that the bubble was big enough to connect to the heater wire or even the opposite channel wall, thus a vapor/gas slug formed blocking the channel, which significantly affected the heat and mass transfer processes in the channel.

Fig. 9 Water with Rhodamine B labeling (heater wire heat flux 1.7 9 105 W/m2, 105°C). a A droplet on the sidewall was going to contact the wire; b, c bubbles formed in the droplet; d droplet shrank back to the sidewall

123

1258


with good wettability. The droplets in Fig. 7b were seen much flatter and with looser boundaries.

The droplets-in-bubble phenomenon was continued in the slugs, as shown in Fig. 7. Now the droplets had larger sizes. The comparison of Fig. 7a, b indicates the influence of the wall wettability on the droplet patterns. The channel walls in Fig. 7a were with poor wettability and in Fig. 7b

3.1.3 Slug dominated At higher heat flux level, 1.3 9 105 to 1.6 9 105 W/m2 with wire temperature 70–90°C, PDMS temperature 38–43°C, the slugs expanded along the channel and merged with each other. The whole channel was then occupied by one or several huge slugs. Continuous liquid bulk was hard to find. Its volume has been replaced by continuous vapor/ gas bulk. That is to say, from slug formation regime to slug dominated regime, liquid phase became minority in the channel. Spherical bubbles cannot be observed, while liquid droplets were left everywhere in the channel, as pointed out in Fig. 8. The liquid film sustained around the heater wire and the liquid droplets helped in heat dissipating from the heater wire.

Fig. 10 Apparent dry out in the channel (heater wire heat flux 2.0 9 105 W/m2, 110°C)

Fig. 11 Unsteady two phase flow at one end of the channel (heater wire heat flux 4.7 9 105 W/m2, 280°C)

Fig. 12 Curves of temperature versus heat flux: a curves in three PDMS microchannels (width 0.2 mm, height 0.12 mm), with one of them using degassed liquid; b comparison to the curve in open pool space (capacity 50 mL, wire length 15 mm); c comparison to the curves in glass tube (diameter 1 mm) and in PDMS tube (diameter 1 mm)

Heat flux ( W/m2)

(a)

3.5x10

5

3.0x10

5

2.5x10

5

2.0x10

5

1.5x10

5

1.0x10

5

5.0x10

4

microchannel 1 microchannel 2 microchannel degassed Slug formation Bubble

Dry out

formation

Slug dominated

Single phase

0.0 0

20

40

60

80

100

120

140

160

Temperature of heater wire ( OC) 3.5x10

6

3.0x10

6

2.5x10

6

2.0x10

6

1.5x10

6

1.0x10

6

5.0x10

5

(c) in open space microchannel 1 microchannel 2 microchannel 3

Heat flux ( W/m2)

Heat flux ( W/m2)

(b)

0.0

5

3.0x10

5

2.5x10

5

2.0x10

5

1.5x10

5

1.0x10

5

5.0x10

4

glass tube 1 glass tube 2 PDMS tube PDMS channel

0.0 0

20

40

60

80 100 120 140 160 O

Temperature of heater wire ( C)

123

3.5x10

0

20 40 60 80 100 120 140 160 180

Temperature of heater wire ( OC)


Further as the heat flux increased to 1.6 9 105 to 2.0 9 105 W/m2 and wire temperature to 80–120°C, the vapor/gas domain was more dominated as shown in Fig. 9. The droplets around the heater wire evaporated up. Only the droplets on the sidewalls were left. Rhodamine B in red color was used as a marker to better distinguish liquid and vapor. A droplet was seen touched the heater wire, and small bubbles then formed on the wire surface as marked in Fig. 9b, c. The droplet was finally evaporated and shrank back to the channel sidewall as shown in Fig. 9d.

1259

its slope at some point. For the glass tubes, it was about 100°C. At that time the bubbles started to block the tube and the heat transfer was deteriorated. For the PDMS tube, it was lowered to about 85°C. For the PDMS microchannel, it was even lowered to about 60°C. We can see that the deterioration of heat transfer in the PDMS microchannels occurred at very early stage. The microscale space, the gas permeable characteristics of PDMS, and the structure of the PDMS channel should all contribute to this result.

3.1.4 Dry out Wire heater

Liquid

Right reservoir

Left reservoir

When the heat flux achieved about 2.0 9 105 W/m2 and the temperature of platinum was about 110°C, PDMS temperature about 46°C, an apparent dry out in the channel started. That was, no obvious liquid phase could be found in the major part of the channel, as shown in Fig. 10. By checking the two ends of the channel, it was seen that the liquid in the reservoirs was difficult to enter the channel due to the obstacle of the vapor slug inside. Sometimes prompt vapor–liquid exchanging processes at the two ends of the channel could be observed which looked like liquid films flashing over as shown in Fig. 11.

Single phase

Wire heater

Liquid

Droplets

Parent bubble

3.2 Boiling curves Bubble formation

With increasing the voltage step by step, the voltage and current on the heater wire were recorded. The heat flux of the wire surface and the averaged wire temperature were calculated and plotted. Figure 12a shows the curves for the experiments in three same-sized microchannels. For the sake of keeping the microchannel from damage, the heat fluxes were controlled below 3.5 9 105 W/m2 and the wire reached as high as 170°C. The five boiling regimes have been classified based on visualizations as discussed in Sect. 3.1. However as seen in Fig. 12a, it was hard to distinguish these five regimes based on the boiling curve, since the curve slope barely changed. Obviously these curves were quite different from typical boiling curves. As the comparison shown in Fig. 12b, for the boiling in open space, the curve slope increased when the wire temperature beyond 100°C, which is well known due to the enhancement of nucleate boiling. For the PDMS microchannels, on the contrary, the curve slope in single phase regime was greater than that after, as indicated by the two dashed lines drawn in Fig. 12a. Obviously the deterioration of heat transfer occurred when bubble formation started. In Fig. 12c, mini tubes, i.e. two glass tubes and a PDMS tube, all in 1 mm inner diameter, were compared to a PDMS microchannel. All these curves are seen decreasing

Liquid

Wire heater

Parent bubble

Droplets

Slug formation

Vapor/gas

Wire heater

Bubbles

Droplets

Slug dominated

Vapor/gas

Wire heater

Apparent dry out

Fig. 13 Summarization of five regimes during heating in uncoated PDMS microchannel with wire heater inside

123

1260

4 Conclusions Uncoated PDMS microchannels were fabricated and microwires were embedded through the channel serving as heater. Pool boiling experiments in the channels were conducted. The evolution of multiphase patterns and heat transfer characteristics were investigated. Five boiling regimes were classified based on direct observations. That is, single phase regime when the channel was fulfilled with only liquid, bubble formation regime when bubbles appeared, slug formation regime when bubbles were big enough to occupy the channel section, slug dominated regime when the channel was occupied by huge slugs, and dry out regime when there was no obvious liquid phase can be found in channel. A summarization for the regimes is made in Fig. 13. Bubbles were found rooting from the PDMS channel walls instead of the heater wire, even though the wire must had higher temperature. Microdroplets were found in the bubbles and repeating a cycle of forming, growing, merging and draining. It is speculated that the droplets were due to the condensation on the channel walls. The boiling curves in the PDMS channels were obtained, which were quite different from typical boiling curves. The curve slope was decreased during bubble formation regime, indicating that the bubbles were worsening the heat transfer in the channel. Acknowledgments This work was supported by the National Natural Science Foundation of China, Grant No. 50706001. The authors thank Laboratory of Phase Change and Interfacial Transport Phenomena at Tsinghua University for the assistances during chip fabrication.

References 1. Lin L, Pisano AP (2005) Thermal bubble powered microactuators. Microsyst Technol 1:51–58 2. Tsai J-H, Lin L (2002) A thermal bubble actuated micro nozzlediffuser pump. IEEE 409–412 3. Laser DJ, Santiago JG (2004) A review of micropumps. J Micromech Microeng 14:35–64 4. Lin L (1998) Microscale thermal bubble formation: thermophysical phenomena and applications. Nanoscale Microscale Thermophys Eng 2:71–85 5. Maxwell RB, Gerhardt AL, Toner M, Gray ML, Schmidt MA (2003) A microbubble-powered bioparticle actuator. J Microelectromech Syst 12:630–640 6. Jiang YY, Wang WC, Wang D, Wang BX (2001) Boiling heat transfer on machined porous surfaces with structural optimization. Int J Heat Mass Transf 44:443–456 7. Zhang Y, Bao N, Yu X-D, Xu J-J, Chen H-Y (2004) Improvement of heat dissipation for polydimethylsiloxane microchip electrophoresis. J Chromatogr A 1057:247–251 8. Li HY, Tseng FG, Pan C (2004) Bubble dynamics in microchannels. Part II: two parallel microchannels. Int J Heat Mass Transf 47:5591–5601

123

Heat Mass Transfer (2010) 46:1253–1260 9. Wu HY, Cheng P (2003) Visualization and measurements of periodic boiling in silicon microchannels. Int J Heat Mass Transf 46:2603–2614 10. Wang G, Cheng P (2008) An experimental study of flow boiling instability in a single microchannel. Int Commun Heat Mass Transf 35:1229–1234 11. Peng XF, Wang BX (1993) Forced-convection and flow boiling heat transfer for liquid flowing through microchannels. Int J Heat Mass Transf 36:3421–3427 12. Peng XF, Wang BX (1994) Evaporating space and fictitious boiling for internal evaporation of liquid. Sci Found China 2:55–60 13. Chabinyc ML, Chiu DT, McDonald JC, Stroock AD, Christian JF, Karger AM, Whitesides GM (2001) An integrated fluorescence detection system in poly(dimethylsiloxane) for microfluidic applications. Anal Chem 73:4491–4498 14. Diercks AH, Ozinsky A, Hansen CL, Spotts JM, Rodriguez DJ, Aderem A (2009) A microfluidic device for multiplexed protein detection in nano-liter volumes. Anal Biochem 386:30–35 15. Pan T, McDonald SJ, Kai EM, Ziaie B (2005) A magnetically driven PDMS micropump with ball check-valves. J Micromech Microeng 15:1021–1026 16. Choban ER, Markoski LJ, Wieckowski A, Kenis PJA (2004) Microfluidic fuel cell based on laminar flow. J Power Sources 128:54–60 17. Shah K, Shin WC, Besser RS (2004) A PDMS micro proton exchange membrane fuel cell by conventional and non-conventional microfabrication techniques. Sens Actuators B Chem 97:157–167 18. Maltezos G, Rajagopal A, Scherer A (2006) Evaporative cooling in microfluidics channels. Appl Phys Lett 89:074107-1–074107-3 19. Shin YS, Cho K, Lim SH, Chung S, Park S-J, Chung C, Han D-C, Chang JK (2003) PDMS-based micro PCR chip with Parylene coating. J Micromech Microeng 13:768–774 20. Kim JA, Lee JY, Seong S, Cha SH, Lee SH, Kim JJ, Park TH (2006) Fabrication and characterization of a PDMS-glass hybrid continuous-flow PCR chip. Biochem Eng J 29:91–97 21. Liu H-B, Gong H-Q, Ramallingam N, Jiang Y, Dai C-C, Hui KM (2007) Micro air bubble formation and its control during polymerase chain reaction (PCR) in polydimethylsiloxane (PDMS) microreactors. J Micromech Microeng 17:2055–2064 22. Garstecki P, Fuerstman MJ, Fischbach MA, Sia SK, Whitesides GM (2006) Mixing with bubbles: a practical technology for use with portable microfluidic devices. Lab Chip 6:207–212 23. Singh SG, Kulkarni A, Duttagupta SP, Puranik BP, Agrawal A (2008) Impact of aspect ratio on flow boiling of water in rectangular microchannels. Exp Therm Fluid Sci 33:153–160 24. Francais O, Jullien MC, Rousseau L, Poulichet P, Desportes S, Chouai A, Lefevre JP, Delaire J (2007) An active chaotic micromixer integrating thermal actuation associating PDMS and silicon microtechnology. In: DTIP of MEMS & MOEMS. arXiv:0711.3290v1 25. Niu ZQ, Chen WY, Shao SY, Jia XY, Zhang WP (2006) DNA amplification on a PDMS-glass hybrid microchip. J Micromech Microeng 16:425–433 26. Toriello NM, Liu CN, Mathies RA (2006) Multichannel reverse transcription-polymerase chain reaction microdevice for rapid gene expression and biomarker analysis. Anal Chem 78:7997–8003 27. Huh C, Kim J, Kim MH (2007) Flow pattern transition instability during flow boiling in a single microchannel. Int J Heat Mass Transf 50:1049–1060 28. Huh C, Kim MH (2006) An experimental investigation of flow boiling in an asymmetrically heated rectangular microchannel. Exp Therm Fluid Sci 30:775–784 29. Duffy DC, McDonald JC, Schueller OJA, Whitesides GM (1998) Rapid prototyping of microfluidic systems in poly(dimethylsiloxane). Anal Chem 70:4974–4984