An Experimental Study on Micro-Hydrodynamics of Evaporating/Boiling Liquid Film

(1)

KTH Engineering Sciences

An Experimental Study on Micro-Hydrodynamics of Evaporating/Boiling Liquid Film

Doctoral Thesis by

Shengjie Gong

Division of Nuclear Power Safety Department of Physics Royal Institute of Technology

Stockholm, 2011

(2)

II

TRITA-FYS 2011:52 ISSN 0280-316X

ISRN KTH/FYS/--11:52--SE

ISBN 978-91-7501-165-3

(3)

I

Abstract

Study of liquid film dynamics is of significant importance to the understanding and control of various industrial processes that involve spray cooling (condensation), heating (boiling), coating, cleaning and lubrication. For instance, the critical heat flux (CHF) of boiling heat transfer is one of the key parameters ensuring the efficiency and safety of nuclear power plants under both operational and accident conditions, which occurs as the liquid layers (microlayer and macrolayer) near the heater wall lose their integrity.

However, an experimental quantification of thin liquid film dynamics is not straightforward, since the measurement at micro-scale is a challenge, and further complicated by the chaotic nature of boiling process.

The object of present study is to develop experimental methods for the diagnosis of liquid film dynamics, and to obtain data for the film instability under various conditions. A dedicated test facility was designed and constructed where micro conductive probes and confocal optical sensors were used to measure the thickness and dynamic characteristics of a thin liquid film on various heater surfaces, while a high speed camera was used to get visual observation. Extensive tests were performed to calibrate and verify the two thickness measuring systems. The micro conductive measuring system was proven to have a high reliability and repeatability with maximum system error less than 5µm, while the optical measuring system is capable of recording the film dynamics with spatial resolution of less than 1 m. The simultaneous measurement on the same liquid film shows that the two techniques are in a good agreement with respect to accuracy, but the optical sensors have a much higher acquisition rate up to 30 kHz, which are more suitable for rapid process.

The confocal optical sensors were therefore employed to measure the dynamic thickness of liquid films (ethanol, hexane and water) evaporating on various horizontal heater surfaces (aluminum, copper, silicon, stainless steel and titanium) to investigate the influences of heat flux, the surface and liquid properties on the film instability and the critical thickness. The critical thickness of water film evaporating on various surfaces was measured in the range of 60-150 m, increasing with the increased contact angle or increased heat flux (evaporating rate) and also with the decreased thermal conductivity of the heater material. The data suggest the conjugate heat transfer nature of the evaporating

(4)

II liquid film dynamics at higher heat fluxes of interest to boiling and burnout. In the case of hexane on the aged titanium surface with contact angle of ~3^o, the liquid film is found resilient to rupture, with film oscillations at relatively large amplitude ensuing as the averaged film thickness decreases below 15 µm.

To interpret our experimental findings on liquid film evolution and its critical thickness at rupture, a theoretical analysis is also performed to analyze the dynamics of liquid films evaporating on heater surfaces. While the influences of liquid properties, heat flux, and thermal conductivity of heater surface are captured by the simulation of the lubrication theory, influence of the wettability is considered via a minimum free energy criterion.

The thinning processes of the liquid films are generally captured by the simulation of the lubrication theory. For the case with ideally uniform heat flux over the heater surface, the instability of the liquid film occurs at the thickness level of tens micro meters, while for the case of non-uniform heating, the critical thicknesses for the film rupture are closer to the experimental data but still underestimated by the lubrication theory simulation. By introducing the minimum free energy criterion to considering the influence of surface wettability, the obtained critical thicknesses have a good agreement with the experimental ones for both titanium and copper surfaces, with a maximum deviation less than ±10%.

The simulations also explain why the critical thickness on a copper surface is thinner than that on a titanium surface. It is because the good thermal conductivity of copper surface leads to uniform temperature distribution on the heat surface, which is responsible for the resilience of the liquid film to rupture.

A silicon wafer with an artificial cavity fabricated by Micro Electronic Mechanical System (MEMS) technology was used as a heater to investigate the dynamics of a single bubble in both a thick and thin liquid layer under low heat flux (<60 kW/m²). The maximum departure diameter of an isolated bubble in a thick liquid film was measured to be 3.2 mm which is well predicted by the Fritz equation. However, in a thin liquid layer with its thickness less than the bubble departure diameter, the bubble was stuck on the heater surface with a dry spot beneath. A threshold thickness of the liquid film which enables the dry spot rewettable was obtained, and its value linearly increases with increasing heat flux.

In addition, another test section was designed to achieve a constant liquid film flow on a titanium nano-heater surface which helps to successfully carry boiling in the liquid film from low heat flux until CHF. Again, the confocal optical sensor was employed to measure the dynamics of the liquid film on the heater surface under varied heat flux

(5)

III conditions. A statistical analysis of the measured thickness signals that emerge in a certain period indicates three distinct liquid film thickness ranges: 0~50 µm as microlayer, 50~500 µm as macrolayer, 500~2500 µm as bulk layer. With increasing heat flux, the bulk layer disappears, and then the macrolayer gradually decreases to ~105 µm, beyond which instability of the liquid film may lose its integrity and CHF occurs. In addition, the high-speed camera was applied to directly visualize and record the bubbles dynamics and liquid film evolution. Dry spots were observed under some bubbles occasionally from 313 kW/m² until CHF with the maximum occupation fraction within 5%. A dry spot was rewetted either by liquid receding after the rupture of a bubble or by the liquid spreading from bubbles’ growth in the vicinity. This implies that the bubbles’ behavior (growth and rupture) and their interactions in particular are of paramount importance to the integrity of liquid film under nucleate boiling regime.

Keywords: liquid film, instability, rupture, liquid film dynamics, critical thickness, boiling, bubble dynamics, dry spot, critical heat flux, confocal optical sensor, micro conductive probe

(6)

IV

Acknowledgements

First of all, I would like to express my sincere gratitude to my supervisor Associate Professor Weimin Ma, for his support, great patience, professional guidance and strong confidence in my doing everything best.

Special appreciation goes to Professor Truc-Nam Dinh who was the head of Nuclear Power Safety Division when I joined KTH, for his vision on boiling research as well as his continuous encourage, insights and comments on my PhD research project.

I want to express my sincere gratitude to my assistant supervisor Dr. Liangxing Li for his valuable help and discussions with regard to my academic development and for his friendship which enriches my student life.

I would like to thank Associate Professor Pavel Kudinov for his assistance in procurement of the confocal optical sensor. I thank the KTH-NPS Lab Group for their technical support in the experimental setup. I thank Sachin Thakre for reading through my first draft manuscript. I thank Dr. Minyue Li and Haopeng Li for helping me to develop the digital image processing program.

My gratitude also goes to Professor Raj Sehgal, Dr. Aram Karbojian, Dr. Tomasz Kozlowski, Dr. Roberta Concilio Hansson, Yaodong Cheng, Storm Lars-Erik and José Galdo, for their valuable assistance and discussions on carrying out my research wok.

I would express my thanks to all my colleagues: Hua, Sean, Thanh, Francesco, Andrei, Joanna,, Viet-An, Walter, Alexander, Ivan, Kaspar and all the others at the Nuclear Power Safety Division for making the working environment so friendly and supportive. I thank all my corridor friends and many Chinese friends in Sweden: it is you who make my leisure time so wonderful.

Last but not least, I would like to express my deepest thank to my dear parents for your endless support, encouragement and love!

This study is initiated by the research Grant VR-2005-5729 from Vetenskapsrådets (Swedish Research Council), and partially supported by MSWI research project at KTH.

(7)

V

List of Publications

Journal articles

● S. J. Gong, W. M. Ma, T. N. Dinh, Diagnostic techniques for the dynamics of a thin liquid film under forced flow and evaporating conditions, Microfluid and Nanofluid, Vol. 9, pp.1077-1089, 2010.

● S. J. Gong, W. M. Ma, T. N. Dinh, An experimental study of rupture dynamics of evaporating liquid film on different heater surfaces, International Journal of Heat and Mass Transfer, Vol.54 (7-8), pp.1538-1547, 2011.

● L. X. Li, S. J. Gong, W. M. Ma, An experimental study on two-phase flow regime and pressure drop in a particulate bed packed with multi-diameter particles, Nuclear Technology, Vol. 177, 2012.

● S. J. Gong, W. M. Ma, T. N. Dinh, Simulation and validation of the dynamics of liquid films evaporating on horizontal heater surfaces, Submitted to Applied Thermal Engineering, 2011.

● S. J. Gong, W. M. Ma, L. X. Li, An experimental study on boiling phenomena in a liquid layer, Submitted to International Journal of Thermal Sciences, 2011.

Peer-reviewed conference papers

● S. J. Gong, W. M. Ma, T. N. Dinh, Measurement of thin liquid film dynamics under forced flow and evaporating condition, Proceeding of ECI International Conference on Boiling Heat Transfer, Florianópolis-SC-Brazil, May 3-7, 2009.

● S. J. Gong, W. M. Ma, T. N. Dinh, An experimental study on boiling phenomenon in a liquid film, 7^th international conference on multiphase flow, ICMF2010, Tampa, FL, May 30 - June 4, 2010.

● S. J. Gong, L. X. Li, W. M. Ma, An experimental study on bubble and film dynamics of boiling in a horizontal liquid layer, ICAPP2011, Nice, France, May 2-5, 2011.

● S. J. Gong, W. M. Ma, T. N. Dinh, Visualization and measurement of bubble and film dynamics in a boiling liquid film, NURETH-14, Toronto, Ontario, Canada, September 25-29, 2011.

(8)

VI

List of Figures

Figure 1.1 Schematic of processes involved in boiling and their relations to burnout [12].

... 2

Figure 1.2 A conceptual picture of the different stages of a bubble life cycle under high heat flux [14]. ... 3

Figure 1.3 Microlayer and macrolayer evaporation model of nucleate boiling [3]. ... 8

Figure 2.1 Schematic of the test facility. ... 10

Figure 2.2 The electric circuit of the micro conductive probe. ... 11

Figure 2.3 The tip of a micro conductive probe. ... 12

Figure 2.4 The position of the tip at the moment when the probe is (a) approaching or (b) leaving the liquid film. The upper white region is the probe’s tip; the lower is the mirror image in the film. ... 13

Figure 2.5 Raw voltage signal of a micro conductive probe at wavy free surface of water with measuring frequency of 10 kHz. ... 13

Figure 2.6 Principle of confocal optical sensor. ... 15

Figure 2.7 CCD view of confocal optical sensor for still liquid film. ... 15

Figure 2.8 Block diagram of the confocal optical sensor system. ... 15

Figure 2.9 Measurement deviation of the confocal optical sensor under different measuring rates for a piece of glass. ... 16

Figure 2.10 Refractive index of water: (a) Refractive index of water at 100˚C; (b) Refractive index error of water with 540 nm wavelength at 100˚C; (c) Refractive index of water for 540 nm light wavelength; (d) Refractive index error of water with 100˚C for 540 nm light wavelength. ... 18

Figure 2.11 Tilting angle between the sensor and the object. ... 19

Figure 2.12 Measurement deviation of the optical sensor with titling angles. ... 19

Figure 2.13 Film thickness without air blow: = 0.73 μm (1 kHz sampling rate). ... 20

(12)

X

Figure 2.14 Film thickness with air blow: = 2.07 μm (1 kHz sampling rate). ... 21

Figure 2.15 Calibration setup. ... 22

Figure 2.16 Adiabatic liquid film thickness. ... 22

Figure 2.17 Evaporating liquid film thickness. ... 22

Figure 2.18 Calibration for micro-pump. ... 23

Figure 2.19 Contact angle measurement: (a) picture of droplet; (b) binary image of droplet. ... 24

Figure 2.20 Test section of evaporating liquid film on a solid surface. ... 25

Figure 2.21 Schematic of the test section. ... 28

Figure 2.22 Silicon wafer (30 mm  20 mm  0.4 mm) with titanium film heater (150 nm of thickness) and artificial cylindrical cavity (100 µm of diameter and 110 µm of depth). ... 28

Figure 2.23 Silicon cavity measured by confocal laser scanning microscope. ... 29

Figure 2.24 Micro conductive probe above the artificial cavity visualized by the stereo microscope. ... 30

Figure 2.25 Schematic of the high-speed visualization. ... 30

Figure 2.26 Schematic of the test section. ... 33

Figure 2.27 Test section: a) assembly of test section and arrangement of instrumentation; b) thickness measurement positions... 34

Figure 2.28 SEM image of titanium film. ... 34

Figure 3.1 Thinning process of a water film evaporating on a Ti_400 surface. Measuring point is at the center of the heater surface, and the critical thickness is 95.4 μm. The wall temperature is ~105 ^oC, and heat flux is 32.4 kW/m². ... 37

Figure 3.2 Hole formation and expansion in an evaporating liquid film: The visual field is 12.5 mm  12.5 mm with spatial resolution of 25 µm, and time interval between frames is 10 ms. ... 37

Figure 3.3 Measuring points (Aat the center; B, C, D, E1 mm, 3 mm, 5 mm and 6.5 mm away from the center). ... 39

(13)

XI Figure 3.4 Thinning curves of a water film at different locations on the Ti_400 surface. 39

Figure 3.5 Lateral profiles of a water film evaporating on a Titanium_400 surface. ... 40

Figure 3.6 Thinning process of water film evaporating on a Titanium_400 surface. Measuring point is 1 mm away from the center, and the critical thickness is 96.1 μm. ... 40

Figure 3.7 Water film evaporation on silicon wafer under averaged heat flux of 56.66 kW/m²± 6% measured by IFS2431-3: (a) Thickness evolution of evaporating liquid film; (b) Critical liquid layer thickness. ... 41

Figure 3.8 Critical thicknesses of water film on titanium surfaces at varied evaporation rate... 43

Figure 3.9 Critical thickness of water film on various metal surfaces (q=32.4 kW/m²). . 44

Figure 3.10 Critical film thicknesses on aged copper and titanium surfaces. ... 45

Figure 3.11 Critical thickness of ethanol film on Ti_200 surface. ... 47

Figure 3.12 Thinning process of a hexane film on heated Ti_200 surface. ... 48

Figure 3.13 Thinning process of a hexane film on adiabatic Ti_200 surface. ... 48

Figure 3.14 Digital image processing for analysis of bubble dynamics. ... 50

Figure 3.15 Binary images of bubbles in a life circle. The spatial resolution is 18 µm and the time interval between frames is 1 ms. ... 50

Figure 3.16 Bubble parameters in a water layer with thickness of ~7.5 mm, under different heat fluxes. ... 52

Figure 3.17 Ratio of latent heat to total heat. ... 53

Figure 3.18 Bubble dynamics in a life cycle under different heat flux. Liquid layer thickness=~7.5 mm. A1=17 kW/m², A2=26 kW/m², A3= 35 kW/m², A4=43 kW/m², A5=52 kW/m², A6=60 kW/m². ... 54

Figure 3.19 Local void measurement by the micro conductive probe (1 kHz). ... 55

Figure 3.20 The images of the bubbles appearing in thin liquid layer of 2 mm ± 0.3 mm under different heat flux of B2-1: 26 kW/m², B3-1: 35 kW/m², B4-1: 43 kW/m², B5-1: 52 kW/m²; the window size is 12 mm  12 mm. ... 57

(14)

XII Figure 3.21 The dynamics of reversible dry spot and the liquid dam after the bubble rupture. Case B3-1: q=35 kW/m², δl=1.964 mm. a) the window size is 11.2 mm  11.2 mm, the resolution is 35 µm; b) rupture initiation of the bubble at 0.2 ms. ... 58 Figure 3.22 Diameter of a reversible dry spot after the bubble rupture (B3-1: q=35 kW/m², δ_l=1.964 mm). ... 59 Figure 3.23 The dynamics of irreversible dry spot and the liquid dam after the bubble rupture. Case B3-2: q=35 kW/m², δ_l=1.519 mm; the window size is 11.2 mm  11.2 mm;

the resolution is 35 µm. ... 59 Figure 3.24 Minimum water layer thickness for keeping the dry spot reversible and corresponding maximum bubble diameter and dry spot diameter. ... 60 Figure 3.25 Boiling under a large bubble in the test B8 (q=78 kW/m², δl=3.545 mm), the window size is 18 mm  18 mm and the resolution is 35 µm. ... 60 Figure 3.26 Averaged temperature values vs. heat flux values for A_tests and B_tests. . 61 Figure 3.27 Averaged liquid film thickness profile along the OO’ direction vs. water flow rate: 1 – 71.0 mm³/s; 2 – 99.4 mm³/s; 3 – 113.6 mm³/s; 4 – 127.8 mm³/s; 5 – 156.2 mm³/s;

6 – 184.6 mm³/s. ... 63 Figure 3.28 Liquid film thickness at liquid flow-rate of 99.4 mm³/s (sampling rate = 1 kHz): a) averaged liquid film thickness profiles in the directions of AA’, OO’ and BB’; b) instantaneous liquid film thickness at 4 points in the OO’ direction. ... 64 Figure 3.29 Liquid film thickness at liquid flow-rate of 127.8 mm³/s (sampling rate = 1 kHz): averaged liquid film thickness profiles in the directions of AA’, OO’ and BB’. ... 65 Figure 3.30 Instantaneous liquid film thickness of test N5 at various heat fluxes at liquid flow-rate of 127.8 mm³/s (sampling rate: 2.5 kHz). ... 67 Figure 3.31 Occurrence of the liquid film thickness during 1 minute. ... 68 Figure 3.32 The liquid thickness values of peaks vs. heat flux. ... 69 Figure 3.33 Bubble dynamics of test N1 under heat flux of 0.313 MW/m² and liquid flow-rate of 99.4 mm³/s: (a) dry spot observed under the bubble; (b) wetting under the bubble. (Sampling frequency: 10 kHz; view window: 3.2 mm × 3.2 mm.) ... 72 Figure 3.34 Bubble dynamics of test N4 under liquid flow-rate of 127.8 mm³/s (Sampling frequency: 10 kHz; view window: 8.0 mm × 8.0 mm). ... 74 Figure 3.35 Bubble dynamics of test N5 under liquid flow-rate of 127.8 mm³/s (Sampling frequency: 10 kHz; view window: 6.0 mm × 6.0 mm). ... 75

(15)

XIII

Figure 3.36 Bubble frequency versus heat flux. ... 75

Figure 3.37 An example of the wall temperature development for test N6 under various heat fluxes during 30 s period (sampling frequency: 1 kHz). ... 76

Figure 3.38 Analyzed characteristic wall temperature values versus heat flux. ... 77

Figure 3.39 Wall temperature development of CHF occurrence for test N6. ... 77

Figure 3.40 Bubble dynamics of N5 under heat flux of 1.25 MW/m² and liquid flow-rate of 127.8 mm³/s (view window: 6.0 mm × 6.0 mm). ... 79

Figure 4.1 The sketch of an evaporating liquid film on a heater surface. ... 80

Figure 4.2 Schematic of hole formation in a liquid film (0  /2). ... 83

Figure 4.3 Test section of liquid film evaporating on a locally heated solid surface. ... 83

Figure 4.4 Schematic of the two-dimensional evaporating liquid film for simulation. .... 84

Figure 4.5 Liquid film evolution on a uniform-temperature heater surface, ΔT=5˚C. (a) h0=300 m; (b) h0=30 m; (c) h0=3 m. ... 86

Figure 4.6 Absence of forces vs. the evolution of a water film at ΔT=5˚C under uniform heating. (a): h₀=300 m; (b): h₀=30 m; (c): h₀=3 m (G – gravity; S – surface tension; TC–thermo-capillary; E – evaporation number). ... 87

Figure 4.7 Temperature profile along the different heater surfaces (q=32.4 kW/m²). ... 89

Figure 4.8 Liquid film evaporating on a titanium heater surface with non-uniform temperature profile (q=32.4 kW/m²): (a) Liquid film thickness profile at rupture for cases with varied initial thicknesses; (b) Rupture time vs. initial film thickness. ... 90

Figure 4.9 Liquid film profiles on different heater surfaces with non-uniform temperature profiles (q=32.4 kW/m²). (a) Liquid thickness profiles at rupture; (b) Calculated critical thickness at liquid film rupture. ... 91

Figure 4.10 Impact of heat flux on the critical thickness of liquid film on non-uniform- temperature heater surfaces (Ti_400 refers to the Ti surface aged at 400^oC). ... 92

Figure 4.11 The measured critical thickness vs. the predicted by Equation (4.11) for different Ti surfaces. ... 93

Figure 4.12 Critical thickness vs. hear flux for copper surface. ... 94

(16)

XIV

List of Tables

Table 1.1 Comparison of optical techniques for the liquid film thickness measurement ... 5

Table 2.1 The main parameters of optical sensor ... 16

Table 2.2 Coefficients of Equation 2.3 ... 17

Table 2.3 Test Matrix ... 26

Table 2.4 Tests in a thick water layer. ... 31

Table 2.5 Tests in a thin water layer. ... 32

Table 3.1 Contact angle of water on titanium surfaces ... 42

Table 3.2 Contact angle of water on surfaces made of different materials ... 44

Table 3.3 Thermal conductivity of materials (25℃) ... 46

Table 3.4 Properties of fluids (20℃ and 0.1MPa) ... 46

Table 3.5 Test results in a thin water layer ... 57

Table 3.6 Boiling tests in liquid films ... 62

Table 4.1 Thermal conductivity of heater surface materials. ... 89

Table 4.2 Contact angle of water on titanium surfaces ... 93

(17)

XV

(18)

XVI

(19)

1

1 Introduction

1.1 Motivation

A thin liquid film spreading over and evaporating on a solid surface is encountered in many engineering applications that involve processes such as spray cooling (condensation), heating (e.g., boiling), coating, cleaning and lubrication. An emerging application of thin liquid film is the cooling of high-power electronics (chips) where the space is constrained. In most cases, the stability and integrity of the liquid film are desired to avoid deteriorating the performance of the processes. If the liquid film ruptures in annular flow boiling, for instance, burnout (dryout) would occur as a threat to equipment safety. Even for pool boiling, it is believed that the behavior of the near-wall liquid layer plays a key role in heat transfer and boiling crisis (DNB). Noteworthy, a good number of analytical models and numerical simulations [1][2][3][4][5] for boiling heat transfer were based on the concepts of near-wall microlayer and macrolayer whose thicknesses were estimated to range from several µm to hundreds µm [6][7][8][9][10][11]. However, an experimental quantification of the behavior of such thin liquid films is not straightforward, since the measurement at micro scale is a challenge, and further complicated by the traditional experimental setups (e.g. pool boiling with heater block) and the chaotic nature of boiling process which impede direct observation and measurement of the thin liquid films, especially under high heat flux conditions.

Advances in micro (nano) fabrication and measurement techniques offer new capabilities to address the challenge. For instance, the dynamic thermal patterns on heater surfaces in pool boiling were investigated through using novel nano-heaters and high-speed high- resolution infrared thermometry [12][13]. The thermal patterns provided the near-wall temperature evolution over a broad range of heat fluxes starting from the onset of nucleation and up to boiling crisis, while an observation of the detailed void pattern was made possible by X-ray radiography. An analysis of the results suggested a “scales- separation” phenomenon (cf. Figure 1.1) which says that high heat-flux boiling and boiling crisis is dominated by micro-hydrodynamics of the near-wall liquid film. More specifically, for a given surface condition and coolant chemistry, boiling crisis can be (and has been) treated as a hydrodynamic phenomenon, and there exist two

(20)

2 hydrodynamic scales: an external one and an internal one (micro-hydrodynamics of liquid microlayer on the heater surface). The external-scale hydrodynamics is irrelevant to boiling crisis in pool boiling. This provides the rationale to perform the BETA-B boiling experiment [14] on a thin liquid film so that the micro-hydrodynamics of the film was visualized directly by a high-speed video camera synchronized with the IR imaging, without losing the key physics of boiling. Based on the observations, the evaporating liquid film dynamics was conceived as shown in Figure 1.2.

Figure 1.1 Schematic of processes involved in boiling and their relations to burnout [12].

Still, it is not a trivial task to qualify the proposed physical picture. The main difficulties for such measurement are the film’s small scale and rapid evolution, as well as randomness of nucleation and bubble growth. We used a “divide-and-conquer” strategy here to overcome these difficulties, and the task will be accomplished through three steps:

i) design and development of an experimental well-controlled platform with diagnostic techniques for detection of liquid film thickness and its variations; ii) application of the developed techniques to measure the dynamics of evaporating liquid films under various conditions; and iii) extension of the techniques to measure the dynamics of a boiling liquid film with bubbles generated from prescribed nucleation sites (over artificial cavities created by MEMS) under low heat flux and from heated thin titanium film under high heat flux, with the bubble dynamics being visualized by high-speed digital cinematography.

(21)

3 Figure 1.2 A conceptual picture of the different stages of a bubble life cycle under high

heat flux [14].

1.2 Diagnostic techniques for liquid film dynamics

A good number of diagnostic techniques have been developed to measure the liquid film thickness, which can be divided to four groups: acoustic methods, nucleonic techniques, electrical methods and optical methods [15]. Acoustic methods are based on the principle that the ultrasonic waves are attenuated and reflected when encountering a gas-liquid interface. The acoustic methods have a high sampling frequency up to 20 MHz, but were reported to be unsuitable to wavy films [16]. Moreover, the ultrasonic wavelength (much larger than the light wavelength) also limits the spatial resolution of the methods and its application to ultra thin films. Nucleonic techniques are based on radiation attenuation (neutron, gamma- and X-ray) during its passage through two-phase system, since the attenuation in liquid is higher than in gas, leading to a distinct loss in radiation intensity when a beam penetrates the two-phase mixture. However, the operational complexity and high cost prevent the method from extensive application. The method is also not suitable for measurement of local liquid film thicknesses.

The electrical methods include conductance-based and capacitance-based methods. The principle of capacitance-based methods is that when metal plates connected an electric current are faced against each other, a capacitance is created whose value will be largely determined by the medium between the plates. The capacitance-based methods are non- intrusive, but only can be used for measurement of averaged values for cross section with the measuring range of 400 m - 23 mm [15]. Among conductance-based methods, the conductive probes are broadly applied in diagnosis of gas-liquid two-phase flow, in term of detection of bubble frequency, local void fraction and film thickness [7][17]. Since it belongs to an intrusive measurement method, the resolution of a conductive probe

(22)

4 depends on both the dimension of its tip and the quality of its positioning system. There were various designs of the conductive probes whose measuring range was of 50 m - 1.2 mm [15]. The micro conductive probes used in the present study are characterized by a miniature tip with the diameter of 2 ~ 3 micrometers (to reduce intrusiveness) and controlled by a manipulator with fine movement resolution (1 m), long-term stability and repeatability. However, the probe is not suitable for a very dynamic film, since the manipulator can not follow rapid thickness variation.

Optical methods are based on the principles of optical phenomena (attenuation, refraction, reflection, scattering, interference, etc.) when light waves pass through different media (solids and fluids) and their interfaces. Generally, the optical methods are the most promising diagnostic techniques for a dynamic liquid film, since most of them are non- intrusive to the fluids and fast-responding with a potential of high spatial resolution.

Table 1.1 lists typical optical methods which found their applications in liquid film measurement. The techniques of interface detection [18] and light attenuation [15] may be simple to use, but their measurement errors are high. The scanning ellipsometry [19]

has a very high resolution and accuracy, but its measuring range is too small for our applications. The interferometry [20] is another precise technique for thickness measurement, but it needs a complex optical and processing system, with measuring range less than 1 mm. The methods with beam laser shadow [21] and fluorescence imaging [22] had limited applications in liquid film measurement, due to measuring range and addition of dye, respectively. The technique of focus displacement of laser [23]

or chromatic lights [24][25] is popular with its good accuracy and operability, especially with the advent of confocal optical sensors. For a rapidly varied liquid film which the conductive probe cannot follow, the confocal optical sensors are an ideal option because the sensors are equipped with data acquisition rate up to 30 kHz, and able to measure thickness ranging from several m to 3 mm with nominal spatial resolution up to less than 1 m.

The confocal optical sensor was therefore chosen as the primary technique in the present study. A micro conductive probe system was also developed for calibration of the optical sensor.

(23)

5 Table 1.1 Comparison of optical techniques for the liquid film thickness measurement

Technique Measuring

range Accuracy Operability Measurement principle Interface detection [18] 20 µm - 1.3 mm low good Light intensity

gradient at interface Light attenuation [15] 1.5 mm - 3 mm low good Fluid absorptivity Scanning ellipsometry [19] 10 nm ~ 1 µm high hard Dynamic imaging

ellipsometry Interferometry [20] 10 µm – 1 mm high hard Interference of light

waves Beam laser shadow [21] 0.4 mm - 0.9 mm Medium - Refraction and

reflection

Fluorescence imaging [22] 5 µm -1.5 mm Medium Need dye Induced fluorescence Focus displacement

[23][24][25] 2 µm ~ 2.8 mm high good Position displacement

1.3 Previous studies on liquid film dynamics

Over the past three decades, a large number of experimental and theoretical studies have been carried out to investigate the liquid film dynamics and instability. Among them, the minimum thickness of liquid film flowing down a vertical or inclined adiabatic solid surface has been predicted theoretically according to force balance or minimum total energy criteria [26][27][28][29][30][31], in a reasonable agreement with experimental results. For a horizontal liquid film, Sharma et al. [32][33] developed a theory to predict the critical thickness at which liquid film rupture occurs by hole formation when the free energy of the film-solid system becomes equal to the free energy of the hole-liquid-solid system. The model predicted film rupture thickness is several hundred micrometers which are in good agreement with the experimental data on selected non-wetting solid surfaces (e.g., Teflon, polyethylene, PMMA, wax). In such thickness range the disjoining pressure cannot account for the film rupture since Van der Waals intermolecular forces are negligible for films thicker than 0.1 m [33]. The predictions also show that critical thickness of film rupture is strongly affected by wettability which can be represented by contact angle; smaller the contact angle, thinner is the critical thickness. Orell et al. [34]

performed experiment to investigate the formation of a dry spot in a non-boiling thin ethanol film creeping on a horizontal surface, and found the downstream film thickness around dry spot is 450 - 660 m in the heat flux range of 3 - 15 kW/m². Benard-type convection cell pattern appears prior to the dry spots. They also concluded that the threshold heat flux for formation of a dry spot at wall temperatures exceeding the saturation temperature was substantially lower than that reported for incipience of pool

(24)

6 boiling of ethanol. For a better wetting solid surface (the equilibrium contact angle is less than a small value), the film may be thinning to such a level that disjoining pressure can rupture the film spontaneously. The film rupture is therefore determined by film-thinning disturbances such as drainage gravity driven and spreading driven by surface tension gradients [35]. Narsimhan et al. [36][37] numerically modeled the rupture of thin stagnant films on a solid surface due to random thermal and mechanical perturbations, towards an understanding of the relations among rupture time, film thickness and perturbation amplitude. The film thickness in their study is less than 1 m which is only applicable to well wetting surfaces.

Oron et al. [38] gave a comprehensive review on the modeling of thin film dynamics, and a unified mathematical system was presented to predict the long-scale evolution of thin liquid films as a result of a long wave theory. The set of mathematical evolution equations is mainly based on the work of Burelbach et al. [39], taking into account the influential factors such as Van der Waal forces, surface tension, gravity, thermo-capillary, mass loss and vapor recoil force. In their review, Oron et al. [38] finally pointed out the importance of verifying the long-wave theory against experiments, and gave a few samples to support the theory of liquid film evolution. Among them, Burelbach et al [40]

investigated the steady thermo-capillary flow of the non-volatile silicone oil with the thickness from 0.125 mm to 1.684 mm on a non-uniformly heated horizontal solid plate, and VanHook et al. [41] performed experiment on the onset of the long-wave instability in a thin layer of silicone oil of thickness ranging between 50 m and 250 m. Elbaum &

Lipson [42] studied the thinning by evaporation of completely wetting water films on clean mica surfaces. It was observed that the thinning was unstable to nucleation of dry patches in the film thickness range from 10 nm to 100 nm, and the spreading of dry patches invoked a dewetting of the substrate. The initial process was described by a simple nucleation theory. Craster & Matar [43] presented a comprehensive review of the work carried out on thin films flows, focusing attention on the studies undertaken after the review by Oron et al. [38]. They pointed out that for the modeling of thin film dynamics, lubrication theory was still used to elucidate a wide variety of flows in which films have small aspect ratios.

In general, the literature survey shows that a good number of analytical models had been developed to simulate dynamics and rupture of liquid films on a wetting solid surface (apparent contact angle near zero) with the film thickness less than 1 m for which long- range intermolecular interactions are pronounced. For non-wetting surfaces (apparent contact angle is much greater than zero), the film instability can be initiated at a

(25)

7 significantly larger film thickness (say, hundreds of micrometers as shown in the above- mentioned works) by external disturbances, and predicted by total free energy criteria. In both cases, high-fidelity experimental data are necessary for understanding of the film rupture phenomenon and validation of the models. As pointed out by Oron et al. [38], there is a clear need for careful experimental investigations to verify phenomena and to give data that can be used to test the theories, and they claimed their review paper stands as a call for such experiments. Remarkably, there is few data for film rupture on a horizontal surface under diabatic (heating) conditions which is important to boiling heat transfer and boiling crisis.

1.4 Previous studies on microlayer and macrolayer for boiling heat transfer

The conception of microlayer and macrolayer are present in a great number of experimental and theoretical investigations on boiling heat transfer since understanding micro hydrodynamics of microlayer and macrolayer are paramount importance to discover physical mechanisms of nucleate boiling and boiling crisis. Generally, as shown in Figure 1.3 [3], the microlayer is described as a thin liquid layer beneath the bubble with the thickness of some micrometers magnitude, while the macrolayer is reported as a superheated liquid layer adjacent heater surface with the thickness from tens micrometers to hundreds micrometers. Stephan and Kern [3] applied the microlayer and macrolayer model schematically to analyze the heat and mass transfer phenomena in nucleate boiling of both pure substance and binary mixtures, which can couple both microscale phenomena (intermolecular forces of adsorption, capillary forces, molecular interfacial phase change resistance, and change of phase equilibrium) and macroscale phenomena (the influence of free and forced convection, transient heat conduction, and latent heat depends on the geometry of the evaporator and the boiling conditions). However, the detailed and quantified expressions of microlayer and macrolayer are still divergent.

A lot of experimental methods were also developed to understand the macrolayer formation and quantify its value. Gaermer [44], Nishio [45] and Bang et. al. [10]

observed the macrolayer in the boiling pool and estimated its thickness by direct visualization. In addition, the conductance-based probe applied by Bhat et. al. [46], Rajvanshi et. al. [6] and Sakashita et. al. [47][48][49] was another main choice to detect the macrolayer and quantify its thickness. Furthermore, the capacitance-based sensor [50]

and temperature pattern based method [51] were developed to deduce the macrolayer

(26)

8 thickness. However, by reviewing some possible mechanisms of macrolayer formation (e.g., Helmholtz instability on vapor stem walls, lateral coalescence of vapor stems and lateral coalescence of bubbles), Sadasivan et. al. [52] concluded that the final resolution of the mechanism of macrolayer formation needs detailed experimental measurements of vapor and liquid flow patterns close to the heater surface.

At the same time, some special experimental methods were developed to investigate on characteristics of the microlayer. Cooper and Lloyd [53] reported a correlation of microlayer thickness through theoretically analyzing experimental synchronized cine photographs and dynamic temperature pattern of heater surface. While Jawurek [54], and Kim & Buongiorno [55] used optical method of interferogram which was recorded by high speed camera and infrared camera respectively to measure the thickness and evolution of the microlayer.

Figure 1.3 Microlayer and macrolayer evaporation model of nucleate boiling [3].

1.5 Research scope

Toward the research goal as mentioned in Section 1.1, the scope of the present study includes i) design and setup of an experimental platform suitable for micro-scale operation; ii) selection, adaption and calibration of a diagnostic technique for the quantification of thin film dynamics; and iii) application of the verified experimental method (the facility and instruments) to various liquid films on different heater surfaces with heat fluxes until occurrence of CHF.

(27)

9 Although experimental characterization of film dynamics was proven to be a formidable task due to the difficulties in traditional techniques to measure the micro-scale film thickness, recent advances in micro (nano) fabrication and diagnostic techniques provide new capabilities to address the challenge. Based on the literature review and survey, an innovative confocal optical sensor system was chosen to characterize the dynamics of a liquid film in question. A micro conductive probe system was also developed.

Verification and calibration of the confocal optical sensor was accomplished by comparative measurements between the sensor and the micro conductive probe. Various tests and measures were carried out to identify and minimize the secondary effect of experimental conditions on measurement accuracy. The confocal optical sensor was then applied to measure liquid films dynamics under various heating conditions.

In the beginning, tests focused on the dynamics and critical thickness of liquid films (water, ethanol, and hexane) evaporating on various heater surfaces (silicon, copper, aluminum, stainless steel and titanium) under heat fluxes until boiling initiation that enable evaluation of factors and properties which govern film dynamics, stability and rupture. According to obtained experimental data, a theoretical method based on the long wave theory and minimum free energy theory was developed to predict the liquid film evolution and its critical thickness. Then, tests investigated the dynamics of a single bubble grows from an artificial cavity both in a thick liquid layer and in a thin liquid layer. Finally, boiling was performed on a coated titanium film heater in a thin liquid layer under wide heat flux range from onset of nucleate boiling till burnout.

(28)

10

2 Experimental method

2.1 Test facility

An experimental platform was designed and constructed to investigate liquid film dynamics and boiling physics at a microscopic level. Figure 2.1 shows the dedicated test facility which consists of an optical table, liquid (water) temperature control and supply systems, power supply and heating systems, stereo microscopic visual systems, a confocal optical sensor system, a micro conductive probe and its control system (including 3D-manipulator), lighting systems, and a test section for liquid film formation on heater surfaces.

The optical table provides the required vibration isolation, and also acts as an ideal operation platform for mounting instrumentation and test sections. Water is preheated in a stainless steel tank by a band heater to a desired temperature and maintained with a temperature controller; the hot water is then supplied to the test section through a micro pump capable of accurate flow rate control. A DC power supplier is used for continuous and uniformed heating. The stereo microscope with maximal magnification of 300 is coupled with a video camera used to well observation. Micro conductive probe and confocal optical sensor are held and controlled by a 3D manipulator for accurate liquid film thickness measurement.

Figure 2.1 Schematic of the test facility.

(29)

11

2.2 Measurement techniques

2.2.1 Micro conductive probe measuring system

The micro conductive probe measuring system includes a conductive probe which is positioned by a micro manipulator (MP-225) and is connected to an electronic circuit (Figure 2.2) whose output voltage is recorded by a computer through a data acquisition card (NI PCI-6251). The electronic circuit has been designed to minimize the noise and magnified the sensitivity of probe through choosing the right electronic components. The probe is made in such a way that only its tiny tip (Figure 2.3) with the diameter of 2 - 3 μm is electrically exposed to fluids. Thus, the output of the circuit differs significantly when the tip shifts from one phase to another (water to air/vapor, or vice versa) due to the distinct electric conductivities of the two phases. This is the working principle for the probe to identify phases and their interface indicated by a positioning mechanism of the manipulator. Besides the size of the probe, the measuring accuracy also depends on the data acquisition frequency (for dynamic signals) and the resolution of manipulator used to control the probe movement. The manipulator driven by a stepping motor with 62.5 nm per step has a fine resolution of 1 μm and a good long-term stability. The NI PCI- 6251 data acquisition card has a maximal data acquisition rate of up to 1.25 MHz.

Figure 2.2 The electric circuit of the micro conductive probe.

Ground DAQ

Probe

R0

R2 R1 DC Power

(30)

12 Figure 2.3 The tip of a micro conductive probe.

By two methods of the probe’s movement (approaching or leaving the film), the film surfaces can be recognized, and the thickness can be obtained. It is found that when approaching the film the probe responds immediately to the contact of the tip with the water surface, i.e., no significant deformation is observed by the microscope; Figure 2.4a.

When leaving the film, however, the probe will not be detached from the flat liquid film until its tip is at some distance away from the film surface. This phenomenon is mainly due to the liquid adhesion to the probe causing the liquid to move upwards with the probe and resulting in a deformation of the surface at the probe’s tip; see Figure 2.4b for microscopic observation of the process. The maximum deformation is less than 30 m at room temperature, and it reduced by increasing the fluid temperature and momentum.

Figure 2.5 shows the raw signal of the micro conductive probe when it is used to measure a wavy water film whose fluctuation amplitude is around  30 m according to the confocal optical sensor’s measurement. One can see that the micro conductive probe has short response time less than 0.1 ms. Considering the resolution and repeatability of manipulator, the precision of conductive probe for measuring the stagnant liquid layer thickness is better than  5 m.

(31)

13

(a) (b)

Figure 2.4 The position of the tip at the moment when the probe is (a) approaching or (b) leaving the liquid film. The upper white region is the probe’s tip; the lower is the mirror image in the film.

Figure 2.5 Raw voltage signal of a micro conductive probe at wavy free surface of water with measuring frequency of 10 kHz.

2.2.2 Confocal optical sensor measuring system

Figure 2.6 shows the measuring principle of the confocal optical sensor. A beam of polychromatic (white) light is dispersed to a series of monochromatic light (denoted by wavelengths from ₀ to _n) through an optical system of multiple lenses. Consequently, the white light source is imaged by the objective lenses on continuous points along the optical axis in the measurement space. When a measured object is placed in the

(32)

14 measurement space, a single of the monochromatic point images (with wavelength of i) is focalized on the object interface. Due to the confocal configuration, only the light of wavelength _i is reflected through the objective lenses and directed towards the spectrometer with high efficiency, all other wavelengths will be out of focus. The spatial peak on the spectrometer indicates the position at which the measured object surface intercepts with the optical axis. When a transparent object is placed in the measuring space, the reflections from the upper and the lower surfaces of the object will be detected by the spectrometer as two peak signals and the thickness of the object is therefore deduced. Figure 2.7 plotted an example of CCD signal for the thickness measurement.

For a liquid layer on a solid surface (e.g., heater), the variation of its thickness between time t1 and t2 can be expressed as

l t t

l n

D D

/ 1 1

1 2



 

 (2.1) where nl is the refractive index of the liquid that has to be input as known parameter in the measurement, and D_t₁and D_t₂ are the positions of the solid surface measured by the confocal optical sensor at time t1 and t2, respectively. This approach can continuously record film evaporation till dryout occurring.

The confocal optical sensors employed in the present study were made by Micro-Epsilon Company (see www.micro-epsilon.com) in Germany. As illustrated in Figure 2.8, the sensor is incorporated with a controller which is also connected to a special Xenon light source. A single controller can support a number of sensors with different measuring ranges and accuracies. The controller optoNCDT2431 is chosen here, which is communicated with the computer through a software package provided by the company or a program further developed by the user depending on a specific application. So far two sensors IFS2431-3 and IFS2431-0.3 have been used for different measuring ranges, as shown in Table 2.1. In principle, the sensors can be applied to a surface with curvature, and have the potential to measure a multilayer transparent system, with further software development for post-processing of the raw signal by users.

Nevertheless, the diagnostic technique was originally developed by the Micro-Epsilon Company to detect grooves, measure glass thickness and trace liquid level with extreme precision. It had little experience to measure the thickness and evolution of a dynamic liquid layer, such as a flowing and/or diabatic (e.g., evaporating) liquid film. To verify and qualify its performance (measurement accuracy and reliability) under such thermal- hydraulic conditions, comprehensive tests under potential influences were performed and some results are presented below.

(33)

15 Figure 2.6 Principle of confocal optical sensor.

Figure 2.7 CCD view of confocal optical sensor for still liquid film.

Figure 2.8 Block diagram of the confocal optical sensor system.

(34)

16 Table 2.1 The main parameters of optical sensor

Sensor model Meas. range Start of meas. range Spot diameter Resolution Max. tilt IFS2431-3 3 mm 16.3 mm 25 µm 0.12 µm ± 22˚

IFS2431-0.3 300 µm 10.5 mm 10 µm 0.012 µm ± 28˚

2.2.2.1 Influence of measuring rate

Since our final goal is to perform measurement on boiling at high heat flux where the bubble life time is around 1 ms, the measuring rate requires the confocal optical system be the technique of choice, whose data sampling rate up to 30 kHz is sufficient for the liquid films under investigation.

Because the measurement accuracy is generally affected by the recording rate due to different time scales for sampling, trials were performed on a thin piece of glass with different sampling rates. It was found the root mean square deviation for every 10000 data sample calculated by Equation 2.2 is no more than 0.5 m if the rate is less than 10 kHz, as seen in Figure 2.9.

(2.2)

Figure 2.9 Measurement deviation of the confocal optical sensor under different measuring rates for a piece of glass.





 



 ⁿ

i

i x

n 1 x

)2

1 ( 1

0 0.5 1 1.5 2

0 5 10 15 20 25 30 35

Frequency (kHz)

Deviation (μm)

(35)

17 2.2.2.2 Effect of liquid properties

The raw signal of the optical system is affected by the measured material through its refractive index which is a function of light wavelength and material temperature. In the present application with water, the refractive index can be obtained by Schiebener’s [56]

correlations as Equation (2.3) and relative coefficients are listed in Table 2.2.

2

* 2 7

* 2

* 6 2

* 2

* 5 2

*

* 4

* 3

* 2

* 1

* 0 2

2 1

2

1 



 

  ^a

a a

T a a T a a n a

n

IR UV

 



 

 (2.3)

where

^*/₀, ₀ 1000kgm^³ ^*/₀, ₀ 0.589m T^*T/T₀, T₀ 273.15K

Table 2.2 Coefficients of Equation 2.3 243905091

.

0 0

a a₅ 2.45733798

3 19.5351809410^

a a₆ 0.897478251

3 2 3.643581110^

a a₇ 1.6306618310^²

4 3 2.6566642610^

a ^*_UV 0.229202

3 4 1.5918932510^

a ^*_IR 5.432937

According to the Equation 2.3, the refractive index of water under 100ºC temperature is calculated as Figure 2.10 related with the wavelength applied for measurement from 450 nm to 700 nm. Equation 2.4 indicates the error rate due to the refractive error influenced by wavelength. Figure 2.10a presents the refractive index of water at 100ºC related with the wavelength, while the error rate is analyzed in Figure 2.10b if the constant refractive index of 540 nm wavelength is considered. It can be seen from Figure 2.10b that the maximum error is very small and no larger than 0.32%. While Figure 2.10c plots the temperature effect on the refractive index and its error rate is illustrated in Figure 2.10d, where the error is no more than 0.2% within 10^oC span. It should be noted that these impacts from material properties can be taken into account in post-processing of the raw signal.

540 540



 





n n Rate n

Error (2.4)

(36)

18 (a) (b)

(c) (d)

Figure 2.10 Refractive index of water: (a) Refractive index of water at 100˚C; (b) Refractive index error of water with 540 nm wavelength at 100˚C; (c) Refractive index of

water for 540 nm light wavelength; (d) Refractive index error of water with 100˚C for 540 nm light wavelength.

2.2.2.3 Effect of the sensor’s orientation

If the optical sensor is not precisely perpendicular to the surface of the measured object (say, with a tilting angle of  as shown in Figure 2.11), the measured thickness may be neither the distance of T through which the light travels in the object, nor the real thickness of To of the object, because of the optical effect at the interface. This is of importance to a wavy liquid film whose surface may not be absolutely flat.

According to the measurement principle of the optical sensor illustrated in Figure 2.6, the surface curvature of measured object will not affect the measurement precision as long as the focus spot is small enough. For the sensors with the spot diameters as shown in Table 2.1, how does the tilting angle influence the measurement accuracy? To answer this question, tests were performed on a thin piece of glass with 995 μm thickness, by regulating the tilting angles. Figure 2.12 shows the deviation of the sensor’s readings

1.31 1.315 1.32 1.325 1.33

400 500 600 700 800

Light wavelength (nm)

Refractive index

-0.004 -0.002 0 0.002 0.004

400 500 600 700 800

Light wavelength (nm)

Error rate

1.305 1.315 1.325 1.335 1.345

0 20 40 60 80 100 120 140

Water temperature (ºC)

Refractive index

-0.012 -0.008 -0.004 0 0.004 0.008

0 30 60 90 120 150

Water Temperature (ºC)

Error rate

(37)

19 when the observation orientation is altered. The maximum tilting angle used in this calibration is ± 15º that is constrained by existing regulating facility, within which the measurement deviation is no larger than 1.5% as shown in Figure 2.12. Moreover, the deviation is not random and can be minimized by the experimental procedure and data processing. This is a great advantage for measurement of a dynamic liquid film which may have curvature. In addition, the measurable angle can be increased further to ± 28º if the sensor IFS2431-0.3 is used.

Figure 2.11 Tilting angle between the sensor and the object.

Figure 2.12 Measurement deviation of the optical sensor with titling angles.

0 0.5 1 1.5 2

0 3 6 9 12 15 18

Tilting angle (º)

Deviation (%)

(38)

20 2.2.2.4 Effect of atmosphere

The sensor fails to work properly if its lens has direct contact with steam, which leads to condensation there. Careful protection of measures (e.g. blowing warm air in front of the lens) is designed and applied in the tests under evaporation/boiling conditions so as to prevent steam from flowing toward the optical sensor but not to disturb the liquid film significantly.

Figure 2.13 shows the natural thickness fluctuation of a water film at room atmosphere, whose amplitude is characterized by a root mean square deviation of 0.73 μm. Figure 2.14 also shows that the blowing air disturbs the liquid film by an increased deviation from 0.73 μm to 2.07 μm, which has to be taken into account into the interpretation of the final measurement results. The major wavy frequency of the film is around 25 Hz.

Figure 2.13 Film thickness without air blow: = 0.73 μm (1 kHz sampling rate).

840 842 844 846 848 850

0 200 400 600 800 1000

Time (ms)

Thickness (μm)

(39)

21 Figure 2.14 Film thickness with air blow: = 2.07 μm (1 kHz sampling rate).

2.2.2.5 Calibration

In order to calibrate and compare the confocal optical sensor and the micro conductive probe for thickness measurement of a liquid film, they are assembled together as shown in Figure 2.15, so as to measure the same water film on a silicon wafer mounted horizontally on the optical table. The micro conductive probe system is also synchronized with the confocal optical sensor system at the sampling frequency of 1 kHz. Figure 2.16 and Figure 2.17 depict the thickness measurements performed for an adiabatic liquid film and an evaporating film, respectively. The data show that the two techniques have a fair agreement with the maximum difference of around 4 m for the cold film and 6 m for the evaporating liquid layer. Figure 2.16 also illustrates the natural thinning of a film at room atmosphere and temperature, while Figure 2.17 shows much faster thinning due to heat input.

810 815 820 825 830 835 840 845 850

0 200 400 600 800 1000

Time (ms)

Thickness (μm)

(40)

22 Figure 2.15 Calibration setup.

Figure 2.16 Adiabatic liquid film thickness.

Figure 2.17 Evaporating liquid film thickness.

600 620 640 660 680 700

0 100 200 300 400 500

Thickness (μm)

Time (s)

Confocal optical sensor Micro conductive probe

0 200 400 600 800 1000

0 50 100 150 200 250

Thickness(μm)

Time (s)

Confocal optical sensor Micro conductive probe

(41)

23

2.2.3 High speed camera

The high-speed visual system used is ‘DRS Lightning RDT plus’ which can capture 512*512 resolution images at 5 kHz full frames per second (fps) and a blazing 100 kHz fps can be achieved at reduced resolution. In addition, more than 13 seconds of full resolution recording at 5 kHz fps are available and longer record time can be achieved by reducing record frequency or frame’s resolution. The camera is supported by MiDAS software from Xcitex allowing to perform vital motion analysis, control up to four cameras simultaneously, and save digital video data in different file formats including AVI, BMP, JPEG and TIFF.

In order to achieve high frequency observation and good image quality, tungsten spot light (DedoCool) and LED spot light are applied and well positioned to illuminate the visualized objectives.

2.2.4 Micro pump

A gear pump with pulse-free pumping is employed to constantly and precisely supply liquid to the test section,. The pump driver ‘REGLO-Z Analog’ is equipped with a DC motor of 50~5000 rpm with 1% resolution. Correspondingly, the gear pump has the flow rate from 0.0142 ml/s ~ 1.42 ml/s with the resolution of 0.0142 ml/s. The in-situ calibration of the pump is shown in Figure 2.18.

Figure 2.18 Calibration for micro-pump.

0 0.3 0.6 0.9 1.2 1.5 1.8

0 20 40 60 80 100

Rotating speed (r/s)

Flowrate (ml/s)

Nominal Calibration

An Experimental Study on Micro-Hydrodynamics of Evaporating/Boiling Liquid Film

KTH Engineering Sciences