Sie sind auf Seite 1von 260

Graduate School ETD Form 9

(Revised 12/07)

PURDUE UNIVERSITY
GRADUATE SCHOOL
Thesis/Dissertation Acceptance
This is to certify that the thesis/dissertation prepared
By Tannaz Harirchian
Entitled Two-Phase Flow and Heat Transfer in Microchannels

For the degree of Doctor of Philosophy

Is approved by the final examining committee:


Professor Suresh V. Garimella

Professor Jayathi Y. Murthy

Chair

Professor Mamoru Ishii

Professor Cagri A. Savran

Professor Chelsey D. Baertsch

To the best of my knowledge and as understood by the student in the Research Integrity and
Copyright Disclaimer (Graduate School Form 20), this thesis/dissertation adheres to the provisions of
Purdue Universitys Policy on Integrity in Research and the use of copyrighted material.

Professor Suresh V. Garimella


Approved by Major Professor(s): ____________________________________

____________________________________
Approved by: Professor E. Daniel Hirleman / Professor Anil K. Bajaj
Head of the Graduate Program

04/27/2010
Date

Graduate School Form 20


(Revised 1/10)

PURDUE UNIVERSITY
GRADUATE SCHOOL
Research Integrity and Copyright Disclaimer

Title of Thesis/Dissertation:
Two-Phase Flow and Heat Transfer in Microchannels

Doctor of Philosophy
For the degree of ________________________________________________________________

I certify that in the preparation of this thesis, I have observed the provisions of Purdue University
Teaching, Research, and Outreach Policy on Research Misconduct (VIII.3.1), October 1, 2008.*
Further, I certify that this work is free of plagiarism and all materials appearing in this
thesis/dissertation have been properly quoted and attributed.
I certify that all copyrighted material incorporated into this thesis/dissertation is in compliance with
the United States copyright law and that I have received written permission from the copyright
owners for my use of their work, which is beyond the scope of the law. I agree to indemnify and save
harmless Purdue University from any and all claims that may be asserted or that may arise from any
copyright violation.

Tannaz Harirchian
______________________________________
Printed Name and Signature of Candidate

04/20/2010
______________________________________
Date (month/day/year)

*Located at http://www.purdue.edu/policies/pages/teach_res_outreach/viii_3_1.html

TWO-PHASE FLOW AND HEAT TRANSFER IN MICROCHANNELS

A Dissertation
Submitted to the Faculty
of
Purdue University
by
Tannaz Harirchian

In Partial Fulfillment of the


Requirements for the Degree
of
Doctor of Philosophy

May 2010
Purdue University
West Lafayette, Indiana

UMI Number: 3418058

All rights reserved


INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

UMI 3418058
Copyright 2010 by ProQuest LLC.
All rights reserved. This edition of the work is protected against
unauthorized copying under Title 17, United States Code.

ProQuest LLC
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106-1346

ii

To my parents who have always believed in me and encouraged me to follow my


dreams.

iii

ACKNOWLEDGMENTS

I would like to thank my advisor, Professor Suresh Garimella, for his


invaluable guidance and support through the past four years. He has been a
great mentor and a caring friend, and has influenced my academic and personal
life deeply. He has set forth an exceptional example of a teacher for me, and I
am honored to have had the opportunity to work with such a passionate and
outstanding person. I am grateful for his faith and confidence in me, and truly
appreciate his continuous inspiration and encouragement.
I am grateful to my thesis review committee, Professor Jayathi Murthy,
Professor Mamoru Ishii, Professor Cagri Savran, and Professor Chelsey
Baertsch, for agreeing to be on my advisory committee and for their helpful
guidance that has led to success of my work.
I received a great deal of help and support during my graduate studies
from my colleagues in the Cooling Technologies Research Center. I would like
to especially thank Dr. Tailian Chen, Dr. Dong Liu, Dr. Poh-Seng Lee, Mr. Ben
Jones, and Mr. John McHale for their help with my research. The great help of
Mr. Bert Gramelspacher with the experimental setup is also appreciated. Special
thanks are due to Mr. Bruce Myers of Delphi Electronics and Safety, Kokomo,
Indiana, for providing the experimental test pieces. Financial support from the
State of Indiana 21st Century Research and Technology Fund and from the
Cooling Technologies Research Center at Purdue University is gratefully
acknowledged.
I most certainly should express my gratitude to my parents, Shayesteh
and Parviz, for their everlasting love, endless sacrifices, and invaluable support
through my entire educational career.

iv

TABLE OF CONTENTS

Page
LIST OF TABLES ................................................................................................vii
LIST OF FIGURES ............................................................................................. viii
NOMENCLATURE ............................................................................................. xiii
ABSTRACT ...................................................................................................... xviii
CHAPTER 1. INTRODUCTION ............................................................................ 1
1.1. Background ................................................................................................ 1
1.2. Objectives................................................................................................... 2
1.3. Unique Features of the Present Study........................................................ 4
1.4. Organization of the Document .................................................................... 5
CHAPTER 2. EXPERIMENTAL SETUP AND PROCEDURES ............................ 8
2.1. Flow Loop ................................................................................................... 8
2.2. Test Section................................................................................................ 9
2.3. Test Chip Calibration ................................................................................ 10
2.4. Experimental Procedures ......................................................................... 11
2.5. Flow Visualization Procedures.................................................................. 12
2.6. Data Reduction......................................................................................... 13
2.6.1. Measurement Uncertainties................................................................ 16
CHAPTER 3. VISUALIZATION OF FLOW BOILING PATTERNS IN
MICROCHANNELS ............................................................................................ 24
3.1. Literature Review...................................................................................... 25
3.2. Results and Discussion ............................................................................ 29
3.2.1. Boiling Flow Patterns .......................................................................... 29
3.2.2. Effect of Channel Dimensions on Flow Patterns................................. 31
3.2.3. Effect of Mass Flux on Flow Patterns ................................................. 35
3.2.4. Effect of Flow Pattern on Heat Transfer Coefficient............................ 37
3.2.5. Microscale Phenomena ...................................................................... 37
3.3. Conclusions .............................................................................................. 40
CHAPTER 4. EFFECT OF CHANNEL SIZE ON BOILING IN
MICROCHANNELS ............................................................................................ 56
4.1. Literature Review...................................................................................... 56

v
Page
4.2. Experimental Setup and Procedures ........................................................ 58
4.3. Results and Discussion ............................................................................ 59
4.3.1. Critical Role of Channel Cross-Sectional Area on Heat Transfer
Coefficient..................................................................................................... 59
4.3.2. Effect of Vapor Confinement on the Heat Transfer Coefficient ........... 62
4.3.3. Boiling Curve ...................................................................................... 63
4.3.4. Pressure Drop and Pumping Power ................................................... 65
4.4. Conclusions .............................................................................................. 66
CHAPTER 5. EFFECT OF FLOW RATE ON BOILING IN MICROCHANNELS . 77
5.1. Literature Review...................................................................................... 77
5.2. Experimental Setup and Procedures ........................................................ 78
5.3. Results and Discussion ............................................................................ 78
5.3.1. Heat Transfer Coefficient.................................................................... 79
5.3.2. Boiling Curve ...................................................................................... 79
5.3.3. Pressure Drop .................................................................................... 80
5.4. Conclusions .............................................................................................. 81
CHAPTER 6. FLOW REGIME MAPS FOR BOILING IN MICROCHANNELS .... 86
6.1. Literature Review...................................................................................... 86
6.2. Results and Discussion ............................................................................ 88
6.2.1. Flow Regime Maps Based on Conventional Approaches................... 88
6.2.2. Effect of Channel Width on Flow Regime Transitions......................... 90
6.2.3. Comprehensive Flow Regime Map..................................................... 91
6.3. Conclusions .............................................................................................. 93
CHAPTER 7. PREDICTION OF HEAT TRANSFER COEFFICIENT AND
PRESSURE DROP WITH EMPIRICAL CORRELATIONS............................... 105
7.1. Literature Review.................................................................................... 105
7.2. Heat Transfer Coefficient........................................................................ 106
7.3. Pressure Drop ........................................................................................ 108
7.4. Conclusions ............................................................................................ 111
CHAPTER 8. REGIME-BASED MODELING OF HEAT TRANSFER AND
PRESSURE DROP .......................................................................................... 125
8.1. Literature Review.................................................................................... 125
8.2. Model Development................................................................................ 128
8.2.1. Bubbly Flow ...................................................................................... 128
8.2.2. Confined Annular Flow ..................................................................... 128
8.2.3. Annular/Wispy-Annular Flow ............................................................ 135
8.2.4. Slug Flow.......................................................................................... 136
8.3. Pressure Drop ........................................................................................ 140
8.3.1. Confined Flow................................................................................... 141
8.3.2. Unconfined Flow............................................................................... 141

vi
Page
8.3.3. Model Assessment ........................................................................... 142
8.4. Conclusions ............................................................................................ 142
CHAPTER 9. CONCLUSIONS AND RECOMMENDATIONS........................... 157
9.1. Conclusions ............................................................................................ 157
9.2. Suggestions for Future Work .................................................................. 160
LIST OF REFERENCES .................................................................................. 164
APPENDICES
Appendix A. Test Chip Calibration................................................................. 175
Appendix B. Degassing Procedures .............................................................. 180
Appendix C. Heat Loss Measurements ......................................................... 182
Appendix D. Measurement Uncertainties and Experiment Repeatability....... 184
Appendix E. Effect of Channel Size on Boiling Heat Transfer and
Pressure Drop ............................................................................................... 191
Appendix F. Effect of Mass Flux on Boiling Heat Transfer and Pressure
Drop............................................................................................................... 204
Appendix G. MATLAB Script for Heat Transfer Calculations in Annular
Flow............................................................................................................... 225
Appendix H. MATLAB Script for Heat Transfer Calculations in Slug Flow..... 229
VITA ................................................................................................................. 232
PUBLICATIONS ............................................................................................... 233

vii

LIST OF TABLES

Table
Page
2.1. Microchannel dimensions used in the current study (channel
dimensions are referred to in the rest of the paper by their nominal
values for convenience)................................................................................ 17
2.2. Mass fluxes tested for different microchannel dimensions and the
corresponding maximum vapor quality achieved at the channel exit (the
four mass fluxes are referred to in the rest of the paper by the nominal
values of 225, 630, 1050 and 1420 kg/m2s). ................................................ 18
3.1 Summary of the experimental data from the literature used in the
comparisons of the confinement transition criterion...................................... 41
7.1. Summary of boiling heat transfer correlations. ......................................... 112
7.2. Deviation of heat transfer coefficient obtained from the experiments
and predicted by Coopers correlation (1984b). (The channel
dimensions are referred to by their nominal values. The deviation is
presented in terms of Mean Absolute Percentage Error.)........................... 116
7.3. Deviation of heat transfer coefficient obtained from the experiments
and predicted by existing correlations. (The deviation is presented in
terms of Mean Absolute Percentage Error.) ............................................... 117
7.4. Studies in the literature from which heat transfer correlations are
selected for comparison against the current experimental data. Mean
absolute error (MAE) and percentage of predictions which fall within
30% of the measurements are listed for each correlation......................... 118
8.1. Proposed values for the empirical parameters in the three-zone model
for slug flow. ............................................................................................... 145

viii

LIST OF FIGURES

Figure
Page
2.1. (a) Photograph and (b) schematic diagram of the experimental setup. ...... 19
2.2. Microchannel test section. .......................................................................... 20
2.3. Integrated heaters and temperature sensors in the microchannel test
piece............................................................................................................. 21
2.4. Silicon test chips with a constant channel depth of 400 m and channel
widths of (a) 100 m, (b) 250 m, (c) 400 m, (d) 700 m, (e) 1000 m,
(f) 2200 m, and (g) 5850 m....................................................................... 22
2.5. Schematic illustration of microchannel heat sink. ....................................... 23
3.1. Description of boiling flow regimes, d = 400 m, (a) w = 400 m, G =
225 kg/m2s, q = 33 kW/m2, (b) w = 400 m, G = 225 kg/m2s, q = 40
kW/m2, (c) w = 400 m, G = 225 kg/m2s, q = 95 kW/m2, (d) w = 400
m, G = 630 kg/m2s, q = 131 kW/m2, (e) w = 400 m, G = 630 kg/m2s,
q = 157 kW/m2, (f) w = 250 m, G = 225 kg/m2s, q = 157 kW/m2. ............. 43
3.2. Flow patterns in the 400 m 400 m microchannels, G = 630 kg/m2s..... 44
3.3. Flow patterns in the 5850 m 400 m microchannels, G = 630
kg/m2s........................................................................................................... 45
3.4. Flow patterns in the 2200 m 400 m microchannels, G = 630
kg/m2s........................................................................................................... 46
3.5. Flow patterns in the 100 m 400 m microchannels, G = 630 kg/m2s..... 47
3.6. Flow patterns in the 400 m 220 m microchannels, G = 630 kg/m2s..... 48
3.7. Summary of boiling flow patterns in the microchannel heat sinks for G
= 630 kg/m2s; the microchannel dimensions are presented as nominal
width (m) nominal depth (m) with the actual single-channel crosssectional area (mm2) in parentheses. ........................................................... 49
3.8. Flow patterns at the onset of boiling in the 400 m 400 m
microchannels for four mass fluxes. ............................................................. 50
3.9. Effect of mass flux on boiling flow patterns in the 400 m 400 m
microchannels, qw 145 kW/m2................................................................... 51
3.10. Variation of heat transfer coefficient and wall temperature with wall
heat flux in the 400 m 400 m microchannels, G = 630 kg/m2s. ............. 52
3.11. Summary of boiling flow patterns in the microchannel test pieces for
four mass fluxes; the microchannel dimensions are presented as
nominal width (m) nominal depth (m) with the actual single-channel
cross-sectional area (mm2) in parentheses. ................................................. 53
3.12. Transition from confined flow to unconfined flow. ..................................... 54

ix
Figure
Page
3.13. Comparison of the confinement criterion with experimental data from
a variety of sources in the literature; solid symbols and open symbols
represent confined and unconfined cases, respectively. .............................. 55
4.1. Effect of microchannel dimensions (m m) on local heat transfer
coefficient as a function of wall heat flux, G = 630 kg/m2s............................ 68
4.2. Variation of exit vapor quality with wall heat flux for different
microchannel widths, d = 400 m, G = 630 kg/m2s. ..................................... 69
4.3. Effect of microchannel cross-sectional area on heat transfer
coefficient, with trend lines added................................................................. 70
4.4. Effect of microchannel width on local heat transfer coefficient as a
function of base heat flux, d = 400 m, G = 630 kg/m2s. ............................. 71
4.5. Effects of physical confinement on heat transfer coefficients for four
mass fluxes; the microchannel dimensions are presented as width (m)
depth (m). ................................................................................................ 72
4.6. Effect of microchannel dimensions (m m) on boiling curves, G =
630 kg/m2s.................................................................................................... 73
4.7. Variation of base heat flux with wall excess temperature for different
microchannel widths, d = 400 m, G = 630 kg/m2s. ..................................... 74
4.8. Effect of microchannel dimensions on pressure drop, G = 630 kg/m2s. ..... 75
4.9. Effect of microchannel width on pumping power, d = 400 m, G = 630
kg/m2s........................................................................................................... 76
5.1. Effect of mass flux on local heat transfer coefficient for 400 m 400
m microchannels. ....................................................................................... 82
5.2. Effect of mass flux on boiling curves for 400 m 400 m
microchannels. ............................................................................................. 83
5.3. Effect of mass flux on boiling curves for 2200 m 400 m
microchannels. ............................................................................................. 84
5.4. Effect of mass flux on pressure drop for 400 m 400 m
microchannels. ............................................................................................. 85
6.1. Flow regime maps on mass flux-vapor quality coordinates with
transition lines for six microchannels widths, d = 400 m............................. 96
6.2. Flow regime maps on superficial velocity coordinates with transition
lines for six microchannels widths, d = 400 m. ........................................... 97
6.3. Effect of channel size on transition from bubbly or slug flow to
intermittent churn/wispy-annular or churn/annular flow on the mass fluxvapor quality coordinate, d = 400 m. Some relevant transition lines
from the literature are also included. ............................................................ 98
6.4. Effect of channel size on transition from bubbly or slug flow to
intermittent churn/wispy-annular or churn/annular flow on the superficial
velocity coordinates, d = 400 m. Some relevant transition lines from
the literature are also included...................................................................... 99
6.5. Effect of channel size on transition from bubbly flow to intermittent
bubbly/slug flow on the mass flux-vapor quality coordinate, d = 400 m.
Some relevant transition lines from the literature are also included............ 100

x
Figure
Page
6.6. Effect of channel size on transition from bubbly flow to intermittent
bubbly/slug flow on the superficial velocity coordinates, d = 400 m.
Some relevant transition lines from the literature are also included............ 101
6.7. Comprehensive flow regime map for FC-77. ............................................ 102
6.8. Comparison of the comprehensive flow regime map with the
experimental data from the literature. ......................................................... 103
6.9. Comprehensive flow regime map for FC-77; modified with the phase
change number........................................................................................... 104
7.1. Comparison of correlations in the literature for heat transfer
coefficients, 250 m 400 m microchannels, G = 225 kg/m2s................. 120
7.2. Comparison of correlations in the literature for heat transfer
coefficients, 250 m 400 m microchannels, G = 1420 kg/m2s............... 121
7.3. Comparison of correlations in the literature for heat transfer
coefficients, 1000 m 400 m microchannels, G = 225 kg/m2s............... 122
7.4. Comparison of correlations in the literature for heat transfer
coefficients, 1000 m 400 m microchannels, G = 1420 kg/m2s............. 123
7.5. Comparison of the experimentally measured pressure drop across
microchannels with predictions from empirical correlations in the
literature. .................................................................................................... 124
8.1. Comparison of bubbly flow experimental heat transfer coefficients with
predictions from the Cooper correlation (1984b). ....................................... 146
8.2. (a) Schematic representation of annular flow in microchannels, and (b)
simplified flow diagram with vapor core and liquid film control volumes. .... 147
8.3. Comparison of confined annular flow experimental heat transfer
coefficients with predictions from the proposed model. .............................. 148
8.4. Variation of heat transfer coefficient with wall heat flux for confined
annular flow. Experimental results and predictions from the proposed
model are included. .................................................................................... 149
8.5. Comparison of the annular/wispy annular flow experimental heat
transfer coefficients with predictions from the proposed model. ................. 150
8.6. Variation of heat transfer coefficient with wall heat flux for
annular/wispy annular flow. Experimental results and predictions from
the proposed model are included. .............................................................. 151
8.7. Comparison of slug flow experimental heat transfer coefficients with
predictions from (a) original three-zone model (Thome et al., 2004),and
(b) modified three-zone model; both predictions use values proposed in
(Dupont et al., 2004) for the five empirical parameters............................... 152
8.8. Comparison of slug flow experimental heat transfer coefficients with
predictions from the modified three-zone model, using empirical
parameters optimized for the current data.................................................. 153
8.9. Variation of heat transfer coefficient with wall heat flux for slug flow:
comparison of experiments with predictions from the modified threezone model, using empirical parameters optimized for the current data..... 154

xi
Figure
Page
8.10. Schematic representation of flow in microchannels for (a) confined
flow, and (b) unconfined flow. ..................................................................... 155
8.11. Comparison of experimental pressure drops with predictions from the
proposed model: (a) including all the experimental data, and (b)
excluding data for which Bo 0.5 Re < 72 .. .................................................... 156
Appendix Figure
A.1. Test piece assembly for the calibration process....................................... 176
A.2. Convection oven (Lindberg Blue). ............................................................ 177
A.3. Calibration of the resistance values of the 25 heating elements
fabricated on the back side of the 400 m 400 m microchannel heat
sink. ............................................................................................................ 178
A.4. Calibration of the voltage drop across the 25 diode temperature
sensors fabricated on the back side of the 400 m 400 m
microchannel heat sink............................................................................... 179
B.1. (a) Photograph and (b) schematic of the expandable reservoir used to
degas the fluid (Chen and Garimella, 2006b). ............................................ 181
C.1. Heat loss measurements of the 25 heating elements fabricated on the
back side of the 400 m 400 m microchannel heat sink........................ 183
D.1. Measurement uncertainties in the heat transfer coefficient for five
microchannel sizes, G = 630 kg/m2s. ......................................................... 185
D.2. Measurement uncertainties in the heat transfer coefficient for four
mass fluxes in the 1000 m 400 m microchannels................................ 186
D.3. Boiling curves from two separate tests, G = 225 kg/m2s, 250 m 400
m microchannels. ..................................................................................... 187
D.4. Boiling curves from two separate tests, G = 630 kg/m2s, 250 m 400
m microchannels. ..................................................................................... 188
D.5. Boiling curves from two separate tests, G = 1050 kg/m2s, 250 m
400 m microchannels. .............................................................................. 189
D.6. Boiling curves from two separate tests, G = 1420 kg/m2s, 250 m
400 m microchannels. .............................................................................. 190
E.1. Effect of microchannel width on boiling curves, d = 400 m, G = 225
kg/m2s......................................................................................................... 192
E.2. Effect of microchannel width on boiling curves, d = 400 m, G = 630
kg/m2s......................................................................................................... 193
E.3. Effect of microchannel width on boiling curves, d = 400 m, G = 1050
kg/m2s......................................................................................................... 194
E.4. Effect of microchannel width on boiling curves, d = 400 m, G = 1420
kg/m2s......................................................................................................... 195
E.5. Effect of microchannel width on local heat transfer coefficient, d = 400
m, G = 225 kg/m2s.................................................................................... 196
E.6. Effect of microchannel width on local heat transfer coefficient, d = 400
m, G = 630 kg/m2s.................................................................................... 197
E.7. Effect of microchannel width on local heat transfer coefficient, d = 400
m, G = 1050 kg/m2s.................................................................................. 198

xii
Appendix Figure
Page
E.8. Effect of microchannel width on local heat transfer coefficient, d = 400
m, G = 1420 kg/m2s.................................................................................. 199
E.9. Effect of microchannel width on pressure drop, d = 400 m, G = 225
kg/m2s......................................................................................................... 200
E.10. Effect of microchannel width on pressure drop, d = 400 m, G = 630
kg/m2s......................................................................................................... 201
E.11. Effect of microchannel width on pressure drop, d = 400 m, G = 1050
kg/m2s......................................................................................................... 202
E.12. Effect of microchannel width on pressure drop, d = 400 m, G = 1420
kg/m2s......................................................................................................... 203
F.1. Effect of mass flux on boiling curves, 100 m 400 m. ......................... 205
F.2. Effect of mass flux on boiling curves, 250 m 400 m. ......................... 206
F.3. Effect of mass flux on boiling curves, 400 m 400 m. ......................... 207
F.4. Effect of mass flux on boiling curves, 700 m 400 m. ......................... 208
F.5. Effect of mass flux on boiling curves, 1000 m 400 m. ....................... 209
F.6. Effect of mass flux on boiling curves, 2200 m 400 m. ....................... 210
F.7. Effect of mass flux on boiling curves, 5850 m 400 m. ....................... 211
F.8. Effect of mass flux on local heat transfer coefficient, 100 m 400 m... 212
F.9. Effect of mass flux on local heat transfer coefficient, 250 m 400 m... 213
F.10. Effect of mass flux on local heat transfer coefficient, 400 m 400
m. ............................................................................................................. 214
F.11. Effect of mass flux on local heat transfer coefficient, 700 m 400
m. ............................................................................................................. 215
F.12. Effect of mass flux on local heat transfer coefficient, 1000 m 400
m. ............................................................................................................. 216
F.13. Effect of mass flux on local heat transfer coefficient, 2200 m 400
m. ............................................................................................................. 217
F.14. Effect of mass flux on local heat transfer coefficient, 5850 m 400
m. ............................................................................................................. 218
F.15. Effect of mass flux on pressure drop, 100 m 400 m. ....................... 219
F.16. Effect of mass flux on pressure drop, 250 m 400 m. ....................... 220
F.17. Effect of mass flux on pressure drop, 400 m 400 m. ....................... 221
F.18. Effect of mass flux on pressure drop, 700 m 400 m. ....................... 222
F.19. Effect of mass flux on pressure drop, 1000 m 400 m. ..................... 223
F.20. Effect of mass flux on pressure drop, 5850 m 400 m. ..................... 224

xiii

NOMENCLATURE

Ab

heat sink base area, m2

Ac

vapor core area, Ac = ( wch 2 )( d ch 2 ) , m2

Acs

cross-sectional area of a microchannel, m2

Af

wetted area of a fin, m2

Aw

total wetted area of microchannels, m2

Bl

boiling number, Bl = q / ( Gh fg )

Bo

Bond number, Bo = g ( f g ) D 2 /

liquid droplet concentration, kg m-3

C 0

correction factor for initial film thickness

Ca

Capillary number, Ca = f u /

Co

convective number, Co = 1 1
x

0.8

0.5

c1 , c 2 constants in heat loss-temperature relation


c fi

correction factor for interfacial friction factor

cp

specific heat of the fluid, J kg-1 K-1

cq

empirical parameter in Eq. (8.21)

microchannel depth, m

length scale,

Dh

hydraulic diameter, m

liquid droplet quality, e = m E / m

Acs , m; diameter, m

xiv
f

friction constant

enhancement factor, Eqs. (7.16), (7.23), (7.29), (7.39)

Fr

Froude number, Fr = G 2 / ( 2 gD )

F fl

fluid parameter, Eq. (7.34)

FPF

pressure correction factor, Eq. (7.13)

mass flux, kg m-2 s-1

heat transfer coefficient, W m-2 K-1

h fg

latent heat of vaporization for FC-77, J kg-1

jf

liquid superficial velocity, j f = (1 x)G / f , m s-1

jg

gas superficial velocity, jg = xG / g , m s-1

thermal conductivity, W m-1 K-1

kd

deposition mass transfer coefficient, m s-1

contraction or expansion coefficient

microchannel length, heat sink width, m

mass flow rate, kg s-1

molecular weight, kg kmol-1

MAE mean absolute error

exponent in the Gorenflo correlation, Eq. (7.13)

nf

empirical parameter in Eq. (8.21)

nq

empirical parameter in Eq. (8.21)

number of microchannels in a heat sink

N pch

phase change number

Nu

Nusselt number

pressure, Pa

pr

reduced pressure, p pcr

pumping power, W

Pc

vapor core perimeter, Pc = 2 ( wch 2 ) + ( d ch 2 ) , m

xv

Pch

channel perimeter, Pch = 2 ( wch + d ch ) , m

PH

heated perimeter, wch + 2 d ch , m

Pr

Prandtl number

qnet

heat dissipated to the fluid, W

q loss

heat loss, W

heat dissipated from the heat sources, W

qb

base heat flux, W m-2

q w

wall heat flux, W m-2

resistance obtained by calibration,

Ra

average surface roughness, m

Re

Reynolds number

Rp

surface roughness parameter, m

suppression factor

heat sink thickness, m; time, s

temperature, C

Tref

reference temperature: T f in single-phase region and Tsat in two-phase


region, C

velocity, m s-1

voltage applied to the heat sources, V

microchannel width, m

wf

fin width, m

We

Weber number, We = G 2 Dh / ( f )

vapor quality

Martinelli parameter

distance from the channel wall, m

distance along the flow, m

xvi
GREEK SYMBOLS

microchannel aspect ratio ( w / d or d / w whichever is larger)

liquid film thickness, m

pressure drop, Pa

psat

psat (Tw ) p f , Pa

Tsat

Tw Tsat , K

two-phase multiplier

deposition mass transfer rate per unit channel length, kg m-1 s-1

fg

evaporation mass transfer rate per unit channel length, kg m-1 s-1

efficiency of a fin in the microchannel heat sink

overall surface efficiency of the microchannel heat sink

dynamic viscosity, kg m-1 s-1

density, kg m-3

homogeneous density, kg m-3

surface tension, N m-1

fluid particle resistance time; pair period, s; shear stress, Pa

frequency of vapor generation

htp hsp

SUBSCRIPTS
0

reference value; initial

annular

contraction; vapor core

cb

convective boiling

ch

channel

cr

critical

cs

channel cross-section

diode temperature sensor

xvii

dev

developing

dry

dryout zone

expansion

end

end

exit

exit

exp

experimental

entrained droplet

liquid; liquid slug

fd

fully developed

film

liquid film between vapor bubble/core and channel wall

vapor

liquid film interface

in

channel inlet

man

manifold

meas measured
min

minimum

nb

nucleate boiling

liquid slug and elongated bubble pair

pl

plenum

pred predicted

slug

sat

saturated liquid

si

silicon

sp

single phase

turbulent

tp

two-phase

tt

turbulent liquid -turbulent vapor

laminar

bottom wall of the channels

xviii

ABSTRACT

Harirchian, Tannaz. Ph.D., Purdue University, May, 2010. Two-Phase Flow and
Heat Transfer in Microchannels. Major Professor: Dr. Suresh V. Garimella.
School of Mechanical Engineering.

Flow boiling in microchannels has been investigated broadly over the last
decade for electronics cooling applications; however, the implementation of
microchannel heat sinks operating in the two-phase regime in practical
applications has lagged due to the complexity of boiling phenomena at the
microscale. In the current study, extensive experimental work has been
conducted to systematically determine the effects of important geometric and
flow parameters on flow regimes and heat transfer in microscale flow boiling.
Local heat transfer measurements obtained with simultaneous, detailed flow
visualizations lead to a better understanding of boiling phenomena and the
governing heat transfer mechanisms in microchannels.
Based on the experimental results obtained with microchannel test pieces
encompassing a wide range of channel dimensions and operating conditions, a
new transition criterion is developed which predicts the conditions under which
microscale confinement effects are exhibited in flow boiling. This criterion
depends on the value of a parameter termed the convective confinement number
in this study, Bo 0.5 Re , which depends not only on the channel dimensions and
fluid properties, but also on the mass flux. It is shown that physical confinement
exists in the microchannels for Bo 0.5 Re < 160 . In this case, thin-film evaporation
contributes to heat transfer in addition to nucleate boiling and results in larger
values of heat transfer coefficient compared to those cases in which no

xix
confinement is observed. For the larger convective confinement numbers where
physical confinement does not occur and nucleate boiling is dominant, the heat
transfer coefficient is independent of channel dimensions.
A comprehensive flow regime map for flow boiling of a perfluorinated
dielectric liquid (FC-77) is developed based on the experimental data. Using the
convective confinement number and a nondimensional form of heat flux as
coordinates, the flow regime map reveals four distinct regions of confined slug,
bubbly, churn/confined annular, and churn/annular/wispy-annular flow regimes
separated by two quantitative transition lines.
Models are proposed for prediction of the heat transfer coefficient in each
of the four regions in the flow regime map. Also, regime-based prediction of
pressure drop in microchannels is discussed by evaluating pressure drop of each
flow regime along the microchannels separately.

CHAPTER 1. INTRODUCTION

1.1. Background
The increase in the functionality and processing speed of microelectronics
has largely followed Moores law, which predicts that the number of transistors on
an integrated circuit is increasing exponentially, doubling approximately every
two years (Moore, 1965). This trend has continued for half a century and is
expected to continue into the coming decades. This continued increase in the
functionality and compactness of microelectronics is associated with increasing
levels of heat dissipation from these devices. The small size of these devices and
the limits on operating temperature necessitates novel methods for effective
removal of the high heat fluxes.
One of the promising cooling approaches for achieving the high heat
removal rates required is the employment of microchannel heat sinks. In addition
to high heat transfer rates, microchannel heat sinks can be integrated directly
into the heat-dissipating substrates, decreasing the volume of the package.
Boiling in microchannel heat sinks is attractive for high-performance electronics
cooling due to the high heat transfer rates that can be achieved with boiling,
while at the same time, maintaining relatively uniform chip temperatures.
Microchannel flow boiling can also serve for interlayer cooling, enabling chips to
be stacked one on top of the other, resulting in an enhancement of performance
with a reduction in volume.
Use of dielectric liquids in microchannel heat sinks has drawn recent
attention since the working fluid in the microchannel heat sinks can then be in
direct contact with the electronics. Perfluorocarbons are particularly suitable for
direct contact cooling due to their high electrical resistivity.

2
Despite the many practical applications of microscale boiling, the physics
of flow boiling in microchannels, the flow patterns present, and the effect of
microchannel size on the boiling heat transfer and pressure drop are not fully
understood, particularly with dielectric fluids such as perfluorocarbon liquids. In
recent years, progress has been made in understanding the pressure drop and
heat transfer characteristics of flow boiling in microchannels; however, a
fundamental knowledge of boiling mechanisms, the effect of microchannel size
on the boiling regimes, and comprehensive flow regime maps for such flows are
as yet unavailable (Garimella et al., 2006).
A review of the literature by Bertsch et al. (2008a) showed that none of the
empirical correlations developed specifically for flow boiling in microchannels can
predict experimental heat transfer coefficient and pressure drop to within a
reasonable error. They pointed to a clear need for the development of physicsbased models based on the prevalent flow regimes to predict microchannel flow
boiling.

1.2. Objectives
Several aspects of boiling in microchannels have been investigated in the
Cooling Technologies Research Center in the past few years. Dong Liu studied
the onset of nucleate boiling and developed an analytical model to predict the
incipience heat flux (Liu et al., 2005). He also investigated boiling of water in
copper microchannels and modified Chens correlation (Chen, 1966) to take into
account the features specific to flow in microchannels (Liu and Garimella, 2007).
Poh-Seng Lee performed experiments with boiling of water in silicon
microchannels and proposed empirical correlations for prediction of heat transfer
coefficient and pressure drop, following a similar approach to that pursued by
Dong Liu (Lee and Garimella, 2008). Tailian Chen conducted experiments to
explore the physics of flow boiling in copper and silicon microchannels using a
dielectric liquid, FC-77 (Chen and Garimella, 2006a, 2006b, 2006c). Stefan
Bertsch studied boiling of refrigerants in copper microchannels (Bertsch et al.,

3
2008b, 2008d). He performed a review of the existing correlations in the literature
for prediction of boiling in small channels and proposed a correlation to predict
boiling heat transfer (Bertsch et al., 2008a, 2008c). Ben Jones and John McHale
are investigating the effects of surfaces enhancement on pool boiling and flow
boiling in microchannels using water and dielectric liquids as the working fluids
(Jones and Garimella, 2007; McHale and Garimella, 2008).
The aim of the current study is to perform a systematic investigation of
convective boiling in microchannels and to address some of the shortcomings of
the previous studies in this group and elsewhere. Hence, experiments are
performed for a wide range of experimental parameters such as heat flux, mass
flux, and channel dimensions to study the effects of these parameters on boiling
heat transfer and pressure drop as well as on the flow regimes.
Effects of several parameters such as heat flux, mass flux, and channel
size on boiling heat transfer and pressure drop in microchannels are studied with
a perfluorinated dielectric liquid, FC-77. Despite the large number of studies in
the literature on convective boiling, the conditions at which the transition from
macroscale boiling to microscale boiling occurs remain unknown. In the current
study, a new transition criterion is developed for the delineation of a regime
where microscale effects become important to the boiling process and a
conventional, macroscale treatment becomes inadequate.
To develop models for the prediction of heat transfer coefficients based on
the actual flow regimes occurring during convective boiling in microchannels, a
thorough understanding of the flow patterns existing under different conditions,
as well as their transitions, is necessary. Hence, flow visualizations are
performed, simultaneously with local measurements of the temperature and
pressure drop, to determine the flow regimes and to study the effects of channel
size and mass flux on boiling flow patterns. The flow patterns also help to
explaining the heat transfer data.
Two types of flow regime maps commonly used in the literature are
developed for boiling of FC77 in microchannels of different sizes, and the effect

4
of microchannel size on transition between different flow regimes is discussed.
Also, a new comprehensive flow regime map is developed for a wide range of
experimental parameters and channel dimensions, along with quantitative
transition criteria based on nondimensional boiling parameters. These flow
regime maps assist in the practical design of microchannel heat sinks that rely on
two-phase heat transport.
The present experimental results are compared to a number of existing
correlations in the literature for prediction of boiling heat transfer coefficients and
pressure drop. It is shown that empirical correlations are not successful in
prediction of microchannel flow boiling; hence, physics-based models are
developed for the prediction of heat transfer and pressure drop in flow boiling in
microchannels.

1.3. Unique Features of the Present Study


A state-of-the-art thermal test chip is used in the experiments which is
capable of dissipating heat uniformly to the microchannels via twenty-five
microheaters fabricated on the chip, simulating the heat dissipation from a real
electronics chip to the heat sink. Also, twenty-five diode microsensors are
fabricated on the chip, enabling local measurements of wall temperature. Hence
a local study of heat transfer coefficients over the microchannel heat sink is
facilitated. The test chips are provided by Delphi Electronics and Safety and the
channels sawn to specifications.
A wide range of microchannel dimensions with widths ranging from 100
m to 5.85 mm and depths ranging and from 100 m to 400 m are tested to
explore the effects of microchannel size on boiling heat transfer and pressure
drop in otherwise identical test pieces. The effect of flow rate is also investigated
by testing four different mass fluxes ranging from 225 to 1420 kg/m2s. As a
result, a large database consisting of approximately 390 data points, covering a
wide range of experimental parameters, is obtained for boiling of fluorocarbon

5
liquids in microchannels in a manner that facilitates direct comparison across the
parameters varied.
To obtain flow boiling patterns, high-speed imaging is performed using a
high-speed camera which is capable of capturing images with frame rates of up
to 120,000 frames per second. High-quality, detailed flow pattern images are
obtained which assist in developing a better understanding of the physics of
microscale boiling.
A review of the literature shows that the effects of channel size and mass
flux on boiling flow patterns, heat transfer, and pressure drop in microchannels
have not been systematically investigated to date. Also, flow regime maps for
boiling in microchannels have been limited to a narrow range of channel sizes,
and were mostly developed for water and refrigerants. Considering the
importance of understanding the boiling of fluorocarbons in microchannels,
boiling heat transfer and flow patterns are studied in detail for boiling of FC-77 as
a function of channel dimension and mass flux in the present work. Also, flow
regime maps are developed for a wide range of channel sizes and flow
parameters. Physics-based predictive models for flow boiling heat transfer and
pressure drop in microchannels are proposed.

1.4. Organization of the Document


Chapter 1 provides a background of convective boiling in microchannels
and its applications, as well as the objectives and the unique features of the
present work.
In Chapter 2, a detailed explanation of the experimental setup and the test
chip is given along with the test procedures and the flow visualization method.
Also, data reduction and measurement uncertainties are explained. Chapters 2
to 8 each start with a literature review on the specific topic under study. The
experimental or analytical results are then discussed and conclusions provided.
Chapter 3 presents the flow visualization results, the boiling flow patterns,
and the effects of channel size and mass flux on flow regimes. Also, new

6
quantitative criteria for the transition between macro- and micro-scale boiling
behavior have been identified. The content of this chapter was presented and
published in the proceedings of the 11th IEEE Intersociety Conference on
Thermal and Thermomechanical Phenomena in Electronic Systems (Harirchian
and Garimella, 2008b) and the Annual IEEE Semiconductor Thermal
Measurement and Management Symposium (Harirchian and Garimella, 2008c).
Also, some of the visualization images were presented in the Heat Transfer
Photogallery at the ASME IMECE2007 and published in the Journal of Heat
Transfer (Harirchian and Garimella, 2008d).
In Chapter 4 and Chapter 5, experimental investigations of the effects of
channel size and mass flux, respectively, on convective boiling heat transfer and
pressure drop are discussed in light of a knowledge of the flow patterns. The
material in these two chapters was presented at the ASME International
Mechanical Engineering Congress and Exposition in 2007 and published in the
proceedings (Harirchian and Garimella, 2007). This material along with a
comparison of the heat transfer results to the predictions of correlations in the
literature (presented in Chapter 7) was later published in the International Journal
of Heat and Mass Transfer (Harirchian and Garimella, 2008a). The effect of
microchannel size on heat transfer and flow patterns and the critical role of
channel cross-sectional area as a geometrical parameter on flow boiling were
published in International Journal of Multiphase Flow (Harirchian and Garimella,
2009b).
In Chapter 6, the absence of adequate flow regime maps for boiling in
microchannels is discussed and two types of flow regime maps using coordinates
common in the literature are developed based on the experimental visualizations
and the effect of channel dimension on flow regime transition is explored. Also, a
new type of comprehensive map is developed encompassing a wide range of
channel dimensions and flow parameters. A paper including the visualization
results of Chapter 3 and the flow regime maps developed on conventional
coordinates is published in the International Journal of Multiphase Flow

7
(Harirchian and Garimella, 2009a). The comprehensive flow regime map and the
quantitative transition criteria for the microscale phenomena discussed in
Chapter 3 were published in the International Journal of Heat and Mass Transfer
(Harirchian and Garimella, 2010).
The experimental results for the heat transfer coefficient and pressure
drop are compared to predictions from a number of existing correlations in the
literature in Chapter 7. A need for physics-based models for prediction of flow
boiling in microscale is discussed. In Chapter 8, analytical models for prediction
of heat transfer and pressure drop in different flow regimes in microchannels are
discussed and compared to the experimental results. Results of these two
chapters are submitted to the International Journal of Heat and Mass Transfer for
possible publication.
Chapter 9 presents a summary of the present study along with
suggestions for future work.
A keynote lecture in the 15th International Workshop on Thermal
Investigation of ICs and Systems (Garimella and Harirchian, 2009) summarized
the comprehensive understanding of flow boiling in microchannels obtained in
this work; this lecture is accepted for publication in the Journal of Electronic
Packaging.

CHAPTER 2. EXPERIMENTAL SETUP AND PROCEDURES

An experimental setup is carefully designed and state-of-the-art thermal


test chips are used to perform systematic experiments of boiling in
microchannels and to investigate the effects of microchannel dimensions and
flow parameters on boiling characteristics. In this chapter, details of the flow loop
and the test section assembly used for the experiments are explained and the
test chip calibration methods are elucidated. The test procedure and flow
visualization method, along with data reduction are presented as well.

2.1. Flow Loop


The experimental setup is shown in Figure 2.1. A magnetically coupled
gear pump drives the dielectric liquid, FC-77, through the closed loop. A
preheater installed upstream of the test section heats the coolant to the desired
subcooling temperature at the entrance of the microchannels, and a liquid-to-air
heat exchanger located downstream of the test section cools the fluid before it
enters the reservoir. The liquid is fully degassed before initiating each test using
the two degassing ports and the expandable reservoir. Details of the expandable
reservoir design and the degassing procedure are available in Chen and
Garimella (2006b). A flow meter with a measurement range of 20-200 ml/min
monitors the flow rate through the loop and five T-type thermocouples measure
the fluid temperature before and after the preheater, before and after the test
section, and after the heat exchanger. The pressure in the outlet manifold of the
test section is maintained at atmospheric pressure. The pressure at the inlet
manifold and the pressure drop across the microchannel array are measured

9
using a pressure transducer (Gems Sensors, 2200 series) and a differential
pressure transducer (Omega, PX2300 series), respectively.

2.2. Test Section


The test section shown in Figure 2.2 consists of a 12.7 mm 12.7 mm
silicon microchannel heat sink with thickness of 650 m, mounted on a printed
circuit board (PCB) 1. The PCB is installed on a quick-connect board with an
insulating G10 piece in between. A polycarbonate top cover above the test piece
provides an enclosed passage for the liquid and is sealed with an o-ring. To
avoid melting of the polycarbonate at high chip temperatures, a 12.7 mm 12.7
mm Pyrex sheet of thickness 0.4 mm with a high melting point (820C) is
sandwiched between the silicon die and the top cover and forms the top wall of
the microchannels.
As detailed in Figure 2.3, parallel microchannels are cut on the top surface
of the silicon chip using a dicing saw. The width (w), depth (d), and number of
microchannels (N), along with the hydraulic diameter (Dh), aspect ration (w/d),
single channel cross-sectional area (Acs), and surface roughness of the bottom
walls of microchannels in each heat sink are listed in Table 2.1. The surface
roughness for the side and bottom walls is measured using a probe-type
profilometer and an optical profilometer, respectively. The bottom wall of the 100
m wide microchannels has an average roughness of 0.22 m. Wider
microchannels were made with a number of cuts for each channel; this process
imparts a waviness to the bottom surfaces resulting in an average roughness of
0.82 to 1.40 m for the different test pieces; the average roughness in the region
of a single cut is 0.2 m. The average surface roughness of the side walls is 0.1
m for all test pieces.

The author thanks Bruce Myers and Darrel Peugh of Delphi Electronics and

Safety, Kokomo, Indiana, for providing the silicon test pieces.

10
A 5 5 array of resistance heat sources and a like array of temperature
sensing diodes are fabricated on the other side of the chip, as shown in Figure
2.3, providing uniform heat flux to the back side of the microchannels and local
measurements of the base temperature. Since the resistances of all the heat
sources are almost identical, they are connected in parallel and are supplied with
a single DC voltage in order to provide a uniform heat flux to the back of the
microchannels. The amount of heat generated by each source is obtained from
the calibrated resistance of the corresponding element and the applied voltage.
The calibration procedure for the fabricated resistors and diodes is explained in
section 2.3.
For a given current passing through a diode temperature sensor, the
voltage drop across the diode depends on the temperature. To obtain the
voltage-temperature relationship for each diode fabricated on the silicon chip, the
test piece is calibrated in a convection oven as will be discussed in the next
section.
The calculated local heat transfer coefficients presented in this document
are based on measurements from the temperature sensor at location 3 in Figure
2.3, which is along the centerline of the test piece near the exit.

2.3. Test Chip Calibration


The test chip calibration procedure is discussed in detail in Appendix A
and key points are included here. To calibrate the resistors and diodes
fabricated on the back side of the silicon chip, the PCB with the mounted silicon
chip is installed on the quick connect board (Figure A.1) and is placed in a
convection oven (Figure A.2). The temperature of the oven is then increased
from room temperature (25C) up to 130C.
The resistors are connected to the data acquisition system and the
resistance of each heating element is measured at seven temperatures ranging
from 25C to 130C, using a two-wire resistance mea surement method. To
correct for the resistance of the leads and connections, the PCB is then replaced

11
with a copper plate and the total resistance of the leads and connections are
measured. These values are then subtracted from the total resistance of the
corresponding resistor to obtain an accurate value for the resistance of each
element. As explained in Appendix A and shown in Figure A.3, the resistance of
each heating element increases quadratically with temperature.
To calibrate the diode temperature sensors, twenty-five individual current
sources are used to apply 100 A to each diode and the voltage drop across
each diode is measured at seven temperatures ranging from 25C to 115C. A
linear relation between the temperature and voltage drop across each diode is
obtained as shown in Figure A.4.
More details of the individual heaters, diode temperature sensors,
calibration procedures, test section assembly, and test loop are available in Chen
and Garimella (2006a), Lee (2007), and Lee and Garimella (2008).

2.4. Experimental Procedures


Experiments are conducted for 12 test pieces with channel depths of 100
m, 220 m, and 400 m and different channel widths ranging from 100 m to
5850 m to explore the effects of geometrical parameters, such as channel
width, height, aspect ratio, and cross-sectional area over a wide range of microto macroscale dimensions on the boiling heat transfer coefficients and pressure
drops as a function of mass flux. Four mass flux values ranging from 225 to
1420 kg/m2s are investigated for each test piece to map the effect of flow velocity
on boiling, as will be discussed in Chapter 5. The mass fluxes used for each of
the test pieces are summarized in Table 2.2. While the actual values are used
for all the calculations in this study, nominal values are used for easier
reference of the dimensions and the flow rates in the text. As listed in this table,
there are ten test pieces with widths ranging from 100 m to 1000 m and two
test pieces with much larger channel widths of 2200 m and 5850 m.
Photographs of the test chips with a depth of 400 m are provided in Figure 2.4.

12
Before initiating each test, the liquid in the test loop is fully degassed,
following a procedure explained in Appendix B, to help reduce flow instabilities
(Chen and Garimella, 2006b). It is then driven into the loop at a constant flow
rate and preheated to approximately 92C, providing 5C of subcooling at the
inlet of the channels. For each test, the flow rate is kept constant throughout the
test and the heat flux to the chip is increased from zero to the point at which the
maximum wall temperature reaches 150C, which is a safe upper-limit operating
temperature for the integrated heaters and temperature sensors. Heat flux
values approaching critical heat flux are avoided since the corresponding
temperatures could cause the solder bumps in the test chip to melt. A throttling
valve positioned upstream of the test section serves to suppress instabilities in
the microchannel heat sink. Nevertheless, mild flow reversals at the inlet of the
microchannels were observed at the highest heat fluxes studied, for
microchannels with small cross-sectional area and for low mass fluxes.
However, these instabilities did not affect the inlet fluid temperature which is held
constant throughout each test.

2.5. Flow Visualization Procedures


A high-speed digital video camera (Photron Fastcam-Ultima APX) is
employed for flow visualizations. The high-speed camera provides full-pixel
resolution of 10241024 for up to 2,000 frames per second (fps) and is capable
of recording at frame rates as high as 120,000 fps with reduced resolution.
Different microscope lenses, depending on the microchannel size, are used for
magnification. An illumination source (Henke-Sass, Wolf) is used to illuminate
the microchannels for visualization. Using the same experimental setup, Chen
and Garimella (2006a) showed that this illumination source does not result in
noticeable additional heat input to the fluid.
At each heat flux, after the system reaches a steady state, high-speed
visualizations are performed simultaneously with the heat transfer and pressure
drop measurements. Movies of the flow patterns are captured with various frame

13
rates ranging from 2000 frames per second (fps) to 24,000 fps, with the higher
frame rates used for the smaller microchannels at the larger mass fluxes. Also,
as the heat flux is increased, the vapor velocity increases and higher frame rates
are required to capture the details of the flow. The images obtained from the
camera are then post-processed using a MATLAB code developed in-house by
Ben Jones and John McHale to enhance the quality of the images, especially for
those captured at higher frame rates.
Flow visualizations have been performed for 11 of the 12 test pieces at
different flow rates. The nominal dimensions of the test pieces and the flow rates
are summarized in Table 2.2.
It is noted that all the values of heat transfer coefficient and wall
temperature reported in this study are based on the local measurement at a
location near the exit and along the centerline of the microchannel array. The
flow visualizations are performed at this location as well.

2.6. Data Reduction


The heat transfer rate to the fluid, qnet , in the microchannels is obtained
from an energy balance conducted for each heating element:
qnet = q qloss

(2.1)

in which q is the total heat dissipated from each of the 25 heat sources and is
calculated as q = V 2 / R . The heat loss, qloss , consists of losses through natural
convection from the test setup surfaces to the surrounding air, radiation from
these surfaces, and conduction through the PCB and the top cover, and is
determined as follows. Before the test section is charged with liquid, a constant
voltage is applied to the heaters. When the readings of the diode temperature
sensors reach a steady state, the temperature of each sensor is recorded and
correlated to the heat dissipated from the corresponding heater at that location.
This procedure is repeated for several levels of input power and a linear relation
in the form of qloss = c1Td + c2 is obtained, where c1 and c 2 are constants and are

14
slightly different at each location and for different test pieces. The heat loss
measurements for the 400 m 400 m microchannel test chip are presented in
Appendix C. The linearity of the heat loss-temperature relationship indicates that
the losses are primarily due to conduction of heat from the test section to the
ambient.
The local heat transfer coefficient corresponding to each heated element
area is then calculated from

h=

qw
o (Tw Tsat )

(2.2)

where Tw is the local microchannel wall temperature as shown in Figure 2.5.


This temperature is corrected for the microchannel base thickness using
Tw = Td

qb ( t d )

ksi

(2.3)

where Td is the temperature measured by an integrated diode. o is the overall


surface efficiency of the microchannel heat sink and is defined as

o = 1

NA f
Aw

(1 )
f

(2.4)

where A f = 2 Ld is the wetted area of a fin , Aw = N ( w + 2d ) L is the total wetted


area of the microchannels, and f =

tanh md
is the efficiency of a fin with an
md

adiabatic tip, with m 2 = 2h ksi w f . In these calculations, the wall between each
microchannel is considered as a fin and the tips of the fins are assumed to be
adiabatic since the heat transfer from the fin tips to the top cover is much lower
than the heat transfer from the fin walls to the liquid inside the channels. As a
first approximation, h is calculated assuming an overall surface efficiency of
100%; this value of h is used to calculate the fin efficiency and overall efficiency,
which then lead to an updated value for h. This iteration is continued till a
converged value is obtained. The overall efficiencies of the heat sinks tested
ranged from 96.5% to 99.9%.

15
The local wall heat flux used in Eq. (2.2) is the net heat flux added to the
fluid as shown in Figure 2.5 and is calculated using the total wetted area as

qw = qnet / ( Aw / 25)

(2.5)

The heat flux used in finding the heat loss through the base of the microchannel
heat sink and hence, calculating the wall temperature in Eq. (2.3), is the base
heat flux and is calculated using the chip base area, Ab = L L , which is the
same as the combined area of all the heat sources

qb = qnet / ( Ab / 25 )

(2.6)

In Eqs. (2.5) and (2.6), the areas are divided by 25 because qnet is defined as the
net heat dissipated from one of the 25 heat sources.
The exit vapor quality is calculated from an energy balance using
xexit =

1
h fg

qnet

c p , f (Tsat Tin )

m f

(2.7)

Important nondimensional parameters often used in flow boiling include


Reynolds number, Re, Bond number, Bo, and Boiling number, Bl. The Reynolds
number is calculated using the liquid phase mass flux as:

Re =

GD

(2.8)

The Bond number represents the ratio of buoyancy force to surface


tension force and becomes important in microscale boiling:
Bo =

g ( f g ) D2

(2.9)

As will be demonstrated in the following chapters, the channel crosssectional area plays a critical role in determining microchannel boiling
mechanisms and heat transfer; therefore, the length scale used in Eqs. (2.8) and
(2.9) is the square root of the cross-sectional area of one channel rather than its
hydraulic diameter.
Boiling number is the nondimensional form of the heat flux and is
calculated using the liquid mass flux and latent heat as follows:

16

Bl =

qw
Gh fg

(2.10)

2.6.1. Measurement Uncertainties


The measurement uncertainties for the flow meter and the pressure
transducers are 1% and 0.25% of full scale, respectively. The uncertainties in
the measurements of the channel dimensions, the T-type thermocouples, and the
diode temperature sensors are 15 m, 0.3C and 0.3C, respectively, while
those for the microheater resistance and applied voltage are 0.002% and
0.004%, respectively. Following a standard uncertainty analysis (Taylor, 1997),
the uncertainties associated with the wall heat flux and the heat transfer
coefficient are estimated to be 2.0 to 11.4% and 2.2 to 11.7%, respectively, for
the cases considered. These uncertainties are primarily governed by
uncertainties in the measurement of the area, since the uncertainties in the net
heat transfer rate, wall temperature, and saturation temperature are relatively
small. In Appendix D, measurement uncertainties for the heat transfer coefficient
are presented for several cases. Repeatability of the experiments is also
discussed in this appendix.

17

Table 2.1. Microchannel dimensions used in the current study (channel dimensions are referred to in the rest of the
paper by their nominal values2 for convenience).

w/d

Channel crosssectional area,


Acs (mm2)

Surface
roughness,
Ra (m)

96

1.05

0.009

0.22

63

134

0.45

0.021

0.22

101.9 (100)

60

159

0.27

0.037

0.22

371.4 (400)

110.8 (100)

35

291

0.64

0.089

0.82

398.2 (400)

64.7 (100)

101.8 (100)

25

111

6.12

0.026

1.28

400 (400)

196.6 (220)

100 (100)

25

264

2.03

0.079

1.33

394.6 (400)

364.9 (400)

105.4 (100)

24

379

1.08

0.144

1.35

686.3 (700)

375.6 (400)

154.1 (150)

14

486

1.83

0.258

1.11

1024 (1000)

225.7 (220)

176 (200)

10

370

4.53

0.231

1.40

978.4 (1000)

373.7 (400)

222.3 (200)

10

541

2.62

0.366

1.03

2202.8 (2200)

370.1 (400)

280 (300)

634

5.95

0.815

1.10

5850.5 (5850)

376.2 (400)

300 (300)

707

15.55

2.201

1.10

Microchannel
width, w (m)

Microchannel
depth, d (m)

Fin width, wf
(m)

# of
channels,
N

Hydraulic
diameter,
Dh (m)

98.5 (100)

93.7 (100)

101.5 (100)

61

96.6 (100)

216.6 (220)

103.4 (100)

101.5 (100)

369.4 (400)

239.1 (250)

The nominal values for channel dimensions are provided inside parentheses in the table.

17

18

Table 2.2. Mass fluxes tested for different microchannel dimensions3 and the corresponding maximum vapor quality
achieved at the channel exit (the four mass fluxes are referred to in the rest of the paper by the nominal values of 225,
630, 1050 and 1420 kg/m2s).

Microchannel
depth, d (m)

Mass flux, G (kg/m2s)

Maximum exit vapor


quality, xe (%)

98.5 (100)

93.7 (100)

660

superheated

96.6 (100)

216.6 (220)

630

78

101.5 (100)

369.4 (400)

214, 621, 1017, 1405

superheated, 74, 48, 32

239.1 (250)

371.4 (400)

226, 611, 1126, 1415

superheated, 49, 31, 23

398.2 (400)

64.7 (100)

615

75

400 (400)

196.6 (220)

637

56

394.6 (400)

364.9 (400)

227, 633, 1031, 1431

77, 35, 21, 15

686.3 (700)

375.6 (400)

225, 641, 1053, 1461

64, 32, 19, 11

1024 (1000)

225.7 (220)

630

20

978.4 (1000)

373.7 (400)

224, 627, 1037, 1440

52, 23, 15, 8

2202.8 (2200)

370.1 (400)

227, 633, 1034, 1427

30, 15, 9, 4

5850.5 (5850)

376.2 (400)

229, 632, 1028, 1289

29, 15 , 9, 5

The nominal values for channel dimensions are provided inside parentheses in the table.

High-speed
visualization

18

Microchannel
width, w (m)

19

High-speed
camera with
microscope
lens

Expandable
reservoir

Pump

Heat
exchanger

Pre-heater
Test section

Differential
pressure
transducer

(a)

P1
Degassing
port 1

Pump
P5

Filter

Filter
Expandable
reservoir

Flow Meter

P4
T1
T5

Degassing
port 2

Preheater
Heat Exchanger

T2

T3

T4

P2
Test
section
P3
dP

(b)

Figure 2.1. (a) Photograph and (b) schematic diagram of the experimental setup.

20

Figure 2.2. Microchannel test section.

21

Figure 2.3. Integrated heaters and temperature sensors in the microchannel test
piece.

22

(a)

(b)

(c)

(d)

(e)

(f)

(g)

Figure 2.4. Silicon test chips with a constant channel depth of 400 m and
channel widths of (a) 100 m, (b) 250 m, (c) 400 m, (d) 700 m, (e) 1000 m,
(f) 2200 m, and (g) 5850 m.

23

L
wf

qw"
d

Tw
Td

qb"
Figure 2.5. Schematic illustration of microchannel heat sink.

24

CHAPTER 3. VISUALIZATION OF FLOW BOILING PATTERNS IN


MICROCHANNELS

Flow boiling regimes in microchannels and the effects of channel size and
mass flux on the flow patterns are explained in this chapter for 11 microchannel
heat sinks with different channel dimensions and four mass fluxes, as listed in
Table 2.2. The detailed investigation of flow regimes in the current study is
aimed at exploring the effects of channel size, mass flux and heat flux on the
boiling flow patterns, and explaining the trends observed in the heat transfer
coefficients which will be discussed in the following chapters. The effects of
channel width and depth as well as the aspect ratio and cross-sectional area on
flow boiling in microchannels are investigated as independent parameters, and
as a result, the important geometrical parameters that influence boiling in
microchannels are identified.
Confinement effects in microchannels are very important as they affect the
heat transfer mechanisms in flow boiling; hence, it is necessary to quantitatively
determine the conditions at which vapor confinement occurs and microscale
effects become present. At the end of this chapter, quantitative criteria for the
transition between macro- and micro-scale boiling behavior are developed and
nondimensional parameters which govern the occurrence and extent of flow
confinement are determined. The results obtained are also employed in the
development of flow regime maps for boiling of FC-77 in microchannels as will be
explained in Chapter 6.

25
3.1. Literature Review
A number of studies in recent years have attempted to better understand
the flow patterns during boiling in microchannels using different working fluids
(Garimella and Sobhan, 2003; Sobhan and Garimella, 2001; and Bertsch et al.,
2008a); however, a systematic investigation into the effects of channel size and
mass flux on the boiling flow patterns in microchannels has not been performed.
High-speed flow visualization has been employed to understand the
physics of boiling in microchannels as well as the differences in the boiling
regimes relative to those in conventional-sized channels. Liu et al. (2005)
studied the onset of nucleate boiling in water flow through microchannels of 275
m 636 m cross-section via high-speed flow visualization and developed an
analytical model to predict the incipience heat flux and bubble size at the onset of
boiling. Peng and Wang (1993) experimentally investigated flow boiling of water
in microchannels of cross-section 600 m 700 m. They showed that unlike
larger tubes, partial nucleate boiling was not observed in the subcooled region,
and fully developed boiling was induced much earlier than at the macroscale.
Kandlikar (2004) observed features such as flow reversal and nucleate boilingdominant heat transfer in microchannel flow boiling as well as flow patterns of
slug flow, annular flow, churn flow, and dryout. Also, two nondimensional groups,
representing some of the important flow boiling characteristics, were derived.
Revellin et al. (2006) studied two-phase flow patterns of R134a in a 500 m glass
channel using an optical measurement method as well as high-speed imaging
and delineated four flow patterns of bubbly, slug, semi-annular, and annular
flows. The frequency of bubble generation, coalescence rate and mean bubble
velocity were also determined. Qu and Mudawar (2004) studied convective
boiling heat transfer, two-phase flow patterns, and pressure drop in parallel
microchannels of 231 m width and 712 m depth using water as the working
fluid. They argued that unlike in the case of fluorochemicals, it is difficult to
sustain bubbly flow regimes in boiling of water in microchannels because of the
high surface tension and large contact angle of water; a slug flow regime was

26
found to develop shortly after the incipience of boiling. Also, the flow patterns
were strongly influenced by the applied heat flux. Chen and Garimella (2006a)
observed bubbly and slug flow patterns at lower heat fluxes and wispy-annular
and churn flow regimes at higher heat fluxes with flow reversal near the
microchannel inlet. Bertsch et al. (2008b) observed slug or intermittent flow in
most of their experiments; only at very high vapor qualities was annular flow
observed.
Kew and Cornwell (1997) experimentally studied two-phase flow regimes
and heat transfer with refrigerant R141b in circular tubes of diameters in the
range of 1.39 to 3.69 mm. They observed flow regimes that differed slightly from
those observed in large channels and found that three flow regimes of isolated
bubble flow, confined bubble flow, and annular-slug flow were sufficient to
describe the observed patterns. They also showed that for narrow channels
having a confinement number of 0.5 and larger, conventional correlations were
not suitable for the prediction of flow boiling heat transfer and that nucleate
boiling correlations such as that of Cooper (1984a) predict the data better than
the established correlations for flow boiling. Huo et al. (2004) investigated boiling
heat transfer with R134a in small-diameter tubes at mass fluxes ranging from
100 to 500 kg/m2s. Their visualizations revealed six flow patterns: dispersed
bubble, bubbly, slug, churn, annular, and mist flow. It was also shown that
nucleate boiling was dominant for vapor qualities less than 40-50% (20-30%) for
tubes of diameter 4.26 (2.01) mm.
Effects of hydraulic diameter and surface roughness on bubble nucleation
and flow patterns were studied by Zhang et al. (2005b) in single silicon
microchannels of 27-171 m hydraulic diameters and different surface
roughnesses with DI water. They showed that in 171 m-diameter channels, a
typical nucleation process occurs and the bubbles formed contribute to an
annular flow pattern. In channels smaller than 50 m in hydraulic diameter, the
bubble nucleation mechanism was found to be that of eruption boiling and mist
flow developed almost right after single-phase flow ended because of the large

27
amount of wall superheat. Hence, no temperature plateau was observed. The
boiling mechanism was found to be determined by the wall surface conditions
rather than the channel dimensions. Bergles et al. (2003) also performed an
analysis and showed that incipience of boiling in a subcooled flow was most
likely governed by the nucleation sites in the walls and not by the channel size.
Lee et al. (2005) investigated the effect of microchannel height on nucleation-site
activity and bubble dynamics using three fluids: water, methanol and ethanol.
The heat sink consisted of ten parallel microchannels fabricated in a silicon
wafer, with width ranging from 150 to 9000 m and height ranging from 5 to 510
m. They found that the bubble nucleation activity was dependent on channel
height for all fluids tested.
Other studies have utilized flow visualization to record the flow patterns
under different flow conditions and with specific geometries. Jiang et al. (2001)
carried out experiments with water in triangular silicon microchannels of hydraulic
diameter 26 and 53 m and observed local nucleate boiling at low input power
and unstable slug flow at intermediate power levels. However, annular flow
developed at a relatively low heat flux, such that evaporation at the liquid
film/vapor core interface was the dominant heat transfer mode over a wide range
of input powers. Lee et al. (2003) explored the effect of channel shape on twophase flow patterns in DI water. The microchannel heat sink they tested
consisted of ten shallow, nearly rectangular silicon microchannels of 120 m
width and 14 m depth. They studied bubble dynamics and two-phase flow
patterns and compared the results to similar work by Jiang et al. (2001) in
triangular microchannels with hydraulic diameter of 26 m. Wang et al. (2008)
performed experiments to investigate the effect of inlet/outlet configurations on
flow boiling instabilities with water in parallel microchannels of hydraulic diameter
186 m. They observed bubbly/slug flow, bubbly/annular alternating flow, and
annular/mist alternating flow when the flow could freely enter and exit the
microchannels. In the case of flow being restricted at the channel inlet but not at
the outlet, steady flow boiling patterns of isolated bubbles, elongated bubbles,

28
and annular flow were seen. Zhang et al. (2005a) studied the flow patterns in
microchannel at different orientations. They observed boiling to occur in isolated
bubble and slug flow regimes and found that the vertical up-flow orientation had
the best thermal performance, highest flow stability, and lowest pressure drop.
Differences between the flow regimes with air-water and steam-water
experiments were observed in the visualizations of Hetsroni et al. (2003) in
parallel triangular channels etched in a silicon substrate with a base dimension of
200 to 310 m.
Adiabatic two-phase flow patterns were investigated through high-speed
visualizations and flow regime maps developed by Field and Hrnjak (2007) and
Hassan et al. (2005). Field and Hrnjak (2007) performed visualizations of
adiabatic two-phase flow in a glass channel of diameter 500 m for R134a,
propane, ammonia, and a mixture of R134a and oil with different concentrations.
They observed four distinct flow regimes of bubbly-slug, slug, slug-annular, and
annular flow and compared these flow patterns to various flow maps.
In flow boiling through channels, as the channel size decreases to
approach the bubble diameter, physical confinement begins to modify the
influence of the different governing forces, resulting in different behavior of boiling
in microchannels compared to that in conventional-sized channels. Kandlikar
(2002) classified channels with hydraulic diameters between 10 and 200 m as
microchannels for flow boiling, based merely on dimensions and not on physical
behavior. Kew and Cornwell (1997) proposed a criterion for a threshold hydraulic
diameter below which microchannel two-phase flow is characterized by confined
single bubbles; the available models for macroscale boiling were found to be
unsuitable for the prediction of heat transfer and pressure drop at these small
channel sizes. Serizawa et al. (2002) recommended a confinement criterion
similar to that of Kew and Cornwell representing the ratio of surface tension and
gravity forces for the channel size below which the influence of surface tension
becomes important. These confinement criteria include channel hydraulic

29
diameter and fluid properties; however, it will be shown from the results obtained
in the current study that mass flux also governs bubble confinement.
The present work seeks to address the limited understanding in the
literature of the effects of channel size and mass flux on boiling flow patterns in
microchannels. The results are also aimed at explaining the effects of channel
size, mass flux and heat flux on the boiling heat transfer coefficients obtained in
the same facility which will be discussed in the following chapters. A new criterion
for the occurrence of vapor confinement in flow boiling in microchannels is
proposed that incorporates mass flux as well as channel cross-sectional area
and fluid properties based on the flow visualization results.

3.2. Results and Discussion


The flow visualizations are first analyzed and the flow patterns observed in
the microchannels characterized. The effects of heat flux, mass flux, and
channel dimensions on the boiling flow patterns are then discussed and the
important geometrical parameters that influence boiling in microchannels
identified. Finally, quantitative criteria for the transition between macro- and
micro-scale boiling behavior are developed.

3.2.1. Boiling Flow Patterns


In the microchannels studied, flow patterns observed via high-speed
visualizations are categorized into five major flow regimes bubbly, slug, churn,
wispy-annular, and annular flows; a sixth post-dryout regime of inverted-annular
flow is also identified. Although these flow patterns have a slightly different
appearance in different channel sizes and for different mass fluxes and flow
rates, each is characterized by certain common features. These flow patterns
are first explained with respect to the 400 m 400 m microchannels.
Representative visualized images in these six flow regimes are illustrated
in Figure 3.1. In all the flow visualization images in this work, the flow direction is

30
from left to right. For better delineation of the liquid and vapor phases in the flow
patterns, a sketch of the flow pattern is added on the right side of each image.
Figure 3.1(a) shows bubbly flow in which isolated round and elongated bubbles
that are smaller than the cross-section of the microchannels move in the flow
direction. Bubbles generally nucleate at the microchannel walls and detach from
the walls after growing. The shape and size of the bubbles vary with flow rate
and heat flux. As the heat flux increases, the bubble generation rate at the walls
increases and bubbles become larger as a result of bubble coalescence. At
higher heat fluxes or in smaller microchannels, bubbles occupy the entire crosssection of the channels, resulting in slug flow as shown in Figure 3.1(b); small
bubbles exist in the liquid slugs between the elongated bubbles. The churn flow
regime is demonstrated in Figure 3.1(c). This flow regime consists of vapor
chunks transported downstream and large bubbles nucleating at a high rate at
the channel walls; however, at high heat fluxes, the nucleation at the walls may
be suppressed. In wispy-annular flow as in Figure 3.1(d), a vapor core is
separated from the channel walls with a relatively thick and unstable liquid film.
Large, irregular-shaped droplets are entrained into the vapor core. Very few
nucleation sites remain in the liquid film and result in small vapor bubbles in the
liquid layer. In annular flow, as illustrated in Figure 3.1(e), the liquid layer is
thinner than in wispy-annular flow, and the interface between the vapor core and
the liquid film can become wavy. The liquid film thickness decreases as the heat
flux increases. Small, round droplets are entrained into the vapor core, while no
vapor bubbles are seen in the liquid annulus. At vey high heat fluxes, when
critical heat flux is reached, the walls can completely dry out under certain
conditions and a vapor blanket forms at the walls around a liquid core flowing
through the center of the channels. This flow regime is called inverted-annular
flow (Figure 3.1 (f)) and was seen in some of the tests. This flow regime is to be
avoided since it is accompanied with a sudden rise in the wall temperature and a
significant drop in the heat transfer coefficient.

31
Under some conditions, these flow patterns may alternate in a single
channel, resulting in an alternating flow. In channels with large width-to-depth
aspect ratios, two different flow patterns may also be present alongside each
other across the width of the channels. In subsequent sections, the conditions
for the occurrence of these flow patterns are discussed for different channel sizes
and different mass fluxes.

3.2.2. Effect of Channel Dimensions on Flow Patterns


The effects of channel dimensions and heat flux on the boiling flow
patterns are discussed in this section for the mass flux of 630 kg/m2s. The effect
of mass flux on flow patterns will be discussed in section 3.2.3.
In the 400 m 400 m (width depth) microchannel heat sink at a mass
flux of 630 kg/m2s, three general flow patterns are observed as the heat flux is
increased bubbly flow, alternating churn and wispy-annular flow, and
alternating churn and annular flow as illustrated in Figure 3.2. Bubbly flow
starts after the onset of boiling at q = 64.8 kW/m2, with isolated round and
elongated bubbles. As the heat flux increases the bubbles coalesce and
generate larger and longer bubbles. At this flow rate, bubbles do not span the
cross-section of the channels and slug flow does not occur. As the heat flux is
increased to 132.6 kW/m2, the flow pattern becomes intermittent, alternating
between churn flow and wispy-annular flow (Figure 3.2 (b)). Very few nucleation
sites remain in the liquid film in the wispy-annular regime. For heat fluxes of
156.5 kW/m2 and higher, annular flow replaces the wispy-annular regime in the
alternating flow as shown in Figure 3.2(c). Small, round droplets are entrained
into the vapor core. As the heat flux is further increased to 220.2 kW/m2, careful
observation of the churn flow shows that bubbles no longer nucleate at the walls
and all the vapor chunks arrive from further upstream with a very thin liquid layer
separating them from the walls. At this heat flux, the nucleate boiling-dominant
regime ends and boiling enters a convective-dominant region. Similar flow
regimes are observed in the 1000 m 400 m (width depth) microchannels.

32
In the microchannels with a larger cross-sectional area and larger aspect
ratio (1000 m 220 m, 2200 m 400 m, and 5850 m 400 m), similar
flow patterns of bubbly, churn/wispy-annular, and churn/annular occur, except
that unlike the 400 m 400 m channels, alternating churn and wispy-annular,
or annular, flow regimes do not necessarily follow each other along the length of
the channel; instead, they are distributed side by side across the width of the
channel due to the large aspect ratio.
Figure 3.3 shows the flow patterns observed in the 5850 m 400 m
microchannels at a mass flux of 630 kg/m2s. Bubbles start nucleating at the
walls at a heat flux of 45 kW/m2 and when the heat flux reaches 59.7 kW/m2,
bubbly flow similar to that in the 400 m 400 m microchannels begins. Figure
3.3(a) shows the bubbly flow at a heat flux of 77.1 kW/m2. As the heat flux
increases, some bubbles grow bigger and fill the height of the channel while
expanding laterally, as shown in Figure 3.3(b). Increasing the heat flux further to
194 kW/m2 leads to the commencement of churn/wispy-annular flow. As can be
seen in Figure 3.3(c) for a heat flux of 221.7 kW/m2, large liquid droplets exist in
the vapor core in the wispy-annular flow and the two flow regimes of churn and
wispy-annular exist side by side across the width of the channel. At the highest
heat flux of 270 kW/m2, annular flow replaces the wispy-annular regime where
small liquid droplets are entrained in the vapor core as shown in Figure 3.3(d). A
careful study of the flow visualizations indicates that bubble nucleation ceases at
the downstream end of the channels at a heat flux of 255 kW/m2. Figure 3.4
shows similar flow regimes in the 2200 m 400 m microchannels. It can be
seen in Figure 3.4(b, c) that churn flow exists on one side of the channel while
wispy-annular or annular flow occupies the other side.
Figure 3.5 shows the flow patterns in the 100 m 400 m
microchannels. In these microchannels, flow patterns are different from those
observed in the larger microchannels, due to the small cross-sectional area and
the resulting confinement effects. Also, the still images are less clear than in the
larger channels the observations described here are based on close

33
observation of the video images. At this microchannel width, bubbly flow is not
established, and instead, slug flow commences early after the incipience of
boiling. Figure 3.5(a) shows the boiling patterns right after the commencement of
boiling. As can be seen in the photograph, the vapor slugs are separated by a
few small bubbles in the bulk liquid. However, these small bubbles disappear at
a heat flux as low as 33.7 kW/m2. At a heat flux of 44.8 kW/m2, the tail of the
slugs breaks up and is no longer very distinguishable. The liquid layer between
the vapor core and the walls also breaks up at the locations in which bubbles
nucleate on the walls. As shown in Figure 3.5(b), the liquid layer breaks into
rivulets forming a discontinuous liquid layer at the walls.
As the heat flux is increased to 62 kW/m2, flow enters an annular regime.
The liquid layer surrounding the vapor core in the annular flow alternates
between a smooth annulus and a discontinuous layer as shown in Figure 3.5(c)
for the heat flux of 71.5 kW/m2. In the discontinuous annular flow, bubble
nucleation at the walls is still seen in the discontinuous liquid annulus up to a
heat flux of 82.1 kW/m2, beyond which bubble nucleation at the walls is
suppressed. Unlike the larger microchannels, no droplets are observed in the
vapor core of the annular flow. At heat fluxes higher than 92.3 kW/m2, an
alternating churn and annular flow appears. This alternating flow regime is
shown in Figure 3.5(d) for a heat flux of 102.7 kW/m2. Although the top wall of
the microchannel partially dries out at heat fluxes exceeding 62 kW/m2, the side
walls do not seem to dry out even at very high heat fluxes.
Bubble confinement is also observed in the 100 m 220 m , 400 m
100 m and 400 m 220 m (width depth) microchannels. Figure 3.6
presents flow patterns in the latter. Although bubbly flow commences at the
incipience of boiling, as the heat flux increases, some bubbles grow bigger and
fill the height of the channel while expanding laterally, resulting in bubbly/slug
flow as shown in Figure 3.6(b). At higher heat fluxes, alternating churn and
annular flow occurs (Figure 3.6(c)). Visualizations in the 100 m 100 m

34
microchannels show the occurrence of annular flow right after the incipience of
boiling, and no slug flow is observed.
In Figure 3.7, the flow patterns observed in all the microchannels
investigated are summarized for a wide range of microchannel widths, depths,
and aspect ratios for the mass flux of 630 kg/m2s. The abscissa shows the heat
flux and the bars represents flow patterns and flow regime transitions for each
microchannel size, with cross-sectional area of the channels increasing from
bottom to top. Microchannel dimensions are provided on the right side of each
bar as width (m) depth (m), with the area of a single channel cross-section
(mm2) in parentheses.
It can be seen from this figure that in the microchannels with a crosssectional area of 0.037 mm2 and smaller, only slug and annular regimes occur at
the visualization location and bubbly flow is not established. In the larger
microchannels with a cross-sectional area of up to 0.089 mm2, bubbly flow exists
at very low heat fluxes and is followed by slug flow and annular flow at higher
heat fluxes. A careful examination of the flow visualizations reveals that in the
five smallest microchannel heat sinks with cross-sectional area smaller than
0.089 mm2, bubbles occupy the whole cross-section of the channels, resulting in
bubble confinement; however, in the 250 m 400 m microchannels (0.089
mm2 area), although slug flow is detected, the elongated bubbles only occupy the
width of the channels and do not span the entire depth; hence, the bubbles are
not confined in the cross-section.
For cross-sectional areas of 0.144 mm2 and larger, bubbly flow
commences at the incipience of boiling and is followed by alternating churn and
wispy-annular flow. As the heat flux is increased, an annular flow regime
replaces the wispy-annular flow. In these microchannels, nucleate boiling at the
walls occurs up to very high heat fluxes. For aspect ratios (width/depth) of 2.6
and smaller, the wispy-annular and annular flow regimes are symmetric with
respect to the channel width, while for aspect ratios of 4.5 and larger, flow
symmetry is lost (Figure 3.3, Figure 3.4).

35
Considering the flow transition from bubbly or slug flow to churn/annular
flow in smaller microchannels and to churn/wispy-annular flow in larger channels
in Figure 3.7, it is seen that this transition occurs at much lower heat fluxes for
microchannels with a cross-sectional area smaller than 0.089 mm2, in which
confinement was also visually observed. Occurrence of slug flow and the early
transition to churn/annular flow in these smaller channels affects the heat transfer
behavior as will be discussed in the next chapter.

3.2.3. Effect of Mass Flux on Flow Patterns


The effect of mass flux on the flow patterns is discussed with respect to
the microchannels of dimensions 400 m 400 m at four mass fluxes of 225,
630, 1050, and 1420 kg/m2s. As depicted in Figure 3.2 and discussed in the
previous section, the main flow regimes for the mass flux of 630 kg/m2s are
bubbly flow, intermittent churn and wispy-annular flow, and intermittent churn and
annular flow. As the mass flux decreases or increases, these flow patterns
change, as does the heat flux at which transitions between the observed flow
regimes occur.
Figure 3.8 shows the flow patterns immediately after the incipience of
boiling at the four mass fluxes tested. It is seen that as the mass flux increases,
boiling commences at a higher heat flux. Also, the bubbly flow regime appears
differently as the mass flux increases. For the lower mass flux of 225 kg/m2s,
vapor confinement is observed and some bubbles grow larger and span the
cross-section of the channel, leading to a bubbly-slug flow regime as soon as
boiling starts, as depicted in Figure 3.8(a). As the heat flux increases at this flow
rate, these vapor slugs coalesce and generate very long vapor slugs with smaller
bubbles following them in the flow (not shown in the figure). As the mass flux
increases, bubbles become smaller but are more elongated at the incipience of
boiling. For mass fluxes of 630 kg/m2s and higher, the bubbles are relatively
small at the incipience of boiling (Figure 3.8(b-d)) and as the heat flux increases,
the size and number of bubbles and the rate of bubble coalescence all increase,

36
resulting in larger and longer bubbles; however, slug flow does not occur at any
heat flux.
As discussed in section 3.2.2, for the mass flux of 630 kg/m2s, an
alternating churn and wispy-annular flow regime arises following the bubbly flow
regime, and annular flow replaces the wispy-annular regime at higher heat
fluxes. A close observation of the annular flow pattern for mass fluxes of 225
and 630 kg/m2s shows that the liquid film thickness decreases and the vaporliquid interface becomes wavy as the mass flux or heat flux increase, due to an
increase in the vapor velocity.
For the mass flux of 225 kg/m2s, the wispy-annular regime is not seen and
the flow enters an alternating churn and confined annular flow regime after the
bubbly/slug regime. At the higher mass flux of 1050 kg/m2s, alternating churn
and wispy-annular flow follows the bubbly flow regime; however, annular flow is
not observed at this mass flux. For the largest mass flux considered of 1420
kg/m2s, bubbly flow changes to churn flow as the heat flux increases and an
intermittent flow and annular regime are not observed. A careful study of the flow
pattern transitions for the 400 m 400 m microchannels, indicates that the
transition from bubbly to alternating churn/annular flow occurs at higher heat
fluxes as the mass flux increases.
Since the transition between specific flow patterns occurs at higher heat
fluxes as the mass flux increases, different flow regimes are seen at a given heat
flux for different mass fluxes. Figure 3.9 presents the flow patterns at a heat flux
of 1452 kW/m2 for four different mass fluxes in the 400 m 400 m
microchannels. It is seen that the alternating annular flow has started for the
mass fluxes of 225 and 630 kg/m2s, while for the mass fluxes of 1050 and 1420
kg/m2s, the flow is still in the bubbly regime. Also, the visualizations reveal that
the flow has entered a convective boiling-dominant regime for the mass flux of
225 kg/m2s at this heat flux, while for the other three mass fluxes, flow is still in
the nucleate boiling regime.

37
In general, the incipience heat flux and the critical heat flux increase with
increasing mass flux. It may also be concluded that bubbly flow replaces slug
flow, and intermittent churn/wispy-annular flow replaces intermittent
churn/annular flow, as the mass flux increases.

3.2.4. Effect of Flow Pattern on Heat Transfer Coefficient


In Figure 3.10, the heat transfer coefficient and wall temperature
corresponding to the visualizations in Figure 3.2 are plotted versus the wall heat
flux for the 400 m 400 m microchannels and a mass flux of 630 kg/m2s. It
can be seen that the onset of boiling and commencement of bubbly flow is
accompanied with a sudden increase in the heat transfer coefficient and a drop in
the wall temperature. In the range of heat flux where nucleate boiling is
dominant, the change in the wall temperature is moderate relative to the single
phase region. At a heat flux of 234.1 kW/m2, partial dry-out is observed on the
side walls in the annular flow regime, leading to a drop in the heat transfer
coefficient and an increase in the wall temperature as denoted in the figure.
In Chapter 4 and Chapter 5, the effects of channel dimensions and mass
flux on heat transfer coefficient will be discussed with the aid of flow
visualizations presented in this chapter.

3.2.5. Microscale Phenomena


There has been a good deal of discussion in the literature regarding the
appropriate definition of a microchannel; however, a clear, physics-based
distinction of microchannels from conventional-sized channels has not emerged.
In general, a microchannel refers to a channel for which the heat transfer
coefficient and pressure drop deviate from the predictions from widely accepted
models for conventional-sized channels. For single-phase flow, Liu and
Garimella (2004) and Lee et al. (2005) showed that channels with hydraulic
diameters as small as 244 m (the minimum considered in the studies) still

38
exhibit heat transfer and pressure drop behavior that is well-predicted by
conventional models. With boiling present in the channels, however, the flow
phenomena differ from those in macroscale channels as the channel approaches
the bubble diameter in size. In these small channels, correlations and models
developed for larger channels no longer apply (Bertsch et al., 2008a). In this
section, a new criterion is developed for delineating microchannels from
macroscale channels based on the presence of vapor confinement.
Figure 3.11 summarizes the existing flow regimes at different
microchannel sizes and different mass fluxes considered in this work. It is seen
that in the smaller microchannels and at low mass fluxes, bubbly flow is not
established; instead, slug flow is observed for low heat fluxes. As the heat flux is
increased, an alternating churn and confined annular flow appears in these
microchannels. In the confined annular flow, the vapor core occupies the whole
cross-section of the microchannels and is separated from the walls by a thin
liquid film.
As the channel cross-sectional area or the mass flux increases, bubbly
flow is observed at low heat fluxes and confinement is not observed. At higher
heat fluxes, alternating churn and wispy-annular or annular flow occurs. In the
wispy-annular or annular flow, the vapor core does not necessarily occupy the
entire cross-section and can instead exist on only one side of the channel; in
other words, the flow is not confined by the channel walls.
The experimental flow visualizations reveal that the flow confinement
depends not only on the channel size, but also on the mass flux since the bubble
diameter varies with flow rate. The different experimental cases listed in Table
2.2 for various channel sizes and mass fluxes are categorized into two groups of
confined and unconfined flow regardless of the heat input, and are represented in
Figure 3.12 on Reynolds number and Bond number coordinates. This plot shows
that for channels of small cross-sectional area and at low mass fluxes, vapor
confinement is observed, while for larger microchannels and at high mass fluxes,
the flow is not confined. The solid line on this plot shows the transition between

39
confined and unconfined flow and is a curve fit to the transition points,
represented by
1 g ( f g )

GD 2 = 160
Bo Re =

0.5

0.5

(3.1)

Bo 0.5 Re , a parameter named the convective confinement number here,

is proportional to the mass flux, G, and the cross-sectional area, D2, and is
inversely proportional to the fluid surface tension. This new flow boiling transition
criterion recommends that for Bo 0.5 Re < 160 , vapor bubbles are confined and
the channel should be considered as a microchannel. For larger convective
confinement numbers, the flow does not experience physical confinement by the
channel walls and the channel can be considered as a conventional (macroscale)
channel. It is important to note that this transition criterion is independent of the
heat flux and is very useful in determining whether a channel behaves as a
microchannel or a conventional, macroscale channel, regardless of the heat
input, for practical applications. A comprehensive flow regime map accounting
for the heat input which determines the specific flow patterns is presented in
Chapter 6.
The proposed criterion for transition between confined and unconfined
flow is compared in Figure 3.13 with available experimental observations from
other studies in the literature for water (Peles et al., 1999; Jiang et al., 2001;
Kandlikar, 2002; Serizawa et al., 2002; Hetsroni et al., 2003; Lee et al., 2003;
Steinke and Kandlikar, 2003; Zhang et al., 2005b; Garimella et al., 2006; Wang
et al., 2008), dielectric liquids (Hetsroni et al., 2002; Mukherjee and Mudawar,
2003; Zhang et al., 2005a; Chen and Garimella, 2006a), and refrigerants
(Coleman and Garimella, 2003). Details of the fluid, geometry, mass flux, and
heat flux of the data points used in this comparison are listed in Table 3.1.
Figure 3.13 shows that the proposed criterion is successful in predicting the
confined or unconfined nature of the flow from a variety of studies in the
literature.

40
3.3. Conclusions
Flow boiling regimes obtained with the perfluorinated dielectric liquid FC77 in 11 different microchannel heat sinks with channel widths ranging from 100
m to 5850 m, and depth ranging from 100 m to 400 m, are investigated as a
function of microchannel dimensions, mass flux and heat flux.
Five boiling flow regimes bubbly, slug, churn, wispy-annular, and
annular flow and a post-dryout regime (inverted annular) are identified. Flow
visualization results show that the microchannel width, depth, or aspect ratio
individually do not determine boiling mechanisms in microchannels; instead, it is
the cross-sectional area of the microchannels that plays a determining role. At a
fixed mass flux, vapor confinement occurs for microchannels with cross-sectional
area below a threshold value.
For each microchannel size, as mass flux increases, the bubbles become
smaller and more elongated in the bubbly region, and the liquid layer thickness in
the wispy-annular and annular regimes decreases. Also, the transition between
specific flow patterns occurs at a higher heat flux for higher mass fluxes.
Based on the flow visualization results, a new transition criterion is
developed which predicts the conditions under which microscale confinement
effects are exhibited in flow boiling. This criterion depends on the value of a
parameter termed the convective confinement number in this study, Bo 0.5 Re ,
which depends not only on the channel dimensions and fluid properties, but also
on the mass flux. It is shown that for flow boiling of FC-77, physical confinement
in the microchannels exists for convective confinement numbers less than 160;
under this condition, thin-film evaporation contributes to heat transfer in addition
to nucleate boiling. For larger convective confinement numbers, vapor
confinement does not occur and nucleate boiling is dominant up to very high heat
fluxes.
A gallery of flow visualization movies obtained in the current study for all
the microchannel test pieces and mass fluxes tested is available at
https://engineering.purdue.edu/CTRC.

41

Table 3.1 Summary of the experimental data from the literature used in the
comparisons of the confinement transition criterion.

Reference

Fluid

Geometry

Wall
Mass flux
heat flux
(kg/m2s)
(kW/m2)

Peles et al.
(1999)

Water

Parallel microchannels
Dh = 157, 207 m

3500,
6000

Jiang et al.
(2001)

DI Water

Parallel grooves
w = 50 m

110-365

Kandlikar (2002)

DI Water

Parallel microchannels
w = 1 mm, d = 1 mm

40

Serizawa et al.
(2002)

DI Water

Circular tube
D = 50 m

24

Lee et al. (2003)

DI Water

Hetsroni et al.
(2003)

Water

Steinke and
Kandlikar (2003)

DI Water

Zhang et al.
(2005)

DI Water

Garimella et al.
(2006)

DI Water

Wang et al.
(2008)

DI Water

Hetsroni et al.
(2002)

Vertrel
XF

FC-72

Mukherjee and
Mudawar (2003)

Parallel microchannels
w = 120 m, d = 14 m
Parallel triangular
microchannels
Dh = 103, 129 m
Parallel microchannels
w = 214 m, d = 200 m

30-60

87-108

80-220

115-467

55-839

Single microchannel
w = 50 m, d = 40 m

160

651

160

91-787

365, 486

Parallel triangular
microchannels
Dh = 130 m

148

36

Parallel mini- and


microchannels in a gap, w
= 200 m, d = 660 m
w = 1.57 mm, d = 3.05 mm

500-1100

Parallel microchannels
w = 275 m, d = 636 m
Parallel trapezoidal
microchannels
w = 208, 427 m, d = 146
m

42

Table 3.1 Continued.

Reference

Fluid

Geometry

Wall
Mass flux
heat flux
(kg/m2s)
(kW/m2)

Zhang et al.
(2005)

FC-72

Parallel microchannels
w = 200 m, d = 2000 m

120

42

Chen and
Garimella (2006)

FC-77

Parallel microchannels
w = 389 m, d = 389 m

345

9-56

Coleman and
Garimella (2003)

R134a

Glass tube
w = 4.9 mm

150, 750

43

Flow direction

Bubbly flow

Vapor bubbles

(a)
Slug flow

Bulk liquid
Elongated vapor bubbles

(b)
Churn flow

Liquid slug

Vapor bubbles

Bulk liquid

Vapor chunks

(c)
Wispy-annular flow

Vapor core

(d)
Annular flow

Liquid droplets
Vapor core

(e)
Inverted annular flow

Thin liquid film

Vapor bubble
Thin liquid film

Liquid droplets
Liquid core

Thick vapor blanket

(f)

Figure 3.1. Description of boiling flow regimes, d = 400 m, (a) w = 400 m, G =


225 kg/m2s, q = 33 kW/m2, (b) w = 400 m, G = 225 kg/m2s, q = 40 kW/m2, (c)
w = 400 m, G = 225 kg/m2s, q = 95 kW/m2, (d) w = 400 m, G = 630 kg/m2s, q
= 131 kW/m2, (e) w = 400 m, G = 630 kg/m2s, q = 157 kW/m2, (f) w = 250 m,
G = 225 kg/m2s, q = 157 kW/m2.

44

Flow direction

Bubbly flow

(a)

q = 81 kW/m2
Churn

Wispy-annular flow

(b)

q = 132.6 kW/m2
Churn

q = 156.5 kW/m2

Annular flow

(c)

Figure 3.2. Flow patterns in the 400 m 400 m microchannels, G = 630


kg/m2s.

45

Bubbly flow

q = 77.1 kW/m2

Bubbly flow

(a)

q = 145.4 kW/m2

Churn/wispy-annular flow

(b)

q = 221.7 kW/m2

(c)

Churn/annular flow

q = 280 kW/m2

(d)

Figure 3.3. Flow patterns in the 5850 m 400 m microchannels, G = 630


kg/m2s.

46

Bubbly flow

q = 145.4 kW/m2

Churn/wispy-annular flow

(a)

q = 221.7 kW/m2

(b)

Churn/annular flow

q = 280 kW/m2

(c)

Figure 3.4. Flow patterns in the 2200 m 400 m microchannels, G = 630


kg/m2s.

47

(Bubbles)

q = 28.1 kW/m2
(Vapor core)

q = 44.8 kW/m2

(Vapor slug)

(a)

Slug flow
(Nucleation sites) (Liquid rivulets)

(b)

Slug flow

Discontinuous annular flow

Annular flow

(c)

q = 71.5 kW/m2
Churn flow

q = 102.7 kW/m2

Annular flow

(d)

Figure 3.5. Flow patterns in the 100 m 400 m microchannels, G = 630


kg/m2s.

48

Bubbly flow

(a)

q = 56.7 kW/m2
Bubbly/slug flow

(b)

q = 69.9 kW/m2
Churn

q = 190.7 kW/m2

Annular flow

(c)

Figure 3.6. Flow patterns in the 400 m 220 m microchannels, G = 630


kg/m2s.

49

Width x Depth (Area)


C/W

B
B

C/W

B/S
B

Acs

C/W

5850 x 400 (2.201)

C/A

2200 x 400 (0.815)

C/A
C/A

1000 x 400 (0.366)


1000 x 220 (0.231)
400 x 400 (0.144)

C/A

250 x 400 (0.089)

C/A

400 x 220 (0.079)

C/A

B/S

100 x 400 (0.037)

C/A

400 x 100 (0.026)

C/A

S
S

C/W

B/S
B

B/S

C/W

C/A

100 x 220 (0.021)

C/A

100 x 100 (0.009)

C/A

100

200
q" (kW/m2)

300

B: Bubbly
S: Slug
C: Churn
W: Wispy-annular
B/S: Intermittent bubbly/slug flow
C/W: Intermittent churn/wispy-annular flow
C/A: Intermittent churn/annular flow
: Single-phase flow

400
A: Annular

Figure 3.7. Summary of boiling flow patterns in the microchannel heat sinks for
G = 630 kg/m2s; the microchannel dimensions are presented as nominal width
(m) nominal depth (m) with the actual single-channel cross-sectional area
(mm2) in parentheses.

50

qONB = 33.6 kW/m2

G = 225 kg/m2s

(a)

qONB = 64.8 kW/m2

G = 630 kg/m2s

(b)

qONB = 90.6 kW/m2

G = 1050 kg/m2s

(c)

qONB = 112.1 kW/m2

G = 1420 kg/m2s

(d)

Figure 3.8. Flow patterns at the onset of boiling in the 400 m 400 m
microchannels for four mass fluxes.

51

Churn

Annular flow

(a)

G = 225 kg/m2s
Churn

Wispy-annular flow

(b)

G = 630 kg/m2s
Bubbly flow

(c)

G = 1050 kg/m2s
Bubbly flow

G = 1420 kg/m2s

(d)

Figure 3.9. Effect of mass flux on boiling flow patterns in the 400 m 400 m
microchannels, qw 145 kW/m2.

52

12
w = 400 m
G = 630 kg/m2s

140

Churn/
wispy-annular

130

ONB

Tw (C)

h (kW/m K)

10

Suppression of
nucleate boiling

120

Partial wall dry-out


4

Churn/annular
110

Tw
0

50

100
150
200
2
q"w (kW/m )

250

100

300

Figure 3.10. Variation of heat transfer coefficient and wall temperature with wall
heat flux in the 400 m 400 m microchannels, G = 630 kg/m2s.

53

Width x Depth (Area)


(m) x (m) (mm2)

1420

B
B

B/S
B

C/W
C/W
C/W

B/S
B
C/W
C/A

B
S

C/A

C/A

G (kg/m s)
630
1050

C/W
C/W
C/W

5850 x 400 (2.201)


2200 x 400 (0.815)
1000 x 400 (0.366)
400 x 400 (0.144)
250 x 400 (0.089)
100 x 400 (0.037)

B/S
C/A

5850 x 400 (2.201)


2200 x 400 (0.815)
1000 x 400 (0.366)
C/W 400 x 400 (0.144)
250 x 400 (0.089)
100 x 400 (0.037)

B
B
B
B/S
C/W

225

B/S
B B/S
S
S
S

S
S

B
B/S
B/S
B/S

B/S

B
S

C/A
C/A
C/A

C/W
C/W
C/W
C/W
C/A
C/A

C/A
C/A

C/W
C/W
C/W
C/A
C/A

C/A
C/A
C/A

C/A

100

200
q" (kW/m2 )

C/A
C/A
C/A
C/A

5850 x 400 (2.201)


2200 x 400 (0.815)
1000 x 400 (0.366)
1000 x 220 (0.231)
400 x 400 (0.144)
250 x 400 (0.089)
400 x 220 (0.079)
100 x 400 (0.037)
400 x 100 (0.026)
100 x 220 (0.021)
100 x 100 (0.009)

5850 x 400 (2.201)


2200 x 400 (0.815)
1000 x 400 (0.366)
400 x 400 (0.144)
250 x 400 (0.089)
100 x 400 (0.037)

300

B: Bubbly
S: Slug
C: Churn
W: Wispy-annular
B/S: Alternating bubbly/slug flow
C/W: Alternating churn/wispy-annular flow
C/A: Alternating churn/annular flow
: Single-phase flow

400

A: Annular

Figure 3.11. Summary of boiling flow patterns in the microchannel test pieces for
four mass fluxes; the microchannel dimensions are presented as nominal width
(m) nominal depth (m) with the actual single-channel cross-sectional area
(mm2) in parentheses.

54

104
Confined Flow
Unconfined Flow

Unconfined Flow

Re

103

10

Confined Flow

Bo 0.5 Re = 160

10
10-2

10-1

100

101

Bo

Figure 3.12. Transition from confined flow to unconfined flow.

55

105

104

Re

10

102

101

100
10-4

10-3

10-2

10-1

100

101

Flow Transition
Peles et al., 1999
Jiang et al., 2001
Kandlikar, 2002
Serizawa et al., 2002
Hetsroni et al., 2003
Lee et al., 2003
Steinke & Kandlikar, 2003
Zhang et al., 2005b
Garimella et al., 2006
Wang et al., 2008
Hetsroni et al., 2002
Mukherjee & Mudawar, 2003
Zhang et al., 2005a
Chen & Garimella, 2006a
Coleman & Garimella, 2003

102

103

Bo

Figure 3.13. Comparison of the confinement criterion with experimental data


from a variety of sources in the literature; solid symbols and open symbols
represent confined and unconfined cases, respectively.

56

CHAPTER 4. EFFECT OF CHANNEL SIZE ON BOILING IN MICROCHANNELS

Heat transfer with liquid-vapor phase change in microchannels can


support very high heat fluxes for applications such as the thermal management
of high-performance electronics. However, the effects of channel cross-sectional
dimensions on the two-phase heat transfer coefficient and pressure drop have
not been investigated extensively. One aim of the present study is to investigate
as independent parameters the channel width and depth as well as the aspect
ratio and cross-sectional area on boiling heat transfer in microchannels, based
on a large database of experimental results.
Heat transfer and pressure drop during boiling in microchannels are
studied in this chapter for microchannels of different sizes and the dependence of
heat transfer coefficient, boiling curve, and pressure drop on channel size is
discussed. Also, the important geometrical parameters that influence boiling
mechanism and heat transfer in microchannels are identified.

4.1. Literature Review


A large number of studies of microchannel flow boiling have been reported
as reviewed by Garimella and Sobhan (2003), Sobhan and Garimella (2001), and
Bertsch, et al. (2008a). However, few have explored the effect of microchannel
dimensions on the thermal transport. While a few studies have considered the
effect of microchannel size on flow boiling patterns and the transition between
different flow regimes, as reviewed in Chapter 3, the effect of microchannel size
on heat transfer coefficient and pressure drop has been largely unexplored.
Lin et al. (2001) performed an experimental study of two-phase flow of
refrigerant R141b in circular tubes of diameter of 1.8, 2.8, 3.6 mm and one

57
square tube of 2 mm 2 mm, with different mass fluxes ranging from 50 to 3500
kg/m2s. They showed that the transition from nucleate to convective boiling at
high heat flux occurred at higher qualities in the smaller diameter tubes. In the
absence of dryout in the saturation boiling region, they found that the mean heat
transfer coefficient varied only slightly with tube diameter and was mainly a
function of heat flux. Saitoh et al. (2005) experimentally investigated the boiling
heat transfer of refrigerant R-134a flow in three horizontal tubes of diameter 0.51,
1.12, and 3.1 mm and mass fluxes ranging from 150 to 450 kg/m2s. Their study
showed that the local heat transfer coefficient increased with increasing mass
flux in larger tubes but was not significantly affected by mass flux in smaller
tubes. The heat transfer coefficient increased with increasing heat flux in all
three tubes. The contribution of forced convective evaporation to the boiling heat
transfer decreased with decreasing tube diameter. Dupont and Thome (2005)
studied the effect of diameter on flow boiling heat transfer and transition from
macro- to microchannel evaporation using a three-zone flow boiling model based
on evaporation of elongated bubbles in microchannels. The model predicted an
increase in heat transfer coefficient with a decrease in diameter for low values of
vapor quality and a decrease in heat transfer coefficient for large vapor qualities.
Use of dielectric liquids in microchannel heat sinks has drawn recent
attention since the working fluid in the microchannel heat sinks can be in direct
contact with the electronics. Despite their poorer thermal properties as
compared to water, perfluorocarbons are particularly suitable for direct contact
cooling due to their high electrical resistivity. Fluorinert FC-77, a perfluorocarbon
liquid manufactured by 3M, is characterized by a specific heat and a latent heat
of vaporization of 1100 J/kg K and 89103 J/kg, respectively, with electrical
resistivity on the order of 1015 ohm-cm. Although there have been a number of
studies on pool boiling of perfluorocarbon liquids (Honda and Wei, 2004; McHale
and Garimella, 2008), investigations of flow boiling in microchannels using
perfluorocarbon liquids have been limited. Chen and Garimella (2006a)
conducted experiments to study the physics of boiling in parallel silicon

58
microchannels with a cross-section of 389 m 389 m using FC-77 as the
working fluid. Zhang et al. (2005a) performed an experimental study of flow
boiling regimes using FC-72 for three different orientations of a microchannel
heat sink: vertical up-flow, vertical down-flow, and horizontal flow. The heat sink
consisted of parallel microchannels 200 m wide and 2 mm deep. Warrier et al.
(2002) performed single-phase and boiling experiments in horizontal parallel
aluminum microchannels of hydraulic diameter 750 m with FC-84 as the test
fluid. They tested different mass flow rates and inlet subcooling conditions and
proposed correlations for both pressure drop and heat transfer.
In view of the absence of any systematic studies in the literature on the
effect of microchannel dimensions on heat transfer coefficient and pressure drop,
one of the objectives of the present work is to investigate the effect of channel
size on boiling heat transfer in microchannel heat sinks with a dielectric fluid, FC77. Boiling curves, heat transfer coefficients, and pressure drops are presented
and compared for a range of flow rates and microchannel widths, and design
criteria for microchannel heat sinks are discussed. The important geometrical
parameters that influence boiling in microchannels are identified based on the
results.

4.2. Experimental Setup and Procedures


The experimental facility explained in Chapter 2 is used to study the effect
of channel size on boiling in microchannels. Twelve test pieces, with
microchannel widths ranging from 100 m to 5850 m and depths ranging from
100 m to 400 m, are included in the experimental investigation. The aspect
ratio and hydraulic diameter of the microchannels in the different test pieces take
values from 0.27 to 15.55 and 96 m to 707 m, respectively. The width (w),
depth (d), and number (N), along with the hydraulic diameter (Dh), aspect ratio
(w/d), and single channel cross-sectional area (Acs) of the microchannels in each
test piece are provided in Table 2.1.

59
4.3. Results and Discussion
Microchannels tested in the present work span a wide range of channel
width, depth, aspect ratio, hydraulic diameter, and cross-sectional area, thus
providing results that help clarify the dominant geometrical factor that governs
boiling heat transfer in microchannels. The effects of physical confinement on
heat transfer coefficient are explained in this section with the aid of flow
visualizations of Chapter 3. The effects of channel size on boiling curve and
pressure drop are also discussed for a fixed mass flux of 630 kg/m2s; however,
similar results were observed for the other three mass fluxes tested. The results
for mass fluxes of 225, 1050, and 1420 are shown in Appendix E. The effects of
mass flux on heat transfer and pressure drop are explained in Chapter 5.

4.3.1. Critical Role of Channel Cross-Sectional Area on Heat Transfer Coefficient


Figure 4.1 illustrates the effect of microchannel dimensions on the heat
transfer coefficient for a mass flux of 630 kg/m2s. In this figure, the heat transfer
coefficient is plotted versus the wall heat flux for both single-phase and twophase flows in all the microchannels considered. As can be seen from this
figure, the onset of boiling is associated with an increase in the wall heat transfer
coefficient.
A careful examination of Figure 4.1 (and review of the dimensions listed in
Table 2.1) reveals that for microchannels with a cross-sectional area of 0.089
mm2 and larger, the heat transfer coefficient is independent of microchannel size.
For smaller cross-sectional areas where bubble confinement was visually
observed, as discussed in section 3.2.2, the heat transfer coefficient behavior is
markedly different, with the heat transfer coefficient being relatively higher at the
lower heat fluxes. As the heat flux increases, the curves cross over, resulting in
lower values of heat transfer coefficient. The largest heat transfer coefficient is
seen in the 100 m 220 m microchannels, with a cross-sectional area of
0.021 mm2, before partial dryout occurs. For the 100 m 100 m

60
microchannels, the heat transfer coefficient is relatively lower at low heat fluxes
since partial dryout occurs even at very low heat fluxes.
The larger heat transfer coefficients in the smaller microchannels are
attributed to the confinement effects caused by bubbles occupying the whole
cross-section of the microchannels due to the small cross-sectional area relative
to the bubble diameter at departure. As discussed in the section 3.2.2, flow
visualizations reveal that in all of the microchannels with a cross-sectional area
smaller than 0.089 mm2, slug flow commences soon after incipience of boiling
and flow enters the churn/annular regime at relatively low heat fluxes. This is
attributed to the high exit vapor quality for these smaller microchannels, as
illustrated in Figure 4.2. Early establishment of annular flow in microchannels of
very small diameter was also reported in other studies (Zhang et al., 2005b;
Jiang et al., 2001). As a result, bubble nucleation at the walls is not the only heat
transfer mechanism, and the evaporation of the thin liquid film at the walls in the
slug and annular flows also contributes to the heat transfer. Therefore, the value
of heat transfer coefficient is larger for these smaller microchannels at lower heat
fluxes. At high heat fluxes, a decrease in heat transfer coefficient is detected,
which is due to an early partial wall dryout in these small channels.
In microchannels with larger cross-sectional areas, nucleate boiling is the
dominant flow regime, and hence, the heat transfer coefficient is independent of
channel size. Similar trends have been reported in the literature for the
dependence of confined pool boiling on plate spacing in parallel-plate
configurations; as the plate spacing was reduced below the bubble departure
diameter, heat transfer was enhanced in the low heat flux region due to
confinement effects. As the spacing was decreased further, the heat transfer
coefficient increased until it reached a maximum, after which it deteriorated with
decreasing channel spacing (Geisler and Bar-Cohen, 2009).
To better illustrate the effects of microchannel cross-sectional area on
heat transfer coefficient, the heat transfer coefficient is plotted versus the crosssectional area for three heat fluxes on a semi-log plot in Figure 4.3. This plot

61
shows that for the lower heat fluxes of 80 and 100 kW/m2, the heat transfer
coefficient is relatively invariant with channel cross-sectional area for areas of
0.089 mm2 and larger where nucleate boiling is dominant. For smaller
microchannels, the heat transfer coefficient increases with decreasing channel
cross-sectional area due to vapor confinement and the contribution of thin liquid
film evaporation to the heat transfer. For a low heat flux of 80 kW/m2, the heat
transfer coefficient reaches a maximum for the 100 m 220 m channels and
decreases with further decreases in cross-sectional area due to early occurrence
of partial wall dryout in smaller channels. At the higher heat flux of 130 kW/m2,
the difference in the heat transfer coefficient values becomes smaller as the heat
transfer curves for smaller microchannels approach those of the larger
microchannels.
It is emphasized that the channel aspect ratio is not the determining
geometric factor affecting boiling heat transfer. For instance, the 400 m 100
m and 2200 m 400 m microchannels both have comparable aspect ratios
(~6), but the former exhibit larger heat transfer coefficients due to the smaller
cross-sectional area. Similarly, the 100 m 100 m microchannels lead to
larger heat transfer coefficients than the 400 m 400 m microchannels
although the aspect ratios of the channels in both chips are identical. It is also
clear from the results that the smallest dimension of the microchannel is not the
key factor; although all the microchannels with a common smallest dimension of
400 m have similar values of heat transfer coefficient, the 100 m 400 m,
100 m 220 m, and 100 m 100 m microchannels (all with a common
smallest dimension of 100 m) exhibit different values of heat transfer coefficient.
As mentioned before, the reason is the dependence of heat transfer coefficient
on cross-sectional area for channels with areas smaller than 0.089 mm2 due to
vapor confinement.
Figure 4.1 illustrates that increasing the cross-sectional area of the
microchannels beyond a threshold value does not affect the heat transfer
coefficient for a fixed wall heat flux, qw (Eq. (2.5)). However, from a design point

62
of view, the dependence of heat transfer coefficient on microchannel size should
be considered in terms of a given amount of heat dissipation from the chip, i.e., a
fixed value of base heat flux, qb (Eq. (2.6)). Plotted in this manner, the heat
transfer coefficient (as a function of base heat flux) for different microchannel
widths and for a fixed channel depth is presented in Figure 4.4. It is evident from
this figure that for a given heat dissipation from the chip, the heat transfer
coefficient increases as the microchannels width increases. However, the
maximum amount of heat that can be removed from the chip increases as the
microchannels become smaller due to the larger surface enhancement with the
smaller microchannels.

4.3.2. Effect of Vapor Confinement on the Heat Transfer Coefficient


The experimentally determined heat transfer coefficients for various
microchannel sizes and four mass fluxes in the range of 225 to 1420 kg/m2s are
plotted in Figure 4.5. At the mass flux of 225 kg/m2s, vapor confinement is
visually observed in microchannels with cross-sectional area of 0.258 mm2
(corresponding to the 700 m 400 m microchannels) and smaller. Figure
4.5(a) shows that the heat transfer coefficients for the microchannels with such
flow confinement are larger in the low heat flux region. As explained in section
3.2.5, for the confined flow (i.e., slug flow at low heat fluxes), evaporation of the
thin liquid film at the walls contributes to the heat transfer, and nucleate boiling is
not the only boiling mechanism. This leads to higher heat transfer coefficient
values. At higher heat fluxes, where annular flow exists in all channel sizes, heat
transfer coefficients in small microchannels approach those of larger channels in
value.
For the mass flux of 630 kg/m2s, flow visualizations reveal confinement in
microchannels smaller than 250 m 400 m. Correspondingly, Figure 4.5(b),
shows that the heat transfer coefficients in these microchannels depend on the
channel dimensions, and also take values that are higher relative to those in

63
larger microchannels, increasing with decreasing channel cross-sectional area.
For microchannels with cross-sectional area of 0.089 mm2 and larger, in which
confinement is not visually observed and nucleate boiling is dominant, the heat
transfer coefficient is independent of channel dimensions.
At a higher mass flux of 1050 kg/m2s, slug flow and vapor confinement are
seen only in the 100 m 400 m microchannel, and all other microchannels
exhibit similar values of heat transfer coefficient regardless of the channel size
(Figure 4.5(c)). For the largest mass flux tested of 1420 kg/m2s, vapor
confinement is not observed for any of the microchannels considered, and Figure
4.5(d) shows an independence of the heat transfer coefficient on channel
dimensions.
The plots in Figure 4.5 show that for the channels in which confinement is
not present and nucleate boiling is dominant up to very high heat fluxes, and for
which the convective confinement number Bo 0.5 Re is larger than 160, the heat
transfer coefficient is independent of microchannel size and all the curves
collapse on to a single curve. For microchannel dimensions and mass fluxes
which result in Bo 0.5 Re < 160 , the heat transfer coefficients are larger due to the
contribution of thin-film evaporation to the heat transfer mechanisms.

4.3.3. Boiling Curve


The effect of microchannel size on the boiling curve for a fixed mass flux
of 630 kg/m2s is shown in Figure 4.6. Both single-phase and two-phase regions
are included in these plots. In the single-phase region, wall excess temperature
(x-axis) is calculated with respect to the mean fluid temperature at the
corresponding heat flux, while in the two-phase region, the saturation
temperature of the liquid is employed. The saturation temperature takes different
values at different points on the plots depending on the local pressure. As can
be seen from these figures, the onset of boiling is associated with a sharp drop in
the wall excess temperature. With sufficiently small heat flux increments, this

64
temperature overshoot was observed with all the test pieces and at all mass
fluxes.
For microchannels of cross-sectional area 0.089 mm2 and larger, the
boiling curves cluster together beyond the onset of nucleate boiling, indicating
the dominance of nucleate boiling. As boiling starts in these microchannels, the
wall temperature shows a weak dependency on the heat flux. This is consistent
with the dominant nucleate boiling flow regime that was observed through the
flow visualizations (Chapter 3). As the heat flux increases, the wall temperature
becomes more dependent on the heat flux and the boiling curves deviate for
different channel sizes as convective boiling dominates.
For the microchannels with smaller cross-sectional areas, the wall
temperature increases with increasing wall heat flux and the boiling curves do not
collapse on to those of the larger microchannels. The strong dependence of the
wall temperature on the heat flux for these microchannels can be explained
based on the flow visualizations of the previous chapter which reveal that thinfilm evaporation and forced convection in the thin liquid film surrounding the
vapor slug or annulus, rather than nucleate boiling, are the main heat transfer
mechanisms in the smaller channels.
The effect of microchannel size on the boiling curves seen in Figure 4.6
was also observed in the experiments conducted with other mass fluxes
(Appendix E).
Further, if the base heat flux (Eq. (2.6)), instead of the wall heat flux (Eq.
(2.5)), were plotted versus wall superheat temperature as is done in Figure 4.7
for 400 m-deep microchannels, it is seen that for a fixed heat dissipation rate
from the chip, the wall temperature increases with channel width and is much
lower for the 100 m-wide microchannels. In other words, more heat can be
removed from the chip at a given wall temperature with the smaller
microchannels. It is emphasized that the number of microchannels incorporated
into the chip base area directly affects the base heat flux value; however, the
heat sinks tested in this work are not optimized in terms of maximizing the

65
number of microchannels that can be accommodated in a given footprint.
Nonetheless, results for the different channel dimensions from the present work
may be readily extrapolated for optimizing microchannel heat sinks by adjusting
the number of microchannels present (by changing the fin width).

4.3.4. Pressure Drop and Pumping Power


In Figure 4.8, the pressure drop is presented for different channel sizes at
a fixed mass flux as a function of average wall heat flux. The two-phase region
can be clearly distinguished from the single-phase region by the sharp change in
slope of the curves. In the single-phase region, the pressure drop decreases
slightly with increasing heat flux due to the reduction in liquid viscosity as the
liquid temperature increases. In the two-phase region, the pressure drop is
strongly dependent on heat flux and increases rapidly and almost linearly with
increasing heat flux due to the acceleration of vapor as well as the two-phase
frictional pressure drop. These trends in pressure drop have been reported in
many studies (Warrier et al., 2002; Qu and Mudawar, 2004; Chen and Garimella,
2006a; Pate et al., 2006; Liu and Garimella, 2007).
In both the single-phase and two-phase regions, the pressure drop
increases with decreasing microchannel cross-sectional area at a given heat flux.
In the two-phase region, the slope of the line also increases as the channel area
decreases, with much larger pressure drops for smaller channels at higher heat
fluxes.
It can also be seen that for the microchannels with similar cross-sectional
areas and different aspect ratios, pressure drops are similar in value (e.g., 250
m 400 m and 400 m 220 m microchannels).
Figure 4.9 shows the pumping power required to manage a given heat
sink base heat flux with different microchannel widths for a fixed channel depth of
400 m. In the single-phase region the pumping power required is almost
constant, independent of the heat flux, while in the two phase region, the
pumping power increases rapidly with heat flux. This figure also shows that for

66
microchannels of width 400 m and larger, the pumping power is not a strong
function of microchannel width. For widths below 400 m, however, the pumping
power increases for a given base heat flux. Therefore, for a given pumping
power, more heat can be removed from the heat source with larger
microchannels, although, as discussed in the previous section, using the larger
microchannels results in higher wall temperatures. The maximum heat that can
be removed from the chip is higher for smaller channels because of greater
surface enhancement. It can be concluded that for a fixed channel depth of 400
m, the 400 m wide microchannel heat sink requires lower pumping power than
for widths of 100 m and 250 m, while at the same time providing a larger heat
removal capability and lower wall temperature for a given base heat flux
compared to channels that are of 700 m and larger widths.

4.4. Conclusions
In this chapter, the effect of channel dimensions on microchannel flow
boiling heat transfer and pressure drop has been investigated through a
comprehensive set of experiments. Flow visualizations from Chapter 3 and heat
transfer results presented in this chapter show that the microchannel width,
depth, or aspect ratio individually do not determine boiling mechanisms and heat
transfer in microchannels; instead, it is the cross-sectional area of the
microchannels that plays a determining role.
For microchannel dimensions and mass fluxes with Bo 0.5 Re > 160 , it is
shown that confinement is not present and nucleate boiling is dominant up to
very high heat fluxes; hence, the heat transfer coefficient and boiling curves are
independent of microchannel size and all the curves collapse on to a single
curve. For microchannel dimensions and mass fluxes which result in
Bo 0.5 Re < 160 , the heat transfer coefficients are larger due to the contribution of
thin-film evaporation to the heat transfer mechanisms.

67
The pressure drop is shown to decrease slightly with heat flux in the
single-phase region and increase rapidly with heat flux in the two-phase region.
For a fixed wall heat flux, pressure drop increases with decreasing channel
cross-sectional area.
For a fixed pumping power, the base heat flux, qb , increases with
increasing width of the microchannels for a fixed channel depth, however, the
maximum heat that can be removed from the chip decreases with increasing
channel width. Also, for a given amount of heat dissipation from the chip, the
wall temperature is lower for the smaller channels.

68

9
8

G = 630 kg/m s

h (kW/m2 K)

7
6
5

5850 x 400
2200 x 400
1000 x 400
700 x 400
1000 x 220
400 x 400
250 x 400
400 x 220
100 x 400
400 x 100
100 x 220
100 x 100

4
3
2
1
0

50

100

150 200 250


q"w (kW/m2)

300

350

Figure 4.1. Effect of microchannel dimensions (m m) on local heat transfer


coefficient as a function of wall heat flux, G = 630 kg/m2s.

69

1
2

G = 630 kg/m s
0.8

xe

0.6

100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

0.4
0.2
0
0

50

100
150
200
2
q"w (kW/m )

250

300

Figure 4.2. Variation of exit vapor quality with wall heat flux for different
microchannel widths, d = 400 m, G = 630 kg/m2s.

70

6
q"w = 80 (kW/m2)
q"w = 100 (kW/m2)
q"w = 130 (kW/m2)

5.5

h (kW/m2K)

5
4.5
4
3.5
3
G = 630 kg/m2s
2.5

10-2

10-1

Acs (mm )

100

Figure 4.3. Effect of microchannel cross-sectional area on heat transfer


coefficient, with trend lines added.

71

9
2

G = 630 kg/m s

h (kW/m2K)

7
6
5
4

100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

3
2
1
0

100

200

300 400
2
q"b (kW/m )

500

600

700

Figure 4.4. Effect of microchannel width on local heat transfer coefficient as a


function of base heat flux, d = 400 m, G = 630 kg/m2s.

72

9
2

h (kW/m K)

4
3
5850 x 400
2200 x 400
1000 x 400
700 x 400
400 x 400
250 x 400
100 x 400

2
1
0

50

100
150
2
q"w (kW/m )

200

4
3

1
0

250

G = 1050 kg/m s

50

100

150 200 250


q"w (kW/m2 )

300

350

G = 1420 kg/m2s

h (kW/m K)

10

10

h (kW/m K)

5850 x 400
2200 x 400
1000 x 400
700 x 400
1000 x 220
400 x 400
250 x 400
400 x 220
100 x 400
400 x 100
100 x 220
100 x 100

11

5850 x 400
2200 x 400
1000 x 400
700 x 400
400 x 400
250 x 400
100 x 400

4
3
2
1
0

G = 630 kg/m2s

h (kW/m K)

G = 225 kg/m s

100

200
q"w (kW/m2)

300

6
5
2200 x 400
1000 x 400

4
3

700 x 400
400 x 400

250 x 400
100 x 400

400

100

200
q"w (kW/m2)

300

400

Figure 4.5. Effects of physical confinement on heat transfer coefficients for four
mass fluxes; the microchannel dimensions are presented as width (m) depth
(m).

73

300
2

G = 630 kg/m s

q"w (kW/m2)

250
200

5850 x 400
2200 x 400
1000 x 400
700 x 400
1000 x 220
400 x 400
250 x 400
400 x 220
100 x 400
400 x 100
100 x 220
100 x 100

150
100
50
0

10

20

30
40
Tw-Tref(K)

50

60

Figure 4.6. Effect of microchannel dimensions (m m) on boiling curves, G =


630 kg/m2s.

74

700
100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

600

q"b (kW/m )

500
400
300

G = 630 kg/m s

200
100
0

10

20
T w-Tref (C)

30

40

Figure 4.7. Variation of base heat flux with wall excess temperature for different
microchannel widths, d = 400 m, G = 630 kg/m2s.

75

80
70

G = 630 kg/m s
5850 x 400
2200 x 400
1000 x 400
700 x 400
1000 x 220
400 x 400
250 x 400
400 x 220
100 x 400
400 x 100
100 x 220
100 x 100

p (kPa)

60
50
40
30
20
10
0

50

100

150 200 250


q"w (kW/m2)

300

350

Figure 4.8. Effect of microchannel dimensions on pressure drop, G = 630


kg/m2s.

76

0.025
100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

0.02

P (W)

0.015

G = 630 kg/m s

0.01

0.005

100

200

300 400
2
q"b (kW/m )

500

600

700

Figure 4.9. Effect of microchannel width on pumping power, d = 400 m, G =


630 kg/m2s.

77

CHAPTER 5. EFFECT OF FLOW RATE ON BOILING IN MICROCHANNELS

The effect of mass flow rate on the boiling heat transfer, boiling curves,
and pressure drop is discussed in this chapter for four mass fluxes ranging from
225 to 1420 kg/m2s. The microchannels with cross-sections of 400 m 400 m
and 2200 m 400 m are selected here for a detailed explanation of variations
with mass flux. Results for the effect of mass flux on boiling heat transfer for
other microchannel heat sinks listed in Table 2.1, are shown in Appendix F.

5.1. Literature Review


The effect of flow rate on microchannel flow boiling has been considered
in a number of studies. However, few have arrived at clear conclusions on the
effect of mass flux on boiling heat transfer and pressure drop. Chen and
Garimella (2006a) compared the boiling heat transfer coefficients and pressure
drops for flow rates of 260, 349, and 445 kg/m2s using fluorinert FC-77. Liu and
Garimella (2007) carried out an experimental study of flow boiling heat transfer in
microchannels (275 m 636 m) cut into a copper block, using DI water with
mass fluxes ranging from 341 to 936 kg/m2s. Chen and Garimella (2006c)
performed an experimental study of flow boiling of FC-77 in a copper
microchannel heat sink (10 channels, 405 m 2.5 mm) at flow rates ranging
from 80 to 133 kg/m2s. Pate et al. (2006) explored two-phase heat transfer in
parallel microchannels with hydraulic diameter of 253 m using FC-72 as the test
fluid. Six re-entrant type cavities, spaced evenly on the base of each channel,
were used to promote controlled nucleation. Mass fluxes in the range of 5352138 kg/m2s were considered. All these studies showed that beyond the onset
of nucleate boiling, the boiling curves collapsed on a single curve for all mass

78
flow rates; as the heat flux was increased further, the curves diverged and
became dependent on mass flux. Lin et al. (2001) observed both nucleate and
convective boiling mechanisms to occur in the tubes of diameter ranging from 1.8
to 3.6 mm and the local heat transfer coefficient was found to be a weak function
of mass flux while the mean heat transfer coefficient was independent of mass
flux. Bertsch et al. (2008b) experimentally studied the flow boiling of refrigerant
R134a in parallel microchannels of hydraulic diameter 1.09 mm for four mass
velocities ranging from 20.3 to 81 kg/m2s. They showed that the heat transfer
coefficient is a strong function of inlet vapor quality and mass flow rate.
This review of the literature highlights the limited understanding of the
effect of mass flow rate on flow boiling heat transfer and pressure drop in
microchannels. The objective of the present work is therefore to study the effect
of mass flux on boiling heat transfer and pressure drop with a dielectric fluid, FC77, for a wide range of channel sizes.

5.2. Experimental Setup and Procedures


The same experimental facility as explained in Chapter 2 is used to study
the effect of mass flux on boiling in microchannels. Four mass flux values of 225,
630, 1050, and 1420 kg/m2s are investigated for seven of the 12 test pieces
listed in Table 2.1 to map the effect of flow velocity on boiling heat transfer and
pressure drop over a wide range of microchannel dimensions.

5.3. Results and Discussion


The effect of mass flux on microscale boiling heat transfer and pressure
drop is discussed in this chapter for the 400 m 400 m and 2200 m 400
m microchannels. Similar results were obtained for the other microchannel
sizes and the complete set of results is provided in Appendix F.

79
5.3.1. Heat Transfer Coefficient
Figure 5.1 illustrates the effect of mass flux on the heat transfer coefficient
as a function of the wall heat flux for the 400 m 400 m microchannels. The
heat transfer coefficient increases with mass flux in the single-phase region for a
fixed wall heat flux. After the onset of nucleate boiling, however, the heat
transfer coefficient becomes independent of mass flux, and increases with heat
flux. At high levels of wall heat flux, as the contribution from convective heat
transfer begins to dominate that of nucleate boiling, the heat transfer coefficient
becomes a function of mass flux and increases with increasing mass flux. Flow
visualizations performed in Chapter 3 show that the plots of heat transfer
coefficient diverge from each other at the heat flux where the bubble nucleation
was observed to be suppressed at the walls. Other microchannel sizes tested
yielded similar trends for the dependence of heat transfer coefficient on mass
flux. These results regarding the dependence of heat transfer coefficient on flow
rate are also consistent with the findings of Chen and Garimella (2006a, 2006c).
At high heat fluxes, a decrease in heat transfer coefficient is detected for
lower mass fluxes. Flow visualizations discussed in Chapter 3 reveal that this is
attributed to a partial wall dryout as also reported in Chen and Garimella (2006a).

5.3.2. Boiling Curve


Figure 5.2 and Figure 5.3 show the boiling curves at four different mass
fluxes for the 400 m 400 m and 2200 m 400 m microchannel test
pieces, respectively. Similar to the boiling curves in Chapter 4, both single-phase
and two-phase regions are included in these plots and the wall excess
temperature (x-axis) is calculated with respect to the mean fluid temperature and
the saturation temperature of the liquid in the single-phase and two-phase
regions, respectively. These two figures show that the onset of nucleate boiling
occurs at a lower temperature and a lower heat flux for a lower mass flux. Also,
the critical heat flux increases with increasing mass flux.

80
For microchannels for which the convective confinement number
Bo 0.5 Re is larger than 160, beyond the onset of nucleate boiling, curves for all

mass fluxes collapse to a single curve, indicating the dominance of nucleate


boiling. This is shown in Figure 5.3 for microchannels of 2200 m 400 m (the
full set of results of boiling curves for all the microchannel dimensions is included
in Appendix F). At higher heat fluxes, as the convective heat transfer starts to
dominate, the curves diverge and, as in the single-phase region, more heat is
dissipated as mass flux increases at a fixed wall temperature. These trends are
also consistent with results in the literature (Liu and Garimella, 2007; Chen and
Garimella, 2006c; Pate et al., 2006). However, the boiling curve for the mass flux
of 225 kg/m2s deviates from those for higher mass fluxes as shown in Figure 5.2
for the 400 m 400 m microchannels. This distinct behavior of the boiling
curve at the lowest mass flux in smaller microchannel sizes is due to an early
transition to slug flow and annular flow regimes since the confinement number
Bo 0.5 Re is smaller than 160, in contrast to the behavior at higher mass fluxes

and in larger channels.

5.3.3. Pressure Drop


The pressure drop across the microchannels is plotted as a function of the
average wall heat flux in Figure 5.4 for the 400 m 400 m microchannels at
four different mass fluxes. In both single-phase and two-phase regions, the
pressure drop increases with increasing mass flux, which agrees with the results
obtained by Pate et al. (2006). Chen and Garimella (2006a), however, found that
the pressure drop was independent of mass flux in the two-phase region. They
related this observation to the balance between the frictional pressure drop and
accelerational pressure drop with the moderate inlet subcooling considered in
their tests (26C), which is in contrast to the very modest subcooling used in the
current work (5 C).

81
5.4. Conclusions
The effect of mass flux on flow boiling heat transfer and pressure drop in
microchannel heat sinks was discussed in this chapter.
For a fixed channel size, boiling curves and heat transfer coefficients are
independent of mass flux in the nucleate boiling region. As the heat flux is
increased further and convective heat transfer becomes dominant, the boiling
curves diverge and more heat is dissipated as the mass flux increases for the
same wall temperatures. The maximum heat transfer coefficient, as well as the
wall heat flux at which this maximum occurs, increase with increasing mass flux.
Pressure drop increases as the heat flux increases in the boiling region.
For a fixed wall heat flux, pressure drop increases with increasing mass flux.

82

10
9

w = 400 m

8
h (kW/m2K)

7
6
5
4

225 kg/m s
2
630 kg/m s
1050 kg/m2s

1420 kg/m s

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure 5.1. Effect of mass flux on local heat transfer coefficient for 400 m 400
m microchannels.

83

350
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

300

1420 kg/m s

q"w (kW/m )

250
200

w = 400 m

150
100
50
0

10

20
T w-Tref (C)

30

40

Figure 5.2. Effect of mass flux on boiling curves for 400 m 400 m
microchannels.

84

400
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

350
300
2

q"w (kW/m )

1420 kg/m s

250

w = 2200 m

200
150
100
50
0

10

20
T w-Tref (C)

30

40

Figure 5.3. Effect of mass flux on boiling curves for 2200 m 400 m
microchannels.

85

12
2

1420 kg/m s

p (kPa)

10

225 kg/m s
2
630 kg/m s
1050 kg/m2s
2

w = 400 m

6
4
2
0

50

100
150
200
2
q"w (kW/m )

250

300

Figure 5.4. Effect of mass flux on pressure drop for 400 m 400 m
microchannels.

86

CHAPTER 6. FLOW REGIME MAPS FOR BOILING IN MICROCHANNELS

Given the complex effects observed on the flow patterns with changes in
channel width, heat flux and mass flux, flow regime maps are essential for the
practical design of microchannel heat sinks that rely on two-phase heat transport.
In this chapter, two types of flow regime maps with coordinates commonly used
for macroscale boiling are developed. However, it is shown that the flow regime
maps represented in these coordinate systems depend on channel dimensions,
and therefore, individual maps need to be developed for each channel size.
Due to the dependence of flow regimes (and regime maps) on channel
dimensions, it is important to include the effects of channel size in the flow
regime maps. Therefore, two types of comprehensive flow regime maps for
microchannel flow boiling are developed in the present study for a wide range of
experimental conditions and microchannel dimensions; quantitative transition
criteria based on nondimensional parameters are also proposed. The capability
of these comprehensive flow regime maps in predicting flow patterns and their
transitions is validated by a comparison to experimental data from a variety of
studies in the literature.

6.1. Literature Review


Flow regime maps are commonly used to determine the flow patterns that
exist under different operating conditions, as well as the conditions for flow
pattern transitions. Such maps are essential to the development of flow regimebased models for the prediction of the heat transfer rate and pressure drop in
flow boiling. The coordinates used to plot these flow regime maps can be
superficial phase velocities or derived parameters containing these velocities;

87
however, the effects of important parameters such as channel size are not
represented in a number of these maps. Baker (1954), Hewitt and Roberts
(1969), and Taitel and Dukler (1976) developed early flow regime maps for
horizontal and vertical two-phase flow in channels with diameters of a few
centimeters.
In recent years, a number of studies have investigated the flow patterns
during boiling in microchannels employing high-speed visualizations as reviewed
in Chapter 3. Only a few of these studies, however, have utilized the observed
flow patterns to develop flow regime maps for boiling in microchannels. Hetsroni
et al. (2003) proposed a flow regime map for transition from a low heat flux
region, characterized by the presence of liquid phase in all microchannels, to a
high heat flux region, characterized by a periodic dryout phenomenon, for boiling
of water in parallel triangular microchannels. Huo et al. (2004) also proposed a
flow regime map for R134a in a small tube. Two regime maps developed for
larger tubes by Taitel (1990) and Mishima and Ishii (1984) were compared to
their map for the small tube and were shown to be inapplicable for predicting flow
regime transitions. Revellin et al. (2006) constructed flow regime maps for
R134a in a 500 m-diameter tube in terms of mass flux versus vapor quality and
superficial liquid velocity versus superficial vapor velocity. Their map included
bubbly, slug, semi-annular, and annular flow regions. They compared their flow
pattern transitions to those of macroscale refrigerant flow and microscale airwater flow, neither of which matched their results well. Revellin and Thome
(2007) and Ong and Thome (2009) proposed flow regime maps for three
refrigerants and three channel dimensions using the same coordinates as used
in Revellin et al. (2006), while classifying flow into three types of isolated bubble,
coalescing bubble, and annular regime.
Flow regime maps for adiabatic two-phase flow in microchannels have
also been proposed through high-speed visualizations (Chung and Kawaji, 2004;
Hassan et al., 2005; Field and Hrnjak, 2007); however, it has been shown
(Revellin et al., 2006) that adiabatic flow regime maps are not suitable for the

88
prediction of microscale boiling. Field and Hrnjak (2007) showed that the flow
maps developed for large channels were not suitable for prediction of the flow
regimes in microchannels; also, flow maps were dependent on the specific fluid
for which they were developed.
Despite the inability of macroscale boiling maps or adiabatic two-phase
flow regime maps to predict the flow patterns for boiling in microchannels, a
review of the literature shows a dearth of investigations into flow regime maps
specifically targeted at microchannels undergoing flow boiling that are applicable
to a wide range of microchannel dimensions and experimental conditions.

6.2. Results and Discussion


In this section, two types of flow regime maps are developed based on the
coordinate systems common in the literature, and the effect of channel width on
the transitions between flow regimes is investigated. Also, a need for
comprehensive flow regime maps is noted and new comprehensive flow regime
maps for microchannel flow boiling are proposed which are capable of flow
regime prediction for a wide range of experimental parameters and channel
dimensions.

6.2.1. Flow Regime Maps Based on Conventional Approaches


As discussed above, existing flow regime maps for boiling in
microchannels are limited to a narrow range of channel sizes and have been
developed with water and refrigerants. Flow regime maps are developed in this
section for boiling of FC-77 in microchannel heat sinks for a wide range of
channel dimensions, and the effect of channel size on the transitions between
flow regimes is investigated.
Two different types of flow regime maps are developed in this section
based on the observed flow patterns. In the first, mass flux is plotted versus
vapor quality, and in the second, liquid and vapor superficial velocities are

89
chosen as the coordinates, as is common in the literature. While in developing
the mass flux-quality flow regime map, the required quantities are measured
directly from the experiments, superficial velocities for the second type of flow
map are obtained by assuming that each phase occupies the entire channel
cross-section and are calculated from j f = (1 x) G / f and jg = x G / g for the
liquid and vapor phases, respectively. In these equations, G and x are mass flux
and vapor quality, and f and g are the liquid and vapor densities,
respectively.
Figure 6.1 shows the mass flux versus vapor quality flow regime maps for
six microchannel widths for a fixed channel depth of 400 m. The experimental
data points are represented by different symbols for different flow patterns and
the transitions between the flow regimes are presented by lines.
In general it is seen that transition to intermittent churn and wispy-annular
or annular flow occurs at a lower vapor quality as the mass flux increases.
Figure 6.1 indicates that the flow patterns, and hence the flow maps, are different
for boiling in microchannels of different sizes. For instance, churn/wispy-annular
flow does not occur in the smaller microchannels. It also appears that slug flow
tends to exist in smaller microchannels with lower mass fluxes and is replaced by
bubbly flow as the microchannel size or mass flux is increased. For the 250 m
400 m and 1000 m 400 m microchannels, the input heat flux exceeded
the critical heat flux and inverted annular flow, which is a post-dryout regime
(Figure 3.1(f)), was observed; transition from the intermittent churn and annular
flow to the post-dryout regime is shown in the corresponding flow maps. As the
mass flux or channel width increases, the transition to post-dryout occurs at a
lower vapor quality.
Figure 6.2 shows the flow maps for different heat sinks using the
superficial velocities as the coordinates. As the liquid superficial velocity
increases, transition to annular flow occurs at higher vapor superficial velocity.
The same trend is seen for transition to the wispy-annular regime in the 400 m
400 m and 1000 m 400 m microchannels; however, in the larger

90
microchannels, this transition occurs at lower vapor superficial velocity as the
liquid superficial velocity increases.
As seen in this section, the flow regime maps represented in these
coordinate systems depend on channel dimensions, and therefore, individual
maps are developed for each channel size. The effect of channel size on the
transitions between different flow regimes is investigated for these maps in the
next section.

6.2.2. Effect of Channel Width on Flow Regime Transitions


In order to readily illustrate the effect of channel size on the transitions
between different flow regimes, the lines representing transition from bubbly or
slug flows to intermittent churn/wispy-annular or churn/annular flows are plotted
for the six microchannel widths in Figure 6.3 and Figure 6.4 for the two types of
flow regime maps. It is seen that as the microchannels become smaller, this
transition occurs at a larger vapor quality or a larger vapor superficial velocity.
Huo et al. (2004) also found that reducing the tube diameter shifts the transition
of slug to churn and churn to annular flow to higher values of gas velocity.
Transition from bubbly flow to intermittent bubbly/slug flow is shown in
Figure 6.5 and Figure 6.6. Since bubbly flow does not occur in the 100 m-wide
microchannels and slug flow is not seen in the 5850 m-wide microchannels, as
discussed in section 3.2.3, the transition lines in these figures are plotted for the
microchannels of width 250 m, 400 m, 1000 m, and 2200 m only, and it is
seen that the flow transitions differ for different microchannel widths.
The dependence of the flow transition on channel size makes the flow
maps in the literature limited to the test conditions and channel dimensions used.
In Figure 6.3 and Figure 6.4, these flow transitions are also compared to those of
Mandhane et al.(1974) and Taitel and Dukler (1976) for adiabatic two-phase flow
in large horizontal channels, Coleman and Garimella (2003) for condensation of
R134a in 4.9 mm tubes, and Revellin et al. (2006) for boiling of R134a in round
tubes of diameter 500 m. It can be seen that only the transition line of Revellin

91
et al. matches the current experimental results well and that of Coleman and
Garimella only predicts the trend. The transition line of Revellin et al. also
matches the current experimental results for the transition from bubbly to
bubbly/slug flow on the mass flux-vapor quality coordinates (Figure 6.5 and
Figure 6.6). It can be concluded that in general the flow maps developed for
boiling or condensation in conventional sized channels or for adiabatic two-phase
flow are not suitable for predicting the boiling in microchannels as also discussed
in the literature (Hetsroni et al., 2003; Huo et al., 2004; Revellin et al., 2006; Field
and Hrnjak, 2007).

6.2.3. Comprehensive Flow Regime Map


Due to dependence of flow regimes (and regime maps) on channel
dimensions, it is important to include the effects of channel size in the flow
regime maps. To address this need, new comprehensive flow regime maps are
proposed in this section.
Figure 6.7 shows the comprehensive flow regime map developed based
on the experimental results and flow visualizations performed with FC-77. The
abscissa in this plot is the convective confinement number, Bo 0.5 Re , which is
proportional to G D 2 . The ordinate is a nondimensional form of the heat flux,
Bl Re , which is proportional to qw D . Plotting all the ~390 experimental data

points obtained for 12 different microchannel test pieces, four mass fluxes, and
heat fluxes in the range of 25 to 380 kW/m2 on Bl Re versus Bo 0.5 Re
logarithmic axes leads to a comprehensive flow regime map with four distinct
regions of confined slug flow, churn/confined annular flow, bubbly flow, and
churn/annular/wispy-annular flow.
The vertical transition line is given by Bo 0.5 Re = 160 , which represents the
transition to confined flow. The other transition line is a curve fit to the points of
transition from bubbly or slug flow to alternating churn/annular or churn/wispyannular flow, given by

92
Bl Re = 0.017 ( Bo 0.5 Re )

0.7

(6.1)

which can be rearranged to give


Bl = 0.017 ( Bo 0.4 Re 0.3 )

(6.2)

This flow regime map shows that for Bo 0.5 Re < 160 vapor confinement is
observed in both slug and churn/annular flow regimes while for Bo 0.5 Re > 160 ,
the flow is not confined. For low heat fluxes with Bl < 0.017 ( Bo 0.4 Re 0.3 ) , flow
patterns of slug (if Bo 0.5 Re < 160 ) or bubbly (if Bo 0.5 Re > 160 ) flow exist in the
microchannels. At higher heat fluxes with Bl > 0.017 ( Bo 0.4 Re 0.3 ) , vapor
bubbles coalesce resulting in a continuous vapor core in the alternating
churn/annular or churn/wispy annular flow regimes.
In Figure 6.8 experimental data from a range of other studies in the
literature are plotted on this comprehensive flow regime map. The map
developed here is clearly able to represent the flow regimes found in the
literature for water and fluorocarbon liquids.
It is noted that the flow regime map in Figure 6.7 is developed for flow
regimes occurring at a specific location along the length of the microchannel heat
sink where the heat transfer measurements are obtained (1.27 mm short of the
exit of the central channel). In order to include the effect of the heated length of
the microchannels on two-phase flow development, this flow regime map is
modified by using the phase change number as the ordinate in the map. The
phase change number was first introduced by Saha et al. (1976) to represent the
rate of phase change due to heat addition and is defined as
N pch =

where =

(6.3)

qw PH f g
L
L
is the frequency of vapor generation and = =
Acs h fg f g
u G / f

is the fluid particle residence time. Hence the phase change number can be
rewritten as

93

N pch =

where DhH =

qw PH f g L
L f g
= Bl
Acs h fg f g G / f
DhH
g

(6.4)

Acs
is the hydraulic diameter based on the heated perimeter of a
PH

microchannel. Using the phase change number, N pch , and the convective
confinement number, Bo 0.5 Re , as the coordinates, the flow regime map in
Figure 6.9 is obtained. Similar to the map in Figure 6.7, the transition lines divide
the map into four distinct quadrants of slug and confined annular flow for
Bo 0.5 Re < 160 and bubbly and alternating churn/annular and churn/wispy-

annular for larger convective confinement numbers.


The vertical transition line on the map, given by Bo 0.5 Re = 160 ,
represents the transition from confined flow to unconfined flow. The other
transition line is a curve fit to the points of transition from bubbly or slug flow to
alternating churn/annular or churn/wispy annular flow and is given by
N pch = 96.65 ( Bo 0.5 Re )

0.258

(6.5)

Substituting Eq. (6.4) in Eq. (6.5), the location along the microchannels at which
the transition from bubbly or slug to annular flow occurs can be determined from
La 0 = 96.65 ( Bo 0.5 Re )

0.258

Bl 1

Acs
f g PH

(6.6)

La 0 will be used in modeling of the heat transfer coefficient and pressure drop in

the annular region, as will be discussed in Chapter 8.

6.3. Conclusions
This review of the literature shows that existing flow regime maps for
boiling in microchannels are limited to narrow ranges of channel sizes and have
been developed for water and refrigerants. In the present work, two types of flow
regime maps with the coordinates conventionally used for flow boiling are
developed on mass flux-vapor quality and superficial velocity coordinates; it is

94
shown that these types of flow regime map depend on microchannel size, hence
for each channel dimension, a separate flow regime map is developed to capture
the flow regime transitions accurately. In these flow regime maps, it is seen that
as the microchannels become smaller, the transition from bubbly or slug flow to
intermittent churn/wispy-annular or churn/annular flow occurs at a larger vapor
quality or a larger vapor superficial velocity. Comparison of flow transitions to
those from the literature shows that flow regime maps developed for large
channels or for adiabatic two-phase flow are not appropriate for predicting boiling
regimes in microchannels.
To include the effects of channel size, a new type of comprehensive flow
regime map for microchannel flow boiling is developed for FC-77 for a wide
range of experimental parameters and channel dimensions with channel crosssectional area in the range of 0.009-2.201 mm2, mass flux in the range of 2251420 kg/m2s, and heat flux in the range of 25-380 kW/m2, with quantitative
transition criteria based on nondimensional parameters. For this comprehensive
flow regime map, the convective confinement number, Bo 0.5 Re , and a
nondimensional form of heat flux, Bl Re , are used as the abscissa, respectively.
Using these coordinates, the flow regime map reveals four distinct regions of
confined slug, bubbly, churn/confined annular, and churn/annular/wispy-annular
flow regimes with two transition lines. One transition line, Bo 0.5 Re = 160 ,
represents the transition to confined flow, while the other transition line,
Bl = 0.017 ( Bo 0.4 Re 0.3 ) ,which includes the effects of heat flux, illustrates the

transition from bubbly/confined slug flow to alternating churn/annular/wispyannular flow.


The proposed comprehensive flow regime map developed is shown to
capture trends observed in the experimental data for water and fluorocarbon
liquids from the literature as well.
A modification to the comprehensive flow regime map is also made, to
include the effect of the heated length of the microchannels on two-phase flow

95
development. The phase change number is used as the ordinate in the modified
comprehensive map, resulting in the same four distinct flow regions on the map
with Bo 0.5 Re = 160 and N pch = 96.65 ( Bo 0.5 Re )

0.258

as transition lines. This

comprehensive flow map facilitates the development of flow regime-based


models for the prediction of boiling heat transfer coefficients and pressure drop
as will be discussed in Chapter 8.

96

1500

1500
Bubbly

w = 100 m

G (kg/m2s)

1000 Bubbly/
Slug

G (kg/m2s)

1000
Churn/Annular

Slug

500

Post-Dryout
500

0.2

0.4

0.6

0.8

Slug

Churn/
Annular

0.2

0.4

0.8

1500
Churn/
Wispy-Annular

w = 1000 m

w = 400 m
1000

Bubbly

G (kg/m2s)

1000
G (kg/m2s)

0.6
x

1500

500

w = 250 m

Churn/
Annular

Bubbly/
Slug

Churn/
Wispy-Annular
Churn/Annular

500

Bubbly

Post-dryout
Bubbly/Slug

0.2

0.4

0.6

0.8

0.2

0.4

x
1500

0.8

1500
w = 2200 m

Bubbly

w = 5850 m

Churn/
Wispy-Annular

Churn/Annular

500

Bubbly/
Slug
0
0.2

Churn/
Wispy-Annular

1000
G (kg/m2s)

G (kg/m2s)

1000

0.6
x

500

Churn/Annular

Bubbly
0.4

0.6

0.8

0.2

0.4

0.6

0.8

500

Transition from Bubbly to Bubbly/Slug or Slug


Transition from Bubbly/Slug to Slug
Transition to Churn/Wispy-Annular

Transition to Churn/Annular
0.2
0.4
0.6
0.8
Transition to Inverted Annular
x

Figure 6.1. Flow regime maps on mass flux-vapor quality coordinates with
transition lines for six microchannels widths, d = 400 m.

97

100

Bubbly/Slug

100

Bubbly
Slug

10

Churn/
Annular

jf (m/s)

jf (m/s)

Slug
-1

10

Churn/
Annular

-1

PostDryout

10-2

w = 100 m

10-1

10-2

100

101

w = 250 m

10-1

100

jg (m/s)

10

Churn/
Wispy-Annular

101

jg (m/s)

10

Bubbly

10

-1

Bubbly/Slug

jf (m/s)

jf (m/s)

Bubbly
Churn/
Annular

10

-1

Churn/
Annular

Bubbly/Slug
Churn/
Wispy-Annular

10-2

w = 400 m

10-1

10-2

100

w = 1000 m

10-1

101

100

Churn/
Annular

-1

10-2

Churn/
Wispy-Annular

Bubbly

Bubbly/Slug

w = 2200 m

10-1

Bubbly
10

101

Churn/
Annular

-1

10-2

100

Churn/
Wispy-Annular

100

jf (m/s)

jf (m/s)

10

101
jg (m/s)

jg (m/s)

10

PostDryout

w = 5850 m

10-1

100

101
jg (m/s)

jg (m/s)

500

Transition from Bubbly to Bubbly/Slug or Slug

Transition from Bubbly/Slug to Slug


Transition to Churn/Wispy-Annular

0
0

Transition to Churn/Annular
0.2
0.4
0.6
0.8
Transition to Inverted Annular
x

Figure 6.2. Flow regime maps on superficial velocity coordinates with transition
lines for six microchannels widths, d = 400 m.

98

1500
w = 100 m
w = 250 m
w = 400 m
w = 1000 m
w = 2200 m
w = 5850 m
Revellin et al. (2006)
Coleman & Garimella
(2003)

G (kg/m2s)

1000

500

0.2

0.4

0.6

0.8

x
Figure 6.3. Effect of channel size on transition from bubbly or slug flow to
intermittent churn/wispy-annular or churn/annular flow on the mass flux-vapor
quality coordinate, d = 400 m. Some relevant transition lines from the literature
are also included.

99

jf (m/s)

10

10

w = 100 m
w = 250 m
w = 400 m
w = 1000 m
w = 2200 m
w = 5850 m
Mandhane et al. (1974)
Taitel & Dukler (1976)
Revellin et al. (2006)

-1

10-2
10

-1

10

10

jg (m/s)
Figure 6.4. Effect of channel size on transition from bubbly or slug flow to
intermittent churn/wispy-annular or churn/annular flow on the superficial velocity
coordinates, d = 400 m. Some relevant transition lines from the literature are
also included.

100

1500
w = 250 m
w = 400 m
w = 1000 m
w = 2200 m
Revellin
al. (2006)
w
= 5850etm

G (kg/m s)

1000

500

0.2

0.4

0.6

0.8

x
Figure 6.5. Effect of channel size on transition from bubbly flow to intermittent
bubbly/slug flow on the mass flux-vapor quality coordinate, d = 400 m. Some
relevant transition lines from the literature are also included.

101

jf (m/s)

100

10

10

-1

w = 250 m
w = 400 m
w = 1000 m
w = 2200 m
Revellin
al. (2006)
w
= 5850etm

-2

10

-1

10

10

jg (m/s)
Figure 6.6. Effect of channel size on transition from bubbly flow to intermittent
bubbly/slug flow on the superficial velocity coordinates, d = 400 m. Some
relevant transition lines from the literature are also included.

102

Confined Slug
Churn/Confined Annular
Bubbly
Churn/Wispy-Annular
Churn/Annular

-2

Bl.Re

101
10

-3

100

Bl = 0.017 ( Bo 0.4 Re -0.3 )


Bo 0.5 Re = 160

-4

10
10-1
10

10

0.5

10

(Bo) .Re

10

Figure 6.7. Comprehensive flow regime map for FC-77.

103

102-1

Churn/Annular/
Wispy-Annular

101-2
Churn/
Confined Annular

Bubbly

Bl.Re

-3
10
100

Flow Transition
Hetsroni et al., 2003 - CB
Hetsroni et al., 2003 - CA
Steinke & Kandlikar, 2003 - CB/S
Steinke & Kandlikar, 2003 - CA
Garimella et al., 2006 - B
Wang et al., 2008 - CB/S
Wang et al., 2008 - CA
Hetsroni et al., 2002 - CB
Zhang et al., 2005a - CB/S
Chen & Garimella, 2006a - CB/S
Chen & Garimella, 2006a - C/CA

-4
10
10-1

-5
10
10-2

Confined Slug

-6
10
10-3
0
10

10

10

10

0.5

10

10

10

(Bo) .Re
B: Bubbly C: Churn
CB: Confined Bubbly

CA: Confined Annular


S: Slug

Figure 6.8. Comparison of the comprehensive flow regime map with the
experimental data from the literature.

104

10

10

N Zu
N
pch

N pch

Bo 0.5 Re = 160
0.258
= 96.65 ( Bo 0.5 Re )

Confined Slug
Churn/Confined Annular
Bubbly
Churn/Wispy-Annular
Churn/Annular
0

10
1
10

10

0.5

10

10

(Bo) .Re

Figure 6.9. Comprehensive flow regime map for FC-77; modified with the phase
change number.

105

CHAPTER 7. PREDICTION OF HEAT TRANSFER COEFFICIENT AND


PRESSURE DROP WITH EMPIRICAL CORRELATIONS

The ability to predict heat transfer coefficient and pressure drop is required
for the design of two-phase microchannel devices. Empirical correlations have
been developed in the literature for prediction of heat transfer coefficient and
pressure drop in these systems. In this chapter, the experimental results for the
heat transfer coefficient and pressure drop are compared to predictions from a
range of correlations from the literature.

7.1. Literature Review


A large number of empirical correlations have been developed in various
studies in the literature for prediction of heat transfer and pressure drop in flow
boiling in microchannels as recently reviewed by. Of the many predictive
correlations for boiling heat transfer proposed in the literature, those by Cooper
(1984b) and Gorenflo (1993) are widely used for predicting nucleate pool boiling
heat transfer coefficients. Flow boiling features simultaneous contributions from
nucleate boiling and forced convection. There have been two main approaches
to model flow boiling: a superposition approach and an extrapolation approach
(Liu and Garimella, 2007). Chen (1966) proposed the first correlation of the
superposition type, in which the nucleate boiling and forced convection
components were linearly summed with the introduction of a suppression factor
for the nucleate boiling term and an enhancement factor for the forced
convection term. Shah (1977) proposed an extrapolation-type correlation which
used a boiling number and a convective number. After these two early studies,
many modifications have been proposed to both approaches to obtain better

106
agreement with different sets of experimental data. Until the last decade, most
experiments were performed in large channels with diameters of the order of a
few millimeters and larger (Gungor and Winterton, 1986; Tran et al., 1996).
Studies adopting the superposition-type correlations typically employed turbulent
flow models for forced convection (Gungor and Winterton, 1986). In the last
decade, experiments have focused on mini- and microchannels, and considering
the low Reynolds numbers for the flow in these small channels, the forced
convection component in flow boiling correlations has been targeted at laminar or
developing flows to better match the flow conditions in mini- and microchannels
(Warrier et al., 2002; Zhang et al., 2004; Liu and Garimella, 2007; Peters and
Kandlikar, 2007; Bertsch et al., 2008c).
In a comprehensive review by Bertsch et al. (2008a), predictions from 25
widely used correlations for boiling heat transfer coefficient were compared
against a large database of 1847 data points from ten different published studies
in the literature. This comprehensive quantitative comparison showed that the
pool boiling correlations evaluated resulted in a better prediction of the
microchannel flow boiling data than those proposed particularly for flow boiling
and that nucleate boiling dominates the heat transfer in microchannels. BarCohen and Rahim (2009) examined the predictions from five classical two-phase
heat transfer correlations for miniature channel flow. They concluded that
although some of these correlations provide good accuracy in prediction of
single-channel refrigerant flow, they fail to predict boiling of water in single
microchannels or refrigerants and dielectric liquids in multiple microchannel
configurations.

7.2. Heat Transfer Coefficient


The experimental results obtained in the present work are compared to
predictions from ten correlations from the literature. The pool boiling and flow
boiling correlations included in the comparisons in the present study are listed in
Table 7.1.

107
These comparisons are illustrated in Figure 7.1- Figure 7.4 for two
different channel dimensions and two mass fluxes. Among all the correlations
considered, including those for pool boiling and flow boiling, Coopers nucleate
pool boiling correlation (1984b) best predicts the experimental results from the
present work. The mean absolute error (MAE) in the predictions from Coopers
correlation for seven test pieces with a fixed channel depth of 400 m and four
mass fluxes is included in Table 7.2. Except for the 100 m and 250 m wide
microchannels for the lower mass flux of 225 kg/m2s, for which the MAE is 25.9%
and 19.5%, respectively, the MAE for all other cases is less than 12.2%. In
general, it can be concluded that for most of the experimental cases with
Bo 0.5 Re > 160 where nucleate boiling is dominant, the nucleate pool boiling

correlation of Cooper (1984b) predicts the experimental results very well, while
for smaller microchannels at lower mass fluxes with Bo 0.5 Re < 160 (which would
exhibit vapor confinement), the error associated with the prediction of heat
transfer coefficient using this correlation are larger.
Deviations from the experimental values of the heat transfer coefficients
predicted by all the studied correlations, for all seven test pieces with a channel
depth of 400 m, are listed in Table 7.3. Correlations proposed by Peters and
Kandlikar (2007) and Warrier et al. (2002) generally overpredict the experimental
results for small microchannels or low mass fluxes and underpredict those for the
large channels with high mass fluxes. The other correlations (Gorenflo, 1993;
Chen, 1966; Shah, 1977; Gungor and Winterton, 1986; Tran et al., 1996; Liu and
Garimella, 2007; Zhang et al., 2004) highly overpredict the experimental results
for all tested microchannels and mass fluxes.
As shown in Figure 7.1 - Figure 7.4 and discussed earlier, except for the
very high heat fluxes, h increases with increasing heat flux. Some of the
correlations (Cooper, 1984b; Gorenflo, 1993; Tran et al., 1996) show this
dependency of h on heat flux correctly, while some others (Chen, 1966; Shah,
1977; Gungor and Winterton, 1986; Zhang et al., 2004; Peters and Kandlikar,
2007) only show this trend for large mass fluxes.

108
The experimental results obtained in the current study were also
compared by Bertsch et al. (2008a) to 25 empirical correlations, ten of which
were discussed above. The values of mean average error with the percentage of
data predicted within 30% for all of the 25 correlations, along with the fluid and
geometry for which each correlation was developed, are listed in Table 7.4. As
seen in this table, none of the examined correlations predict the heat transfer
measurements adequately, except Coopers pool boiling correlation which
predicts the experimental results in the 400 m-deep channels with an MAE of
7.3%. Including the data for the other microchannel test pieces listed in Table
2.1, an MAE of 11.9% is obtained for correlation of Cooper. Kew and Cornwell
(1997) also showed that the correlation of Cooper (1984a) predicted their
experimental results better than the existing correlations for flow boiling.
It may be concluded that for the range of parameters considered in the
present study, heat transfer coefficient predictions from Coopers correlation are
very satisfactory for flow boiling in microchannels for nucleate boiling dominant
regime.

7.3. Pressure Drop


The experimental results for pressure drop are compared in this section
with predictions from empirical correlations in the literature.
The pressure drop measured between the inlet and outlet manifolds
located upstream and downstream of the microchannels includes the pressure
drop across the microchannels and the inlet and outlet manifolds as well as the
pressure loss and recovery due to the inlet contraction and the outlet expansion.
For the comparisons in this work, the pressure drop across the microchannels
alone is extracted as follows:
Pch = Pmeas Pe Pc

(7.1)

109
The working fluid enters the microchannels in a purely liquid state. The
pressure loss associated with the flow contraction at the inlet of the
microchannels is obtained from (Blevins, 1992; Liu and Garimella, 2004)
NA

Pc = 1 cs
Apl

2
1 G2

+ K c
2 f

(7.2)

where K c is the contraction coefficient given by


K c = 0.0088 2 0.1785 + 1.6027

(7.3)

Before entering the microchannels, the liquid enters a plenum with a crosssectional area smaller than that of the inlet manifold. The pressure loss at the
entrance of this plenum is calculated similarly, using the appropriate values for
the cross-sectional areas and mass flux in Eq. (7.2) and aspect ratio in Eq.(7.3).
A two-phase mixture of liquid and vapor exits the microchannels and the
pressure recovery resulting from the flow expansion at the exit for two-phase flow
is calculated from (Chisholm and Sutherland, 1969)
Pe,tp =

G 2 NAcs

f Aman

NAcs

5
1
2
1 (1 xexit ) 1 +
+ 2

Aman

X vv X vv

(7.4)

in which X vv is the Martinelli parameter for laminar liquid and laminar vapor
phases and is given by

X vv = f

0.5

(1 xexit )

xexit

0.5

0.5

(7.5)

The corrected value of Pch so obtained is compared to predicted values.


In cases where both single-phase and two-phase flow exists in the
microchannels, the predicted values are calculated separately for the singlephase and two-phase regions. The single-phase region can be divided into a
developing and a fully developed region for which the lengths can be obtained
from (Lee and Garimella, 2008)

Lsp ,dev

Lsp
if L+sp < 0.05
=
if L+sp 0.05
0.05 ( Re Dh )

(7.6)

110
Lsp , fd = Lsp Lsp ,dev

(7.7)

where L+sp = Lsp / ( Re Dh ) in which the overall length of the single-phase region,
Lsp , can be obtained from a heat balance

Lsp =

GAcs c p (Tsat T f ,in )

(7.8)

qw PH

The friction factor associated with the developing region is then obtained from

f sp ,dev

0.05
2

2
+ 0.57

3.2 / ( 0.5 Lsp ) + ( f fd Re ) / Re

=
0.05
2

0.57 2
+ ( f fd Re )

3.2
/
0.5
0.05
/ Re
(
)

if L+sp < 0.05

(7.9)
if L+sp 0.05

where the fully developed friction constant for a rectangular channel is


96 1.3553 1.9467 1.7012 0.9564 0.2537
f fd = 1
+

2
3
4
5
Re

(7.10)

The total single-phase pressure drop in then obtained from


Psp =

G2
2 f

(f

sp , dev

Lsp , dev + f sp , fd Lsp , fd )


Dh

(7.11)

The pressure drop in the two-phase region of the microchannels is the


sum of the frictional and the accelerational components. A large number of
empirical correlations are available in the literature for prediction of these two
components and many studies in the literature have compared these correlations
to experimental data (Lee and Garimella, 2008; Qu and Mudawar, 2003a; Lee
and Mudawar, 2005; Mauro et al., 2007; Cheng et al., 2008). Six of these
correlations that predicted the experiments of Lee and Garimella (2008), Qu and
Mudawar (2003a), and Lee and Mudawar (2005) better than other correlations
are chosen here for comparison to the experimental data. One of these
correlations is a widely used macrochannel correlation (Lockhart and Martinelli,
1949), while the others were developed for mini/microchannels (Lee and Lee,
2001; Lee and Garimella, 2008; Mishima and Hibiki, 1996; Qu and Mudawar,
2003a; Lee and Mudawar, 2005). Predictions from these correlations are

111
compared with the measured pressure drops in Figure 7.5. The MAEs listed in
this figure ranging from 84.7% to 394.2% reveal the failure of these empirical
correlations in providing a suitable prediction of the experimental results, mainly
because the correlations were developed for specific fluids and ranges of
operating parameters that differ from those of the current experimental data.

7.4. Conclusions
The heat transfer coefficients obtained from the current experiments are
compared to predictions from a number of correlations in the literature. The
correlation of Cooper (1984b) for nucleate pool boiling predicts the experimental
heat transfer coefficients very well, with an average mean absolute percentage
error of 11.9%. This error is larger for smaller microchannels and lower mass
fluxes where confinement occurs.
The experimental results for pressure drop are also compared with
predictions from empirical correlations in the literature. The empirical
correlations are shown to fail to adequately predict experimental pressure drops
in microchannels.
Although the pool boiling correlation of Cooper was shown to predict the
experimental heat transfer data well, none of the empirical correlations
developed specifically for flow boiling in microchannels were found to predict
experimental heat transfer coefficient and pressure drop to within a reasonable
error. Hence, it is essential to develop physics-based models based on the
relevant flow regimes to predict both heat transfer coefficient and pressure drop
in microchannel flow boiling. Physics-based models are expected to be
applicable to a wider range of parameters, and not just to specific data sets.

112

Table 7.1. Summary of boiling heat transfer correlations.

Reference

Fluid

Cooper
(1984b)

Water,
refrigerant,
organic
fluids

Geometry

Parameters range/
Flow regime

Pool boiling

0.001< pr<0.9
2<M<200
Nucleate pool boiling

Correlation for h

hnb = 55 pr

0.120.4343ln Rp

( 0.4343ln pr )

hnb = h0 FPF ( q q0 ) ( Rp Rp 0 )
n

Gorenflo
(1993)

Large
selection of
fluids

Pool boiling

0.0005< pr<0.95
Nucleate pool boiling

for RC318
FPF = 1.2 p

0.27
r

n = 0.9 0.3 p

Chen
(1966)

Water &
organic
fluids

Vertical tubes
& annuli

Shah
(1977)

Water,
refrigerants
(R11, R12,
R22, R502)

Horizontal &
vertical tubes
D=6-41 mm

0.55 < P < 34.8 bar


0.06 < ui< 4.5 m/s
1 < x < 71 %
6.2 < q < 2400 kW/m2
Saturated flow boiling
0.004< pr<0.8
Saturated flow boiling

0.133

0.55

M 0.5q0.67

(7.12)

, h0 = 4200 W / m 2 K

+ 2.5 pr + pr (1 pr )

(7.13)
(7.14)
(7.15)
(7.16)

0.3
r

htp = Shnb + Fhsp


hsp = 0.023 (1 x ) Re f

0.8

Pr f0.4 ( k f D )

(7.17)

k 0.79 c 0.45 0.49


0.24
0.75
hnb = 0.00122 0.5f 0.29p f 0.24 f 0.24 ( Tsat ) ( psat ) (7.18)
f h fg g

(7.19)
F = f ( X tt ) , S = f ( Re )

htp hsp = max ( nb , cb )

hsp = 0.023 (1 x ) Re

0.8

(7.20)
Pr f0.4 ( k f D )

nb = f ( Bl , Co ) , cb = f ( Co )

(7.21)
(7.22)

112

113

Table 7.1 Continued.

Reference

Gungor &
Winterton
(1986)

Tran et al.
(1996)

Warrier et
al.
(2002)

Fluid
Water,
refrigerants,
ethylene
glycol

Refrigerants
(R12, R113)

FC84

Geometry
Horizontal &
vertical tubes
& annuli
D=3-32 mm
Horizontal
circular &
rectangular
channels of
Dh ~ 2.4 mm
Horizontal
rectangular
parallel
channels of
Dh=0.75 mm

Parameters range/
Flow regime
0.08 < Psat < 203 bar
12 < Gi< 61518 kg/m2s
0 < x < 173 %
1 < q < 91534 kW/m2
Saturated & subcooled
flow boiling
Psat = 0.8 Mpa
44 < Gi< 832 kg/m2s
x < 94 %
3.6 < q < 129 kW/m2
Nucleation dominant
region in small
channels
0.00027 < Bl < 0.00089
3 < x < 55 %
Saturated flow boiling

Correlation for h
(7.23)

htp = Shnb + Fhsp


hsp = 0.023 (1 x ) Re

0.8

Pr 0.4
f (k f D)

hnb = 55 pr 0.12 ( 0.4343ln pr )

0.55

M 0.5 q0.67

F = f ( Bl , X tt ) , S = f ( Re )

htp = 8.4 105 Bl 2We f

htp hsp , fd = 1 + 6 Bl

1
16

) (
0.3

(7.24)
(7.25)
(7.26)

g )

0.4

5.3 (1 855 Bl ) x 0.65

(7.27)

(7.28)

113

114

Table 7.1 Continued.

Reference

Zhang et
al.
(2004)

Fluid

Water,
refrigerants
(R11, R12,
R113)

Geometry
Horizontal/
vertical single
circular/rectan
gular
channels
Dh = 0.78-6
mm

Parameters range/
Flow regime
0.101 < P < 1.21 MPa
23.4 < Gi< 2939 kg/m2s
2.95 < q < 2511
kW/m2
Saturated flow boiling
in mini channels

Correlation for h
htp = Shnb + Fhsp

(7.29)

hsp = ( k f Dh ) Nusp ,t for Re f 2300

(7.31)

hsp = ( k f Dh ) max( Nusp ,v , Nu sp ,t ) for Re f < 2300 (7.30)

0.45 0.49
k0.79

0.24
0.75
f cp f f
(7.32)
hnb = 0.00122 0.5 0.29 0.24 0.24 ( Tsat ) ( Psat )
f hfg g

(7.33)
F = max ( 0.64 f ,1) , S = f ( Re )

htp hsp = c1Coc2 (1 x )

0.8

( 25Fr )

c5

+ c3 Bl c4 (1 x ) Ffl (7.34)
0.8

htp hsp = c3 Bl c4 (1 x ) Ffl for Re f < 100


0.8

Peters &
Kandlikar
(2007)

R123

Circular &
rectangular
channels of
Dh ~ 0.21 mm

400 < Gi< 800 kg/m2s


25 < q < 85 kW/m2
0 < x < 70 %
Flow boiling in
mini/micro channels

hsp =

hsp =

( Re

1000 ) Pr f ( f 2 ) ( k f D )

1 + 12.7 Pr f 3 1 ( f 2 )
Re f Pr f ( f 2 ) ( k f D )

)( f

1 + 12.7 Pr f 3 1

2)

0.5

0.5

(7.35)

for 104 < Re f < 5 106


(7.36)

for 3000 < Re f < 104 (7.37)

hsp = ( k f Dh ) Nu for Re f < 1600

(7.38)

114

115

Table 7.1 Continued.

Reference

Liu &
Garimella
(2007)

Bertsch et
al.
(2008c)

Fluid

Water

12 different
fluids

Geometry

Parameters range/
Flow regime

Rectangular
channels
0.30.6 &
0.41 mm

221 < Gi< 1283 kg/m2s


q < 1290 kW/m2
x < 20 %
Saturated flow boiling
in micro channels

0.16 mm < Dh
< 2.92 mm

20 < Gi< 3000 kg/m s


4 < q < 1150 kW/m2
0 < x < 100 %

Correlation for h
(7.39)

htp = Shnb + Fhsp


hsp = 1.86 (1 x ) Re f Prf Dh L

1
3

hnb = h0 FPF ( q q0 ) ( R p R p 0 )
n

w )( k f Dh )

0.133

F = f ( f , , c p , k , Pr ) , S = f ( Re, F )

(7.41)
(7.42)

htp = hnb (1 x) + Fhconv,tp 1 + 80 x 2 x 6 e 0.6 Co

hconv ,tp = hconv , f (1 x ) + hconv , g x

hnb = 55 pr

0.120.2log10 Rp

( log10 pr )

(7.40)

(7.43)
(7.44)

0.55

M 0.5q0.67 (7.45)

115

116

Table 7.2. Deviation of heat transfer coefficient obtained from the experiments
and predicted by Coopers correlation (1984b). (The channel dimensions are
referred to by their nominal values. The deviation is presented in terms of Mean
Absolute Percentage Error.)

Microchannel Microchannel
width, w (m) depth, d (m)

100

250

400

700

1000

2200

5850

Mass flux, G
(kg/m2s)

Deviation between hexp and


hpred (Cooper 1984b) (%)

400

214
621
1017
1405

25.9
12.1
8.1
4.7

400

226
611
1126
1415

19.5
12.2
10.7
9

400

227
633
1031
1431

9.4
7
9.3
10.8

400

225
641
1053
1461

10.3
6.1
5.1
3

400

224
627
1037
1440

10.7
6.9
3.8
4.1

400

227
633
1034
1427

6.8
5.9
6.7
7.7

400

229
632
1028
1289

10
7.2
6.1
6.9

117

Table 7.3. Deviation of heat transfer coefficient obtained from the experiments and predicted by existing correlations.
(The deviation is presented in terms of Mean Absolute Percentage Error.)4

w (m)
Gungor & Tran Warrier Zang Peters &
Liu &
Betsch et
Cooper Gorenflo Chen Shah
al.
(nominal
Winterton et al.
et al.
et al. Kandlikar Garimella
(1984b)
(1993)
(1966) (1977)
(1986)
(1996) (2002) (2004)
(2007)
(2007)
(2008c)
values)
100

13

183

203

307

464

36

369

211

239

428

26

250

13

217

164

284

531

80

124

161

159

381

43

400

177

84

198

445

60

61

76

86

253

19

700

213

87

209

512

92

58

82

80

266

34

1000

214

77

179

519

102

55

85

75

269

37

2200

230

42

191

549

105

64

62

21

129

46

5850

214

54

122

524

120

46

108

63

290

35

All data

207

102

213

506

85

97

112

103

288

34

Comparison is made to the experimental data in the 400 m-deep microchannels only.

117

118

Table 7.4. Studies in the literature from which heat transfer correlations are
selected for comparison against the current experimental data. Mean absolute
error (MAE) and percentage of predictions which fall within 30% of the
measurements are listed for each correlation.5

Correlation

Fluid,
Geometry

MAE
(%)

Percentage of
pred. within
30%

Cooper (pool boiling)


(1984b)

Water, refrigerants, organic fluids,


cryogens

7.3

100

Several refrigerants, water and


cryogenics

154

Chen (1966)

Water, Methanol, Pentane,


Heptane, Benzene,

161.1

Bennett and Chen


(1980)

Water, Methanol, Pentane,


Heptane, Benzene,

186.3

1.6

R11, R12, R22, R502,

836

R113; Dh = 3.1 mm

149.6

Water, R11, R12, R113,...


Dh = 2.95-32.0 mm

484.1

Kandlikar (1991)

Water, R11, R12, R22, R113,


Nitrogen,...
Dh = 4.632 mm

106.7

29.7

Liu and Winterton


(1991)

Water and refrigerants; Dh = 2.9532.0 mm

79.0

28.1

Steiner and Taborek


(1992)

Water, refrigerants, cryogenics


Dh = 1 - 32 mm

148.1

Tran et al. (1996)

R12, R113; Dh = 2.4-2.92 mm

61.4

14.4

Gorenflo (pool
boiling) (1993)

Shah (1977)
Lazarek and Black
(1982)
Gungor and Winterton
(1986)

Comparison is made to the experimental data in the 400 m-deep

microchannels only.

119

Table 7.4 Continued.

Fluid,
Geometry

MAE
(%)

Percentage of
pred. within
30%

Yan and Lin (1998)

R134a; Dh = 2.0 mm

16248.6

2.6

Lee and Lee (2001)

R113; Dh = 0.78-3.6 mm

295.4

18.2

Warrier et al. (2002)

FC-84; Dh = 0.75 mm

68.1

27.2

Yu et al. (2002)

Water; Dh = 2.98 mm

4131.5

Haynes and Fletcher


(2003)

R11, R123; Dh = 0.92-1.95 mm

126.6

0.6

Sumith et al. (2003)

Water; Dh = 1.45 mm

101.1

35.1

Balasubramanian and
Kandlikar (2004)

Water, R113, R123, R141b,


Dh = 0.19 - 2.92 mm

103.1

29.7

Thome et al. (2004)

R11, R12, R113, R123, R134a,


R141b, CO2
Dh = 0.7-3.1 mm

43.4

39.3

Lee and Mudawar


(2005)

R134a, water; Dh = 0.35 mm

461.2

6.1

Zhang et al. (2004)

Water, R11, R12, and R113


Dh = 0.78-6.0 mm

100.2

11.8

Yun et al. (2007)

R410A; Dh = 1.36, 1.44 mm

502.2

Liu and Garimella


(2007)

Water; Dh = 0.38, 0.59 mm

83.7

31.9

Saitoh et al. (2007)

R134a; Dh = 0.5-11.0 mm

211.7

Lee and Garimella


(2008)

Water; Dh = 0.16-0.57 mm

339.9

Correlation

120

30
w = 250 m
2
G = 225 kg/m s

25

h (kW/m K)

20
Experiments
Cooper, 1984
Gorenflo, 1993
Chen, 1966
Shah, 1977
Gungor & Winterton, 1986
Tran et al., 1996
Warrier et al., 2002
Zhang et al., 2004
Peters & Kandlikar, 2007
Liu & Garimella, 2007

15
10
5
0

50

100 150 200


q"w (kW/m2)

250

300

Figure 7.1. Comparison of correlations in the literature for heat transfer


coefficients, 250 m 400 m microchannels, G = 225 kg/m2s.

121

30
w = 250 m
G = 1420 kg/m2s

25

h (kW/m K)

20
Experiments
Cooper, 1984
Gorenflo, 1993
Chen, 1966
Shah, 1977
Gungor & Winterton, 1986
Tran et al., 1996
Warrier et al., 2002
Zhang et al., 2004
Peters & Kandlikar, 2007
Liu & Garimella, 2007

15
10
5
0

100

200 300 400


q"w (kW/m2)

500

600

Figure 7.2. Comparison of correlations in the literature for heat transfer


coefficients, 250 m 400 m microchannels, G = 1420 kg/m2s.

122

30
w = 1000 m
2
G = 225 kg/m s

25

h (kW/m K)

20
Experiments
Cooper, 1984
Gorenflo, 1993
Chen, 1966
Shah, 1977
Gungor & Winterton, 1986
Tran et al., 1996
Warrier et al., 2002
Zhang et al., 2004
Peters & Kandlikar, 2007
Liu & Garimella, 2007

15
10
5
0

50

100 150 200


q"w (kW/m2)

250

300

Figure 7.3. Comparison of correlations in the literature for heat transfer


coefficients, 1000 m 400 m microchannels, G = 225 kg/m2s.

123

30
w = 1000 m
2
G = 1420 kg/m s

25

h (kW/m K)

20
Experiments
Cooper, 1984
Gorenflo, 1993
Chen, 1966
Shah, 1977
Gungor & Winterton, 1986
Tran et al., 1996
Warrier et al., 2002
Zhang et al., 2004
Peters & Kandlikar, 2007
Liu & Garimella, 2007

15
10
5
0

100

200 300 400


q"w (kW/m2)

500

600

Figure 7.4. Comparison of correlations in the literature for heat transfer


coefficients, 1000 m 400 m microchannels, G = 1420 kg/m2s.

124

350

350
+30%

300

250

250
-30%

200

150

P pred (kPa)

P pred (kPa)

+30%

300

100

150
100

Lockhart & Martinelli, 1949


MAE
MAE == 186.0%
186.0%
12.8%
of predictions
12.8%
of predictions
within within
error of 30%
30%

50

-30%

200

50

100

150

200

250

300

Mishima & Hibiki, 1996


MAE
MAE= =84.7%
84.7%
23.2%
predictions
within
23.2%
of of
predictions
within
error of30%
30%

50

350

50

100

P exp (kPa)

150

350

250

250
-30%

200
150

P pred (kPa)

P pred (kPa)

300

100

350

-30%

200
150

100

Lee & Lee, 2001


MAE
MAE == 108.6%
108.6%
21.1%
of predictions
21.1%
of predictions
within within
error of 30%
30%

50

50

100

150

200

250

300

Qu & Mudawar, 2003a


MAE
= =163.3%
MAE
163.3%
7.4%
of predictions
7.4%
of predictions
withinwithin
error of30%
30%

50
0

350

50

100

P exp (kPa)

150

200

250

300

350

P exp (kPa)

350

350
+30%

+30%

300

300

250

250

P pred (kPa)

P pred (kPa)

300

+30%

300

-30%

200
150

100

-30%

200
150

100

Lee & Mudawar, 2005


MAE
MAE == 394.2%
394.2%
4.7%
of predictions
4.7%
of predictions
within within
error of 30%
30%

50
0

250

350
+30%

200

P exp (kPa)

50

100

150

200

P exp (kPa)

250

300

Lee & Garimella, 2008


MAE
MAE == 193.0%
193.0%
4.4%
of predictions
4.4%
of predictions
within within
error of 30%
30%

50

350

50

100

150

200

250

300

350

P exp (kPa)

Figure 7.5. Comparison of the experimentally measured pressure drop across


microchannels with predictions from empirical correlations in the literature.

125

CHAPTER 8. REGIME-BASED MODELING OF HEAT TRANSFER AND


PRESSURE DROP

The comprehensive flow regime maps developed in Chapter 6 for a wide


range of channel dimensions and experimental conditions, consisted of four
distinct regions of bubbly, slug, confined annular, and alternating
churn/annular/wispy-annular flow regimes. In this chapter, physics-based
analyses of local heat transfer in each of the four regimes of the comprehensive
map are formulated. Flow regime-based models for prediction of heat transfer
coefficient in slug flow and annular/wispy-annular flow are developed and
compared to the experimental data. Also, regime-based prediction of pressure
drop in microchannels is discussed by evaluating pressure drop of each flow
regime along the microchannels separately.

8.1. Literature Review


Few regime-based models exist in the literature for the prediction of heat
transfer coefficient and pressure drop in microchannel flow boiling. Thome et al.
(2004) proposed a three-zone boiling model to predict the local dynamic and
time-averaged heat transfer coefficient in the elongated bubble regime. This
model assumed the passage of a liquid slug, confined elongated bubble, and
vapor slug at a fixed point in the microchannel, with transient evaporation of the
thin liquid film surrounding the elongated bubble being the dominant heat transfer
mechanism (rather than nucleate boiling). This model illustrates the strong
dependency of the heat transfer on bubble frequency, the minimum liquid film
thickness at dryout, and the liquid film formation thickness, all of which are
obtained from experiments due to the difficulty in obtaining them theoretically.

126
The authors compared the time-averaged local heat transfer coefficient predicted
by the three-zone model to the experimental measurements from seven
independent studies in the literature, including six refrigerants and CO2 (Dupont
et al., 2004), and obtained a set of general empirical parameters to be used in
the model. The model predicted 67% of the database within a mean average
error (MAE) of 30%. Ribatski et al. (2007) compared predictions from the threezone slug flow model (Thome et al., 2004) to experimental results for boiling heat
transfer of pure Acetone. Using the general empirical parameters developed in
(Dupont et al., 2004), 69% of the experimental data were predicted to within
30%, while using a new set of empirical parameters optimized for Acetone data,
the model predicted 90% of the heat transfer data to within 30%. Predictions
from this model were also compared to experimental data for flow boiling of
R254fa and R236fa (Agostini et al., 2008). Adjusting the empirical parameters of
the model to this experimental dataset, the model predicted 90% of the
measurements to within 30% of error. Shiferaw et al. (2007) compared their
experimental data with R134a to the predictions from the three-zone model of
Thome et al. (2004) as well as from other empirical correlations and suggested
that the three-zone model based on convective heat transfer performs at least as
well as empirical correlations that interpret the data in terms of nucleate boiling.
Qu and Mudawar (2003b) performed experiments in water-cooled
microchannel heat sinks and showed an abrupt transition to an annular regime
upon the onset of boiling. They concluded that the dominant heat transfer
mechanism in microchannels is forced convective boiling corresponding to
annular flow. Comparison of their experimental results to predictions from 11
empirical correlations which were developed for both macrochannels and
microchannels revealed deviations from 19.3% to 272.1% in terms of mean
absolute errors due to the unique features of water-cooled microchannel boiling
and the operating conditions that fall outside the recommended range for most
correlations. Qu and Mudawar (2003c) developed a model to predict the
saturated heat transfer coefficient in the annular regime, incorporating features

127
relevant to boiling of water in microchannels such as laminar liquid and vapor
flow, smooth interface, and strong droplet entrainment and deposition effects.
Their model predicted their experiments with an MAE of 13.3%. Their model also
allowed the calculation of pressure drop over the length of the annular region,
considering the entire two-phase length of the channel as being in annular flow
(Qu and Mudawar, 2003a). This led to an MAE of 12.7%, matching the accuracy
of the best of ten empirical correlations that were also tested.
Quiben and Thome (2007b) performed an analytical investigation of
pressure drop during boiling in horizontal single tubes. They proposed a flow
pattern-based model for prediction of the frictional pressure drop, treating each
flow regime separately and assuming that only one flow regime exists in the
complete test section. Their model ensured a smooth transition in the predicted
pressure drop at the transitions between flow regimes and predicted 82.3% of the
experimental data with three refrigerants (Quiben and Thome, 2007a) to within
30%.
A review of the literature reveals only a few studies that have focused on
modeling of flow boiling based on the existing flow regimes and taken into
account the interfacial structure between the liquid and vapor phases. Also, even
these studies have assumed the existence of a single regime in the channels. It
has been shown in the current study as well as in the literature (Huo et al., 2004;
Kandlikar, 2004; Revellin et al., 2006; Chen and Garimella, 2006a), however, that
several flow regimes can be present in microchannels for different operational
and geometric conditions or even concurrently in a streamwise direction.
In the present study, an analytical model is developed for the annular flow
regime to predict local heat transfer coefficient. An empirical parameter is
introduced for calculating the interfacial shear stress in the liquid film surrounding
the vapor core. This model also enables calculation of pressure drop in the
annular flow. For the slug flow regime, the three-zone model of Thome et al.
(2004) is compared against the measured data in the slug region. The model is
then modified by using a different method for prediction of the initial liquid film

128
thickness surrounding the elongated bubble in order to improve the original
model for better agreement with the measurements.

8.2. Model Development


In this section, three different analytical models are proposed for three of
the four quadrants of the flow regime map in Figure 6.9, i.e., confined annular,
annular/wispy-annular, and slug flow. The three models are then validated by a
comparison to the experimental data. For the fourth region of bubbly flow, use of
an existing empirical correlation the Cooper correlation (1984b) is suggested.
The proposed physical models for heat transfer are also used to predict pressure
drop in these flow regimes.

8.2.1. Bubbly Flow


An analytical model for the bubbly flow is not attempted in this study since
it has been shown in Chapter 7 that the empirical correlation of Cooper (1984b)
for pool boiling predicts the experimental data very well in this nucleate boiling
dominant region. Figure 8.1 depicts predictions from Coopers pool boiling
correlation for the bubbly flow data in the current study; the MAE is 13.9% and
86.5% data points are captured to within 30%.

8.2.2. Confined Annular Flow


A model for prediction of heat transfer coefficient and pressure drop in
confined annular flow is developed based on the conservation of mass,
momentum, and energy, using an approach similar to that presented by Carey
(1992) for large vertical tubes.

129
8.2.2.1. Model Development
Figure 8.2 shows a schematic representation of annular flow in
microchannels. A continuous vapor core flows along the center of the
microchannel and is surrounded by a thin liquid film along the channel walls.
Liquid droplets can be entrained into this vapor core. The model discussed here
assumes that the two-phase flow is steady, the pressure is uniform across the
channel cross-section, the liquid film-vapor core interface is smooth, the
thickness of the liquid film is uniform along the channel circumference, and
evaporation occurs only at the liquid film-vapor core interface and evaporation
from the entrained droplets is neglected.
The mass flow rates of the vapor core, the liquid film, and the liquid
droplets can be found from
m g = x0 m

(8.1)

m film = (1 x0 e0 ) m

(8.2)

m E = e0 m

(8.3)

where x0 is the vapor quality at the onset of annular flow. Knowing the location
along the channel at which annular flow commences, La 0 , from the flow regime
map discussed in section 6.2.3 (Eq. (6.6)), x0 can be obtained from an energy
balance over the length of La 0
x0 =

1 qw PH La 0

c p (Tsat Tin )

h fg m

(8.4)

In Eq. (8.2), e0 is the liquid droplet quality at the onset of annular flow. Qu and
Mudawar (2003c) discussed different correlations to determine this parameter
and developed an expression of the form e0 = 0.951 0.15 We f 0 . The total mass
flow rate in each microchannel is the sum of the three components of the flow in
Eqs. (8.1)-(8.3): m = m g + m film + m E .

130
The mass transfer rate due to evaporation per unit channel length is defined as

fg =

qw PH
and the mass transfer rate due to deposition is d = kd C Pc (Qu and
h fg

Mudawar, 2003c), where C =


C
and kd = 47.8 Bo

m E
is the liquid droplet concentration
m g / g + m E / f

0.147

jg is the deposition mass transfer coefficient proposed

by Qu and Mudawar (2003c) based on a correlation originally developed by


Paleev and Filippovich (1966). Using these mass transfer rates, the variation of
each component of the mass flow rate with distance along the channel can be
calculated from
dm film
dz

= fg + d

dm E
= d
dz
dm g
dz

= fg

(8.5)
(8.6)
(8.7)

Momentum conservation in vapor core:


A control volume of length z covering the vapor core and extending to
the liquid film interface as depicted in Figure 8.2(b), is used to apply the
momentum and force balance to the vapor core in the flow direction which results
in (Qu and Mudawar, 2003c)
d
d
H uc2 Ac ) + d uc fg ui = ( PAc ) i Pc
(
dz
dz

where H = x / g + (1 x ) / f

(8.8)

is the homogeneous density of the vapor core.

From this equation, the pressure gradient for the vapor core (and the liquid
film) is obtained

1
dP Pc
= i ( fg ui d uc )
dz Ac
Ac

(8.9)

131
where i is the interfacial shear stress. The interfacial velocity ui is
approximated to be twice the mean liquid film velocity, ui = 2u film = 2

m film

f ( Acs Ac )

The validity of the approximation used for ui was discussed by Qu and Mudawar
(2003c). The mean velocity of the vapor core is evaluated assuming
homogeneous flow for the vapor core, uc =

m E + m g

H Ac

Interfacial shear stress:


To determine the interfacial shear stress, an approach by Wallis (1969) is
used to incorporate the influence of evaporation mass transfer at the interface on
interfacial friction, as expressed in Qu and Mudawar (2003c) in the form

i =

fg
1
2
f i H ( uc ui )
( uc ui )
2
2 Pc

(8.10)

For the interfacial friction factor, f i , a simple correlation proposed by Wallis


(1969) is used in the current model. Since this correlation is developed from airwater data in large tubes, a correction factor, c fi , is introduced in the current
model which is optimized based on the current experimental data for annular flow
in microchannels as discussed further below in the model assessment


f i = c fi 0.005 1 + 300

Dh

(8.11)

Momentum conservation in liquid film:


Applying momentum conservation to a control volume in the liquid film as
shown in Figure 8.2(b), with the shape of a rectangular ring, leads to (Qu and
Mudawar, 2003c)

1
dP
( fg ui d uc )
+i
Pch
dz

= ( y )

(8.12)

Substituting the shear stress in the laminar liquid film with an expression in terms
of the local velocity gradient, = f

du f
dy

, in Eq. (8.11) and integrating the

132
resulting equation, using the no-slip boundary condition at the wall, the local
velocity in the liquid film is obtained as

uf =

1
1
y 2 dP 1

y
y i
y ( fg ui d uc )


+
f
f Pch
2 dz f

(8.13)

Integrating the local liquid velocity over the film thickness, conservation of mass
in the liquid film requires that:

Pch f 3 dP Pch f 2
f 2
= f Pch u f dy =
i
( fg ui d uc ) (8.14)

+
3 f dz
2 f
2 f
0

m film

Solution procedure:
For known values of qw , G , x0 , channel dimensions, and fluid properties,
the equations developed above give a closed system to obtain m F , i ,

dP
, and
dz

; however, due to complexity of the equations, they need to be solved


numerically according to the following procedure:
1. The location of the onset of annular flow is determined first using
Eq. (6.6). A one-dimensional grid with sufficient number of cells is
then assumed along the channel length in the annular region. The
solution is initiated at the upstream boundary node.
2. The mass flow rates of the vapor core, liquid film, and entrained
droplets are determined at the upstream boundary, using Eqs. (8.1)
-(8.3).
3. A value of is guessed at this node.
4. Ac and Pc can be calculated using the guessed value of . ui and
uc are obtained knowing the mass flow rates and the geometrical

parameters. The interfacial shear stress, i , is then evaluated from


Eqs. (8.10) and (8.11).
5. Eq. (8.9) is solved to obtain dP / dz .
6. The pressure gradient obtained in step 5 is substituted in Eq. (8.14)
to evaluate the integral of u f across the film thickness. Knowing

133

the value of m F from previous steps, if the mass conservation in


Eq. (8.14) is satisfied, the solution is complete at this node.
Otherwise, the calculations must be repeated from step 3 by
guessing a new value of . This process is continued until Eq.
(8.14) is satisfied, at which time the values of parameters from the
last iteration are adopted for this node.
7. Now the solution for the next downstream node is sought. The
mass flow rates for the three components of the flow are calculated
using Eqs. (8.5)-(8.7). The numerical procedure in steps 3 to 6 are
then repeated for this new node to complete the solution by
satisfying Eq. (8.14). This procedure is then repeated by marching
downstream to finally obtain a local solution for all the nodes in the
annular region.
After obtaining the film thickness from the procedure described above, the
local heat transfer coefficient in the annular region can be obtained, assuming a
laminar liquid film and that all the heat input to the fluid is transferred to the vapor
core, from
h( z ) =

kf

(8.15)

The validity of this equation for annular flow heat transfer has been discussed by
Collier and Thome (1994).

8.2.2.2. Model Assessment


Flow visualizations in Chapter 3 reveal that droplet entrainment in the
vapor core is negligible for the confined annular flow in case of the perfluorinated
liquids discussed here. Hence, e0 , m E , and d are all set to zero in the model.
This simplifies the solution procedure explained in the previous section, due to
the linear variation of the vapor and film mass flow rates in the flow direction
when no droplet entrainment is present. Hence, in order to find the local heat

134

transfer coefficient at location L along the channel and downstream of La 0 , it is


not necessary to start the solution at the upstream boundary; instead, the
calculations are performed at location L following steps 2 to 6 of section 8.2.2.1,
by substituting L for La 0 in Eq. (8.4). In Appendix G, a MATLAB script is given
that calculates the heat transfer coefficient in the annular region.
The heat transfer coefficient values were obtained from the proposed
numerical model at the same location along the microchannels where the
experimental measurements were performed. Only data for the confined annular
flow are included in this section. The value of c fi in Eq. (8.11) was optimized by
a comparison of the numerical values to the experimental values from the current
study. The optimized value obtained for the correction factor in the friction
coefficient for confined annular flow is expressed as
c fi = 4 10 5 ( Bo 0.5 Re )

(8.16)

where the expression in the parentheses is the convective confinement number


proposed in section 3.2.5. For different geometries and mass fluxes where
confined annular flow is present, c fi takes values in the range of 0.01 to 1.11.
Eq. (8.16) indicates that the correction factor for the interfacial friction factor is
smaller for smaller microchannels and lower mass fluxes.
In Figure 8.3, predictions from the proposed model for annular flow are
compared to the experimental data from the present work for confined annular
flow. The experiments are predicted with an MAE of 16.4% with 87.2% of the
data predicted to within 30%. Also, good agreement is seen in prediction of the
trend seen in the variation of heat transfer coefficient with heat flux, as depicted
in Figure 8.4(a)-(c) for microchannels of dimension 100 m 400 m for three
mass fluxes of 225, 1050, and 1420 kg/m2s.

135
8.2.3. Annular/Wispy-Annular Flow
Alternating annular/churn flow and alternating wispy-annular/churn flow
occurs in the channels for Bo 0.5 Re > 160 and N pch > 96.65 ( Bo 0.5 Re )

0.258

. In this

study it is assumed that the effect of film evaporation in the annular/wispyannular flow is more dominant than the nucleate boiling heat transfer in the churn
flow in determining the heat transfer coefficient. Hence, the same model
developed for confined annular flow is used to predict the heat transfer coefficient
in the annular/wispy-annular/churn region in the microchannels with an optimized
value of c fi specific for the data in this region.

8.2.3.1. Model Assessment


An optimized value of c fi is obtained for the annular/wispy-annular data as
c fi = 32 105 ( Bo 0.5 Re )

(8.17)

which results in c fi values in the range of 7.4 to 67.5 for different channel
dimensions and mass fluxes, which are much larger than the values obtain for
confined annular flow and increase with increasing the channel cross-sectional
area and mass flux.
Predictions from the proposed annular model, using the interfacial friction
factor correction in Eq. (8.17) and neglecting droplet entrainment, are compared
to the experimental data in Figure 8.5. This plot shows an MAE of 24.8% with
71.9% of the data predicted to within 30%. Figure 8.6 illustrates the heat
transfer coefficients as a function of heat flux for annular/wispy-annular flow data
in 400 m 400 m microchannels and shows that the model predicts the trends
very well.
Although entrained droplets are seen in flow visualizations of the wispyannular flow as reported in section 3.2.1, the model gives better predictions of
the heat transfer coefficient without considering this effect rather than taking into
account droplet deposition. It should also be noted that in channels with very

136
large aspect ratios, i.e., channels with width of 2200 m and 5850 m, flow loses
symmetry and churn and annular flow exist side-by-side in the channels as
shown in section 3.2.2; hence, the assumption of a circumferentially uniform film
thickness does not hold anymore in these cases and the simplified model
proposed in the current study does not agree well with the data; data for these
very large aspect ratios are excluded from the comparisons reported in this
section.

8.2.4. Slug Flow


Thome et al. (2004) proposed a model for prediction of the transient local
heat transfer coefficient in a slug flow regime, based on the cyclic passage of a
liquid slug, an elongated bubble, and a vapor slug triplet. The model is briefly
explained here and reference may be made to Thome et al. (2004) and Dupont
et al. (2004) for more detailed descriptions.
At a fixed location along the channel, an elongated bubble follows a liquid
slug. In the elongated bubble, heat transfer is characterized by evaporation of a
thin liquid film surrounding the vapor bubble at the walls. If the liquid film
evaporates completely and local dry-out occurs, a vapor slug follows the
elongated bubble. The time-averaged local heat transfer coefficient over the
three zones is given by
h( z ) =

tf

hf ( z) +

t film

h film ( z ) +

tdry

hg ( z )

(8.18)

In this equation, h f and hg are the heat transfer coefficients for the liquid and
vapor slugs and are obtained from the local Nusselt number using correlations of
Shah and London (VDI-Warmeatlas, 1997) for laminar flow and Gnielinski (VDIWarmeatlas, 1997) for transitional and turbulent flow. The mean heat transfer
coefficient of the evaporating thin liquid film of the elongated bubble is obtained
by assuming one-dimensional heat conduction in a stagnant liquid film, using the
averaged value of the film thickness

137

h film ( z ) =

2k f

(8.19)

0 ( z ) + end

Here, 0 is the initial film thickness at the formation of the elongated bubble and

end is the film thickness at dryout or at the beginning of the next cycle. In case
of dryout, end = min , which is the minimum possible film thickness before dryout
occurs.
To find the initial film thickness, Thome et al. (2004) used a prediction
method proposed by Moriyama and Inoue (1996), who experimentally measured
the thickness of a liquid film of R-113 formed by a bubble growing radially in a
gap between two parallel heated plates. For large superheat or bubble velocity,
the film formation was shown to be controlled by the viscous boundary layer,
while at low bubble speed or small gap between plates, the surface tension force
was dominant. Two different expressions were proposed for the film thickness
for each of these conditions. Thome et al. (2004) used these empirical
correlations and proposed the following asymptotic expression to calculate the
film thickness

f
= C 0 3
u p Dh f
Dh

0.84

( 0.07 Bo 0.41 ) 8 + 0.18

1/8

(8.20)

where C 0 is an empirical correction factor.


The triplet (or feature pair, if dryout does not occur) period, , in Eq. (8.18) is
predicted empirically as a function of the process variables as follows
c Pr nq
= q
q

nf

(8.21)

In this model, three parameters are obtained empirically: the minimum


liquid film thickness at dryout, min , the pair period, , and the correction factor in
the initial film thickness, C 0 . The pair period in turn contains three parameters
that need to be determined: cq , nq , and n f . In order to determine these five
parameters, Dupont et al. (2004) compared the three-zone model to 1591

138
experimental data points from the literature and performed a parametric study to
determine the optimum values of these parameters. An optimized set of values
for these parameters from least-squares fits were proposed as listed in Table 8.1.

8.2.4.1. Model Development


In the present study, model proposed by Thome et al. (2004) is modified
by using a different approach in determining the initial film thickness that is more
relevant to microchannel flow boiling. Aussillous and Quere (2000) investigated
the thickness of the liquid film left behind when a drop moves inside a capillary
tube for wetting liquids with a range of liquid viscosities. They observed three
regimes: a visco-capillary regime where the film thickness only depends on
Capillary number, a visco-inertial regime where inertia has a thickening effect on
the film and the thickness depends on both Capillary number and Weber number,
and a viscous boundary layer regime where the film thickness is limited by the
viscous boundary layer.
For the visco-capillary regime, which occurs at very low Capillary
numbers, they proposed the following correlation for film thickness:

0 (z)
Dh

0.66Ca 2/3
=
1 + 3.33Ca 2/3

(8.22)

In the viscous boundary layer regime, the film thickness is obtained by balancing
inertia and viscosity which yields

0 ( z)
Dh

L
= f f
u
f

1/ 2

(8.23)

where L f is the liquid slug length and u is the film deposition velocity. In case of
small velocity, a viscous fluid, or a long drop where the boundary layer thickness
is larger than the thickness obtained from Eq.(8.22), capillary effects are
dominant. Otherwise, the boundary layer limits the fluid deposition and Eq.
(8.23) should be used to find the film thickness.

139
In the modified model proposed in current study, the smaller value of the
film thickness obtained from Eqs. (8.22) and (8.23) is used as the film thickness
at formation, with a correction factor that takes into account the difference in
channel shape and fluid properties:
0.66 D Ca 2/3 L

h
, f f
0 ( z ) = C 0 min
2/3

1 + 3.33Ca f u p

1/2

(8.24)

In calculating the viscous boundary layer, the liquid slug and elongated bubble
pair velocity is used for the deposition velocity with the same definition as in
Thome et al. (2004).
The minimum possible film thickness is assumed to be of the same order
of magnitude as the surface roughness since the film breaks up and a dry zone
appears as the film thins to the height of the surface roughness. Agostini et al.
(2008) used the actual surface roughness instead of the value of 0.3 m
proposed in Dupont et al. (2004) and obtained much better predictions. In the
current study, the actual surface roughness values for each test piece are used
for min ; these values are listed in Table 2.1.

8.2.4.2. Model Assessment


The Thome et al. model with the original values recommended for the
adjustable parameters is compared to the slug flow data from the present work in
Figure 8.7(a). The original model is seen to generally underpredict the
experiments with an MAE of 41.2%, and only 35.9% of data are predicted to
within 30%.
In Figure 8.7(b), experimental data are compared to predictions from the
modified model proposed in this study, using Eq. (8.24) to find the liquid film
thickness. Similar to the comparison in Figure 8.7(a), the values proposed by
Thome et al. (2004) are again used for all five empirical parameters. This
comparison shows some improvement by using the modified model for film

140
thickness, with an MAE of 33.4%, and 56.4% of the data predicted to within
30%.
Next, the actual values of surface roughness are used for min , and the
values of the other four parameters C 0 , cq , nq , and n f are optimized in the
modified model to match the current experimental data, using the MATLAB
scripts in Appendix H. The optimized parameters are listed in Table 8.1. Figure
8.8 shows the comparison between the experimental data and the predictions
from this modified model with the optimized parameters. The predictions from
the modified model are found to be in good agreement with the slug flow
experimental data, with an MAE of 17.8% and 82.1% of the data predicted to
within 30%.
In Figure 8.9, the heat transfer coefficients for slug flow in microchannels
of dimensions 250 m 400 m and 400 m 400 m are plotted versus the
wall heat flux. Both the experimental data and the predictions from the modified
model with current optimized parameters are shown in this figure which illustrates
the capability of the model for prediction of the trends in the heat transfer
coefficient.

8.3. Pressure Drop


As discussed in section 7.2 and 7.3, although the empirical correlation of
Cooper (1984b) predicts the heat transfer data as well as do the flow regimebased models, the empirical correlations for pressure drop fail to predict
experimental values in microchannels. Flow regime-based modeling of the
pressure drop is discussed for confined and unconfined flow in this section,
corresponding to the heat transfer model developed above for annular flow.
Since several flow regimes co-exist along the microchannels at once, the
pressure drop of each region is calculated separately for each regime, as
discussed below.

141
8.3.1. Confined Flow
A possible arrangement of flow in each microchannel for confined flow
( Bo 0.5 Re < 160 ) is depicted in Figure 8.10. The single-phase length and the
length of the onset of annular flow can be determined from Eqs. (7.8) and (6.6),
respectively. The total pressure drop in the microchannel is the sum of the
pressure drop in the single-phase region, the slug region, and the annular region:
Pch = Psp + Ps + Pa

(8.25)

The single-phase pressure drop is calculated from Eq. (7.11). The pressure drop
over the annular region can be evaluated by integrating Eq. (8.9). The pressure
drop in the slug region cannot be readily calculated using the three-zone heat
transfer model discussed in section 8.2.4. Hence, the pressure drop in the slug
region is assumed to be similar in magnitude to the annular pressure drop, if
annular flow existed over the length of La 0 Lsp . In other words, assuming that
transition to annular flow occurs at Lsp , a grid is superposed over both the slug
and annular regions along the microchannel length and the numerical procedure
developed in section 8.2.2.1 is followed to calculate dP / dz from Eq. (8.9) at
each node. The two-phase pressure drop, Ps + Pa , is then calculated by
integrating the pressure gradient along the slug and annular flow regions. It
should be noted that Eq. (8.16) is used to determine the friction factor at the
interface.

8.3.2. Unconfined Flow


For the unconfined flow ( Bo 0.5 Re > 160 ), the two-phase flow in the
microchannels consists of bubbly flow and annular/wispy-annular flow as
illustrated in Figure 8.10(b). The total pressure drop across the microchannel is
then:
Pch = Psp + Pb + Pa

(8.26)

142
For the annular/wispy-annular region, the pressure drop is calculated from Eq.
(8.9) along with Eq. (8.17), following the numerical procedure developed earlier
for annular flow. For the bubbly flow region, pressure drop is calculated using
the single-phase methodology as in Eq. (7.11), using the homogeneous density,

H , and the homogeneous viscosity, H = x g + (1 x ) f (Cicchitti et al., 1960).

8.3.3. Model Assessment


Regime-based pressure drop predictions along the microchannels as
discussed above are compared to the experimental values for pressure drop in
Figure 8.11(a). It can be concluded that the regime-based models predict the
experiments much better than the empirical correlations reviewed in section 7.3,
with an MAE of 37.7%. A closer examination reveals that the errors are much
larger for some of the confined annular flow data in the smallest microchannels
with Bo 0.5 Re < 72 . This may be due to the approximation made in the
calculation of pressure drop in the slug flow region. Excluding these data from
the comparison, as depicted in Figure 8.11(b), results in an improved MAE of
29.2%.
Although experimental pressure drop has been predicted in the literature
(Qu and Mudawar, 2003; Lee and Garimella, 2008) with an MAE as low as
11.4%, the empirical correlations used in these predictions were developed by
fitting curves to the specific experimental data considered in the studies and the
accuracy of prediction is limited to the range of operating conditions and fluids
considered. The regime-based models, on the other hand, are expected to
extrapolate to a wider range of parameters with better accuracy.

8.4. Conclusions
Although the pool boiling correlation of Cooper was shown to predict the
experimental heat transfer data well, none of the empirical correlations

143
developed specifically for flow boiling in microchannels were found to predict
experimental heat transfer coefficient and pressure drop to within a reasonable
error (Chapter 7). Hence, it is essential to develop physics-based models based
on the relevant flow regimes to predict both heat transfer coefficient and pressure
drop in microchannel flow boiling. Physics-based models are expected to be
applicable to a wider range of parameters, and not just to specific data sets.
In the current study, models are proposed or identified for prediction of
heat transfer coefficient in each of the four regions in the comprehensive flow
regime map developed in section 6.2.3. For the bubbly flow region, the empirical
correlation of Cooper (1984b), originally developed for pool boiling, is suggested
as it results in excellent agreement with the experiments (MAE of 13.9%). For
the other three regions, physics-based models are developed. For the annular
region, an analytical model is developed which predicts the heat transfer
coefficient in confined annular flow with an MAE of 16.4% and that in
annular/wispy-annular flow with an MAE of 24.8, while capturing correct trends
for heat transfer coefficient. For slug flow, the three-zone model of Thome et al.
(2004) is modified in terms of the prediction of liquid film thickness in the
elongated bubble. This modified model predicts the experiments with an MAE of
17.8% and is able to capture trends in variation of heat transfer coefficient with
heat flux.
Knowing the location along the microchannels at which the transitions to
bubbly, slug, and annular flow occur, the pressure drop in each region can be
calculated separately. The annular flow model developed in this work is used to
calculate the pressure drop across the length of the channel where confined
annular, annular, or wispy-annular flow exists. Pressure drop in the slug flow
region of the channel is estimated with the annular flow model, while pressure
drop in the bubbly flow region is calculated using the homogeneous model. It is
shown that the regime-based prediction of pressure drop results in much better
agreement with experiment than is possible with the empirical correlations.

144
To improve these regime-based models, it is necessary to determine the
bubble generation frequency and liquid film thickness in the slug region
analytically, and account for the vapor-liquid film interfacial phenomena in the
annular flow. Pressure drop predictions, using regime-based methods, are very
sensitive to the length of different flow regimes in the microchannels; hence,
regime maps capable of accurately determining the transition points should be
used.

145

Table 8.1. Proposed values for the empirical parameters in the three-zone model
for slug flow.

Empirical
parameters

Dupont et
al. (2004)

Current
study

min

0.3

N/A

C 0

0.29

1.31

cq

3328

1.18108

nq

-0.5

3.26

nf

1.74

1.64

146

10000
+30%

h pred (W/m2K)

8000

6000

-30%

4000

2000
MAE = 13.9%
86.5% of predictions within error of 30%
0

2000

4000

6000
2

8000

10000

h exp (W/m K)

Figure 8.1. Comparison of bubbly flow experimental heat transfer coefficients


with predictions from the Cooper correlation (1984b).

147

(a)

Vapor core
CV

Liquid film
CV

(b)
Figure 8.2. (a) Schematic representation of annular flow in microchannels, and
(b) simplified flow diagram with vapor core and liquid film control volumes.

148

8000
+30%
7000

h pred (W/m2K)

6000
5000

-30%

4000
3000
2000
MAE = 16.4%

1000

87.2% of predictions within error of 30%


0

1000

2000

3000

4000

5000

6000

7000

8000

h exp (W/m K)

Figure 8.3. Comparison of confined annular flow experimental heat transfer


coefficients with predictions from the proposed model.

149

7000

4000

G = 1050 kg/m2s
w = 100 m
d = 400 m

G = 225 kg/m s
w = 100 m
d = 400 m

6000
5000

h (W/m2 K)

h (W/m2 K)

3000

2000

3000
2000

1000

hexp
hpred

MAE = 8.7%
0

4000

20

40

60

80

100

1000

120

hexp
hpred

MAE = 11.4%
0

50

100

150

200

250

q"w (kW/m2 )

q"w (kW/m2)

(a)

(b)

7000
G = 1420 kg/m2 s
w = 100 m
d = 400 m

6000

h (W/m2 K)

5000
4000
3000
2000
1000
0

hexp
hpred

MAE = 4.1%
0

50

100

150

200

250

q"w (kW/m2)

(c)
Figure 8.4. Variation of heat transfer coefficient with wall heat flux for confined
annular flow. Experimental results and predictions from the proposed model are
included.

150

12000
+30%

h pred (W/m2K)

10000

8000
-30%
6000

4000

2000

MAE = 24.8%
71.9% of predictions within error of 30%

2000

4000

6000

8000

10000

12000

h exp (W/m K)

Figure 8.5. Comparison of the annular/wispy annular flow experimental heat


transfer coefficients with predictions from the proposed model.

151

7000
G = 630 kg/m2s
w = 400 m
d = 400 m

6000

h (W/m K)

5000
4000
3000
2000
1000
0

hexp
hpred

MAE = 17.6%
0

50

100

150

200

250

300

q"w (kW/m )

Figure 8.6. Variation of heat transfer coefficient with wall heat flux for
annular/wispy annular flow. Experimental results and predictions from the
proposed model are included.

152

5000
+30%

h pred (W/m K)

4000

3000

-30%

2000

1000
MAE = 41.2%
35.9% of predictions within error of 30%
0

1000

2000

3000
2

4000

5000

h exp (W/m K)

(a)
5000
+30%

h pred (W/m K)

4000

3000

-30%

2000

1000
MAE = 33.4%
56.4% of predictions within error of 30%
0

1000

2000

3000
2

4000

5000

h exp (W/m K)

(b)
Figure 8.7. Comparison of slug flow experimental heat transfer coefficients with
predictions from (a) original three-zone model (Thome et al., 2004),and (b)
modified three-zone model; both predictions use values proposed in (Dupont et
al., 2004) for the five empirical parameters.

153

5000
+30%

h pred (W/m2K)

4000

3000

-30%

2000

1000
MAE = 17.8%
82.1% of predictions within error of 30%
0

1000

2000

3000
2

4000

5000

h exp (W/m K)

Figure 8.8. Comparison of slug flow experimental heat transfer coefficients with
predictions from the modified three-zone model, using empirical parameters
optimized for the current data.

154

4
G = 225 kg/m2s
w = 250 m
d = 400 m

3.5

2.5

h (kW/m K)

2
1.5
1
0.5
0

hexp
hpred

MAE = 11.4%
0

20

40

60

80

100

q"w (kW/m )

(a)
4
G = 225 kg/m2s
w = 400 m
d = 400 m

3.5

2.5

h (kW/m K)

2
1.5
1
0.5
0

hexp
hpred

MAE = 7.7%
0

20

40

60

80

100

q"w (kW/m )

(b)
Figure 8.9. Variation of heat transfer coefficient with wall heat flux for slug flow:
comparison of experiments with predictions from the modified three-zone model,
using empirical parameters optimized for the current data.

155

Lsp

La0

(a)
Lsp

La0

(b)
Figure 8.10. Schematic representation of flow in microchannels for (a) confined
flow, and (b) unconfined flow.

156

60
+30%
50

Ppred (kPa)

40
-30%
30

20

10
MAE = 37.7%
46.7% of predictions within error of 30%
0

10

20

30

P exp (kPa)

40

50

60

(a)
60
+30%
50

P pred (kPa)

40
-30%
30

20

10
MAE = 29.2%
58.7% of predictions within error of 30%
0

10

20

30

40

50

60

P exp (kPa)

(b)
Figure 8.11. Comparison of experimental pressure drops with predictions from
the proposed model: (a) including all the experimental data, and (b) excluding
data for which Bo 0.5 Re < 72 .

157

CHAPTER 9. CONCLUSIONS AND RECOMMENDATIONS

In this chapter, the main conclusions of the present work are summarized
and recommendations for future work provided.

9.1. Conclusions
In the present study, boiling in microchannels is investigated with a
dielectric liquid, FC-77. Extensive experimental work has been conducted with
microchannel test pieces encompassing a wide range of channel dimensions and
over a broad range of operating conditions to systematically determine the
effects of important geometric and flow parameters on flow regimes, heat
transfer, and pressure drop associated with microscale flow boiling. Local heat
transfer measurements obtained with simultaneous, detailed flow visualizations
have led to a better understanding of boiling phenomena in microchannels and
the governing heat transfer mechanisms, and facilitated physics-based analyses
of local heat transfer and pressure drop. The extensive microscale boiling
experiments and analyses have resulted in a comprehensive understanding of
boiling, with some of the significant findings listed below:

A large database of boiling flow pattern visualizations is obtained for a


wide range of channel dimensions and flow parameters, leading to a
good understanding of microscale flow regimes.

A clear knowledge of the effects of microchannel dimensions on flow


boiling is obtained; the cross-sectional area of the microchannels is
found to play a determining role in boiling mechanisms and heat
transfer.

158

Vapor confinement and emergence of microscale effects are shown to


depend not only on channel size and fluid properties, but also on the
flow rate. Based on the experimental results, a new transition criterion
is developed which predicts the conditions under which microscale
confinement effects are exhibited in flow boiling.

For conditions under which flow confinement does not occur, the heat
transfer and boiling curves are independent of, and pressure drop and
pumping power have only minor sensitivity to, channel size and flow
rate.

In confined flow, the heat transfer coefficient increases as the crosssectional area of the microchannels decreases. Thin liquid film
evaporation in confined boiling results in larger values of heat transfer
coefficients relative to unconfined flow where nucleate boiling is
dominant.

Two types of flow regime maps with coordinate systems conventionally


used for flow boiling are developed on mass flux-vapor quality and
liquid-vapor superficial velocity coordinates. It is shown that these types
of flow regime maps depend on microchannel size; the effects of
channel dimensions on flow transitions are discussed.

Two comprehensive flow regime maps using new coordinate systems


are developed for flow boiling of FC-77, along with quantitative regimetransition criteria, based on approximately 390 data points
encompassing a wide range of microchannel dimensions, mass fluxes,
and heat fluxes. These maps are presented in terms of four regions
slug, confined annular, bubbly, and alternating churn/annular/wispyannular flow. The comprehensive flow maps facilitate the development
of flow regime-based models for the prediction of boiling heat transfer
coefficients.

The heat transfer coefficients obtained from the current experiments are
compared to predictions from a number of empirical correlations in the

159
literature. The correlation of Cooper (1984b) for nucleate pool boiling
predicts the experimental heat transfer coefficients very well, especially
in the nucleate boiling region, with an average mean absolute
percentage error of 11.9%.

The experimental results for pressure drop are also compared with
predictions from empirical correlations in the literature. Large errors
obtained in these comparisons reveal the failure of these empirical
correlations in providing a suitable prediction of the experimental
results, mainly because the correlations were developed for specific
fluids and ranges of operating parameters that differ from those of the
current experimental data.

Although the pool boiling correlation of Cooper was shown to predict the
experimental heat transfer data well, none of the empirical correlations
developed specifically for flow boiling in microchannels were found to
predict experimental heat transfer coefficient and pressure drop to
within a reasonable error. Hence, it is essential to develop physicsbased models based on the relevant flow regimes to predict both heat
transfer coefficient and pressure drop in microchannel flow boiling.

Models are proposed or identified for prediction of heat transfer


coefficient in each of the four regions in the comprehensive flow regime
maps. For the bubbly flow region, the empirical correlation of Cooper
(1984b), originally developed for pool boiling, is suggested as it results
in excellent agreement with the experiments (MAE of 13.9%).

For the annular region, an analytical model is developed which predicts


the heat transfer coefficient in confined annular flow with an MAE of
16.4% and that in annular/wispy-annular flow with an MAE of 24.8%,
while capturing correct trends for heat transfer coefficient. For slug flow,
the three-zone model of Thome et al. (2004) is modified in terms of the
prediction of liquid film thickness in the elongated bubble. This modified

160
model predicts the experiments with an MAE of 17.8% and is able to
capture trends in variation of heat transfer coefficient with heat flux.

Knowing the location along the microchannels at which the transitions to


bubbly, slug, and annular flow occur from the developed flow regime
map, the pressure drop in each region can be calculated separately.
The annular flow model developed in this work is used to calculate the
pressure drop across the length of the channel where confined annular,
annular, or wispy-annular flow exists. Pressure drop in the slug flow
region of the channel is estimated with the annular flow model, while
pressure drop in the bubbly flow region is calculated using the
homogeneous model. It is shown that the regime-based prediction of
pressure drop results in much better agreement with experiment than is
possible with the empirical correlations.

9.2. Suggestions for Future Work


Recent experimental investigations and analyses in the authors group
(Chen and Garimella, 2006a; Chen and Garimella, 2006c; Liu and Garimella,
2007; Lee and Garimella, 2008; Bertsch et al., 2008b; Bertsch et al., 2009; Jones
et al., 2009) including the present work have led to a more comprehensive
understanding of the physical mechanisms and parameter dependencies in
microchannel flow boiling for several liquids, and have facilitated development of
predictive methods for flow boiling heat transfer over a wide parameter space.
The effects of important geometric parameters, operating conditions, and fluid
properties on the flow regimes and thermal performance of microchannels have
been determined through extensive experimental work. In this section, some
suggestions for the future work are presented based on the work done in the
current study.
While the convective confinement number criterion for microscale effects
and the comprehensive flow regime map developed in the current work have

161
been demonstrated to be very promising tools for predicting the flow regimes and
their transitions in microchannel boiling, only limited studies are available for
comparison, as both visualized flow boiling patterns in microchannels as well as
heat flux data are necessary for such comparisons. As more well-characterized
data become available, it will be possible to further generalize this confinement
criterion and to expand the applicability of the comprehensive flow regime map to
other fluids.
In the present study, the three-zone model of Thome et al. (2004) for slug
flow is modified for prediction of heat transfer coefficient in the slug region. In
this model, the initial liquid film thickness at the formation of the elongated
bubble, the minimum liquid film thickness at dryout, and the bubble frequency
need to be determined; however, these parameters are obtained empirically
without a fundamental analysis. For the initial film thickness surrounding the
elongated bubble, Thome et al. (2004) used an empirical relation developed by
Moriyama and Inoue (1996) for thickness of a liquid film formed by a bubble
growing radially in a gap between two parallel heated plates. In the current
modified model, a different method for prediction of the initial liquid film thickness,
originally developed for the film thickness in capillary tubes by Aussillous and
Quere (2000), is used which is more relevant to microchannel flow boiling; the
modified model shows some improvement comparing to the original model.
However, in order to improve the analytical models for heat transfer, a careful
experimental and analytical study is necessary to develop methods for prediction
of the liquid film thickness surrounding an elongated bubble in microchannels for
a wide range of operating conditions and fluid properties.
Although expressions are available to determine the bubble nucleation
frequency in macrochannels, methods are not available in the literature to
determine the bubble departure frequency in microchannels. To improve the
regime-based models for heat transfer in the slug regime, it is necessary to
determine the bubble generation frequency in the slug region analytically.
Bubble departure frequency is complicated by the nucleation process; hence,

162
detailed experimental investigation needs to be conducted to develop a better
understanding of the nucleation process and propose predictive models for
bubble departure frequency.
In the current study, a simplified analytical model is developed to predict
heat transfer and pressure drop in the annular flow regime. To improve this
model, the vapor-liquid film interfacial phenomena in the annular flow should be
investigated carefully and taken into account in the model. The current study has
also shown that the model results in better accuracy without considering the
liquid droplet deposition even in the wispy-annular region in which flow
visualizations reveal the presence of entrained droplet in the vapor core. The
criteria under which droplet deposition to the liquid film occurs have not been
studied for microchannels and needs to be further investigated.
Flow visualizations of the current study reveal that annular or wispyannular flow are not always steady and often alternate with churn flow. In the
annular model developed in the current study, it is assumed that the effect of film
evaporation in the annular/wispy-annular flow is more dominant than the nucleate
boiling heat transfer in the churn flow in determining the heat transfer coefficient.
To examine the effects of churn flow on heat transfer and possibly improve the
model in the alternating churn/annular flow, a superposition approach for the heat
transfer mechanisms in the annular and churn flow regimes should be
developed. Due to complexity of the heat transfer mechanisms in churn flow, a
nucleate boiling model such as correlation of Cooper (1984b) may be used as an
approximation to estimate the heat transfer.
A regime-based approach has been presented here to predict pressure
drop in microchannel flow boiling. Once the location along the microchannels at
which the transitions to different flow regimes occur is determined from the
developed flow maps, the pressure drop in each region can be calculated
separately. The annular flow model developed in this work is used to calculate
the pressure drop across the length of the channel where confined annular,
annular, or wispy-annular flow exists; however, pressure drop in the slug flow

163
region of the channel is estimated with the annular flow model, while pressure
drop in the bubbly flow region is calculated using the homogeneous model. To
improve the pressure drop predictions, models need to be developed for
pressure drop predictions in the slug and bubbly flow regimes.
Void fraction is one of the key parameters that characterize flow regimes;
hence, the measurement of void fraction and its temporal and spatial variation is
of great importance in experimental investigations. It is important to develop void
fraction measurement techniques for microchannels to aid in the development
and improvement of theoretical models.

LIST OF REFERENCES

164

LIST OF REFERENCES

3M Fluorinert Liquids Product Manual, 1999.


Agostini, B., Thome, J. R., Fabbri, M., Michel, B., Calmi, D., and Kloter, U., 2008,
High Heat Flux Flow Boiling in Silicon Multi-Microchannels - Part II: Heat
Transfer Characteristics of Refrigerant R235fa, International Journal of Heat and
Mass Transfer, 51, pp. 5415-5425.
Aussillous, P. and Quere, D., 2000, Quick Deposition of Fluid on the Wall of a
Tube, Physics of Fluids, 12(10), pp. 2367-2371.
Baker, O., 1954, Design of Pipe Lines for Simultaneous Flow of Oil and Gas,
Oil and Gas Journal, 26.
Bar-Cohen, A. and Rahim, E., 2009,Modeling and Prediction of Two-Phase
Microgap Channel Heat Transfer Characteristics, Heat Transfer Engineering, 30
(8), pp. 601-625.
Bennett, D. L. and Chen, J. C., 1980, Forced Convective Boiling in Vertical
Tubes for Saturated Pure Components and Binary Mixtures, American Institute
of Chemical Engineers Journal, 26 (3), pp. 454-461.
Bergles, A. E., Lienhard, J. H., Kendall, G. E., and Griffith, P., 2003, Boiling and
Evaporation in Small Diameter Channels, Heat Transfer Engineering, 24 (1), pp.
18-40.
Bertsch, S. S., Groll, E. A., and Garimella, S. V., 2008a, Review and
Comparative Analysis of Studies on Saturated Flow Boiling in Small Channels,
Nanoscale and Microscale Thermophysical Engineering, 12 (3), pp. 187-227.
Bertsch, S. S., Groll, E. A., and Garimella, S. V., 2008b, Refrigerant Flow Boiling
Heat Transfer in Parallel Microchannels as a Function of Local Vapor Quality,
International Journal of Heat and Mass Transfer, 51, pp. 4775-4787.
Bertsch, S. S., Groll, E. A., and Garimella, S. V., 2008c, A Composite Heat
Transfer Correlation for Saturated Flow Boiling Heat Transfer in Small Channels,
International Journal of Heat and Mass Transfer, in press.

165
Bertsch, S. S., Groll, E. A., and Garimella, S. V., 2008d, Effects of Heat Flux,
Mass Flux, Vapor Quality, and Saturation Temperature on Flow Boiling Heat
Transfer in Microchannels, International Journal of Multiphase Flow, in press.
Bertsch, S. S., Groll, E. A., and Garimella, S. V., 2009, Effects of Heat Flux,
Mass Flux, Vapor Quality, and Saturation Temperature on Flow Boiling Heat
Transfer in Microchannels, International Journal of Multiphase Flow, 35, 142-154.
Blevins R. D., Applied Fluid Dynamics Handbook, Krieger Pub. Co., 1992, pp.
7778.
Carey, V. P., 1992, Liquid-Vapor Phase Change Phenomena, Hemisphere
Publishing Corp., Washington, pp. 483-490.
Chen, J. C., 1966, Correlation for Boiling Heat Transfer to Saturated Fluids in
Convective Flow, Industrial and Engineering Chemistry Process Design and
Development, 5 (3), pp. 322-329.
Chen, T. and Garimella, S. V., 2006a, Measurements and High-Speed
Visualization of Flow Boiling of a Dielectric Fluid in a Silicon Microchannel Heat
Sink, International Journal of Multiphase Flow, 32 (8), pp. 957-971.
Chen, T. and Garimella S. V., 2006b, Effect of Dissolved Air on Subcooled Flow
Boiling of a Dielectric Coolant in a Microchannel Heat Sink, Journal of Electronic
Packaging, 128 (4), pp. 398-404.
Chen, T. and Garimella, S. V., 2006c, Flow Boiling Heat Transfer to a Dielectric
Coolant in a Microchannel Heat Sink, IEEE Transactions on Components and
Packaging Technologies, 30 (1), pp. 24-31.
Cheng, L., Ribatski, G., Quiben, J. M., and Thome, J. R., 2008, New Prediction
Methods for CO2 Evaporation Inside Tubes: Part I A Two-Phase Flow Pattern
Map and a Flow Pattern Based Phenomenological Model for Two-Phase Flow
Frictional Pressure Drops, International Journal of Heat and Mass Transfer, 51,
pp. 111-124.
Chisholm, D. and Sutherland, L. A., 1969, Prediction of Pressure Gradients in
Pipeline Systems during Two-Phase Flow, in: Symposium in Two-phase Flow
Systems, University of Leeds.
Cicchitti, A., Lombardi, C., Silverstri, M., and Zavattarelli, G. S. R., 1960, TwoPhase Cooling Experiments - Pressure Drop, Heat Transfer and Burnout
Measurements, Energia Nucleare, 7 (6), pp. 407-425.

166
Coleman, J. W. and Garimella S., 2003, Two-Phase Flow Regimes in Round
Square and Rectangular Tubes during Condensation of Refrigerant R134a,
International Journal of Refrigeration, 26, pp. 117-128.
Collier J. G. and Thome J. R., Convective Boiling and Condensation, Third ed.,
Oxford University Press, Oxford, 1994.
Cooper, M. G., 1984a, Saturated Nucleate Pool Boiling - a Simple Correlation,
1st UK National Heat Transfer Conference, IChemE Symposium Series, 2 (86),
pp. 785-793.
Cooper, M. G., 1984b, Heat Flow Rates in Saturated Nucleate Pool Boiling - a
Wide-Ranging Examination Using Reduced Properties, Advances in Heat
Transfer, 16, pp. 157-239.
Dupont, V., Thome, J. R., and Jacobi, A. M., 2004, Heat Transfer Model for
Evaporation in Microchannels. Part II: Comparison with the Database,
International Journal of Heat and Mass Transfer, 47, pp. 3387-3401.
Dupont, V. and Thome, R., 2005, Evaporation in Microchannels: Influence of the
Channel Diameter on Heat Transfer, Microfluidics and Nanofluidics, 1 (2), pp.
119-127.
Field, B. and Hrnjak, P., 2007, Visualization of Two-Phase Refrigerant and
Refrigerant-Oil Flow in a Microchannel, ASME International Mechanical
Engineering Congress and Exposition, IMECE2007-43471, Seattle, WA.
Garimella, S. V. and Harirchian, T., 2009, Boiling Heat Transfer and Flow
Regimes in Microchannels a Comprehensive Understanding, Keynote lecture
at Thermal Investigations of ICs and Systems, THERMINIC 2009, October 7-9,
Leuven, Belgium.
Garimella, S. V., Singhal, V., and Liu, D., 2006, On-Chip Thermal Management
with Microchannel Heat Sinks and Integrated Micropumps, Proceedings of the
IEEE, 94 (8), pp. 1534-1548.
Garimella, S. V. and Sobhan, C. B., 2003, Transport in Microchannels - a Critical
Review, Annual Review of Heat Transfer, 13, pp. 1-50.
Geisler, K. J. L. and Bar-Cohen, A., 2009, Confinement Effects on Nucleate
Boiling and Critical Heat Flux in Buoyancy-Driven Microchannels, International
Journal of Heat and Mass Transfer, 52 (11-12), 2427-2436.
Gorenflo, D., 1993, in VDI Heat Atlas, Chapter Ha.

167
Gungor, K. E. and Winterton, R. H. S., 1986, General Correlation for Flow
Boiling in Tubes and Annuli, International Journal of Heat and Mass Transfer, 29
(3), pp. 351-358.
Harirchian, T. and Garimella, S. V., 2007, Microchannel Size Effects on TwoPhase Local Heat Transfer and Pressure Drop in Silicon Microchannel Heat
Sinks with a Dielectric Fluid, Proceedings of the ASME International Mechanical
Engineering Congress and Exposition, IMECE2007, 11, pp. 437-446.
Harirchian, T. and Garimella, S. V., 2008a, Microchannel Size Effects on Local
Flow Boiling Heat Transfer to a Dielectric Fluid, International Journal of Heat and
Mass Transfer, 51, pp. 3724-3735.
Harirchian, T. and Garimella, S. V., 2008b, An Investigation of Flow Boiling
Regimes in Microchannels of Different Sizes by Means of High-Speed
Visualization, 11th IEEE Intersociety Conference on Thermal and
Thermomechanical Phenomena in Electronic Systems, I-THERM, pp. 197-206.
Harirchian, T. and Garimella, S. V., 2008c, Flow Boiling in Silicon Microchannel
Heat Sinks, Proceedings of the Annual IEEE Semiconductor Thermal
Measurement and Management Symposium, SEMI-THERM, pp. 1-8.
Harirchian, T. and Garimella, S. V., 2008d, Flow Patterns during Convective
Boiling in Microchannels, Journal of Heat Transfer, 130 (8), 080909-1.
Harirchian, T. and Garimella, S. V., 2009a, Effects of Channel Dimension, Heat
Flux, and Mass Flux on Flow Boiling Regimes in Microchannels, International
Journal of Multiphase Flow, 35, pp. 349-362.
Harirchian, T. and Garimella, S. V., 2009b, The Critical Role of Channel CrossSectional Area in Microchannel Flow Boiling Heat Transfer, International Journal
of Multiphase Flow, 35, pp. 904-913.
Harirchian, T. and Garimella, S. V., 2010, A Comprehensive Flow Regime Map
for Microchannel Flow Boiling with Quantitative Transition Criteria, International
Journal of Heat and Mass Transfer, 53, pp. 2694-2702.
Harirchian, T. and Garimella, S. V., Boiling Heat Transfer and Flow Regimes in
Microchannels a Comprehensive Understanding, Journal of Electronic
Packaging, accepted for publication.
Hassan, I., Vaillancourt, M., and Pehlivan, K., 2005, Two-Phase Flow Regime
Transitions in Microchannels: a Comparative Experimental Study, Microscale
Thermophysical Engineering, 9, pp. 165-182.

168
Haynes, B. S. and Fletcher, D. F., 2003, Subcooled Flow Boiling Heat Transfer
in Narrow Passages, International Journal of Heat and Mass Transfer, 46, pp.
3673-3682.
Hetsroni, G., Mosyak, A., Segal, Z., and Pogrebnyak, E., 2003, Two-Phase Flow
Patterns in Parallel Microchannels, International Journal of Multiphase Flow, 29,
pp. 341-360.
Hetsroni, G., Mosyak, A., Segal, Z., and Ziskind, G., 2002, A Uniform
Temperature Heat Sink for Cooling of Electronic Devices, International Journal
of Heat and Mass Transfer, 45, pp. 3275-3286.
Hewitt, G. F. and Robert, D. N., 1969, Studies of Two-Phase Flow Patterns by
Simultaneous X-Ray and Flash Photography, AERE-M 2159, HMSO.
Honda, H. and Wei, J. J., 2004, Enhanced Boiling Heat Transfer from Electronic
Components by Use of Surface Microstructures, Experimental Thermal and
Fluid Science, 28, pp. 159-169.
Huo, X., Chen, L., Tian, Y.S., and Karayiannis, T. G., 2004, Flow Boiling and
Flow Regimes in Small Diameter Tubes, Applied Thermal Engineering, 24, pp.
1225-1239.
Incropera, F. P., DeWitt, D. P., Bergman, T. L., Lavine, A. S., 2007,
Fundamentals of Heat and Mass Transfer, 6th Ed., John Wiley & Sons.
Jiang, L., Wong, M., and Zohar, Y., 2001, Forced Convection Boiling in a
Microchannel Heat Sink, Journal of Microelectromechanical Systems, 10, pp.
80-87.
Jones, B. J. and Garimella, S. V., 2007, Effects of Surface Roughness on the
Pool Boiling of Water, Proceedings of the ASME/JSME Thermal Summer Heat
Transfer Conference, 2, pp 219-227.
Jones, B. J., McHale, J. P., and Garimella, S. V., 2009, The Influence of Surface
Roughness on Nucleate Pool Boiling Heat Transfer Journal of Heat Transfer,
131, 121009.
Kandlikar, S. G., 1991, A Model for Correlating Flow Boiling Heat Transfer in
Augmented Tubes and Compact Evaporators, Journal of Heat Transfer, 113, pp.
966-972.
Kandlikar, S., 2002, Fundamental Issues Related to Flow Boiling in
Minichannels and Microchannels, Experimental Thermal and Fluid Science, 26
(2-4), pp. 389-407.

169
Kandlikar, S., 2004, Heat Transfer Mechanisms During Flow Boiling in
Microchannels, Journal of Heat Transfer, 126, pp. 8-16.
Kew, P. A. and Cornwell, K., 1997, Correlations for the Prediction of Boiling Heat
Transfer in Small-Diameter Channels, Applied Thermal Engineering, 17, pp.
705-715.
Lazarek, G. M. and Black, S. H., 1982, Evaporative Heat Transfer, Pressure
Drop and Critical Heat Flux in a Small Vertical Tube with R-113, International
Journal of Heat and Mass Transfer, 25 (7), pp. 945-960.
Lee, P-S., 2007, Heat Transfer in Silicon Microchannel Arrays, PhD Thesis,
Purdue University.
Lee, M., Cheung, L. S. L., Lee, Y., and Zohar, Y., 2005, Height Effect on
Nucleation-Site Activity and Size-Dependent Bubble Dynamics in Microchannel
Convective Boiling, Journal of Micromechanics and Microengineering, 15, pp.
2121-2129.
Lee, P-S. and Garimella, S. V., 2005, Investigation of Heat Transfer in
Rectangular Microchannels, International Journal of Heat and Mass Transfer,
48, pp. 1688-1704.
Lee, P-S. and Garimella, S. V., 2008, Saturated Flow Boiling Heat Transfer and
Pressure Drop in Silicon Microchannel Arrays, International Journal of Heat and
Mass Transfer, 51 (3-4), pp. 789-806.
Lee, H. J. and Lee, S. Y., 2001, Heat Transfer Correlation for Boiling Flows in
Small Rectangular Horizontal Channels with Low Aspect Ratios, International
Journal of Multiphase Flow, 27, pp. 2043-2062.
Lee, J. and Mudawar, I., 2005, Two-Phase Flow in High-Heat-Flux MicroChannel Heat Sink for Refrigeration Cooling Applications: Part I Pressure Drop
Characteristics, International Journal of Heat and Mass Transfer, 48, pp. 928940.
Lee, J. and Mudawar, I., 2008, Fluid Flow and Heat Transfer Characteristics of
Low Temperature Two-Phase Micro-Channel Heat Sinks Part 1: Experimental
Methods and Flow visualization Results, International Journal of Heat and Mass
Transfer, 51, pp. 4315-4326.
Lee, M., Wong, Y. Y., Wong, M., and Zohar, Y., 2003, Size and Shape Effects
on Two-Phase Flow Patterns in Microchannel Forced Convection Boiling,
Journal of Micromechanics and Microengineering, 13, pp. 155-164.

170
Lin, S., Kew, P. A., and Cornwell, K., 2001, Flow Boiling of Refrigerant R141b in
Small Tubes, Chemical Engineering Research and Design, 79 (4), pp. 417-424.
Liu, D. and Garimella, S. V., 2004, Investigation of Liquid Flow in
Microchannels, Journal of Thermophysics and Heat Transfer, 18 (1), pp. 65-72.
Liu, D. and Garimella, S. V., 2007, Flow Boiling Heat Transfer in Microchannels,
Journal of Heat Transfer, 129 (10), pp. 1321-1332.
Liu, D., Lee, P. S., and Garimella, S. V., 2005, Prediction of the Onset of
Nucleate Boiling in Microchannel Flow, International Journal of Heat and Mass
Transfer, 48, pp. 5134-5149.
Liu, Z. and Winterton, R. H. S., 1991, A General Correlation for Saturated and
Subcooled Flow Boiling in Tubes and Annuli, Based on a Nucleate Pool Boiling
Equation, International Journal of Heat and Mass Transfer, 34 (11), pp. 27592766.
Lockhart, R. W. and Martinelli, R. C., 1949, Proposed Correlation of Data for
Isothermal Two-Phase, Two-Component Flow in Pipes, Chemical Engineering
Progress, 45, pp. 3948.
McHale, J. P. and Garimella, S. V., 2008, Measurement of Bubble Nucleation
Characteristics in Pool Boiling of a Wetting Liquid on Smooth and Roughened
Surfaces, 2008 ASME Summer Heat Transfer Conference, Jacksonville, Fl,
USA.
Mandhane, J. M., Gregory, G. A., and Aziz, K., 1974, A Flow Pattern Map for
Gas-Liquid Flow in Horizontal Pipes, International Journal of Multiphase Flow, 1,
p. 537-553.
Mauro, A. W., Quiben, J. M., Mastrullo, R., and Thome, J. R., 2007, Comparison
of Experimental Pressure Drop Data for Two-Phase Flows to Prediction Methods
Using a General Model, International Journal of Refrigernation, 30, pp. 13581367.
Mishima, K. and Hibiki, T., 1996, Some Characteristics of AirWater Two-Phase
Flow in Small Diameter Vertical Tubes, International Journal of Multiphase Flow,
22, pp. 703712.
Mishima, K., Ishii, M. 1984, Flow Regime Transition Criteria for Upward TwoPhase Flow in Vertical Tubes, International Journal of Heat and Mass Transfer,
27 (5), 723-737.

171
Moore, G. E., 1965, Cramming more components onto integrated circuits,
Electronics Magazine, 38 (8).
Moriyama, K. and Inoue, A., 1996 , Thickness of the Liquid Film Formed by a
Growing Bubble in a Narrow Gap between Two Horizontal Plates, Journal of
Heat Transfer, 118, pp. 132139.
Mukherjee, S. and Mudawar I., 2003, Smart Pumpless Loop for Micro-Channel
Electronic Cooling Using Flat and Enhanced Surfaces, IEEE Transactions on
Components and Packaging Technologies, 26 (1), 99-109.
Ong, C. L. and Thome, J. R., 2009, Flow Boiling Heat Transfer of R134a,
R236fa and R245fa in a Horizontal 1.030 mm Circular Channel, Experimental
Thermal and Fluid Science, 33, 651-663.
Paleev, I. I. and Filippovich, B. S., 1966, Phenomena of Liquid Transfer in TwoPhase Dispersed Annular Flow, International Journal of Heat and Mass
Transfer, 9, pp. 10891093.
Pate, D. P., Jones, R. J., and Bhavnani, S. H., 2006, Cavity-Induced Two-Phase
Heat Transfer in Silicon Microchannels, Proceedings of the Intersociety
Conference on Thermal and Thermomechanical Phenomena and Emerging
Technologies in Electronic Systems, pp. 71-78.
Peles, Y. P., Yarin, L. P., and Hetsroni, G., 1999, Evaporating Two-Phase Flow
Mechanism in Microchannels, Proceedings of SPIE 3680, 226-236.
Peng, X. F. and Wang, B. X., 1993, Forced Convection and Flow Boiling Heat
Transfer for Liquid Flowing Through Microchannels, International Journal of
Heat and Mass Transfer, 36 (14), pp. 3421-3427.
Peters, J. V. S. and Kandlikar, S. G., 2007, Further Evaluation of a Flow Boiling
Correlation for Microchannels and Minichannels, Proceedings of the Fifth
International Conference on Nanochannels, Microchannels and Minichannels,
ICNMM2007, June 18-20, 2007, Puebla, Mexico.
Qu, W. and Mudawar, I., 2003a, Measurement and Prediction of Pressure Drop
in Two-Phase Micro-Channel Heat Sinks, International Journal of Heat and
Mass Transfer, 46, pp. 2737-2753.
Qu, W. and Mudawar, I., 2003b, Flow Boiling Heat Transfer in Two-Phase
Microchannel Heat Sinks I. Experimental Investigation and Assessment of
Correlation Methods, International Journal of Heat and Mass Transfer, 46, pp.
2755-2771.

172
Qu, W. and Mudawar, I., 2003c, Flow Boiling Heat Transfer in Two-Phase
Microchannel Heat Sinks II. Annular Two-Phase Flow Model, International
Journal of Heat and Mass Transfer, 46, pp. 2773-2784.
Qu, W. and Mudawar, I., 2004, Transport Phenomena in Two-Phase MicroChannel Heat Sinks, Journal of Electronic Packaging, 126, pp. 213-224.
Quiben, J. M. and Thome, J. R., 2007a, Flow Pattern Based Two-Phase
Frictional Pressure Drop Model for Horizontal Tubes, Part I: Diabatic and
Adiabatic Experimental Study, International Journal of Heat and Fluid Flow, 28,
1049-1059.
Quiben, J. M. and Thome, J. R., 2007b, Flow Pattern Based Two-Phase
Frictional Pressure Drop Model for Horizontal Tubes, Part II: New
Phenomenological Model, International Journal of Heat and Fluid Flow, 28, pp.
1060-1072.
Revellin, R., Dupont, V., Ursenbacher, T., Thome, J. R., and Zun, I., 2006,
Characterization of Diabatic Two-Phase Flows in Microchannels: Flow
Parameter Results for R-134a in a 0.5 mm Channel, International Journal of
Multiphase Flow, 32, pp. 755-774.
Revellin, R. and Thome, J. R., 2007, A New Type of Diabatic Flow Pattern Map
for Boiling Heat Transfer in Microchannels, Journal of Micromechanics and
Microengineering, 17, 788-796.
Ribatski, G., Zhang, W., Consolini, L., Xu, J., and Thome, J. R., 2007, On the
Prediction of Heat Transfer in Micro-Scale Flow Boiling, Heat Transfer
Engineering, 28 (10), pp. 842-851.
Saha, P., Ishii, M., and Zuber, N., 1976, An Experimental Investigation of the
Thermally Induced Flow Oscillations in Two-Phase Systems, Journal of Heat
Transfer, 98, pp. 616-622.
Saitoh, S., Daiguji, H., and Hihara, E., 2005, Effect of Tube Diameter on Boiling
Heat Transfer of R-134a in Horizontal Small-Diameter Tubes, International
Journal of Heat and Mass Transfer, 48 (23-24), pp. 4973-4984.
Saitoh, S., Daiguji, H., and Hihara, E., 2007, Correlation for Boiling Heat
Transfer of R-134a in Horizontal Tubes Including Effect of Tube Diameter,
International Journal of Heat and Mass Transfer, 50, pp. 5215-5225.
Serizawa A., Feng Z., and Kawara, Z., 2002, Two-phase Flow in
Microchannels, Experimental Thermal and Fluid Science, 26 (6-7), 703-714.

173
Shah, M. M., 1977, General Correlation for Heat Transfer During Subcooled
Boiling in Pipes and Annuli, ASHRAE Transactions, 83 (Pt 1), pp. 202-217.
VDI-Warmeatlas, Springer-Verlag, Berlin, Heidelberg, 1997.
Shiferaw, D., Huo, X., Karayiannis, T. G., and Kenning, D. B. R., 2007,
Examination of Heat Transfer Correlations and a Model for Flow Boiling of
R134a in Small Diameter Tubes, International Journal of Heat and Mass
Transfer, 50, pp. 5177-5193.
Sobhan, C. B. and Garimella, S. V., 2001, A Comparative Analysis of Studies on
Heat Transfer and Fluid Flow in Microchannels, Microscale Thermophysical
Engineering, 5, pp. 293-311.
Steiner, D. and Taborek, J., 1992, Flow Boiling Heat Transfer in Vertical Tubes
Correlated by an Asymptotic Model, Heat Transfer Engineering, 13 (2), pp. 4368.
Steinke, M. E. and Kandlikar S. G., 2003, Flow Boiling and Pressure Drop in
Parallel Flow Microchannels, First International Conference on Microchannel
and Minichannels, Rochester, NY, ICMM2003-1070.
Sumith, B., Kaminaga, F., and Matsumura, K., 2003,Saturated Flow Boiling of
Water in a Vertical Small Diameter Tube, Experimental Thermal and Fluid
Science, 27, pp. 789-801.
Taitel, Y., 1990, Flow Pattern Transition in Two Phase Flow, Keynote lecture,
9th International Heat Transfer Conference, Jerusalem, 19-24 August 1990, pp.
237-254.
Taitel, Y. and Dukler, A. E., 1976, A Model for Predicting Flow Regime
Transitions in Horizontal and Near Horizontal Gas-Liquid Flow, AIChE Journal,
22, pp. 47-55.
Taylor, J. R., 1997, An Introduction to Error Analysis, 2nd Ed., University Science
Books.
Thome, J. R., Dupont, V., and Jacobi, A. M., 2004, Heat Transfer Model for
Evaporation in Microchannels. Part I: Presentation of the Model, International
Journal of Heat and Mass Transfer, 47, pp. 3375-3385.
Tran, T. N., Wambsganss, M. W., and France, D. M., 1996, Small Circular- and
Rectangular-Channel Boiling with Two Refrigerants, International Journal of
Multiphase Flow, 22 (3), pp. 485-498.

174
Wallis G. B., One-Dimensional Two-Phase Flow, McGraw-Hill, New York, 1969.
Wang, G., Cheng, P., and Bergles, A. E., 2008, Effects of Inlet/Outlet
Configuration on Flow Boiling Instability in Parallel Microchannels, International
Journal of Heat and Mass Transfer, 51 (9-10), pp. 2267-2281.
Warrier G. R., Dhir, V. K., and Momoda, L. A., 2002, Heat Transfer and
Pressure Drop in Narrow Rectangular Channels, Experimental Thermal and
Fluid Science, 26 (1), pp. 53-64.
Yan, Y. and Lin, T., 1998, Evaporation Heat Transfer and Pressure Drop of
Refrigerant R-134a in a Small Pipe, International Journal of Heat and Mass
Transfer, 41, pp. 4183-4194.
Yu, W., France, D. M., Wambsganss, M. W. and Hull, J. R., 2002, Two-Phase
Pressure Drop, Boiling Heat Transfer, and Critical Heat Flux to Water in a Small
Diameter Horizontal Tube, International Journal of Multiphase Flow, 28, pp. 927941.
Yun, R., Heo, J. H., and Kim, Y., 2007, Erratum to Evaporative Heat Transfer
and Pressure Drop of R410A in Microchannels, [International Journal of
Refrigeration 29 (2006) 92-100], International Journal of Refrigeration, 30, pp.
1468.
Zhang, W., Hibiki, T., and Mishima, K., 2004, Correlation for Flow Boiling Heat
Transfer in Mini-Channels; International Journal of Heat and Mass Transfer, 47
(26), pp. 5749-5763.
Zhang, H. Y., Pinjala, D., and Wong T. N., 2005a, Experimental Characterization
of Flow Boiling Heat Dissipation in a Microchannel Heat Sink with Different
Orientations, Proceedings of 7th Electronics Packaging Technology Conference,
EPTC, 2, pp. 670-676.
Zhang, L., Wang, E. N., Goodson, K. E., and Kenny, T. W., 2005b, Phase
Change Phenomena in Silicon Microchannels, International Journal of Heat and
Mass Transfer, 48, pp. 1572-1582.

APPENDICES

175
Appendix A. Test Chip Calibration
The calibration results for the resistor heating elements and diode
temperature sensors are presented in this appendix for the 400 m 400 m
microchannel heat sink. Pictures of the test piece assembly and the convection
oven are provided in Figure A.1 and Figure A.2, respectively.
The calibration results for the 25 resistors fabricated on the back side of
the chip are depicted in Figure A.3. In this figure, symbols show the measured
data points and the curves represent quadratic polynomial curves fit to the
measured values. For instance, the R-T relation for the heating element 3
(Figure 2.3) is R = 0.0003 T 2 + 0.0148 T + 34.55 .
The voltage drop measurements across the 25 temperature sensing
diodes are shown in Figure A.4 along with linear curves fit to the measured
values. The V-T relation for diode 3 in Figure 2.3 is V = .0105 T + 3.509 .

176

Figure A.1. Test piece assembly for the calibration process.

177

Figure A.2. Convection oven (Lindberg Blue).

178

55

50

R ()

45

40

35

30

20

40

60

80

100

120

140

T (C)

Figure A.3. Calibration of the resistance values of the 25 heating elements


fabricated on the back side of the 400 m 400 m microchannel heat sink.

179

3.2

V (Volt)

2.8

2.6

2.4

2.2

20

40

60

80

100

120

140

T (C)

Figure A.4. Calibration of the voltage drop across the 25 diode temperature
sensors fabricated on the back side of the 400 m 400 m microchannel heat
sink.

180
Appendix B. Degassing Procedures
At standard pressure and temperature conditions, FC-77 contains 41% of
air by volume (3M 1999), which is equivalent to an air concentration of 283 parts
per million. A custom-built degassing scheme, designed by Tailian Chen (Chen
and Garimella, 2006b), is used in order to fully degas the liquid. Figure B.1
shows an expandable reservoir and a constraint mechanism used to facilitate the
degassing of the liquid. After testing the reservoir to be leak-free, it is filled with
the working fluid, FC-77. Degassing is performed by repeatedly evacuating the
reservoir of air from the top, while the container is expanded, using a vacuum
pump. A period of 10 minutes is allowed between evacuations for air to diffuse
from the liquid to the open volume in the container. When the pressure in the
reservoir remains at its maximum vacuum level and is unchanged with time, the
fluid is considered to be fully degassed. The constraint mechanism on the
reservoir is then adjusted to allow the bellows to contract such that the pressure
in the reservoir is maintained at atmospheric pressure.
An additional degassing port is also provided in the test loop (Figure 2.1)
to facilitate the evacuation of air form the flow loop prior to charging it with the
coolant.

181

(a)

(b)
Figure B.1. (a) Photograph and (b) schematic of the expandable reservoir used
to degas the fluid (Chen and Garimella, 2006b).

182
Appendix C. Heat Loss Measurements
To determine the heat loss in the microchannel test section, before the
test section is charged with liquid, a constant voltage is applied to the heaters.
When the readings of the diode temperature sensors reach a steady state, the
temperature of each sensor is recorded and correlated to the heat dissipated
from the corresponding heater at that location. This procedure is repeated for
several levels of input power and a linear relation in the form of qloss = c1Td + c2 is
obtained, where c1 and c 2 are constants and are slightly different at each
location and for different test pieces. The heat loss measurements for the 400
m 400 m microchannel test chip are presented in Figure C.1.

183

0.5

0.4

q loss (W)

0.3

0.2

0.1

0
20

40

60

80

T d (C)

100

120

140

Figure C.1. Heat loss measurements of the 25 heating elements fabricated on


the back side of the 400 m 400 m microchannel heat sink.

184
Appendix D. Measurement Uncertainties and Experiment Repeatability
Figure D.1 shows the heat transfer coefficients for five microchannel sizes
for a mass flux of 630 kg/m2s, including the uncertainties in the measurement of
heat transfer coefficient. As discussed in section 4.3.1, heat transfer coefficient
is independent of channel size for microchannels with a cross-sectional area
larger than 0.089 mm2 (shown for 1000 m 400 m microchannels in Figure
D.1), and increases with decreasing the channel cross-sectional area for smaller
microchannels. Figure D.2 depicts the effect of mass flux on heat transfer
coefficient for the 1000 m 400 m microchannels, including the error bars.
This figure shows that heat transfer coefficient is independent of mass flux as
discussed in 5.3.1.
To ensure repeatability of the experiments, all the measurements are
repeated twice or more, and the results are compared. Figure D.3 to Figure D.6
present the boiling curves in the 250 m 400 m microchannels. In each of
these figures, results of two repeated tests are included for a fixed mass flux. It
can be seen that repeating the experiments under the same operating conditions
leads to very similar results.

185

9
8

G = 630 kg/m s

h (kW/m2 K)

7
6
5
4
3

1000 x 400
400 x 220
100 x 400
400 x 100
100 x 100

2
1
0

50

100
150
200
2
q"w (kW/m )

250

300

Figure D.1. Measurement uncertainties in the heat transfer coefficient for five
microchannel sizes, G = 630 kg/m2s.

186

10
9

1000 m x 400 m

8
h (kW/m2K)

7
6
5
4

225 kg/m2 s
630 kg/m2 s

3
2

1050 kg/m2s
1420 kg/m2s

1
0

50

100

150 200 250


q"w (kW/m2)

300

350

Figure D.2. Measurement uncertainties in the heat transfer coefficient for four
mass fluxes in the 1000 m 400 m microchannels.

187

400
350

q"w (kW/m2)

300
250

Test 1
Test 2

250 m x 400 m
G = 225 kg/m2s

200
150
100
50
0
80

100

120
T w (C)

140

160

Figure D.3. Boiling curves from two separate tests, G = 225 kg/m2s, 250 m
400 m microchannels.

188

400
350

q"w (kW/m2)

300
250

Test 1
Test 2

250 m x 400 m
G = 630 kg/m2s

200
150
100
50
0
80

100

120
T w (C)

140

160

Figure D.4. Boiling curves from two separate tests, G = 630 kg/m2s, 250 m
400 m microchannels.

189

400
350

q"w (kW/m2)

300
250

Test 1
Test 2

250 m x 400 m
G = 1050 kg/m2s

200
150
100
50
0
80

100

120
T w (C)

140

160

Figure D.5. Boiling curves from two separate tests, G = 1050 kg/m2s, 250 m
400 m microchannels.

190

400
350

q"w (kW/m2)

300
250

Test 1
Test 2

250 m x 400 m
G = 1420 kg/m2s

200
150
100
50
0
80

100

120
T w (C)

140

160

Figure D.6. Boiling curves from two separate tests, G = 1420 kg/m2s, 250 m
400 m microchannels.

191
Appendix E. Effect of Channel Size on Boiling Heat Transfer and Pressure Drop
The effects of channel size on heat transfer coefficient, boiling curve, and
pressure drop have been discussed in Chapter 4 for a fixed mass flux of 630
kg/m2s. In this appendix, results for mass fluxes of 225, 1050, and 1420 are
presented for a fixed channel depth of 400 m.

192

300
100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

200

q"w (kW/m )

250

150

G = 225 kg/m s

100
50
0

10

20
30
T w-Tref (C)

40

50

Figure E.1. Effect of microchannel width on boiling curves, d = 400 m, G = 225


kg/m2s.

193

300
100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

200

q"w (kW/m )

250

150

G = 630 kg/m s

100
50
0

10

20
T w-Tref (C)

30

40

Figure E.2. Effect of microchannel width on boiling curves, d = 400 m, G = 630


kg/m2s.

194

400
100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

350

q"w (kW/m )

300
250
200

G = 1050 kg/m s

150
100
50
0

10

20
T w-Tref (C)

30

40

Figure E.3. Effect of microchannel width on boiling curves, d = 400 m, G =


1050 kg/m2s.

195

400
100 m
250 m
400 m

350

q"w (kW/m )

300

700 m
1000 m
2200 m

250
200

G = 1420 kg/m s

150
100
50
0

10

20
T w-Tref (C)

30

40

Figure E.4. Effect of microchannel width on boiling curves, d = 400 m, G =


1420 kg/m2s.

196

6
2

G = 225 kg/m s

h (kW/m2K)

5
4
3

100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

2
1
0

50

100
150
2
q"w (kW/m )

200

250

Figure E.5. Effect of microchannel width on local heat transfer coefficient, d =


400 m, G = 225 kg/m2s.

197

9
2

G = 630 kg/m s

h (kW/m2K)

7
6
5
100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

4
3
2
1
0

50

100
150
200
2
q"w (kW/m )

250

300

Figure E.6. Effect of microchannel width on local heat transfer coefficient, d =


400 m, G = 630 kg/m2s.

198

11
10

G = 1050 kg/m s

h (kW/m2K)

8
7
6
100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

5
4
3
2
1
0

100

200
2
q"w (kW/m )

300

400

Figure E.7. Effect of microchannel width on local heat transfer coefficient, d =


400 m, G = 1050 kg/m2s.

199

10
9

G = 1420 kg/m s

8
h (kW/m2K)

7
6
5

100 m
250 m
400 m

4
3

700 m
1000 m
2200 m

2
1
0

100

200
2
q"w (kW/m )

300

400

Figure E.8. Effect of microchannel width on local heat transfer coefficient, d =


400 m, G = 1420 kg/m2s.

200

14
2

G = 225 kg/m s

12

p (kPa)

10
8

100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

6
4
2
0

50

100
2
q"w (kW/m )

150

200

Figure E.9. Effect of microchannel width on pressure drop, d = 400 m, G = 225


kg/m2s.

201

30
2

G = 630 kg/m s

25

p (kPa)

20
15

100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

10
5
0

50

100
150
200
2
q"w (kW/m )

250

300

Figure E.10. Effect of microchannel width on pressure drop, d = 400 m, G =


630 kg/m2s.

202

35
30

G = 1050 kg/m2s

p (kPa)

25
20

100 m
250 m
400 m
700 m
1000 m
2200 m
5850 m

15
10
5
0

100

200
2
q"w (kW/m )

300

400

Figure E.11. Effect of microchannel width on pressure drop, d = 400 m, G =


1050 kg/m2s.

203

40
2

G = 1420 kg/m s

35

p (kPa)

30

100 m
250 m
400 m
700 m
1000 m
2200 m

25
20
15
10
5
0

100

200
300
2
q"w (kW/m )

400

Figure E.12. Effect of microchannel width on pressure drop, d = 400 m, G =


1420 kg/m2s.

204
Appendix F. Effect of Mass Flux on Boiling Heat Transfer and Pressure Drop
The effects of mass flux on heat transfer coefficient, boiling curve, and
pressure drop have been discussed in Chapter 5 for microchannels with crosssections of 400 m 400 m and 2200 m 400 m. In this appendix, results
for other microchannel test pieces are shown.

205

300
250
200

1420 kg/m s

q"w (kW/m )

225 kg/m s
2
630 kg/m s
1050 kg/m2s

w = 100 m

150
100
50
0

10

20
30
T w-Tref (C)

40

50

Figure F.1. Effect of mass flux on boiling curves, 100 m 400 m.

206

350
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

300

1420 kg/m s

q"w (kW/m )

250
200

w = 250 m

150
100
50
0

10

20
30
T w-Tref (C)

40

50

Figure F.2. Effect of mass flux on boiling curves, 250 m 400 m.

207

350
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

300

1420 kg/m s

q"w (kW/m )

250
200

w = 400 m

150
100
50
0

10

20
T w-Tref (C)

30

40

Figure F.3. Effect of mass flux on boiling curves, 400 m 400 m.

208

400
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

350
300
2

q"w (kW/m )

1420 kg/m s

250

w = 700 m

200
150
100
50
0

10

20
T w-Tref (C)

30

40

Figure F.4. Effect of mass flux on boiling curves, 700 m 400 m.

209

400
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

350
300
2

q"w (kW/m )

1420 kg/m s

250

w = 1000 m

200
150
100
50
0

10

20
T w-Tref (C)

30

40

Figure F.5. Effect of mass flux on boiling curves, 1000 m 400 m.

210

400
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

350
300
2

q"w (kW/m )

1420 kg/m s

250

w = 2200 m

200
150
100
50
0

10

20
T w-Tref (C)

30

40

Figure F.6. Effect of mass flux on boiling curves, 2200 m 400 m.

211

400
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

350

q"w (kW/m )

300
250

w = 5850 m

200
150
100
50
0

10

20
T w-Tref (C)

30

40

Figure F.7. Effect of mass flux on boiling curves, 5850 m 400 m.

212

9
w = 100 m

h (kW/m2K)

7
6
5
4
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

3
2
1
0

1420 kg/m s
0

50

100
150
2
q"w (kW/m )

200

250

Figure F.8. Effect of mass flux on local heat transfer coefficient, 100 m 400
m.

213

10
9

w = 250 m

8
h (kW/m2K)

7
6
5
4

225 kg/m s
2
630 kg/m s
1050 kg/m2s

1420 kg/m s

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.9. Effect of mass flux on local heat transfer coefficient, 250 m 400
m.

214

10
9

w = 400 m

8
h (kW/m2K)

7
6
5
4

225 kg/m s
2
630 kg/m s
1050 kg/m2s

1420 kg/m s

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.10. Effect of mass flux on local heat transfer coefficient, 400 m 400
m.

215

10
9

w = 700 m

8
h (kW/m2K)

7
6
5
4

225 kg/m s
2
630 kg/m s
1050 kg/m2s

1420 kg/m s

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.11. Effect of mass flux on local heat transfer coefficient, 700 m 400
m.

216

10
9

w = 1000 m

8
h (kW/m2K)

7
6
5
4

225 kg/m s
2
630 kg/m s
1050 kg/m2s

1420 kg/m s

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.12. Effect of mass flux on local heat transfer coefficient, 1000 m 400
m.

217

10
9

w = 2200 m

8
h (kW/m2K)

7
6
5
4

225 kg/m s
2
630 kg/m s
1050 kg/m2s

1420 kg/m s

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.13. Effect of mass flux on local heat transfer coefficient, 2200 m 400
m.

218

10
9

w = 5850 m

8
h (kW/m2K)

7
6
5
4

225 kg/m s
2
630 kg/m s
1050 kg/m2s

3
2
1
0

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.14. Effect of mass flux on local heat transfer coefficient, 5850 m 400
m.

219

40

225 kg/m s
2
630 kg/m s
2

1050 kg/m s
2
1420 kg/m s

p (kPa)

30

w = 100 m

20

10

50

100
150
2
q"w (kW/m )

200

250

Figure F.15. Effect of mass flux on pressure drop, 100 m 400 m.

220

25
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

20

p (kPa)

1420 kg/m s
15

w = 250 m

10

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.16. Effect of mass flux on pressure drop, 250 m 400 m.

221

12
2

1420 kg/m s

p (kPa)

10

225 kg/m s
2
630 kg/m s
1050 kg/m2s
2

w = 400 m

6
4
2
0

50

100
150
200
2
q"w (kW/m )

250

300

Figure F.17. Effect of mass flux on pressure drop, 400 m 400 m.

222

10
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

p (kPa)

1420 kg/m s
6

w = 700 m

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.18. Effect of mass flux on pressure drop, 700 m 400 m.

223

10
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

p (kPa)

1420 kg/m s
6

w = 1000 m

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.19. Effect of mass flux on pressure drop, 1000 m 400 m.

224

10
2

225 kg/m s
2
630 kg/m s
1050 kg/m2s

p (kPa)

w = 5850 m

50

100

150 200 250


2
q"w (kW/m )

300

350

Figure F.20. Effect of mass flux on pressure drop, 5850 m 400 m.

225
Appendix G. MATLAB Script for Heat Transfer Calculations in Annular Flow
As discussed in section 8.2.2.1, to calculate the heat transfer coefficient in
the annular regime using the analytical model proposed in the current study, a
numerical procedure should be followed. The following MATLAB script
calculates the local heat transfer coefficient in the confined annular region with
no droplet entrainment as discussed in section 8.2.2.2.
%---------------------------------------------------------------------%Annular.m
%script by Tannaz Harirchian, Purdue University, Apr. 2010
%for PhD work under Dr. Garimella
%
%caculating the heat transfer coefficient using the
%model developed for confined annular regime in Chapter 8
%
clear
fluidproperties
%reading fluid properties
data
%reading experimental data
GG=GqxhP(:,1);
%mass flux
qq=GqxhP(:,2)*1000;
%heat flux, W/m2
xx_exp=GqxhP(:,3);
%exit vapor quality
hh_exp=GqxhP(:,4)*1000; %experimental heat transfer coefficient, W/m2k
PP_sat=GqxhP(:,5)*1000; %saturation pressure, Pa
TT_in=GqxhP(:,6);
%inlet temperature
TT_sat=GqxhP(:,7);
%saturation temperature
nrow=nNwd(:,1);
N=nNwd(:,2);
w=nNwd(:,3);
d=nNwd(:,4);
L_ch=0.012667;
%chip length, channel length, also chip width
s=0;
n=0;
m=length(nrow);
%# of cases
within30=0;
for i=1:7
%loop over different cases
s=sum(nrow(1:i-1));
n=sum(nrow(1:i));
%# of data for each case
wch=w(i)*1e-6;
%channel width; m
dch=d(i)*1e-6;
%channel depth; m
Nch=N(i);
%# of channels
%Geometry calculations
A_cs=wch*dch;
P_h=wch+2*dch;
%heated perimeter of one channel
P_ch=2*wch+2*dch;
%total perimeter of channel
D_h=sqrt(A_cs);
R=D_h/2;
zzz=0;
for z=s+1:n
%loop over data in each case
width(z)=wch;

226
depth(z)=dch;
G=GG(z);
q=qq(z);
x_exit=xx_exp(z);
h_exp=hh_exp(z);
P_sat=PP_sat(z);
T_in=TT_in(z);
T_sat=TT_sat(z);
Re=G*D_h/mu_l;
Bo=g*(rho_l-rho_v)*D_h^2/sigmal; %Bond number
Bl=q/(G*h_fg);
%Boiling number
C1=.00004*(Re^2*Bo);
%correction factor in f_i
%STEP 1
L_sp(z)=G*A_cs*C_p_l*(T_sat-T_in)/(q*P_h); %onset of boiling
%location of onset of annular flow; from flow regime map
L_a0(z)= 96.65*(A_cs/P_h)*(rho_v/(rho_l-rho_v))*Bl^(-1)*
(Re*sqrt(Bo))^(-0.26);
%quality at onset of annular regime
x_a0(z)=(L_a0(z)-L_sp(z))*x_exit/(L_ch-L_sp(z));
L_a=L_ch-L_a0(z); %length of channel under annular regime
h_dry=0;
L_dry=0;
if (L_a>0)
%STEP 2
L=.01143; %distance from channel inlet at which
%experimental data are obtained
x0=(L-L_sp(z))*x_exit/(L_ch-L_sp(z));
m_dot=G*A_cs;
%total mass flow rate through one channel
m_Ff=m_dot*(1-x0);
%liquid film flow rate
m_g=m_dot*x0;
%vapor flow rate
delta_init=1e-6;
%initial guess for delta
d_delta=1e-8;
delta=delta_init;
zzz=zzz+1;
balance=false;
zz=0;
while ~balance
%STEP 3
delta=delta+d_delta; %increasing delta
zz=zz+1;
gama_fg=q*P_h/h_fg; %evaporation rate per unit length
%STEP 4
P_c=2*((wch-2*delta)+(dch-2*delta));
rho_H=rho_v;
A_c=(wch-2*delta)*(dch-2*delta);
alpha=A_c/A_cs;
u_c=G*x0/(rho_v*alpha);
u_Ff=G*(1-x0)/(rho_l*(1-alpha));
u_i=2*u_Ff;
beta_c=(wch-2*delta)/(dch-2*delta);
d_hc=4*A_c/P_c;
Re_c=rho_H*(u_c-u_i)*d_hc/mu_v;
%interfacial friction factor
f_i=.005*(1+300*delta/D_h)*C1;

227
%interfacial shear stress
Tau_i=f_i*(0.5*rho_H*(u_c-u_i)^2)(gama_fg/(2*P_c))*(u_c-u_i);
dxdz=4*q/(G*D_h*h_fg);
%STEP 5
dpdz=-Tau_i*P_c/A_c+gama_fg*u_i/A_c;
%mass conservation in Eq. (8.14)
m_Ff_n=((P_ch*rho_l*delta^3)/(3*mu_l))*(-dpdz+3*Tau_i/
(2*delta)-3*gama_fg*u_i/(2*delta*P_ch));%[M]
%STEP 6: balancing m_Ff_n with m_Ff
RHS(zz)=m_Ff_n;
LHS(zz)=m_Ff;
error1(zz)=LHS(zz)-RHS(zz);
sign1(zz)=sign(real(error1(zz)));
if (zz~=1)&&(sign1(zz)~=sign1(zz-1))&&(1e-6<delta)
balance=true;
%break out of while loop
deltabalance=delta;
end
if (delta>200e-6)
balance=true;
%break out of while loop
deltabalance=0.;
end
end
%calculating heat transfer coefficient
if deltabalance~=0
h_z(z)=lambda_l/delta; %conduction through liquid film
else
%vapor convection in the case of dryout
L_dry=L_a;
Re_v=G*D_h/mu_v;
Pr_v=C_p_v*mu_v/lambda_v;
Nus_lam_v=0.455*Pr_v^(1/3)*sqrt((D_h*Re_v)/L_dry);
xi_v= (1.82*log10(Re_v)-1.64)^(-2);
Nus_trans_v=(xi_v/8*(Re_v-1000)*Pr_v)/(1+12.7*
sqrt(xi_v/8)*(Pr_v^(2/3)-1))*(1+(1/3)*(D_h/L_dry)^(2/3));
h_z(z)=(lambda_v/D_h)*(Nus_lam_v^4+Nus_trans_v^4)^0.25;
'dried';
end
error_h(z)=abs(h_exp-h_z(z))/h_exp*100;
if (error_h(z) <= 30)
within30=within30+1; %# of data points within 30%
end
end
%check if entrainemt occurs:
%min liquid film G for onset of entrainment (Carey 1992)
G_Ff0=(mu_l/D_h)*exp(5.8504+.429*(mu_v/mu_l)*sqrt(rho_l/rho_v));
G_Ff=m_Ff/(A_cs-A_c);
if G_Ff>G_Ff0
ent(z)=1;
else
ent(z)=0;
end
end

228
end
MAE_h=mean(error_h); %mean average error
within30=within30/n*100; %percentage of data within 30%
%----------------------------------------------------------------------

229
Appendix H. MATLAB Script for Heat Transfer Calculations in Slug Flow
The following MATLAB script calculates the heat transfer coefficient in the
slug region, using the modified model of Thome et al. (2004) developed in
section 8.2.4.1. Four empirical parameters of C 0 , cq , nq , and n f are optimized
with the least-square error method, using the nlinfit tool in MATLAB.
%---------------------------------------------------------------------%nonlinfit.m
%script by Tannaz Harirchian, Purdue University, Apr. 2010
%for PhD work under Dr. Garimella
%
%using MATLAB tool nlinfit to estimate the coefficients of
%a nonlinear regression function using least squares
%
clear
%reading experimental data
data
G=GqxhP(:,1);
%mass flux
q=GqxhP(:,2);
%heat flux
x_exp=GqxhP(:,3);
%exit vapor quality
h_exp=GqxhP(:,4);
%experimental heat transfer coefficient
P_sat=GqxhP(:,5);
%saturation pressure
X=[G,q,x_exp,P_sat];
%matrix of independent variables
y=h_exp;
%vector of dependent variable
c0=[1,1.74,3.3,1e8];
%coefficient vector; initial values
options=statset('MaxIter',400);
c = nlinfit(X,y,@htcoef,c0,options);
%---------------------------------------------------------------------%---------------------------------------------------------------------%htcoef.m
%Script for calculating the heat transfer coefficient using the
% modified model of Thome et al. (2004) developed in Chapter 8
%
function yhat=htcoef(c,X)
C_delta_o=c(1);
eta_f=c(2);
n_q=c(3);
alpha_q=c(4);
fluidproperties
%reading fluid properties
data
%reading experimental data
G=GqxhP(:,1);
q=GqxhP(:,2)*1000;
x_exp=GqxhP(:,3);
h_exp=GqxhP(:,4)*1000;
P_sat=GqxhP(:,5)*1000;
nrow=nNwd(:,1);
N=nNwd(:,2);
% # of data in each case

230
w=nNwd(:,3);
d=nNwd(:,4);
roughness=nNwd(:,5); %using max roughness for delta_min
%calculations:
s=0;
n=0;
m=length(nrow);
%# of cases; each case comprises of data points for
%one microchannel test piece and a fixed mass flux
mm=sum(nrow);
%total # of data points
h_z=zeros(mm,1);
error_h=zeros(mm,1);
f=zeros(mm,1);
within30=0;
for i=1:m
%loop over different cases
n=nrow(i)+s;
wch=w(i)*1e-6; %channel width; m
dch=d(i)*1e-6; %channel depth; m
Nch=N(i);
%# of channels
delta_min=roughness(i); %using max roughness for delta_min
%Geometry calculations
A_cs=wch*dch;
D_h=sqrt(A_cs);
R=D_h/2;
for z=s+1:n
%loop over data in each case
%delta_o calculation
q_ref=alpha_q*(P_sat(z)/P_crit)^n_q;
f_opt=(q(z)/q_ref)^eta_f;
tau=1/f_opt;
G_tot=G(z)+(4/3)*R*(rho_v/tau);
U_p=G_tot*(x_exp(z)/rho_v+(1-x_exp(z))/rho_l);
L_p=U_p*tau;
L_v=(tau*G_tot)/rho_v*x_exp(z);
L_l=(tau*G_tot)/rho_l*(1-x_exp(z));
Bo=(rho_l*D_h)/sigmal*U_p^2;
Ca=mu_l*U_p/sigmal;
%calculating liquid film thickness
delta_T=D_h*(0.66*Ca^(2/3))/(1+3.33*Ca^(2/3));
delta_BL=sqrt((nu_l*L_l)/U_p); %boundary layer thickness
delta_small=min(delta_T,delta_BL);
delta_o_z=C_delta_o*delta_small;
t_v=tau/(1+(rho_v/rho_l)*((1-x_exp(z))/x_exp(z)));
t_l=tau/(1+(rho_l/rho_v)*(x_exp(z)/(1-x_exp(z))));
t_dryfilm=(rho_l*h_fg)/q(z)*(delta_o_z-delta_min);
if (t_dryfilm < 0.)
t_dryfilm=0;
end
if (t_dryfilm > t_v)
delta_end=delta_o_z-(q(z)/(rho_l*h_fg))*t_v;
t_film=t_v;
t_dry=0;
L_dry=0;
else
delta_end=delta_min;
t_film=t_dryfilm;
t_dry=t_v-t_film;

231
L_dry=U_p*t_dry;
end
%calculating h for liquid slug
Re_l=rho_l*U_p*D_h/mu_l;
Pr_l=C_p_l*mu_l/lambda_l;
Nus_lam_l=0.455*Pr_l^(1/3)*sqrt((D_h*Re_l)/L_l);
xi_l= (1.82*log10(Re_l)-1.64)^(-2);
Nus_trans_l=(xi_l/8*(Re_l-1000)*Pr_l)/(1+12.7*sqrt(xi_l/8)*
(Pr_l^(2/3)-1))*(1+1/3*(D_h/L_l)^(2/3));
h_l=(lambda_l/D_h)*(Nus_lam_l^4+Nus_trans_l^4)^0.25;
%calculating h for vapor slug
if (t_dryfilm < t_v)
Re_v=rho_v*U_p*D_h/mu_v;
Pr_v=C_p_v*mu_v/lambda_v;
Nus_lam_v=0.455*Pr_v^(1/3)*sqrt((D_h*Re_v)/L_dry);
xi_v= (1.82*log10(Re_v)-1.64)^(-2);
Nus_trans_v=(xi_v/8*(Re_v-1000)*Pr_v)/(1+12.7*sqrt(xi_v/8)*
(Pr_v^(2/3)-1))*(1+1/3*(D_h/L_dry)^(2/3));
h_v=(lambda_v/D_h)*(Nus_lam_v^4+Nus_trans_v^4)^0.25;
else
h_v=0;
end
%calculating h for annulus
h_film=2*lambda_l/(delta_o_z+delta_min);
%calculating local time-averaged h of a triplet passing z
h_z(z)=(t_l/tau)*h_l+(t_film/tau)*h_film+(t_dry/tau)*h_v;
(t_l/tau)*h_l;
(t_film/tau)*h_film;
(t_dry/tau)*h_v;
error_h(z)=abs(h_exp(z)-h_z(z))/h_exp(z)*100;
if (error_h(z) <= 30)
within30=within30+1; %# of data points within 30%
end
end
s=n;
end
yhat=h_z/1000;
MAE_h=mean(error_h); %mean average error
within30=within30/mm*100; %percentage of data within 30%
%----------------------------------------------------------------------

VITA

232

VITA

Tannaz Harirchian was born in Tehran, Iran. She obtained her BS degree
in mechanical engineering in the area of thermal fluid sciences from the
University of Tehran in 2003. During the last two years of her studies, she
worked part-time in a pharmaceutical machinery designing company where she
continued working for 18 months after her graduation. Tannaz attended
Southern Illinois University at Edwardsville and obtained her MS degree in
mechanical engineering in 2006. Her MS thesis was focused on heat transfer
enhancement using electrohydrodynamic technique. She then joined the PhD
program in mechanical engineering at Purdue University. She has been a
research assistant in the NSF Cooling Technologies Research Center,
performing experimental and analytical work on two phase flow and heat transfer
in microchannel heat sinks for cooling of power electronics. She received the
EPPD Student Member of the Year Award from the Electronic and Photonic
Packaging Division of the ASME in 2009 as well as the 2009 Harvey Rosten
Award for Excellence in the physical design of electronics.

PUBLICATIONS

233

PUBLICATIONS

Garimella, S. V. and Harirchian, T., Microchannel Coolers for Electronic


Components. Bar-Cohen, A. (Editor), The Encyclopedia of Electronic Packaging.
World Scientific, in preparation.
Harirchian, T. and Garimella, S. V., Phenomenological Modeling of Heat
Transfer and Pressure drop for Boiling in Microchannels, International Journal of
Multiphase Flow, in preparation.
Harirchian, T. and Garimella, S. V., Boiling Heat Transfer and Flow Regimes in
Microchannels a Comprehensive Understanding, Journal of Electronic
Packaging, accepted for publication.
Harirchian, T. and Garimella, S. V., 2010, A Comprehensive Flow Regime Map
for Microchannel Flow Boiling with Quantitative Transition Criteria, International
Journal of Heat and Mass Transfer, 53, pp. 2694-2702.
Garimella, S. V. and Harirchian, T., 2009, Boiling Heat Transfer and Flow
Regimes in Microchannels a Comprehensive Understanding, Keynote lecture
at Thermal Investigations of ICs and Systems, THERMINIC 2009, October 7-9,
Leuven, Belgium.
Harirchian, T. and Garimella, S. V., 2009, Effects of Channel Dimension, Heat
Flux, and Mass Flux on Flow Boiling Regimes in Microchannels, International
Journal of Multiphase Flow, 35, pp. 349-362.
Harirchian, T. and Garimella, S. V., 2009, The Critical Role of Channel CrossSectional Area in Microchannel Flow Boiling Heat Transfer, International Journal
of Multiphase Flow, 35, pp. 904-913.
Harirchian, T. and Garimella, S. V., 2009. A Systematic Investigation of the
Effects of Channel Width, Depth, and Aspect Ratio on Boiling in Microchannels,
ASME Summer Heat Transfer Conference, July 19-23, San Francisco, CA.

234
Holcomb, B. T., Harirchian, T. and Garimella, S. V., 2009, An Experimental
Investigation of Microchannel Size Effects on Flow Boiling with De-Ionized
Water, ASME Summer Heat Transfer Conference, July 19-23, San Francisco,
CA.
Harirchian, T. and Garimella, S. V., 2008, Microchannel Size Effects on Local
Flow Boiling Heat Transfer to a Dielectric Fluid, International Journal of Heat and
Mass Transfer, 51, pp. 3724-3735.
Harirchian, T. and Garimella, S. V., 2008, Flow Patterns during Convective
Boiling in Microchannels, Journal of Heat Transfer, 130 (8), 080909-1.
Harirchian, T. and Garimella, S. V., 2008, An Investigation of Flow Boiling
Regimes in Microchannels of Different Sizes by Means of High-Speed
Visualization, 11th IEEE Intersociety Conference on Thermal and
Thermomechanical Phenomena in Electronic Systems, I-THERM, pp. 197-206.
Harirchian, T. and Garimella, S. V., 2008, Flow Boiling in Silicon Microchannel
Heat Sinks, Proceedings of the Annual IEEE Semiconductor Thermal
Measurement and Management Symposium, SEMI-THERM, pp. 1-8.
Harirchian, T. and Garimella, S. V., 2007, Microchannel Size Effects on TwoPhase Local Heat Transfer and Pressure Drop in Silicon Microchannel Heat
Sinks with a Dielectric Fluid, Proceedings of the ASME International Mechanical
Engineering Congress and Exposition, IMECE2007, 11, pp. 437-446.