.

a...

a ‘ I v
. ‘ ‘ 1.533,} In.
‘3:(:..!.’I.ILK . in).
nri .l .

in.

. 212:1
‘ . . 4%.: .0
A tunuxﬁﬂizl‘. 14.
ct SAL“. 9O
. .CH 1
n a
.3“ a 5 .w
:22: 52.12.. ”I! A
x. 33:} iii! {In K

{v.3 ., V 1-“. £1.55
.9”... a .19} Allyn"
i .2: if: .1 i .11... .3
_ v 3.4 v I
1L.» :.J : 3. ;
Ea

, .

$353: A. .. v . w.

 

 

LIBRARY
Michigan State
University

"W «nu-4 v v __ v'

 

 

 

This is to certify that the

thesis entitled
Velocity Field Measurement for a
Uni-Directional Solidification of an
Ammonium Chloride (NHaCl) Solution
using Molecular Tagging Velocimetry (MTV)

presented by
CHEE LEONG LUM

has been accepted towards fulﬁllment
of the requirements for

Masters of ScienceJegree in Mechanical Engineering

 

 

 

Date July 31, 2001

 

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 chlRC/DateDuepes-p. 15

VELOCITY FIELD MEASUREMENT FOR A UNI-DIRECTIONAL
SOLIDIFICATION OF AN AMMONIUM CHLORIDE (NH4CL) SOLUTION USING
MOLECULAR TAGGING VELOCIMETRY (MTV)

By

Chee Leong Lum

AN ABSTRACT OF A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Mechanical Engineering

2001

Dr. Manoochehr M. Koochesfahani

ABSTRACT

THE VELOCITY FIELD MEASUREMENT IN A UNI-DIRECTIONAL
SOLIDIFICATION OF AMMONIUM CHLORIDE (NH4C1) USING MOLECULAR
TAGGIN G VELOCIMETRY (MTV)

By

Chee Leong Lum

During the solidiﬁcation of a binary alloy under off-eutectic conditions, solutal
and thermal forces can produce imperfections in the form of solute-rich channels. Many
researchers observe that the use of uni-directional solidiﬁcation for alloys can provide
increased resistance to creep and thermal fatigue in the ﬁnal solidiﬁed ingot. Most
experimental methods applied to date have been poorly suited to study the complex
dynamic ﬁelds associated with these convective phenomena. In this work, the novel
technique of Molecular Tagging Velocimetry was extended and applied to investigate the
transient velocity ﬁelds present during the uni-directional solidiﬁcation of a transparent
alloy analog (ammonium chloride). The whole-ﬁeld two-component velocity map near
the mushy layer and above the chimney, where lighter ﬂuid rises, was measured at
different times during the solidiﬁcation process and at different locations. The mushy
region exhibits a maximum velocity of order 1mm/s in both components while the
plumes have a maximum velocity of order 7mm/s in the upward component. The
measured velocity ﬁelds are expected to enhance understanding of the diverse transport
phenomena present during the solidiﬁcation process and provide quantitative veriﬁcation

of future computational analysis.

DEDICATION

This thesis is dedicated to my parents who have been so supportive of me during my

entire academic career.

iii

ACKNOWLEDGEMENTS

It is a pleasure to have been working under the supervision of Dr. Manoochehr
Koochesfahani who has taught me many things over the last two years. I thank him and
other members of my committee; Dr. John J. McGrath and Dr. Andre Benard, whose
advice and assistance have proved to be invaluable. I would also like to extend my
gratitude to Doug Bohl, who has taught me many things during my stay in the lab and
Frederic Boudesseul, whose help made the completion of this project possible.

This work has been made possible with funding from the National Science
Foundation under Grant No. CTS-9901040 and made use of shared facilities of the

MRSEC Program of the National Science Foundation, Award No. DMR-9809688.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... vii
LIST OF FIGURES .......................................................................... viii
LIST OF SYMBOLS AND ABBREVIATIONS ......................................... xii
Chapter 1
Introduction ................................................................................... 1
Chapter 2
Experimental Procedure
2.1: Unidirectional solidiﬁcation and Experimental Setup .................... 4
2.1.1 Details of Ammonium Chloride .................................. 4
2.1.2 Experimental Setup ................................................. 6
2.2: Details of Molecular Tagging Velocimetry ................................ 9
2.2.1 MTV Chemicals ..................................................... 9
2.2.2 Optics Setup ......................................................... 12
2.2.3 Correlation of MTV Images ....................................... 19
2.3: Image Acquisition System ................................................... 22
2.4: Experimental Procedure ...................................................... 26
Chapter 3
Results and Discussion ....................................................................... 31
3.1: Visualization ................................................................... 3 1
3.2: MTV Measurements of Mushy Region .................................... 38
3.3: MTV Measurements above Chimney ....................................... 49
3.4: Flow Rate Calculation above the Mush ..................................... 62
Chapter 4
Conclusion .................................................................................... 66
Appendix A
Molecular Tagging Velocimetry (MTV) and Ammonium Chloride Chemical Mix
Formulas ....................................................................................... 70
Appendix B
Timing Diagrams of Gate Signals to the Xybion Camera and Laser .................. 73
Appendix C
X-Y Ratio of The Xybion Camera .......................................................... 77

Appendix D
Calculation of temperature increase of aqueous ammonium chloride due to

irradiation by Excimer Laser ................................................................ 79
Bibliography ................................................................................... 81
General References ........................................................................... 84

vi

LIST OF TABLES

Table 1. Instantaneous maximum/minimum velocity measurements in the
ISI FOV near the mushy zone ...................................................... 47

Table 2. Instantaneous maximum/minimum velocity measurements in the
2nd FOV near the mushy zone ...................................................... 48

Table 3. Instantaneous maximum/minimum velocity measurements in the
3rd FOV near the mushy zone ...................................................... 48

Table 4. Instantaneous maximum/minimum velocity measurements for the
plume ................................................................................. 61

vii

LIST OF FIGURES

Figure 2.1.1 Equilibrium diagram of ammonium chloride .............................. 5
Figure 2.1.2 Schematic of test section ..................................................... 6
Figure 2.1.3 Schematic of base plate and embedded rectangle nucleators. . . . . . . . 7
Figure 2.1.4 Schematic of nucleator ....................................................... 8

Figure 2.2.1 Plot of lifetime as a function of weight percentage of ammonium

chloride + MTV chemicals (Cyclohexanol) ................................ 11
Figure 2.2.2 Optics Setup ................................................................... 13
Figure 2.2.3 Beam blocker design .......................................................... 14
Figure 2.2.4 Measurement ﬁeld of views for mushy region ............................ 16
Figure 2.2.5 Refracted laser lines for plume measurement ............................. 17
Figure 2.2.6 Schematic of optics setup for plume measurement ....................... 18
Figure 2.2.7 Minimum height of grid pattern from mushy zone surface .............. 18
Figure 2.2.8 Displacement sub-pixel accuracy versus 6 for different

values of S/N (Gendrich et. al., 1996) ....................................... 21
Figure 2.2.9 Effect of intersection angle and S/N on the ratio of x- and y-

displacement error magnitudes .............................................. 22
Figure 2.3.1 Physical setup of camera and electronics .................................. 23

Figure 2.3.2 Schematic of typical timing used for gating the camera and laser ...... 24

Figure 2.3.3 Schematic of gating signals with respect to the vertical drive of the
camera ........................................................................... 25

Figure 2.4.1 Setup for Visualization ...................................................... 29

Figure 3.1.1. Small ﬁngers and ammonium chloride crystals at 8 minutes into the
solidiﬁcation ................................................................... 32

Figure 3.1.2. Images show visualization of solidiﬁcation process from 4-30
minutes .......................................................................... 33

viii

Figure 3.1.3.

Images show visualization of solidiﬁcation process from 40-120

minutes ......................................................................... 34
Figure 3.1.4. Image of solidiﬁcation process at 2 hours. ............................... 35
Figure 3.1.5. Plot of Height of Mushy Region as a function of time ................ 37
Figure 3.2.1. Tagging region of the ISI FOV at 8 minutes into the solidiﬁcation
process .......................................................................... 38
Figure 3.2.2. Velocity ﬁelds of mushy region at 4 minutes (instantaneous) and
from 8-16 minutes (averaged) in the 1St F OV ............................ 39
Figure 3.2.3. Averaged velocity ﬁelds of mushy region from 20-35 minutes
(1St F OV) ...................................................................... 40
Figure 3.2.4. Tagging region of the 1St F OV at 8 minutes into the solidiﬁcation
process ......................................................................... 42
Figure 3.2.5. Averaged velocity ﬁelds of mushy region from 4-16 minutes
(2“d FOV) ....................................................................... 42
Figure 3.2.6. Averaged velocity ﬁelds of mushy region from 20-45 minutes
(2“(1 FOV) ....................................................................... 43
Figure 3.2.7. Averaged velocity ﬁelds of mushy region from 20-45 minutes
(2“d FOV) ....................................................................... 44
Figure 3.2.8. Tagging region of the 1St F OV at 8 minutes into the solidiﬁcation
process ........................................................................... 45
Figure 3.2.9. Averaged velocity ﬁelds of mushy region at 4 and 12 minutes
(3rd FOV) ........................................................................ 45
Figure 3.2.10 Averaged velocity ﬁelds of mushy region from 16-70 minutes
(3’dFOV) ....................................................................... 46
Figure 3.2.11. Averaged velocity ﬁelds of mushy region at 80, 100 minutes
(3rd FOV) ...................................................................... 47
Figure 3.3.1. Velocity proﬁle of plume (Wirtz et. al., 1998) ........................... 49
Figure 3.3.2. Delayed image using single line tagging (Wirtz et. al., 1998) .......... 49

ix

Figure 3.3.3. Typical location of multiple line tagging of single component
measurements ................................................................ 50

Figure 3.3.4. Single component velocity measurement of plume at 28 minutes. . 51
Figure 3.3.5. Single component velocity measurement of plume at 30 minutes. . 52
Figure 3.3.6. Velocity proﬁle of plume at 35 minutes into the solidiﬁcation

process .......................................................................... 53

Figure 3.3.7. Velocity proﬁle of plume at 40 minutes into the solidiﬁcation
process .......................................................................... 54

Figure 3.3.8. Velocity proﬁle of plume at 45 minutes into the solidiﬁcation
Process ......................................................................... 54

Figure 3.3.9. Visualization of tagging region at 79 minutes into the solidiﬁcation
process ........................................................................... 55

Figure 3.3.10. Velocity vectors measured at t0 = 79 min of plume from t0 to
t0 + 1.335 ..................................................................... 56

Figure 3.3.11. Velocity measurements at t0 = 75 minutes into the solidiﬁcation
process ...................................................................... 57

Figure 3.3.12. Instantaneous velocity measurements of plume from 71 to 95
minutes ....................................................................... 59

Figure 3.3.13. Instantaneous velocity measurements of plume from 100 to 120
minutes ........................................................................ 60

Figure 3.3.14. Flow rate plots at 4 and 8 minutes of three ﬁeld of views ............. 63
(presented in color)

Figure 3.3.15. Flow rate plots at 12 and 16 minutes for three ﬁeld of views ......... 64
(presented in color) -

Figure B1. Timing signals of gate signals with respect to the vertical drive .......... 74

Figure BZ. Timing diagram for different delay times with respect to the
vertical drive ..................................................................... 75

Figure B3. Sample image of gating the camera in an odd or even ﬁeld ............... 76

Figure C1. Scale image taken with the Xybion camera and the corresponding
measurements .................................................................... 78

xi

LIST OF SYMBOLS AND ABBREVIATIONS

5y - y component of displacement error
5x - x component of displacement error
Sq - error in displacement magnitude

9 - grid intersecting angle

u - micro

CCD - Charged Coupled Device

F OV - Field of View

Hz - Hertz

K - kilo

min - minutes

mC D - mass of cyclodextrin

mhzo - mass of water

m MTV -mass of MTV chemicals

m NH4C1 - mass of ammonium chloride
M - Molar

MTV - Molecular Tagging Velocimetry
MSU - Michigan State University
NH4C1 - ammonium chloride

RMS - Root Mean Squared

S/N - Signal to noise ratio

5 - seconds

xii

 

TTL - Transistor - Transistor Logic

Valcohol - volume of alcohol

VD - Vertical Drive

xiii

CHAPTER 1

INTRODUCTION

Uni-directional solidiﬁcation presents a unique method in the cooling of metals
due to the increased resistance to creep and improved thermal fatigue behavior of the
resultant cooled solid. However, the cooling orientation greatly inﬂuences the resultant
imperfections within the product. In a side cooled casting, imperfections in the form of
slanted rods known as A- and V- segregates form, while in a bottom chilled procedure, as
utilized in this study, longitudinal channels oriented parallel to the direction of gravity
form (Magirl and Incropera (1993)). These defects translate to the formation of a weaker
material thus providing an impetus to the current research project.

Previous researchers used a myriad of techniques to better understand the convective
mechanisms present during solidiﬁcation and which cause the imperfections to occur.
Magirl and Incropera (1993) used a dye visualization technique to track the ﬂow of inter-
dendritic ﬂuid exiting the channels and also particle tracking to further reveal ﬂow
patterns within the melt, while Chen and Chen (1991) used a computer tomography
technique to determine the porosity of the mush. Hellawell (1987) and Hellawell, Serazin
and Steube (1993) conjectured that the onset of plume convection is caused by ﬁnite
amplitude disturbances in the boundary layer at the mush-liquid interface. The onset of
plume due to instabilities in the mushy layer is reinforced by linear stability theory
(Worster, 1992).

Past experimental, analytical and numerical studies have established that the
convective phenomena present can best be described in terms of complex dynamic

velocity ﬁelds. Most experimental methods applied to date have been poorly suited to

 

accurately measure these phenomena and does not provide for whole-ﬁeld measurements
of the velocity (Chen et. al., 1991; Hellawell et. al., 1993; Magirl et. al., 1993; Prescott et.
al., 1996; Beckerrnan et. al., 1996). One implementation of whole ﬁeld measurements of
the velocity was performed using PIV (Wang et. al., 1998) but the density gradients
within the ﬂuid itself may have made it difficult to select an appropriate seed particle. A
recent effort at Michigan State University (MSU) utilized a novel technique called
Molecular Tagging Velocimetry (MTV) in which a molecular tracer was used to non-
intrusively measure the velocity ﬁelds above the mush using single line tagging that
yielded one-component velocity measurements of the plume (Wirtz, et. a1, 1998).

The ﬁrst attempt at measuring the velocity of the plume at Michigan State
University was meant to gauge the capabilities and application of the MTV technique in
measuring velocity ﬁelds within an aqueous ammonium chloride solution. It was shown
that the MTV technique is well suited to perform measurements on these scales. The
present study attempts to extend the previous effort and provide comparisons and
experimental validation for a future numerical simulation in the study of convective
mechanisms of small-scale ﬂows measuring velocities on the order of mm/s.
Measurements were performed of the transient velocity ﬁelds above the mushy zone and
in and around the plume where lighter ﬂuid is rejected from within the solid mush during
the uni-directional solidiﬁcation of a transparent metal analog (ammonium chloride,
NH4C1). Many researchers (Magirl and Incropera (1993), Beckerrnann and Wang (1996),
Chen and Chen (1991)) use ammonium chloride in their study of convective mechanisms
as it is a low-temperature analog of metal alloys and its transparent nature allows for

optical access. The velocity ﬁelds obtained are expected to provide insight into the

porosity of the mush in future numerical analysis as part of the overall goal of our
research group.

The better understanding of the convective mechanisms present during the
solidiﬁcation of a binary alloy would lead to the creation of a stronger alloy. In an effort
to quantify the velocity ﬁelds within these convective mechanisms, two methods were
utilized to study the solidiﬁcation process, one being quantitative and the other
qualitative. An adapted shadowgraph technique was used as a means of obtaining a visual
picture of the process, from the formation of ﬁnger-like instabilities during the early
stages, to the formation of plumes at later stages of the solidiﬁcation.

The second method, Molecular Tagging Velocimetry (MTV), was utilized to
obtain instantaneous velocity ﬁeld measurements of the ﬂow ﬁeld above the mushy zone
and above the chimney. The current study improves on the results of the previous work
with the use of a two-component measurement and an improvement of the spatial

resolution and data density by a factor of two.

CHAPTER 2: EXPERIMENTAL PROCEDURE

2.1: Unidirectional solidification and Experimental Setup

This study explores the convective mechanisms present during solidiﬁcation by
way of unidirectional cooling of a ﬂuid from the bottom. A hypereutectic aqueous
ammonium chloride solution (26% weight percentage) was used which is known as a
popular metal analog to metal alloys (Liu et. a1. (1999), Wirtz et. a1. (1998), Beckermann

et. al. (1996)).

2.11 Details of Ammonium Chloride

Ammonium chloride is a popular low temperature analog due to the transparent
nature of the ﬂuid that allows for optical access. Also the temperature range at which
solidiﬁcation occurs provides a means for studying the convective mechanisms present in
metal alloys. The equilibrium diagram for ammonium chloride is shown in Figure 2.1.1
where the dashed lines indicate the initial concentration and the ﬁnal equilibrium
temperature of the aqueous ammonium chloride used in the experiment while the dot
above the liquidus line indicates the initial starting conditions. A detailed calculation for
the preparation of the aqueous solution and the molecular tracers used in this experiment
is provided in Appendix A.

The aqueous ammonium chloride solution was initially kept at 23°C. Once
cooling of the ﬂuid commences, the temperature of the ﬂuid decreases with no change in
concentration until the liquidus line is reached. The process will continue along the
liquidus line, which implies that as the ﬂuid gets cooler, ammonium chloride crystals

precipitate out of the solution since the weight percentage of the ﬂuid has to reduce to

maintain equilibrium. This process continues until the ﬂuid reaches a temperature of
-14°C (above the eutectic solid line). Limiting the minimum temperature that the ﬂuid
achieves prevents the solidiﬁcation process from forming an eutectic solid and simpliﬁes

a numerical simulation.

 

- ,. .- -.. .-, ,--- -.
45" -l

 

 

 

 

 

i LiquidusLine :
35- .
6 25? NH‘O'HZO salmon Initial starting .
° ’ terrperature j
T; and concentration .
5
9—0
‘3
8 NH4Clcrystalsolution
E
o
F 1
1
________ T_-_--________________é.
EutecticSolid
"250‘ U ‘4‘ ‘ ”é‘HATa‘Hie” ‘ ‘2‘o‘w‘2i4‘ ‘2'3‘ ‘ T33 ‘ ‘ ‘3‘6‘ ”‘40
Werght peroentageofNH4C|

Figure 2.1.1 Equilibrium diagram of ammonium chloride.

During the cooling of the aqueous ammonium chloride solution, a vertical
temperature gradient is established within the salt solution. This temperature gradient
causes a stable density gradient to occur and results in a buoyancy force. As the
ammonium chloride concentration near the mushy zone decreases due to the precipitation
of ammonium chloride when the temperature is reduced, a solutal buoyancy force is
established within the mushy zone, which is opposed to the buoyancy force induced by

the temperature gradient. However, the solutal buoyancy force becomes a dominant

factor despite the presence of the stabilizing thermal force. This then causes large-scale

convection to be induced in the bulk ﬂuid.

2.1.2 Experimental Setup

A schematic of the test section used to conduct experiments is shown in Figure
2.1 .2. The solidiﬁcation system utilized is similar to the one described in Magirl and
Incropera (1993) and is a modiﬁed version of a setup that was originally designed and

used by Wirtz et. a1 (1998).

‘ljll

shrirkage compensdor

  
 
 

 

Optlcs + Laser

lllllll

 

   
        
  
   
 
   
  
  
 
 
  

 

 

/
i/

/tleld of view

   
 

 

 

canera
UUWS hi

otmrved plume

lh uld iagbn N
solidiﬁcation front

' ' - - . E Opﬂcs+Lasar
solldlag‘nn \‘ '.~ 5 _— ________.__

coolarl outlet ,
9

 

 

 

 

    

 

nuclea ors

plexidas box /0 '
stainless steel base plate

/c oolmt lrlet

   

r

Figure 2.1 .2. Schematic of test section.

The cooling of the ﬂuid within the test section is achieved with the use of a base
stainless steel plate as a heat transfer medium, which is exposed to a coolant bath on one

side to provide for the cooling and aqueous ammonium chloride on the other within the

test section. An enclosure made of 1/4" quartz glass measuring 52 x 89 x 100mm was
placed over the stainless steel base plate and sealed with high vacuum grease and gaskets
to form the, test section. Quartz glass was used due to the fact that it absorbs very little of
the laser energy and thus is suitable for the transmission of laser beams and results in less
attenuation.

During initial experimental trials, it was observed that convective plumes due to
the density instability develop within the test section in a random manner. However, each
experimental trial would yield approximately 11 chimneys within the test section. To
better predict the location of these chimneys, 11 square nucleators were embedded in a
regular manner in the base stainless steel plate to induce ﬁnite disturbances to the
formation of the mush, which would allow the channels to form at those approximate
locations. A schematic of the stainless steel plate and the coordinates of the stainless steel

nucleators are provided in Figure 2.1.3.

89mm 11.1lmm11.17mm

r

H .(_*

A

 

F'l
J

 

El

Row 3 --*---—f

r 11.71mm
r t: {5——
i 14.29mm

 

mt
I"!
11.!

Row 2 —-> {E E

/ f .2....

 

I‘ll
Ill

Rowl —>

 

 

 

 

 

 

 

 

 

 

I

Mushy zone Plume tagging
‘38ng “381011 region

 

Figure 2.1.3. Schematic of base plate and embedded rectangle nucleators (shaded grey)

Experiments with and without the nucleators were performed to study the effects on the
solidiﬁcation process of adding nucleators to the base plate to control the formation of the
plumes. No signiﬁcant effect was discernible and the growth of the mushy zone with
respect to time remains consistent with experiments done without the use of nucleators.
To ensure that the nucleator positions remain ﬁxed within the test section and do
not change between experiments, the nucleators were shaped so that an additional round
base protrudes below the nucleator. The round base is embedded into the base stainless
steel plate and glued-in with Loctite© Black Max glue. Figure 2.1.4 shows the schematic

of the nucleator.

1.59mm

   

1.59mm

Embedded
base plate

    

V

3.18mm (above base plate)

Figure 2.1.4. Schematic of nucleator

The nucleators act to induce the formation of channels associated with the
rejection of buoyant ﬂuid from the mushy zone. By ﬁxing the arrangement of the
channels in a regular manner by the use of nucleators, the formation of plumes is
predictable - facilitating optical data acquisition for velocity measurements using MTV.

The stainless steel base plate is attached to a cooling chamber made of Plexiglas
(see Figure 2.1.2). The cooling chamber has an inlet and outlet to allow the cooling ﬂuid

to pass. The cooling ﬂuid used is a 50% mix of water and generic automotive anti-freeze

which is refrigerated by an external cooling bath (N eslab, RTE-140) and pumped through
insulated tubing into the cooling chamber. To reduce the heat transfer to the
surroundings, the cooling chamber was covered with weather strips. The cooling
provided to the test section would then initiate the crystallization process.

Once the quartz enclosure is placed above the stainless steel plate, the test section
is ﬁlled with a binary alloy; a hypereutectic (26%) ammonium chloride. A Plexiglas lid is
then placed over it to seal the test section. Shrinkage compensators were ﬁtted on the lid
to allow additional ﬂuid to be supplied during the experiment to compensate for
solidiﬁcation shrinkage. A window made of quartz glass was also ﬁtted to allow
penetration of the laser lines from the top.

However, this setup proved to be inadequate due to the amount of condensation
that forms on the outer quartz walls whenever an experiment was performed. This is
currently partially circumvented by running the experiments on low humidity days only,

which highly restricts the progress of this work.

2.2 Details of Molecular Tagging Velocimetry

2.2.1 MTV Chemicals

A phosphorescent compound used in the MTV technique was added to the
aqueous ammonium chloride solution to serve as molecular level tracers. When excited
with an excimer laser (308nm), these tracers emit light that has a lifetime of ~5ms. A grid
pattern created with the laser would be imaged at two successive times within the lifetime
of the tracer to capture two consecutive images. The two images are correlated to obtain

the velocity vectors and thus the velocity ﬁeld measurements.

A molecular complex is suitable for molecular tagging applications if its lifetime as a
tracer is long enough relative to the ﬂow convection time scale to allow sufﬁcient
displacement of the tagged regions. The MTV technique has successfully been applied to
the characterization of a variety of ﬂows at the Turbulent Mixing and Unsteady
Aerodynamics Laboratory at Michigan State University including the ﬂow ﬁeld
measurement within an internal combustion engine, the wake of an unsteady airfoil and
vortex ﬂows and it is now being used for small-scale applications in the solidiﬁcation of
an alloy. The MTV Triplex consists of a lumophore, l-Br-Np; an alcohol, cyclohexanol;

and a sugar, cyclodextrin; and is collectively known as l-Ber-GB-CD.ROH. The

concentrations of the MTV triplex used in this study are 2 x 10'4M for GB-Cyclodextrin,
0.06M for alcohol and a saturated amount of bromo-napthalene was added (Gendrich et.
aL,1997)

The MTV technique involves mixing a tracer into the ammonium chloride solution.
Therefore the presence of the tracer could potentially affect the solidiﬁcation by changing
the crystallization process, the thermal behavior or the ﬂow behavior. The typical
concentration of the ammonium chloride solution is 26% by weight, whereas the highest
concentration of the three components of the MTV triplex is less by at least three orders
of magnitude (in molar concentration). Due to the small concentrations of the tracer used,
it is expected that the physical properties do not change signiﬁcantly, so that the thermal
and ﬂow behavior of both systems with and without the tracer would be equal. To
conﬁrm this, experiments with and without using the tracer were performed. A
comparison between the experiments did not indicate any change in behavior including

the temperature range in which crystallization commenced or the development of the

10

mushy zone. Thus, the solidiﬁcation and convective behavior were not noticeably
inﬂuenced by the presence of the tracer molecules.

During initial experimental trials, it was observed that the phosphorescence of the
tagged ﬁeld of view appears brighter than tagged ﬁeld of views of conventional
applications of MTV within water and yielded longer delay times between the undelayed
and delayed images. This prompted the investigation of the lifetime of the mixture of
aqueous ammonium chloride and the MTV chemicals to determine how the addition of
ammonium chloride to the solution has affected the lifetime values. The lifetime of the
MTV Triplex in the presence of different concentrations of ammonium chloride was

determined and the results are shown in Figure 2.2. 1.

 

7_o.---,----,---l...--,..-f,
.

65-v

.
.
p .

55. _. ,, ,. I. ,,,,, .

Lifetime, I (ms)

50 ._ . ......
4.5.. ,,

4.0. .. ,,,,, . _

 

 

 

l 1 ; .
3.5 A L ‘ L 4 ‘ ‘ L ‘ ‘ ‘ ‘ r ‘ ‘ 4 - A . l - - . - r . . r -
0 5 10 15 20 25 30

Weight Percentage of Anmonium Chloride

Figure 2.2.1. Plot of lifetime as a function of weight percentage of ammonium chloride +
MTV chemicals (Cyclohexanol)

11

In the ﬁgure, it can be seen that the lifetime value without ammonium chloride
matches the previously published lifetime value of the MTV triplex of ~3.7ms at the
same mixture concentration (Gendrich et. al., 1997). By adding ammonium chloride to
the MTV solution, it can be seen that the lifetime increases signiﬁcantly before tapering
off at ~ 6.7ms. At this point, the weight percentage of ammonium chloride present in the
solution (~20-25%) is approaching the saturation limit of ammonium chloride solubility.
Due to the increased lifetime of the molecular tracers in the presence of ammonium
chloride, longer time delays and thus measurements of slower velocities are possible. It is
theorized that the presence of ammonium chloride shields the molecular tracers from
being quenched after tagging by the laser lines.

However, it should be noted that lifetime is measured as the time when the
intensity drops to e‘1 of the original intensity. Measurements can still be made at longer
time delays with the use of proper detectors, which in this case is an image-intensiﬁed
camera. Using delays in the range of 50-60ms, the MTV Triplex alone would not yield
enough light to capture the delayed image when the intensity would have dropped to

approximately 6'16

of the original intensity. It is through the presence of ammonium
chloride, which allows the intensity to be raised to approximately e'9 of the original

intensity, and the use of an image intensiﬁed camera that it is possible to measure such

slow ﬂows. More details of the MTV triplex can be found in Gendrich, et. al. (1997).
2.2.2 Optics Setup

Creating the appropriate tagging pattern for velocity measurements involves an

elaborate setup of optics. During the setup, many lenses were used to manipulate the laser

12

beam. To illuminate the ﬁeld of View effectively, we need to make a laser sheet that is as
thin as possible in the physical world so that the tagged ﬁeld of view is planar. Thinning
the laser beam requires an accurate setup of optics and how thin it can get is limited by
the diffraction limits of the optics.

To achieve this, a 20ns, lOOmJ/pulse beam from an excimer laser (Lambda Physik
LPX 220 iCC ﬁlled with XeCl) was initially passed through a VFL (Variable Focal
Length) setup consisting of a 150mm and 50mm cylindrical lens to increase the aspect
ratio. This, in effect, reduces the outgoing rectangular laser beam from the laser to a thin

sheet that is less than 1mm thick. The optics setup is illustrated in Figure 2.2.2.

5050 Beam splitter

' l 1%

1 50mm
----------------- cylindrical CYIIIldl‘ical

 

 

 

 

 

 

  

99.04.0900 9 o
Ooeeoe§§¢e§o~
0090 ooo§+0¢e .
066909966000
..

Q

o

 

OOO++ 06*.
I 9+0.

0‘.-
oeeooo§¢eo$eo

......¢ooo+¢oo
‘ O O...§O o

 

 

 

 

 

 

1.

 

 

 

 

 

 

 

 

 

 

 

 

--------.1 ll

 

 

 

 

L
\ Test

section

Figure 2.2.2. Optics Setup

To achieve two intersecting laser lines, the main laser beam is split into two using a 50:50

beam splitter. Aﬁer this, the laser sheet is basically reﬂected using mirrors and passed

13

through beam blockers to create the intersecting grid pattern. The optics setup in Figure
2.2.2 was used to measure the ﬂow ﬁeld near the mushy zone where the velocities are
slow - on the order of 1mm/s. In a ﬁeld of view of ~14mm x 10.5mm, a total of 192
intersections were achieved with 12 horizontal and 16 vertical laser lines at a 90°
intersection. This corresponds to a spatial resolution of 775nm for the velocity
measurements above the mushy zone. The ﬁeld of View just mentioned corresponds to a
1.6:1 image ratio since the CCD array size is 8.8x6.6mm.

In designing the beam blockers, the goal was to create an array of slots that are
very thin in order for the smaller features within the ﬂow ﬁeld to be resolved. However,
due to equipment limitations, the smallest slot achievable was 0.015cm (0.006") with a

spacing of 0.058cm (0.023"). A schematic of the beam blocker is shown in Figure 2.2.3.

~0.015cm
slot

~0.058
cm
spacing

 

Figure 2.2.3. Beam blocker design

If the slots or spacing between them were cut any smaller, warping would occur and the
beam blockers would yield non-uniform laser lines. This area could be improved upon in
the future by employing more cutting edge machinery or chemical etching to create
thinner slots.

In previous experiments, it was found that as the mushy zone grows and nears the
grid pattern, the intensity of the lower laser lines closest to the mushy zone increases due
to the lower temperature present, causing the image to saturate. Reducing the gain causes
the upper lines farther from the mushy zone to be too dark for correlation. Thus the lower
laser lines (closest to the base plate) were intentionally made to be less intense so that as
the mushy zone nears the grid pattern, the image is still below saturation while
maintaining sufﬁcient signal to noise ratio over the whole image. One concern in
irradiating the ﬂow ﬁeld with a laser is the potential heating of the ﬂuid. The heating
effect of the laser was calculated and a temperature rise of less than 0.09K/pulse (worst
case where all the energy is absorbed) over the whole ﬁeld of view was determined. This
is negligibly small and is not expected to affect the ﬂow ﬁeld within the bulk ﬂuid.

Images from three ﬁelds of view were acquired at different times to capture the
ﬂow ﬁeld near the mushy zone. Figure 2.2.4 shows the location of the ﬁelds of view
where measurements were performed. Each ﬁeld of view covers an area of ~14.5 x
10.5mm with the bottom of each ﬁeld of view being y = 10.5mm, 21.0mm and 31.5mm

respectively from the base plate.

15

 

Note: Grey nucleator
marks a single nucleator.
Black nucleator marks two
nucleators set back to
back. Refer to Figure 2.1.3

  

 

 

 

 

Quartz
/ Wall
Y
x

 

Figure 2.2.4. Measurement ﬁeld of views for mushy region

Note that the measurements made above the mushy region are located between
two nucleators where no plumes would form. The ﬁeld of views are located on the same
plane as Row 1 (refer to Figure 2.1.3 for row designations) and the nucleator in this
region is located on Row 2. However, the grid pattern used for measurements in the
mushy zone could not be applied to measurements of the plume. As the plume rises
vertically, the edges of the plume causes the vertical laser lines to refract and result in a
"fuzzy" pattern in certain areas of the image, which may not correlate well. An example

of a refracted grid line is shown in Figure 2.2.5.

16

 

 

Good grid Laser lines
intersections refracted as it
strikes the plume

Figure 2.2.5. Reﬁ'acted laser lines for plume measurement

To circumvent this difficulty, the optics setup was rotated by an angle of ~ 45° so
that the laser lines penetrate from the side of the quartz test section. By having the grid
pattern intersect the plumes from the side, the possibility of having the laser lines
refracted by the rising plumes would decrease and thus produce better quality images that
correlate better. However, when the laser lines penetrate the side quartz walls at an angle,
the spacing between the laser lines increases and causes the grid density to decrease,
thereby lowering the data spatial density. The spatial density of the measurements is
960nm. Future considerations to improve the data spatial density are to utilize an
additional lens of the appropriate focus length to focus down the beam and reduce the
width of the laser lines and also the distance between them. Beam blockers with smaller

slots would also contribute to increasing the data density.

 
 
 
 
 
 
 
 
  

50:50

Beam

Splitter

50mm
Cylindrical

1 50mm
Cylindrical

 

 

~ll.17mm
m

Figure 2.2.7. Minimum height of grid pattern from mushy zone surface

The schematic of the modiﬁed optics setup is provided in Figure 2.2.6.With this setup,
the bottom of the grid pattern is always at a distance from the tip of the chimney. Figure
2.2.7 shows the minimum height that the grid pattern would be able to approach the
chimney before the mushy region blocks penetration to the laser line.

From the schematic above, it is clear that the diagonal grid pattern can only be as
close as ~7mm to the chimney. Thus an additional method needs to be devised to
measure the plume velocity very near the chimney. Since the vertical laser lines could not
penetrate the vertically rising plumes and the ﬂuid is mainly rising vertically, horizontal
laser lines were passed just right above the chimney to measure the V-component of

velocity using simple line tagging, before the plume takes on a helical motion.

2.2.3 Correlation of MTV images

The images obtained from the experiments were correlated using an in-house
correlation code written in C. This was developed at the Turbulent Mixing and Unsteady
Aerodynamics Laboratory (Gendrich et. al., 1997). To ensure that the displacements
obtained from the correlation are physically meaningful, a scale image needs to be
obtained after each experiment. This scale image acts to rescale the data from the original
pixel units to physical units and thus yield velocity measurements in mm/s.

The measurements obtained need to be placed onto a regular grid as the original
MTV grid is most likely arranged in an irregular fashion. A local least squares ﬁt of a 2"°
order polynomial was performed on the irregularly spaced velocity points. In using the
second order polynomial ﬁt, a minimum of 6 and a maximum of 8 surrounding velocity

points were used to remap a single velocity data on the regular grid. When the velocity

19

data are remapped onto a regular grid, the grid spacing is kept to be approximately the
same as the average grid spacing in the original MTV grid so that the data density in the
irregular grid is the same after remapping it onto a regular grid. The accuracy of the
remapped data is discussed ﬁrrther in Cohn et. a1. (2000).

In the measurements of the mushy region, the laser lines were orthogonal with
respect to each other and thus the errors in the estimated x- and y- components of the
displacement vector (5y and ﬁx) in the correlation procedure are comparable. However, as
the intersection angle increases, as in the case of measurements above the chimney, |5x|

values decrease at the expense of increasing |5y|. As expected, in the limit of the grid

intersecting angle, 0 = 180°, the determination of Ay becomes ill posed (Gendrich et. al.
1996)

Based on the results published in Gendrich et. a1. (1996), at intersection angles of
less than 130°, the error in the y displacement and image signal to noise ratio (S/N) of
better than 4 should produce displacement estimates with better than 0.1 pixel accuracy in
both the x- and y- directions. Figure 2.2.8 is taken from Gendrich et. al (1996) which
illustrates the accuracy expected from the correlated images with tagging patterns at
different intersecting angles.

Most of the data yielded a S/N ratio of 8 and an intersection angle of ~1 10°. This
translates to a displacement uncertainty of less than 0.05 pixels which in physical units
relates to velocity measurements with an uncertainty of : 0.02mm/sec within the plume.
Measurements above the mushy region would result in a slightly higher accuracy due to
the orthogonal intersection. The displacement uncertainty corresponds to a 95%

conﬁdence level, which implies that 95% of the computed displacements have errors less

20

than or equal to the displacement uncertainty. This is a more conservative measurement
compared to the RMS value. If the error distribution is Gaussian, then the RMS error

estimate is 8q)RMS=O.5(5q)o,95

 

 

 

 

 

 

 

 

()e8 1 l I l l
SIN = 2 6—
0.7 - SIN =4 --l-----
S/N=8 --El--
0.6 . SIN: 15 -~><--
2? 0.5 -
.2
9;" 0.4 ~
.3: 0.3 »-
0.2 V "
0 «newsm—
90 100 110 l20 130 I40 150

Grid intersection angle, 6 (degrees)
Figure 2.2.8. Displacement sub-pixel accuracy versus 0 for different values of S/N
(Gendrich et. al., 1996)

Another aspect that warrants discussion is the effect the intersecting angle has on
the individual component displacement errors. This is illustrated in Figure 2.2.9. It is
apparent from Figure 2.2.9 that as the intersecting angle increases, the error magnitude in
one component of the velocity measurement will increase accordingly (40% worse in one
component in our case). In the measurements of the plume, the acute angle is in the x-
direction and results in the x- component having a higher magnitude of error than the y-

component. A more detailed discussion of sub-pixel accuracy and other correlation issues

are provided in Gendrich et. a1. (1996).

21

 

x I T I I I

7 - SIN : 2 9_
S/N = 4 -+....
(l l' 5m 3 8 .....E
S/N = ]5 «xx...
Iawllufx

 

 

 

 

90 Hill Ill) I20 130 I411 151)

(irirl intersection angle. 9 ldegrccsl
Figure 2.2.9. Effect of intersection angle and S/N on the ratio of x- and y— displacement
error magnitudes.

2.3 Image Acquisition System

In imaging the ﬁelds of view, an intensiﬁed camera and various electronics were
used in order to detect the tagging pattern in low light conditions. A 105mm Micro
Nikkor ( f / 2.8 ) lens with an additional close-up lens was used in conjunction with an
intensiﬁed camera, a Xybion Gated Intensiﬁed CCD Video Camera (Model ISG-350) and
the Video Camera Control Unit (Model CCU-01). To provide the electronically gated
camera with a gate signal, a Stanford Research System, Inc. (SRS) Four Channel Digital
Delay/Pulse Generator was used. Two TTL pulses were generated and combined using an

"AND" gate and fed to the camera. The physical setup is illustrated in Figure 2.3.1.

22

838330 was 8083 mo 953 ”BREE A MN 33mm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

89>.

.833
Sam

 

 

«Hugo
“39:0
83, sesame
a 8083
, E O O O 0 839m
0 O O O
r T 393
3:9:
«$0
9.2 Rama _
.. _ l _
$.3me ...—EH. 0 0 MW 0
0
0000000000 can > > oomuooomaonoo
gunman. nun I 00000 Wmoumm 000E
DDDDDD ODD DD UDU.I\ < — J
D nmn -OQ
_ an - on J O 0 00 mm 33an seesaw seesaw
“anmhw soaouom 303.5 oaooaozmoao “.3an m “ﬂotsam?
—
.8 mmCH commCH

23

Figure 2.3.2 details the input and output signals fed to the camera and laser for
synchronization:

 

Delay
Generator #1

 

 

 

 

Delay
Generator #2

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

E 6115 pulse width E E 8ms pulse widthi
E i E E i E .......... i r .......... -
i ii I I l I I 2ms pulse width E
: <——> : 5<————> 5 5 l 5
: T+0.064s E : T+0.ll7s : g , ;
5 <——> L 5
AND E T + 0.063958 E
Gate : ..................... .
E 8ms pulse width 5
ST + 0.0645 l 5
I (—-) : V
I(———> — I Laser
5 r + 0.1 175 5

(__

 

Camera

 

 

 

Figure 2.3.2. Schematic of typical timing used for gating the camera and laser

24

With reference to the timing diagram in Figure 2.3.2, the ﬁrst gate (6115) allows
the camera to capture the undelayed image when the laser ﬁrst ﬁres while the second gate
(8ms) allows the delayed image to be captured at a later time. In selecting the time delay
between the delayed and undelayed images, two factors need to be considered, one being
the range of timing which allows for full frame capture of the image and the other being
the amount of light available for detection by the CCD.

The delay time used is constrained by the timing of the camera, which only
enables double gating in a speciﬁc range due to the RS-170 standard utilized by the
camera. In order for the two images (undelayed and delayed) to be captured in full frame
mode, the gating signals need to be positioned within the ﬁeld overlap so that both the
even and odd ﬁelds would be exposed in that period to form a complete image. This is

illustrated in Figure 2.3.3.

 

 

 

 

 

 

 

 

Odd E
ﬁeld W“
4 < i” ) ﬁeld %
Vertical . .
Drive (VD)
Field» .
Overlap
Gating H
Signal to
Camera Acquires Acquires
Undelayed delayed
Image Image

Figure 2.3.3. Schematic of gate signals with respect to the vertical drive of the camera

25

If the gates were positioned in either the odd or even ﬁeld only and not in the
overlap region, the image acquired would only show every other line, as the other ﬁeld
was not exposed to the ﬁeld of view in the gate period. To measure the low speeds
present in this experiment, the delay time between the undelayed and delayed image
needs to be more than 50ms. Since the time between vertical drives is ~16.66ms, the two
gate periods need to be separated by three vertical drives to attain that delay time. This is
shown in Figure 2.3.3. Longer delay times (an additional 16.66ms) could not be used, as
by then, there would not be enough light left to capture the image. Since the gating
signals are separated by three vertical drives, one overlap period in between the
undelayed and delayed image would be unshuttered. Thus there would be an empty frame
in between each image captured, which reduces the acquisition rate by half. For every 60
images that were captured, 30 would be blank frames and of the other 30 remaining, half
are undelayed and the rest delayed leaving 15 pairs for correlation.

The time delay between the undelayed and delayed image used is in the range of
54.995ms-56.995ms measured from the center of both gates. At a later stage in the
experiment, when bleaching effects come into play, the gate period needs to be increased

accordingly to allow sufﬁcient light for image acquisition.

2.4 Experimental Procedure

One of the main concerns in any experiment is the proper characterization of the
boundary conditions. In this experiment, it is found to be even more crucial since it is a
temperature-based experiment and the surrounding temperature and humidity in our

location changes with the season. Thus all experiments were completed in a very narrow

26

time frame to ensure that the boundary conditions remain essentially the same since the
experimental setup did not take into account environmental changes.

Before starting the experiment, the Neslab external cooling bath was initially set
to 23°C and the ﬂuid circulated within the Plexiglas box (see Figure 2.1.2). This was
done each time before the experiment to ensure that the base plate is always at the same
temperature so that the operating conditions remain the same in between experiments.
Any entrained air within the cooling box is evacuated to the external cooling bath by
tilting the test section. Alter the coolant has been circulating within the cooling box for
some time, the ﬂow of ﬂuid was impeded and the Neslab coolant bath was brought down
to a temperature of -21 .5°C.

The test section is then ﬁlled with aqueous ammonium chloride and sealed. Once
the coolant in the Neslab coolant bath has achieved a temperature of -21.5°C, the valves
were reopened to allow the experiment to commence. Since the ﬂow lines and the
cooling box were set to 23°C initially, the chilled coolant naturally experiences a drop in
temperature when the ﬂow is restored. This is taken into account by intentionally setting
the temperature lower than the desired temperature of -18.6°C. When the ﬂow is restored,
the temperature of the external cooling bath drops a few degrees to around -18°C and the
temperature was reset to -18.6°C which is the desired ﬁnal temperature. This procedure
reduces the time needed for the temperature of the base plate to reach -14°C.

Countless experimental trials and calculations done by Kay Wirtz in his thesis
arrived upon the temperature value of -18.6°C. In the previous effort at MSU by Kay

Wirtz, temperature measurements were made at different locations of the test section and

27

it was concluded that the base plate temperature would attain a value of -l4°C if the
coolant bath was set to a temperature of -l8.6°C (Wirtz et. al, 1998).

One of the ﬁrst experiments that were done was a visualization of the
solidiﬁcation process to provide a means for understanding the complex dynamic ﬂow of
a uni-directional solidiﬁcation. In visualizing the convective plumes that develop during
solidiﬁcation, a Shadowgraph type visualization was utilized. This shadowgraph
technique is slightly modiﬁed from that normally used by other researchers. Most
shadowgraph techniques utilize a parallel light source from behind the interrogation area,
which would project the shadow of the density gradients within the ﬂuid onto a screen.

The method used here utilizes ambient room light for illumination and an
alternating vertical black and white background to contrast the edges of the convective
plumes. One of the main factors in this setup that allows the plume edges to contrast
against the alternating black and white lines is the direction of the scattered light. If the
test section was exposed to room light, the illumination from above the test section would
be stronger than that from behind it making it difﬁcult to discern the presence of the
plumes. However, when the test section is enclosed so that illumination from above is
impeded, the main illumination is from the back of the test section, which allows this
technique to work very well. Another factor that needs to be considered in producing
good image contrast is in the wavelength and the width of the alternating black and white
lines placed a distance away from the test section. These lines were placed approximately
2 meters away from the test section and cover the whole background of the test section.

Figure 2.4.1 illustrates the setup that was used in the visualization.

28

/, Shrinkage

\ A// compensator
r, ,
*3 3:4

’D
p . .\ltern;lting
‘ 9‘ black and

5'

“hile lines
I

I

. section .._. -
1: .1 . ‘
.. I y
F.- , man-i .. in

Figure 2.4. 1. Setup for Visualization

 

When the images are taken, the width of these lines needs to be such that it is the
same as the distance separating two adjacent plumes in the ﬁeld of view. This in effect
allows the edges of the alternating lines to coincide with the plumes and provide for the
contrast necessary to discern the plumes. Other litre widths and wavelengths were
experimented with but these provided unsatisfactory results.

However, this method would display density gradients well only in the horizontal
direction due to the edges of the plume being mainly in the vertical direction. Double
diffusive layers are less obvious compared to the standard Shadowgraph technique
although still very much visible. If the vertical alternating black and white lines were

rotated to be in the horizontal direction, the vertical density gradient i.e. double diffusive

29

layers would be more obvious but sacriﬁces the ability to contrast the horizontal density
gradients. Before starting the experiment, a reference image was taken of the test section.
This was used to subtract out the background so that the images do not show the
alternating black and white background. To acquire the images used for visualization, a
Pulnix TM-1040 (1K x 1K) camera was used in conjunction with a Coreco data
acquisition board.

In making whole ﬁeld velocity measurements with the MTV technique, various
aspects of the experiment need to be kept in check to ensure that the laser tagging process
is unimpeded. The quartz test section needs to be exceptionally clean so that the laser
penetrates the test section cleanly and does not cause reﬂection, which would deteriorate
the image quality. In the process of aligning the laser beam, it is imperative that the laser
beam is always "square" with the optics so that minimal reﬂection would occur and the
shape of the laser beam is not distorted. In the event that the laser beam is not "square",
the laser-tagging pattern may cause reﬂection within the test section or induce irregular

illumination of the tagging ﬁeld and cause a reduction in the signal to noise ratio.

30

Chapter 3: Results and Discussion

3.1 Visualization

One of the ﬁrst sets of data to be taken for this experiment was the visualization
of the process from when the cooling starts to 2 hours into the solidiﬁcation. The
visualization serves as a means to better understand the ﬂow phenomena present during
the solidiﬁcation process and also to illustrate the location where MTV measurements
were made in this study. The method used to obtain high quality pictures was an adapted
Shadowgraph method as described in Section 2.4: Experimental Procedure.

During the experiment, many time ticks were recorded which include images of
the solidiﬁcation process at 0, 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 25, 30, 35, 40, 45, 50, 55,
60, 70, 80, 90, 100, 110 and 120 minutes. At each time tick, 120 images were taken in a
time frame of 20 seconds at an acquisition rate of 6H2 (Four frames are skipped for every
' frame acquired). Although the video sequence appears jerky due to the low acquisition
rate, it is the only way to save disk space and have the ability to acquire a longer
sequence. A longer sequence allows us to capture a more complete picture of the
evolving plumes and other characteristics of the ﬂow.

During the initial stages of cooling, rapid growth of the mushy layer indicated the
presence of substantial ﬂuid under-cooling. This results in the formation of many
ascending water-rich ﬁngers and the formation of ammonium chloride crystals, which
would settle to the bottom of the test section to make up the mushy region. Figure 3.1.1

shows the regime of small ﬁngers and the formation of the mushy region.

31

 

0
1591:” Superposed image of

nucleator location

Figure 3.1.1. Small ﬁngers and ammonium chloride crystals at 8 minutes into the
solidiﬁcation.

It is observed that at 4 minutes into the solidiﬁcation process, small ﬁngers start
to form at the center of the base plate and can be seen rising, although they quickly lose
their solutal identity and mix with the surrounding ﬂuid. At 6 minutes, more plumes
maintain their solutal identity and are able to traverse higher. Due to the large Lewis
number of ammonium chloride, the water-rich ﬁngers retained their compositional
identity and were clearly visible to the top of the test section before they would begin to
collapse. (Magirl and Incropera, 1993)

A substantial amount of ammonium chloride crystals can be seen falling from
~5mm from the mushy region due to the substantial ﬂuid undercooling present just above
the mushy region during the early stage. As solidiﬁcation continues, the amount of

ammonium chloride crystals descending onto the mush lessens but the ﬁngers become

32

more vigorous and would sometimes advect small crystals upward into the ﬂow ﬁeld. A
sequence of images is shown in Figures 3.1.2 to illustrate the evolution of the mushy

region as well as the plumes.

 

18 minutes 22 minutes 26 minutes

Figure 3.1.2. Images show visualization of solidiﬁcation process from 4-30 minutes.

Notice that during the initial stages of solidiﬁcation, water-rich ﬁngers are manifest
throughout the entire ﬂow ﬁeld. At ~18 minutes, stronger ﬁngers start to appear that
reach all the way to the top of the test section. As the plumes strengthen, surrounding
ﬁngers start to decrease in quantity as neighboring plumes compete for the supply of
water rich ﬂuid from within the mush. The strengthening of the plumes can be seen from

18 minutes to 30 minutes in Figure 3.1.2

33

 

80 minutes 90 rrrinutes 100 minutes 120 minutes

Figure 3.1.3. Images show visualization of solidiﬁcation process from 40-120 minutes.

At around 40 minutes as illustrated in Figure 3.1.3, it becomes apparent that the
plumes are developing above the nucleation sites although there are approximately 14
plumes at this stage compared to the 11 nucleators. At a later stage in the solidiﬁcation
process (~60 minutes), due to the ﬁnite disturbance induced by the square nucleators,
most of the ﬁngers have died down and consolidated into stronger plumes at locations
where the nucleators are positioned. These plumes originate from within the mushy layer
where lighter ﬂuid is rejected due to the formation of the mush. At later stages of the
experiment, double diffusive layers form, which are indicated by the horizontal diffusive

interfaces at 90 minutes in Figure 3.1.3.

34

Small-scale ﬁngers become less visible late in the solidiﬁcation process and
completely disappear at around 120 minutes. One point of interest can be observed at 70
minutes where plume buckling is visible at the top of the test section. However, the
plumes seem to recover and take on a three-dimensional helical motion again later in
time. Magirl and Incropera (1993) noted this phenomenon in their experiments and
concluded that as the plume oscillated in its plane, ﬂuid motion in the outer bends is
retarded by viscous drag but as ﬂuid continues to ascend from below the bend, it induced

a biﬁrrcation of ﬂow in the bend causing the plume to fold onto itself.

shedding

  
 

Double ‘3,“ =7 ' .
diffusive I , T . .
layer '

\Ill\ll_\

region

\ :llle} elllmllw

ll.llm

+-»—>
a ' I A f

Figure 3.1.4. Image of solidiﬁcation process at 2 hours.

35

As the plumes rise above the chimney, at about 10m away, the plumes start to take on a
three-dimensional helical motion, which has been reported by many researchers (Stuebe
and Hellawell, 1992; Magirl and Incropera, 1993). When the plumes begin to oscillate,
wisps of ﬂuid separate from the plume and descend as shown. Signiﬁcant shedding can
be seen at the plume edges as it makes the turn, with the shedding becoming more
pronounced as the solidiﬁcation continues. Figure 3.1.4 shows a few important features
present in the ﬂow. (Note: There are two plumes, overlapping each other starting on the
far left and alternates to the far right due to the nucleator positions, see Figure 2.1.2 for
nucleator positions)

As the solidiﬁcation continues, the growth rate of the mushy zone decreases due
to a slower cooling rate. The mush that has formed acts as an insulated boundary between
the ﬂuid and the base plate thus in effect shielding it from being cooled further. The
height of the mushy zone at a given time is measured to be the distance ﬁom the base
stainless steel plate to the valley of the mush as shown Figure 3.1.4. From the plot in
Figure 3.1.5, signiﬁcant growth of the mush occurs in the early stages of the experiment
from ~5 minutes to 40 minutes before it starts to taper off and level out at a maximum
height of approximately 27mm. Two separate trials were conducted showing that the
growth rate of the mushy zone was consistent. Magirl and Incropera's experiments
utilized a cooling temperature of -30°C, which resulted in an earlier formation of the
ﬁngers as compared to the present study, which cooled the base plate to a minimum

temperature of -14°C.

36

Height of mushy region (mm)

 

v v v ' v v 7 1+7 1 1 v r1 ' v v v 1 v v v I r v 1 f 'ﬁ'f' r v v

vvvvvvvvvvvv

................................

 

 

kg; 4‘

m———o Trial #1

 

 

 

 

.i...I..

A 4

 

 

Figure 3.1.5. Plot of Height of Mushy Region vs.

37

90 100 '110‘ 120

Time

3.2 MTV Measurements of Mushy Region

One of the features in the solidiﬁcation process that is of interest is the small
ﬁngers present at the onset of the solidiﬁcation process. The MTV measurements of these
water-rich ﬁngers yield quantitative information describing the velocity ﬁeld just above
the mushy zone. Velocity measurements were made in these regions at various times to
better understand the permeability of the mush. During the experiment, a total of 60
images were taken at each time tick, of which 15 pair of images would contain the image
of the tagged ﬁeld of view. This yields 15 velocity realizations over a 2-second period.
For the ﬁrst ﬁeld of view, measurements were made from 0 minutes to 40 minutes before
the mushy region covered the ﬁeld of view completely. At about 20 minutes, the mushy
zone would have grown to the base of the laser line tagging and would continue to grow
until the entire ﬁeld of view was covered. Figure 3.2.1 shows the tagging region with

respect to the solidiﬁcation process at 8 minutes.

13.88mm

10.4lmm

 

Figure 3.2.1. Tagging region of the lSt FOV at 8 minutes into the solidiﬁcation process

38

 

Figure 3.2.2 and Figure 3.2.3 illustrate the instantaneous realizations of the velocity ﬁelds

obtained from an MTV measurement of the region above the mushy layer for the ﬁrst

ﬁeld of view where the bottom is 10.5mm from the base stainless steel plate. The y-axis

indicates the location of the ﬁeld of View with respect to the base stainless steel plate. In

Figure 3.2.2, the velocity vectors displayed at 4 minutes is an instantaneous realization

due to signiﬁcant variation between the 15 realizations. All other velocity vectors are

averaged over 15 realizations to reduce the amount of noise.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

reference vector (0.5mmlsec) —- reference vector (0.5mmlsec) _.
18 18
0 \ \ x \ - _ ‘ ' - \ t \ I 17 I \\\ I I \ ’o—‘z / / / I
17 . ._ - \ I , t ‘ \ , , s \ \\ \ , \ \, ..., / / / I
16 I ‘ r - - I 16 -' \ \ ‘ I \ \‘- - / / / / I
15 """"" " 15""/I“"l////
>14 1 / I r r I I - .... \ I , >14 \ s, / I ‘v I / / / or
I . I r ‘ I I ~ ‘ - \ '- / / I 1’ r/ / I \ I
13 \ I s \ \ I - . \ ‘ .- a 13 ’ \ o- I ‘ ‘ V/A/ / I a
12 I I / I r l I ~ \ ‘1 I 12 I I I r l r W// / " I
11 , I I I I \\ \| I 11 ‘ 1 I I I / \/// / ' \ \
10.11.? i ’ ‘.r’.‘.'i‘ LET ‘. 10.1.1 ’.{L/./l‘.(.(m/.1.{..\.‘n
2 4 6 8 10 12 2 6 8 10 12
.L “L
4 minutes 8 nunutes
reference vector (0.5mmlsec) - reference vector (0.5mmlsec) —
18:- 18?
:I ‘ ‘ r . r \ - : , ~ \ v /
,, .ttk mil- Milli;
16%' ‘ \ ‘ r I \ \ \ \ \ \ 16%! I / v I I l I 1 \ \ \ ’
15%! -‘— \ I I I ‘. \ \\ \ \ I 15%! I I / I \ I I l \ \ \ \ I
>14E."""'"““" >14f_lll/"lll\\\"
E’,,, ~~~\\\\I :IlI/IIII\\\\"’
13-° / / I i \ \ I \ \ \I 135.1 I I / y I I I I I \ \ ’ ’
12%“ I I I I . \ \ \ \ \ . 122., ’l / / I ‘ I I J \ ‘ / /
11E-‘llII/'\\\\\" 11f.r/////////\";
10.’. ’..‘..'l I rl'..‘.n§l.\.\.l‘..'..'l 10.1.4.4.’i/..’-/././i.l..\r‘..l...
2 6 8 10 12 2 4 6 10 12
.x _x
12 minutes 16 minutes

Figure 3.2.2. Velocity ﬁelds of mushy region at 4 minutes (instantaneous) and from 8-16

minutes (averaged) in the 1St FOV

39

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

reference vector (0.5mmlsec) —- reference vector (0.5mmlsec) -
18 18
I ,1 .. I/ I, .. I I
// ’l/ ‘ 1 / I ‘ 1‘ I/
17 7
'/‘/I‘ //III 1 II 1 ' ‘//"’I 1 I,
16 ' /\1‘ I ,//"’I 16 \ I , I, -II, \ a 'l
15 1 ’\ \ \ Ill!’ /\- 15 ‘ ’i/l’lll’ ‘ I “
1. , 1 r . . /\1 11 - , ‘11,
>14 1‘ “ >14, ’ /’ l’ ‘ ‘
13", \\-\ ’\\-/\ \ 13 \///-/ ’ ‘ \ \ I,
’I"\“\1\"'” "///"11“’/
12 "“"‘//""‘ 12’////"I\“’//
I\\\\ , ///1\ / I
11 x- ' ., nix/1N H'H/
10111’ 1l\/1/1/ 1! ‘1111I‘l 10111111111 111111111111111 J
2 4 6 8 10 12 2 4 6 8 10 12
x x
20 minutes 25 minutes
reference vector (0.5mmlsec) — reference vector (0.5mmlsec) -
18 18
17llllll//""// 17'////I'\11\\\1
I”/////",‘// I///11111//111
16,-,,,,/,-11-,/ 15II111111/IIII
15,///////-\\,// 15.- . \ »_.—-~.11/’--
>~141vz/z/ II ::’/1 >.14
.-\\,— . . . ‘1.
13 13
12 12
11 11
11l111111111111111111111141l 1111 111111l11npl11L1lL111l
1° 2 4 6 8 1o 12 1° 2 4 6 8 1o 12
x x
30 minutes 35 minutes

Figure 3.2.3. Averaged velocity ﬁelds of mushy region from 20-35 minutes (1St FOV)

Even at an early stage in the experiment, a well-deﬁned convective flow has
already formed. However, the velocities were not solely in the upward direction, which
was originally thought, as depicted by the u component of the velocity (note: u
component of velocity is in the x direction), which has the same order of magnitude as
the v component of velocity. It is apparent from the measurements that the velocities of
the small ﬁngers are on the order of 1mm/s and the velocity components in both the x and

y directions are comparable. Note that for measurements on this order of magnitude, the

40

maximum displacement of the laser grid is about 4 pixels in the ﬁeld of view of 10.41 x
13.88 mm which translates to a velocity of 1.6 mm/s with a delay time between the
delayed and undelayed images to be 56.995ms. The maximum velocities are documented
in Table 1 on page 47.

By looking at Figure 3.2.2, at 8 minutes, it can be seen that well-deﬁned
convective ﬂow patterns have formed at this early stage in the experiment. At 16 minutes,
the ﬂow is generally in the downward direction with slight movement of the ﬂuid in the
upward direction at the far right of the ﬁeld of view while at 20 minutes, the ﬂow pattern
seems to return to the initial chaotic pattern that was exhibited at 4 minutes, albeit with a
higher magnitude of velocity. One feature that is of interest is shown at 35 minutes where
the velocity vectors at the very bottom of the ﬁeld of view seem to congregate to one
location. Salt-rich ﬂuid descends to ﬁll the void left by the ascending ﬁngers. This salt-
rich ﬂuid would approach the mush and due to the undercooling that the ﬂuid
experiences, precipitates ammonium chloride and ﬁnally ascends as water-rich ﬁngers.
This process continues as the mushy region grows in height.

In the second data set, the same image ratio was used but at a height of 21mm
from the base stainless steel plate. Figure 3.2.4 shows the location of the tagging region
for the 2nd FOV relative to the solidiﬁcation process at 8 minutes. The second ﬁeld of
view acts to document the velocities above the mushy zone aﬁer 40 minutes when the
ﬁrst ﬁeld of view is completely blocked and also at earlier times to observe the ﬂuid
movement at a farther distance from the mushy region. The following ﬁgures illustrate
the measured velocity vectors at different times into the solidiﬁcation process. All the

velocity vectors shown are averaged over 15 realizations in a time frame of 2 seconds.

41

13.88mm

 

Figure 3.2.4. Tagging region of the 2“d FOV at 8 minutes into the solidiﬁcation process

 

—.

reference vector (0.5mrnleec)

///y/v/|.\u.\l/
//Io|lo lllll
I I/p/v/p/////
\ \ p /////I'/
loll \ a . , I .\
l\\\\\\
. x x \ \\ x a \ a

...v/ //’IOI.IOIDII

 

 

8 minutes

reference vector (0.5mmlsec) ..

’1’...\.

. .

Illlltllli

\ l O -

\\\\\\\\

\\II;.'~

IO‘IIII\\

n.11I/

\\
\\

\ I

///ll.alf~
\I.,.III/’

‘\f|lI////

.

 

 

 

reference vector (0.5mmleec) _.

    

--\,\\",

28-.\\\-.
27\\\\\\

654321
222222

12

10

 

4 minutes
reference vector (0.5mmleec)

 

—.

f \ \
I I 1 . \ ‘ . a r .
l I Q . I . I I I I
\ \ \ 4 \ . . | \ I

\\\\\.l/,...

\\‘\\I|"‘l I ‘ \

I'll/ll If,“ 1

////////II.

I.

.\ ///////..H

I/Ia..r.\\

till\\\\\\
Isllll\\\ \
III/It . (\Io\

Ill/ll" ‘l‘\|\§

l

Eh

8765432
2222222ﬂ

>

 

 

 

16 minutes

Figure 3.2.5. Averaged velocity ﬁelds of mushy region from 4-16 minutes (2“d FOV)

12 minutes

42

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>

 

 

 

 

 

 

>

 

.12 2 12
It'll."‘\ \ s 11 0 O I I o, l I #1 l:l"\\ III'I'II'IL1
1 L
‘ III/'I"!|I'l|0lf|'1 ‘ \ \I I, i a I I I L P I’ll/IIIII'I1
) ) 1 1
m "1\\\\\I\‘\Lo m II.“\\\\\\\[‘OI m I \\\ \\ \ ‘lm
11 1
m //I'II'I:////H km. ///////I\Io1 :mu //////.\//1
we I l ’ ’/ I I I I o d8 S we \ \ \ \ I I l I I I '18 S we I I \ I I I S I I1_8
w \u\\\\s\lI/I # mm \1\\\lI/I.II m N \\ \ \ \\ \ .\\
W \\. \.1.\\ 16 DO \\\llollol4// 16 mm \I/\\\1\\\\16
I v 01 L
m IIIllIcI/I/g. Mm xxa. .t..111d m m /////,. 1,1.
I. 1 5 1 5 l ‘ I a I l“\ 1
m /r///// I // 1 2 m z / III/It’ll 1 3 m I 1
h \ \ . I e I \ \ \ \ '4 h \ o \ 0 I I ll’llnu4 b III! ‘6' I I ‘II I1.4
m \ \ \ \ \ \ x \ a r m \ \ \ \ \ I \ I I I” m \‘III \\‘1\\\ H
\ \ \ \ \ II I, 1 's\ I /, OI I I 1 // I \. II I c 12
I1 12 .l
I I . \ \\\\.lv ul'l'12 lIII'”\ I’ll, 1 I I'I'IIIII'\II'I'II/1
El, EE’tLI h-l- .— b - h b b n-Pb b n - n > b th rbkb b bPL—P - b n — h ~ P Fl
8 7 6 5 3 2 1 O 8 7 6 5 4 3 2 1 0 8 7 6 5 4 3 2 1 0
2 2 2 2 «M.— 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
> > >
2 12 I2
\ \ \ \\ a 1 l I I I1_1 llllllll I I 11 I I / I f I s s \ I! 11
— \\\\I\'III\\\\H — s s . I I \ \ ‘ \ .11 ‘ I I |. I \ \ \ \ \U
) \II 1 ) 1
m I'IfIIII-I'IIO“\QI.§140| m \\\\\\|\ \\ \ \\ lanlv m I a . \ If I! \Q \\ IO 1m
m I I I I 1’ 1’ I ‘ ‘ 1 m l'l".\\‘I'I'//l'1 m / I I I / a I I I,”
5 1 5 1
e I. e I. 0 I18
0 Illl/Ials\\\ 18 0 / z . s \ . . I 18 0 II/ I/ / I . I / 1
( S ( S H 1
m \\ \\\\.|I.Il‘ H m m III” """ \ 1 w o I I I I ll/I‘\ 1
w II.“II'\II\I \L um I\\I,II/," 16 mm / / I I / / LI\\ \ 46
6 e In .1 I.
1 m \\\r\\ x . x I 1 m / I . . . / I |\ \ 1
n \\ x s . . . 1 I /1 m n IIIIIOI'IV\\\‘IOI1 mm \\‘\‘\QIIO\\1II
m 0| I \ c v . . . o '14 m ll/I'I'll'“.l'l'/l4 h nI'Il'II'I'I'I'I'I'I/lh4
m \\\\!I§I0II I m IIIIIIIIIIIIII1 m \ ‘I////’// 1
\‘I“."////”h2 I..II'II.“‘.\"‘II.\I11.2 I ’ I I o O O \ \ I...2
’ [III/[I'll I“\\\\‘ ‘‘‘‘ 4 ‘ I], I ‘ \ \ \ \ 1
8 7 6 5 4 3 2 4| 0 8 7 6 5 4 3 2 1 o 8 7 6 5 4 3 2 1 0
2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

 

 

45 minutes

Figure 3.2.6. Averaged velocity ﬁelds of mushy region from 20-45 minutes(2nd FOV)

40 minutes

43

 

 

 

 

 

 

 

 

 

reference vector (0.5mmlsec) — reference vector (0.5mmlsec) _.
28///"’//\:"Il 28ll""’ll' ”
27/ll"ll\\’[\ 27:1' 'Iia~"1/
26/""’/l‘\"/‘ 26"""Ilt~‘axr
\\ 1

’,1\,,/‘\ //I \l' 'I/[\‘ 1”
>25t til/1\‘\//’ >25\I .I/II‘ 0",
24\ \;\//\‘\//’ 24!! I/[l‘ ’lll
23\"IH//""/w 23/"'I/1/' "\l
22\"/lI//"‘/// zzll’llllll “\\
tl/lll/l‘lll llllllllll'l\\
21:!I/J/I/“:I/ 21’11[\I'\\;1111
201L11111111111111111L1111111 20,11##111111PL111P1L1PP111141
4 6 81012 2 4 6 81012

x x

50 minutes 60 minutes

Figure 3.2.7. Averaged velocity ﬁelds of mushy region from 50-60 minutes(2Ind FOV)

From Figure 3.2.5, it can be observed that at 8 minutes into the solidiﬁcation, an
obvious downﬂow is occurring in the middle of the ﬁeld of view coupled by a slight rise
of the ﬂuid on the left side. This is typical of the ﬂuid ﬂow phenomena at this stage of the
experiment where lighter ﬂuid rise as water-rich ﬁngers and salt-rich ﬂuid descends to
ﬁll the void present. All the instantaneous realizations of the velocity measurements
made in the FOV showed a well-deﬁned convective ﬂow pattern with velocities on the
order of 1mm/sec, similar to velocity measurements made in the 1St FOV. At around 65
minutes into the solidiﬁcation process, the mushy region would have covered part of the
ﬁeld of view. A third ﬁeld of view was taken to measure the ﬂuid velocity at a time later
than 65 minutes. The 3rd FOV is located at 31.5 mm from the base stainless steel plate
while the maximum height of the mushy zone is only ~27mm. Thus the laser-tagging
pattern would not be obstructed by the mushy region in this ﬁeld of view. Figure 3.2.8
shows the relative location of the tagging region with respect to the solidiﬁcation process

at 8 minutes.

44

 

Figure 3.2.8. Tagging region of the l8t F 0V at 8 minutes into the solidiﬁcation process

Figures 3.2.9, 3.2.10 and 3.2.11 illustrate the velocity ﬁeld measurements made in

 

     

 

 

 

 

 

 

this ﬁeld of view.
reference vector (0.5mmluc) _. reference voctor(0.5mmluc) -
39.;.\---,-,\-.. 39\\\\\\/I\\\\\\
38 ‘ r g I a - — ‘ I v ‘ . I 38 , K \ A.— //\ \ ‘\ ' '
37 I \ o a o ‘ ‘ - o - \ ‘ ’ 37 ‘ ' I /\’ \ ’\ ‘ I \ I I
36 I I / t / - | \ \ \ \ c ' | 36 I I
>35"””'"""‘ >35 H
c I , . \ - I - a a , - ‘ ‘
34 , / \ , ~ ~ - I \ - - . . \ 34 ‘ \
33 r r ‘ I‘- i l - \ sssss 33 - \ \
o - s , . - . - ‘ ‘ \ I I \ \ \
32 ’ 32 \
' ' ' ". ‘ ‘ .’ ’ '. ‘ ‘ ' ‘ ‘ \ ‘
3‘ 4 6 8 1o 12 3‘ 2 4 6 8 1o 12
x x
4 minutes 12 minutes

Figure 3.2.9. Averaged velocity ﬁelds of mushy region at 4 and 12 minutes (3“i FOV)

45

 

reference vector (0.5mmlsec) —

l//’

o\\\.\\ 1

I

I

t
'

t

\

s

I
‘

t

\

l

I I I t ' I I \l

\\\\I

\\\\l//

\ \

\

II’1

I L
l

\ u \ \ \1

‘m

If I'll/III Ill
‘ .I .I ‘ \ \ \ c I ’1

I’ll'llll\‘\\

\\\\r\

9876
3

1O

 

5%
3333 333
>

25 minutes

 

 

 

 

reference vector (0.5mmlsec) ~

I ' ‘ I ' \l

\ I L

//////////H
\\\\\\\\\ Tr

' l \

M

10

6

I’ll/I'll...
1

\ c ‘ ’ t I '14

\
\ \ \ \ —

I 44 J 1

r

t //

 

31

 

 

16 minutes

 

 

 

 

 

 

 

 

70 minutes

 

 

 

 

2 12
\\\\\\\\\\IU1 ‘ ///r/.r/I’/i1
r n
P II'\II'|..’I"'II'I.“1 ) //I'/ IIIII A
m ‘l‘I‘l‘l’l‘l‘lﬁlﬂuw m //'//rlfl"\ . .lm
W lvleo//////J W /////// x I J
.m .,I\1\\\,‘n .m //,.\\\\\-H
m OIOI\\\\‘.\\\\I._8%m.\ ///'\\\\\\Il_8
l ..l
m \\\\\\\\\51 mm /////.\\1-
w \\\\\1I/,L6.mw ///////,\-h6
m /./\/\\...,\ A <43 m ////////x l”
/” L 1
m / xxxx. m //////////
am / /!\\\\\\n4 m //////////1.4
x///\\\\\\ /////////K
\\- III‘ 4 ..
\\ \\\~L2 ~'////////..2
\ \\ ...,/m \\\\. .////
—LPP»-»-phhtb-bbhp-puhnhpbprhp-hh—phbb -h~bb-rhh’b—pphb~»b»Pl—rbhhPhhhhphbnru-hpbp
wmnmwwuﬂwmuw ”mm;wwumwnm
> %
E -m
\ \olvlv\‘\\lfl.vlfl1 loll/’I/l/I-
m \\\\\llll\.1 w . ,:\\\\\
m \\\\\\ll,@ m ll\\\\\\\m
‘00. \\\\\\\\ \\\18 S M \ \s n///// ’18
( i e ( 1
.l \\\\ 1 t r I //////// I.
o . \ \ x \ \l u o 1
d //I\ . .m d ’ . x , ///// .
m ///t\\\a\,. 5m \\\\\//,/T
n \“\\\////H 3n \\\\\\\./,
m s \I I ’ ////..H4 m \t\\\‘\\\lva ILA.
m I], m \‘r‘l'nlfll'lll

’I’\\\~.

 

 

 

L
I ’ I / / .I

,///h
//////..\-.
EE
9876 4
3333 3

5
3
P

 

321
333

55 minutes

 

 

Figure 3.2.10 Averaged velocity ﬁelds of mushy region from 16-70 minutes(3'dFOV)

46

 

 

 

 

 

 

 

 

 

 

reference vector (0.5mmlsec) - reference vector (0.5mmlsec) —.
39 39
I \ \ \ \ \\\\\\\\ \ t ’ ’ 5 I I
38 I \ \ \\\—0~.\—" 38 ~ 0 r I I ' ‘ I ’ . g z
\\

37 .- - -. \ \ —.~.\-—o// \\ 37 0 \ \ l ‘ I ‘~ \ ‘ \ - a
36 .- .. \ ‘ \ _ /_.\ I l/ / \\ 36 ‘ ‘ ' . ’ ’ I s \ \ \ \ \
x x x - .. I / _. / I , , -

>35 \ \ \ x x I l / :l l ’ ‘ i ’35 .‘ j ’l i I] ‘\ : 1 \\ I I i ' I
34 .. s \ \ \ l l /‘ . \ , ‘1 x 34 ' I I \ ' \ \ \ \ \ \ \ \ \
33H‘::§',',/’\'11 3311/;;\\\\\\\\\
32’” /'\‘/1 32/1/1/r\\\\\\\\

.’. i; .‘. .1: 1: /. 1.1.24.1. .(./../../../..L.1.1.\..\.\.\.\A.
3‘ 2 4 6 10 12 31 2 4 6 8 10 12
X X
80 minutes 100 minutes

Figure 3.2.11. Averaged velocity ﬁelds of mushy region at 80, 100 minutes (3rd FOV)

Tables 1, 2 and 3 condense the respective maximum or minimum U and V component
velocities present in the three ﬁelds of view. In all three ﬁelds of view, the velocity ﬁeld
is seen to have a velocity magnitude in the U and V component of order 1mm/sec. In the
ﬁrst ﬁeld of View, the maximum velocity magnitude in the V component generally occurs

in the downward direction.

 

 

 

 

 

 

 

 

 

 

Time U - Component V - Component
(minutes) (mm/sec) (mm/sec)
8 -O.76 -O.43
12 0.44 -O.43
16 0.46 -O.60
20 -O.42 -O.48
25 -O.32 -0.54
30 -O.24 -0.47
35 -O.38 -O.58
40 -0.24 -O.39

 

 

 

 

Table 1. Instantaneous maximum/minimum velocity measurements in the 1St FOV near
the mushy zone.

47

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Time U - Component V - Component
(minutes) (mm/sec) (mm/sec)

8 -0.39 0.61
12 -0.43 0.73
16 -0.35 0.61
20 -0.35 0.61
25 -0.42 0.67
30 -0.31 -0.70
35 -0.47 -0.70
40 -0.34 -0.52
45 -0.32 -0.53
50 -0.36 -0.44
55 -0.36 —0.70
60 -0.30 053

 

Table 2. Instantaneous maximum/minimum velocity measurements in the 2nd FOV near
the mushy zone.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Time U - Component V - Component
(minutes) (mm/sec) (mm/sec)
12 0.60 0.79
16 0.52 0.58
20 1.03 -0.67
25 -0.66 0.64
30 0.49 -0.82
35 0.33 -0.58
40 0.62 -0.63
45 0.32 -0.66
50 0.65 -0.86
55 0.43 -0.59
60 0.64 -0.53
70 -0.47 -0.60
80 0.59 -0.74
90 0.47 -0.32
100 0.34 -0.44
110 0.41 -0.44
120 -0.43 -0.45

 

 

 

Table 3. Instantaneous maximum/minimum velocity measurements in the 3rd FOV near
the mushy zone.

48

3.3 MTV Measurements above Chimney

In the regions above the chimney, previous research obtained a peak velocity in
the vicinity of 8mm/s. Figures 3.3.1 and 3.3.2 illustrate the velocity proﬁle and the

delayed image using single line tagging just above the chimney. (Wirtz et. al., 1998)

 

10

O
l

imam»

 

U

 

 

 

b

Distance (mm)

Figure 3.3. 1. Velocity proﬁle of plume (Wirtz et. al., 1998)

 

Figure 3.3.2. Delayed image using single line tagging (Wirtz et. al., 1998)
The measurements of the velocity ﬁeld of the plume in this study are divided into two
parts. In regions near the chimney, two component measurements of the velocity were

not performed due to the way the experimental setup was conﬁgured. As shown in Figure

49

2.1.1, the nearest the two-component measurement technique can get to the mushy region
is 7mm. In order for velocity measurements at distances lower than this, single laser lines
were passed through the ﬁeld of views similar to the ones used for measurements just
above the mushy region. It should be clariﬁed that the chimney itself is part of the mushy
region but measurements of the mushy region apply only to locations where plumes do
not exist while all plume measurements relate to the location above the chimney.

In the single component velocity measurements, it is assumed that the main
velocity component is in the upward (V) component. This however is not entirely true but
the U component is assumed to be negligibly small thus allowing us to validate the
assumption of single component velocity very near the chimney. Figure 3.3.3 illustrates

the region where single line tagging was performed.

   
 
 

Single line

tagging typical
ﬁeld of view

  
  

Laser + Optics Setup

.\

Figure 3.3.3. Typical location of multiple line tagging of single component measurements

Laser Lines

50

The single velocity component measurements were performed in a ﬁeld of view with the
lowest laser line being 21.0mm away from the base stainless steel plate. Typically, 12

laser lines are used in a single ﬁeld of view. The velocity measurements are illustrated in

 

 

 

 

 

 

 

the following ﬁgures.
7
6.
Sr-
4+
is)? .
e 3'
2 . Q g .
g Oht‘ ff; ’ ...; v-57" ..
~ at»; we '49,? e- .
'2' " w———o Line1(~21rrmfnombaseplate,~7nrnfmmsolidiﬁcationfront)
’ ,___-, Line3(~22.6rrmfmmbaseplate,~8.6rrmfmmsolidiﬁcation front) 1
.3...- a ------ a Ljne5(~24.2nrnfrombaseplate.10.2rrmfromsolidiﬁmtionfront)

 

 

 

Length (mm)

Figure 3.3.4. Single component velocity measurement of plume at 28 minutes

In Figure 3.3.4, the velocity measurements showed that at three locations above the
chimney, the plume exhibited a velocity component of ~ 7mm/s. This is in agreement
with the results obtained by Wirtz et. a1 (1998) which exhibited a velocity component of
~8mm/s. Magirl and Incropera (1993), reported velocity measurements of ~10mm/s due

to their much lower cooling temperature of -30°C which resulted in a higher rate of

51

cooling and growth of the mushy region. This contributed to the need for a higher ﬂow

for rejection of water-rich ﬂuid from within the mushy region.

 

7

‘ : : ‘
F . i . . . 1 . 4
' . . , j 1

 

 

Velocity (mm/sec)

 

 

'2 ..- _ -+ Line 11 (~28.97rrm from base plate, ~13.97rrm from solidiﬁmtion front) E
. ----- - Line 7 (~25.71n'm from base plate, ~10.71n'm from solidiﬁwtion front) ‘
-3 F- — - 1 Line 3 ( ~22.6rrm from base plate, ~7.6nrn from solidiﬁcation front)

Length (rrm')

 

 

 

 

Figure 3.3.5. Single component velocity measurement of plume at 30 minutes

In Figure 3.3.5, the velocity measurements are noticeably lower than in Figure 3.3.4. The
maximum velocity in Figure 3.3.5 is less than 5mm/sec. However, one noticeable
phenomenon in both Figures 3.3.4 and 3.3.5 is the lateral displacement of the tagging
lines at different vertical locations. This indicates that the plume has a lateral component
of velocity, which was not taken into account during the correlation of the velocity. The
correlation technique used assumes a pure v component of velocity. Thus there is an
inherent error in the correlation technique used in the single line tagging which would

tend to compute a lower velocity. However it provides a good estimate of the velocity

52

component. Another feature of interest in the velocity measurements as shown in Figure
3.3.5 is the double peak present on line 3 and 11. This feature has been repeatedly
observed especially in regions of lower speeds. As the plume traverses the plane of
illumination, the velocities can be seen to increase from line 3 to line 7 and decrease
again on line 11. Due to the planar nature of the tagging method, all velocity
measurements can only be made over a plane. Whenever the plume moves out of plane,
the laser lines would be tagging the edge of the plume and not inside it, which may result
in lower velocity measurements. Figures 3.3.6, 3.37 and 3.38 illustrate measurements
made at different times within the ﬁeld of view that was used in the one-component

velocity measurement with line tagging. Uncertainty calculations are under progress.

 

 

 

 

 

 

 

G .
a) ...... .
£’ .
E 3
g ................. j
E :
o i
2 l
o .- _ . 3 , ,
> w ,j . ﬁi
'2T F-_-. Une7(~25.7rrmfrornbaseplate,~8.2rrmfromsoli<iﬁcationfront) ‘
; a ------ a Line5 ~24.15rrmfromthebaseplate,~6.7n'mfromsolidﬁmﬁon front) 1
-3 --------- o-———o Line3 ~22.62rrmfrombaseplate,~5.2rrmfmmsolidiﬁmtimfront) ------ 1

 

 

 

 

Length (nrn)

Figure 3.3.6. Velocity proﬁle of plume at 35 minutes into the solidiﬁcation process

53

Velocity (mm/sec)

 

 

 

'ZW' o——-—-o Line7(~25.6rrmfromthebaseplate,~6.6nmfromthesolitiﬁcationfront)
..-—-7 Line5(~24.1rrmfromthebaseplate.~5.1nmfromthesolidiﬁmtionfront)
.3-..- a —————— a Line3(~225rrmfromthebaseplate.~3.5mnfromthesolidiﬁwtionfront)

 

 

 

 

 

 

Length (mm)

Figure 3.3.7. Velocity proﬁle of plume at 40 minutes into the solidiﬁcation process
7 .

 

Velocity (mm/sec)

 

 

 

 

 

 

 

 

'2 ’ - ..... . Line 7 ( ~25.6rrm from the base plate. ~5.6rrm from the solidiﬁcation front)

k , - — -- Line 5 (~24.1rrm from the base plate. ~4.1rrm from the solidiﬁcation front)

-3 v— - — - v Line 3 ( ~22.5rrm from the base plate, ~2.5rmi from the solidiﬁcation front)
Length (rrrn)

Figure 3.3.8. Velocity proﬁle at 45 minutes into the solidiﬁcation process

54

At a larger distance away from the chimney, two-component velocity
measurements were implemented. The tagging pattern for measurements of the plume is
rotated by ~45 degrees compared to the tagging pattern of the mushy region. This is due
to problems encountered with the index of refraction of the plume when vertical laser
lines intersect the edge of the plumes as shown in Figure 2.2.5. From the two—component
velocity measurements obtained above the chimney, it can be seen that the ﬂow is
generally moving in the upwards direction with velocities of order 6.5mm/s. The U-
component velocities are seen to be typically less than 2mm/s. Figure 3.3.9 illustrates the
location of the plume with respect to the ﬁelds of view at t0=79 minutes into the

solidiﬁcation process. The bottom of the ﬁeld of view is 35mm from the base plate.

 

13.88mm
Figure 3.3.9. Visualization of tagging region at 79 minutes into the solidiﬁcation process

In Figure 3.3.9, in the tagging region of 10.41mm x 13.88mm, by observing a sequence
of images, ﬂuid can be seen descending on the upper right comer from shedding that
occurred at a distance above the ﬁeld of view while slight shedding can be seen on the
left edge of the plume. This phenomenon is captured in the velocity measurements and

shown in Figure 3.3.10.

55

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m
-...t...,i\_,//, w /\\\\_...~\.//// m \\\\\~_..\\,////
_ iiiii 2 t . ,,/I _ il\\\_._....//Il _ \iiir\\\_/..,/li\\
\l 8
\va lllll \i ,z// m m lfl\\\\\~..i///\ m m \\\\\ \_//I/\I\\\ 7
m llirl .,/I/ lllll \\\\,/lI/// I rrrrr \\_ltlllrilili
m l- t w s m
m [rill/I‘ll. I'll "m H” m ///l.ir\\r../lirl//l. m 8 m iiil/I1x./i.litl|illri ..m
U 6 H. 9 U /
“ l/rillrllllltc\\r\r|1l 2 a ’l’rllll..\ilrl\\l\rl 7 a III/Ii_\i\l.l\\b\.ll\,
. I//
d ///l.l///,.\\\tili-\ M 0 d ///ll/_\\\\t\t\r\‘r| u 0 m ////..\\o\r\\\t\t\ai N
x //
W I/lll/’\\\R\b\h\\r\ xv W ////.\\\\b\t\r“\\lii M m //.\\\\t\~\u\\\\t\r/ J
m . \ . .\\~\t\t\~\\\\\\\ 2 m .\\\.\\\\t\\\\\\ .. 2 = n \\\\\\\\\\\ . . . i n
m illv __ n .iilii... m :iiil\\\.\.
h §\\\\\ \ z t h §§\~\\\\ \ \ t h L\L\\\\ \
mg\\ \ . I /‘\\\4\1 n §\\ \ \ z / //|r|rilo\9\t m Oilbir\\ \ _ / /I|r/l|.t|t.lr\elb 0
{f/I\\ . llirilllt.1.\1\\m \‘il. .l/lll/lr\\l\r\| w rii\. ill’tl-llili\rlrilr\t 7
,Irii.,.\t\rlll\ \tii...llllill\\ i.._.-lii.llll\lll
Ill/iii. \ ///// w I\ir\\i .\ llllll w ii\\\,\\\lIi/// w
. . . _ . . . . . . ...: . . . . . . . . . . . . . . . . . . . . . . . .
1r 5
wuuauuwuwwmuu wwuwauwmmwwuy mauuu4wwwvwsu
> a >
o m
;\\.r//, w /\\~\._.,\\~///. 8 \\\\\~_.\\\4///l
_ lllllll . i\ i/’/ P, l\\\\e_...t.//// \_I \\\t\\‘\,.._1//I\ 8
NW lllll \|, i\l n m \\\\\\\\\\t_.llll m m \\\\\\\,IIi//l\\ 7
m lr\ilil\i\\l..‘e.i\/ W lllll \\\\_.rll\i W Iilli\\\\~////I\\
m Iiii/i/li\\li-. ..il w m lllllll \\.,i\\ll K. «mm». m llI/l\\\‘I//I\\\ w
1 1 1
( 111., ill 0 ( llI/iiiii.i\\\ili 3 ( liiiz...i\llliil/
W iilo//// .\ a 5 W //
a llu//J//lli.\\\r\~\ H M a ////Ill,~\\r\r\n\ll1| H 0 m l/Iili.\\~\\r\r\u\¢i N
X I// X
M ///////,\\\\b\\liiv\ 1: W /////,‘\\\u\s\s\l\t\rl M m ///,\\\\\t\\\\b\rli
\.\
n I.._.\\.\.\t\.\.\r\t\\\ n t m ..\\\\.\t\t\i\\.\\\\. u __ M .. \.\.\.\\\\\\\ \r\tn
7T7, // \//|r\r\ lilit
m 1§\~\\\LI m3§t\.\\x I am rirld\r\.\\\\2_/
h \ lllliiii rillliix _ ,iI/Iliiiiiil
m§\\\\,ll¢ it m§\\czl 0 m 0
lliilrl\\\\ . ,/irlrl1lril m iii/[It‘I/ll/liliilll 7 ril’l , .Iliiillrlil\\1\t\ 7
‘I///\\\..i///ll \\\\\\‘il\.lllriilir IIIIII l\\\l/l.il\ii
till. it M i\i.\i//,\\i///i w airi\\,_\\\I//II &
- _ _ — - _ — _ - _ _ _ - _ . _ . _ _ - . _ _ _ . _ . _ _ — _ - . - — -
wﬁuanuwwwnuwu mwuauuwwwmwuu wauuuuwwuuwuu
> >

 

 

 

 

 

 

 

t = to + 1.333s
79 min ofplume from to to t0 + 1.33s

56

t = t0 + 1.067s
Figure 3.3.10. Velocity vectors measured at t0

From Figure 3.3.10, at to = 0, the maximum velocity within the ﬁeld of view occurs at the
very top of the measurement grid. The velocity vector translates to a velocity of ~4mm/s
in the upward component. However, this may not be the maximum velocity within the
plume itself as the laser tagging plane may not be entirely coincident with the plume. As
we observe the sequence of images from t = to to t = to + 1.3335, the plume is slightly
swinging to the right of the ﬁeld of view as it rejects ﬂuid in the upward direction. In the
process, ﬂuid in the upper right comer descends which may be caused by shedding at a
higher location as shown in Figure 3.3.10 and some gets entrained within the plume.
Some downward movement of the ﬂuid can also be seen on the lower left corner of the
ﬁelds of view as some shedding of the ﬂuid occurs as the plume tries to traverse a bend.

The velocities measured in this region are typically less than 1mm/s.

 

 

 

 

 

 

 

 

46 reference vector (1 mmlsec) - 46 reference vector (1 mmlsec) —-
45 45
4‘ ’ 'I i ‘ I l l i l i i I ‘“ ’ ' : i i ‘ I I i l l i i i
l . l I \
43 I ' \ / I l I I l l i 43 ’ ' ‘ \ \ \ I l l l l i l l
’ / 1 I / / l i . ‘ \ \ I I / l i I l l
42 ’ ‘ ; x I I / l l l 42 ’ ' \ \ / / 1 l i I
' \ \ \ I I I 1 g r s ’ \ \ \ ' I I I | \ x \
41 ' \ \ ’ I I t 1 ' ‘ 41 ‘ ' ‘ \ ‘ ' I I \ \ \ ~
‘ \ ' I I I r x \ r i I ' \ \ ‘ \ \ \ \ "
~ 40 , x \ \ \ , l | , . \ b 40 ‘ | ‘ \ \ \ \ . I a _. -
s \ \ \ - \ a - _ \ \ \ a - I I ’
. , t ‘ I l -
39 ‘ ‘ , \ \ \ \ . . I l 39 ‘ I : I , \ \ \ 1 I I I r
38 , | I I \ \ \ \ \ z 38 ‘ I / I / \ \ 1 \ i I z -
, I I I ,, \ \ \ \ I ‘ I l ‘ I \ \ \ \ i l I .. .
37 , , I I , \ \ \ \ . 37 , I ‘ x I \ \ \ \ 1 1 . . .
| \ \ \ \ \ e \ 1 1 I \ \ i \
36 \ i l I \ 36 ‘ i \ i i \ ‘
| \ i i i x \ \ \ f | \ \ \ . l 1 1 I I 1
35 , l I I l . \ \ \ \ i 35 | ‘ \ \ \ l i i I r r I t
4 1 l L l l l l I l J L l 1 l L 1 l l l l L J_1 I I l l I l l l l l l l l l 1441 L 1 J l l l L l I l l l 1 L J 1 J L
34 68 7O 72 74 76 78 80 34 68 70 72 74 76 78 80
t=to+0 t=to+0.5325

Figure 3.3.11. Velocity measurements at t0 = 75 minutes into the solidiﬁcation process

57

 

 

 

 

 

 

 

 

 

 

46 reference vector (1 mmlsec) — 46 reference vector (1 mmlsec) —
45 45
44 - . \ § § \ I I I I I I I 44 _ , . \ h \ , I . I I , I i
.. . \ \ \ I I I I I I I I .. ~ . \ \ x t I I I
43 - . . \ \ - , 43 . \ \ \ ..
\ \ I I I | \ \ \ ' i \ \
42 \ x a ‘ \ \ \ I / I | i \ 42 \ \ I l \ \ \ x . _ I \ \
\ \ i ‘ \ ‘ ' \ \ \ \ \ | l I r \ \ \ ‘ ‘ \ ~
41 \ l l , \ \ \ \ \ , \ \ \ - 41 l I l I I \ \ \ \ t - ‘ ‘ _
I I I I - \ \ \ \ . - \ \ - I I I I I t \ \ \ \ i - I -
'~ 40 \ \ \ \ i I , _. - ~ 40 t \ \ \ t \ t I I
' / / I ’ I I / l I
I \ \ \ l l I I , I I \ \ \ l \ \ l r
39 I I I I 39 I I I
I I I I I t \ \ l \ i I I I I I I \ \ i \ \ \ \ .
33 I , I I I - \ \ \ i \ i , 33 , , I I I l i \ \ \ t .
I I I I I \ \ \ \ l i , - , , I I I i l t r \ i i .
37 I , I I I \ I \ i I , , . 37 \ . I I I . . II \ \ l l l i
36 I I I l l l I I i r I i i 36 I I i I l \ I I ‘l i I I I I
I l I I I I l I I I l l l \ l I I \ I I I I
35 I I l I l I i l l i I i I \ 35 I I l I l \ I / I i , i t \
u l l l 1 I A L l 14 J l I l l l l I l l l L L L4 1 L L 1 l l .L l L M , 1 J 144 l 1 l l l l l l I 1 L4 14 l l l I A L 1 L: J 1 L L
68 70 72 74 76 78 80 68 70 72 74 76 78 80
J 4.
t=to+1.064s t=to+l.596

Figure 3.3.11. Velocity measurements at t0 = 75 minutes (continued)

As the movement of the plume becomes helical, tagging the plume becomes
difﬁcult as the plume moves in and out of the tagging plane. In the image sequence taken
at a speciﬁc time, it is often observed that the plume is tagged at an earlier time tick but
slowly moves out of the tagging plane later in time. This makes it more difﬁcult to
determine the velocity of the plume itself. In Figure 3.3.11, an example of this difﬁculty
is illustrated in the velocity measurements.

This difﬁculty in tagging the plume is particularly apparent as the solidiﬁcation
process continues and the plume becomes more three dimensional in nature. The
measurement technique employed in this study can only measure the in—plane velocity
ﬁelds. Any out of plane velocity ﬁeld can only be measured with a stereoscopic
implementation of MTV. The following ﬁgures are velocity measurements at different
times within the same ﬁeld of view. All the ﬁgures are instantaneous realizations and are
not representative of the 15 realizations taken at each time tick. As the ﬂow ﬁeld changes
dramatically within the time frame of 2 seconds, an average velocity vector would not

yield any useful information.

58

 

reference vector (1 mmleec) —

ll\\\~_:,. .I‘\\

A|\\\\\Il\(:1|||

7B
78

p
..... \.\\\1Il/II \\II-I. ./w_.\. rIl.\\\\.',..://
-1..V./\\\\\Izz, \\\-......____.\ '11:..__,,,,‘\,_
:...~\\\|\\\\\tl m \\\::;:r. ...,l\.ﬁlc r/Iu. _,//,.._
I . ,.\\\\\\\l\o\1|.ll Illllololtlll I I |\ l/Illi | \ . z I III!)

\Ilultlvlrl/ _;I.Il.

I,.\\\\\\

%III\ |.v|
.\\\\\\\\\IlI/// \\\

74
X
75min
I
/
/
7
X
85min

\\\\\\\\~.(:-\\\ \I/(III/IIIJ}!

72

reference vector (1 mmleec) -
/
/
4
reference vector (1 mmlsec) —
\
\
\

AAAlJllAlJALAlAALAlLAAAl‘AJAJ_J_I

 

 

78

95min

 

 

 

 

reference vector (1 mmlsec) —

\\\ \ , III/Illal..|\|\ l\\|.\\. . . . ‘_/ (x u , , I’lltl‘rllrl

\\.:IIII/////I..Ia| Il\\\\::|..~\... Ill\\\\ 11111,!
0 0

\.\|Il///Il/llll 7 4;\\|ntun,. \xxe 7 \\\\\\ ~\\;,,_./

.~\\|I//// \\\\\\ \II,,-.,,,,-\\I\ \Itar._ __ \‘_,/

__\\\\lI/// ||||| 1w \\\ 11111 ..t\\\|\ W .\ :1__;III\\‘.;

. . . . . . . . . . 1 _ . . . . . . . . . _. . . . . . . . . _ _ T.

1 109 5

uuunwmuumuu uuuun4wmmmmﬁu «wuuu443mmw3

h h >
m M ..t .\t .\\

.......\l\\\\lll

» I\\Il\\ \\‘\\\I: ....a/. \
B a
.nuvu‘Il/z/t 7 IIIIII 1;, ,AII/A 7 _. --,___\
,,....\1u\\|...\ l‘ltlI/ul ‘|\.I|\ ._ _. .. ‘\\\
////,, ,,.\\ . A ..m tll1|lll74///,.~\|Illn ‘ /.-., \t

/////////.. 1.
_////e/4./‘.///,.|\~ .
\\.////4//4///l\\-
\\\_///4A%/llntlu/

7
80min

reference vector (1 mmlcec) —-
/
/
I
\
|
A
reference vector (1 mmleec) -—
, .

\\\~.,//o{' n I|\\\\\\. ..l‘otl. n _ ulI///’/IIIIII
\\\\.,,///./¢JJUA.H|/ ((((( \\1....e\\\ .. .\ .,/ 111111
\\\\_,////////// m |\ 1111111 . \\\\\ m »~\u,~. .,I..,...
_\\._4,,//////// \|\\ .,,.\\\\I\ \1. e c. _\..
,,‘~,,.,,////l// w -1..\-,._\\\|n'- u ix:

. _ _ ... _ _ . . . . . . ___A:_:

 

 

4e .—
45 -
u _
43
‘2
41
>- 40

 

59

72 74 76 78
90min

70

68

 

a.1.l....l...1111..l1.11l....l....l..
74
X
.
71mm

 

___ .7.
”ﬂux nuwuw

_
9
3

‘5.
45.
u.
43
42
41
>40-
39-
38
37
36-
35-
34
45-
45-
“—
43-
42-
41
>40-

 

 

 

 

 

Figure 3.3.12. Instantaneous velocity measurements of plume from 71 to 95 minutes.

 

a.1.l11..l..1.l..1.]....l.|.xl..1.l..

.n

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

45 - reference vector (1 mmlsec) — 45 - reference vector (1 mmlsec) —
45- 45- ‘
44- I | ‘ s - _. \ \ ‘ _ \ \ \ 44- _ . \ \ f I / x x I _.
‘3_ I , \ , - .. - \ \ I , \ \ \ ‘3 _ \ \ \ \ . / I I / I ‘
._ , , \ _ .. a \ I ‘ _ \ \ \ \ \ \ ‘ ‘ l I I I \ _ I
42—” ‘ ' \ ’ ’ 7 \ l I ' 7 ‘ ‘ 42- x \ \ I \ \ ’ ‘ ' “ - /
__ \ I . I z _ \ l l I _ __ .— \ \ \ ‘ \ \ , I \ x x _ _ 1
41... x \ Q I r / , \ l I ‘- _ _ 41- I \ \ \ \ \ - I \ ‘ ._ _ _ ,
I \ I l . _ a - ._ _ _
>. _' I I \ \ ' ~ >, _I I I \ \ x ‘ '
407 ' \ \ \ ‘ I / | \ r r ‘ 40 I I I \ l \ \ ' ' " ‘ ' ‘
397. ‘ ’ ‘ \ \ I I I I I \ ' 39. I ( | I I \ \ I 1 - ‘ ‘

, ~ ~ ' \ ‘ I I I l \ ‘ ‘ l I I I I \ I | I r ‘ ‘ ‘ ‘
38— — — . \ \ , . , . \ ~ aa— , I . . . . , . .- \ ~ - —

’ I I \ \ \ _

, I . . . I ~ - — , I I I . , . (, , —
37: ‘ ' | ' Q I \ I . ‘ 37" — , l . ‘ \ l’ " I / /
36;- ' ‘ ' i I \ | I I I 36-- ~ - l ’ I I \ I I z '

, ' ' ‘ I I I I I l ‘ ‘ ‘ ‘ I I I \ I I / -
35... I I I \ , , . | \ \ 35:. I \ I x I l _ I ' I , - r
14 1L4; I 111- .1111111 1 , I. “all II..I.,,.II. II. .I 11

68 70 72 74 76 7B 80 w 70 72 74 76 78 80
x x
100mm 105mm
45 - reference vector (1 mmlsec) — 45 - reference vector (1 mmlsec) —
45- 45-
44- I I I l i \ r \ \ ‘ \ l 44. ._ _ . t \ \ x , ‘ I I \ ‘ ‘
/ I l . l l \ , \ | ’ _ _ . , \ \ \ , I I \ \ a ‘
43' . , , l I t \ 43' - \\ . _. ..
— . ‘ \ \ ‘— - , \\ . I - ~
42- ‘ ‘ ~ ‘ ‘ \ ‘ ‘ I \ \ \ 42- - \ \ I . \\ { I x - ~ ~ \
a _ a I . \ - . \ a —. — \ \ . . \ A - — .. \ \
41- a t \ I . l \ - . , — —. 41- , a. I .§ I , — \ \ .
l \ I I I , _ I I
> _ \ \ l I I -‘ >- _ ‘ l I I ‘ \ I r
w . I , I I . l \ \ I \ l , ~ 40 \ , , . \ I I . . , , .
39_ . I / I I . 1 l I I I c \ 39_ \ | ‘ I ‘ I I I , . 4
, I I | \ I I I I I - r I / ‘ I ‘ . l \ I / /
38- \ I I ‘ \ . I I I I I . y 33. I I I I I . I \ I I x ,
I I \ I \ I I / -

_ ‘ ‘ I \ \ ~ ' ' ‘ ‘ _ / I I I I \

37 , \ \ \ _ . . \ a t r 37 I I , . ‘ . I I . I I I , _
36- l ‘ 1 I \ l | \ \ \ \ \ \ 36- I I I \ \ V I 4 / I I \ I

I I I ‘ ‘ I \ . . \ \ \ I I I I | \ \ \ l l I I I \ \ x
35- I I I I I \ \ I ' ' ‘ ' ’ ’ 35-—- / I I \ ‘ ' ‘ ‘ '7 ‘ '-
.....I...III...I....I....I.I..II...I,. 34.1. 1....11..I....I....I.I..II.
'7 68 7o 72 14 7e 78 so 68 7o 72 74 76 7a 80

x x
110mm 120mm

Figure 3.3.13. Instantaneous velocity measurements of plume from 100 to 120 minutes.

Many convective features that were apparent in the ﬂow visualization were
measured by the MTV technique. At 71 minutes into the solidiﬁcation process, it can be
seen that the plume has a velocity of ~7mm/s with surrounding ﬂuid moving at a velocity
of ~ lmm/s. Plume shedding was also measured with velocities on the order of 1mm/s.
Table 4 condenses the maximum or minimum velocity components of measurements

made.

60

 

Time

U - Component

V - Component

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(minutes) (mm/sec) (mm/sec)
71 1.77 6.49
75 -0.97 4.98
76 -0.97 4.98
77 -1.66 4.75
78 -0.89 4.03
79 -0.88 5.40
80 1.17 4.40
81 1.30 3.19
82 -1.10 3.18
83 1.20 1.77
84 1.72 4.44
85 -1.33 4.18
86 1.87 4.88

87.5 1.28 4.53
88 1.03 4.24
90 1.10 2.70
91 1.08 5.25
92 -l .00 3.02
93 -1.49 4.28
94 -1.11 4.28
95 -1.32 3.76
96 1.44 4.14
97 0.74 3.96
99 -0.59 3.69
100 -0.59 4.71
101 1.46 4.49
105 -1.39 3.65
110 0.70 3.39
120 1.41 3.83

 

61

Table 4. Instantaneous maximum/minimum velocity measurements for the plume

 

3.4 Flow Rate Calculation above the Mush

In trying to determine the porosity of the region just above the mush, the average
speed was computed for all y locations for select time ticks of the three ﬁelds of view of
measurements of the mushy region. The V-component of velocity is integrated over the
entire constant y location to obtain an average speed moving across the particular y
location. This procedure is performed at multiple locations and time ticks to produce
contour plots. The equation utilized for the calculation of the average speed is as shown

below:

I I V - dxdz
Ax
Since velocity information is not available in the 2 plane, we assume the dz term to be

unity. By dividing the integrated value of velocity, V, by Ax, we essentially obtain an

average speed in mm/s.

The following ﬁgures show the contour plots that were obtained from the

calculation of the average speed above the mush at the three ﬁelds of View.

62

 

V(mmls)
-0.16 -0.05 0.06 0.17 0.28 0.39 0.50

 

 

 

‘, V (mm/s) ‘
-0.66 —0.45 -0.25 —0.04 0.17 0.37 0.58.
\

 

 

 

29 -0.16 -0.05 0.06 0.17 0.28 0.39 0.50

 

 

 

!\7(mm/s)
-0.16 -0.05 0.06 0.17 0.28 0.39 0.50

Y (5an #44

 

Li‘,w

 

 

i100" “ois'H‘i” «15” 1
i. , aﬁiﬂttsecionmw..- 74’
4min lSt FOV 8min lSl FOV g
V ("TM/8) 1 V (mm/s)

-0.66 -0.45 -0.25 -0.04 0.17 0.37 0.58

      

 

Lama“) a,
8mmzﬂrov

[_ i _. —_.___ 7!

.\7(mmls)
-O.66 045 -0.25 —0.04 0.17 0.37 0.58

 

l '* ......

l
1.5

 

A .1.
iii, t(secongs_)## ,,,,,,
8mm3dFOV

Figure 3.3.14. Flow rate plots at 4 and 8 minutes of three ﬁeld of views.

63

<11

(mm/s

\7(mmls)
‘1 -0.45 —0.36 -0.27 -0.18 —0.08 0.01 0.10

-037 -021 -0.05 0.11 0.27 0.43 0.58

    

 

    

11111

 

 

 

 

 

 

 

 

1 1 1 l 1 1 1 1
0 0.5 1 1.5
L t(seconds)
,2- a, ,,,1,,6,min2£€8 15' 597V..- - , A
|V(mm/s)
29 0.45 -0.36 0.27 -0.18 0.08 0.01 0.10
28-
‘ 27- 4
A26“ I
l g 25 -
‘>‘ 24 -
I
i 23—
; 22—
' 1 1 1 1 1 1 1 14 1 n 1 1
1 1i5 21 0:5 1 11-5
1,, ﬂ Vai¥giseconds) W , MJ L_,,._,,-W,,,i 11881300“) 7 g
7 *1; minutes 2"d Fov ,- .. 7 h 16 minutes 2na FOV
i 171111111131— ‘ 17 (mm/s)

, '0-37 '021 ‘0-05 0-11 0'27 0-43 0-58 I 0.45 —0.36 0.27 0.18 -0.08 0.01 0.10

   

l 1

 

. 1 1 L I v 1 1 1 1 I
0 0.5 1 1.5
t (seconds)

16 minutes 3“d FOV

1 1 l 1 1 1 l 1 1 l 1
O 0.5 1 1.5
t (seconds)
12 minutes 3rd FOV

 

 

Figure 3.3.15. Flow rate plots at 12 and 16 minutes for three ﬁeld of views.

64

In Figures 3.3.14 and 3.3.15, Y designates the location where the speed was
computed with respect to the base stainless steel plate. This procedure was repeated at
different Y locations and at different time ticks to yield the plots above for average ﬂow
rate in mm/s. At 4 minutes in the 1St FOV, it can be seen that the average ﬂow rate is
initially negative at all locations before becoming positive at around 1.5 seconds later.
This characteristic of the ﬂow at this time is exhibited in the second and third FOV as
well. At 8 minutes into the solidiﬁcation process however, this phenomenon is no longer
observed and the variation of the ﬂow in time is minimal. In the third ﬁeld of View, at a
lower location, a positive ﬂow rate is evident while a negative ﬂow rate seems to
dominate at a higher location. At 12 minutes and 16 minutes, ﬂow variations in time are
also minimal. The ﬂuctuation in the magnitude of the ﬂow rate is observed to be highest

at 8 minutes.

65

Chapter 4: Conclusion

In this study, the uni-directional solidiﬁcation of a binary alloy was studied. The
test facility used in conducting the experiments is the same as that utilized by Kay Wirtz
at Michigan State University (Wirtz, et. al. 1998) with some modiﬁcations added to allow
the ﬁrll two-component implementation of the MTV technique. However, this setup
proved to be inadequate due to the amount of condensation that forms on the outer quartz
walls whenever an experiment was performed. This is detriment to the penetration of the
laser lines used for tagging purposes during the summer season as the humidity of the
ambient air increases substantially. It is also observed that the solidiﬁcation process is
affected by the air temperature, especially the height of the mushy region, which peaked
during the winter months. To ensure that the different measurements were essentially
measuring the same process, the entire data set needs to be coordinated and completed in
a very narrow time frame so the ﬂuctuations in room temperature are negligible. This
difﬁculty can be circumvented in future work by enclosing the entire test section in a
controlled environment with preset temperature and humidity controls.

To visualize the convective mechanisms present during the solidiﬁcation process,
a modiﬁed Shadowgraph technique was utilized. This method resulted in a much higher
contrast in the visualization of the density gradients within the plume. With a
conventional Shadowgraph method, the images are two-dimensional as the resultant
images are integrated over the whole depth of the test section and projected to a ﬂat
screen. Also, the need to implement a parallel light source makes the setup for a

Shadowgraph technique more difﬁcult and the generally non-uniform illumination

66

doesn’t provide for very appealing images. With the visualization method used in this
study, the implementation procedure is highly simpliﬁed and high-resolution images of
the plumes and other convective phenomena present in the ﬂow ﬁeld are clearly visible.
This allows for better determination of the location of the plumes and chimneys during
the solidiﬁcation process and provides a marked improvement in image quality compared
to previous visualizations.

For quantitative measurements of the convective mechanisms within the ﬂow, a
non-intrusive measurement technique called MTV was implemented. The advantage of
this approach compared to other velocity measurement techniques like particle-based
method is that this technique uses a molecular scale probe which produces a minimal
effect on the very sensitive ﬂow ﬁeld within the bulk ﬂuid. This technique was used to
measure one and two-component velocities within the ﬂow ﬁeld. Above the mushy
region where there are no plumes present, a two-component velocity measurement was
implemented which yielded velocities on the order of 1mrn/s. For measurements of the
plume, a combination of the one-component and two-component techniques was
necessary to characterize the whole plume. Measured velocities within the plume
exhibited a magnitude of ~ 7mm/s which is consistent with previous measurements made
by Wirtz et. al. (1998).

The velocity measurements obtained in this study and the previous one conducted
by Wirtz et. a1. (1998) are dissimilar with the work done by Magirl and Incropera (1993)
due to the different cooling rates used in the two experiments. The initial goal of our
study was to provide a database for comparison with a numerical simulation to

characterize ﬂow conditions present during the solidiﬁcation process. In order to simplify

67

the numerical simulation in future work, the ﬁnal temperature of the base plate is set to
only -14°C, so that the solidus front does not form, as compared to -30°C in the
measurements made by Magirl and Incropera. However, due to the lower cooling rate
used in this study, the velocity measurements are inherently lower than those reported by
other researchers using a lower temperature.

To better understand the porosity of the mush, the average speed of the ﬂuid that
traverses different locations were computed and showed that the speed is typically higher
at a higher location within the test section. This is especially obvious at 8 and 16 minutes
into the solidiﬁcation process. The information provided here could then be used in future
computation analysis for the calculation of porosity within the mush.

The current work uses a single plane measurement technique. Thus all velocity
measurements are made in two dimensions. To account for out of plane motion and the
three-dimensional nature of the plume, measurement of three components of the ﬂow
ﬁeld would require stereoscopic MTV. A stereoscopic implementation of MTV is

discussed in Bohl et. a1. (2001).

68

APPENDICES

69

APPENDIX A

Molecular Tagging Velocimetry (MTV) and Ammonium Chloride Chemical

Mix Formulas

7O

Formulas for calculating weight percentage and MTV Triplex concentration

Ammonium chloride (NI-14Cl) amount corresponding to a 26 Weight Percentage

For 26 Weight Percentage:

mNH Cl
4 X100 = 26%
mNH4Cl + mhzo + mMTV

 

mNH4C, = 0.26mNH4C, + 0.26mh20 + 0.26mMTV
0.74mNH4Cl = 026th + 0.26mMTV

0.26

mNH4CI: 0 7 4 —(mh20 + mMTV)

The equation above can be simpliﬁed to exclude the weight of the MTV chemicals, since
it is added in small amounts. Below show the calculations for different weight percentage
used for the lifetime measurements.

0'06 =w0500g): 94 745g

4% Weight: mNH4CI : O 94 mhzo O 94

71

0.08 0.08

8% Weight: mNH4Cl : 6—9—2'mhzo = "665(1400g) : 121-739g

0.10 0.10

10% Weight: mma = 030mb”) = 63604008) =1551556g

0.12 0.12

12% Weight: mNH4CI = mmhzo = HES—(13008.) = 177-273g

0.16 0.16

16% Weight: mNH4CI : mmhzo : 63203003) = 247-619g

0.20 0.20
20% Weight: mNH4CI = mmhzo = ma 2008) = 300-000g

0.24 0.24

24% Weight: mNH4Cl = ﬁmhzo = 6%(11008) = 347-368g

In calculating the amounts of the MTV chemicals to add, the following formulas were
used:

Known concentration of the triplex chemicals for effective tagging:
Cyclodextrin Concentration: 2.0--4.0 x10“t M

Alchohol Concentration: 0.06 M
Bromo-naphtalene: For 11iter of volume, use less than lml, about an inch height of the

small vial.

Formulas for calculating amount of chemicals to add to a speciﬁc volume of ﬂuid

 

 

"1CD = Volume(liter) * Molar — Concentration * 130018
mo
Valcoho, = Volume(liter) * Molar — Concentration * 00g :1: 1m]
mol 0.9624 g

Bromonapthalene Concentration: Add to saturation

72

APPENDIX B

Timing Diagram of Gate Signals to the Xybion Camera

And Laser

 

73

.98me 8am «Em E: Someo— to...
BE 2: comma: 3 3m: 2 .636ch .336 383 d1 @9533 on 8 omen: vague 2t 0.30:5 Saw 9393 2: 0:5» omaﬁz «093393
2: £300 8 52:3 at 883% Sam SE SE. “2:16 Rat? mo Hum 28 3202 3 was a x02: Row M .9216 E 631:5 “onto“. 0%
mo Cowman $953 «$85 2: Eat 823m. 0H. «Rama Sam 05 mo 35:68 mEE: 2: £95qu 8083 Sumo 93% 1.53.6.» BE.

02.8 3038.; of 8 3on 8 5:5 .0.qume 3mm. mo 23m; 9:83. am ocnmmm

gin— _au.ntw> n E Suez

Smut H
.5an

930 «.H

3ND EN

Uas:
Ema

 

74

«8&8 vegan 2.: 8&3» sum Hoe 3w cocoa mcuwm 2: “ﬂ: 8mm 2: 8. one 3893 ﬂ Row 28
Eco “ﬁt 30% 0.890 “a mamas: uﬁaﬁmxm .uvoﬁ 08am :8 8 wBEmBE on o“ Eamon vegan 2t GEE» 2am 8 we as mﬁqma mcumm SE.

92% He once» 2: 8. 30%“: ﬁg? 883 maﬁow acouubwv com 88mg 9339. .mm 85mm

hﬂou mEocH
awn—av mEom
ha—ou mEcw
ha—ou mEE.
emu—mu 25¢
3% 25m
and—mu mEov

ha—ou qum

32. 2:8

mm ow 1g

22.. 35
22c germ

 

75

Alternating lines
due to one ﬁeld not
being exposed
within the ﬁeld
overlap region

 

m
m
“
__——————————_
—_—-————-—_
w
_
_
————_—-——————
w
m
*—
w
W
u———-———————_
—_—_—
_
_——u_——-
*
—-—————_
m
m
m
——_——————
w
w
_
_——-_——-
m
w
W
_
w
_
w
m
w
_
m
w
_
w
m
w
W
. . -....-_
. - I .
“—1....
.—
-— .-1
_ . .
--—— .-.. 7
-—1
———— .

 

Figure B3. Sample image of gating the camera in an odd or even ﬁeld

APPENDIX C

X-Y Ratio of the Xybion Camera

77

uuuﬁuusﬁoﬁ 1.33 8 gouache m 2.: 8 ﬂexinm a donoohw x 05 E mﬁoxdanm “at human 3833 on can a “BS. .uEMmamoa
a “an“. Huang uaomo 83.35% m at? - .8283 a 8058 qumN on“ do 900 oﬁmo 035 an 05 “an“ even m“ a 63% 0.363 2.: 88m

 

$388388 mcmvaoamotoo 2: 93 $083 2353““ 2: ﬁt: :83 amen: 38m .HO 0.3mm

 

 

 

 

 

 

 

 

.8535 a a 53:3. ”.../05.5.1.5! .

.5586
w .8 N uﬁ .556 E on 8 mvuoq 2:0 «.683 05 mam: ﬂangeuaxo com owns: 3.8m 23 .083 2: a 038 Mon Rm 3 .3: ﬁt: 8
«guano coBmN 2.3 we“? 5&3 was» omen: ad donoohw M van N 2% E 322886 Hana 08mm oﬁ we: wow: 8250 on» $5 335 ch.

  

78

APPENDIX D

Calculation of temperature increase of aqueous ammonium chloride due to

irradiation by Excimer Laser

79

Temperature Increase of Fluid Due to UV Laser Illumination

 

0 Laser energy used Emax = 100mJ/pulse

0 Size ofmain laser beam = 3.8 X 1.3 cm

0 Approximate size of laser beam aﬁer VFL: 60mm x 1mm

0 Using a 50/50 beam splitter, the energy of the laser is divided into 2 beams of equal
intensity

0 Assuming 21 thick lines (worst case scenario where the lines are much thicker than
the actual beam blockers)

0 Total area where laser passes through = (21 x 0.015cm)x 1mm = 3.15 x 10""m2

Fraction of energy that passed by = 3.15 x 10'6/60 x 10'6 ~ 0.05

Volume of ﬂuid irradiated, V = 21 x 1mm x 0.015cm = 3.15 x 10'8m3
mAmc1= pV = 1078 x v = 3.39 x10'5 kg

Q = meAT = 3.39 x 10'5kg x (3.249kJ/kg.K) x AT = 10m]

AT = 0.09 K

we have assumed that a = 1 whereby all the laser is absorbed by the ﬂuid.

80

BIBLIOGRAPHY

81

Bibliography

Beckerman, C.; Wang, CY, 1996, "Equiaxed dendritic solidiﬁcation with convection:
Part III. Comparisons with NI—I4Cl-H20 experiments," Metallurgical and Materials
Transactions A, 27A, pp. 2784-2795.

Bohl, D.G., Koochesfahani, M.M. and Olson, B.J., 2001, "Development of stereoscopic
molecular tagging velocimetry," Experiments in Fluids, 30, pp. 302-308.

Chen, CF. and Chen, R, 1991, "Experimental study of directional solidiﬁcation of
aqueous ammonium chloride solution," J. Fluid Mech., 227, pp. 567-5 86.

Cohn, R.K. and Koochesfahani, M.M., "The accuracy of remapping irregularly spaced
velocity data onto a regular grid and the computation of vorticity" Experiments in
Fluids," Experiments in Fluids, 29, pp. 61-69

Gendrich, CF. and Koochesfahani, M.M., 1996, "A spatial correlation technique for
estimating velocity ﬁelds using molecular tagging velocimetry (MTV)," Experiments in
Fluids, 22, pp. 67-77.

Gendrich, C.P., Koochesfahani, M.M. and Nocera, D.G, 1997, "Molecular Tagging
Velocimetry and other novel applications of a new phosphorescent supramolecule,"
Experiments in Fluids, 23, pp. 361-372

Hellawell, A, Sarazin, JR. and Steube, RS, 1993, "Channel convection in partly
solidiﬁed systems," Phil. Trans. R. Soc. Lond. A, 345, pp. 507-544.

Incropera RP. and Magirl, CS, 1993, "Flow and Morphological Conditions associated
with Unidirectional Solidiﬁcation of Aqueous Ammonium Chloride," Transactions of the
ASME, 115, pp.1036-1043.

Koochesfahani, M.M., 1999, "Molecular Tagging Velocimetry (MTV) Progress and
Applications (Invited),AIAA Paper 99-3786," 30"1 AIAA Fluid Dynamics Conference, 28
June-1 July 1999.

Koochesfahani, M.M., Cohn. R.K., Gendrich, CF. and Nocera, D.G, 1996, "Molecular
tagging diagnostics for the study of kinematics and mixing in liquid phase ﬂows",
Presented at the Plenary Session of the Eight International Symposium on Applications of
Laser Techniques to Fluid Mechanics, Lisbon, Portugal, July 8-11, 1-12

Liu, S. and Hellawell, A, 1999, "Experiments with constrained chimney-plume ﬂows in
the system ammonium chloride-water: comparison with unconstrained case," J. Fluid
Mech., 388, pp. 21-48.

82

Magirl, CS. and Incropera, FR, 1993, "Flow and morphological conditions associated
with unidirectional solidiﬁcation of aqueous ammonium chloride", J. Heat Transfer, 115,
1036-1043.

Prescott, R]. and Incropera, ER, 1996, "Convective heat and mass transfer in alloy
solidiﬁcation," Advances in Heat Transfer, 28, pp.231-238.

Solomon, TH. and Hartley, RR, 1998, "Measurements of the temperature ﬁeld of mushy
and liquid regions during the solidiﬁcation of aqueous ammonium chloride", J. Fluid
Mech., 358, pp 87-106.

Schulze, TR and Worster, M.G, 1998, "A numerical investigation of steady convection in
mushy layers during the directional solidiﬁcation of binary alloys", J. Fluid Mech., 356,
pp.199-220.

Wang, S.Y., Liu, Y.Z., Lin, C.X. and Ebadian, M.A., 1998, "Double diffusive velocity
measurement during the solidiﬁcation process using particle image velocimetry"
Proceedings of the ASME Heat Transfer Division, HTD-Vol 361-4, pp. 111-118.

Wang, S.Y., Lin, C.X. and Ebadian, M.A., 1999, "Study of double diffusive velocity
during the solidiﬁcation process using particle image velocimetry Int. J. Heat Mass
Transfer, 42, pp. 4427-4445.

Wang, S.Y., Lin, C.X. and Ebadian, M.A., 1999, "Vortex ﬂow of low concentration
NH4Cl-H20 solution during the solidiﬁcation process" Int. J. Heat Mass Transfer, 42, pp.
4153-4163.

Wﬁllhover, K., Kérber, Ch., Scheiwe, M.W. and Hartmann, U., 1985, "Unidirectional
freezing of binary aqueous solutions: an analysis of transient diffusion of heat and mass"
Int. J. Heat Mass Transfer, 28-4, pp. 761-769.

Worster, M.G, 1997, "Convection in mushy layers" Annu. Rev. Fluid Mech. 29, pp. 91-
122.

Wirtz, K., Koochesfahani, M., McGrath, U. and Benard, A., 1998, "Molecular Tagging
Velocimetry applied to buoyancy-driven convective phenomena during solidiﬁcation,"
Proceedings of the ASME Heat Transfer Division, 4, pp. 103-110.

Wirtz, Kay, 1998, "Experimental study of convection phenomena associated with

solidiﬁcation processes", Diplomarbeit thesis, RWTH-Aachen, conducted at Mechanical
Engineering Department, Michigan State University

83

General References

Chen, Y. -J. and Davis, SH, 1999, "Directional solidiﬁcation of a binary alloy into a
cellular convective ﬂow: localized morphologies", J. Fluid Mech., 395, pp. 253-270.

Schulze, TB and Worster, MG, 1999, "Weak convection, liquid inclusions and the
formation of chimneys in mushy layers", J. Fluid Mech. 388, pp. 197-215.

Liu, S. and Hellawell, A., 1999, "Experiments with contrained chimney-plume ﬂows in
the system ammonium chloride-water: comparison with the unconstrained case", 388, pp.
21-48.

Cummings, L.M., Hohlov, Y.E., Howison, SD. and Komev, K, 1999, "Two-dimensional
solidiﬁcation and melting in potential ﬂows", 378, pp. 1-18.

Grifﬁths, R.W. and Fink, J.H., 1997, "Solidifying Bingham extrusions: a model for the
growth of silicic lava domes", 347, pp. 12-36.

Wettlaufer, 1.8. and Worster, M.G, 1997, "Natural convection during solidiﬁcation of an
alloy from above with application to the evolution of sea ice", 344, pp. 291-316.

Schiafﬁno, S. and Sonin, A.A., 1997, "Formation and stability of liquid and molten beads
on a solid surface", 343, pp. 95-110.

84