Mechanism of the Novel Antisolvent Vapour Precipitation ... · Mechanism of the Novel Antisolvent...
Transcript of Mechanism of the Novel Antisolvent Vapour Precipitation ... · Mechanism of the Novel Antisolvent...
Jiunn Yuan Tan (21637520) Page 1
Mechanism of the Novel Antisolvent Vapour
Precipitation (AVP) Process
A thesis in fulfillment of the requirements for
Master in Engineering Science (Research) Degree
By
Jiunn Yuan Tan
Department of Chemical Engineering
Monash University, Clayton Campus, Australia
March 2015
Jiunn Yuan Tan (21637520) Page 2
Notice 1
Under the Copyright Act 1968, this thesis must be used only under the normal
conditions of scholarly fair dealing. In particular no results or conclusions should be
extracted from it, nor should it be copied or closely paraphrased in whole or in part without
the written consent of the author. Proper written acknowledgement should be made for
any assistance obtained from this thesis.
Jiunn Yuan Tan (21637520) Page 3
Declaration:
In accordance with Monash University Doctorate Regulation 17 / Doctor of
Philosophy and Master of Philosophy (MPhil) regulations, I hereby declare this thesis
contains no material which has been accepted for the award of any other degree or diploma
at any university or equivalent institution and that, to the best of my knowledge and belief,
this thesis contains no material previously published or written by another party, except
where due reference is made.
Signed: ………………………………………………
Date: ………............................................
Jiunn Yuan Tan (21637520) Page 4
Table of Contents ABSTRACT..................................................................................................................................7
ACKNOWLEDGEMENT...............................................................................................................9
LIST OF FIGURES......................................................................................................................11
LIST OF TABLES........................................................................................................................15
ABBREVIATIONS......................................................................................................................16
NOMENCLATURES...................................................................................................................17
CHAPTER 1: INTRODUCTION
1.1 Background ..................................................................................................................... 23
1.2 Drying and Analysis of the Mass Change Profiles of Different Maltodextrin Particles
using the Modified Single Droplet Rig .................................................................................. 26
1.3 Modelling of the AVP Drying Process ............................................................................. 27
1.4 Research Aim .................................................................................................................. 28
CHAPTER 2: LITERATURE REVIEW
2.1 Spray Drying ................................................................................................................... 30
2.2 Precipitation Method ..................................................................................................... 32
2.2.1 Liquid Antisolvent (LAS) Precipitation ..................................................................... 32
2.2.2 Supercritical Antisolvent (SAS) Precipitation........................................................... 33
2.3 Antisolvent Vapour Precipitation (AVP) ......................................................................... 35
2.4 Current Vapour Generation and Humidity Measurement Technique ........................... 39
2.4.1 Vapour Generation Method ..................................................................... …………….39
2.4.2 Vapour Humidity Measurement Technique ............................................................ 44
2.5 Modelling the Drying of a Droplet ................................................................................. 45
2.5.1 Lump Versus Distributed Model .............................................................................. 46
2.5.2 Effect of Solute on Droplet Drying........................................................................... 47
2.5.3 Mass Depression Phenomenon ............................................................................... 48
2.5.4 Multicomponent Modelling .................................................................................... 51
2.5.5 Modelling Absorption .............................................................................................. 52
2.6 Summary and Remarks .................................................................................................. 53
Jiunn Yuan Tan (21637520) Page 5
CHAPTER 3: MATERIALS & METHODS
3.1 Materials ........................................................................................................................ 56
3.1.1 Preparation of Maltodextrin/Maltose Samples ...................................................... 56
3.1.2 Solvent and Antisolvent ........................................................................................... 56
3.2 Original AVP Single Droplet Drying Experiment ............................................................. 56
3.2.1 Experimental Method .............................................................................................. 56
3.2.2 Ethanol Vapour Humidity Measurement ................................................................ 58
3.2.3 Control Experiments ................................................................................................ 65
3.3 Modified AVP Single Droplet Rig .................................................................................... 65
3.3.1 Vapour Generation System ..................................................................................... 65
3.3.2 Safety Considerations .............................................................................................. 67
3.3.3 Control Systems ....................................................................................................... 68
3.4 Mass Change and Temperature Measurements of the Droplet during Drying ............. 84
3.5 Solubility Measurement ................................................................................................. 85
CHAPTER 4: RESULTS & DISCUSSIONS
4.1 Unveiling the Mechanism of AVP in Producing Porous and Spherical Particles ............ 87
4.1.1 Control experiment using nitrogen gas ................................................................... 87
4.1.2 Ethanol Vapour Precipitation .................................................................................. 88
4.1.3 Effect of Relative Humidity and Absolute Humidity ................................................ 91
4.1.4 Effect of Initial Concentration and Chain Length .................................................... 91
4.1.5 Porous and Microspheres Formation ...................................................................... 94
4.1.6 Crystallisation and Precipitation .............................................................................. 97
4.1.7 Application ............................................................................................................. 100
4.2 Analysis of the Single Droplet Drying of Maltodextrin under AVP .............................. 101
4.2.1 Single Droplet Drying of Maltodextrin under AVP using the Modified Single
Droplet AVP Rig .............................................................................................................. 101
4.2.2 Discussion .............................................................................................................. 105
4.3 Modelling of the Simultaneous Absorption and Evaporation Process of the Droplet
under AVP ........................................................................................................................... 108
4.3.1 Theoretical Modelling Method .............................................................................. 108
4.3.2 Comparison of Mass Change Profile of Pure Water Droplet and Maltodextrin
Solution Droplet Dried under AVP .................................................................................. 114
Jiunn Yuan Tan (21637520) Page 6
4.3.3 Comparison of Experimental and Theoretical Modelling of the Mass Change
Profile of AVP Process..................................................................................................... 116
4.3.4 Comparison of Experimental and Theoretical Modelling of the Mass Change and
Temperature Profile of Pure Droplet ............................................................................. 119
4.3.5 Comparison of Experimental and Theoretical Modelling of the Mass Change of
Water-Ethanol Droplet ................................................................................................... 123
4.3.6 Comparison of Experimental and Theoretical Modelling with Mass Transfer
Depression of the Mass Change Profile of AVP Process ................................................ 125
4.3.7 Discussion .............................................................................................................. 127
CHAPTER 5: CONCLUSIONS & RECOMMENDATION
5.1 Conclusions ...................................................................................................................... 130
5.2 Recommendations ........................................................................................................... 132
5.3 List of Publications………………………………………………………………………………………………………135
REFERENCES……………………………………………………………………………………………………………………..136
APPENDIX…………………………………………………………………………………………………………………………153
Jiunn Yuan Tan (21637520) Page 7
ABSTRACT
Ultrafine spherical maltodextrin and maltose particles were successfully produced
with the Antisolvent Vapour Precipitation (AVP) technique. Comparison between lactose
and maltodextrin reaffirmed that a key requirement for the process is its ability to inhibit
crystallization of the material. The precipitation process consists of: (1) an initial phase
separation forming an emulsion, (2) phase inversion and (3) finally a water-maltodextrin
shrinkage phase which forms the spherical particles driven by interfacial surface tension.
Dehydrating the droplet at different stages of the process resulted in different particle
morphologies; porous, smooth, microsphere network and microspheres. Higher ethanol
relative humidity, higher ethanol absolute humidity and lower initial weight concentration
were found to favour the formation of amorphous microspherical particles upon drying. A
unique liquid phase separation was observed which leads to the proposed particle
formation mechanism for the AVP process.
Further quantitative study were conducted using a newly built vapour generation
system, incorporating the LabView control and monitoring system and a new humidity
measurement technique based on fundamental mass and energy balance. Analysis of the
mass change of the droplet throughout the AVP drying process revealed a trend which
suggests that the maximum ethanol concentration within the droplet may be the prevailing
factor governing microsphere formation. In addition, an interesting observation on the final
solid mass recorded for the porous and microsphere network structures showed that these
structures exhibit liquid retention behaviour which could be useful for encapsulation
applications.
Jiunn Yuan Tan (21637520) Page 8
In order to better understand the mechanism of the AVP process, additional
research was conducted to develop an AVP drying model to describe the simultaneous
absorption and evaporation of ethanol and water within the droplet. This model was
developed based on fundamental heat and mass transfer analysis and the incorporation of
Raoult’s Law and UNIFAC model to account for the binary interaction between water and
ethanol. Comparison between the model and the experimental measurements revealed
overestimation of ethanol absorption and total drying time. Further analysis suggests that
this may be attributed to the counter diffusion of water and ethanol during the drying
process and possibly non-Fickian diffusion behaviour of ethanol within the droplet. This
work provides a fundamental basis for future work on modelling of this physical
phenomenon.
The mechanism and analysis provided in this work have contributed to a
fundamental understanding of the AVP process in forming microspherical particles, which
has potential application in drug delivery. Based on the mechanism proposed, the
underlying principles of particle formation under AVP drying can be applied to other
materials. The formation of porous and microsphere network, with high liquid retention
behaviour suggests possible encapsulation application. In addition, the analysis and
considerations employed for the AVP drying model provide a fundamental basis for further
model development which would be useful in scaling up the AVP process.
Jiunn Yuan Tan (21637520) Page 9
ACKNOWLEDGEMENT
I would like to take this opportunity to communicate my deepest gratitude to all who
have helped me throughout this challenging yet amazing research experience. It is of
certainty that without your contribution whether directly or indirectly, I will not be able to
pull through the tough times and persevere forward to complete this research work.
First and foremost, I would like to express my sincerest gratitude towards my
supervisors Dr. Meng Wai Woo, Dr. Shahnaz Mansouri, Professor Karen Hapgood and
Professor Xiao Dong Chen for their patience, guidance and encouragement throughout my
entire candidature. Their opinions, ideas and feedbacks during this entire research course
have proven to be invaluable and essential in making this thesis possible. When I am faced
with a challenge in my research, it is their ongoing encouragement and assurance that
reignited my confidence. They are always proud of my achievements no matter how trivial it
might be and that has truly motivated me to work harder. Special thank you to my main
supervisor, Dr. Meng Wai Woo, for his prompt feedbacks, creative ideas and invaluable
advices. It has been truly an amazing experience to be able to work and learn from a brilliant
yet down to earth supervisor like yourself.
I would also like to thank all my colleagues in the Food Engineering Group in the
Department of Chemical Engineering in Monash University: Dr. Jia Han Chew, Dr. Wenjie Liu,
Ruohui Lin, Sean Chew, Kathryn Waldron, Peter Tsirikis and Paurnami Chandran. They have
been helpful and fun colleagues who brightened my research life. Thank you for organizing
all the fun events for our research group; it has truly been an enjoyable experience spending
time with all of you.
Jiunn Yuan Tan (21637520) Page 10
Thank you Dr. Xi Ya Fang and Dr. Ma Jisheng from MCEM for their assistance in
conducting the training and help on using the Scanning Electron Microscope (SEM).
Besides that, I would also like to acknowledge the administrative and technical staff
within the Department of Chemical Engineering, particularly Ron Graham, Kim Phu, Jill
Crisfield, Lilyanne Price, Gamini Ganegoda, Ross Ellingham, Harry and Rebecca Bulmer.
Thank you for your help in making the university a conducive and safe place for research
work. Thank you for the financial support for the project given by the Australian Research
Council via a Discovery Grant: DP 130104836 and the scholarship from the Faculty of
Engineering, Monash University.
Last but not least, I would like to thank my family and my friends for always being
there for me throughout my entire research period. Your support and words of
encouragement have been my pillars of strength in this research project.
Jiunn Yuan Tan (21637520) Page 11
LIST OF FIGURES
Chapter 2
Figure 2.1: Schematic Diagram of a Typical Spray Drying Process
Figure 2.2: Schematic diagram of SAS apparatus
Figure 2.3: Schematic diagram of the physical phenomenon of AVP
Figure 2.4: Sketch of a lab-scale humidity generator
Figure 2.5: Images of hygrometer, alcohol breathalyser and gas detector (left to right)
Figure 2.6: Drying stages of a liquid droplet containing solid
Chapter 3
Figure 3.1: Schematic figure of preliminary experimental set-up for single-droplet glass filament
drying system using ethanol vapor as antisolvent. (a) Nitrogen tank; (b) Ethanol chambers; (c) Water
bath; (d) Drying chamber; (e) Suspending glass filament; (f) Camcorder
Figure 3.2: Wet bulb and dry bulb measurement set-up at the bleed line
Figure 3.3: Comparison of theoretical and experimental WB and DB temperatures for (a)
ethanol; (b) water and (c) acetone
Figure 3.4: Schematic figure of the proposed experimental set-up for single droplet ethanol vapour
distributor system. (a) Peristaltic pump; (b) Load cell; (c) Round bottom flask heating mantle; (d) Tee;
(e) Convective gas heater
Figure 3.5: Programming Flow for the Execution of Humidity Control
Jiunn Yuan Tan (21637520) Page 12
Figure 3.6: Ethanol AH Control Stabilization for On/Off Heating Method for AHSP = 0.1
Figure 3.7: Ethanol AH Control Stabilization PID Control within the Heating Mantle for AHSP
= 0.06
Figure 3.8: Ethanol AH Control Stabilization for Set-Points (a) 0.09; (b) 0.065 and (c) 0.038
Figure 3.9: Water AH Control Stabilization for Set-Points (a) 0.015; (b) 0.011 and (c) 0.006
Figure 3.10: Acetone AH Control Stabilization for Set-Points (a) 0.13; (b) 0.1 and (c) 0.07
Figure 3.11: Dual-stream AH and RH control stabilization
Figure 3.12: Schematic diagram of mass change single droplet rig
Chapter 4
Figure 4.1: Drying behaviour over time of a single droplet using (a) nitrogen gas; (b) nitrogen
gas and ethanol vapour
Figure 4.2: (a) (i) Clustered crystal structure; (ii) Smooth solid structure; (b) (i) Smooth
surface; (ii) Patchy porous network; (iii) Round porous network; (iv) Microsphere network; (v)
Microsphere
Figure 4.3: Graph of ethanol relative humidity against initial weight concentration for: (a)
Maltodextrin DE 10; (b) Maltodextrin DE 18 and (c) Maltose (solid line – apparent boundary
for microspheres formation, dash line – apparent boundary of microsphere network
formation)
Figure 4.4: Microscope images of the cloudy droplet
Figure 4.5: Schematic diagram of the proposed mechanism of phase separation
Jiunn Yuan Tan (21637520) Page 13
Figure 4.6: SEM images of particle structures obtained under AVP drying at EAH of (a) 0.05
kg/kg db; (b) 0.065 kg/kg db and (c) 0.08 kg kg/db
Figure 4.7: Particle structure map for maltodextrin DE 10
Figure 4.8: Mass change profile of the AVP drying of (a) Microspheres; (b) Microsphere
network and (c) Porous network particles
Figure 4.8: Comparison of wet bulb temperature by experimental measurement wet bulb
temperature predicted by the model for a water droplet with dry nitrogen gas at 35oC
Figure 4.9: Solubility curve of maltodextrin DE 10 in water-ethanol mixture
Figure 4.10: Mass change profile of the drying of water and maltodextrin DE 10 solution (5
wt%) at EAH: (a) 0.05 kg/kg db; (b) 0.065 kg/kg db and (c) 0.08 kg/kg/db
Figure 4.11: Comparison of droplet mass change profile for drying of a water droplet under
AVP measured experimentally and predicted by the model for ethanol absolute humidity: (a)
0.08 kg/kg db; (b) 0.065 kg/kg db and (c) 0.05 kg/kg db
Figure 4.12: Comparison of mass change and droplet temperature by experimental
measurement and model prediction for evaporation of a pure water droplet with dry
nitrogen at 40oC
Figure 4.13: Comparison of mass change and ethanol droplet temperature by experimental
measurement and model prediction for drying of a water droplet with dry nitrogen gas at
40oC
Jiunn Yuan Tan (21637520) Page 14
Figure 4.14: Comparison of mass change by experimental measurement and model
prediction for drying of an ethanol-water droplet with dry nitrogen gas at 25oC
Figure 4.15: Comparison of mass change by experimental measurement and model
prediction (with depression) for drying of water droplet under AVP
Appendix
Figure A.1: 3-D Plot of the Experimental Matrices Undertaken for Maltodextrin DE 10
Figure A.2: 3-D Plot of the Experimental Matrices Undertaken for Maltodextrin DE 18
Figure A.3: 3-D Plot of the Experimental Matrices Undertaken for Maltose
Figure A.4: Spherical Particle Size Distribution for Maltodextrin DE 10
Figure A.5: Spherical Particle Size Distribution for Maltodextrin DE 18
Figure A.6: Spherical Particle Size Distribution for Maltose
Jiunn Yuan Tan (21637520) Page 15
LIST OF TABLES
Chapter 2
Table 2.1: Range of Particle Sizes Obtained in Spray Dryers of Different Products
Table 2.2: List of Methods Used to Generate Vapours for Laboratory Applications
Table 2.3: List of Typical Mass Transfer Coefficient Expressions in Literatures for Different
Droplet Evaporation Conditions
Appendix
Table A.1: Summary Results of Maltodextrin DE 10
Table A.2: Summary Results of Maltodextrin DE 18
Table A.3: Summary Results of Maltose
Jiunn Yuan Tan (21637520) Page 16
ABBREVIATIONS
ERH – Ethanol Relative Humidity
EAH – Ethanol Absolute Humidity
AVP – Antisolvent Vapour Precipitation
LAS – Liquid Antisolvent Precipitation
SAS – Supercritical Antisolvent Precipitation
APIs – Active Pharmaceutical Ingredients
MCP – Mixture Critical Pressure
HHG – Hybrid Humidity Generator
LIVG – Liquid Injection Vapour Generator
WB – Wet Bulb Temperature
DB – Dry Bulb Temperature
VLE – Vapour-Liquid Equilibrium
DE – Dextrose Equivalent
WPI – Whey Protein Isolate
Jiunn Yuan Tan (21637520) Page 17
NOMENCLATURES
Chapter 3
Twb = Wet bulb temperature (oC)
Ta = Dry bulb temperature at humidity box (oC)
Ta’ = Dry bulb temperature at drying chamber (oC)
ht = Heat transfer coefficient (W/m2)
A = Surface area (m2)
Heva = Heat of vaporization (J/kg)
hm = Mass transfer coefficient (m/s)
Cs = Concentration of vapour at the surface (kg/m3)
Ca = Concentration of vapour in the bulk convective medium (kg/m3)
σ = Boltzmann constant (W/m2K4)
ε = Emissivity
Sh = Sherwood number
Dab = Binary mass transfer diffusion coefficient (m2/s)
L = Critical length scale (m)
Nu = Nusselt number
Jiunn Yuan Tan (21637520) Page 18
k = Thermal conductivity (W/m.K)
Re = Reynolds number
Sc = Schmidt number
Pr = Prandtl number
C1 = Sherwood number coefficient
C2 = Nusselt number coefficient
n1 = Sherwood number exponent
n2 = Nusselt number exponent
= Density of nitrogen (kg/m3)
V = Velocity of nitrogen flow (m/s)
µ = Viscosity of nitrogen (N.s/m2)
Cpv = Heat capacity of vapour (J/kg.K)
Cp = Heat capacity of nitrogen (J/kg.K)
C’ = 0.683 (Nusselt number coefficient for a cylinder)
Bm = Mass transfer number
Bt = Heat transfer number
= Non-dimensional parameter defined by
R = Universal gas constant (J/K.mol)
Jiunn Yuan Tan (21637520) Page 19
m = Mass of vapour (kg)
V’ = Volume of vapour (m3)
Pv =Partial pressure of vapour (Pa)
Pa =Partial pressure of bulk convective medium (Pa)
Psat = Saturation pressure of vapour (Pa)
Ma = Molecular mass of convective medium (g/mol)
Mv = Molecular mass of vapour (g/mol)
AH = Vapour Absolute Humidity (kg / kg db)
AHSP = Vapour Absolute Humidity Set Point (kg / kg db)
RH = Vapour Relative Humidity (%)
RHSP = Vapour Relative Humidity Set Point (%)
v = Input Signal Voltage (V)
vsp = Voltage that Corresponds to Set-point Temperature (V)
w = Rate of Change of Voltage
ΔAH = AH – AHSP (kg / kg db)
Jiunn Yuan Tan (21637520) Page 20
Chapter 4
= Total droplet mass at a given time, n (g)
= Predicted total droplet mass based on the instantaneous mass change (g)
= Net mass transfer of the droplet at a given time, n (g)
= Net mass transfer of water at a given time, n (g)
= Net mass transfer of ethanol at a given time, n (g)
A = Area of the droplet (m2)
= Mass transfer coefficient of water (m/s)
= Mass transfer coefficient of ethanol (m/s)
= Water activity
= Ethanol activity
= Concentration of water at the convective medium (kg/m3)
= Concentration of water at the surface (kg/m3)
= Concentration of water at the convective medium (kg/m3)
= Concentration of water at the surface (kg/m3)
= Mol fraction of water
= Mol fraction of ethanol
Jiunn Yuan Tan (21637520) Page 21
= Partial saturation pressure of water (Pa)
= Partial saturation pressure of ethanol (Pa)
RHw = Relative humidity of water
RHe = Relative humidity of ethanol
R = Universal gas constant (J/K.mol)
= Molecular weight of water (g/mol)
= Molecular weight of ethanol (g/mol)
= Ambient temperature (oC)
= = Wet bulb temperature (oC)
Sh = Sherwood number
Dab = Binary mass transfer diffusion coefficient (m2/s)
r = Radius of the droplet (m)
Re = Reynolds number
Sc = Schmidt number
= Density of convective medium (kg/m3)
= Velocity of convective medium flow (m/s)
µ = Viscosity of convective medium (N.s/m2)
= Droplet temperature at a given time, n (oC)
Jiunn Yuan Tan (21637520) Page 22
= Predicted droplet temperature based on the instantaneous mass change (oC)
= Net change in temperature of the droplet at a given time, n (oC)
= Total mass of droplet (g)
= Mass of water (g)
= Mass of ethanol (g)
Cp = Heat capacity of droplet (J/kg.K)
ht = Heat transfer coefficient (W/m2)
= Latent heat of vaporization of water (J/kg)
= Latent heat of vaporization of ethanol (J/kg)
σ = Boltzmann constant (W/m2K4)
ε = Emissivity
Nu = Nusselt number
k = Thermal conductivity (W/m.K)
= Thermal conductivity of convective medium (W/m.K)
D = Diameter of droplet (m)
Pr = Prandtl number
Jiunn Yuan Tan (21637520) Page 23
CHAPTER 1
INTRODUCTION
1.1 Background
Spray drying has received wide attention in the food and pharmaceutical industry
and is used extensively to produce solid products from a solution or suspension
commercially. The technique of spray drying takes advantage of rapid solvent evaporation
from atomized droplets. The liquid stream is atomized into fine droplets in order to increase
the surface area and to accelerate the heat transfer as well as the evaporation process
when it is mixed with the drying gas in the drying chamber. Many reviews can be found on
the technique 1, application 2 and particle formation 3 of spray drying. In a conventional
spray dryer, a single particle, typically in the range of tens or hundreds of micron is
produced from each atomized droplet. The final particle size is predominantly controlled by
the size of the initial atomized droplet. In commercial applications, the product fluid is
normally atomized into droplets with a range of sizes. The ability to produce uniform
droplets with commercial nozzle or rotating atomizers is still a challenge. This inevitably
leads to dried particles of varying sizes. In a similar vein, it is a challenge to produce
sufficiently fine droplet sizes to produce final dried particles in the sub-micron range.
Although there are commercially available ’nano-scale’ spray dryers 4, these units are mainly
designed for lab scale applications 5 with relatively low flow rates.
Jiunn Yuan Tan (21637520) Page 24
Recently, a novel antisolvent vapour precipitation has been developed for intended
application in spray dryers 6. This approach, which incorporates ethanol vapour as the
antisolvent convective drying medium instead of hot air allows the production of large
numbers of uniform microparticles from within a single droplet at normal atmospheric
conditions. The premise of the process is in allowing the aqueous droplets to absorb ethanol
vapour. This new approach was first explored with aqueous lactose droplets in which the
absorption of ethanol reduced the solubility of lactose; precipitating the lactose as drying
proceeds. As a result, the particles produced are not directly determined by the size of the
initial droplet. Furthermore, relatively smaller particles are produced without the need to
generate very fine initial droplets.
Surprisingly, in contrast to the conventional precipitation process, amorphous
lactose microspheres were obtained. Scenario based analysis showed that the conventional
supersaturation based mechanism which typically describes crystallization or precipitation
processes might not be adequate to describe this phenomenon. Previous work elucidated a
unique 'pinched off' mechanism, in which the lactose phase separates within the droplet,
shrinks and eventually forms very fine spherical particles due to surface tension, could be
the driving force for the process 7. A unique observation in the previous report is that the
occurrence of any crystallization within the droplet, under certain operating conditions, will
negate the pinch off mechanism leading to relatively large crystalline particles. It was then
proposed that a prerequisite in controlling the antisolvent vapour precipitation process to
generate ultrafine particles is to prevent crystallization of the solute in the presence of
progressively increasing antisolvent concentration within the droplet.
Jiunn Yuan Tan (21637520) Page 25
Therefore, it was hypothesized that dissolved polymeric materials might have a
higher potential to precipitate into ultrafine uniform particles as they do not exhibit strong
crystallization behaviour. In past experiments, only the disaccharide (lactose) was
investigated. There is a need to examine the feasibility of the antisolvent vapour process for
polymeric materials. Polymeric materials have many applications, as a large range of
pharmaceutical and food products exist in the form of synthetic or natural polymers with
varying chain length. Polymers structure, molecular weight, linearity, intra- and
intermolecular interactions determine its thermal, physical and mechanical properties 8.
These are crucial in producing cellulose-based polymers, hydrocolloids and particularly
polymers for drug delivery applications. In the current work, maltodextrin was used as a
model polymer, as it represents a mixture of amorphous saccharides with broad molecular
weight distribution (i.e. varying chain length), with a wide range of application based on its
hydrolyzed polymer chain length. Maltodextrin is also widely used in the food and
pharmaceutical industry for functionalities such as dispersing aid, flavour carrier, bulking
agent, viscosifier and fat replacement .
The first part of the study examined the behaviour of maltodextrin in the ethanol
AVP method and the particle structures obtained at different ethanol vapour absorption
rates and initial weight concentration. Further experiments with different polymeric chain
length coupled with an in-situ observation of the precipitation process shed more light on
the fundamental observation and mechanism of the AVP process.
Jiunn Yuan Tan (21637520) Page 26
1.2 Drying and Analysis of the Mass Change Profiles of Different
Maltodextrin Particles using the Modified Single Droplet Rig
The drying of maltodextrin under AVP resulted in three different particle structures:
porous, microsphere network and microsphere. It is of interest to understand the drying
behaviour for the formation of these three particle structures. Experiments on maltodextrin
were conducted using the modified single droplet rig and the mass change of the droplet
was recorded throughout the drying process. The modified single droplet rig utilises a new
design of the vapour generation system which is more efficient and controllable for accurate
vapour generation in order to obtain the desired particle structures.
There was a need to build and design a new vapour generation system as most of
the available vapour generation techniques are vapour specific and limited to low vapour
concentration generation, typically in the ppm concentration range. In some other cases,
the control method is only limited to water vapour and lacks continuous control and
monitoring system. In other words, the operating conditions are pre-calibrated to produce
the desired humidity. This pre-calibrated approach is of limited use for applications where
the humidity needs to be ramped up or ramped down. In addition, constant monitoring and
control of process conditions is required to improve the safety and accuracy of vapour
generation, particularly for highly flammable vapours. Typically, vapour humidity
measurement is conducted using a vapour specific sensor such as hygrometer, alcohol
breathalyser or acetone sensor. Some applications make use of optical gas sensing for
detection of low concentration gasses 9. Vapour saturators require pre-calibrated conditions
to produce the desired vapour humidity 10. Due the limitations in the pre-existing laboratory
vapour generators, it was of interest to design a controllable and more versatile vapour
Jiunn Yuan Tan (21637520) Page 27
generation system. In this work, a simple temperature controller system was designed to
control humidity of the resultant gas mixture for a lab-scale vapour generator. The modified
humidity measurement technique introduced in this work allowed continuous humidity
measurement of various binary gas mixtures. The control principle applied in this design can
be extended to interchanging different gas vapours humidity applications. This new system
generates vapour through evaporation of liquid. The LabView control system incorporates
volume, relative humidity (ERH), absolute humidity (EAH) and temperature control.
The second part of this study examines the drying profile of each maltodextrin
droplet resulting in different particle structures. Further quantitative analysis of the ethanol
concentration within the droplet for three different particle structures will provide a better
understanding of particle structure formation in the AVP process.
1.3 Modelling of the AVP Drying Process
Understanding the effect of various drying conditions on the drying behaviour of the
atomised droplets is of great importance to optimize a spray drying process. Typically, the
drying behaviour of a single droplet is modelled through the single droplet drying
experiment to provide a fundamental basis for further implementation in an actual spray
dryer. Previous modelling work mainly focuses on the evaporation of water from the droplet
11. The resultant particle structures are affected by the absorption rate of ethanol and the
maximum concentration of ethanol within the droplet which are dependent on the ethanol
humidity in the convective drying stream 7. In order to have a better insight of this unique
precipitation mechanism, it is of interest to develop a simultaneous absorption and
evaporation model within the droplet. This model will provide a good representation of the
mechanism of AVP.
Jiunn Yuan Tan (21637520) Page 28
The final aim of this work is to investigate the feasibility of developing a
simultaneous absorption and evaporation model into a droplet using the Raoult’s Law and
UNIFAC equations. By measuring the mass change of droplet with time using the single
droplet experiment, the theoretical model was then compared to the experimental results.
The newly built vapour generator, incorporating the LabView control system is use to
generate ethanol vapour efficiently at adjustable ethanol humidity conditions. The present
study examined the mass change profile and temperature profile of a liquid droplet at
different conditions. Further analysis is conducted to compare these experimental results
with the theoretical model.
1.4 Research Aim
Utilising polymer based material with various compositions dried under different
drying conditions by the single droplet drying rig under the novel Antisolvent Vapour
Precipitation (AVP) approach, the general aim of this research is to gain a fundamental
understanding of the mechanism of the simultaneous absorption and evaporation in the
AVP process.
The specific research aims are:
1. To investigate the precipitation behaviour of polymer-based material under varying
conditions using the AVP approach. Polymeric materials were chosen because it will
precipitate even better, as it negates the possible crystallisation process, which was
identified as the key process impeding the precipitation of microspheres.
2. To analyse the drying profile and the effect of ethanol concentration of a droplet on
the formation of different maltodextrin particle structures under AVP.
Jiunn Yuan Tan (21637520) Page 29
3. To investigate the feasibility of developing a model using heat and mass transfer
considerations, Raoult's Law and UNIFAC equation for the simultaneous absorption
and evaporation process within the droplet during AVP drying.
Jiunn Yuan Tan (21637520) Page 30
Chapter 2
LITERATURE REVIEW
2.1 Spray Drying
Spray drying is the most commonly used method in the pharmaceutical industry to
produce particles of desired morphologies and functionalities from solutions and
suspensions 12. It has also been widely used for microencapsulation of food ingredients,
vaccines and microbial cells 13. The five main steps that constitutes a spray drying process
are feedstock concentration, atomization of feedstock solutions into small droplets, droplet-
air contact, droplet drying and product separation 14. Figure 2.1 shows the schematic
diagram of a typical spray drying process. Three crucial elements that affect the design and
outcome of a spray dryer are the atomizer, spray drying chamber and air flow 14. The
specifications of these elements are designed to cater for the desired product properties of
a spray drying process 1. The air flow conditions and the design of a spray drying chambers
are mainly determined through rigorous simulations and mathematical modelling work for a
particular spray drying process 15.
Jiunn Yuan Tan (21637520) Page 31
FeedstockConcentration
Atomization of Feedstock
Feed Flow
Droplet-air Contact &
Droplet Drying
Product Separation
Heater
Cyclone
Collection Vessel
Spray Dryer
Figure 2.1: Schematic Diagram of a Typical Spray Drying Process.
Atomization of the feedstock solutions is the key step in a spray drying process to
produce sufficiently fine droplet, with high surface/mass ratio for the optimum evaporation
of solvent to take place in a spray dryer 2. In a typical spray dryer, each atomized droplet is
dried to form a single particle. The bigger the atomized droplet, the bigger the particle
produced 16. The quality and morphology of the resultant particles are greatly influenced by
the atomizer and spray method 17. For an ideal atomization process, all the droplet should
have the same size, which would result in a uniform particle size. However, in practice it is
satisfactory if a narrow particle size distribution is achieved 18. The range of mean particle
sizes produced by conventional atomizers such as rotary wheel, pressure nozzle and two-
fluid nozzle used in spray dryers are 10-100 µm, 30-200 µm and 3-75 µm respectively 2.
Commercially available atomization system are mainly classified based on the type of energy
employed to generate the spray: pressure nozzles, centrifugal atomizers, kinetic energy
nozzles and sonic energy atomizers 19. Higher energy results in lower surface tension and
Air
Jiunn Yuan Tan (21637520) Page 32
viscosity of the feed; hence, forming smaller droplet 18. Table 2.1 shows the range of particle
sizes of different commercial products produced via spray drying. Evidently, the particle size
and uniformity of products produced through spray drying is limited by the atomization
process. Therefore, other avenues such as the liquid antisolvent (LAS) and supercritical
antisolvent (SAS) precipitation method are investigated.
Table 2.1: Range of Particle Sizes Obtained in Spray Dryers of Different Products 20.
Products Particle Sizes (µm)
Milk 30-250
Coffee 80-400
Pigments 10-200
Ceramics 30-200
Pharmaceutics 5-50
Chemicals 10-1000
2.2 Precipitation Method
2.2.1 Liquid Antisolvent (LAS) Precipitation
Liquid Antisolvent Precipitation (LAS) is a widely used precipitation method that
takes advantage of the rapid and high supersaturation due to the addition of liquid
antisolvent 21. It is mainly coupled with spray drying to produce various active
pharmaceutical ingredients (APIs) such as atorvastatin calcium, amphotericin B and other
poorly water soluble APIs 22. Zu et. al. has also reported that the solubility, antioxidant
ability and bioavailability of taxifolin can be improved via LAS. In addition, LAS application
has also been extended to produce nanoparticles for the food industry and nanocomposites
for batteries 23. The mechanism of LAS precipitation consists of 21:
Mixing of solution and antisolvent, generation of supersaturation.
Nucleation and growth by coagulation and condensation.
Agglomeration in case of uncontrolled growth.
Jiunn Yuan Tan (21637520) Page 33
The mixing of solution and antisolvent is shown to be an effective way of controlling
the particle size. Usage of enhanced mixing such as ultrasound was shown to be able to
reduce the particle size. This is because enhanced mixing decreases the mixing time which
increases the nucleation rate and resulted in smaller particles 24. Other important
parameters to control the particle size distribution and stability of the growth process are
the addition of solute to antisolvent, antisolvent to solvent ratio, compound concentration,
temperature, antisolvent selection and stabilizers 23a. Generally, increasing the addition of
solute rate, increasing the ratio of antisolvent to solvent, reducing the compound
concentration, decreasing temperature and appropriate selection of antisolvent as well as
stabilizers were found to produce smaller particle size and a more stable process 25.
Evidently, the control of LAS precipitation is complicated and the usage of liquid antisolvent
and stabilizers will require complex post separation steps.
2.2.2 Supercritical Antisolvent (SAS) Precipitation
Supercritical Antisolvent (SAS) precipitation is a micronization method used in the
food and pharmaceutical industry to produce particles with various functionalities 26. A
more general application of the micronization via the SAS process is compiled and reviewed
by Reverchon 27. The SAS precipitation is a complex mechanism that is the result of an
interplay of fluid dynamics, mass transfer, precipitation kinetics and thermodynamics which
is still not well understood 28. SAS precipitation primarily incorporates the use of
supercritical CO2 gas as the antisolvent. Solute material is firstly dissolved into the solvent,
and the mixture is then atomized into a pressurized tank where it is exposed to the
supercritical CO2. Reviews have been conducted on the various applications and
technologies available for the SAS precipitation 29. Figure 2.2 shows a schematic diagram of
a typical SAS apparatus set-up.
Jiunn Yuan Tan (21637520) Page 34
Figure 2.2: Schematic diagram of SAS apparatus 30.
There are many factors affecting the particle size in the SAS process: pressure,
temperature, concentration, chemical composition of organic solvent, chemical composition
of solute, nozzle geometry and flow rates of CO2 and liquid phase. The effects of these
paramaters are dependent on the type of feedstock and other operating conditions within a
SAS process 26a. The use of impinging jets to improve the mixing of a SAS process was found
to be able to reduce the particle size 31. Arnaud et. al. proposed a 3-D simulation accounting
for the hydrodynamics, phase equilibrium, crystallization kinetics and mass transfer of a SAS
process that is able to predict the particle size formed from the crytallisation process 32.
Besides that, the particle morphology is also tunable by altering the conditions
within the SAS process. Generally, the final particle morphology is dependent on the
interplay between the phase equilibria, jet fluid dynamics and mass transfer during the SAS
Jiunn Yuan Tan (21637520) Page 35
process 28, 33. It was reported by Braeuer et. al. that there are two ways to switch between
precipitation of microparticles, nanoparticles or expanded nanoparticles which is by
increasing the pressure to above the mixture critical pressure (MCP) or by operating above
the MCP but increasing the solute concentration 28. Reverchon et. al. studied the influence
of pressure, temperature and concentration on the mechanism of SAS particle precipitation.
As pressure increased, expanded microparticles, microparticles and nanoparticles were
produced in sequence while the effect of increasing concentration had the reverse effect 34.
The use of supercritical fluid in the SAS process requires high pressure condition which
increases the technology cost and requires thorough safety considerations.
LAS has been shown to require complex post separation due to the addition of
stabilizers to control the precipitation process while SAS is a technique which requires high
pressure to induce the supercritical conditions for the precipitation process. Therefore, an
alternative precipitation method which utilises antisolvent vapour instead of liquid
antisolvent and operates under atmospheric conditions is proposed.
2.3 Antisolvent Vapour Precipitation (AVP)
Recently, a novel Antisolvent Vapour Precipitation (AVP) was found to be able to
produce multiple micron range spherical particles within a single droplet at atmospheric
conditions 35. Therefore, the particle size produced is not dependent on the initial size of the
droplet. Contrary to the LAS precipitation, this method utilizes vapour antisolvent instead of
liquid. Previous work have shown that AVP is able to produce microspherical particles from
lactose and whey protein 7, 36. The microparticles produced has been reported to be able to
encapsulate oil 37.
Jiunn Yuan Tan (21637520) Page 36
The physical phenomenon for this method of precipitation involves the simultaneous
absorption of antisolvent (ethanol) into the droplet and evaporation of solvent (water) from
the droplet which led to rapid precipitation and dehydration of droplet 35. Figure 2.3 depicts
the physical phenomenon of the AVP process. Instead of being mixed in the fluid phase, the
antisolvent for this technique is in the gas stream and diffuses into the droplet. The inward
absorption of ethanol was caused by the internal gradient between the surface of the
ethanol absorbed droplet surface and the centre of the droplet. 35.
Initially, the droplet gained mass due to the absorption of ethanol. Subsequently, at
the point of ethanol saturation, the droplet underwent mass loss due to the continuous
evaporation of water. It was speculated by Mansouri et. al. that the loss of water would
induce a progressively higher solute concentration, that might have accelerated the
precipitation process. By analysing the solubility of lactose at maximum ethanol
concentration in the droplet, it was suggested that the precipitation might not be solely
attributed by supersaturation-based mechanism, as the entire period of the initial increase
in droplet mass was below or at most slightly above the supersaturation of lactose 35.
Jiunn Yuan Tan (21637520) Page 37
Figure 2.3: Schematic diagram of the physical phenomenon of AVP 6.
In another work, Mansouri et. al. 7 were able to precipitate smooth amorphous or
pollen structure microparticles which are composed of straight needle-like or short
entwined dendrites using the same convective antisolvent and dehydration method but
varying the initial concentration of lactose solution and relative humidity ethanol. The
regions in which smooth amorphous or pollen structures microparticles were formed was
mapped onto a 3-D graph of initial solute weight concentration, ethanol relative humidity
and ethanol absolute humidity. There appeared to be a 'pinched off' mechanism, where the
microspheres seemed to be ‘pinched’ out from the network structure. This smooth
amorphous microparticles precipitation mechanism was further elucidated using simple
interfacial energy analysis suggesting that a surface tension particle size gradient
Jiunn Yuan Tan (21637520) Page 38
mechanism might possibly be the dominant factor leading to the narrow size distribution of
the microspheres 7. A unique observation in the previous report is that the occurrence of
any crystallization within the droplet, under certain operating condition, will negate the
pinch off mechanism leading to relatively large crystalline particles. It was then proposed
that a prerequisite for controlling the antisolvent vapour precipitation process to generate
ultrafine particles is to first prevent crystallization of the solute in the presence of
progressively increasing antisolvent within the droplet. Therefore, it was hypothesized that
dissolved polymeric materials might offer a greater potential to precipitate into ultrafine
uniform particles precipitate into as it does not exhibit strong crystallization behaviour. In
past experiments, only the disaccharide (lactose) was investigated. There is need to
examine the feasibility of the antisolvent vapour process for polymeric materials. Mansouri
et. al. extended their research to investigate the effect of pH on the precipitation of whey,
lactose and composite whey-lactose. Interestingly, segregated precipitation was found for
the composite whey-lactose precipitation and precipitation of semi-uniform whey particles
occurred close to its isoelectric pH 36. It is evident from previous work that antisolvent
humidity is one of the key parameters controlling particle precipitation in the AVP process.
Therefore, it is of interest to explore the available vapour generation methods and humidity
measurement techniques.
Jiunn Yuan Tan (21637520) Page 39
2.4 Current Vapour Generation and Humidity Measurement Technique
2.4.1 Vapour Generation Method
There is an emerging need for different vapour gas mixture for various laboratory
applications. In the AVP process, ethanol vapour is used as an antisolvent for efficient drying
to obtain desired particle morphology 6. Acetone vapour is used for metal-organic
frameworks application while toluene vapour is used to induce structural organization in
syndiotactic polystyrene film 10, 38. Therefore, it is of interest to establish general and
versatile humidity control principles to cater for different type of gas vapours applications.
Organic vapours such as ethanol and acetone have a wide-range of applications; however,
their high flammability and volatility pose high safety risks. Moreover, different laboratory
work is sensitive to different ranges of humidity. For instance, variations in the of ethanol
vapour produces different types of particle morphology when used during drying 7. Hence, it
is crucial to incorporate continuous humidity control and monitoring within any vapour
generation system to ensure a safe and accurate vapour humidity generation. Table2.2
shows the list of methods used to generate vapours for lab-scale applications. The dynamic
vapour generator concept was adopted with modification to be applied for our application.
Jiunn Yuan Tan (21637520) Page 40
Table 2.2: List of Methods Used to Generate Vapours for Laboratory Applications.
Method
Vapour
Generation
Range
Principle Gap
Vapour
saturator
10
0.03 - 0.2 kg/kg
db
- Nitrogen gas as bulk flow
saturated with acetone by
changing the temperature or
pressure in the saturator
(chamber filled with liquid
vapour with changeable
pressure and temperature).
- The control of vapour
humidity is not
continuously monitored
and controlled.
- The saturator
conditions are pre-
calibrated.
Bubbling
through liquid
vapour
7
0 - 100%
0 - 0.15 kg/kg
db
- Nitrogen gas is bubbled
through two conical flasks.
- The control of vapour
humidity is not
continuously monitored
and controlled.
- Inconsistency in
bubbling.
Hybrid humidity
generator
(HHG)
39
2.5μ mol/mol
and 0.57
mol/mol
- Air is dried and saturated
with water by altering the
temperature and pressure in
the saturator (chamber filled
with liquid vapour with
- Only applicable to
water.
Jiunn Yuan Tan (21637520) Page 41
changeable pressure and
temperature).
Dynamic vapour
generator
40
Ppb - low ppm - Vapour is produced
through evaporation and
driven by air supply.
- Low concentration of
vapour.
FI-vapour
generation
system
41
- - Cross-flow nebulizer with
standard Scott type spray
chamber is used to generate
vapour and sample injection
is used to optimize vapour
generation.
- Have a detection limit.
- The control of vapour
concentration is not
continuously monitored
and controlled.
Vapour
evaporation
chamber
42
10-200 ppm - Pre-determined
concentrations of liquid is
injected into the chamber
and evaporated to achieve
the desired vapour
conditions in the chamber.
- Batch process (non-
continuous)
Humidity
generator
43
0 - 100%
0 - 4 (kg/kg db)
- Water is evaporated by
heating and driven by dry
air.
- Only can be used for
water.
PAH vapour
generator
44
0.3 - 30 ppbv - Dual stream flow of
nitrogen gas through vapour
generation chamber and
- The control of vapour
concentration is not
continuously monitored
Jiunn Yuan Tan (21637520) Page 42
purified atmospheric air.
- Solid sublimated via water
bath heating to produce
vapour, swept by the
nitrogen gas.
and controlled.
- Low concentration of
vapour.
Liquid injection
vapour
generator
(LIVG)
45
Ppb - ppt - Liquid vapour is injected
with syringe needle through
cartridge heaters and driven
through by air.
- Low concentration of
vapour.
Dioxin vapour
generator
46
3-100 ppm - Dual stream through 2
glass vessels for generation
and dilution.
- Solid sublimated via
heating to produce vapour
and swept by the inert gas
flow.
- Low concentration of
vapour generated.
- The control of vapour
concentration is not
continuously monitored
and controlled.
Jiunn Yuan Tan (21637520) Page 43
Figure 2.4: Sketch of a lab-scale humidity generator 43.
Jiunn Yuan Tan (21637520) Page 44
2.4.2 Vapour Humidity Measurement Technique
Humidity is an important parameter in many laboratory applications. Efforts have
been put into improving the current humidity sensing devices, particularly in the case of
water vapour 47. In the case of other vapours, many studies have been focusing on
improving the gas sensing of various vapours 9, 48. Sophisticated system and platform has
been design to characterize and evaluate these sensors 49. However, these gas sensors are
usually limited to low vapour concentrations in the range of ppm. For higher humidity
vapour generation, some applications utilise saturators which requires pre-calibration of
temperature and pressure to achieve the desired humidity 10, 39. Commercially available
vapour sensor such as hygrometer, alcohol breathalyzer and gas detectors are generally
vapour specific.
Figure 2.5: Images of hygrometer, alcohol breathalyser and gas detector (left to right).
The concept of obtaining vapour humidity through the measurement of wet bulb
(WB) and dry bulb (DB) temperature has been widely established 50. This conventional
method uses a psychrometer, which is vapour specific or reading off the values based on a
psychrometric chart, which is done manually. Fundamentally, the generation of
psychrometric chart for different vapours are done by adopting a simple mass and energy
Jiunn Yuan Tan (21637520) Page 45
balance analysis 51. Many factors such as the vapour velocity, wick contamination and heat
radiation can affect the accuracy of the humidity measurement due to variation in wet bulb
temperature 52. By analysing the relationship between wet bulb, dry bulb and vapour
properties with respect to the humidity fundamentally, the measurement of humidity
through the wet bulb and dry bulb temperature can be conducted more accurately.
The use of an efficient vapour generation system and accurate humidity
measurement technique are crucial in order to accurately model the drying behaviour of the
droplet under AVP conditions.
2.5 Modelling the Drying of a Droplet
Drying process is a combination of heat and mass transfer process which have been
explained by many researchers 53 . Various drying models have been developed for different
applications to gain a better insight into the drying mechanism for different products and
techniques 15c, 54. Some of these models are incorporated in large scale commercial spray
dryer models 55. Ongoing research has been devoted to modelling the drying profile of a
droplet as well as evaluating the different behaviours observed within the droplet during
the drying process 56. This allows a better understanding of the effect of various process
conditions on the atomised droplets during spray drying 57. Having such understanding
allows for optimisation and better control of a spray drying process. The study of droplet
drying is typically classified into three categories: evaporation of pure droplet 58, drying of
droplet containing soluble material 59 and drying of droplet containing insoluble material 59b,
60.
The use of ethanol vapour has been gaining much attention for various applications,
mainly for post-treatment of organic and inorganic materials 61. Only recently the use of
Jiunn Yuan Tan (21637520) Page 46
ethanol vapour has been applied in the field of drying through the AVP process, which
involves a simultaneous absorption and evaporation phenomenon 6. Understandably, there
is no literature available that explains the absorption of ethanol vapour into a water droplet.
Nevertheless, many studies have been conducted to evaluate the vapour-liquid equilibrium
(VLE) of an ethanol + water system 62, mainly to model operations within a distillation
column 63. Based on fundamental heat and mass transfer considerations 64, the
simultaneous absorption and evaporation process can be modelled. Drawing analogy from a
gas particle partitioning system 65, vapour-liquid equilibrium evaluation based on Raoult’s
Law and UNIFAC equation is used to account for the interaction between the ethanol and
water mixture as well as the mixture deviation from ideal condition. With that in mind, a
review was undertaken covering the many different aspects of the modelling approach
reported in the literature.
2.5.1 Lump Versus Distributed Model
A drying process can be evaluated based on a lump model 66 or a distributed model
67. A lump model evaluates the drying behaviour of a droplet by lumping together several
parameters that may affect the drying process. On the other hand, the distributed model
attempts to evaluate the effect of these drying factors separately. Some modelling work
utilises a combination of both of these approaches to provide an economical yet accurate
analytical solution 68. An example of a lump model is the well-established D2 law (Equation
2.1) that can be used to describe the mass transfer of evaporating droplets under still
conditions. However, this approach is only applicable to non-convective ambient condition
69. In most water droplet evaporation applications, such non-convective ambient condition
is rarely encountered. Nevertheless, the lump model approach is the preferred method as it
is simpler and time efficient.
Jiunn Yuan Tan (21637520) Page 47
2.5.2 Effect of Solute on Droplet Drying
In a spray drying application, the droplets contain some dissolved or undissolved
solids. The crust formation of these solid particles with the solvent within the droplet affects
the evaporation process. Liquid diffusion, vapour diffusion, hydrodynamic flow and other
mass transfer mechanisms represent an overall mass transport of water in the material 70.
There are four different approaches typically used to describe the drying of the droplet
containing solid: study based on the characteristic drying curve (CDC) 71, study based on the
formation of crust 59b, study based on receding interface model 60 and reaction engineering
approach 72. In some cases where the particles are precipitating out and there is no crust
formation, the evaporation of the droplet containing solids can be approximated to the
evaporation of a pure liquid droplet 59b, 73. Typically, the concentration of solute in the
droplet determines its significance in affecting the drying process. The typical drying stages
of liquid droplet containing solids are shown in Figure 2.6. The initial droplet is first heated
to the wet bulb temperature for droplet with low solids concentration. It then experiences
shrinkage due to the loss of water through evaporation. Depending on the solid material
within the droplet, some solids will form an outer crust layer during drying. As drying
continues, the formation of crust progresses until the entire droplet is dried. The final dried
droplet will eventually reach a final gas dry-bulb temperature.
Jiunn Yuan Tan (21637520) Page 48
Initial Droplet
Droplet Heated to Wet Bulb
Temperature
Drying with Crust Formation
End of Drying
Sensible Heat of Dried Droplet
Sensible Heat
ShrinkageDrying with
Droplet Shrinkage
Crust ProgressingSensible
Heat
Figure 2.6: Drying stages of a liquid droplet containing solid.
2.5.3 Mass Depression Phenomenon
Other studies have been conducted to produce an analytical expression for the
evaporation of droplet under different conditions 74. Under the convective regime with
temperature up to 200oC and 0<Re<200, the heat and mass transfer correlations are shown
in Equation 2.2 and 2.3. 58b. Other studies have also shown that when the mass transfer flux
is high, the heat and mass transfer boundary layer might be significantly distorted 58a, 75.
Table 2.3 shows a list of mass transfer coefficient expressions under different droplet
evaporation conditions.
Dry-bulb gas
temperature
Jiunn Yuan Tan (21637520) Page 49
Table 2.3: List of Typical Mass Transfer Coefficient Expressions in Literatures for Different
Droplet Evaporation Conditions.
Correlation Conditions
800 K
10 < Re < 2000
76
324 – 502 K
Turbulent flow
0.5 – 10 MPa
77
27 – 340oC
24 < Re < 325
58a
0 < Re < 1000
78
30 – 95oC
2 < Re < 2631
11
Room temperature
2 < Re < 600
79
Air Jet
38oC
Jiunn Yuan Tan (21637520) Page 50
65 < Re < 320
80
16 mm < Diameter < 38 mm
Room temperature
100 < Re < 1050
81
Free fall
Wind tunnel
82
-40 – 750oC
25 < Re < 625
83
296 – 364 K
32 < Re < 328
0.6 < Sc < 1.66
84
Free fall
5 – 20oC
0.1 mm < Diameter < 0.4 mm
85
Jiunn Yuan Tan (21637520) Page 51
2.5.4 Multicomponent Modelling
Raoult's Law is a thermodynamic law describing the vapour-liquid equilibrium (VLE)
of a binary liquid mixture based on the vapour pressure of each component 86. For an ideal
solution, the Raoult's Law can be mathematically expressed as:
where is the partial vapour pressure of component A
is the vapour pressure of the pure component A
is the mole fraction of component A in the mixture
In the case of non-ideal solutions, Raoult's Law can be modified by taking into
consideration the fugacity coefficient and the activity coefficient. Fugacity coefficient
accounts for the deviation of gas from the ideal gas law while the activity coefficient is used
to account for the interactions in the liquid phase between the different molecules 86. Well-
developed model such as UNIFAC, UNIQUAC, NRTL and Wilson are selectively used to
calculate the activity coefficient, of a liquid mixture.
The dependence of activity coefficient of liquid mixtures on temperature and
composition can be described using different well-developed activity coefficient models.
Selection of the type of model used is dependent on the type of mixtures. For electrolyte
solutions, Davies equation 87, Pitzer equation 88, TCPC model 89 and SIT theory 90 may be
used. For non-electrolyte solutions, activity coefficient models such as UNIQUAC 91, Wilson
92 and NRTL 93 are Gibbs energy models obtained by fitting the parameters simultaneously
Jiunn Yuan Tan (21637520) Page 52
to binary experimental data. In the case where no experimental data is available, group
contribution methods such as UNIFAC, PSRK and ASOG are used to predict the required
activity coefficients for components at a specific temperature and composition 94.
The UNIFAC activity coefficient model is an extension of the UNIQUAC model using
group interaction parameters obtained from data reduction to predict activity coefficients
of binary and multicomponent non-electrolyte liquid mixtures 95. It has been used to predict
the activity coefficients of heavy alkanes and light gases 96, alkane-alcohol system 97 and
fundamental biochemicals in water 98. Due to its popularity, more and more group
interactions parameter has been presented over the years in an effort to extend its
applicability to different mixture systems 99. A comprehensive description of the theory
behind the UNIFAC model on VLE has been explained by Fredenslund et. al. 100. With some
modification, the application of UNIFAC model has also been extended to predicting high
pressure and temperature VLE 101, VLE of reactive system 102, gas solubilities at high and low
pressures 103 as well as VLE and VLLE of binary and ternary mixtures 104. The versatility and
proven reliability of the UNIFAC model makes it a suitable model to predict the activity
coefficients of most systems.
2.5.5 Modelling Absorption
The absorption of vapour into liquid usually involves two processes: the
condensation of vapour on the droplet and the diffusion of the condensed vapour into the
liquid 105. Analogous to the evaporation of a droplet, the vapour absorption can be
described using the reverse of the mass transfer equation for an evaporation of a droplet.
The instantaneous concentration of absorbed vapour is determined by the combined effect
of diffusion transfer of vapour into the droplet and convective transfer away from the
Jiunn Yuan Tan (21637520) Page 53
droplet 106. However, the interaction between the condensed vapour and liquid has to be
considered as it exists as a non-ideal liquid mixtures 107. A comprehensive study on the
absorption of SO2 into an evaporating water droplet has been presented by Huckaby and
Ray 105a. From hindsight, the VLE of the absorbed vapour and liquid droplet system can be
described by Raoult's Law and an activity coefficient model can be adopted to account for
the interaction between the two components.
2.6 Summary and Remarks
The ability of a spray dryer to produce sufficiently small and uniform particle size is
limited by its atomizer. Therefore, it is of interest to investigate other precipitation avenues
such as LAS and SAS. LAS has been shown to require complex post separation due to the
addition of stabilizers to control the precipitation process while SAS is a technique which
requires high pressure to induce the supercritical conditions for the precipitation process.
Recently, a novel precipitation method known as AVP which in this case utilises ethanol
vapour as an antisolvent instead of liquid has been discovered. Multiple lactose and WPI
microspherical particles in the micron range were produced within a single droplet using
this technique under atmospheric conditions. The physical phenomenon of this precipitation
technique involves simultaneous absorption and evaporation of antisolvent vapour within
the liquid droplet. 7. A unique observation in the previous report is that the occurrence of
any crystallization within the droplet, under certain operating condition, will negate the
pinch off mechanism leading to relatively large crystalline particles. It was then proposed
that a prerequisite in controlling the antisolvent vapour precipitation process to generate
ultrafine particles is to firstly prevent crystallization of the solute in the presence of
progressively increasing antisolvent within the droplet. Therefore, it was hypothesized
Jiunn Yuan Tan (21637520) Page 54
dissolved polymeric materials might offer a higher potential to precipitate into the ultrafine
uniform particles precipitate into as it does not exhibit strong crystallization behaviour. In
past experiments, disaccharide (lactose) was investigated. There is a need to examine the
feasibility of the antisolvent vapour process for polymeric materials. Previous work has also
revealed a 'pinched off' mechanism for this process and also highlights the effect of the
ethanol humidity on the produced particle structure which indicates that one of the key
parameter controlling this process could possibly be the absorption behaviour of the
antisolvent.
Therefore, to better understand this unique process, it is of interest to model the
simultaneous absorption and evaporation of ethanol vapour phenomenon. In order to do
that, an efficient vapour generation system and accurate humidity measurement technique
are required to generate reliable experimental data. Most vapour generation systems are
limited to low concentration vapour or requires pre-calibration of the system conditions
while humidity measurement devices are generally vapour specific. This creates the need to
design a more controllable vapour generation system and establish a more general vapour
measurement technique.
Theoretically, the drying model of a droplet can be evaluated using fundamental
heat and mass transfer equations to describe the absorption and evaporation process
independently. The heat and more significantly the mass transfer correlation were found to
deviate depending on the system and conditions of the process. Raoult's Law can be used to
obtain the vapour-liquid equilibrium properties of the water-ethanol system while the
UNIFAC equation can be adopted to account for the deviation of the ethanol-water system
from ideal solutions and the interaction between ethanol and water. The versatility and
Jiunn Yuan Tan (21637520) Page 55
proven reliability of the UNIFAC model makes it a suitable model to predict the activity
coefficients of the water-ethanol system. The detail mechanism and modelling of the
physical phenomenon of the AVP process will be further elucidated in subsequent chapters.
Jiunn Yuan Tan (21637520) Page 56
Chapter 3
MATERIALS AND METHODS
3.1 Materials
3.1.1 Preparation of Maltodextrin/Maltose Samples
Maltodextrins and maltose are products derived from acid and enzymatic hydrolysis
of starch. The degree of hydrolysis are described in terms of their 'dextrose equivalent' (DE)
value. The DE value is inversely proportional to the polymer chain length and hence, the
molecular weight. Maltodextrin DE 10 (F03220, The Melbourne Food Ingredient Depot),
maltodextrin DE 18 (F03380, The Melbourne Food Ingredient Depot) and maltose (M-5885,
Sigma Chemical Company) solutions at concentrations of 2.5 wt%, 5 wt%, 10 wt% and 15 wt%
were prepared by dissolving tapioca maltodextrin powder, maltodextrin powder and
maltose hydrate grade 1, respectively, in Mili-Q water.
3.1.2 Solvent and Antisolvent
Pure Mili-Q water is used as the solvent as well as sample of pure water. Liquid
ethanol, 99% purity (100983, Merck Millipore) is used as the antisolvent.
3.2 Original AVP Single Droplet Drying Experiment
3.2.1 Experimental Method
A detailed explanation of the experimental set-up and working principle of the single
droplet drying technique is provided by Lin and Chen 108. This technique was further
Jiunn Yuan Tan (21637520) Page 57
modified by Mansouri et. al. to incorporate the AVP process into the single droplet drying 6.
Brief details are given here for completeness. The schematic diagram of the single droplet
rig used in this experiment is shown in Figure 3.1. A standard initial single droplet size of 2
µL was generated using a 5 µL gas chromatograph micro syringe (5FX, Part # 001100, SGE
Analytical Science Pty Ltd, Australia) and suspended onto a glass filament positioned in the
drying chamber using a separate transferring glass filament. When generating and
transferring the droplet, a bypass barrier plate was used to divert the conditioned nitrogen-
ethanol stream away, in order to minimize droplet evaporation before the monitoring
begins. The bulk of the convective medium was supplied by compressed nitrogen, bubbled
through two conical flasks connected in series filled with. The ethanol vapour and nitrogen
mixture was pre-heated within a heating coil submerged in a water bath held at 70oC. It is
important to note that the resultant ethanol vapour entering the chamber was
approximately 30oC due to heat loss in transit. Concentration of ethanol vapour was
controlled by adjusting the volume (level) of ethanol in the conical flasks. Video monitoring
was used to track and record the visual changes within the droplet with time during drying.
The dried product was scraped onto a carbon stub for scanning electron microscopy
(PhenomTM SEM) imaging and the sample was coated with gold/paladium (Sputter Coater
Quorum SC77620).
Experiments were repeated under five different ethanol relative humidity and three
different initial solute weight concentrations. Each drying conditions were repeated three
times. The resulting particle morphology observed under SEM was recorded and collated for
all drying runs.
Jiunn Yuan Tan (21637520) Page 58
Figure 3.1: Schematic figure of original experimental set-up for single-droplet glass filament
drying system using ethanol vapor as antisolvent. (a) Nitrogen tank; (b) Ethanol chambers; (c)
Water bath; (d) Drying chamber; (e) Suspending glass filament; (f) Camcorder.
3.2.2 Ethanol Vapour Humidity Measurement
There are various methods used to measure vapour humidity or concentration. In
some devices, the vapour generation device is pre-calibrated for temperature and pressure
based on the vapour-liquid properties to obtain the desire humidity 10. Others required a
batch process set-up where the experimental chamber is pre-conditioned to achieve the
desired humidity conditions 42. The concept of obtaining vapour humidity through the
measurement of wet bulb (WB) and dry bulb (DB) temperature has been widely established.
This conventional method uses a psychrometer, which is vapour specific or reading off the
values based on a psychrometric chart, which is done manually.
The wet bulb and dry bulb measurements are taken using a humidity box
configuration, where the gas mixture is allowed to bleed-off a confined tube, with two
mounted temperature sensors. This is a simple and continuous way of obtaining the vapour
(b)
(a)
(c) (d)
(e)
(f)
Jiunn Yuan Tan (21637520) Page 59
humidity. Theoretically, the WB temperature is the measurement of nitrogen gas
temperature if it were cooled to saturation (100% vapour relative humidity) and the DB
temperature is the measurement of the freely exposed nitrogen –vapour gas temperature,
shielded from external radiation and moisture. While the DB temperature measurement is
simply the measurement of the temperature of the ethanol-nitrogen gas mixture via first
temperature sensor, the WB temperature was obtained by measuring the temperature of
liquid ethanol absorbing through a wick with one end dipped into a pure liquid ethanol
solution and the other end attached to the second temperature sensor. The premise of this
concept is the absorption of liquid vapour by the wick from the liquid reservoir, which
creates the saturation condition through the evaporation of vapour when the gas mixture
flows through the wick. The system takes approximately 10 minutes to reach steady-state.
The set-up for this method of measurement is shown in Figure 3.2. It is noteworthy that this
vapour humidity measurement concept can be used for any vapour of interest for lab-scale
application. This is done by simply changing the liquid reservoir to coincide with the desired
vapour type. Additional modifications on humidity model to account for different vapours
will be outlined in subsequent discussions.
101
TT
102
TT
Figure 3.2: Wet bulb and dry bulb measurement set-up at the bleed line.
Gas
mixture
in
Gas
mixture
out
LabView
thermocouple
sensor Thermocouple
Wick
Confined
tube
Liquid
reservoir Humidity
box
Jiunn Yuan Tan (21637520) Page 60
Part of the humidity measurement and control is in translating the WB and DB
transmitted signal to the AH and RH reading, this was computed adopting a simple mass
and energy balance analysis 51, with slight modifications. At equilibrium, the energy balance
for the wetted wick, incorporating convective heating and evaporative cooling takes the
form:
(1)
The heat and mass transfer coefficients can be expressed in terms of the Nusselt and
Sherwood numbers respectively:
(2)
(3)
Assuming cylindrical shape thermocouples, the Sherwood and Nusselt number can be
expressed in terms of Reynold, Schmidt and Prandtl numbers:
(5)
(6)
The equations for Re, Pr and Sc are:
(7)
(8)
(9)
Jiunn Yuan Tan (21637520) Page 61
Cs is obtained using the ideal gas law:
(10)
Antoine equation is used to calculate the Psat of the vapour.
Rearranging equation 1 and substituting Equations 2-10, the theoretical relationship
between WB temperature and DB temperature for zero vapour concentration (i.e. Ca = 0)
can be expressed as:
(11)
The heating of pure dry nitrogen stream over a range of temperature was conducted to
verify the WB to DB relationship when the concentration of vapour in the stream is zero.
Figure 3.3 shows the comparison between the experimental and theoretical values of WB
and DB temperatures.
Jiunn Yuan Tan (21637520) Page 62
0
5
10
15
20
25
30
35
0 1 2 3 4 5 6 7 8 9
Ta (0C)
Twb (oC)
Experimental Data
Linear (Model data)
0
5
10
15
20
25
30
35
40
0 2 4 6 8 10 12 14 16
Ta (0C)
Twb (oC)
Experimental data
Linear (Model data)
(a)
(b)
Jiunn Yuan Tan (21637520) Page 63
Figure 3.3: Comparison of theoretical and experimental WB and DB temperatures for (a)
ethanol; (b) water and (c) acetone.
The coefficients C1, C2, n1 and n2 are empirical coefficients depending on the system.
Assuming a cylindrical shape thermocouple, n1 and n2 are 0.466 109. In order to account for
the effect of high mass flux on the heat and mass transfer coefficients, the C values had to
be adjusted to fit the experimental data. C1 can be expressed in terms of the equation below,
similar to the concepts from other reports 76, 109:
(12)
C’ is dependent on the Reynold number and the shape of the thermocouple which is
assumed to be cylindrical. In our case, Reynolds number in the range of 2000-3000 was
obtained. Therefore, a value of 0.683 is used as suggested by Incropera and Dewitt.109.
According to Abramzon and Sirigano 110, Bm for a high mass flux system can be approximated
as:
(13)
0
5
10
15
20
25
30
35
40
-14 -12 -10 -8 -6 -4 -2 0
Ta (0C)
Twb (oC)
Experimental data
Linear (Model data)
(c)
Jiunn Yuan Tan (21637520) Page 64
(14)
(15)
In order for the model to fit the data, it was found that C2 values varied with respect to DB
temperature. Through the wet bulb and dry bulb temperatures calibration, a relationship of
C2 in terms of C1 and DB temperature was established. The general form of this expression is:
It is noteworthy this C2 expression is system and vapour specific. However, for a similar
system, this calibration and fitting method is only done once for each vapour of interest.
Once the relation between the wet bulb and dry bulb temperatures were calibrated, the
concentration of vapour in the convective stream can be calculated using the following
equation:
(17)
Using the ideal gas law, the partial pressure of the vapour is calculated.
(18)
Based on the partial pressure, the respective humidities are obtained using the following
equations:
(19)
(20)
Jiunn Yuan Tan (21637520) Page 65
The humidity of different vapours with known properties can be calculated using this
model simply by changing the liquid reservoir and altering the vapour properties accordingly.
This measurement and calibration method allows a simpler and more versatile approach to
measure the humidities of different vapours.
3.2.3 Control Experiments
The premise of the technique is in introducing ethanol into the droplet via the
vapour absorption mechanism in a controlled manner. Therefore, as control experiments, it
is important to gauge on the morphology of the particles attained if: (1) ethanol vapor was
not used leaving only nitrogen as the convective drying medium and (2) liquid ethanol was
added directly into the aqueous sample of the maltodextrin. The first control experiment
was carried out by flowing nitrogen gas through empty conical flasks, directly heated by the
water bath and into the drying chamber continuously for a similar drying time of 30 minutes.
The second control experiment involved adding liquid ethanol gradually into the sample
solution until precipitation occurred. The solution was sieved to obtain the precipitated
products and left to dry in an oven overnight. Similarly, the products obtained were sent for
SEM imaging.
3.3 Modified AVP Single Droplet Rig
3.3.1 Vapour Generation System
In order to have a more independent and efficient control of vapour humidity, a
dual-stream vapour generation system using nitrogen was adopted. Nitrogen was used as
the bulk convective medium for safety reasons. It is safer to heat up a separate nitrogen gas
stream and allow the gas mixture to equilibrate to the desired temperature compared to
heating the volatile vapour directly. In addition, it provides an additional temperature and
Jiunn Yuan Tan (21637520) Page 66
RH control while maintaining the desired concentration of vapour (i.e. AH) in the overall gas
mixture. It is noteworthy that the second stream is only used when the control of RH or final
vapour temperature is required. A liquid pumping system ensures liquid is fed into the
vapour generation chamber to ensure continuous vapour generation. The vapour sensors
are located at the outlet streams for accurate analysis of the vapour conditions. A bleed-off
is introduced to allow the sampling and measurement of vapour humidity without affecting
the outlet vapour-gas mixture. This enables continuous monitoring and control of humidity
and temperature to be conducted throughout the vapour generation process. A schematic
diagram of the new vapour generator is shown in Figure 3.4. The entire system is an open
system that runs at atmospheric conditions with no need for pressure adjustment.
Nitrogen sourceV-2
Vapour-Nitrogen
101
TT
101
RHC
Feed Liquid VapourFrom Tubing Pump
101
FI
Liquid Reservoir
V-3
102
FI
V-1
Nitrogen source
Nitrogen source
101
WT
101
WC
101
AHC
103
TT
102
TT
Vapour-Nitrogen Bleed
101
TIC
102
TIC
V-4
Figure 3.4: Schematic figure of the proposed experimental set-up for single droplet ethanol vapour
distributor system. (a) Peristaltic pump; (b) Load cell; (c) Round bottom flask heating mantle; (d)
Tee; (e) Convective gas heater.
(a)
(c)
(b)
(e)
(d)
Jiunn Yuan Tan (21637520) Page 67
Four different parameters are automatically controlled within this system: reservoir
level, final gas mixture temperature, absolute humidity and relative humidity. The need for
reservoir level is to automate the addition of liquid into the vapour generation system,
avoiding disruption during a process and also avoiding the risk of reservoir running dry. Two
types of temperature measurements were taken. TT 103 and TT 102 measure the dry bulb
temperatures while TT 101 measures the wet bulb temperature. The final gas mixture
temperature measured by TT 103 is controlled by heating up the nitrogen only stream using
the air heater and allowed the resultant gas mixture to equilibrate until the desired set-
point final temperature is achieved. The absolute humidity is controlled by altering the
temperature of the heating mantle while the relative humidity is controlled by altering the
temperature of the convective gas heater. Although the AH control and RH control shares a
common temperature signal (TT 101), the controllers act independently. When the system
is initiated, the control of absolute humidity will be first activated and allowed to stabilize.
Subsequently, the relative humidity control is enabled. This control strategy capitalises on
the phenomenon where relative humidity is sensitive to the variation in temperature while
absolute humidity is only dependent on the amount of vapour being generated and hence
will not change regardless of the heating of the second stream.
3.3.2 Safety Considerations
Due to the high volatility and self-ignition potential of some gas vapours, the vapour
generation poses a high safety risk. Therefore, the experiment is conducted using nitrogen
gas as the bulk convective medium instead of air. The temperature upper limit is pre-set for
the heating mantle and air heater is fixed into the temperature controller system to prevent
self-ignition of volatile vapour. Ambient pressure conditions and an open system prevent
issues such as ethanol flashing and pressure build-up. A check valve is installed to prevent
Jiunn Yuan Tan (21637520) Page 68
the backflow of ethanol vapour into the heating element. Constant monitoring of the
vapour conditions through the LabView data acquisition device is conducted to ensure no
extreme deviation of process parameters occur throughout the drying run.
3.3.3 Control Systems
The control system makes use of the LabView Automated System to execute the
appropriate control response. The LabView CompactDAQ Chassis (cDAQ-9174) is used to run
different LabView modules through LabView programming on the computer. The 24-Bit
Bridge Input Module (NI 9237) is connected to the load cell to read the volume of liquid
vapour in the round bottom flask while the 24-Bit Thermocouple Differential Analog Input
Module (NI 9211) is connected to the thermocouples for temperature measurements. Lastly,
16-Bit Analog Output Module (NI 9263) is connected to the heating mantle, air heater and
peristaltic pump to execute the appropriate control response based on an analog voltage
output signal. The National Instruments LabView 2012 (32-bit) program is used as the
interface to read the input readings and execute the appropriate control output response.
3.3.3.1 Reservoir Volume/Level Control
As mention earlier, it was important to maintain the volume of liquid inside the
vapour generation system to ensure continuous process throughout the experiment. This
also helps mitigate problems such as liquid spillage and overheating of the heating mantle
(which can occur if heating continues after the liquid after all the liquid has evaporated). A
load cell feeds the mass change signal into the LabView system. When the mass of liquid
reduces, the load cell detects the reduction in mass below the set-point and sent an
analogue voltage signal to the peristaltic pump to top-up liquid vapour until the set-point
mass (i.e. volume) was achieved. This volume control makes use of an on/off control on the
Jiunn Yuan Tan (21637520) Page 69
peristaltic pump. An Autoclude Peristaltic Pump Type AU UV EZ (Serial number: 30017) was
used due to its compatibility with computer systems and its low liquid pumping rate, which
was sufficient for laboratory scale applications.
3.3.3.2 Temperature Control
Temperature control of the convective drying medium is achieved by adjusting the
heating of the convective gas heater using a temperature controller, connected to the
thermocouple located in the drying chamber. This is a much safer alternative compare to
heating the vapour stream directly. An Omega Air Heater (AHPF-061) was connected to the
Eurotherm PID Temperature Controller (Model 2116/2132) to execute the appropriate
heating, with a proportional band of 20, integral time of 360 and derivative time of 60.
3.3.3.3 Absolute Humidity (AH) Control
AH was controlled by adjusting the temperature of heating mantle (MRC Lab, K-1D),
with FY 400 Digital PID Temperature Controller. By inputting the vapour humidity model
into the LabView system, the vapour humidity was calculated accurately and
instantaneously based on the wet and dry bulb temperature measured. The resultant ΔAH
vapour humidity signal was sent through a voltage model to establish a relation between
ΔAHSP and degree of heating. Based this difference, the voltage model produced a voltage
signal to control the degree of heating by the heating mantle until the desired absolute
humidity set point was achieved. The overall signal programming flow within the humidity
controllers is shown in Figure 3.5. A heating limit was set for the heating mantle to ensure
no runaway heating which may cause vapour ignition. The heating mantle operated at a
proportional only control with a proportional band of 5 for fast and stable heating response.
As expected for proportional gain control, there was an offset which was overcome by
Jiunn Yuan Tan (21637520) Page 70
inputting a slightly higher set-point value. It was found that this is a better alternative than
introducing an integral time or derivative time term in the control system, as it will cause
the system to take longer to stabilize. This was due to the fact that the heating mantle takes
a while to heat up the system and generate enough evaporation in order to achieve the set-
point humidity, particularly in high humidity conditions. An integral time control will
accumulate the errors (i.e. deviation between AH and AHSP) over a period of time and cause
large overheating. The resulting control using the heating mantle with a PID setting is shown
in Figure 3.7.
Wet Bulb Measurement
Humidity Model
AH Reading
RH ReadingDry Bulb
Measurement
AH Voltage Model
RH Voltage Model
AH Voltage Signal
RH Voltage Signal
Air Heater Voltage Input
Heating Mantle
Voltage Input
Degree of Heating
Degree of Heating
AH Set-Point
RH Set-Point
Humidity BoxTemperatureMeasurement
Figure 3.5: Programming Flow for the Execution of Humidity Control.
The incorporation of the voltage model was crucial in this control system as it allows
faster and more efficient control response, with a AH fluctuation of ±0.002 for the vapour
humidity range in our application. Using only a basic on-off heating control with heating
mantle resulted in large degree of overshoot and undershoot from the set point humidity,
as shown in Figure 3.6. The premise of this control was to input a temperature signal into
the temperature controller to perform the required heating until the set point humidity was
achieved. It is noteworthy that this temperature signal is not a direct temperature
measurement, but a falsified signal generated through a voltage input in order to perform
Jiunn Yuan Tan (21637520) Page 71
the required heating or cooling based on a pre-set set point temperature within the
temperature controller. As a proportional only control was used for this heating, the set-
point temperature had to be high enough to induce the required heating for vapour
generation. The voltage input corresponds to the degree of heating which then affect the
humidity; lower voltage input will correspond to a lower temperature input signal inducing
higher heating rate. Higher heating rate causes high vapour evaporation rate; hence, a
higher AH. The voltage input can be modelled in terms of the deviation of the absolute
humidity from its set-point prior to being input to the temperature controller. This was done
by using the MathScript function in LabView by inputting a model to alter the voltage signal
based on the AH change:
Using this model, a relationship between the temperature control and the vapour
humidity was established. The heating rate can be controlled and modelled in terms of
various ΔAH. Vsp is the voltage signal corresponding to the pre-set set-point temperature of
the temperature controller while v is the falsify voltage input signal corresponding to the
measured temperature. Based on the model, when ΔAH reduces (i.e. AH approaching
AHSP), the voltage input signal, v increases (i.e. less heating). When the absolute humidity
has reached its set-point value (ΔAH = 0) or exceed its set-point value (ΔAH > 0), v is
equivalent to vsp or higher than vsp and the heating will stop. This voltage model relates ΔAH
to the temperature change in order to perform the required control action. When the ΔAH
is large the second term is larger causing the overall v to reduce, corresponding to a lower
temperature signal; thus, inducing more heating. On the other hand, when ΔAH is small the
Jiunn Yuan Tan (21637520) Page 72
second term reduces, resulting in a higher overall v that corresponds to a higher
temperature signal; thus, inducing less heating. The model reduces the overheating and
creates a temperature change cause by the voltage signal change which corresponds to the
ΔAH. The term w was used to scale the ΔAHSP reading into the voltage range. This is
dependent on the voltage signal range corresponding to the temperature reading on the
temperature controller.
This control using the voltage model is applicable to any indirect control, whereby
the controlled parameter is not a direct measurement of the measured parameter. Most
equipment executes a control response based on an electronic voltage or current signal. By
linking the controlled parameter and measured parameter through a voltage or current
model based on the approach described above, the appropriate control response could be
executed efficiently by the equipment.
Figure 3.6: Ethanol AH Control Stabilization for On/Off Heating Method for AHSP = 0.1.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0 1000 2000 3000 4000 5000 6000
Ethanol Absolute Humidity
(kg/kg db)
Time (s)
Overshoot
Undershoot
Jiunn Yuan Tan (21637520) Page 73
The EAH control response for a set-point of 0.1 using an on-off heating method with
the heating mantle, without the voltage model is shown in Figure 3.6. During the initial
stage, the heating mantle just started heating up the system. Hence, there was insufficient
heat to drive the liquid vapour evaporation, resulting in a slight decrease in EAH. As heating
continued, sufficient heat had been transferred to the system for the liquid evaporation to
gradually increase the EAH. When the set-point value of 0.1 was achieved, the heating
stopped. However, the remaining heat within the system caused continuous evaporation of
the liquid resulting in the EAH overshoot. As it reached the peak when sufficient heat was
released from the system, the EAH gradually dropped. When EAH dropped below 0.1, the
heating response was initiated again. In this case, there was insufficient build-up of heat
within the system to induce sufficient liquid vapour evaporation, resulting in EAH
undershoot. Based on this observation, it was apparent that a relation between the degree
of heating relative to the ΔEAH had to be established in order to achieve a more efficient
control response. This resulted in the voltage model proposed above to control the heating
response based on ΔEAH. Incorporating the voltage model has shown to achieve a better
control response, as shown in Figure 3.8.
Jiunn Yuan Tan (21637520) Page 74
Figure 3.7: Ethanol AH Control Stabilization PID Control within the Heating Mantle for
AHSP = 0.06.
Most temperature controllers utilise a default PID heating control setting. Figure 3.7
shows the EAH control response for a set-point of 0.06 using a PID heating control within
the heating mantle after incorporating the voltage model. It could be seen that the
overshoot and undershoot from EAHSP had been significantly reduced compared to the on-
off heating method. However, the fluctuation from the set-point value was still fairly large
which was ±0.006. It is noteworthy that at higher humidity set-point, this fluctuation will be
amplified. This was due to integral time control which is the gain from the accumulation of
errors (i.e. deviation between AH and AHSP) over the entire heating duration. As observed
in Figure 3.7, it takes approximately 1400s for each heating cycle to heat up the system
when the EAH dropped below the set-point. This considerably long duration of time caused
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0 1000 2000 3000 4000 5000 6000 7000 8000
Ethanol Absolute Humidity
(kg/kg db)
Time (s)
Heating
Cycle 1
Heating
Cycle 2
Jiunn Yuan Tan (21637520) Page 75
a high gain from the integral time control due to the large accumulation of error, resulting in
a slight overshoot in each heating cycle. Therefore, a proportional gain only control was
used for the heating of the heating mantle, resulting in the control response shown in Figure
3.8 below.
0
0.02
0.04
0.06
0.08
0.1
0.12
0 1000 2000 3000 4000 5000 6000 7000
Ethanol Absolute Humidity
(kg/kg db)
Time (s)
(a)
Jiunn Yuan Tan (21637520) Page 76
Figure 3.8: Ethanol AH Control Stabilization for Set-Points (a) 0.09; (b) 0.065 and (c) 0.038.
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0 500 1000 1500 2000 2500 3000 3500 4000
Ethanol Absolute Humidity
(kg/kg db)
Time (s)
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.045
0.05
0 500 1000 1500 2000 2500 3000 3500
Ethanol Absolute Humidity
(kg/kg db)
Time (s)
(b) (c)
Jiunn Yuan Tan (21637520) Page 77
0
0.005
0.01
0.015
0.02
0.025
0.03
0 2000 4000 6000 8000 10000 12000 14000
Water Absolute Humidity
(kg/kg db)
Time (s)
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Water Absolute Humidity
(kg/kg db)
Time (s)
(a)
(b)
Jiunn Yuan Tan (21637520) Page 78
Figure 3.9: Water AH Control Stabilization for Set-Points (a) 0.015; (b) 0.011 and (c) 0.006.
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0 1000 2000 3000 4000 5000 6000 7000 8000
Water Absolute Humidity
(kg/kg db)
Tims (s)
0
0.05
0.1
0.15
0.2
0.25
0.3
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Acetone Absolute Humidity
(kg/kg db)
Time (s)
(a)
(c)
Jiunn Yuan Tan (21637520) Page 79
Figure 3.10: Acetone AH Control Stabilization for Set-Points (a) 0.13; (b) 0.1 and (c) 0.07.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0 500 1000 1500 2000 2500 3000
Acetone Absolute Humidity
(kg/kg db)
Time (s)
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0 500 1000 1500 2000 2500 3000 3500 4000
Acetone Absolute Humidity
(kg/kg db)
Time (s)
(b)
(c)
Jiunn Yuan Tan (21637520) Page 80
Experiments were conducted for three different vapours: ethanol, water and
acetone at three different set-points. Acetone is the most volatile and flammable among the
three followed by ethanol and lastly water. The AH readings with time for all the
experimental runs are shown in Figure 3.8, 3.9 and 3.10 for ethanol, water and acetone
respectively. AH was the measure of the mass ratio of gas vapour to nitrogen within the
vapour-nitrogen gas mixture. Higher temperature induced higher evaporation rate which
resulted in higher AH.
Figure 3.8 and Figure 3.10 show the AH control stabilization of ethanol and acetone
vapour respectively. Initially, the sharp increase of AH was due to the natural entrainment
of liquid vapour within the evaporation chamber by nitrogen, which depended on the liquid
level in the vapour generation chamber. Higher liquid level resulted in higher entrainment
and hence, a higher initial increase of AH. It could also be seen that acetone has a higher
initial increase compared to ethanol due to its higher volatility. In the case of ethanol,
vapour entrainment was not sufficient to achieve the set-point AH. Therefore, heating was
induced to evaporate more vapour until the set-point humidity was achieved. Based on
Figure 3.8, higher AH set-point takes a longer time to achieve the steady state set-point
simply because more heat was required to generate the required vapour. As for acetone, its
high volatility resulted in a large degree of vapour entrainment causing a sharp AH increase
above the set-point AH as shown in Figure 3.10. Nevertheless, heating was required to
maintain the AH at the desired set-point. This was because the vapour generated through
entrainment was inconsistent and difficult to control. As the liquid level within the
evaporation chamber reduced, the vapour entrainment reduced significantly as well. When
Jiunn Yuan Tan (21637520) Page 81
the humidity dropped below the set-point AH, heating was induced to generate and
maintain sufficient vapour generation to achieve the set-point AH.
The AH control of water is shown in Figure 3.9. It is noteworthy that for the water
runs, the system was pre-heated beforehand before allowing the nitrogen flow and
initiating the control and measurement system. This was due to the fact that water has a
low volatility and flammability level. Therefore, more heat was required to induce the
evaporation of water in order to generate sufficient water vapour. Since water is not
flammable, it was safe and more efficient to pre-heat the system beforehand to allow the
accumulation of heat within the system before initiating the control actions. This reduced
the amount of nitrogen used in these runs. As can be seen in Figure 3.9-b and 3.9-c, the pre-
heated system generated slightly higher water vapour compared to the required AH set-
point. It was then allowed to cool below the set-point AH and stabilize through heating,
similar to the acetone runs. For the high humidity (Figure 3.9-a), the pre-heating was not
sufficient to generate enough water vapour. Therefore, more heating was required to
induce higher evaporation until the set-point AH was achieved.
The AH stabilization time was dependent on the volatility of the liquid vapour and
the set-point humidity. Water took the longest time to achieve the required set-point AH
followed by ethanol and lastly acetone. This was due to the fact that higher volatility
increased vapour generation rate allowing the set point humidity to be achieved faster.
Evidently, the higher the set point humidity, the longer the time needed for the system to
stabilize to the set point. Longer time was needed as more heating was required for higher
AH set-point.
Jiunn Yuan Tan (21637520) Page 82
3.3.3.4 Relative Humidity (RH) Control
The same heating and control principle was used to manipulate the RH by controlling
the temperature of the air flow heater or the heating mantle. Similarly, the same principles
were employed to optimise the control of RH through the heating and cooling process.
Using the same technique, the voltage model for RH is shown below. The air heater
operated at a proportional only control with proportional band of 5 for fast heating
response. A similar voltage input model as the absolute humidity was employed:
In order to establish an independent control between RH and AH, the control of RH
was done by heating up the second nitrogen only stream until the desired RH was achieved.
In order to control both the AH and RH, a dual stream nitrogen flow was required. In similar
manner, the AH is controlled and allowed to achieve its set point first using the heating
mantle. Subsequently, the RH control was turned on by heating up the second stream using
the air heater in order to achieve the desire RH set point at a specific AH. Both the RH and
AH control operates concurrently and independently through the air heater and heating
mantle respectively .The dual stream control of AH and RH is shown in Figure 3.11. As RH is
inversely proportional with temperature, the limitation in our system was the relative
humidity can only be reduced from its corresponding absolute humidity condition. In order
to increase the relative humidity, a cooling unit is required which is beyond the scope of this
system.
Jiunn Yuan Tan (21637520) Page 83
Figure 3.11: Dual-stream AH and RH control stabilization.
Figure 3.11 depicts the concurrent dual stream control of AH and RH. RH was the
amount of gas vapour over the maximum amount of gas vapour within the gas mixture at a
particular temperature. The change in relative humidity is inversely proportional to
temperature. Higher temperature allowed higher maximum amount of gas vapour within
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0.018
0 1000 2000 3000 4000 5000
Water Absolute Humidity (kg/kg db)
Time (s)
0
10
20
30
40
50
60
70
80
90
0 1000 2000 3000 4000 5000
Water Relative
Humidity (%)
Time (s)
Jiunn Yuan Tan (21637520) Page 84
the gas mixture, resulting in lower RH. Similarly, the system was pre-heated and the
nitrogen flow and AH control were initiated. The system was allowed to cool and stabilize at
0.04 AH set-point. Subsequently, the RH control was enabled to a set point of 20% at time
1000s. The initial sharp decrease in RH was corresponding to the decrease in AH. As the AH
stabilized, the gradual decrease in RH was due to the heating of nitrogen only stream by the
air heater. The heated nitrogen stream increased the temperature of the overall gas mixture;
thus, decreasing the overall RH to the set-point required. As mention earlier, in order to
increase the RH, a cooling unit is required which is outside the scope of this work.
3.4 Mass Change and Temperature Measurements of the Droplet during
Drying
A water droplet was evaporated on the single droplet rig using ethanol vapour at
known humidity and temperature conditions. The flow rate of nitrogen used as the bulk
convective medium is 10 L/min. Throughout the drying process, the mass change and
temperature of the droplet were measured and recorded against time. The detail method of
this experiment is outline by Chen and Lin 108. Figure 3.12 shows the schematic diagram of a
single droplet mass change experiment. This was done by measuring the deflection of the
droplet from a marker point. This deflection was correlated to the amount of mass change
of the droplet during the drying. By comparing the deflection of the liquid droplet to the
deflection of standard glass beads with known mass, the mass change of the droplet was
calculated throughout the experiment. It is noteworthy the mass change of this droplet was
attributed by both the simultaneous absorption of ethanol as well as the evaporation of
water. The temperature measurement was done by placing a thermocouple within the
liquid droplet. From the data obtained, the mass change and temperature profiles at
Jiunn Yuan Tan (21637520) Page 85
different humidity were collated and plotted. The mass change profiles produced from
these experimental works were then compared to the mass change profiles produced by a
theoretical model obtained from equation derivation.
Ethanol-Nitrogen Vapour
Drying Chamber
Glass filament Chamber
Figure 3.12: Schematic diagram of mass change single droplet rig.
3.5 Solubility Measurement
Maltodextrin exists as glucose units with variable chain length. Typically, a dextrose
equivalent (DE) value is used to classify the type of maltodextrin. However, maltodextrin
with similar DE value may not have similar physio-chemical characteristics due to the
different proportions of saccharides of high and low molecular mass. In addition, there is no
solubility data available for maltodextrin DE 10 in a water-ethanol mixture in the literature.
Therefore, it was required to determine the solubility of the sample maltodextrin material
experimentally.
Camcorder
Jiunn Yuan Tan (21637520) Page 86
The solubility experiment was conducted by adding maltodextrin DE 10 powder into
the solvents gradually at ambient temperature until the point where it becomes insoluble
and the amount of solute mass was recorded. Upon addition, the solution was stirred
continuously to avoid the formation of clumps. A small amount of the saturated solution
was then left to dry in an oven until there was no detectable mass change for the dried
solute. All solutions were prepared in triplicates. Solubility experiments were first conducted
by adding maltodextrin DE 10 into pure ethanol and pure water. The solubility of
maltodextrin DE 10 in pure water solvent was found to be 0.103 g / g solution while it was
insoluble in pure ethanol solvent. Next, the solubility of maltodextrin was determined in a
solvent mixture with volume proportion of 80% water and 20% ethanol was prepared. The
solubility was measured to be 0.094 g / g solution. It was found that for mixtures at lower
proportion of water compared to ethanol, the addition of maltodextrin in the solvent
mixture resulted in the formation of clumps. In that case, a fixed amount of maltodextrin DE
10 mass was diluted in a known volume of water. Liquid ethanol was then added into the
solution gradually until precipitation occurred, the volume of liquid ethanol added was
recorded. Similarly, the saturated solution was then left to dry in an oven until there was no
detectable mass change for the dried solute. The solubility points of maltodextrin DE 10 at
different proportion of ethanol-water mixture was used to generate a solubility curve for
further analysis discussed in Chapter 4.
Jiunn Yuan Tan (21637520) Page 87
Chapter 4
RESULTS AND DISCUSSION
4.1 Unveiling the Mechanism of AVP in Producing Porous and Spherical
Particles
4.1.1 Control experiment using nitrogen gas
It is important to note that the usage of nitrogen gas as the drying medium is similar
to drying with air due to the high proportion of nitrogen in air. We could observe the drying
behaviour of the droplet throughout the drying process in Figure4.1 (a). The size of the
droplet decreased over time, the droplet turned cloudy gradually and solidification was
observed. The droplet shrinkage was due to the convective drying by the flow of nitrogen
gas in the chamber at approximately 30oC and 10 L/min, which resulted in the evaporation
of water from the droplet. The cloudiness of the droplet was due to the formation of solids
within the droplet, as the dehydration process took place. The final particle morphology
obtained under SEM was a chunk of discrete solids shown in Figure 4.2 (a) (i).
Jiunn Yuan Tan (21637520) Page 88
Figure 4.1: Drying behaviour over time at ambient temperature of a single droplet using (a)
pure nitrogen gas; (b) nitrogen gas and ethanol vapour (0.09 kg/kg db).
4.1.2 Ethanol Vapour Precipitation
Figure 4.1 (b) depicts the drying behaviour of the sample droplet over time using
nitrogen/ethanol vapour gas mixture as the convective drying medium. The droplet initially
expanded for a short period of time and started to shrink over time. It gradually turned
cloudy throughout the entire process and finally formed a mushroom like particle. The initial
expansion of the droplet was due to the absorption of ethanol vapour into the droplet. As
the droplet reached a maximum size, the droplet began to shrink due to evaporation. The
cloudiness of the droplet could be attributed by the precipitation process, resulting in solids
formation. There was a clear distinction observed between the runs with and without
ethanol vapour in Figure 4.1; the droplet for the ethanol run turned cloudy much faster as
compared to the nitrogen only run even when the droplet is very liquid like. This is
attributed to the solid precipitation process when ethanol was added into the system in
contrast to the solidification process. Another interesting observation was the droplet
(b) (a)
(b)
Jiunn Yuan Tan (21637520) Page 89
gradually shrivelled into a mushroom like shape during the drying process. It is noteworthy
that this mushroom like shape was only observed for the drying run which resulted in
microspheres precipitation.
Figure 4.2: (a) (i) Clustered crystal structure; (ii) Smooth solid structure; (b) (i) Smooth
surface; (ii) Patchy porous network; (iii) Round porous network; (iv) Microsphere network;
(v) Microsphere.
(i) (ii) (iii)
(iv) (v)
(i) (ii) (a)
(b)
Jiunn Yuan Tan (21637520) Page 90
Figure 4.2 (b) shows the type of particle morphology obtained from this drying
approach for maltodextrin and maltose. Various morphologies were obtained: smooth
surface; porous network; microsphere network and microspheres. These particle
morphologies differed greatly from the nitrogen only run. The mechanisms of formation of
these structures are discussed later on. In most runs, both round and patchy porous
network were observed in the same sample. It was found that the concentration of the
ethanol vapour, initial concentration of the solute and the length of the carbohydrates
strongly influenced the morphology produced (Figure 4.3). The data from the experiments
are collated and included in the supplementary information.
For all the drying runs, the suspended droplet precipitated into a spherical 'droplet
like' shape due to the fact that the solutes were still dissolving within the droplet. However,
in the case of the drying runs resulted in microsphere precipitation, it was observed that the
particle formed a mushroom like shape upon dehydration. In view that the density of
saccharides is higher than water, the precipitated particles aggregate from the bottom and a
mushroom like shape was formed due to surface tension effect. This observation indicates
that the microspherical particles were ‘precipitating out’ of the water droplet throughout
the drying run.
Instead of ethanol vapour, liquid ethanol was used to precipitate the sample solution.
Figure 4.2 (a) (ii) shows the SEM image of the particles obtained, which have a smooth solid
structure. This result is of great significance, as it indicates the degree of ethanol absorption
into the system and the manner at which it occurred has a major influence on the
precipitation process. More discussions will be outlined later to further explain this
occurrence.
Jiunn Yuan Tan (21637520) Page 91
4.1.3 Effect of Relative Humidity and Absolute Humidity
The drying of all three samples of different initial solute weight concentrations were
conducted at varying ethanol relative humidity (ERH) and ethanol absolute humidity (EAH).
It was observed that microspheres precipitation occurred at high ERH condition (Figure 4.3).
Collating the experimental observations in an ERH versus initial droplet concentration plot
and an EAH versus initial droplet concentration plot showed similar trend. To avoid
repetition, only the plot of ERH versus initial droplet concentration was presented. A higher
ERH would induce a higher concentration gradient between the surface of the droplet and
its surroundings, which will increase the rate of absorption of ethanol into the droplet. This
higher concentration driving force would allow larger amount of ethanol to absorb into the
droplet. Similarly, higher EAH correlates to a higher ambient temperature and higher
amount of ethanol in the gas mixture. Consequently, the absorption rate of ethanol into the
droplet and amount of ethanol in the droplet was also increased by the higher
concentration gradient. Therefore, it can be deduced that higher rate of ethanol absorption
into the droplet or higher ethanol concentration in the droplet favoured microsphere
precipitation. It is also interesting to point out that there seemed to be a transition of the
resultant particle structures for the maltodextrin samples from porous, microsphere
network and finally microspheres, as the ambient ethanol humidity was increased.
4.1.4 Effect of Initial Concentration and Chain Length
Lower initial solute concentration seemed to allow a larger range of ethanol vapour
humidity which can lead to the formation of microspheres for maltodextrins (10DE) with the
longer polymeric chain length (Figure 4.3). Precipitation of microspheres at higher initial
weight concentration required higher ERH. If the precipitation process was driven by
supersaturation, it is expected that higher concentration of solute would more likely to
Jiunn Yuan Tan (21637520) Page 92
induce precipitation. This result showed otherwise; hence, it further reinforces the notion
that supersaturation is not the mechanism driving thes process of ultrafine particle
formation. In view that higher amount of maltodextrin required higher ethanol vapour
concentration; this suggests that a certain ‘threshold’ ratio of ethanol to maltodextrin is
required to facilitate the unique precipitate process.
This trend, however, was not observed for maltodextrin (18DE) with the shorter
polymeric chain length. It is unclear at the moment on why the shorter polymeric chain
produced a different trend. In addition, for maltose which is a dissacharide, formation of
the ultrafine uniform particles was only observed at the lowest initial concentration tested.
At the higher concentration, conventional precipitation in which smooth chunks of the
maltose was produced, which were not observed for the maltodextrins. This can be
attributed to the relatively high propensity to crystallize for maltose when compared to the
maltodextrins and will be discussed in detail later on. Nevertheless, with the exception of
maltose at 10 wt% and 15 wt% initial weight concentration, all the samples produced a
similar particle structure transition from: porous network, microsphere network and lastly
microsphere, with increasing ERH condition.
Jiunn Yuan Tan (21637520) Page 93
Microsphere Region (a)
(b)
Microsphere
Network Region
Microsphere
Network Region
Microsphere Region
Jiunn Yuan Tan (21637520) Page 94
Figure 4.3: Graph of ethanol relative humidity against initial weight concentration for: (a)
Maltodextrin DE 10; (b) Maltodextrin DE 18 and (c) Maltose (solid line – apparent
boundary for microspheres formation, dash line – apparent boundary of microsphere
network formation).
4.1.5 Porous and Microspheres Formation
The observation of the formation of the porous network structure at low ethanol
vapour concentration is an interesting finding, delineates the formation process of the
microspheres (i.e. low ERH or EAH conditions). One possible explanation for the formation
of the porous structure could be phase separation of ethanol with the droplet when the
droplet was still in the liquid stage. Phase separation of a polymer solution upon the
addition of antisolvent has also been reported by Erbil et. al.111. The addition of an
(c) Microsphere
Network
Region
Microsphere
Region
Jiunn Yuan Tan (21637520) Page 95
antisolvent into a polymer mixture solution has been reported to produce smaller polymer
aggregates and increases particle formation rate due to the higher rate of evaporation.
Upon dehydration, the porous structure was formed. In view that maltodextrin is more
soluble in water and conversely insoluble in ethanol, the matrix surrounding the bubbles
would be the phase predominantly consisting of water. However, it is unsure whether the
bubble phase consisted of ethanol or ethanol-water before dehydration. The phase
separation and subsequent dehydration into the porous structures was mainly observed for
the low ethanol vapour conditions. On the other hand, solid microsphere formation
occurred at conditions which promoted high absorption rate and large amount of ethanol
within the droplet. In view that the attainment of higher ethanol concentration within the
droplet will firstly have to undergo the initial period of low ethanol concentration, from
hindsight, we hypothesize that the formation of the phase separation could be a precursor
stage leading to the microsphere formation.
Testing this hypothesis, we further extended our experimental work by conducting
microscope imaging of the droplet during the absorption process. Experiments were
undertaken for maltodextrin DE 10 (90% ethanol relative humidity and 2.5 wt% initial
weight concentration), as these parameters were most favourable for microspheres
precipitation. During the absorption process, when the droplet just turned cloudy (Figure
4.1b – 3), the liquid droplet was removed and placed between two microscope glass slides
for observation. It was important to ensure that there was no entrapped air between the
glass slides, so that the distinctive regions viewed through the microscope represents that
of the bubble and not entrapped air bubbles. Distinctive bubbles were observed within the
liquid supporting our hypothesis (Figure 4.4). Separation of miscible liquids into separate
Jiunn Yuan Tan (21637520) Page 96
layers in the presence of solute112 and surfactant113 has been reported before in the
literatures. In the current experiments, however, we were able to observe ethanol-water
separation forming an apparent ‘emulsion’ within the droplet. It is noteworthy that the
formation of the meso-scale bubbles are not molecular-scale supra molecules typically
observed in ethanol-water systems 114. Typically, the formation of such ‘self-emulsification’
phenomenon is observed in a water-oil mixture which is immiscible115, resulting in two-
dimensional spherical colloidal structures under certain conditions116. The presence of
maltodextrin, which was reported to decrease the surface tension of water117, seemed to
have induced a phase separation between water and ethanol into possibly a water-ethanol
phase and predominantly water, water-maltodextrin phase. Interestingly, the formation of
an emulsion (phase separation) from the fully miscible alcohol and water through the
addition of lactose sugar has also been reported in brief without detailed investigation by
Herrington 118. It is also interesting to take note of the bubble-like shape observed in the
current experiments of the second phase, which the shape and size is similar to the porous
network.
Observation on this initial phase separation then raised another question. In the
initial separation of the phases forming the continuous phase is predominantly the water
phase with the dissolved maltodextrin. However, the formation of microspheres at the end
of the precipitation process delineates that predominantly water phase eventually becomes
the discrete in the droplet. How does this inversion could have occurred? Based on the
current observations combining with the previously deduced ‘pinch off’ phenomenon, the
following mechanism is proposed in Figure 4.5.
Jiunn Yuan Tan (21637520) Page 97
At low ethanol concentration (i.e. low ERH or EAH conditions), the absorption and
concentration of ethanol within the droplet was limited, an accumulation of small amount
of ethanol within the particle resulted in the phase separation of small water-ethanol
bubbles across the bulk water-maltodextrin phase. When particle was fully dried, the water-
ethanol bubble phase was fully evaporated and the porous structure was formed. We
speculate that continue increase in ethanol within the system caused more ethanol to be
absorbed and expand resulting in the shrinkage of the predominantly water-maltodextrin
phase to shrink, which was referred to as the 'deduced transition'. At some point of ethanol
concentration, the ethanol-water phase then becomes the bulk phase with the water-
maltodextrin phase forming a network like phase within the droplet. When high enough
ethanol absorbed into the droplet, further shrinkage of the network like phase then resulted
in discontinuity which forms the discrete spherical water-maltodextrin phase due to surface
tension. This was previously observed and deduced for the ‘pinch off’ mechanism proposed
by Mansouri et. al. 7. Figure 4.5 illustrates this framework put forward.
4.1.6 Crystallisation and Precipitation
For the precipitation of maltose (Figure 4.3 - c), microspheres precipitation could be
obtained at low initial weight concentration. Similarly, high ERH and EAH (i.e. high ethanol
absorption rate & high maximum ethanol concentration in the droplet) were conditions
favourable for microspheres precipitation. However, it is interesting to note that the
precipitation of maltose at high initial weight concentration at high ERH and EAH seemed to
produce smooth structure instead of the expected microspheres. This is due to the higher
crystallisation potential of maltose at higher humidity and high initial weight concentration
119. For maltose, higher initial concentration could have facilitated crystallization of the
material as opposed to maltodextrins which do not crystallize. The crystallisation process
Jiunn Yuan Tan (21637520) Page 98
impedes the initial formation of the phase separated bubbles which is the precursor for the
subsequent formation of microspheres. Although direct comparison cannot be made,
antisolvent crystallisation study in the precipitation of ibuprofen showed that under certain
precipitating temperature and ibuprofen concentration, phase separation was observed in
place of precipitated crystals. On the other hand, in conditions when crystals are formed,
phase separation is not observed 120. The observation suggests that the prevention of
crystallisation is the key step in generating the microspheres, as suggested by Mansouri et.
al. 7. Although the conventional liquid antisolvent21 and the vapour antisolvent technique
shares the same underlying principle, in manipulating the solubility of the solute by the
addition of antisolvent, the key difference lies in controlling the rate in which the
antisolvent is introduced into the system. Contrary to the liquid ethanol control run where
the solids `crash out’ due to supersaturation, the gradual absorption of ethanol vapour into
the droplet seemed to induce a different type of precipitation mechanism resulting in the
formation of amorphous microspherical particles.
Jiunn Yuan Tan (21637520) Page 99
Figure 4.4: Microscope images of the cloudy droplet.
Ethanol -
water
water-
maltodextrin
overview of
microscope
image
Water -
maltodextrin
ethanol -
water
20 µm 20 µm
Jiunn Yuan Tan (21637520) Page 100
Porous Network
MicrosphereMicrosphere
Network
Liquid Stage Prior to Dehydration
Dehydrated Particle
Current observation
Deduced transition
Pinch-off mechanism
Figure 4.5: Schematic diagram of the proposed mechanism of phase separation.
4.1.7 Application
The structural-functional properties of saccharide products have received a
significant degree of attention from the pharmaceutical and food industry recently. It will be
interesting to gauge how these micro-maltodextrin spheres can be used as ultrafine
encapsulants for the food and pharmaceutical industry. The porous particles could be
extended for encapsulation applications. Besides that, the precipitated particles have the
advantage of a smooth surface which could be used to improve flow properties and packing
properties in capsule and tableting technology. The bubble and microsphere network are
merely precursors to the formation of porous structure and microspheres. These work
provided the basis required to extend the application of the antisolvent vapour precipitation
technique to control the particle structures of saccharides with varying chain length.
Jiunn Yuan Tan (21637520) Page 101
4.2 Analysis of the Single Droplet Drying of Maltodextrin under AVP
4.2.1 Single Droplet Drying of Maltodextrin under AVP using the Modified Single
Droplet AVP Rig
Figure 4.6: SEM images of particle structures obtained under AVP drying at EAH of (a) 0.05
kg/kg db; (b) 0.065 kg/kg db and (c) 0.08 kg kg/db at gas velocity of 0.1 m/s and ambient
gas temperature of 25oC.
Maltodextrin DE 10 with an initial weight concentration of 5 wt% was dried under
AVP using the modified single droplet AVP rig, incorporating the vapour generator system
described in Chapter 3. The particle structure obtained at different ethanol humidity
conditions are shown in Figure 4.6. The porous network particle was obtained at EAH of
(a) (b)
(c)
Jiunn Yuan Tan (21637520) Page 102
0.05 kg/kg db while the microsphere network and microsphere particles were produced at
EAH of 0.065 kg/kg db and 0.08 kg/kg db respectively. It is noteworthy that the
corresponding ERH conditions were 55%, 65% and 80% for the EAH conditions of 0.05 kg/kg
db, 0.065 kg/kg db and 0.08 kg/kg db respectively. These results agreed well with the
mapping of particle structure obtained in the original single droplet AVP rig, as shown in
Figure 4.7. This shows the reproducibility of the work done and the efficacy of the vapour
generation system built and described in Chapter 3. In view that the three different particle
structures of interest were obtained under these three EAH conditions, these drying
conditions were adopted for further analysis on the effect of mass and temperature change
on the precipitation process.
Figure 4.7: Particle structure map for maltodextrin DE 10.
Microsphere Region
Microsphere
Network Region
Jiunn Yuan Tan (21637520) Page 103
The mass change profiles of the drop corresponding to the three conditions were
measured, as shown in Figure 4.8. Only the drying run producing microsphere particles
showed an observable ethanol absorption peak of approximately 0.5 mg during the initial
stage of drying. At this ethanol humidity (0.08 kg/kg db), the ethanol concentration gradient
between the bulk medium and the droplet was high enough to induce a rate of absorption
of ethanol that was higher than the rate of evaporation of water during the initial stage of
the AVP drying process. There is a slight absorption peak of approximately 0.05 mg
measured for the microsphere network drying run while the mass change of droplet
resulting in the porous network decreased since the beginning of the drying run. This
implied that the ethanol concentration gradient between the bulk medium and the droplet
was almost equal or lower than the rate of absorption of ethanol was lower than the rate of
evaporation of water.
Another interesting observation was the final mass recorded for each drying run.
Referring to Figure 4.8 (a), it was clearly shown that the final mass of maltodextrin droplet
forming the microspheres was approximately 0.1 mg, which corresponded well with the
initial weight of maltodextrin diluted to form the solution. However, drying of maltodextrin
droplet for microsphere network and porous network particles resulted in a significantly
higher final mass of 0.9 mg and 0.5 mg, as shown in Figure 4.8 (b) and Figure 4.8 (c). This
indicates that the microsphere network and porous structure have moisture (ethanol)
retention behaviour.
Jiunn Yuan Tan (21637520) Page 104
-0.5
0
0.5
1
1.5
2
2.5
3
-200 0 200 400 600 800 1000
Mass (mg)
Time (s)
0
0.5
1
1.5
2
2.5
-200 0 200 400 600 800 1000
Mass (mg)
Time (s)
(a)
(b)
Jiunn Yuan Tan (21637520) Page 105
Figure 4.8: Mass change profile of the AVP drying of (a) Microspheres; (b) Microsphere
network and (c) Porous network particles.
4.2.2 Discussion
4.2.2.1 Effect of Maximum Ethanol Concentration in the Droplet
Figure 4.9: Solubility curve of maltodextrin DE 10 in water-ethanol mixture.
0
0.5
1
1.5
2
2.5
-200 0 200 400 600 800 1000
Mass (mg)
Time (s)
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0 0.2 0.4 0.6 0.8 1 1.2
Solubility (g/g solution)
Mass Fraction of Ethanol
(c)
Jiunn Yuan Tan (21637520) Page 106
Mass change measurement of the droplets during AVP drying clearly indicated that
microsphere particles were formed when the droplet experienced the highest ethanol
absorption (i.e. highest ethanol absorption rate & higher maximum concentration of ethanol
within the droplet). Subsequently, at slightly lower ethanol absorption the microsphere
network was formed and lastly the porous network was produced at the lowest ethanol
absorption condition. These results further reinforced the notion and mechanism put forth
in Section 4.1.
Based on the mass change profiles of the different particle structures, we further
attempted to analyze the implication of higher absorption rate and maximum ethanol
concentration within the droplet on the resulting particle precipitation. Assuming the mass
transfer of nitrogen was negligible at the droplet interface and the water evaporation was
negligible in the initial part of the drying process, the increase of 0.5 mg during the
microspheres drying run was solely attributed to the absorption of ethanol. This gives an
apparent maximum ethanol concentration of 20.8 wt%. On the basis of the solubility curve
generated (Figure 4.9), the solubility of maltodextrin DE 10 can be approximated to 0.078 g
maltodextrin / g solution at room temperature. At the peak of the droplet mass, the
concentration of maltodextrin can be approximated to be 0.042g maltodextrin / g solution,
which was lower than the saturation concentration. At the initial concentration of
maltodextrin of 0.05 g maltodextrin / g solution, without the presence of ethanol, the
solubility of maltodextrin is 0.103 g maltodextrin / g solution. This analysis was further
extended to the microsphere network and porous network drying runs. For the microsphere
network run, the maltodextrin concentration was about 0.051 g maltodextrin / g solution at
the peak of the droplet mass. The apparent maximum concentration of ethanol was merely
Jiunn Yuan Tan (21637520) Page 107
2.6% and the corresponding solubility was about 0.1 g maltodextrin / g solution. For the
porous network run, no observable absorption peak was detected. This observation agrees
well with previous analysis conducted on the drying of lactose particles 6. The indicative
trend based on this analysis suggests that sufficiently high maximum ethanol concentration
within the droplet was the prevailing factor resulting in the formation of microspheres.
However, this analysis presented is a lower bound case of ethanol concentration in the
droplet. The effect of simultaneous water evaporation from the droplet on the maximum
ethanol concentration is still unknown. Therefore, it is of interest to develop a drying model
for this the simultaneous absorption and evaporation of ethanol and water within the
droplet in order to better understand the fundamental mechanism of the process. This
modelling work will be the subject of interest explored and discussed in Section 4.3.
4.2.2.2 Liquid Retention within the Solid Microsphere Network and Porous Network
Structure
The high final mass recorded for the microsphere network and porous network
drying runs due to the high retention of liquid within the solid was an interesting
observation. The solid precipitation seemed to occur at the outer region of the droplet,
hence forming a solid crust during the drying process. This solid crust viewed under SEM
consisted of microsphere network and porous network structures which are so densely
packed that it was able to impede the outward liquid migration and retain a large amount of
liquid at the end of the drying process. However, this observation was not observed for the
microspheres drying run as the formation of microspheres particles were well-dispersed,
thus allowing room for outward liquid migration and continuous evaporation. Such
retention ability has been reported for drying of maltodextrin solution albeit in trace
amounts 121. In addition, the liquid retention could also indicate the occurrence of liquid
Jiunn Yuan Tan (21637520) Page 108
phase separation phenomenon proposed in Section 4.1. During the drying process, the
droplet components were separated into two distinct phases of ethanol-water and water-
maltodextrin. As the drying continued, precipitation of maltodextrin solid occurred forming
the outer crust layer. The prevailing liquid retain within the maltodextrin solid could very
likely be the ethanol-water phase that was not evaporated due to solid formation. It is of
great interest to explore the drying behavior and retention capability of maltodextrin via the
AVP drying approach in more detail as it allows for more efficient encapsulation
applications. Nevertheless, this study is not within the scope of this project and will be
subjected to future work in this field.
4.3 Modelling of the Simultaneous Absorption and Evaporation Process of
the Droplet under AVP
4.3.1 Theoretical Modelling Method
A theoretical model is developed based on a mass and energy balance analysis of the
droplet. By evaluating the heat and mass transfer of the droplet, the mass and temperature
of the droplet at any given time can be modelled. It is noteworthy that in this analysis the
mass and heat transfer of ethanol and water are evaluated independently. The overall mass
of the droplet at any given time (s) can be calculated using the Euler method:
where is the total droplet mass at a given time, n
is the predicted total droplet mass based on the instantaneous mass
change
Jiunn Yuan Tan (21637520) Page 109
is the net mass transfer of the droplet at a given time, n
Similarly, for the temperature of the droplet at any given time (s) can be calculated:
Mass Balance Analysis:
The instantaneous mass change of the droplet is determined by the mass transfer of
water and ethanol within the droplet, which can be expressed as:
The mass transfer equation for the evaporation of water from droplet is:
As the driving force of ethanol absorption into the droplet during the initial stage of the
drying process is due to the higher concentration of ethanol in the surroundings compared
to the droplet, the mass transfer of ethanol into the droplet is:
The term A is the area of the droplet which is calculated based on a spherical droplet. The
term represents the activity coefficient of ethanol and water at the surface of the droplet.
This term was obtained using the UNIFAC model provided by the UNIFAC group contribution
Jiunn Yuan Tan (21637520) Page 110
through Dortmund Data Bank and it is to account for the interaction between water and
ethanol at the surface of the droplet 99.
The surface and ambient water and ethanol vapour concentration is given by:
RH represents the relative humidity of water and ethanol in the ambient bulk convective
medium. The partial saturation pressures can be determined using the Antoine equation
based on the wet bulb temperature. The instantaneous concentration of water and ethanol
on the surface of the droplet is determined by the partial saturation pressures of each
component at any given time evaluated based on Raoult’s Law.
where is the mol fraction of water
is the mol fraction of ethanol
The heat transfer coefficient, hT and mass transfer coefficient, hm for a spherical droplet is
expressed as:
From the Ranz-Marshall correlation 58b,
Jiunn Yuan Tan (21637520) Page 111
where
Substitution of equations 2-13 into equation 1 allows for the calculation of the mass of a
droplet with a known initial mass throughout the drying process.
Energy Balance Analysis:
The overall energy balance of the droplet involves the combined effect of heat
transfer, evaporation of water, evaporation or absorption of ethanol and the effect of
radiation within the droplet.
Jiunn Yuan Tan (21637520) Page 112
Therefore, the instantaneous temperature change of the droplet is determined by the heat
transfer of water and ethanol within the droplet, which can be expressed as:
where ε is the emissivity of the droplet which is approximate to 0.95
σ is the Stefan-Boltzmann Constant which is equals to 5.6703 x 10-8 (W/m2K4)
is the latent heat of vaporization (J/kg)
is the specific heat of the droplet (J/kg.K)
The heat transfer coefficient, hT for a spherical droplet is expressed as:
where k is the thermal conductivity of convective medium (W/m.K)
From the Ranz-Marshall correlation 58b,
where
Jiunn Yuan Tan (21637520) Page 113
Substitution of equations 15-18 into equation 2 allows for the calculation of the
temperature of a droplet, with ambient temperature as the drying temperature throughout
the drying process.
Jiunn Yuan Tan (21637520) Page 114
4.3.2 Comparison of Mass Change Profile of Pure Water Droplet and Maltodextrin
Solution Droplet Dried under AVP
-0.5
0
0.5
1
1.5
2
2.5
0 200 400 600 800 1000 1200
Mass (mg)
Time (s)
-0.5
0
0.5
1
1.5
2
2.5
0 200 400 600 800 1000 1200
Mass (mg)
Time (s)
(a)
(b)
Maltodextrin
Water
Maltodextrin
Water
Maltodextrin
Water
Jiunn Yuan Tan (21637520) Page 115
Figure 4.10: Mass change profile of the drying of water and maltodextrin DE 10 solution (5
wt%) at EAH: (a) 0.05 kg/kg db; (b) 0.065 kg/kg db and (c) 0.08 kg/kg/db.
The first part of the work was to experimentally assess if the presence of 5 wt%
solute significantly affect the drying behaviour of the droplet or not. The mass change
experiments were first conducted on water and maltodextrin DE 10 solution with initial
weight concentration of 5 wt % at three different EAH conditions. Based on the results
shown above, the mass change profile of the water and maltodextrin solution of 5 wt% at a
particular EAH was similar. The deviation in the final mass after complete drying was due to
the presence of the fully dried maltodextrin solid, as described earlier in Section 4.2.1.
Nevertheless, it was evident that the precipitation of maltodextrin did not affect the
absorption and evaporation profile of water and ethanol. Therefore, it was justifiable that a
water-ethanol system (without the influence of solute) was being considered in the analysis
on the modelling of the simultaneous absorption and evaporation process of the droplet.
-0.5
0
0.5
1
1.5
2
2.5
3
0 200 400 600 800 1000 1200
Mass (mg)
Time (s)
(c) Maltodextrin
Water
Jiunn Yuan Tan (21637520) Page 116
4.3.3 Comparison of Experimental and Theoretical Modelling of the Mass Change
Profile of AVP Process
-0.0005
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0.0035
0.004
0.0045
0 200 400 600 800 1000 1200 1400 1600 1800
Mass (g)
Time (s)
Experimental
Model
(a)
Jiunn Yuan Tan (21637520) Page 117
Figure 4.11: Comparison of droplet mass change profile for drying of a water droplet
under AVP measured experimentally and predicted by the model for ethanol absolute
humidity: (a) 0.08 kg/kg db; (b) 0.065 kg/kg db and (c) 0.05 kg/kg db.
-0.0005
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0.0035
0 200 400 600 800 1000 1200 1400 1600 1800
Mass (g)
Time (s)
Experimental
Model
-0.0005
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0 500 1000 1500 2000
Mass (g)
Time (s)
Model
Experimental
(b)
(c)
Jiunn Yuan Tan (21637520) Page 118
Mass change experiments were conducted on a pure water droplet under AVP
drying for three different ethanol absolute humidity conditions. Figure 4.11 shows the
comparison between the mass change by experimental measurement and predicted by the
model. Based on the experimental measurement, only the high ethanol absolute humidity
of 0.08 kg/kg db resulted in observable ethanol absorption, as the rate of absorption of
ethanol was higher than the rate of evaporation of water during the initial stage of the AVP
drying process. For the mid and low ethanol absolute humidity of 0.065 kg/kg db and 0.05
kg/kg db, no obvious peak was observed for the experimentally measured mass change
profile. The droplet continuously experienced a reduction in mass since the beginning of the
AVP drying process, as the ethanol concentration gradient between the bulk medium and
the droplet was not high enough that the rate of absorption of ethanol was lower than the
rate of evaporation of water. Droplet exposed to lower ethanol humidity dried faster as
smaller amount of ethanol absorbed into the droplet resulted in a shorter evaporation time.
As observed, the model greatly overestimated the ethanol absorption into the droplet.
Higher ethanol absolute humidity resulted in a more significant overestimation. This large
overestimation of ethanol absorption within the droplet by the model resulted in a similar
overestimation for the drying time, as longer evaporation time is required for the larger
amount of absorbed ethanol within the droplet. In view of this deviation, further analysis
were conducted to analyse the evaporative behaviour of each component independently.
Some factors which may have affected the prediction were: mass transfer depression for
both components and UNIFAC model.
Jiunn Yuan Tan (21637520) Page 119
4.3.4 Comparison of Experimental and Theoretical Modelling of the Mass Change
and Temperature Profile of Pure Droplet
Figure 4.12: Comparison of mass change and droplet temperature by experimental
measurement and model prediction for evaporation of a pure water droplet with dry
nitrogen at 40oC.
-0.0005
0
0.0005
0.001
0.0015
0.002
0.0025
0 100 200 300 400 500 600 700 800 900
Mass (g)
Time (s)
Model 3 micro L
Model 2 micro L
Exprimental 3 micro L
Experimental 2 micro L
0
5
10
15
20
25
30
35
40
0 100 200 300 400 500 600
Temperature ( °C )
Time (s)
Model
Experimental 2 micro L
Jiunn Yuan Tan (21637520) Page 120
Firstly, the effect of mass transfer depression for water was assessed. Mass change
experiments and droplet temperature measurement were conducted on pure water droplet
of two different initial sizes. Figure 4.12 shows the comparison of the mass change profile
and droplet temperature profile during drying under dry nitrogen gas condition at 40oC
respectively between the experimental data and the model. The rate of evaporation was the
highest during the initial stage of drying and gradually reduced as the droplet shrunk and
eventually dried over time. This was due to the higher surface area of droplet exposed to
the drying medium in the beginning. The shrinkage of water droplet over time due to
evaporation reduced the droplet exposed surface area resulting in a gradual reduction of
evaporation rate. Initial comparison has shown that the Ranz-Marshall correlation has over
predicted the evaporation of the water droplet. It was found that a mass transfer
depression factor was required to account for expansion of thermal and mass boundary
layer due to the high mass flux evaporation of water droplet. The thickening of thermal and
mass boundary layer created a resistance to and from droplet; hence, depressing the overall
evaporation rate. This phenomenon has been reported by Woo et. al. 11and elucidated by
Kar and Chen 75. The high concentration gradient between the dry nitrogen gas and the
water droplet could have induced the high mass flux evaporation. Further analysis revealed
a depression factor of 0.65 was required to account for this occurrence. As observed, the
model provided a good prediction of the mass change and temperature of the droplet over
time.
Jiunn Yuan Tan (21637520) Page 121
Figure 4.13: Comparison of mass change and ethanol droplet temperature by
experimental measurement and model prediction for drying of a water droplet with dry
nitrogen gas at 40oC.
Mass change experiments and droplet temperature measurement were then
conducted on pure ethanol droplet of three different initial sizes. Figure 4.13 shows the
-0.0005
0
0.0005
0.001
0.0015
0.002
0 50 100 150 200 250 300 350 400 450
Mass (g)
Time (s)
Experimental 4 micro L
Experimental 3 micro L
Experimental 2 micro L
Model 4 micro L
Model 3 micro L
Model 2 micro L
0
5
10
15
20
25
30
35
40
-10 10 30 50 70 90 110 130 150
Temperature ( °C )
Time (s)
Model
Experimental 3 micro L
Experimental 4 micro L
Jiunn Yuan Tan (21637520) Page 122
comparison of the droplet temperature profile and mass change profile during drying under
dry nitrogen gas condition at 40oC respectively between the experimental data and the
model. Similar to the pure water droplet, the rate of evaporation was the highest in the
initial stage of drying and gradually decreased due to the shrinkage to the droplet exposed
surface are. The wet bulb temperature recorded was much lower due to the higher rate of
evaporation of ethanol droplet compare to water droplet, as ethanol is more volatile.
Contrary to the pure water droplet, there was no depression for the mass transfer of the
evaporation of ethanol.
Jiunn Yuan Tan (21637520) Page 123
4.3.5 Comparison of Experimental and Theoretical Modelling of the Mass Change
of Water-Ethanol Droplet
Figure 4.14: Comparison of mass change by experimental measurement and model
prediction for drying of an ethanol-water droplet with dry nitrogen gas at 25oC.
We then further assessed the effect of using the UNIFAC approach. Experimental
work was then extended to investigate the evaporation of a pure ethanol-water droplet of
three different ratios. Figure 4.14 shows the comparison of the droplet mass change profile
during drying under dry nitrogen gas condition at 25oC respectively between the
experimental data and the model. The model evaluates the evaporation of water and
-0.0005
0
0.0005
0.001
0.0015
0.002
0 100 200 300 400 500 600 700 800 900 1000
Mass (g)
Time (s)
Model (70% water, 30% ethanol)
Model (65% water, 35% ethanol)
Model (60% water, 40% ethanol)
Experimental (70% water, 30% ethanol)
Experimental (65% water, 35% ethanol)
Experimental (60% water, 40% ethanol)
Jiunn Yuan Tan (21637520) Page 124
ethanol component within the droplet independently. The same depression factors were
adopted from the mass transfer of pure water (0.65) and pure ethanol (1) droplet for the
simultaneous mass transfer of both components. As observed, the model predicted the
mass change profile of the ethanol-water droplet fairly well for all three different ratios.
Interestingly, it was found that the use of ideal solution assumption instead of UNIFAC
provided a better fit for the modelling of the evaporation of ethanol-water droplet. This is
probably due to the fact that the ethanol-water droplet mixture was thoroughly mixed,
resulting in an ideal solution state. Based on this result, an attempt was done to remove the
UNIFAC model from the original AVP predictive model. However, the assumption of ideal
solution caused the AVP predictive model to deviate even further from the experimental
data as it overestimated the absorption of ethanol into the droplet by a larger magnitude.
This observation indicated that there is a possibility the use of the UNIFAC model might be
inadequate or inappropriate to account for the water-ethanol interaction during the AVP
drying process.
Jiunn Yuan Tan (21637520) Page 125
4.3.6 Comparison of Experimental and Theoretical Modelling with Mass Transfer
Depression of the Mass Change Profile of AVP Process
-0.0005
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0.0035
0 200 400 600 800 1000 1200 1400 1600 1800
Mass (g)
Time (s)
Overall Model
Experimental
-0.0005
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0 200 400 600 800 1000 1200 1400 1600 1800
Mass (g)
Time (s)
Overall Model
Experimental
(a)
(b)
Jiunn Yuan Tan (21637520) Page 126
Figure 4.15: Comparison of mass change by experimental measurement and model
prediction (with depression) for drying of water droplet under AVP for ethanol absolute
humidity: (a) 0.08 kg/kg db; (b) 0.065 kg/kg db and (c) 0.05 kg/kg db.
In view of that the models are in agreement with the evaporative behaviour of each
component, it was hypothesize that the deviation could be due to the absorption behaviour
of ethanol during AVP. This could be due to the fact that the evaporation of water created a
diffusion barrier which impeded the absorption of ethanol. With that in mind, the mass
transfer for the absorption of ethanol was depressed to account for this phenomenon.
Figure 4.15 shows the comparison of mass change measured experimentally and the mass
change predicted by the model with a depression factor. The depression factor for all three
ethanol humidity condition varied with a factor of 0.5, 0.35 and 0.2 used for ethanol
absolute humidity of 0.08 kg/kg db, 0.065 kg/kg db and 0.05 kg/kg db respectively. Despite
-0.0005
0
0.0005
0.001
0.0015
0.002
0.0025
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Mass (g)
Time (s)
Overall Model
Experimental
(c)
Jiunn Yuan Tan (21637520) Page 127
introducing the depression factor, the model still did not provide a good fit with the
experimental data.
4.3.7 Discussion
Initial comparison of the model developed shows a great deviation from the
experimental measurements. The model largely over predicted the absorption of ethanol
into the droplet. Further analysis have shown that the model predicted the evaporation of
pure droplets of water, ethanol and water-ethanol very well with the incorporation of a
mass depression factor for the mass transfer of water due to the high mass flux evaporation.
The incorporation of a mass transfer depression for the absorption of ethanol into the
droplet was insufficient to account for the overestimation. Therefore, it was concluded that
merely evaluating the mass transfer of ethanol and water to and from the droplet
independently does not provide a good model to predict the absorption and evaporation
behaviour of this ethanol-water system. The incorporation of the UNIFAC model was
insufficient to account for the interaction between water and ethanol for this simultaneous
absorption and evaporation process. Subsequent discussions attempt to further elucidate
the reasons for this deviation and outline several other possible considerations to be
incorporated in the model.
Firstly, it is beneficial to break down the analysis of the overall physical phenomenon
to the ethanol absorption and the ethanol evaporation stage. During the ethanol absorption
stage, the actual absorption of ethanol into the droplet was much smaller than the
absorption predicted by the model. This could be attributed to the counter diffusion of
water which may have impeded the absorption of ethanol into the droplet. As the ethanol
and water vapour diffused against each other at the interfacial of the liquid droplet, the
Jiunn Yuan Tan (21637520) Page 128
evaporation of water resulted in a slight diffusion barrier that reduced the effective
diffusion of ethanol vapour into the droplet. Evidently, the introduction of the mass transfer
depression factor was insufficient to account for this phenomenon. A correction factor was
developed by Pierre Grenier for the counter diffusion of ammonia and water vapour
through stagnant air in an absorption tower system based on the Stefan-Maxwell diffusion
equation 122. The development of such correction factor for the physical phenomenon of
AVP could provide a better description of the ethanol absorption stage.
For the ethanol evaporation stage, the large overestimation of ethanol absorption
had resulted in a longer evaporation time due to higher volume of ethanol in the droplet. In
addition, the use of ethanol vapour during the process could have induced a more rapid
evaporation process. The effect of liquid ethanol concentration in increasing the
evaporation rate of an ethanol-water droplet has been reported by several studies 123.
However, the effect of ethanol vapour condensing and absorbing onto a liquid water droplet
on the overall evaporation rate has not been studied. Besides that, the diffusion of water
and ethanol within the droplet was assumed to obey FIck's Law. The diffusion behaviour of
the components could deviate from the Fickian theory due to the presence of nitrogen and
the counter-diffusion of water and ethanol vapour. This would introduce curious
phenomenon such as osmotic diffusion, reverse diffusion and diffusion barrier which was
previously anticipated by Toor 124. Such ambiguity could be accounted for using the
Maxwell-Stefan approach to mass transfer which was reviewed in great detail by Krishna
and Wesselingh 125. One possible method is to introduce an experimentally determined
Maxwell-Stefan factor to account for the mass transfer depression, resulted by this
phenomenon. It is important to note that as the AVP drying model is a transient process,
Jiunn Yuan Tan (21637520) Page 129
this factor could vary over the drying period and might be dependent on the concentrations
of water and ethanol within the droplet. Another consideration to improve the accuracy of
the model is to use a distributed-parameter drying model instead of the lump dyring model
approach, particularly when evaluating the absorption stage of ethanol into the droplet. The
spatial distribution of ethanol vapour within the bulk convective medium as well as within
the droplet should be evaluated as they might affect the overall absorption of ethanol into
the droplet. An analysis could also be done to experimentally determine the effective
droplet surface absorption area during the drying process. Extensive experimental work to
evaluate the applicability of incorporating these theories in the application of AVP drying
should be conducted in order to generate a more accurate model to describe the unique
physical phenomenon.
Jiunn Yuan Tan (21637520) Page 130
Chapter 5
CONCLUSION & RECOMMENDATIONS
This thesis explored the fundamental mechanism for micro-particle formation by the
antisolvent vapour precipitation (AVP) drying technique for potential application in
encapsulation and drug delivery 126. The resultant particle morphology can be controlled by
altering the formulation and drying conditions, particularly the antisolvent vapour humidity,
thus establishing a direct relationship between the process parameters and the type of
functional particles produced. An attempt was made to develop a model for the AVP drying
process to describe the physical phenomenon. The main scientific outcomes from this work
are outlined as follows.
5.1 Conclusions
Saccharides can be precipitated under a wider operating range with the antisolvent
vapour precipitation technique when compared to disaccharides. This is due to the long
chain structure preventing crystallization which inhibits the antisolvent vapour precipitation
process. Smooth, porous, microsphere network and microspheres particle structures were
obtained by the AVP method depending on the ethanol absorption rate and the maximum
concentration of ethanol in the droplet. A three stage mechanism was proposed for the
formation of the ultrafine spherical particles. The initial stage is phase separation step
resulting in the formation of bubbles or an emulsion in the ethanol-water system. This is an
Jiunn Yuan Tan (21637520) Page 131
interesting observation due to the fact that ethanol and water should be fully soluble,
contrary to the emulsion formation resulting in the bubbles. Therefore, it is possible that the
presence of maltodextrin within the system has induced a liquid phase separation between
water and ethanol into a water-maltodextrin phase and water-ethanol phase. The second
stage involved inversion of the phases due to further absorption of the antisolvent. The
third stage involved shrinkage of the water-maltodextrin phase leading to the formation of
spherical particles related to the surface tension ‘pinch-off’ mechanism. Higher ERH, higher
EAH and lower initial weight concentration are parameter trends that were found to favour
the formation of microspherical particles upon drying. This work has provided qualitative
insight into the antisolvent vapour precipitation process to produce ultrafine spherical
particles. The effect of particle size on the drying behaviour is still unclear. An observable
particle size trend from this work is a decreasing particle size produced for maltodextrin DE
10, maltodextrin DE 18 and lastly maltose, which could be due to the physio-chemical effect
of these materials such as the polymer chain length. It will be of interest to expand the
range of polymer chain length investigated in this work or to extend the drying of AVP to
other materials in order to better understand this drying behaviour. Another approach
could be to conduct the AVP drying of the similar saccharide materials within the
microsphere region at increasing ethanol humidity. It will be interesting to see the effect of
increasing ethanol concentration on the resulting particle size produced. Subsequent work
in this thesis includes a quantitative analysis on the droplet drying behaviour resulting in the
precipitation of the porous, microsphere network and microsphere particles.
Droplet mass measurement were conducted throughout the entire AVP drying
process for all three conditions resulting in the porous, microsphere network and
Jiunn Yuan Tan (21637520) Page 132
microsphere particles. Only the microsphere run exhibited an observable ethanol
absorption peak, further reinforcing the postulate that higher absorption rate and higher
maximum concentration of ethanol within the droplet were the key factors in producing the
microsphere. A simple analysis suggests the primary factor dictating the type of particles
produced from the AVP process was the maximum concentration of ethanol within the
droplet. In addition, a unique liquid retention behaviour was also observed for the porous
network and microsphere network particles.
Quantitative analysis was further extended to developing an AVP drying model to
describe the simultaneous absorption and evaporation of ethanol within the droplet. The
effect of solute with initial weight concentration of 5 wt% on the droplet drying behaviour
was found to be negligible. A drying model developed based on the lump model approach
using heat and mass transfer evaluations, Raoult's Law and UNIFAC equation was found to
overestimate the absorption behaviour of ethanol. Further experimental work and mass
transfer analysis indicated that the model provides a good prediction for the evaporative
behaviour of the droplet. Therefore, it was concluded that the deviation between the model
prediction and the experimentally measured mass change of the droplet was largely due to
the overestimation of the ethanol absorption into the droplet that may have caused by the
counter diffusion of water and ethanol within the droplet. Several considerations were then
proposed to account for this deviation.
5.2 Recommendations
The results from this study warrant further investigation of the fundamental of
antisolvent vapour precipitation (AVP) drying and its relevant applications. Future work in
research and development of this field may need to address the following:
Jiunn Yuan Tan (21637520) Page 133
1. Exploring the applicability of AVP technique on a wider range of materials: Previous
work and current work have shown AVP's applicability in lactose, protein based
material (WPI) and saccharides in producing microspheres. It will be of interest to
expand the use of AVP to a larger variety of materials such as drugs, vitamins and
minerals for a wider range of applicability in the food and pharmaceutical industry.
2. Investigating the application of the AVP produced particles: It will be beneficial to
investigate the application of the microspherical particles produced under AVP
drying as an adjuvant or in drug delivery applications. Besides that, the retention
behaviour exhibited by the microsphere network and porous network structure
obtained from the drying of saccharides under AVP is worth exploring in greater
detail for encapsulation applications.
3. Further development on the model governing the physical phenomenon of the AVP
process: A model should be developed to describe the simultaneous absorption and
evaporation behaviour of the antisolvent within the droplet based on the
recommendations and considerations outline in this work. This model will be the key
to understanding the fundamental mechanism of the AVP process and also provides
a basis for future scaling up application.
4. Scaling up of the AVP technique: Once the fundamental understanding of the AVP
process has been well established, it will be of great interest to scale up this process
for industrial application. It will be interesting to investigate the feasibility of
incorporating the AVP technique in conventional spray drying units. This will include
Jiunn Yuan Tan (21637520) Page 134
experiments involving the usage of dimensionless groups in order to describe the
effect of different drying regimes on the drying behaviour of the AVP process for
better control within a commercial spray dryer.
Jiunn Yuan Tan (21637520) Page 135
5.3 List of Publications
Journal Publications:
1. J.Y.Tan, V.M. Tang, J. Nguyen, S. Chew, S. Mansouri, K. Hapgood, X.D.
Chen, M.W. Woo; Unveiling the Mechanism of Antisolvent Vapour
Precipitation in Producing Ultrafine Spherical Particles. Powder Technology,
2015, doi: 10.1016/j.powtec.2015.01.059
Conference Proceedings:
1. J.Y.Tan, V.M. Tang, J. Nguyen, S. Chew, S. Mansouri, K. Hapgood, X.D.
Chen, M.W. Woo; Formation of Microspherical Particles from Carbohydrate
Polymers via Antisolvent Vapour Precipitation. International Drying
Symposium, 2014, Lyon, France.
2. J.Y.Tan, L.C. Lum, M.G. Lee, S. Mansouri, K. Hapgood, X.D. Chen, M.W.
Woo; Improving the Dissolution Rate of Folic Acid via the Antisolvent Vapour
Precipitation. International Conference of Pharmaceutical Science
Engineering, 2014, Melbourne, Australia.
Jiunn Yuan Tan (21637520) Page 136
REFERENCES
1. Cal, K.; Sollohub, K., Spray drying technique. I: Hardware and process parameters.
Journal of Pharmaceutical Sciences 2010, 99 (2), 575-586.
2. Susan, C. W.; Metin, C., Spray Drying and Pharmaceutical Applications. In Handbook
of Pharmaceutical Granulation Technology, Second Edition, CRC Press: 2005; pp 129-158.
3. Vehring, R.; Foss, W. R.; Lechuga-Ballesteros, D., Particle formation in spray drying.
Journal of Aerosol Science 2007, 38 (7), 728-746.
4. Li, X.; Anton, N.; Arpagaus, C.; Belleteix, F.; Vandamme, T. F., Nanoparticles by spray
drying using innovative new technology: The Büchi Nano Spray Dryer B-90. Journal of
Controlled Release 2010, 147 (2), 304-310.
5. Lee, S. H.; Heng, D.; Ng, W. K.; Chan, H.-K.; Tan, R. B. H., Nano spray drying: A novel
method for preparing protein nanoparticles for protein therapy. International Journal of
Pharmaceutics 2011, 403 (1–2), 192-200.
6. Mansouri, S.; Fu, N.; Woo, M. W.; Chen, X. D., Uniform Amorphous Lactose
Microspheres Formed in Simultaneous Convective and Dehydration Antisolvent
Precipitation under Atmospheric Conditions. Langmuir 2012, 28 (39), 13772-13776.
7. Mansouri, S.; Chin, G. Q.; Ching, T. W.; Woo, M. W.; Fu, N.; Chen, X. D., Precipitating
smooth amorphous or pollen structured lactose microparticles. Chemical Engineering
Journal 2013, 226 (0), 312-318.
8. Sinko, P. J., Martin's Physical Pharmacy and Pharmaceutical Sciences: Physical
Chemical and Biopharmaceutical Principles in the Pharmaceutical Sciences. Lippincott
Williams & Wilkins: 2006.
Jiunn Yuan Tan (21637520) Page 137
9. Hodgkinson, J.; Tatam, R. P., Optical gas sensing: a review. Measurement Science and
Technology 2013, 24 (1), 012004.
10. Terencio, T.; Di Renzo, F.; Berthomieu, D.; Trens, P., Adsorption of Acetone Vapor by
Cu-BTC: An Experimental and Computational Study. The Journal of Physical Chemistry C
2013, 117 (49), 26156-26165.
11. Woo, M. W.; Fu, N.; Che, L.; Chen, X. D., Evaporation of Pure Droplets in the
Convective Regime Under High Mass Flux. Drying Technology 2011, 29 (14), 1628-1637.
12. (a) Sollohub, K.; Cal, K., Spray drying technique: II. Current applications in
pharmaceutical technology. Journal of Pharmaceutical Sciences 2010, 99 (2), 587-597; (b)
Vehring, R., Pharmaceutical Particle Engineering via Spray Drying. Pharm Res 2008, 25 (5),
999-1022.
13. (a) Estevinho, B. N.; Rocha, F.; Santos, L.; Alves, A., Microencapsulation with chitosan
by spray drying for industry applications – A review. Trends in Food Science & Technology
2013, 31 (2), 138-155; (b) Gharsallaoui, A.; Roudaut, G.; Chambin, O.; Voilley, A.; Saurel, R.,
Applications of spray-drying in microencapsulation of food ingredients: An overview. Food
Research International 2007, 40 (9), 1107-1121; (c) Rathore, S.; Desai, P. M.; Liew, C. V.;
Chan, L. W.; Heng, P. W. S., Microencapsulation of microbial cells. Journal of Food
Engineering 2013, 116 (2), 369-381.
14. Patel, R.; Patel, M.; Suthar, A., Spray drying technology: an overview. Indian Journal
of Science and Technology 2009, 2 (10), 44-47.
15. (a) Keshani, S.; Daud, W. R. W.; Nourouzi, M. M.; Namvar, F.; Ghasemi, M., Spray
drying: An overview on wall deposition, process and modeling. Journal of Food Engineering
2015, 146 (0), 152-162; (b) Mujumdar, A. S.; Huang, L.-X.; Chen, X. D., An overview of the
recent advances in spray-drying. Dairy Science & Technology 2010, 90 (2-3), 211-224; (c)
Jiunn Yuan Tan (21637520) Page 138
Tsotsas, E.; Mujumdar, A. S., Modern Drying Technology Vol. 1 Computational Tools at
Different Scales. Drying Technology 2008, 26 (6), 812-814.
16. Elversson, J.; Millqvist-Fureby, A.; Alderborn, G.; Elofsson, U., Droplet and particle
size relationship and shell thickness of inhalable lactose particles during spray drying.
Journal of Pharmaceutical Sciences 2003, 92 (4), 900-910.
17. Walzel, P., Influence of the Spray Method on Product Quality and Morphology in
Spray Drying. Chemical Engineering & Technology 2014, 37 (6), 906-906.
18. Parikh, D. M., Handbook of pharmaceutical granulation technology. CRC Press: 1997;
Vol. 198.
19. Marshall, W. R., Atomization and Spray Drying. American Institute of Chemical
Engineers: 1954.
20. Filkova, I.; Mujumdar, A. S., Industrial spray drying systems. Handbook of industrial
drying 1995, 1, 263-308.
21. Thorat, A. A.; Dalvi, S. V., Liquid antisolvent precipitation and stabilization of
nanoparticles of poorly water soluble drugs in aqueous suspensions: Recent developments
and future perspective. Chemical Engineering Journal 2012, 181–182 (0), 1-34.
22. (a) Hu, J.; Ng, W. K.; Dong, Y.; Shen, S.; Tan, R. B. H., Continuous and scalable process
for water-redispersible nanoformulation of poorly aqueous soluble APIs by antisolvent
precipitation and spray-drying. International Journal of Pharmaceutics 2011, 404 (1–2), 198-
204; (b) Zhang, H.-X.; Wang, J.-X.; Zhang, Z.-B.; Le, Y.; Shen, Z.-G.; Chen, J.-F., Micronization
of atorvastatin calcium by antisolvent precipitation process. International Journal of
Pharmaceutics 2009, 374 (1–2), 106-113; (c) Zu, Y.; Sun, W.; Zhao, X.; Wang, W.; Li, Y.; Ge, Y.;
Liu, Y.; Wang, K., Preparation and characterization of amorphous amphotericin B
Jiunn Yuan Tan (21637520) Page 139
nanoparticles for oral administration through liquid antisolvent precipitation. European
Journal of Pharmaceutical Sciences 2014, 53 (0), 109-117.
23. (a) Joye, I. J.; McClements, D. J., Production of nanoparticles by anti-solvent
precipitation for use in food systems. Trends in Food Science & Technology 2013, 34 (2), 109-
123; (b) Su, K.; Liu, F.; Chen, J., Preparation of high performance carbon-coated LiMnPO4
nanocomposite by an acetate-assisted antisolvent precipitation method. Journal of Power
Sources 2013, 232 (0), 234-239.
24. Dalvi, S. V.; Dave, R. N., Controlling Particle Size of a Poorly Water-Soluble Drug Using
Ultrasound and Stabilizers in Antisolvent Precipitation. Industrial & Engineering Chemistry
Research 2009, 48 (16), 7581-7593.
25. (a) Beck, C.; Sievens-Figueroa, L.; Gärtner, K.; Jerez-Rozo, J. I.; Romañach, R. J.; Bilgili,
E.; Davé, R. N., Effects of stabilizers on particle redispersion and dissolution from polymer
strip films containing liquid antisolvent precipitated griseofulvin particles. Powder
Technology 2013, 236 (0), 37-51; (b) Kakran, M.; Sahoo, N. G.; Tan, I.-L.; Li, L., Preparation of
nanoparticles of poorly water-soluble antioxidant curcumin by antisolvent precipitation
methods. J Nanopart Res 2012, 14 (3), 1-11.
26. (a) Kalani, M.; Yunus, R., Application of supercritical antisolvent method in drug
encapsulation: a review. International Journal of Nanomedicine 2011, 6, 1429-1442; (b)
Pasquali, I.; Bettini, R.; Giordano, F., Supercritical fluid technologies: An innovative approach
for manipulating the solid-state of pharmaceuticals. Advanced Drug Delivery Reviews 2008,
60 (3), 399-410; (c) Yang, L.; Sun, Z.; Zu, Y.; Zhao, C.; Sun, X.; Zhang, Z.; Zhang, L.,
Physicochemical properties and oral bioavailability of ursolic acid nanoparticles using
supercritical anti-solvent (SAS) process. Food Chemistry 2012, 132 (1), 319-325.
Jiunn Yuan Tan (21637520) Page 140
27. Reverchon, E., Supercritical antisolvent precipitation of micro- and nano-particles.
The Journal of Supercritical Fluids 1999, 15 (1), 1-21.
28. Braeuer, A.; Dowy, S.; Torino, E.; Rossmann, M.; Luther, S. K.; Schluecker, E.; Leipertz,
A.; Reverchon, E., Analysis of the supercritical antisolvent mechanisms governing particles
precipitation and morphology by in situ laser scattering techniques. Chemical Engineering
Journal 2011, 173 (1), 258-266.
29. (a) Jung, J.; Perrut, M., Particle design using supercritical fluids: Literature and patent
survey. The Journal of Supercritical Fluids 2001, 20 (3), 179-219; (b) Yeo, S.-D.; Kiran, E.,
Formation of polymer particles with supercritical fluids: A review. The Journal of
Supercritical Fluids 2005, 34 (3), 287-308.
30. Imsanguan, P.; Pongamphai, S.; Douglas, S.; Teppaitoon, W.; Douglas, P. L.,
Supercritical antisolvent precipitation of andrographolide from Andrographis paniculata
extracts: Effect of pressure, temperature and CO2 flow rate. Powder Technology 2010, 200
(3), 246-253.
31. Careno, S.; Boutin, O.; Badens, E., Drug recrystallization using supercritical anti-
solvent (SAS) process with impinging jets: Effect of process parameters. Journal of Crystal
Growth 2012, 342 (1), 34-41.
32. Erriguible, A.; Fadli, T.; Subra-Paternault, P., A complete 3D simulation of a
crystallization process induced by supercritical CO2 to predict particle size. Computers &
Chemical Engineering 2013, 52 (0), 1-9.
33. (a) Reverchon, E.; De Marco, I., Mechanisms controlling supercritical antisolvent
precipitate morphology. Chemical Engineering Journal 2011, 169 (1–3), 358-370; (b)
Reverchon, E.; De Marco, I.; Adami, R.; Caputo, G., Expanded micro-particles by supercritical
antisolvent precipitation: Interpretation of results. The Journal of Supercritical Fluids 2008,
Jiunn Yuan Tan (21637520) Page 141
44 (1), 98-108; (c) Reverchon, E.; Torino, E.; Dowy, S.; Braeuer, A.; Leipertz, A., Interactions
of phase equilibria, jet fluid dynamics and mass transfer during supercritical antisolvent
micronization. Chemical Engineering Journal 2010, 156 (2), 446-458.
34. De Marco, I.; Reverchon, E., Influence of pressure, temperature and concentration
on the mechanisms of particle precipitation in supercritical antisolvent micronization. The
Journal of Supercritical Fluids 2011, 58 (2), 295-302.
35. Mansouri, S.; Fu, N.; Woo, M. W.; Chen, X. D., Uniform Amorphous Lactose
Microspheres Formed in Simultaneous Convective and Dehydration Antisolvent
Precipiatation under Atmospheric Conditions. American Chemical Society 2012, 10 (1021).
36. Mansouri, S.; Woo, M. W.; Chen, X. D., Making Uniform Whey, Lactose, and
Composite Lactose–Whey Particles from the Dehydration of Single Droplets with Antisolvent
Vapor. Drying Technology 2013, 31 (13-14), 1570-1577.
37. Yue, S.; Mansouri, S.; Woo, M. W.; Chen, X. D., Antisolvent Vapour Precipitation of
Droplets with Multi-Components: Non-soluble Encapsulation and Simultaneous
Precipitation of Soluble Materials. Monash University, 2012.
38. Tsutsui, K.; Katsumata, T.; Yamamoto, Y.; Fukatsu, H.; Yoshimizu, H.; Kinoshita, T.;
Tsujita, Y., The structural organization in syndiotactic polystyrene film induced by toluene
vapour sorption. Polymer 1999, 40 (13), 3815-3819.
39. Meyer, C. W.; Miller, W. W.; Ripple, D. C.; Scace, G. E., Uncertainty Budget for the
NIST Hybrid Humidity Generator. Int J Thermophys 2012, 33 (8-9), 1488-1499.
40. Statheropoulos, M.; Pallis, G. C.; Mikedi, K.; Giannoukos, S.; Agapiou, A.; Pappa, A.;
Cole, A.; Vautz, W.; Thomas, C. L. P., Dynamic Vapor Generator That Simulates Transient
Odor Emissions of Victims Entrapped in the Voids of Collapsed Buildings. Analytical
Chemistry 2014, 86 (8), 3887-3894.
Jiunn Yuan Tan (21637520) Page 142
41. (a) Hwang, T.-J.; Jiang, S.-J., Determination of Cadmium by Flow Injection Isotope
Dilution Inductively Coupled Plasma Mass Spectrometry With Vapour Generation Sample
Introduction. Journal of Analytical Atomic Spectrometry 1997, 12 (5), 579-584; (b) Yang, H.-J.;
Jiang, S.-J., Hydride generation inductively coupled plasma mass spectrometric detection of
lead compounds separated by liquid chromatography. Journal of Analytical Atomic
Spectrometry 1995, 10 (11), 963-967.
42. Campos, M., Electrical response of polypyrrole films doped with dodecylbenzene
sulfonic acid to acetone vapor. Journal of Applied Polymer Science 2011, 121 (5), 2518-2525.
43. Lin, S. X. Q.; Chen, X. D., An effective laboratory air humidity generator for drying
research. Journal of Food Engineering 2005, 68 (1), 125-131.
44. Temime, B.; Francois, S.; Monod, A.; Wortham, H., An experimental set up of a PAH
vapour generator and its use to test an annular denuder. Environmental Pollution 2002, 120
(3), 609-616.
45. Collins, G. E.; Giordano, B. C.; Sivaprakasam, V.; Ananth, R.; Hammond, M.; Merritt, C.
D.; Tucker, J. E.; Malito, M.; Eversole, J. D.; Rose-Pehrsson, S., Continuous flow, explosives
vapor generator and sensor chamber. Review of Scientific Instruments 2014, 85 (5), -.
46. Bonvalot, L.; Mercury, M.; Zerega, Y., Experimental device for dioxin vapour
generation: A steady and adjustable source of gaseous 2,3-DCDD. Chemosphere 2014, 102
(0), 48-54.
47. (a) Alwis, L.; Sun, T.; Grattan, K. T. V., Fibre optic long period grating-based humidity
sensor probe using a Michelson interferometric arrangement. Sensors and Actuators B:
Chemical 2013, 178 (0), 694-699; (b) Gong, M.-S.; Kim, J.-U.; Kim, J.-G., Preparation of water-
durable humidity sensor by attachment of polyelectrolyte membrane to electrode substrate
by photochemical crosslinking reaction. Sensors and Actuators B: Chemical 2010, 147 (2),
Jiunn Yuan Tan (21637520) Page 143
539-547; (c) Han, D.-S.; Gong, M.-S., Zwitterionic sulfobetaine-containing polyelectrolyte for
controlling humidity-sensitivity. Sensors and Actuators B: Chemical 2010, 145 (1), 254-258;
(d) Yao, W.; Chen, X.; Zhang, J., A capacitive humidity sensor based on gold–PVA core–shell
nanocomposites. Sensors and Actuators B: Chemical 2010, 145 (1), 327-333.
48. (a) Yong, Z.; Xuejun, Z.; Tong, Z.; Jing, S.; Yan, B.; Jie, S., Gas sensing properties of
coral-like Bi 0.5 K 0.5 TiO 3 powders synthesized by metal-organic decomposition.
Measurement Science and Technology 2011, 22 (11), 115205; (b) Yoo, K. P.; Hong, H. P.; Lee,
M. J.; Min, S. J.; Park, C. W.; Choi, W. S.; Min, N. K., Fabrication, characterization and
application of a microelectromechanical system (MEMS) thermopile for non-dispersive
infrared gas sensors. Measurement Science and Technology 2011, 22 (11), 115206; (c)
Zoolfakar, A. S.; Ahmad, M. Z.; Rani, R. A.; Ou, J. Z.; Balendhran, S.; Zhuiykov, S.; Latham, K.;
Wlodarski, W.; Kalantar-zadeh, K., Nanostructured copper oxides as ethanol vapour sensors.
Sensors and Actuators B: Chemical 2013, 185 (0), 620-627.
49. (a) Christopher, R. F.; Cy, R. T.; Morgan, W.; Susan, L. R.-P., A versatile sensor
performance evaluation platform with an impactor-inspired sample chamber and virtual pin
grid array. Measurement Science and Technology 2014, 25 (6), 065901; (b) Nikolai, H.;
Marco, S.; Christian, B.; Andreas, S.; Tilman, S., Gas mixing apparatus for automated gas
sensor characterization. Measurement Science and Technology 2014, 25 (5), 055903.
50. Shallcross, D. C., Handbook of Psychrometric Charts: Humidity Diagrams for
Engineers. Blackie Academic & Professional: 1997.
51. Threlkeld, J. L., Thermal environmental engineering. Prentice-Hall: 1962.
52. Lee, C. H., Effects of Wick Contamination and Thermal Component Variation on
Thermal Indices. Texas Tech University: 1986.
Jiunn Yuan Tan (21637520) Page 144
53. (a) Fusco, A. J.; Avanza, J. R.; Aguerre, R. J.; Gabitto, J. F., A DIFFUSIONAL MODEL FOR
DRYING WITH VOLUME CHANGE. Drying Technology 1991, 9 (2), 397-417; (b) Jayaraman, K.
S.; Das Gupta, D. K., DEHYDRATION OF FRUITS AND VEGETABLES - RECENT DEVELOPMENTS
IN PRINCIPLES AND TECHNIQUES. Drying Technology 1992, 10 (1), 1-50; (c) Mulet, A., Drying
modelling and water diffusivity in carrots and potatoes. Journal of Food Engineering 1994,
22 (1–4), 329-348; (d) Waananen, K. M.; Litchfield, J. B.; Okos, M. R., CLASSIFICATION OF
DRYING MODELS FOR POROUS SOLIDS. Drying Technology 1993, 11 (1), 1-40.
54. Ratti, C., Shrinkage during drying of foodstuffs. Journal of Food Engineering 1994, 23
(1), 91-105.
55. (a) Chen, X. D.; Putranto., A., Modelling Drying Processes. Cambridge University Press:
2013; (b) Grasmeijer, N.; Waard, H. d.; Hinrichs, W. L. J.; Frijlink, H. W., A User-Friendly
Model for Spray Drying to Aid Pharmaceutical Product Development. PLoS One 2013, 8 (9).
56. (a) Kadja, M.; Bergeles, G., Modelling of slurry droplet drying. Applied Thermal
Engineering 2003, 23 (7), 829-844; (b) Lin, S. X. Q.; Chen, X. D., Changes in Milk Droplet
Diameter During Drying Under Constant Drying Conditions Investigated Using The Glass-
Filament Method. Food and Bioproducts Processing 2004, 82 (3), 213-218; (c) Sloth, J.; Kiil, S.;
Jensen, A. D.; Andersen, S. K.; Jørgensen, K.; Schiffter, H.; Lee, G., Model based analysis of
the drying of a single solution droplet in an ultrasonic levitator. Chemical Engineering
Science 2006, 61 (8), 2701-2709.
57. Adhikari, B.; Howes, T.; Bhandari, B. R.; Truong, V., Experimental studies and kinetics
of single drop drying and their relevance in drying of sugar‐rich foods: A review.
International Journal of Food Properties 2000, 3 (3), 323-351.
58. (a) Downingm, C. G., The evaporation of drops of pure liquids at elevated
temperatures: Rates of evaporation and wet-bulb temperatures. AIChE Journal 1966, 12 (4),
Jiunn Yuan Tan (21637520) Page 145
760-766; (b) Ranz, W.; Marshall, W., Evaporation from drops. Chem. Eng. Prog 1952, 48 (3),
141-146.
59. (a) Charlesworth, D. H.; Marshall, W. R., Evaporation from drops containing dissolved
solids. AIChE Journal 1960, 6 (1), 9-23; (b) Nešić, S.; Vodnik, J., Kinetics of droplet
evaporation. Chemical Engineering Science 1991, 46 (2), 527-537.
60. Cheong, H. W.; Jeffreys, G. V.; Mumford, C. J., A receding interface model for the
drying of slurry droplets. AIChE Journal 1986, 32 (8), 1334-1346.
61. (a) Chen, Y.; Zhang, H.; Fox, D.; Faulkner, C. C.; Jeng, D.; Bari, M., Microstructural and
chemical variation of TiO2 electrodes in DSSCs after ethanol vapour treatment. Materials
Science and Engineering: B 2013, 178 (1), 71-76; (b) Yanuriati, A.; Savage, G. P.; Rowe, R. N.,
The effects of ethanol treatment on the metabolism, shelf life and quality of stored
tomatoes at different maturities and temperatures. Journal of the Science of Food and
Agriculture 1999, 79 (7), 995-1002; (c) Zhang, W.; Li, X.; Wang, X.; Wang, G.; Zheng, J.;
Abeysinghe, D. C.; Ferguson, I. B.; Chen, K., Ethanol vapour treatment alleviates postharvest
decay and maintains fruit quality in Chinese bayberry. Postharvest Biology and Technology
2007, 46 (2), 195-198.
62. (a) Arce, A.; Arce Jr, A.; Martínez-Ageitos, J. M.; Soto, A., Isobaric vapor–liquid
equilibria of 1,1-dimethylethoxy-butane + methanol or
ethanol + water at 101.32 kPa. Fluid Phase Equilibria 2007, 259 (1), 57-65;
(b) Cristino, A. F.; Rosa, S.; Morgado, P.; Galindo, A.; Filipe, E. J. M.; Palavra, A. M. F.; Nieto
de Castro, C. A., High-temperature vapour–liquid equilibrium for the water–alcohol systems
and modeling with SAFT-VR: 1. Water–ethanol. Fluid Phase Equilibria 2013, 341 (0), 48-53.
63. (a) Oliveira, S. B. M.; Marques, R. P.; Parise, J. A. R., Modelling of an ethanol–water
distillation column with vapour recompression. International Journal of Energy Research
Jiunn Yuan Tan (21637520) Page 146
2001, 25 (10), 845-858; (b) Voutsas, E. C.; Pamouktsis, C.; Argyris, D.; Pappa, G. D.,
Measurements and thermodynamic modeling of the ethanol–water system with emphasis
to the azeotropic region. Fluid Phase Equilibria 2011, 308 (1–2), 135-141.
64. Welty, J. R.; Wicks, C. E.; Wilson, R. E., Fundamentals of momentum, heat, and mass
transfer. Wiley: 1976.
65. Bowman, F. M.; Melton, J. A., Effect of activity coefficient models on predictions of
secondary organic aerosol partitioning. Journal of Aerosol Science 2004, 35 (12), 1415-1438.
66. Seth, D.; Sarkar, A., A Lumped Parameter Model for Effective Moisture Diffusivity in
Air Drying of Foods. Food and Bioproducts Processing 2004, 82 (3), 183-192.
67. Gac, J. M.; Gradoń, L., A distributed parameter model for the spray drying of
multicomponent droplets with a crust formation. Advanced Powder Technology 2013, 24 (1),
324-330.
68. Barati, E.; Esfahani, J. A., Mathematical modeling of convective drying: Lumped
temperature and spatially distributed moisture in slab. Energy 2011, 36 (4), 2294-2301.
69. Law, C. K., Recent advances in droplet vaporization and combustion. Progress in
Energy and Combustion Science 1982, 8 (3), 171-201.
70. Karathanos, V. T.; Villalobos, G.; Saravacos, G. D., Comparison of Two Methods of
Estimation of the Effective Moisture Diffusivity from Drying Data. Journal of Food Science
1990, 55 (1), 218-223.
71. Sano, Y.; Keey, R. B., The drying of a spherical particle containing colloidal material
into a hollow sphere. Chemical Engineering Science 1982, 37 (6), 881-889.
72. Fu, N.; Wai Woo, M.; Qi Lin, S. X.; Zhou, Z.; Dong Chen, X., Reaction Engineering
Approach (REA) to model the drying kinetics of droplets with different initial sizes—
experiments and analyses. Chemical Engineering Science 2011, 66 (8), 1738-1747.
Jiunn Yuan Tan (21637520) Page 147
73. (a) Dalmaz, N.; Ozbelge, H. O.; Eraslan, A. N.; Uludag, Y., Heat and Mass Transfer
Mechanisms in Drying of a Suspension Droplet: A New Computational Model. Drying
Technology 2007, 25 (2), 391-400; (b) Jayanthi, G. V.; Zhang, S. C.; Messing, G. L., Modeling
of Solid Particle Formation During Solution Aerosol Thermolysis: The Evaporation Stage.
Aerosol Science and Technology 1993, 19 (4), 478-490.
74. Schwarz, J.; Smolík, J., Mass transfer from a drop—I. Experimental study and
comparison with existing correlations. International Journal of Heat and Mass Transfer 1994,
37 (14), 2139-2143.
75. Kar, S.; Chen, X. D., Effect of high mass fluxes on heat and mass transfer through a
flat surface. Proceedings of the Institution of Mechanical Engineers 2004, 218 (4), 213-220.
76. Renksizbulut, M.; Nafziger, R.; Li, X., A mass transfer correlation for droplet
evaporation in high-temperature flows. Chemical Engineering Science 1991, 46 (9), 2351-
2358.
77. Birouk, M.; Abou Al-Sood, M. M.; Gökalp, I., Droplet Evaporation in a Turbulent
Environment at Elevated Pressure and Temperature Conditions. Combustion Science and
Technology 2008, 180 (10-11), 1987-2014.
78. Lage, P. L. C.; Rangel, R. H.; Hackenberg, C. M., Multicomponent heat and mass
transfer for flow over a droplet. International Journal of Heat and Mass Transfer 1993, 36
(14), 3573-3581.
79. Frössling, N., Über die Verdunstung fallender Tropfen. Gerlabds Beitr. Geophys. 1938,
52, 170-216.
80. N. T. Hsu, K. S. a. B. H. S., Material transfer in turbulent gas streams : Influence of
shape on evaporation of drops of n-heptane. Industr. Engng Chem. 1954, 46, 870-876.
Jiunn Yuan Tan (21637520) Page 148
81. P. N. Rowe, K. T. C. a. J. B. L., Heat and mass transfer from a single sphere in an
extensive flowing fluid. Trans. Instn Chem. Engrs 1965, 43, T14-T31.
82. Pruppacher, H. R.; Beard, K. V., A wind tunnel investigation of the internal circulation
and shape of water drops falling at terminal velocity in air. Quarterly Journal of the Royal
Meteorological Society 1970, 96 (408), 247-256.
83. Dobrovolskaya, V. I. B. a. L. P., 0 vlianii temperatury na isparenie kapel zhidkosti v
vozdushnom potoke. Inzh. Fiz. Zh. 1972, 23, 686-691.
84. Chuchottaworn, P.; Fujinami, A.; Asano, K., EXPERIMENTAL STUDY OF EVAPORATION
OF A VOLATILE PENDANT DROP UNDER HIGH MASS FLUX CONDITIONS. Journal of Chemical
Engineering of Japan 1984, 17 (1), 7-13.
85. Wedding, J. B.; Kim, Y. J.; Dennison, R. S., Mass transfer from water droplets under
simulated free-fall conditions. Atmospheric Environment (1967) 1986, 20 (5), 1039-1045.
86. Jenkins, H. D. B., Chemical Thermodynamics at a Glance. Wiley: 2008.
87. King, E. L., Ion Association. C. W. Davies. Butterworth, Washington, D.C., 1962. viii +
190 pp. Illus. $7.50. Science 1964, 143 (3601), 37.
88. Millero, F., Estimation of the Partial Molar Volumes of Ions in Mixed Electrolyte
Solutions Using the Pitzer Equations. J Solution Chem 2014, 43 (8), 1448-1465.
89. Ge, X.; Zhang, M.; Guo, M.; Wang, X., Correlation and Prediction of Thermodynamic
Properties of Nonaqueous Electrolytes by the Modified TCPC Model. Journal of Chemical &
Engineering Data 2007, 53 (1), 149-159.
90. Sipos, P., Application of the Specific Ion Interaction Theory (SIT) for the ionic
products of aqueous electrolyte solutions of very high concentrations. Journal of Molecular
Liquids 2008, 143 (1), 13-16.
Jiunn Yuan Tan (21637520) Page 149
91. Abrams, D. S.; Prausnitz, J. M., Statistical thermodynamics of liquid mixtures: A new
expression for the excess Gibbs energy of partly or completely miscible systems. AIChE
Journal 1975, 21 (1), 116-128.
92. Wilson, G. M., Vapor-Liquid Equilibrium. XI. A New Expression for the Excess Free
Energy of Mixing. Journal of the American Chemical Society 1964, 86 (2), 127-130.
93. Renon, H.; Prausnitz, J. M., Local compositions in thermodynamic excess functions
for liquid mixtures. AIChE Journal 1968, 14 (1), 135-144.
94. Gmehling, J., From UNIFAC to modified UNIFAC to PSRK with the help of DDB. Fluid
Phase Equilibria 1995, 107 (1), 1-29.
95. Fredenslund, A.; Jones, R. L.; Prausnitz, J. M., Group-contribution estimation of
activity coefficients in nonideal liquid mixtures. AIChE Journal 1975, 21 (6), 1086-1099.
96. Haghtalab, A.; Mahmoodi, P., Vapor–liquid equilibria of asymmetrical systems using
UNIFAC-NRF group contribution activity coefficient model. Fluid Phase Equilibria 2010, 289
(1), 61-71.
97. Gmehling, J.; Weidlich, U., Results of a modified unifac method for alkane-alcohol
systems. Fluid Phase Equilibria 1986, 27 (0), 171-180.
98. Kuramochi, H.; Noritomi, H.; Hoshino, D.; Nagahama, K., Representation of activity
coefficients of fundamental biochemicals in water by the UNIFAC model. Fluid Phase
Equilibria 1997, 130 (1–2), 117-132.
99. Wittig, R.; Lohmann, J.; Gmehling, J., Vapor−Liquid Equilibria by UNIFAC Group
Contribution. 6. Revision and Extension. Industrial & Engineering Chemistry Research 2002,
42 (1), 183-188.
100. Fredenslund, A., Vapor-liquid equilibria using UNIFAC: a group-contribution method.
Elsevier: 2012.
Jiunn Yuan Tan (21637520) Page 150
101. Santacesaria, E.; Tesser, R.; di Serio, M., Simple and predictive approach for
calculating the high pressure and temperature vapour—liquid equilibria of binary mixtures
by applying a UNIFAC equation of state method. Fluid Phase Equilibria 1991, 63 (3), 329-340.
102. Santacesaria, E.; Di Serio, M.; Tesser, R.; Triunfo, P., An application of the UNIFAC-
RKS method to a reactive system. Fluid Phase Equilibria 1993, 84 (0), 111-122.
103. Dahl, S.; Fredenslund, A.; Rasmussen, P., The MHV2 model: a UNIFAC-based
equation of state model for prediction of gas solubility and vapor-liquid equilibria at low and
high pressures. Industrial & Engineering Chemistry Research 1991, 30 (8), 1936-1945.
104. Hong, G.-B.; Hsieh, C.-T.; Lin, H.-m.; Lee, M.-J., Multiphase Equilibrium Calculations
from Soave Equation of State with Chang-Twu/UNIFAC Mixing Rules for Mixtures Containing
Water, Alcohols, and Esters. Industrial & Engineering Chemistry Research 2012, 51 (13),
5073-5081.
105. (a) Huckaby, J. L.; Ray, A. K., Absorption of sulfur dioxide by growing and evaporating
water droplets. Chemical Engineering Science 1989, 44 (12), 2797-2808; (b) Ray, A. K.;
Huckaby, J. L.; Shah, T., Thermal effects of condensation on absorption of gases in growing
droplets. Chemical Engineering Science 1987, 42 (8), 1955-1967.
106. Smolík, J.; Vítovec, J., Quasistationary evaporation of a droplet into a
multicomponent gaseous mixture. Journal of Aerosol Science 1984, 15 (5), 545-552.
107. Krishna, R., Mass transport from a spherical body into a multicomponent fluid
stream. Letters in Heat and Mass Transfer 1976, 3 (3), 257-266.
108. Qi Lin, S. X.; Chen, X. D., Improving the Glass-Filament Method for Accurate
Measurement of Drying Kinetics of Liquid Droplets. Chemical Engineering Research and
Design 2002, 80 (4), 401-410.
Jiunn Yuan Tan (21637520) Page 151
109. Incropera, F. P.; Bergman, T. L.; Lavine, A. S.; DeWitt, D. P., Fundamentals of heat and
mass transfer. Wiley: 2011.
110. Abramzon, B.; Sirignano, W. A., Droplet vaporization model for spray combustion
calculations. International Journal of Heat and Mass Transfer 1989, 32 (9), 1605-1618.
111. Erbil, H. Y.; Demirel, A. L.; Avci, Y.; Mert, O., Transformation of a simple plastic into a
superhydrophobic surface. Science 2003, 299 (5611), 1377-80.
112. Takamuku, T.; Yamaguchi, A.; Matsuo, D.; Tabata, M.; Yamaguchi, T.; Otomo, T.;
Adachi, T., NaCl-Induced Phase Separation of 1,4-Dioxane−Water Mixtures Studied by Large-
Angle X-ray Scattering and Small-Angle Neutron Scattering Techniques. The Journal of
Physical Chemistry B 2001, 105 (41), 10101-10110.
113. Imae, T.; Sasaki, M.; Abe, A.; Ikeda, S., Liquid-liquid phase separation in dilute
aqueous solutions of surfactants: the effect of added salt. Langmuir 1988, 4 (2), 414-418.
114. Manrique, J.; Martinez, F., Solubility of ibuprofen in some ethanol+ water cosolvent
mixtures at several temperatures. Latin American Journal of Pharmacy 2007, 26 (3), 344.
115. López-Montilla, J. C.; Herrera-Morales, P. E.; Shah, D. O., New Method to
Quantitatively Determine the Spontaneity of the Emulsification Process. Langmuir 2002, 18
(11), 4258-4262.
116. González-Ochoa, H.; Arauz-Lara, J. L., Spontaneous Two-Dimensional Spherical
Colloidal Structures. Langmuir 2007, 23 (10), 5289-5291.
117. Hoorfar, M.; Kurz, M. A.; Policova, Z.; Hair, M. L.; Neumann, A. W., Do
Polysaccharides Such as Dextran and Their Monomers Really Increase the Surface Tension of
Water? Langmuir 2005, 22 (1), 52-56.
118. Herrington, B. L., Some Physico-Chemical Properties of Lactose: II. Factors
Influencing the Crystalline Habit of Lactose. Journal of dairy science 1934, 17 (8), 533-542.
Jiunn Yuan Tan (21637520) Page 152
119. Hostetler, S. E. Amorphous to crystalline transformation in maltose: Implications for
orally disintegrating tablets. Ph.D., Purdue University, Ann Arbor, 2002.
120. Khalifeh, I. M. Thermodynamic evaluation of ibuprofen solubility in aqueous and
non-aqueous cosolvent systems. Ph.D., Purdue University, Ann Arbor, 2000.
121. Furuta, T.; Tsujimoto, S.; Oazaki, M.; Toei, R., EFFECT OF DRYING ON RETENTION OF
ETHANOL IN MALTODEXTRIN SOLUTION DURING DRYING OF A SINGLE DROPLET. Drying
Technology 1983, 2 (3), 311-327.
122. Grenier, P., Solvent counter-diffusion in gas absorption. The Canadian Journal of
Chemical Engineering 1966, 44 (4), 213-216.
123. (a) O'Hare, K. D.; Spedding, P. L., Evaporation of a binary liquid mixture. The Chemical
Engineering Journal 1992, 48 (1), 1-9; (b) Sefiane, K.; Tadrist, L.; Douglas, M., Experimental
study of evaporating water–ethanol mixture sessile drop: influence of concentration.
International Journal of Heat and Mass Transfer 2003, 46 (23), 4527-4534.
124. Toor, H. L., Diffusion in three-component gas mixtures. AIChE Journal 1957, 3 (2),
198-207.
125. Krishna, R.; Wesselingh, J. A., The Maxwell-Stefan approach to mass transfer.
Chemical Engineering Science 1997, 52 (6), 861-911.
126. Woo, M. W.; Mansouri, S.; Chen, X. D., Antisolvent vapor precipitation: the future of
pulmonary drug delivery particle production? Expert Opinion on Drug Delivery 2014, 11 (3),
307-311.
Jiunn Yuan Tan (21637520) Page 153
APPENDIX
Figure A.1: 3-D Plot of the Experimental Matrices Undertaken for Maltodextrin DE 10.
Figure A.2: 3-D Plot of the Experimental Matrices Undertaken for Maltodextrin DE 18.
Jiunn Yuan Tan (21637520) Page 154
Figure A.3: 3-D Plot of the Experimental Matrices Undertaken for Maltose.
Figure A.4: Spherical Particle Size Distribution for Maltodextrin DE 10.
-2
0
2
4
6
8
10
12
14
16
18
0 1 2 3 4 5 6
Percentage (%)
Particle size (µm)
Jiunn Yuan Tan (21637520) Page 155
Figure A.5: Spherical Particle Size Distribution for Maltodextrin DE 18.
Figure A.6: Spherical Particle Size Distribution for Maltose.
-5
0
5
10
15
20
25
0 0.5 1 1.5 2 2.5 3 3.5
Percentage (%)
Particle size (µm)
-2
0
2
4
6
8
10
12
14
16
18
0 0.5 1 1.5 2 2.5 3
Percentage (%)
Particle size (µm)
Jiunn Yuan Tan (21637520) Page 156
Table A.1: Summary Results of Maltodextrin DE 10.
Initial weight concentration (wt%)
Ethanol absolute humidity (kg/kg db)
Ethanol relative humidity (%)
Particle Morphology
2.5 0.068 57 Porous
2.5 0.074 66 Microsphere Network
2.5 0.090 85 Spherical
2.5 0.092 87 Spherical
2.5 0.094 90 Spherical
5 0.080 70 Microsphere Network
5 0.090 73 Spherical
5 0.096 80 Spherical
5 0.100 87 Spherical
5 0.100 90 Spherical
10 0.080 65 Porous
10 0.090 79 Porous
10 0.092 80 Porous
10 0.098 85 Microsphere Network
10 0.112 93 Spherical
15 0.078 70 Porous and bubble
15 0.088 75 Porous and bubble
15 0.090 80 Porous
15 0.100 89 Spherical
15 0.106 92 Spherical
Jiunn Yuan Tan (21637520) Page 157
Table A.2: Summary Results of Maltodextrin DE 18.
Initial weight concentration (wt%)
Ethanol absolute humidity (kg/kg db)
Ethanol relative humidity (%)
Particle Morphology
2.5 0.074 70 Microsphere Network
2.5 0.088 75 Microsphere Network
2.5 0.090 80 Spherical
2.5 0.100 89 Spherical
2.5 0.102 94 Spherical
5 0.074 65 Microsphere Network
5 0.084 70 Microsphere Network
5 0.090 80 Spherical
5 0.100 89 Spherical
5 0.110 93 Spherical
10 0.088 65 Porous and bubble
10 0.090 70 Porous and bubble
10 0.098 80 Spherical
10 0.100 89 Spherical
10 0.102 94 Spherical
15 0.078 70 Porous and bubble
15 0.088 75 Microsphere network
15 0.090 80 Spherical
15 0.094 88 Spherical
15 0.102 94 Spherical
Jiunn Yuan Tan (21637520) Page 158
Table A.3: Summary Results of Maltose.
Initial weight concentration (wt%)
Ethanol absolute humidity (kg/kg db)
Ethanol relative humidity (%)
Particle Morphology
2.5 0.070 65 Porous and smooth surface
2.5 0.090 70 Porous and smooth surface
2.5 0.096 82 Microsphere Network
2.5 0.100 89 Microsphere
2.5 0.110 93 Microsphere
5 0.076 68 Porous and smooth surface
5 0.090 76 Porous and smooth surface
5 0.092 83 Microsphere Network
5 0.098 88 Microsphere Network
5 0.110 93 Microsphere
10 0.074 65 Porous and smooth surface
10 0.088 70 Porous and smooth surface
10 0.090 80 Smooth surface
10 0.100 89 Smooth surface
10 0.106 92 Smooth surface
15 0.078 70 Porous and smooth surface
15 0.086 73 Porous and smooth surface
15 0.090 80 Smooth surface
15 0.100 89 Smooth surface
15 0.106 92 Smooth surface