mechanism of the novel antisolvent vapour precipitation ... · mechanism of the novel antisolvent...

Jiunn Yuan Tan (21637520)

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


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.


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: ………............................................


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


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


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


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.


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.


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.


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.


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


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


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


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


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


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


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


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


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)


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)


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


= 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)


= 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


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.


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.


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.


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


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.


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.


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.


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.


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


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.


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.


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


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.


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.


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


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.


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.


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.


humidity is not


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.


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.


concentration is not


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





purified atmospheric air.

- Solid sublimated via water

bath heating to produce

vapour, swept by the

nitrogen gas.

and controlled.


vapour.

Liquid injection

vapour

generator

(LIVG)

45

Ppb - ppt - Liquid vapour is injected

with syringe needle through

cartridge heaters and driven

through by air.


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.


vapour generated.




and controlled.


Figure 2.4: Sketch of a lab-scale humidity generator 43.


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


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


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.


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.


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


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


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


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


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


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


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


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.


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


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.


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)


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


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)


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.


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)


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)


(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)


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


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)


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


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


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


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


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


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


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.


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


(kg/kg db)

Time (s)

Heating

Cycle 1

Heating

Cycle 2


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


(kg/kg db)

Time (s)

(a)


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


(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


(kg/kg db)

Time (s)

(b) (c)


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


(kg/kg db)

Time (s)

(a)

(b)


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


(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)


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


(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


(kg/kg db)

Time (s)

(b)

(c)


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


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.


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.


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)


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


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


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.


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).


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)


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)


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.


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


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.


Microsphere Region (a)

(b)

Microsphere

Network Region

Microsphere

Network Region

Microsphere Region


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


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


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.


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


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.


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


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.


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)


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


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.


-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)


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)


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


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


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


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


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,


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.


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


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.


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


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



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)


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)


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.



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



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.


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)




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




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.



of Water-Ethanol Droplet


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)



Experimental (70% water, 30% ethanol)




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.


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)



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)


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


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,


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.


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


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


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:


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


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.


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.


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.


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)


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.


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


nanoparticles for oral administration through liquid antisolvent precipitation. European


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.


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,


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.


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),


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.


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),


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


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.


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.


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.


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.


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.


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.


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.


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.

-2

0

2

4

6

8

10

12

14

16

18

0 1 2 3 4 5 6

Percentage (%)

Particle size (µm)


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)


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



5 0.080 70 Microsphere Network

5 0.090 73 Spherical




10 0.080 65 Porous

10 0.090 79 Porous

10 0.092 80 Porous



15 0.078 70 Porous and bubble


15 0.090 80 Porous




Table A.2: Summary Results of Maltodextrin DE 18.




Particle Morphology

















15 0.088 75 Microsphere network





Table A.3: Summary Results of Maltose.




Particle Morphology

2.5 0.070 65 Porous and smooth surface

2.5 0.090 70 Porous and smooth surface


2.5 0.100 89 Microsphere

2.5 0.110 93 Microsphere

5 0.076 68 Porous and smooth surface




5 0.110 93 Microsphere



10 0.090 80 Smooth surface








mechanism of the novel antisolvent vapour precipitation ... · mechanism of the novel antisolvent...

Documents