Particle fabrication techniques for pharmaceutical applications

It was found that nitrogen flow rate, solution flow rate and voltage difference between the nozzle and ring can significantly affect the particle collection efficiency of the EHDA proces

PARTICLE FABRICATION VIA

ELECTROHYDRODYNAMIC ATOMIZATION FOR PHARMACEUTICAL

APPLICATIONS

ALIREZA REZVANPOUR

(M.S., B.S., Sharif University of Technology, Iran)

A THESIS SUBMITED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF CHEMICAL AND BIOMOLECULAR ENGINEERING

NATIONAL UNIVERSITY OF SINGAPORE

2012

Acknowledgements

First and foremost, I would like to truthfully express my gratitude to my highly respected supervisor Professor Chi-Hwa Wang, for his constant guidance, moral support, supervision, comments and suggestions throughout my whole PhD studies Definitely, without his support, the completion of this thesis would not have been possible I would also like to sincerely thank Professor William B Krantz (Department of Chemical & Biological Engineering, University of Colorado, Boulder) for his invaluable guidance and comments during my studies I don’t hesitate to say that I have learnt many precious points during several discussions with him

I am specially grateful to Professor Wuqiang Yang (School of Electrical and Electronic Engineering, The University of Manchester) and Professor Yung C Liang (Department

of Electrical and Computer Engineering, National University of Singapore) for their invaluable discussions and comments during the work The support from Mr Kok Hong Boey, Ms How Yoke Leng, Ms Fengmei Li and other lab technicians and administrative staff is greatly appreciated

I would like to thank Dr Lim Wee Chuan, Dr Sudhir Hulikal Ranganath, Dr Nie Hemin,

Dr Davis Yohanes Arifin, Dr Lim Liang Kuang, Dr Jingwei Xie, Dr Cheng Yongpan,

Ms Lei Chenlu, Mr Qiao Gian, Mr Xu Qingxing Noel and other group members for helpful technical support and discussions And I would like to thank students Mr Tan Guowei, Ms Amalina Bte Ebrahim Attia, Mr Lee Teng Yong Jeffrey, Mr Luo Yu, Ms Nancy Liliana Setiawan and Mr Jun Wei for their participation on various experiments and simulations in this project The research scholarship from the National University of

Singapore, Department of Chemical and Biomolecular Engineering, is deeply acknowledged

Lastly but definitely not the least, I would like to truthfully express my thanks to my lovely wife, Mrs Shima Najafi Nobar, and other family members for providing me inexhaustible support and love during my PhD study

Chapter 2: Experimental and Computational Studies of Electrohydrodynamic Atomization process in Encapsulation Chamber for Pharmaceutical Particle Fabrication to Enhance the Particle Collection Efficiency

2.2.3.4 Effect of electrical conductivity and polymer material 37

2.2.3.6 Concentration of residual DCM in particles 45

2.3.1.4 Electric Force Model 54 2.3.1.5 Initial Conditions for Particle Trajectory Simulations 55

4.3 Results and Discussion 134

4.4 Applying scaling analysis to electrospray deposition process 151

Chapter 5: Investigation of droplet distribution in EHDA encapsulation chamber using AC-based Electrical Capacitance Tomography (AC-ECT) system with internal-external electrode sensor

Summary

In the present work, Electrohydrodynamic Atomization was employed to produce biodegradable polymeric microparticles in a new generation of shuttle glass chamber The effects of different operational parameters on the particle collection efficiency and residual amount of organic solvent in collected particles were investigated systematically The Taguchi method was used to design the experiments It was found that the important factors affecting particle collection efficiency were given in the following orders: solution flow rate, nitrogen flow rate, ring, and nozzle voltage It was found that solution flow rate and nozzle voltage can considerably affect the size of fabricated particles For all the trials, the residual DCM content of the particles fabricated using the EHDA method was well within the limit of safety standards at the end of process without engaging any additional freeze-drying process

A computational model (using FLUENT and COMSOL as computational fluid dynamic software) was developed in this study to simulate the fluid and particle dynamics in an EHDA chamber It was found that nitrogen flow rate, solution flow rate and voltage difference between the nozzle and ring can significantly affect the particle collection efficiency of the EHDA process Electric field and electric potential profiles in the chamber were significantly affected by the combined voltages of the nozzle and ring The computational model developed in this study provided a means of understanding the various processes involved in particle fabrication using the EHDA methodology

In a new set of experiments, an additional aluminum plate was located a few centimeters above the collecting plate in EHDA chamber which was connected to positive high voltage generator This work aimed to investigate the effect of the auxiliary electric field

on particle collection efficiency, morphology and size distribution The final results show that application of the auxiliary electric field can clearly enhance particle collection efficiency in comparison to the EHDA process without auxiliary electric field Additionally, it was established that the particle size distribution was not considerably influenced by the auxiliary electric field On the contrary, the smoothness of the particles can be affected by the auxiliary electric field especially when a high voltage is applied to the flat plate

Scaling analysis was used to assess the relative importance of the terms in the particle force balance The collection efficiency of the EHDA process was determined from a force balance on the particles that in turn depends on the fluid dynamics and electric field It led to a unique dimensionless group that permits collapsing all the experimental data for the effect on the particle collection efficiency of the carrier gas flow rate, liquid solution flow rate and electric field strength onto a generalized plot for which a cubic trendline fits the data with a coefficient of determination 2

1

R  Electrospray deposition on a substrate through a mask, to generate biodegradable polymeric particle patterns, was also considered in this study to investigate the effect of different operational parameters Moreover, a mathematical model was developed to track the particle trajectories and focusing effect in electrospray deposition process on the substrate The final results confirm that the clearest particle pattern and the best focusing effect on the substrate can be achieved with long distance between the nozzle and the substrate, high voltage difference between the nozzle and the mask, short process time and low solution flow rate Furthermore, micro-fibers can be observed on the mask when the voltage difference between the nozzle and substrate is not high

In the last part of this study, Electrical Capacitance Tomography (ECT) was used for situ monitoring of very dilute pharmaceutical droplet and particle trajectories in different regions of EHDA encapsulation chamber A new type of ECT sensor with internal and external electrodes was used to improve the sensitivity of ECT measurement for detection

in-of the objects in the central area in-of the EHDA encapsulation chamber The water-air and dicholoromethane (DCM)-air systems in the dripping and spray modes were investigated

to determine the feasibility of imaging water and DCM droplets of low concentration in the encapsulation chamber using ECT

Table 2.3 Summary of factors, levels, obtained results (collection efficiency)

and S/N ratio in each trial

23

Table 2.4 S/N ratio for levels of each factor and S/N differences between

these levels where particle fabrication has been optimized

26

Table 2.5 Results of analysis of variance (ANOVA) for particle collection

efficiency for each factor

27

Table 2.6 Estimated performance at the optimum conditions 28

Table 2.7 Summary of factors, levels, obtained results (particle size) and S/N

ratio in each trial

40

Table 2.8 Results of analysis of variance (ANOVA) in particle size

measurements for each factor

40

Table 2.9 Summary of factors, levels, obtained results (concentration of

DCM) and S/N ratio in each trial

45

Table 2.10 Physical properties of DCM+PLGA solution used for calculation

of droplet diameter and current

56

Table 2.11 Experimental conditions, setup characteristics and physical

properties of the used materials

77

Table 2.12 Controllable factors and their levels used in the experiments 77

Table 2.13 Comparison between particle collection efficiencies in the trials

with and without AEF at 35 Lit/min nitrogen flow rate and 7 kV auxiliary voltage

79

Table 2.14 Physical and geometrical properties and operating conditions 84 Table 3.1 Physical and geometrical properties and operating conditions 109 Table 4.1 Operating conditions considered for deposition on the substrate 126 Table 4.2 Operating conditions considered for deposition on the mask 126

Table 4.3 Physical and geometrical parameters for the particle deposition

List of Figures Figure 1.1 Schematic diagram of overview of this thesis 11

Figure 2.2 Variation of the collection efficiency versus solution flow rate at

optimum conditions

29

Figure 2.3 Variation of the collection efficiency versus nitrogen flow rate at

(a) optimum conditions shown in Table 2.6, (b) Nozzle voltage 8.5

kV, ring voltage 5.5 kV and solution flow rate 1.5 mL/h (c) Nozzle voltage 8.5 kV, ring voltage 6.5 kV and solution flow rate 2.5 mL/h

32

Figure 2.4 Variation of collection efficiency for four example nitrogen flow

rates versus solution flow rate

32

Figure 2.5 Variation of the collection efficiency versus (a) nozzle and (b) ring

voltages respectively at optimum conditions

34

Figure 2.6 Variation of collection efficiency for each nozzle (a) and ring (b)

voltages versus the solution flow rate

34

Figure 2.7 Variation of collection efficiency for four example ring voltage

settings versus nozzle voltages

36

Figure 2.8 Effect of voltage difference between nozzle and ring on the particle

size and particle collection efficiency (a) trial 7: voltage difference: 2.5 kV; average CE: 78.6%; particle size: 4-5 μm (b) trial 1; voltage difference: 2 kV; average CE: 75.8%; particle size:

5-6 μm (c) trial 15; voltage difference: 1.5 kV; average CE: 62.2%;

particle size: 6-8 μm (d) trial 9; voltage difference: 1 kV; average CE: 63.5%; particle size: 8-10 μm Detailed operating parameters for the respective trial numbers are shown in Table 9

37

Figure 2.9 Variation of particle collection efficiency versus electrical

conductivity of the solution for two different polymers, PLGA and PLA at 8.5 kV nozzle voltage, 7 kV ring voltage, 2 mL/h solution flow rate and 25 L/min nitrogen flow rate

38

Figure 2.10 Particle size distribution and related SEM photo for the following

conditions: Nozzle voltage 8.5 kV, Ring voltage 7 kV, solution flow rate 2 mL/h and nitrogen flow rate 25 L/min (a) and (b), Nozzle voltage 7.5 kV, Ring voltage 6 kV, solution flow rate 2 mL/h and nitrogen flow rate 20 L/min (c) and (d), Nozzle voltage

42

8.5 kV, Ring voltage 5.5 kV, solution flow rate 1.5 mL/h and nitrogen flow rate 30 L/min (e) and (f)

Figure 2.11 Variation of arithmetic particle mean size versus solution flow rate

for different nozzle voltages at ring voltage 6 kV and nitrogen flow rate 25 L/min

43

Figure 2.12 Variation of arithmetic particle mean size versus nozzle voltage for

different solution flow rates at ring voltage 6 kV and nitrogen flow rate 25 L/min

44

Figure 2.13 Variation of residual concentration of DCM in collected particles

versus nitrogen flow rate

47

Figure 2.14 Electric field lines (V/m) in the computational domain obtained

from CFD simulations carried out using COMSOL with the following conditions: nozzle voltage 9 kV and ring voltage 6 kV

60

Figure 2.15 (a) Flow field within the chamber for 35 L/min nitrogen flow rate,

(b) Electric field within the chamber for 9 kV nozzle voltage and 6

kV ring voltage

61

Figure 2.16 Particle residence time profiles in the computational domain

obtained from the CFD simulations with the following conditions:

(a) solution flow rate 1.0 mL/h, nitrogen flow rate 35 L/min, voltage difference between nozzle and ring 3 kV (b) solution flow rate 2.5 mL/h, nitrogen flow rate 25 L/min, voltage difference between nozzle and ring 2 kV (c) solution flow rate 2.5 mL/h, nitrogen flow rate 20 L/min, voltage difference between nozzle and ring 1 kV

62

Figure 2.17 Forces acting on the Taylor cone in the EHDA process 65

Figure 2.18 Comparisons of particle collection efficiency variations with

solution flow rate obtained experimentally with CFD simulations under the following conditions: nozzle voltage 9 kV, (a) nitrogen flow rate 35 L/min, voltage difference between nozzle and ring 3.5

kV (b) nitrogen flow rate 25 L/min, voltage difference between nozzle and ring 3 kV (c) nitrogen flow rate 20 L/min, voltage difference between nozzle and ring 2 kV

67

Figure 2.19 Comparisons of particle collection efficiency variations with

nitrogen flow rate obtained experimentally with CFD simulations under the following conditions: nozzle voltage 9 kV, (a) solution flow rate 1.0 mL/h, voltage difference between nozzle and ring 3.5

68

nozzle and ring 3 kV (c) solution flow rate 2.5 mL/h, voltage difference between nozzle and ring 2 kV

Figure 2.20 Turbulence intensity in the EHDA chamber at nitrogen flow rate of

25 L/min

70

Figure 2.21 Velocity contours for nitrogen flow rates of (a) 20 L/min, (b) 25

L/min, (c) 30 L/min and (d) 35 L/min

71

Figure 2.22 (a) Electric field profile for 9 kV nozzle voltage, (b) Electric field

profile for 9 kV nozzle voltage and 6 kV ring voltage

72

Figure 2.23 Electric field profile (shown by electric field contours in V/m) with

9 kV nozzle voltage and (a) 6 kV ring voltage, (b) 7 kV ring voltage, (c) 8 kV ring voltage and (d) 9 kV ring voltage

73

Figure 2.24 Comparisons of particle collection efficiency variations with

voltage difference between nozzle and ring obtained experimentally with CFD simulations under the following conditions: nozzle voltage 9 kV, (a) solution flow rate 2.5 mL/h, nitrogen flow rate 35 L/min (b) solution flow rate 2.0 mL/h, nitrogen flow rate 30 L/min (c) solution flow rate 1.5 mL/h, nitrogen flow rate 25 L/min

81

Figure 2.28 Variation of collection efficiency versus flat plate voltage observed

in the following operating conditions: (a) nozzle voltage 10 kV and solution flow rate 1 mL/h, (b) ring voltage 6 kV and solution flow rate 1 mL/h, and (c) nozzle voltage 10 kV and ring voltage 5 kV

82

Figure 2.29 FE-SEM images of collected particles at nozzle voltage 9kV, ring

voltage 6kV, flat plate voltage 5 kV and solution flow rate, (a) &

(b) 1 mL/h, (c) & (d) 2.5 mL/h

87

Figure 2.30 FE-SEM and SEM images of collected particles at: (a) nozzle

voltage 10 kV, ring voltage 6kV, Solution flow rate 1 mL/h,

88

particle size range: 6-10 μm, (b) nozzle voltage 10 kV, ring voltage 6kV, flat plate voltage 5 kV, solution flow rate 1 mL/h, particle size range: 3-7 μm, (c) nozzle voltage 10 kV, ring voltage 6kV, solution flow rate 2.5 mL/h, particle size range: 8-15 μm, (d) voltage 10 kV, ring voltage 6kV, flat plate voltage 7 kV, solution flow rate 2.5 mL/h, particle size range: 6-11 μm

Figure 2.31 Particle size distribution at various flat plate voltages (Vaef) Other

operating parameters are given by, nozzle voltage 10 kV, ring

voltage 6 kV, solution flow rate 1 mL/h

89

Figure 2.32 Particle size distribution at nozzle voltage 9 kV, flat plate voltage 5

kV, solution flow rate 1 mL/h and different ring voltages: (a) 7 kV, (b) 6 kV, (c) 5 kV

91

Figure 2.33 Particle size distribution at nozzle voltage 10 kV, ring voltag: 5 kV,

flat plate voltage 6 kV, and different solution flow rates (a) 2.5 mL/h, (b) 1 mL/h

92

Figure 2.34 Rosin-Rammler curve fitting for the particle size distribution

obtained at the following process parameters: nozzle voltage 9 kV, Ring voltage 7 kV, solution flow rate 2.5 ml/hr, and nitrogen flow rate 35 lit/min

93

Figure 3.1 Geometry for the EHDA deposition chamber showing the axial

component of the laminar nitrogen carrier gas flow and the relevant coordinates and geometrical parameters

108

Figure 3.2 Generalized plot of the particle collection efficiency as a function

of the dimensionless group defined in equation (5-33); ● data for varying the nitrogen flow rate; ■ data for varying the liquid flow rate; ▲data for varying the electric potential

Figure 4.4 Equipotential lines (blue lines) in (a) 2D electrospray deposition

geometry close to the substrate and (b) 3D electrospray deposition geometry

130

Figure 4.5 Electric field lines without considering the effect of space charge in

(a) 3D, and (b) 2D

134

Figure 4.6 The shape of individual particles in the patterns formed on the

substrate corresponding to following operational conditions: mask voltage 4 kV (a) nozzle voltage 8 kV, solution flow rate 0.4 mL/hr, distance between nozzle tip and substrate 20 mm, process time 30 min, (b) nozzle voltage 10 kV, solution flow rate 0.2 mL/hr, distance between nozzle tip and substrate 40 mm, process time 10 min, (c) nozzle voltage 9 kV, solution flow rate 0.1 mL/hr, distance between nozzle tip and substrate 30 mm, process time 20 min, (d) nozzle voltage 7 kV, solution flow rate 0.3 mL/hr, distance between nozzle tip and substrate 10 mm, process time 40 min

136

Figure 4.7 Particle patterns on the substrate observed in the following

operating conditions: Nozzle voltage 8 kV, mask voltage 3 kV, solution flow rate 0.3 mL/hr, process time 10 min and distance between tip of the needle and substrate (a) 10 mm, PDE: 92% (b)

20 mm, PDE: 89% (c) 30 mm, PDE: 87.7% and (d) 40 mm, PDE:

86.3%

137

Figure 4.8 Particle patterns on the substrate observed in the following

operating conditions: Nozzle voltage 7 kV, mask voltage 3 kV, solution flow rate 0.4 mL/hr, distance between tip of the needle and substrate 40 mm and process time (a) 10 min, PDE: 85.8%, (b) 20 min, PDE: 86.6%, (c) 30 min, PDE: 85.5%, and (d) 40 min, PDE:

85.9%

139

Figure 4.9 Particle patterns on the substrate observed in the following

operating conditions: Nozzle voltage 9 kV, mask voltage 4 kV, distance between tip of the needle and substrate 30 mm, process time 10 min and solution flow rate (a) 0.1 mL/hr, PDE: 87.9%, (b) 0.3 mL/hr, PDE: 87.6%, and (c) 0.4 mL/hr, PDE: 87.4%

140

Figure 4.10 Particle patterns on the substrate observed in the following

operating conditions: distance between tip of the needle and substrate 30 mm, process time 20 min and solution flow rate 0.3 mL/hr and (a) nozzle voltage 10 kV, mask voltage 4 kV, PDE:

88.7%, (b) nozzle voltage 8 kV, mask voltage 3 kV, PDE: 87.5%, and (c) nozzle voltage 7 kV, mask voltage 3 kV, PDE: 85.8%, and (d) nozzle voltage 9 kV, mask voltage 7 kV, PDE: 84.2%

141

Figure 4.11 Particle trajectories (left) and particles captures on the substrate

(right) at the following operating conditions: fixed process time, (a) and (b) nozzle voltage 8 kV, mask voltage 4 kV, solution flow rate 0.3 ml/hr, PDE 86.6%, (c) and (d) nozzle voltage 9 kV, mask

143

voltage 4 kV, solution flow rate 0.4 ml/hr, PDE 88.3%, (e) and (f) nozzle voltage 9 kV, mask voltage 3 kV, solution flow rate 0.2 ml/hr, PDE 90.1%, (g) and (h) nozzle voltage 10 kV, mask voltage

3 kV solution flow rate 0.3 ml/hr, PDE 91.4%

Figure 4.12 PLGA particle patterns formed on the mask observed in the

following operating conditions: Nozzle voltage 10 kV, substrate voltage -2 kV, distance between nozzle and substrate 40 mm, process time 20 min and solution flow rate (a), (b), (c) 0.1 mL/hr, (d), (e), (f) 0.3 mL/hr, and (g), (h), (i), (j) 0.4 mL/hr

146

Figure 4.13 PLGA particle patterns formed on the mask observed in the

following operating conditions: Nozzle voltage 9 kV, substrate voltage -3 kV, distance between nozzle and substrate 30 mm, solution flow rate 0.2 mL/hr and process time (a), (b), (c) 10 min, (d), (e), (f) 20 min, and (g), (h), (i), (j) 40 min

147

Figure 4.14 PLGA particle patterns formed on the mask observed in the

following fabrication conditions: Nozzle voltage 8 kV, distance between nozzle and substrate 30 mm, solution flow rate 0.4 mL/hr and process time 30 min, substrate voltage (a), (b), (c) -4 kV, (d), (e), (f) -3 kV, and (g), (h), (i), (j) -1 kV

148

Figure 4.15 PLGA particle patterns formed on the mask observed in the

following operating conditions: Nozzle voltage 9 kV, substrate voltage -2 kV, solution flow rate 0.3 mL/hr, process time 20 min and distance between nozzle and substrate (a), (b), (c) 10 mm, (d), (e), (f) 20 mm, and (g), (h), (i) 40 mm

150

Figure 5.1 Cylindrical sensor with an internal electrode known as

Internal-External Electrode (IEE)

168

Figure 5.2 ECT sensor installed on chamber Dashed line show internal

electrode installed in chamber

168

Figure 5.3 Experimental setup for water-air/DCM-air system (with syringe)

and EHDA system

170

Figure 5.4 Sensitivity maps for IEE sensor (a) electrode pair 2-5, (b) electrode

pair 3-5, (c) electrode pair 4-5, (d) electrode pair 12-5, (e) electrode pair 13-5, (f) electrode pair 14-5

172

using LBP algorithm with volume ratio of (a) dripping mode, (b) spray mode

Figure 5.6 Results of DCM-air system with charge/discharge ECT system

using LBP algorithm with volume ratio of (a) dripping mode, (b) spray mode

174

Figure 5.7 Results of water-air system in dripping and double dripping modes

close to chamber’s wall with ACECT system using LBP and Landweber (3 iterations) algorithms

175

Figure 5.8 Results of water-air system in dripping and double dripping modes

in central region of chamber with ACECT system using LBP and Landweber (3 iterations) algorithms

176

Figure 5.9 Results of water-air system in spray mode close to chamber’s wall

(EE sensor) and in center of chamber (IEE sensor) with ACECT system using LBP algorithm

177

Figure 5.10 Results of DCM-air system in dripping and spray modes close to

chamber’s wall (EE sensor) and in center of chamber (IEE sensor) with ACECT system using LBP and Landweber (3 iterations) algorithms

177

Figure 5.11 Results of EHDA system (peripheral area) in dripping and spray

modes with ACECT system using Landweber (3 iterations) algorithm

179

Figure 5.12 Results of EHDA system (central area) in dripping and spray

modes with ACECT system using Landweber (3 iterations) algorithm

180

Nomenclature

uf = fluid velocity vector, m s-1

up = particle velocity vector, m s-1

ud = droplet velocity vector, m s-1

m p = mass of the particle, kg

m d = mass of the droplet, kg

 = the droplet density, kg m-3

f

 = fluid density, kg m-3

g = gravitational acceleration vector, m s-2

E= electrical field vector, V m-1

 = the space charge density, C m-3

μ = the electric mobility of a charged particle

 = the viscous stress tensor

 = the fluid viscosity, Pa.s

ε 0 = the permittivity of vacuum, C S2 kg-1 m-3

ε r = the relative permittivity of the medium

ε l = the electrical permittivity of the liquid solution

D i,m = the diffusion coefficient of vapor in the bulk, m2 s-1

Dvap = the diffusion coefficient of the vapor in the ambient surrounding, m2 s-1

k c = is the mass transfer coefficient, m s-1

T = the absolute temperature, K

M = the molecular weight

A d = surface area of the droplet, m2

cvap = the specific heat capacity of mixture, J Kg-1 K-1

H L = the latent heat of vaporization of liquid solvent, J K-1

T∞ = the ambient temperature, K

G = the shear rate, s-1

U = the average velocity of the carrier gas flow, m/s

L e = the entrance length required to achieve fully developed flow, m

L c = the vertical position of the particle above the collecting plate, m

S ij (x, y) = the field sensitivity corresponding to electrode pair i and j

Chapter 1 Introduction

1.1 Background

Many recent studies are concentrated on the invention of new biomaterials for biomedical applications such as in controlled drug delivery systems, tissue engineering, biosensors and biochips However, most of them are concentrating on the material part and do not cover the process part

Recently several novel methods were developed for the preparation of biodegradable controlled release drug delivery devices Firstly, direct analysis method is a process in which includes the exchange of organic solvents and water via diffusion through a dialysis membrane with certain cut-off molecular weight The polymer molecules are self-assembled to form nanoparticles or microparticles inside the dialysis membrane Direct dialysis method was developed for the simple preparation of drug carriers such as lipsomes and polymeric micelle (Jeong et al 2001) More recently, surfactant free nanoparticles of Poly (D, L-lactide-co-glycolic acid) or PLGA were prepared for controlled released drug delivery systems (Jeong et al 1998; Jeong et al 2003) Their work mainly focused on particle fabrication, characterization and in vitro release studies However, paclitaxel was not employed in their system which is considerably used in this thesis

Secondly, which is used for fabricating pharmaceutical micro and nanoparticles, is atomizing a bulk liquid to fine droplets The interest in atomizing liquids comes from a wide variety of industries Automotive engine fuel injection, pesticide spraying, spray

painting and spraying of medicine are some of the fields where a liquid is needed in atomized form instead of the bulk volume Better transport, better deposition and better combustion are some of the benefits of an atomized liquid Many devices can be used to generate sprays from a bulk liquid These devices, commonly called atomizers, can be found in many industrial, agricultural and propulsion systems

The transportation of the bulk liquid and consequent spraying can be achieved with different sources of energy (Bayvel and Orzechowski 1993) Most often high pressure is applied to a nozzle causing the liquid to disintegrate and form a spray In electrohydrodynamic Atomization (EHDA), or electrospray, an electric force is the main source of energy that is used to atomize the liquid In electrospray, an electric field is applied between a nozzle, to which a liquid is supplied, and a counter electrode The liquid at the tip of the nozzle is accelerated through the columbic interaction includes both the formation of a liquid jet and subsequent disruption into charged droplets With electrospray a spray of charged droplets from nanometers up to tens of micrometers is generated In the electrospraying, no additional mechanical energy, other than that from the electric field alone, is needed for liquid atomization Electrospraying allows generation of fine droplets of charge magnitude close to one-half of the Rayleigh limit The Rayleigh limit is the magnitude of charge on a drop that overcomes the surface tension force, which leads to fission of the droplet The charge and size of the droplet can

be easily controlled to some extent by adjusting the flow rate and voltage applied to the nozzle Depending on the spray conditions, monodisperse droplets can be produced The charge on the droplets prevents agglomeration of the droplets and promotes dispersion of the droplets These qualities make the technique of electrospray interesting for many

applications The electrospraying has some advantages over conventional mechanical spraying systems with droplets charged by induction:

 Droplets have size smaller than those available from conventional mechanical atomizers, and can be smaller than 1 μm

 The size distribution of the droplets is usually narrow, with low standard deviation; droplets can be of equal size only for dripping and micro-dripping modes, or for Rayleigh jet breakup due to varicose wave instability

 Charged droplets are self-dispersing in the space, which results in absence of droplet agglomeration and coagulation

 The motion of charged droplets can be easily controlled (including deflection or focusing) by electric fields

 The deposition efficiency of charged spray on an object is much higher than for un-charged droplets

In the literature, the term electrospraying is often used to describe the process of electrostatic spraying However, this term also refers to processes where atomization takes place by other sources of energy, with simultaneous or subsequent electrical charging The electric field is primarily used to charge the liquid and to transport the charged droplets Electrospray has been also applied to spray liquid metals As these liquids have a much higher conductivity compared to normal liquids, the spraying mechanism is different Therefore, these applications are not included in this thesis The pharmaceutical applications are mainly discussed in this thesis

The phenomenon of an electric effect on liquid menisci is known since sixteenth century The interaction between a water droplet and a piece of amber, leading to a conical shape

droplet was reported at 1600 (Gillbert 1600) In the beginning of the 20th century, Zeleny showed that liquid menisci subjected to a high enough electric field change to a conical shape and emit a mist of very small droplets (Zeleny 1916; Zeleny 1917) In the 1950’s, Vonnegut and Neubauer tried to relate the properties of a liquid to their ability of being sprayed (Vonnegut and Neubauer 1952)

Taylor made the first mathematical description of the cone by balancing the forces that are present At the beginning of the nineties of the 20th century, scaling laws for electrospraying of liquids were developed, to predict the characteristics of the produced droplets on forehand (Taylor 1964) Moreover, a physical model that describes the shape

of the cone and the droplet size of the droplets in the spray was derived (Hartman et al 1999) These models refer to a so called cone-jet system, one of the modes that can be obtained in electrospray

Electrospray is a powerful technique for easy scaling up trough the use of multiplexed microfabricated source spray (Ding et al 2005) Fuerthermore, it could be also miniaturized to the portable devices which may encapsulate and delivery drugs on demand (Yeo et al 2005) Electrospray technique was widely employed in many areas such as fabrication of inorganic nanoparticles (Lenggoro et al 2002), preparation of thin films (Buchko et al 1999; Uematsu et al 2004), deposition of nanoparticle clusters (Jayasinghe et al 2004), micro/nano encapsulation (Loscertales et al 2002; Berkland et

al 2004) and production of pharmaceutical particles (Ijsebaert et al 2001)

Due to the recent emergence of the application of electrospray, only a few researchers are studying polymeric particle fabrication using electrospray It is reported that PLGA particles could be obtained by flow-limited field injection electrostatic spraying and the

smallest particle size was around 300 nm (Berkland et al 2004) However, in their experimental setup the liquid was sprayed in the atmosphere and the solvent evaporation rate was not optimized and particle shape was not spherical Poly-lactic acid (PLA) microparticle fabrication using AC electrospray was investigated in which particles were collected in a collector tube In this study the collector was rinsed with deionized water and its contents passed through a porous membrane filter to recover the microspheres (Yeo et al 2005) However, it seemed that the particle size distribution was wide and the procedure was tedious In addition, bovine serum albumin solid particles were encapsulated in polymeric microparticles by extending the application of electrostatic extrusion (electrospray in the dripping mode) A suspension of protein particles within a hydrophobic polymer in organic solvent was employed to generate microbead (average diameter around 0.5-1.5 μm) containing solid protein particles with the diameter around several tens microns to 100 microns (Amsden and Goosen 1997) Therefore, the major problem was that the particle size was very large and was hard to administer the dosage form Moreover, the protein was not uniformely distributed in the microbead and the release rate was not obtained in a controlled manner

Micro-patterning is an important application of electrospray process It is an essential process in the fabrication of semiconductor devices, micro-electro mechanical systems (MEMS), flat panel displays, biosensors, biochips and so on Micro-patterning with nanometer resolution is already well established in patterning processes of metals or inorganic materials, especially for semiconductors or MEMS devices In most of these processes, evaporation, sputtering or chemical vapor deposition (CVD) is used to form thin films which are patterned by wet or dry etching using photoresist masking On the

other hand, micro-patterning processes for organic materials, synthetic polymers or bioactive materials including DNA or protein require different technologies because these materials are usually not compatible with the combination of photoresist masking and

wet/dry etching processes that use solvents, acidic/alkaline solutions, or plasma made

from corrosive gases A variety of processing techniques have been developed for the formation of micropatterns, such as sputtering, chemical and physical vapor deposition, the sol–gel process, spray pyrolysis, dip coating, spin coating, screen printing, and tape casting In a spray deposition process, precursor solutions or colloidal suspensions are atomized and transported onto a substrate to deposit thin and thick solid films or patterns Spray deposition using a precursor solution is often processed with a heated substrate to induce chemical reactions Compared with other film and pattern formation techniques, the spray deposition method has the advantages of a simple setup, relatively low cost for equipment, excellent chemical composition control, a continuous process, ambient atmosphere operation, a high growth rate, and a wide choice of inexpensive and non-toxic precursors Electrospray is one of the most significant methods employed for micropatterning The main advantage of electrospray deposition over other types of spray deposition is its high deposition efficiency, as the charged droplets are attracted toward a substrate by Coulomb force (Siefert 1984) The unipolar charges on electrosprayed droplets also prevent the coagulation of droplets during flight and lead to uniform self-dispersion over the substrate due to the mutual electrical repulsion between droplets Moreover, electrospray deposition does not require a carrier gas to transport the aerosols, which is particularly advantageous in tailoring the film morphology (Chen et al 1996)

Electrospray has a variety of functioning modes, such as microdripping, spindle, cone-jet, and the multi-jet mode, depending on the applied voltage, flow rate, electrode configuration, and the physical properties of the liquid, such as electrical conductivity, viscosity, surface tension, and density (Cloupeau and Prunetfoch 1994)

An important issue in EHDA process is particle collection efficiency or particle deposition efficiency in electrospray deposition process It was previously reported that EHDA technique is able to produce microparticles from 1 to 15 μm but the yield was sometimes lower than 10% (Ding et al 2005) However, the operating parameters in the fabrication process were not fully optimized

Although many studies have been reported in the literature about different productions and applications of EHDA process but there are a few researches concentrated on processing part In other words, most of the works performed in this area have focused on downstream and there is an important lack in scientific works focused on upstream The most important factor affecting the feasibility of EHDA process is particle collection efficiency This factor is exceedingly important for designing EHDA system especially from the point of view of operating costs Since the materials used for pharmaceutical particle fabrication are expensive, higher particle collection efficiency is extremely desirable Therefore, research gap for the current study of pharmaceutical particle fabrication by electrohydrodynamic atomization (EHDA) process is particle collection efficiency

The present work mainly aimed to develop a deeper understanding of the EHDA process for polymeric particle fabrication and micropatterning and also to achieve better

collection and particle deposition efficiency by adjusting some of the operational and geometrical parameters The specific objectives of this research are:

 Applying EHDA process in a new generation of shuttle chamber which helps to enhance the collection efficiency and decrease the residual solvent in collected particles

 Using Taguchi as a statistical experimental optimization method was used to design the experiments and analyze the results to obtain the optimum conditions for gaining the highest collection efficiency

 Employing Computational Fluid Dynamics (CFD) to simulate EHDA process in the chamber and also in electrospray deposition process to track the particle trajectories in order to predict particle collection or deposition efficiency in different operating and geometrical parameters

 Developing scaling analysis on both EHDA process in the chamber and electrospray deposition to model these processes and finding the most important forces acting in these systems which can be effectively used in process optimization

 Employing Charge/Discharge Electrical Capacitance Tomography (ECT) and AC-based Electrical Capacitance Tomography (ACECT) to consider the particle flow and trajectories in cross-section of the chamber

The results of the present study may have significant impact on the quality of the fabricated particles from shape, morphology and size points of view More importantly, enhancing the particle collection efficiency and decreasing the residual solvent can be

two significant developments, which are now two important issues in fabricating particles

by EHDA process Moreover, enhancing the particle deposition efficiency and also improving the quality of micropattarns are the other considerable achievements of this research work

It is really understood that studying on Electrohydrodynamic Atomization process cannot

be independent of cone-jet formation at the tip of high voltage nozzle However, the process of forming cone-jet and different types of that is a very complicated process and requires very deep and independent study Therefore, this topic is not central in the current study and hence is beyond the scope of this thesis Moreover, investigating the characteristics of the final products of EHDA process such as shape and morphology is not covered in this PhD project

1.2 Overview of this thesis

Reffering to its title, this thesis deals with studying different aspects of EHDA process which are affecting the particle collection efficiency Several methods such as designed experiments (by Taguchi method), CFD simulation, scaling analysis and ECT are employed in this work to study the EHDA process and enhance the particle collection yield Figure 1.1 shows the schematic diagram for the overview of the present study

Chapter 1 of this thesis gives an overview of EHDA process General introduction,

objectives and organization of this study are presented in this chapter Taguchi method as

a statistical method is used in Chapter 2 to design the experiments in which the optimum

operating conditions are obtained for having higher collection efficiency in EHDA process in the chamber Moreover, a new generation of EHDA encapsulation chamber is

designed in this chapter FLUENT and COMSOL as computational fluid dynamics (CFD) softwares are also used in this chapter to simulate the EHDA process in the chamber The simulation results obtained are validated against experimental data collected using the actual EHDA setup with the same geometry and applying the same operating conditions Furthermore, this chapter aims to investigate the effect of Auxiliary Electric Field (AEF) on particle motion and deposition in the EHDA encapsulation chamber Furthermore, the influences of the auxiliary electric field on particle collection efficiency, particle size distribution and morphology are examined Scaling analysis is

employed in Chapter 3 to model the EHDA process in the chamber in both fluid

mechanic and electrostatic modes to find the most effective forces in the process

Chapter 4 determines a set of parameters that can help to achieve most focused

micropatterns of drug particles on the substrate The unresolved objectives in traditional electrospray deposition such as the reduction on the loss of raw materials, enhancement

on the deposition efficiency, increase in the spot density, and reduction on the large inherent inter-spot distance are also investigated in this chapter A new type of AC-based

ECT (ACECT) sensor is designed in Chapter 5 to study the particle motion in the

chamber because the particle distribution in cross-section of the chamber is important to

control the particle collection efficiency Finally, Chapter 6 gives general conclusions

and recommendations for future works

Figure 1.1: Schematic diagram of overview of this thesis

Chapter 2

Experimental and Computational Studies of Electrohydrodynamic

Atomization process in Encapsulation Chamber for

Pharmaceutical Particle Fabrication to Enhance the Particle

Collection Efficiency

2.1 Introduction

Since the last few decades in the field of pharmacy and medicine, many new products and technologies have been developed Biodegradable polymeric micro- and nanoparticles are found among a group of pharmaceutical products which are broadly used in drug delivery applications (Jiang et al 2005; Kreuter 2001) Different methods, such as solvent evaporation, single emulsion and double emulsion, are typically employed to fabricate biodegradable polymeric micro- and nanoparticles However, the most obvious drawback

in these methods is the size distribution which is wide and the process is hard to scale up Most significantly, bio-functionalities of drugs were possibly compromised as a result of exposure to organic solvents, high shear stress and aqueous organic interfaces (Tamber et

al 2005; Freitas et al 2005)

EHDA is a well-known phenomenon for production of fine droplets of uniform size (Almekinders and Jones 1999; Tatemoto et al 2007) EHDA in the cone-jet mode has been studied for many years (Hartman et al 1999) It has been demonstrated to be useful for fabricating particles with better morphologies than other conventional methods such

as spray drying (Almekinders and Jones 1999; Tatemoto et al 2007) EHDA has various

applications such as electrospray ionization in mass spectroscopy (Fenn et al 1989), electrospray deposition of thin films (Buchko et al., 1999), pharmaceutical productions (Ijsebaert et al 2001; Tang and Gomez 1994), and polymeric particle fabrications for drug encapsulation (Ding et al 2005; Rezvanpour et al 2010 ; Yao et al 2008 ; Farook et

al 2007) With controlled solvent evaporation during the particle fabrication process, further enhancements in terms of narrow polydispersities and smooth, spherical morphologies of particles can be achieved (Hartman et al 1999) The underlying concept

of Electrohydrodynamic Atomization is in the application of an electric field to a liquid that is flowing out of a capillary tube so as to induce hydrodynamic instabilities in the liquid stream which then gives rise to the phenomenon of atomization As such, the electric potential difference applied needs to be sufficiently high in order to generate a strong electric field which is responsible for inducing instability within the liquid stream before atomization occurs Both physical properties and operating conditions such as the electrical conductivity, surface tension, viscosity, flow rates and surrounding gas play important roles in determining the onset of hydrodynamic instabilities in liquids flowing through capillary tubes (Hayati et al 1987) In the EHDA process, such physical properties also have significant effects on the formation of the liquid cone (Barrero 1996) The formation of a stable liquid cone is essential for operation in the cone-jet mode which produces droplets that are roughly one order of magnitude smaller than the diameter of the capillary tube (Lopez-Herrera et al 1999; Noymer and Garel 2000; Ragucci et al 2000; Farook et al 2007; Lastow and Balachandran 2007)

The stability and atomization characteristics of electrohydrodynamic jets can experimentally be investigated (Lopez-Herrera et al 1999; Farook et al 2007; Lastow

and Balachandran 2007) According to some research results, the cone-jet mode exists at lower level of applied voltage, while the multi-jet mode exists at higher voltage levels (Noymer and Garel 2000; Ragucci et al 2000) Stable operation in the cone-jet mode produces droplets that are roughly three orders of magnitude smaller than the capillary diameter (Ku and Kim 2003; Lenggoro et al 2007) One of the most important applications of EHDA process is production of controlled size and controlled physical property polymer particles The research results demonstrate that water soluble and water insoluble, low dispersity polymer particles can be readily prepared by EHDA process with geometric mean diameters in the micron size It shows that this method can be used for production of advanced polymer materials (Hogan et al 2007) Therefore, EHDA can

be used in fabrication of biodegradable polymeric micro- and nanoparticles which are widely used in drug delivery systems (Kreuter 2001; Jiang et al 2005) The practical aspects of production of polymeric particles loaded with a drug by electrohydrodynamic atomization method are presented in some research articles The research results state that quickly evaporating solvents have a tendency to form hollow particles Drug release rate from hollow particles is substantially higher than from solid ones (Ciach 2006) The electrospray process in the dripping mode can be also used for cell microencapsulation Microencapsulation of living cells with controllable size and narrow size distribution can

be obtained using electrospray in a dripping mode and can be stabilized by a ring electrode The findings of this process are very useful in the scaling up of the mentioned process to generate large quantities of microbeads These microbeads can be employed as biosorbents for removal of metal ions in water and as delivery devices for controlled release of drugs Moreover, the effect of operating conditions like electrical field

strength, liquid and inert gas flow rate on particle size and morphology are studied and the results are reported in some research articles (Xie et al 2006; Xie and Wang 2007; Hong et al 2008; Yao et al 2008)

Apart from experimental studies, various aspects of the EHDA process have been investigated computationally An air-assisted electrostatic induction charging spray nozzle was modeled for both flat and spherical targets(Zhao et al 2007) The airflow, liquid droplets as well as the electrostatic field were considered for calculation of trajectories of the charged droplets The effects of parameters such as droplet size, charge-to-mass ratio and nozzle-to-target distance on the motion of the charged droplets were investigated The results confirmed that the spray cloud expanded with increasing droplet charge-to-mass ratio and nozzle-to-target distance due to increasing space charge(Zhao et al 2007) Nonlinear breakup of charged liquid jets was analyzed numerically in the limit of very small electrical Strouhal numbers T T e/ b 1(i.e negligible charge relaxation effects, applicable to highly conducting liquids), where T is the electric e

relaxation time of charges, and T is the breakup time in a Lagrangian framework b

following the liquid jet at its average axial velocity (Lopez-Herrera et al 1999) The influence of the electrical Bond's number and viscosity on the capillary Rayleigh's most probable breakup length, the breakup time, the volume of the satellite, and the charge of both main drop and satellite were analyzed The electrical Bond's number is a dimensionless number expressing the ratio of body forces (electric forces) to surface tension forces The findings confirmed that the influence of the electrical Bond's number

on Rayleigh's length was small within the usual parametric limits of stability of a steady Taylor cone-jet at atmospheric pressure (Lopez-Herrera et al 1999) The commercial

Computational Fluid Dynamics (CFD) software known as CFX 4.4 had also been used to simulate the EHDA process (Lastow and Balachandran 2007) The heat conduction equation was modified to represent the electrostatic field and electric body forces were determined during the calculations The calculated velocity fields for an EHDA process using heptane and ethanol as the operating fluids were found to be consistent with published results The model applied did not include a droplet break-up model and droplet size calculated based on jet diameter was found to compare well with experiments(Lastow and Balachandran 2007) The atomization of water had also been studied both experimentally and computationally using CFD(Lastow and Balachandran 2006) Based

on the experimental results the atomization of water occurred outside the applicability range of the scaling laws due to the high dielectric constant of water and the low flow rate used The experimental results also showed that droplet size remained approximately constant with increasing flow rate In other recent studies, the effects of various operating conditions such as electric field strength, liquid and inert gas flow rate on particle size and morphology were reported (Xie et al 2006; Xie and Wang 2007; Hong et al 2008; Yao et al 2008)

As discussed, the characteristics of final products of EHDA process such as shape and morphology are extensively investigated in the literature However, particle collection efficiency and residual solvent in collected particles are two important issues in particle fabrication using EHDA process which are not addressed in the previous articles These two factors are exceedingly important for designing EHDA system especially from the point of view of operating costs Since the materials used for pharmaceutical particle fabrication are expensive, higher particle collection efficiency is extremely desirable

Moreover, allowable concentration of residual solvent in collected particles after EHDA process and before freeze-drying is desirable because it eliminates freeze-drying as an additional energy-consuming process The present chapter aims to employ EHDA technology to fabricate particles in a new generation of shuttle chamber Here, we investigated the production process with focus on the particle collection efficiency and residual solvent in the collected particles The experiments are designed based on the Taguchi (combined statistical and experimental) optimization method Subsequently, the results are summarized in different tables and figures In other words, we intend to develop a deeper understanding of the EHDA process for polymeric particle fabrication and also to achieve better collection efficiency and solvent evaporation by adjusting some

of the operational and geometrical parameters

At the next section, CFD simulations of both particle and fluid phases within the EHDA chamber were carried out in this chapter using FLUENT 6.3 and COMSOL 3.5 The simulation results obtained were validated against experimental data collected and employed to better justify the experimental observations In the last section of this chapter, one additional flat plate is positioned a few centimeters above the collecting plate, which is connected to a positive high voltage generator Indeed, the target in this process is the enhancement of the particle deposition on the collecting plate using the additional electric field In summary, this section intends to examine the effect of Auxiliary Electric Field (AEF) on particle motion and deposition in the EHDA encapsulation chamber Furthermore, the influences of the auxiliary electric field on particle collection efficiency, particle size distribution and morphology are examined