university of groningen polymer melt micronisation using supercritical … · 2016-03-07 ·...

University of Groningen

Polymer melt micronisation using supercritical carbon dioxide as processingNalawade, Sameer

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2005

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Nalawade, S. (2005). Polymer melt micronisation using supercritical carbon dioxide as processing.[Groningen]: s.n.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 05-07-2020

https://www.rug.nl/research/portal/en/publications/polymer-melt-micronisation-using-supercritical-carbon-dioxide-as-processing(67607fc4-231d-4e15-af75-008f66475b4e).html

https://www.rug.nl/research/portal/en/publications/polymer-melt-micronisation-using-supercritical-carbon-dioxide-as-processing(67607fc4-231d-4e15-af75-008f66475b4e).html

Polymer Melt Micronisation using Supercritical Carbon Dioxide as Processing Solvent

Sameer P. Nalawade

This research was financially supported by Technologiestichting STW, Postbus 3021, 3502 GA Utrecht, under the project DWT 4939.

RIJKSUNIVERSITEIT GRONINGEN

Polymer Melt Micronisation using Supercritical Carbon Dioxide as Processing Solvent

Proefschrift

ter verkrijging van het doctoraat in de

Wiskunde en Natuurwetenschappen

aan de Rijksuniversiteit Groningen

op gezag van de

Rector Magnificus, dr. F. Zwarts,

in het openbaar te verdedigen op

vrijdag 2 december 2005

om 13.15 uur

door

Sameer P. Nalawade

geboren op 19 juni 1978

te Mumbai, India

Promotor : Prof.dr.ir. L.P.B.M. Janssen Copromotor : Dr. F. Picchioni Beoordelingscommissie : Prof.dr.ir. P.J. Jansens Prof.dr.ir. J.T.F. Keurentjes Prof.dr.ir. H.J. Heeres

ISBN 90-367-244-06 ISBN 90-367-244-14 (electronic version) © 2005 by Sameer P. Nalawade All rights reserved. No part of this publication may be reproduced, stored in a retrieval system of any nature, or transmitted in any form or by any means, electronic, mechanical, now known or hereafter invented, including photocopying or recording, without prior written permission of the author.

CONTENTS 1. Supercritical CO2 as a green solvent for processing polymers 1 1.1. Introduction 2 1.2. Supercritical carbon dioxide 3 1.2.1. Solubility of CO2 in polymers 3 1.2.2. Viscosity reduction 6 1.3. Supercritical CO2 in the production of micron-size particles 7 1.3.1. Rapid expansion of supercritical solution 8 1.3.2. Supercritical anti solvent methods 9 1.3.3. Particles from gas saturated solution 10 1.4. Selection of a method for the production of polymer particles 11 1.5. Purpose and out line of the thesis 11 References 13 2. FT-IR studies on the interactions of CO2 and polymers having different chain groups 17 2.1. Introduction 18 2.2. Experimental 19 2.2.1. Materials 19 2.2.2. Apparatus 20 2.2.3. Film preparations 21 2.2.4. Experimental method 22 2.3. Results and discussion 22 2.3.1. PPO 22 2.3.2. PPB and PEB 24 2.3.3. PEG 26 2.3.4. Discussion 28 2.4. Conclusions 31 References 31 3. Solubilities of sub- and supercritical CO2 in polyester resins: measurements and prediction 33 3.1. Introduction 34 3.2. Experimental 35 3.2.1. Materials 35 3.2.2. Apparatus and method 35 3.3. Sanchez-Lacombe equation of state 39 3.4. Results and discussion 40 3.5. Conclusions 45 References 45 4. Batch production of micron size particles from poly(ethylene glycol) using supercritical carbon dioxide as a solvent 47 4.1. Introduction 48 4.2. Experimental 49 4.2.1. Materials 49 4.2.2. Apparatus 49 4.2.3. Experimental method 50 4.2.4. Particle analysis 50 4.3. Results and discussion 50

4.3.1. Literature: CO2 solubility and viscosity of PEG-CO2 50 4.3.2. Batch production of PEG particles 52 4.3.3 Solidification 59 4.4. Conclusions 61 References 62 5. Prediction of the viscosity reduction of PPB due to dissolved CO2 and an

elementary approach in the supercritical CO2 assisted continuous particle production 63

5.1. Introduction 64 5.2. Viscosity reduction theory 65 5.3. Experimental 67 5.3.1. Materials 67 5.3.2. Apparatus 67 5.3.3. Experimental method 71 5.3.4. Particle analysis 71 5.4. Results and discussion 72 5.4.1. Viscosity prediction results 72 5.4.2. Twin screw extruder results 75 5.4.3. Static mixer (Kenics type) results 75 5.5. Conclusions 83 References 83 6. An engineering study of supercritical CO2 assisted continuous polymer micron-size particles production using an SMX mixer 85 6.1. Introduction 86 6.2. Experimental 87 6.2.1. Modified set up 87 6.2.2. Experimental method 88 6.2.3. Particle analysis 89 6.3. Results and discussion 89 6.3.1. Principle component analysis 89 6.3.2. Vital roles of processing parameters in the particle production 95 6.3.3. Shape and morphology of PPB particles 101 6.3.4. Solidification 109 6.3.5. Properties of polymer after and before processing 110 6.3.6. Dimensional analysis 111 6.4. Conclusions 113 References 113 7. Technological assessment and prospects of supercritical CO2 115 7.1. A qualitative technological evaluation 116 7.2. Prospects of supercritical CO2 118 7.2.1. A continuous production of particles from polyester epoxy

resin melt 118 7.2.2. Supercritical CO2 in microcellular foaming 118 7.3. Conclusions of the thesis and future outlook 122 References 123

List of symbols 125 Summary 126 Samenvatting 128 Acknowledgements 131 Publications 133

Chapter 1: Supercritical carbon dioxide as a green solvent for processing polymers 1

Chapter 1

Supercritical carbon dioxide as a green solvent for processing

polymers Supercritical fluids are well established as a processing solvent in various polymer applications such as polymer modification, formation of polymer composites, polymer blending, microcellular foaming, particle production and polymerization. As carbon dioxide (CO2) is quite soluble in many polymers, it can be used as a solvent or plasticizer. Apart from an inert nature and easily attainable supercritical conditions, gas-like diffusivity and liquid-like density in the supercritical phase allow replacing conventional, often noxious, solvents with supercritical CO2. Dissolved CO2 causes a considerable reduction in the viscosity of molten polymer, a very important property for the applications stated above. In this chapter, solubility and viscosity measurement studies and various particle production methods which use supercritical CO2 have been discussed. Finally, the purpose of the thesis and its contents have been discussed. Sameer P. Nalawade, Francesco Picchioni, L. P. B. M. Janssen, Supercritical carbon dioxide as a green solvent for processing polymer melts: processing aspects and applications, a review, Progress in Polymer Science, in press.


1.1. Introduction Polymers have become an inseparable part of daily life. Not only the synthesis but also the processing of polymers has been given a major attention. Processed polymers are used for particular applications in particular forms as foam or blend or powder. For example, powder industries such as paint, toner and drug always seek for the solutions that provide particles of micron or nano size with a narrow particle size distribution. There are classical methods that use environmentally hazardous volatile organic solvents (VOC) and chlorofluorocarbons (CFC) for processing and synthesis of polymers. Due to the enormous increment of VOC/CFC emissions and also the generation of aqueous waste streams a number of chemical engineers and chemists have already been looking for new and cleaner alternatives. One of these methods is the use of supercritical fluids as a processing solvent. A supercritical fluid is defined as a substance for which both pressure and temperature are above the critical values, Fig. 1.1.

Temperature

Pres

sure

T

SolidLiquid

Gas

SCF

P C

Tc

c

Fig. 1.1. A phase diagram of a substance approaching the supercritical phase, C: critical point Though a supercritical fluid doesn’t contain two phases such as gas and liquid, it possesses the properties of both gas and liquid. The special combination of gas-like viscosity and liquid like-density of supercritical fluid results in it being an excellent solvent. The density of supercritical fluid can be tuned easily by small changes in pressure. It has successfully been used as a solvent in the processing of polymers such as blending, microcellular foaming and particle production, in


extraction applications and in polymer synthesis [1-4]. In addition, the improved product quality is also an important concern for its selection [1-4]. 1.2. Supercritical carbon dioxide One of the supercritical fluids, which can be used, is carbon dioxide (CO2). Supercritical CO2 is a clean and versatile solvent and a promising alternative to noxious organic solvents and chlorofluorocarbons. It has attracted particular attention as a supercritical fluid in the synthesis as well as processing areas for polymers due to the following properties: (a) CO2 is non- toxic, non-flammable, chemically inert, and inexpensive. A large amount is available as a by-product from NH3 and ethanol industries and refineries. (b) Supercritical conditions are easily achieved: Tc = 304 K and Pc = 7.38 MPa. (c) The solvent may be removed by simple depressurization. (d) The density of the solvent can be tuned by varying the pressure. (e) Many polymers become highly swollen and plasticized in the presence of CO2. Moreover, the use of supercritical CO2 doesn’t create a problem with respect to the green house effect as it is being conserved during the processes.

CO2 is a good solvent for many non-polar (and some polar) molecules with low molecular weight [5]. It is a very poor solvent for most high molecular weight polymers under readily achievable conditions. Very few polymers have shown a good solubility in pure CO2 under mild conditions like certain amorphous fluoropolymers and silicones [4,6-10]. Though the solubility of most polymers in supercritical CO2 is extremely low, the solubility of supercritical CO2 in many polymers is substantial. The concentration of dissolved CO2 in polymer mainly depends on the processing temperature and pressure. Now, the question arises is the dissolved CO2 beneficial from a processing point of view? The dissolved CO2 causes a considerable reduction in viscosity due to increase in a free volume of polymer. Thus, less energy is consumed during the process. The dissolved CO2 also alters the other physical properties such as reduction in density and increase in diffusion coefficient. Therefore, it has a tremendous potential as a plasticizer in polymer processing. 1.2.1. Solubility of CO2 in polymers The knowledge of gas solubility in molten polymers is crucial for the commercial success of supercritical-polymer processes. The dissolved supercritical CO2 in molten polymers alters most physical properties of the polymers like the viscosity, density, diffusivity and swollen volume. A lot of attention has been paid to the situations where polymers are dissolved in


supercritical CO2. Only a limited part of this concerns molten polymers, the most likely form for processing. The reviews by Cooper [1] and Tomasko et al. [2] cover extensive information on the various applications of supercritical CO2 to polymer synthesis and processing. Recently, Kendall et al. [3] have nicely reviewed polymerizations in supercritical CO2. In general, an increase in pressure increases the solubility of a gas in a solvent. The same law is applicable also to a polymer and CO2. The density of CO2, which is a strong function of temperature and pressure plays a vital role in deciding its solubility in a polymer. However, the quantity of CO2 dissolved in different polymers also differs depending on the available chemical groups. A difference in the solubility can be explained with the specific intermolecular interaction between CO2 and the chemical groups available in the polymers. Several studies have been carried out to reveal the interactions between a polymer and CO2. CO2 does not have a dipole moment due to its structural symmetry. However, a quadrupole moment and Lewis acidity contribute to its solubility in a polymer. Using FT-IR spectroscopy Kazarian et al. [11,12] have given spectroscopic evidence of the Lewis acid-base interactions between CO2 and a polymer. They used the bending mode rather than the stretching mode of CO2 as it is much more sensitive to the Lewis acid-base interaction. The splitting of the bending mode of CO2 signified the interaction between CO2 and the polymers. The possible interactions of CO2 with different chemical groups are shown in Fig. 1.2.

δ+

δ-

δ+

δ-

δ+

δ-

a) b) c)

Fig. 1.2. Weak interactions of CO2 with different chemical groups: a) ether, b) carbonyl, and c) aromatic ring Fortunately in the last decade, various experimental methods for solubility measurements at elevated temperatures and pressures have been made available in the literature. These are applied to both solid (below glass transition temperature or melting point) as well as molten (above glass transition temperature or melting point) polymers. An overview of various methods and their applications to various polymer-CO2 systems are given in Table 1.1 [13-29].


Table 1.1. Sorption studies of various supercritical CO2-polymer systems (N.R.: not reported)

Method Polymers/CO2 Press. (MPa)

Temp. (K)

Features of

equipment Ref.

Pressure decay PMMA PS

13.78 293, 473 N.R. [13]

Phase separation PEG 400 600 1000

1-15 313, 323 373 K 35 MPa

[14]

Phase separation PDMS 1-26 323, 353, 373

N.R. [15]

Pressure decay PS 1-20 373, 453 N.R. [16] Pressure decay PP, HDPE 1-17 433, 453,

473 N.R. [17]


35000

2-30 316-373 N.R. [18]


3-26 313-348 423 K 50 MPa

[19]

Phase separation PEG 200 1500 4000 8000

5-30 323-393 423 K 50 MPa

[20]

Piezoelectric-quartz sorption

PVAc, PBMA 1-10 313, 333, 353

N.R.

[21]

Chromatographic PMMA 1.5-9 236-453 N.R.

[22]

Chromatographic PDMS 1.5-10 308-393 N.R. [23] High-pressure

optical cell PDMS 13.8-

27.8 303, 323,

343 50 MPa [24]

Gravimetric PBS, PBSA 1-20 308-393 523 K 35 MPa

[25]

Gravimetric PVAc PS

1-17.5 1-20

313-373 373-473

523 K 35 MPa

[26]

Gravimetric PPO PPO/PS

1-20 373,427, 473

523 K 35 MPa

[27]

Gravimetric LDPE/TiO2

1-15 423,448, 473

N.R.

[28]

Gravimetric PPO PPO/PS

1-15 423-473

523 K 35 MPa

[29]


The modelling of solubility data is as crucial as experimental measurements for understanding of the processes. It is an inexpensive way compared to the experiments. Thermodynamics plays a vital role here. Phase equilibria of pure components or solutions are generally determined by equating the chemical potential of a component in the existing phases. The theories used for the thermodynamics are the lattice, the cubic equation of state (EOS) and the off-lattice. Among the theories above, the lattice theory has frequently been used. According to this theory, polymer molecules are ordered in a lattice structure with holes (represent free volume). The lattice fluid theory doesn’t require separate parameters to account for the flexibility of the molecule. The lattice fluid theory can be used to calculate heat and volume of mixing, lower critical solution temperature and enthalpic and entropic components of the chemical potential. The lattice fluid theory reduces to the Flory-Huggins theory at low temperatures. The EOS, a molecular lattice theory of classical fluids based on a well-defined statistical mechanical model, has been presented by Sanchez and Lacombe [30-31]. A detailed description of several versions of this model that extends the basic Flory-Huggins theory can be found in the literature [32-34]. 1.2.2. Viscosity reduction The processing of high molecular weight polymers is not an easy task. High viscosity is a major obstacle in processing a high molecular weight polymer. An option is the processing of a polymer at elevated temperatures since viscosity decreases with increasing temperature. But at elevated temperatures degradation of polymers is an important concern. In addition, it consumes a large amount of energy. The use of organic solvents can avoid this problem by reducing the viscosity of polymers at low temperatures. Nevertheless, emissions of the solvents into the environment, the separation of the solvents, and the reactive nature of the solvents are the major problems. An alternative to this option is the use of supercritical CO2 as a plasticizer or a solvent for viscosity reduction in various processing. The dissolved CO2 causes plasticization at a low temperature. This plasticization is due to the reduction in the glass transition or melting point of a polymer [35-39]. Plasticization is generally referred to the reduction in viscosity due to the dissolved gas [40]. In this way, the processing of polymers can be carried out at low temperatures and hence, the degradation of polymers can be avoided. An increased attention has been given to understand the rheological properties of various polymer-CO2 solutions. Capillary/wedge/slit die extrusion rheometers have been modified and reported for high pressure rheological measurements for various polymer-CO2 systems


by several research groups [41-48]. In addition to an extruder, a static mixer has been used to form a single phase solution. A pressure drop along the die has been used for viscosity calculations. The results show that the viscosity decreases with increasing concentration of CO2. Though the solubility of CO2 increases with pressure, the viscosity increases at high pressures due to reduction in the free volume (hydraulic pressure effect). A large pressure drop can lead to phase separation if the concentration of supercritical CO2 is above the solubility limit at the final pressure and temperature. However, such devices are mostly useful when the concentration of CO2 is far away from equilibrium. This concentration limitation can be overcome by using drag-driven devices. Since the pressure is constant during measurements, it allows carrying out measurements at or near equilibrium concentrations of CO2. The viscosity measurement using a drag-driven device, a magnetically levitated sphere rheometer (MLSR), has been reported for the low viscosity polymer [49]. The viscosity curves (shear viscosity ( η ) vs shear rate ( γ )) for various concentrations of CO2 in a molten polymer are usually of a similar shape as those of a pure molten polymer. The curves obtained at different concentrations can be shifted (scaled) to the curve obtained from a pure polymer. This curve is known as the master curve. Extensive research has been done for theoretical viscosity prediction of polymer-CO2 solutions using the free volume theory [50-51]. 1.3. Supercritical CO2 in the production of micron-size particles Milling, grinding, crystallization and spray drying are the particle formation methods commonly used in the coating, toner and drug delivery industries. Narrow particle size distribution, solvent recovery and avoiding the emissions of VOCs are the major challenges associated with these methods. In addition, milling and grinding are not suitable for thermally instable and low glass transition temperature or melting point compounds due to the frictional heat dissipated during processes. Therefore, the industries have been looking for new technologies, which would provide micron size particles with a narrow particle size distribution using as small as possible quantity of VOC. This has motivated chemical engineers as well as chemists to apply a supercritical technology rather than classical methods. The supercritical technology utilizes the solubility of supercritical CO2 in a polymer or vice versa. In the last decade, the research on particle production using supercritical CO2 has rapidly been growing. Various methods already exist that use supercritical CO2 as a solvent or anti-solvent. Recently, these methods have been broadly reviewed [52]; rapid expansion of supercritical solutions (RESS), gas anti-solvent crystallization (GAS), supercritical anti-solvent precipitation (SAS), precipitation by compressed anti-solvent (PCA), solution enhanced dispersion


by supercritical fluid (SEDS) and particles from gas saturated solutions (PGSS). To have a brief idea over the listed methods, a comparison of all of them is summarized in Table 1.2. Moreover, a brief introduction has been provided for most commonly used supercritical methods. Table 1.2. The comparison of various supercritical methods

RESS

GAS/SAS/PCA

SEDS

PGSS

Process

Discontinuous

Semicontinuous

Continuous

Continuous

Gas quantity

High

Medium

Medium

Medium

Organic solvent

Absent

Present

Present

Absent

Pressure

High

Medium

Medium

Low

Separation of

gas

Easy

Easy

Easy

Easy

1.3.1. Rapid expansion of supercritical solution

The rapid expansion of supercritical solution (RESS) method utilizes a dramatic change in the dissolving power of a solvent, when it is rapidly expanded from a supercritical pressure to a low pressure. After expansion, the solvent exist as a gas that makes the collection of the resulting particles (solute) much easier. RESS is based on crystallization or precipitation of a solute in order to facilitate the powder production. The method can generally be used if the solubility of a solute (polymer) in supercritical CO2 or another appropriate fluid is high. Fig. 1.3 represents a schematic drawing of the RESS method. A fluid is pressurized and heated to ascertain the supercritical conditions needed for the process and passed through an extractor containing a solute in order to form single phase solution. Following this, the solution is depressurized over a nozzle to atmospheric pressure. The rapid depressurisation leads to nucleation of the solute caused by the lowering of the solvation power and therefore particles are formed. After the depressurization, CO2 turns into the gas phase and is purged out of the collecting device.


P

P

CO 2

Extraction unit Expansion unit

Vent

Pump

Polymer

Fig. 1.3. A schematic drawing of the rapid expansion of supercritical solution method (RESS) 1.3.2. Supercritical anti-solvent methods Supercritical anti solvents methods are applicable to materials whose solubility in a supercritical fluid is very low. In these methods, a supercritical fluid is used as an anti-solvent. The operating principle is the same for all supercritical anti-solvent methods. A schematic drawing of the precipitation with a compressed anti-solvent (PCA) method, one of the anti-solvent methods, is shown in Fig. 1.4. A solute is first dissolved in a solvent and then, exposed to a supercritical fluid in order to generate particles. The selected solvent has a good affinity for the supercritical fluid. The solvating power of the solvent is reduced after the exposure to CO2 and the solution becomes supersaturated with the solute. Consequently, the precipitation of the solute takes place and micron size particles are formed. The nozzle through which the supercritical fluid is added is an important factor in order to control the morphology and size of the particles. At the end of the process, the precipitator is washed with the anti-solvent to remove the solvent completely.


PP

P

Pump

CO 2 Polymersolution

Pump

Precipitator

Filter

Fig. 1.4. A schematic drawing of the precipitation from a compressed antisolvent method (PCA) 1.3.3. Particles from gas saturated solution Unlike RESS, a gas is dissolved in a solute under sub- or supercritical conditions in the particles from gas saturation (PGSS) method. A schematic drawing of the PGSS method is shown in Fig. 1.5. A solution is formed by saturating a solute with a gas. The gas-saturated solution possesses a low viscosity due to an increase in free volume. Moreover, the interfacial tension between the gas and the liquid phase is lowered as the surface tension of the gas in the supercritical state is zero. These properties ease the expansion of the solution. The solution is then expanded over a nozzle from a supercritical pressure to ambient pressure. It causes a supersaturation of the gas and an intense expansion of the nucleated gas bubbles leads to explosion of the molten material into fine particles. The particles are solidified due to the cooling effect of an expanded gas. The method has been patented by Mandel et al. [53] and Weidner et al. [54]. Very few polymers have been processed using PGSS in a batch and continuous mode. In a batch mode a solution is formed using a stirrer while in a continuous mode a static mixer is used to saturate a molten polymer with a gas. The batch process has been applied for the generation of powder of poly(ethylene glycol) [55]. Conventional coating systems like acrylic coatings, polyester-epoxy systems and low-melting polyester coatings have been produced using a continuous process [56].


PP

Batch vessel Expansion unit

CO 2

Pump

Vent

Polymer+CO2

Fig. 1.5. A schematic drawing of the particles from gas saturated solution method (PGSS) 1.4. Selection of a method for the production of polymer particles Different morphologies and particle sizes have been produced using different methods for several polymers [52]. The selection of a method is dependent on several factors such as the solubility of CO2 in a polymer, solubility of the polymer in CO2, solubility of CO2 in a solvent and solubility of the polymer in a solvent and amount of solvent. Since the anti-solvent methods use organic solvents, a last preference may be given to these methods. RESS and PGSS are most preferable methods as they do not use an organic solvent. In case of RESS and PGSS, the major difference between them is the solubility term. In RESS, a polymer is dissolved in a supercritical fluid while in PGSS a supercritical fluid is dissolved in a polymer. Using CO2 as a supercritical fluid, RESS is only applicable to very few polymers due to low solubility of many polymers in supercritical CO2. PGSS is a better choice as it utilizes the solubility of supercritical CO2, which is high in several polymers. However, the method has yet not been tested for different polymers of high viscosity and may require some modifications. 1.5. Purpose and out line of the thesis This thesis mainly focuses on the production of particles using CO2 as a supercritical fluid for different polymers such as polyester resins and poly (ethylene glycol) (PEG), and aspects relevant to the process such as solubility, viscosity and processing parameters. The polyester resins and PEG in particle form are used in toner and drug delivery applications. The particle size, the


particle size distribution and the morphology are the important properties of the particles depending on the applications. In a drug delivery application, PEG particles are generally used for encapsulating a drug where a release rate of the drug is dependent on the properties of the particles. It is possible to generate particles with different properties by tuning process conditions using supercritical CO2 as a solvent or plasticizer, which allows a good control over the release of the drug. Moreover, the absence of an organic solvent in the supercritical CO2 assisted process is the major advantage in the production PEG particles containing a drug. In a toner application, polyester resins are generally used due to their low melting point or glass transition temperature and hydrophobic nature. The resins are always mixed with colour pigments in order to have a desired colour prior to the production of particles. In this application, the particle size and size distribution play a very important role in determining the texture of the film and adhesion of the film to the surface as shown schematically in Fig. 1.6. A smooth film is obtained when the particles of a small size and with a narrow particle size distribution are used. Moreover, a better adhesion of the film to the surface is also achieved.

a)

b)

Fig. 1.6. A schematic representation of the effect of the particle size on the film formed using a) big particles with a broad particle size distribution b) small particles with a narrow particle size distribution In this thesis the PGSS method has been adopted for low viscous PEG and modified for a high viscous polyester resin based on propoxylated bisphenol (PPB) in order to produce micron size particles. The PGSS method has already been reported for the production of micron size particles for a few polymers. Viscosity reduction caused by the dissolved CO2 plays a vital role in the production of particles using PGSS. The higher the dissolved amount of CO2 in a polymer the higher is the reduction in the viscosity of the polymer. For this purpose, knowledge of the solubility of CO2 in a polymer is essential a priori. The solubility of CO2 varies from one polymer to another depending on the available chemical groups. Chapter 2 discusses the relationship between the solubility of CO2 and the strength of the interactions between polymers and CO2. Fourier transform infrared spectroscopy (FT-IR) has been used as a screening tool for the selection of the polymers to be processed. As the FT-IR study gives only qualitative information, Chapter 3 investigates the solubility of CO2 in the polymers experimentally using a gravimetric method. In order to incorporate the buoyancy correction, a requisite in gravimetric measurements,


the swelling of the polymers using an optical cell has been measured separately. Chapter 3 also describes the modelling of phase equilibrium data using Sanchez-Lacombe equation of state, which can be used for interpolating and extrapolating the solubility data at different processing conditions. Chapter 4 describes a batch PGSS process for the production of particles from PEG of different molecular weights. In this chapter, flow and solidification models, and the effect of various parameters such as temperature, pressure, nozzle diameter and molecular weight have been studied in detail. Like PEG, shear viscosity data in the presence of dissolved CO2 for PPB is not available in the literature. In Chapter 5, a model based on the free volume theory has been applied to predict the reduction in the viscosity of PPB which appeared to be considerable. The PGSS method has been tested for the particles production of PPB. It has not been possible to produce particle of PPB using PGSS mainly due to its very high viscosity compared to PEG. The method has been modified in terms of the amount of CO2 used. Chapter 5 also discusses the results obtained from scouting particles production experiments, which have been carried out using a continuous set up in which CO2 and polymer have been mixed together in a Kenics type static mixer. It is possible to produce particles from the high viscous polymer melt by modifying the PGSS method. To overcome practical limitations of the set up, a gear pump and a SMX static mixer for higher through put and a better mixing, respectively, have been introduced. Chapter 6 describes a detailed engineering study of the continuous production of polymer particles using the modified set up for various processing conditions. The effect of various parameters such as temperature, pressure, gas to polymer mass ratio, core-slot width and nozzle diameter on the particle size, shape and particle size distribution has been studied. Chapter 6 also discusses the solidification model and the dimensionless analysis model to determine the contribution of various parameters. Finally, a technological assessment by comparing the modified method with the traditional methods, prospects of supercritical CO2, and conclusions with future outlook of the thesis work have been discussed in Chapter 7. References [1] Cooper A I. Synthesis and processing of polymers using supercritical carbon

dioxide. Journal Materials Chemistry 2000;10:207-234. [2] Tomasko D L, Li H, Liu, Han X, Wingert M J, Lee L J, Koelling K W. A review

of CO2 applications in the processing of polymers. Industrial Engineering Chemistry and Research 2003;42:6431-6456.

[3] Kendall J L, Canelas D A, Young J L, De Simone J M. Polymerizations in Supercritical Carbon Dioxide. Chemical reviews 1999;99:543-564.

[4] Mchugh M A, Krukonis V J. Supercritical fluid extraction. Boston: Butterworth-Heinemann, 1994.

[5] Hyatt J A. Liquid and Supercritical Carbon Dioxide as Organic Solvents. The Journal of Organic Chemistry 1984;49:5097-5101.


[6] De Simone J M, Guan Z. Synthesis of fluoropolymers in supercritical carbon dioxide. Science 1992;257:945-947.

[7] Yilgor I, McGrath J E, Krukonis V J. Novel supercritical techniques for polymer fractionation and purification. 2. Fractionation and characterization of functional siloxane oligomers. Polymer bulletin 1984;12:499-506.

[8] Krukonis V J. Processing of polymers with supercritical fluids. Polymer news 1985;11:7-16.

[9] Guan Z, Combs J R, Menceloglu Y Z, DeSimone J M. Homogeneous free radical polymerizations in supercritical carbon dioxide: 2. Thermal decomposition of 2,2'-Azobis(isobutyronitri1e). Macromolecules 1993;26:2663-2669.

[10] Hoefling T A, Newman D A, Enick R M, Beckman E J. Effect of structure on the cloud-point curves of silicone-based amphiphiles in supercritical carbon dioxide. The Journal of Supercritical Fluids 1993;6:165-171.

[11] Kazarian G S, Vincent M F, Bright F V, Liotta C V, Eckert A C. Specific intermolecular interaction of carbon dioxide with polymers. The Journal of American Chemical Society 1996;118:1729-1769.

[12] Kazarian G S, Brantley N H, West B L, Vincent M F, Eckert C A. In situ spectroscopy of polymers subjected to supercritical CO2: plasticization and dye Impregnation. Applied Spectroscopy 1997;51:491-494.

[13] Elkovitch M D, Lee L J, Tomasko D L. Supercritical carbon dioxide assisted blending of polystyrene and poly(methyl methacrylate). Polymer Engineering and Science 1999;39:2075-2084.

[14] Daneshvar M, Kim S, Gulari E. High-pressure phase equilibria of poly(ethylene glycol)-carbon dioxide. The Journal of Physical Chemistry 1990;94:2124-2128.

[15] Garg A, Gulari E, Manke C W. Thermodynamics of polymer melts swollen with supercritical gases. Macromolecules 1994;27:5643-5653.

[16] Sato Y, Yurugi M, Fujiwara K, Takishima S, Masuoka H. Solubilities of carbon dioxide and nitrogen in polystyrene under high temperature and pressure. Fluid Phase Equilibria 1996;125:129-138.

[17] Sato Y, Fujiwara K, Takikawa T, Sumarno, Takishima S, Masuoka H. Solubilities and diffusion coefficients of carbon dioxide and nitrogen in polypropylene, high-density polyethylene, and polystyrene under high pressures and temperatures. Fluid Phase Equilibria 1999;162:261-276.

[18] Weidner E, Wiesmet V, Knez Z, Skerget M. Phase equilibrium (solid-liquid-gas) in polyethyleneglycol-carbon dioxide systems. The Journal of Supercritical Fluids 1997;10:139-147.

[19] Gourgouillon D,.Ponte M N d. High pressure phase equilibria for poly(ethylene glycol)s+CO2: experimental results and modelling. Phys. Chem. Chem. Phys. 1999;1:5369-5375.

[20] Wiesmet V, Weidner E, Behme S, Sadowski G, Arlt W. Measurement and modelling of high-pressure phase equilibria in the systems polyethyleneglycol (PEG)-propane, PEG-nitrogen and PEG-carbon dioxide. The Journal of Supercritical Fluids 2000;17:1-12.

[21] Wang N-H, Takishima S, Masuoka H. Measurement and correlation of solubility of a high pressure gas in a polymer by piezoelectric quartz sorption-CO2+PVAc and CO2+PBME systems. International Chemical Engineering 1994;34:255-262.

[22] Edwards R R, Tao Y, Xu S, Wells P S, Yun K S, Parcher J F. Chromatographic investigation of CO2-polymer interactions at near-critical conditions. The Journal of Physical Chemistry B 1998;102:1287-1295.


[23] Xu S, Wells P S, Tao Y, Yun K S, Parcher J F. Stationary-and mobile-phase interactions in supercritical fluid chromatography. ACS Symp Ser 2000;748:96-118.

[24] Royer J R, DeSimone J M, Khan S A. Carbon dioxide-induced swelling of poly(dimethylsiloxane). Macromolecules 1999;32:8965-8973.

[25] Sato Y, Takikawa T, Sorakubo A, Takishima S, Masuoka H, Imaizumi M. Solubility and diffusion coefficient of carbon dioxide in biodegradable polymers. Industrial Engineering Chemistry and Research 2000;39:4813-4819.

[26] Sato Y, Takikawa T, Takishima S, Masuoka H. Solubilities and diffusion coefficients of carbon dioxide in poly(vinyl acetate) and polystyrene. The Journal of Supercritical Fluids 2001;19:187-198.

[27] Sato Y, Takikawa T, Yamane M, Takishima S, Masuoka H. Solubility of carbon dioxide in PPO and PPO/PS blends. Fluid Phase Equilibria 2002;194-197:847-858.

[28] Areerat S, Hayata Y, Katsumoto R, Kegasawa T, Egami H, Ohshima M. Solubility of carbon dioxide in polyethylene/titanium dioxide composite under high pressure and temperature. J. App. Poly. Sci. 2002; 86:282-288.

[29] Areerat S, Hayata Y, Katsumoto R, Kegasawa T, Egami H, Ohshima M. Measurement and prediction of diffusion coefficient of supercritical CO2 in molten polymer. Poly. Eng. Sci. 2004; 44:1915-1924.

[30] Sanchez I C, Lacombe R H. An elementary molecular theory of classical fluids. pure fluids. The Journal of Physical Chemistry 1976;80:2352-2362.

[31] Sanchez I C, Lacombe R H. Statistical thermodynamics of polymer solutions. Macromolecules 1978;11:1145-1156.

[32] Sanchez I C, Rodgers P A. Solubility of gases in polymers. Pure Applied Chemistry 1990;62:2107-2114.

[33] Panayiotou C, Sanchez I C. Statistical thermodynamics of associated polymer solutions. Macromolecules 1991;24:6231-6237.

[34] Sanchez I C. Relationship between polymer interaction parameters. Polymer 1989;30:471-475.

[35] Wissinger R G, Paulaitis M E. Glass transitions in polymer/CO2 mixtures at elevated pressures. Journal of Polymer Science: Part B: Polymer Physics 1991;29:631-633.

[36] Condo P D, Sanchez I C, Panayiotou C G, Johnston K P. Glass transition behaviour including retrograde vitrification of polymers with compressed fluid diluents. Macromolecules 1992;25: 6119-6127.

[37] Goel S K, Beckman E J. Plasticization of poly(methyl methacrylate) (PMMA) networks by supercritical carbon dioxide. Polymer 1993;34:1410-1417.

[38] Condo P D, Paul D R, Johnston K P. Glass transitions of polymers with compressed fluid diluents: type II and III behaviour. Macromolecules 1994;27:365-371.

[39] Zhang Z,.Handa Y P. An in situ study of plasticization of polymers by high-pressure gases. Journal of Polymer Science: Part B: Polymer Physics 1998;36:977-982.

[40] Gerhardt L J, Gulari E, Manke C W. Rheology of polydimethylsiloxane swollen with supercritical carbon dioxide. Journal of Polymer Science: Part B: Polymer Physics 1997;35:523-534.

[41] Kwag C, Manke C W, Gulari E. Rheology of molten polystyrene with dissolved supercritical and near-critical gases. Journal of Polymer Science: Part B: Polymer Physics 1999;37:2771-2781.


[42] Lee M, Park C B, Tzoganakis C. On-line measurements of PS/CO2 viscosities. SPE ANTEC 1997;231:1991-1995.

[43] Lee M, Park C B, Tzoganakis C. Measurements and modeling of PS/supercritical CO2 solution viscosities. Polymer Engineering and Science 1999;39:99-109.

[44] Lee M, Park C B, Tzoganakis C. Extrusion of PE/PS blends with supercritical carbon dioxide. Polymer Engineering and Science 1998;38:1112-1120.

[45] Lee M, Tzoganakis C, Park C B. Effects of supercritical CO2 on the viscosity and morphology of polymer blends . Advances in Polymer Technology 2000;19:300-311.

[46] Royer J R, Gay Y J, DeSimone J M, Khan S A. High-pressure rheology of polystyrene melts plasticized with CO2: Experimental measurement and predictive scaling relationship. Journal of Polymer Science: Part B: Polymer Physics 2000;38:3168-3180.

[47] Royer J R, Gay Y J, DeSimone J M, Khan S A. High-pressure rheology and viscoelastic scaling predictions of polymer melts containing liquid and supercritical carbon dioxide. Journal of Polymer Science: Part B: Polymer Physics 2001;39:3055-3066.

[48] Lan H-Y, Tseng H-C. Study on the rheological behaviour of PP/supercritical CO2 mixture. Journal of Polymer Research 2002;9:157-162.

[49] Royer J R, Gay Y J, Adam M, De Simone J M, Khan S A. Polymer melt rheology with high-pressure CO2 using a novel magnetically levitated sphere rheometer. Polymer 2002;43:2375-2383.

[50] Gerhardt L J, Garg A, Manke C W, Gulari E. Concentration-dependent viscoelastic scaling models for polydimethysiloxane melts with dissolved carbon dioxide. Journal of Polymer Science: Part B: Polymer Physics 1998;36:1911-1918.

[51] Kwag C, Manke C W, Gulari E. Effects of dissolved gas on viscoleastic scaling and glass transition temperature of polystyrene melts. Industrial Engineering Chemistry and Research 2001;40:3048-3052.

[52] Jung J, Perrut M. Particle design using supercritical fluids: Literature and patent survey. The Journal of Supercritical Fluids 2001;20:179-219.

[53] Mandel F S, Green C D, Scheibelhoffer A S. Method of preparing coating materials. (Ferro Corporation, Cleveland, USA). US:5548004, 1996

[54] Weidner E, Knez Z, Novak Z. Process for the production of particles or powders. (USA). US:6056791, 2000

[55] Weidner E, Knez Z, Steiner R. Powder generation from polyethyleneglycols with compressible fluids. High Pressure Chemical Engineering, proceedings of process technology 1996;223-228.

[56] Weidner E, Petermann M, Blatter K, Rekowski V. Manufacture of Powder Coatings by Spraying of Gas-Enriched Melts. Chemical Engineering and Technology 2001;24:529-533.

Chapter 2: FT-IR studies on the interactions of CO2 and polymers having different chain groups 17

Chapter 2

FT-IR studies on the interactions of CO2 and polymers having

different chain groups A Fourier transform-infrared spectroscopy (FT-IR) set up has been successfully modified in order to characterize different polymeric materials under sub- and supercritical CO2 conditions. Polymers used in this study are polyesters (PPB and PEB), poly (ethylene glycol) (PEG) and polyphenylene oxide (PPO). Analysis of the corresponding spectra shows evidences of weak interaction (Lewis acid-base) between CO2 and polymers. In particular, shifts to higher wavelengths of the chain groups of the polymer and the modification of the absorption band of CO2 represent qualitative evidence of such interactions. Analysis of CO2 absorption bands allowed ranking of the polymeric materials. Polymers with ether group display high interaction strength, in terms of width of the band, than polyesters. The effect of the dissolved CO2 on the depression of the melting point or the glass transition temperature can also be studied by using FT-IR depending on the enhancement in the free volume. The spectrum of PEG, unlike the other polymers, is completely modified above the critical pressure (7.38 MPa). Sameer P. Nalawade, F. Picchioni, Jan H. Marsman, L. P. B. M. Janssen, FT-IR studies on the interactions of CO2 and polymers having different chain groups, Journal of supercritical Fluids, in press.


2.1. Introduction Nowadays, CO2 has become a potential solvent for various polymer applications like polymer processing and polymer synthesis [1-3]. An inert nature and a low cost are the most attractive advantages of CO2 as a solvent. The dissolved CO2 in a polymer reduces the viscosity and allows the process or the synthesis to be carried out at a lower temperature. It makes the process or the synthesis less energy consumptive. In the applications, the solubility of CO2 in a polymer is the vital parameter for its selection as a solvent. The higher the solubility of CO2 in a polymer the higher is the reduction in the viscosity [3]. The solubility represents the amount of the CO2 that can be dissolved at equilibrium conditions. The solubility measurement data of various polymer-CO2 systems are well described in the literature [3]. The available data have mostly been interpreted thermodynamically in terms of the effect of temperature and pressure on the solubility of CO2 in a polymer. Earlier, it was a general notion that the CO2 solubility in a polymer is a function of pressure and temperature only. In fact, the amount of dissolved CO2 varies from one polymer to another even at the same conditions. The specific intermolecular interactions between CO2 and polymer are responsible for the difference in solubility. Fourier transmission infrared spectroscopy (FT-IR) is a good tool to study such interactions [4-11]. Albeit CO2 lacks a dipole moment, it has a large quadrupole moment, and both a Lewis acid and base site. In CO2, the Lewis acidity results from the electropositivity of the carbon atom due to deficiency of electron density compared to the oxygen atoms. Thus, an electron acceptor-donor interaction is present when CO2 is contacted with a polymer containing Lewis base site. Few authors have revealed the specific intermolecular interactions between CO2 and different polymers by carrying out IR spectroscopy studies. The FT-IR spectra of cellulose acetate (CA) and poly (methyl methacrylate) (PMMA) in the presence and the absence of CO2 have been recorded by Fried and Li [4]. In this study, the carbonyl stretching vibrations for CA and PMMA were shifted to higher wavenumbers in the presence of CO2. Since the observed shifts were rather small, dipole-dipole interactions between the CO2 and the carbonyl groups were suggested instead of the Lewis acid-base interactions. Later on, Kazarian et al. [6-7] have also used FT-IR to study the intermolecular interactions between CO2 and polymers. They used the bending mode ( 2ν , around 660 cm-1) of CO2 to show the interactions with the polymers. In case of polymers containing carbonyl group, e.g. PMMA, splitting of the 2ν was observed. The splitting is due to the Lewis acid-base interaction where the carbonyl group act as an electron donor and CO2 act as an electron acceptor. No such splitting has been reported for the polyethylene (PE) and polystyrene (PS). The reason is the absence of carbonyl group in PS as well as PE. However, some


distortion of the 2ν was observed only in PS. This is due to the interaction between CO2 and the phenyl ring in PS. Meredith et al [8] have also used FT-IR spectroscopy to test the intermolecular interaction of CO2 with different Lewis bases: triethylamine (TEA), pyridine (PYR), and tributyl phosphate (TBP). In case of PYR, a base having phenyl ring, no splitting was observed similar to PS and PE. Recently, both the anti symmetric stretching mode ( 3ν ) and the 2ν of CO2 have been used to show its intermolecular interaction with PMMA, PS or polycarbonate (PC) [12]. The recorded spectra for the 2ν of CO2 in case of PMMA and PS has shown similar results, as explained earlier. In case of the 3ν of CO2 for PMMA and PC, an absorption band having a weak shoulder appeared in the spectra. The width of the absorption band was larger in PC compared to PMMA. Both the carbonyl groups as well as the benzene rings present in PC provide more sites for CO2 molecules as compared to PMMA. In PS the recorded absorption band was not comparable, in terms of width, to those observed for PMMA and PC. In PS, the interactions are present only between CO2 and the aromatic rings. These interactions are weaker than the ones present between CO2 and carbonyl groups in PMMA and PC. In this study, we have modified a golden gate FT-IR set up, generally used at ambient pressure, to a high pressure FT-IR setup. The main aim of the study is to reveal the intermolecular interactions between CO2 and polymers having different chemical groups along the main chain. The chemical groups which are studied here include ester, ether, and aromatic rings. The CO2 bending mode has been studied for all the polymers to reveal the interactions. 2.2. Experimental 2.2.1. Materials Polyester based on propoxylated bisphenol (PPB) and based on ethoxylated bisphenol (PEB) (CAS:177834-94-5 and 170831-75-1) having a weight average molecular weight (Mw) of 7000 and 20000, respectively, were supplied by Akzo Nobel, The Netherlands. Chloroform (CHCl3) and poly (ethylene glycol) (PEG) having a Mw of 6000 were purchased from Aldrich, The Netherlands. Polyphenylene oxide (PPO) of a Mw of 20000 was supplied by GE Corporations, The Netherlands. Glass transition temperatures (Tg)/melting point (Tm) of the polymers are provided in Table 2.1. High purity CO2, 99.99 %, was used in the FT-IR experiments. The materials were used without any further purification.


Table 2.1. Glass transition temperature (Tg) or melting point (Tm) of different polymers obtained experimentally

Polymers Tg or Tm (K)

PPB 327 PEB 330 PEG 331 PPO 503

2.2.2. Apparatus For spectroscopic measurements in the presence of CO2, the existing FT-IR golden gate apparatus (Perkin Elmer, England ) was modified to a high pressure FT-IR set up. The apparatus is generally used to record spectra of a solid or molten material, which is usually provided in the form of a film or powder. Due to better uniformity, a film rather than powder is preferred. The film is kept on the surface of the crystal (Quartz) mounted in the heating plate in conjunction with a temperature controller. The temperature of the material is raised using a heating plate in order to record the spectra of a molten or solid material. The maximum temperature to which crystal can be heated is 473 K. For the better contact of the solid film with the crystal, a groove-bolt arrangement is provided to apply a certain force on the film. The CO2 absorbance becomes too high at higher pressures if a crystal of long path length is selected. Therefore, ATR (attenuated total reflection) IR spectroscopy with FT-IR transmission was used in the study. ATR spectroscopy provides the path length of only several micrometers [13]. A schematic drawing of the high pressure FT-IR unit is shown in Fig. 2.1. The high pressure unit consists of a golden gate FT-IR apparatus, a CO2 cylinder, a high pressure syringe pump (Isco, USA), a high pressure cell, and valves. The high pressure cell, tubes and valves are made of stainless steel. The high pressure cell was built up at the University of Groningen (The Netherlands). The same groove-bolt arrangement was used to ensure the good contact of an O-ring with the cell as well as the surface. The O-ring between the cell and the surface of the plate makes the system leak-proof. Using the groove-bolt arrangement the force is transferred from the cell to the O-ring by tightening the bolt. The cell can withstand pressures up to 10 MPa. CO2 inlet and exit lines are provided along with the needle valves. The cell, inlet, exit lines and valves are completely insulated with a glass wool to prevent heat losses to the surrounding. The heating source is sufficient to raise the temperature of a polymer film as well as gas to a desired value as the internal volume of the cell is very small. To measure a pressure value, a bourdon pressure gauge meter is connected in between the CO2 inlet line and the cell. The pressure gauge meter (Swagelok, Germany) can measure a pressure in bar (Pmax= 10 MPa). The spectra are


measured on the Spectrum 2000 spectrometer. The scanning range available in the FT-IR set up is 500-4000 cm-1.

P FP

T

CO 2

Syringe Pump

IR beam

Groove-bolt

ATR crystal

To detector

Cell

Fig. 2.1. A high pressure FT-IR golden gate unit 2.2.3. Film preparations The polymers were used in the form of a thin film. Different procedures were adopted to prepare thin films of polymers. Very small amounts of polymers were used for the film preparation. The use of organic solvent in the production of films was avoided whenever possible. In case of polyester resins and PEG, the films were prepared without using a solvent. The films were prepared directly on the crystal by melting the polyester resins and PEG. To obtain the molten polymer, the temperature of the crystal was raised above the melting point (Tm) or glass transition temperature (Tg). The molten polymer was then smeared as a very thin film using a thin glass slit. The film was then pressed against the crystal using the groove-bolt arrangement having a flat glass surface attached at the bottom of the nut. It ensures a uniform surface and a good contact of the film with the crystal. The thin film adheres to the surface of the crystal as it cools down. Since the Tg of PPO is very high, the method above was not suitable. A solvent cast method was used for PPO. A dilute solution of PPO and CHCl3 was cast on a glass plate and kept in a vacuum oven at 318 K for three days to obtain a dried film. As PPO cannot be molten at a low temperature, the film was the pressed against the crystal using the flat surface attached at the bottom of the nut.


2.2.4. Experimental method The cell and the polymer film were heated to the desired temperature. The spectra were recorded for the polymer film in an absence of CO2. The cell was then flushed with CO2 for a few minutes using a syringe pump. After this, the desired pressure in the system was achieved by adjusting the flow rate in the pump and closing the exit valve. The inlet valve was closed as soon as the desired pressure in the cell was achieved. The system was kept under an isobaric and isothermal condition for at least 3 hrs. After recording the spectrum, the pressure was increased to a higher value. Sufficient time (at least 1 hr) was provided for the CO2 dissolution before recording the new spectrum. The spectra were scanned for the complete range with a resolution of 2 cm-1. The average number of scans was 200. The spectra were recorded on a computer using the software, Spectrum for Windows, provided along with the FT-IR set up. 2.3. Results and discussion

2.3.1. PPO The FT-IR experiments for PPO have been carried out at 313 K. The spectra were recorded in the absence of CO2 and at different CO2 pressures (3-8 MPa) under isothermal conditions. In PPO, the CO2 interactions are possible with the ether group as well as the aromatic ring. The effect of the dissolved CO2 on the stretching vibrations of C-O group (sigma bond, ~1180 cm-1) and the aromatic out-of plane bending vibrations of C-H (~855 cm-1) [14] can be seen in Fig. 2.2 and Fig. 2.3, respectively.

Fig. 2.2. The spectra of the stretching vibrations of C-O group (sigma bond) of PPO at 313 K

750 770 790 810 830 850 870 890 910 930 950

wavenumber, cm-1

Only PPO, 854.8 cm-1

3 MPa, 855.4 cm-1

6 MPa bar, 855.7 cm-1

8 MPa, 855.8 cm-1


1100 1120 1140 1160 1180 1200 1220 1240

wavenumber, cm-1

Only PPO, 1180.2 cm-1

6 MPa, 1181.3 cm-1

3 MPa, 1181.1 cm-1

8 MPa bar, 1181.3 cm-1

Fig. 2.3. The spectra of the out of plane bending vibrations of C-H group (aromatic ring) of PPO at 313 K From Fig. 2.2 and 2.3, it is clear that the C-O (sigma bond) stretching vibrations and aromatic C-H bending vibrations were shifted slightly to higher wavenumbers in presence of CO2. The wavenumber was shifted to higher values with the increasing CO2 pressures in case of the ether as well as aromatic group. The shifts are attributed to the complexes formed between CO2 and the above mentioned groups. Actually, the evidence of interaction between CO2 and the polymer can be visualized also in terms of the bending vibration of CO2 as shown in Fig. 2.4. An extra band around 655 cm-1 was observed for PPO in the presence of CO2 compared to the single-band observed for CO2 only at 667 cm-

1. This is in a full agreement with what reported earlier in the literature for polymers containing similar groups [7-8].


600 620 640 660 680 700 720

wavenumber, cm-1

Only PPO

Only CO2, 6 MPa

3 MPa, PPO+CO2

6 MPa, PPO+CO2

8 MPa, PPO+CO2

Fig. 2.4. The spectra of the bending vibrations of CO2 entrapped into PPO at 313 K 2.3.2. PPB and PEB In case of aromatic polyesters, the chemical groups that most probably interact with the CO2 are the carboxyl one and the aromatic ring as before. Fig. 2.5 and Fig. 2.6 present the absorption profiles relative to the stretching vibration of the carboxyl group (C=O, around 1720 cm-1 [15]) in PPB and PEB for various CO2 pressures. The wave numbers were shifted to higher values with maximum shifts of 3.7 (PPB) and 2 cm-1 (PEB). The shifts, in case of C=O, are in agreement with the strength of the interaction predictable on the basis of the chemical structure and with the data reported in the literature in the comparison of PMMA (4 cm-1) [4,7]. While, in the region of the aromatic absorption (730 cm-1 [15]) relatively smaller shifts of 1.2 cm-1 (maximum) are observed for both materials (only PPB is reported in Fig. 2.7). Also in this case changes in the CO2 bending vibration (Fig. 2.8 for PPB) can be observed along with the free CO2 at 667 cm-1. It indicates the presence of interactions of CO2 with base sites and the phenyl rings available in PPB.


1620 1670 1720 1770 1820

wavenumber, cm-1

Only P120, 1716.8 cm-1

3 MPa, 1717.0 cm-1

6 MPa, 1720.3 cm-1

8 MPa, 1720.5 cm-1

Fig. 2.5. The spectra of stretching vibrations of C=O group in PPB at 313 K

1650 1670 1690 1710 1730 1750 1770 1790

wavenumber, cm-1

Only P130, 1717.9 cm-1

3 MPa, 1719.1 cm-1

6 MPa, 1720.1 cm-1 8 MPa, 1719.9 cm-1

Fig. 2.6. The spectra of the stretching vibrations of C=O group in PEB at 313 K


710 715 720 725 730 735 740 745 750

wavenumber, cm-1

Only P120, 728.3 cm-1

3 MPa, 728.8 cm-1

6 MPa, 729.2 cm-1

8 MPa, 729.5 cm-1

Fig. 2.7. The spectra of the out of plane bending vibrations of in the aromatic region of PPB at 313 K

600 620 640 660 680 700

wavenumber, cm-1

8 MPa, P120+CO2

Only P120

3 MPa, P120+CO2

6 MPa, P120+CO2

Fig. 2.8. The spectra of the bending vibrations of CO2 entrapped into PPB at 313 K 2.3.3. PEG In PEG, the ether (C-O) group is able to interact with CO2. When CO2 is sorbed into PEG, a shift in the C-O stretching vibration [16-17] is observed, Fig. 2.9. Albeit the shifts are smaller at sub-critical pressures, a complete modification in the spectrum is observed at a supercritical pressure. This effect can be related to an increase in the free volume in the presence of CO2. For PEG at 8 MPa the spectrum is representing the absorption bands of the polymer much closer to its melting point (Tm). This is in agreement with the thermal properties of PEG reported in the literature and with the Tm reduction of approximately 2 K/MPa (starting Tm of about 331 K) [18]. While in case of PPO, which also contains C-O group, the starting value of the glass transition temperature (Tg) (503 K) is too


high to observe such effect on the spectrum (reduction of Tg for PPO is reported to be 5 K/MPa [3]).

1000 1050 1100 1150 1200 1250 1300

wavenumber, cm-1

Only PEG, 1103.9 cm-1

3 MPa, 1104.0 cm-1

6 MPa, 1104.5 cm-1

8 MPa, 1095.5 cm-1

Fig. 2.9. C-O stretching vibration in PEG at different CO2 pressures Indeed, also the bending vibration of CO2 entrapped in PEG (Fig. 2.10) displays a completely different spectrum with a very broad absorption at 8 MPa compared to what observed for the other polymers (Fig. 2.4 and 2.8). At 3 MPa, a relatively small, but broad, peak was observed. At 6 MPa, it was possible to observe the distinguished splitting of a band around 660 cm-1 due to weak interactions between CO2 and the ether group. Such splitting has already been reported in case of polymers not containing aromatic rings [7].

600 620 640 660 680 700 720

wavenumber, cm-1

Only PEG

3 MPa, PEG + CO2

6 MPa, PEG + CO2

8 MPa, PEG + CO2

Fig. 2.10. The spectra of the bending vibrations of CO2 entrapped in PEG


2.3.4. Discussion Possible factors that can influence the shift of chemical groups present in the polymer as a function of CO2 pressure are: kind of chemical group, steric hindrance at the interaction site and an accessible free volume. The polymers studied in this work have in common the presence of chemical groups with the same structure. Upon interaction with CO2, the IR absorption of every single group shifts to a higher wavelength, as already reported for Lewis acid-base kind of interaction. To our knowledge, no attempt has been described in the literature to correlate the observed shifts with the type of polymer used. We have also tried to find a correlation by taking into account the most similar materials, i.e. PPB and PEB. The shifts of the C=O groups in PPB and PEB with increasing CO2 pressure are shown in Fig. 2.5 and 2.6. It is observed that the shifts for PPB are slightly higher than for PEB. Since these two materials have the same chemical group (here shift of the carboxyl is considered) and nearly similar thermal properties, we thought at first instance that the different behaviour could be accounted for by the difference in the steric hindrance close to the carboxyl group. A schematic representation of the steric hindrance in the two cases is shown in Fig. 2.11. As can be seen, for PPB, the presence of the methyl group in α position with respect to the C-O bond, dramatically increases the steric hindrance, thus making in principle the Lewis acid-base interaction with CO2 much more difficult. In the case of PEB, the methyl group is substituted with a hydrogen which of course displays much less hindrance for the interaction with CO2. Obviously, since the observed shifts are actually higher for PPB (more hindrance) than PEB (less hindrance), other factor may play a more determining role. In this case, since the only other difference is the one regarding molecular weight (higher for PEB than for PPB), we hypothesize that the difference in an accessible free volume in the polymers may be more important than the one caused by the steric hindrance. In polymers, effect of the molecular weight on the free volume is generally observed in terms of the viscosity. The higher the molecular weight the higher is the viscosity i.e. the lower is the accessible free volume. Thus, it is more difficult for CO2 to access the C=O groups in PEB. Shah et al. [19] have found that the accessible free volume of a polymer has a greater effect on the solubility than the CO2-polymer interactions. However, individual contributions of these different factors to the observed IR shifts require in our opinion a detailed theoretical study.


a) b)

Fig. 2.11. Schematic representation of the steric hindrance for interaction of CO2 with carboxyl groups in a) PPB and b)PEB (Gray: C, white: H, and Black: O) The situation becomes even more complicated if one compares, for example, the shifts relative to the bending vibrations in the aromatic region, where no significant shifts could be observed for PPB, PEB and PPO (all of these contain aromatic rings in the backbone). In this case, indeed, the shift differences between the three polymers are less than 0.3 cm-1. The latter value is in our opinion too small to allow drawing conclusions on the reported behaviour. These data show that the same chemical group, if present in polymers with different properties (molecular weight, thermal behaviour, and steric hindrance) result in different FT-IR shifts upon interaction with CO2. In order to define a clear trend, theoretical simulation should be carried out to calculate ab initio the strength of the interaction. Since the shifts obtained with the chemical groups of the polymers are small, we have tried to determine the interaction strength by analysing the bending vibrations of the CO2 entrapped into the polymer. A nice correlation, as described in the literature [7, 8] for other polymers, can be found considering the bending mode of CO2. A width of the CO2 band in the bending mode is used to estimate the strength of the interaction. A splitting of the band in the 2ν mode has been observed for PEG. In case of PMMA such splitting, as it does not contain a phenyl ring, has been reported [7]. For the absorption band around 660 cm-1, it is possible to deconvolute (see Fig. 2.12 for PEG as example) the spectrum in different contributions: one is located at about 650 cm-1 and the other is at about 660 cm-1. The bands around 650 and 660 cm-1 are actually assigned to the bending vibration of CO2 (in-plane bending and out-of-plane modes, respectively), which form a doublet. Despite the lack of an evident splitting in the CO2 band for the other polymers, it is possible to deconvolute their spectra by estimating the two peaks around 650 cm-1 and 660 cm-1 as reported by Meredith et al. [8]. However, only the width at half-maximum of the 2ν band (Table 2.2), without splitting, are considered. The


widths are reported for 6 MPa, since the considerable shifts are observed in the chemical groups of the polymers at this pressure. If one compares different materials (by comparing again the band width), it is then possible to rank them according to their interaction strength with CO2:

PEG ≈ PPO >> PPB ≈ PEB

The ranking above is clearly reflective of the chemical structure: ether groups (PEG and PPO) interact stronger than carboxyl one (PPB and PEB). One would like to compare this ranking with the solubility of CO2 into the polymer, expecting a higher solubility according to the ranking. Beckman and co-workers have reported strong CO2 interactions for polymers with ether functionalities [20].

0

0.02

0.04

0.06

0.08

0.1

0.12

600 620 640 660 680 700

wavenumber, cm-1

A

Fig. 2.12. The deconvolution of CO2 absorption band in PEG Table 2.2. Width of the band in the CO2 bending mode in polymers

Polymer Pressure

(MPa) Width (cm-1)

Wavenumber (cm-1)

PPO 6 8.2 658.8

PPB 6 6.0 653.9

PEB 6 6.1 653.2

PEG 6 8.6 652.1

The effect of CO2 solubility on the Tm or Tg depression could also be observed in the FT-IR studies depending on an increment in the free volume of the polymer. As an example one may compare the spectra of PEG and PPO at 8 MPa (Fig. 2.2 and 2.9). The spectra of PPO is basically the same as the one at 3 and 6


MPa, i.e. no significant differences are observed in the transition from sub-critical to supercritical CO2. On the contrary, the shape of the PEG spectra is changed completely because of the larger enhancement in the free volume. In addition, despite of a similar glass transition temperature such results were not observed with PPB and PEB. It can also be related to the smaller free volume available in PPB and PEB.

2.4. Conclusions FT-IR equipment can be conveniently modified to work under elevated CO2 pressures. Shifts of FT-IR absorption bands of the polymer are indicative of interaction between the polymers and CO2 but only on a qualitative level. This has been confirmed by the absorption bands of CO2 itself. Analysis of the band width allows ranking the polymers in terms of their interaction strength with CO2. The interaction strength is higher in the polymers containing ether group than the polymers containing ester group. A correlation between FT-IR spectra and solubility of CO2 is not possible just on the basis of the shifts for the chain groups of polymers containing a similar chemical structure. Free volume effects must be taken into account. Theoretical studies are needed to properly distinguish the different effects and possibly obtain a more quantitative correlation between FT-IR and solubility data. Shifts are dependent on structure of chemical group, thermal property of the polymer (Tg or Tm) and in general free volume effects. References [1] Cooper A I. Synthesis and processing of polymers using supercritical carbon

dioxide. J. Mater. Chem. 2000;10:207-234. [2] Kendall J L, Canelas D A, Young J L, DeSimone J M. Polymerizations in

supercritical carbon dioxide. Chem. Rev. 1999;99:543-564. [3] Tomasko D L, Li H, Liu D, Han X, Wingert M J, Lee L J, Koelling K W. A review

of CO2 applications in the processing of polymers. Ind. Eng. Chem. Res. 2003; 42:6431-6456.

[4] Fried J R, Li W J. High-pressure FT-IR studies of gas-polymer interactions. J. App. Poly. Sci. 1990;41:1123-1131.

[5] Briscoe B J, Kelly C T. Optical studies of polymers in high pressure gas environments. Mat. Sci. Eng. 1993;A168:111-115.

[6] Kazarian S G, Vincent M F, Eckert C A. Infrared cell for supercritical fluid-polymer interactions. Rev. Sci. Instrum. 67 (1996) 1586-189.

[7] Kazarian S G, Vincent M F, Bright F V, Liotta C L, Eckert C A. Specific intermolecular interaction of carbon dioxide with polymers. J. Am. Chem. Soc. 1996;118:1729-1769.

[8] Meredith J C, Johnston K P, Seminario J M, Kazarian S G, Eckert C A. Quantitative equilibrium constants between CO2 and Lewis bases from FTIR spectroscopy, J. Phys. Chem. 1996;26:10837-10848.


[9] Flichy N M B, Kazarian S G, Lawrence C J, Briscoe B J. An ATR-IR study of poly(dimethylsiloxane) under high-pressure carbon dioxide: simultaneous measurement of sorption and swelling. J. Phys. Chem. B 2002;106:754-759.

[10] Kazarian S G. Polymers and supercritical fluids: Opportunities for vibrational spectroscopy. Macromol. Symp. 2002;184:215-228.

[11] Kazarian S G, Andrew Chan K L. FTIR imaging of polymeric materials under high-pressure carbon dioxide. Macromolecules 2004;37:579-584.

[12] Shieh Y-T, Liu K-H, The effect of carbonyl group on sorption of CO2 in glassy polymer. J. Sup. Fluids 2003;25:261-268.

[13] Harrick N J. Internal reflection spectroscopy. Harrick Scientific Corp: Ossining, NY, 1987.

[14] Dikshit A K, Kaito A. Structure formation during the isothermal crystallization of an oriented blend of isotactic polystyrene with poly(2,6-dimethyl-1,4-phenylene oxide). Macromol. Rap. Comm. 2004;25:1019-1023.

[15] Holland B J, Hay J N. The thermal degradation of PET and analogous polyesters measured by thermal analysis-Fourier transform infrared spectroscopy. Polymer 2002;43:1835-1847.

[16] Jung D-H, Ko Y K, Jung H-T. Aggregation behaviour of chemically attached poly(ethylene glycol) to single-walled carbon nanotubes (SWNTs) ropes. Mater. Sci. Eng. 2004;C24:117-121.

[17] Nie F-Q, Xu Z-K, Yang Q, Wu J, Wan L-S. Surface modification of poly(acrylonitrile-co-maleic acid) membranes by the immobilization of poly(ethylene glycol). J. Memb. Sci. 2004;235:147-155.

[18] Kukova E, Petermann M, Weidner E. Phase behaviour (S-L-G) and fluid dynamic properties of high viscous poly (ethylene glycols)s in the presence of compressed carbon dioxide. Proceedings of 6th ISSF, materials processing, vol. 3, 2003, p1547-1552.

[19] Shah, V M, Hardy B J, Stern S A. Solubility of carbon dioxide, methane, and propane in silicone polymers. Effect of polymer backbone chains. J. Poly. Sci.: Poly. Phys. 1993;31:313-317.

[20] Sarbu T, Stranec T J, Beckman E J. Design and synthesis of low cost, sustainable CO2-philes. Ind. Eng. Chem. Res. 2000;39:4678-4683.

Chapter 3: Solubilities of sub- and supercritical CO2 in polyester resins: measurements and prediction

33

Chapter 3

Solubilities of sub- and supercritical CO2 in polyester resins:

measurements and prediction

The solubilities of carbon dioxide (CO2) in polyester resins based on propoxylated bisphenol (PPB) and ethoxylated bisphenol (PEB) have been measured using a magnetic suspension balance at temperatures ranging from 333 to 420 K and pressures up to 30 MPa. An optical cell has been used to independently determine the swelling of the polymers which has been incorporated in the buoyancy correction. In both polyester resins, the solubility of CO2 increases with increasing pressure and decreasing temperature as a result of variations in CO2 density. The experimental solubility has been correlated successfully to the Sanchez-Lacombe equation of state. Sameer P. Nalawade, Vishal E. Patil, Reiner Staudt, Francesco Picchioni, Jos. T. F. Keurentjes, and L. P. B. M. Janssen, Solubilities of sub- and supercritical CO2 in polyester resins: measurements and prediction, Polymer Engineering and Science, accepted.


34

3.1. Introduction Polyester resins in powder form are frequently used in the paint and toner industry. Milling, grinding and spray drying are the particle formation methods commonly used in the industry. Narrow particle size distribution, solvent recovery and the prevention of volatile organic components (VOC) emission are the major challenges associated with these methods. Moreover, the classical methods have clear disadvantages in terms of energy requirement due to expensive cryogenic cooling and problems with product quality due to heat dissipation during milling that causes agglomeration of molten polymer particles. Recently developed supercritical methods overcome the disadvantages of the classical methods [1]. Unusual solvent properties above the critical point like gas-like diffusivities and liquid-like densities make supercritical technology attractive. PGSS (Particles from gas saturated solution) is one of the particle production methods using a supercritical fluid [1]. Supercritical CO2 is generally used as a solvent or a plasticizer as it has a high solubility in many polymers. In PGSS, the viscosity of the polymer, particle size, and morphology of the particles are mainly determined by the amount of CO2 dissolved in the polymer. Therefore, it is important to determine solubilities of CO2 in a polymer at different conditions in order to define a processing window. Fourier transform infrared spectroscopy is a good tool to determine the interaction strength of CO2 and a polymer. However, it does not provide quantitative information on the solubility of CO2. Various experimental methods exist to determine the solubility of CO2 in solid and in molten polymers. Phase separation [2], volumetric [3], and gravimetric [4] methods are commonly used. In the first two methods, the amount of polymer required is large compared to gravimetric methods and hence, the time required to reach equilibrium is substantially longer. Moreover, high accuracy in pressure sensors and volume measurements are required in the first two methods for solubility calculations. These disadvantages are overcome by a gravimetric method that uses a microbalance [4]. The principle behind the gravimetric method is the weight difference between a gas-free and a gas-sorbed polymer sample. With a microbalance of high accuracy, even a small change in the weight of the polymer sample due to dissolved gas can be measured. Recently, a magnetic suspension balance (MSB), developed by Kleinrahm and Wagner [5], has been used to measure the solubility of CO2 in various polymers [6-8]. A major advantage of using the MSB is that measurements can be carried out at elevated temperatures and pressures without having direct contact between a sample and a balance. The main objective of this work is to determine the solubilities of CO2 in polyester resins using a MSB. An important parameter in such gravimetric


35

measurements is the swelling of polymer due to dissolved CO2. A buoyancy correction due to swelling has to be taken into account while calculating the dissolved quantity of CO2 in a polymer [6-8]. Therefore, an optical cell has been used in separate experiments to observe the swelling of polymers in the presence of CO2 at similar conditions. The second objective of the study is to describe the solubility data using a thermodynamic model, the Sanchez-Lacombe (S-L) equation of state (EOS) [9-10]. 3.2. Experimental 3.2.1. Materials Polyester resins based on propoxylated bisphenol (PPB, CAS: 177834-94-5), and ethoxylated bisphenol (PEB, CAS: 170831-75-1) were obtained from Akzo Nobel, The Netherlands. The physical properties of the polymers are provided in Table 3.1. Both polymers are amorphous which was confirmed by DSC measurements. Dry grade (> 99.5 %) carbon dioxide was used for the measurements. All chemicals were used as received without further purification. Table 3.1. Physical properties of the polymers

3.2.2. Apparatus and method Magnetic suspension balance A magnetic suspension balance (MSB), Fig.3.1, was used for measuring the solubility of CO2 in both polymers. The MSB can be used at temperatures up to 473 K and pressures up to 50 MPa. A polymer sample was kept in a basket which was not directly connected to the weighing balance (microbalance), but was kept in place using a so-called suspension magnet. The suspension consists of a measuring load, a sensor core, and a permanent magnet. The measured weight of the basket containing the polymer was transmitted by a magnetic suspension coupling to an external microbalance and thus, leak-proof measurements can be performed. In the MSB apparatus, the microbalance can be tared and calibrated during measurements as the sorption times are generally long. Using the MSB, the amount of CO2 dissolved in a polymer was determined from the following relationship,

Polymers Mn ( g/mol )

Mw (g/mol )

Tg ( K )

ηo at 363 K (Pa s)

PPB 2700 7000 325 -329 2965 PEB 8500 20000 328 -332 47540


36

2COW = W∆ +

2COρ ( )( )BP VSTPV +,, 1) where W∆ is the weight difference between a CO2 equilibrated polymer sample and the polymer sample without CO2 at similar temperature T and pressure P.

2COρ , VP (P, T, S), and VB are the density of CO2, the volume of the polymer containing dissolved CO2 with a solubility S, and the volume of the basket, respectively. The second term in equation 1)

Balance

Electromagnet

CO2 inlet

Sample basket

Permanentmagnet

Oil heater

Heater

P

Position sensor

Fig. 3.1. The magnetic suspension balance (MSB) apparatus used for the solubility measurements is a buoyancy correction term, which is required as polymers swell considerably when exposed to high CO2 pressures. As it was not possible to observe the polymer swelling simultaneously during the solubility measurements, an optical cell was used separately for the swelling measurements. CO2 solubility measurements were carried out above the glass transition temperature (Tg) of both polyesters. Temperature and pressure were varied from


37

333-420 K and 5-30 MPa, respectively. A polymer sample was first exposed to a vacuum for 1 hr at the measurement temperature and an initial reading was recorded. This was followed by addition of CO2 in the chamber until the desired pressure was attained. The sample was allowed to attain sorption equilibrium (in terms of the weight difference) before the final reading was recorded. Subsequently, more CO2 was introduced to attain a higher pressure and the new reading was recorded after equilibrium. Thus, sorption isotherms as a function of pressure were obtained. Optical cell A high-pressure optical cell used for swelling measurements is shown in Fig. 3.2. The cell can be used at temperatures up to 473 K and pressures up to 35 MPa. The inner volume of the cell is 20 ml. The temperature of the cell was controlled within a difference of 0.1 K using an oil bath. CO2 was pumped to the cell at elevated pressures using an HPLC pump. A 10 ml glass cuvette having a square cross section was used for holding a polymer sample. In the cuvette, the swelling of the polymer occurred only in one direction as the other directions were confined by the walls of the cuvette as shown in Fig. 3.2. The swelling of the polymer was viewed through a quartz window of the optical cell. A cathetometer having a precision of 0.01 mm was used to measure the difference in the height of the sample from which a fractional change in the volume of the polymer was calculated. Here, the change in the volume of the polymer sample is termed as fractional swelling, oVV /∆ . V∆ and oV are the increment in the volume of the polymer sample due to swelling and the volume of the polymer sample in the absence of dissolved CO2, respectively. For swelling measurements, a sample was prepared by pouring a molten polymer into a mould having a shape similar to the cuvette. After weighing the molded sample, it was fitted into the cuvette. The sample was again heated slightly above its Tg and pressed against the cuvette walls by a metal rod. The cylinder was then kept inside the cell and was heated to the desired temperature for about 3 hours. Subsequently, the polymer surface was marked with a cathetometer followed by the addition of CO2 to the desired pressure. Due to a rapid initial swelling, it was difficult to mark the surface immediately after the addition of CO2 with the cathetometer. Pressure-volume-temperature (PVT) data of the polymer (i.e. specific volume of the polymer) were used to correct the initial cathetometer reading. The sample was then allowed to attain equilibrium. As the sorption equilibrium was reached, no further swelling of the polymer occurred. Then, the new surface was marked with the cathetometer. The difference in the sample volume was used to calculate the swelling of the


38

polymer. For subsequent measurements at higher pressures an additional amount of CO2 was introduced stepwise and a similar procedure was repeated.

a)

b)

Fig. 3.2. a) The optical cell used for measuring the swelling of a polymer b) One dimensional swelling of a polymer The PVT data of the polymers were obtained using a high-pressure GNOMIX PVT apparatus (DatapointLabs, USA) for a temperature range from 317 K to 473 K and pressures up to 40 MPa. The PVT data of CO2 were obtained from the Span and Wagner EOS [11].


39

3.3. Sanchez-Lacombe equation of state To predict the solubility of CO2 at equilibrium conditions, the Sanchez-Lacombe equation of state (S-L EOS) was used. Measured solubilities of sub- and supercritical CO2 in various polymers have successfully been correlated to the S-L EOS [6-8]. According to this theory, the polymer molecules are ordered according to a lattice structure. The theory accounts for the change in volume due to the presence of “holes” in the lattice and hence, it does not require separate parameters to account for the flexibility of the molecule. The S-L EOS is given by,

( ) ( )[ ] 0~/11~1ln~~~ 2 =−+−++ ρρρ rTP 2)

,~/1~ ρυ = 3) where ρ~ , υ~ , P~ , T~ and r are the reduced density, specific volume, pressure, temperature, and the number of the lattice sites occupied by a molecule, respectively. An assumption used in the S-L EOS is that the polymer is monodisperse. The reduced parameters are defined as

***

*

*

*

*

/,/~,/~

,/~,/~

ρ

υυυ

ρρρ

RTMPrTTT

PPP

=

=

=

=

=

4)

where *ρ (the corresponding mass density in the close-packed state at 0 K), *υ (the corresponding specific volume in the close-packed state), *P (the hypothetical cohesive energy density in the close-packed state), and *T (related to the depth of the potential energy well) are the characteristic parameters of components. These parameters are obtained by fitting experimental PVT data of pure components to equations 2)-4) [12]. The EOS used for a mixture is similar to equation 2). The characteristic parameters used in the EOS for a mixture are obtained using the following mixing rules.

**ijj

i ji PP φφ∑∑= 5)

( )( ) 5.0*** 1 jiijij PPkP −= 6)


40

( ) **0** / ii

ii PTPT ∑= φ 7)

**0 υυ iii rr = 8)

∑=

iii*0* υφυ 9)

i

ii rr //1 ∑= φ 10)

( ) ( )∑=

jjjjiiii TPTP ****0 /// φφφ 11)

( ) ( )∑=j

jjiii ww ** /// ρρφ 12)

where φ and w represent the volume and weight fraction of components in two phases, respectively. Superscript ‘0’ denotes the pure state of a component. Along with equations 1)-12), the chemical potential (µ ) of a component (i) in the available phases is used to predict the solubility of CO2 in a polymer. At equilibrium,

polymeri

gasi µµ = 13)

Here, CO2 is termed as component 1 while a polymer as component 2. The chemical potential of 1 in the polymer phase is given by

( )[ ]( ) ( )( )[ ]ρρρρυυρ

φρφφµ~ln/~~1ln~1~~/~~/~

))~/()2((~/1ln0

11110

1

1*

1*

12*

2*

122

0122111

rTPTRTr

TPPPPrrrRTpolymer

+−−++−

+−++−+= 14)

Equation 14) is also used to calculate gas

1µ by considering only the gas phase. For polymers of high molecular weight, it is safe to assume that no polymer is present in the gas phase. Thus, the experimental solubility data are regressed with an adjustable interaction parameter, kij, which measures the deviation of ijP from the geometric mean of iP and jP using equations 2)-14). 3.4. Results and discussion The PVT data of the polymers and CO2 are essential for the interpretation of the swelling measurements and in the S-L EOS as discussed above. Therefore, the PVT and swelling studies will be described before the solubility results. The PVT data have been successfully modelled using the S-L EOS for a wide range of temperatures, pressures, and densities. The results for CO2, PPB, and PEB are shown in Fig. 3.3, 3.4, and 3.5, respectively. The characteristic parameters of the


41

pure components obtained using the S-L EOS are given in Table 3.2. These parameters have been used to calculate the solubility. Table 3.2. The characteristic parameters of the polymers and CO2 obtained using the S-L EOS

Component P* (MPa)

T* (K)

ρ * (kg/m3)

CO2 427.70 338.7 1405.5 PPB 439.72 683.2 1242.7 PEB 640.27 728.6 1271.0

0

200

400

600

800

1000

320 345 370 395 420 445

T, K

ρ, k

g/m

3

4.5 MPa10 MPa

15 MPa

20 MPa30 MPa

35 MPa

S-L EOS

Fig. 3.3. Densities of CO2 obtained from Span and Wagner EOS and predicted using the S-L EOS

1070

1090

1110

1130

1150

1170

1190

325 345 365 385 405 425 445

T, K

ρ,

kg/m

3

4.5 MPa

10 MPa

20 MPa30 MPa

35 MPa

S-L EOS

Fig. 3.4. Densities of PPB obtained experimentally and predicted using the S-L EOS


42

1120

1140

1160

1180

1200

1220

325 345 365 385 405 425 445

T, K

ρ, k

g/m

3

4.5 MPa

10 MPa

20 MPa

30 MPa

35 MPa

S-L EOS

Fig. 3.5. Densities of PEB obtained experimentally and predicted using the S-L EOS. When CO2 is dissolved into a polymer, the mobility of the polymer chains is increased due to disentanglement of the polymeric chains. As a result, the free volume of the polymer is increased and swelling of the polymer takes place. Experimentally obtained swelling isotherms of PPB and PEB in the presence of CO2 using the optical cell apparatus are shown in Fig. 3.6 and 3.7, respectively. It can be seen from these figures that the fractional swelling increases with increasing pressure for both PPB and PEB. In general, the higher the dissolved amount of CO2 in a polymer the larger is the swelling of the polymer. This effect is the result of increasing CO2 density upon an increase in pressure. Since the CO2 density decreases with increasing temperature, a reduction in the swelling is expected at higher temperatures in both polymers. However, this is observed only for PPB. The swelling of PEB increases with increasing temperature. This behaviour has also been reported for poly (dimethylsiloxane) (PDMS) [13], poly (ethylene-terephthalate) (PET) and bisphenol-A polycarbonate (PC) [14]. It has been suggested that the CO2 density is not the only parameter, which affects the swelling of the polymer. A positive temperature influence on chain mobility is pronounced compared to the influence of CO2 density for an inverse swelling behavior [14]. This effect is not present in PPB, which is most probably due to an easily accessible free volume even at low temperatures due its low molecular weight.


43

0

0.04

0.08

0.12

0.16

0 5 10 15 20 25 30

P, MPa

frac

tiona

l sw

ellin

g

333 K

368 K

420 K

Fig. 3.6. Fractional swelling isotherms of PPB in the presence of CO2

0

0.04

0.08

0.12

0.16

0 5 10 15 20 25

P, MPa

frac

tiona

l sw

ellin

g

334 K

373 K

418 K

Fig. 3.7. Fractional swelling isotherms of PEB in the presence of CO2 CO2 solubilities in both polymers have been measured and have been predicted using the S-L EOS. The results are shown in Fig. 3.8 and 3.9. The experimental data from the MSB has been corrected for buoyancy effects using the results of the swelling measurements. The CO2 solubility has been represented in terms of the weight fraction of CO2 dissolved in the polymer. The solubilities of CO2 have been corrected with the experimental swelling data. For both polymers, it is found that the solubility increases with increasing pressure, whereas it decreases with increasing temperature. This relates to high CO2 densities at high pressures and low temperatures, and vice versa. At a low temperature, 333 K, the solubility behaviour is not linear above 15 MPa for both polymers. Probably at elevated pressures, the free volume available in the polymer is reduced due to compression of the polymer. Such effects are more pronounced at the low temperature due to relatively small free volume. Moreover, Fig. 3.8 and 3.9 also show that this non-linear behaviour is absent at high temperatures.


44

0

0.05

0.1

0.15

0.2

0.25

0 5 10 15 20 25 30 35

P, MPa

CO 2

wt.

frac

tion

333 K

368 K

420 K

S-L EOS

Fig. 3.8. The CO2 solubility isotherms of PPB, k12(333 K)= 0.065, k12(368 K)= 0.09, and k12(420 K)= 0.108

0

0.02

0.04

0.06

0.08

0.1

0 5 10 15 20 25 30 35

P, MPa

CO 2

wt f

ract

ion 334 K

373 K418 K

S-L EOS

Fig. 3.9. The CO2 solubility isotherms of PEB, k12(334 K)= 0.905, k12(373K)= 0.826, and k12(418 K)= 0.119 The CO2 solubilities in PPB are higher than in PEB. The importance of the free volume available in the polymer [15] and minor changes in the groups present in a polymer [16-17] for the CO2 solubility has already been discussed in the literature. Since the structure is nearly similar for both polyesters, the accessible free volume in PEB that is smaller due to higher chain entanglements (high molecular weight) compared to PPB is probably responsible for different solubility. The higher chain entanglements make CO2 more difficult to access the carboxyl groups in PEB than in PPB. Albeit the swelling increases with an increase in temperature in PEB, the relatively low CO2 densities at high temperatures reduce the solubility. The solubility results have been correlated with the S-L EOS for both PPB and PEB, see also Fig. 3.8 and 3.9, respectively. In order to fit the S-L EOS to the


45

experimental solubility data equations 2)-14) have been solved using the characteristic parameters determined from the PVT data (Table 2). A non-linear regression optimization procedure (Levenberg-Marquardt, MATLAB 7) has been used for minimizing the difference between the chemical potential of CO2 in the gas phase and the polymer phase, and also between the experimental and predicted solubilities using the interaction parameter, k12. In the case of PPB, k12 varies linearly with temperature so that a simple linear relationship between k12 and T can be used to interpolate and possibly extrapolate the solubilities at different temperatures and pressures. Though the S-L EOS has already been reported for several polymers to predict the swelling due to dissolved CO2, it has not been tested together with the experimental swelling and solubility data for molten polymers. The density of a mixture can be determined using the S-L EOS, which has been used to predict the swelling of both polymers. The swelling has been poorly predicted by this EOS for PPB and PEB. The linear mixing rule for the volume of the mixture in the S-L EOS may be responsible for this poor predictability [13]. Recently, over prediction of swelling using the S-L EOS has been reported for EVA polymers by Jacobs et al. [18]. 3.5. Conclusions The solubilities of CO2 in PPB and PEB in the molten state have been measured using the MSB. It has been necessary to correct the data obtained from the MSB with independent swelling data. The polymers swell considerably in the presence of CO2 when they are exposed to elevated pressures. The CO2 solubility in the polymers increases with an increase in pressure and decreases with an increase in temperature. The CO2 solubility in PEB is lower than in PPB, which is probably due to its smaller free volume. At 333 K, a non linear trend for solubility-pressure has been observed at elevated pressures for both polymers due to compression effect. The experimental solubility data have been correlated to the S-L EOS using the pure component parameters and an adjustable interaction parameter. Although the S-L EOS has often been used to predict the swelling of a polymer in the presence of CO2, it is a poor model for predicting the swelling of the polymers investigated here. References [1] Perrut M, Jung J. Particle design using supercritical fluids: Literature and patent

survey. J. Sup. Fluids 2001;20;179-219. [2] Wiesmet V, Weidner E, Behme S, Sadowski G, Arlt W J. Measurement and

modeling of high-pressure phase equilibria in the systems polyethylene glycol (PEG)-propane, PEG-nitrogen and PEG-carbon dioxide. J. Sup. Fluids 2000;17:1-12.

[3] Sato Y, Yurugi M, Fujiwara K, Takishima S, Masuoka H. Solubilities of carbon dioxide and nitrogen in polystyrene under high temperature and pressure. Fluid Phase Equilibria 1996;125:129-138.


46

[4] Kamiya Y, Mizoguchi K, Terada K, Fujiwara Y, Wang J-S. CO2 sorption and dilation of poly(methyl methacrylate). Macromolecules 1998;31:472-478.

[5] Kleinrahm R, Wagner W. Measurement and correlation of the equilibrium liquid and vapour densities and the vapour pressure along the coexistence curve of methane. J. Chem. Thermodyn. 1986;18:739-760.

[6] Sato Y, Takikawa T, Sorakubo A, Takishima S, Masuoka H, Imaizumi M. Solubility and diffusion coefficient of carbon dioxide in biodegradable polymers. Ind. Eng. Chem. Res. 2000;39:4813-4819.

[7] Sato Y, Takikawa T, Takishima S, Masuoka H. Solubilities and diffusion coefficients of carbon dioxide in poly(vinyl acetate) and polystyrene. J. Sup. Fluids 2001;19:187-198.

[8] Sato Y, Takikawa T, Yamane M, Takishima S, Masuoka H. Solubility of carbon dioxide in PPO and PPO/PS blends. Fluid Phase Equilibria 2002;194-197:847-858.

[9] Sanchez I C, Lacombe R H. An elementary molecular theory of classical fluids. Pure fluids. J. Phys. Chem. 1976;80:2352-2362.

[10] Sanchez I C, Lacombe R H. Statistical thermodynamics of polymer solutions. Macromolecules 1978;11:1145-1156.

[11] Span R, Wagner W. A new equation of state for carbon dioxide covering the fluid region from the triple-point temperature to 1100 K at pressures up to 800 MPa. J. Phys. Chem. Ref. Data 1996;25:1509-1596.

[12] Garg A, Gulari E, Manke C W. Thermodynamics of polymer melts swollen with supercritical gases. Macromolecules 1998;27:5643-5653.

[13] Royer J R, DeSimone J M, Khan S A. Carbon dioxide-induced swelling of poly(dimethylsiloxane). Macromolecules 1999;32:8965-8973.

[14] Schnitzler J V, Eggers R. Mass transfer in polymers in a supercritical CO2-atmosphere. J. Sup. Fluids 1999;16:81-92.

[15] Shah V M, Hardy B J, Stern S A. Solubility of carbon dioxide, methane, and propane in silicone polymers:effect of polymer backbone chains. J. Poly. Sci.: Poly. Phys. 1993;31:313-317.

[16] Kazarian S G, Vincent M F, Eckert C A. Infrared cell for supercritical fluid-polymer interactions. Rev. Sci. Instrum. 1996;67:1586-1589.

[17] Kazarian S G, Vincent M F, Bright F V, Liotta C L, Eckert C A. Specific intermolecular interaction of carbon dioxide with polymers. J. Am. Chem. Soc.1996;118:1729-1736.

[18] Jacobs M A, Kemmere M F, Keurentjes Jos T F. Foam processing of poly(ethylene-co-vinyl acetate) rubber using supercritical carbon dioxide. Polymer 2004;45:7539-7547.

Chapter 4: Batch production of micron size particles from poly(ethylene glycol) using supercritical carbon dioxide as a solvent

47

Chapter 4

Batch production of micron size particles from poly(ethylene

glycol) using supercritical carbon dioxide as a solvent

The major advantage of using supercritical CO2 as a solvent in polymer processing is an enhancement in the free volume of a polymer due to dissolved CO2, which causes a considerable reduction in the viscosity. This allows spraying the polymer melt at low temperatures to produce micron size particles. We have used supercritical CO2 as a solvent for the generation of particles from poly(ethylene glycol) (PEG) of different molecular weights. Since PEG is a hydrophilic compound, it is a most commonly used polymer for encapsulating a drug. PEG particles with different properties may allow keeping a good control over the release of the drug. It has been possible to produce particles with different size, size distribution, porosity and shape by playing with the various process and product related parameters such as molecular weight, temperature, pressure, and nozzle diameter. A flow and a solidification model have been applied to gain a theoretical insight into the role of the different parameters. Sameer P. Nalawade, F. Picchioni, L.P.B.M. Janssen, Batch production of micron size particles from poly(ethylene glycol) using supercritical carbon dioxide as a solvent, Chemical Engineering Science, submitted.


48

4.1. Introduction Pharmaceutical industries are always in an utmost need of methods or technologies, which can produce fine particles without using organic solvents. This is possible with traditional methods such as grinding or milling. However, the cost in case of cryogenic milling or the heat dissipation during grinding constraint their applications to particular materials. During the last two decades, various methods using supercritical fluids as solvents or anti solvents have become available. Supercritical fluids could replace the organic solvents due to their gas-like and liquid-like properties, which play a vital role in the dissolution process. It is possible to produce nano or micron size particles not only for low but also for high molecular weight materials using supercritical fluids [1]. Among various fluids, carbon dioxide (CO2) has already been touted as a supercritical fluid for many applications as it is inexpensive and inert in nature. It has a critical temperature close to room temperature (304 K) and a moderate critical pressure (7.38 MPa). Moreover, since it is a gas at ambient conditions, it is easy to separate it from the final product. PGSS (particles from gas saturated solution) is one of the particle production methods, which uses supercritical fluid as a solvent. PGSS utilizes the solubility of a supercritical fluid in a material. The dissolved supercritical fluid reduces the viscosity of the material to be micronised and hence, allows expansion at low temperatures. The principle of PGSS is actually very simple. Thermodynamics instability caused by depressurisation leads to a supersaturation due to reduction in the solubility of a gas. Supersaturation is defined as the ratio of the actual solution concentration when precipitation or crystallization takes place to the equilibrium concentration. The supersaturation leads to nucleation of micron size gas bubbles and a vigorous expansion of these bubbles breaks up the solution into small particles. CO2 is generally used as a supercritical fluid in PGSS for polymers mainly due to its high solubility [2-6]. PGSS has already been reported for the production of particles from poly(ethylene glycol) of different molecular weights of 1500, 4000, and 35000 [2-3]. In this work, particle production of poly(ethylene glycol) (PEG) of different molecular weights using supercritical CO2 has been studied in detail. PEG is one of the commonly used compounds in the pharmaceutical industry for controlled drug delivery because of its hydrophilic nature. PEG is available in different states such as liquid or solid depending on the molecular weight. PEG in the solid state is mainly used for the pharmaceutical applications as drug carriers in particle form. In the application above, the particle shape, size, density, and particle size distribution are very important. Since the melting point (Tm) of PEG is not high (around 335 K), milling or grinding is not an easy task.


49

In this study, the batch production of particles from PEG with a weight average molecular weight (Mw) of 6000 and 10000 has been reported. The experimental results have been presented in terms of particle size, shape, morphology, and particle size distribution, which are dependent on the molecular weight, nozzle diameter, and process conditions. Moreover, a flow and a solidification model have been applied to study the effect of the rate of nucleation on the particle size distribution and the effect of different parameters on the particle shape, respectively. 4.2. Experimental 4.2.1. Materials Poly(ethylene glycol) (PEG) having a weight average molecular weight (Mw) of 6000 and 10000 were purchased from Aldrich, The Netherlands. Digital scanning calorimetric (DSC) measurements were carried out to determine the melting points (Tm) and the heat of melting (∆Hm) of PEG 6000 and 10000. The crystallinity of the PEG calculated using ∆Hm of 100 % crystalline PEG (196.8 J/g [7]) are provided in Table 4.1. High purity CO2, 99.99 %, was used in the experiments. The materials were used without any further purification. Table 4.1. The physical properties of PEG of different weight average molecular weights (Mw)

Mw Tm (K)

∆Hm (J/g)

Crystallinity %

6000 332 185.1 94 10000 338 181.4 92

4.2.2. Apparatus The batch set up is shown schematically in Fig. 4.1. The set up, designed and constructed, in our laboratory can withstand a pressure of 25 MPa at 473 K. A cylinder (150 ml), tubes, nozzle and valves were obtained from Swagelok, The Netherlands. In the set up, a pressure sensor (Dynisco, USA) is mounted in the CO2 line using a connector just before to the entrance of the cylinder. Before the nozzle, a ball valve (open-close) is used as it can be opened or closed more quickly. The cylinder and the tubes are heated using a heating element, whose temperature is controlled using a Eurotherm controller, The Netherlands. The set up is mounted on the supports, which oscillate, very slowly, over 180 o in order to achieve better mixing. Additionally, an insert having left-right elements in the cylinder also improves the mixing. CO2 is added using a membrane pump (Lewa, USA) and heated in a double pipe heat exchanger prior to its addition. A


50

drum is used for the expansion of the polymer solution and collection of the particles.

Nozzle

T P CO2 in

Vacuum Heating element

Fig. 4.1. A schematic drawing of the batch particle production set up 4.2.3. Experimental method A known quantity of PEG, ~ 75 g, was first added to the cylinder and the set up was then subjected to a vacuum. Subsequently, the temperature of the polymer was set to a desired value and CO2 was added using a membrane pump to the cylinder until the desired pressure was reached. The set up was then disconnected from the pump and mounted on a horizontal support. The polymer was always allowed to be in contact with CO2 at least for 4 hrs despite the fact that the equilibrium, in terms of a pressure reduction, was achieved in a relatively short time. Then, the set up was mounted vertically to expand the PEG-CO2 single phase solution. Before expanding the solution, the CO2 inlet line was connected to the CO2 pump to avoid that the pressure inside the cylinder would decrease below the saturation pressure. Finally, the solution was expanded over a nozzle (length of 1.2 mm) for a short time in a spray drum and solid powder was collected. 4.2.4. Particle analysis A wet laser diffraction (WLD) apparatus, Malvern Mastersizer®, was used to measure the particle size and PSD. Toluene was used as a solvent for the particle size measurements. An average particle diameter (dp,0.5) was determined from the cumulative volume fraction. Scanning electron microscope (SEM) was used to observe the morphology and shape of the particles. 4.3. Results and discussion 4.3.1. Literature: CO2 solubility and viscosity of PEG-CO2 In PGSS, the shear viscosity plays an important role in the break up of a molten polymer into particles. The higher the viscosity the more difficult is the particle production [2-3]. Therefore, not only the supersaturation but also the shear


51

viscosity is important. Since the viscosity reduction due to dissolved CO2 is dependent on the amount of CO2 dissolved, a detailed knowledge of the solubility and viscosity reduction is essential in particle production. Weidner et al. [8] have published the CO2 solubility data for PEG (Mw 1500-35000 g/mol) over a pressure range from 0.5-30 MPa at different temperatures in the range of 323-393 K. It was found that the CO2 solubility is independent of the molecular weights investigated, Fig. 4.2. Recently, viscosities of several PEG-CO2 solutions have been measured by Kukova [9]. The viscosities curves for different molecular weights are shown in Fig. 4.3. From Fig. 4.3, it is clear that the viscosity increases in the multiple of 2 with an approximately 60 % increment in the molecular weight.

0

0.05

0.1

0.15

0.2

0.25

0.3

0 5 10 15 20 25 30

P, MPa

CO 2

wt f

ract

ion

328 K338 K

343 K

348 K

Fig. 4.2. CO2 sorption isotherms for PEG 6000 and PEG 10000 [8] (Reprinted from Journal of Supercritical Fluids, Wiesmet V, Weidner E, Behme S, Sadowski G, Arlt W., Measurement and modelling of high-pressure phase equilibria in the systems polyethyleneglycol (PEG)-propane, PEG-nitrogen and PEG-carbon dioxide, 17:1-12 (2000), with permission from Elsevier)

0

500

1000

1500

2000

2500

0 5 10 15 20 25

P, MPa

η, m

Pa-s

328 K

338 K

343 K

Fig. 4.3. Viscosity reductions in the presence of CO2 for PEG 6000 (solid lines) and PEG 10000 (dashed lines) [9] (Reprinted from PhD Thesis, Kukova E., Phasenverhalten und Transporteigenschaften binärer Systeme aus hochviskosen Polyethylenglykolen und Kohlendioxid (2000) with permission from Ruhr-Universität Bochum, Germany)


52

4.3.2. Batch production of PEG particles Several experiments have been performed at different temperatures and pressures in the supercritical state using two different nozzles, 0.81 and 0.36 mm, for PEG 6000 and 10000. All experimental results of powder generation from PEG 6000 and 10000 using different diameter nozzles are provided in Table 4.2. These results have been explained using various parameters such as the molecular weight, temperature, pressure, and nozzle diameter. Table 4.2. The average diameter of the particles obtained from PEG 6000 and 10000 using different nozzles

PEG (Mw)

dn (mm)

T (K)

P (MPa)

dp,0.5

( µm) 6000 0.81 328 20 297.2

15 307.2 12.7 321.0 0.81 338 20 246.9 15.6 213.7 12.5 255.1 0.81 343 15.8 324.4 12.6 316.5 0.36 328 22 266.9 13 233.4 0.36 338 22 200.1 13 198.1

10000 0.81 328 20.0 318.2 14.5 330.6 0.81 338 21.1 357.1 14.8 345.7 0.81 348 20.7 321.1 15.5 354.8 0.36 338 21.0 292.7 12.0 314.2 0.36 348 20.5 309.2 12.2 280.1

Since a few experiments have also been carried out below the melting point of the polymers, it is requisite to ensure that PEG is present in the liquid state under these conditions. It has already been reported in literature [8] that the dissolved CO2 in PEG causes a considerable decrease in the melting point. Minimum CO2 pressures around 2 and 5.5 MPa at 328 K have been reported for PEG 6000 and 12000, respectively, above which they are present in the liquid state. Therefore, the pressures above 5.5 MPa have always been selected in these experiments.


53

During experiments, PEG has always been exposed to the temperatures specified above for a long duration to ensure CO2 saturation. This may lead to thermal degradation of the polymer. Therefore, a thermal degradation test has been performed for the low molecular weight PEG. No weight loss during a thermo gravimetric analysis (TGA) of the low molecular weight PEG confirms the thermal stability of both PEG (Fig. 4.4).

96

97

98

99

100

101

102

103

104

0 25 50 75 100 125 150 175

Time, min

wt.

loss

, %

PEG 6000

Fig. 4.4. Weight loss (%) as a function of time of PEG 6000 at 350 K determined using a thermo gravimetric apparatus Effect of molecular weight The shear viscosity of the polymer plays an important role in PGSS during the break up of polymer/gas solution into particles. The lower the polymer solution viscosity, the easier is the break up of the solution. The viscosity of polymer can in general be related to its molecular weight, the viscosity increases with increasing molecular weight. Therefore, PEG of different molecular weights have been selected for this study. The effect of the molecular weight on the shape of the particles for PEG 6000 and 10000 can be seen in Fig. 4.5. Relatively long fibre shape particles are obtained for PEG 10000 compared to PEG 6000 for similar processing conditions below the melting point of both PEG. Since the CO2 solubility is independent of the molecular weight in case of PEG, the viscosity is responsible for the different morphology of the products. In the high viscous polymer, expansion force is not sufficient to overcome viscous force and hence, break up of the solution is delayed compared to the solidification. Above the melting point (pure polymer), no considerable effect of the molecular weight on the shape of the particles has been found. However, the results can be compared with the particle size as smaller particles are obtained with the low molecular weight PEG, Table 4.2.


54

a) PEG 6000 b) PEG 10000

Fig. 4.5. Scanning electron microscope pictures of the particles obtained using 0.81 mm diameter nozzle at 20 MPa and 328 K Effect of pressure and temperature The higher the dissolved amount of CO2 in a polymer, the higher is the supersaturation and hence, a larger nucleation of CO2 bubbles is achieved during the expansion [1]. Since the dissolved amount of CO2 is a function of temperature and pressure, the effect of temperature and pressure are discussed together for both PEG. It is possible to produce particles from both PEG 6000 as well as PEG 10000 even below their melting points. The decrease in the melting point due to dissolved CO2 allows the expansion of the polymers. The higher the dissolved amount of CO2, the higher is the decrease in the melting point. A considerable effect of temperature on the particle size is observed for PEG 6000, Table 4.2. The particle size increases with increasing temperature above its melting point. The effect can be related to a decrease of the CO2 solubility with increasing temperature. However, converse to the result above, large particles of PEG 6000 are obtained below its melting point probably due to foaming of the particles. In Fig. 4.6, foaming can be seen at the temperature below the melting point for both nozzles in case of PEG 6000. This might be explained by the fact that CO2 cannot escape easily due to rapid solidification of the particles as the polymer is already below its melting point. Since the solidification is fast, the diffusion of CO2 from the particles is reduced due to an instantaneous increase in the viscosity of the polymer. Consequently, bubbles of CO2 are captured inside the particles resulting in foamed particles. The effect of temperature on the particle size is, probably due to high viscosity, absent for PEG 10000.


55

a) b)

c) d) Fig. 4.6. Scanning electron microscope pictures of the PEG 6000 particles obtained using 0.81 mm diameter nozzle at a) 12.7 MPa, 328 K, b) 20 MPa, 328 K c) 12.5 MPa, 338 K, and d) 20 MPa, 338 K A noticeable effect of temperature has been observed on the shape of the particles. The particles obtained from PEG 6000 and 10000 under different processing conditions are shown in Fig. 4.7. At high temperatures nearly-spherical shaped particles are obtained. The shape of the particles is mainly dependent on the available solidification time. As the temperature is increased, more sensible heat of the material needs to be removed, which delays the solidification process. Moreover, less energy is utilized for evaporation of CO2 as the amount of CO2 dissolved decreases with increasing temperature. Such a delayed solidification facilitates retraction of molten polymer into a spherical shape both by visco-elastic relaxation and surface tension. Unlike temperature, the effect of pressure on the particle size and shape has been absent for both PEG. Similar results have been reported in the literature for


56

PEG of different molecular weights [2]. Pressure affects mainly the bulk density of particles. For example, the bulk densities of the particles obtained at 13 and 22 MPa (338 K) using 0.36 mm nozzle are 270 and 195 kg/m3, respectively. The higher the pressure, the higher is the CO2 solubility. Therefore, more foamed (less dense) particles are produced at the elevated pressures.

a) b)

c)

Fig. 4.7. Scanning electron microscope pictures of the particles obtained using a nozzle of 0.36 mm diameter a) PEG 6000: 22 MPa, and 328 K, b) PEG 6000: 22.5 MPa, and 338 K; and c) PEG 10000: 12.2 MPa, and 348 K In addition to the particle size, shape, and morphology, the effect of temperature and pressure on the particle size distribution has also been tested. The particle size distributions of the particles of PEG 6000 and 10000 obtained under different processing conditions are shown in Fig. 4.8. A bimodal distribution with a very small peak in the range of 0.1-1 micron size is obtained. In PGSS, generally, a cyclone separator is used in order to remove very fine dust. Such arrangement is not present in our set up. The effect of pressure on the particle


57

size distribution is absent while relatively narrow particle size distributions are obtained at the low temperature for PEG 6000. In case of PEG 10000, the effect of pressure and temperature on the particle size distribution is absent.

0

20

40

60

80

100

0 200 400 600 800 dp,0.5, µm

q, %

12.5 MPa, 338 K, PEG6000

20.0 MPa, 338 K, PEG6000

12.6 MPa, 343 K, PEG6000

21.1 MPa, 338 K, PEG10000

20.7 MPa, 348 K, PEG10000

Fig. 4.8. The particle size distributions of the PEG particles obtained using a 0.81 mm diameter nozzle at different processing conditions Effect of nozzle diameter From Table 4.2, it is clear that the average diameter of the particles is decreased with decreasing nozzle diameter. Apart from the size of the nozzle, this may be explained using a pressure drop rate depending of the nozzle diameter. The classical homogeneous nucleation theory [10] is not applicable here as it includes only the pressure drop and not the pressure drop rate. The effect of pressure drop rate on the nucleation of CO2 bubbles has already been discussed theoretically for a microcellular foaming study [10]. Authors conjectured that pressure drop rate decides the solubility drop rate, which in turn determines the nucleation rate of CO2 bubbles. The approach above has been considered also for this study. Thus, it is first necessary to calculate the average residence time of a polymer solution inside a nozzle under different expansion pressures. The Fanning equation, equation 1), has been used to calculate the velocity of solution, which in turn has been used to calculate the residence time. Here, it has been assumed that the single phase polymer solution is incompressible. The assumption is valid as the concentration of the polymer is always much higher than the concentration of CO2.

⎟⎟⎠

⎞⎜⎜⎝

⎛⎟⎠⎞

⎜⎝⎛=∆

n

ns d

lfP 2

214 υρ 1)

In the equation above, P∆ , f, sρ , υ , ln, and dn are the pressure drop, the friction factor, the density of the polymer solution, the velocity of the polymer solution, the length of the nozzle, and the diameter of the nozzle, respectively. The


58

density of the polymer solution has been approximated using a linear mixing rule based on the weight fraction of CO2. Equation 1) is applicable to Newtonian fluids. Viscosity measurements have been performed for PEG 6000 at 338 and 343 K in a shear rate range of 10-200 1/s. In this region, the polymer behaves like a Newtonian fluid at both temperatures. In the equation above, the friction factor is different for laminar to turbulent flow. Therefore, the Reynolds numbers (NRe) have been calculated for different nozzle diameters. Because the velocity is not known, a limiting value of 300 m/s (being the velocity of sound in vacuum) has been used. It has been found that the flows are in the laminar regime (< 2000), despite the high velocity, because of high viscosity of PEG. To calculate the friction factor, the f-NRe relationship for the laminar regime has been used, equation 2).

Re

64644Nd

fsn

s =⎟⎟⎠

⎞⎜⎜⎝

⎛=

υρη 2)

In equation 2),

sη is the viscosity of the polymer solution. The average pressure drop rate is calculated using equation 3), where t∆ is the average residence time of the polymer solution.

nlP

tP

dtdp υ∆

−≈∆∆

−≈− 3)

Pressure drop rates obtained using equations 1)-3) for PEG 6000 are given in Table 4.3. It is clear from the obtained values that the pressure drop rate is higher in the smaller diameter nozzle. An order of magnitude of the pressure drop rate for the smaller nozzle is around 25 times higher than for the bigger nozzle, while it is only around 3 in case of the different pressures for the same nozzle. Table 4.3. The pressure drop rates calculated for different nozzles

P (MPa)

tP∆∆

(MPa/s ) tP∆∆

(MPa/s) 0.36 mm 0.81 mm

20 26917.7 1100.4 12.5 8467.2 278.7

Since the CO2 solubility is linearly proportional to pressure for PEG, the solubility drop rate may be considered proportional to the pressure drop rate. A high solubility drop rate in the smaller diameter nozzle results into a high


59

thermodynamic instability, supersaturation, and hence, a large number of nuclei is formed. Moreover, due to short residence time in the nozzle with smaller diameter a growth of the existing nuclei is a limiting factor. A better expansion takes place with this large number of nuclei and hence, smaller diameter particles are produced. On the other hand, a relatively longer time is available for a growth of the existing nuclei in the bigger nozzle. In addition to the particle size, a positive effect of the nozzle diameter on the particle size distribution has been found for PEG 6000, Fig. 4.9. A narrow particle size distribution is obtained for the smaller diameter nozzle. For PEG 10000, due to high viscosity, an improvement in the particle size distribution is not comparable to PEG 6000.

0

20

40

60

80

100

0 200 400 600 800

dp,0.5, µm

q, %

13 MPa, 338 K, 0.36mm nozzle

12.5 Mpa, 338 K, 0.81mm nozzle

Fig. 4.9. The particle size distributions of the particles (PEG 6000) obtained using a 0.81 and 0.36 mm diameter nozzle 4.3.3. Solidification In PGSS, the effect of temperature and pressure on the shape and morphology (porosity) can be explained using solidification of molten particles (droplets). The effect of temperature and pressure on the shape and morphology have qualitatively been discussed in detail. In addition to convection, the amount of CO2 dissolved, which is dependent on temperature and pressure, contributes to the solidification in the form of the heat of evaporation of dissolved CO2 ( vH∆ ). A solidification model has been applied to a single droplet. An average time (tavg) required for cooling down the droplet to the melting point with a complete phase change has been calculated using this model. The estimated solidification time provides an insight into the dependency of the particles shape and morphology on different parameters. The assumption used in the model are 1. The droplet is spherical in shape.


60

2. The velocity difference between droplet and air present in the spray drum is negligible. 3. CO2 expansion within the droplet is instantaneous after the nozzle exit. 4. No under cooling of the droplet takes place. 5. The droplet/particle does not shrink or expand during the solidification. A liquid droplet is cooled by losing heat by evaporation of CO2 and convection to the surrounding. For a crystalline material, the heat of crystallization ( mH∆ ) also needs to be taken into account. According to an energy balance over the droplet, heat in excess to the heat of evaporation ( vH∆ ) must be removed by the convection. Since the convection occurs over a complete solidification period, the average solidification time can be calculated using a convection term. The amount of heat needs to be removed in order to cool the droplet to the melting point of the polymer (Tm) is given by

⎟⎟⎠

⎞⎜⎜⎝

⎛

−∆−∆+∆−=−

2

2

1)( ln

co

covppmppavgmixmppp w

wHVHVtThSTTCV ρρρ 4)

where, h, S, lnT∆ , pρ , Vp, Cp, Tmix, and 2cow are the heat transfer coefficient due to convection, the surface area of the droplet, the logarithmic mean temperature difference, density of the polymer, the volume of the droplet (exclusive of the volume occupied by the pores), specific heat of the polymer, the temperature of the mixture, the weight fraction of CO2, respectively. The lnT∆ is calculated using the Tmix, Tair and Tm. The h is calculated by using equation 5) proposed by Ranz-Marshall [11].

( )3PrRe

5.0,

6.00.2 NNd

hp

a +=λ 5)

where, aλ and NPr are the thermal conductivity of air and the Prandtl number. Before applying the heat balance equation, it is necessary to check the condition that there is no temperature gradient present within the droplet (NBi (Biot number) << 1) [12]. For calculations, the physical properties reported elsewhere for PEG 6000 have been used [9]. The heat of crystallization, the specific heat, and the thermal conductivity ( pλ ) of PEG 6000 are 181.5 kJ/kg, 2.1-2.5 kJ/kgK (293-393 K), and 0.23 w/mK, respectively. The solidification times have been estimated for different isothermal conditions used in the case of nozzle diameter of 0.81 mm. Constant values of average particle diameter as 300 and 320 micron for different pressures at 328 and 338 K, respectively, have been used in the calculations. For porous particles, the volume fraction occupied by the bubbles has been


61

approximated to 0.4 for all calculations. A constant value for the temperature of air has been considered (Tair = 296 K). The results obtained using the model above are shown in Fig. 4.10.

0.4

0.5

0.6

0.7

0.8

0.9

1

12 14 16 18 20

P, MPa

t avg, s

328 K

343 K

Fig. 4.10. The estimated average solidification time for PEG 6000 particles at different temperatures In Fig. 4.10, the solidification time is decreased with increasing pressure and decreasing temperature. A high CO2 solubility is mainly responsible for a decrease in the solidification time. This is related to an amount of energy used for the evaporation of CO2, which is large at higher pressures and low temperatures. At low temperature, a large amount of CO2 is captured inside particles due to a rapid solidification and hence, porous particles are formed. On the other hand, at high temperatures an amount of sensible heat that needs to be removed from the droplet is increased. Moreover, the heat of evaporation is low due to low CO2 solubility. Therefore, the solidification is delayed at higher temperatures. A more time is available at high temperatures in order to retract molten polymer into a spherical shape. 4.4. Conclusions The batch production of particles from PEG of different molecular weights is possible using supercritical CO2 as a processing solvent even below the melting point. Relatively larger particles have been produced for PEG 10000 compared to PEG 6000. The higher the molecular weight the higher is the viscosity of the polymer and hence, the more difficult is the expansion of the high molecular weight polymer melt despite the same CO2 solubility. The decrease in temperature and the nozzle diameter result into smaller particles. This effect is explained by a large nucleation caused due to high CO2 solubility and high depressurisation rate. The effect of pressure on the particle size and particle size distribution has been found absent in all PEG. Not only the size but also the shape and the density of the PEG particles can be controlled in the process


62

studied. Depending on the processing conditions, both hollow and dense particles can be produced that can easily be related to high and low CO2 solubility, respectively. Nearly spherical particles have been produced from PEG 6000 and 10000 as the temperature has been increased above the melting point. The solidification theory is a good tool to explain the roles of temperature and the CO2 solubility in determining the shape and morphology of the particles. Finally, the process studied provides an opportunity of processing various other pharmaceutical and polymeric compounds that are difficult to process by the conventional methods. References [1] Jung J, Perrut M. Particle design using supercritical fluids: Literature and patent

survey. The Journal of Supercritical Fluids 2001;20:179-219. [2] Weidner E, Knez Z, Steiner R. Powder generation from polyethyleneglycols with

compressible fluids. High Pressure Chemical Engineering, Proceedings of process technology 1996;223-228.

[3] Weidner E, Knez Z, Novak Z. Process for the production of particles or powders. (USA). US:6056791, 2000.


[5] Mandel F S, Wang J D, McHugh M A. Pharmaceutical material production via supercritical fluids employing the technique of the particles from gas-saturated solutions (PGSS). Polym. Mater.: Sci. and Eng. 2001;84:39-44.

[6] Liu H, Finn N, Yates M Z. Encapsulation and sustained release of a model drug, indomethacin, using CO2-based microencapsulation. Langmuir 2005;21:379-385.

[7] Pielichowski K, Flejtuch K. Differential scanning calorimetry studies on Poly(ethylene glycol) with different molecular weights for thermal energy storage materials, Polymers for Advanced Technology 2002;13:690-696.

[8] Wiesmet V, Weidner E, Behme S, Sadowski G, Arlt W. Measurement and modelling of high-pressure phase equilibria in the systems polyethyleneglycol (PEG)-propane, PEG-nitrogen and PEG-carbon dioxide. The Journal of Supercritical Fluids 2000;17:1-12.

[9] Kukova E. Phasenverhalten und Transporteigenschaften binärer Systeme aus hochviskosen Polyethylenglykolen und Kohlendioxid, Ruhr-Universität Bochum, Germany, PhD Thesis (2003).

[10] Park C B, Baldwin D F, Suh N P. Effect of the pressure drop on cell nucleation in continuous processing of microcellular polymers. Polymer Engineering and Science 1995;35:432-440.

[11] Ranz W E, Marshall W R. Evaporation from drops: Parts II. Chem. Eng. Prog. 1952;48:173-180.

[12] Bergmann D, Fritsching U, Bauckhage K. A mathematical model for cooling and rapid solidification of molten metal droplets. Int. J. Therm. Sci. 2000;39:53-62.

Chapter 5: Prediction of the viscosity reduction of PPB caused by dissolved CO2 and an elementary approach in the supercritical CO2 assisted continuous particle production

63

Chapter 5

Prediction of the viscosity reduction of PPB due to dissolved CO2 and an elementary approach in the supercritical CO2

assisted continuous particle production The dissolution of CO2 in a polymer causes its plasticization and hence, the viscosity is reduced. A model based on the free volume theory has been used for a polyester resin, which shows a considerable reduction in the viscosity due to dissolved CO2. Therefore, supercritical CO2 has been used as a processing solvent in the continuous production of micron size particles of the resin. Despite the viscosity reduction caused by the dissolved CO2, an excess of CO2 with respect to its solubility limit has been used for micronisation of the polymer due to its high viscosity. As the mixing of CO2 and the polymer has not been possible in an extruder at high gas to polymer mass ratios, a simplified Kenics type static mixer has been used for the mixing purpose. In this study, the effect of various parameters such as temperature, pressure, nozzle diameter, and gas to polymer mass ratio on the particle shape, morphology, and size has been studied. The experimental results manifest the technological as well as theoretical insight into the particles production from a high viscosity material on an elementary level. Sameer P. Nalawade, Vincent H. J. Nieborg, F. Picchioni, L.P.B.M. Janssen, Prediction of the viscosity reduction of PPB due to dissolved CO2 and an elementary approach in the supercritical CO2 assisted continuous particle production, Powder Technology, submitted.


64

5.1. Introduction Polyester resin materials in particle form are generally used the toner and paint industries. The production of small size particles (micrometer to nanometer) on a large scale with a narrow particle size distribution is a major challenge in these industries. Various traditional particle production methods, such as milling or grinding, spray atomization, and crystallization are still used. Heat dissipation during grinding or milling, emission of volatile organic components (VOCs), and separation of particles from a solvent are the major problems associated with the traditional methods. Various new methods to produce micron size particles using supercritical CO2 as a solvent have been developed in the last decade [1-2]. Among them, the PGSS (the particles from gas saturated solution) method has been reported in Chapter 4 for the batch production of particles from poly(ethylene glycol) (PEG) with different molecular weights. Recently, the continuous production of particles using the modified PGSS method has been reported by Kappler et al. for PEG having a molecular weight of 6000 [3]. The viscosity of PEG 6000 has been reported as 0.574 Pa-s at 373 K [4]. Unlike the batch process, a gas in excess to the solubility limit has been used in the continuous process. It had a positive influence in terms of the reduction in the particle size [3].

As a reduction of the viscosity of a polymer due to dissolved CO2 is one of the important parameters in the particle production, considerable attention has been given to experimental as well as to theoretical rheological studies. Various research groups have shown that the viscosity reduction of polymer melts in the presence of dissolved CO2 may be understood in terms of the classical viscosity scaling theory. Gerhardt et al. [5] modified the expression developed by Kelly and Bueche [6] based on Doolittle’s [7-8] free volume theory to calculate scaling factors for different CO2 concentrations in poly(dimethyl siloxane) (PDMS). Kwag et al. [9] used the free volume theory of Gerhardt et al. [5] to obtain a master curve that relates the shear viscosity to the shear rate (η -γ ) for polystyrene(PS)-CO2 system. The predicted scaling factors using the modified free volume theory are in good agreement with experimental scaling factors for both PDMS-CO2 and PS-CO2 systems. Lee et al. [10] also proposed a model using the well known generalized Cross-Carreau equation [11] and Doolittle’s free volume theory to describe the viscosity of PS-CO2 system theoretically. Later, Royer et al. [12-13] developed a free volume model using the Williams-Landel-Ferry (WLF) equation [14] directly obtained from Doolittle’s free volume theory and Chow’s [15] model. The theoretical studies suggest that the free volume mechanism plays an important role in the reduction of polymer melt viscosity by dissolved gas.


65

In this work, we have used the free volume theory for predicting the viscosity reduction of a polyester resin as a function of temperature, pressure, and weight fraction of dissolved CO2 using the physical properties of the components. The second objective of this study is to produce particles from a high viscous propoxylated polyester resin in a continuous manner using supercritical CO2 as a processing solvent. The polyester resin in powder form is known for its application in toner industries. A continuous set up has been designed and constructed for the particle production using a new method. The method applied is a modification of PGSS as CO2 in excess to its solubility limit has been used. A few exploratory experiments have been performed in order to test the method and to have qualitative information over the effect of various processing parameters on the product quality. 5.2. Viscosity reduction theory

Scaling (shifting) of isothermal shear viscosity data as a function of shear rate to a reference temperature is a general practice in polymer rheology. The curve resulting from such scaling is termed as a master curve. A similar concept has successfully been used in combination with a depression in glass transition temperature (Tg) or melting point (Tm) and free volume theory to generate a master curve using pressure and CO2 weight fraction for various polymer-CO2 systems[9,12-13]. The master curve allows rescaling of viscosity data from a reference situation to different processing conditions.

The fact that the viscosity of polymer is a function of free volume has first been suggested by Doolittle [7-8]. The widely used William, Landel, and Ferry (WLF) equation [14] can be obtained from the Doolittle theory.

( )

r

r

Tr

TT TTC

TTCa

−+−−

=⎟⎟⎠

⎞⎜⎜⎝

⎛=

2

1loglogηη 1)

Where Ta and Tη are the shift factor and shear viscosity at temperature T. In case of a reference temperature (Tr) different from the Tg, the WLF parameters, C1 and C2, can be reduced to C1g and C2g using the following equations.

grg TTC

CCC+−

=2

211 2)

grg TTCC +−= 22 3) The WLF equation is generally applicable to amorphous thermoplastic polymers from its gT up to 373+gT K. In order to predict the viscosity reduction due to dissolved CO2, a set of shift factor equations analogous to the WLF equation must be derived to account for the pressure effect and the Tg depression. These


66

can be derived based on the assumption that the fractional free volume at Tg is a material constant independent of CO2 concentration and pressure. The Chow model predicts the depression in the gT of a polymer upon the dissolution of a solvent (plasticizer). The Chow model is based on classical as well as statistical thermodynamics and has been found to be in good agreement with the reported experimental data [15-17].

( ) ( )[ ]θθθθ ln1ln1ln , +−−Ψ=⎥⎥⎦

⎤

⎢⎢⎣

⎡

g

mixg

TT

4)

2

2

1 CO

CO

dil

pol

zMM

ωω

θ−

= 5)

Tgppol CMzR

,Δ=Ψ 6)

In this equation mixgT , is the glass transition temperature of the polymer-diluent mixture, 2COω is the weight fraction of diluent (in this case CO2), z is the lattice coordination parameter for the polymer repeat unit, TgpC ,Δ is the change in heat capacity associated with the gT , and polM and dilM are the molecular monomeric masses of the polymer and the diluent, respectively. To estimate the value for

polM , the monomeric unit masses ( M ) and their molar fractions ( x ) are used.

332211 MxMxMxM pol ++= 7)

Using equations 4)-7), the depression in the gT can be computed and in turn the concentration shift factor, Ca , can be calculated using the following equation in which the subscript “co” refers to the zero concentration of CO2.

( ) ( ) ( )mixgg

mixgg

gg

gg

coT

cTC TTC

TTCTTC

TTCa

,2

,1

2

1

,

,loglog−+

−−

−+

−=⎟

⎟⎠

⎞⎜⎜⎝

⎛=

ηη

8)

The effect of pressure on the viscosity of polymer is opposite to the effect of the concentration of CO2 and temperature. An increase in pressure decreases the free volume of the polymer and increases its Tg. The effect of pressure can be incorporated in to the WLF equation by a shift in the free volume based upon a change in gT . The gT at different pressures, in the absence of CO2, can be been determined from the pressure-volume-temperature (PVT) data of the polymer. Analogy similar to the temperature effect can be used to calculate the pressure effect on the viscosity in terms of pressure shift factor, Pa .


67

( ) ( ) ( )Pgmixg

Pgmixg

Pgmixg

Pgmixg

PT

PTP TTC

TTCTTC

TTCa

,2

,1

0,2

0,1

0,

,loglog−+

−−

−+

−=⎟

⎟⎠

⎞⎜⎜⎝

⎛=

ηη

9)

Finally, the product of the shift factors is used to generate a master curve by multiplying it with shear rates and dividing it with the viscosities obtained at the corresponding shear rates. 5.3. Experimental 5.3.1. Materials Propoxylated polyester resin based on bisphenol (PPB, CAS: 170834-94-5) having a molecular weight around 7000 (Mw-weight average) was supplied by Akzo Nobel, The Netherlands. The glass transition temperature (Tg) of the polymer is around 327 K. The zero shear viscosities of the polymer at different temperatures are provided in Table 5.1. High purity CO2, 99.99 %, used in the experiments was supplied by Hoekloos, The Netherlands. The materials were used without any further purification. Table 5.1. Zero shear viscosities of PPB obtained from the rheological measurements carried out using a cone and plate rheometer

T (K)

η0 (Pa-s)

363 2965.0 383 216.21 403 23.961

5.3.2. Apparatus In the absence of a static mixer, CO2 is added inside an extruder. To inject CO2 at elevated pressures, the knowledge of a pressure profile and different zones in the extruder is essential. Fig. 5.1 shows the different zones and a typical pressure profile applicable to the addition of CO2. The type of extruder used is twin screw counter-rotating (Rollepaal, The Netherlands). The length to diameter ratio (Le/De) of the extruder is 12. The total extruder is divided into five sections including the feed. Each section consists of a barrel surrounding the screws with a heating arrangement. The screws are fully intermeshing i.e. the gap between the core of one screw and a flight of the other screw is very small. A nozzle is attached directly at the end of the extruder. Dividing an extruder in different zones is a normal practice in extrusion. In the low pressure (injection) zone, screw elements having a high pitch value are used. To prevent backflow of CO2 and a polymer a high pressure build up before


68

the injection zone is required, which can be created by a dynamic polymer melt seal (melt plug). There are various ways to create a melt plug using reverse screw elements, low pitch screw elements or a disc with two circular slots. In this study, a disc with circular slots has been used to create a melt plug before the CO2 injection. The disc is fitted between two barrels in such a way that very small gaps are present between the circular slots made in the disc and the core of the screws. The disc is located between the two barrels distance from the feed zone. The first three barrel sections (out of five) are used to feed, transport, melt, and create the melt plug. The fourth and fifth sections are used for the CO2 injection and the mixing, respectively.

Feed zone Melting zone Plug zone Meltseal

Low pressurezone Mixing zone Nozzle

CO2 injection

1st barrel 2nd barrel 3rd barrel 4th barrel 5th barrel

Fig. 5.1. Different zones and a pressure profile in an extruder After the injection zone, a pressure build up is required in order to dissolve the desired amount of CO2 in a polymer. The diameter of the nozzle determines the pressure build up in the last section. In the presence of a static mixer, CO2 is added before the static mixer as shown in Fig. 5.2. A filter and one-way valve are connected before the injection of CO2. The filter is used to prevent choking of a nozzle caused by fine dirt, if any, present in the polymer. The one-way valve prevents a backflow of the polymer towards the extruder after addition of CO2.


69

CO2

Polymer

Hot air in

Hot air out

Frequency controlller

PoTo

F

P1

T1 T2 T3 T4 Td

P2

P3

T5

T6

TMotor

Fig. 5.2. A schematic drawing of the continuous particle production set up containing a static mixer Fig. 5.3 shows a schematic drawing of the CO2 set up used in the continuous production of particles. In the set up, CO2 is supplied from a bottle at 6 MPa. To prevent cavitation inside the CO2 pump (Lewa, USA), CO2 is first cooled down to 273 K in a heat exchanger (a cooler) (Huber, The Netherlands). The mass flow capacity of the pump is 60 kg/hr at 35 MPa. To prevent instabilities in the flow, a buffer cylinder is provided after the pump. After the buffer cylinder, a mass flow meter is installed (Danfoss: Type mass 2100, USA). The desired flow rate is controlled by using a frequency converter connected to the mass flow meter as well as to the pump. A safety relief valve is set at 34 MPa at the exit of the pump. Before CO2 enters into the static mixer, it is heated to a desired temperature using two heat exchangers located after the pump. The temperature can be adjusted from 303 to 473 K. A pressure transducer (Dynisco, USA) is located in a CO2-line after the third heat exchanger. The CO2-line from the heat exchanger to the continuous particle production setup is heated using a heating element (Tmax= 473 K).

F

P

TT

P

T

CO2

Cooler

Membranepump

Buffercylinder

Flowcontroller

Heat exchanger

Fig. 5.3. A schematic drawing of the CO2 set up used in the particle production


70

A simplified Kenics type static mixer is constructed by twisting a RVS-blade of 0.9 mm thickness in our laboratory. The house-constructed and an actual Kenics static mixer are shown Fig. 5.4. The model mixer consists of thirteen elements perpendicular to each other with a total length of 270 mm. The diameter of the mixer is 9.0 mm and a tube around the mixer has a diameter of 9.1 mm. The main difference between the actual Kenics static mixer and the simplified one is that the elements are not alternatively turned left and right in case of the simplified one. Moreover, the length of the mixer is short. Of course, this may affect the mixing efficiency depending on the viscosity of the material.

a) b) Fig. 5.4. a) The constructed static mixer b) the Kenics static mixer with an enlarged view of right and left element arrangement

A complete-nozzle body (Spraying systems, USA) is connected using a male-female coupling at the end of the tube surrounding the mixer. The complete-nozzle body consists of a cap, nozzle, core, ring, and body as shown in Fig. 5.5. A nozzle has a micron size capillary hole of a very short length (1.2 mm). An enlarged view of a core (a slot can be seen at the front side of the core) is also shown separately in Fig. 5.5. A core provides the extra mixing before the nozzle entrance. Different flow behaviours can be expected in the presence and in the absence of a core. The width of the slots in different cores are provided in Table 5.2. The cores will be referred here with their names rather than with the dimensions.

a) b)

Fig. 5.5. a) A complete nozzle body b) an enlarged view of a core


71

Table 5.2. The width of the slots in different cores used in the particle production

Core slot width (mm)

A 0.41 B 0.51 C 0.64

5.3.3. Experimental Method Initially experiments were carried out in the absence of a static mixer using the set up shown in Fig. 5.1, where a nozzle was attached to a die connected to the exit of an extruder. An experimental procedure for a continuous operation was very simple. First, a polymer was fed by a hopper (K-Tron Feeder, The Netherlands) that was calibrated for the mass flow of 1-10 kg/hr. The four sections and the die at the end of the extruder were kept at 363, 368, 373, 378, and 378 K, respectively. This is the minimum temperature profile possible with the present extruder without cooling. It was observed that the selected temperature profile could provide a minimal backflow in the extruder due to high viscosity of the polymer. In the experiments, the temperature of the mixture was varied between 363 to 423 K. A pressure build up in the set up was decided by a mass flow rate of the polymer, a nozzle resistance, and the viscosity of the polymer. The constant pressure at the last section of the extruder indicated a stable flow rate. Then, CO2 was injected initially at a gas to polymer mass ratio smaller than unity using the CO2 pump and the mass controller. In the absence of and presence of the static mixer CO2 was added inside the extruder and after the extruder, respectively. Subsequently, the ratio was increased to higher values until particles were produced. Pressure and CO2 flow rate stabilities in the CO2-line were an indication of a steady condition. The CO2-polymer mixture was continuously expanded into a Perspex drum where the particles were collected.

5.3.4. Particle analysis

A wet laser diffraction (WLD) apparatus, Malvern Mastersizer®, was used for the particle size measurement. During the measurements, a continuous recirculation of a liquid containing a suspension of particles was provided in order to scan the particles for a large number of times (around 2000). A few drops of a surfactant were added to demi-water (solvent) to prevent agglomeration of particles during the measurements. However, the possibility of agglomeration can not be avoided due to a hydrophobic nature of the polymer. The average diameter of the particles (dp,0.5) was determined from the cumulative volume fraction. A scanning electron microscope (SEM) was specifically used to observe the morphology and shape of the particles.


72

5.4. Results and discussion The model based on the free volume theory has been used to calculate the reduction in the shear viscosity of PPB as a function of the weight fraction of the dissolved CO2. Since the residence times of CO2 and polymer are generally very short in a continuous operation, the degree of mixing achieved during the process is very important. There have been two possibilities available for the mixing. The first is to mix inside an extruder in the absence of a static mixer. In the extruder, high shear generated between the screws and the wall and the screws and the core of the screws is used to break up the dispersed (gas) phase. A further break up of these gas bubbles results into enhanced mass transfer of gas into polymer. The second possibility is using a static mixer that is also widely used for mixing materials having different viscosities.

5.4.1. Viscosity prediction results Recently, it has been reported by several research groups that a master curve exist for the experimental rheological data of a polymer-CO2 solution if the effect of temperature, pressure, and CO2 weight fractions is considered together [5, 9-10,12-13]. The theory discussed here has been successfully applied to several polymer-CO2 solutions in order to generate a master curve using the physical properties of the pure components. In this work, it has been assumed that a master curve exists for PPB-CO2 solution at a reference temperature and the data has been rescaled to different processing conditions using the shift factors calculated from the physical properties of the pure components.

In the absence of CO2, a master curve exists for the rheological data obtained at different temperatures for pure PPB melt, Fig. 5.6. This scaling has been used to calculate the WLF parameters. The fractions of the monomeric units in the polymer have been determined by 1H-NMR. The polymer repeat number, z , can be estimated as, the closest integer number, the ratio of the molecular size of the average monomeric unit relative to CO2 [5]. The size of the monomeric units and CO2 have been estimated with the help of a computer program (ACD/3D). Table 5.3 provides the estimated values of Mpol and z for PPB. The ac and ap predicted using the model based on the WLF theory and the Chow model are shown in Fig. 5.7 and 5.8.


73

100

1000

10000

0.00001 0.001 0.1 10 1000

γ*aT, 1/s

η/a

T, Pa

-s

363 K

383 K

403 K

423 K

Fig. 5.6. The master curve generated using the experimental viscosity data (0 % CO2) for PPB

Table 5.3. Estimated polymer repeat unit (z) and average monomer mass (Mpol) for PPB

Polymer z (-)

polM ( g/mol )

PPB 4 226.57

0

0.2

0.4

0.6

0.8

1

0 0.02 0.04 0.06 0.08 0.1

wt fraction of CO2

a c

353 K

373 K383 K

393 K

403 K413 K

423 K

Fig. 5.7. Predicted ac (concentration) shift factors under different isothermal conditions for PPB


74

1

1.5

2

2.5

3

3.5

4

0 10 20 30 40

P, MPa

a p

368 K, 0.01393 K 0.01

413 K, 0.01

368 K, 0.1

393 K, 0.1413 K, 0.1

Fig. 5.8. Predicted ap (pressure) shift factors under different isothermal conditions and CO2 weight fractions for PPB

The predicted results suggest that the ap increases with increasing pressure and ac decreases with increasing CO2 weight fraction. Thus, the theory predicts that the CO2 dissolution decreases the glass transition temperature, while an increase in pressure increases the glass transition temperature. An increment and a reduction in free volume, due to dissolved CO2 and pressure, respectively, are responsible for such effects. The predicted zero-shear viscosities using ap and ac values at different CO2 weight fractions under isothermal conditions are shown in Fig. 5.9. It can be seen from the predicted viscosity curves that a considerable reduction in the zero-shear viscosity takes place with increasing CO2 weight fraction. However, the reduction in the viscosity is smaller at higher CO2 concentration as the pressure effect starts to dominate at higher pressures. Though the theoretical estimated data have not been tested here with experimental data, the estimates favour the possibilities of processing PPB with supercritical CO2.

0.1

10

1000

100000

10000000

0 0.02 0.04 0.06 0.08 0.1

wt fraction of CO2

η, m

Pa-s 368 K

393 K

413 K

Fig. 5.9. PPB-CO2 zero-shear viscosity data rescaled using shift factors from the viscosity measured at 363 K in the absence of CO2


75

5.4.2. Twin screw extruder results The mixing in an extruder is influenced by a number of parameters such as the number of screws, screw configuration, number of mixing elements, residence time, etc. A twin screw extruder is a suitable choice compared to a single screw extruder. In twin screw extruders the mixing efficiency is higher due to more shear regions (between barrel (wall) and screws, and between the screws). In this study, a high viscous material (polymer) has been mixed with a very low viscous fluid (CO2) at elevated pressures. Since the Le/De ratio of the available extruder is small, problems such as a backflow and short residence time are expected. A few experiments have been carried out to test the viability of the extruder setup. During the experiments, the amount of CO2 added has been varied from a low to a high mass flow rate. It has been possible to add CO2 only up to 1 kg/h. The added amount of CO2 has not been sufficient to produce particles. However, micron size fibrous particles along with foam has been produced around the supercritical pressure. At higher CO2 flow rates, a backflow of the polymer and CO2 has been observed due to a large reduction in the viscosity of the polymer-CO2 solution. The set up is suitable for low CO2 flow rates, which are required in case of a microcellular foaming application [18-21]. In further experiments, CO2 has been added after the extruder before the static mixer to increase the flow stability. 5.4.3. Static mixer (Kenics type) results The experimental results obtained using different nozzle diameters in the presence of the static mixer are given in Table 5.4, 5.5 and 5.6. From the Tables, it can be seen that it is not possible to produce particles for all experimental conditions. A principal component analysis, a descriptive statistical technique, has been used to differentiate products as a function of various parameters. Also, the effect of various parameters such as nozzle diameter, core-slot width, gas to polymer mass ratio (GTP), temperature (T), and pressure (P) on the product shape and size have been discussed. Principal component analysis Principle component analysis (PCA) [22] allows clustering of data in different groups using multiple variables. The analysis uses the combination of standardized dimensionless variables to produce indices that are uncorrelated. The indices are called principal components and are linear function of variables. The standardized dimensionless variables are obtained from the original data using the variables mean values and the standard deviations. The comprehensive details of PCA with different examples has been given elsewhere [23]. At the end of the analysis, two principal components with larger variances are selected


76

Table 5.4. Experimental results obtained with the 0.40 mm diameter nozzle

Exp. mP (kg/hr)

GTP

(P

CO

mm

2 )

Core

T (K)

P (MPa)

Product

1 2.00 3.00 C 393 18.40 Foam

2 2.97 1.00 C 404 14.55 Foam 3 4.32 3.00 C 405 17.40 Agg. fibers 4 4.32 4.00 C 404 19.20 Agg. fibers

5 4.32 0.95 C 412 13.50 Foam 6 4.32 1.40 C 413 13.60 Foam 7 4.32 2.00 C 413 14.40 Foam 8 4.32 3.00 C 413 16.50 Agg. fibers

9 3.80 0.33 C 422 12.30 Foam 10 3.80 1.00 C 421 12.80 Foam 11 5.00 1.00 C 421 12.75 Foam 12 5.00 1.50 C 424 12.95 Foam 13 3.80 1.60 C 422 14.10 Foam 14 5.00 2.00 C 423 14.75 Foam + agg. fibers 15 3.80 2.10 C 422 13.60 Foam + agg. fibers 16 5.00 3.00 C 423 17.75 Less agg. fibers

17 3.6 3.00 A 423 16.20 Less agg. fibers


77



Exp. mP (kg/hr)

GTP

(P

CO

mm

2 )

Core

T (K)

P (MPa)

5.0,Pd ( mμ )

Product

30 4.75 4.21 - 373 10.7 - Foam

31 6.38 2.00 - 392 8.35 - Foam + particles 32 6.38 3.00 - 392 9.40 - Foam + particles

33 4.75 3.00 B 395 9.55 137.4 Particles

34 4.75 3.00 B 399 9.30 157.9 Particles 35 6.38 3.00 B 398 11.65 133.5 Particles 36 4.75 3.00 B 404 9.20 150.9 Particles 37 6.38 3.50 B 402 11.85 128.8 Particles

Exp. mP (kg/hr)

GTP

(P

CO

mm

2 )

Core

T (K)

P (MPa)

Product

18 4.75 1.40 - 395 12.65 foam 19 4.75 2.10 - 394 13.05 foam 20 4.75 2.80 - 393 13.90 foam with fibers

21 3.33 4.00 - 404 11.25 Less agg. fibers 22 4.75 3.00 - 406 11.95 Less agg. fibers 23 4.75 4.00 - 405 16.45 Less agg. fibers

24 6.38 3.75 - 414 14.55 Less agg. fibers 25 6.38 3.00 - 416 12.40 Less agg. fibers

26 4.75 3.00 B 397 12.50 Less agg. fibers 27 4.75 4.00 B 398 14.75 Less agg. fibers

28 4.75 3.00 B 403 12.35 Less agg. fibers

29 3.33 3.00 C 421 7.60 Less agg. fibers


78

to represent the dependency of an output on the variables. The important thing in this analysis is that a large number of variables are represented by dimensionless indices, the principal components. Moreover, it uses correlativity, positive or negative, between the variables. In this study, the PCA has been used to describe the possibility of foam or fibers or particles production. Cs, Ts, GTPs, Ps, dns are the standardized dimensionless variables of core-slot width, temperature, gas to polymer mass ratio, pressure, and nozzle diameter, respectively used in this analysis. Two principal components as a function of standardized variables are given in equation 10) and 11). Z1=-0.491*Cs-0.534*Ts+0.418*GTPs -0.187*Ps+0.514*dns 10) Z2=0.084*Cs+0.041*Ts+0.0506*GTPs+0.857*Ps+0.024*dns 11) Z1 and Z2 determine the morphology of the final product. Using the Z1 and Z2, a possible grouping of different morphologies is shown in Fig. 5.10. It is clear from Fig. 5.10 that the particles are mostly produced for the positive values of Z1 (> 2) and Z2 (> 0) and the foam is mostly produced for the negative values of Z1 (< 0) and the positive values Z2 (> 0). The fibers (agglomerated fibers) are present between the foam and the particles. These results are sufficient to preliminarily choose conditions depending on the desired product.

-1.5

-1

-0.5

0

0.5

1

1.5

-3 -2 -1 0 1 2 3

Z1

Z2

fibers

foamparticles

Fig. 5.10. Plot of the PCA results in the absence of a core, particles, fibers and foam, for the two principal components, Z1 and Z2 Elementary study of various parameters affecting the product quality Table 5.4, 5.5, and 5.6 clearly show the dependence of product quality on the core and nozzle diameter. In the presence of a core, the quantity of foam produced along with particles is reduced compared to similar conditions in the absence of a core. The slots present in the core, causes splitting of the stream before the nozzle entrance, provide an extra homogenization of excess CO2 and


79

a CO2-saturated polymer solution. This results into a better expansion. However, a small amount of foam has been present in the product. The product has always been screened through a 900 mμ size sieve.

In PGSS, a single-phase solution of a polymer and CO2 is expanded over a nozzle in order to produce particles. Such expansion is dependent on the shear viscosity of the polymer solution. In case of high molecular weight polymers both shear and extensional viscosity inhibit the break up of the solution. Since the shear viscosity of PPB is very high, an excess of CO2 is necessary in order to break up the polymer melt. For PPB, it is possible to have a continuous expansion at a gas to polymer mass ratio equal to or higher than 3 only. Since an excess of CO2, unlike PGSS, is used in this method for the particle production, the method has been termed as expansion of gas-saturated solution with excess gas (EGSEG). The effect of increasing gas to polymer mass ratio (GTP) on the flow behaviour is shown schematically in Fig. 5.11. Increasing the ratio, a transition from one flow regime to another takes place in the nozzle. From Table 5.4, 5.5, and 5.6 it can be seen that at smaller ratios (<3) mostly foam is present while the particles are produced as the ratio is increased. At higher ratios, the expansion of excess CO2 causes intense instabilities at the surface of the polymer melt present in the form of a thin film and hence, the break up of the polymer melt is enhanced. However, it should be noted that the foam is produced even at higher ratios if the temperature and pressure are low, for example experiments 1 and 30.

Fig. 5.11. Different flow regimes in a nozzle with the increasing gas to polymer mass ratio

A considerable effect on the product quality is also observed in terms of the nozzle diameter. For example, the scanning electron microscope results of experiments 3, 28, 33, and 36 are shown in Fig. 5.12. As the nozzle diameter is increased, the product is changed from agglomerated fibers to irregular shaped particles. The results can be explained considering the extensional viscosity (viscoelasticity) of the polymer, which also plays an important role in the breakup of a polymer solution [24-25]. A schematic drawing of the elongation of polymer molecules is shown in Fig. 5.13. Mansour and Chigier [24] have concluded from their study that a very large stretching motion takes place before the break up in case of viscoelstaic materials compared to visco-inelastic


80

materials. Before the break up of a viscoelastic material, long threads are always formed along with droplets and thus, the breakup is inhibited. This is mainly caused by the molecular orientation under an extensional effect.

a) b)

c) d)

Fig. 5.12. Scanning electron microscope pictures of the product obtained using different diameter nozzles a) 0.4 mm (exp.3) b) 0.57 mm (exp.28) c) 0.81 mm (exp.36) d) 0.81 mm (exp.33)

Fig. 5.13. Schematic drawings of the elongation of polymer molecules in a small diameter nozzle

Before entering a nozzle Extensional and shear effect


81

Polymer molecules, which are long chains intermingled into each other, entering a nozzle are subjected to elongation if the flow converges from a large to small cross section. The smaller the diameter of the nozzle the higher is the extensional effect experienced by the polymer molecules. Moreover, a high shear experienced by the polymer molecules inside a small diameter nozzle keeps the molecules align. Such elongation prevents the break up of a polymer solution into particles even at elevated pressures and hence, agglomerated fibers are formed. Not only extensional viscosity but shear viscosity also plays an important role in the particle production. The lower the shear viscosity of a polymer the easier is the break up of the polymer melt into particles. It has already been mentioned in the text that it is not possible to produce particles at low temperatures even with high gas to polymer mass ratios with a bigger diameter nozzle where the extensional effect is a less pronounced. It clearly indicates that shear viscosity must also be reduced sufficiently in order to produce particles. The viscosity model predicts a considerable reduction in the viscosity in the presence of dissolved CO2. The higher the temperature the lower is the shear and extensional viscosity. Thus, it is easier to break up a polymer melt into particles at high temperatures. Less agglomerated product is obtained at a high temperature for 0.4 mm diameter nozzle as shown in Fig. 5.14. It is also expected that at higher temperatures the mixing efficiency increases because of a low shear viscosity, not only caused by an increase in temperature but also by the dissolved CO2, which results into a better expansion. In case of 0.4 and 0.57 mm diameter nozzle, it is difficult to check the effect of pressure and temperature on the particle size due to an agglomerated fibrous product. The agglomerated fibers are hollow in a longitudinal direction, which also confirms the extensional effect. Due to an elongation of polymer molecules it is difficult for CO2 to escape out from the surface of the melt and consequently, a longitudinal path is preferred by the CO2 for its diffusion. The fibrous product is easily crushable to very fine powder even with a very low shear. The low-density product is mainly due to the CO2 captured inside the fibers during solidification.


82

Fig. 5.14. Scanning electron microscope picture of the product obtained using the 0.4 mm diameter nozzle (experiment 16) Relatively smaller differences are observed in terms of the particle diameter affected by temperature or pressure for a nozzle with a diameter of 0.81 mm, Table 5.6. The particle size distributions obtained using the traditional grinding and the EGSEG method under different conditions are shown in Fig. 5.15. No considerable effect of the temperature and the pressure are present. The key advantage of the new method is the single step process, whereas the traditional method requires a number of recycles. The pressure and the mixing inside the static mixer are the limiting factors of the set up studied. Better results, i.e. a further reduction in the particle size and a narrower particle size distribution, are expected at higher pressures. This study gives a good insight in the process, which is applicable to high viscosity materials.

0

1

2

3

4

5

6

7

0 200 400 600 800 1000

dp,0.5, μm

ln(q

)

grinding

exp 33

exp 34

exp 35

exp 36

exp 37

Fig. 5.15. Particle size distributions obtained using 0.81 mm nozzle diameter under different processing conditions and grinding method


83

5.5. Conclusions A model based on the free volume theory clearly indicates the plasticization of the polymer due to dissolved CO2. The continuous micron size particles production from the high viscous polymer using CO2 as a supercritical fluid is possible by modifying PGSS. Unlike PGSS, CO2 in excess to its solubility are essential in the break up of the high viscous polymer solution. A principle component analysis predicts different regions for formation of foam, fibers, and particles. The extensional and shear viscosity, solubility, nozzle diameter, pressure, and temperature play important roles in the break up of polymer solution. No effects of pressure and temperature on the particle size and particle size distribution have been observed. A few modifications are required in the set up to enlarge the processing window. References [1] Jung J, Perrut M. Particle design using supercritical fluids: Literature and patent.

Journal of Supercritical Fluids 2001;20(3):179-219. [2] Yeo S-D, Kiran E. Formation of polymer particles with supercritical fluids: A

review. Journal of Supercritical Fluids 2005;34(3):287-308. [3] P. Kappler, W. Leiner, M. Petermann, E. Weidner, Size and morphology of

particles generated by spraying polymer-melts with carbon dioxide, Proceedings of 6th ISSF, materials processing, vol. 3, 2003, p1891-1896.

[4] E. Kukova, M. Petermann, E. Weidner, Phase behaviour (S-L-G) and fluid dynamic properties of high viscous poly (ethylene glycols)s in the presence of compressed carbon dioxide, Proceedings of 6th ISSF, materials processing, vol. 3, 2003, p1547-1552.

[5] Gerhardt L J, Garg A, Manke C W, Gulari E. Concentration-dependent viscoelastic scaling models for polydimethysiloxane melts with dissolved carbon dioxide. Journal of Polymer Science: Part B: Polymer Physics 1998;36:1911-1918.

[6] Kelley F N, Bueche F. Viscosity and glass temperature relations for polymer-diluent systems. Journal of Polymer Science 1961;L:549-556.

[7] Doolittle A K. Studies in Non-newtonian flow. I. The dependence of the viscosity of liquids on free-space. Journal of Applied Physics 1951;22:9-26.

[8] Doolittle A K. Studies in Newtonian flow. II. The dependence of the viscosity of liquids on free-space. J. Non-Newt. Fluid Mech. 1951;22:1471-1475.

[9] Kwag C, Manke C W, Gulari E. Effects of dissolved gas on viscoleastic scaling and glass transition temperature of polystyrene melts. Industrial Engineering Chemistry and Research 2001;40:3048-3052.

[10] Lee M, Park C B, Tzoganakis C. Extrusion of PE/PS blends with supercritical carbon dioxide. Polymer Engineering and Science 1998;38:1112-1120.

[11] Hieber C A, Chiang H H. Shear-rate-dependence modeling of polymer melt viscosity. Polymer Engineering and Science 1992;32:931-938.

[12] Royer J R, Gay Y J, DeSimone J M, Khan S A. High-pressure rheology of polystyrene melts plasticized with CO2: Experimental measurement and predictive scaling relationship. Journal of Polymer Science: Part B: Polymer Physics 2000;38:3168-3180.


84

[13] Royer J R, Gay Y J, DeSimone J M, Khan S A. High-pressure rheology and viscoelastic scaling predictions of polymer melts containing liquid and supercritical carbon dioxide. Journal of Polymer Science: Part B: Polymer Physics 2001;39:3055-3066.

[14] Williams M L, Landel R F, Ferry J D. The temperature dependence of relaxation mechanisms in amorphous polymers and other glass-forming liquids. The Journal of American Chemical Society 1955;77:3701-3707.

[15] Chow T S. Molecular interpretation of the glass transition temperature of polymer-diluent systems. Macromolecules 1980;13:362-364.

[16] Wissinger R G, Paulaitis M E. Swelling and sorption in polymer-carbon dioxide mixtures at elevated pressures. J. Polym. Sci. Polym. Phys. 1987;25(12):2497-2510.

[17] Chiou J S, Barlow J W, Paul D R. Polymer crystallization induced by sorption of carbon dioxide gas. J. Appl. Polym. Sci. 1985;30(9):3911-24.

[18] Park C B, Baldwin D F, Suh N P. Effect of the pressure drop on cell nucleation in continuous processing of microcellular polymers. Poly. Eng. Sci. 1995;35:432-440.

[19] Park C B, Suh N P. Rapid polymer/gas solution formation for continuous production of microcellaur plastics. J. Manuf. Sci. Eng. 118 (1996) 639-645.

[20] Siripurapu S, Gay Y J, Royer J R, DeSimone J M, Spontak R J, Khan S A. Generation of microcellular foams of PVDF and its blends using supercritical carbon dioxide in a continuous process. Polymer 2002;43:5511-5520.

[21] Han X, Zeng C, Lee J, Koelling K W, Tomasko D L. Extrusion of polystyrene nanocomposite foams with supercritical CO2 using supercritical carbon dioxide in a continuous process. Polym. Eng. Sci. 2003:43:1261-1275.

[22] Pearson K. On lines and planes of closest fit to a system of points in space. Philos. Magaz. 1901;2:557-572.

[23] Manly B F J. Multivariate statistical methods. 2nd ed., Chapman and Hall, London, UK, 1998.

[24] Mansour A, Chigier N. Air-blast atomization of non-Newtonian liquids. J. Non-Newt. Fluid Mech. 1995;58:161-164.

[25] Christanti Y, Walker L M. Surface tension driven jet break-up of strain-hardening polymer solutions. J. Non-Newt. Fluid Mech. 2001;100:9-26.

Chapter 6: An engineering study of supercritical CO2 assisted continuous polymer micron-size particles production using an SMX mixer

85

Chapter 6

An engineering study of supercritical CO2 assisted continuous polymer micron-size particles production using an SMX mixer

A continuous production of micron-size particles from a high viscous polyester resin melt is possible using supercritical CO2 as a solvent. CO2 in excess to its solubility limit has been used for micronisation of the polymer due to its high viscosity. Effects of various parameters such as temperature, pressure, nozzle diameter, core-slot width, and gas to polymer mass ratio on the particle shape, morphology, and size have been studied in detail. Foam, agglomerated fibers, and particles have been produced by tuning the process conditions. A solidification model and dimensional analysis have been applied to study the shape of the particles and to determine the contribution of different forces in the particle production, respectively. Sameer P. Nalawade, F. Picchioni, L.P.B.M. Janssen, An engineering study of supercritical CO2 assisted continuous polymer micron-size particles production using an SMX mixer, Journal of Supercritical Fluids, submitted.


86

6.1. Introduction The higher the dissolved amount of CO2 in a polymer the higher is the viscosity reduction of the polymer. This allows processing of high molecular weight materials without using volatile organic components and other solvents hazardous to environment. Unlike in a batch particle production process, the residence times of CO2 and a polymer in a continuous process are very short. Therefore, the contact time of components becomes very important in continuous processes. In that situation, uniform mixing is a viable solution in order to achieve equilibrium composition of CO2 and the polymer in a short time. It has already been found in the exploratory work, Chapter 5, that a static mixer is a good choice for mixing CO2 and a molten polymer. SMX type mixers have been used for mixing CO2 with different polymers by various research groups in microcellular foaming and particle production applications [1-10]. Despite the use of supercritical CO2 as a plasticizer or a solvent for various applications in the last few decades, less attention has been given to continuous particles production. Detailed engineering study of a continuous particle production from high viscous polymer melts is still lacking. The previous experimental work, Chapter 4 and 5, gives a clear indication of the supercritical CO2 assisted process’s applicability to polymers having viscosity in a range of 0.1-2500 Pa-s. In the continuous process studied previously for a polyester resin, it has not been possible to portray a complete picture of the continuous particle production over a wide processing range due to practical limitations. For this purpose, a modification of the set up used in the exploratory work is needed. In this study, the set up used in the exploratory work has been modified by using a gear pump and an SMX static mixer. Various parameters affecting the morphology, shape, and size of the particles of a polyester resin such as temperature, pressure, gas to polymer mass ratio, nozzle diameter, and core-slot width have been studied in detail. To differentiate the product quality in terms of foam and fibers/particles depending on various processing parameters a statistical method, principle component analysis, has been used. A solidification model has been applied in order to relate the shape and the morphology of particles to pressure and temperature. Moreover, a dimensional analysis has been used to find the relation between various forces and the particle size for the particles produced in the absence of core.


87

6.2. Experimental 6.2.1. Modified set up A schematic drawing of the set up used in a continuous particle production is shown in Fig. 6.1. Apart from a few modifications, the set up is similar to the one used in case of the Kenics type mixer. The modifications mainly involve a gear pump (MAAG, USA) and an SMX mixer (Sulzertech, Switzerland) to overcome practical limitations of the old set up. In the old set up, it has not been possible to run the extruder at high polymer feed rates due to a high pressure build up, which causes a high backflow of material. Therefore, the gear pump was implemented between the extruder and the static mixer. The pump can work up to 35 MPa with a maximum pressure difference of 25 MPa for a material having viscosity as low as 5 Pa-s. This construction allows keeping a low pressure on the extruder side even at high polymer feed rates and a high pressure at the downstream of the pump. A CO2 injection tube is located after the pump just before the first element of the mixer.

MotorPo

CO2Polymer

To

F

Motor

P1

T1 T2 T3 T4Td Tp

P2

P3

T5

P4

T6

P2 set

T Heatingelement

Fig. 6.1. A schematic drawing of the modified continuous particle production set up

The static mixer used in this study is shown in Fig. 6.2. It consists of an array of 18 similar stationary elements placed behind each other inside a pipe having a diameter of 22.5 mm. The diameter of the elements is 22 mm. In the mixer the flow inside the mixing elements is split up in multiple streams where every element is rotated by 90° relative to its previous element. Consequently, the mixing occurs through continuous redirecting, splitting, and stretching of the fluids as they pass through available openings. This type of mixer is known for its excellent mixing and high dispersion effect with a narrow residence time distribution. In the SMX mixer, the striation thickness is reduced due to frequent splitting and reorientation, which in turn reduces the diffusion distance for CO2


88

considerably. The striation thickness is defined as an average distance between interfaces of two components in the mixture. Because of a smaller striation thickness, better mixing can be expected in an SMX mixer than in a Kenics type mixer. Pressure and temperature sensors are mounted at the beginning as well as at the end of the pipe.

Fig. 6.2. The SMX static Mixer used in the continuous particle production process

In the old setup, a temperature has been controlled using a hot gun where hot air has been circulated through a tube surrounding the tube of the Kenics type mixer. Although it is possible to achieve a desired temperature, the time required to achieve a constant value is long and increased at high flow rates of polymer and CO2. Therefore, electrical heating elements are provided in the new set up, which allows a better control over a temperature within a short time. Different cores and nozzles used in this study have already been reported in Chapter 5.

6.2.2. Experimental method Except a few gear pump adjustments, an experimental procedure similar to the exploratory work was adopted in this study. First, a polymer was fed at a particular flow rate to the extruder. A low pressure was possible at the end of the extruder even at high flow rates using the gear pump. The pump was always kept at the lowest possible temperature because a basic working principle of the gear pump is that the higher the viscosity of polymer melt the higher is the pressure difference possible over the pump. The rest of the set up after the pump was kept at a desired processing temperature using temperature controlled heating elements. Then, CO2 was injected into the mixer at a desired gas to polymer mass ratio. Stable pressures and CO2 flow rate were the indications of a steady condition. After the nozzle the particles were collected in a drum.


89

6.2.3. Particle analysis A wet laser diffraction (WLD) apparatus, Malvern Mastersizer®, was used for the particle size measurement. Better results are obtained when the particles are spherical. During measurements, a continuous recirculation of a liquid containing a suspension of particles is provided in order to scan the particles for a large number of times . A few drops of a surfactant (a commercial soap) were added to demi-water (solvent) to prevent agglomeration of particles during the measurements. However, the possibility of agglomeration can not be avoided due to a hydrophobic nature of the polymer. The average diameter of the particles (dp,0.5) was determined from the cumulative volume fraction. A scanning electron microscope (SEM) was used to observe the morphology and shape of the particles. 6.3. Results and discussion The results obtained using different diameter nozzles in the absence and in the presence of a core are listed in Table 6.1 and 6.2, respectively. The CO2 dissolved in the polymer is not sufficient to produce the particles due to its very high viscosity. An excess of CO2 is required in order to break up the polymer solution into particles. As discussed in Chapter 5, a flow pattern inside a nozzle ultimately determines the product quality. Unlike with the Kenics type mixer, a relatively low quantity of a byproduct (foam) along with the particles has been obtained in the absence of a core using the SMX mixer at nearly same conditions. The expansion product has always been screened through a 900 mμ size mesh. With the SMX mixer, this byproduct is not present if the nozzle is equipped with a core depending on processing conditions. Therefore, it can be conjectured that the mixing efficiency of the SMX mixer is better tha/n that of the Kenics type mixer. 6.3.1. Principle component analysis It has been found that the production of particles or agglomerated fibers using different diameter nozzles has not been possible under the same processing conditions. The principle component analysis (PCA) method, a data clustering tool, has been used to test the effect of various processing parameters on the morphology of the product. Two principal components as a function of standardized variables are given in equation 1)-2) and 3)-4) for different nozzles in the absence and presence of a core, respectively. The results for different diameter nozzles in the absence and presence of a core are shown in Fig. 6.3. The detailed discussion of the PCA method has been given in Chapter 5.


90

Table 6.1. The operating conditions and the results obtained using different nozzles in the absence of core Exp

dn

(mm) mP

(kg/hr) GTP

(mCO2/mP) T

(K) P

(MPa) dp,0.5

( μ m) Product

1 0.4 3.40 10 373 24.30 - long fibers with particles 2 0.4 6.38 2 379 16.30 - foam 3 0.4 3.33 1 381 10.25 - foam 4 0.4 3.33 2 381 12.35 - foam 5 0.4 3.33 3 381 14.15 - foam 6 0.4 6.38 1 381 13.25 - foam 7 0.4 3.40 10 380 22.2 85.93 particles 8 0.4 6.38 4 381 21.95 104.10 particles 9 0.4 6.38 5 381 24.95 92.99 particles

10 0.4 6.38 3 382 20.00 117.90 particles 11 0.4 7.11 4 390 23.20 116.47 particles 12 0.4 4.89 3 393 18.50 216.35 particles 13 0.4 4.59 5 394 19.50 157.91 particles 14 0.4 6.38 4 394 21.00 134.75 particles 15 0.4 3.49 7 392 25.80 124.93 particles 16 0.4 4.89 3 397 15.90 - foam 17 0.4 4.89 1 402 10.05 - foam 18 0.4 7.86 1 403 13.85 - foam 19 0.4 6.38 2 407 15.05 - foam 20 0.4 6.38 3 403 18.30 280.38 particles 21 0.4 6.38 5 401 24.30 161.15 particles 22 0.4 5.04 7 409 24.30 231.73 particles

23 0.57 3.33 1 362 8.85 - foam 24 0.57 6.38 2 359 11.15 - foam 25 0.57 3.33 2 362 6.95 - foam 26 0.57 8.21 2 368 11.10 - foam 27 0.57 3.33 3 356 6.95 - foam with particles 28 0.57 3.33 3 362 7.85 - foam with particles 29 0.57 6.38 3 366 11.60 - foam with particles 30 0.57 8.21 3 367 13.90 - foam with particles 31 0.57 7.27 3 368 13.15 - foam with particles 32 0.57 6.38 3 371 12.20 - foam with particles 33 0.57 6.38 3 371 12.15 - foam with particles 34 0.57 6.38 4 369 14.70 116.23 particles 35 0.57 6.38 1 381 9.90 - foam 36 0.57 6.38 2 380 10.55 - foam 37 0.57 6.38 3 380 8.20 - foam with particles 38 0.57 3.33 0.9 386 5.40 - foam


91

Exp

dn

(mm) mP

(kg/hr) GTP

(mCO2/mP) T

(K) P

(MPa) dp,0.5

( μ m) Product

39 0.57 3.33 1 386 10.30 - foam 40 0.57 3.33 2 389 7.05 - foam 41 0.57 7.23 3 382 12.35 - foam with particles 42 0.57 5.07 3 385 9.55 - foam with particles 43 0.57 3.33 3 388 7.35 - foam with particles 44 0.57 6.73 3 391 11.45 - foam with particles 45 0.57 8.46 3 394 14.35 118.78 particles 46 0.57 8.46 5 394 14.65 108.94 particles 47 0.57 4.59 10 392 14.70 93.54 particles 48 0.57 8.46 5 398 14.60 137.76 particles 49 0.57 8.46 3 404 13.45 153.10 particles 50 0.57 8.01 7 412 21.40 199.78 particles

51 0.81 5.63 10 374 11.80 - fibers with particles 52 0.81 7.86 7 391 11.90 73.62 particles 53 0.81 9.04 5 395 10.50 115.44 particles


92

Table 6.2. The operating conditions and the results obtained using different nozzles in the presence of core

Exp

dn (mm)

mP (kg/hr)

GTP (mCO2/mP)

Core -

T (K)

P (MPa)

dp,0.5 ( μ m)

Product

54 0.4 6.38 3 B 378 20.60 - aggl. fibers with particles55 0.4 6.38 3 B 383 19.80 - aggl. fibers with particles56 0.4 6.38 3 B 395 19.30 - aggl. fibers with particles57 0.4 6.38 3 C 393 19.95 - aggl. fibers with particles58 0.4 7.86 3 C 395 22.95 - aggl. fibers with particles59 0.4 7.86 4 C 394 23.30 - aggl. fibers with particles60 0.4 7.11 4 B 403 27.30 - aggl. fibers with particles61 0.4 7.86 3 C 404 23.50 - aggl. fibers with particles62 0.4 6.38 3 C 404 19.20 - aggl. fibers with particles63 0.4 7.11 4 C 404 26.10 - aggl. fibers with particles64 0.4 5.04 10 C 410 22.80 78.45 particles (fibrous) 65 0.4 8.45 3 B 413 18.40 150.10 particles (fibrous) 66 0.4 5.04 10 B 413 21.50 110.64 particles (fibrous) 67 0.4 8.45 3 C 413 23.40 127.10 particles (fibrous) 68 0.4 8.45 3 C 413 22.80 121.84 particles (fibrous)

69 0.57 6.38 1 C 368 11.40 - foam 70 0.57 6.38 2 C 363 12.55 - foam 71 0.57 6.38 3 C 362 13.15 - Foam with particles 72 0.57 6.38 2 C 374 9.95 - foam 73 0.57 6.38 2 C 376 9.30 - foam 74 0.57 6.38 2 C 377 10.00 - foam 75 0.57 6.38 3 B 377 14.21 - aggl. fibers with particles76 0.57 8.16 4 B 395 17.30 - aggl. fibers with particles77 0.57 4.59 10 B 394 17.20 - aggl. fibers with particles78 0.57 4.59 10 C 394 15.30 - aggl. fibers with particles79 0.57 8.60 3 B 401 16.60 155.01 particles (fibrous) 80 0.57 4.59 10 B 399 16.65 141.46 particles (fibrous) 81 0.57 5.34 10 B 405 19.80 85.47 particles (fibrous) 82 0.57 8.16 3 B 405 15.10 137.56 particles (fibrous) 83 0.57 8.45 3 B 415 12.30 155.81 particles 84 0.57 5.34 10 C 413 16.60 76.32 particles 85 0.57 8.46 3 C 418 11.00 140.45 particles

86 0.81 9.35 3 B 375 13.10 - aggl. fibers with particles87 0.81 9.35 3 C 378 10.10 - aggl. fibers with particles88 0.81 9.35 3 A 378 15.20 - aggl. fibers with particles89 0.81 9.35 3 B 388 11.70 - aggl. fibers with particles90 0.81 9.35 5 B 387 15.80 - aggl. fibers with particles


93

Exp

dn (mm)

mP (kg/hr)

GTP (mCO2/mP)

Core -

T (K)

P (MPa)

dp,0.5 ( μ m)

Product

91 0.81 9.35 2 B 390 9.80 - foam 92 0.81 9.35 3 B 392 11.70 116.92 particles (fibrous) 93 0.81 9.35 1 B 394 7.75 - foam 94 0.81 9.35 3 A 394 14.20 - aggl. fibers with particles 95 0.81 7.86 7 C 394 12.80 84.12 particles (fibrous) 96 0.81 9.35 2 A 395 12.00 - foam 97 0.81 9.35 3 B 403 11.40 127.50 particles 98 0.81 7.86 7 B 403 18.40 110.21 particles 99 0.81 5.48 10 B 401 18.80 114.80 particles

100 0.81 7.86 7 C 406 13.50 75.20 particles 101 0.81 9.35 1 B 406 6.80 - foam 102 0.81 9.35 2 B 407 9.10 - foam 103 0.81 7.86 7 A 412 25.65 79.89 particles (fibrous) 104 0.81 9.04 5 B 413 16.40 117.80 particles 105 0.81 5.48 10 C 412 13.60 64.27 particles 106 0.81 5.48 10 B 414 18.30 80.04 particles 107 0.81 9.35 3 A 416 13.20 123.83 particles 108 0.81 5.48 10 A 414 25.10 83.03 particles 109 0.81 9.35 3 B 419 11.00 156.67 particles


94

Z1=-0.033*Ts+0.739*GTPs+0.026*Ps +0.673*dns 1) Z2=-0.434*Ts-0.401*GTPs -0.673*Ps+0.445*dns 2) Z1=-0.152*Cs-0.389*Ts-0.434*GTPs-0.648*Ps+0.458*dns 3) Z2=-0.589*Cs+0.457*Ts+0.385*GTPs-0.015*Ps+0.544*dns 4)

-4

-3

-2

-1

0

1

2

3

-4 -2 0 2 4

Z1

Z2

ParticlesFoam

Fibers

a)

-3

-2

-1

0

1

2

3

4

-3 -2 -1 0 1 2 3

Z1

Z2

Particles

Foam

Fibers

b)

Fig. 6.3. Plots of the results obtained using the PCA method for the two principal components, Z1 and Z2 a) core is absent b) core is present It is clear from Fig. 6.3 that foam are mostly produced for the Z1<0 and Z2>0 in the absence of a core. In the presence of a core foam is present for the positive values of Z1 (>1). Though a clear transition can not be observed between particles and agglomerated fibers, distinct regions in which only foam is produced can be clearly seen. Accountability of multivariable dependency of the product quality on several processing parameters is the major advantage of the PCA method.


95

6.3.2. Vital roles of processing parameters in the particle production As it has not been possible to perform the experiments under the same conditions for all the nozzles, a simple data-fitting model (equation 5)) has been used to present the particle size data in terms of different processing parameters. The model predicts the average particle diameter (dp,0.5) as a function of temperature (T), pressure(P), nozzle diameter (dn), and gas to polymer mass ratio (GTP). dp,0.5 = 93+ exp(a+b*T+d*GTP+e*P+m*dn) 5) Where, a, b, d, e, and m are the fitting parameters. The value of a, b, d, e, and m are 28.712, 0.0934, -0.0803, -0.0666, and -0.007, respectively. A non-linear least-square regression procedure (MatLab 7) has been used in order to predict the average particle diameter. A procedure involves a minimization of the sum of the deviations between predicted and experimental data. The model fits to the experimental data with an average relative deviation of ~ 9 %, Fig. 6.4.

50

100

150

200

250

300

50 100 150 200 250 300dp,0.5 exp., μm

d p,0

.5 p

red.

, μm

Fig. 6.4. The fitting of the experimental data obtained in the absence of the cores (for 0.4 and 0.57 mm nozzle) to the proposed model Three dimensional curves, the average particle diameter as a function of two other variables, can be obtained using this model. This model has not been applied to the data obtained using cores. The amount of data is not sufficient as the particles are produced only at high temperatures. Some of the results predicted using the model for 0.41 mm and 0.57 mm diameter nozzles in an absence of a core are shown in Fig. 6.5. Effect of nozzle diameter and core-slot width From Fig. 6.5a-b and Table 6.2 it is clear that the particle size decreases with increasing nozzle diameter in the presence as well as in the absence of a core.


96

With decreasing nozzle diameter, one would expect that the break up of a polymer melt should be possible at low pressures as the shear viscosity is decreased and hence, the diameter of the molten polymer film inside the nozzle is reduced. Nevertheless, opposite results have been obtained due to an increase in extensional (elongational) viscosity that resists the break up of the polymer melt. High pressures are required for smaller diameter nozzles under same isothermal conditions. Moreover, a large foaming in particles due to more amount of CO2 dissolved at high pressures may also be responsible for bigger particles.

50100150200250300350400450500

0.08 0.085 0.09 0.095 0.1 0.105 0.11

CO2 wt. fratcion

d p,0

.5, μ

m

3 GTP, 17.5 MPa5 GTP, 17.5 MPa

10 GTP, 17.5 MPa

50

150

250

350

0.108 0.118 0.128 0.138 0.148

CO2 wt. fraction

d p,0

.5, μ

m

3 GTP, 25 MPa

5 GTP, 25 MPa10 GTP, 25 MPa

Fig. 6.5a. The particle size predicted at different processing conditions for a 0.4 mm diameter nozzle in an absence of a core


97

80

100

120

140

160

180

200

0.08 0.085 0.09 0.095 0.1 0.105 0.11

CO2 wt fratcion

d p,0

.5, μ

m3 GTP, 17.5 MPa

5 GTP, 17.5 MPa

10 GTP, 17.5 MPa

80

100

120

140

160

0.107 0.1145 0.122 0.1295 0.137 0.1445

CO2 wt fraction

d p,0

.5, μ

m

3 GTP, 25 MPa

5 GTP, 25 MPa

10 GTP, 25 MPa

Fig. 6.5b. The particle size predicted at different processing conditions for a 0.57 mm diameter nozzle in an absence of a core Relatively higher temperatures are required in the presence of cores compared to that in the absence of cores to produce particles. An additional elongation caused by the slots on the cores is mainly responsible for such results. The particles produced using a 0.4 mm nozzle in the absence and the presence of a core are shown in Fig. 6.6. Particles together with fibers are present in the presence of a core even at a high temperature. However, the major advantage of using a core is that the quantity of byproduct is substantially reduced and is relatively absent at high pressures. This is due to extra mixing caused inside the core-slot before the nozzle entrance.


98

a) dn=0.4 mm, GTP=5, T=381 K, and P=24.95 MPa (experiment 9)

b) dn=0.4 mm, GTP=3, T=404 K, and P=23.50 MPa (experiment 61)

Fig. 6.6. The SEM pictures of the products obtained in a) absence and b) presence of the core, C Smaller particles are obtained in the presence of a core at elevated temperatures close to 413 K at the CO2 pressures lower than the pressures used in the absence. Fig. 6.7 shows the influence of the core on the particle size for a 0.4 mm nozzle. In Fig. 6.7, the values without a core are predicted from equation 5) and the values with a core are experimentally obtained for similar processing conditions. Similar to the nozzle diameter effect, the smaller the core-slot width the higher is the elongation. Such effect is absent at high temperatures most probably due to low extensional viscosity and low shear viscosity that allow a better distribution of excess CO2 in the molten polymer. A large improvement over the particle size distribution can be seen in Fig. 6.8 when compared to that obtained using the traditional method.


99

0

100

200

300

400

500

600

d p,0

.5, μ

mno core core

Fig. 6.7. Influence of the core on the particle size for various conditions using a 0.4 mm nozzle. (P refers to the size predicted using the model, equation 5))

0

1

2

3

4

5

6

7

0 200 400 600 800 1000dp,0.5, μm

q ln,

%

grinding

core B

core A

core C

Fig. 6.8. The particle size distributions obtained in the presence of different cores for a 0.81 mm nozzle (core A: exp 108, core B: exp 106, and core C: exp 105) and obtained using a grinding method Effect of gas to polymer mass ratio With increasing gas to polymer mass ratio, the expansion product transforms from foam (microcellular) to particles or agglomerated fibers. Also, the particle size decreases with increasing ratio in the absence as well as in the presence of cores. As the ratio is increased above the solubility, CO2 bubbles are present and their size increases with increasing ratio. Consequently the thickness of the CO2 saturated polymer film decreases. A sudden depressurization of the excess CO2 leads to disturbances on the surface of the film and helps in better atomization of the melt. A positive effect of the ratio is present also on the particle size distribution in the absence of a core, Fig. 6.9. The effect is insignificant in the presence of cores probably due to a better distribution of CO2 bubbles.

65P exp 65 66P exp 66 67P exp 67 68P exp 68 64P exp 64


100

0

2

4

6

8

10

12

0 200 400 600 800 1000

dp,0.5, μm

q ln,

%

3 GTP

5 GTP

10 GTP

a) dn=0.57 mm, T ≈ 393 K, P~14.5 MPa: GTP=3 (exp 45), GTP=5 (exp 46), and GTP=10 (exp 47)

0

1

2

3

4

5

6

7

0 200 400 600 800 1000dp,0.5, μm

q ln,

%

7 GTP

10 GTP

10 GTP(repeat)

b) dn=0.81 mm, T ≈ 402 K, P ≈ 18.5 MPa: GTP=7 (exp 98) and GTP=10 (exp 99 and 99(repeat)) Fig. 6.9. The effect of gas to polymer mass ratio (GTP) on the particle size distribution obtained a) in the absence of core b) in the presence of core B Effect of temperature and pressure Fig. 6.5 clearly shows that the temperature and pressure contribute significantly in the particle production process. The particle size is decreased with decreasing temperature and increasing pressure, which can be related to high CO2 solubility. However, the effect of pressure at low temperatures, means high CO2 weight fraction, is insignificant, which is due to high shear and extensional viscosity of the material. The lower the shear or extensional viscosity of a polymer melt the easier is the expansion of the polymer melt. It has been shown in Chapter 5 that a significant reduction in the shear viscosity can be achieved due to dissolved CO2. Indeed, this assists in the production of particles. In the absence of CO2, it has been observed that the particle production


101

has not been possible despite very high pressure and temperature conditions (~30 MPa and ~413 K) both in the absence as well as in the presence of a core. In the particle production, the break up of polymer melt is caused not only by the expansion of dissolved CO2 but also by the expansion of excess of CO2. The density of CO2 increases with increasing pressure and decreases with increasing temperature. A sudden depressurization of dense CO2 to atmospheric condition causes an abrupt enhancement in its specific volume. This leads not only to a rapid nucleation of CO2 bubbles inside the polymer melt due to supersaturation but also to intense disturbances on the surface of the polymer film. At low temperatures, a resistance from viscous forces to the disturbances and the expansion of dissolved CO2 is high despite a significant shear viscosity reduction caused by the dissolved CO2. Therefore, the effect of pressure comes into the picture when a temperature is increased above a certain value. For isobaric conditions, an increment in the particle size with respect to increasing temperature in the absence of a core can be related to low CO2 solubility and low CO2 density. The particle size distributions obtained at different temperature and pressure conditions for different nozzles are shown in Fig. 6.10. In the absence of core, a narrower particle size distribution is obtained at the low temperature. It suggests that expansion of the gas saturated polymer solution at different temperatures is determined by the density and solubility of CO2. In the presence of core, a completely different picture can be seen in Fig. 6.10. At the high temperature, a relatively narrow particle size distribution is obtained. However, at the low temperature a positive effect of the pressure on the particle size distribution is present. One cannot compare the two results above as the diameter of the nozzle is different. A better expansion of uniformly dispersed excess CO2 due to low viscosity at the high temperature and a high super saturation of the dissolved CO2 at the high pressure are responsible for the narrow particle size distribution in both cases. 6.3.3. Shape and morphology of PPB particles The shape and morphology of particles is as important as the particle size in any powder application. The spherical shape is always preferred in various applications because of its good flow ability and of a good heat transfer. However, the shape becomes less important as the particle size reduces to micron level. The effect of nozzle diameter on the shape of the particles has already been discussed earlier in this chapter. Particles with different shapes (irregular, nearly-spherical and fibrous) and different morphologies (dense or foam) can be created by playing with processing conditions. The particles with different morphologies and shapes obtained in this study under different processing conditions for different diameter nozzles are shown in Fig. 6.11a-e.


102

0123456789

10

0 200 400 600 800 1000dp,0.5, μm

q ln,

%

16.3 MPa

19.8 MPa

0

1

2

3

4

5

6

7

0 200 400 600 800 1000

dp,0.5, μm

q ln,

%

381 K

401 K

a) dn=0.41 mm, GTP=5, and P ≈ 24.62 MPa : T=381 K (exp 9) and T=401 K (exp 21)

0

1

2

3

4

5

6

7

0 200 400 600 800 1000dp,0.5, μm

q ln,

%

403 K

413 K

b) dn=0.81 mm, core=B, P ≈ 18.55 MPa, and GTP=10: T= 401 K (exp 99) and T= 414

K (exp 106)

c) dn=0.57 mm, core=B, T ≈ 402 K, and GTP=10: P=16.65 MPa (exp 80) and P=19.8 MPa (exp 81)

Fig. 6.10. The effect of temperature (a and b) and pressure (c) on the particle size distribution obtained using different diameter nozzles


103

dn=0.4 mm, GTP=5, T=401 K, and P=24.3 MPa (exp 21)



Fig. 6.11a. Scanning electron microscope pictures of different morphologies and shapes obtained using different diameter nozzles


104




Fig. 6.11b. Scanning electron microscope pictures of different morphologies and shapes obtained using different diameter nozzles


105

dn=0.81 mm, core=C, GTP=7, T=394 K, and P=12.80 MPa (exp 95)

dn=0.81 mm, core=B, GTP=7, T= 403 K, and P=18.4 MPa (exp 98)

dn=0.81 mm, core=C, GTP=10, T=412 K, and P=13.6 MPa (exp 105) Fig. 6.11c. Scanning electron microscope pictures of different morphologies and shapes obtained using different diameter nozzles


106

dn=0.81 mm, core=A, GTP=3, T=416 K, and P=13.20 MPa (exp 107)

dn=0.81 mm, core=B, GTP=3, T=419 K, and P=11.00 MPa (exp 109)

dn=0.57 mm, core=B, GTP=10, T=405 K, and P=19.80 MPa (exp 81) Fig. 6.11d. Scanning electron microscope pictures of different morphologies and shapes obtained using different diameter nozzles


107


dn=0.40 mm, core=B, GTP=3, T=413 K, and P=18.40 MPa (exp 65)


Fig. 6.11e. Scanning electron microscope pictures of different morphologies and shapes obtained using different diameter nozzles


108

From the scanning electron microscope pictures shown in Fig. 6.11a-e and the processing conditions, morphology, and shape of the expanded product as a function of temperature and pressure can be represented schematically as shown in Fig. 6.12. At low pressures, because of low CO2 density, expansion of excess as well as dissolved CO2 does not overcome the viscous forces if the temperature is low. Consequently, a break up of polymer melt is not possible and hence, it results into foam. At low temperatures and at high pressures irregular shaped foamed particles are formed in the absence of core. Foaming of particles is due to a large amount of CO2 dissolved at high pressures. The irregular shape acquired by the particles is due to a rapid solidification caused by the evaporation of the dissolved CO2. The fibre shown along with the nearly spherical particle is the situation where a core is present. The additional elongation caused by the slot in the core is mainly responsible for the production of agglomerated fibers. A quantity of fibers reduces with an increase in the diameter of the nozzle, the core-slot width, and the temperature due to a decrease in the extensional viscosity. At high temperatures and low pressures, a reduction in the viscosity of a polymer caused by the temperature leads to an improved expansion. However, foam is formed along with particles due to a low pressure. At high temperatures and high pressures, nearly-spherical dense (less-foamed) particles are formed. The low CO2 solubility at high temperatures is responsible for such shape and morphology. Moreover, the dissolved CO2 can diffuse out easily from the particles due to low viscosity of the polymer. The solidification theory may provide a better insight in the proposed scheme.

T

P

Foamed irregular particle/fiber

Foam

Less foamed nearly spherical particle

Foam and Dense Particle

+

Fig. 6.12. A schematic representation of the effect of temperature and pressure on morphology and shape


109

6.3.4. Solidification Atomization of molten metals by a high pressure gas is a well established technique for continuous production of solidified metal particles. In such process, a high pressure gas passing through a small orifice or nozzle serves two functions. One is to break up the molten stream into particles by kinetic energy (pressure energy transformation) and second is to solidify the particles by the heat transfer due to forced convection. In the process used here, CO2 serves also similar functions, but in different manners, to break up a molten polymer into particles. The break up is caused by both expansion of excess CO2 and supersaturation of the dissolved CO2. Moreover, in the solidification of particles an extra cooling is achieved by the expansion of the dissolved CO2 from an elevated pressure to an ambient pressure. The extra cooling induces a rapid solidification, which has a substantial effect on the morphology and shape of particles. The effect of temperature and pressure on the morphology and shape of the particles could clearly be observed from the experimental results. However, the solidification model, used in Chapter 4, has been applied to the particles obtained under different processing conditions in order to have a theoretical understanding. As the particle size is in the order of micron, an assumption of the zero relative velocity (between the expanded CO2 and particles) is valid in this process. Compared to poly(ethylene glycol) particles, a faster solidification is expected in case of PPB as there is no heat of crystallization involved. The solidification model results are shown Fig. 6.13 in terms of the average solidification time, tavg,, as a function of pressure for different isothermal conditions. The calculations have been performed for a particle size of 80 μm assuming the gas volume fraction of 0.5 (foamed particles) and 0 (dense/solid particles) for temperatures 363 and 413 K, respectively.

1

10

100

1000

7 10 13 16 19 22 25 28

P, MPa

t avg, m

s 368 K

413 K

Fig. 6.13. The solidification times calculated for different pressures under isothermal conditions


110

The solidification time is found to be increased with the increasing temperature and the decreasing pressure. However, a pressure dependent behaviour is found only at the low temperature. At the low temperature, the expansion of dissolved CO2 plays a vital role in the solidification of particles. The higher the pressure and the lower the temperature the faster is the cooling caused by the expansion of dissolved CO2 due to high CO2 solubility. Consequently polymer particles in the molten state freeze immediately after the expansion. Thus, irregular shape particles containing CO2 bubbles are formed. While at the high temperature, the solidification process is delayed i.e. the average solidification time is increased. This is because of high sensible heat acquired by the polymer melt and low heat of evaporation of CO2 due to low CO2 solubility. The delay provides sufficient time to retract the molten particle to a spherical shape both due to the surface tension and enhanced relaxation time for polymer molecules before the cooling takes place.

6.3.5. Properties of polymer after and before processing Indeed it is necessary that the physical properties of the material should not be altered after processing with CO2. As CO2 is an inert gas, no chemical attack is expected for PPB. However, the polymer might have been subjected to chain degradation caused by high shear generated inside the nozzles. Therefore, the FT-IR and gel permeation chromatography (GPC) measurements have been carried out for the processed and un-processed material. The FT-IR spectra, Fig. 6.14, and the molecular weights obtained from GPC (polystyrene standard based) measurements, Table 6.3, provide evidences for an absence of such degradation.

0.01

0.11

0.21

0.31

0.41

600 800 1000 1200 1400 1600 1800 2000

wavenumber, cm-1

abso

rptio

n

Fig. 6.14. FT-IR spectra of unprocessed PPB and supercritical CO2 processed PPB

exp 99exp 83

exp 46PPB


111

Table 6.3. Molecular weights of unprocessed and processed PPB obtained from GPC analysis

Exp.

Mn (kg/kmol)

Mw (kg/kmol)

PPB 2960 7150 45 2950 7240 67 2940 7280

106 2930 7220 Induced crystallization has been reported for some amorphous polymers after treating with supercritical CO2 [12-14]. The plasticization of a polymer caused by the dissolved CO2 increases the mobility of the polymer chains. This allows the chains to rearrange into a more ordered conformation and results in induced crystallization. Therefore, digital scanning calorimetric measurements have been performed for the polymer samples processed under different conditions (experiments 45, 46, 81, 99 and 106). No evidences of a crystalline nature and a change in the glass transition temperature have been found. Thus, the process allows the continuous production of particles without any chemical or structural changes in the polymer. 6.3.6. Dimensional analysis In atomization studies, it is a very common practice to relate the diameter of liquid droplets or particles to physical properties of fluids and processing parameters. Detailed information on different atomization methods and the dimensional analysis of several studies has been provided by Lefebvre [15]. In dimensional analysis, the physical properties and the parameters are arranged in such a way that a dimensionless empirical correlation is obtained. It allows estimating the contributions of various forces as well as different parameters in determining the diameter of droplets or particles. Here, a dimensionless correlation has been developed and verified for 0.4 and 0.57 mm diameter nozzles, equation 6). The equation is regressed to the particle size data obtained using equation 5) with a nonlinear least-square fit procedure for a pressure range of 18-25 MPa and a temperature range of 378-413 K, Fig. 6.15.

4.0

2

***

7.0

**

2.1*

4.111*7.2)/(*91.55.0,

⎟⎟⎟

⎠

⎞

⎜⎜⎜

⎝

⎛

⎟⎟

⎠

⎞

⎜⎜

⎝

⎛

⎟⎟

⎠

⎞

⎜⎜

⎝

⎛⎟⎠⎞

⎜⎝⎛ +=

s

snds

ndPs

c

sGTPndnl

ndpd

η

ρσσ

ρ

ρ 6)

In the equation above, dn, ln, sρ , cρ , sσ , and sη are the diameter of nozzle, the length of nozzle, the density of polymer solution, the density of CO2, the surface tension of polymer solution, and the viscosity of polymer solution,


112

respectively. The surface tension of the solution has been calculated using the mixing rule given by Reid et al. [16].

0

100

200

300

400

500

600

0 100 200 300 400 500 600

dp,0.5,exp, μm

d p,0

.5,fi

t, μm

Fig. 6.15. The average particle size data fitted to the empirical correlation The exponentials in equation 6) clearly suggest that the diameter of the particles is mostly influenced by the gas to polymer mass ratio, the nozzle length to diameter ratio, and the density ratio. The higher the diameter of a nozzle and the density of CO2 the smaller is the size of the particle, which is related to low extensional viscosity and high supersaturation, respectively. Moreover, it can be seen from equation 6) that the particle size decreases with an increase in pressure and a decrease in surface tension. It is generally observed in the atomization process where a gas is not dissolved in polymer solutions that the droplet size increases with increasing shear viscosity ( pη ). However, an inverse relationship is obtained for the supercritical CO2 assisted expansion where CO2 is dissolved in the polymer melt. This suggests that the particle production is also controlled by the expansion of dissolved CO2. The last term in the equation represents the effect of temperature. At low temperature, the amount of CO2 dissolved in a polymer is high and hence, a high supersaturation is achieved upon the expansion despite the high shear viscosity. The deviation in the predicted and experimental particle size data is reasonable since theoretically estimated shear viscosity and surface tension data have been used. A better fit can be expected with actual, measured, shear viscosity and surface tension. Moreover, an inclusion of extensional viscosity (in the presence of dissolved CO2) of polymer may also improve the fitting.


113

6.4. Conclusions An extruder and the SMX static mixer is a good combination for a continuous production of micron size particles from a high viscous polymer melt using supercritical CO2. It is not possible to break up the polymer melt into particles even at elevated pressures in the absence of CO2. Not only the CO2 solubility but also the nozzle diameter, core-slot width, gas to polymer mass ratio, temperature, and pressure are important processing parameters of the continuous process to control the particle size, shape, particle size distribution, and morphology. The negative effect of extensional viscosity on the particle production has clearly been observed through the nozzle diameter and the core-slot width. In general, smaller particles and a better particle size distribution have been obtained at low temperatures and high pressures in the absence of a core. The high CO2 solubility at low temperatures and high pressures are responsible for such results. On the other hand, smaller particles have been produced at high temperatures in the presence of a core due to uniform distribution of excess CO2 caused by the core. Nearly spherical shaped particles have been obtained at high temperatures as a result of a delayed solidification which is in a good agreement with the solidification theory. Overall study opens up the possibility of processing different polymers, used in powder form, with supercritical CO2 without altering their physical properties using the designed set up and playing with the various parameters to obtain micron size particles. References [1] Baldwin D F, Park C B, Suh N P. An extrusion systems for the processing of

microcellular polymer sheets: shaping and cell growth control. Polymer Engineering and Science 1996;36:1425-1435.

[2] Park C B, Baldwin D F, Suh N P. Effect of the pressure drop on cell nucleation in continuous processing of microcellular polymers. Polymer Engineering and Science 1995;35:432-440.

[3] Park C B, Suh N P. Rapid polymer/gas solution formation for continuous production of microcellaur plastics. Journal of Manufacturing Science and Engineering 1996;118:639-645.

[4] Michaeli W, Heinz R. Foam extrusion of thermoplastics polyurethanes(TPU) using CO2 as a blowing agent. Macromolecular Materials Engineering 2000;284-285:35-39.

[5] Jacobsen K, Pierick D. Microcellular foam molding: advantages and applications ANTEC 2000;2:1929-1933.

[6] Siripurapu S, Gay Y J, Royer J R, DeSimone J M, Khan S A, Spontak R J. Microcellular polymeric foams (MPFs) generated continuously in supercritical carbon dioxide. Material Research Society Symposium Proceeding 2000;629:FF9.9.1-FF9.9.6.

[7] Jacobsen K, Pierick D. Injection molding innovation: the microcellular foam process Plastics Engineering 2001;57:46-51.


114

[8] Siripurapu S, Gay Y J, Royer J R, DeSimone J M, Spontak R J, Khan S A. Generation of microcellular foams of PVDF and its blends using supercritical carbon dioxide in a continuous process. Polymer 2002;43:5511-5520.

[9] Han X, Zeng C, Koelling K W, Tomasko D L, Lee J. Continuous microcellular polystyrene foam extrusion with supercritical CO2. Polymer Engineering and Science 2002;42:2094-2106.

[10] Han X, Zeng C, Lee J, Koelling K W, Tomasko D L. Extrusion of polystyrene nanocomposite foams with supercritical CO2. Polymer Engineering and Science 2003;43:1261-1275.


[12] Ma W, Yu J, He J. Direct Formation of γ form crystal of syndiotactic polystyrene from amorphous state in supercritical CO2. Macromolecules 2004;37:6912-6917.

[13] Handa Y P, Zhang Z, Wong B. Effect of compressed CO2 on phase transitions and polymorphism in syndiotactic polystyrene. Macromolecules 1997;30:8499-8504.

[14] Chiou J S, Barlow J W, Paul D R. Polymer crystallization induced by sorption of carbon dioxide gas. Journal Applied Polymer Science 1985;30:3911-3924.

[15] Lefebvre A H. Atomization and sprays. Hemisphere publishing corporation, USA,1989.

[16] Reid R C, Prausnitz J M, Poling B E. The properties of gases and liquid. 4th edition. McGraw Hill book corporation, New York, 1986.

Chapter 7: Technological assessment of the supercritical CO2 assisted particle production methods and prospects of supercritical CO2

115

Chapter 7

Technological assessment of the supercritical CO2 assisted particle production methods and prospects of supercritical CO2


116

7.1. A qualitative technological evaluation Interest in applications involving supercritical fluids and their potential for process improvement over traditional methods has significantly increased in the last decades. Supercritical fluids have been adopted or are being explored in various applications such as extraction, blending, microcellular foaming, fractionation, crystallization, particle production, and polymerizations. Among them, particle formation using supercritical fluids is still in its infancy on a commercial level. Ferro Corporation, USA, is a leading company, which has been exploring the use of supercritical fluids in particle production of several polymers including biopolymers. There is a no doubt that of all supercritical solvents CO2 is mostly preferred due to its easily achievable critical conditions and harmless nature. The high solubility of CO2 in many polymers provides an opportunity to process polymers in the absence (or a very small presence) of harmful and hazardous organic solvents. Moreover, the process can be carried out at low temperatures due to plasticization of polymer caused by the dissolved CO2. Thus, supercritical CO2 assisted processes ease the handling of polymers without. This is beneficial from a commercial point of view. In this work, we have adopted and modified the PGSS method for the production of particles of poly(ethylene glycol) and polyester resins, respectively. To commercialize supercritical processes, it is requisite to know their potential also from an economical point of view. Here, a qualitative technological assessment has been carried out on the methods used in this work. It is presented in terms of a block diagram shown in Fig. 7.1. From Fig. 7.1 it is clear that the methods used in this work are often superior to the traditional methods. A cost factor due to the high pressure conditions in a supercritical process can be compensated with a number of benefits obtained over the traditional methods. Compared to the cost of cryogenic milling the cost of a supercritical process is not so high considering the product quality obtained with the supercritical process. Looking at increasing amount of emissions of hazardous solvents in the environment and their consequences, CO2 as a solvent may be accepted on a commercial level more easily. Moreover, it is very easy to separate CO2 from the final product just by depressurization. A good control over the size and morphology of particles is possible in the supercritical process by tuning the processing conditions. Dense and foamed particles can be produced in a supercritical process depending on the amount of CO2 dissolved in the polymer. Production of particles using supercritical CO2 is a single step process while in the processes based on the traditional methods various steps such as separation of the solvent and fractionation of the particles of different sizes are necessary in order to get a desired product size.


117

Fig. 7.1. A qualitative technological assessment of the supercritical CO2 and the traditional methods To make the supercritical processes more effective, the synthesis of polymer may also be carried out in the presence of supercritical CO2. The dissolved CO2 reduces the viscosity of the synthesized polymer melt and consequently, the mass transfer of monomers to the active sites of the catalyst enhances, which leads to a high conversion. The processes studied here are not applicable to polymers in which the solubility of CO2 is very low. The reason is that the lower the dissolved amount of CO2 in a polymer the lower is the potential of CO2 as a plasticizer or solvent.

Operating cost

Organic solvent

Separation cost

No. of steps

Control overproduct quality

PGSS/EGSEG

Processingconditions

Environmentalbenefits

Grinding/milling/spraydrying/crystallization

high

mild

absent

no

high

better

low

Mild/elevated

present/ absent

yes

low

poor

1 >1


118

7.2. Prospects of supercritical CO2 7.2.1. The continuous production of particles from a polyester epoxy resin melt The expansion of gas-saturated solution with excess gas (EGSEG) has been adopted for the production of micron size particles from a polyester epoxy resin (Holland colors, The Netherlands) melt. Polyester epoxy resin (PER) has a shear viscosity slightly lower than a polyester based on propoxylated bisphenol (PPB). A few experiments have been performed using a nozzle diameter of 0.81 mm in the presence of a core using a modified set up. Like PPB, it is not possible to produce particles only with the dissolved amount of CO2 for PER. A minimum gas to polymer mass ratio of 3 is required for the continuous particle production. It clearly suggests that an excess CO2 plays a pivotal role in the production of particles from high viscous polymer melts. The scanning electron micrographs of the particles of PER obtained under different processing conditions are shown in Fig. 7.2. Irregular shaped particles are formed in both cases.

a) b) Fig. 7.2. Scanning electron micrographs of polyester epoxy resin particles using 0.81 mm nozzle and core C : a) P = 16 MPa, T= 393 K, and gas to polymer mass ratio= 4 and b) P = 15.7 MPa, T= 403 K, and gas to polymer mass ratio= 4 7.2.2. Supercritical CO2 in microcellular foaming A microcellular polymer is defined as a porous material having cell sizes less than or equal to 10 microns. Compared to dense polymers, microcellular foamed polymers have lower weight combined with the high impact strength, toughness, and fatigue life. Such materials are used in applications such as separation media, adsorbents, and catalyst supports [1, 2]. The most common techniques are foaming by the thermal induced phase separation (TIPS) and the use of chemical blowing agents. In the TIPS process, a single phase solution of a polymer and an organic solvent is subjected to a phase separation by a temperature quench. After careful removal of the solvent by freeze drying or


119

supercritical extraction, a microcellular structure is obtained. The major drawbacks of the process are possible structure collapsing due to the liquid-vapour phase interface and the separation of the solvent. Though such problems are not encountered with the other technique, the depletion of the ozone layer in the atmosphere due to chlorofluorocarbons (CFC), generally used as a blowing agent, has now enforced to put a ban on their use.

Microcellular foaming of polymers using supercritical CO2 as a blowing agent has already been touted as a revolutionary invention in polymer industries [3-8]. Recently, the use of CO2 is also found in various medical applications involving controlled release devices, scaffolds, and medical devices where foamed biodegradable or biocompatible polymers are used. The replacement of various traditional blowing agents used for foaming process such as chlorofluorocarbons (CFC), hydrochlorofluorocarbons (HCFC), and volatile organic components (VOCs) with supercritical CO2 allows working in a clean and safe environment. Moreover, a narrow cell size distribution, easy solvent recovery, good plasticizing ability, and high diffusivity are the advantages of using supercritical CO2 in microcellular foaming processes. Several research groups have reported batch foaming of various polymers such as poly(methyl methacrylate) (PMMA), polycaprolactone (PCL), poly (D,L-lactic acid) (PD,LLA), poly(ethylene terephthalate) (PET), polystyrene (PS), polycarbonate (PC), poly(lactic-co-glycolic acid) (PLGA) [9-25].

An exploratory study of the batch foaming of poly(vinyl chloride) (generally used in an electrical insulation application) with sub- and supercritical CO2 has been carried out. In this study, the effect of pressure and temperature on the foam morphology and the density have been investigated.

Experimental method and analysis

The experimental setup is shown schematically in Fig. 7.3. Poly(vinyl chloride) (PVC), in the form of cylinder (4 mm diameter and 25 mm length), was enclosed in a stainless steel high pressure bomb. The cylindrical shape samples were prepared by extruding the polymer through a small capillary die. The bomb was heated using a heating element and a temperature is controlled using an Eurotherm (The Netherlands) controller. The bomb was pressurized with CO2 using a CO2 pump.

P

T

Heating element

High pressure bomb

Fig. 7.3. A schematic drawing of the set up used in foaming experiments


120

In a typical experiment, the bomb was flushed with CO2 for a few minutes. Then, a pressure was increased to the desired value while the temperature was held constant. The system was kept for the desired saturation time. The pressure was then rapidly quenched over a very small period to atmospheric pressure. The system was allowed to stabilize at that pressure and then the bomb was cooled down to room temperature.

The foamed samples were analysed for the bulk density of the samples and morphology of the microstructure. A difference between the bulk density of a sample before and after foaming was used to calculate a percentage reduction in the bulk density caused by the foaming. A scanning electron microscope was used for a qualitative assessment of the microstructures of the samples. Results and discussion In foaming, the dissolved CO2 plasticizes the polymer and reduces the apparent glass transition temperature or melting point to the processing temperature or to a near ambient temperature. Venting the CO2 upon depressurization, a thermodynamic instability causes supersaturation of the CO2 dissolved in the polymer matrix and hence, nucleation of cells occurs. The growth of the cells continues until the polymer vitrifies. The saturation pressure, the saturation temperature, the saturation time, and the depressurization rate are the key parameters in determining properties and morphology of the foam. In this study we have given attention to the first two parameters while keeping the rest constant.

Fig. 7.4 shows the foam structures obtained at different temperatures and pressures in our experiments. The effect of pressure and temperature can be explained together in terms of the solubility of CO2. There is marked difference between the structure obtained at 6 MPa and 15 MPa. A microcellular structure is only seen in case of the process conditions above the critical pressure (7.38 MPa). The results are related to the solubility of CO2, which increases with increasing pressure. The solubility of CO2 in the polymer is probably very low at 6 MPa. Homogeneous nucleation theory can be used to explain the effect of pressure on the microcellular structure. According to the theory, the energy barrier to nucleation decreases with increasing pressure drop.

The effect of temperature on morphology can be seen in terms of the size of the cells and the cell distribution. An increase in the cell size and a non uniform distribution of the cells are observed with an increase in temperature. Since the solubility of CO2 decreases with increasing temperature, a degree of supersaturation is low at a high temperature that leads to a low nucleation density. Moreover, due to high temperatures the dissolved CO2 has a longer time


121

available to diffuse to the already existing nuclei and to grow the nuclei to a big size before the polymer is vitrified.

a) 6 MPa, 333 K b) 15 MPa, 333 K

c) 15 MPa, 373 K Fig. 7.4. Scanning electron micrographs of the foam structures a), b), and c) obtained at different conditions (saturation time = 24 hrs, depressurization time ≈ 20 sec) Percentage reductions in the bulk densities of the samples foamed at different temperatures and pressures are given in Table 7.1. The bulk density is decreased with increasing temperature and pressure. The solubility of CO2 increases with increasing pressure and hence, the nucleation rate increases leading to a low bulk density. An opposite result is expected when a temperature is increased as the solubility of CO2 decreases with increasing temperature. The bulk density decreases at 373 K mainly due to bigger size cells. In order to have a complete microcellular process study, the effect of other parameters such as saturation time and depressurization rate must be taken into account.


122

Table 7.1. Percentage reductions in the bulk densities of the samples foamed at different processing conditions

7.3. Conclusions of the thesis and future outlook In this thesis, the PGSS method has been tested and modified for the production of particles from low viscous and high viscous polymers, respectively, using supercritical CO2 as a solvent. Knowledge of the solubility of CO2 in the polymer is essential for both methods. The An FT-IR studies gave qualitative information about the solubility while gravimetric measurements gave quantitative information. Dissolved CO2 has a dramatic effect on the viscosity of a polymer due to the plasticization effect. The experimental rheological data show and a model based on the free volume theory predicts a considerable reduction in the viscosity of poly(ethylene glycol) and polyester based on propoxylated bisphenol, respectively. The production of micron size particles is possible for both polymers using supercritical CO2 as a solvent. A better control over the properties of particles such as particle size, particle shape, particle size distribution, porosity, and bulk density is possible by tuning the various parameters such as pressure, temperature, nozzle diameter, core, and gas to polymer mass ratio. Both methods provide an opportunity to process several low and high viscosity polymers in which the solubility of CO2 is adequate. Moreover, a qualitative technological evaluation suggests that the methods used for the particle production are a promising solution for the already available traditional methods. Apart from the economical and technological advantages, the methods are very good also from an environmental concern. However, before replacing the traditional methods with PGSS or EGSEG, it is necessary to look carefully at overlooked segments. It has already been discussed that the shear and extensional viscosity play important roles in determining the product morphology. To find a better correlation between the particle size and different processing parameters, a detailed experimental investigation of shear and extensional viscosity in the presence of dissolved CO2 has to be done. Since the extensional effect is less pronounced in large diameter nozzles, various experiments have to be carried out using the large nozzles at elevated pressures for low gas to polymer mass ratios. Modelling of the flow behaviour inside a static mixer and a nozzle for high pressure gas-polymer systems can provide theoretical insights in the

P

(MPa)

T

(K) ⎟⎟⎠

⎞⎜⎜⎝

⎛ −100*

0

0

ρρρ t

(%) 6 333 0

15 333 38.4 15 373 54.7


123

mechanism of the particle production. Recently, a positive effect of a back pressure on the shape of the particles has been reported for the PGSS method [26]. An expansion of a polymer-gas solution from elevated to moderate pressures may result into different products due to differences in supersaturation, break up mechanism, and solidification. The EGSEG method has to be tested for several polymers including biopolymers, which are difficult to grind due to their low glass transition or melting points. Various other supercritical methods must be tested for these materials before finalizing the new method. Since the solubility of many polymers in CO2 is low, a method in which supercritical CO2 is used as an antisolvent for a polymer solution consisting of a polymer and an organic solvent may be another choice. Looking at rapidly increasing interest for supercritical technologies, attention must also be given to various other applications such as microcellular foaming, extraction, polymer modification, and polymerization in which supercritical CO2 replaces an organic solvent. References [1] Klempner D, Frisch K C. Handbook of polymeric foams and foam Technology.

Munich: Hanser Publishers, 1991. [2] Shutov F A. Integral/ Structural Polymeric Foams: Technology, Properties and

Applications. Berlin Heidelberg: Springer-Verlag, 1986. [3] Kishbaugh L A, Levesque K J, Guillemette A H, Chen L, Xu J, Okamoto K T.

Fiber-filled molded foam articles, molding, and process aids. (USA) WO:2002026482, 2002.

[4] Farrar P A. Method of forming foamed polymer insulators for use in high density circuits. (USA) US:2002168872, 2002.

[5] Devine J N, Kemmish D J, Wilson B, Griffiths I, Seargeant K M. Biocompatible polymeric materials for medical devices. (UK) WO:2002000275, 2002.

[6] De Simone J M, Paisner S N. Methods of forming porous polymeric structures using carbon dioxide and polymeric structures formed thereby. (USA) US:2003180522, 2003.

[7] Clarke A J. Novel pharmaceutical dosage forms produced by injection molding. US:2003057197, 2003.

[8] Liu H-C, Shih H-H, Tsai C-C, Wu C-T, Liu W-B. Manufacturing polymeric foam using supercritical fluids. US:2003057197, 2003.

[9] Baldwin D F, Park C B, Suh N P. A microcellular processing study of poly(ethylene terephthalate) in the amorphous and semicrystalline state. part I. Polymer Engineering and Science 1996;36:1437-1446.

[10] Baldwin D F, Park C B, Suh N P. A microcellular processing study of poly(ethylene terephthalate) in the amorphous and semicrystalline state. part II: cell growth and process design. Polymer Engineering and Science 1996;36:1446-1453.

[11] Mooney D J, Baldwin D F, Suh N P, Vacanti J P, Langer R. Novel approach to fabricate sponges of poly(D,L-lactic-co-glycolic acid) without the use of organic solvents. Biomaterials 1996;17:1417-1422.

[12] Sparacio D, Beckman E J. Generation of microcellular biodegradable polymers using supercritical carbon dioxide. ACS symposium series 1998;11:181-193.


124

[13] Stafford C M, Russell T P, McCarthy T J. Expansion of polystyrene using supercritical carbon dioxide: Effects of molecular weight, polydispersity, and low molecular weight. Macromolecules 1999;32:7610-7616.

[14] Japon S, Leterrier Y, Manson J E. Recycling of poly(ethylene terephthalate) into closed cell foams. Polymer Engineering and Science 2000;40:1942-1952.

[15] Kokturk G, Howdle S M. Supercritical fluids: new solvents for polymer synthesis and polymer processing. DECHEMA Monographs 2001; 137:79-95.

[16] Mizumoto T, Sugimura N, Moritani M. CO2-induced stereo complex formation of stereoregular poly(methyl methacrylate) and microcellular foams. Macromolecules 2000;33:6757-6763.

[17] Jin W, Xingguo C, Mingjun Y, Jiasong H. An investigation on the microcellular structure of polystyrene/LCP blends prepared by using supercritical carbon dioxide. Polymer 2001;42:8265-8275.

[18] Quirk R A, France R M, Shakesheff K M, Howdle S M. Supercritical fluid technologies and tissue engineering scaffolds. Current Opinion in Solid State and Materials Science 2004;8:313-321.

[19] Barry J J A, Gidda H S, Scotchford C A, Howdle S M. Porous methacrylated scaffolds: supercritical fluid fabrication and in vitro chondrocyte responses. Biomaterials 2004;25:3559-3568.

[20] Lee S-J, Kim M-S, Chung J G. Preparation of porous polycarbonates membranes using supercritical CO2 enhanced miscibility. Journal of industrial and Engineering Chemistry 2004;10:877-882.

[21] Xu Q, Ren X, Chang Y, Wang J, Yu L, Dean K. Generation of microcellular biodegradable polycaprolactone foams in supercritical carbon dioxide. Journal of Applied Polymer Science 2004;94:593-597.

[22] Hile D D, Pishko M V. Solvent-free protien encapsulation within biodegradable polymer foam. Drug Delivery 2004;11:287-293.

[23] Fujiwara T, Yamaoka, K Y, Wynne K J. Poly(lactide) swelling and melting behaviour in supercritical carbon dioxide and post-venting porous material. Biomacromolecules 2005 accepted.

[24] Xu Q, Pang M, Peng Q, Li J, Jiang Y. Application of supercritical carbon dioxide in the preparation of biodegradable polylactide membranes. J. Appl. Pol. Sci. 2004;94: 2158-2163.

[25] Asandei A D, Erkey C, Burgess D J, Saquing C, Saha G, Zolnik B S. Synthesis of PLGA foams by high temperature supercritical CO2 expanded ring opening copolymerization of D,L-lactide and glycolide. Polymer Preprints 2004;45:1014-1015.

[26] Hao J, Whitaker M J, Serhatkulu, G Shakesheff, Kevin M, Howdle S M. Supercritical fluid assisted melting of poly(ethylene glycol): a new solvent-free route to microparticles. Journal of Materials Chemistry 2005;15: 148-1153.

125

List of symbols a Shift factor d Particle diameter f Friction factor h Heat transfer coefficient k Interaction parameter l Length t Solidification time

t∆ Residence time Ci WLF constant Cp Specific heat of polymer

H∆ Heat GTP Gas to polymer mass ratio M Molecular weight P Pressure S Surface area T Temperature V Volume W Weight fraction Z Lattice coordination parameter λ Thermal conductivity µ Chemical potential η Viscosity ρ Density ν Specific volume

2ν Bending mode

3ν Antisymmetric stretching mode * Characteristic ~ Reduced Subscripts a Air B Basket c Critical g Glass transition i, j Component m Melting point mix Mixture n Nozzle o Pure p Polymer p,0.5 Average s Solution v Vaporization w Weight average

126

Summary Due to unusual, gas-like and liquid-like, properties of supercritical CO2, it has become a replacement for environmentally hazardous organic solvents in a variety of synthesis and processing. Presently, many polymer and pharmaceutical industries are taking an initiative to use CO2 on a commercial level. Supercritical CO2 has already been used as a solvent or an anti solvent in the production of small size particles varying from nanometers to micrometers. Its use also allows a better control over the particle morphology, shape, density, and the particle size distribution. The particle from gas saturated solution (PGSS) method has been used for the production of particles from different polymer melts. The PGSS method is applicable only to low viscous polymer melts. In this method, an expansion of a polymer-CO2 solution takes place by a supersaturation of the dissolved gas due to a sudden depressurization and hence, the break up of the solution results in small particles. In this thesis, the PGSS method has been tested and modified for the production of particles from low viscous poly(ethylene glycol) and high viscous polyesters using supercritical CO2 as a solvent. The polyesters are used in powder form for toner applications, while poly(ethylene glycol) is used in a controlled drug release application. The modified method is called expansion of gas saturated solution with excess gas (EGSEG). Before applying these methods, the knowledge of the solubility of CO2, the viscosity reduction due to the dissolved CO2 and the influence of the various parameters on the product quality is essential. The solubility of CO2 in a polymer is not just a pure physical phenomenon. The solubility of CO2 varies from one polymer to another depending on weak interactions between CO2 and chain groups in the polymers. We have used Fourier transform–infrared spectroscopy (FT-IR) to reveal the interaction strengths of different polymers with CO2. Shifts in wave numbers, however small, have been observed with the chain groups of the polymers due to a Lewis-acid base kind of interaction. On the other hand, the bending mode of CO2 shows a significant modification in its spectra after interacting with the chain groups. From the width of the spectra, it has been concluded that ether group containing polymers show a higher interaction with CO2 compared to the polyesters. FT-IR spectroscopy is a good screening tool in a preliminary selection of a polymer. With the FT-IR spectroscopy only qualitative information over the solubility of CO2 in the polymers has been obtained. Therefore, a magnetic suspension balance has been used for measuring the solubility of CO2 in polyesters based on propoxylated- and ethoxylated-bisphenol (PPB and PEB). In order to incorporate the buoyancy correction, which is requisite due to the swelling of polymer caused by dissolved CO2, swelling measurements for similar experimental conditions have been performed using an optical cell separately. The solubility of CO2 increases with increasing pressure and decreasing temperature for both polyesters. As the solubility of CO2 is high in PPB, PPB has been used as a model compound for the particle production using the EGSEG method.

Batch experiments have been performed to test whether the PGSS method is applicable to PPB for the particle production. Due to high viscosity of PPB it has not been possible to expand the PPB-CO2 solution using the CO2 dissolved in PPB only. A batch set up has also been used for the generation of particles from poly(ethylene glycol) (PEG) melts of different molecular weights (6000 and 10000). PEG is mostly used in a drug delivery application for encapsulating a drug. The PGSS method is applicable to PEG due to its relatively low viscosity. The effect of various parameters such as pressure, temperature, nozzle diameter, and molecular weight on the particle size and particle size distribution

127

has been studied in detail. Different morphologies and particle sizes are possible by tuning the processing conditions. A solidification model has been used to explain the effect of temperature and pressure on the shape of the particles. For the PEG particles, with different morphologies such as foam or dense and different sizes, produced using the PGSS method, a good control over a drug release may be obtained.

The dissolved CO2 causes a considerable reduction in the viscosity of the polymer due to an enhancement in its free volume. As the viscosity of PPB is very high, it is essential to know in advance about a reduction in the viscosity that can be caused by the dissolved CO2. A model based on the free volume theory has been used to estimate the reduction in the viscosity. The model shows a considerable reduction in the viscosity. For the particle production, a continuous set up containing mainly an extruder and a Kenics type static mixer has been designed and exploratory experiments have been performed. The static mixer has been used for mixing the polymer and CO2 as it is more economical than increasing the extruder length. Moreover, it has not been possible to mix them adequately in an extruder with high gas mass ratio. CO2 in excess to its solubility has been used in the particle production. An expansion of the excess CO2 enhances a break up of a polymer melt by creating intense disturbances on its surface. The product quality is also dependent on the nozzle diameter. The smaller the nozzle diameter the higher is the extensional effect and hence, the product morphology varies from agglomerated fibers to particles as the nozzle diameter increases. A core containing slots has been inserted before the nozzle, which provides a better homogenization. In this configuration a by product such as a small amount of foam is absent. On the other hand, a high temperature is necessary to overcome the additional extensional viscosity caused by the slots.

The set up used in the exploratory experiments has been modified for high polymer feed rates and pressure conditions. A gear pump has been attached after the extruder to keep a low pressure on the extruder side (input of the pump) and a high pressure on the other side. An SMX static mixer has been used for a better mixing of the polymer and CO2. Several experiments have been performed using different diameter nozzles and cores under different processing conditions. Different morphologies, particle sizes and particle size distributions have been obtained by playing with different parameters such as nozzle diameter, core-slot width, pressure, temperature and gas to polymer mass ratio. In the absence of a core, the particle size reduces with an increasing pressure and a decreasing temperature. High CO2 solubility under these conditions is responsible for such results. Narrow particle size distributions have been obtained using this method compared to the grinding method. The solidification model has been used to reveal the role of temperature and pressure in determining the shape of the particles. An attempt has been made to empirically correlate the particle size with different contributions such as geometric, shear viscosity, surface tension, pressure, and flow using a dimensional analysis. Apart from the economical and technological advantages, the EGSEG method is very good from an environmental concern. A qualitative technological evaluation has been done which suggests that the method developed for the continuous process is a promising solution for the already available traditional methods. However, before replacing the traditional methods with EGSEG it is necessary to test also the other supercritical methods. Looking at rapidly increasing interest for supercritical technologies, various other applications of supercritical CO2 emerge such as microcellular foaming, extraction, polymer modification and polymerization in which supercritical CO2 replaces an organic solvent.

128

Samenvatting

Door de bijzondere gas en vloeistof eigenschappen van superkritisch CO2, is het een alternatief geworden voor milieuonvriendelijk organische oplosmiddelen in een scala aan syntheses en processen. Tegenwoordig nemen de polymeer industrie en farmaceutische industrie steeds meer initiatief om CO2 op een commercieel niveau te gebruiken. Superkritisch CO2 wordt al gebruikt als oplosmiddel of anti-oplosmiddel in de productie van kleine deeltjes, variërend in grootte van nanometers tot micrometers. Het gebruik ervan zorgt ook voor een betere controle over de morfologie van de deeltjes, de vorm, de dichtheid en de deeltjes-grootte verdeling. De PGSS (deeltjes uit met gas verzadigde oplossing) methode werd gebruikt voor de productie van deeltjes uit verschillende gesmolten polymeren. Deze methode is alleen toepasbaar bij laag viskeuze gesmolten polymeren. Bij deze methode, vindt een expansie van het polymeer-CO2 mengsel plaats. Door superverzadiging van het opgeloste CO2, als gevolg van de plotselinge vermindering druk, resulteert de expansie van de oplossing, in het opbreken van de polymeerdeeltjes in veel kleinere deeltjes. In dit proefschrift is de PGSS methode getest en gemodificeerd voor de productie van deeltjes uit laag viskeus polyethyleenglycol en hoog viskeuze polyesters, waarbij superkritisch CO2 als oplosmiddel gebruikt is. De polyesters worden gebruikt als poeder voor toner applicaties, terwijl polyethyleenglycol gebruikt wordt voor gecontroleerde dosering van geneesmiddelen. De gemodificeerde methode wordt EGSEG (expansie van met gas verzadigde oplossing met een overmaat aan gas) genoemd. Voordat deze methode toegepast kan worden, is het essentieel om kennis te vergaren over de oplosbaarheid van CO2, de reductie viscositeit van het polymeer door opgelost CO2 en de invloed van de verschillende parameters op de kwaliteit van het product. De oplosbaarheid van CO2 in een polymeer is niet alleen maar een fysisch fenomeen. De oplosbaarheid van CO2 verschilt per polymeer omdat het afhankelijk is van de zwakke interacties van CO2 en de ketengroepen in de polymeren. Fourier transformatie-infrarood spectroscopie (FT-IR) werd gebruikt om de interactie tussen verschillende polymeren en CO2 te meten. Er zijn kleine verschuivingen in de golflengtes van de polymeerketengroepen waargenomen, dit wijst op een soort Lewis zuur/base interactie. Aan de andere kant toont de buigmodus van CO2 een significante verandering in de spectra na interactie met de polymeerketengroepen. Uit de wijdte van de spectra, is geconcludeerd dat polymeren die ether groepen bevatten een grotere interactie vertonen met CO2 dan polyesters. FT-IR spectroscopie is een geschikte methode analyse gebleken voor de initiële selectie van een polymeer. Met FT-IR is alleen kwalitatieve informatie verkregen over de oplosbaarheid van CO2 in de verschillende polymeren. Daarom is er een magnetische suspensie balans gebruikt om de oplosbaarheid van CO2 in polyesters, gebaseerd op propoxylaten- en ethoxylaten-bisfenol (PPB and PEB), te meten. Om de buoyancy correctie toe te voegen, wat nodig is vanwege de zwelling van het polymeer veroorzaakt door opgelost CO2, zijn er met behulp van een optische cel metingen zwelling uitgevoerd onder dezelfde experimentele condities. Bij stijgende druk en dalende temperatuur neemt de oplosbaarheid van CO2 toe voor beide polyesters. Doordat de oplosbaarheid van CO2 hoog is in PPB, is PPB als voorbeeld materiaal gebruikt voor de deeltjes productie met de EGSEG methode.

Batch experimenten zijn uitgevoerd om te testen of de PGSS methode toepasbaar is voor de deeltjes productie van PPB. Door de hoge viscositeit van PPB is het niet mogelijk geweest om het PPB-CO2 mengsel te laten expanderen met alleen het opgeloste CO2 in

129

PPB zodat een overmaat CO2 gebruikt moet worden. Er is ook een batch opstelling gebruikt om deeltjes te produceren uit vloeibaar polyethyleenglycol met verschillende molgewichten (6000 en 10000). PEG wordt voornamelijk gebruikt om geneesmiddelen die gecontroleerd gedoseerd moeten worden in te kapselen. De PGSS methode is toepasbaar op PEG doordat het een relatief lage viscositeit heeft. Het effect van verschillende parameters zoals druk, temperatuur, spuitmond diameter en molecuul gewicht op grootte deeltjes en grootte deeltjes verdeling is zeer gedetailleerd bestudeerd. Verschillende morfologie en groottes deeltjes zijn mogelijk door het veranderen van de procescondities. Een stollingsmodel is gebruikt om het effect van temperatuur en druk op de vorm van de deeltjes te verklaren. Met de PGSS methode kunnen PEG deeltjes met verschillende morfologie, zoals een geschuimde structuur of dichte structuur en verschillende groottes, geproduceerd worden die zeer gecontroleerd een geneesmiddel kunnen doseren.

Het opgeloste CO2 zorgt voor een aanmerkelijke daling van de viscositeit van het polymeer door het vergroten van het vrije volume. Omdat de viscositeit van PPB heel hoog is, is het essentieel om te weten hoe groot de daling van de viscositeit, veroorzaakt door opgelost CO2, kan zijn. Een model gebaseerd op de vrije volume theorie is gebruikt om de viscositeit in te schatten. Het model toont een aanmerkelijke daling in de viscositeit. Voor de productie deeltjes, is een continue opstelling ontwikkeld. Deze opstelling bestaat voornamelijk uit een extruder en een Kenics statische menger, eerste experimenten zijn hiermee uitgevoerd. Er is gekozen om de statische menger te gebruiken voor het mengen van het polymeer en CO2, omdat dit stabiliteit van het proces verhoogt. Bovendien is het niet mogelijk geweest om het mengsel bij hoge gas-massa verhouding goed in de extruder te mengen. In de deeltjes productie is een overmaat aan CO2 gebruikt. De expansie van het overmaat CO2 vergroot het afbreken van het vloeibare polymeer door het creëren van intense agitatie op het oppervlak. De kwaliteit van het product is ook afhankelijk van de spuitmond diameter. Hoe kleiner de spuitmond diameter des te groter de verstrekking en dus varieert de product morfologie van samengeklonterde vezels tot deeltjes naarmate de spuitmond diameter toeneemt. Er wordt betere homogeniteit verkregen door het inbrengen van een kern met gleuven voor de spuitmond. Door deze verandering ontstaan er geen bijproducten zoals deeltjes met een geschuimde structuur. Aan de andere kant is een hogere bewerkingstemperatuur nodig om de extra extensie viscositeit, veroorzaakt door de gleuven, te overbruggen.

De opstelling die gebruikt werd voor de eerste experimenten is aangepast zodat een hogere polymeervoeding en hogere druk gebruikt konden worden. Een tandrad pomp werd gekoppeld aan de extruder om een lage druk aan de extruder kant (toevoer van de pomp) te houden en een hoge druk aan de andere kant. Een SMX statische menger werd gebruikt voor betere menging van het polymeer met CO2. Verschillende experimenten werden uitgevoerd met verschillende spuitmond diameters en kernen onder verschillende procesomstandigheden. Door de verschillende parameters, zoals spuitmond diameter, kern-gleuf breedte, druk, temperatuur en gas/polymeer massa ratio te veranderen, zijn er deeltjes verkregen met verschillende morfologie, grootte deeltjes en grootte deeltjes verdeling. Zonder kern, neemt de deeltjes grootte af bij toenemende druk en afnemende temperatuur. Deze resultaten zijn toe te schrijven aan een hoge CO2 oplosbaarheid onder deze omstandigheden. Vergeleken met het vermalen van de polymeren zijn er met deze methode deeltjes verkregen met een smalle grootte deeltjes verdeling. Het stollingsmodel werd gebuikt om de invloed van temperatuur en druk op de vorm van de deeltjes te bepalen. Er is een poging gewaagd om een empirische correlatie te vinden die de grootte

130

deeltjes kan verklaren aan de hand van dimensie analyse, met verschillende parameters, zoals geometrie, afschuifviscositeit, oppervlakte spanning, druk en vloeistofsnelheid. Behalve economische en technologische voordelen heeft de EGSEG methode ook in milieu opzicht veel voordelen. Een kwalitatieve technologische evaluatie heeft aangetoond dat de ontwikkelde methode voor het continue proces een veelbelovende oplossing is voor de huidige traditioneel beschikbare methodes. Desalniettemin, is het nog steeds noodzakelijk om andere superkritische methodes te testen voordat de traditionele methodes worden vervangen door het EGSEG proces. Vooruitkijkende naar de snelgroeiende interesse voor superkritische technologieën, verschijnen er verschillende andere toepassingen voor superkritisch CO2, zoals microcellulair schuimen, extracties, polymeer modificaties en polymerisaties waar de organische oplosmiddelen vervangen worden door superkritisch CO2.

131

Acknowledgements First, I would like to thank prof. dr. ir. L.P.B.M. Janssen for his belief in my capabilities and his relentless inspiration and encouragement. Leon, I have gone through different learning phases, scientific as well as personal, as a researcher under your guidance. The freedom given by you was a vital tool for me to shape myself as an independent researcher. I am also very grateful to Dr. F. Picchioni for being my copromotor for the last two years. I appreciate his enthusiasm and the fruitful discussions, which helped me to expedite the Ph.D. work and complete it on the right time. Also, I would like to acknowledge Dr. Frank Wubbolts for his valuable suggestions and technical guidance that provided a platform to build up this research. Apart from the scientific knowledge, the financial support also played an important role in this Ph.D. work. This research would have not been possible without the financial support of STW and its commercial members. I am also very grateful to Océ-Nederland, Akzo Nobel Chemicals, and Holland Colours for the polymers provided in this work. I thank the “Beoordelingscommissie”, prof. dr. ir. P. J. Jansens, TUDelft, prof. dr. ir. J. T. F. Keurentjes, TU/e, and prof. dr. ir. H. J. Heeres, RuG for spending their precious time in reading my thesis and their scientific and general comments. Working in a different country with a different language and different culture (coffee breaks and a lunch break) was not an easy task. Though the acquaintance time was a little long, I could successfully manage to cope up with it. The technical department was always used by me as a bridge between practice and theory. I am very thankful to Laurence, Marcel, Anne, Erwin, and Marsman for their timely technical help and solutions. I would like also to acknowledge Alberta and Nijland for their contribution in terms of the analysis of polymers, which is an inseparable part of this thesis. I am also thankful to Ineke for her guidance in my first year and to Marya for her help in the official work. In the small world of the chemical engineering department, I enjoyed working with Indonesian, Italian, Indian, Chinese, Polish, Russian, and Dutch colleagues. As a master student, the contribution of Vincent to the experimental work was invaluable. I appreciate endurance of my colleagues, especially office-mates (Francesca, Jan Peter, and Zhang), to listen Hindi songs. I especially thanks to a MATLAB expert, Asaf, for his regular help in explaining and solving the MATLAB problems. I also thanks to a local and an international Dutch language specialist, Vincent and Nidal, for helping me in translating the summary. Anant and Zhang receive special attention for checking my thesis and for providing valuable suggestions in the final phase. For enthusiastic Zhang, I must say that I enjoyed non-stop working with him on different projects for the last six months. I am also thankful to Vishal Patil, TU/e, for the collaborative project and the research discussions that were useful in various aspects. I deeply acknowledge Han Scherpenkate for his timely help in verhuizen activities without excuses. In the social world of Groningen, Groningen-desis (Indians) made me feel like staying with a family. I have enjoyed both frictions and happy moments during my stay. Different foods from and different moods with all madarasis (south Indians), aai, tais, babas, dadas, bacchas, families, senior and junior kakas, and past and present raos always kept the Indian atmosphere alive even in the strong winters.

132

Finally, I want to thank my family members, a triangle of Nalawade, Bhonsale and Raorane, and my friends, Mandar, Sumit, Anand, Milind, Mehul, Anil, Navin, Sameer, Vilas, Nitesh, Bapat, and Shashank, for their constant support in climbing another step in my life in the right direction.

I want to dedicate this work to my aaji, Suwarnalata and mawashi, Sujata.

Sameer P. Nalawade Groningen, December 2005

133

Publications

[1] L.P.B.M. Janssen, Sameer. P. Nalawade, Chapter 12, Polymer extrusion with supercritical carbon dioxide, Supercritical carbon dioxide in polymer engineering, Edts, Maartje Kemmere and Thierry Meyer, Wiley-VCH (ISBN: 3-527-31092-4), Aug 2005.

[2] Sameer. P. Nalawade, L.P.B.M. Janssen, Production of polymer particles using

supercritical carbon dioxide as a processing solvent in an extruder, Proceedings of the 6th int. symp. on supercritical fluids (Versailles France, April 2003),3,1559-1563, Poster presentation..

[3] Sameer. P. Nalawade, F. Picchioni, L.P.B.M. Janssen, Prediction and measurements of

CO2-solubilities in and swelling of polyester resins using S-L equation of state (EOS), Proceedings of the 7th int. symp. on supercritical fluids (Florida USA, May 2005), N153, Poster presentation.

[4] Sameer. P. Nalawade, F. Picchioni, L.P.B.M. Janssen, The FT-IR studies of the

interactions of CO2 and polymers having different chain groups, Proceedings of the 7th int. symp. on supercritical fluids (Florida USA, May 2005), N167, Oral presentation.

[5] Sameer P. Nalawade, I. Ganzeveld, L.P.B.M. Janssen, Supercritical carbon dioxide in

polymer processing, NPT Procestechnologie (National Journal), 10(1), 18-20 (2003). [6] Sameer P. Nalawade, F. Picchioni, L.P.B.M. Janssen, A green solvent for polymer

processing polymer melts: review, Progress in Polymer Science, in press. [7] Sameer P. Nalawade, F. Picchioni, L.P.B.M. Janssen, Jan H. Marsman, The FTIR studies

of the interactions of CO2 and polymers having different chain groups, Journal of Supercritical Fluids, in press.

[8] Sameer. P. Nalawade, Vishal Patil, F. Picchioni, R. Staudt, L.P.B.M. Janssen, J. T. F.

Keurentjes, Phase equilibrium studies of sub- and supercritical CO2 in polyester resins: experiments and theory, Polymer Engineering science, accepted.

[9] Sameer P. Nalawade, F. Picchioni, L.P.B.M. Janssen, An engineering approach towards

the production of micon-size particles from PEGs using supercritical CO2, Manuscript in preparation for Chemical Engineering Science, submitted.

[10] Sameer P. Nalawade, Vincent H. J. Nieborg, F. Picchioni, L.P.B.M. Janssen, Prediction of

the viscosity reduction of PPB due to dissolved CO2 and an elementary approach in the supercritical CO2 assisted continuous particle production, Powder Technology, submitted.

[11] Sameer P. Nalawade, F. Picchioni, L.P.B.M. Janssen, Vital roles of different processing

parameters on the continuous polymer particles production using supercritical CO2 as a processing solvent, Journal of Supercritical Fluids, submitted.

[12] Sameer P. Nalawade, Youchun Zhang, F. Picchioni, A.A. Broekhuis, L.P.B.M. Janssen,

Microcellular foaming of recyclable PET by using supercritical carbon dioxide, Journal of Supercritical Fluids, manuscript in preparation.

university of groningen polymer melt micronisation using supercritical … · 2016-03-07 ·...

Documents