Ionic liquid performance in pilot plant contactors for aromatics extraction

(1)

Ionic liquid performance in pilot plant contactors for aromatics

extraction

Citation for published version (APA):

Onink, S. A. F. (2011). Ionic liquid performance in pilot plant contactors for aromatics extraction. Technische Universiteit Eindhoven. https://doi.org/10.6100/IR715602

DOI:

10.6100/IR715602

Document status and date: Published: 01/01/2011 Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

(2)

Ionic Liquid Performance in Pilot Plant

Contactors for Aromatics Extraction

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de rector magnificus, prof.dr.ir. C.J. van Duijn, voor een

commissie aangewezen door het College voor Promoties in het openbaar te verdedigen op woensdag 31 augustus 2011 om 16.00 uur

door

Steven Adrianus Ferdie Onink

(3)

Dit proefschrift is goedgekeurd door de promotor:

prof.dr.ir. A.B. de Haan

Copromotor:

dr.ir. G.W.Meindersma

A catalgue record is available from the Eindhoven University of Technology Library.

ISBN: 978-90-386-2538-6

Printed by: Gildeprint Drukkerijen, Enschede, The Netherlands Cover design: Ferdy Onink

The research described in this work has been financially supported by BASF.

(4)

Voor pa en ma

(5)

Voorzitter Prof. dr. ir. J.C. Schouten

Promotor Prof. dr. ir. A.B. de Haan

Copromotor Dr. ir. G.W. Meindersma

Leden Prof. Dipl.-Ing. Dr. techn. H.J. Bart

Prof. dr. G.J. Witkamp

Prof. dr. ir. M. van Sint Annaland

Reserve-lid Prof. dr. J. Meuldijk

Adviseurs Dr. U. Vagt

Dr. ir. J. van der Schaaf

“…if I could start again a million miles away I would keep myself I would find a way…”

Johnny Cash, Hurt, (2002)

(6)

6.4.2 Regeneration of [3-mebupy][DCA] ... 137 6.5 Conclusions ... 144 6.6 Nomenclature list ... 144 6.7 References ... 145 7 Contactor comparison ... 147 7.1 Abstract ... 147 7.2 Introduction ... 147 7.3 Experimental Section ... 148 7.3.1 Materials ... 148 7.3.2 Analysis ... 149 7.3.3 Experimental Setup ... 149

7.3.4 Column Operation and Characterization ... 152

7.4 Results and discussion ... 159

8 Conclusions and Outlook ... 177

8.1.1 Applicability of RTILs as solvent for extraction ... 177

8.1.2 Contactor comparison ... 179

8.2 Outlook ... 179

Appendices ... 181

Appendix A: Data physical properties ... 181

Appendix B: Process scheme of pilotplant RDC ... 185

Appendix C: Overview of sample ports on RDC ... 186

Appendix D: Aspen Custom Modeler model ... 187

Appendix E: Aspen Custom Modeler output ... 201

Dankwoord ... 207

Publications ... 209

(9)

Summary

The main objectives of this study were an investigation into the applicability, in this case extraction capacity and equipment performance, of room temperature ionic liquids as solvent in the extraction of aromatics from aliphatics and a comparison of three types of contactors (a rotating disc contactor (RDC), a Kühni contactor and a pulsed disc and doughnut column (PDDC)) for this extraction. The separation of aromatic hydrocarbons (benzene, toluene, ethylbenzene and xylenes) from C4 – C10 aliphatic hydrocarbon mixtures is challenging since these hydrocarbons have boiling points in a close range and several combinations form azeotropes, ruling out conventional distillation processes. The contactors, RDC, PDDC and Kühni were selected since they are amongst the most commonly used extractors, or are most promising for (aromatics) extraction in the (petro)chemical industry. Early research showed that ionic liquids are promising solvents for the extraction of aromatics from aliphatics, but these studies are mainly based on a thermodynamic approach and only a conceptual process design was suggested. However, successful introduction of RTILs into extraction operations also requires knowledge on their physical properties, hydrodynamics, mass transfer characteristics, stability after long term usage and, since the costs of replacement of lost solvent plays an important role, recovery of the RTIL from the raffinate stream.

Model

To fully understand the behavior of Room Temperature Ionic Liquids as solvent for aromatics extraction, a theoretical model was developed using existing theory on rotating disc contactors. This model described the operational and mass transfer characteristics of an RDC for the extraction of toluene from n-heptane with RTILs. Mass transfer characteristics were modeled with a differential axial dispersion model, with correlations for the mass transfer coefficient and the axial dispersion coefficient. The modeled hydraulic characteristics covered the Sauter mean diameter, the hold-up of the dispersed phase and the operational window, being correlated using physical properties, operational parameters, and geometrical characteristics of the column. Since all applied equations were originally derived for conventional solvents, corrections might be necessary, which need to be confirmed by pilot experiments. All equations are based on the physical properties and operational parameters of the system, geometrical characteristics of the column and internals as well as certain fit parameters, describing the influence of each variable.

(10)

Physical properties and Liquid-Liquid Extraction Equilibiria (LLE)

The input of above mentioned model consisted of physical properties and data on the liquid-liquid extraction equilibrium. Density data were correlated well using a linear relation for the influence of the concentration of solute, whereas the volumetric thermal expansion coefficient is used to describe the temperature influence on the system. The absolute values of the relative error (AARE’s) varied between 0.10 and 0.76%. The density of sulfolane was the largest, followed by that of [4-mebupy]BF4 and followed by that of [3-mebupy][DCA]. Viscosity data were correlated using the Nissan and Grunberg equation for the influence of concentration, taking into account the influence of temperature on the binary interaction parameter. The Vogel equation was used to accurately describe the influence of temperature on the viscosity of the pure solvents, leading to AARE’s of 0.30 to 4.03%. The viscosities of both investigated ionic liquids exceeded the viscosity of the commonly used solvent sulfolane, but with increasing toluene content the viscosities of both RTILs decreased dramatically. Interfacial tension data were fitted with the Szyzkowski equation and the Jasper equation for the temperature influence on the binary system. The influence of concentration on the ternary systems was described via the concentration of toluene in the organic phase, which resulted in good fits. An influence of the temperature on the Szyzkowski coefficients for [3-mebupy][DCA] was observed, but for [4-mebupy]BF4, this influence was not clearly present. The resulting AARE’s were between 2.5 and 3.2%, hence indicating that a good representation of the interfacial tension for all systems was obtained using this approach.

LLE phase compositions for the the systems n-heptane + toluene + [4-mebupy]BF4 and n-heptane + toluene + [3-mebupy][DCA] were determined at 313 K. Regarding extraction capacity, [3-mebupy][DCA] outperformed [4-mebupy]BF4 and sulfolane when applied as solvent for the extraction of toluene from a mixture of toluene and n-heptane with a toluene content below 40 wt%. The weight based distribution coefficient decreased from 0.45 to 0.34 and from 0.32 to 0.25 for [3-mebupy][DCA] and [4-mebupy]BF4, respectively and increased from 0.26 to 0.34 for sulfolane for concentrations of 6 to 40 weight percent toluene in n-heptane. The selectivity decreased from 65 to 29, from 30 to 13 and from 25 to 11 for [3-mebupy][DCA], sulfolane and [4-mebupy]BF4, respectively.

Solvent comparison

The use of the ionic liquids [4-mebupy]BF4 and [3-Mebupy][DCA] as solvents for the extraction of toluene from n-heptane on a pilot plant rotating disc contactor has been investigated and benchmarked against the conventional solvent sulfolane. It has been found that both RTILs could be applied as solvent for the extraction of toluene from n-heptane. Furthermore, the studied hydrodynamic parameters, drop size, hold-up and operational area, indicated that the use of the ionic liquids [4-mebupy]BF4 and [3-mebupy][DCA] as extraction solvents was not limited by their higher viscosity of 80 mPa s and 20 mPa s for pure [4-mebupy]BF4 and [3-mebupy][DCA], respectively. The pure IL viscosities

(11)

decreased with increasing toluene content in the ionic liquid phase to 25 mPa s and 12 mPa s for a toluene content of 15 wt%, whereas the viscosity of sulfolane only decreased from 8 to 6 mPa s.

At fluxes of 3 to 8 m hr-1_{, the resulting drop sizes ranged from 1.1 to 0.6 mm, for the} ILs as well as sulfolane at rotor speeds between 200 to 800 rpm.

At first sight it seemed that unexpected behavior for the hold-up was observed in experiments when the RTILs were applied as solvent: the hold-up decreased with increasing rotor speed, whereas an increase was expected, because of formation of smaller droplets (and consecutively a smaller down going velocity) with increasing energy input. This phenomenon could be explained by the existence of three operational regimes for a rotating disc contactor; which clearly depend on flux and on rotor speeds.

The operational window is largest for [4-mebupy]BF4 and smallest when [3-mebupy][DCA] is used as solvent. This is caused by a higher interfacial tension, smaller viscosity and lower density difference when [3-mebupy][DCA] is applied as solvent.

The mass transfer capacities, expressed in HETS, were comparable for [3-mebupy][DCA] and sulfolane with HETS of 1.5 m and 4.5 m for [4-mebupy]BF4 for the extraction of 10 wt% toluene from n-heptane at 313K. It was concluded that the overall performance of [3-mebupy][DCA] exceeded the extraction capacities of sulfolane and [4-mebupy]BF4, because the highest interfacial area was obtained when [3-mebupy][DCA] was used as solvent.

After optimization of the hydrodynamic behavior parameters, which were selected on basis of a sensitivity analysis, by minimizing the AARE’s and with a correction for the diffusion coefficients of the RTILs, the resulting model consisted of equations capable to describe the Sauter drop diameter, hold-up, operational regimes and mass transfer coefficients resulting in a good qualitative description of the hydrodynamic behavior and overall separation performance.

Solubilities of RTILs in aromatic/aliphatic mixtures

Since also recovery of the solvent plays an important role in the design of an extraction process, because of costs and environmental aspects, an ion chromatographic (IC) method for the quantification of the RTILs concentrations in aromatic solvents has been developed. With an analysis time of only 20 minutes, linear standard curves for both RTILs (R2_{varying from 0.9980 to 0.9998), precisions (relative standard deviations) being} better than 2.9%, accuracy ranging from 98.5 to 105.4% and finally a limit of detection for both RTILs in all solvents varying between 0.7 and 2.6 mg kg-1_{, the method was proven to} be simple and efficient with excellent accuracy and precision. Results for the solubilities of two RTILs, [4-mebupy]BF4 and [3-mebupy][DCA], in four aromatic solvents and in mixtures of n-heptane and toluene showed that with an increase of the aromatic character of the solvent (ethylbenzene ≈ o-xylene < toluene < benzene) more RTIL dissolves. In every solvent or mixture of n-heptane and toluene used, but most notable for benzene, more

(12)

[4-mebupy]BF4 is dissolved than [3-mebupy][DCA]. This latter effect is primary caused by the nature of the anion. Below concentrations of 50 weight percent toluene in n-heptane for both RTILs, the solubilities were below the detection limits (±1 mg kg-1_).

Recovery and regeneration

When applying ionic liquids as solvent for the extraction of toluene from n-heptane also the recovery of these solvents from the raffinate, n-heptane stream, and the long-term stability play a major role. Although the solubility of [3-mebupy][DCA] in toluene/n-heptane mixtures was found below the detection limit, in the raffinate stream higher concentrations were found due to entrainment. For the recovery of the ionic liquid from the raffinate phase, a mixer settler with a coalescer with water as solvent was suggested, because of the high affinity for ionic liquids indicated by a distribution coefficient above 56,000. However, water has a negative influence on the extraction performance of [3-mebupy][DCA] and should therefore be totally removed.

Except for some colorization, the physical properties and extraction capacity of [3-mebupy][DCA] after extensive usage (approximate 2000 hrs) and regeneration for more than 1200 hrs at elevated temperatures (up to 400 K) were not changed. This was verified by comparing the density, viscosity and interfacial tension of new, fresh, [3-mebupy][DCA] and after extensive usage/regeneration. Furthermore, hydrodynamic and mass transfer profiles ‘before’ and ‘after’ were compared. It was shown that an improvement of the extraction capacity after extensive usage was due to a removal of an impurity, present in ‘fresh’ [3-mebupy][DCA].

Contactor comparison

In industry, the most commonly used extractor for the separation of aromatics from aliphatics is the rotating disc contactor. Other contactors like the pulsed disc and doughnut column and the Kühni contactor are, to a lesser extent, also used for this type of separation. It had been found that the extraction of toluene from n-heptane is successful for all three contactors. Furthermore, the studied hydrodynamic behavior indicated that for the PDDC the largest operational window and the largest hold-up were obtained. The mass transfer capacity, expressed in HETS, was also largest for the PDDC (HETS = 1.2 m), since for this column the highest specific interfacial area (0.69 mm-1_{) was achieved as result of a larger} hold-up. Hence, the overall performance of the PDDC exceeded the extraction capacities of the RDC and Kühni contactor, which could be explained by the fact that in the PDDC the energy input was distributed more homogeneous than when the RDC or Kühni were applied.

(13)

1

1.

pro res wi of pet sep ba Ha Ho op pro aro res be are in the the pla Do ion sul sel for hy det sep fro

Genera

.1 Overvie

Previous stu omising solvents searches mainly ith subsequent an a suitable ion trochemical stre paration behavio sed on a thermo ansmeier suggest owever, successfu erations also r operties, hydrody As Anjan5 omatics extractio search of the beh cause of the phy e commonly beli aromatics extra e results of an ext e extraction of a ants: a rotating d oughnut Column nic liquids, [4-m lfolane® were us The contac lected since they r (aromatics) extr ydrodynamics (dr termine the co paration toluene/ om 10 to 75 wt%

l Introduc

ew

udies1, 2_{, demonst} s for the extrac

focused on extr n RDC evaluatio nic liquid for th

ams, followed b r of an RDC2_{. T} dynamic approac ted a rather con ful introduction o equires knowled ynamics, mass tra

noted, implem on processes req havior of latter s ysical properties ieved to hamper ction. Therefore tensive theoretica aromatics from a disc contactor (R n (PDDC) and a mebupy]BF4 and sed as solvents. ctors, RDC, PD

are amongst the raction in the (pe rop size, hold-up olumn performan

/n-heptane mixtur %. The results reg

ction

trated that Room ction of aromat action of aromat n1_{, and on the se} he extraction o by a description These studies are

ch and Meinders nceptual process of RTILs into ex dge on their p ansfer characteris mentation of RT quires a more th solvents in real p of ionic liquids usage of latter s , this research p al and practical s aliphatics in thre RDC), Pulsed D a Kühni Column [3-mebupy][DCA DDC and Kühn most commonly etro)chemical ind p and operational nce4_{. As mode} res were chosen, garding the hydro

m Temperature Ion tic hydrocarbons tics from aliphat election of three n of the mainly sma and design. xtraction physical stics3, 4 TILs in horough practice, , which solvents presents study on ee pilot Disc and n. Two A], and ni were y used extractors, dustry1, 3, 4, 6-9_{. In} l window) and m el components , in which the to odynamic behavi Figure column from di solvent raffinate

nic Liquids (RTI s. and stability. tics with ionic li

, or are most prom the extraction pr mass transfer effi

for aromatic/ali luene content is ior of the column

1-1: Liquid-Liq , C as component t ispersed to continuo , F as feed, E as e e flow IL) are Both iquids, mising rocess, ciency iphatic varied ns and quid extraction to be transferred ous phase, S as extract and R as

(14)

mass transfer profiles over the columns are benchmarked against the commonly used solvent Sulfolane®10_.

Beneath the capacities and the hydrodynamic behavior of the RTILs extraction system, also attention is paid to the stability of ionic liquids after long term usage and/or regeneration and the feasibility of recovering RTILs from the raffinate stream with water.

In this chapter first some fundamentals of liquid-liquid extraction will be discussed, followed by an overview of current aromatics extraction processes. An introduction into the investigated contactors, RDC, Kühni contactor and PDDC will be given as well as some background information about ionic liquids. This chapter will be concluded with the objectives and the structure of this thesis.

1.2 Liquid-Liquid Extraction

Liquid–liquid or also known as solvent extraction is the separation method of choice where distillation fails, e.g., for azeotropic mixtures or temperature sensitive components. Separation is achieved by adding a liquid solvent phase (S) to the original liquid, feed (F) carrying the component (C) to be extracted. One of the phases must be dispersed into droplets in the other, continuous phase to achieve a sufficiently large mass-transfer interface. The feed and solvent flows leave the contactor as the raffinate (R) and extract (E) phase, respectively. Extraction is performed in mixer–settler equipment or extraction columns, as can be seen in Figure 1-1,where the mass-transfer direction is chosen from the dispersed to the continuous phase. These extractors are frequently equipped with rotating internals or pulsators for energy input to positively influence droplet size3_.

Important parameters in extraction are defined as the distribution coefficient8, 11, 12 and selectivity8_{, see Equations 1-1, 1-2and 1-3, respectively.}

1-1

1-2 With f as main component of the feed flow.

(15)

1.2.1 Aromatics Extraction

The separation of aromatic hydrocarbons (benzene, toluene, ethylbenzene and xylenes) from C4 – C10 aliphatic hydrocarbon mixtures is challenging since these hydrocarbons have boiling points in a close range and several combinations form azeotropes, ruling out conventional distillation processes. Therefore several extraction or advanced distillation processes are designed, depending on the aromatic content13_{, see} Table 1-1:

Table 1-1: Type of processes for the separation of aromatics from

aliphatics

Separation aromatics - aliphatics Aromatic content Proces < 20 % - 20 – 65 % Liquid-Liquid extraction 65 – 90 % Extractive distillation > 90 % Azeotropic distillation

The extractive separation of aromatic from aliphatic hydrocarbons is thus an important application in the petrochemical industry14_{. Current processes mostly use polar} solvents15, 16_{such as sulfolane (UOP}10_{, SHELL}17_{), diethylene or triethylene glycols}18_and n-methyl pyrrolidone (Lurgi)19, 20_.

As suggested by Bailes15_{as model for the extraction of aromatics from aliphatics the} separation of toluene from n-heptane is chosen. As solvent benchmark Sulfolane17_is adopted.

1.2.2 Liquid-Liquid Extraction Equipment

As Van Delden wrote in his thesis21_{, the first commercially available, continuous} extraction equipment was nothing more than a series of mixer-settlers and open-spray columns, originating from a column, patented in 19358_{. It is believed that the first type of a} packed liquid-liquid extraction column was based on the design of a gas absorption and gas washer, in use already in the nineteenth century22_{. Since then over 25 types of extractors} are in use worldwide, mainly because for every specific type of extraction a different contactor is developed and designed by the companies23_{themselves, see Figure 1-2.}

(16)

Figure 1-2: Overview of commercial extractors8

Important criteria for the selection of a proper contactor are: • Stability and residence time

• Settling characteristics • Number of stages required • Capital and maintenance costs • Available space

• Throughput.

The focus of the choice for an extractor lies mainly on the settling characteristics (physical properties of the system) and on the number of stages required to obtain sufficient separation performance8, 24_{. Thus, when describing the performance of an extraction} column, two criteria are important: capacity25-33_{and mass transfer efficiency}25, 29, 34-37_. These two are, however, related to each other, e.g. by the influence of the drop diameter on both. The rate of mass transfer in an extractor can be expressed in general terms as a function of drop sizes, axial dispersion and diffusion.35, 38-41_{In general, drop diameters are a} function of physical properties like density (or density differences between the dispersed and continuous phases), viscosity, interfacial tension and the power input.

(17)

1.2.3 Rotating Disc Contactor

The rotating disc contactor (RDC) was developed by Shell28, 35_{in de mid 1950’s, and has since} found numerous applications. Some hundreds of RDC’s are present in use worldwide, ranging in diameter from 1 to 4.5 m. It has the advantage, that because extensive research into the performance of an RDC is conducted, well proven relationships29, 33, 39, 42-45 and scale-up rules are developed, thus providing the opportunity of scaling up a pilot plant RDC with a diameter of 64 mm directly to a diameter of 4-4.5 m.35

Figure 1-3 schematically shows a rotating disc contactor. The column consists of vertical segments, in which horizontal stator rings are installed. These stator rings are flat plates with a central opening. In the middle of the compartments formed by the stator rings rotor disks are installed. The rotor disks are flat plates attached to a central shaft and driven by an electric motor. Above the top stator ring and below the bottom

stator ring settling compartments are installed. The feed and solvent inlets are, in general, arranged tangentially, in order not to disturb the flow pattern in the inlet compartments.

In case of above (1.2.2) described capacity several regimes could be limiting for the operation of a rotating disc contactor. At low rotor speeds flooding limits column operation. At high rotor speeds flooding, due to high agitation, phase inversion or entrainment is limiting28, 35

1.2.4 Pulsed Disc and Dougnut Column

The pulsed disc and doughnut column (PDDC) is based on the pulse packed and pulse sieve plate column. In industry at first it was designed for metal extraction 46-48_{, but} research by Van Delden21_{indicated it might also be used for the extraction of caprolactam.} Research by Lerner et al.9_{showed its application for the extraction of aromatics from} mineral oil by furfural. PDDCs with diameters ranging from 0.5 m to 3.0 m and an effective height of up to 35 m have been used in industrial applications49_.

The PDDC, see Figure 1-4, consists of a large diameter vertical pipe filled alternatively with disc and annular shaped (doughnuts) baffles which facilitate contact between the immiscible liquids passing through the column. The heavy (aqueous) phase

(18)

enters through a disperser at the top of the column and the light (organic solvent) phase enters through a similar device near the bottom.

A settler at each end of the column permits the liquids to coalesce and be decanted separately. When the solvent phase is continuous the interface between the phases is in the lower settler and when the aqueous phase is continuous, it is in the upper settler. The columns are pulsed by blowing air at the required amplitude and frequency into the pulse leg. The air pressure is controlled to provide pulses of the required amplitude in the column while the frequencies of the pulses are controlled using three-way solenoid valves.

The operating regimes are generally described as mixer-settler, dispersion and emulsion.50, 51_The mixer-settler regime is limited by flooding due to a too low amount of energy applied. The other two regimes are limited either by flooding due to a too small relative velocity between the phases, phase inversion in which the original continuous phase becomes dispersed or when the maximal allowable entrainment level is reached.

1.2.5 Kühni Column

The Kühni column was described in 1965 by Mögli6, 22_{. Its design resembles that of} an RDC, but instead of flat rotors, turbine rotors are mounted on the central shaft, as can be seen in Figure 1-5.

Worldwide over 300 Kühni columns are in operation ranging from lab scale (diameters of 60 mm, length of 0.5 m)) to industrial applications (diameter 2.5 m)6_.

Compared to RDCs shaft speeds are relatively low; in the range of 80-200 RPM for pilot columns and 5-20 RPM in industrial applications, thus leading to a moderate power consumption of the shaft.

Figure 1-4: Pulsed disc and doughnut

(19)

Figure 1-5: Internals Kühni column, schematically(left), in operation52_(right)

1.3 Room Temperature Ionic Liquids

Ionic liquids (ILs) are a relative new class of solvents. They were first suggested to be applicable in “green” and industrial chemistry in the mid-1990’s, although already in 1914, Paul Walden described the physical properties of ethylammonium nitrate.53 Even earlier, in the mid-19th century, an ionic liquid was described by chemists as a separate phase occurring in Friedel-Crafts synthesis, the so called “red oil”. It was not until the availability of NMR spectroscopy that this substance could be identified as such.54

In most literature53-56_{, the definition of ILs is that they are salts which are liquid at} temperatures below the boiling point of water. Many are liquid at room temperature and even below; sometimes explicitly called Room Temperature Ionic Liquids. A main feature not contained in the definition is that ionic liquids consist of bulky organic cations and (in)organic anions, see Figure 1-6 for the structural formulas of 4-methyl-N-butylpyridinium tetrafluoroborate and 3-methyl-N-4-methyl-N-butylpyridinium dicyanamide.

Besides their additional advantage of a low melting point they share numerous advantageous properties53-55, 57-61_{with their high temperature inorganic molten salt} counterparts like a wide liquid range (~300K), negligible vapor pressure, good thermal, chemical and electrochemical stability, good solubility in water (although this does not count for not all RTILs) and non-flammability. Although not abundant, there are several commercial processes developed because of the attractive properties that ionic liquids possess.62

In principle, RTILs have the same kind of combination between a cation and anion as an inorganic salt like NaCl for example. The dominant force between such a salt is coloumbic in nature. However, in RTILs, the asymmetric organic ions have delocalized charges with less coloumbic attraction between them. This decrease in attraction between

(20)

the cations and anions lead to the desired lower melting points for RTILs compared to normal salts. There is another difference between an ionic solution and an IL. When an ionic salt (NaCl for example) is dissolved in a polar solvent like water, the crystal lattice of the salt is broken up to form an ionic solution. This solution contains besides ions also the solvent molecules. These solvent molecules surround the positive and negative ions. This is very different compared to an RTIL, which only consists of ions.

Figure 1-6: Structural formulas of [4-mebupy]BF4 (left) and

[3-mebupy][DCA] (right)

The chemical properties of an RTIL are dependent on the combination of cation and anion. Because the ions can be tailored by using different combinations of cation and anion, varying the length and degree of branching of groups attached to it, the properties exhibited by the RTIL can be tailored as well. More than 1018_{number of combinations can be made.}63 Theoretically, a specific RTIL could be designed for a specific application; the reason that RTILs are often called designer solvents.

An overview of possible (dis)advantageous properties of RTILs is presented in Table 1-2 below. Most of the disadvantageous properties are due to the nature of the anions: corrosiveness might occur if the anions consist of only halogens, like I3-_{or Cl}-_, moisture instability is caused by halogen compounds like AlCl4-_{or PF6, additionally these} halogens contribute to the significant higher viscosity of RTILs-_.64, 65_{Water solubility is a} profitable property, because it enables an easy recovery of the RTIL in a process.1_Ionic liquids containing, e.g. the anion Tf2N-_{and comparable components, are water-insoluble,} and, additionally exhibit a significant high molecular weight. A high molecular weight is an unfavorable property, since the high molecular weight causes the need for large quantities (tons) of ionic liquid, which negatively influences process economics and economic feasibility of the process.66

(21)

Table 1-2: Advantageous and disadvantageous properties of RTILs65

(Physical) properties RTILs

Advantageous Disadvantageous

non-volatile corrosive

tailoring possible viscous

soluble in water moisture instability

high molecular weight

insoluble in water

1.4 Objectives and Structure of Thesis

As mentioned in the first paragraph of this introduction, the aim of this work is to investigate whether ionic liquids can be applied as solvent for the extraction of aromatics from aliphatics. Commonly adopted ideas that ionic liquids are not suitable for aromatics extraction because of their physical properties, (un)stability and problems concerning recovery of these solvents from raffinate streams, will be refuted.

Although in general as contactor for aromatics extraction an RDC17, 19, 35_is suggested and applied other contactors like the Kühni (IFP6_{) are also used, and further,} Lerner et al.9_{suggested the application of a PDDC. Therefore, the second aim of this thesis} will be a research on the performance of a pilot plant RDC, Kühni contactor and PDDC for the extraction of toluene from n-heptane with an ionic liquid ([3-mebupy][DCA]) as solvent. Investigated will be the hydrodynamic parameters, drop size, hold-up and operational region and the mass transfer performance.

In Chapter 2 a mathematical model, based on plug flow with axial dispersion, will be developed describing the hydraulic characteristics and mass transfer performance of a RDC for the extraction of toluene from n-heptane. The suggested models are developed based on existing column independent equations and equations derived for RDC, when conventional solvents were applied. In this chapter, the mathematical model for mass transfer will be described, followed by the for this model required parameters such as the holdup, the axial dispersion coefficients, the mass transfer coefficient and the specific surface area, described by the Sauter mean diameter. To complete the hydraulic characteristics, also the operating regime is discussed. Latter parameters are, among other things, depending on the physical properties of suggested solvents.

Therefore in Chapter 3, next to phase composition data, physical properties, being density and viscosity of the separate liquid phases and interfacial tension of the liquid-liquid system, are investigated. Although the data for pure liquid-liquids and pure liquid-liquid-liquid-liquid systems have been reported earlier67_{, the influence of toluene on these physical properties is} unknown and therefore examined in this chapter. Furthermore, density and viscosity data of

(22)

the separate phases and interfacial tension data of the systems studied will be determined and correlated.

In Chapter 4, firstly, the hydrodynamic behavior, such as drop size, hold-up and the operational window, will be compared and evaluated with the models developed in Chapter 2. Hereafter, mass transfer profiles and the resulting HETS of the extraction of toluene from

n-heptane with the selected solvents in the RDC will be determined to asses the capacities

of the used solvents. Finally, the extraction of toluene from n-heptane will be modeled with an axial dispersion model. This model will be optimized for all solvents.

Since it is necessary to have a very low amount of RTILs in the raffinate, because of fouling and cost efficiency1_{, a very sensitive and robust analytical method is required to} quantify the solubilities of these solvents in organic streams. Therefore in Chapter 5, a chromatographic method that can be used for latter quantification will be developed and used to estimate possible losses in the raffinate stream.

Previous research1, 2_{concentrated on the recovery of extracted aromatics from the} ionic liquid, solvent, stream. Important for the development of a conceptual design is the recovery of the solvent, ionic liquid, from the raffinate stream. Chapter 6, therefore, focuses on the recovery of the ionic liquid from the raffinate with the use of water. Introducing water as the back extraction agent for the removal of RTILs from the raffinate stream could have an influence on the extraction efficiency of the RTIL, in this case the distribution coefficient and selectivity. Hence, latter influence will be investigated in this chapter. Because for the application of RTILs as solvent in the extraction processes, not only the capacity or performance is important, but most certain also the stability of used ionic liquid, the stability of the solvent after long-term usage and regeneration will be investigated and reported in this chapter. To test this stability, phase composition data, hydrodynamic performance and mass transfer characteristics of the used ionic liquid will be determined again and compared with the results as obtained in Chapter 4.

Although extraction of aromatics from aliphatics is mostly carried out with an RDC as contactor, also other contactors like the Kühni contactors and PDDC can be applied. Hence in Chapter 7, the hydrodynamic behavior, such as drop size, hold-up and the operational window, of mentioned contactors will be evaluated. Furthermore, in this chapter mass transfer profiles and the resulting HETS of the extraction of toluene from n-heptane with [3-mebupy][DCA] in the RDC, Kühni contactor and PDDC will be compared to asses the capacities of the used contactors.

Finally, this thesis will be concluded in Chapter 8, presented together with a future outlook and recommendations for further research.

(23)

1.5 References

1. Meindersma, G. W., Extraction of Aromatics from Naphtha with Ionic Liquids : from Solvent Development to Pilot RDC Evaluation. University of Twente, Enschede, 2005.

2. Hansmeier, A., Ionic Liquids as Alternative Solvents for Aromatics Extraction. Eindhoven University of Technology, Eindhoven, 2010.

3. Müller, E.; Berger, R.; Blass, E.; Sluyts, D.; Phennig, A., Liquid-Liquid Extraction Equipment. In Ullman's Encyclopedia of Industrial Chemistry (online version), Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

4. Godfrey, J. C.; Slater, M. J., Liquid-liquid extraction equipment. John Wiley and Sons Ltd.: Chichester, 1994.

5. Anjan, S. T., Ionic liquids for aromatic extraction: Are they ready? Chemical

Engineering Progress 2006, 102, (12), 30-39.

6. Mögli, A.; Bühlmann, U., The Kühni Extraction Column. In Handbook of Solvent

Extraction, Reprint 1991 ed.; Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds. John

Wiley & Sons: New York, 1983; pp 441-447.

7. Schweitzer, P. A., Handbook of separation techniques for chemical engineers.

McGraw-Hill: New York, 1997.

8. Stevens, G. W.; Lo, T. C.; Baird, M. H. I., Liquid-Liquid Extraction. In Kirk-Othmer

Encyclopedia of Chemical Technology (online version), John Wiley & Sons Inc.,

2007.

9. Lerner, O.; Efraim, A.; Kleinberger, R.; Chitterer, R.; Azvolinsky, I., Mineral Oils Aromatics Extraction in a Bateman Pulsed Column, International Solvent Extraction

Conference, Beijing, China, 2005; pp 746-753.

10. https://www.uop.com/objects/55%20Sulfolane.pdf, accessed March 2006.

11. Berthelot, Sur le lois qui président au partage d'un corps entre deux dissolvants (théorie). Ann. Chim. Phys. 1872, 4, (26), 408-417.

12. Berthelot; Jungfleisch, Sur le lois qui président au partage d'un corps entre deux dissolvants (experiences). Ann. Chim. Phys. 1872, 4, (26), 396-407.

13. Arpe, H. J.; Hawkins, S., Industrial Organic Chemistry. John Wiley & Sons: 2004; p 313-336.

14. Hamid, S. H.; Ali, M. A., Comparative study of solvents for the extraction of aromatics from naphtha. Energy Sources 1996, 18, 65-84.

15. Bailes, P. J., Aromatics-Aliphatics Separation. In Handbook of Solvent Extraction, Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds. John Wiley & Sons: New York, 1983; pp 519-521.

(24)

16. Bailes, P. J., Other extraction Processes. In Handbook of Solvent Extraction, Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds. John Wiley & Sons: New York, 1983; pp 547-548.

17. Kosters, W. C. G., Sulfolane Extraction Processes. In Handbook of Solvent

Extraction, Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds. John Wiley & Sons: New

York, 1983; pp 541-545.

18. Vidueira, J. A., Union Carbide Tetra Process. In Handbook of Solvent Extraction, Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds. John Wiley & Sons: New York, 1983; pp 531-539.

19. Krishna, R.; Goswami, A. N.; Nanoti, S. M.; Rawat, B. S.; Khanna, M. K.; Dobhal, J., Extraction of aromatics from 63-69oC naphtha fraction for food grade hexane production using sulpholane and NMP as solvents. Indian Journal of Technology

1987, 25, 602-606.

20. Müller, E., NMP (Arosolvan) Process for BTX Separation. In Handbook of Solvent

Extraction, Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds. John Wiley & Sons: New

York, 1983; pp 523-529.

21. van Delden, M. L. Caprolactam Extraction in a Pulsed Disc and Doughnut Column wit a Benign Mixed Solvent. University of Twente, Enschede, 2005.

22. Hampe, M. J.; Hartland, S.; Slater, M. J., Historical Background. In Liquid-Liquid

Extraction Equipment, Godfrey, J. C.; Slater, M. J., Eds. John Wiley & Sons Ltd.:

Chichester, England, 1994; pp 7-14.

23. Baird, M. H. I., Solvent extraction. The challenges of a 'mature' technology.

Canadian Journal of Chemical Engineering 1991, 69, (6), 1287-1301.

24. Pratt, H. R. C., Computation of Stagewise and Differential Contactors: Plug Flow. In Handbook of Solvent Extraction, Reprint 1991 ed.; Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds. John Wiley & Sons: New York, 1983; pp 151-198.

25. Strand, C. P.; Olney, R. B.; Ackerman, G. H., Fundamental aspects of rotating disk contactor performance. AIChE Journal 1962, 8, (2), 252-261.

26. Bailes, P. P., Hydrodynamic behavior of packed, rotating-disk and Kühni liquid-liquid-extraction columns. Chemical Engineering Research & Design 1986, 64, (1), 43-55.

27. Kamath, M. S.; Rau, M. G. S., Prediction of operating range of rotor speeds for rotating disc contactors. Canadian Journal of Chemical Engineering 1985, 63, (4), 578-584.

28. Korchinsky, W. J. J., Rotating Disc Contactors. In Liquid-Liquid Extraction

Equipment, Godfrey, J. C.; Slater, M. J., Eds. John Wiley & Sons Ltd.: Chichester,

England, 1994; pp 247-275.

29. Laddha, G. S.; Degaleesan, T. E.; Kannappan, R., Hydrodynamics and mass transport in rotary disk contactors. The Canadian Journal of Chemical Engineering

(25)

30. lstrok; Odzimierz, N.; Krzysztof, K., Recommended operating range for a rotating disc contactor. Chemical Engineering & Technology 1995, 18, (1), 63-67.

31. Onink, F.; Drumm, C.; Meindersma, G. W.; Bart, H.-J.; de Haan, A. B., Hydrodynamic behavior analysis of a rotating disc contactor for aromatics extraction with 4-methyl-butyl-pyridinium·BF4 by CFD. Chemical Engineering

Journal 2010, 160, (2), 511-521.

32. Bastani, D.; Shahalami, S. M., New approach in the prediction of RDC liquid-liquid extraction column parameters. Chemical Engineering and Technology 2008, 31, (7), 971-977.

33. Jeffreys, G. V.; Al-aswad, K. K. M.; Mumford, C. J., Drop-Size Distribution and Dispersed Phase Hold-up in a Large Rotating Disc Contactor. Separation Science

and Technology 1981, 16, (9), 1217 - 1245.

34. Al-Aswad, K. K.; Mumford, C. J.; Jeffreys, G. V., The application of drop size distribution and discrete drop mass transfer models to assess the performance of a rotating disc contactor. AIChE Journal 1985, 31, (9), 1488-1497.

35. Kosters, W. C. G., Rotating-Disk Contactor. In Handbook of Solvent Extraction, Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds. John Wiley & Sons: New York, 1983; pp 391-405.

36. Logsdail, D. H.; Thornton, J. D.; Pratt, H. R. C., Liquid-Liquid Extraction. Part XI: Flooding Rates and Performance Data for a Rotary Disc Contactor. Transactions of

the Institution of Chemical Engineers. 1957, 35, 301-315.

37. Thornton, J. D., Spray liquid-liquid extraction columns: Prediction of limiting holdup and flooding rates. Chemical Engineering Science 1956, 5, (5), 201-208. 38. Kumar, A.; Hartland, S., Correlations for prediction of mass transfer coefficients in

single drop systems and liquid-liquid extraction columns. Chemical Engineering

Research and Design 1999, 77, (5), 372-384.

39. Kumar, A.; Hartland, S., Prediction of Axial Mixing Coefficients in Rotating-Disk and Asymmetric Rotating-Disk Extraction Columns. Canadian Journal of Chemical

Engineering 1992, 70, (1), 77-87.

40. Sarkar, S. S.; Mumford, C. J.; Phillips, C. R., Liquid-Liquid Extraction with Interphase Chemical Reaction in Agitated Columns - 1. Mathematical Models.

Industrial and Engineering Chemistry Process Design and Development 1980, 19,

(4), 665-671.

41. Sarkar, S. S., Liquid-Liquid Extraction with Interphase Chemical Reaction in Agitated Columns - 2. Hydrodynamics and Mass Transfer in Rotating Disk and Oldshue Rushton Contactors. Industrial and Engineering Chemistry Process Design

and Development 1980, 19, (4), 672-679.

42. Kumar, A.; Hartland, S., Emperical Prediction of Operating Variables. In

Liquid-Liquid Extraction Equipment, Godfrey, J. C.; Slater, M. J., Eds. John Wiley & Sons

(26)

43. Kumar, A.; Hartland, S., A unified correlation for the prediction of dispersed-phase hold-up in liquid-liquid extraction columns. Industrial and Engineering Chemistry

Research 1995, 34, (11), 3925-3940.

44. Kumar, A.; Hartland, S., Unified Correlations for the Prediction of Drop Size in Liquid-Liquid Extraction Columns. Ind. Eng. Chem. Res. 1996, 35, 2682-2695. 45. Kung, E. Y.; Beckman, B. B., Dispersed-phase holdup in a rotating disk extraction

column. AIChE journal 1961, 7, (2), 319-324.

46. Kleinberger, R., Zinc Sulfate Extraction by Bateman Pulsed Column Technology,

Proceedings of the Copper, Cobalt, Nickel and Zinc Recovery Conference

Johannesburg, South Africa, 2001; South African Institute of Mining and Metallurgy; pp K1-K13.

47. Buchalter, E.; Kleinberger, R.; Grinbaum, B. Copper Solvent Extraction using a Bateman Pulsed Column, Proceedings of the Extraction Metallurgy Africa

Conference, Johannesburg, South Africa, 1998; South African Institute of Mining

and Metallurgy; pp 185-199.

48. Jahya, A. B.; Stevens, G. W.; Pratt, H. R. C. In Proceedings of the International

Congress on Mineral Processing and Extractive Metallurgy, Mass transfer Studies

for a Pulsed Disc and Dougnut Extraction Column, Melbourne, 2000; Australasian Institute of Mining and Metallurgy; pp 281-284.

49. http://www.bateman.co.il/files/bateman/download/1311200535896297.pdf, accessed January 2010.

50. Logsdail, D. H.; Slater, M. J., Pulsed Perforated-Plate Columns. In Handbook of

Solvent Extraction, Reprint 1991 ed.; Lo, T. C.; Baird, M. H. I.; Hanson, C., Eds.

John Wiley & Sons: New York, 1983; pp 355-372.

51. Klička, V.; Čermák, J., Zweihphasenströmung in der Pulsier-Extraktionskolonne. Verfahrenstechnik 1971, 5, 320-327.

52. http://www.kuehni.biz/pages/uoex_f3e.html, accessed January 2010.

53. Gordon, C. M., New developments in catalysis using ionic liquids. Applied

Catalysis A: General 2001, 222, (1-2), 101-117.

54. Wilkes, J. S., A short history of ionic liquids - From molten salts to neoteric solvents. Green Chemistry 2002, 4, (2), 73-80.

55. Heintz, A., Recent developments in thermodynamics and thermophysics of non-aqueous mixtures containing ionic liquids. a review. Journal of Chemical

Thermodynamics 2005, 37, (6), 525-535.

56. Meindersma, G. W.; Maase, M.; Haan, A. B. d., Ionic Liquids. In Ullmann's

Encyclopedia of Industrial Chemistry (electronic version, 7th_{edition), 2008.}

57. Chiappe, C.; Pieraccini, D., Ionic liquids: Solvent properties and organic reactivity.

Journal of Physical Organic Chemistry 2005, 18, (4), 275-297.

58. Domanska, U., Thermophysical properties and thermodynamic phase behavior of ionic liquids. Thermochimica Acta 2006, 448, (1), 19.

(27)

59. Holbrey, J. D.; Seddon, K. R., Ionic Liquids. Clean Technologies and

Environmental Policy 1999, 1, (4), 223-236.

60. Marsh, K. N.; Boxall, J. A.; Lichtenthaler, R., Room temperature ionic liquids and their mixtures—a review. Fluid Phase Equilibria 2004, 219, 93-98.

61. Zhao, H.; Xia, S.; Ma, P., Use of ionic liquids as 'green' solvents for extractions.

Journal of Chemical Technology and Biotechnology 2005, 80, (10), 1089-1096.

62. Maase, M., Industrial Applications of Ionic Liquids. In Ionic Liquids in Synthesis, Wasserscheid, P.; Welton, T., Eds. Wiley-VCH: Weinheim, 2008; pp 663-687. 63. Rogers, R. D.; Seddon, K. R., Ionic Liquids - Solvents of the Future? Science 2003,

302, (5646), 792-793.

64. Olivier-Bourbigoe, H.; Vallee, C., Catalysis in Non-aqueous Ionic Liquids; Ionic Liquids: Non-Innocent Solvents. In Multiphase Homogeneous Catalysis, Cornils, B.; Hermann, W. A.; Horvath, I. T., Eds. Wiley-VCH: Weinheim, 2005; pp 415-418.

65. Mantz, R. A., Viscosity and Density of Ionic Liquids. In Ionic Liquids in Synthesis, Wasserscheid, P.; Welton, T., Eds. Wiley-VCH: Weinheim, 2008; pp 72-87.

66. Meindersma, G. W.; Hansmeier, A. R.; de Haan, A. B., Ionic Liquids for Aromatics Extraction. Present Status and Future Outlook. Industrial & Engineering Chemistry

Research 2010, 49, (16), 7530-7540.

67. Galán Sánchez, L. M.; Espel, J. R.; Onink, F.; Meindersma, G. W.; De Haan, A. B., Density, viscosity, and surface tension of synthesis grade imidazolium,pyridinium, and pyrrolidinium based room temperature ionic liquids. Journal of Chemical and

Engineering Data 2009, 54, (10), 2803-2812.

(28)

2 Modeling of a Rotating Disc Contactor

2.1 Abstract

Extraction of aromatics from aliphatics is challenging due to close boiling points; in industry polar solvents like sulfolane and NMP are applied with a Rotating Disc Contactor as extractor. In this chapter, a model is established for the description of the hydraulic and mass transfer characteristics of a Rotating Disk Contactor for the extraction of toluene from n-heptane with Room Temperature Ionic Liquids. Mass Transfer Characteristics are modeled with a differential axial dispersion model, with correlations for the mass transfer coefficient and the axial dispersion coefficient. The hydraulic characteristics cover the Sauter mean diameter, the hold-up of the dispersed phase and the operational window, which is described by flooding due to a too small relative velocity between both phases. Finally, two equations are presented for the description of the Sauter drop diameter and hold-up at each operational point and one equation for the prediction of the operational window. The hydraulic characteristics are correlated using physical properties, operational parameters, and geometrical characteristics of the column.

2.2 Introduction

Room Temperature Ionic Liquids (RTILs) are promising solvents for extraction processes1-4_{. The extractive separation of aromatic from aliphatic hydrocarbons is an} important application in the petrochemical industry5. Current processes mostly use polar solvents such as sulfolane (UOP6_{, SHELL) and n-methyl pyrrolidone (Lurgi)}7_{. RTILs are a} new class of solvents and next to their almost negligible vapor pressure, they also offer the opportunity to be tailored for the targeted separation. With RTILs higher distribution ratios and selectivities are achieved compared to the common solvents used8, 9_{. Of course, this has} a major impact on the column design. Smaller columns can be used to achieve the requirements compared to the commonly used solvents. The main difference between RTILs and common solvents is their often considerably higher viscosity10, 11_{. For this} reason, successful introduction of RTILs into extraction operations requires knowledge on their hydrodynamic characteristics in extraction equipment. In the design of an extraction column, besides the separation performance, the hydraulic characteristics, being hold-up and drop size, which determine the operating regimes and the operational window, are key parameters in order to determine the column capacity and the required column diameter to provide the desired throughput12, 13_{Although the hydrodynamic behavior of classical} solvents has been extensively studied12, 14-18_{, to our knowledge no studies have been}

(29)

reported for RTILs, except for our previous preliminary work19, 20_{. In this chapter a} mathematical model is developed describing the hydraulic characteristics and mass transfer performance of a RDC for the extraction of toluene from n-heptane. For this study, we have selected a pilot plant rotating disc contactor (RDC) because it is the most commonly used extractor for aromatics extraction in the petrochemical industry21_{. The suggested models are} developed based on existing column-independent equations and equations derived for the RDC, when conventional solvents were applied. In the following paragraphs, the mathematical model for mass transfer will be described, followed by the required parameters for this model such as the holdup, the axial dispersion coefficients, the mass transfer coefficient and the specific surface area, described by the Sauter mean diameter. To complete the hydraulic characteristics, also the operating regime is discussed.

2.3 Modeling Approach

In order to describe the separation performance in extraction processes, different models, are developed22_{. These models range from very simple models, e.g. the plug flow} model, to more sophisticated ones, e.g. population based modelling23_{. The models that} assume plug flow are the simplest models available. The plug flow model assumes that both phases have uniform velocities and are in pure counter current flow. This model can be easily computed, but cannot be used for scale up or design since the modeling approach is too simplistic and the results cannot always be trusted. The overall flow pattern in extraction equipment, however, is complex and based on experience, it became evident that the description of mass transfer in column extractors with the plug flow model was oversimplified. Therefore, in general, axial dispersion (backmixing) of one or both phases needs to be included in the description, since it shows a major influence on the separation performance13_{. Within the concept of axial mixing, the solute concentrations in any plane} perpendicular to the flow are uniform, and the model also implies that the solute can actually move, or ‘diffuse’, upstream of the main current. Axial dispersion is described by Fick’s law, the only difference between diffusion by Fick’s law and axial dispersion is the fact that the molecular diffusivity has been replaced by an empirical constant, the axial dispersion coefficient. Within the concept of axial mixing, several different model approaches have been developed, namely: stage wise, tanks in series and differential models, all approaches giving comparable results22_{. Calculations with these types of models} are fast and these models are able to estimate parameters or design a column for a large variety of systems and column types. These types of models should, therefore, always be considered before more sophisticated models are used. A differential model with axial dispersion is presented in Figure 2-1. The application of axial mixing on the dispersed phase is no longer valid if drops persist for a long time, without breakage or coalescence.

(30)

Since larger drops will move faster than smaller drops, they will experience different mass transfer rates and, therefore, the concentrations within the individual drops will differ. The assumption in axial mixing that in every plane perpendicular to the flow direction the concentrations are uniform no longer holds, since the concentrations in the larger drops will differ from those in the smaller drops, even at the same height in the column. For these kinds of ‘forward mixing’ models the mass transfer rate and drop velocity have to be known as a function of the drop diameter. Although several methods for predicting these values have been developed, most of them are developed for single drops moving through a stationary continuous phase, and they produce severe errors when applied to drop swarms in a non stationary continuous phase22_{. The missing parameters in predicting these} functions cannot be obtained from fitting these equations to the experimental values, since too many parameters are involved. Due to the inaccuracy or unavailability of these parameters and functions, these models are only suited for prediction of general trends and parameter studies. Another class of models are the population balance-based models22-25_. All previous mentioned models rely on experimental observation or empirical relations to predict their hydrodynamic parameters such as hold-up, drop size and axial dispersion.

Figure 2-1: Graphical representation of a differential axial dispersion model.

These models do not account for velocity profiles within the continuous phase. These velocity profiles will cause drops to coalesce and break-up more readily. The influence of these velocity profiles in the continuous phase is different for different drop diameters, therefore, a population balance for the dispersed phase has to be set up and solved, giving a previously unknown drop-size distribution. Simultaneously, the mass

(31)

transfer equations have to be solved as well. This is possible because if the amount of drops and their diameter is known, the various hydrodynamic parameters can be calculated as well, without the need for these hydrodynamic parameters to be made explicit. This type of model requires many equations, since all mass transfer equations need to be solved for all drop diameter classes individually. With the help of powerful computers this is possible, but the outcome remains questionable since the parameters for drop breakage, coalescence and drop velocity are not accurately known. The last class of models to be pointed out is the Computational Fluid Dynamics (CFD) type of models 26_{. These models work on the basis} of the Navier stokes equation. CFD models are the most complex class of models and require the longest computational times.

Based on an evaluation by Steiner22_{, it was decided to model the RDC as a} differential plug flow model with superimposed axial dispersion coefficients. This choice was made by the following process of elimination of the available alternatives. Plug flow models are rather simplistic and are reported to result in large errors. Other types of models are either based on parameters that are difficult to correlate or are complicated as a whole, resulting in large computation times22_{. Therefore, the choice was limited to an axial} dispersion model. Within this class of models, the tanks in series and stage wise models have also been dismissed because these are based on parameters that need to be fitted to experimental data, whereas some parameters such as the axial dispersion coefficients could not be measured in the available set-up. For the differential plug flow model,on the other hand, a lot of correlations to describe the individual parameters are available in literature.27-32

2.3.1 Mass Balances

The core of the model consists of two partial differential equations, describing the mass-balances of toluene in the organic and ionic liquid phase, Equations 2-1 and 2-2. These balances have been obtained from Hartland et al.27

1 2-1

2-2

(32)

In these balances represents the equilibrium concentration of toluene in the organic phase, calculated in Equation 2-3.

It was assumed that mass-transfer of n-heptane and [3-mebupy][DCA] from one phase to the other played no significant role; the distribution coefficient of n-heptane was found to be very low33_{, it is assumed that when the ionic liquid phase enters the column} instantaneously this phase is saturated with n-heptane and research showing that the amount of [3-mebupy][DCA] being transferred to the organic phase is only in the order of a few PPM is reported in Chapter 7. In Equations 2-1 and 2-2, and involve the axial dispersion, with E the axial dispersion coefficient, where accounts for mass transfer of toluene between the ionic liquid, extract, and organic, raffinate, phase, with k as mass transfer coefficient. The differential equations are solved using 4 boundary conditions. Each phase has one boundary condition at the point where it is fed to the column and one at the point where it exits the column, for the organic phase Equations 2-4 and 2-5.

| 2-4

0 2-5

For the ionic liquid phase the boundary conditions are given by Equations 2-6 and 2-7. 2-6

0 2-7

Equations 2-4 and 2-6 define the starting concentrations and Equations 2-5 and 2-7 declare that the final concentration of toluene cannot change anymore.

2.4 Hydraulic Characteristics

According to the theory for general hydrodynamic design of extraction columns24_, rotating disc contactor operation can be characterized by the Sauter mean drop diameter, d32 and the dispersed phase hold-up, φ, of which the latter is dependent on the operational regime in the column. The column can be operated within a certain window covering a

(33)

range of operational parameters, which is limited by flooding due to too low agitation and by flooding, phase inversion or entrainment at high agitation24_{, see Figure 2-2. Equations} describing the characteristics within the operational window, being drop diameter, hold-up and operational regime are selected first, after which the equations describing the limitations of this window are selected.

Figure 2-2: Operational window of a rotating disc contactor, limited by

flooding at low agitation and flooding, phase inversion and entrainment at high agitation

The parameters on which the equations are based can be divided into three categories: physical properties, operational parameters and geometrical characteristics of the column and internals. The physical properties cover the densities of the phases, ρc and ρd, and thus the corresponding density difference, Δρ = │ρc - ρd│, the dynamic viscosities of the phases, ηc and ηd, and the interfacial tension, γ. These properties are dependent on the compositions of the extract and raffinate flows and the temperature, T, as operational parameter. The conditions used in the experimental pilot set-up and the corresponding physical properties of the two binary systems toluene + n-heptane and toluene + RTIL ([4-mebupy]BF4 and [3-mebupy][DCA]) are described in Chapter 3. The physical properties were correlated by the equations as presented in Chapter 3. The operational parameters cover furthermore the phase velocities, Vc and Vd, and thus the corresponding flux and flow ratio, phase continuity and the rotational speed. The geometrical characteristics are determined by the internals as presented in Figure 2-3. The model is developed for the hydraulic situation and therefore assumes equilibrium conditions in order to rule out the influence of mass transfer. A schematic overview of for in the model needed dimensions is depicted in Figure 2-3.

(34)

Figure 2-3: For model needed dimensions of RDC: 1: stator diameter (Ds), 2 and 9:

compartment height (Hc), 3, and 7: rotor and stator thickness, 4: shaft diameter, 5: rotor

diameter (DR), 6: column diameter (Dc), 8: section height.

2.5 Hold-up

From drop size and hold-up of the dispersed phase, the total interfacial area of the dispersed phase can be calculated. This area is important in mass transfer calculations. Theoretical and empirical equations for the description of the dispersed phase hold-up in a RDC are well reported14, 15, 25, 34, 35_{. The description of the hold-up was, however, based on} unified equations derived by Kumar and Hartland36_{, because these equations were based} on a large number of published experimental data and were applied by Moreira et al.17_for the description of a water/n-heptane system in a rotating disc contactor. The final form of the correlation describing hold-up as a function of various dimensionless groups is:

2-8 in which Π allows for the power input per unit mass, ε, Φ the effect of the phase flow rates, Vd and Vc, Ψ the physical properties, and Γ the geometrical characteristics of the column. These quantities are defined as follows:

(35)

, , . 2-9 , . , . 2-10 , 2-11 , , . 2-12

The constant CΠ,h is included in Π since the hold-up is finite even when the power input per unit mass, ε, is zero. An exponential term is used to describe the effect of Vc, since the hold-up is also finite at zero continuous phase velocity. Such an exponential variation describes the effect of Vc better than a simple power function. The parameters CΨ,h and CΓ,h allow for the effects of mass transfer and geometrical characteristics of columns, respectively. The value of CΨ,h was taken as unity for the case of no mass transfer. The values of CΠ,h, CΨ,h and CΓ,h were held constant at 0.19, 1 and (DR/Hc)0.62_{, respectively.} The values for n1 to n7, original being 0.67, 0.69, 7.13, -0.65, 0.14, -0.26 and -0.1, respectively, were investigated in Chapter 4.

2.6 Axial Dispersion Coefficient

Mass transfer performance of RDC-columns and other liquid-liquid extraction equipment is impaired by axial dispersion effects i.e. the departure from pure plug flow25, 28_. In this model the axial dispersion is governed by an analogy to Fick’s law of diffusion, which is superimposed on the flow. The driving forces behind this process are molecular diffusion and convective diffusion. Axial mixing in the continuous phase, in this thesis the organic phase, differs from axial dispersion in the dispersed ionic liquid phase, therefore, they are treated individually. In the continuous phase axial mixing is mainly caused by circulatory flow caused by the rotating action of the discs, molecular and turbulent diffusion, channeling and non-uniform velocity distributions28_{. Axial mixing in the} dispersed phase in is caused by the distribution of droplet velocities, which is in turn caused by the range of droplet sizes. Entrainment of the small droplets by the continuous phase is also an important factor in the axial mixing behavior of the dispersed phase28_.