at SciVerse ScienceDirect

Journal of Organometallic Chemistry 724 (2013) 102e116

Contents lists available

Journal of Organometallic Chemistry

journal homepage: www.elsevier .com/locate/ jorganchem

Review

Solubility of organometallic complexes in supercritical carbon dioxide: A review

Wen Hui Teoh a,b, Raffaella Mammucari b, Neil R. Foster b,*aDepartment of Chemical Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysiab School of Chemical Engineering, Chemical Science Building (F10), University of New South Wales, Sydney, NSW 2052, Australia

a r t i c l e i n f o

Article history:Received 1 December 2010Received in revised form28 September 2012Accepted 2 October 2012

Keywords:SolubilityOrganometallicSupercritical fluidCarbon dioxideThermodynamic modelLigands

* Corresponding author. Tel.: þ612 9385 4341; fax:E-mail address: n.foster@unsw.edu.au (N.R. Foster

0022-328X/$ e see front matter � 2012 Elsevier B.V.http://dx.doi.org/10.1016/j.jorganchem.2012.10.005

a b s t r a c t

The solubility and solubility trends of organometallic complexes in supercritical carbon dioxide arereviewed. The influence of intermolecular forces, physical properties and the metal chelates on solubilityis explored. A number of thermodynamic models used to predict the solubility behavior of organome-tallic complexes in supercritical carbon dioxide, and the advantages and limitations to these thermo-dynamic models are also discussed.

� 2012 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1022. Solubility and the factors affecting solubility in supercritical fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104

2.1. The effects of intermolecular forces on solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1042.2. The effects of the free volume difference on solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1052.3. Clustering and the solubility enhancement factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1052.4. The effects of pressure and temperature on solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1062.5. The effects of organometallic complexes on solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

3. Thermodynamic modeling of soluteeSCF solubility behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1113.1. Solubility parameter, d and the regular solutions theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1113.2. Empirical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1123.3. Equations of state (EOSs) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

4. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

1. Introduction

A substance above its critical temperature and pressure is knownas a supercritical fluid (SCF). Supercritical fluids exist in a singlehomogenous phase where the liquid and gas phases are

þ612 9385 5966.).

All rights reserved.

indistinguishable. As such, SCFs have characteristics that are inter-mediate to those of gases and liquids. Typically, a liquid substance ischanged to its gaseous phase bymoving it through the vaporizationcurve, given by path AeBeC in Fig. 1. This involves heating and/ordecompression which is accompanied by drastic or abrupt changesto its physical properties. Ameniscus is observed asA is changed toCthrough B. However, a liquid can also be changed to its gaseous formwithout having to go through phase transition. This is carried out by

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116 103

manipulating both temperature and pressure changes; allowing thesubstance to pass through the supercritical region [path AeDeC]where its physical properties are varied continuously.

The two most distinctive traits of a supercritical fluid are itsenhanced mass transfer properties compared to liquids and itsvariable density. The density of a SCF is highly tunable as it issensitive to changes in temperature and pressure. As densitycontributes directly to the solvent strength of a given SCF, thesolvating power of a SCF is also easily manipulated. The easymanipulation allows for the maneuvering of process designs tofacilitate separation, extraction and deposition of compounds. Inmost cases, the ease of control and the tune-ability of the liquid-likedensity of SCFs enable the need for toxic organic solvents to bebypassed. Themass transport properties of SCFs are similar to thoseof conventional gases and add to the desirability of SCF processes, astheir gas-like diffusivity and viscosity allow for faster penetrationinto solid matrices and surfaces. These characteristics facilitate thedeposition of compounds into, and the extraction of compoundsfrom, matrices. In addition, the low operating temperatures typicalof SCF processes allows for the processing of thermally-labilecompounds, while the sterilizing ability of SCFs adds to the advan-tages of using SCFs in biomedical and pharmaceutical applications.

Carbon dioxide is the most widely used SCF due to its mildcritical temperature (31.1 �C) and pressure (7.38 MPa). It is non-flammable, non-toxic, the second least expensive solvent afterwater, fairlymisciblewith awide variety of organic solvents, easy torecover after processing due to high volatility and is considered‘environmentally friendly’ as it can be obtained from existingindustrial processes without adding to the greenhouse effect [1,2].Review on the advantages of supercritical fluids as green processingmedium can be found in Beckman [2]. Compared to conventionallarge molecular-sized solvents, CO2 with its small linear structurediffuses more quickly [1].

Supercritical fluid technologies have garnered the interest andattention of researchers and industries alike since the late 1970sdue to their unique properties [3]. These green technologies havebeen successfully employed in the decaffeination of coffee [3] andthe extraction of spices, hops, tobacco, aromas, essential oils,acetone residues from antibiotics and pharmaceuticals frombotanicals [4e12]. In recent years, research efforts in SCF technol-ogies have expanded from mere extractions to sterilizations,micronizations, encapsulations and depositions for biomedical,pharmaceutical, chemical, food, energy and agriculture applica-tions. Current research includes fabrication of three-dimensionalporous scaffolds for tissue engineering applications [13], prepara-tion of electrocatalysts for membrane fuel cells through supercrit-ical fluid deposition [14], extraction and in-situ chelation of metalsfrom aqueous and waste solutions [15e19], and encapsulation of

Supercritical region

solid

Critical Temperature

Pressure

C

Triple point

liquid

gas

Temperature

B

ADCritical

Pressure

Critical point

Vaporization curve

Fig. 1. Phase diagram for a pure substance.

active pharmaceutical ingredients (APIs) for controlled deliverypurposes [20]. Chemical reactions, properties and applications inSCFs are extensively reviewed in the literature and can be obtainedfrom Refs. [3,21e29].

The use of metallic, organometallic and inorganic compounds inchemical and biomedical applications is not new. Metal-basedchemotherapeutic drugs such as Cisplatin have been used sincethe late 1970s and have been proven to be effective in anticancertreatments [30,31]. The success of Cisplatin, or cis-dia-minedicholoroplatinum(II), has spawned a domino effect on thedevelopment of platinum and palladium based anticancer drugs[32e34]. The diverse and varying coordination numbers andstructures of organometallics, and the accessible redox statesprovide many opportunities in the design and application of novelmedical therapeutics [35]. Moreover, the intrinsic properties of thecationic metal and ligand allow for control over kinetic and ther-modynamic characteristics through ligand design [35e38]. Thus,recent years have seen concerted efforts focused toward thedevelopment of metal-based compounds, particularly for diag-nostic and therapeutic purposes. Recent developments include theuse of ruthenium, yttrium, indium, rhodium, osmium, gold, galliumand cobalt in anticancer therapies [32,39e44] while cyclo-pentadienyls, arene, carbonyl and carbene are typical classes ofligands used in medicinal therapeutics [36]. Transition metals suchas technetium, manganese and iron are also increasingly used inmedical imaging as only low dosages are required of these ‘highlysensitive materials’. The discovery of molecular targets and thedevelopment of nanocarriers carrying metal-based anticancerdrugs that selectively target tumors are creating new opportunitiesfor a drug combination therapywhere two ormore types of therapy(such as chemotherapy and radiotherapy), or two or more APIs, aredelivered toward a specific target [45,46]. The use of organometalliccompounds in medicinal therapeutics has been reviewed ina number of publications and can be found in Refs. [34,36,47e50].

The significant use of metal-based compounds in imaging,therapeutics and catalysis, and the advantages of SCF technologieshave resulted in the initiation of numerous studies of SCF processingof organometallics. These include the deposition of copper films foradvanced interconnecting structures using copper(I)(1,1,1,5,5,5,-hexafluoro-2,4-acetylacetonate)(1,5-cyclooctadiene) [51], synthesisof ultra high molecular weight polyethylene (UHMWPE)/silvernanocomposites for total joint replacement components [52],synthesis of gold nanoparticles from gold(I) perfluorooctanoate insupercritical carbon dioxide [53], copolymerization of carbonmonoxide and styrene in supercritical CO2 with palladiumcomplexes as catalysts [54], and the mironization of titanocenedichloride (a metallocene catalyst) via rapid expansion of super-critical solvent (RESS) [55]. However, most SCF related studies onorganometallic compounds are heavily concentrated on catalysisandwastemetal extractionswhile research in the area of processingmetal-based therapeutics has not received the same level of interest.The exploitations of SCF technologies for homogeneous andheterogeneous catalysis have been reviewed and can be found inRefs. [56e62]. Review on supercritical fluid extraction of metalsfrom aqueous media can be obtained from Refs. [15,63].

While the use of SCFs to process metal-based compounds isgaining ground, a limitation of the technology is a lack of funda-mental data. The complexity, the non-ideality and the non-linearityof SCF characteristics make it difficult to predict their behavior andthat of the compounds involved under supercritical conditions. Thelack of predictability hence, requires extensive, expensive andtime-consuming trial and error experimental studies [64]. Phasebehavior predictions of soluteeSCF systems also become morechallenging as the molecular complexity of the solutes increases[65]. While the current thermodynamic models such as the Penge

Factors that affect solubility

Intermolecular (attractive) forces

Physical forces

Momentary dipole-dipole potential

energy

Permanent dipole-dipole potential

energy

Chemical forces

Hydrogen bonding, electron aceptor-

donor complexing

Free volume difference

Density

Temperature and pressure

Fig. 2. Factors affecting the solubility of a solute in a solvent. Extracted from Ref. [3].

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116104

Robinson EOS (equation of state) and the SoaveeRedlicheKwongEOS are highly regarded, limitations remain in their ability topredict highly non-ideal thermodynamic behavior, particularly formetal-based complexes, in the supercritical region. The challengesinvolved in predicting solid-SCF phase equilibria include: (1) thelack of vapor pressure data for relatively non-volatile solids, (2) theextremely asymmetric sizes and energies of the components in SCFsolutions, and (3) the highly condensable solutions that lead tosolvent clustering around the solute [65]. Hence, accurate ther-modynamic models for highly non-ideal and highly compressiblemixtures are essential.

In the present paper, the solubility of organometallic complexesin supercritical carbon dioxide (SCCO2) and a number of thermo-dynamic models used in organometallic-SCCO2 systems arereviewed. Solubility studies are fundamental to quantifying theability of a SCF to act as a solvent and consequently, the feasibility ofa particular process for a particular application [66,67]. Solubilitydata influences the type of a SCF process to be used e be it an anti-solvent or a solvent based SCF technology. Solubility also shows thelimit to extraction or the maximum concentration available fordeposition, and provides the thermodynamic limit to a process[68]. There is therefore, a need to investigate the factors that affectsolutes solubility in supercritical fluids to allow for a continualdevelopment in the many potential applications, design and scaleup of SCF processes. Reviews on the solubility and dissolution ofsolutes in sub- and supercritical fluids have been covered in varyingdepths and degrees by various authors [3,69e73]. The experi-mental methods for solubility determination generally involveanalytical or synthetic methods, and have been systematicallyreviewed by several authors over the last few decades [74e79]. Thephase behavior of binary and ternary mixtures in supercriticalcarbon dioxide warrant its own review, and can be found in Refs.[3,22,80e83].

2. Solubility and the factors affecting solubility insupercritical fluids

Solubility is the analytical composition of a solution saturatedunder equilibrium with one of the components of the solution ata particular temperature and pressure [84,85]. It quantifies thedynamic equilibrium state achieved where the dissolution rateequals that of the precipitation rate in a solideliquid system.Solubility is very dependent on the intermolecular forces betweensoluteesolvent, soluteesolute and solventesolvent [3]. Fora system consisting of a solute, i and a solvent, j, the ability of i todissolve in j is given by the strength of interactions between i and j.If the attractive forces of iej are higher than that of iei and jej, thenthe solute i would dissolve in the solvent j. The ancient heuristic of“like dissolves like” is still applicable in many soluteesolventsystems since intermolecular forces between similar chemicalcompounds lead to smaller endothermic enthalpy of solution thandissimilar compounds [86]. The factors that contribute to thesolubility of a given solute in a solvent are summarized in Fig. 2.

2.1. The effects of intermolecular forces on solubility

The fundamental source of most physical properties of a mole-cule is its intermolecular forces [64]. Intermolecular forces inoperation between molecules are governed by the physical andchemical forces existing between themolecules in the solution. Thephysical forces of attraction between molecules are governed bythe momentary (induced) dipoleedipole potential energy and thepermanent dipoleedipole potential energy. In the momentarydipoleedipole potential energy interaction, a molecule witha momentary dipole causes a dipole in neighboring molecules,

producing net attractive forces between the molecules thatsubsequently governs the solubility of a non-polar solute in a non-polar solvent [3]. The momentary dipoleedipole potential energydepends on the polarizability of a molecule and the distancebetween two molecules. The polarizability of a solvent (whichincreases with its molecular size) indicates the strength of thesolvent [3]. The momentary dipoleedipole potential energy is,however, not affected by temperature changes.

The permanent dipoleedipole potential energy provides anadditional force of attraction between molecules. Large, complexmolecules generally have weakened dipole interactions due to thelarge volume of the molecules [87]. A molecule with high polarmoments and high dielectric constant is considered polar. Thepermanent dipoleedipole potential energy changes with inter-molecular distance while it decreases with increasing temperature.Hence, a polar liquid solvent (which generally has a high criticaltemperature) exhibits non-polarity in the supercritical region [3].At very high pressures and densities, repulsive intermolecularforces could also dominate and reduce the solubility of a solute[88]. Higher-order polar moments also affect the solvating power ofa solvent although these effects are less significant than a dipolemoment [3]. A good example is carbon dioxide which has a largequadrupole moment that plays an important role in its solvatingpower [3,89] and greatly enhances the solubility of a number ofpolar solutes in supercritical carbon dioxide [90].

Chemical forces of attraction such as hydrogen bonding andelectron acceptoredonor complexing also play an important role insolubility, although these are harder to quantify [3]. The effects ofchemical forces in a solution are generally categorized as solvation,where the molecules of the solutes and solvents tend to formcomplexes [91]. In contrast to physical forces, chemical forces canbecome saturated and are very dependent on temperature [91].Chemical forces generally decrease with increasing temperaturesuch as the observed weaker H-bonds in sub-critical and super-critical water [92,93]. Generally, H-bonds are easily broken athigher temperatures when the molecules involved have sufficientenergy to break loose as the strength of H-bonds are lower thanthat of covalent bonds [91].

Supercritical carbon dioxide is generally considered a solventcomparable to dioxane, with solvation in carbon dioxide largelyattributed to its quadrupole moment [89,94]. In spite of its zerodipole moment, carbon dioxide has partial negative charges on theelectronegative oxygen atoms and a large partial positive charge onits electropositive carbon atom, allowing for the electron-deficientcarbon atom to act as a Lewis acid (LA) and the oxygen atoms asLewis base (LB) moieties. Recent studies have shown that thecharge separation and electronic structure of CO2 allow for a LewisacideLewis base (LAeLB) interaction between CO2 with a carbonyl[95e97]. Spectroscopy and theoretical studies also found weaker,shorter CeH/O hydrogen bond [98,99] that interacts in

-4.50

-4.00

-3.50

-3.00

-2.50

-2.00

-1.50

-0.150 -0.100 -0.050 0.000 0.050 0.100 0.150 0.200 0.250 0.300

log

S

log r

313 K - Cr(thd)3 323 K -Cr(thd)3 343 K - Cr(thd)3

313 K - Co(thd)3 323 K - Co (thd)3 343 K - Co(thd)3

Fig. 4. Linear relationships between the logarithms solute solubility with the loga-rithms of reduced SCCO2 density at various temperatures for Co(thd)3 and Cr(thd)3.Data extracted from Haruki et al. [113].

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116 105

cooperation with the LAeLB bond that further stabilizes thesoluteesolvent interactions [95,96]. A resultant of the LAeLBinteraction and the CeH/O hydrogen bond is the increasedpolarization of one of the CaO bonds of CO2 that subsequentlyenhances the solubility of a number of polar solutes in SCCO2 [96].Recent 1H, 13C and 19F NMR studies conducted on b-diketones andUO2(b-diketonato)2dmso [dmso ¼ dimethyl sulfoxide] complexesfound that their solvations in SCCO2 are attributed to a balance ofvan der Waals, LAeLB and the CeH/O interactions [95,100].Further reading on the polar attributes of CO2 and their contribu-tion to solvation can be obtained from Refs. [94e96].

Various other studies have also found that thepresenceof inter- orintra-molecular hydrogen bonds in a SCF can enhance solubility[101e106]. The addition of a modifier into a SCF that can formhydrogenbondswith a solute, generally increases the solubilityof thesolute. Fig. 3 shows the solubility of benzoic acid in SCCO2, and SCCO2with methanol. The solubility of benzoic acid increases by almost anorder of magnitude with the addition of methanol as a co-solvent.Anderson and Siepmann [102] attributed this effect to the hydrogenbond formedbetweenbenzoic acidandmethanol.Hydrogenbondinghas also been found to occur between metal-containing and/ororganometallic complexes in SCFs [90,107] although the effects ofsuch bonding on solubility are rarely quantified. In a study conductedby Tenorio et al. [108], the solubility of nickel hexa-fluoroacetylacetonate dehydrate [Ni(hfa)2$2H2O] was found to behigher in SCCO2 with ethanol as a modifier than that in SCCO2 alone.The authors suggested that the better solvation of the nickel hexa-fluoroacetylacetonate dehydrate complex in CO2eethanol mixturewas due to a possible rupturing of the hydrogen bond networkbetween the Ni(hfa)2$2H2O molecules and the formation of newhydrogen bonds between the nickel complex and ethanol [108].

2.2. The effects of the free volume difference on solubility

Free volume is defined as the total integral over the part of thepotential energy due to the thermal displacements of the center ofgravity of the molecule from its equilibrium [110]. It is the “empty”volume available to the molecule after subtracting the hard-corevolume of the molecule itself [111]. Therefore, free volumereflects the compressibility or expansivity of a substance [3]. Thefree volume difference between a solute and a solvent is related tothe density of the solvent as this affects the interaction levelbetween molecules in a solution. A solute in a SCF must be in closeenough proximity to interact with the solvent molecules to affectsolubility [3]. As the density of a SCF increases, so does the ability ofthe SCF to solvate a compound. Therefore, the solubility of solutes

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

10.00

11.00

100 150 200 250 300

Solu

bilit

y, S

××10

3(m

ole

frac

tion

)

Pressure (bar)

Supercritical carbon dioxide Supercritical carbon dioxide + methanol

Fig. 3. Solubilities of benzoic acid in SCCO2 and SCCO2 with 3.5 mol% methanol as a co-solvent at 308 K. Data extracted from Ref. [109].

can be directly related to the density of the SCF [1,112] althoughMcHugh and Krukonis [3] cautioned that this “heuristic is only trueto a first approximation”.

The solubility of a solute has been shown to increase directlywith the density of a SCF at constant temperature. In a study con-ducted by Haruki et al. [113], the logarithms of solubility (log S) oftris(2,2,6,6-tetramethyl-3,5-heptanedionato)cobalt(III) [Co(thd)3]and tris(2,2,6,6-tetramethyl-3,5-heptanedionato)chromium (III)[Cr(thd)3] were found to increase linearly with the logarithms ofreduced SCCO2 density [logrr] at different isotherms, as shown inFig. 4. The increase in solute solubility with density, however,cannot be oversimplified and generalized to the whole supercriticalregion. A comparison between the solubilities and reduced densi-ties of Cr(thd)3 at various temperatures is shown in Table 1. Itdemonstrates that while solubility increases with density ata constant temperature, the solubility of a solute at a highertemperature but lower solvent density, can be higher than thesolubility of a solute at a lower temperature but of higher solventdensity. Thus, other competing factors affected by the rise intemperature may also influence the solubility of a solute in a SCF.The influence of temperature on solubility is discussed further inSection 2.4.

2.3. Clustering and the solubility enhancement factor

The solubility behavior of a solute in a SCF is also enhanced bythe clustering of solvent molecules around the solutes. Local clusterstructures have been observed in the lower density region of neatSCCO2 between 0.2 g/cm3 and 0.6 g/cm3 while at higher densityregion of 0.6e0.9 g/cm3, cluster structures with increasing amountof CO2 were observed [95]. Since the solubility of a heavy solute ina dense gas solution rarely exceeds a few mole percent [114,115],the number of solvent molecules surrounding a solute molecule is

Table 1Solubility of Cr(thd)3 in SCCO2 at various temperatures and reduced densities. Dataextracted from Haruki et al. [113].

Reduced density, r Solubility, S 103 (mole fraction) Temperature (K)

1.47 3.23 313

1.35 3.44 323

1.20 5.24 343

1.20 5.24 343

1.09 2.18 343

0.975 1.14 343

incr

ease

s

incr

ease

s

incr

ease

s

incr

ease

s

incr

ease

s

cons

tant

ρ ×

Pressure

Solu

bilit

y

Ret

rogr

ade

regi

on

T3

T1

T2

T4>T3>T2>T1

PL PU

T4

Fig. 5. Plot of solubility versus pressure at different temperatures.

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116106

in excess. As a SCF solution is more compressible, and has a largerfree volume, attractive forces can induce molecules into energeti-cally favorable locations to form clusters [65]. Since the solventesolute net attractive forces in a solution are stronger than that ofa solventesolvent or soluteesolute attractive forces, excess solventmolecules form clusters around a solute molecule. Cluster forma-tion has been shown to enhance the solubility of a solute in dilutebinary dense gas mixtures [114,116]. The high enhancement factor(ratio of actual to ideal gas mixture predicted solubility) of densegas systems is typically in the range of 102e105 [114]. In a binarysystem consisting of a solid solute and a SCF, the enhancementfactor, E is given by

E ¼ y2P

Psat2(1)

where y2 is the mole fraction of the solute, P is the pressure of thesystem, and Psat2 is the vapor pressure of the pure solvent. Theenhancement factor can be equated to the fugacity coefficient of thepure solid through equation (2)

E ¼ fs2bf2

exp

"vs2�P � Psat2

�RT

#(2)

where f2s is the fugacity coefficient of the pure solute, b42 is the

fugacity coefficient of the solute in the solution, and n2s is the molar

volume of the solid solute. Since vapor pressure of pure solids areusually very low, P[Psat2 and f2

s is generally near unity [117],equation (2) can be simplified into

E ¼ 1bf2

exp�vs2PRT

�(3)

An alternative approach to correlating the enhancement factoris via the density of a SCF. Since linearity is often observed betweenthe logarithm of solubility with density along isotherms, equation(1) has been used to correlate the enhancement factor intoempirical equations that conform with equation (4) [118]

E ¼ expðAþ B$rÞ (4)

where A and B are temperature dependent, empirical constants andr is the density of the SCF. Further discussion on the solubilityenhancement factor can be found in Refs. [112,117e120].

Organometallic complex

Metal

Distance between ligand and centre atom

Oxidation state : M+ < M2+ < M3+

Ligand

Phenyl-substituted < aliphatic-substituted < fluorinated

Heteroleptic < homoleptic

Sterical hindrances

Fig. 6. Factors within an organometallic complex that affect solubility in SCFs.

2.4. The effects of pressure and temperature on solubility

The isothermal solubility of solids in SCFs is not a direct functionof pressure where the solubility initially decreases with increasingpressure, reaches a minimum, and followed by a steep increase inthe critical pressure region [121]. However, as the pressureincreases further from the critical point, the increase in thesolvating power of the SCF diminishes. Fig. 5 shows the effects ofpressure and temperature on solubility.

Solubility is primarily affected by vapor pressures of the soluteswhile other soluteesolvent interactions play secondary roles in SCFmixtures [65]. Since vapor pressure is determined from the temper-ature of a system, the effect of temperature on solubility is morecomplicated, as temperature affects both the density of the super-critical solvent and the vapor pressure of the solute [73]. In theretrograde region, i.e. the region between the lower crossover pres-sure (PL) and the upper crossover pressure (PU), the solubility ofa solute decreaseswith increased temperature. Thedecrease in solutesolubility in the retrograde region is attributed to the dominant effectof a rapid decrease in density. At PL and PU, the solubility of various

isotherms intersect and the competing effects of solvent density andsolute vapor pressure balance each other out [121]. Above and belowthe retrograde region, solubility increases with temperature as thevapor pressure of a solute dominates. Hence, in the regions above PUand below PL, the solubility of a solute in a SCF increases withtemperature. In the work of Haruki et al. [113], measured solubilitiesof Co(thd)3 and Cr(thd)3 in a relatively lower pressure region werefound to decrease with increased temperature while at a relativelyhigher pressure region, the solubilities of Co(thd)3 and Cr(thd)3werefound to increase with temperature. Such observations commensu-rate with the behavior of a solute in the retrograde region, and theregion above the upper crossover pressure, PU.

2.5. The effects of organometallic complexes on solubility

The metal and ligand side-chains as well as the crystal structureof a metal chelate are factors that contribute to the solubility ofmetal chelates in SCCO2 [70,122e125]. The contributing effects ofthe organometallic complexes to their solubilities in SCFs aresummarized in Fig. 6.

2.5.1. The effects of ligand side-chains on the solubility oforganometallic complexes in SCCO2

Ligands and ligand design are important in optimizing theelectronic characteristics and the solubility of metal chelates in

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116 107

SCCO2 [58]. A number of solubility studies has been conducted onorganometallic complexes in SCCO2 in which the influence ofligands on solubility has been systematically studied. A list ofligands and their acronyms is given in Table 2 while the solubilitystudies conducted on metal complexes in supercritical carbondioxide by various authors are shown in Table 3. While a number ofstudies is available for various classes of ligands, b-diketonates arethe most extensively studied ligands in SCCO2. A number of ligandsthat were systematically studied and their influence on the order ofsolubility in metal complexes are shown in Table 4.

Generally, metal complexes containing fluorinated ligands arerelatively more soluble, and these are followed by aliphatic- andphenyl-substituted systems. Rows (a) and (c) in Table 4, illustratethe higher solubility of fluorinated species such as hfa and fddc, incomparison to their non-fluorinated species. The higher solubilityobserved for fluorinated organometallic complexes are alsoobserved in other studies [70,122,123,125,134]. The observationthat fluorination of a solute increases the CO2-philicity of the soluteis not new, and the mechanism by which solvation occurs forfluorinated compounds has been a source of various theoretical andspectroscopic studies. It has been suggested that the higher vola-tility of the fluorinated chelates causes the observed higher solu-bility in SCCO2 [131]. Another possible reason is the reduction in thepolarity of the complexes with the addition of fluorine [123]. UVeVIS and FTIR spectra conducted on fddc and edcmetal complexes inSCCO2 confirmed the polarity reduction in the fddc complex, inwhich the shift of charge transfer bands to lower energies in fluo-rinated metal chelates reflected the high electron withdrawingnature of the fluorinated methyl groups [122].

Several other spectroscopic and theoretical studies have alsobeen conducted to elucidate the mechanism surrounding thesolvation of fluorinated solutes in SCCO2, although conflictingresults were reported. Evidence of van der Waals interactionsbetween afluorinated solute (1,1-dihydroperfluorooctylpropionate)andCO2wasfirst observed in a 1H and 19FNMRchemical shifts studyby Dardin et al. [146]. The higher solubility observed in fluorinatedsolutewas attributed to themagnetic shielding originating fromvander Waals interactions [146]. The findings by Dardin and Samulski

Table 2Acronyms of ligands.

Acronym Ligands

acac Acetylacetonateacac-Br 3-Bromopentane-2,4-dionatobzac 1-Phenylpentane-1,3-dionatobdc DibutyldithiocarbamateCO Carbonylcod 1,5-Cyclooctadienecp Cyclopentadienyldc Dithiocarbamatedctp Dichlorobis(triphenylphosphine)dfh 4H,4H-Devafluoroheptane-3,5-dionatedibm 2,6-Dimethylheptane-3,5-dionatodmhd 1,1-Dimethylhexane-3,5-dionatoDMSO Dimethyl sulfoxideedc Diethyldithiocarbamatehdc Dihexyldithiocarbamatehfa Hexafluoroacetylacetonatefddc (Trifluoroethyl)dithiocarbamateme Methylpdc Pyrrolidinedithiocarbamatep3dc Dipropyldithiocarbamatep5dc Dipentyldithiocarbamatethd 2,2,6,6-Tetramethyl-3,5-heptanedionatotfa 1,1,1-Trifluoropentane-2,4-dionatotfbzm 1,1,1-Trifluoro-4-phenylbutane-2,4-dionatotod 2,2,7-Trimethyl-3,5-octanedionato

were in agreement with a number of other authors [95,140,147e149]. A specific intermolecular interaction between fluorine andCO2 was also observed in an NMR study on fluorinated and non-fluorinated hydrocarbons in CO2. Excess chemical shift effectswere observed in 19F NMR for fluorinated compounds in CO2, incontrast with the bulk susceptibility-dominated 1H NMR observedfor hydrocarbons [146]. Similarly, IR spectroscopy of polymers withCO2 indicated interaction between the dipole of CeF bondswith CO2[150] which point to a possible causality to the higher solubility offluorinated organometallic compounds.

Enhanced interaction between hexafluoroethane and carbondioxide interactionwas also calculated ab initio, inwhich itwas foundthat the positively charged carbon atom has a strong attraction to thenegatively charged fluorine atom [151]. However, the ab initio calcu-lation brought about comments in which the use of a restrictedHartreeeFock level of calculationwas questioned [152]. Recalculationby Diep et al. [153] found that the hydrocarboneSCCO2 interactionsare stronger than the fluorocarboneSCCO2 interactions. Yee et al. [90]used various FTIR techniques to probe the molecular interactions ofheptafluoro-1-butanol where they found no evidence of a specificinteraction of carbon dioxide with the perfluorinated portion of thesolute. Further investigationsbyYonkerandPalmer [154,155]using 1Hand 19F high pressure NMR chemical shifts and NMR relaxationtechniques corroborate the lackof specific CO2/fluorine interactions insolution.

While the exact solvating mechanism of a fluorinated complexin SCCO2 is still under debate and constant study, the presence offluorine generally increases the solubility of organometalliccomplexes. The exception to the “rule”was observed by Laintz et al.[123] inwhich the solubility of Bi(fddc)3 was found to be lower thanthat of Bi(edc)3 at 323 K and 100 atm. The cause for the lowersolubility observed in the fluorinated Bi(fddc)3 complex wasattributed to a possible conformational difference between the twocomplexes [123]. At a higher pressure of 230 atm, the order ofsolubility was observed to revert to edc < fddc [123].

The solubility of organometallic complexes also increases withincreasing alkyl chain lengths substitution of the ligands, as illus-trated in part (b) of Table 4. The increasing solubility trend ofcopper complexes with progressive alkyl substitution of the edcligand shown in Table 4(b) was also observed in the increasingsolubility of Cu(acac)2, Cu(dmhd)2 and Cu(thd)2 [124]. The methylgroup in the acac ligand was progressively substituted by tert-butylgroups, with the structures of Cu(acac)2 and Cu(thd)2 shown inFig. 7. The higher solubility of the thd ligand in comparison to theacac ligand, was also observed in the work of other authors[113,130,134]. All three Cu(acac)2, Cu(dmhd)2 and Cu(thd)2 crys-tallize into square planar structure [156]. While the similar crystalstructures of the three complexes do not provide additional infor-mation as to the causality of the order of solubility, it has beensuggested that the additional alkyl group in the thd ligand providesa better shielding of the center atom that subsequently increasesthe solubility of the metalethd complexes [130,136]. It has beenfurther suggested that the addition of the alkyl reduces the cohe-sional energy density, and hence, increases the solubility of the thdcomplexes [130,134].

In a recent study, the solubilities and charge densities ofCo(acac)3, Cr(acac)3, Co(thd)3 and Cr(thd)3 were estimated byquantum chemical calculation and COSMO-RS to determine theeffect of molecular structure of organometallic complexes on theaffinity of CO2. In the work of Haruki et al. [113], the solubilities ofCo(thd)3 and Cr(thd)3 were found to be between 50 and 70 timesmore soluble than that of Co(acac)3 and Cr(acac)3. The lowersolubilities of Co(acac)3 and Cr(acac)3 were attributed to thehydrocarbon part in the acac ligand that were too small to cover upthe high charge density of the metal and oxygen atoms. In contrast,

Table 3Solubility studies conducted on a number of organometallic complexes in supercritical carbon dioxide.

Substance Temperature range (�C) Pressure (bar) or density (kg/m3)range [density denoted r]

Solubility range(in mole fractionunless statedotherwise)

Ref

Li(acac) 60 98e294 �0.01 mg/La [126]Ag(acac) 40 103e300 (0.89e3.46) � 10�7 [127]UO2(acac) 40 100e250 (4.18e26.7) � 10�4M [100]Co(acac)2 40 160 0.53 � 10�5 kg/m3 [128]

40 161e300 (3.91e8.59) � 10�5 [129]Cu(acac)2 60 100e300 (0e3) � 10�5 mol/mol [130]

150e170 120e220 (6.20e155) � 10�4 [131]40 103.4e344.7 0.750e2.307 [124]

35e55 r: 652.0e902.3 (0.508e9.126) � 10�5 [68]60 98e294 �0.21 mg/La [126]40 160 7.3 � 10�5 kg/m3 [128]

40e70 143e300 (0.997e4.74)) � 10�5 [129]Fe(acac)2 40e60 100e201 (0.108e5.97) � 10�4 [132]Mn(acac)2$2H2O 60 98e294 �0.40 mg/La [126]Ni(acac)2 60 100e300 (0e6) � 10�5 mol/mol [130]

40 160 3.4 � 10�5 kg/m3 [128]Pd(acac)2 60 100e300 (0e6) � 10�5 mol/mol [130]

40 100e300 (0.93e5.75) � 10�5 [127]Pt(acac)2 40 105e290 (0.93e5.75) � 10�5 [127]Zn(acac)2 60 98e294 �1.01 mg/La [126]

60 101e232 (9.5e90) � 10�4 mol/L [122]40 160 7.2 � 10�5 kg/m3 [128]

Co(acac)2$2H2O 60 98e294 �0.25 mg/La [126]Co(acac)3 60 98e294 �0.62 mg/La [126]

40e60 97e40 (0.830e10.669) � 10�5 [133]Cr(acac)3 40 103.4e344.7 (1.716e19.09) � 10�5 [124]

60 203e405 (2.0e3.5) � 10�3 mol/L [125]Fe(acac)3 40e60 98.2e353.7 (0.06e0.30) � 10�3 [134]

40e60 90.1e275.8 (8.75e576) � 10�6 [135]40e60 101e176 (0.0509e2.89) � 10�4 [132]

Ga(acac)3 60 98e294 �3.01 mg/La [126]In(acac)3 60 98e294 �2.63 mg/La [126]Mn(acac)3 60 98e294 �1.26 mg/La [126]Rh(acac)3 40 100e298 (3.32e10.30) � 10�5 [127]Ru(acac)3 40 100e300 (1.70e9.50) � 10�5 [127]Y(acac)3 150e170 120e220 (0.47e3.40) � 10�5 [131]Cu(bdc)2 60 101e232 (1.3e72) � 10�5 mol/L [122]Hg(bdc)2 60 101e232 (5.6e56) � 10�5 mol/L [122]Zn(bdc)2 60 101e232 (8.2e69) � 10�5 mol/L [122]

55 240.5 12.42 � 10�6 g/ml [136]Co(cp)2 60 100e175 (26e222) � 10�5 mol/mol [130]Cr(cp)2 60 100e175 (19e206) � 10�5 mol/mol [130]Fe(cp)2 60 100e175 (24e250) � 10�5 mol/mol [130]

40e70 97.5e366.3 (2.22e240) � 10�4 [137]35e50 80.2e403.4 (0.098e3.986) � 10�3 [133]40e60 100e200 (0.233e2.41) � 10�3 [132]

Mn(cp)2 60 100e300 (8e9) � 10�5 mol/mol [130]Ni(cp)2 60 100e175 (18e218) � 10�5 mol/mol [130]Os(cp)2 60 100e250 (1e13) � 10�5 mol/mol [130]Ru(cp)2 60 100e170 (4e29) � 10�5 mol/mol [130]Ti(cp)2Cl2 64 179e300 (0.86e3.39) � 10�5 mol/mol [138]Na(edc) 50 101 1.5 � 10�4 mol/L [123]Ni(edc)2 50 101 8.5 � 10�7 mol/L [123]Cu(edc)2 50 101 1.1 � 10�6 mol/L [123]

60 101e232 (1.4e11) � 10�6 mol/L [122]35e55 r: 759.93e898.45 (1.95e13.9) � 10�7 [68]

Hg(edc)2 50 152 8.2 � 10�6 mol/L [123]60 101e132 (6.8e53) � 10�6 mol/L [122]

35e55 100e300 4.086e7.181 mg/L [139]Bi(edc)3 50 101 6.0 � 10�6 mol/L [123]Co(edc)3 50 101 2.4 � 10�6 mol/L [123]Zn(edc) 55 240.5 0.59 � 10�6 g/ml [136]Na(fddc) 50 101 4.7 � 10�4 mol/L [123]Ni(fddc)2 50 101 7.2 � 10�4 mol/L [123]Cu(fddc)2 50 101 9.1 � 10�4 mol/L [123]

60 101e232 (9.1e10) � 10�4 mol/L [122]Hg(fddc)2 50 152 5.0 � 10�3 mol/L [123]

60 101e232 (3.0e14) � 10�3 mol/L [122]Bi(fddc)3 50 101 7 0.3 � 10�4 mol/L [123]Co(fddc)3 50 101 8.0 � 10�4 mol/L [123]Cu(hdc)2 60 101e232 (2.1e28) � 10�4 mol/L [122]

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116108

Table 3 (continued )

Substance Temperature range (�C) Pressure (bar) or density (kg/m3)range [density denoted r]

Solubility range(in mole fractionunless statedotherwise)

Ref

Hg(hdc)2 60 101e232 (1.6e38) � 10�4 mol/L [122]Zn(hdc)2 60 101e232 (3.2e58) � 10�4 mol/L [122]UO2(hfa) 40 100e250 (1.21e3.73) � 10�2 M [100]Ni(hfa)2 60 203e405 (8.0e9.9) � 10�3 mol/L [125]Ni(hfa)2$2H2O 40e60 94e251 (2.95e20.23) � 10�5 [108]Ba(hfa)2 150e170 120e220 (1.30e24.0) � 10�5 [131]Cu(hfa)2 60 203e405 (8.0e9.9) � 10�3 mol/L [125]

40 103.4e344.7 (6.173e74.15) � 10�5 [124]40 100 0.029 mol/dm3 [140]

Cu(hfa)2$H2O 40 103.4e344.7 (152.0e414.0) � 10�5 [124]UO(hfa)2$H2O 40 100 0.040 mol/dm3 [140]VO(hfa)2$H2O 40 100 0.025 mol/dm3 [140]Cr(hfa)3 60 203e405 (8.0e9.9) � 10�3 mol/L [125]

40 100 0.03 mol/dm3 [140]Y(hfa)3 150e170 120e220 (1.30e24.0) � 10�5 [131]Zr(hfa)4 40 100 0.044 mol/dm3 [140]Cu(pdc)2 60 101e232 (4.1e40) � 10�7 mol/L [122]Hg(pdc)2 60 101e232 (3.5e34) � 10�7 mol/L [122]Zn(pdc)2 60 101e232 (3.2e90) � 10�7 mol/L [122]Cu(p3dc)2 60 101e232 (6.3e120) � 10�6 mol/L [122]Hg(p3dc)2 60 101e232 (1.2e23) � 10�5 mol/L [122]Zn(p3dc)2 60 101e232 (7.9e150) � 10�6 mol/L [122]Cu(p5dc)2 60 101e232 (0.90e18) � 10�4 mol/L [122]Hg(p5dc)2 60 101e232 (1.0e20) � 10�4 mol/L [122]Zn(p5dc)2 60 101e232 (1.6e32) � 10�4 mol/L [122]Ag(thd) 60 100e300 0 [130]K(thd) 60 100e200 0 [130]Rb(thd) 60 100e200 0 [130]Co(thd)2� 40e70 100e159 (2.10e12.8) � 10�4 [129]Cu(thd)2 60 100e175 (1e82) � 10�5 [130]

40 103.4e344.7 (6.173e74.15) � 10�5 [124]40e70 119e280 (2.71e33.9)) � 10�4 [129]

Ni(thd)2 60 100e250 (0e28) � 10�5 [130]Zn(thd)2 60 100e150 (7e>500) � 10�5 [130]Co(thd)3 60 100e170 (2e213) � 10�5 [130]

40e70 100e190 (0.368e2.83) � 10�3 [113]Cr(thd)3 60 100e150 (4e>500) � 10�5 [130]

40 103.4e344.7 (400e604.9) � 10�5 [124]40e70 101e163 (0.233e5.24) � 10�3 [113]

Fe(thd)3 60 100e150 (3e>500) � 10�5 [130]40e60 94.8e306.7 (1.64e10.13) � 10�3 [134]

Mn(thd)3 60 100e150 (3e>500) � 10�5 [130]Ru(thd)3 60 100e150 (1e174) � 10�5 [130]Tb(thd)3 40e60 124.6e352.3 (0.07e2.34) � 10�3 [134]Ti(thd)3 60 100 >500 [130]Ce(thd)4 40e60 101.2e350.9 (0.05e0.62) � 10�3 [134]Zn(thd)4 60 100e175 (1e109) � 10�5 [130]Cu(tod)2 40 103.4e344.7 (26.03e269.7) � 10�5 [124]Fe(tod)3 40e60 90.6e177.4 (1.39e12.67) � 10�3 [134]Tb(tod)3 40e60 124.3e318.4 (0.53e3.45) � 10�3 [134]Ce(tod)4 40e60 98.0e243.0 (1.53e7.84) � 10�3 [134]Cu(acac-Br)3 40 103.4e344.7 (1.70e9.50) � 10�5 [124]Rh(acac)(cod) 60 100e200 (0e23) � 10�5 [130]Cu(bzac)2 40 103.4e344.7 (0.179e1.047) � 10�5 [124]Pt(cod)(me)2 60 100e200 (3e132) � 10�5 [130]

40e80 125.9e296.4 (6.453e34.36) � 10�4 [141]Zn(dc)2 55 240.5 0.089 � 10�6 g/ml [136]Cu(dibm)2 40 103.4e344.7 (9.17e88.41) � 10�5 [124]Cu(dmhd)2 40 137.9e344.7 (3.696e36.21) � 10�5 [124]Cu(tfa)2 40 103.4e344.7 (29.60e59.38) � 10�5 [124]UO(tfa) 40 100e250 (2.20e13.7) � 10�3 M [100]Cu(tfbzm)2 40 137.9e344.7 (0.702e4.269) � 10�5 [124]trans-Cr(tfa)3 40 103.4e344.7 (147.9e272.1) � 10�5 [124]Ru(thd)2(cod) 60 100e150 (1e56) � 10�5 [130]cis-Cr(tfa)3 40 103.4e344.7 (67.30e190.8) � 10�5 [124]Mo(CO)6 40e60 76.5e113.9 (9.5e145.1) � 10�4 [142]Ni(dctp)2Cl2 35e55 108e279 (1.32e5.33) � 10�6 [143]Ti(Oi-Pr)2(dpm)2 40e60 60e200 (0.025e1.1) � 10�2 [144]UO2(dfh)2DMSO 40 100 0.12 mol/dm3 [140]trans-Co2(CO)6 [3,5-bis(CF3)C6H3P(i-C3H7)2] 40e70 100e260 (0.8e16.7) � 10�3 M [145]

a Numerical solubility values were provided only at 294 bar.

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116 109

Table 4A number of organometallic complexes and the influence of ligands on their solu-bility in supercritical carbon dioxide.

Solubility of organometalliccomplexes in increasing order

Reference

(a) Cu(bzac)2 < Cu(acac)2 < Cu(tfbzm)2 < Cu(dmhd)2 < Cu(tfa)2< Cu(thd)2 < Cu(dibm)2 < Cu(tod)2 < Cu(hfa)2$H2O < Cu(hfa)2

[124]

(b) Cu(pdc)2 < Cu(edc)2 < Cu(p3dc)2 < Cu(bdc)2 < Cu(p5dc)2< Cu(hdc)2

[122]

(c) Cu(edc)2 < Cu(fddc)2 [123](d) Ni(hfa)2$2H2O < Ni(hfa)2 [108](e) Ru(thd)3 < Ru(thd)2(cod) [130](f) Rh(acac)2 < Rh(acac)(cod) [130](g) Co(acac)3 < Co(thd)3; Cr(acac)3 < Cr(thd)3 [113](h) Fe(acac)3 < Fe(acac)2 < Fe(cp)2 [132]

0

20

40

60

80

100

120

140

160

180

200

250 350 450 550 650 750 850

Solu

bilit

y, S

10

5(m

ole/

mol

e C

O2)

Density (kg/m3)

Rh(acac)(cod)

Pd(acac)2

Cu(acac)2

Ni(acac)2

Ru(thd)3

Ru(thd)2(cod)

×

Fig. 8. The solubilities of a number of homoleptic and heteroleptic organometalliccomplexes at 333 K and various SCCO2 densities. Data extracted from Ref. [130].

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116110

thd ligand has larger surface area and was found to have lowercharge densities; hence, higher solubilities.

Low solubility trends were observed in studies involvingaromatic substituted ligands such as that observed for Cu(bzac)2,Cu(pdc)2 and Ni(dctp)2Cl2 [122,124,143]. In the work of Wai et al.[122] where the solubilities of various metal in a number ofdithiocarbamate ligands were compared, metal chelates with thering structure of pdc exhibit the lowest solubility. Typically, suchstructures also exhibit very high melting point [122].

Comparisonbetweenhfaand itshydrated ligand [rows (a) and (d)in Table 4] shows reduced solubility effect on the hydrated metalcomplexes. The results obtained are in agreement with otherexperimental observations [108,126]. Tenorio et al. [108] suggestedthat the lower solubility values observed for Ni(hfa)2$2H2O couldpossibly be due to the formation of intermolecular hydrogen bondsbetween the hydrated complex molecules. In terms of the highersolubility of Cu(hfa)2 than its monohydrated ligand, Lagalante et al.[124] suggested that thefive-coordinate, squarepyramidal structureof the monohydrated Cu(hfa)2 complex (as opposed to the four-coordinate square planar of Cu(hfa)2) as the cause of the differencein solubility. Investigation by Kanakubo et al. [157] also foundapparent sensitivity of the solvent densities to solute molecularstructure that affects the solubility of the solute in SCCO2.

Changing homoleptic ligands with another class of ligand intoheteroleptic ligands also influences the solubility of a metalcomplex. The solubility of heteroleptic Rh(acac)(cod) complex isobserved to be one order of magnitude higher than a number ofacac metal complexes, as shown in Fig. 8. The increase shielding ofthe metal center through the substitution of the cod ligand wassuggested as the reason for the increase solubility of Rh(acac)(cod)[130]. On the other hand, the solubility of the homoleptic Ru(thd)3was found to be higher than the heteroleptic Ru(thd)2(cod). Noreasons were provided by the authors as to the causality of thelower solubility observed for Ru(thd)2(cod).

Fig. 7. Molecular structures of cobalt (III) acetylacetonate [Co(acac)3], and tris(2,2,6,6-tetramethyl-3,5-heptanedionato)chromium (III) [Cr(thd)3].

The solubility of the metal complex is generally lower than pureligands although such influence is usually neglected in thermody-namic models [15,70,122,124]. The influence of ligands on metalcomplexes solubility has been predicted via calculated solubilityparameter of the free ligand. Generally, an increase in the solubilityparameter of a free ligand corresponds to a decrease in the complexsolubility [122,124]. However, the calculated solubility parametercouldnot offer an accuratepredictionon theorderof solubility for allinvestigated free ligands. In a study conducted by Lagalante et al.[124], the solubility parameters of a number of free ligands werecalculated in the order of: acac < dmhd < dibm < tod < thd. Exper-imentally, copper complex of tod exhibits a higher solubility thanthd, as shown in row (a) of Table 4. One possible explanation for thediscrepancy is the existence of a small permanent dipolemoment inthe Cu(tod)2 complex due to its cis square planar configuration thatis not accounted for in the calculation of the solubility parameter ofthe free ligands [124]. While both thd and tod are isomers, thd issterically rigid whereas tod is more flexible and provides a bettershielding of the metal atom compared to thd [134]. Thus, highersolubility is experimentally observed in Cu(tod)2.

2.5.2. The effects of the central metal atom on the solubility oforganometallic complexes in SCCO2

The solubilities of organometallic complexes with identicalvalence electron configuration generally decrease with increasingatomic numbers [130], as illustrated by the solubilities of Fe(cp)2,Ru(cp)2 and Os(cp)2 (Table 5). The decreasing solubilities of thethree metal cyclopentadienyl complexes from Group 8 of thePeriodic Table with increasing molecular weights are shown inTable 5. Aschenbrenner et al. [130] attributed the decrease insolubility to increased intermolecular cohesion and increased

Table 5Solubilities of various metal complexes in supercritical carbon dioxide at 333 K and150 bar (in mole/mole CO2), and their corresponding molecular weights. Solubilitydata extracted from Ref. [130].

Compound Molecular weight Solubility, S � 105

Cr(cp)2 182.18 77Mn(cp)2 185.12 9Fe(cp)2 186.03 176Co(cps)2 189.12 149Ni(cp)2 188.88 142Ru(cp)2 231.26 27Os(cp)2 320.42 9

0

50

100

150

200

250

300

350

400

200 300 400 500 600 700

Solu

bilit

y, S

×10

5(m

ole/

mol

e)

Density (kg/m3)

Ag(thd)K(thd)Rb(thd)Cu(thd)2Ni(thd)2Co(thd)3Cr(thd)3Fe(thd)3Mn(thd)3

Fig. 9. Solubility of various metalethd complexes with different oxidation states. Dataextracted from Ref. [130].

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116 111

distance between ligands and the center metal atom that result inreduced shielding of the metal by the ligands.

Examinationoffivemetal cp complexes (Cr24,Mn25, Fe26, Co27, andNi28) across Period 4 of the Periodic Table indicated that their order ofsolubility varies differently from theirmolecularweights, as shown inTable 5. The actual solubilities of the five complexeswere found to bein the order of: Mn(cp)2< Cr(cp)2 < Ni(cp)2 < Co(cp)2< Fe(cp)2. Thelow solubility of Mn(cp)2 was attributed to its polymeric chainstructurewhere aMn atom is linkedwith two other neighboringMnatoms by one bridging cyclopentadienyl ligand, resulting in a higherpolarity metaleligand bonds [130,158]. The other four metal cpcomplexes are monomeric cyclopentadienyls. Aschenbrenner et al.[130] suggested a causal relationship between the solubility of metalcomplexes with their magnetic moment and the distance betweenthe ligand and the center atom, although further investigation isrequired to substantiate the causality. They further stated thatmostofthe cp complexes have an electron deficiency or electron excess thatcontributes to the electron imbalance in the complex molecule. Theincreasing electron imbalance causes decreasing strength betweenmetal and ligand; thereby, increasing thedistancebetweenmetal andligand which reduces the shielding of the metal center [130].

A solubility study conducted byM’Hamdi et al. [131] on Cu(hfa)2,Y(hfa)3 and Ba(hfa)2 shown in Table 6, also concur with the findingsof Aschenbrenner et al. Of the three complexes, Ba(hfa)2 has thelowest solubility in SCCO2 even though it has a similar planarstructure to Cu(hfa)2. It has been suggested that the high electro-positivity of barium increases the ionic characteristic of the metaleligand bond, subsequently rendering the interactions between thenon-polar SCCO2 with Ba(hfa)2 unfavorable [131]. Y(hfa)3 on theother hand, has a lower solubility than Cu(hfa)2 due to its hex-acoordinated structure with a closer ligand cage [131,159,160].

Metal-based complexes with higher oxidation states generallyshow higher solubilities in SCCO2 [126,130]. The solubility measure-ment of a number of thd complexes shown in Fig. 9 was found to begenerally in the order of M(thd) < M(thd)2 < M(thd)3 [M ¼ metals][130]. Saito et al. [126] compared the solubilities of Mn(acac)2$2H2O,Mn(acac)3, Co(acac)2$2H2O, and Co(acac)3 with the conclusion thathigher oxidation states produce higher solubilities in SCCO2. Both thetrivalent Co and Mn complexes show higher solubilities than theirbivalent counterparts. The higher oxidation state in a metal alsocorresponds with an increase in the number of ligands that subse-quently, provides a better shielding of the center metal atom.However, the increasing solubilities of metal complexes withincreasing oxidation states of the metals, and the number of ligandssurrounding the center atom, could be tapered by the subsequentincrease in molecular mass that result in a decrease in solubility[130,161]. An example is the solubility of a tetravalent complex,Zr(thd)4 which was found to be of the same order of magnitude as itsbivalent complexes [130]. The lower than expected solubility wasreasoned to be caused by the increased steric hindrance of the ligands,due to the large space required for four ligands [130].

3. Thermodynamic modeling of soluteeSCF solubility behavior

Classical thermodynamic models can generally be categorizedinto three types, with the models being based on (a) the excessGibbs free energy or activity coefficient, (b) equations of state, and

Table 6Solubility (in mole fraction) of various hfa complexes in supercritical carbon dioxideat 170 �C and 170 bar.

Compound Solubility Crystal structure

Cu(hfa)2 >0.1 PlanarY(hfa)3 4.20 � 10�4 HexacoordinatedBa(hfa)2 8.60 � 10�6 Planar

(c) empirical methods. Generally, the solubility of metal chelates inSCCO2 is correlated either through equations of state (EOS) orempirical models. While a number of thermodynamic models isavailable, only a small number was utilized for organometalliccomplexes-SCCO2 systems. In the following sections, the thermo-dynamic models used in a number of studies related to organo-metallic complexes in SCCO2 are reviewed.

3.1. Solubility parameter, d and the regular solutions theory

Hildebrand and Scott [162] first introduced the solubilityparameter, d, to categorize soluteesolvent behavior in regularsolutions. The solubility parameters for the solute and solvent arecalculated individually, and are then compared. Heuristically,a solute can be predicted to dissolve in a solvent when the differ-ence between the solubility parameters of the solute and thesolvent is approximately�1. The solubility parameter is formulatedon the grounds that the soluteesolvent molecules mainly exhibitdispersion forces with weak chemical interactions, and is given asthe square root of the cohesional energy density, given by theequation

d ¼ffiffiffiffiffiffiffi�Ev

r(5)

where E is the cohesive energy of the solute or solvent, and v is themolar volume. E is essentially the energy of vaporization of liquid tovapor at infinite separation; representing the total energy holdingthe molecules of the liquid together. Dissolution can occur whenthe Gibbs free energy of mixing (DGm) is negative. The Gibbs freeenergy of mixing is given by

DGm ¼ DHm � TDSm (6)

where DHm is the energy change associated with intermolecularinteractions, and DSm is the energy change associatedwith a changein the molecular arrangements. Mixing occurs if DHm is negative, orless than TDSm. This reasoning forms the basis for the developmentof the Hildebrand solubility parameter. While all heats of mixing inthe Hildebrand solubility parameter calculations were assumed to

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116112

be positive, in reality, DHm can be positive or negative [163,164].Hence, given that DSm multiplied with temperature promotesa more negative DGm, and that dissolution occurs when DGm isnegative, a low DHm is beneficial to the dissolution process. SinceDHm f (d1 � d2)2, a low DHm can be attained by the differencebetween the solubility parameters of the solute and solvent.Therefore, the solubility of a solute in a solvent is maximized whenthe solubility parameters are similar. The Hildebrand solubilityparameter is built upon the regular solutions theory [162] whichassume (1) no chemical effects between molecules, (2) that theentropy of solution at constant volume is equal to an ideal solution,and (3) that the interaction energy between two molecules ofdifferent species is the geometric mean of their interaction energies[165]. The regular solutions theory however, cannot be generalizedto supercritical fluids since SCF solutions are highly compressibleand non-ideal with activity coefficients usually above 100 [65].Supercritical solutions are also highly asymmetric due to the largedisparities in the sizes and energies of the solutes and SCF. As such,the use of the geometric mean is an approximation at best [91].Hence, the solubility parameter and the activity coefficient definedin the Hildebrand solubility parameter are limited to temperaturesbelow the critical point of a component [165].

Giddings et al. [166] extended the Hildebrand solubilityparameters to supercritical fluids on the condition that the solu-bility parameter cannot be used for esters, ketones, alcohols andother polar liquids. They assumed that “(i) the solubility parameteris relevant to the physical effect as well as the chemicaleffect.and (ii) the gas obeys the van der Waals Equation of State”[3,166]. The solubility parameter formulated by Giddings et al.[166] is given as

d ¼ 1:25rrrr;1

ffiffiffiffiffiPc

p(7)

where d is the solubility parameter, Pc is the critical pressure, rr isthe reduced density of the fluid, and rr,1 is the reduced density ofthe fluid at its normal boiling point. The term 1:25P1=2c accounts forthe chemical effects and the term (rr/rr,1) accounts for the stateeffects. The solubility parameter by Giddings was found to bea good starting point for a qualitative forecast although its quan-titative prediction was poor [64]. Over the years, a number ofextended and modified solubility parameters has also been devel-oped through various modifications by Prausnitz [165], Hansen[167], Allada [64] and Fedors [168,169]. The utilization of thesolubility parameters on the order of solubility for free ligands hasbeen discussed in Section 2.5.1.

Solute solubility can be estimated via regular solutions theorythrough the ScatchardeHildebrand equation, given by

RTlng2 ¼ V2Ø21ðd2 � d1Þ2 (8)

g2 is the activity coefficient of the solute, y2 is the mole fraction ofthe solute, V is the molar volume, f1 is the volume fraction of thesolvent and T is the temperature of the system. The activity coef-ficient of the solute is related to its fugacity through

f S2f L2

¼ g2y2 (9)

where f2L is the fugacity of pure liquid of the solute and f2

S is thefugacity of the solid solute at system temperature. The fugacityratio of Equation (9) [91] can be expressed as

ln

f S2f L2

!¼ � DHf

RTm2

�Tm2

T� 1

�� lng2 (10)

where DHf is the enthalpy of fusion, R is the gas constant and Tm2 isthe melting temperature of pure solute 2. The solubility of thesolute can therefore, be calculated from the expression

y2 ¼ exp�� DHf

RTm2

�Tm2

T� 1

�� V2Ø

21

RTðd2 � d1Þ2

�(11)

Equation (11) was used by Lagalante et al. [124] to predict thesolubility of Cr(acac)3 in SCCO2. While the solubility of Cr(acac)3 atpressures above 30 MPa was adequately predicted, the forecastedsolubility at pressures below 30 MPa deviated by one to two ordersof magnitude. The high deviation from experimental resultscompelled Lagalante et al. [124] to conclude that the regular solu-tions theory framework is a poor predictive model for organome-tallic complexes in SCCO2.

3.2. Empirical methods

Empirical methods are used widely in the calculation oforganometallic complexes solubility in SCCO2. The popularity of theempirical methods is due, chiefly, to a lack of thermophysical datafor organometallic complexes required in EOS models. Whilea number of empirical methods is available in the literature[119,170e176], the Chrastil correlation [177] is the most widelyused correlation in organometallic-SCCO2 systems. The Chrastilcorrelation is a density-based model in which the solubility ofa solute is related to the density of SCFs based on the associationlaws and the entropies of the solutes in SCFs given by

S ¼ rkexpaTþ b

(12)

S is the solubility of the solute, r is the density of the SCF, k is theassociation number, a ¼ DH/R, and b is a constant. DH representsthe total heat of solvation and heat of vaporizationwhile R is the gasconstant. A linear plot of log S against log r (as in Fig. 4) would yieldk, and b. The correlation by Chrastil is widely used in organome-tallic-SCCO2, in part, due to its ease of use and accuracy[70,108,127,133,134]. The absolute standard deviation (ASD)between calculated and experimental data generally range from 2%to 20%. Density-based models such as the Chrastil correlationgenerally yield good results within 10e30 MPa (the conditionswhere most processes are conducted) but fail at extreme densityregions [69,178]. However, the Chrastil correlation is limited inscope in that its dependence on temperature allows only limitedextrapolation to a wider density range. While not rigorous, itprovides a way to compare the solubilities of organometalliccompounds over a range of densities [134].

Another empirical method used in organometallic-SCCO2systems is the density-based correlation proposed by Mendez-Santiago and Teja [170]. The Mendez-Santiago and Teja (MST)correlation, given by Equation (12), uses the enhancement factor, E,to remove the effect of sublimation pressure with the constants, Cand D independent of temperature.

TlnðEÞ ¼ C þ Dr1 (13)

Solubility data at various temperatures would fall into a straightline when TlnE is plotted against solvent density, r1. The MSTcorrelation can also be used as a consistency test on experimentaldata [170]. A number of authors correlated their experimental

Table 7PengeRobinson EOS and its parameters.

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116 113

solubility data with the MST correlation that yielded deviations(ASD) that range from 2% to 19% [108,141].

PengeRobinson EOS

P ¼ RTv� b

� avðvþ bÞ þ bðv� bÞ (15)

can be rewritten in the form of

Z3 � ð1� BÞZ2 þA� 3B2 � 2B

Z �

AB� B2 � B3

¼ 0 (16)

where

A ¼ aPR2T2

(17)

B ¼ bPRT

(18)

Z ¼ PvRT

(19)

aðTÞ ¼ aca ¼ 0:45724R2T2cPc

a (20)

a ¼"1þ k

1�

ffiffiffiffiffiTTc

s !#2(21)

k ¼ 0:37464þ 1:54226u� 0:26992u2 (22)

b ¼ bc ¼ 0:07780RTcPc

(23)

Mixing rules:

am ¼Xi

Xj

xixjaij (24)

bm ¼Xi

Xj

xixjbij (25)

Note: Tr ¼ reduced temperature; Tc ¼ critical temperature; Pc ¼ critical pressure;u ¼ acentric factor; R ¼ gas constant; v ¼ volume; a ¼ cohesion energy parameter.

3.3. Equations of state (EOSs)

Equations of state are used in engineering practice since theseequations can describe phase behavior of mixtures at wide range oftemperatures and pressure. Cubic EOSs are used widely in thechemical industry for phase equilibria calculations due to theirsimplicity, reliability and the large availability of data bases for purecompounds. Although highly regarded, a limitation of cubic EOSs isthat these equations do not represent the behavior of pure fluidsnear their critical points [179]. The earliest notable cubic EOS is thevan der Waals EOS which is expressed as the summation of anattraction pressure and a repulsion pressure, given as

P ¼ RTv� b

� av2

(14)

where P is the pressure of the fluid, R is the gas constant, T is theabsolute temperature and v is the specific volume of the system. RT/v � b accounts for the repulsion pressure while a/v2 accounts forthe attraction pressure. Parameter a is regarded as a measure of theintermolecular attraction force and treated as a constant. Parameterb is temperature independent, is related to the size of the hardspheres (molecules) and corrects for the finite volume occupied bygas molecules. As opposed to real molecules that are neither hardnor spheres, the van der Waals equation treated molecules aselastic, impenetrable, free moving hard spheres that are attractedto each other with weak, long-range forces [180]. While the van derWaals EOS is a first noteworthy attempt that accounts for thedeviation of a real gas from the ideal gas law, the simplicity of theequation renders it inappropriate for rigorous calculations.

The more robust and common cubic EOS that is also used inorganometallic-SCCO2 systems is the PengeRobinson EOS (PR-EOS)[181]. The PR-EOS is a modification of the van der Waals EOS withimprovements made to the attractive pressure term, and theinclusion of a temperature-dependent parameter a. The PR equa-tions and parameters are defined in Table 7.

In the PR-EOS, parameters a and b are regularly calculated frompure component parameters. The calculations of a and b presenteda problem for organometallic complexes as critical properties data,as well as solutes sublimation pressure data, are rarely available.Organometallic complexes usually degrade at temperatures aroundthe normal boiling point, preventing the measurement of ther-mophysical data beyond those boiling points [135]. In order toovercome the lack of thermophysical data, Cross et al. [68] usedmixture solubility data to regress for Pc and Tc. However, the use ofregressed Pc, Tc and kij values left little confidence for extrapolationand interpolation [135]. The Joback’s group contribution method[182] was used by Roggeman et al. [135] to estimate the criticalparameters of organometallic complexes. The contribution of themetal center was neglected by the authors, resulting in deviationsbetween experimental and calculated data in the range of 5e28%.

In a mixture, the parameters a and b of both pure componentsare combined into am and bm [m ¼mixture] with the use of mixingrules. Mixing rules are employed as composition dependence lawsfor both am and bm while combination rules are used to relatebinary parameters, aij and bij. A number of mixing rules andcombination rules is available in the literature [91,183e190]although in many cases, the classical van der Waals mixing rules[equations (23)e(24)] are used. The van der Waals mixing ruleshowever, are limited to molecules of similar sizes and chemicalnature. For an organometallic complex-SCCO2 system, the disparityin the sizes, energies and component critical temperatures between

the solute and the solvent are so large that kij, is required to accountfor a huge correction to the geometric mean [191]. In addition, theunlike-pair interaction parameter, kij, is correlated from existingdata and corresponding states theory based on critical properties.Thus, the variation in kij is also unpredictable [191].While a numberof mixing rules has been proposed for highly non-ideal mixtures,these models either fail for simple mixtures or suffer from theMichelseneKistenmacher syndrome where the mixing rule is notinvariant when dividing a component into a number of identicalsubcomponents [183,192]. A detailed discussion on mixing rulescan be found in Refs. [183,187,189,193e195].

Johnston et al. [65] made it clear that a single model cannot beused to treat all solubilitycases, and that a lackof vaporpressuredatameans that thermodynamicmodels based on the ideal gas referencestate are limited. The highly variable chemical potential of thesolutes in the dense gas region also makes predictions and correla-tions challenging [65]. Many other thermodynamic models such asthe perturbed-hard sphere EOS, the augmented van der Waalsmodel, the mean field lattice model, and the KirkwoodeBuff solu-tion theory have been utilized in the dense gas regions for organiccompounds. While the applicability of these thermodynamic

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116114

models to organometallic systems in SCCO2 would be of interest tomany, a lack of thermophysical data will continue to prevent theirutilization.

4. Conclusion

The potential applications of supercritical fluid technologies(SFT) to process organometallic compounds for biomedical, phar-maceutical and chemical purposes are extensive. The focus offundamental solubility studies however, has generally been gearedtoward applications in catalysis and waste metal extractions. Whilequite an extensive number of solubility studies has been conductedon organometallic complexes in SCCO2, most of the ligandsinvolved in the studies are the b-diketonates which are not thetypical classes of ligands used in medicinal therapeutics. A lack offundamental thermodynamic studies on metal-based therapeuticsin supercritical fluids will therefore, continue to impede thedevelopment of SFT in the biomedical field.

The solubility behavior of metal chelates in SCCO2 is affected bytheir ligand side-chains and themetal involved. Fluorinated ligandsin general, are more CO2-philic, showing higher solubility whilearomatic substituted ligands elucidate poor solubility in SCCO2.Increasing alkyl chain lengths in the ligand side-chains also tend toincrease shielding of the metal center; thereby increasing themetalchelates solubility. Metal chelates with higher oxidation states alsoshow higher solubility in SCCO2; partly due to an increasedshielding provided by the additional ligands. The polarity of themetal chelates (in which the charge of the metal atom, their crystalstructure and the distance between the metal center and its ligandscontribute to their polarity) also affect their solubilities in SCCO2.Challenges remain to accurately model the solubility behavior ofsolutes at dense gas conditions given that the organometallic-SCCO2 system is highly complex due to the extremely compressibleSCF, the asymmetric sizes and energies of the solute and solvent,and a lack of thermophysical data that will further impede the useof available thermodynamic models.

Acknowledgment

This work was supported in part by the Ministry of HigherEducation, Malaysia.

Nomenclature

a, b parameters in SRK-EOS and PR-EOSE enhancement factorGfE excess Gibbs energy at infinite pressure

GEatt attractive excess Gibbs energy

DGm Gibbs free energy of mixingDHm heat of mixingk association numberkij unlike-pair interaction parameterP system pressurePc critical pressurePsat2 vapor pressure of pure solid solutePL lower crossover pressurePU upper crossover pressureR gas constantS solubilityDSm entropy of mixingT absolute temperaturev specific volumeVm volume of mixturex, y mole fraction

Greek symbolsd solubility parameter4 volume fractionlij empirically determined, adjustable interaction parameterr densityrr reduced densityf2s fugacity coefficient of pure soluteb42 fugacity coefficient of solute in solution

References

[1] D.L. Tomasko, Z. Guo, Kirk-Othmer Encyclopedia of Chemical Technology,John Wiley & Sons, Inc., 2006, pp. 3e29.

[2] E.J. Beckman, The Journal of Supercritical Fluids 28 (2004) 121e191.[3] M.A. McHugh, V.J. Krukonis, Supercritical Fluid Extraction: Principles and

Practice, Butterworth, Stoneham, MA, 1986.[4] V. Krukonis, G. Brunner, M. Perrut, in: Proceedings of the 3rd International

Symposium on Supercritical Fluids, Strassbourg, France, 1994, pp. 1e22.[5] M.M. Esqúıvel, M.A. Ribeiro, M.G. Bernardo-Gil, The Journal of Supercritical

Fluids 14 (1999) 129e138.[6] M. Ondarza, A. Sanchez, Chromatographia 30 (1990) 16e18.[7] V.M. Rodrigues, P.T.V. Rosa, M.O.M. Marques, A.J. Petenate, M.A.A. Meireles,

Journal of Agricultural and Food Chemistry 51 (2003) 1518e1523.[8] L.T. Danh, P. Truong, R. Mammucari, N. Foster, Chemical Engineering Journal

165 (2010) 26e34.[9] P. Hubert, O.G. Vitzthum, Angewandte Chemie International Edition in

English 17 (1978) 710e715.[10] O.J. Catchpole, N.B. Perry, B.M.T. da Silva, J.B. Grey, B.M. Smallfield, The

Journal of Supercritical Fluids 22 (2002) 129e138.[11] B. Díaz-Reinoso, A. Moure, H. Domínguez, J.C. Parajó, Journal of Agricultural

and Food Chemistry 54 (2006) 2441e2469.[12] M. Herrero, J.A. Mendiola, A. Cifuentes, E. Ibáñez, Journal of Chromatography

A 1217 (2010) 2495e2511.[13] H. Tai, V.K. Popov, K.M. Shakesheff, S.M. Howdle, Biochemical Society

Transactions 35 (2007) 516e521.[14] A. Bayrakçeken, A. Smirnova, U. Kitkamthorn, M. Aindow, L. Türker, I. Eroglu,

C. Erkey, Journal of Power Sources 179 (2008) 532e540.[15] C. Erkey, The Journal of Supercritical Fluids 17 (2000) 259e287.[16] H.Q. Yu, Y. Zhang, B. Wu, N.N. Yang, X.N. Yang, X.J. Liu, Chemical Research in

Chinese Universities 27 (2011) 850e853.[17] A. Ohashi, K. Ohashi, Solvent Extraction Research and Development 15

(2008) 11e20.[18] C.M.Wai, Y. Lin, R. Brauer, S.Wang,W.F. Beckert, Talanta 40 (1993) 1325e1330.[19] Y.-C. Sun, Y.-T. Chung, M.-H. Yang, Analytical Communications 34 (1997)

333e334.[20] M.J. Cocero, Á. Martín, F. Mattea, S. Varona, The Journal of Supercritical Fluids

47 (2009) 546e555.[21] R.W. Shaw, T.B. Brill, A.A. Clifford, C.A. Eckert, E. Ulrich Franck, Chemical and

Engineering News 69 (1991) 26e39.[22] J.A. Darr, M. Poliakoff, Chemical Reviews 99 (1999) 495e542.[23] C.C. Beh, R. Mammucari, N.R. Foster, Chemical Engineering Journal 188

(2012) 1e14.[24] M.O. Balaban, X. Liao, in: M.O. Balaban, G. Ferrentino (Eds.), Dense Phase

Carbon Dioxide: Food and Pharmaceutical Applications, Wiley-Blackwell,USA, 2012.

[25] P.E. Savage, S. Gopalan, T.I. Mizan, C.J. Martino, E.E. Brock, AIChE Journal 41(1995) 1723e1778.

[26] G. Ferrentino, G. Ferrari, in: M.O. Balaban, G. Ferrentino (Eds.), Dense PhaseCarbon Dioxide: Food and Pharmaceutical Applications, Wiley-Blackwell,Oxford, UK, 2012.

[27] N.R. Foster, R. Mammucari, L.T. Danh, W.H. Teoh, in: M.O. Balaban,G. Ferrentino (Eds.), Dense Phase Carbon Dioxide: Food and PharmaceuticalApplications, Wiley-Blackwell, Oxford, UK, 2012.

[28] J.T.C. Yuan, J.S. Novak, in: M.O. Balaban, G. Ferrentino (Eds.), Dense PhaseCarbon Dioxide: Food and Pharmaceutical Applications, Wiley-Blackwell,Oxford, UK, 2012.

[29] C.A. Eckert, B.L. Knutson, P.G. Debenedetti, Nature 383 (1996) 313e318.[30] D.B. Zamble, S.J. Lippard, Trends in Biochemical Sciences 20 (1995) 435e439.[31] R.A. Alderden, M.D. Hall, T.W. Hambley, Journal of Chemical Education 83

(2006) 728.[32] J.L. Sessler, S.R. Doctrow, T.J. McMurry, S.J. Lippard, in: J.W. Kozarich (Ed.),

Medicinal Inorganic Chemistry: Promises and Challenges, American Chem-ical Society, Washington DC, 2005.

[33] R.R. Barefoot, Journal of Chromatography B: Biomedical Sciences andApplications 751 (2001) 205e211.

[34] S. Komeda, A. Casini, Current Topics inMedicinal Chemistry 12 (2012) 219e235.[35] P.C.A. Bruijnincx, P.J. Sadler, Current Opinion in Chemical Biology 12 (2008)

197e206.[36] G.Gasser, I.Ott,N.Metzler-Nolte, JournalofMedicinalChemistry54 (2011)3e25.[37] M. Sawamura, Y. Ito, Chemical Reviews 92 (1992) 857e871.[38] R.Noyori, Catalysis inOrganic Synthesis, JohnWiley and Sons,NewYork, 1994.

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116 115

[39] D. Peer, J.M. Karp, S. Hong, O.C. Farokhzad, R. Margalit, R. Langer, NatureNanotechnology 2 (2007) 751e760.

[40] S.H. van Rijt, P.J. Sadler, Drug Discovery Today 14 (2009) 1089e1097.[41] W.H. Ang, P.J. Dyson, European Journal of Inorganic Chemistry 2006 (2006)

4003e4018.[42] K. Palanichamy, N. Sreejayan, A.C. Ontko, Journal of Inorganic Biochemistry

106 (2012) 32e42.[43] Z. Liu, A. Habtemariam, A.M. Pizarro, G.J. Clarkson, P.J. Sadler, Organome-

tallics 30 (2011) 4702e4710.[44] S. Craig, L. Gao, I. Lee, T. Gray, A.J. Berdis, Journal of Medicinal Chemistry 55

(2012) 2437e2451.[45] F. Greco, M.J. Vicent, Advanced Drug Delivery Reviews 61 (2009) 1203e1213.[46] F. Schmitt, L. Juillerat-Jeanneret, Anticancer Agents in Medicinal Chemistry

12 (2012) 500e525.[47] K.B. Garbutcheon-Singh, M.P. Grant, B.W. Harper, A.M. Krause-Heuer,

M. Manohar, N. Orkey, J.R. Aldrich-Wright, Current Topics in MedicinalChemistry 11 (2011) 521e542.

[48] P. Horcajada, R. Gref, T. Baati, P.K. Allan, G. Maurin, P. Couvreur, G. Férey,R.E. Morris, C. Serre, Chemical Reviews 112 (2012) 1232e1268.

[49] G. Jaouen, P. Dyson, in: H.C. Robert, D.M.P. Mingos (Eds.), ComprehensiveOrganometallic Chemistry III, Elsevier, Oxford, 2007, pp. 445e464.

[50] C.G. Hartinger, P.J. Dyson, Chemical Society Reviews 38 (2009) 391e401.[51] A. Cabañas, J.M. Blackburn, J.J. Watkins, Microelectronic Engineering 64

(2002) 53e61.[52] K.S. Morley, P.B. Webb, N.V. Tokareva, A.P. Krasnov, V.K. Popov, J. Zhang,

C.J. Roberts, S.M. Howdle, European Polymer Journal 43 (2007) 307e314.[53] K. Esumi, S. Sarashina, T. Yoshimura, Langmuir 20 (2004) 5189e5191.[54] A. Campos-Carrasco, C.T. Estorach, A. Bastero, M. Reguero, A.M. Masdeu-

Bultó, G. Franciò, W. Leitner, A. D’Amora, B. Milani, Organometallics 30(2011) 6572e6586.

[55] J.Wang, J. Chen, Y. Yang, The Journal of Supercritical Fluids 33 (2005) 159e172.[56] P.E. Savage, Chemical Reviews 99 (1999) 603e621.[57] A. Baiker, Chemical Reviews 99 (1999) 453e473.[58] P.G. Jessop, T. Ikariya, R. Noyori, Chemical Reviews 99 (1999) 475e494.[59] W. Leitner, Accounts of Chemical Research 35 (2002) 746e756.[60] W. Leitner, Pure and Applied Chemistry 76 (2004) 635e644.[61] T. Ikariya, Y. Kayaki, Catalysis Surveys from Japan 4 (2000) 39e50.[62] J.D. Grunwaldt, R. Wandeler, A. Baiker, Catalysis Reviews: Science and

Engineering 45 (2003) 1e96.[63] C.M. Wai, S. Wang, Journal of Chromatography A 785 (1997) 369e383.[64] S.R. Allada, Industrial & Engineering Chemistry Process Design and Devel-

opment 23 (1984) 344e348.[65] K.P. Johnston, D.G. Peck, S. Kim, Industrial & Engineering Chemistry Research

28 (1989) 1115e1125.[66] S.J. Macnaughton, D.L. Tomasko, N.R. Foster, C.A. Eckert, Process Safety and

Environmental Protection 71 (B2) (1993) 124e128.[67] F.P. Lucien, The University of New South Wales, Sydney, 1996, p. 46.[68] W. Cross, A. Akgerman, C. Erkey, Industrial & Engineering Chemistry

Research 35 (1996) 1765e1770.[69] M. Skerget, Z.e. Knez, M.a. Knez-Hrncic, Journal of Chemical & Engineering

Data 56 (2011) 694e719.[70] N.G. Smart, T. Carleson, T. Kast, A.A. Clifford, M.D. Burford, C.M. Wai, Talanta

44 (1997) 137e150.[71] C.M. Wai, B. Waller, Industrial & Engineering Chemistry Research 39 (2000)

4837e4841.[72] O. Kajimoto, Chemical Reviews 99 (1999) 355e390.[73] F.P. Lucien, N.R. Foster, in: P.G. Jessop, W. Leitner (Eds.), Chemical Synthesis

Using Supercritical Fluids, Wiley-VCH, Weinheim, 1999.[74] R. Dohrn, J.M.S. Fonseca, S. Peper, Annual Review of Chemical and Biomo-

lecular Engineering 3 (2012) 343e367.[75] J.M.S. Fonseca, R. Dohrn, S. Peper, Fluid Phase Equilibria 300 (2011) 1e69.[76] R. Dohrn, S. Peper, J.M.S. Fonseca, Fluid Phase Equilibria 288 (2010) 1e54.[77] M. Christov, R. Dohrn, Fluid Phase Equilibria 202 (2002) 153e218.[78] R.E. Fornari, P. Alessi, I. Kikic, Fluid Phase Equilibria 57 (1990) 1e33.[79] U.K. Deiters, G.M. Schneider, Fluid Phase Equilibria 29 (1986) 145e160.[80] C.J. Peters, K. Gauter, Chemical Reviews 99 (1999) 419e432.[81] J.A. Schouten, Physics Reports 172 (1989) 33e92.[82] G. Schneider, Berichte der Bunsengesellschaft für physikalische Chemie 70

(1966) 10e16.[83] G. Schneider, Berichte der Bunsengesellschaft für physikalische Chemie 70

(1966) 497e520.[84] H. Gamsjager, J.W. Lorimer, P. Scharlin, D.G. Shaw, Pure and Applied

Chemistry 80 (2008) 233e276.[85] M.Z. Özel, K.D. Bartle, A.A. Clifford, M.D. Burford, Analytica Chimica Acta 417

(2000) 177e184.[86] U. Domanska, in: T.M. Letcher (Ed.), Developments and Applications in

Solubility, RSC Publishing, UK, 2007.[87] G.W. Castellan, Physical Chemistry, Addison-Wesley Publishing, Reading, MA,

1971.[88] C.T. Lira, in: B.A. Charpentier, M.R. Sevenants (Eds.), Supercritical Fluid

Extraction and Chromatography: Techniques and Applications, ACS Sympo-sium Series, 1988, pp. 1e25.

[89] J.F. Kauffman, The Journal of Physical Chemistry A 105 (2001) 3440e3442.[90] G.G. Yee, J.L. Fulton, R.D. Smith, The Journal of Physical Chemistry 96 (1992)

6172e6181.

[91] J.M. Prausnitz, R.N. Lichtenthaler, E.G. de Azevedo, Molecular Thermody-namics of Fluid-phase Equilibria, Prentice-Hall Inc., Englewood Cliffs, NJ, 1969.

[92] R.C. Dougherty, The Journal of Chemical Physics 109 (1998) 7372e7378.[93] C. Czeslik, J. Jonas, Chemical Physics Letters 302 (1999) 633e638.[94] P. Raveendran, Y. Ikushima, S.L. Wallen, Accounts of Chemical Research 38

(2005) 478e485.[95] T. Tsukahara, Y. Kachi, Y. Kayaki, T. Ikariya, Y. Ikeda, The Journal of Physical

Chemistry B 112 (2008) 16445e16454.[96] P. Raveendran, S.L. Wallen, Journal of the American Chemical Society 124

(2002) 12590e12599.[97] Y. Kachi, T. Tsukahara, Y. Kayaki, T. Ikariya, J. Sato, Y. Ikeda, The Journal of

Supercritical Fluids 40 (2007) 20e26.[98] J.D. Sutor, Nature 195 (1962) 68e69.[99] T. Steiner, Crystallography Reviews 6 (1996) 1e51.

[100] T. Tsukahara, Y. Kachi, Y. Kayaki, T. Ikariya, Y. Ikeda, The Journal of Super-critical Fluids 39 (2006) 6e12.

[101] S.L. Draper, G.A.Montero, B. Smith, K. Beck, Dyes andPigments 45 (2000) 177e183.

[102] K.E. Anderson, J.I. Siepmann, The Journal of Physical Chemistry B 112 (2008)11374e11380.

[103] D.Tuma, B.Wagner,G.M.Schneider, FluidPhaseEquilibria182 (2001) 133e143.[104] J. Fasihi, Y. Yamini, F. Nourmohammadian, N. Bahramifar, Dyes and Pigments

63 (2004) 161e168.[105] B. Guzel, A. Akgerman, Journal of Chemical & Engineering Data 44 (1998)

83e85.[106] H.-D. Sung, J.-J. Shim, Journal of Chemical & Engineering Data 44 (1999)

985e989.[107] S.G. Kazarian, P.A. Hamley, M. Poliakoff, Journal of the American Chemical

Society 115 (1993) 9069e9079.[108] M.J. Tenorio, A. Cabañas, C. Pando, J.A.R. Renuncio, The Journal of Super-

critical Fluids 70 (2012) 106e111.[109] J.M. Dobbs, J.M. Wong, R.J. Lahiere, K.P. Johnston, Industrial & Engineering

Chemistry Research 26 (1987) 56e65.[110] J.F. Kincaid, H. Eyring, The Journal of Chemical Physics 6 (1938) 620e629.[111] G.M. Kontogeorgis, in: C.M. Hensen (Ed.), Hansen Solubility Parameters: A

User’s Handbook, CRC Press, Boca Raton, 2007, pp. 77e94.[112] S.K. Kumar, K.P. Johnston, The Journal of Supercritical Fluids 1 (1988) 15e22.[113] M. Haruki, F. Kobayashi, M. Okamoto, S.-i. Kihara, S. Takishima, Fluid Phase

Equilibria 297 (2010) 155e161.[114] W.J. Schmitt, R.C. Reid, Journal of Chemical & Engineering Data 31 (1986)

204e212.[115] X. Wang, L.L. Tavlarides, Industrial & Engineering Chemistry Research 33

(1994) 724e729.[116] P.G. Debenedetti, Chemical Engineering Science 42 (1987) 2203e2212.[117] B.C.-Y. Lu, D. Zhang, W. Sheng, Pure and Applied Chemistry 62 (1990) 2277e

2285.[118] A. Vetere, The Journal of Supercritical Fluids 12 (1998) 129e139.[119] D.H. Ziger, C.A. Eckert, Industrial & Engineering Chemistry Process Design

and Development 22 (1983) 582e588.[120] A.H. Harvey, The Journal of Physical Chemistry 94 (1990) 8403e8406.[121] N.R. Foster, G.S. Gurdial, J.S.L. Yun, K.K. Liong, K.D. Tilly, S.S.T. Ting,

H. Singh, J.H. Lee, Industrial & Engineering Chemistry Research 30 (1991)1955e1964.

[122] C.M. Wai, S. Wang, J.-J. Yu, Analytical Chemistry 68 (1996) 3516e3519.[123] K.E. Laintz, C.M. Wai, C.R. Yonker, R.D. Smith, The Journal of Supercritical

Fluids 4 (1991) 194e198.[124] A.F. Lagalante, B.N. Hansen, T.J. Bruno, R.E. Sievers, Inorganic Chemistry 34

(1995) 5781e5785.[125] M. Ashraf-Khorassani, M.T. Combs, L.T. Taylor, Talanta 44 (1997) 755e763.[126] N. Saito, Y. Ikushima, T. Goto, Bulletin of the Chemical Society of Japan 63

(1990) 1532e1534.[127] S. Yoda, Y. Mizuno, T. Furuya, Y. Takebayashi, K. Otake, T. Tsuji, T. Hiaki, The

Journal of Supercritical Fluids 44 (2008) 139e147.[128] C. Kersch, G.J. Witkamp, G.F. Woerlee, Solvent Extraction and Ion Exchange

23 (2005) 189e212.[129] M. Haruki, F. Kobayashi, S.-i. Kihara, S. Takishima, Journal of Chemical &

Engineering Data 56 (2011) 2230e2235.[130] O. Aschenbrenner, S. Kemper, N. Dahmen, K. Schaber, E. Dinjus, The Journal

of Supercritical Fluids 41 (2007) 179e186.[131] R. M’Hamdi, J.F. Bocquet, K. Chhor, C. Pommier, The Journal of Supercritical

Fluids 5 (1992) 55e59.[132] M. Haruki, F. Kobayashi, S.-i. Kihara, S. Takishima, Fluid Phase Equilibria 308

(2011) 1e7.[133] C.Y. Kong, K. Sone, T. Sako, T. Funazukuri, S. Kagei, Fluid Phase Equilibria 302

(2011) 347e353.[134] W.C. Andersen, R.E. Sievers, A.F. Lagalante, T.J. Bruno, Journal of Chemical &

Engineering Data 46 (2001) 1045e1049.[135] E.J. Roggeman, A.M. Scurto, J.F. Brennecke, Industrial & Engineering Chem-

istry Research 40 (2000) 980e989.[136] J. Wang, W.D. Marshall, Analytical Chemistry 66 (1994) 1658e1663.[137] C.M. Cowey, K.D. Bartle, M.D. Burford, A.A. Clifford, S. Zhu, N.G. Smart,

N.D. Tinker, Journal of Chemical & Engineering Data 40 (1995) 1217e1221.[138] J. Wang, J. Chen, Y. Yang, Fluid Phase Equilibria 220 (2004) 147e151.[139] Y. Ding, H. Huang, Y. Ding, C. Yang, J. Ji, Advanced Materials Research 236e238

(2011) 411e416.

W.H. Teoh et al. / Journal of Organometallic Chemistry 724 (2013) 102e116116

[140] N. Miyamoto, T. Tsukahara, Y. Kachi, M. Harada, Y. Kayaki, T. Ikariya, Y. Ikeda,Journal of Nuclear Science and Technology 49 (2012) 37e46.

[141] M. Türk, M. Crone, G. Upper, The Journal of Supercritical Fluids 58 (2011) 1e6.[142] R.P. Warzinski, C.-H. Lee, G.D. Holder, The Journal of Supercritical Fluids 5

(1992) 60e71.[143] D.R. Palo, C. Erkey, Journal of Chemical & Engineering Data 43 (1998) 47e48.[144] H. Uchida, K. Sekino, Y. Hayakawa, S. Koda, The Journal of Supercritical Fluids

66 (2012) 59e65.[145] N. Dahmen, P. Griesheimer, P. Makarczyk, S. Pitter, O. Walter, Journal of

Organometallic Chemistry 690 (2005) 1467e1473.[146] A. Dardin, J.M. DeSimone, E.T. Samulski, The Journal of Physical Chemistry B

102 (1998) 1775e1780.[147] T. Umecky, M. Kanakubo, Y. Ikushima, The Journal of Physical Chemistry B

115 (2011) 10622e10630.[148] P. Raveendran, S.L. Wallen, The Journal of Physical Chemistry B 107 (2003)

1473e1477.[149] M. Kanakubo, T. Umecky, P. Raveendran, T. Ebina, Y. Ikushima, Journal of

Solution Chemistry 33 (2004) 863e874.[150] S.G. Kazarian, M.F. Vincent, F.V. Bright, C.L. Liotta, C.A. Eckert, Journal of the

American Chemical Society 118 (1996) 1729e1736.[151] A. Cece, S.H. Jureller, J.L. Kerschner, K.F. Moschner, The Journal of Physical

Chemistry 100 (1996) 7435e7439.[152] Y.-K. Han, H.Y. Jeong, The Journal of Physical Chemistry A 101 (1997) 5604.[153] P. Diep, K.D. Jordan, J.K. Johnson, E.J. Beckman, The Journal of Physical

Chemistry A 102 (1998) 2231e2236.[154] C.R. Yonker, B.J. Palmer, The Journal of Physical Chemistry A 105 (2001)

308e314.[155] C.R. Yonker, The Journal of Physical Chemistry A 104 (2000) 685e691.[156] E.G. Cox, K.C. Webster, Journal of the Chemical Society (Resumed) 162 (1935)

731e733.[157] M. Kanakubo, T. Umecky, C.C. Liew, T. Aizawa, K. Hatakeda, Y. Ikushima, Fluid

Phase Equilibria 194e197 (2002) 859e868.[158] W. Bünder, E. Weiss, Zeitschrift für Naturforschung B 33 (1978) 1235e1237.[159] N.P. Kuz’mina, G.N. Kupriyanova, S.I. Troyanov, Russian Journal of Coordi-

nation Chemistry/Koordinatsionnaya Khimiya 26 (2000) 367e372.[160] G.V. Girichev, V.V. Rybkin, N.V. Tverdova, S.A. Shlykov, N.P. Kuz’Mina,

I.G. Zaitseva, Journal of Structural Chemistry 48 (2007) 811e820.[161] C.J. Powell, E.J. Beckman, Industrial & Engineering Chemistry Research 40

(2001) 2897e2903.[162] J.H. Hildebrand, R.L. Scott, The Solubility of Nonelectrolytes, third ed., Rein-

hold, New York, 1950.[163] G. Delmas, D. Patterson, T. Somcynsky, Journal of Polymer Science 57 (1962)

79e98.[164] C. Hansen, Hansen Solubility Parameters: A User’s Handbook, second ed.,

CRC Press, Boca Raton, 2007.[165] J.M. Prausnitz, AIChE Journal 4 (1958) 269e272.[166] J.C. Giddings, M.N. Myers, L. McLaren, R.A. Keller, Science 162 (1968) 67e73.[167] C.M. Hansen, Journal of Paint Technology 39 (1967) 505e510.[168] R.F. Fedors, Polymer Engineering & Science 14 (1974) 147e154.[169] R.F. Fedors, Polymer Engineering & Science 14 (1974) 472.[170] J.Méndez-Santiago, A.S. Teja, Fluid Phase Equilibria 158e160 (1999) 501e510.[171] J.M. Del Valle, J.M. Aguilera, Industrial & Engineering Chemistry Research 27

(1988) 1551e1553.[172] Z.-R. Yu, B. Singh, S.S.H. Rizvi, J.A. Zollweg, The Journal of Supercritical Fluids

7 (1994) 51e59.[173] M.D. Gordillo, C. Pereyra, E.J. Martinez de la Ossa, The Journal of Supercritical

Fluids 27 (2003) 31e37.[174] M.D. Gordillo, M.A. Blanco, A. Molero, E. Martinez de la Ossa, The Journal of

Supercritical Fluids 15 (1999) 183e190.[175] G.M. Anderson, S. Castet, J. Schott, R.E. Mesmer, Geochimica et Cosmochi-

mica Acta 55 (1991) 1769e1779.[176] F.E. Wubbolts, O.S.L. Bruinsma, G.M. van Rosmalen, The Journal of Super-

critical Fluids 32 (2004) 79e87.[177] J. Chrastil, The Journal of Physical Chemistry 86 (1982) 3016e3021.[178] C.B. Kautz, G.M. Schneider, J.-J. Shim, B. Wagner, D. Tuma, Journal of

Chemical & Engineering Data 53 (2008) 2356e2371.[179] J.M.H.L. Sengers, R.F. Chang, G. Morrison, in: Equations of State, American

Chemical Society, 1986, pp. 110e131.[180] A.I. Kipnis, B.E. Efimovich, J.S. Rowlinson, Van der Waals and Molecular

Sciences, Clarendon Press, Oxford, 1996.

[181] D.-Y. Peng, D.B. Robinson, Industrial & Engineering Chemistry Fundamentals15 (1976) 59e64.

[182] K.G. Joback, R.C. Reid, Chemical Engineering Communications 57 (1987)233e243.

[183] P. Ghosh, Chemical Engineering & Technology 22 (1999) 379e399.[184] A.Z. Panagiotopoulos, R.C. Reid, Fluid Phase Equilibria 29 (1986) 525e534.[185] M.-J. Huron, J. Vidal, Fluid Phase Equilibria 3 (1979) 255e271.[186] J. Ahlers, J. Gmehling, Industrial & Engineering Chemistry Research 41 (2002)

3489e3498.[187] R.J. Sadus, AIChE Journal 40 (1994) 1376e1403.[188] S.J. Park, T.Y. Kwak, G.A. Mansoori, International Journal of Thermophysics 8

(1987) 449e471.[189] J. Chen, K. Fischer, J. Gmehling, Fluid Phase Equilibria 200 (2002) 411e429.[190] H.A. Lorentz, Annalen der Physik 12 (1881) 127e136.[191] J.M. Wong, R.S. Pearlman, K.P. Johnston, The Journal of Physical Chemistry 89

(1985) 2671e2675.[192] M.L. Michelsen, H. Kistenmacher, Fluid Phase Equilibria 58 (1990) 229e230.[193] J. Vidal, Chemical Engineering Science 33 (1978) 787e791.[194] P.M.Mathias,H.C. Klotz, J.M. Prausnitz, Fluid PhaseEquilibria 67 (1991) 31e44.[195] J. Chen, J.-G. Mi, K.-Y. Chan, Fluid Phase Equilibria 178 (2001) 87e95.

Wen Hui Teoh B(Hons.) Chem. Eng. Universiti KebangsaanMalaysia, MPhil Cambridge.Wen Hui Teoh is a fellow of the University of Malaya and iscurrently a PhD student under Prof. Neil Foster and Dr.Raffaella Mammucari.

Raffaella Mammucari B.E. University of Naples (Italy),PhD University of New South Wales (Australia) RaffaellaMammucari is a Senior Research Associate at the Univer-sity of New South Wales, Australia. Her research field is insupercritical fluids and gas expanded media technology.Her research interests include particle engineering appliedto active pharmaceutical ingredients, polymer bio-blendsand scaffolds, the production of magnetically responsiveparticles; the development of polymer vesicles for medicalapplications, material processing by sub-critical water, andbiosynthesis in non-conventional media.

Neil R. Foster UNSW Scientia Professor, BSc PhD DSc FRACIFIEAust FTSE Chart. Chem. CPEng NPER, HonoraryProfessor (BUCT), PRC Qian Ren 1000 Talents ProgrammeProfessor Neil Foster is a Scientia Professor at the Univer-sity of New South Wales, Australia. He is also a People’sRepublic of China “Qian Ren” 1000 Talents NationalProfessor based at Beijing University of Technology. He hasbeen involved in supercritical fluids and dense gasresearch for 30 years, more recently in the area of nano-medicine and re-engineering of pharmaceuticals. He has260 publications, 130 of which are in peer reviewed scien-tific media. Neil is also a named inventor on 13 patentapplications, 4 of which have been granted. Neil is alsothe founder and Director of Bioparticle Technologies P/L.