Co

Ca

b

a

ARRA

KHSECSI

1

nparaipcioapCwwaaovc

0d

J. of Supercritical Fluids 61 (2012) 191– 198

Contents lists available at SciVerse ScienceDirect

The Journal of Supercritical Fluids

jou rn al h om epage: www.elsev ier .com/ locate /supf lu

ombination of supercritical carbon dioxide and ionic liquid in a novel assemblyf carvacrol

atarina I. Meloa, Rafał Bogel-Łukasika,b, Ewa Bogel-Łukasika,∗

REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, PortugalLaboratório Nacional de Energia e Geologia, I.P., Unit of Bioenergy, Estrada do Pac o do Lumiar 22, 1649-038 Lisbon, Portugal

r t i c l e i n f o

rticle history:eceived 2 September 2011eceived in revised form 4 October 2011ccepted 5 October 2011

a b s t r a c t

Exploitation of neoteric solvents in the efficient total synthesis of carvacrol, an antimicrobial reagent,has been accomplished. The first reaction aimed supercritical mediated heterogeneous hydrogenationof carvone catalysed by palladium. The second step guided to an application of great salvation capacityof ionic liquids in separation of carvacrol from the post-hydrogenation mixture. The process has beencompleted by use of a high pressure CO2 potential to extract carvacrol from the ionic liquid reservoir.

eywords:ydrogenationeparationxtractionarvacrolupercritical carbon dioxide

© 2011 Elsevier B.V. All rights reserved.

onic liquids

. Introduction

Carvacrol is a natural plant chemical, one of the major compo-ents of oregano and thyme essential oils. Due to its antimicrobialroperties and low toxicity accompanied by odour of oregano and

pizza-like taste, carvacrol is a food additive to prevent bacte-ial contamination [1]. Among antimicrobials, carvacrol has beenpplied to potato puree, table grapes, chicken and sprouts [2]. Anncorporation of carvacrol into the Gelidium corneum film used forrocessed food and pharmaceutical products improved an antimi-robial activity as well as a mechanical property of the film. Due tots specific properties, carvacrol is an important chemical and meth-ds of synthesis different than extraction from the natural plantsre studied extensively. Up to now, the moderate to high tem-erature methods for production of carvacrol have been reported.arvacrol was produced by Friedel–Crafts alkylation of o-cresolith isopropanol using superacidic catalysts at 180 ◦C within 2 hith selectivity of 82% at almost complete conversion [3]. In arom-

tization of carvone, in presence of 6 M H2SO4 with Amberlyst 15cid catalyst started at 75 ◦C and proceeded by an exothermic effect

f the reaction at 180 ◦C within 1 min, carvone was completely con-erted into carvacrol [4]. In other acid catalysed isomerisation ofarvone to carvacrol, use of 6 M sulphuric acid was reported [5,6].

∗ Corresponding author. Tel.: +351 212948500 ext. 10941; fax: +351 212948550.E-mail address: ewa@dq.fct.unl.pt (E. Bogel-Łukasik).

896-8446/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.supflu.2011.10.005

The isomerisation of carvone in water at 250 ◦C yielded a 95% of car-vacrol [7]. All of the mentioned methods of carvacrol synthesis arecharacterised by relatively high yields of the reaction (60–100%),but unfortunately their drawbacks are the use of hazardous chemi-cals, of moderate or high temperature and the production of wastes’stream.

Supercritical CO2, a promising solvent for organic synthesis [8],applied as a medium in hydrogenation facilitates reaction [9] due tototal miscibility of gases and reduction in mass transfer limitationsand viscosities of the obtained mixtures. The favourable solubilitiesof CO2 in terpenes [10–12] and terpenoids [13] allowed executingthe hydrogenation reaction in the high-pressure CO2 [14].

Ionic liquids (ILs) are other solvents offering the flexible physi-cal properties. They are characterised by near-zero vapour pressure[15], thermal stability [16] and widely tuneable properties asregards polarity, hydrophobicity and miscibility with other sol-vents [17,18] through appropriate modification of the cation andanion. The negligible vapour pressure of ILs is an interesting featurewhich makes them easily recyclable and reusable. Furthermore, avery low vapour pressure of ILs reduces the risk of explosion that isa clear advantage over the use of the classical volatile solvents [15].Ionic liquids have found an application in many chemical reactions[19–22] and processes including extraction and separation [23–27].

In this work, we attempted to produce carvacrol for the first time

in mild conditions using alternative solvents in a three-step processaiming at: (1) synthesis of desired carvacrol in the heterogeneouslycatalysed hydrogenation of cheap carvone occurred in scCO2; (2)separation of carvacrol from the post-hydrogenation mixture using

1 ritical

al

h5seoAgtnd

creIt

2

2

wavBwa1[rITa(wwmc2

GacgH2ahps

LuHuotvtiS

92 C.I. Melo et al. / J. of Superc

n ionic liquid as a solvent; (3) extraction of carvacrol from the ioniciquid reservoir by scCO2.

The first aim was accomplished by the palladium catalysedydrogenation of carvone executed in scCO2. Carvone, 2-methyl--(prop-1-en-2-yl)cyclohex-2-enone, is one of the most availableources to obtain carvacrol as is produced yearly in the scalexceeding of 1500 tonnes [28,29]. Carvone is a major constituentf a spearmint essential oil of Menthe spicata (Spearmint) [30] andnethum graveolens (Dill) [28]. Carvone exhibits a significant anal-esic activity [31] and is an important chiron for the synthesis ofhe complex natural products [32]. Carvone is a terpenoid in whichon-polar CO2 is highly soluble thus carvone is a promising candi-ate for hydrogenation in scCO2.

The second target was achievable due to the specific solventapability of IL to disclose a subtle difference in solubilities of theeaction products. The last aim was gained due to the specific prop-rties of ionic liquids and scCO2. The near-zero vapour pressure ofLs in combination with scCO2 allowed for separation and extrac-ion of carvacrol.

. Materials and methods

.1. Chemicals

Hydrogen and carbon dioxide were supplied by Air Liquideith a stated purity of 99.998 mol%. R-(−)carvone (purity: >99.0%)

nd carvacrol (purity: 95.0%) were supplied by TCI Europe. Car-omenthone, sum of isomers (purity: 98.0%) was delivered fromedoukian Research, Inc. These substances were used as receivedithout any further purification. A 0.5 wt% Pd/Al2O3 catalyst

s pellets was bought from Aldrich. The studied ionic liquids-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide,bmim][NTf2], and 1-butyl-3-methylimidazolium trifluo-omethanesulfonate, [bmim][OTf], were supplied by Iolitec,onic Liquids Technologies with a stated purity of 99% (mass).hey were further thoroughly degassed, dried, and freed fromny small traces of volatile compounds by applying a vacuum0.1 Pa) at moderate temperature (60 ◦C). All drying proceduresere performed for 48 h, and fresh samples of the componentsere used to prepare solutions, always immediately prior toeasurements. Karl Fischer titrations revealed the following

ontents of water: [bmim][NTf2], 80 ppm and [bmim][OTf],60 ppm.

The qualitative identification of terpenes was performed byC/TOFMS using a few representative reaction samples. The sep-ration was performed with a 30 m × 250 �m i.d. fused silicaapillary column coated with a 0.25 �m thickness film of 5% phenylroups in the dimethylpolysiloxane (CP-Sil 8 CB) from Varian Inc.elium was used as a carrier gas at a flow rate between 1.5 and.0 mL min−1. The oven temperature was programmed from 50 ◦Ct 4 ◦C min−1 to 130 ◦C, from 130 ◦C at 20 ◦C min−1 to 240 ◦C, andeld isothermal at 240 ◦C for 5 min. The injector and detector tem-erature were set at 250 ◦C. The 0.5 �L of sample was injected in aplit mode, with a split ratio of 10.

The presented retention indices (linear retention indices –RIs) were calculated according to van den Dool and Kratz [33]sing a mixture of C8 to C27 hydrocarbons (AccuStandard, Newaven, CT, USA) in pentane. The hydrocarbons C9–C12 weresed for the calculations based on the GC/TOFMS results. Thebtained LRIs were compared with the literature data (cis- and

rans-carvomenthone, carvone, carvotanacetone, carvacrol, car-omenthol, carvenone) [34]. The targeted terpenes were thusentatively identified using the LRI calculations and a match of thendividual spectra with the NIST 2005 MS library using the Chem-tation software [34].

Fluids 61 (2012) 191– 198

2.2. Hydrogenation

Hydrogenations were performed in an apparatus with asapphire-windowed cell connected by a pump to a tubular reac-tor which encloses a catalyst. This apparatus has already beendescribed in details in the literature [9–11,35]. The reactants (CO2-expanded liquid terpene and hydrogen) circulate continuouslyfrom the bottom of the view cell, through the catalyst bed, andback to the upper entrance of the cell. The flow rate of 3.3 mL min−1

employed guarantees that during 1–2 min the whole reactionmixture recirculates at least once [36,37]. Due to this, the feedcomposition changed rapidly that simulates a process performedin a batch mode. The reaction samples were taken at regular inter-vals through a HPLC valve equipped with a 100 �L sampling loop.The contents of the loop were dissolved in a measured amount ofpentane with nonane used as an internal standard for the sampleanalysis. CO2 was vented to the atmosphere. For each reaction 1.5 gof the fresh catalyst and 2 mL of carvone were used. The qualita-tive identification of products was based on comparisons betweenthe obtained GC-FID LRIs and those obtained by GC/TOFMS, andauthentic samples. The quantitative analysis was carried out bythe gas chromatography by means of a Varian Chrompac CP-3800gas chromatograph. Hydrogen and air were used as carrier gasesat a flow rate of 30 mL min−1 and 300 mL min−1, respectively. Themethod of analysis and column was described before in the qualita-tive identification of terpenes. The response factor for the analysedterpenes was determined as follows: cis- and trans-carvomenthone– 1.1, carvone – 1.1, carvotanacetone – 1.1, carvacrol – 1.0, carvo-menthol – 1.0, carvenone – 1.1.

2.3. Liquid–liquid equilibrium (LLE)

The temperature–composition phase diagrams were obtainedat ambient pressure of 0.1 MPa and in a temperature range of273.15–350.15 K using a visual method (observation, by eye, ofthe turbidity (cloud point) that preceded phase separation). Exper-iments were performed in a Pyrex-glass cell that was equippedwith a stirrer. The cell could be open/close by a Teflon valve atthe end of a long, capillary-thin (inner diameter of 0.1 mm) neck. Itallowed the cell to be deeply immersed in a temperature-controlledbath while, at the same time, diminished losses due to an evapo-ration. The solutions were prepared in the cell by adding liquidsthrough the neck using Hamilton syringes with the lengthy needle.The liquid inside the cell always occupied almost its entire volume(0.5 mL). The compositions were measured gravimetrically using aMettler Model AT201 analytical microbalance with a stated accu-racy (repeatability) of 2 × 10−5 in mass fraction. A temperature wasmonitored using a Pt100 temperature probe with the accuracy of0.03 K.

In each LLE experiment, a heterogeneous solution was continu-ously heated with constant stirring from room temperature until itbecame homogeneous. Then, starting from the one-phase region,the solution was cooled continuously and the temperature at whichit became turbid was taken as a cloud point. The experiments wereperformed in 3–4 runs; with the last run being performed veryslowly (the rate of temperature change near the expected cloudpoint was no more than 5 K h−1). The uncertainties in tempera-ture determination taken from the repeated measurements neverexceeded 0.5 K.

The ternary LLE investigations were performed in 5 mL vialswith the closed lid. A known quantity of the liquid containing car-vacrol and carvone mixture was placed inside and solvent (ILs or

water) were added. Various compositions of solutes and solventswere tested to achieve broad range of experimental points. Next,the vials placed in a constant temperature (275.15 K) in the orbitalshaker set for 400 rpm for 1 h and samples from ketone (upper)

C.I. Melo et al. / J. of Supercritical Fluids 61 (2012) 191– 198 193

O

O

O

O

HO

carvone

carvacrol

carvenone

trans-carvomenthone

carvotanacetone

O

cis-carvomenthone

HO

H2

H2

-H2

+

H2

H2

nation

aa

tvu

2

s2ecpwtJ(msA

s(wtgp

carvomenthol

Fig. 1. The hydroge

nd IL-rich phase (bottom) were taken and analysed by a GC using method for the analysis of the hydrogenation reaction samples.

The separation studies were performed in the same manner asernary LLE studies mentioned above; however instead of the car-acrol and carvone mixture, the post-hydrogenation mixture wassed.

.4. Extraction study

The extraction studies were performed in the apparatus pre-ented elsewhere [38]. A sapphire-windowed cell (2.9 cm high,.4 cm internal diameter) is a core of this installation. For thextraction the sample of the IL phase from the separation studiesontaining IL (x[bmim][OTf] = 0.33), carvacrol and other minor com-ounds extracted to the IL phase during the separation studiesas used. The sample was introduced into cell using a syringe, and

he cell was placed into a water-bath. The temperature controllerULABO PC was used in order to maintain the required temperature25 ◦C). The temperature stability during experiments was deter-

ined at level ±0.1 K. The mixture was stirred driven by a magnetictirrer. The pressure in the system was measured by a manometerMPHENOL B12 with the accuracy of 0.1%.

CO2 was initially pumped into the cell through a manual pres-ure generator of a 60 mL capacity to achieve 12 MPa of pressureabove critical pressure for CO2 + carvacrol). The pressurised cell

as left for the 1 h of mixing. Afterwards, by careful opening

wo venting valves, the extract started to be collected in twolass cold traps placed in a liquid nitrogen bath. The drop ofressure in the high pressure cell was continuously compensated

reaction pathway.

by the introduction of fresh CO2 from the pressure generator.The procedure was continued until the mass of traps remainedconstant. The compositions of the liquid in the cold trap andin the high pressure reactor were analysed by a GC using amethod described in the section related to the hydrogenation reac-tion.

3. Results

3.1. Hydrogenation reaction

The hydrogenation of carvone was catalysed by a commerciallyavailable 0.5 wt% Pd/Al2O3 catalyst. Chemical structures of carvoneand reaction products are presented in Fig. 1. The reaction was car-ried out at 323.15 K, 4 MPa of H2 and 12.5 MPa of total pressure for2 h. The analysis of obtained results showed that in the reactioncourse, the concentration of carvone decreases rapidly and after25 min of the reaction, concentration of carvone is below 1 mol%.Among products of the reaction two isomers of carvomenthoneare produced chiefly with yield of 44.77% for trans- and 31.52%for cis-isomer. Apart of the two mentioned compounds, carvacrol,as a targeted product of this work, is obtained in the concentra-tion equalled to 21.95 mol%. Furthermore, in the initial stage of thereaction (first 15 min) two intermediate products (carvomenthol

and carvotanacetone) are formed and their concentrations peak upto 5% for carvomenthol and 18% for carvotanacetone, and decay inthe later stage. The third intermediate product, carvenone is formedduring the reaction; however, the concentration of this compound

194 C.I. Melo et al. / J. of Supercritical Fluids 61 (2012) 191– 198

12090603000

25

50

75

100

%mol

time/min

Fig. 2. The mole percentage of carvone (�), trans-carvomenthone (�), cis-chs

ip

3

3

iagwpbo1[flLfalaictiavltab[tmcw[mtovt

1.000.750.500.250.00270

280

290

300

310

320

330

340

350

T/K

x1

Fig. 3. The LLE phase envelopes for systems containing water (1) + carvone (2)

arvomenthone (©), carvacrol (�), carvomenthol (�), carvotanacetone (�) inydrogenation of carvone over 0.5 wt% Pd/Al2O3 catalyst. Carvenone is not pre-ented in the figure. The solid lines are provided as guide for the eye.

s lower than 1.5 mol%. The concentration profiles of substrate androducts as a function of time are illustrated in Fig. 2.

.2. The carvacrol separation

.2.1. Phase equilibriaThe liquid–liquid (LLE) phase equilibria for systems contain-

ng environmentally benign solvents (water or ionic liquids)nd carvacrol or carvone were studied. In the LLE investi-ations carvone as cheap and commercially available ketoneas used to imitate other ketones (carvomenthone isomers)resent in the post-hydrogenation mixture. Among solvents foroth hydrocarbons, carvone and carvacrol, water and two typesf ionic liquids were tested. Along with water, hydrophobic,-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide,bmim][NTf2], and hydrophilic, 1-butyl-3-methylimidazolium tri-uoromethanesulfonate, [bmim][OTf], ionic liquids were used forLE studies to examine conditions for the carvacrol separationrom the mixture obtained in the hydrogenation reaction. Thechieved results showed that LLE with water and tested ioniciquids display miscibility gap with upper critical solution temper-ture (UCST). In general, the solubility of carvacrol and carvonen water and in the hydrophobic ionic liquid shows no signifi-ant difference in solubility (Fig. 3); however the solubility of bothested compounds reveals a noticeable difference of 10 K in UCSTn a hydrophilic ionic liquid. The system with carvacrol displays

UCST at 274.85 K while for the corresponding system with car-one the critical temperature is equalled to 284.41 K. The dashedines in Fig. 3 were provided as a guide for the eye to highlighthe difference in the UCST. To facilitate separation of carvacrolnd ketones the optimal conditions had to be selected on thease of the following criteria: (1) miscibility gap for a system ofbmim][OTf] + carvone should be as broad as possible, and (2) theemperature at which separation was planned to be performed

ust guarantee complete mutual miscibility of ionic liquid andarvacrol. Both criteria were found to be accomplished at 275.15 Khich was selected arbitrary. At this temperature, on one hand, the

bmim][OTf] + carvacrol system was found to show off completeiscibility, and on the other hand, the miscibility gap for a sys-

em of [bmim][OTf] + carvone was wide enough being in the rangef 0.042 < x1 < 0.973, where x1 is the ionic liquid mole fraction. Toerify the potential of carvacrol recovery, the study of ternary sys-em containing carvacrol, carvone and [bmim][OTf] was essential.

(�), water (1) + carvacrol (2) (�), [bmim][NTf2] (1) + carvone (2) (�), [bmim][NTf2](1) + carvacrol (2) (�), [bmim][OTf] (1) + carvacrol (2) (�), and [bmim][OTf](1) + carvone (2) (©).

Results of this investigation at 275.15 K at ambient pressure aredepicted in Fig. 4.

The LLE data for ternary system for various concentrations ofsolution containing [bmim][OTf], carvacrol and carvone and thecomposition of the ketone and ionic liquid-rich phase are presentedin Table 1.

3.2.2. ModellingThe correlation of the experimental ternary LLE data was per-

formed using the Peng–Robinson equation of state (PR-EOS) [39]and the Mathias–Klotz–Prausnitz mixing rule [40]. The PE soft-ware [41] was used in modelling. The procedure of the modellingis described in details in Supporting Information. It was found thatthe equation of state and mixing rule were able to describe accu-rately the phase boundary for the examined range of composition(tie-lines and phase boundary solid lines in Fig. 4) except the criti-cal region due to the lack of the experimental points in this zone. Itis worth to underline that critical zone is one of the most difficultto model. Nevertheless, for the purpose of this work the region farfrom the critical is the most interesting and this one is describedaccurately by the experimental data.

3.2.3. Separation and extraction studiesThe separation studies were preceded by calculation of parti-

tion coefficients and separation factors. Details of the methodologyare provided in Supporting Information. The results of hydrogena-tion of carvone showed that in the post-reaction mixture the molarratio of carvacrol to ketones is like 21.95 to 78.05 which is repre-sented as open square (�) in Fig. 4. The experiments with differentcompositions of the post-hydrogenation sample and [bmim][OTf]were performed considering the obtained ternary phase equilib-rium data and calculated separation factors. The results of thisinvestigation are demonstrated in Fig. 4 and in Table 2. The sep-aration factor for the studied samples containing [bmim][OTf] andmixture from the hydrogenation reaction was found to be as high as17 and increases to 140 as the concentration of separation solvent(ILs) reaches maximum according to Table 2.

The experiments performed with the post-hydrogenation sam-ples and the [bmim][OTf] ionic liquid in various ratios (open circles

(©) in Fig. 4) confirmed the calculated results obtained from the LLEdata for ternary systems (closed circles (�) in Fig. 4). It was foundthat minimal amount of IL (x[bmim][OTf] = 0.0521) allows for thesignificant enrichment (from 21.95 mol% to 95 mol%) of carvacrol

C.I. Melo et al. / J. of Supercritical Fluids 61 (2012) 191– 198 195

[bmim][OTf]

1.00.90.80.70.60.50.40.30.20.10.0

carvacrol

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

carvone

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Fig. 4. The ternary phase diagram for carvone + [bmim][OTf] + carvacrol obtained at ambient pressure and 275.15 K. (�) – the mutual solubility of carvone and [bmim][OTf];(�) – the composition of the initial mixture and corresponding ionic liquid and ketone-rich phase composition; (�) – the composition of the mixture after the hydrogenationreaction; (©) – compositions of the mixture of the post-hydrogenation samples and [bmim][OTf]. The solid lines illustrate tie-lines and phase envelope obtained from thestudied ternary mixture and confirmed by used model. The dashed line denotes the change in the composition depending on the IL fraction in the post-hydrogenation sample.

Table 1The ambient pressure liquid–liquid equilibrium (LLE) data obtained at 275.15 K for the ternary mixture constituted by carvone, carvacrol and [bmim][OTf]. The partitioncoefficients (K) and separation factors (˛) for the mother solutions calculated from the tie-lines of the phase diagram.

Initial composition Ketone-rich phase IL-rich phase Kcarvacrol Kcarvone ˛

xcarvacrol xcarvone xcarvacrol xcarvone xcarvacrol xcarvonexIL-rich phase

carvacrol

xketone-rich phasecarvacrol

xIL-rich phasecarvone

xketone-rich phasecarvone

KcarvacrolKcarvone

0.355 0.199 0.192 0.758 0.458 0.095 2.38 0.13 19.00.353 0.356 0.139 0.813 0.396 0.080 2.85 0.10 29.10.289 0.386 0.139 0.813 0.396 0.080 2.85 0.10 29.10.188 0.636 0.127 0.825 0.379 0.076 2.98 0.09 32.4

cah

c

TT

0.151 0.425 0.057 0.896 0.226

0.129 0.656 0.075 0.878 0.275

0.049 0.775 0.034 0.919 0.143

oncentration. A higher concentration of IL (x[bmim][OTf] > 0.33)

llowed increasing the purity of carvacrol above 98 mol% or evenigher as depicted in Fig. 4 as dashed line.

The extraction studies were performed on the example sampleontaining IL (x[bmim][OTf] = 0.33) and hydrogenation products after

able 2he LLE data, partition coefficients (K) and separation factor (˛) for the separation of carv

Initial composition Ketone-rich phase IL-rich phase

xcarvacrol xcarvone xcarvacrol xcarvone xcarvacrol

0.208 0.740 0.207 0.744 0.472

0.193 0.688 0.152 0.800 0.413

0.184 0.656 0.127 0.825 0.379

0.174 0.620 0.105 0.848 0.340

0.168 0.598 0.094 0.859 0.320

0.147 0.522 0.066 0.887 0.251

0.087 0.311 0.027 0.927 0.114

0.007 0.025 0.004 0.950 0.017

0.051 3.95 0.06 69.00.058 3.68 0.07 55.70.041 4.22 0.05 94.2

separation with [bmim][OTf] (chiefly carvacrol and other minor

compounds which were co-extracted during liquid–liquid sepa-ration with IL). The extraction was performed at 12 MPa becauseat this condition CO2 and carvacrol form one phase (above criticalpressure). The GC analyses showed no traces of carvacrol remained

acrol from the post-hydrogenation mixture using examined ionic liquid.

Kcarvacrol Kcarvone ˛

xcarvonexIL-rich phase

carvacrol

xketone-rich phasecarvacrol

xIL-rich phasecarvone

xketone-rich phasecarvone

KcarvacrolKcarvone

0.099 2.28 0.13 17.10.083 2.72 0.10 26.20.076 2.98 0.09 32.40.069 3.25 0.08 40.20.065 3.40 0.08 44.80.055 3.82 0.06 62.00.038 4.27 0.04 104.00.029 4.25 0.03 139.2

1 ritical

iara

4

saatff3ov�fsbCplvlahaphnwwao2pCt

cmplaBmf

plfsmmptfas(sth

96 C.I. Melo et al. / J. of Superc

n the IL phase in the high pressure reactor and the quantitativend qualitative analyses of the cold trap demonstrated completeemoval of carvacrol from the IL phase with purity significantlybove 98 mol%.

. Discussion

The objective of the hydrogenation of carvone was to achieveuch a selectivity of carvacrol to could consider further separationnd extraction of carvacrol from the post-reaction mixture usinglternative solvents. Among several catalyst tested in hydrogena-ion of carvone the Pd/Al2O3 was selected as the most promisingor the carvacrol synthesis [42]. The hydrogenation of carvone per-ormed in the presence of supercritical CO2 was carried out at23.15 K at relatively low H2 pressure (4 MPa) and at 12.5 MPaf overall pressure. The selected conditions were taken from pre-ious studies with similar terpenes such as limonene [9] and-myrcene [11]. Similarly to limonene or myrcene there are data

or system consisting of CO2 + carvone [43,44]; however ternaryystems containing hydrogen as a third component have nevereen studied. Nevertheless based on our preceding studies forO2 + H2 + terpene [9,12] it can be assumed that at 12.5 MPa totalressure system is biphasic (gaseous CO2 + H2 and CO2-expanded

iquid carvone). This assumption was confirmed by a visual obser-ation performed during progress of the reaction. Additionally,iquid-like CO2 is not only a good solvent for hydrocarbons butlso maintains the properties characteristic for gases thus bringsydrogen to the CO2-expanded liquid phase facilitating reactionnd reducing conditions of the reaction [9]. Reactions at high-ressure conditions are relatively fast and in this particular case,ydrogenation achieves equilibrium within 30 min from the begin-ing of the reaction. The major products of carvone hydrogenationith a Pd/Al2O3 are two stereoisomers of carvomenthone. Bothere formed with concentrations equalled to 44.77 mol% for trans-

nd 31.52 mol% for cis-carvomenthone. Nevertheless, in the coursef the reaction carvacrol was obtained in the concentration of1.95 mol% as a third product. One of the main advantages of theerformed reaction is that carvacrol produced via supercriticalO2-assisted hydrogenation was obtained in reaction at modestemperature.

The pathway leading to the formation of the obtainedompounds are generally consistent with the Horiuti–Polanyiechanism [45] and Klabunovskii et al. [46]. Klabunovskii et al.

ostulated that hydrogenation of carvone catalysed by Pd-Blackeads to the first intermediate product which is carvotanacetonend in the further steps carvomenthones and carvacrol are formed.oth compounds are formed in the parallel pathways and carvo-enthones are created via hydrogenation while a second one is

ormed by the irreversible catalysis [46].One of the most common methods of recovery of a com-

ound is the liquid–liquid extraction. In a conventional way of theiquid–liquid extraction volatile solvents are used (i.e. hexane) toacilitate further purification of the final product. Recently, use ofustainable solvents such as ionic liquids becomes more popularostly because of the environmental concerns as well as afore-entioned due to a higher selectivity of ionic liquids in the specific

rocesses [20]. The LLE phase equilibrium investigations of sys-ems containing ionic liquids and studied mixture are tools suitableor development of the optimal conditions of the required sep-ration. In a case presented in this work, a subtle difference inolubilities between carvone (ketone representative) and carvacrol

analogous alcohol) was found thus optimal solvent for furthereparation was selected. Both carvacrol and carvone are based onhe hydrocarbon skeleton and contain polar groups (carbonyl andydroxyl). The fine dissimilarity between both groups guided us to

Fluids 61 (2012) 191– 198

test either hydrophobic or hydrophilic solvents. As a representa-tive of hydrophobic solvents, [bmim][NTf2] ionic liquid was usedas one of the most popular ionic liquids [47,48] being a precursorin new applications [26]. The hydrophilic solvents were repre-sented by water and [bmim][OTf]. The obtained LLE data allowedassuming that combination of the hydrophobic skeleton of car-vone and carvacrol as well as hydrophilic groups present in bothcompounds determines specific interactions between solvent andsolute. The LLE phase envelopes shown in Fig. 3 clearly indicatethat both solutes are strongly hydrophobic and therefore solubilityin hydrophobic [bmim][NTf2] is almost complete (UCSTs < 274 K)in the range of temperatures studied. Contrary to the solubility inthe hydrophobic IL, carvone and carvacrol exhibit limited affinitytowards different hydrophilic solvent such as water. For exam-ple, at temperature as high as 350 K the solubility of carvone andcarvacrol in water is negligible and is lower than 0.0005 of molefraction of solute. This behaviour can be explained by a very strongintermolecular hydrogen bond network formed by water and thusfurther reduction in solubility of hydrocarbons. The remedy fora very low solubility of hydrocarbons in water and quasi com-plete miscibility with [bmim][NTf2] is [bmim][OTf] which seemsto be ideal solvent for nonpolar skeleton of carvone and carvacrolbut at the same time is known as one of the hydrophilic ionicliquids [48]. The LLE data showed that solubility of carvone andcarvacrol in [bmim][OTf] exhibits higher UCST than in the case ofhydrophobic [bmim][NTf2] and dramatically lower than theoret-ical value of a UCST for water systems. The observed differencein solubility between both compounds can be explained by dis-similarity in structure thus difference in the chemical potential ormore precisely in the activity coefficient of both compounds inthe ionic liquid studied. There is no activity coefficient data forstudied pairs (carvone or carvacrol + [bmim][OTf]); however theactivity coefficients available for other example of pair containingketone (acetone) and alcohol (2-propanol) in [bmim][OTf] showthat at room temperature activity coefficients are 1.48 and 0.85 for2-propanol and acetone [49]. These values may indicate that in gen-eral alcohols such as 2-propanol and probably as carvacrol are moresoluble in [bmim][OTf] than ketones such as acetone and carvone.

The LLE studies for ternary samples simulating compoundspresented in the hydrogenation sample and [bmim][OTf] with dif-ferent initial composition allowed drawing ternary phase diagram.From phase diagram the partition coefficients and separation fac-tors were calculated. The obtained LLE data disclosed significantdifference between partition coefficients of carvone and carvacrolbetween [bmim][OTf] and ketone-rich phases and high values ofseparation factors. These results are caused by large differencesin solubilities of both solutes in ILs as it was explained above.These intriguing data drove us to test the selected IL with areal mixture of products from hydrogenation reaction of carvone.Analysing obtained results of real post-hydrogenation sample itwas found that carvacrol separation from the product-rich phase ispossible even with small additive (x1 = 0.0521) of the hydrophilic[bmim][OTf]. In other words carvacrol was completely recoveredfrom the ketone phase in the presence of minimal amount of ionicliquid and thus the carvacrol’s concentration in the sample from thehydrogenation reaction was enriched from 21.95 mol% to 95 mol%.The separation was repeated for different concentrations of IL. It canbe stated that even at mole fraction of IL equalled to 0.3, carvacrolwas extracted quantitatively (>98 mol%). These results confirm high

as it increases significantly from 17 to 140 at maximal mole frac-tion at which still two phases are present.

The further step of recovery of carvacrol from the IL phase

was performed using the CO2 extraction technique previously pre-sented in our other work [38]. The extraction with a high dense CO2was selected as more favourable than e.g. distillation due to highboiling point of carvacrol and existence of the risk of the partial or

ritical

ctc9

5

vatacpetsolt

oa

A

na

A

t

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[[

[

[

[

[

[

C.I. Melo et al. / J. of Superc

omplete IL decomposition. The extraction with CO2 proved againhat CO2 is a good solvent due to its ability to extract quantitativelyompounds such as carvacrol with purity significantly higher than8 mol% equalled to the purity after the separation step.

. Conclusions

An efficient, advantageous synthesis and separation of car-acrol recognised as antimicrobial and antioxidant agent has beenccomplished. The process presented herein has several notewor-hy features: (1) potential production of carvacrol from cheap andbundant plant rich in carvone being recognised as a building blockhemical; (2) hydrogenation performed at drastically lower tem-erature compared with the known classical processes due to theffect of CO2 which acts as a H2 co-solvent and liquid expander forerpene; (3) fine tuning of ionic liquid properties that differentiateolubility of ketones and carvacrol guiding to a successful recoveryf carvacrol with a minimal involvement of IL; (4) the phase equi-ibrium studies confirmed the IL utility as a tool for separation ofhe product being of the interest.

Concluding, the aforementioned advantages prove the potentialf alternative solvents in the innovative processes towards sustain-ble development.

cknowledgement

This work was supported by the Fundac ão para a Ciência e a Tec-ologia (FCT, Portugal) through the grants SFRH/BPD/26356/2006nd PEst-C/EQB/LA0006/2011.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at doi:10.1016/j.supflu.2011.10.005.

eferences

[1] P. Rattanachaikunsopon, P. Phurnkhachorn, Assessment of factors influencingantimicrobial activity of carvacrol and cymene against Vibrio cholerae in food,J. Bioscience and Bioengineering 110 (2010) 614–619.

[2] G.O. Lim, Y.H. Hong, K.B. Song, Application of Gelidium corneum edible filmscontaining carvacrol for ham, J. Food Science 75 (2010) C90–C93.

[3] G.D. Yadav, S.B. Kamble, Synthesis of carvacrol by Friedel–Crafts alkylationof o-cresol with isopropanol using superacidic catalyst UDCaT-5, J. ChemicalTechnology and Biotechnology 84 (2009) 1499–1508.

[4] C. Gozzi, A. Convard, M. Husset, Heterogeneous acid-catalysed isomerization ofcarvone to carvacrol, Reaction Kinetics and Catalysis Letters 97 (2009) 301–306.

[5] J.J. Ritter, D. Ginsburg, The Action of t-butyl hypochlorite on �-pinene, J. theAmerican Chemical Society 72 (1950) 2381–2384.

[6] R.A. Kjonaas, S.P. Mattingly, Acid-catalyzed isomerisation of carvone to car-vacrol, J. Chemical Education 82 (2005) 1813–1814.

[7] J.Y. An, L. Bagnell, T. Cablewski, C.R. Strauss, R. Trainor, Applications of high-temperature aqueous media for synthetic organic reactions, The J. OrganicChemistry 62 (1997) 2505–2511.

[8] J.M. Lopes, Z. Petrovski, R. Bogel-Łukasik, E. Bogel-Łukasik, Heterogeneouspalladium-catalyzed telomerization of myrcene with glycerol derivatives insupercritical carbon dioxide: a facile route to new building blocks, Green Chem-istry 13 (2011) 2013–2016.

[9] E. Bogel-Łukasik, I. Fonseca, R. Bogel-Łukasik, Y. Tarasenko, M. Nunes da Ponte,A. Paiva, G. Brunner, Phase equilibrium-driven selective hydrogenation oflimonene in high-pressure carbon dioxide, Green Chemistry 9 (2007) 427–430.

10] E. Bogel-Łukasik, A. Serbanovic, R. Bogel-Łukasik, A. Banet-Osuna, V.Najdanovic-Visak, M. Nunes da Ponte, Hydrogenation of CO2-expanded liquidterpenes: phase equilibrium-controlled kinetics, in: K. Hutchenson, A. Scurto,B. Subramaniam (Eds.), Gas-Expanded Liquids and Near-Critical Media: GreenChemistry and Engineering ACS Symposium Series 1006, Washington, DC, USA,2008, pp. 191–201.

11] E. Bogel-Łukasik, I. Gomes da Silva, I.D. Nogueira, R. Bogel-Łukasik, M. Nunes

da Ponte, Study on selectivity of �-myrcene hydrogenation in high-pressurecarbon dioxide catalysed by noble metal catalysts, Green Chemistry 11 (2009)1847–1856.

12] E. Bogel-Łukasik, A. Szudarska, R. Bogel-Łukasik, M. Nunes da Ponte,Vapour–liquid equilibrium for �-myrcene and carbon dioxide and/or hydrogen

[

Fluids 61 (2012) 191– 198 197

and the volume expansion of �-myrcene or limonene in CO2 at 323.15 K, FluidPhase Equilibria 282 (2009) 25–30.

13] M. Chatterjee, Y. Ikushima, T. Yokoyama, M. Sato, Density-dependent formationof the pure trans-isomer of the unsaturated alcohol by selective hydrogenationof citral in supercritical carbon dioxide, Advanced Synthesis and Catalysis 350(2008) 624–632.

14] E. Bogel-Łukasik, R. Bogel-Łukasik, M. Nunes da Ponte, Pt and Pd catalysedlimonene hydrogenation in high-density carbon dioxide, Monatshefte fürChemie 140 (2009) 1361–1369.

15] Y.U. Paulechka, G.J. Kabo, A.V. Blokhin, O.A. Vydrov, J.W. Magee, M. Frenkel,Thermodynamic properties of 1-butyl-3-methylimidazolium hexafluorophos-phate in the ideal gas state, J. Chemical and Engineering Data 48 (2003)457–462.

16] U. Domanska, R. Bogel-Łukasik, Physicochemical properties and solubility ofalkyl-(2-hydroxyethyl)-dimethylammonium bromide, The J. Physical Chem-istry B 109 (2005) 12124–12132.

17] E. Bogel-Łukasik, C. Lourenc o, M.E. Zakrzewska, R. Bogel-Łukasik, Insightinto the phase equilibrium phenomena of systems containing dienes anddicyanamide ionic liquids as a new potential application, The J. Physical Chem-istry B 114 (2010) 15605–15609.

18] A. Forte, E. Bogel-Łukasik, R. Bogel-Łukasik, Miscibility phenomena in systemscontaining polyhydroxy alcohols and ionic liquids, J. Chemical and EngineeringData 56 (2011) 2273–2279.

19] P. Virtanen, H. Karhu, G. Toth, K. Kordas, J.P. Mikkola, Towards one-pot synthesisof menthols from citral: modifying supported ionic liquid catalysts (SILCAs)with Lewis and Brønsted acids, J. Catalysis 263 (2009) 209–219.

20] E. Bogel-Łukasik, S. Santos, R. Bogel-Łukasik, M. Nunes da Ponte, Selectivityenhancement in the catalytic heterogeneous hydrogenation of limonene insupercritical carbon dioxide by an ionic liquid, The J. Supercritical Fluids 54(2010) 210–217.

21] M.E. Zakrzewska, E. Bogel-Łukasik, R. Bogel-Łukasik, Solubility of carbohydratesin ionic liquids, Energy and Fuels 24 (2010) 737–745.

22] M.E. Zakrzewska, E. Bogel-Łukasik, R. Bogel-Łukasik, Ionic liquid-mediatedformation of 5-hydroxymethylfurfural-a promising biomass-derived buildingblock, Chemical Reviews 111 (2011) 397–417.

23] A. Berthod, M. Ruiz-Angel, S. Carda-Broch, Ionic liquids in separation tech-niques, J. Chromatography A 1184 (2008) 6–18.

24] R. Bogel-Łukasik, Sustainable processes employing ionic liquids for secondaryalcohols separation, Monatshefte für Chemie 138 (2007) 1137–1144.

25] D. Hand, K.W. Row, Recent applications of ionic liquids in separation technol-ogy, Molecules 15 (2010) 2405–2426.

26] J. Le Bideau, L. Viau, A. Vioux, Ionogels, ionic liquid based hybrid materials,Chemical Society Reviews 40 (2011) 907–925.

27] J.G Huddleston, H.D. Willauer, R.P. Swatloski, A.E. Visser, R.D. Rogers, Roomtemperature ionic liquids as novel media for ‘clean’ liquid–liquid extraction,Chemical Communications 16 (1998) 1765–1766.

28] C.C.C.R. de Carvalho, M.M.R. Da Fonseca, Carvone: why and howshould one bother to produce this terpene, Food Chemistry 95 (2006)413–422.

29] J.O. Bledsoe Jr., in: J.I. Kroschwitz, M. Howe-Grant (Eds.), Kirk-Othmer Ency-clopedia for Chemical Technology, vol. 23, 4th ed., J. Wiley & Sons, New York,1997, pp. 833–882.

30] J. Gershenzon, M. Maffei, R. Croteau, Biochemical and histochemical localiza-tion of monoterpene biosynthesis in the glandular trichomes of Spearmint(Mentha spicata), Plant Physiology 89 (1989) 1351–1357.

31] D.P. de Sousa, E. Raphael, T.J. Brocksom, Synthesis of (R)-(−)-carvone deriva-tives, Zeitschrift für Naturforschung B 65 (2010) 1381–1383.

32] J. Wang, M. Froeyen, C. Hendrix, C. Andrei, R. Snoeck, E. Lescrinier, E. de Clereq,P. Herdewijn, (D)- and (L)-Cyclohexenyl-G, a new class of antiviral agents: syn-thesis, conformational analysis, molecular modeling, and biological activity,Nucleosides, Nucleotides and Nucleic Acids 20 (2001) 727–730.

33] H. Van Den Dool, P.D. Kratz, A generalization of retention index system includ-ing linear temperature programmed gas–liquid partition chromatography, J.Chromatography 11 (1963) 463–471.

34] Mass Spectra Library NIST version 2.0, 2005.35] E. Bogel-Łukasik, R. Bogel-Łukasik, M. Nunes da Ponte, Effect of flow rate of

a biphasic reaction mixture on limonene hydrogenation, Industrial and Engi-neering Chemistry Research 48 (2009) 7060–7064.

36] E. Bogel-Łukasik, R. Bogel-Łukasik, K. Kriaa, I. Fonseca, M. Nunes da Ponte,Limonene hydrogenation in high-pressure CO2: effect of hydrogen pressure,The J. Supercritical Fluids 45 (2008) 225–230.

37] E. Bogel-Łukasik, J. Wind, R. Bogel-Łukasik, M. Nunes da Ponte, The influenceof hydrogen pressure on the heterogeneous hydrogenation of �-myrcene in aCO2-expanded liquid, The J. Supercritical Fluids 54 (2010) 46–52.

38] R. Bogel-Łukasik, V. Najdanovic-Visak, S. Barreiros, M. Nunes da Ponte, Distri-bution ratios of lipase-catalyzed reaction products in ionic liquid supercriticalCO2 systems: resolution of 2-octanol enantiomers, Industrial and EngineeringChemistry Research 47 (2008) 4473–4480.

39] D.Y. Peng, D.B. Robinson, New 2-constant equation of state, Industrial and Engi-neering Chemistry Fundamentals 15 (1976) 59–64.

40] P.M. Mathias, H.C. Klotz, J.M. Prausnitz, Equation-of-state mixing rules for mul-

ticomponent mixtures – the problem of invariance, Fluid Phase Equilibria 67(1991) 31–44.

41] O. Pfohl, S. Petkov, G. Brunner, PE V2.9.9a – software for phase equi-libria calculations, Technische Universität Hamburg-Harburg, Hamburg,1998.

1 ritical

[

[

[

[

[

[

[

(2010) 4850–4857.

98 C.I. Melo et al. / J. of Superc

42] C.I. Melo, R. Bogel-Łukasik, M. Gomes da Silva, E. Bogel-Łukasik, Advantageousheterogeneously catalysed hydrogenation of carvone with supercritical carbondioxide, Green Chemistry (2011), doi:10.1039/c1gc15495h.

43] K.H. Kim, J. Hong, Equilibrium solubilities of spearmint oil components insupercritical carbon dioxide, Fluid Phase Equilibria 164 (1999) 107–115.

44] T. Gamse, R. Marr, High-pressure phase equilibria of the binary systemscarvone–carbon dioxide and limonene–carbon dioxide, at 30, 40 and 50 ◦C,Fluid Phase Equilibria 171 (2000) 165–174.

45] I. Horiuti, M. Polanyi, Exchange reactions of hydrogen on metallic catalysts,Transactions the Faraday Society 30 (1934) 1164–1172.

46] E.I. Klabunovskii, L.F. Godunova, L.K. Maslova, Catalytic-hydrogenation of (+)-carvone on palladium and platinum catalysts, Akademiia nauk SSSR IzvestiiaSeriia khimicheskaia (1972) 1063–1067.

[

Fluids 61 (2012) 191– 198

47] J.L. Anderson, J.K. Dixon, J.F. Brennecke, Solubility of CO(2), CH(4),C(2)H(6), C(2)H(4), O(2), and N(2) in 1-hexyl-3-methylpyridiniumbis(trifluoromethylsulfonyl)imide: comparison to other ionic liquids,Accounts Chemical Research 40 (2007) 1208–1216.

48] C.A.S Trindade, Z.P. Visak, R. Bogel-Łukasik, E. Bogel-Łukasik, M. Nunes da Ponte,Liquid–liquid equilibrium of mixtures of imidazolium-based ionic liquids withpropanediols or glycerol, Industrial and Engineering Chemistry Research 49

49] M.-L. Ge, L.-S. Wang, Activity coefficients at infinite dilution of polarsolutes in 1-butyl-3-methylimidazolium trifluoromethanesulfonate usinggas–liquid chromatography, J. Chemical and Engineering Data 53 (2008)846–849.