j our na l ho me page: www.elsev ier .com/ locate /supf lu
rospective pathway for a green and enhanced friedelin productionhrough supercritical fluid extraction of Quercus cerris cork
li S ena, Marcelo M.R. de Melob, Armando J.D. Silvestreb, Helena Pereiraa,arlos M. Silvab,∗
Centro de Estudos Florestais, Instituto Superior de Agronomia, Universidade de Lisboa, Tapada da Ajuda, 1349-017 Lisboa, PortugalCICECO, Department of Chemistry, University of Aveiro, Aveiro 3810-193, Portugal
r t i c l e i n f o
rticle history:eceived 17 October 2014eceived in revised form1 December 2014ccepted 11 December 2014vailable online 20 December 2014
eywords:
a b s t r a c t
Supercritical fluid extraction (SFE) was applied for the first time to Quercus cerris cork, and comparedwith Soxhlet with dichloromethane (DCM). Novel triterpenes, viz. betulin and squalene, and �-sitosterolwere identified for the first time in the lipophilic extracts.
The SFE at 300 bar and 40–80 ◦C provided extracts much richer in friedelin–the major compound fromboth SFE and DCM extracts – with concentrations up to 40.6 wt%, against 26.0 wt% for DCM. The SFE yieldsranged between 0.97 and 1.81 wt% with pure CO2, and attained 2.83 wt% when ethanol was introducedas cosolvent (10 wt%). In this case, however, the friedelin concentration dropped significantly due to the
additional removal of non-target compounds. In general, the experimental data and their trends were inaccordance with the theoretical predictions of kinetic and equilibrium properties estimated in this workfor the friedelin/SC-CO2 system.
This study demonstrates that the SFE of Q. cerris cork arises as a prospective pathway for a green andenhanced friedelin production process under the biorefinery concept.
. Introduction
Cork is one of the plant tissues forming the barks of somerees species. A great popularity of this natural material is alreadyredited to the cork oak (Quercus suber L.), whose transforma-ion is responsible for a myriad a products such as cork stoppers,nsulation, surfacing and paneling materials, engine joints, etc1]. Nevertheless, the Turkey oak (Quercus cerris) is also a poten-ial important provider of cork that can be found in regionsuch as Eastern Europe and Minor Asia [2]. Since in countriesike Turkey, the bark of Q. cerris is not used except for fuel, thextraction of chemicals from this forest residue may offer oppor-unities towards its integrated utilization under the biorefineryoncept.
Within oak species, the barks vary in terms of cork proportionnd spatial distribution. In the cork oak, the cork producing cellsthe phellogen or cork cambium) make up a cylindrical contin-
ous envelope of the stem and branches and are active throughhe tree’s lifetime, making up a thick layer of cork [3,4]. In mostree species the phellogen is discontinuous and short-lived, being
replaced at intervals by a new phellogen in an inner part ofthe bark, giving rise to a layered rhytidome-type bark structure.This is the case of the Turkey oak (Q. cerris) [3,4]. As cork has arather unique set of properties and offers many utilization pos-sibilities, the study of cork proportion in the Q. cerris rhytidomehas recently raised interest. Chemical summative characterizationof Q. cerris cork showed lesser suberin and higher lignin quan-tities than those of Q. suber [2]. The extractives contents weresimilar in both but the non-polar fraction was up to 11 wt% inQ. cerris cork [2] while in Q. suber cork it is around 3.6–6.0 wt%[5,6].
In the last decade the supercritical carbon dioxide (SC-CO2) hasgained importance for the extraction of chemicals from lignocellu-losic materials [7–10]. The supercritical fluid extraction (SFE) is anintensively studied separation technology that commonly allows aselective removal of compounds from complex mixtures and veg-etable matrices [8,10]. One may cite, for example, the extraction oftriterpenoids from Eucalyptus globulus bark [11,12], whose bioac-tive properties have been demonstrated [13], and the SFE of spentcoffee residues, which gives rise to extracts enriched in diterpenes
like cafestol and kahweol [14].
When the target solutes are polar, SC-CO2 can sometimes betuned through the addition of a third party polar compound tothe process, which acts as a cosolvent. In this respect, ethanol is
AbbreviationsAR aromaticCp heat capacityD12 tracer diffusion coefficientDCM dichloromethaneFA fatty acidGC–MS gas chromatography–mass spectrometrykf,i convective mass transfer coefficientLCAA long chain aliphatic alcoholM molecular weightP pressurePOL polyol� gas constantRt retention timeS entropySC-CO2 supercritical carbon dioxideSFE supercritical fluid extractionST phytosteroltr. tracesT temperatureTT triterpeneV molar volumey solubilityw Pitzer acentric factorZ compressibility factorwcork mass of cork
Greek letters� fugacity coefficient�t total extraction yield�friedelin friedelin extraction yield� viscosity� density
Subscriptb boilingc criticalCO2 relative to carbon dioxidebp at normal boiling pointi relative to species i; friedelinm melting
Superscript* equilibriumsat saturation
tp[c
rdt2aaTuas
SCF supercritical fluid
he most used cosolvent in SFE research [8] and is able to com-ensate the rather low polarity of CO2. The SFE works on grapes15], strawberries [16], and red pepper [17] are examples of suchosolvent usage.
While no SFE results have yet been published for Q. cerris, theeported extraction yields of both SFE and Soxhlet extraction withichloromethane (DCM) of Q. suber cork revealed that friedelin ishe main component of both extracts with concentrations around0 wt%, among five other triterpenes detected by 13C NMR [5]. Inddition Sousa et al. [6] identified cerine, friedelin and betulinic acids major compounds in DCM-extracted natural and boiled corks.
hese findings induced us to investigate the potential of Q. cerrisnder the context of green solvents like SC-CO2, with a particularttention to triterpenic compounds. Accordingly, the aim of the thistudy was to report for the first time the SFE of Q. cerris cork, with
luids 97 (2015) 247–255
the respective composition of extracts and the impact of operatingconditions on extraction yields and composition. These results arecompared and contrasted with available SFE and Soxhlet results ofQ. suber cork [5,6,18].
Concerning the structure of the article, Section 2 is devotedto “Materials and methods”, followed by the “Results and discus-sion” part (Section 3), which is subdivided into: Section 3.1, for theassessment of Soxhlet extraction results in terms of yield and com-position; Section 3.2 for the theoretical study of how kinetic andequilibrium properties of the SC-CO2 and friedelin evolve with tem-perature; and finally Section 3.3, for the analysis of the SFE resultsin terms of yield and friedelin concentration. The conclusions of thework are drawn in Section 4.
2. Materials and methods
2.1. Chemicals and plant material
Nonacosan-1-ol (98% purity) and �-sitosterol (99% purity)were purchased from Fluka Chemie (Madrid, Spain); ursolicacid (98% purity), betulinic acid (98% purity), and oleano-lic acid (98% purity) were purchased from Aktin Chemicals(Chengdu, China); betulonic acid (95% purity) was purchasedfrom CHEMOS GmbH (Regenstauf, Germany); palmitic acid(99% purity), dichloromethane (99% purity), pyridine (99%purity), N,O-bis(trimethylsilyl)trifluoroacetamide (99% purity),trimethylchlorosilane (99% purity), and tetracosane (99% purity)were supplied by Sigma Chemical Co. (Madrid, Spain). CO2 (99.95%)was supplied by Praxair (Porto, Portugal).
Q. cerris bark was obtained from Kahramanmaras, Turkey, andwas granulated with a hammer-type industrial mill. The resultinggranules were separated by density difference in distilled water in10 min mixing time. The floating fraction of cork-enriched granules(subsequently named cork) was dried and grinded into 20–40 mesh(0.42–0.84 mm).
2.2. Soxhlet and supercritical fluid extractions
Soxhlet extractions were carried out with DCM (120 mL) using1 g of 20–40 mesh Q. cerris cork during 8 h. The SFE experimentswere performed at constant pressure (300 bar) and CO2 flow rate(11.0 gCO2
min−1). All operating conditions may be found in Table 1:pressure (P), temperature (T), extraction time (t), mass of cork(wcork), and ethanol concentration. The influence of temperature(40–80 ◦C) and ethanol content (0 and 5 wt%) upon total extractionyield (�t) and extract composition were evaluated. One experimentwas conducted sequentially with the same biomass in four suc-cessive steps (SFE1.1, SFE1.2, SFE1.3 and SFE1.4; see Table 1) inorder to disclose the impact of varying temperature and ethanolconcentration.
2.3. Analyses of extracts by GC–MS
The composition of the SFE and Soxhlet extracts was deter-mined by gas chromatography–mass spectrometry (GC–MS).Approximately 20 mg of dried extract were converted intotrimethylsilyl (TMS) derivatives according to the literature [19].Tetracosane was used as internal standard, and the calibra-tion of the equipment for the various families of compoundswas performed using nonacosan-1-ol, �-sitosterol, ursolic acid,betulinic acid, oleanolic acid, betulonic acid, and palmiticacid.
The analyses were performed using a Trace Gas Chromatograph2000 Series equipped with a Thermo Scientific DSQ II mass spec-trometer, using helium as carrier gas (35 cm s−1), and a DB-1 J&Wcapillary column (30 m × 0.32 mm i.d., 0.25 �m film thickness).
he chromatographic conditions were as follows: initial temper-ture: 80 ◦C for 5 min; heating rates of 4 ◦C min−1 up to 260 ◦C andhen 2 ◦C min−1 until the final temperature of 285 ◦C; maintain-ng the last temperature for 10 min; injector temperature: 250 ◦C;ransfer-line temperature: 290 ◦C; split ratio: 1:50. The mass spec-rometer was operated in the electron impact mode with electron
mpact energy of 70 eV and data collected at a rate of 1 scan s−1
ver a m/z range of 33–700. The ion source was maintained at50 ◦C.
able 2ain compounds present in dichloromethane (DCM) and SC-CO2 extracts of Q. cerris cork
3.1. Soxhlet extraction of Q. cerris cork with dichloromethane(DCM)
The DCM extraction yield measured in this work for Q. cerris cork
was 4.02 wt% (Table 2), in agreement with our previous studies inwhich yields of 2.5 and 10.9 wt% were attained [2,20]. This result isalso concordant with those of Q. suber cork using the same solvent
. SFE conditions: 300 bar, 80 ◦C, 6 h, 11.0 gCO2min−1 (experiment SFE1.1 in Table 1).
DCM): Castola et al. [5,18] reported values of 6–7 wt%, Sousa et al.6] obtained 3.6 wt%, and Pereira [21] a variation between 3.5 and.4 wt%.
In terms of chemical composition, five fatty acids (FA), two longhain aliphatic alcohols (LCAA), an aromatic (AR), a polyol (POL), ahytosterol (ST) and three triterpenes (TT) were identified in theCM extract (see Table 2). It is worth noting that, like in Q. suberork case, friedelin is the most abundant compound among thosehat were successfully identified by GC–MS, making up approxi-
ately one-fourth of the total extract weight (26.02 wt%; Table 2).n addition, �-sitosterol was found as the second most abundantomponent, although with a much lower concentration: 1.25 wt%.
In comparison to previous results on Q. cerris cork DCM extracts,he detection of �-sitosterol, betulin, and squalene is a novelesult for this species. Additionally, compounds such as glycerol,enzaldehyde, hexadecanoic acid, octadec-11-enoic acid, octade-anoic, icosan-1-ol, docosan-1-ol, docosanoic and tricosanoic acidsere found in minor quantities in the extracts, with concentrations
etween 0.07% and 3.53% (wt).As far as individual molecule yields are concerned, the DCM
xtract of Q. cerris cork indicates that this biomass is a richer sourcef friedelin (10458.6 mg kg−1
cork vs. 2308 mg kg−1cork) and betulin
414.0 mg kg−1cork vs. 324 mg kg−1
cork), but a slightly scarcer one in-sitosterol (504.9 mg kg−1
cork vs. 539 mg kg−1cork) in comparison to
he Q. suber cork extract [6]. In the whole, our DCM extracts of Q.erris cork totalize 10958.4 mg kg−1
cork of triterpenes, an yield that is% greater in relation to Q. suber [6]. This advantageous result of Q.erris cork is mostly at the expenses of the substantial abundancef friedelin.
.2. Friedelin + SC-CO2 system: kinetic and equilibriumheoretical predictions
In this section the estimation of equilibrium and kinetic prop-rties of interest to interpret and understand our experimentalesults is focused. Despite being a process with considerable vari-bility concerning its operating conditions and the composition andtructure of the biomass, the discussion of the SFE of Q. cerris corkay be enriched following a theoretical approach to the extraction
esults. Accordingly, the following properties and variables werestimated: SC-CO2 density (�CO2 ) and viscosity (�CO2 ), the vaporressure of friedelin (Psat
i), the solubility (y∗
i) of pure friedelin in
C-CO2, the tracer diffusion coefficient of friedelin in SC-CO2 (D12),nd the convective mass transfer coefficient in the extractor (kf,i).
The density of carbon dioxide was calculated by Pitzer andchreiber’s model [22]; its viscosity was estimated by the empiricalquation of Altunin and Sakhabetdinov [23]; the diffusion coef-cient of friedelin was calculated by an improved hydrodynamicodel of Vaz et al. [24]; the convective mass transfer coefficientas calculated by Puiggené et al. correlation [25,26]; finally, the
riedelin solubility was estimated according to the methodologyescribed in detail by Silva et al. [27] for lycopene, and de Melot al. [28] for ursolic and 3-acetylursolic acids. Summarily, thesofugacity condition was adopted, the friedelin vapor pressure
as calculated using an integrated form of the Clausius–Clapeyronquation proposed by Sepassi et al. [29], and the fugacity coef-cient was computed with the Peng–Robinson equation of state30]; the latent and critical properties needed were estimated byhe group contribution method proposed by Marrero et al. [31] andy expressions given by Sepassi et al. [29]; the Pitzer acentric fac-or of friedelin was calculated by Lee-Kesler correlation using Aspen
roperties version 7.3. Table 3 lists the estimated values of severalroperties of friedelin, namely: molar volume (Vi), melting tem-erature (Tm), boiling temperature (Tb), critical temperature (Tc),ritical pressure (Pc), critical volume (Vc), Pitzer acentric factor (w),
�Cpm(J mol K ) 67.8�Cpb(J mol−1 K−1) −104.0
entropy of melting (�Sm), entropy of boiling (�Sb), heat capacitychange on melting (�Cpm), and heat capacity change on boiling(�Cpb).
In light of the constant pressure (300 bar) at which the SFE stud-ies were performed, Table 3 presents the influence of temperatureon �CO2 , �CO2 , Psat
i, y∗
i, kf,i and D12. The equations used for the cal-
culations of these properties may be consulted in Appendix A. Foran eased assessment of the trends, four of the previous proper-ties are presented in dimensionless form, taking the 40 ◦C resultsas reference: Psat
i (T)/Psati (40 ◦C), y∗
i (T)/y∗i (40 ◦C), kf,i(T)/kf,i (40 ◦C),
and D12(T)/D12 (40 ◦C)As far as SC-CO2 properties are concerned, one should empha-
size the very high density of the fluid at 40 ◦C, i.e. 910.6 kg m−3
which is a quite liquid-like value. In every 10 ◦C increase of T, thedensity decreases at the constant pace of ca. 40 kg m−3 until reach-ing a final value of 746.2 kg m−3 when 80 ◦C is attained. A decreasingprofile along T is also observed on the viscosity of SC-CO2, thoughthe dependence on temperature is naturally greater: in this case,�CO2 is 9.52×10−7 Pa s at 300 bar/40 ◦C and suffers a 32% loss whenreaches 300 bar/80 ◦C.
While �CO2 and �CO2 play an effective role on diffusivities andmass transfer coefficients (also on hydrodynamics), their observeddegree of variation with T, particularly �CO2 , can anticipate animpact on the process kinetics. Accordingly, the last two rows ofTable 3 disclose the major changes of kf,i and D12: both increase withtemperature, taking ca. 12% increases every 10 ◦C. From a transportpoint of view, the operation at higher temperatures is recommend-able. However, when one observes how equilibrium evolves withtemperature, extremely pronounced increases of friedelin vaporpressure ratio are attained between 40 ◦C and 80 ◦C, leading to ajump from 1.0 to 157.4. Such massive increment of Psat
ireflects con-
sequently on the solubility of friedelin in SC-CO2: a 324% increaseof y∗
iis predicted between 40 ◦C and 80 ◦C. These results demon-
strate that equilibrium is the major effect driving the supercriticalfluid extraction in opposition to kinetic arguments.
According to the previous insights, the SFE results can be ana-lyzed with an enlightened perspective of how the supercriticalsolvent and the major solute of the biomass (i.e. friedelin) wouldbehave under different temperature conditions at 300 bar. In anycase attention should be paid to the fact that in the SFE carried outin this work we are dealing with natural biomass which means theresults may be affected by the coextraction of several solutes and/orby their solute-matrix interactions.
3.3. Supercritical fluid extraction of Q. cerris cork
3.3.1. Total extraction yieldIn Fig. 1A the total extraction yield (�t) achieved with SC-CO2
at 40, 50 and 80 ◦C – experiments SFE4, SFE3 and SFE2 of Table 1– at constant pressure (300 bar), CO2 flow rate (11.0 g min−1) andextraction time (12 h) are plotted. At 80 ◦C there is an additional
A. S en et al. / J. of Supercritical Fluids 97 (2015) 247–255 251
F tial ar
mi
fI5
TI
ig. 1. Total extraction yield results (�t) for: (A) individual SFE runs, and (B) sequenate (11.0 gCO2
min−1). �t is expressed in kg per 100 kg of cork.
easurement (SFE1.1) for half the extraction time of the others,.e. 6 h.
According to Fig. 1A the best operating temperature is 50 ◦C,or which a maximum yield of 1.81 wt% was found (see run SFE3).t is worth noting that the increment of temperature from 40 to0 ◦C increased the yield from 0.97 to 1.81 wt%, while a further
able 4nfluence of temperature on kinetic and equilibrium properties of the supercritical system
T (◦C) 40 50
�CO2 (kg m−3) 910.6 871.2
�CO2 (×10−7 Pa s) 9.52 8.61
Psati
(T)/Psati
(40 ◦C) 1.00 4.07
y∗i(T)/y∗
i(40 ◦C) 1.00 1.55
kf,i(T)/kf,i (40 ◦C) 1.00 1.11
D12(T)/D12 (40 ◦C) 1.00 1.12
nd cumulative SFE runs. Experiments at constant pressure (300 bar) and mass flow
increase to 80 ◦C reduced the yield down to 1.42 wt%. The high-est uptake obtained at 50 ◦C must be interpreted under the light of
the trade-off between the loss of solvent power of SC-CO2 (throughdensity reduction) and the increase of vapor pressure of the avail-able extractives (friedelin and all the others) in Q. cerris cork, astemperature is increased. If from a transport point of view 80 ◦C is
CO2 + friedelin at 300 bar. Model equations in Appendix A.
252 A. S en et al. / J. of Supercritical Fluids 97 (2015) 247–255
F (B) se( O2
min
tytdtl
aoSoae(ies
SQt
ig. 2. Friedelin concentration in the extracts of the: (A) individual SFE runs, and
right axis). Experiments at constant pressure (300 bar) and mass flow rate (11.0 gC
he most favorable T for the SFE (see Table 3 and Section 3.2), theield results evidence that the jump to such temperature is coun-erproductive in relation to 50 ◦C. This observation means that theensity diminution above 50 ◦C assumes greater importance andhus penalizes the total extraction yield (see the calculated valuesisted in Table 4).
Despite being the best yield result, the removal of 1.7 wt%ccomplished by SC-CO2 at 80 ◦C (SFE2) is still far from the yieldbtained by Soxhlet extraction with DCM (4.02 wt%; Table 2).uch poor performance may be due, in part, to the usage of non-ptimized P–T–t conditions in the experiments. However, it shouldlso be referred that sometimes selectivity advantages of SC-CO2xtraction are achieved at the price of lower extraction yieldsexample: SFE of diterpenes from spent coffee grounds [14]) whilen other situations SFE selectivity is attained concomitantly to anxtraction yield increase (example: SFE of phytosterol from roselleeeds [32]).
With regard to literature results on the performance of SFE vs.oxhlet extraction assays, Castola et al. [5] achieved equivalent. suber extraction yields for both SFE (under optimum condi-
ions of 50 ◦C and 220 bar) and Soxhlet extraction, namely, 7 and
quential SFE runs. The total yield (�t) of each individual run is also superimposed−1).
6 wt%. On the chosen conditions, these authors refer that 6 h (i.e.76.8 kgCO2
kg−1cork) was enough for the extraction, and that further
extension of the process did not lead to a significant incrementon extraction yield. Accordingly, through the overlap of SFE2 (12 hlong assay) and SFE1.1 (6 h long assay) runs it is clear in Fig. 1A thatmost of the extraction at 80 ◦C also takes place in the first 6 h (i.e.79.2 kgCO2
kg−1cork), which agrees with the referred insight of Castola
et al. [5]. However, in our case, a non-negligible amount of extract(22%) was still obtained in the period 6–12 h.
These observations are reinforced in Fig. 1B, where the influ-ence of successive extractions (for the same Q. cerris cork) usingsupercritical solvent under different conditions is graphed: (i) incomparison to SFE2 (12 h at 80 ◦C), the results for SFE1.2 (6 h at80 ◦C & 6 h at 50 ◦C) were better due to the positive effect of work-ing 6 h at 50 ◦C: 1.42 vs. 1.61 wt%. However, when the 12 h wererun always at 50 ◦C (SFE3, Fig. 1A), the highest yield of 1.81 wt%was reached as expected. (ii) The transition from SFE1.2 to SFE1.3
comprises additional 6 h of extraction at 40 ◦C. In this case, theimprovement found was very modest (from 1.61 to 1.75 wt%) whichis connected to the discussed weaker performance at 40 ◦C in the12 h assay (SFE4).
With regard to the use of ethanol, it enhanced significantly �t:rom SFE1.3 to SFE1.4, the modification of the SC-CO2 with 5 wt%thanol during 6 h was sufficient to increment the yield from 1.75o 2.83 wt% (Fig. 1B), which is much closer to the value obtainedy Soxhlet extraction with DCM (4.02 wt%; Table 2). This effect wasue to the polarity communicated by ethanol to the non polar CO2,ith final positive influence upon the supercritical solvent capac-
ty, and disclosed evident intermolecular interactions establishedetween ethanol and the extract components [33–35].
.3.2. Friedelin concentration in the supercritical extractsAs discussed above, lower global yields in SFE assays are some-
imes accompanied by better extraction selectivities. In fact, theower SFE yields in our work led effectively to a concentrationnhancement of some individual compounds in the supercriticalxtracts. Looking specifically to SFE1.1, whose GC–MS results areresented in Table 2, significant differences were observed in theuality of extracts. By summing the results of triterpenes in eachxtracts, it is observed that the total concentration of the triter-enes family is 1.37 times higher in SFE1.1 (at 300 bar and 80 ◦C)han in DCM extract, with an important selectivity towards friedelinfrom Table 2: 35.87/26.02 = 1.38) on the side of the SFE run. Look-ng in greater detail to the triterpenes uptake, other compoundsncreased their concentrations in the SFE extract: squalene jumpedrom 0.21% to 1.40 wt%, and �-sitosterol from 1.25% to 3.38 wt%.n the other hand, though the DCM extract indicates that 414.0g kg−1
cork of betulin was available in the biomass for extraction, itas only found in trace amounts in the SFE extract.
When working with Q. suber cork, Castola et al. [5] did not findignificant differences in friedelin concentration between SFE andCM extracts. In their case, they ranged nearly around 20.4 wt%
SFE) and 21.0 wt% (Soxhlet extraction), i.e. the concentration ratioas ca. 0.97. This fact presents an advantage of the SFE of Q. cer-
is that may be exploited with interest for commercial applicationfter careful optimization and economic evaluation of the pro-ess. As far as we know no sound commercial application has yeteen established for friedelin, despite the evidences of its effectiveerformance as anti-tumor [36], anti-inflammatory [37], analgesic37], and antipyretic [37] agent. Furthermore, a method compris-ng conventional solid–liquid extraction followed by purification ofriedelin has been patented for extracts from cork and cork-derived
aterials by Corticeira Amorim (a Portuguese cork industry com-any) [38], which is a signal of the commercial interest around thisompound.
In Fig. 2A the friedelin concentrations are plotted for SFE1.1,FE2, SFE3 and SFE4 runs, along with the respective total yields (�t).or the series 40–50–80 ◦C at 12 h, the determined friedelin con-entrations were 30.4, 40.3 and 40.6 wt%, and the total yields were.97, 1.81 and 1.42 wt%. These results showed that the friedelinoncentration in the supercritical extracts increased with increas-ng temperature, mainly near 40 ◦C (see bars), while �t exhibited alear parabolic profile (see diamonds). These experimental obser-ations are in accordance with the qualitative trend found for theolubility of pure friedelin in SC-CO2 (see Table 4). In the whole theest operating temperature was 50 ◦C, since the combined puritynd yield favors friedelin productivity (�friedelin = 0.73 wt%).
Taking into account the two assays at 80 ◦C for 6 h (SFE1.1) and2 h (SFE2) it was possible to compute �t, �friedelin, and friedelin con-entration for an intermediate step of 6 h of extraction after SFE1.1,.e. from t = 6 h to t = 12 h. These results were plotted in Fig. 2Ander the label of “SFE2”, being �t = 0.31 wt%, �friedelin = 0.18 wt%nd friedelin concentration of 57.5 wt%. This computation revealed
he remarkable concentration (57.5 wt%) of the extract obtained inhe 6–12 h period of SFE2, though the respective friedelin yield isust 0.18 wt% under that circumstance, against a 2.2 times greaterield in the first half of the process.
luids 97 (2015) 247–255 253
With respect to the addition of ethanol, Fig. 2B evidences thatthis cosolvent penalized significantly the friedelin concentration(small bar for SFE1.4) notwithstanding the total yield increasedconsiderably in this last step. The great increment of the co-extraction of remaining components (i.e., non friedelin molecules)enhanced total yield but dropped friedelin purity. For instance, thiswas also observed in the case of SFE of yellow mustard [39].
4. Conclusion
The present study comprises the first report of supercritical CO2extractions of Q. cerris cork in terms of yields and composition ofextracts, allowing thus a comparison with conventional Soxhletextraction with dichloromethane (DCM).
Novel compounds were identified by GC–MS for the first time inQ. cerris cork extracts, namely the triterpenes betulin and squalene,and �-sitosterol. Furthermore, friedelin was found in significantamounts in both SFE (30.4–40.6 wt%) and DCM (26.0 wt%) extracts.
The SFE experiments were carried out at 300 bar and between 40and 80 ◦C. In general, the total extraction yields were lower in SFE(0.97–1.81 wt%) than in DCM (4.02 wt%), but as mentioned abovethe SC-CO2 ensured a much selective removal of friedelin, with con-centrations up to 60% higher. These results are also very positivein relation to Q. suber cork for which the extracts concentrationsreported in the literature attained only ca. 20 wt%.
In conclusion, the SFE of Q. cerris cork can be a prospective path-way for a green and enhanced friedelin production.
Acknowledgments
Authors thank funding from the European Union SeventhFramework Programme (FP7/2007–2013) under grant agree-ment no. CP-IP 228589 AFORE, and Fundac ão para a Ciênciae a Tecnologia (Portugal) for funding Associate LaboratoryCICECO (Pest-C/CTM/LA0011/2013), CEF Research Center (Pest-OE/AGR/UI0239/2014) and the post doc scholarship of Ali S en.
Appendix A.
The solubility of friedelin in SC-CO2 was computed based on theisofugacity condition, as follows [28,30]:
y∗i = Psat
i
�SCFi P
exp
[V sat
i (P − Psati )
�T
](1)
where its fugacity coefficient was calculated by Peng Robinson EoS[28,30]:
ln �SCFi = − ln(Z − B) − A
B√
8ln
[Z + (1 +
√2)B
Z + (1 −√
2)B
]+ Z − 1 (2)
The friedelin vapor pressure was calculated from an integratedform of the Clausius–Clapeyron equation proposed by Sepassi et al.[29]:
log Psati = −�Sm(Tm − T)
2.3�T+ �Cpm
2.3�[
Tm − T
T− ln
(Tm
T
)]− �Sb(Tb − T)
2.3�T+ �Cp,b
2.3�[
Tb − T
T− ln
(Tb
T
)](3)
As Psati demands the knowledge of several properties (Tm, Tb, Tc,
Pc, Vc, �Sm, �Sb, �Cpm, �Cpb), they were estimated from group
contribution methods proposed by Marrero et al. [31] and Sepassiet al. [29] as follows:
Tm = Tm0 exp(∑
iNiTm1i
+∑
jMjTm2j
+∑
kOkTm3k
)(4)
2 itical F
T
T
P
V
�
�
�
�
wmiPcp
a[
D
wcR
oir
R
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
54 A. S en et al. / J. of Supercr
b = Tb0exp
(∑iNiTb1i
+∑
jMjTb2j
+∑
kOkTb3k
)(5)
c = Tc0 exp(∑
iNiTc1i
+∑
jMjTc2j
+∑
kOkTc3k
)(6)
c = Pc4 +(
Pc5 +∑
iNiPc1i
+∑
jMjPc2j
+∑
kOkPc3k
)−0.5(7)
c = Vc0 +∑
iNiVc1i
+∑
jMjVc2j
+∑
kOkVc3k
(8)
Sm = 56.5 − 19.2 log(�) + 9.2 (9)
Sb = 86 + 0.4 + 1421HBN (10)
Cpm = �Sm (11)
Cpb = −56 − 4 − 40HBP (12)
here is the molecular flexibility number, � is the molecular sym-etry parameter, HBN is the hydrogen bond density number, HBP
s the hydrogen bonding parameter, and N, M, O, Pc1, Pc2, Pc3, Pc4,c5, Vc0, Vc1, Vc2, Vc3, Tc0, Tc1, Tc2, Tc3, Tb0, Tb1, Tb2, Tb3 are groupontribution parameters of the method (reported in the originalapers).
The diffusion coefficient of friedelin in SC-CO2 was estimatedccording to the modified Scheibel equation proposed by Vaz et al.24]:
12 = 1.98 × 10−7
(T
�CO2
)0.86041
V0.2561bp,friedelin
×[
1 +(
3Vbp,CO2
Vbp,friedelin
)2/3]
(13)
here, Vbp,i is the molar volume at normal boiling point ofarbon dioxide and friedelin, both computed according to theef. [23].
The convective mass transfer coefficient ratios were calculatedn the basis of the correlation of Puiggené et al. [25,28]. Accord-ngly, considering two distinct operating conditions 1 and 2, thisatio is:
(kf,i)1(kf,i)2
=(
T1
T2
)2/3(
�CO2,1
�CO2,2
)7/15(�CO2,1
�CO2,2
)−17/15
(14)
eferences
[1] H. Pereira, Cork: Biology Production and Uses, Elsevier Science B.V., Amsterdam,2007.
[2] A. S en, I. Miranda, S. Santos, J. Graca, H. Pereira, The chemical composition ofcork and phloem in the rhytidome of Quercus cerris bark, Industrial Crops andProducts 31 (2010) 417–422.
[3] J. Grac a, H. Pereira, The periderm development in Quercus suber, IAWA J. 25(2004) 325–335.
[4] A. S en, T. Quilhó, H. Pereira, Bark anatomy of Quercus cerris L. var. cerris fromTurkey, Turkish J. Botany 35 (2011) 45–55.
[5] V. Castola, B. Marongiu, A. Bighelli, C. Floris, A. Lai, J. Casanova, Extrac-tives of cork (Quercus suber L.): chemical composition of dichloromethaneand supercritical CO2 extracts, Industrial Crops and Products 21 (2005)65–69.
[6] A.F. Sousa, P.C.R.O. Pinto, A.J.D. Silvestre, C.P. Neto, Triterpenic and otherlipophilic components from industrial cork byproducts, J. Agricultural and FoodChemistry 54 (2006) 6888–6893.
[7] H. Sovová, R.P. Stateva, Supercritical fluid extraction from vegetable materials,Reviews in Chemical Engineering 27 (2011) 79–156.
[8] M.M.R. de Melo, A.J.D. Silvestre, C.M. Silva, Supercritical fluid extraction of veg-etable matrices: applications, trends and future perspectives of a convincinggreen technology, J. Supercritical Fluids 92 (2014) 115–176.
[9] M.M.R. de Melo, R.M.A. Domingues, A.J.D. Silvestre, C.M. Silva, Extraction andpurification of triterpenoids using supercritical fluids: from lab to exploitation,Mini-Reviews in Organic Chemistry 11 (3) (2014) 362–381.
10] E. Reverchon, I. De Marco, Supercritical fluid extraction and fractionation ofnatural matter, J. Supercritical Fluids 38 (2006) 146–166.
[
luids 97 (2015) 247–255
11] D.J.S. Patinha, R.M.A. Domingues, J.J. Villaverde, A.M.S. Silva, C.M. Silva, C.S.R.Freire, C.P. Neto, A.J.D. Silvestre, Lipophilic extractives from the bark ofEucalyptus grandis x globulus, a rich source of methyl morolate: selectiveextraction with supercritical CO2, Industrial Crops and Products 43 (2013)340–348.
12] R.M.A. Domingues, M.M.R. de Melo, E.L.G. Oliveira, C.P. Neto, A.J.D. Silvestre,C.M. Silva, Optimization of the supercritical fluid extraction of triterpenic acidsfrom Eucalyptus globulus bark using experimental design, J. Supercritical Fluids74 (2013) 105–114.
14] H.M.A. Barbosa, M.M.R. de Melo, M.A. Coimbra, C.P. Passos, C.M. Silva, Opti-mization of the supercritical fluid coextraction of oil and diterpenes from spentcoffee grounds using experimental design and response surface methodology,J. Supercritical Fluids 85 (2014) 165–172.
15] L. Fiori, D. de Faveri, A.A. Casazza, P. Perego, Grape by-products: extraction ofpolyphenolic compounds using supercritical CO2 and liquid organic solvent –a preliminary investigation, Cyta-J. Food 7 (2009) 163–171.
16] S. Akay, I. Alpak, O. Yesil-Celiktas, Effects of process parameters onsupercritical CO2 extraction of total phenols from strawberry (Arbutusunedo L.) fruits: an optimization study, J. Separation Science 34 (2011)1925–1931.
17] A. Romo-Hualde, A.I. Yetano-Cunchillos, C. Gonzalez-Ferrero, M.J. Saiz-Abajo,C.J. Gonzalez-Navarro, Supercritical fluid extraction and microencapsulation ofbioactive compounds from red pepper (Capsicum annum L.) by-products, FoodChemistry 133 (2012) 1045–1049.
18] V. Castola, A. Bighelli, S. Rezzi, G. Melloni, S. Gladiali, J.-M. Desjobert, J.Casanova, Composition and chemical variability of the triterpene fraction ofdichloromethane extracts of cork (Quercus suber L.), Industrial Crops and Prod-ucts 15 (2002) 15–22.
19] M.M.R. de Melo, R.M.A. Domingues, M. Sova, E. Lack, H. Seidlitz, F. Lang Jr., A.J.D.Silvestre, C.M. Silva, Scale-up studies of the supercritical fluid extraction oftriterpenic acids from Eucalyptus globulus bark, J. Supercritical Fluids 95 (2014)44–50.
20] A. S en, I. Miranda, H. Pereira, Temperature-induced structural and chemicalchanges in cork from Quercus cerris, Industrial Crops and Products 37 (2012)508–513.
21] H. Pereira, Variability of the chemical composition of cork, BioResources 8(2013) 2246–2256.
22] K.S. Pitzer, D.R. Schreiber, Improving equation-of-state accuracy in the criticalregion: equations for carbon dioxide and neopentane as examples, Fluid PhaseEquilibria 41 (1988) 1–17.
23] V.V. Altunin, M. Sakhabetdinov, Viscosity of liquid and gaseous carbon dioxideat temperatures of 220–1300 K and pressures up to 1200 bar, Thermal Engi-neering 19 (1972) 124–129.
24] R.V. Vaz, A.L. Magalhães, C.M. Silva, Improved Stokes–Einstein based modelsfor diffusivities in supercritical CO2, J. Taiwan Institute of Chemical Engineers45 (2014) 1280–1284.
25] J. Puiggené, M.A. Larrayoz, F. Recasens, Free liquid-to-supercritical fluidmass transfer in packed beds, Chemical Engineering Science 52 (1997)195–212.
26] E.L.G. Oliveira, A.J.D. Silvestre, C.M. Silva, Review of kinetic models for super-critical fluid extraction, Chemical Engineering Research and Design 89 (2011)1104–1117.
27] A.F. Silva, M.M.R. de Melo, C.M. Silva, Supercritical solvent selection (CO2 versusethane) and optimization of operating conditions of the extraction of lycopenefrom tomato residues: innovative analysis of extraction curves by a ResponseSurface Methodology and Cost of Manufacturing hybrid approach, J. Supercrit-ical Fluids 95 (2014) 618–627.
28] M.M.R. de Melo, E.L.G. Oliveira, A.J.D. Silvestre, C.M. Silva, Supercritical fluidextraction of triterpenic acids from Eucalyptus globulus bark, J. SupercriticalFluids 70 (2012) 137–145.
29] K. Sepassi, P.B. Myrdal, S.H. Yalkowsky, Estimating pure-component vaporpressures of complex organic molecules: Part II, Industrial and EngineeringChemistry Research 45 (2006) 8744–8747.
31] J. Marrero, R. Gani, Group-contribution based estimation of pure componentproperties, Fluid Phase Equilibria 183 (2001) 183–208.
32] K.L. Nyam, C.P. Tan, O.M. Lai, K. Long, Y.B. Che Man, Optimization ofsupercritical fluid extraction of phytosterol from roselle seeds with a cen-tral composite design model, Food and Bioproducts Processing 88 (2010)239–246.
33] Ö. Güc lü-Üstündag, F. Temelli, Solubility behavior of ternary systems of lipids,cosolvents and supercritical carbon dioxide and processing aspects, J. Super-critical Fluids 36 (2005) 1–15.
34] R.M.A. Domingues, E.L.G. Oliveira, C.S.R. Freire, R.M. Couto, P.C. Simões, C.P.Neto, A.J.D. Silvestre, C.M. Silva, Supercritical fluid extraction of Eucalyptus glo-bulus bark—a promising approach for triterpenoid production, International J.
Molecular Sciences 13 (2012) 7648–7662.
35] E.C. de Vasconcelos, J.H.Y. Vilegas, F.M. Lanc as, Comparison of extraction andclean-up methods for the analysis of friedelan-3-ol and friedelin from leaves ofMaytenus aquifolium Martius (Celastraceae), Phytochemical Analysis 11 (2000)247–250.
36] B. Lu, L. Liu, X. Zhen, X. Wu, Y. Zhang, Anti-tumor activity of triterpenoid-rich
extract from bamboo shavings (Caulis bamfusae in Taeniam), African J. Biotech-nology 9 (2010) 6430–6436.
37] P. Antonisamy, V. Duraipandiyan, S. Ignacimuthu, Anti-inflammatory, anal-gesic and antipyretic effects of friedelin isolated from Azima tetracantha Lam.in mouse and rat models, J. Pharmacy and Pharmacology 63 (2011) 1070–1077.
[
luids 97 (2015) 247–255 255
cation of friedelin, EP2070906A1, 2009.39] V. Barthet, J. Daun, An evaluation of supercritical fluid extraction as an analytical
tool to determine fat in canola, flax, solin, and mustard, J. American Oil ChemistsSociety 79 (2002) 245–251.
