Www.sciencedirect.com Science MiamiImageURL Cid=271146& User=5254909& Pii=S0260877411005516& Check=y& Origin=& Cove

8/3/2019 Www.sciencedirect.com Science MiamiImageURL Cid=271146& User=5254909& Pii=S0260877411005516& Check…

http://slidepdf.com/reader/full/wwwsciencedirectcom-science-miamiimageurl-cid271146-user5254909-piis0260877411005516 1/8

Effect of boldo (Peumus boldus M.) pretreatment on kinetics of supercriticalCO2 extraction of essential oil

Edgar Uquiche a,b, Elizabeth Huerta a, Alicia Sandoval a, José Manuel del Valle c,d,⇑

a Department of Chemical Engineering, Universidad de La Frontera (UFRO), Temuco, Chileb Center of Food Biotechnology and Bioseparations, Scientific and Technological Bioresource Nucleus (BIOREN), UFRO, Temuco, Chilec Department of Chemical and Bioprocess Engineering, Pontificia Universidad Católica (UC) de Chile, Santiago, Chiled ASIS-UC Interdisciplinary Research Program on Tasty and Healthy Foods, UC, Santiago, Chile

a r t i c l e i n f o

Article history:

Received 2 May 2011

Received in revised form 3 October 2011

Accepted 8 October 2011

Available online xxxx

Keywords:

Boldo

Essential oils

Mass-transfer

Microstructure

Milling

ModelingRapid decompression

Supercritical CO2 extraction

a b s t r a c t

This work examined the effect of the solid matrix on supercritical carbon dioxide (SC CO2) extraction of

essentials oils from boldo leaves (Peumus boldus M.) subjected to rapid decompression of a CO2-impreg-

nated sample, conventional milling, and low-temperature milling. Low-temperature conditioning prior to

milling decreased heat-driven losses of volatile compounds during milling, as attestedby a higher extract

yield for low-temperature (10.8 g extract/kg extract-free substrate) than conventionally (9.63 g/kg)

milled sample. Extract yield was even larger for the rapidly decompressed sample (11.4 g/kg). Results

of SC CO2 extraction experiments carried out at 40 °C and 10 MPa were adjusted to a diffusional model

using the effective diffusivity of the extract in the solid matrix ( De) and a single partition of essential oils

between solid substrate and SC CO2 as best-fitting parameters. A microstructural factor (F M), which was

estimated as the ratio between the binary diffusion coefficient of the essential oil in SC-CO2 under extrac-

tion conditions and De, was used to characterize the effect of sample pretreatment on extraction rates.

Values of F M for rapidly decompressed (202) and low-temperature milled (1740) samples were smaller

than value for conventionally milled samples (2200), which revealed that the two first treatments dis-

rupted secretory cavities in boldo leaves more effectively than the third. This was confirmed by light

microscopy observations. The work included also measurements using oregano bracts (Origanum vulgare

L.) as the substrateto confirmliterature reports on the SC CO2 extraction of pretreated bracts and to serveas a reference to our main results with boldo leaves. Trends with oregano coincided with those of boldo.

Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Boldo (Peumus boldus M.) is a native plant from Chile whoseleaves contain bioactive essential oils, alkaloids, and flavonoidshaving interesting medicinal properties (Bisset, 1994; Speiskyand Cassels, 1994). Essential oils are complex mixtures of terpenes,a widespread and chemically diverse group of natural compoundsthat derive from isoprene, a five-carbon, unsaturated hydrocarbonmolecule (Brielmann et al., 2006). They generally include the 10-

carbon (or 2-isoprene) monoterpenes and the 15-carbon (or 3-iso-prene) sesquiterpenes, as well as oxygen-containing derivatives(e.g., oxygenated monoterpenes) that are responsible for character-istic plant aromas (del Valle et al., 2011).

Supercritical (SC) fluid extraction has emerged as a viable alter-native of conventional extraction techniques for plant essentialoils, such as hydrodistillation, steam distillation, and solvent

extraction, because it circumvents the use of organic solvents,and/or high temperatures (Quirin and Gerard, 2007; Reverchonand De Marco, 2007). Carbon dioxide (CO2) is the fluid most widelyused in SC fluid extraction applications because of it is non-corro-sive, non-flammable, non-reactive, and innocuous; it is availablehighly purified at low cost; it is effective at room temperature;and it can be removed completely from treated substrates and ex-tracts (del Valle and Aguilera, 1999).

del Valle et al. (2004, 2005) attempted the SC CO2 extraction of

natural antioxidants in boldo barks and leaves. Main boldo antiox-idants are boldine in the alkaloid fraction, and catechin in the fla-vonoid fraction (Schmeda-Hirschmann et al., 2003; Quezada et al.,2004). On the other hand, boldo essential oils and other fatty com-pounds have pro-oxidant activity (Quezada et al., 2004). Obtainingantioxidant extracts from boldo leaves using SC CO2 as the solventrequires >70 MPa and/or polar ethanol as co-solvent (del Valleet al., 2004, 2005) because of the low solubility in SC CO2 of boldineand especially catechin (Berna et al., 2001; de la Fuente et al.,2005). An alternative to isolate boldo antioxidants is to removearoma compounds using moderate-pressure CO2 (del Valle et al.,2005; Mazutti et al., 2008), and then extract the antioxidants using

0260-8774/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.

doi:10.1016/j.jfoodeng.2011.10.013

⇑ Corresponding author at: Department of Chemical and Bioprocess Engineering,

Pontificia Universidad Católica (UC) de Chile, Avenida Vicuña Mackenna 4960,

Macul, Chile. Tel.: +56 2 3544418; fax: +56 2 354580.

E-mail address: [email protected] (J.M. del Valle).

Journal of Food Engineering xxx (2011) xxx–xxx

Contents lists available at SciVerse ScienceDirect

Journal of Food Engineering

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j f o o d e n g

Please cite this article in press as: Uquiche, E., et al. Effect of boldo (Peumus boldus M.) pretreatment on kinetics of supercritical CO2 extraction of essentialoil. Journal of Food Engineering (2011), doi:10.1016/j.jfoodeng.2011.10.013



ethanol, water, or their mixtures (Pasquel et al., 2000; Yoda et al.,2003). Boldo essential oil, however, may have commercial value onits own because it has the characteristic aroma of boldo leaves andcontains some unusual compounds such as ascaridole (Bruns andKöhler, 1974; Bisset, 1994; Miraldi et al., 1996; Vila et al., 1999;

Mazutti et al., 2008), a bicyclic monoterpene with a peroxidebridge having sedative, pain-relieving, antifungal, and antitumoreffects (Pare et al., 1993; Efferth et al., 2002; Dembisky, 2008).

A factorthat affects the SCCO2 extraction rate andyieldof essen-tial oilsis thepretreatment of thesubstrate.In most studies, thesub-strate is milled, which is eventually complimented by separationaccording to particle size by sieving. These pretreatments are inef-fective when they do not selectively disrupt essential-oil-holdingstructures (Gaspar et al., 2000, 2001), or cause essential oil lossesby evaporation and/or thermal/oxidative degradation (Pruthi,

1980; Pesek et al., 1985; Murthy and Bhattacharya, 2008). Gasparet al. (2000) proposed cryogenic milling to reduce thermal degrada-tion and evaporation of essential oils during sample pretreatment.

Effective treatments prior to essential oil extraction should beaimed at a selective destruction of cellular barriers holding essen-tial oils in the herb but not at the expense of causing losses of vol-atile compounds. Many essential oils are produced in glandularcells on the surface of leaves, bracts, petals, and other organs of aromatic herbs, and are stored in the subcuticular space formedat the gland anex (Bosabalidis and Tsekos, 1982, 1984; Gasparet al., 2001; Zizovic et al. 2005). The secretory structures thatencapsulate plant essential oils in vegetable tissue depend on theplant family and species, with other specialized encapsulatingstructures including ducts and cavities (Zizovic et al., 2007; Stame-nic et al., 2008). Gaspar et al. (2000, 2001) proposed disrupting

glandular trichomes of oregano (Origanum virens L.) bracts by con-tact with high-pressure CO2 followed by a rapid decompression.

This work evaluatedthe effect of the pretreatment(conventionalmilling, low-temperature milling, and fast decompression of CO2-impregnated sample) on theSC CO2 extraction of essential oils fromboldo (P. boldus M.) leaves. Because the effect of sample pretreat-menton extraction rateand yield canbe quantifiedby mathematicalmodeling using the effective diffusivity of a relevant pseudo-solutein the solid matrix (De) as a fitting parameter (Araus et al., 2009),authors evaluated the effectiveness of the pretreatments by lightmicroscopy and estimationof De values from cumulative extractionplots of essential oil yield versus time. To validate the methods pro-posed by Gaspar et al. (2000) additional experiments were carriedout using oregano (Origanum vulgare L.) bracts as the substrate.

2. Materials and methods

2.1. Substrates and pretreatments

Dry boldo (P. boldus M.) leaves and oregano (O. vulgare L.) bractswere locally acquired in a market in Temuco (Chile). They werehand-picked to eliminate damaged parts, and subjected to conven-tional milling, low-temperature milling, and rapid decompression.Samples were placed in a desiccator with silica gel prior to treat-ments to reduce and homogenize their moisture content. For con-ventional milling, substrates were treated in 100-g batches in ahammer mill (M2 Assa, Assa Tecnología S.A., Santiago, Chile) with1-mm openings for 15 min (boldo leaves) or 10 min (oreganobracts).

Low-temperaturemillingwascarriedout in a similar fashion, but

in this case samples were pre-cooled prior to milling. Forthat, 200-gbatchesof eachsubstrate were placed in a metallicbasketthat wasinturn placed inside a 10-dm3 vessel containing 1 dm3 of liquid nitro-gen for 3 min (thetime thattook evaporating all the liquidN2). Con-ventionally milled and low-temperature milled samples werepacked in polyethylene bags, and stored at 5 °C up to analysis.

For the rapid decompression of CO2-impregnated matrices, ca.12 g of boldo leaves (weighed accurately) manually cracked to passthrough a mesh Tyler number 7 screen (openings of 2.36 mm), orca. 5 g of oregano bracts (weighed accurately) of approximatelythe same size, were loaded in a 50-cm3 extraction vessel (innerdiameter = 1.4 cm; height = 32.5 cm). The extraction vessel wasplaced in the air-convection oven of a Spe-ed SFE unit (AppliedSeparations, Allentown, PA), pressurized with high-purity (99.95%pure) CO2 (AGA S.A., Santiago, Chile), and held at 40 °C and

7 MPa for 60 min (Gaspar et al., 2003). After substrates were fullyswollen with SC CO2, glands or cavities were disrupted by a fastdecompression step; CO2 was released by opening a valve whichlet the pressure diminish to the normal atmospheric level in120 s (boldo leaves) or 80 s (oregano bracts).

2.2. Substrate characterization

Density, particle size distribution, and microstructural charac-teristics of pretreated substrates were evaluated. True density(qs) was estimated by weighing the material loaded into a 1-dm3

graduated cylinder using a standard tapping procedure (del Valleet al., 2003). Apparent density (qb) was estimated by weighing a fi-nely milled sample of the material that could be tightly packed into

Nomenclature

C fo solute concentration in the SC CO2 phase at the begin-ning of the experiment (g/kg CO2)

C so solute concentration in the solid phase at the beginningof the experiment (g/kg substrate)

dpi size of fraction i of particles (mm)

(dp)ref particle size parameter of the Rosin–Rommler–Bennettsize distribution (mm)dpS mean Sauter diameter of solid substrate particles (mm)De effective diffusivity of solute in the solid matrix (m2/s)DL axial dispersion coefficient (m2/s)D12 binary diffusion coefficient of the solute in SC CO2 (m2/

s)F M microstructural factor (À)i ( j) index characterizing particles of a given size (À)kf film mass transfer coefficient (m/s)K solute partition coefficient between the solid matrix and

SC CO2 phase (À)l half-thickness of the slab-shaped solid particle (mm)

mi mass of fraction of particles of size i

MW 2 molecular weight of the extract (g/mol)n uniformity coefficient of the Rosin–Rommler–Bennett

particle size distribution (À)T process temperature (K)

T c critical temperature of SC CO2 (304.3 K)T r reduced temperature of SC CO2 (À)V c2 critical volume of the extract (cm3/mol)

Greek letterse void fraction in the bed (À)l viscosity of pure CO2 at high pressure (kg mÀ1 sÀ1)q density of pure CO2 at high pressure CO2 (kg/m3)qb apparent (bulk) density of the substrate (kg/m3)qc critical density of CO2 (468 kg/m3)qr reduced density of CO2 (À)qs true density of the substrate (kg/m3)

2 E. Uquiche et al./ Journal of Food Engineering xxx (2011) xxx–xxx

Please cite this article in press as: Uquiche, E., et al. Effect of boldo ( Peumus boldus M.) pretreatment on kinetics of supercritical CO2 extraction of essentialoil. Journal of Food Engineering (2011), doi:10.1016/j.jfoodeng.2011.10.013



a 10-cm3 graduated cylinder (del Valle et al., 2003). The particlesize distributions of the milled substrates were determined in aRo-Tap test sieve shaker (W.S. Tyler, Mentor, OH) using 80, 48,35, 28, 24, and 20 mesh Tyler screens. These measurements wereperformed in triplicate.

Light microscopy of pretreated boldo leaves and oregano bracts

was carried out using the procedures of Uquiche et al. (2004). Fol-lowing sample fixing, samples were embedded with paraffin prior

to cutting thin slices (30-lm thick) using a manual microtome Jung(Heidelberg, Germany). Following removal of paraffin with xylol,samples were stained with safranine and fast green. Photomicro-graphs of the stained samples were captured using a light micro-scope Olympus BX50 (Tokyo, Japan) equipped with a camera JVCTK-1280E (Yokohama, Japan).

2.3. Supercritical carbon dioxide extraction experiments

Pretreated samples were extracted in the Spe-ed SFE unitusing 7.2 g/min SC-CO2 at 40 °C and 10 MPa (Reverchon et al.,

1993). The temperature of the expansion valve was kept con-stant at 100°C, and the total extraction time was 120 min inall experiments. In the case of conventionally or low-tempera-

ture milled boldo leaves and oregano bracts, ca. 15 g samples(weighed accurately) were loaded in the extraction vessel,whereas those samples subjected to rapid decompression wereextracted under the aforementioned conditions following 10-min standing in the extraction vessel at normal atmosphericpressure after pretreatment.

Ten extract aliquots were periodically collected during extrac-tion in pre-weighed glass vials (15-cm3 capacity) and dried in adesiccator with silica gel to remove co-extracted water, and the

recovered extract was assessed gravimetrically by difference withcleaned and dried vials. The weight loss (initial minus final) of the substrate was measured and compared to the total recoveryin the ten glass vials to detect random experimental errors, andthen the extraction yield (or mass of recovered solute per unitmass of water-free substrate) was computed as a function of

extraction time. Extraction experiments were carried out induplicate.

Pretreated and SC-CO2-extracted samples of boldo leaves andoregano bracts were finely milled with a mortar and pestle beforequantification of moisture content. Sample moisture was deter-mined gravimetrically by drying in an air-convection oven

(Memmert model UM-400, WTB Binder, Tutlingen, Germany) setat 105 °C to a constant final weight (15 h).

3. Results and discussion

This section describes the effect of sample pretreatment on thephysical properties of boldo leaves and oregano bracts, includingdensities, particle sizes and size distributions, and microstructural

features (Section 3.1), and then focus attention to the effect of those pretreatments on SC CO2 extraction rate and yield (Sec-tion 3.2), and the parameters of a mathematical model of theextraction process (Section 3.3).

3.1. Characterization of pretreated herbs

Fig. 1 presents Rosin–Rommler–Bennett (RRB) function plots of cumulative undersize distribution, F (dp), for conventional and low-temperature milled samples of boldo and oregano. The RRB func-tion plot assumes the following relationship between F and dp

(Manohar and Sridhar, 2001; Nagy et al., 2008):

F ðdpÞ ¼ 1À exp Àdp

ðdpÞref

n

; ð1Þ

where (dp)ref is a particle size parameter (the 36.8th percentile of the cumulative size distribution), and n, a uniformity coefficient re-lated to the steepness of the size distribution (as the value of n in-creases the scattering of the distribution decreases). Linearization of Eq. (1) suggest presenting particle size distribution data in a log–logplot of ln [1À F (dp)]À1 versus dp, as done in Fig. 1. The RRB functionfitted the size distribution of the milled samples more closely thanalternative models such as the log–normal or log–log (or Gaudin–Schumann) plots, as also observed by Manohar and Sridhar (2001)for conventionally- and cryogenically-milled turmeric samples.Fig. 1 shows differences in particle size distribution between the

substrates, and also between milling treatments for oregano bracts,but not for boldo leaves.

The particle size distribution can be also reduced to a singlemean value such as the Sauter diameter (dpS) that represents thediameter of a sphere having the same surface-to-volume ratio asthe entire population of particles in a milled sample, and can beestimated using Eq. (2):

dpS ¼

Pi

mi þP j

m j

Pi

mi

31t þ 2

dpi

þ

P j

m j

dp j

; ð2Þ

where i represents a fraction of particles (mass mi) having a thick-ness (t ) smaller than the average sieve-opening (dpi), which weretreated as disks of diameter dpi and height (2l), and j, a fraction of

particles (mass m j) whose thickness was above the average sieve-opening and which were treated as spheres of diameter dp j. Themean Sauter diameter of rapidly-decompressed samples was esti-mated by assuming they were shaped as 2.36-mm-diameter disksof height (2l), where thickness (2l) was estimated as the averagecross section height of seven light-microscopy images of untreatedboldo leaf and oregano bract samples.

Table 1 summarizes physical properties of the milled samples,including the RRB model parameters derived from Fig. 1, meanSauter diameters, and total porosities (e) estimated using Eq. (3)(Nagy et al., 2008):

e ¼ 1Àqb

qs

: ð3Þ

10-1

100

101

Low-temperature milled oregano

Conventionally milled oregano

Low-temperature milled boldo

Conventionally milled boldo

0.2 0.4 0.5 0.6 0.7 0.8

Particle diameter ( , mm)d p

)sselnoisnemi

d(

)

-1(nl

F

- 1

Fig. 1. Rosin–Rommler–Bennett plot of cumulative undersize population versus

particle size for conventionally and low-temperature milled boldo leaf and oregano

bract particles, where F represents the weight fraction of particles of size dp or less.

E. Uquiche et al./ Journal of Food Engineering xxx (2011) xxx–xxx 3




There were only minor differences in the uniformity coefficientof the particle size distribution, which ranged from 3.5 to 3.9, indi-cating a similar data scattering for all samples, whereas the particlesize parameter was approximately the same for the boldo samples(480 lm), but varied between the conventionally (450 lm) andlow-temperature (410lm) milled samples of oregano. These dif-ferences translated into similar difference between mean Sauterdiameters that were slightly smaller than particle size parametersin all cases. In any case, unlike Manohar and Sridhar (2001), whofound that cryogenic milling produced turmeric particles that were

about 10 times smaller than those produced by conventional mill-ing (8.56 lm versus 88.2lm Ferret diameter), authors did not ob-serve large differences between conventional- and low-temperature-milled samples. This was probably due to differencesin mechanical properties of the treated samples, and in processtemperatures between the two studies. Indeed, it is reasonable toassume that turmeric stalk samples were more brittle in the cryo-genic milling experiments of Manohar and Sridhar (2001) thanoregano bracts and boldo leaves used in our low-temperature mill-ing experiments, which resulted in smaller particles in their casethan ours (8.56 lm versus 339–380 lm). On the other hand, theaverage particle diameters were much larger for the rapidly

decompressed than either of the milled samples for both oreganoand boldo (Table 1) because those samples were manually crackedprior to treatment, and the rapid decompression treatment did notresult in a further reduction in particle size but a selective destruc-

tion of essential-oil-containing structures. The larger particles of rapidly decompressed samples resulted in smaller apparent densi-ties than those of milled samples, because of the increased diffi-culty in accommodating large as compared to small particles inthe void spaces between particles in packed beds. This, in turn, re-sulted in larger total porosities for rapidly decompressed thanmilled samples especially in the case of oregano bracts (cf. Table 1).It is important stressing that values of e in Table 1 incorporate both

the inner porosity of the particles (that should change little withherb pretreatment) and the interparticle porosity of the packedbed (that should decrease considerably as a result of particle sizereduction during milling).

There is limited information on the specialized structures thatstore essential oils in boldo, a plant of the Monimiaceae family.Martínez-Laborde (1988) described the glands or hairs on the sur-

face of boldo leaves as being stellate (star-shaped), Bisset (1994)presented microphotographs of stellate trichomes on the surfaceand oil cells in the spongy parenchyma of boldo leaves, and Uauy(1998) presented microscopic evidence of secretory cavities withinboldo leaves. Secretory cavities are spherical structures whereplants store the essential oils (Zizovic et al. 2007; Stamenic et al.2008), resins, and/or other secondary metabolites used to fightagainst, e.g., microorganism and insects. These substances are syn-thesized in the epithelial cells that line the cavities, and are subse-quently secreted into them (Salisbury and Ross, 1992). Uponrupture of the cavities, resins and essential oils are secreted to offerprotection and/or provide the characteristic plant scent.

On the other hand, there is extensive information in literatureabout glandular trichomes or glands; the specialized storage struc-

tures for essential oils on the surface of oregano bracts (Bosabalidisand Tsekos, 1982, 1984; Gaspar et al., 2001) and other Lamiaceaeherbs (Zizovic et al., 2005, 2007; Krstic et al., 2006). These glandshave the same function as secretory cavities in boldo leaves, andtheir amount depends on the species and exhibits plant-to-plantvariability (Krstic et al., 2006).

Fig. 2 presents light microscopy images of conventionally milled(Fig. 2A and B), low-temperature milled (Fig. 2C and D) and rapidlydecompressed (Fig. 2E and F) boldo leaves (Fig. 2A, C, and E) andoregano bracts (Fig. 2B, D, and F). There is a stellate trichomeclearly visible on the upper surface of the conventionally milledboldo leaf in Fig. 2A. The presence of unruptured cavities/glandscould explain a limited extraction rate due to microstructural lim-itations to intraparticle essential oil transport in the conventionallymilled samples. This is very important considering that the extrac-tion rate of plant essential oils has been claimed to be controlled byinternal resistance to mass transfer in the vegetable structure (Gas-par et al., 2001; Reverchon and De Marco, 2007). Fig. 2C showsdetachment of trichome, superficial shrinkage, and a pronouncedchange in inner microstructure in the low-temperature milled bol-do sample which may have freed essential oils and favor theirintraparticle transport during SC CO2 extraction. Advantages of

cryogenic-milling for sample preparation prior to extractions havebeen reported, but there is limited microscopy evidence in litera-ture about the advantages of this size-reduction method. Fig. 2Eshows the disruption of a stellar trichome on the upper surface

of a rapidly-decompressed boldo leaf treated with CO2, and disrup-tion of the intraparticle cellular structure (cf. lower left side andcompare with Fig. 2A) which suggest specificity of this treatmentto free entrapped essential oils so as to favor their extraction withSC CO2. Microphotographs showed similar improvements in sam-ple disruption when applying low-temperature milling (Fig. 2D)or rapid-decompression (Fig. 2F) instead of conventional milling(Fig. 2B) to oregano bracts. Shrinking, detachment, and/or disap-

pearance, of the glands seen in the upper and lower face of thebracts in Fig. 2B suggest freeing of essential oil during low-temper-ature milling or rapid-decompression treatments. The effect of ra-pid decompression on superficial glands of aromatic herbs wasexplained by the slow penetration of CO2 into the glands when ex-posed to high-pressure CO2 until the oil becomes saturated in CO2

which may result in bursting of a fraction of the glands, e.g., 39.2%

of the trichomes in mentha leaves (assessed by SEM) following 1 hexposure to SC CO2 at 40 °C and 10 MPa, as compared to 34.9% tri-chome disruption by conventional milling (Zizovic et al., 2007).Stamenic et al. (2010) provided visual evidence of the swellingon a mint leaf upon exposure to SC CO2 at 40 °C and 10 MPa. Theeffectiveness of the treatment increases if the exposure to high-pressure CO2 is followed by a rapid decompression because of the pressure gradient that builds up across the walls and epidermisof glands and opposes free flow of CO2 thus resulting in the disrup-tion of the glands when the pressure drop across those barrierssurpasses their outbreak pressure (Gaspar et al., 2001). As anexample, Gaspar et al. (2003) reported a 85% efficiency in the dis-ruption of oregano bract glands following rapid (113 s) decompres-sion of samples held 1 h is SC CO2 at 40 °C and 7 MPa.

Table 1

Physical properties of pretreated boldo leaves and oregano bracts.

Sample pretreatment Moisture (g water/100 g ) 2l (mm) n (À) (dp)ref (mm) dpS (mm) qs (kg/m3) qb (kg/m3) e (À)

Conventionally-milled boldo leaves 5.07 ± 0.08 0.432 ± 0.037 3 .71 0.480 0.385 1057.3 ± 4.1 2 84.7 ± 4.2 0 .730

Low-temperature-milled boldo leaves 5.82 ± 0.18 0.432 ± 0.037 3.52 0.476 0.380 1057.3 ± 4.1 283.3 ± 7.2 0.732

Rapidly-decompressed boldo leaves 5.51 ± 0.12 0.432 ± 0.037 – – 0.949 1057.3 ± 4.1 209.7 ± 2.5 0.801

Conventionally-milled oregano bracts 5.84 ± 0.12 0.369 ± 0.061 3.58 0.445 0.354 984.6 ± 1.2 283.3 ± 4.2 0.713

Low-temperature-milled oregano bracts 5.63 ± 0.38 0.369 ± 0.061 3.88 0.408 0.339 984.6 ± 1.2 283.3 ± 5.3 0.713

Rapidly-decompressed oregano bracts 5.25 ± 0.26 0.369 ± 0.061 – – 0.844 984.6 ± 1.2 98.0 ± 3.5 0.900





3.2. Effect of sample pretreatment on SC CO 2 extraction rate and yield

Fig. 3 describes the kinetics of extraction of pretreated boldoleaves (Fig. 3A) and oregano bracts (Fig. 3B) using 7.2 g/min of SCCO2 at 40 °C and 10 MPa (U = 1.24 mm/s). Curves approach hori-

zontal asymptotes (not shown) that represent the yield of the pro-cess (C so) in grams of extract per kilogram of dry substrate. Basedon longer preliminary SC CO2 extraction experiments, the valuesof C so reported in Table 2 were adopted as 5% above those observedafter 2-h treatment that can be read from Fig. 3. Reported values of essential oil content in boldo leaves range from 4 g/kg (del Valleet al., 2005) to 31 g/kg (Quezada et al., 2004), which illustrates typ-ical variations in biologic samples due to genetic variations or dif-ferences in harvest time, drying treatment, and storage condition(Miraldi et al., 1996; Vogel et al., 1999). The essential oil contentalso depends on sample pretreatment being larger in the low-tem-perature milled (10.8 g/kg) or rapidly decompressed samples(11.4 g/kg) than in the conventionally milled samples (9.63 g/kg).This was as expected from microscopy results in Section 3.1 that

suggested a more pronounced liberation of essential oils in low-temperature milled or rapidly decompressed than conventionallymilled samples. The heat generated during grinding results inlosses of volatile compounds, which can be reduced by loweringthe initial temperature of the material prior to grinding, as done

in cryogenic grinding operations (Murthy and Bhattacharya,2008). Indeed, the use of liquid nitrogen for sample cooling allowssamples to absorb the heat generated during grinding withoutexperiencing a temperature that will result in volatile losses orthermal degradation reactions (Krejc ová et al., 2008).

Data in Fig. 3A also suggests a faster initial extraction of essen-tial oils from rapidly decompressed boldo leaves than either milledsample, and only following a short initial extraction period doesthe yield from the low-temperature milled sample approach thatof the rapidly decompressed sample. These differences in initialextraction rate may be due to the selective destruction of masstransfer barriers to essential oil extraction during the rapid decom-pression of the CO2-impregnated samples (Gaspar et al., 2000,2001, 2003). Gaspar et al. (2001) reported the effectiveness of

Fig. 2. Light micrographsof cross sectional areas of (A,C, E) boldo leaves and(B, D, F) oregano bractsexposed to (A,B) conventional milling, (C,D) low-temperature milling, or

(E, F) rapid decompression. The magnification of all images is the same (bar = 200lm).





the rapid decompression pretreatment in disrupting glandular tric-homes in oregano bracts, and our results suggest that this pretreat-ment in also effective to disrupt secretory cavities in boldo leaves.

An alternative to characterize initial extraction rates is using the

so-called apparent solubility concept (C fo), which is the initial slopeof a plot of cumulative yield (gram of extract per kilogram of drysubstrate) versus specific CO2 consumption (kilogram of CO2 usedper kilogram of dry substrate) (data not shown). Units of theseslopes (gram extract per kilogram of CO2) coincide with the onesof true solubilities, but the property is named ‘‘apparent’’ solubilitybecause it is influenced by extract availability and binding to the

solid matrix. Because thecritical pressure of binary (CO2 + essentialoil component) systems at 40 °C in slightly below 10 MPa (del Valleet al., 2011), they are miscible in SC CO2 under process conditions.

Thus, values of C fo reported in Table 2 do not correspond to truesolubilities; they represent the fraction of essential oils freed bysample pretreatment and released during the static extraction per-iod. Thus, the slightly higher value of C fo of the rapidly decom-pressed sample as compared to the milled samples in spite of a20% reduction in the load of the extraction vessel supports the

claim of an improved release of essential oil components fromsecretory cavities in boldo leaves by that treatment.

Qualitatively, the trends observed for the pretreated oreganosamples (Fig. 3B) are quite similar to those observed for the boldosamples (Fig. 3A), with the only differences being due to the loweressential oil content in oregano bracts than boldo leaves. Thus, ourresults confirm trends already reported by Gaspar et al. (2000).Conditioning to low-temperature prior to milling reduces lossesof volatile compounds as a result of heat generation during grind-ing, whereas rapid decompression of CO2-impregnated samplesselectively disrupts glandular trichomes holding essential oils inselected plants, thus facilitating SC CO2 extraction processes. Thereis another difference between oregano bracts and boldo leavesregarding apparent solubility that was considerably lower in rap-idly decompressed than milled samples (Table 2). In this case, C fo

is probably defined by essential oil availability instead of essentialoil solubility, which paralleled a decrease in the loading capacity of the extraction vessel and an increase in total porosity for oreganobracts as compared to boldo leaves. The negative impact on essen-tial oil availability of the rapid decompression pretreatment of thedecrease in oregano bract load as compared to boldo leaf load wascompounded by the lower essential oil content in the former thanthe latter.

3.3. Mathematical modelling of SC CO 2 extractions

Cumulative extraction curves in Fig. 3 were best-fitted to thediffusion model of Araus et al. (2009), which describes the concen-tration of extract in the substrate and the SC CO2 phases as a func-tion of extraction time and distances from the particle axis andentrance of the extraction vessel. This model assumes non-porous,homogeneous, slab-shaped particles of thickness (2l) containing asingle pseudo-solute; a constant partition coefficient (K ) of thepseudo-solute between the solid matrix and SC CO2; and constantphysical properties of pretreated substrate and SC CO2 in thepacked bed during extraction.

In this work, authors solved numerically the differential massbalance equations of the model using a single value of K for eachsubstrate, and a pretreatment-dependent effective diffusivity coef-ficient (De) as adjustable model parameters. K is sometimes esti-mated as the ratio between the initial extract content C so andapparent solubility of the extract C fo. Although for a single extrac-tion condition K should be a function of the extract and substrate

surface chemistries only, which for pretreatments such as those

studied here do not change, values of the ratio C so/C fo varieddepending on herb pretreatment in this work (cf. Table 2). Thus,K was made independent of the C so/C fo ratio in this work as

0

2

4

6

8

10

12

0 20 40 60 80 100 120

0

1

2

3

4

5

Conventional-milling

Cryogenic-milling

Rapidly-expanded

Extraction time (min)

Extractyield

(g

/kg

drysubstrate)

A

B

Fig. 3. Effect of substrate conditioning on cumulative extraction plots of essential

oil versus time for (A) boldo leaves and (B) oregano bracts using supercritical CO2 at

40 °C and 10MPa (U = 1.24 mm/s). Samples were exposed to (} –Á–Á–) conventional

milling, (s –Á–Á–) low-temperature milling, or (4 –Á–Á–) rapid decompression.

Symbols and bars represent average and standard deviation of experimental data(duplicates) and lines best-fitting curves.

Table 2

Substrate characteristic, extraction yield and model parameters for supercritical CO 2 extraction experiments (40 °C and 10 MPa) of essential oils from boldo leaves and oregano

bracts.

Sample pretreatment C so (g/kg) C fo (g/kg CO2) K (À) kf Â 105 (m/s) DL Â 108 (m2/s) De Â 1012 (m2/s) F M (À)

Conventionally-milled boldo leaves 9.63 0.855 20.2 10.8 9.27 6.02 2204

Low-temperature-milled boldo leaves 10.8 0.887 20.2 10.8 9.28 7.64 1737

Rapidly-decompressed boldo leaves 11.4 0.934 20.2 9.24 9.55 65.6 202

Conventionally-milled oregano bracts 5.06 0.507 23.4 11.1 9.24 6.22 2162

Low-temperature-milled oregano bracts 5.33 0.475 23.4 11.0 9.30 6.95 1935

Rapidly-decompressed oregano bracts 5.42 0.127 23.4 9.50 9.67 38.5 349





proposed by Perrut et al. (1997) in their sorption isotherm/isobarmodel.

Additional parameters of the mathematical model include thefilm mass transfer coefficient (kf ), that was estimated using thedimensionless relationship of Puiggené et al. (1997), and the axialdispersion coefficient (DL ), that was estimated using the equation

of del Valle et al. (2011). These calculations required in turn esti-mates of the physical properties of the SC CO2 phase. The effect

of dissolved essential oil on the density (q) and viscosity (l) of the SC CO2 was neglected, being these properties estimated usingNIST Standard Reference Database (2007) for pure CO2 (q =628.6 kg/m3; l = 4.78Â 10À5 kg mÀ1 sÀ1 at 40 °C and 10 MPa). Val-ues of the binary diffusion coefficient of essential oil in CO2 (D12)were estimated using the correlation of Catchpole and King(1994) as a function of the reduced temperature (T r = T /T c = 1.029,where T c is the critical temperature of CO2, 304.3 K) and reduceddensity (qr = q/qc = 1.351, where qc is the critical density of CO2,467.6 kg/m3) of the SC CO2 phase, and the molecular mass (MW 2)and critical volume (V c2) of the essential oil. Using values of MW 2and V c2 reported by del Valle et al. (2011), that were in turn basedon literature reports on essential oil composition of boldo leaves

and oregano bracts, the estimated binary diffusion coefficients atprocess conditions were D

12= 1.33Â 10À8 m2/s for boldo essential

oil (MW 2 = 166.6 g/mol, V c2 = 521.0 cm3/mol) and D12 = 1.34Â10À8 m2/s for oregano essential oil (MW 2 = 151.9 g/mol,

V c2 = 514.5 cm3/mol).Fig. 3 includes best-fit curves for the extraction of herb essential

oils using CO2 at 40 °C and 10 MPa. It is apparent that the diffu-sional model of Araus et al. (2009) adequately describes the extrac-tion of essential oils from boldo leaves (Fig. 3A) and oregano bracts(Fig. 3B) when using K and De as fitting parameters. Best-fitted val-ues of K and De were of similar order of magnitude independent of the substrate and substrate pretreatment (Table 2). There were

small but consistent differences in effective diffusivity values as afunction of sample pretreatment, with the largest value being asso-ciated with rapid decompression, and the smallest with conven-tional milling, which mimics the effect of sample pretreatment in

extraction rate. This reveals the effectiveness of rapid decompres-sion as pretreatment to modify herb microstructure (rupturingsecretory cavities and glandular trichomes thus releasing essentialoils in boldo leaves and oregano bracts, respectively) so as to im-prove diffusion mechanisms controlling extraction rates, as at-tested by microscopy evidence (Fig. 2).

It is better to discuss the effect of substrate pretreatment onextraction rate as a function of the so-called microstructural factor

F M (Araus et al., 2009; del Valle et al., 2011):

F M ¼D12

De

: ð4Þ

Values of F M in Table 2 suggest that values of D12, which dependonly on essential oil composition and extraction conditions, are200–2000 times larger than values of De, which depend also on sub-

strate pretreatment. In the study of Araus et al. (2009) values of F Mwere from comparable to 10 times as large as those reported in Ta-ble 2 (ca. 2.5Â 104 in the case of oregano). Values covered an everwider interval (102 < F M < 109) in the review of del Valle et al.(2011), but values of F M depend critically on the assumptions of mass transfer models fitted to SC CO2 extraction data, that may dif-fer from actual conditions under which extractions takes place.Being the model applied the same in the two works, values of F M re-ported in Table 2 approach closely those reported by Araus et al.(2009).

Values of D12 in Eq. (4) may be questioned because authors didnot measure the chemical make up of boldo and oregano extracts.The composition of extracts typically changes during extractionbecause more soluble monoterpene hydrocarbons are extracted

at the beginning, whereas less soluble sesqueterpene and oxygen-ated compounds are extracted later (del Valle et al., 2011). How-ever, changes in D12 differ little between typical essential oilcomponents. For example, D12 of a heavy sesquiterpene ester (far-nesyl acetate) is only 75% that of a monoterpenene hydrocarbon(limonene), and only when a large amount of low solubility waxes

and triglycerides is incorporated to the extract would D12 decreaseto ca. 45% the value of a typical monoterpene hydrocarbon (del

Valle et al., 2011). When authors recalculated kf and DL usingD12 = 1.00Â 10-8 m2/s to simulate extraction curves using best fitvalues of De informed in Table 2, differences to curves in Fig. 3 wereminimal (data not shown). Authors concluded that small differ-ences in D12 associated to small differences in chemical makeupof extract should not affect best-fit values of De to a great extent.Of course the values of F M should change, but not the pretreat-ment-associated differences between samples.

4. Concluding remarks

Low-temperature milling and rapid decompression of CO2-impregnated samples were effective pretreatments to increasethe rate and/or yield of SC CO2 extraction processes for herb essen-

tial oils. Low-temperature conditioning may reduce losses of vola-tile compounds during milling and other thermally-induceddeteriorative reactions, thus increasing extraction yield andimproving extract quality. Cryogenic conditions can make a vege-table matrix more brittle thus improving particle size reductionduring milling, but this was not observed in our experiments.The reduction in particle size increases the specific surface area

thus allowing faster extractions, but there are hydrodynamic limi-tations to the flow of SC CO2 in a packed bed that may preclude theuse of very small particles in extraction experiments.

Rapid decompression allows not only a selective disruption of glandular trichomes in the surface of Lamiaceae herbs that holdessential oils, but also of the secretory cavities within the plant tis-sue having a similar function in other herbs (e.g., boldo leaves).This tissue disruption facilitates mass transfer during SC CO2

extraction as attested by a large decrease of the microstructuralfactor that characterizes the effect of the solid matrix on theextraction rate. The extraction of the essentials oil from aromaticherbs can be described by a diffusional model which assumes thatthe partition of essential oils between the solid substrate and CO 2

phases is constant, and which uses the effective diffusivity of theextract in the solid matrix as fitting parameter.

Acknowledgment

Funding by Comisión de Investigación Cientifica y Tecnológica(Fondecyt project 105-0675) from Chile is greatly acknowledged.

References

Araus, K., Uquiche, E., del Valle, J.M., 2009. Matrix effects in supercritical CO2

extraction of essential oils from plant material. Journal of Food Engineering 92,438–447.

Berna, A., Cháfer, A., Montón, J.B., Subirats, S., 2001. High-pressure solubility data of system ethanol (1) +catechin(2) +CO2 (3). The Journal of Supercritical Fluids 20,157–162.

Bisset, N.G. (Ed.), 1994. Herbal Drugs and Phytopharmaceutical. A Handbook forPractice on a Scientific Basis. Medpharm GmbH Scientific Publ., Stuttgart,Germany.

Bosabalidis, A.M., Tsekos, I., 1982. Glandular scale development and essential oilsecretion in Origanum dictamnus L. Planta 156, 496–504.

Bosabilidis, A.M., Tsekos, I., 1984. Glandular hair formation in Origanum species.Annals of Botany 53, 559–563.

Brielmann, H.L., Setzer, W.N., Kaufman, P.B., Kirakosyan, A., Cseke, L.J., 2006.Phytochemicals: the chemical components of plants. In: Cseke, L.J., Kirakosyan,A., Kaufman, P.B., Warber, S.L., Duke, J.A., Brielmann, H.L. (Eds.), NaturalProducts from Plants, second ed. Taylor & Francis Group, Boca Raton, FL,pp. 1–49.





Bruns, K., Köhler, M., 1974. Über die Zusammensetzung des Boldoblätteröls.Parfümerie und Kosmetik 55, 225–227.

Catchpole, O.J., King, M.B., 1994. Measurement and correlation of binary diffusioncoefficients in near critical fluids. Industrial and Engineering ChemistryResearch 33, 1828–1837.

de la Fuente, J.C., Quezada, N., del Valle, J.M., 2005. Solubility of boldo leaf antioxidant components (boldine) in high-pressure carbon dioxide. Fluid PhaseEquilibria 235, 196–200.

del Valle, J.M., Aguilera, J.M., 1999. Extracción con CO2 a alta presión. Fundamentosy aplicaciones en la industria de alimentos. Food Science and Technology

International 5, 1–24.del Valle, J.M., Jiménez, M., Napolitano, P., Zetzl, C., Brunner, G., 2003. Supercriticalcarbon dioxide extraction of pelletized Jalapeño. Journal of the Science of Foodand Agriculture 83, 550–556.

del Valle, J.M., Godoy, C., Asencio, M., Aguilera, J.M., 2004. Recovery of antioxidantsfrom boldo (Peumus boldus M.) by conventional and supercritical CO2

extraction. Food Research International 37, 695–702.del Valle, J.M., Rogalinski, T., Zetzl, C.,Brunner, G., 2005.Extraction of boldo (Peumus

boldus M.) leaves with supercritical CO2 and hot pressurized water. FoodResearch International 38, 203–213.

del Valle, J.M., de la Fuente, J.C., Uquiche, E., Zetzl, C., Brunner, G., 2011. Masstransfer and equilibrium parameters on high-pressure CO2 extraction of plantessential oils. In: Aguilera, J.M., Simpson, R., Welti-Chanes, J., Bermudez-Aguirre,D., Barbosa-Canovas, G. (Eds.), Food Engineering Interfaces. Springer, New York,NY, pp. 393–470.

Dembisky, V.M., 2008. Bioactive peroxides as potential therapeutic agents.European Journal of Medicinal Chemistry 43, 223–251.

Efferth, T., Olbrich, A., Sauerbrey, A., Ross, D.D., Gebhart, E., Neugebauer, M., 2002.Activity of ascaridol from the anthelmintic herb Chenopodium anthelminticum L.Against sensitive and multidrug resistant tumor cells. Anticancer Research 22C,4221–4224.

Gaspar, F., Santos, R., King, M.B., 2000. Extraction of essential oils and cuticularwaxes with compressed CO2: effects of matrix pretreatment. IndustrialEngineering Chemistry Research 39, 4603–4608.

Gaspar, F., Santos, R., King, M.B., 2001. Disruption of glandular trichomes withcompressed CO2: alternative matrix pre-treatment for CO2 extraction of essential oils. The Journal of Supercritical Fluids 21, 11–22.

Gaspar, F., Leeke, G.A., Al-Duri, B., Santos, R., 2003. Modelling the disruption of essential oils glandular trichomes with compressed CO2. Journal of SupercriticalFluids 25, 233–245.

Krejc ová, A., Pouzar, M., C ernohorsky, T., Pešková, K., 2008. The cryogenic grindingas the important homogenization step in analysis of inconsistent food samples.Food Chemistry 109, 848–854.

Krstic, L., Malencic, D., Anackov, G., 2006. Structural investigations of trichomes andessential oil composition of Salvia verticillata. Botanica Helvetica 116, 159–168.

Manohar, B., Sridhar, B., 2001. Size and shapecharacterizationof conventionally andcryogenically ground turmeric (Curcuma domestica) particles. PowderTechnology 120 (3), 292–297.

Martínez-Laborde, J.B., 1988. Some comments on a recent classification of the

Monimiaceae. Taxon 37, 834–837.Mazutti, M., Mossi, A.J., Cansian, R.L., Corazza, M.L., Dariva, C., Oliveira, J.V., 2008.

Chemical profile and antimicrobial activity of boldo (Peumus boldus Molina)extracts obtained by compressed carbon dioxide extraction. BrazilianJournal of Chemical Engineering 25, 427–434.

Miraldi, E., Ferri, S., Franchi, G.G., Giorgi, G., 1996. Peumus boldus essential oil: newconstituents and comparison of oils from leaves of different origen. Fitoterapia67, 227–230.

Murthy, C.T., Bhattacharya, S., 2008. Cryogenic grinding of black pepper. Journal of Food Engineering 85, 18–28.

Nagy, B., Simándi, B., András, C.D., 2008. Characterization of packed beds of plantmaterials processed by supercritical fluid extraction. Journal of FoodEngineering 88, 104–113.

NIST Standard Reference Database, 2007. 23 (REFPRO), Gaithersburgh, MD.<www.nist.gov>.

Pare, P.W., Zajicek, J., Ferracini, V.L., Melo, I.S., 1993. Antifungal terpenoids fromChenopodium ambrosiodies. Biochemical Systematics and Ecology 21, 649–653.

Pasquel, A., Meireles, M.A.A., Marques, M.O.M., Petenate, A.J., 2000. Extraction of stevia glycosides with CO2 plus water, CO2 plus ethanol, and CO2 plus waterplus ethanol. Brazilian Journal of Chemical Engineering 17, 271–282.

Perrut, M., Clavier, J.Y., Poletto, M., Reverchon, E., 1997. Mathematical modeling of sunflower seed extraction by supercritical CO2. Industrial and EngineeringChemistry Research 36, 430–435.

Pesek, C.A., Wilson, L.A., Hammond,E.G., 1985. Spice quality: effect of cryogenic andambient grinding on volatiles. Journal of Food Science 50, 599–601.

Pruthi, J.S., 1980. Spices and Condiments: Chemistry, Microbiology, Technology.Academic Press, New York, NY.Puiggené, J., Larrayoz, M.A., Recasens, F., 1997. Free liquid-to-supercritical fluid

mass transfer in packed beds. Chemical Engineering Science 52, 195–212.Quezada, N., Asencio, M., del Valle, J.M., Aguilera, J.M., Gómez, B., 2004. Antioxidant

activity of crude extract, alkaloid fraction, and flavonoid fraction from boldo(Peumus boldus Mol.) leaves. Journal of Food Science 69, C371–C376.

Quirin, K.-W., Gerard, D., 2007. Supercritical fluid extraction (SFE). In: Ziegler, H.(Ed.), Flavourings: Production, Composition, Applications, Regulations, seconded. Wiley–VCH, Weinheim, Germany, pp. 49–65.

Reverchon, E., De Marco, I., 2007. Essential oil extraction and fractionation usingsupercritical fluids. In: Martínez, J.L. (Ed.), Supercritical Fluid Extraction of Nutraceuticals and Bioactive Compounds. CRC Press, Boca Raton, FL, pp. 305–335.

Reverchon, E., Donsi, G., Sesti Osséo, L., 1993. Modeling of supercritical fluidextraction from herbaceous matrices. Industrial Engineering ChemistryResearch 32, 2721–2726.

Salisbury, F.B., Ross, C.W., 1992. Plant Physiology, fourth ed. Wadsworth Publ. Co.,Belmot, CA.

Schmeda-Hirschmann, G., Rodríguez, J.A., Theoduloz, C., Astudillo, S.L., Feresin, G.E.,Tapia, A., 2003. Free radical scavengers and antioxidants from Peumus boldusMol. (‘‘Boldo’’). Free Radicals Research 37, 447–452.

Speisky, H., Cassels, B.K., 1994. Boldo and boldine: an emerging case of natural drugdevelopment. Pharmacological Research 29, 1–10.

Stamenic, M., Zizovic, I., Orlovic, A., Skala, D., 2008. Mathematical modelling of essential SFE on the micro-scale. Classification of plant material. Journal of Supercritical Fluids 46, 285–292.

Stamenic, M., Zizovic, I., Eggers, R., Jaeger, P., Heinrich, H., Rój, E., Ivanovica, J., Skala,D., 2010. Swelling of plant material in supercritical carbon dioxide. Journal of Supercritical Fluids 52, 125–133.

Uauy V., J.N., 1998. Extracción de aceite esencial de plantas nativas chilenas.Unpublished B.Sc. thesis, Pontificia Universidad Católica de Chile, Santiago,Chile.

Uquiche, E., del Valle, J.M., Ortiz, J., 2004. Supercritical carbon dioxide extraction of red pepper (Capsicum annuum L.) oleoresin. Journal of Food Engineering 65, 55–66.

Vila, R., Valenzuela, L., Bello, H., Cañigueral, S., Montes, M., Adzet, T., 1999.Composition and antimicrobial activity of the essential oil of Peumus boldusleaves. Planta Medica 65, 178–179.

Vogel, H., Razmilic, I., Muñoz, M., Doll, U., San Martín, J., 1999. Studies of geneticvariation of essential oil and alkaloid content in boldo ( Peumus boldus). PlantaMedica 65, 90–91.

Yoda,S.K., Marques, M.O.M., Petenate, A.J., Meireles, M.A.A., 2003. Supercritical fluidextraction from Stevia rebaudiana Bertoni using CO2 andCO2 + water: extractionkinetics and identification of extracted components. Journal of FoodEngineering 57, 125–134.

Zizovic, I., Stamenic, M., Orlovic, A., Skala, D., 2005. Supercritical carbon dioxideessential oil extraction of Lamiaceae familyspecies: mathematical modelling onthe micro-scale and process optimization. Chemical Engineering Science 60,6747–6756.

Zizovic, I.T., Stamenic, M.D., Orlovic, A.M., Skala, D.U., 2007. Supercritical carbon-dioxide extraction of essential oils and mathematical modelling on the micro-scale. In: Berton, L.P. (Ed.), Chemical Engineering Research Trends. Nova Sci.Publ. Inc., New York, NY, pp. 221–249.


Please cite this article in press as: Uquiche, E., et al. Effect of boldo ( Peumus boldus M.) pretreatment on kinetics of supercritical CO2 extraction of essentialil J l f F d E i i (2011) d i 10 1016/j jf d 2011 10 013

Documents

Www.sciencedirect.com Science MiamiImageURL Cid=271146& User=5254909& Pii=S0260877411005516& Check=y& Origin=& Cove