Original article

Fruits, vol. 65 (6) 387

Influence of curing procedures on sensory quality of vanilla beans.Abstract — Introduction. During maturation, vanillin is accumulated in green vanilla beans as glu-covanillin; it is hydrolyzed to free vanillin by endogenous glucosidases during curing and gives the cha-racteristic flavor of vanilla. The objective of our study was to investigate methods of curing that couldgreatly reduce the time to complete the process and yield cured beans that retain high concentrationsof vanillin and other flavor compounds with high sensory quality rating. Materials and methods.Mature green beans were obtained from a commercial grower (Cairns, Queensland, Australia). Onebatch of beans was continuously sweated at 35 °C at high relative humidity (RH) for 12 d. Two furtherbatches were blanched in water at 67 °C for 3 min, and then sweated at 45 °C at high RH for 4 d orat 35 °C for 5 d. The beans were sweated until they turned brown. Three methods of drying were eva-luated: a heat pump dryer at 40 °C and RH 15%, tunnel dryer at 60 °C and RH 20%, and tunnel dryerat 60 °C and RH 10%. Vanillin was extracted from powdered samples of beans with n-pentane anddichloromethane (1:1 v/v) and assayed by HPLC. Glucovanillin was measured as total vanillin after acidhydrolysis of powdered samples of beans. Results and discussion. About 90% of the glucovanillinwas converted to vanillin in non-blanched beans continuously sweated at 35 °C, but there was only70% conversion in beans blanched at 67 °C for 3 min and sweated at 45 °C for 4 d or at 35 °C for 5 d.The sensory quality of cured beans was assessed by untrained panelists. Profiling showed that the beanssweated continuously at 35 °C had superior aroma compared with beans blanched in hot water andsweated at 45 °C or at 35 °C but the appearance of non-blanched beans was less attractive. Conclusion.The study revealed that a mild hot water blanching treatment followed by sweating at 35–45 °C andrapid drying is required to produce cured beans with excellent appearance and attractive aroma.

Australia / Vanilla planifolia / vanilla (spice)/ vanillin / processing / foodtechnology / hot air drying / quality controls / organoleptic analysis

Influence des méthodes de préparation des gousses de vanille sur leurqualité sensorielle.Résumé — Introduction. Au cours de la maturation, la vanilline est accumulée dans les gousses devanille verte sous forme de glucovanilline ; elle est hydrolysée en vanilline par des glucosidases endo-gènes pendant le processus de préparation et donne la saveur caractéristique de la vanille. L'objectifde notre étude a été d'étudier des méthodes de préparation de la vanille qui pourraient réduire consi-dérablement le temps du déroulement du processus et produire des gousses à fortes concentrationsde vanilline et autres composés d'arôme avec de hautes qualités sensorielles. Matériel et méthodes.Des gousses de vanille mûres ont été obtenues chez un producteur commercial (Cairns, Queensland,Australie). Un lot de gousses a été continuellement étuvé à 35 °C et à une humidité relative (HR) élevéependant 12 j. Deux autres lots ont été blanchis dans l'eau à 67 °C pendant 3 min, puis étuvés à 45 °Cavec une HR élevée pendant 4 j ou à 35 °C pendant 5 j. Les gousses ont été étuvées jusqu'à ce qu'ellesbrunissent. Trois méthodes de séchage ont été évaluées : à l’aide d’une pompe à chaleur sèche à 40 °Cet 15 % HR, d’un tunnel de séchage à 60 °C et 20 % HR, et d’un tunnel de séchage à 60 °C et 10 % d’HR.La vanilline a été extraite avec du n-pentane et du dichlorométhane (1:1 v/v) et dosée par HPLC, à partird'échantillons de poudre obtenue des gousses traitées. La glucovanilline a été mesurée par le taux devanilline totale obtenue après hydrolyse acide des échantillons de poudre. Résultats et discussion.Environ 90 % de la glucovanilline ont été convertis en vanilline dans les gousses non blanchies et conti-nuellement étuvées à 35 °C mais il y a eu seulement 70 % de conversion dans les gousses blanchiesà 67 °C pendant 3 min et étuvées à 45 °C pendant 4 j ou à 35 °C pendant 5 j. La qualité sensorielle desgousses traitées a été évaluée par des panélistes inexpérimentés. Leur analyse a montré que les goussesétuvées continuellement à 35 °C avaient un arôme supérieur à celui des gousses blanchies dans l'eauchaude et étuvées à 45 °C ou à 35 °C, mais l'apparence des gousses non blanchies était moins attrayante.Conclusion. L'étude a révélé qu’un traitement modéré de blanchiment à l’eau chaude, suivi par unétuvage à 35–45 °C et un séchage rapide était nécessaire pour obtenir des gousses préparées ayant untrès bel aspect et un arôme intéressant.

Australie / Vanilla planifolia / vanille / vanilline / traitement / technologiealimentaire / séchage par air chaud / contrôle de qualité / analyse organoleptique

1 Univ. Western Sydney, Cent.Plants Environ., Locked Bag1797, Penrith South DC, NSW1797, Australias.vandyk@uws.edu.au

2 Univ. Western Sydney,School Nat. Sci., Locked Bag1797, Penrith South DC, NSW1797, Australia

3 Former., Sensory SolutionsLtd., Northmead, NSW 2145,Australia;current address, 12 PartridgeAve., Castle Hill, NSW 2154,Australia

Influence of curing procedures on sensory quality of vanilla beansSahar VAN DYK1*, William Barry MCGLASSON1, Mark WILLIAMS2, Cathy GAIR3

* Correspondence and reprints

Received 14 September 2009Accepted 21 April 2010

Fruits, 2010, vol. 65, p. 387–399© 2010 Cirad/EDP SciencesAll rights reservedDOI: 10.1051/fruits/2010033www.fruits-journal.org

RESUMEN ESPAÑOL, p. 399

Article published by EDP Sciences and available at http://www.fruits-journal.org or http://dx.doi.org/10.1051/fruits/2010033

388 Fruits, vol. 65 (6)

S. van Dyk et al.

1. Introduction

Vanilla (Vanilla planifolia, Andrews.) is aclimbing orchid indigenous to Mexico [1].The distinctive flavor and aroma of curedbeans is derived mainly from the phenoliccompound, vanillin, plus other aromaticcompounds that comprise less than 2% of thefresh weight of vanilla beans. Green, maturevanilla beans contain glycosyl precursors ofthese aroma compounds, of which fifteenhave been identified [2, 3]. The most impor-tant flavor precursor is glucovanillin, whichaccumulates in the fruit from approximately(15 to 30) weeks after pollination when thebeans are mature [4–6]. The distribution ofglucovanillin and β-glucosidase activity inbean tissue is a contentious subject [7]. Joelet al. showed that the enzymes involved invanillin biosynthesis are present in the secre-tory tissue that is composed of closely-packed, unicellular hairs, located betweenthe placentae along the central fruit cavity [8].Odoux et al. showed that glucovanillin andβ-glucosidase activity are present in the cen-tral placental tissue of the fruit [9]. Morerecently, Odoux and Brillouet confirmedthat more than 90% of the stored glucova-nillin is located in the placentae [10].

Mature green pods generate their charac-teristic flavor or aroma and turn brownwhen they are cured. The purpose of thecuring process is to bring together the flavorprecursors and the enzymes that catalyzethe hydrolysis of these compounds to flavorproducts [11]. The traditional curing processconsists of four major stages: killing, sweat-ing, drying and conditioning, during whichglucosidases are released that hydrolyzeglucovanillin and other substrates to releasethe aroma compounds. All four stages areimportant, but the killing stage is the mostimportant [12]. Killing stops further physio-logical activity and promotes cell disorgan-ization which, in turn, causes a mixing offlavor and aroma precursors with theirrespective enzymes [13]. The most commonkilling methods involve exposure to sun-light, the use of a controlled temperatureoven or hot water [14]. The traditional curingprocess can take 5 weeks to 5 monthsdepending on the production region andprocesses used [11].

In the last century, many studies weredevoted to the determination of conditionsfor producing good-quality, cured vanillabeans [15–17], isolation of aroma glucosidesresponsible for vanilla flavor production [18,19] and assaying the activity of enzymes thatare involved in the production of flavorcompounds [11, 17–22]. The aim of ourstudy was to examine methods for optimiz-ing the curing and drying of vanilla beansand to assess the effects of these methodson the sensory quality of cured beans. Thestudy evaluated different combinations ofcuring times, curing temperatures and dry-ing methods, including a comparison ofblanched and non-blanched beans.

2. Materials and methods

2.1. Plant materials and treatments

Mature green vanilla beans were suppliedby a commercial grower from Cairns,Queensland, Australia. The beans were har-vested on 1 September, 2006, and sent byair to the Postharvest Laboratory at the Uni-versity of Western Sydney (HawkesburyCampus), New South Wales. Any beansshowing browning or splitting at the blos-som end were trimmed with a sharp knifeto remove the damaged tissue. The beanswere graded into nine matching groups of30 beans plus one group of 6 beans thatserved as an unprocessed control. The initialfresh mass of each bean was recorded andthe dry matter of fresh beans (18–20% FW)was determined by drying to constantweight using a microwave oven.

2.2. Blanching, curing and sweatingmethods

Three groups of 30 non-blanched greenbeans were cured by placing them in plasticbags in an incubator at 35 °C for 12 days.The beans were examined daily until allbeans turned dark brown. Generally, thisbrowning started at the blossom end of thebeans. Six groups of 30 green beans wereblanched by immersion in tap water main-tained at 67 °C for 3 min. After blanching,the beans were enclosed in polyethylene

Curing procedures of vanilla beans

Fruits, vol. 65 (6) 389

bags and transferred to incubators. Threegroups of beans were incubated (sweated)at 45 °C for 4 days and the remaining threegroups were sweated at 35 °C for 5 days.Sweating was terminated when all beansturned dark brown. The processed beanswere stored at –20 °C until drying. Theweight of individual beans was recorded onarrival at the laboratory, after sweating andfollowing drying.

2.3. Drying methods

The nine groups (figure 1) of frozen beanswere transported in an insulated containerto the University of New South Wales,Department of Food Science and Technol-ogy, for drying. One group of beans fromeach curing treatment was partially dried toabout 45% of their fresh weight (dry matter40–44%) using one of two methods of dry-ing. The first used a heat pump dryer, inwhich the beans were placed in single layerson trays in the dryer. The beans that wereblanched at (45 or 35) °C were dried at 40 °Cand RH 15% for 5 days, and non-blanchedbeans that were cured at 35 °C were driedat 40 °C and RH 15 % for 10 days. The sec-ond method of drying used a tunnel dryer,in which groups of beans were placed in sin-gle layers on trays and dried at 60 °C ateither 20% or 10% RH for 2 days. At the endof the drying period, the nine groups ofbeans were sealed separately in glass jars fit-ted with screw lids. The dried beans werestored at 20 °C pending further analyses.

2.4. Extraction of vanillin

The beans were sampled for vanillin contentat the green stage, at the end of sweatingand at the end of drying. Six representativebeans from each stage of the curing and dry-ing treatment were stored at –80 °C pendinganalysis. Beans were cut into pieces approx-imately 20 mm long, frozen in liquid N2 andmilled to a fine powder using a domesticcoffee blender. This powder was allowed towarm to room temperature. Standard prac-tice was to suspend a sample (200 mg) ofpowder in 10 mL distilled water and to stirfor 10 min at lab temperature 22 °C (pH 5.1)before solvent extraction. Each aqueous

suspension was transferred to a separatingfunnel, 10 mL of n-pentane and dichlo-romethane (1:1 v/v) were added to extractthe vanillin and the mixture was shaken andallowed to settle. The upper, organic phasewas recovered and the remaining aqueouslayer was extracted three more times withthe n-pentane dichloromethane mixture.The four organic phases were combined,dried over anhydrous Na2SO4 and filtered(polyamide: polymer, 6:6). The filtrate wasadjusted to 50 mL in a volumetric flask withn-pentane and dichloromethane (1:1 v/v).

A supplementary experiment was con-ducted to determine whether there was anyglucosidase activity in these aqueous sus-pensions. Powdered samples were sus-pended in water for up to 120 min beforesolvent extraction.

2.5. Hydrolysis of glucovanillin

The beans were also assayed for total vanillincontent (vanillin obtained after hydrolysisplus free vanillin) when still green and at theend of sweating and drying. Beans for assaywere powdered as above and 200 mg ofpowder suspended in 10 mL distilled water.After addition of 0.5 mL of sulfuric acid(18 M), the suspension was thoroughlymixed and placed in a steam bath at 60 °C

Figure 1.Plan of the experiment done forstudying the influence of curingprocedures on sensory qualityof vanilla beans.

390 Fruits, vol. 65 (6)

S. van Dyk et al.

for 3 h. The mixture was cooled to room tem-perature and 1 mL KOH (9.4 M) was addedto neutralize the mixture. The resulting aque-ous mixture was extracted as outlined above.

2.6. Determination of vanillin

Vanillin (12.1 mg) (Aldrich Chemical Com-pany, 99%) was dissolved in 10 mL of HPLCgrade n-pentane and dichloromethane (1:1v/v). The solution was transferred into a100-mL volumetric flask and diluted to themark with n-pentane and dichloromethane(1:1 v/v). This standard contained 121 µg·L–¹

vanillin. Five mL of the vanilla extract waspipetted into four 10-mL volumetric flasks.Aliquots [(0.0, 1.0, 2.0 and 3.0) mL] of thestandard vanillin solution were pipetted intothese flasks and the extracts made up to vol-ume with n-pentane and dichloromethane(1:1 v/v). Vanillin was separated from otherorganic compounds extracted from thepowdered vanilla beans by HPLC. The chro-matograph used was a Varian 9012equipped with a 20-µL sample loop, a Varian9050 variable wavelength UV-visible detec-tor and a reverse-phase C18 column(Gemini 5U C18 110A, 250 mm × 4.60 mm,Phenomenex, Australia). The mobile phaseconsisted of 10% water, 10% acetonitrile and80% methanol. Vanillin was identified bycomparison with authentic vanillin at271 nm. The data for six replicates from thenine treatments were subjected to ANOVAusing SPSS (version 12) and the means wereseparatedby theRyan-Einot-Gabriel-Welschrange test at P ≤ 0.05.

2.7. Sensory analysis

The aroma and sensory quality of driedcured beans were assessed by 64 untrainedpanelists. The sensory analysis was con-ducted by Sensory Solutions Ltd. (North-mead, New South Wales). Each of the ninegroups of treated beans was divided intothree matched samples of six beans. Thethree samples from each treatment wererequired to ensure that enough sampleswere available to accommodate sessions ofup to 25 consumers conducting simultane-ous evaluations. Each sample of six beanswas enclosed in a 2.5-L amber bottle (Plas-

dene Glass Pack) labeled with a three-digitcode for identification about 24 h prior toconsumer assessment. In addition, threevanillin reference samples were alsoenclosed in 1-L amber bottles. A total often samples, one bottle of the vanillin refer-ence sample and one bottle from each of thenine treatments, were evaluated by eachpanelist. The reference sample was testedfirst, followed by the nine treatment samples,that were presented one at a time in randomorder to reduce the effects of fatigue andpositional bias. The session took 45 min.

At the start of the consumer evaluationsession and in between each sample, pan-elists were asked to sniff tissue paper to pro-vide a neutral odor from which to start eachevaluation. The panelists were asked tofocus on the odor and answer two types ofquestions: hedonic (liking) and diagnostic(strength) questions. All questions wereanswered using a 100-mm unstructured linescale with labeled anchor points at 0 mmand 100 mm. The left hand anchor pointlabels corresponded to either “dislikeextremely” or “very weak”, for the hedonicand diagnostic scales, respectively. Theright hand anchor point labels were “likeextremely” or “very strong”. The question-naire for the sensory assessment wasdivided into three sections: in the first part,the panelists were asked to answer shortquestions about personal information andusage, and attitudes about vanilla beforetesting the fragrances. The second sectioncontained three hedonic and diagnosticquestions about the reference sample ofvanillin. The third section contained a totalof 30 hedonic and diagnostic questions thatwere repeated for each of the nine samples.Within the third section, panelists wereasked to smell each of the samplesthree times over the course of the 30 ques-tions. The initial group of questions askedafter the first smell included overall liking ofthe vanilla fragrance, the strength of thearoma and its sweetness. The next group ofquestions asked after smelling the beans asecond time included the level of liking andstrength of aroma described as fruity, floral,cereal, spicy, fermented and acidic. Theremaining questions asked after smellingthe beans a third time covered the qualityof the vanilla odor for usage as a fragrance

Curing procedures of vanilla beans

Fruits, vol. 65 (6) 391

or for cooking and an overall liking or dis-liking of the fragrance they perceived.

The data extracted from the 100-mm linescales were entered into Excel spreadsheets(Microsoft Corporation). The data were sub-jected to ANOVA using SPSS (version 12,SPSS 2004).

3. Results and discussion

All powdered samples were routinelysoaked for 10 min in distilled water beforesolvent extraction of vanillin. A pilot studyshowed that prolonged soaking of pow-dered samples of green beans in distilledwater at 22 °C for up to 120 min did notaffect the concentrations of free vanillin,indicating that β-glucosidase was inactiveunder these conditions. The pH of the aque-ous suspensions was measured at 5.1. Puri-fied β-glucosidase has been shown to beinactive and unstable below pH 6.5 [23].

The concentrations of free vanillin werelow in green beans but increased signifi-cantly during curing (figure 2). Percentageconversion of glucovanillin to vanillin washighest in non-blanched beans sweatedcontinuously at 35 °C, and least in beansblanched for 3 min in water at 67 °C fol-lowed by sweating at (45 or 35) °C. Thesedata suggest that the mixing of glucovanil-lin and glucosidases was more efficient inbeans sweated continuously at 35 °C, butalso that blanching at 67 °C partiallyreduced glucosidase activity [24]. The lattersuggestion was supported by the data forcured beans following drying (figure 3).There was a further much larger increase inthe percentage of free vanillin during dry-ing in beans sweated continuously at 35 °Ccompared with beans that were blanchedthen sweated at (45 or 35) °C (figure 3).These data highlight the large variation inglucovanillin concentrations among beans.Producing green beans of consistent sizeand glucovanillin content appears essentialfor the production of uniform, high-qualitycured beans. Although less glucovanillinwas hydrolyzed in blanched beans com-pared with those sweated continuously at35 °C, their appearance was more attrac-tive. Their color was shiny brown-black

Figure 2.Total and free vanillin concentrations in green beans and in blanched and non-blanched vanilla beans after sweating. Total vanillin was measured following acidhydrolysis of powdered samples of beans. The inserted numbers show the %conversion from glucovanillin to vanillin. Bars marked by the same letter are notsignificantly different at P ≤ 0.05 (n = 6).

Figure 3.Concentrations of total and free vanillin in cured vanilla beans at the end of dryingstages of each of nine curing treatments. Note that cured beans were dried to anaverage of 45% of their harvest fresh weight. Bars for each treatment marked by thesame letter are not significantly different at P ≤ 0.05 (n = 6). HPD: heat pump dryer,TDH: tunnel dryer with high RH, TDL: tunnel dryer with low RH.

392 Fruits, vol. 65 (6)

S. van Dyk et al.

compared with non-blanched beans,which were dull brown. The method ofdrying appeared not to have had any inde-pendent effects on hydrolysis of glucova-nillin (figure 3).

3.1. Sensory analysis

Odor profiling showed that beans curedcontinuously at 35 °C without blanching

had higher concentrations of impact aro-mas, including those giving vanilla, sweet,floral, fruity and spicy smells compared withbeans that were blanched before sweating(figure 4, 5). Beans that were blanched andsweated at 45 °C had the strongest spicy andacidic odors (figure 6). However, beanscured continuously at 35 °C had the highestquality scores for use in cooking or as a fra-grance (figure 7).

Figure 4.Comparison of sensory scoresof liking of sweet, fruity, floraland spicy odors of vanillabeans cured using ninedifferent processes. Barsmarked by the same letterwithin each quality attribute arenot significantly different atP ≤ 0.05. HPD: HPD: heatpump dryer, TDH: tunnel dryerwith high RH, TDL: tunnel dryerwith low RH.

Figure 5.Comparison of sensory scoresof strength of sweet, fruity andfloral odors of vanilla beanscured using nine differentprocesses. Bars marked by thesame letter within each qualityattribute are not significantlydifferent at P ≤ 0.05. HPD: heatpump dryer, TDH: tunnel dryerwith high RH, TDL: tunnel dryerwith low RH.

Curing procedures of vanilla beans

Fruits, vol. 65 (6) 393

3.2. Correlation between vanillinconcentrations and sensory scores

Overall liking of the fragrance of the curedvanilla beans among the nine curing anddrying treatments was highly correlated

with quality of the aroma for cooking or asa fragrance, liking of sweetness and thepresence of fruity, straw and spicy odors.Liking of acidic fragrance of the curedvanilla beans was not correlated with likingof sweet, fruity and floral odors (table I).

Figure 6.Sensory scores of strength ofstraw, spicy, fermented andacidic odors of vanilla beanscured using nine differentprocesses. Bars marked by thesame letter within each qualityattribute are not significantlydifferent at P ≤ 0.05. HPD: heatpump dryer, TDH: tunnel dryerwith high RH, TDL: tunnel dryerwith low RH.

Figure 7.Sensory scores of odor qualityfor cooking or fragrance andoverall liking of vanilla odor ofvanilla beans cured using ninedifferent processes. Barsmarked by the same letterwithin each quality attribute arenot significantly different atP ≤ 0.05. HPD: Heat pumpdryer, HPD: heat pump dryer,TDH: tunnel dryer with high RH,TDL: tunnel dryer with low RH.

394 Fruits, vol. 65 (6)

S. van Dyk et al.

Tab

leI.

Cor

rela

tion

coef

ficie

nts

(r)f

rom

the

sens

ory

eval

uatio

nof

cure

dva

nilla

bea

ns.V

alue

s>

0.50

are

the

mos

tim

por

tant

corr

elat

ions

.

Sen

sory

eval

uatio

nO

vera

lllik

ing

vani

llafr

agra

nce

Like

swee

tfr

agra

nce

Like

frui

tfr

agra

nce

Like

flora

lfr

agra

nce

Like

bea

nyfr

agra

nce

Like

stra

wfr

agra

nce

Like

spic

yfr

agra

nce

Like

acid

icfr

agra

nce

Qua

lity

frag

ranc

efo

rco

okin

g

Qua

lity

forf

ragr

ance

Ove

rall

like

the

frag

ranc

ep

erce

ived

Ove

rall

likin

gva

nilla

frag

ranc

e_

Like

swee

tfr

agra

nce

0.61

2_

Like

frui

tfr

agra

nce

0.62

90.

733

_

Like

flora

lfr

agra

nce

0.31

80.

582

0.51

3_

Like

bea

nyfr

agra

nce

0.56

70.

565

0.56

50.

409

_

Like

stra

wfr

agra

nce

0.61

00.

623

0.62

80.

625

0.73

5_

Like

spic

yfr

agra

nce

0.61

60.

517

0.67

00.

360

0.51

40.

515

_

Like

acid

icfr

agra

nce

0.20

20.

116

0.19

10.

178

-0.1

10.

233

0.20

2_

Qua

lity

frag

ranc

efo

rco

okin

g0.

587

0.58

60.

595

0.38

60.

628

0.57

10.

585

0.09

3_

Qua

lity

for

frag

ranc

e0.

634

0.45

60.

537

0.20

10.

524

0.44

90.

583

-0.1

00.

678

_

Ove

rall

like

the

frag

ranc

ep

erce

ived

0.73

00.

613

0.68

80.

465

0.72

90.

786

0.63

7-0

.10

0.74

80.

704

_

Curing procedures of vanilla beans

Fruits, vol. 65 (6) 395

The strength of the spicy odor was highlycorrelated with the strength of the acidicodor. Similarly, the strength of the sweetnessodor from cured vanilla beans among thetreatments was highly correlated with thestrength of spicy and acidic odors. The qual-ity of the vanilla fragrance for cooking washighly correlated with the quality of thebeans for fragrance and overall liking of thefragrance of the cured vanilla beans(table II).

The scores for vanilla, floral, fruity andsweet fragrances were high in beans curedcontinuously at 35 °C, due in part to highervanillin concentrations (groups 6, 7 and 9)(figure 8). The fermented odor scores werealso high in beans cured continuously at35 °C, presumably because respiratoryenzymes remained active. The spicy odorscores were high in beans blanched andsweated at 45 °C (groups 1, 3 and 4) (fig-ure 8); however, these beans had lowerconcentrations of vanillin. These data showthat the sensory quality of vanilla beans doesnot depend only on vanillin content but isalso affected by the formation of other vol-atile compounds that may include vanillicacid, p-hydroxybenzoic acid and p-hydroxybenzaldehyde [11, 25, 26]. Adedejiet al. identified 250 constituents that mayimpart cured vanilla beans with their char-acteristic aroma and flavor; chief amongthem is vanillin [27]. Arana found that sweat-ing and drying at 45 °C gave superior qualitybeans [28], whereas Rivera and Hagemanfound that sweating of beans at 38 °C pro-duced a better product than at 45 °C [29].Arana emphasized the potential importanceof peroxidases in particular and oxidativeenzymes in general during curing, becauseperoxidase is resistant to inactivation by var-ious killing and conditioning treatments.Peroxidase activity in vanilla beans wasfound to be high during curing [30]. Accord-ing to Arana, vanillin formed during sweat-ing and drying may be further oxidized toproduce quinone compounds with morecomplex structures and different aroma.This process occurs slowly during condi-tioning and this could explain the delayeddevelopment of the characteristic vanillaaroma in some treatments. Peroxidase maybe responsible for further oxidation to pro-

duce different aroma compounds and, thus,partially contribute to the overall aroma ofvanilla beans [31].

Our research showed that beansblanched at 67 °C for 3 min and sweated at45 °C for 4 d had relatively low vanillin con-centrations but high concentrations of aro-matic odors. The relatively large proportionof glucovanillin remaining in these beansafter curing may have been due to rapidinactivation of β-glucosidase or to retentionof some cellular compartmentation that keptsome glucovanillin separate from β-glucosi-dase. The differential effects of blanchingand sweating at 45 °C for 4 d on the pro-duction of aroma volatiles may be explainedby the retention of the activity of enzymesother than glucosidases such as peroxidases(22, 31, 32).

Figure 8.Principal component analysisof vanillin measurement andsensory attributes given byuntrained panelists regardingsamples of vanilla beans curedusing nine different processes.HPD: heat pump dryer, TDH:tunnel dryer with high RH, TDL:tunnel dryer with low RH.

396 Fruits, vol. 65 (6)

S. van Dyk et al.

Tab

leII.

Cor

rela

tion

coef

ficie

nts

(r)f

rom

the

sens

ory

eval

uatio

nof

cure

dva

nilla

bea

ns.V

alue

s>

0.50

are

the

mos

tim

por

tant

corr

elat

ions

.

Sen

sory

eval

uatio

nO

vera

lllik

ing

vani

llafr

agra

nce

Ove

rall

likin

gva

nilla

frag

ranc

est

reng

th

Sw

eet

frag

ranc

est

reng

th

Frui

tfr

agra

nce

stre

ngth

Flor

alfr

agra

nce

stre

ngth

Str

awfr

agra

nce

stre

ngth

Sp

icy

frag

ranc

est

reng

th

Ferm

ente

dfr

agra

nce

stre

ngth

Aci

dic

frag

ranc

est

reng

th

Qua

lity

frag

ranc

efo

rco

okin

g

Qua

lity

for

frag

ranc

e

Ove

rall

like

the

frag

ranc

ep

erce

ived

Ove

rall

likin

gva

nilla

frag

ranc

e_

Ove

rall

likin

gva

nilla

frag

ranc

est

reng

th

0.55

9_

Sw

eet

frag

ranc

est

reng

th0.

341

0.48

5_

Frui

tfr

agra

nce

stre

ngth

0.08

30.

136

0.43

8_

Flor

alfr

agra

nce

stre

ngth

-0.0

610.

121

0.11

40.

285

_

Str

awfr

agra

nce

stre

ngth

0.00

60.

156

0.13

3-0

.040

0.33

3_

Sp

icy

frag

ranc

est

reng

th0.

258

0.33

50.

571

0.30

00.

353

0.47

0_

Ferm

ente

dfr

agra

nce

stre

ngth

-0.2

020.

100

0.25

30.

129

0.14

80.

370

0.46

8_

Aci

dic

frag

ranc

est

reng

th0.

170

0.16

50.

515

0.21

6-0

.070

0.48

80.

764

0.46

0_

Qua

lity

frag

ranc

efo

rco

okin

g0.

587

0.44

70.

405

0.22

20.

136

-0.1

120.

439

0.09

30.

289

_

Qua

lity

for

frag

ranc

e0.

634

0.46

10.

328

0.19

50.

049

-0.0

610.

289

-0.1

020.

107

0.67

8_

Ove

rall

like

the

frag

ranc

ep

erce

ived

0.73

00.

403

0.38

70.

224

0.03

5-0

.074

0.28

0-0

.101

0.03

10.

748

0.70

4_

Curing procedures of vanilla beans

Fruits, vol. 65 (6) 397

This research confirmed that curing ofvanilla beans is a complex process thatinvolves many more variables than simplythe hydrolysis of glucovanillin. About 90%of glucovanillin was converted to vanillin inbeans cured continuously (non-blanched)at 35 °C for 12 d, much higher than in beansthat were blanched at 67 °C for 3 min andthen sweated. The lower percent conver-sion of glucovanillin to vanillin in beansblanched at 67 °C for 3 min and sweated at45 °C for 4 d or at 35 °C for 5 d suggests thatthe blanching treatment drastically reducedβ-glucosidase activity [14, 18, 22, 32]. Dryingmethods may also affect vanillin levels bydirect effects on enzyme activities and bysublimation of vanillin. Overall, tunnel dry-ing at 60 °C at either 20% or 10% RH waspreferred because the rate of drying wasmuch faster and the moisture content of thedried beans was uniform. Since sensoryanalysis showed that the flavor of beanssweated continuously at 35 °C was betterbut appearance was inferior to beans thatwere blanched then sweated, further workis required to develop a mild blanchingtreatment that disrupts the tissue enough tofacilitate mixing of glucovanillin and glu-cosidases without denaturing glucosidasesand other enzymes involved in flavor for-mation. Nevertheless, the quality of curedvanilla beans will be higher if the curingstarts with mature beans of consistent sizethat have the highest glucovanillin concen-tration and the highest β-glucosidase activ-ity. The data presented in this paper showedthat higher concentrations of vanillin can bemaintained in cured dried vanilla beans thanthose reported in previous publications.

Acknowledgements

We gratefully acknowledge Daintree Vanilla& Spice, Cairns, Queensland, and Dr.George Srzednicki, University of New SouthWales, Department of Food Science andTechnology, for conducting the drying treat-ments. Sensory Solutions Ltd. advised us onthe design of the sensory evaluations andconducted these evaluations in their labo-ratory, and Mr. Oleg Nicetic, Center forPlants and the Environment, UWS, provided

advice on the statistical analysis of the data.We also thank Australian Vanilla Plantationsand the UWS Hawkesbury Trust for fundingthis project.

References

[1] Childers N.F., Cibes H.R., Hernandez-Medina E., Vanilla – the orchid of commerce,in: Withner C.L., Robert E. (Eds.), TheOrchids, A scientific survey, Krieger Publ.Co., Malabar, Fla., U.S.A., 1959, 477–508.

[2] Kanisawa T., Flavor development in vanillabeans, Kouryou 180 (1993) 113–123.

[3] Dignum M.J.W., Van Der Heijden R., KerlerJ., Winkel C., Verpoorte R., Identification ofglucosides in green beans of Vanilla planifo-lia Andrews and kinetics of vanilla β-glucosi-dase, Food Chem. 85 (2004) 199–205.

[4] Havkin-Frenkel D., Podstolski A., WitkowskaE., Molecki, P., Mikolajczyk M., Vanillin bio-synthetic pathways, an overview, in: Fu T.J.,Singh G., Curtis W.R., (Eds.), Plant cell andtissue culture for the production of foodingredients, Kluwer Acad. Press / PlenumPubl., N.Y., U.S.A., 1999, 35–43.

[5] Kanisawa T., Tokoro k., Kawahara S., Flavordevelopment in the beans of Vanilla planifo-lia Andrews, in: Kurihara K., Suzuki N.,Ogawa H. (Eds.), Olfaction taste XI, Proc. Int.Symp. Tokyo, Springer, Japan, 1994,pp. 268–270.

[6] Ranadive A.S., Szkutnica K., Guerrera J.G.,Frenkel C., Vanillin biosynthesis in vanillabeans, in: Proc. IX Int. Congr. Essent. Oils,Singap., 1983, pp. 147–154.

[7] Havkin-Frenkel D., French J.C., Pak F.E.,Frenkel C., Inside vanilla: Vanilla planifolia'sbotany, curing options and future marketprospects, Perfum. Flavour. 30 (2005) 36–55.

[8] Joel D.M., French J.C., Graft N., Korteva G.,Dixon R.A., Havkin-Frenkel D., A hairy tissueproduces vanillin, Isr. J. Plant Sci. 51 (2003)157–159.

[9] Odoux E., Escoute J., Verdeil J.L., BrillouetJ.M., Localization of β-D-glucosidase acti-vity and glucovanillin in vanilla bean, Ann.Bot. 92 (2003) 437–444.

[10] Odoux E., Brillouet J.M., Anatomy, histoche-mistry and biochemistry of glucovanillin,oleoresin and mucilage accumulation sites in

398 Fruits, vol. 65 (6)

S. van Dyk et al.

green mature vanilla pod (Vanilla planifolia;Orchidaceae): a comprehensive and criticalreexamination, Fruits 64 (2009) 221–241.

[11] Havkin-Frenkel D., French J.C., Graft N.M.,Pak F.E., Frenkel C., Joel D.M., Interrelationof curing and botany in vanilla (Vanilla plani-folia) bean, Acta Hortic. 629 (2004) 93–102.

[12] MariezcurrenaM.D.,ZavaletaH.,WaliszewskiK.N., Sanchez V., The effect of killing condi-tions on the structural changes in vanilla(Vanilla planifolia, Andrew) pods during thecuring process, Int. J. Food Sci. Technol. 43(2008) 1452–1457.

[13] Purseglove J.W., Brown E.G., Green C.L.,Robins S.R.J., Spices, in: Tropical agricultu-ral, Longman Inc., N.Y., U.S.A., 1981,pp. 644–735.

[14] Havkin-Frenkel D., Frenkel C., Postharvesthandling and storage of cured vanilla beans,Stewart Postharvest Rev. 4 (2006) 1–9.

[15] Jones M.A., Vicente G.C., Criteria for testingvanilla in relation to killing and curingmethods, Agric. Res. 78 (1948) 425–434.

[16] Theodose R., Traditional methods of vanillapreparation and their improvement, Tropic.Sci. 15 (1973) 47–57.

[17] Dignum M.J.W., Kerler J., Verpoorte R.,Vanilla curing under laboratory conditions,Food Chem. 79 (2002) 165–171.

[18] Odoux E., Changes in vanillin and glucova-nillin concentrations during the variousstages of the process traditionally used forcuring of Vanilla fragrans beans in Réunion,Fruits 55 (2000) 119–125.

[19] Dignum M.J.W., Kerler J., Verpoorte R.,Vanilla production: technological chemicaland biosynthetic aspects, Food Rev. Int. 17(2001) 199–219.

[20] Wild-Altamirano C., Enzymatic activityduring growth of vanilla fruit. 1. Proteinase,glucosidase, peroxidase and polyphenoloxi-dase, J. Food Sci. 34 (1969) 235–238.

[21] Hanum T., Changes in vanillin and activity ofβ-glucosidase and oxidases during posthar-vest processing of vanilla beans (Vanilla pla-nifolia), Bull. Teknol. Ind. Pangan 8 (1997)46–52.

[22] Marquez O., Waliszewski K.N., The effect ofthermal treatment on β-glucosidase inacti-vation in vanilla bean (Vanilla planifolia

Andrew ), Int. J. Food Sci. Technol. 43 (2008)1993–1999.

[23] Odoux E., Chauwin A., Brillouet J.-M., Purifi-cation and characterization of vanilla bean(Vanilla planifolia Andrews) ß–D-glucosidase,J. Agric. Food Chem. 51 (2003) 3168–3173.

[24] Odoux E., Escoute J., Verdeil J.L., The rela-tion between glucovanillin, ß-D-glucosidaseactivity and cellular compartmentationduring the senescence, freezing and traditio-nal curing of vanilla beans, Ann. Appl. Biol.149 (2006) 43–52.

[25] Cervantes L., Azuara E., Beristain, C.I., Animprovement in the curing process of vanillabeans, Food Chem. Conf.: Chemical effectsof food processing and preservation, LasVegas, U.S.A., 2004.

[26] Perez-Silva A., Osouz E., Brat P., Riveyre F.,Rodriguez-Jimenes G., Robles-Olvera V.,Garcia-Alvarado M.A., Gunata Z., GC-MSand GC-olfactometry analysis of aromacompounds in a representative organicaroma extract from cured vanilla (Vanilla pla-nifolia G. Jackson) beans, Food Chem. 99(2006) 728–735.

[27] Adedeji J., Hartman T.G., Ho C.-T., Flavourcharacterization of different varieties ofvanilla beans, Perfum. Flavour. 18 (1993)25–33.

[28] Arana F.E., Vanilla curing and its chemistry,USDA Fed. Exp. Stn., Mayaguez, USDA, P.R.Bull. No. 42, Wash. D.C., U.S.A., 1944.

[29] Rivera J.G., Hageman R.H., Vanilla curing,USDA Rep. Fed. Exp. Stn., Mayaguez,USDA, P.R., 1950.

[30] Sreedhar R.V., Roohie K., Venkatachalam L.,Narayan M.S., Bhagyalakshmi N., Specificpre-treatments reduce curing period ofVanilla (Vanilla planifolia) beans, J. Agric.Food Chem. 55 (2007) 2947–2955.

[31] Gatfield I., Reib I., Krammer G., SchmidtC.O., Kindel G., Bertram H.J., Divanillin:novel taste-active component of fermentedvanilla beans, Perfum. Flavour. 1 (2006)18–20.

[32] Dignum M.J.W., Kerler J., Verpoorte R., ß-glucosidase and peroxidase stability incrude enzyme-extracts from green beans ofVanilla planifolia Andrews, Phytochem. Anal.12 (2001) 174–179.

Curing procedures of vanilla beans

Fruits, vol. 65 (6) 399

Influencia de los métodos de preparación de las vainas de vainilla en sucalidad sensorial.

Resumen — Introducción. A lo largo de la maduración, la vainillina se acumula en las vai-nas de vainilla verde en forma de glucovainillina, y es hidrolizada en vainillina mediante glu-cosidasas endógenas durante el proceso de preparación y aporta el sabor característico de lavainilla. El objetivo de nuestro trabajo fue el estudio de métodos de preparación de la vainillaque podrían reducir considerablemente el tiempo del desarrollo del proceso y producir vai-nas con concentraciones fuertes, así como otros compuestos de aroma con altas calidadessensoriales. Material y métodos. Se obtuvieron vainas de vainilla maduras por un productorcomercial (Cairns, Queensland, Australia). Un lote de vainas fue continuamente incubado a35 °C y con una humedad relativa (HR) elevada durante 12 días. Otros dos lotes se blan-quearon en agua a 67 °C durante 3 min, a continuación fueron incubados a 45 °C con unaHR elevada durante 4 d, o a 35 °C durante 5 d. Las vainas fueron incubadas hasta quedarbronceadas. Se evaluaron tres métodos de secado: con ayuda de una bomba de calor seco a40 °C y 15% HR, de un túnel de secado a 60 °C y 20% HR, y de un túnel de secado a 60 °C y10% HR. Se extrajo la vainillina con n-pentano y diclorometano (1:1 v/v) y se dosificó porHPLC, a partir de muestras de polvo obtenido de las vainas tratadas. Se midió la glucovainil-lina por el índice de vainillina total obtenida tras la hidrólisis ácida de las muestras de polvo.Resultados y discusión. Cerca del 90% de la glucovainillina se convirtió en vainillina en lasvainas no blanqueadas e incubadas continuamente a 35 °C, mientras que en las vainas blan-queadas a 67 °C durante 3 min e incubadas a 45 °C durante 4 d o a 35 °C durante 5 d perosólo hubo un 70% de conversión. Se evaluó la calidad sensorial de las vainas tratadas por ungrupo especial inexperimentado. Su análisis mostró que las vainas incubadas continuamentea 35 °C poseían un aroma superior a aquél de las vainas blanqueadas en agua caliente e incu-badas a 45 °C o a 35 °C, pero la apariencia de las vainas no blanqueadas era menosatrayente. Conclusión. El estudio mostró que un tratamiento moderado de blanqueamientocon agua caliente, seguido por una incubación a 35–45 °C y un secado rápido era necesariopara obtener vainas preparadas con un muy buen aspecto y un aroma interesante.

Australia / Vanilla planifolia / vainilla / vanillina / procesamiento / tecnología dealimetos / secado por aire caliente / control de calidad / análisis organoléptico