MASS TRANSFER KINETICS AND EFFECTIVE DIFFUSIVITIES DURING COCOA ROASTING

Transcription

1 Journal of Enginring Scinc and Tchnology Vol. 12, No. 1 (2017) School of Enginring, Taylor s Univrsity MASS TRANSFER KINETICS AND EFFECTIVE DIFFUSIVITIES DURING COCOA ROASTING Y. M. BAGHDADI, C. L. HII Cntr for Food and Bio-product Procssing, Manufacturing and Industrial Procsss Rsarch Division, Faculty of Enginring, Univrsity of Nottingham, Malaysia Campus, Jalan Broga, Smnyih, Slangor DE, Malaysia *Corrsponding Author: Ching-Lik.Hii@nottingham.du.my Abstract Th currnt studis invstigatd th ffcts of tmpratur and moistur addition on th mass transfr kintics of cocoa nibs during roasting. Exprimnts wr carrid out by roasting 500 gm of cocoa nibs insid an air vntilatd ovn at thr tmpratur lvls (120 C, 140 C and 160 C) undr mdium air flowrat for on hour. Two typs of sampls wr prpard namly th raw and soakd nib sampls. Th soakd nib sampls wr prpard by soaking th raw nibs in 200 ml of watr at room tmpratur for 5 and 10 hours. Mathmatical modlling was carrid out to modl th mass transfr procss using smi-mpirical modls. Modlling showd that both Pag and two-trm modls wr abl to giv clos fitting btwn th xprimntal and prdictd valus. Effctiv diffusivity valus wr stimatd in th ordr of magnitud of 10-5 m 2 /s for th mass transfr procss. Rsults obtaind from ths studis fill th currnt knowldg gap on th mass transfr kintics of cocoa roasting. Kywords: Cocoa, Diffusivity, Mass Transfr, Modlling, Roasting. 1. Introduction Th procssing of cocoa bans (Thobroma cacao) bgins with frmntation and drying in ordr to dvlop th ncssary flavour prcursors for aroma dvlopmnt during roasting [1]. Prior to roasting, th ban is brokn and th shll is rmovd to obtain only th nib (or known as drid cocoa cotyldon) using a winnowr and brakr machin. Th cocoa nibs ar usually roastd insid a batch drum roastr (tmpratur C) or somtims but not widly usd, insid a continuous roastr. Th duration of roasting is typically minuts in industrial opration dpnding on th rcip and th dgr of roasting. 127

2 128 Y. M. Baghdadi and C.L. Hii Nomnclaturs a, b, g, k, n Constants in smi-mpirical modl D Effctiv diffusivity, m 2 /s D o Pr-xponntial factor, m 2 /s E Activation nrgy, kj/mol L Half thicknss, m MR Moistur ratio, dimnsionlss N Numbr of data R Univrsal gas constant, J/mol.K RMSE Root man squar rror SD Standard dviation T Tmpratur, K t Tim, s Moistur contnt, g H 2 O/g dry solid z Numbr of constants Subscripts Exp o Pr Exprimntal Equilibrium Initial Prdiction Upon roasting, th roastd nibs ar ground into past form insid a grinding machin. Th past is commonly known as cocoa liquor and this is on of th major ingrdints usd in chocolat manufacturing. Th cocoa liquor can b furthr procssd into cocoa powdr and cocoa buttr which ar also usd in chocolat making [1]. Th flavour of cocoa liquor and powdr is gratly affctd by roasting but not for cocoa buttr as it is usually dodorizd upon buttr prssing. Roasting is a simultanous hat and mass transfr oprations whrby hat is supplid to nabl moistur diffusion and vaporation from th cocoa nibs. Th amount of moistur prsnt in th nibs will affct th rat of mass (moistur) transfr and hnc th procssing tim; which is crucial as this will affct not only th colour but also th flavour of th roastd nibs. Typically, th roastd nibs hav intns chocolaty aroma and th moistur contnt should b lss than 2 % (d.b) bfor grinding. In th past, various studis hav bn rportd for ffct of roasting on nibs colouration [2, 3]; physical/chmical changs [4-6] and flavour changs [7-9] Howvr, rportd studis on th ffct of roasting on th mass transfr kintics ar rlativly scarc. Schrodr t al. [10] had rportd th changs in moistur contnts of cocoa bans and nibs during convctiv and microwav assistd roasting but without furthr attmpt to modl th mass transfr kintics. In addition, ffctiv diffusivity is a transport proprty which is important in th dsign of thrmal hating quipmnts such as dryr and roastr. Typically, ffctiv diffusivity is a strong function of tmpratur which govrns th progrss of moistur diffusion upon hating [11]. Th diffusion procss also govrns th vaporation of flavour volatils and th dvlopmnt of flavour compounds, i.., in Millard raction [1]. Thrfor, this proprty affcts gratly

3 Mass Transfr Kintics and Effctiv Diffusivitis during Cocoa Roasting 129 th procssing tim rquir to complt a roasting procss which is vital not only in quipmnt dsign but also in aroma dvlopmnt Hnc, th objctivs of th rsarch wr to invstigat th ffcts of tmpratur and moistur addition on th mass transfr kintics of cocoa nibs during roasting. Th purpos of moistur addition was to imitat th conditions of th nibs which ar vry moist and partly rhydratd aftr th alkalizing procss bfor roasting [1]. Rsults from ths studis fill th currnt knowldg gap on th mass transfr kintics of cocoa roasting which ar scarcly rportd in publishd litraturs. 2. Matrials and Mthods 2.1. Roasting Cocoa nibs (also known as drid cocoa cotyldons) wr obtaind from Malaysian Cocoa Manufacturing Sdn. Bhd. (Srmban, Malaysia). About 500 gm sampls wr prpard and placd on a wir mshd tray in thin layr (~1 cm nibs thicknss) prior to roasting. Roasting was carrid out by placing th tray in th middl sction of an air vntilatd ovn (Mmmrt, Grmany) with airflow adjustd to mdium stting. Two typs of sampls wr prpard namly th raw and soakd nib sampls (Tabl 1). Tabl 1. Roasting trials conductd for raw and soakd nib sampls. Sampl Tmpratur Raw nibs 120,140 and 160 C Soakd nibs (0h, 5h, 10h) 140 C Th soakd sampls wr prpard by soaking th raw nibs in 200 ml of watr at room tmpratur for 5 and 10 hours. This is to imitat th partly rhydratd and moist conditions of th nibs aftr alkalization. In a typical procssing plant, alkalization is carrid out using alkalin solutions in ordr to achiv diffrnt dgr of colourization undr prssurizd condition. It is thrfor proposd that this can b furthr invstigatd in futur rsarch. During roasting, th wights of th nibs wr masurd vry 10 minuts using an analytical balanc (Mtlr Toldo, USA) for fixd roasting duration of 60 minuts. Moistur contnts of th nibs () wr dtrmind by using Eq. (1) with rfrnc to th bon dry wight [12]. Th bon dry sampls wr prpard by placing th final roastd nibs in an ovn ovrnight at 105 C for 24 hour. All roasting trails wr conductd in duplicat. Wt.of sampl(g) Wt.of bon dry sampl(g) (1) Wt.of bon drysampl(g) 2.2. Roasting rats Roasting rats wr calculatd basd on th moistur contnt data according to Eqs. (2) to (4) [13]:

4 130 Y. M. Baghdadi and C.L. Hii d dt d dt At t = t 0 (initial tim), t 1 0 (2) 1 t 0 At t = t i (intrmdiat points) t i 1 i-1 (3) i1 t i-1 whr i = 1,., N-1 d dt At t = t N (final tim), t N N-1 (4) N t N Mathmatical modlling Th mass transfr kintics was analyzd by fitting thr smi-mpirical modls to th xprimntal data as shown in Tabl 2. Statistical paramtrs such as R 2, chisquar, Eq. (5) and root man squar rror (RMSE, Eq. (6)) wr usd to dtrmin th goodnss of fitting [12]. SOLVER was usd (MS Excl, USA) for th rgrssion analyss. Modl nam Nwton Pag Two-trm modl Tabl 2. Smi-mpirical modls. Equation MR xp MR xp kt n kt MR axp kt bxp whr a, b, k, g and n ar constants to b dtrmind from rgrssion. N 2 (MRpr, i MRxp, i ) i 1 Chi squar (5) N z gt RMSE 1 N N i1 (MR 2 pr, i MRxp, i ) 1/2 (6) whr N = numbr of data, z = numbr of constants in modl Effctiv diffusivitis This was dtrmind by using th analytical solution of Fick s scond law [11] as shown in Eq. (7). Th shap of th nibs was assumd in a slab form with avrag thicknss (2L) of about 3.66 mm. Fiv trms of th solution (n = 5) wr usd in th stimation.

5 Mass Transfr Kintics and Effctiv Diffusivitis during Cocoa Roasting n 1 (2n 1) π Dt MR xp[ (7) 2 n1 2 2 π (2n 1) 4L 0 whr th trm is known as th moistur ratio (MR). 0 Th Arrhnius quation was usd to rlat th ffctiv diffusivitis with tmpraturs for th dtrmination of activation nrgy (E) in Eq. (8). o E RT D D xp (8) Th activation nrgy was dtrmind by taking natural logarithm at both sids, and th slop of th linar graph (ln D vs. 1/T) was usd to dtrmin this valu Txtur analyss Txtur analyss wr carrid out by dtrmining th hardnss of th nibs bfor and aftr soaking. Th tst was conductd by using txtur analysr (Stabl Microsystms TA.T TEE32, UK) fittd with a circular comprssion prob (Diamtr = 7.5 cm). About 10 g of nibs wr placd on a solid mtal bas and comprssion tst was carrid out using th prob at tst spd of 0.5 mm/s and comprssion distanc of 2 mm. Hardnss was dtrmind as th maximum forc rgistrd from th forc dformation curv (forc vs tim). This masurmnt was prformd in triplicat Statistical analyss Statistical analyss wr carrid out using on-way ANOVA and Duncan s Multipl Rang Tst for man comparison at 95% confidnc lvl. SAS for Window softwar was usd (SAS Institut, USA). 3. Rsults and Discussion 3.1. Mass transfr kintics Figur 1 shows th mass transfr kintics of th cocoa nibs basd on th moistur ratio profil. It can b obsrvd that th kintics follows xponntial dcay and only falling rats ar obsrvd with rgards to moistur transfr. This is in agrmnt with th roasting rats (Fig. 2) dtrmind from th moistur contnt data, whr no constant rat can b obsrvd. This shows that th mass transfr procss is dominatd by intrnal diffusion [14-16]. Howvr, it is intrsting to not that in ovrall th drop in moistur ratios for th soakd sampls (5 h and 10 h) ar fastr than th raw sampls (Fig. 1b). This could b du to th slightly xpandd/softn structur within th solid matrix upon soaking that improvs diffusion, in addition to th gratr moistur gradint as th main driving forc for mass transfr. This is in agrmnt with th hardnss tst carrid out whr th longr is th soaking duration; th softr is th nibs upon soaking (Fig. 3). It can b sn that both soakd sampls ar significantly softr than th raw sampls (p < 0.05).

6 132 Y. M. Baghdadi and C.L. Hii In addition, comparison of initial drying rats show significant diffrnt (p < 0.05) btwn th soakd and raw sampls (Fig. 4). Although no significant diffrnt (p > 0.05) can b obsrvd btwn th soakd sampls (5 h and 10 h) but thr is a tndncy for th 5 h soakd sampls to dry fastr as compard to th 10 h soakd sampls which could b du to th lssr amount of watr rhydratd into th nibs upon soaking. a. Raw sampls b. Soakd sampls Fig. 1. Moistur ratio profils for raw and soakd nib sampls (solid lin: modl prdiction; markr: xprimntal data).

7 Mass Transfr Kintics and Effctiv Diffusivitis during Cocoa Roasting 133 a. Raw sampls b. Soakd sampls Fig. 2. Roasting rats profils for raw and soakd nib sampls. Th radings ar xprssd as mans ± SD Fig. 3. Comparison of hardnss btwn th raw and soakd nib sampls.

8 134 Y. M. Baghdadi and C.L. Hii Th radings ar xprssd as mans ± SD Fig. 4. Comparison of initial drying rats btwn th raw and soakd nib sampls Mathmatical modlling Tabl 3 shows that both Pag and two-trm modls can b usd to dscrib th xprimntal data during roasting dpnding on th xprimntal tratmnt (also as shown in Fig. 1). Ths sampls wr slctd basd on th modl that gav th highst R 2, lowst chi-squar and RMSE valus. Smi-mpirical modls hav bn usd widly in modlling of moistur transfr procss during drying [14]. In this xprimnt, similar modls hav bn usd as moistur vapour is diffusd and vaporatd from th cocoa nibs during roasting undr high tmpraturs Effctiv diffusivity Tabl 4 shows th ffctiv diffusivity valus obtaind basd on th analytical solutions of Fick s law [11]. It can b sn that diffusivity stimatd from th soakd sampls (5h and 10 h) ar highr as compard to th raw sampls. Th diffusivity valus stimatd ar gnrally in th ordr of magnitud of 10-5 m 2 /s, which ar much highr than thos obtaind from convntional drying at m 2 /s [17]. This could b du to th much highr tmpratur usd in roasting (> 100 C) as compard to hot air drying (60 C - 80 C). Activation nrgy to initiat moistur diffusion during roasting was dtrmind at 28.6 kj/mol, Eq. (9). This valu is much lowr than that dtrmind for cocoa drying (45 kj/mol) [18] du to th highr tmpratur usd which corrsponds to gratr driving forc for mass transfr. In addition to that, roasting was carrid out using nib fragmnts instad of using whol bans as in th cas of cocoa drying. Hating rat is thrfor fastr for th nibs du to th smallr particl siz and absnc of th shll layr. D RT xp (9)

9 Mass Transfr Kintics and Effctiv Diffusivitis during Cocoa Roasting 135 Nwton: k = Chi-squar = RMSE = R 2 = Tabl 3. Statistical rsults of modlling. Effcts of tmpratur 120 C C Nwton: k = Nwton: k = Chi-squar = Chi-squar = RMSE = RMSE = R 2 = R 2 = Pag: k = , n = Chi-squar = RMSE = R 2 = Two-trm: a = 0.17, b = 0.83, k = 1.105, g = Chi-squar = RMSE = R 2 = Nwton: k = Chi-squar = RMSE = R 2 = Pag: k = , n = Chi-squar = RMSE = R 2 = Two-trm: a = , b = , k = , g = Chi-squar = RMSE = R 2 = Pag: k = , n = Chi-squar = RMSE = R 2 = Two-trm: a = , b = , k = , g = Chi-squar = RMSE = R 2 = Effcts of moistur addition (T = 140 C) 0 h 5 h 10 h Nwton: k = Chi-squar = RMSE = R 2 = R 2 = Pag: k = , n = Chi-squar = RMSE = R 2 = Two-trm: a = , b = , k = , g = Chi-squar = , RMSE = R 2 = Pag: k = , n = Chi-squar = RMSE = R 2 = Two-trm: a = , b = , k = , g = Chi-squar = , RMSE = R 2 = Tabl 4. Effctiv diffusivitis during roasting. Nwton: k = Chi-squar = RMSE = Pag: k = , = Chi-squar = RMSE = R 2 = Two-trm: a = , b = , k = , g = Chi-squar = RMSE = R 2 = Effcts of tmpratur 120 C C m 2 /s m 2 /s m 2 /s Effcts of moistur addition (T = 140 C) 0 h 5 h 10 h m 2 /s m 2 /s m 2 /s n 4. Conclusion Highr roasting rats wr obsrvd in th soakd cocoa nibs as compard to th raw nibs which could b du to th softn structur of th nibs and gratr moistur gradint. Only falling rats wr obsrvd for moistur transfr during

10 136 Y. M. Baghdadi and C.L. Hii roasting. Both Pag and two-trm modls can b usd to dscrib th roasting kintics for th raw and soakd nib sampls. Effctiv diffusivity valus wr dtrmind in th ordr of 10-5 m 2 /s whil activation nrgy was stimatd at 28.6 kj/mol during roasting. Rsults from th studis hav providd an insight into th mass transfr kintics of cocoa nibs during roasting which ar bnficial to th cocoa industry spcially to undrstand furthr th intraction btwn moistur diffusion and aroma dvlopmnt in futur rsarch. Rfrncs 1. Bcktt, S.T. (1994). Industrial chocolat manufactur and us. (2 nd Ed.). Glasgow: Chapman and Hall. 2. Wisława, K. (2006). Influnc of roasting conditions on coloration of roastd cocoa bans. Journal of Food Enginring, 77(3), Wahidu, Z; and Tajul A.Y. (2013). Effct of suprhatd stam and convction roasting on changs in physical proprtis of cocoa ban (Thobroma cacao), Food Scinc and Tchnology Rsarch, 19(2), Jinap, S.; Wan Rosli, W.I.; Russly, A.R.; and Nordin, L.M. (1998). Effct of roasting tim and tmpratur on volatil componnt profils during nib roasting of cocoa bans (Thobroma cacao). Journal of th Scinc of Food and Agricultur, 77(4), Millr, K.B.; Hurst, W.J.; Payn, M.J.; Stuart, D.A.; Apgar, J.; Swigart, D.S.; and Ou, B. (2008). Impact of alkalization on th antioxidant and flavanol contnt of commrcial cocoa powdrs. Journal of Agricultural and Food Chmistry, 56(18), Ioannon, F.; Di Mattia, C.D.; D Grgorio, M.; Srgi, M.; Srafini, M.; and Sacchtti, G. (2015). Flavanols, proanthocyanidins and antioxidant activity changs during cocoa (Thobroma cacao L.) roasting as affctd by tmpratur and tim of procssing. Food Chmistry, 174, Misnawi; Jinap, S.; Jamilah, B.; and Nazamid, S. (2004). Snsory proprtis of cocoa liquor as affctd by polyphnol concntration and duration of roasting. Food Quality and Prfrnc, 15(5), Nazaruddin, R.; Osman, H.; Mamot, S.; Wahid, S.; and Nor Aini, I. (2006). Influnc of roasting conditions on volatil flavor of roastd Malaysian cocoa bans, Journal of Food Procssing and Prsrvation, 30(3), Yu, L; Yun, F; Song, Z; Changrong, L; Jianguo, M; and Fang, Z. (2012). Th ffct of alkalization on th bioactiv and flavor rlatd componnts in commrcial cocoa powdr. Journal of Food Composition and Analysis, 25(1), Schrödr, J.; Gaukl, V.; and Schuchmann, H.P. (2013). Influnc of th tim-tmpratur-profil during convctiv and microwav assistd roasting on physical proprtis of cocoa. Procdings of Eurodrying 2013, Paris, 2-4 Octobr. 11. Crank, J. (1975). Th mathmatics of diffusion (2 nd d.). Oxford: Clarndon Prss. 12. Hii, C.L.; Law, C.L.; and Clok, M. (2008). Modlling of thin layr drying kintics of cocoa bans during artificial and natural drying. Journal of Enginring Scinc and Tchnology (JESTEC), 3(1), 1-10.

11 Mass Transfr Kintics and Effctiv Diffusivitis during Cocoa Roasting Guiné, R.P.F. and Frnands, R.M.C. (2006). Analysis of th drying kintics of chstnuts. Journal of Food Enginring, 76(3), Jayas, D.S.; Cnkowski, S.; Pabis, S.; and Muir, W.E. (1991). Rviw of Thin Layr Drying and Wtting Equations. Drying Tchnology, 9(3), Kulshrshtha, M.; Singh, A.; and Vipul, D.A. (2009). Effct of drying conditions on mushroom quality. Journal of Enginring Scinc and Tchnology (JESTEC), 4(1), Kuy, A.; Oko, C.O.C; and Nnamchi, S.N. (2007) Simulation of th drying charactristics of ground nm sds in a fluidisd bd. Journal of Enginring Scinc and Tchnology (JESTEC), 2(1), Zogzas, N.P.; Maroulis, Z.B.; and Marinos-Kouris, D. (1996). Moistur diffusivity data compilation in foodstuffs. Drying Tchnology, 14(10), Hii, C.L.; Law, C.L.; and Clok, M. (2009). Modling using a nw thin layr drying modl and product quality of cocoa. Journal of Food Enginring, 90(2),