A NOVEL OPTIMIZED ENERGY-SAVING EXTRACTION PROCESS ON COFFEE

Wag, C.-C., et al.: A Novel Optimized Eergy-Savig Extractio Process o Coffee S53 A NOVEL OPTIMIZED ENERGY-SAVING EXTRACTION PROCESS ON COFFEE by Cheg-Chi WANG a*, Shae-Rog SHEU b, Ya-Ye CHOU c, Mig-Jyi JANG b a, ad Li-Che YANG a Departmet of Mechaical Egieerig, Far East Uiversity, Taia, Taiwa b Departmet of Automatio ad Cotrol Egieerig, Far East Uiversity, Taia, Taiwa c Departmet of Marketig ad Logistics Maagemet, Far East Uiversity, Taia, Taiwa Origial scietific paper UDC: 66.061.3:547.857.4:66.084.8 DOI: 10.2298/TSCI11S1053W I this paper, Taguchi method is applied to optimize ultrasoud thermal process for extractig caffeie ad flavor from coffee. The use of ultrasoud ca abridge experimets i cost, eergy loss ad time; the differet operatig coditios for extractio experimets are executed ad the results are also compared. The results show that the best desig factors for caffeie are 95 of extractio temperature, 28 khz of operatig frequecy ad 30 s of extractio time. The proposed optimized extractio method is efficiet ad eergy-savig compared with the geeral process for makig coffee. Key words: eergy-savig, Taguchi, ultrasoud, coffee Itroductio Traditioal experimetal desig focused o the desig of the factors; as log as the cotrol factor icreases, however, it ofte leads to the icremet of the umbers of the experimets as well as the complexity of the process, however, lots of experimetal factors ca be simplified or removed by The Taguchi method, which is applied to decrease the umbers of experimets by usig orthogoal table: firstly, select the experimetal variables, secodly, determie the level of each variable, ad fially, solve the aalysis of variace [1]. It focuses o how to decrease the variability of experimetal results, so as to fid out the best combiatio of parameters i accordace with the beeficial results of cost [2, 3]. Ultrasoic extractio is the way of usig the powerful eergy which is produced by the burst of the bubbles-produced by the chage of the pressure whe ultrasoud is i the liquid-to accelerate the cotact speed of the solvet ad the extract. The mai feature of the approach is that it ca rapidly ad evely bled the extract ad the solvet; i the process of extractig useful costituets from the atural substaces, it ca brig the extract from the matrix to the solvet without damagig the structure of the extract [4, 5]. *Correspodig author; e-mail: wccpip@yahoo.com.tw

S54 Wag, C.-C., et al.: A Novel Optimized Eergy-Savig Extractio Process o Coffee The igrediets of coffee iclude caffeie, aroma, proteis, taic acid, fat, etc.. Whe used i moderatio, caffeie ca stimulate the brai, ehace memory, ispire ethusiasm, clarify thought, ad reduce fatigue ad sleepiess. However, excessive amouts of caffeie would result i axiety, ueasiess, heart palpitatios, headaches, diarrhea, isomia ad may more. Oe will fid the accouts o the fact that caffeie ca have detrimetal effects o huma health i literature; these egative effects iclude exacerbatig coroary artery disease, raisig high blood pressure, ad heighteig the higher risk of heart attack ad kidey disease [6, 7]. This paper applies Taguchi method to cosider the ew extractio equipmet that is desiged to optimize the cosistecy of the coffee powder ad extractio liquid to allow for a highly efficiet extractio process. Experimetal method This experimet uses a ultrasoic equipmet to extract caffeie ad coffee flavor from coffee. By cotrollig the temperature, wave frequecy, ad time duratio of each extractio, these variable effects o the caffeie cocetratio of each extractio ca be measured as well as a comparative aalysis of the extractio liquid by high performace liquid chromatography (HPLC) ad gas chromatography (GC). Descriptio of experimet The materials used i the experimet are RO water ad Italia coffee powder. The equipmet eeded icludes a gridig machie, a water bath, a refrigerated circulatio bath, ad a ultrasoic device that fuctios at 28 khz ad 42 khz with GC ad HPLC. To esure a cosistet ad accurate temperature for each experimet, the circulatio bath is utilized to either raise or lower the temperature of the coffee/ro solutio. Whe the correct temperature for the specific experimet is reached, the coffee powder will be mixed ito the RO water. This mixture is the processed through oe of the two separate ultrasoic machies. Oe ra at 28 khz for the duratio of 15 secods ad the other ra at 40 khz for the same amout of time. The results of these extractios are collected i collectio tak 1 ad collectio tak 2, respectively. The process is ext repeated for the duratio of 30 secods. The extractio results are also gathered ito the taks. Each collectio tak the holds the solutio of the two extractios. Process of experimet First, a grider is employed to grid the coffee beas ito fie grais ad the filter them through a 40 mesh filter. Whe the temperature reaches the operatig temperature, the mixture of coffee powder ad RO water is placed ito the ultrasoic equipmet, ad processed uder various extractio coditios. The extractio liquid is iitially filtered through a 40 mesh filter ad collected. To achieve a ratio of extractio liquid to water of 1:9 for a fial dilutio of 10 times, further the extractio liquid is filtered through a 0.45 μm filter paper. Ad the, the coffee flavor cocetratio is used as the base for the comparative aalysis of HPLC. The caffeie cocetratio, o the other had, is used as the base for the comparative aalysis of GC.

Wag, C.-C., et al.: A Novel Optimized Eergy-Savig Extractio Process o Coffee S55 The itegrity of the process ad aalysis is verified by usig the kow coffee flavor cocetratio of the stadardized coffee. The results of the caffeie cocetratio of the stadardized coffee are compareed with the aalyzed caffeie cocetratio of the sample provided by the issuig laboratory. If these results are ot similar, the experimet will be repeated with adjustmet to the operatig coditios. Fially, these results are tabulated ad the comparesos idicate the optimum values for time/frequecy ad temperature would produce the greatest cocetratios of caffeie. The experimetal process is show i fig. 1. Figure 1. Process of experimet Experimetal desig by Taguchi method The orthogoal arrays of L4(2 3 ) are used with three factors for two levels. These three operatig parameters i this experimet iclude: solutio temperature, operatig frequecy ad oscillatio periods. The solutio temperature is either 85 C or 95 C. The operatig frequecy is either 28 khz or 40 khz. The oscillatio periods are either 15 or 30 secods. Each experimet uses 300 ml of RO water ad 30 grams of coffee powder. Each extractio is the tested for caffeie cocetratio. The experimet without applyig Taguchi Method has to be doe 40 times, which meas there will be 160 times of performaces of the experimet if we do 4 rouds of repetitio; the experimet with Table 1. Experimetal data of extractio Taguchi Method desig, however, Exp. A (temperature) B (frequecy) C (time) oly eeds half of the performaces. 1 85 C 28 khz 15 s Taguchi method is ot oly more 2 85 C 40 khz 30 s efficiet but also more coveiet for the experimetal plaig, as show 3 95 C 28 khz 30 s i tab. 1. 4 95 C 40 khz 15 s Equatios of Taguchi method The experimetal results are aalyzed by specific equatios, icludig average value, stadard deviatio, sigal to oise (SN) ratio, total variatios ad errors, respectively. Average value: y = 1/ i 1 yi, Stadard deviatio: S = [1/ 1 i 1 (y i y ) 2 ] 1/2 (1)

S56 Wag, C.-C., et al.: A Novel Optimized Eergy-Savig Extractio Process o Coffee S 1 SN ratio (the smaller, the better 10log y2 i N i 1 (2) Total variatios: DOF T = r 1, T r SS y r y (3) i 1 j 1 2 2 ij Results ad discussio L r A Factor effects:, DOF 1, ( ) 2 A L SS A yk y L k 1 A Errors: DOF ( r 1), SS S 2( r 1) (5) e e i i 1 The optimized combiatio parameter of caffeie by Taguchi method We do the 4 experimets i tab. 2 respectively ad repeatedly for 3 times ad thus gai the data of caffeie P1, P2 ad P3; ext, eq. (1) is used to calculate the average value of each experimet; the, eq. (2) is applied to obtai each error of stadard. Accordig to differet demads, we determie the characteristics of the S/N ratio the smaller the values, the better. Therefore, the S/N ratio data are obtaied, 43.26, 43.39, 43.96 ad 42.32, by substitutig the eq. (3) with the average value ad stadard value. The results are show i tab. 2. Table 2. Table of experimet records of extractio temperature at 85 C ad 95 C Exp. P1 [ppm] P2 [ppm] P3 [ppm] y [ppm] S S/N 1 150.10 144.28 141.97 145.45 4.19 43.26 2 149.35 149.90 144.06 147.77 3.22 43.39 3 153.36 156.05 163.91 157.77 5.48 43.96 4 148.61 146.51 144.65 146.59 1.98 42.32 (4) Table 3. Table of factor effects of extractio temperature at 85 C ad 95 C A B C Level 1 146.61 151.61 146.02 Level 2 152.18 147.18 152.77 Effect 5.57-4.43 6.75 Table 3 is the respose table. The values of A1, B1 ad C1 are the average umbers from the level 1 value they are 146.61, 151.61 ad 146.02, respectively. The values of A2, B2 ad C2 are the average umbers from the level 2 value they are 152.18, 147.18 ad 152.77, respectively. A factor is 5.57; B factor is 4.43; C factor is 6.75; the optimized factor desig is A2B1C2 parameter combiatio (extractio temperature is 95 C, operatig frequecy is 28 khz ad extractio time is 30 s). Table 4 is about the use of the eq. (7) to solve factor effects of A, B ad C. Equatio (5) is used to get the total square umber of the variatios 412.39. Equatio (6) solves the DOF of the three factors. Equatio (4) is used to get the DOF of total variatios 11.

Wag, C.-C., et al.: A Novel Optimized Eergy-Savig Extractio Process o Coffee S57 Cotributio is proportioal of the square of each factor, like (93.07/412.39) 100% = 22.57%. Because the cotributio of experimet errors of C factor is 30%, the ifluece of C factor is larger for the experimet. The optimized parameter of coffee flavor by Taguchi method Error Total 123.75 412.39 8 11 30.0083% 100% Tab. 5 displays the optimized parameters of coffee flavors, icludig 2-Methoxy-3-Methylpyrazie, 2-Pheylethaol, Capric Acid, ad 4-Tridecaoe by the kow factor effects. For 2-Methoxy-3- Methylpyrazie, 2-Pheylethaol ad Capric Acid, they have the optimized parameters whe the temperature is 85 C, the operatig frequecy is 28 khz ad the extractio time is 15 s. For 4-Tridecaoe, evertheless, it has the best parameters whe the temperature is 95 C, the operatig frequecy is 40 khz ad the extractio time is 15 s. Table 5. Table of the best parameters of coffee flavors Table 4. Table of the aalysis of extractio variace at the temperature of 85 C ad 95 C Factor SS DOF Cotributio A (temperature) 93.07 1 22.5697% B (frequecy) 58.87 1 14.2766% C (time) 136.69 1 33.1454% Coffee flavor 2-Methoxy- 3-Methylpyrazie 2-Pheylethaol Capric acid 4-Tridecaoe Temperature Operatig frequecy (28 khz) Operatig frequecy (40 khz) Extractio time (15 s) 85 C 95 C 85 C 95 C 85 C 95 C 85 C 95 C Extractio time (30 s) Work of extractio Time of Extractig is maitaied for 120 secods; the, we compare the geeral coffee machie with ultrasoud thermal method to extract the caffeie ad flavor with differet operatig temperatures. The frequecy is maitaied at 28 khz whe ultrasoud thermal method is used. The curret cosumed by the geeral coffee machie at 65 C is over the iterval of 5.51 to 5.55, ad by the ultrasoud thermal method, which is show i tab. 6. The powers of extractio by these methods are derived from eq. 10 ad the results are show i fig. 5. where I is curret ad V is voltage Power I V 2 (6)

S58 Wag, C.-C., et al.: A Novel Optimized Eergy-Savig Extractio Process o Coffee From fig. 2, the power of applyig ultrasoud thermal process is more tha usig geeral coffee machie for extractio of coffee i 30 secods, but decreases right after the time. Also, it ca be foud that the differeces of powers at 35 C ad 65 C by ultrasoud thermal method are tiy, so we may raise the operatig temperature higher to icrease the quatity of caffeie without too much power. Accordigly, the work of extractio ca the be obtaied by calculatig the areas uder the differet lies i fig. 2. The cosumed work is show i tab. 6 ad be able to lear that cosumed work obtaied by ultrasoud thermal method is always more efficiet ad eergy-savig tha usig the geeral coffee machie. Especially, the work of the geeral coffee machie is 1.84 ad 1.81 times the stregth of ultrasoud thermal process at 35 C ad 65, respectively. Figure 2. Power of differet extractio methods (M1: geeral coffee machie, M2: ultrasoud extractio at 5 C, M3: ultrasoud extractio at 35 C, M4: ultrasoud extractio at 65 C) Table 6. Work of extractio Work Method Geeral coffee machie Ultrasoud thermal method (35 ) Ultrasoud thermal method (65 C) Extractig 60 s 51,709 54,426 54,882 Extractig 120 s 103,045 55,974 56,641 Coclusios From the results of these experimets, it ca be cocluded that the ifluece of oscillatio periods ad temperature are the most ad the secod importat factors i determiig caffeie cotet of a extractio by Taguchi method. From the table of S/N ratio ad factor effects, at 85 C of extractio temperature, 28 khz of operatig frequecy ad 15 secods of extractio time by usig Taguchi Method produces the best result i the experimet of 2-Methoxy-3-Methylpyrazie, 2-Pheylethaol ad Capric Acid. If the extractio time factor accouts for the most of total quality loss, that meas the extractio time factor has the greatest ifluece o this kid of situatio. The cosumed work of the geeral coffee machie is 1.84 ad 1.81 times the stregth of ultrasoud thermal process at 35 C ad 65 C, respectively. So, the proposed

Wag, C.-C., et al.: A Novel Optimized Eergy-Savig Extractio Process o Coffee S59 extractio method is a kid of optimized eergy-savig process ad ca be applied to largescale maufacture. The results idicate the ifluece of temperature, extractio duratio, ad ultrasoud frequecy as well. It provides the isight ito creatig a more efficiet, i terms of time ad eergy cosumptio, maufacturig process for caffeie ad flavor cocetratio extractio from coffee. Ackowledgmets The fiacial support of this research by Natioal Sciece Coucil of the R.O.C., uder Grat No.NSC-98-2622-E-269-004-CC3 ad NSC-98-2221-E-269-012 is greatly appreciated. Nomeclature I V curret, [A] voltage, [V] y y i average of y i, [ppm] caffeie or flavor sigal, [ppm] Refereces [1] Madal, V., Moha, Y., Hemalatha, S., Microwave Assisted Extractio of Curcumi by Sample Solvet Dual Heatig Mechaism Usig Taguchi L9 Orthogoal Desig, Joural of Pharmaceutical ad Biomedical Aalysis, 46 (2008), 2, pp. 322-327 [2] Gruau, D., Che, S., Eggert, K., A Lattice Boltzma Model for Multiphase Fluid Flows, Phys. Fluids A, 5 (1993), 10, pp. 2557-2562 [3] Yi, Y., Yu, H., Zhag, H., A Feature Extractio Method Based o Wavelet Packet Aalysis for Discrimiatio of Chiese Viegars Usig a Gas Sesors Array, Sesors ad Actuators B: Chemical, 134 (2008), 2, pp. 1005-1009 [4] Mätysalo, M., Mätysalo, E., Extractio ad Filterig i Ultrasoic Field: Fiite Elemet Modellig ad Simulatio of the Processes, Ultrasoics, 38 (2000), 6, pp. 723-726 [5] Babaei, R., Jabbari, A., Yamii,Y., Solid-Liquid Extractio of Fatty Acids of Some Variety of Iraia Rice i Closed Vessel i the Absece ad Presece of Ultrasoic Waves Asia, Joural of Chemistry, 18 (2006), 1, pp. 57-64 [6] Sigh, D. K., Sahu, A., Spectrophotometric Determiatio of Caffeie ad Theophyllie i Pure Alkaloids ad its Applicatio i Pharmaceutical Formulatios, Joural of Aalytical Biochemistry, 349 (2006), 2, pp. 176-180 [7] Zhag, Q. L., et al., Separatio of Caffeie ad Theophyllie i Poly(dimethylsiloxae) Microchael Electrophoresis with Electrochemical Detectio, Joural of Chromatography A, 1098 (2005), 2, pp. 172-176 Paper submitted: July 11, 2010 Paper revised: September 30, 2010 Paper accepted: November 11, 2010