Optimization of decoloring conditions of crude fatty acids recovered from crude glycerol by acid-activated clay using response surface method

Transcription

1 Koren J. Chem. Eng., 31(11), (2014) DOI: /s INVITED REVIEW PAPER INVITED REVIEW PAPER pissn: eissn: Optimiztion of decoloring conditions of crude ftty cids recovered from crude glycerol by cid-ctivted cly using response surfce method Hong Su*, Xue Wng*, You Geun Kim**, Sung Be Kim*, Yng-Gon Seo*, Jin Seog Kim***, nd Chng-Joon Kim*, *Deprtment of Chemicl nd Biologicl Engineering & ERI, Gyeongsng Ntionl University, Jinju , Kore **KB Cosmetics, Jinju , Kore ***Deprtment of Applied Sttistics nd Informtion Science, Dongguk University, Gyeongju , Kore (Received 18 April 2014 ccepted 2 June 2014) Abstrct Crude glycerol, by-product of the biodiesel production process, contins high mount of ftty cids tht cnnot be used directly without removing the strong drk color. This study ims to remove the color impurities in crude ftty cids, prepred by cid-precipittion nd hexne extrction, using cid-ctivted cly s n dsorbent. The effects of bleching temperture, contct time, the mount of cid-ctivted cly, nd concentrtion of color impurities were investigted. No significnt effects of bleching temperture or contct time were observed. In n optimiztion study using the centrl composite design, complete decoloriztion ws chieved t n optimized condition, in which 3.5 g of cly ws dded to 10 ml of crude ftty cid solution with n opticl density t 373 nm of 20. Notbly, more thn 80% of the peroxides ws concurrently removed. However, 37% loss in ftty cids ws observed during decoloriztion. The cly ws repetedly used without loss of dsorption cpbility if it ws clcined t 600 o C fter use. Keywords: Biodiesel, Crude Glycerol, Ftty Acid, Decoloriztion, Response Surfce Design INTRODUCTION To whom correspondence should be ddressed. E-mil: cj_kim@gnu.c.kr Copyright by The Koren Institute of Chemicl Engineers Crude glycerol is the principl by-product during the biodiesel production process [1]. Generlly, 10 kg of glycerol is generted for every 100 kg of biodiesel produced [2]. As the demnd nd production of biodiesel re growing extensively, worldwide surplus of glycerol hs resulted nd, thus, the price of crude glycerol hs plummeted. Although refined glycerol cn be obtined by purifying crude glycerol nd then utilized in the cosmetic, pint, utomotive, food, phrmceuticl, nd pulp industries [1], it is not complete solution becuse crude ftty cids re lso generted. Intensive studies on bioconversion of crude glycerol to vlue-dded products by mens of fermenttion hve found tht crude ftty cids re toxic to microorgnisms nd, thus, prtilly purified (ftty cid-free) crude glycerol is better fermenttion substrte thn tht of crude glycerol [3,4]. Therefore, we cnnot void the genertion of crude ftty cids s wste when utilizing crude glycerol. We hve found high mount of ftty cids slts ( 17%, w/w) in crude glycerol. The presence of these ftty cids my be due to neutrliztion of the free ftty cids present s impurities in oils by lkline ctlysts nd concurrent sponifiction of oils during the biodiesel production process. No industril pplictions of these crude ftty cids hve been proposed. Environmentl problems would be creted if these seprted crude ftty cids were disposed. We found tht plmitic cid (C16:0), oleic cid (C18:1), nd linoleic cid (C18:2) re predominnt in crude ftty cids generted from this process s mentioned in text. Fortuntely, these long-chin sturted nd unsturted ftty cids cn be used directly or s precursors for rw mterils in cosmetics, phrmceuticls, nd fine chemicls [5-7]. However, the seprted crude ftty cids hve strong drk brown color. If these crude ftty cids re intended for cosmetic pplictions, it is essentil to remove the color impurities [8] but no processes re vilble for this purpose. Adsorption processes re conventionlly used to decolorize vegetble oils, in which the oil is contcted with dsorbents in vcuo in the presence of het, followed by filtrtion [8,9]. Vrious kinds of dsorbents hve been tested for refining vegetble oils. Although decoloring performnce of ctivted crbon nd silic-bsed dsorbents ws better thn other dsorbents, it ws found tht cid-ctivted bentonites (clcium-montomorillonite) re the most effective in removing color from the refined oils [8,10]. Therefore, this cidctivted cly ws pplied to decolorize the crude ftty cids in this study. Operting fctors criticlly influence the performnce of the decoloriztion process, so n evlution nd optimiztion of these fctors is required. However, there is little informtion on optimiztion to obtin the highest decoloring performnce. Furthermore, the excessive use of ctivted cly cn cuse environmentl problems due to incresed lnd-fill disposl [11]. This study ims to chieve complete decoloriztion of crude ftty cids ( 99%) using cid-ctivted cly with minimum genertion of wste cly. In preliminry experiments, the effects of temperture nd contct time were investigted. Centrl composite design (CCD) ws pplied s sttisticl pproch to determine the optiml mount of cly nd initil color intensity of the crude ftty cid solution. It ws lso exmined whether cly could remove toxic peroxides from ftty cid solution. Finlly, simple regenertion method is proposed for the repeted use of cly. MATERIALS AND METHODS 1. Mterils Crude glycerol ws obtined from Eco Solution (Seoul, Kore).

2 Decoloriztion of crude ftty cids 2071 Tble 1. Physicl properties of cid-ctivted cly Free moisture (%) 6-10 ph (5%, sol.) 3-5 Acidity (KOH, mg/g) <2.5 Prticle size (%) (200 mesh pss) >90 Specific surfce re (m 2 /g) Bulk density (g/l) Dt were obtined from supplying compny Acid-ctivted cly (DC-A3) ws purchsed from Donghe Chemicls (Seoul, Kore). This cly is powder-type prticle (200 mesh pss) with high surfce re. It ws produced by cid tretment of clcium-montmorillonite nd hs been populrly used for refining edible oils (niml nd vegetble oils). The physicl properties of this cly re summrized in Tble 1. Hexne (95% n-hexne) nd HCl (purity, 35%) were purchsed from J. T. Bker (Phillipsburg, NJ, USA) nd Smchun Chemicls (Pyeongtek, Kore), respectively. 2. Preprtion of Crude Ftty Acids from Crude Glycerol Ftty cids were prepred from crude glycerol following the methods described by Wnsundr nd Shhidi [12] with some modifiction; 200 ml of crude glycerol ws mixed with 200 ml of distilled wter. The ph of the mixture ws djusted to 1.0 by dding HCl to convert the soup (ftty cid slt) to free ftty cids, which were precipitted from solution. Hexne (400 ml) ws dded to the cidified solution, the mixture ws trnsferred to seprtory funnel, nd the liberted ftty cids were extrcted into the hexne. A hexne lyer contining ftty cids ws collected, nd the hexne ws removed in rotry evportor (EYELA N-1100, Rikkiki Co., LTD., Tokyo, Jpn) t 55 o C for 40 min under vcuum. A sticky, drkbrown colored crude ftty cid solution (38 ml) ws obtined, nd this concentrted solution ws suitbly diluted with hexne for decoloriztion. 3. Decoloriztion of the Crude Ftty Acid Solution All experiments were crried out in 250 ml Erlenmeyer flsks contining the crude ftty cid solution nd cid-ctivted cly by mintining the mixture t 100 rpm in shking incubtor (Chngshin Scientific Co. Ltd., Seoul, Kore). To exmine the effect of temperture nd contct time, 2 g of cly nd 10 ml of suitbly diluted crude ftty cid solution (opticl density vlue t 375 nm [OD 375 ]=38) were mixed nd mintined t two different tempertures (30 nd 70 o C). After 30 or 120 min, 1 ml of smple ws tken, the cly ws seprted by centrifugtion (9,800 g for 10 min), nd the superntnt color intensity ws mesured. CCD ws pplied to determine the optiml mount of cly nd initil color intensity of the ftty cid solution. Fourteen experiments were performed in duplicte ccording to this design. Regression nlysis of the dt ws conducted using SAS version 9.1 softwre (SAS Institute, Cry, NC, USA), nd the optiml vlues of these fctors were obtined by solving second-order polynomil eqution nd nlyzing the response surfce contour plots. The cly nd ftty cids were seprted by pssing the slurry through 0.2 mm nylon membrne filter (47 mm, Whtmn Interntionl Ltd., Midstone, Englnd) under vcuum nd the recovered cly ws dried t 80 o C for 12 h. This spent cly ws dded to n unbleched crude ftty cid solution with or without clcintion t 600 o C for 6 h, nd the decoloriztion cycles were repeted to determine reusbility. 4. Anlyses The color intensity of the crude ftty cid solution ws mesured with UV-visible spectrophotometer (Hewlett-Pckrd, Plo Alto, CA, USA). The concentrted crude ftty cid stock solution ws diluted in hexne, nd the OD 375 ws determined using hexne s reference fter mesuring the full spectrum of the smple between nm. Decoloriztion efficiency (%) ws clculted using the following formul: Decoloriztion efficiency (%)=(A unbleched A bleched ) 100/A unbleched where A unbleched nd A bleched re the OD 375 vlues of the smples before nd fter decoloriztion with cly, respectively. Ftty cid composition ws determined for the crude ftty cid solution with or without decoloriztion. After esterifiction to ftty cid methyl esters (FAME) using methnolic HCl [13], the FAME smples were nlyzed by gs chromtogrphy (M600D, Yong- Lin, Seoul, Kore) equipped with flme-ioniztion detector nd fused silic cpillry column (100 m 0.25 mm I.D. 0.2 μm film thickness, SP TM -2560, Supelco, Belfonte, PA, USA). Nitrogen ws used s the crrier gs t flow rte of 2.7 ml/min. The oven temperture ws initilly 140 o C for 5 min nd ws then rised to 240 o C t rte of 4 o C/min nd held for 10 min. The injector nd detector tempertures were kept t 240 nd 260 o C, respectively. Smple volume ws 1 μl with split rtio of 10 : 1. The ftty cids were identified by compring the retention times with those of stndrd ftty cids (Supelco 37 Component FAME Mix). Chnges in the re of the smple before nd fter decoloriztion were clculted to estimte the loss of ftty cids during the bleching step. Concentrtions of hydrogen peroxide (H 2 O 2 ) nd tert-butyl hydroperoxide (t- BuOOH) were determined by colorimetric rection using commercil kit supplied by Sigm (St. Louis, MO, USA). Aqueous solutions of hydrogen peroxide (30%) nd t-buooh (70%) (Sigm) were used for prepring the clibrtion curve, respectively. RESULTS AND DISCUSSION 1. Selection of Suitble Wvelength for Monitoring Color Removl The crude ftty cid solution ws strong drk brown color which Fig. 1. UV/Visible spectr of crude ftty cids solution before () nd fter (b) decoloriztion. Crude ftty cids were recovered from crude glycerol nd diluted in hexne. Acid-ctivted cly (2 g) ws dded to the crude ftty cid solution (10 ml) nd the slurry ws shken t 30 o C nd 100 rpm for 30 min. Koren J. Chem. Eng.(Vol. 31, No. 11)

3 2072 H. Su et l. Tble 2. Effect of temperture nd contct time on decoloriztion Time (min) Decoloriztion efficiency (%) b 30 o C 70 o C ± ± ± ±0.7 Ten ml ftty cid solution (OD 375 =38) nd 2 g cid-ctivted cly were mixed for decoloriztion b Dt re shown s removl (%)±S.D. disppered when the smple ws treted with cly. Treted nd untreted smples were scnned on UV-visible spectrophotometer. As shown in Fig. 1, the crude ftty cid solution demonstrted strong bsorption t wvelengths of nm, wheres peks corresponding to wvelengths >365 nm disppered fter decoloriztion. Considering tht most peks pper t nm in purified ftty cids [14,15], this result clerly indictes tht colored impurities existing in the crude ftty cids contributed to bsorb light t wvelengths >365 nm nd tht cly cn remove them effectively by dsorption. These impurities my be non-voltile decomposed compounds such s oxidized tricylglycerols nd free ftty cids, which re generted during the biodiesel production process [16], nd they were not seprted when the crude glycerol ws recovered. The color intensity of the crude ftty cid solution decresed when it ws diluted with hexne. Notbly, plot of bsorbnce t 375 nm vs. dilution fctor provided stright line (dt not shown). This result clerly indictes tht removing color cn be monitored by mesuring bsorbnce t this wvelength. 2. Effect of Temperture nd Contct Time on Decoloriztion The efficiency of decolorizing crude ftty cid solution cn be influenced by severl fctors such s bleching temperture, contct time, concentrtions of color impurities nd crude ftty cids, nd the mount of cly. In preliminry experiments, the effect of temperture nd contct time ws investigted. Generlly, oil bleching processes using dsorbents re crried out t bleching tempertures of o C nd contct times of min depending on the type of oil nd dsorbent [9,17]. Room temperture (30 o C) nd 70 o C were selected s low nd high level. Higher temperture could not be tested becuse boiling point of solvent (hexne) is close to 70 o C. Conventionl (30 min) nd longer contct time (120 min) ws selected s low nd high level. Successful decoloriztion ( 90%) ws chieved in 30 min when the crude ftty cid solution ws treted t 30 o C (Tble 2). No dditionl improvement occurred despite incresing the temperture or extending the contct time. In fct, the temperture effect during the oil bleching process is relted to oil viscosity. When oil viscosity is high, trnsport mechnisms slow down due to decresed diffusion rtes of the colored mterils to the dsorbent surfces. Therefore, the time required to rech dsorption equilibrium decreses s temperture increses [18,19]. Another possibility is tht n increse in dsorption rte with temperture cn be ttributed to the ctivtion of more dsorption sites by het [20]. The former suggestion is more plusible for decoloriztion of the crude ftty cid solution with cly. However, the viscosity of this crude ftty cid solution ws much lower thn tht of oils for refining in conventionl processes becuse the concentrted ftty cid solution ws suitbly diluted with hexne. Therefore, contct between the color impurities nd cly ws not prevented even t 30 o C. This result indictes tht bleching temperture nd contct time were not criticl fctors for removing color impurities in crude ftty cid solution with cly in this study. Therefore, other fctors were further optimized to chieve complete decoloriztion t 30 o C for 30 min. 3. Determining the Optiml Amount of Cly nd Initil Color Intensity of the Crude Ftty Acid Solution Although cly is good dsorbent for removing color impurities from crude ftty cid solution, it hs limited dsorption cpcity for these impurities. Therefore, the optimum rtio between the mount of cly required nd concentrtion of impurities to be removed should be determined to chieve mximum decoloriztion. These two fctors were therefore optimized in experiments ccording to CCD, in which incubtion temperture nd contct time were fixed t 30 o C Tble 3. Experimentl runs of centrl composite design for decoloriztion of the crude ftty cid solution Symbol X 1 (Cly, g) X 2 (OD 375 of ftty cids solution) Decoloriztion efficiency (%) Run Coded Uncoded Coded Uncoded Exp. Clc ± ± ± ± ± ± ± ± X 2 is the reltive concentrtion of color impurities in the initil crude ftty cid solution nd is represented s OD 375 November, 2014

4 Tble 4. Estimted regression coefficients nd significnce test for the second-order model Prmeter Coefficient estimte Stndrd error t-vlue Decoloriztion of crude ftty cids 2073 Prob> t Intercept <.0001 X <.0001 X <.0001 X 1 *X <.0001 X 2 *X X 1 *X <.0001 Tble 5. Anlysis of vrince for decoloriztion using cid-ctivted cly Source Degree of freedom Sum of squres Men squre F vlue Prob>F Model < Error Totl R 2 =0.987 nd 30 min, respectively. Ftty cid solutions with different concentrtions of color impurities were prepred by diluting the concentrted drk colored ftty cid solution with hexne. As mentioned bove, reltive concentrtions of color impurities cn be monitored by mesuring OD 375. The design mtrix nd corresponding results from seprte experiments re shown in Tble 3. A regression nlysis of the dt performed using n SAS progrm generted second-order polynomil eqution. The sttisticl significnce of the model eqution ws evluted by nlysis of vrince (ANOVA), which indicted tht the model terms, X 1, X 2, X 1 X 2, X 12, nd X 22 were significnt (P<0.005), s shown in Tble 4. The overll second-order polynomil eqution for decoloriztion ws: Y= X X X X X 1 X 2 where Y is decoloriztion efficiency (%), nd X 1 nd X 2 re the coded vlues of the mount of cly nd the OD 375 vlue of the ftty cid solution, respectively. Tble 5 shows the verifiction of the model bsed on the results of n F-test nd ANOVA. The vlue of the P model > F ws less thn , indicting tht the results produced by the model were significnt. The coefficient of determintion (R 2 ) ws when goodness of fit of the model ws tested, which indicted tht only 1.3% of the totl vrition ws by chnce nd, thus, ws not explined by the model [21,22]. The ccurcy of the model ws verified by the dt in Tble 3, which showed tht the observed nd predicted vlues for decoloriztion were lmost equl. The signs of the coefficients of the fctors in the model eqution indicted their reltive effects on decoloriztion efficiency. The positive sign for the mount of cly (X 1 ) indicted tht higher response could be obtined if the pplied mount of cly ws greter thn tht of the center point, wheres the negtive sign for OD 375 (X 2 ) indicted the opposite [21,22]. The model eqution showed the presence of the interction between the two fctors. Therefore, the model eqution suggests tht mximum decoloriztion cn be obtined using Fig. 2. Contour plot of decoloriztion of crude ftty cid solution in response to vrying the mount of cid-ctivted cly nd color intensity of the initil crude cid solution. higher mount of cly thn tht of the center point while lso using ftty cid solution with OD 375 tht is lower thn the center point. A contour plot ws drwn bsed on the model eqution, which clerly shows the effect of the two fctors on decoloriztion efficiency (Fig. 2). This plot indictes tht the color ws not removed completely even though the mximum mount of cly ws loded to tret the ftty cid solution with OD 375 higher thn the center points. In contrst, mny experimentl conditions cn be used to chieve complete decoloriztion (100%) depending on the OD 375 of the ftty cid solution if its OD 375 is less thn tht of the center point following dilution with hexne. It is necessry to consider the opertionl nd economic spects to choose suitble opertion point to chieve complete decoloriztion. A smll mount of cly is required for decolorizing wekly colored ftty cid solution (low OD 375 ), but lrge volume of hexne is required to dilute ftty cid solution. In contrst, lrger mount of cly is required for decolorizing strongly colored ftty cid solution (high OD 375 ) lthough the required volume of hexne is smller thn forml one. Dilution of crude ftty cids with hexne leds to concomitnt decrese in ftty cid concentrtion s well s OD 375, thereby reducing solution viscosity. Furthermore, it is esier to seprte cly from slurry mixture by centrifugtion or filtrtion if the crude ftty cid solution is less viscous. Cly will be disposed of s wste fter use but hexne cn be recovered for reuse by distilltion. Considering the economic, environmentl, nd opertionl spects, use of the lest mount of cly rther thn sving hexne is better pproch. Therefore, we selected the point ( 0.5, 1.0) t which complete decoloriztion ws observed on the contour plot. This condition corresponded to 3.5 g cly nd strting with crude ftty cid solution of modest color (OD 375 =20). When, triplicte experiments were performed under these conditions, the observed decoloriztion ws 99.7±0.1%, which corresponded firly well to the predicted vlue. Smples with nd without decoloriztion re compred in Fig. 3. The results clerly indicte tht cid-ctivted cly cn completely remove color impurities in crude Koren J. Chem. Eng.(Vol. 31, No. 11)

5 2074 H. Su et l. Fig. 3. Crude ftty cid solution before () nd fter (b) decoloriztion by cid-ctivted cly under optiml conditions. Fig. 4. Removl of peroxides from crude ftty cids during decoloriztion using cid-ctivted cly. ftty cid solution under optiml condition. 4. Removl of Peroxides during Decoloriztion Some biodiesels re produced by trnsesterifiction of spent vegetble oils, which my contin oxidized products such s peroxides. More peroxides re generted during biodiesel production due to the high temperture rection. Therefore, crude ftty cids derived from this process my contin peroxides. Contct of peroxides with the skin cn cuse severe skin dmge or premture skin ging [23-25]. Furthermore, they cuse dmge to the cellulr molecules such s DNA, proteins, lipids [25] even t low concentrtion if pssed through cytoplsmic membrne. These highly rective mterils my lso cuse side rections during chemicl processes. Therefore, crude ftty cids cnnot be used s cosmetic ingredients or rw mterils for chemicl industries without removing these toxic compounds. Thus, the bility of cly to remove peroxides ws evluted. Decoloriztion ws performed under optiml conditions; 3.5 g of cly ws dded to 10 ml of ftty cid solution (OD 375 =20), nd the mixture ws incubted t 30 o C for 30 min. Then, the mount of peroxides ws compred in the ftty cid solution before nd fter decoloriztion. H 2 O 2 nd t-buooh were used s representtive inorgnic nd orgnic peroxides, respectively. As expected, high mount of peroxides ws detected in the ftty cid solution, nd t-buooh (132 mm) ccumulted much more thn H 2 O 2 (37 mm), s shown in Fig. 4. Notbly, H 2 O 2 nd t-buooh were 4.4 μm nd 26 μm, respectively in the decolored smple, indicting tht more thn 80% of the peroxides were removed. This result clerly indictes tht cly cn remove peroxides concomitntly s well s color impurities from crude ftty cid solution. The destruction of peroxides in oils by cid-ctivted clys hs been demonstrted in previous reports [17,18,26,27]. 5. Loss of Ftty Acids during Decoloriztion Cly is n dsorbent, so we wondered whether it dsorbed not only color impurities but lso ftty cids. To nswer this question, gs chromtogrphy nlysis ws performed to nlyze the ftty cids in the smples. By compring with the chromtogrm of n uthentic smple, we discovered tht plmitic cid (C16:0), oleic cid (C18:1), nd linoleic cid (C18:2) were the predominnt crude ftty cids (>79%). Therefore, pek res of these components in smples before nd fter decoloriztion were compred to estimte the reltive chnges in these ftty cids in the smple fter decoloriztion. As shown in Tble 6, the mount of these ftty cids decresed by pproximtely 37% fter decoloriztion. This result indictes tht cly lso removes ftty cids. Adsorption of free ftty cids onto cid-ctivted cly hs been demonstrted, nd dsorption cpcity vries depending on the cid tretment conditions [18,28,29]. However, it ppered tht cly removed the color impurities nd toxic peroxides much more selectively thn the ftty cids becuse they were lmost completely removed, wheres only one-third of the ftty cids were removed. 6. Reusbility of Cly It is more cost-effective nd minimizes wste genertion if cly is reused. Accordingly, repeted decoloriztion cycles were performed using cly to evlute the reusbility of the cly. Decoloriztion efficiency (%) decresed rpidly s the cycles were repeted nd, thus, efficiency ws only 17.6% fter the third cycle (Fig. 5). It ws suspected tht the loss of decoloriztion cpcity ws due to sturtion of the dsorption sites on the cly with orgnic mterils, including color impurities during the recycling process. Therefore, cly ws clcined t 600 o C for 6 h to burn these orgnic mterils off fter every cycle, nd then reusbility ws evluted in the next Tble 6. Chnge in ftty cid concentrtion during decoloriztion Ftty cid Ares Chnge (%) Strting smple Decolored smple Plmitic cid (C16:0) ± ± ±3.70 Oleic cid (C18:1) ± ± ±11.5 Linoleic cid (C18:2) ± ± ±3.80 Chnge (%)=(re of strting smple re of decolored smple) 100/re of strting smple November, 2014

6 Decoloriztion of crude ftty cids 2075 glycerol derived from the biodiesel production process. Color impurities in the ftty cids were removed by cid-ctivted cly. Bleching temperture nd contct time hd no significnt influence on decoloriztion of the crude ftty cids. An optimiztion study ws performed using CCD to determine the miniml mount of cly to chieve complete decoloriztion. Almost complete decoloriztion (99.7%) ws chieved in 30 min t room temperture when 3.5 g cly nd crude ftty cid solution with modest color (OD 373 =20) were pplied. Notbly, more thn 80% of the peroxides were concomitntly removed during decoloriztion under these conditions. We lso showed tht cly cn be reused if it is regenerted by burning dsorbed mterils fter use. This is the first demonstrtion of decolorizing crude ftty cids recovered from crude glycerol. Our results will contribute to promote biodiesel production by reducing cost. Fig. 5. Decrese in the performnce of cid-ctivted cly following repeted use for decolorizing crude ftty cid solution. Tble 7. Performnce of clcined cly fter use in every cycle Cycles Color intensity of ftty cid solution (OD 375 ) Before After cycle. Notbly, 99% of decoloriztion ws mintined with no decrese during five repeted cycles, s shown in Tble 7. This indictes tht the cly could be used repetedly to remove color impurities with no loss in performnce if dsorbed mterils re removed by burning. Decoloriztion efficiency of regenerted cly increses with clcintion temperture; 90% decoloriztion ws ttined in 15 min with spent cly clcined t 600 o C, wheres decoloriztion efficiency ws lower with spent cly clcined t <500 o C [30]. The optiml clcintion temperture is 500 o C to regenerte spent erth to decolorize edible oil, nd this temperture corresponds to the beginning of crboniztion of orgnic mtter contined in the spent cly [31]. Sintering or destructuring of the cly my occur if spent cly is clcined t >600 o C, leding to lower performnce. Tken together, clcintion t 600 o C is suitble for regenerting cly for repeted use. Generlly, cly from the vegetble oil refining industry is recycled through clcintion process t tempertures of o C, followed in some cses by ctivtion with cids [30-32]. However, different from the oil refining process, further cid ctivtion step ws not necessry to regenerte cly in this study nd this ws probbly becuse ftty cids formed n cidic environment round the cly, which my hve contributed to ctivte the cly. CONCLUSIONS Decoloriztion efficiency (%) b 1 st ± ± nd ± ± rd ± ± th ± ± th ± ±0.10 Cly ws clcined fter use in ech step b Dt re shown s removl (%)±S.D. Drk-brown colored crude ftty cids were recovered from crude ACKNOWLEDGEMENTS This study ws supported by the Business for Coopertive R & D between Industry, Acdemy, nd the Reserch Institute funded by the Kore Smll nd Medium Business Administrtion in 2010, nd the Bsic Science Reserch Progrm (NRF-2012R1A1A ), Pioneering Reserch Center for Nno-morphic Biologicl Energy Conversion nd Storge ( ) through the Ntionl Reserch Foundtion (NRF) of Kore. REFERENCES 1. J. C. Thomson nd B. B. He, Appl. Eng. Agr., 22, 261 (2006). 2. F. Yng, M. A. Hnn nd R. Sun, Biotechnol. Biofuels, 5, 13 (2012). 3. D. J. Pyle, R. A. Grci nd Z. Wen, J. Agric. Food Chem., 56, 3933 (2008). 4. K. P. Venktrmnn, J. J. Botmn, Y. Kurniwn, K. A. Tconi, G. D. Bothun nd C. Scholz, Appl. Microbiol. Biotechnol., 93, 1325 (2012). 5. A. M. R. Alvrez nd M. L. G. Rodriguez, Fsc., 51, 74 (2000). 6. S. A. Ibrhim nd S. K. Li, Phrm. Res., 27, 115 (2010). 7. P. Mki-Arvel, J. Kuusisto, E. M. Sevill, I. Simkov, J.-P. Mikkol, J. Myllyoj, T. Slmi nd D. Y. Murzin, Appl. Ctl. A Gen., 345, 201 (2008). 8. P. Gonus nd H.-J. Wille, US Ptent, 5,401,862 (1995). 9. N. Worsith, B. A. Goodmn, N. Jeyshoke nd P. Thirvetyn, J. Am. Oil Chem. Soc., 88, 2005 (2011). 10. B. Mkhoukhi, M. A. Didi, D. Villemin nd A. Azzouz, Grss Y. Aceites, 60, 343 (2009). 11. F. Hussin, M. K. Arou nd W. M. A. W. Dud, Chem. Eng. J., 170, 90 (2011). 12. U. N. Wnsundr nd F. Shhidi, Food Chem., 65, 41 (1999). 13. S. Hu, X. Lu, C. Wn nd Y. Li, J. Agric. Food Chem., 60, 5915 (2012). 14. R. L. Arudi, M. W. Sutherlnd nd B. H. J. Bielski, J. Lipid Res., 24, 485 (1983). 15. R. T. Holmn, W. O. Lundberg, W. M. Luer nd G. O. Burr, J. Am. Chem. Soc., 67,1285 (1945). 16. C. Mohmmed, Y. Alhssn, G. I. Yrgmji, S. Grb, Z. Bello nd A. I. Ifeyinw, J. Bsic. Appl. Chem., 1, 80 (2011). 17. M. A. Usmn, V. I. Ekwueme, T. O. Alje nd A. O. Mohmmed, ISRM Cermics, 5 (2012). Koren J. Chem. Eng.(Vol. 31, No. 11)

7 2076 H. Su et l. 18. N.-A. B. Joy, K. Richrd nd N. J. Pierre, J. Appl. Sci., 7, 2462 (2007). 19. T. Lngmck nd R. Eggers, Eur. J. Lipid Sci. Technol., 104, 98 (2002). 20. E. C. Achife nd J. Ibemesi, J. Am. Oil Chem. Soc., 66, 247 (1989). 21. S. K. Ahuj, G. M. Ferreir nd A. R. Moreir, Biotechnol. Bioeng., 85, 666 (2004). 22. L. Hung, Z. Lu, Y. Yun, F. Lu nd X. Bie, J. Ind. Microbiol. Biotechnol., 33, 55 (2006). 23. A. J. P. Klein-Sznto nd T. J. Slg, J. Invest. Dermtol., 79, 30 (1982). 24. R. Gnceviciene, A. I. Likou, A. Theodoridis, E. Mkrntonki nd C. C. Zouboulis, Dermto-Endocrinology, 4, 308 (2012). 25. J. A. Lewis II, J. C. Dinrdo nd D. H. McDniel, Cosmet. Dermtol., 22, 576 (2009). 26. L. Meesuk nd S. Semmi, ScienceAsi., 36, 33 (2010). 27. P. Flrs, I. Kovnis, F. Lezou nd G. Seirgkis, Cly Minerls, 34, 221 (1999). 28. A. Sri nd O. Ipyldk, Bull. Chem. Soc. Ethiop., 20, 259 (2006). 29. K. R. Rogn, Colloid. Polym. Sci., 272, 82 (1994). 30. E. L. Foletto, C. C. A. Alves, L. R. Sgnzerl nd L. M. Porto, Ltin Americn Applied Reserch, 32, 205 (2002). 31. A. Boukerroui nd M.-S. Ouli, J. Chem. Technol. Biotechnol., 75, 773 (2000). 32. K.-S. Low, C.-K. Lee nd L.-Y. Kong, J. Chem. Technol. Biotechnol., 72, 67 (1998). November, 2014