Available online at www.sciencedirect.com
Bioresource Technology 99 (2008) 6360–6364
Eﬀect of heat on the adsorption capacity of an activated carbon for decolorizing/deodorizing yellow zein
D.J. Sessa a,*, D.E. Palmquist b
Plant Polymer Research, National Center for Agricultural Utilization Research, US Department of Agriculture, Agricultural Research Service, 1815 N. University Street, Peoria, IL 61604-3902, USA b Biometrical Services, National Center for Agricultural Utilization Research, US Department of Agriculture, Agricultural Research Service, 1815 N. University Street, Peoria, IL 61604-3902, USA Received 21 June 2007; received in revised form 29 November 2007; accepted 29 November 2007 Available online 29 January 2008
Abstract The Freundlich model was evaluated for use to assess the eﬀect of heat on the adsorption capacity of an activated carbon for decolorizing/deodorizing corn zein. Because zein protein and its color/odor components are all adsorbed by activated carbon, a method to monitor their removal was needed. Yellow color is due to xanthophylls; a contributor to oﬀ-odor is diferuloylputrescine. The oﬀ-odor component absorbs ultraviolet (UV) light at about 325 nm and its removal coincides with removal of yellow color. A spectrophotometric method based on UV absorbances 280 nm for protein and 325 nm for the oﬀ-odor component was used to monitor their adsorptions onto activated carbon. Equilibrium studies were performed over temperature range from 25 to 60 °C for zein dissolved in 70% aqueous ethanol. Runs made at 55 °C adsorbed signiﬁcantly more of the color/odor components than the protein. Published by Elsevier Ltd.
Keywords: Activated carbon; Adsorption; Freundlich isotherm; Thermal eﬀects; Zein decolorization/deodorization
1. Introduction Zein, a class of proteins known as prolamins, is the major group of proteins in corn as well as corn gluten meal, a co-product of the ethanol industry generated by wet-milling corn. Its use in the medical, pharmaceutical, and cosmetic ﬁelds and also in applications as a paper coating, packaging material and biodegradable chewing gum base is limited by its inherent yellow color and oﬀ-odor. Sessa et al. (2003) devised a colorimetric assay to quantify residual color in decolorized zein products and performed a statistical assessment for color removal by a variety of processing methods. Color removal with an activated carbon was the best decolorization method. Activated carbons have been used in the past to decolorize zein (Mason and Palmer, 1934; Swallen, 1938; Pearce,
Corresponding author. Tel.: +1 309 681 6351; fax: +1 309 681 6686. E-mail address: David.Sessa@ARS.USDA.GOV (D.J. Sessa).
1941; Starling et al., 1951; McInnis and Tang, 2003). Activated carbon is a non-speciﬁc adsorbent that not only binds the color components but also the protein components. The objective of this investigation is to determine if a change in temperature can be used to enhance the binding capacity of color/odor components onto activated carbon while limiting that of protein adsorption. A means of monitoring protein versus color/odor components is required to measure the eﬀect of heat on the kinetics of adsorption of those components onto activated carbon. In a column chromatographic scheme for decolorizing zein in aqueous ethanol (Sessa et al., 2003) column eluates were monitored by ultraviolet (UV) spectroscopy at wavelengths 280 and 325 nm. Those researchers demonstrated that pooled eluates with low UV absorbance at 325 nm possessed slight yellow coloration, whereas, pooled eluates with increased absorbance at 325 nm were intensely yellow. The colored components coelute with a polyamine
0960-8524/$ - see front matter Published by Elsevier Ltd. doi:10.1016/j.biortech.2007.11.076
D.J. Sessa, D.E. Palmquist / Bioresource Technology 99 (2008) 6360–6364
conjugate characterized as diferuloylputrescine, a compound found in corn oil by extracting ground corn with ethanol (Moreau et al., 2001; Moreau and Hicks, 2005). Yellow zein extracts processed with activated carbon (Sessa et al., 2003) yielded a white zein product with trace UV absorbance at wavelength 325 nm. Odor evaluations by two international companies interested in a decolorized/ deodorized zein declared that the white zein produced was odor free (unpublished data). Therefore, monitoring adsorption of yellow aqueous alcohol solutions of zein onto an activated carbon at wavelength 280 nm for tyrosine in zein protein and 325 nm for diferuloylputrescine provided the means to evaluate the kinetics of adsorption at several temperatures. We applied the Freundlich isotherm model (Crittenden et al., 1985) for this evaluation. 2. Experimental 2.1. Materials Commercial yellow zein (FC4000) with proximate analysis, dry-basis: 93.09% crude protein (Dumas N Â 6.25), 5.26% crude fat, 0.04% crude ﬁber, 0.05% ash, was purchased from Freeman Industries, Tuckahoe, NY. A granular activated carbon from coal, type CPG-LF 12 Â 40 (acid-washed) was provided by Calgon Carbon Corp., Pittsburgh, PA. Ethanol, ACS reagent, 200 proof, P99.5% was obtained from Sigma–Aldrich Inc., St. Louis, MO. The aqueous ethanol concentrations used in this investigation were measured on a weight/volume basis. 2.2. Preparation of colorless zein product A highly puriﬁed zein product was essential in order to develop the analytical procedure for our model system. To accomplish that task yellow zein was processed using a combination of the activated carbon treatment and ultraﬁltration/diaﬁltration on a tangential ﬂow system (Sessa et al., 2003). An activated carbon ﬁlter device with pleated HEPA (i.e. high eﬃciency particulate air) ﬁlter, CarbonCap 150 with area equivalent 82,000 m2, Whatman Inc., Clifton, NJ, was hooked in tandem with a Pall Centramate equipped with a 5000 Da MWCO OMEGA membrane. A yellow, 10% zein solution in 70% ethanol (350 mL) was diaﬁltered at room temperature (i.e. 25 °C) with 12 1-L batches of 70% aqueous ethanol at a rate of 800 mL/h by use of a feed pressure of 220 kPa and retentate pressure of 110 kPa. Solvent was stripped from the retentate on a rotary evaporator until zein started to precipitate. Remaining alcohol was removed by dialyzing the milky zein dispersion contained in 1000 Da MWCO dialysis casing against water; the precipitated zein was freeze-dried. 2.3. Sample protocol The activated carbon granules were sieved through a number 30 screen to remove ﬁne carbon powder. That por-
tion retained on the screen was washed three times with an excess volume of 70% aqueous ethanol to further remove powdered carbon adhering to the surface of the carbon granules. The drained, washed granules, after evaporation of ethanol, were dried overnight in a hot air oven at 105 °C. The dried carbon granules were measured into eleven 250 mL Erlenmeyer ﬂasks as follows: 0, 0.1, 0.2, 0.3, 0.4, 0.5, 1.0, 2.0, 4.0, 7.0, and 10.0 g. A stock solution of 0.4% commercial zein, 100 mL each ﬂask, was added and the ﬂasks then sealed with a rubber stopper. The ﬂasks were gently shaken at 50 rpm on an Innova 4000 Incubator Shaker, New Brunswick Scientiﬁc Co. Inc., Edison, NJ, where equilibrium conditions were established at 24 h. Multiple adsorption experiments were run at 25, 40, 45, 50, 55, 60 °C with shaking. The volume of the liquid portion from each ﬂask was measured to ensure accurate concentration calculation, then decanted into separate tared evaporating dishes. The respective carbon solids from each ﬂask were likewise transferred to tared evaporating dishes. The ultraviolet absorbances of each liquid fraction were measured at wavelengths 280 and 325 nm. Solvents from the zein decantates as well as residual solvents remaining on the carbon solids were each evaporated by air drying followed by completely drying in a hot air oven at 105 °C for 3 h. The mass of zein as well as carbon solids in each of the evaporating dishes were measured. 2.4. Analyses The test adsorptions were designed to permit a general linear model regression consisting of three replicate trials measuring log (qe) as a function of log (Ce) at two wavelengths and six temperatures. The dependent variable qe (equilibrium solid phase concentration from the Freundlich isotherm qe = KfCe1/n) was examined as a function of Ce (liquid phase solute concentration for weights of activated carbon) for zein at two wavelengths (A280 nm and A325 nm) and six temperatures (25, 40, 45, 50, 55, 60 °C). Data from three replicate trials were averaged and log (qe) and log (Ce) values were used for developing simple linear regression equations (SLR). Protein (A280 nm) and contaminants (i.e. color/odor assessed at A325 nm) equations were compared at each temperature using full and reduced model regression techniques (Neter et al., 1990). The original Freundlich isotherm equation qe = KfCe1/n was linearized with the resulting SLR form: log (qe) = log (Kf) + (1/n)log (Ce). Adsorption capacity estimates (Kf) were calculated as the antilog of the intercept from the linearized SLRs. Adsorption bond strength estimates (1/n) were obtained directly as the slope from the linearized SLRs. Values, where Ce = 0, were deleted from the analyses because the log transformation is undeﬁned at that point. For each wavelength, intercepts and slopes of the linearized equations for all six temperatures were compared to determine which temperature had the largest adsorption
D.J. Sessa, D.E. Palmquist / Bioresource Technology 99 (2008) 6360–6364
capacity and bond strength. When the 95% conﬁdence intervals of the regression parameters overlap, the intercepts or slopes for each temperature are not signiﬁcantly diﬀerent from one another. 3. Results and discussion 3.1. Yellow and white zein standards Processing yellow zein using a combination of activated carbon treatment and ultraﬁltration/diaﬁltration on a tangential ﬂow system yielded a white zein product with an average Dumas nitrogen of 15.31, which when adjusted for 4.90% moisture, gave 100% protein. Spectral analysis of that product (Fig. 1), dissolved in 70% aqueous ethanol from 250 to 400 nm showed one major absorbing peak at wavelength 280 nm and no absorbance at wavelength 325 nm. The UV absorbing peak at about 325 nm found in yellow zein solutions (see Plot #1, Fig. 1) is caused by a polyamine conjugate which has been isolated and identiﬁed as diferuloylputrescine by UV and mass spectroscopy (Sessa, unpublished data). The UV absorption spectrum of diferuloylputrescine gave a dual peak with maximum absorbance at 317 nm and a smaller peak at 293 nm (data not shown) which UV absorbances are similar to that reported in the literature (Moreau et al., 2001). The spectrum of yellow zein is a convolution of all the components, whereas, the spectrum of pure zein (see Plot #2 in Fig. 1) lacks the impurities found in yellow zein. Because the UV spectrum of diferuloylputrescine overlaps with the protein component, we monitored its contribution at wavelength 325 nm, where the zein absorbance contribution is essentially zero. Standard curves were developed using pure zein and yellow zein at a variety of weight percents from 0.4% to 0.05%. With our pure zein as a reference, subtraction of the impure yellow zein was used to quantify the amount of impurities present in zein solutions.
1.8 1.5 1.3 #1 #2
Based on these ﬁndings the spectral trend line for puriﬁed zein was used as the default calculation concerning the A280 nm and A325 nm values. Hence, the A280 nm used in this investigation for zein protein standard is deﬁned by a straight line y = 4.7466x + 0.0281 with an R2 = 0.9994 and the A325 nm contaminant is deﬁned by y = 3.0384x + 0.0032 with an R2 = 0.9987, where x represents the concentration. Removal of the xanthophylls, that cause the yellow color in zein (Sessa et al., 2003), accompanies the removal of diferuloylputrescine. Therefore, x or Ce in our linearized Freundlich isotherm model for the protein components is deﬁned by Eq. (1) and the Ce for the color/odor components is deﬁned by Eq. (2): Ce Ce at k280 nm ¼ at k325 nm y À 0:0281 4:7466 y À 0:0032 ¼ 3:0384 ð1Þ ð2Þ
3.2. Adsorption isotherms Kinetics of adsorptions of protein and color/odor components onto activated carbon at two temperatures, 25 and 60 °C, were performed over a 27 h period to establish the equilibration times needed to maximize their adsorptions. At either temperature, maximum adsorption of protein measured by A280 nm and odor components represented by A325 nm occurred at 24 h and thereafter. With that equilibration time at 25 °C, the A280 nm diminished from 3.34 to 1.41 while A325 nm dropped from 1.78 to 0.17; at 60 °C, the A280 nm changed from 3.29 to 1.07 and A325 nm decreased from 1.75 to 0.07. Table 1 data demonstrate the adsorption capacity (Kf) comparisons and bond strength (1/n) comparisons between temperatures for each wavelength (k). The adsorption capacity for zein protein did not change signiﬁcantly over the entire temperature range from 25 to 60 °C. However, adsorption capacities for the color/odor components did increase signiﬁcantly at 55°, where the Kf value is well above all other values while at all other temperatures evaluated only slight increases were observed.
1.0 0.8 0.5 0.3 0.0 250 275 300 325 350 375 400
Table 1 Adsorption capacity (KF) comparisons and bond strength (1/n) comparisons between temperatures for each wavelength (k) Temp. (°C) A280 k (nm) KF 25 40 45 50 55 60 166.5a 170.8a 190.9a 182.6a 218.2a 193.8a
A325 k (nm) KF 155.3d 177.5cd 224.1bc 196.4bcd 291.2a 233.1b
A280 k (nm)
A325 k (nm)
Fig. 1. Spectral analysis of yellow zein (#1) at 20 mg/ml and colorless/ odorless zein (#2) at 24 mg/ml each in 70% aqueous ethanol over the wavelength region 250 to 400 nm. (For interpretation of the references in colour in this ﬁgure legend, the reader is referred to the web version of this article.)
1.4791a 1.0850ab 1.1477ab 0.8878b 1.1079ab 0.7550b
0.7137aB 0.5335bc 0.6659ab 0.4681c 0.6998a 0.4362c
A KF values within a column (k) followed by the same letter are not signiﬁcantly diﬀerent based on overlap of the 95% conﬁdence intervals. B Slopes (1/n) within a column (k) followed by the same letter are not signiﬁcantly diﬀerent based on overlap of the 95% conﬁdence intervals.
D.J. Sessa, D.E. Palmquist / Bioresource Technology 99 (2008) 6360–6364 Table 2 Simple linear regression equations from linearized Freundlich isotherms for zein protein (A280 nm) and color/odor components (A325 nm) adsorbed to activated carbon at diﬀerent temperatures Temp. (°C) 25 25 40 40 45 45 50 50 55 55 60 60 k (nm) 280 325 280 325 280 325 280 325 280 325 280 325 Equationa Log (qe) = 2.2215 + 1.4791 Log (Ce) Log (qe) = 2.1911 + 0.7137 Log (Ce) Log (qe) = 2.2325 + 1.085 Log (Ce) Log (qe) = 2.2491 + 0.5335 Log (Ce) Log (qe) = 2.2808 + 1.1477 Log (Ce) Log (qe) = 2.3504 + 0.6659 Log (Ce) Log (qe) = 2.2614 + 0.8878 Log (Ce) Log (qe) = 2.2932 + 0.4681 Log (Ce) Log (qe) = 2.2338 + 1.1079 Log (Ce) Log (qe) = 2.4642 + 0.6998 Log (Ce) Log (qe) = 2.2874 + 0.7550 Log (Ce) Log (qe) = 2.3675 + 0.4362 Log (Ce) Adj. R2 0.94 0.98 0.94 0.98 0.93 0.98 0.90 0.98 0.94 0.99 0.91 0.99
a Within each temperature level there were signiﬁcant diﬀerences (p 6 .01) between equations for protein (A280 nm) and color/odor (A325 nm).
SLR equations from linearized Freundlich isotherm model are given in Table 2. Within each temperature level there were signiﬁcant diﬀerence (p 6 0.01) between equations for protein (A280 nm) and color/odor (A325 nm). Figs. 2 and 3 are visual depictions of trend lines from the tabulated data in Tables 1 and 2 for each temperature and their respective correlation coeﬃcients at each wavelength 280 and 325 nm. An increase in adsorption capacity with increased solution temperature denotes an endothermic process. Because diﬀusion is an endothermic process we can conclude that the increased adsorptions observed for the color/odor components of zein with increased heat may be governed by either pore diﬀusion (Anoop Krishnan and Anirudhan, 2002) or the energy required to adsorb onto a nonporous
carbon especially as it saturates. The viscosity of zein solutions in aqueous ethanol decreases with increased temperatures (Fu and Weller, 1999). Diminished viscosity of the aqueous ethanol solution of zein should allow transport of the low molecular weight color/odor components into the interior of the carbon micropores. Bond strength governed by the slope (1/n) for protein showed considerable overlap of values with temperatures, where bond strengths for proteins were highest at 25 °C and lowest at 50 and 60 °C. Bond strength for the odor component at wavelength 325 nm was signiﬁcantly higher at 25 and 55 °C than all other temperatures evaluated. The speciﬁc adsorption temperature of 55 °C for the odor component absorbing at wavelength 325 nm is unusual and may indicate a change in surface phenomenon attributed to the zein molecule. Selling et al. (2007) demonstrated with circular dichroism that changes in the secondary structure of zein’s primary structure occurred upon heating 80% aqueous ethanol solution to 70 °C because mean residue ellipticity decreased 20% at wavelengths 208 and 222 nm. Those changes were reversed upon cooling. The two negative peaks at 208 and 222 nm are indicative of the amount of a-helix present in the protein (Sreerama and Woody, 1994). A decrease in mean residue ellipticity with heating denotes protein denaturation accompanied by a loss of ahelix character. Momany et al. (2006) proposed a model for zein based on molecular dynamics simulations where three lutein molecules ﬁt into the core of the three triad helical segments helping to stabilize the conﬁguration. Conceivably, the loss of a-helix character upon heating would release those xanthophylls, thereby, making them available for adsorption on the activated carbon. Because the binding of the zein color/odor components deﬁned as 1/n is signiﬁcantly higher at 55 °C than those in all other
Linear (60—280) Linear (55—280) Linear (50—280) Linear (45—280) Linear (40—280) Linear (25—280)
y = 0.755x y = 1.1079x y = 0.8878x y = 1.1477x y = 1.085x y = 1.4791x
+ 2.2874 + 2.3388 + 2.2614 + 2.2808 + 2.2325 + 2.2215
R2 = 0.9075 R2 = 0.9355 R2 = 0.8976 R2 = 0.9315 R2 = 0.9404 R2 = 0.9348
1.5 1.0 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0
Fig. 2. Eﬀect of temperatures on the adsorption of zein protein at k280 nm by a coal-based activated carbon with linearized trend lines from Freundlich isotherms.
D.J. Sessa, D.E. Palmquist / Bioresource Technology 99 (2008) 6360–6364
Linear (60—325) Linear (55—325) Linear (50—325) Linear (45—325) Linear (40—325) Linear (25—325)
y = 0.4362x y = 0.6998x y = 0.4681x y = 0.6659x y = 0.5335x y = 0.7137x
+ 2.3675 + 2.4642 + 2.2932 + 2.3504 + 2.2491 + 2.1911
R2 = 0.9905 R2 = 0.9942 R2 = 0.984 R2 = 0.9771 R2 = 0.9776 R2 = 0.98
1.5 1.0 -2.0 -1.8 -1.6 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0
Fig. 3. Eﬀect of temperature on the adsorption of color/odor components of zein at k325 nm by a coal-based activated carbon with linearized trend lines from Freundlich isotherms.
zein aqueous ethanol solutions heated above 25 °C, the unusual adsorption behavior at 55 °C may result from a combination of events involving solvent environment, viscosity changes as well as changes in structure of the zein molecule. Further investigations are needed to prove this hypothesis. 4. Conclusions The Freundlich isotherm model for adsorptions of zein protein and its color/odor components onto an activated carbon proved to be a satisfactory model based on statistical analyses. This model can now be used to investigate the adsorption phenomena of other adsorbents. Application of this model to the zein decolorization/deodorization process to examine the impact of heat on the adsorption capacity of an activated carbon medium demonstrated that processing zein aqueous ethanol solutions at 55 °C signiﬁcantly enhanced the adsorptions of the color/odor components relative to the adsorption of protein. Disclaimer Names are necessary to report factually on available data; however, the USDA neither guarantees nor warrants the standard of the product, and the use of the name by USDA implies no approval of the product to the exclusion of others that may also be suitable. Acknowledgements We thank Mardell L. Schaer and David Baudo for their excellent technical assistance.
Anoop Krishnan, K., Anirudhan, T.S., 2002. Removal of mercury II from aqueous solutions and chlor-alkali industry eﬄuent by steam activated and sulphurised activated carbons prepared from bagasse pith: kinetics and equilibrium studies. J. Hazard. Mater. B92, 161–183. Crittenden, J.C., Luft, P., Hand, D.W., Orvaitz, J.L., Loper, S.W., Arl, M., 1985. Prediction of multicomponent adsorption equilibria using ideal adsorbed solution theory. Environ. Sci. Technol. 19, 1037–1043. Fu, D., Weller, C.L., 1999. Rheology of zein solutions in aqueous ethanol. J. Agric. Food Chem. 47, 2103–2108. Mason, I.D., Palmer, L.S., 1934. Preparation of white zein from yellow corn. J. Biol. Chem. 107, 131–132. McInnis, J., Tang, Q., 2003. Methods and apparatus for recovering zein from corn. US Patent 6,610,831. Momany, F.A., Sessa, D.J., Lawton, J.W., Selling, G.W., Hamaker, S.A.H., Willett, J.L., 2006. Structural characterization of a-zein. J. Agric. Food Chem. 54, 543–547. Moreau, R.A., Nunez, A., Singh, V., 2001. Diferuloylputrescine and pcoumaroylferuloyl putrescine abundant polyamine conjugates in lipid extracts of maize kernels. Lipids 36, 839–844. Moreau, R.A., Hicks, K.B., 2005. The composition of corn oil obtained by the alcohol extraction of ground corn. J. Am. Oil Chem. Soc. 82, 809– 815. Neter, J., Wasserman, W., Kutner, M.H., 1990. Applied Linear Statistical Models: regression, analysis of variance, and experimental designs, third ed. Richard D. Irwin Inc, 1974, 1985, and 1990. Pearce, L.O.G., 1941. Preparation and puriﬁcation of zein. US Patent 2,229,870. Selling, G.W., Hamaker, S.A.H., Sessa, D.J., 2007. Eﬀect of solvent and temperature on secondary and tertiary structure of zein by circular dichroism. Cereal Chem. 84, 265–270. Sessa, D.J., Eller, F.J., Palmquist, D.E., Lawton, J.W., 2003. Improved methods for decolorizing corn zein. Ind. Crops Prod. 18, 55–65. Sreerama, N., Woody, R.W., 1994. Poly (Pro) II helixes in globular proteins: identiﬁcation and circular dichroic analysis. Biochemistry 33, 10022–10025. Starling, D.S., Pinner, S.H., Whitehead, A.D., 1951. Decolorization of zein. GB Patent No. 651,396. Swallen, L.C., 1938. Process for the production of zein. US Patent 2,120, 946.