Translate this page into:
Adsorption of aflatoxin B1 on magnetic carbon nanocomposites prepared from bagasse
⁎Corresponding author. Mobile: +92 3416062388. mohammadzahoorus@yahoo.com (Muhammad Zahoor),
-
Received: ,
Accepted: ,
This article was originally published by Elsevier and was migrated to Scientific Scholar after the change of Publisher.
Abstract
A novel adsorbent for the removal of aflatoxin from poultry feed was prepared from bagasse and was characterized by surface area analyzer, SEM, XRD, FTIR, TG/DTA and EDX. A specially designed chamber was used for the preparation of the adsorbent. SEM, XRD and FTIR analysis showed the iron oxide presence on the adsorbent surface. The adsorption parameters were determined for aflatoxin adsorption using Freundlich and Langmuir isotherms. The equilibrium time was 115 min for 200 ppm at pH 3 while 150 min at pH 7. At high pH there was a decline in percent adsorption. Best fit was obtained with pseudo first order kinetics model for the kinetics data of adsorption. The value of ΔS0 (30.67 kJ mol−1 deg−1) was positive while that of ΔH0 (−5.9 kJ mol−1) and ΔG0 (−9.303, −9.610, −9.916 and −10.226 kJ mol−1 correspond to 30, 40, 50 and 60 °C) was negative. The increase in ΔG0 values with temperature showed that the adsorption process was favorable at high temperature.
Keywords
Aflatoxins
Adsorption
Equilibrium time
Bagasse
1 Introduction
Mycotoxin contamination in feed is a worldwide problem. Among mycotoxins, aflatoxins are considered one of the most hazardous toxins encountered in animal feeds, throughout the world (Williams et al., 2004). Aflatoxins are derivative of difurocoumarin, synthesized by various toxigenic species of Aspergillus. Natural contamination of aflatoxin in the poultry feed results in severe economic losses to the poultry industry as poultry birds are much susceptible to this toxin as compared to other animals. Fungal growth and mycotoxin production is very common and frequent in poultry feeds as reported by survey where the range was from 1.0 to 900 mg/kg (Devi et al., 2002; Mohanamba et al., 2007).
Chemically aflatoxins are difurocoumarin and divided into two main classes. One class includes difurocoumarolactone series of aflatoxin G1 and G2, while the other is difurocoumarocyclopentenone series containing aflatoxin B1, B2, B2A, M1, M2, M2A and aflatoxicol. Though various methods have been employed to reduce or detoxify aflatoxins in poultry feed however no single treatment has been found to be effective against aflatoxins (Binder, 2007; Varga and Toth, 2005). Several chemicals like oxidizing and reducing agents, acids, bases, salts and chlorinating agents have been investigated for their activity to degrade or detoxify aflatoxin in animal feeds (Amezqueta et al., 2009; Awad et al., 2010; He et al., 2010; Jouany, 2007; Kabak et al., 2006; Varga and Toth, 2005). However these treatments are associated with negative effects also. A recent approach to detoxify mycotoxins contaminated products has been the use of nonnutritive adsorbents in the diet to reduce the absorption of mycotoxins from the gastrointestinal tract. It has been reported by many workers that adsorbent in feed can selectively bind mycotoxins during digestion and pass harmlessly from the gastrointestinal tract of animals. Major advantages of adsorbents are low cost, safety and easiness to add in animal feed (Avantaggiato et al., 2005; Dakovic et al., 2003; Khadem et al., 2012; Teleb et al., 2004).
Magnetic adsorbents have gained importance in water treatment processes due to the fact that they can be easily separated from the media after treatment by simple magnetic process (Oliveira et al., 2000; Zahoor and Mahramanlioglu, 2011).
Due to rising number of information on contamination of aflatoxin in feeds, there is great demand for simple, practical and cost effective detoxification methods. In present study magnetic carbon nanocomposites were prepared, characterized and used for the detoxification of aflatoxin B1 in in vitro.
2 Material and methods
Bagasse was used as biomass to produce magnetic carbon nanocomposite. The biomass was shredded and transferred to ethanolic FeCl3·6H2O (10% w/v). After 30 min biomass was separated and air dried at room temperature for 24 h. Thereafter, biomass was charred in a specially designed chamber (designed by the author) consisting of a stainless steel container equipped with an electric heater, wire gauze, nitrogen gas (to provide a nitrogen rich atmosphere) inlet and exhaust outlet (Fig. 1).
Diagram of chamber used for preparation of Iron oxide carbon nanocomposites.
To determine pore distribution and surface area of the adsorbent 0.1 g of adsorbent was taken and analyzed by Surface Area Analyzer (NOVA 2200e Quantachrome, USA) using N2 as purge Gas.
The adsorbent was also characterized by Joel X-ray Diffractometer JDX-3532 with Ni filter, using monochromatic Cu Kα radiation of wave length 1.5418 Å. The X-ray generator was operated at 40 kV and 30 mA. The scanning range 2θ/θ was selected. The scanning speed 10 min−1 was employed for precise determination.
Infrared spectra were collected by using Fourier transform infrared spectrometer (IR Prestige-21 Fourier transform infrared spectrophotometer, Shimadzo Japan). The prepared adsorbent was scanned ranging 750–525 cm−1 and 4000–600 cm−1.
A topographic analysis of adsorbent was done placing the prepared sample on SEM grid and gold coated through sputter coater (SPI, USA) AT 30 mA for 2 min. The images were then taken by Joel JSM-5910 type scanning electron microscope at an accelerating voltage of 20 KV.
Thermo gravimetric and differential thermal analysis of the prepared magnetic carbon nanocomposites was done by Diamond Series TG/DTA Perkin Elmer, USA analyzer using Al2O3 as reference.
The prepared magnetic carbon nanocomposite elemental study was carried out with EDX coupled with SEM JSM-5910 (JEOL. Japan) model INCA 200, X-sight Oxford Instrument U.K. The dispersed particles were sprinkled onto a double sided sticky tape which was mounted on a microscopic stub of aluminum.
The kinetics of aflatoxin B1 adsorption was carried out as per method of Stroka et al. (2000). Briefly, a series of 25 ml flasks (two sets) were taken and each added with 5.4 ml ethanol and 13 ml distilled water (one set having pH 3 while other having pH 7). Each flask was spiked with known amount of standard aflatoxin B1 (to achieve 200 ppm), thereafter previously prepared adsorbent was added so as to achieve a level of 0.5% (w/v) in each flask. All the flasks were orbit rotated at 300 rpm at room temperature. The sorbents were separated from each flask at predetermined time intervals using magnetic bar followed by final filtration using Whatman No 1 prior to HPLC analyses.
The HPLC system consisted of Hitachi model L-200 equipped with two pumps L-2130, auto injector L-2200 and fluorescence detector L-2458 (Macoa, Japan) that were used for aflatoxin B1 analyses. Degassing of the mobile phase acetonitrile/methanol/water (8:27:65, v/v/v) was done by sonication. The column inertsil ODS-3 (25 cm × 4.5 mm I.D., 5 μm, GL science, Tokyo, Japan) was connected as LC column. The column was maintained at 40 °C with a flow rate of 0.8 ml/min, while the injection volume was 20 μl. The aflatoxin B1 was detected at the excitation and emission wavelengths of 365 nm and 450 nm, respectively. For quantification purpose, a separate calibration curve was established. Triplicate samples were used for setting the calibration curve, determining LODs and extraction recovery.
A separate series of 25 ml flasks were prepared as mentioned in Section 2.2, except with variation in aflatoxin B1 standard so as to achieve the desired levels (150, 175, 200, 225, 250, 275, 300, 325, and 350 ppm). All the flasks were shaken at 300 rpm for 480 min at room temperature. The sorbents were separated and analyzed as mentioned above.
To optimize pH for aflatoxin B1 adsorption a series of 25 ml flasks were prepared as mentioned in Section 2.2, with pH ranging from 1 to 14 and spiked with standard aflatoxin B1 to achieve concentration of 200 ppm prior to addition of sorbent. All the flasks were shaken at 300 rpm for 240 min followed by sorbent removal and LC analyses as discussed above.
A series of flasks were prepared as mentioned in Section 2.2 and subjected to similar experimental conditions except for temperatures that were 30, 40, 50, and 60 °C for each set for a period of 240 min. Thereafter, separation and analyses were carried on as mentioned above.
3 Results and discussion
3.1 Characterization of magnetic carbon nanocomposites prepared from bagasse
The magnetic carbon nanocomposites were prepared from bagasse using the specially designed assembled chamber depicted in Fig. 1.
The prepared nanocomposites were subjected to a bar magnet. Attraction toward the bar magnet indicated the formation of magnetic carbon nanocomposite.
3.1.1 Characterization of the prepared adsorbents by surface area analyzer
The surface area and pore distribution of the prepared adsorbent are depicted in Figs. 2 and 3 respectively while different surface parameters are given in Table 1. The BET surface of the prepared adsorbent is lower as compared to the commercially available activated carbons. The reduction in surface area in the case of this novel adsorbent was due to deposition of iron oxide. The reduction in surface area due to deposition of iron oxide has already been reported by Oliveira et al. (2000) and Zahoor and Mahramanlioglu (2011).
Graphical representation of BET surface area of the prepared adsorbent from bagasse.
Graphical representation of pore distribution in the prepared adsorbent from bagasse.
BET surface area (m2/g)
97.07
Langmuir surface area (m2/g)
502.16
Total pores volume (cm3/g)
1.71
Micropore volume (cm3/g)
0.65
Average pore diameter (A0)
135.91
3.1.2 Characterization of the prepared adsorbents by X-ray diffractometer (XRD)
The XRD analysis of adsorbent prepared from bagasse showed that the resulting material contained the nano-crystalline iron oxide (Fig. 4). The XRD patterns of the carbon nanocomposite indicated that iron oxide is deposited on the activated carbon. The X-ray diffraction method is used for structure determination as well as for the measurement of particle size (Cullity, 1974). The results of Fig. 3 revealed the synthesis of Fe3O4 and Fe2O3 as impurities. The diffraction peaks at 2θ of 29.7, 35.7, 44.9, 54.15, 57.55 and 62.5 represent the corresponding indices 220, 311, 400, 422, 511 and 440 planes of cubic unite cells which correspond to the magnetite structure reported by Krehula and Music (2007), Liu and Kim (2009). The diffraction peaks at 2θ; 33.2, 40.95 and 49.55 are related to goethite (α FeOOH) and hematite as impurities (Krehula and Music, 2007). The magnetic carbon nanocomposite size was obtained using Debye–Scherer’s equation (3.1). By using the formula the composite sizes were found to be in the range of 70–350 nm.
XRD pattern of iron oxide carbon nano composite prepared from bagasse.
3.1.3 Characterization of the prepared adsorbents by Infrared Spectroscopy
Infrared spectroscopy is the most widely used technique for iron oxide characterization. It is a useful technique to show information about crystal morphology, nature of surface hydroxyl groups and adsorbed H2O. The magnetic carbon nanocomposites prepared from bagasse show broad band in the region of 1000–1200 cm−1 attributed to C–C and C–O stretching while 596.65 cm−1 can be attributed to Fe–O stretching of deposited magnetite in the carbon nanostructure (Figs. 5 and 6). The IR spectroscopy demonstrated the presence of Fe3O4 in the carbon phase of the prepared magnetic carbon nanocomposites (Kahani et al., 2007).
Far IR spectra of iron oxide carbon nanocomposite prepared from bagasse.
Mid IR spectra of iron oxide carbon nanocomposite prepared from bagasse.
3.1.4 Characterization of the prepared adsorbents by scanning electron microscope
The morphology of iron oxide carbon nanocomposites prepared from bagasse is shown in Fig. 7 with low and high magnification. Scanning electron microscopy observation shows some differences in size and shape of the composites. The white patches in the images show the crystallization of the iron oxide, while the black portions represent the carbon. SEM monograph also reveals aggregation of particles which was due to the moisture contents absorbed in the sample. It is also observed from the images that the shape of the Fe3O4 appears somewhat cubical whereas the sizes of iron oxide carbon nanocomposites estimated were found in the range of 60–300 nm.
SEM images of iron oxide carbon nano composite prepared from bagasse at different magnification.
3.1.5 Characterization of the prepared adsorbents by thermo gravimetric and differential thermal analyzer
Differential thermal analysis of adsorbent prepared from bagasse (Fig. 8) shows an endothermic and exothermic peak; also indicated mass losses at two stages. First mass loss was observed in the range of 30 °C–70 °C with DTA endothermic peak, while further mass loss was observed in the range of 250–720 °C with DTA exothermic peak.
TG/DTA curves of iron oxide carbon nano composite prepared from bagasse.
3.1.6 Characterization of the prepared adsorbent by electron dispersive X-ray
Energy dispersive X-ray analysis of iron oxide carbon nanocomposite prepared from bagasse is represented in Fig. 9. It clearly indicates the presence of Iron (Fe), Oxygen (O) and Carbon (C). A small peak of Calcium was also observed as impurity. As stated earlier O-Kα, Fe-Lα, Fe-Kα and Fe-Kβ peaks show the presence of magnetite deposition in the composite which is in conformity to Kahani et al. (2007).
EDX spectra of iron oxide carbon nano composite prepared from bagasse.
3.2 In vitro study of magnetic carbon nanocomposite prepared from bagasse
3.2.1 Adsorption isotherm
The adsorption of aflatoxin B1 on the adsorbents prepared from bagasse was studied using Giles isotherm (Giles et al., 1960) that was obtained by plotting C (concentration) verses q (amount of aflatoxin adsorbed) as shown in Fig. 10. Based on initial slope and curvature Giles classified the isotherms into constant partition (C), high affinity (H), Langmuir (L) and sigmoidal (S) types. Isotherm in Fig. 10 is C type in which the availability of adsorption sites remains constant at all concentrations up to saturation. It is characterized by the constant partition of contaminant between solution and substrate up to maximum possible adsorption. The linearity of curves indicate that the number of adsorption sites remains constant and as adsorption progresses more and more sites are created; which is the case when strong attraction of solute instead of solvent exists for adsorbent. The solute then break enters the substrate bonds thereby interfering with solvent penetration.
Giles isotherm for the adsorption of aflatoxin B1 on iron oxide carbon nanocomposites prepared from bagasse.
The adsorptions of aflatoxin B1 on the prepared adsorbent was quantified by using Langmuir (1918) and Freundlich (1906) adsorption isotherms. Langmuir adsorption isotherm is based on the assumption that the maximum adsorption corresponds to a saturated monolayer of solute molecules on the adsorbent surface, having no interaction with molecules adsorbed from lateral sides. The linear form of Langmuir isotherm is given as:
Langmuir plot for the adsorption of aflatoxin B1 on magnetic carbon nanocomposite prepared from bagasse.
The Langmuir constants Q0 and b were obtained from the slope and intercept of the plot of specific adsorption (C/q) against equilibrium concentration and found to be 66.68 and 0.25 respectively with R2 value of 0.991.
The Freundlich isotherm is usually used to describe heterogeneous systems and is represented by the following equation.
The values of Freundlich constants K and 1/n were calculated from the slope and the intercept of the ln C verses ln q plot. The K and 1/n values from Fig. 12 were 9.3 and 0.714 respectively with R2 value 0.998 for adsorbent prepared from bagasse (Table 2).
Freundlich plot for the adsorption of aflatoxin B1 on magnetic carbon nanocomposite prepared from bagasse.
Adsorbent prepared from:
Langmuir isotherm
Freundlich isotherm
Q0 (mgg−1)
b
R2
K
1/n
R2
Bagasse
66.68
0.25
0.991
9.3
0.714
0.998
The result shows that Freundlich adsorption isotherm for both adsorbents fitted the data better than Langmuir isotherm (Table 2), as demonstrated by higher coefficients of determination values (R2) obtained.
3.2.2 Adsorption kinetics
The time of contact required to reach equilibrium for an adsorbent is an important factor in adsorption processes. The fast uptake of aflatoxin occurs within the first few minutes as initially the adsorbent sites are free and more available for the adsorption of aflatoxin as depicted in Fig. 13 for bagasse based nanostructures that gives time t verses C plot for 200 ppm aflatoxin solutions at pH 3 and pH 7. As time progresses more and more sites are occupied and adsorption process becomes slow. Finally a saturation point is reached that corresponds to equilibrium time of adsorption. The equilibrium time for the adsorption of aflatoxin B1 (200 ppm) on the adsorbent prepared from bagasse at pH 3 and 7 is 115 and 150 min respectively. The reason for carrying out the experiments at pH 3 and pH 7 was that the adsorbents will also be subjected to in vivo studies at the stomach and intestine.
Effect of contact time of aflatoxin adsorption on magnetic carbon nanocomposites prepared from bagasse.
Pseudo first order (Lagergren, 1898) and second order (Ho and Mckay, 1998) adsorption kinetics equations were used to analyze the adsorption kinetics data. The pseudo first order equation is given as:
Pseudo first order kinetics plots for the adsorption of aflatoxin B1 on iron oxide carbon nanocomposites prepared from bagasse.
Adsorbent prepared from:
Concentration (ppm)
pH
Pseudo first order kinetics model
Pseudo second order kinetics model
ka
R2
K2
R2
Bagasse
200
3
0.023
0.99
0.00074
0.986
200
7
0.034
0.98
0.0053
0.973
The pseudo second order kinetics equation is given as follows
By plotting t/q versus t a straight line was obtained and the values of K2 and q were calculated from intercepts and slopes of the plot as depicted in Fig. 15 for adsorbent prepared from bagasse. The values of K2 and R2 for the prepared adsorbents are given in Table 3.
Pseudo second order kinetics plots for the adsorption of aflatoxin B1 on iron oxide carbon nanocomposites prepared from bagasse.
Table 3 shows that the adsorption kinetics of aflatoxin B1 at two different pHs and same agitation speed could be best explained in terms of the pseudo first order rate equation with precision in the correlation coefficients, while pseudo second order rate equation did not reflect the experimental results.
3.2.3 Effect of pH
pH is an important factor that affects the adsorption process since it affects the surface charge of the adsorbent, and the degree of ionization and speciation of the adsorbate. The effect of pH on aflatoxin B1 adsorption is shown in Fig. 16 for adsorbent prepared from bagasse. Fig. 16 shows that the amount of aflatoxin B1 adsorbed is not much affected from pH 1 to pH 9 whereas above this a decline was noted.
Effect of pH on adsorption of aflatoxin B1 on iron oxide carbon nanocomposites prepared from bagasse.
3.2.4 Adsorption thermodynamics
To determine the adsorption thermodynamics, adsorption experiment was carried out at 30, 40, 50 and 60 °C. The Vant Hoff equation was utilized to determine ΔH0 and ΔS0 of the adsorption process.
K is the distribution constant of adsorption, ΔH0 is the enthalpy change, ΔS0 is the entropy change, T is temperature in Kelvin while R is universal gas constant. The value of ΔH0 was calculated from the slope while ΔS0 was calculated from intercept of the ln K and 1/T plot (Fig. 17) for adsorbent prepared from bagasse was noted to be −5.9 kJ mol−1 and 30.67 kJ mol−1 deg−1. The positive value of ΔS0 shows that there is an increase in the randomness in the system solid/solution interface during the adsorption process while the negative value of ΔH0 indicates that the adsorption of aflatoxin B1 on the prepared carbon nanocomposites is an exothermic process.
Vant Hoff plot for the adsorption of aflatoxin B1 on magnetic carbon nanocomposites prepared from bagasse.
The values of standard free energy ΔG0 were calculated from equation:
The values calculated from equation, −9.303, −9.610, −9.916 and −10.226 kJ mol−1 correspond to 30, 40, 50 and 60 °C respectively for nanocomposites of bagasse. The negative values of ΔG0 at various temperatures signify the spontaneous nature of the process and a high affinity of aflatoxin B1 for the prepared adsorbents. The increase in ΔG0 with the rise in temperature indicates that the process of adsorption is more favorable at high temperatures.
From the in vitro results it is concluded that the prepared adsorbents can be used as an alternative of powdered activated carbon for the detoxification of aflatoxin in poultry feed as the former causes dehydration and salt deficiencies when administered to poultry birds. In the subsequent in vivo experiments the adsorbent was tested for the purpose mentioned.
4 Conclusions
In this study magnetic carbon nanocomposites were prepared from bagasse and characterized by SEM, XRD, FTIR, TG/DTA and EDX. SEM, XRD and FTIR analysis showed the iron oxide presence on the adsorbent surface. The equilibrium time was 115 min for 200 ppm at pH 3 while 150 min at pH 7. At high pH (9) there was a decline in percent adsorption. Best fit was obtained with pseudo first order kinetics model for the kinetics data of adsorption. The value of ΔS0 was positive while those of ΔH0 and ΔG0 were negative. The increase in ΔG0 values with temperature showed that the adsorption process was favorable at high temperature. From the results it is concluded that the prepared adsorbent can be used as an alternative of powdered activated carbon for the detoxification of aflatoxin in poultry feed as the former causes dehydration and salt deficiencies when given to poultry birds. In our in vivo study the adsorbent will be tested for the purpose mentioned.
References
- Recent advances on the use of adsorbent materials for detoxification of Fusarium mycotoxins. Food Addit. Contam.. 2005;22:379-388.
- [Google Scholar]
- Decontamination and detoxification strategies for the Fusarium mycotoxin deoxynivalenol in animal feed and the effectiveness of microbial biodegradation. Food Addit. Contam. Part A. 2010;27:510-520.
- [Google Scholar]
- Managing the risk of mycotoxins in modern feed production. Anim. Feed Sci. Technol.. 2007;133:149-166.
- [Google Scholar]
- Element of X-ray Diffraction. Reading, MA: Addison Wesley; 1974.
- Adsorption of ochratoxin A on octadecyldimethyl benzyl ammonium exchanged–clinoptilolite–heulandite tuff. Colloids Surf. B Biointerfaces.. 2003;30:157-165.
- [Google Scholar]
- Occurrence of aflatoxins B and ochratoxin A in Indian poultry feeds. J. Food Prot.. 2002;65:1338-1340.
- [Google Scholar]
- Über die adsorption in lösungen (Adsorption in solution) Z. Phys. Chem.. 1906;57:384-470.
- [Google Scholar]
- Studies in adsorption. Part IX. A system of classification of solution adsorption isotherms and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids. JCS. 1960;30:3973-3993.
- [Google Scholar]
- Chemical and biological transformations for detoxification of trichothecene mycotoxins in human and animal food chains: a review. Trend Food Sci. Technol.. 2010;21:67-76.
- [Google Scholar]
- Methods for preventing, decontaminating and minimizing the toxicity of mycotoxins in feeds. Anim. Feed Sci. Technol.. 2007;137:342-362.
- [Google Scholar]
- Strategies to prevent mycotoxin contamination of food and animal feed: a review. Crit. Rev. Food Sci. Nutr.. 2006;46:593-619.
- [Google Scholar]
- Kahani, S.A., Hamadanian, M., Vandadi, O., 2007. Deposition of Magnetite Nanoparticles in Activated Carbons and Preparation of Magnetic Activated Carbons. Nanotechnology and Its Applications, First Sharjah International Conference. American Institute of Physics (978-0-7354-0439).
- Evaluation of the effectiveness of yeast, zeolite and active charcoal as aflatoxin absorbents in broiler diets. Global Vet.. 2012;4:426-432.
- [Google Scholar]
- Krehula, S., Music, S., 2007. Formation of Magnetite in Highly Alkaline Media in the Presence of Small Amounts of Ruthenium, Croatia Chemica ACTA CCACAA.80,517–527.
- Zurtheorie der sogenannten adsorption geloesterstoffe. Kungliga Svenska Vetenskapsakad. Handlingar. 1898;24:1-39.
- [Google Scholar]
- The adsorption of gases on plane surfaces of glass, mica and platinum. J. Am. Chem. Soc.. 1918;40:1361-1403.
- [Google Scholar]
- Solvothermal synthesis and magnetic properties of magnetite nanoplatelets. Mater. Lett.. 2009;63:428-430.
- [Google Scholar]
- Activated carbon/iron oxide magnetic composites for adsorption of contaminants in water. Carbon. 2000;40:2177-2183.
- [Google Scholar]
- Immunoaffinity column cleanup with liquid chromatography using post-column bromination for determination of aflatoxins in peanut butter, pistachio paste, fig past, paprika powder. Collaborative study. J. Assoc. Anal. Chem. Inter.. 2000;83:320-340.
- [Google Scholar]
- Novel cubic magnetite nanoparticle synthesis using room temperature ionic liquid. Eur. J. Chem.. 2012;9:1070-1076.
- [Google Scholar]
- Efficiency of kaolin and activated charcoal to reduce the toxicity of low level of AF in broilers. Sci. J. King Faisal Univ.. 2004;5:1425-1428.
- [Google Scholar]
- Novel strategies to control mycotoxins in feeds: a review. Acta Vet. Hung.. 2005;53:189-203.
- [Google Scholar]
- Human aflatoxicosis in developing countries: a review of toxicology, exposure, potential health consequences and interaction. Am. J. Clin. Nutr.. 2004;80:1106-1122.
- [Google Scholar]
- Removal of phenolic substances from water by adsorption and adsorption–ultrafiltration. J. Sep. Sci. Technol.. 2011;46:1482-1494.
- [Google Scholar]