Application of Soft Computing for Evaluation of Bismuth Adsorption onto Coconut Skin in Kinetic Aspects

Mohammad Gheibi^*
*Correspondence: Mohammad Gheibi, Department of Civil Engineering, Ferdowsi University of Mashhad, Iran, Tel: 989152198721, Email:

Keywords

Bismuth • Coconut skin • Adsorption synthetic • Nonlinear regression • Simulink • Genetic algorithm

Introduction

Bismuth is a weak, trivalent, white crystalline, heavy and brittle metal with a slight pink tinge and is chemically similar to arsenic and antimony. Bismuth is the most magnetic of all metals and has less thermal conductivity than all elements except mercury. Bismuth compounds are used in cosmetic, pigment and medicinal manufacturing [1]. Bismuth has unusually low toxicity compared to other heavy metals, and bismuth alloys have been used as substitutes for lead since lead toxicity was identified in humans. However, it should be noted that bismuth itself can lead to many acute and chronic epidemiological effects. Therefore, different adsorbents and methods are used to remove or pre-concentrate bismuth from water and wastewater sources. Madrakian et al. [2] used a fast, accurate and sensitive method based on spectrophotometry to measure bismuth ions. In this method, bismuth ion is adsorbed on the activated carbon surface after reacting with thiourea and bromide ion in an acidic environment. After concentrating the complex, it is absorbed by a solution of bromide ion in dimethylformamide solvent and measured at 375 nm. The results showed that the calibration curve for measuring bismuth ion is in the range of 1x10^-7-1.5x10^-7 mol/L and its detection limit is 8x10^-10 mol/L. Also, the effect of other disturbing ions on the bismuth ion absorbance was studied and finally this method was used to measure very small amounts of bismuth in aqueous samples. Afkhami et al. [3] also used cloud point extraction and spectrophotometric measurements to measure small amounts of bismuth. In this method, after mixing the bismuth ion with the mixing agent of red bermogallol, it was extracted in 114X ion-ion micelles surfactant medium. By optimizing various parameters that affect the extraction efficiency, it was found that the calibration curve of bismuth measurement is linear in the range of 4.6-120 μg/L and the detection limit of its measurement is 2 micrograms per liter. Also, the disturbance effect of other ions was investigated and finally, this method was used to pre-concentrate and measure bismuth in urine samples. Zanavaras et al. [4] used a flow injection method with a spectrophotometric detection system to measure small amounts of bismuth ions. In this method, methylene blue has been used as a mixing agent with bismuth, and after the formation of the methylene blue-bismuth complex; the color intensity was measured at a wavelength of 548 nm. Parameters affecting extraction efficiency were studied and it was found that under optimal conditions the calibration curve of bismuth measurement is linear in the range of 0-75 mg/L. Finally, this method was used to measure bismuth in drug samples and to evaluate the accuracy of the method; the results were compared with results obtained from the flame atomic absorption device. Rastegarzadeh et al. [5] used a liquid-liquid extraction method based on scattering and ultravioletvisible spectrophotometric measurement system to pre-concentrate small amounts of bismuth. In this method, after the complexion of bismuth with iodide ion, the ionic complex of bismuth-iodide is formed; this becomes non-ionic after reaction with methyl triactylammonium chloride and is extracted in carbon tetrachloride solvent. By optimizing different parameters affecting extraction efficiency, it was found that the calibration curve of bismuth measurement is linear in the range of 5-4 micrograms per liter and has a detection limit of 1.6 micrograms per liter. Finally, this method was used to measure bismuth ions in tablets and blood serum samples. Due to its high selectivity and the availability of a flame atomic absorption device, this technique is widely used to measure bismuth. In 2015, Chamsaz et al. [6] used a liquid-liquid dispersion microextraction method based on ion-pair formation and a graphite furnace atomic absorption measurement system to pre-concentrate very small amounts of bismuth. In this method, thiocyanate ion was used to form an anionic complex with bismuth, and after reaction with the cationic surfactant steel pyridinium chloride, the complex was ionized and extracted in carbon tetrachloride extracting solvent. Various parameters affecting the extraction efficiency were studied and in optimal conditions, it was found that the calibration curve of bismuth measurement is in the range of 0.8-3 micrograms per liter. Finally, this method was used to measure very small amounts of bismuth in aqueous and serum samples. Shemirani et al. [7] used the cloud point extraction method to pre-concentrate small amounts of bismuth. In this method, after mixing with di-tizone ligand, bismuth ion is extracted in Michelle Triton 114 X-medium. Also 20 μl is injected and measured into the graphite furnace atomic absorption device after dilution by tetrahydrofuran solvent. When parameters affecting the extraction efficiency were optimized, results showed that the calibration curve of bismuth measurement is linear in the range of 0.4-0.6 μg/L and the standard deviation of bismuth measurement is less than 5%. Finally, this method was used to pre-concentrate and measure small amounts of bismuth in aqueous and biological samples (blood and hair serum). Tekman et al. [8] used solid phase extraction method and atomic absorption measurement system of graphite furnace in order to pre-concentrate and measure small amounts of bismuth, lead and nickel. In this method, a syringe containing silica gel modified with 3-aminopropyl trimethoxysilane was used as the adsorbent. After adsorption, the ions were adsorbed by a decomposition column with HCl 2 mol/L solution and injected into a graphite furnace atomic absorption device for measurement. Shakerian et al. [9] used solid phase extraction method and hydride production measurement method for pre-concentration and measurement of small amounts of bismuth. In this method 2-mercaptobenzothiaz placed on the surface of alumina coated with sodium dodecyl sulfate, which traps bismuth ions, used as the decomposition column. The bismuth ions are then adsorbed by a 2 mol/L hydrochloric acid solution containing thiourea and transferred to the reduction cell to form hydride and reduced by the addition of sodium borohydride. By optimizing different parameters, it was determined that the calibration curve of bismuth measurement is linear in the range of 40-1250 ng/L and the standard deviation is 2.3%. Finally, this method was used to measure small amounts of bismuth in different water samples.

In a fundamental study, Chaudhari [10] also removed bismuth from water sources using coconut shells. It also evaluated the chemical kinetics at different times. This study aims; (i) first, using nonlinear regression relations determine the bismuth adsorption reaction order (based on Chaudhari research) and then (ii) calculate the bismuth adsorption reaction constant using a combination of MATLAB software simulation environment and fixed genetic algorithm.

Materials and Methods

Bismuth Adsorption Using Coconut Shell

As mentioned earlier; Chaudhari provided a method to remove bismuth ion (Bi+3) using the coconut shell. Curves regarding the capacity and reaction kinetics of bismuth adsorption by coconut shell are shown in (Figures 1, 2).

Calculating the Order of Adsorption Reaction

To calculate the order of the chemical reaction, regression analysis of changes in the remained bismuth concentration over time was used as described in (Table 1).

Determination of Bismuth Adsorption Kinetic Coefficient Using a Combination of Simulink and Genetic Algorithm

There are several methods for analyzing a differential equation; In this study, the simulation of ODE 45 method in MATLAB 2013b Simulink environment and optimization of the total squared error have been implemented using a genetic algorithm (GA) (Figure 3) [11].

In the next step, the initial values for the synthetic coefficients of each reaction are considered and the difference between these values and the coefficients calculated from the real measurements is optimized by the genetic algorithm using the cost function of the total squared error.

Genetic algorithm → min (Ro-Rd)²

Initial population = 50

Number of iteration = 50

Mutation rate = 0.2

Crossover probability = 0.8

Results and Discussion

At the beginning of the analyzes in this section, the data depicted in Figure (2) were separated by Extraction (x-y) software, and the initial calculations of the zero, first, and second reaction order was performed, the results of which are summarized in (Table 2). Also, diagrams related to the distribution of residual concentration over time in reactions with orders of zero, first, and second are shown in (Figures 4-6) respectively.

In the following, regression analyzes have been performed in the ToolPak Analysis environment of Excel2013 software and the results for grades zero, first and second are described in (Table 3).

As shown in the table above; the quadratic reaction produces the highest values of the regression coefficient (R-Square) and correlation. On the other hand, the error values resulting from data fitting (Standard Error) in the quadratic reaction create better conditions than the other two cases. It should be noted that the P-value of the second-order adsorption reaction is less than 0.001 and fulfills the appropriate conditions in terms of analysis of variance. The comparison of the predicted values of the models (zero, first, and secondorder) with the real values is shown in (Figures 7-9).

Considering the three above figures, it can be concluded that the data in the quadratic regression function have the highest convergence with the real data obtained from the experiment and the quadratic model can be considered as the basis for the order of reaction of bismuth adsorption by coconut skin.

Results of the kinetic coefficient evaluation of the bismuth adsorption reaction showed a value of 0.0076 L/mg.h, while in the linearization of the measured data for second order reactions this coefficient was estimated to be equal to 0.007 L\mg.h. As a result, an error of 8% can be considered for the linearization method. The results of the convergence of the objective function in the GA are depicted in (Figure 10).

Conclusion

Bismuth with the Bi symbol in the periodic table has an atomic number of 83 and a mass of 98/208. Its density is equal to 10.05 grams per cubic centimeter and the melting and boiling temperatures are 271 and 1564 degrees Celsius, respectively. Bismuth is a weak, trivalent, white crystalline, heavy and brittle metal with a slight pink tinge that is chemically similar to arsenic and antimony. This mineral has significant epidemiological effects on humans which make the point of being treated and removed from the water or any other sources which are consuming by human beings necessary. This study investigated the kinetics of bismuth removal by coconut shell based on the Chaudhari research. The results showed that the process of adsorption of bismuth by coconut shell is of second order. Also, a constant reaction coefficient of 0.0076 was determined after soft calculations in the combination of MATLAB and GA Simulink environment. Meanwhile, an 8% error was observed in the linearization calculation of the constant reaction coefficient.

References

1. Teepker, Michael, Hajo M. Hamer, Susanne Knake, Oliver Bandmann, Wolfgang H. Oertel, and Felix Rosenow. "Myoclonic encephalopathy caused by chronic bismuth abuse." Epileptic Disord 4 (2003): 229-233.
2. Madrakian, Tayyebeh, Abbas Afkhami, and Akram Esmaeili. "Spectrophotometric determination of bismuth in water samples after preconcentration of its thiourea–bromide ternary complex on activated carbon." Talanta 60 (2003): 831-838.
3. Afkhami, Abbas, Tayyebeh Madrakian, and Hajar Siampour. "Cloud point extraction spectrophotometric determination of trace quantities of bismuth in urine." J Braz Chem Soc 17 (2006): 797-802.
4. Tzanavaras, Paraskevas D., Demetrius G. Themelis, and Anastasios Economou. "Sequential injection method for the direct spectrophotometric determination of bismuth in pharmaceutical products." Anal Chim Acta 505 (2004): 167-171.
5. Rastegarzadeh, Saadat, Nahid Pourreza, and Arash Larki. "Dispersive liquid–liquid microextraction for the microvolume spectrophotometric determination of bismuth in pharmaceutical and human serum samples." Anal Methods 6 (2014): 3500-3505.
6. Vakilzadeh, Samira, Mohammad Eftekhari, Mahmoud Chamsaz, and Farzaneh Javedani-Asleh. "Ion pair based dispersive liquid–liquid microextraction for the preconcentration of ultra-trace levels of bismuth (III) and its determination by electrothermal atomic absorption spectroscopy." Anal Methods 7 (2015): 7653-7658.
7. Shemirani, Farzaneh, Majid Baghdadi, Majid Ramezani, and Mohammad Reza Jamali. "Determination of ultra trace amounts of bismuth in biological and water samples by electrothermal atomic absorption spectrometry (ET-AAS) after cloud point extraction." Anal Chim Acta 534 (2005): 163-169.
8. Tokman, Nilgun, Suleyman Akman, and Mustafa Ozcan. "Solid-phase extraction of bismuth, lead and nickel from seawater using silica gel modified with 3-aminopropyltriethoxysilane filled in a syringe prior to their determination by graphite furnace atomic absorption spectrometry." Talanta 59 (2003): 201-205.
9. Shakerian, Farid, Ali Mohammad Haji Shabani, Shayessteh Dadfarnia, and Mahboubeh Kazemi Avanji. "Hydride generation atomic absorption spectrometric determination of bismuth after separation and preconcentration with modified alumina." J Sep Sci 38 (2015): 677-682.
10. Chaudhari, U. E. "Coconut shell: a carrier for the removal of bismuth from aqueous solution." Int J Chem Sci 7 (2009): 71-79.
11. Goharimanesh, Masoud, and Aliakbar Akbari. "Optimum parameters of nonlinear integrator using design of experiments based on Taguchi method." J Comput Appl Mech 46 (2015): 233-241.