Scribd Logo
Upload

Welcome to Scribd!

  • Statistcal and Responce Surfaces استخدام زيت السمسم و زيت الخروع

    Uploaded by

    Walaaa Mhm
    5 views12 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    5 views12 pages

    Statistcal and Responce Surfaces استخدام زيت السمسم و زيت الخروع

    Uploaded by

    Walaaa Mhm

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 12
    Materials Today Communications 21 (2019) 100601 Contents lists available at ScienceDisect materalsioday Materials Today Communications journal homepage: www.elsevier.comilocatelmtcomm Corrosion inhibition of mild steel using binary mixture of sesame and castor oil in brine solution car Daniel T. Oyekunle”,, Tomiwa I. Oguntade”, Christiana S. Ita’, Temiloluwa Ojo, Oyinkepreye D. Orodu” hp f Chia Enger Calle: of Bieri. Conn Unvesy, genia sparen of Pac cri, Calle Eager. Coan Une, ena — Fay ar of cane td ara becs veg oon hiro we cea ‘satus byl bn, pouaton cme td Soming econ mre EN) eas Gs canal ronanpeny ha Specmany (OCS) tala showed psec of unre enous he joerdonyad ‘oleic, steric and palmitic acid which are responsible for the corrosion inhibitory properties of the binary mixture. ett on pina non ee of 6% wr predic by Mina 1 atc ware ewe oe ana ‘method at 0.79 M brine concentration, 22.1 mg/L concentration of binary inhibitor and at a time period of 14 Spf ples af a ny i of se an eter alc fe mee intitor ran cnet eso car roan nde non In gh co {Caunoncl ie te ease NeSEM ae nde eb cen te ray mito ‘mise 1. nro inte mates dtrrate aa rest of tech ration wth Corrosion isan irreversible and rapid degradation of metal and alloy ‘withthe environment through an electrochemical or chemical reaction. Corrosion isa destructive phenomenon that affects the beauty of an ‘object, and oftentimes causing structural failure. Corrosion of metas affects its impact on the environment and its economic value [1]. It ‘causes a huge waste of metallic materials which cesults in high e¢o: nomic losses globally. As a consequence ofthis, corrosion has led to significant industrial and academic awareness. Corrosion is a world: wide problem, negatively affecting the advancement of developed and developing counties, According toa study by the National Association ‘of Corrosion Engineers (NACE), in 2011, more than US $2.2 tilion was spent on corrosion. The cost of corrosion in India was estimated to be above US $100 billion, also about US $96 billion was expended on corrosion in South Affica. It was reported that the amount spent on corrosion can be reduced by 35% using the prevailing methods of preventing corrosion (21 ‘Typically, erosion is the degradation of metallic materials by phy seal factors but eorrosion is restricted to harmful attacks on metals only [8]. Most non-metallic materials are susceptible to deterioration by ‘chemical reaction with the immediate environment. Often times, non "Corresponding author. ‘mal adres: daniel oyehunleeovenantuniversy dg (LT. Oyekunle). hnup://do.org/10.1016/.mtomm.2019.100691 the surroundings, for instance, cracks and bulges are observed in plasties, eroding ofthe outer layer is observed on granite surfaces, split observed on wood and leaching of Portland. Corresion is restricted to destructive attack on metals by electrochemical and /or chemical attack [2], However, a frequent and effective means of protecting metallic substrates from corrosion occurs when organic coatings are applied. Organic coatings confer metal protection by the eteation ofa protective layer against oxygen so as to reduce the corrosion rate by lowering the presence of corrosion-causing factors ike water hydrogen and oxygen fon on metal surfaces. Although, as a result of the permissiveness of lorganie coatings to water and oxygen, some active ingredients such as Inorganic additives are added to these inhibitors to further hinder the ‘corrosion rate especially once the protective covering has been bres- ched [4,51 Different types of organic compounds have been applied as an in hibitor to form organie coatings. Often times, organic compounds possess significant influence onthe rate at which adsorption occurs on 1a metal surface, hence they can effectively inhibit corrosion. The eff ciency of organic inhibitors is due to the existence of polar Funetions with §, 0 oF N atoms in the molecule, x eleetons, and heteroeylie compounds. Also, the polar function is referred to as the center for the Received 26 Jay 2019; Received in revised form 6 October 2019; Accepted 8 October 2019 ‘Available onlin 13 October 2019, 252-4998/ © 2019 Evie Lid All rights reserved Dr Opens a Inception of the adsorption process [6]. Most synthetic organte in- hibitors have several hazardous effects and controlled environmental regulations, which had made researchers focus on developing cheap, non-toxie and environmentally favorable natural prodcts as corrosion inhibitors. These natural onganie compounds can be synthesized or ‘extracted from @ medicinal plant, aromatic herbs, and spices. Plants ‘extracts are seen as a remarkable source of chemical compounds that ‘ean be extracted easily at a Tow cost and are biodegradable in nature (7). The use of natural produets extracted from leaves, seeds, and roots ‘of plants as a corresion inhibitor has been documented in te literature [8-111 Numerous extracts from plants have been used as corrosion in hibitors this includes Hibiscus sabdarffa, Azadirachia indica, Teferia ‘occidentalis, Ocimum viridis and Garcinia kola extract [12]. Other plant ‘extracts that have been used are succinic acid (13), ascorbic acid (141, ‘aminoacids 15}, Ponnyroyal ol [15], Justicia gendarussa [7] and caffeic ‘acid (171, In previous studies, castor oll has been used as a corrosion inhibitor on mild stee (18-20) as wel as Sesame oil [21,22]. Castor oil has proved to be an efficient inhibitor as a result ofthe presence of @ functional group with 126, The existence of OH groups in fatty acid ‘chains makes the oil unusually polar {19}. AS a result ofthis inhibiive nature, this study was performed to combine the corrosion inhibitive property ofboth ols (sesame and castor oi) ina binary mixture on mild sel. The morphology and compositions ofthe precipitated film on the nil stel surface in brine solutions containing different concentrations ‘ofthe binary inhibitor (sesame and eastor oil) were studied by SEM and FAR analysis. Polarization tests were also carried out 10 examine corrosion inhibition effets at different immersion times. Statistical models such as Artificial Neural Network (ANN) and Response Surface Methodology (RSM) have been applied to optimize various processes in numerous fields. [23-25]. Response surface methodology’ is a mathematical and statistical tool used to study in dividual and interactional effeets of variables on responses and mostly fitted by polynomial equations such as quadratic, cubic or higher-order functions. Box Behnken design (BBD) was selected from ‘other RSM designs as a result of its efficiency, feasibility, simplicity and also to reduce the numberof test runs required compared o other RSM. ‘design. On the other hand, a computational technique such as ANN imitates biological neural systems like the human brain. RSM and ANN are computer software programs designed to model the correlatic between the independent and dependent variables and are proficient in the modeling of non-linear and complex relationships directly from rave data. The benefit of using ANNs is its capacity to control complex, noisy, incomplete and less constructive data and is ability of parallel processing [25] In this study, response surface methodology (RSM) Box-Behnken ‘design was used to design experimental sets and to investigate the ef fects of main snd interaction of process variables sich as concentration ‘of inhibitor, concentration of the corrosive medium (brine), and time using binary mixture of sesame and castor oil, and the same sets of ‘experimental data was applied for training artificial neural network “The results acquired from both models (RSM and ANN) were compared to ensuce better predictability of inhibition efficiency of the binary in hibitor (sesame and eastor ol) on mild steel in brine solutions 2. Experimental procedures 2.1. Materials 2.1.1. Preparation of plant extract and coupon preparation Sesame and castor oil were obtained from Lagos Stat, Nigeria, Both oils (Sesame and castor ofl) were viscous, the sesame oll was a dark brown liquid while the castor oil was alight yellow liquid. Equal vo: umes of both oils were used throughout this experiment, Mild stel ‘coupons (Mn = 0.03; Cr = 0012; C= 0.253; $ = 0.024; P= 0.013; Si = 0.12; and remainder Fe) of dimensions 20 x 20x 3mm size were Monts Toey Communica 21 (2019) 100691 sed for weight loss measurements and polarization studies, Prior to the experiments, the coupons were polished with 600 grade of emery pa- pers. They were immersed in acetone for 20min to remove grease, washed with distilled water and died in air before immersing in Uhe corrosion medium, 2.1.2. Gas chromatography- mass spectrometry (GEMS) 8 ul ofthe binary inhibitor with equal volumes of sesame and eastor oil were sonicated using n-hexane and analyzed by GCMS-QP2010 Plas Shimadzu, Japan equipped with electron impact ionization mass spec trometer. The oven temperature was programmed at 70-280°C at a linear velocity of 49.2em/see and held for a hold-up time of Smin. Other operating conditions are: split ratio 20.0, detector temperature 280 °, injector temperature 250 °C and Helium (99.99%) was used as the cartier gas. 2.1.3. Fourier wansform infrared spectroscopy (FTIR) FIR instrument, Shimadzu FTIR-84008 was used 10 perform structural characterization of the binary inhibitor. The functional groups were analyzed inthe range of 750 ~ 4000 em” (wavenumbers), and established by cortelation with the standard peak placement of the groups. 22, Methods 22.1. BBD and ANN optimization and design of experiments using weight loss method | Box-Behnken Design (BBD) and Artifical Neural Network (ANN) were applied to give a precise prediction of the correlation between the process input variables (Concentration of brine, Concentration of binary Inhibitor and time) and the output (Inhibition efficiency). Mild steel coupons of 20 x 20 x 3mm size were immersed in 0.5 - 0.9 M of brine, 19-26mg/L of binary inhibitor and 8-21 days as designed by RSM to determine the corrosion rate generated by Box Behnken Design Mita 17 (Fables 1 and 2). Each experimental rum was in tripliate with and without the addition of the binary inhibitor at 27 °C. At the end of the exposure period, the coupons were cleansed and their weight was noted. The following equation by Satapathy [7] was used to determine the percentage inhibition efficiency (IE), 100 Initio tony (1) = (4 ~ a4) 4% : Where M - weight loss of mild stel in uninhibited brine solutions (g) Mz weight loss of mild stee in inhibited brine solutions (3). 22.1.1. Box-Rehnken design. The input process variables were varied into separate levels to note their effect on the corrosion inhibition process. The results obtained from the BBD experimental design matrix was used to derive mathematical model by applying the RSM (quadratic model, The experimental data derived from BBD were used for ANN training using a multilayer normal feed-forward neural network. The design of experiments was built on three variables and four levels as displayed in Table 1, which discussed the actual and coded factors, the total amount of experiments generated by Box: Behnken design was 20. The correlation and interaction between the ‘Table 1 Experimental range an factors sod in Box Behnken design. Cincantaton OF Hie (| a 03 05 a7 09 1 Concentration ofthe Bary Ink Bo 3% w tre.) Tie a) & 2 own Dr Opens a ‘Table 2 Monts Toey Communica 21 (2019) 100691 Experimental design of the weigh los experiment on mild see! in binary solution. on Inhisor (me (ay Eiceney 0) ene (9) on variables were determined by iting the equation of the second-order polynomial to the resulting data obtained. The model quality was ‘evaluated by applying the analysis of variance (ANOVA) and the test of significance. The fitted quadratic response model is given as Eq, (2) YE dy # dN + AY + YH GHP + day} + dk} + dH FdoYH+doUB+e 2 Where ¥is value predicted for the response (inhibition efficiency), dp is the offset term, d;, da, and ds, are the linear coefficients, dy, dz, and ‘day are the quadcatic coefficients, dia, day and das are cross products ‘and e represents the error term. All the experiments were carried out in triplicates and the mean values were used, The design matric of the BBD ‘experimental sets and the responses observed were tabulated in Table 2. Minitab 17 was used for experimental design and data analysis was Performed by BBD response surface design 22.1.2, Anifcial neural networks. In the taining technique, information was performed in the forward direction from the input layer tothe hidden layer, then from the hidden layer tothe output layer derived asthe assigned network (ig. 1), The computational application ‘of the ANN program is to conduct simulation as well as prediction. The intemal variables and structure can be configured and determined as desired in the designed simulation. In this study, multilayer normal feed-forward neural network training with QuickProp (QP) learning algorithm was developed using Neural Power version 2.5 (CPCX software, USA. The connection strength between the inputs, hidden and output layers was determined by ANN parameters such as weights and biases 22.2. Linear polarization resistance tests Linear polarization resistance (LPR) measurements were potento- dynamically executed in + 10mV potential range with reference to ‘open circuit potential applied ata sean rate of 0.5 mV/s. The system response (inhibited and uninhibited mild steel in brine solution) was analyzed through Autolab software, Polarization resistance (P) values were obtained from the current corrosion density and equilibrium corrosion potential. From the measured polarization resistance values, the inhibition efficiency (IES) was determined from the equation: BaR ym ia B @ Where Py and Py are the polarization resistance values with the pre sence and absence of inhibitors 23, Surface analysis ‘The morphologies of the mild steet surfaces in inhibited and unin hibited brine solution were analyzed by scanning electron microscope. Dias Dias Concentration of Brine (M4) Inhibition Volume of Binary cen ais ta 1} Efcimey ae%) Time (days) —+ 1} — Inputlayer “Hidden ayer’ Output layer Fig 1. Structure of ANN architecture fo Inhibition eficeney (8) of il te Dr Opens a Monts Toey Communica 21 (2019) 100691 ‘Transmittance (%) ‘Waventimbers (em?) Fig. 2, FTIR spectra of sesame and castor ol binary inkbitr 3. Results and discussion 3.1 Results 3.1.1. Characterization ofthe binary inhibitor SLLLL. FTIR spectrum and resuls. Fig. 2 illustrates the FT-IR spectrum ‘of the binary mixture of sesame and castor oll. Absorption band ‘observed at 3556.85, 3510.56, 3379.40 and 3286.81 em~" bands are ‘associated with the broad hydroxyl group. The 2194.23em" bands fndicate the presence of ammonium ion in the organie mixture. The peak at 2931.90 em and 2854.74 cm can be assigned to the aliphatic ‘and aromatic stretching mode of C-H groups respectively (71. The 2559.84 em" band is associated with SH. stretching band. The 2191.21 em" band can be identified with the CC= band. The peak at 4748.71 em ‘coreresponds tothe stretching modes of carbonyl groups in organic acids (RCOOH) or ketones (RCOR) and aldehyde (RCOH) or ‘ester (R-———— 00-0). The peak at 1450.52 cm Tis attributed tothe methyl C-H asymmetric bend. The peak for phenol or tertiary aleohol bend and aromatic phosphate stretch are noticed at 1365.65 em~tand 1234.48 em" respectively. The peak at 1157.33cm" implies the stretching mode of C-N. The peaks at 864.14 em! and 717.54em"! implies the presence of aromatic C-H bend. The occurtence of various bands implies thatthe binary inhibitor contains a mixture of various ‘compounds like oils, flavonoids, and alkaloids [7,25] 31.1.2, GC-MS results. Gas chromatography-mass spectroscopy’ (GC-MS) spectra of the binary inhibitor used in this study isshoven in aot Fig. 8 It comprise of 7 major peaks as well as many other small peaks indicating the presence of more than 7 major compounds, The small Peaks may be associated with compounds present in minite quantities fn addition to possible disintegrated major compounds. The 7 major peaks identified are oleic, stearic, methyl ricinoleate, behenic, palmitic, tundecylenic, and nonadceyclic acid. The most predominant was oleic acid. Other peaks with low retention time are largely plant compounds with low polarity such as lupenol, friedelin, and besitosterol, ete Unsaturated compounds such as oleic steric and palmitic acid have been reported for its corrosion inhibitory properties [271, hence, the binary misture was used to perform corrosion inhibition studies. 3.2, Weight loss measurements Corrosion inhibition occurs as result of surface film stabilization fm the steel. The effect of chemically stable surface-active inhibitors increases the activation energy of iron dissolution reaction, thus red cing the surface available for corrosion [7]. The increase weightloss with time resulted in a corresponding increase in the inhibition eff ciency [28] in the presence and absence of binary inhibitor. This in- dicates that insoluble suxface film was not ereated on the electrode surface during mild stee! corrosion in brine solutions [13 Ie implies thatthe binary inhibitor was frst adsorbed on the surface ofthe elec trode impeding corrosion either by blocking the reaction sites or by altering the mechanism of anodic and eathodie corrosion process [28] ‘Therefore, the inhibitive nature of binary mixture in brine solutions results from electrosatic adsorption of a negatively charged Abundance Time (min) Fig. 8. GC-MS of seme and castor oil binary inhibit. Dr Opens a deprotonated mixture of the binary inhibitor to the positively charged ‘electrode surface, thus forming a barrier on the stel surface. The su. face film protects the mild steel against corrosive Tons, such as Cl~ of on” [3] “The design matrix obtained from the Box Behnken response surface ‘optimization tool is used to derive a quadratic model that emphasizes the effect of interactional and individual effect of process variables and their level of significance in the process. The design matrix obtained from the Box Behnken design was utilized forthe ANN training 3.2.1. Molding using Box Behiken design — response surface methodology From the experimental results, an empirical relationship between the response (Inhibition Efficiency) and independent variable (Concentration of brine, Concentration of binary inhibitor and time) in ‘coded terms is expressed in the equation below. Inhibition Eflieney() = -22627 + 31017 Za + 10106 + 10.205Z¢ 198.30 20? — 023409 202 — 049m Ze? = 001668 24 Ze ” ‘Where Za represents the concentration of brine, Zb represents the ‘concentration of Binary inhibitor and Ze represents immersion time, ‘The predicted response values (Inhibition efficiency %) have been determined by using the quadratic model in Eq, (4) and are given in ‘Table 2. From the results presented in Table 2 it ean be affirmed that there is a good agreement between the experimental values and Box Behnken RSM predicted values, therefore validating the accuracy of the proposed model, The analysis of variance (ANOVA) results were sum: marized by the quadratic model demonstrate the adequacy and sig- nificance ofthe model, presented in Table 3, According to Table 3, the F-value ofthe model (6609.52) indicates that the model is statistically significant [29-21]. The coefficient of ‘determination (R) was determined to be 99.97%, which implies that more than 999% of the experimental data were compatible and only about 0.039% ofthe total variation cannot be described by the mode High R? value implies that the tst variations of 99.97% could be fitted by the model Eq, (4). Adjusted R? i used to calculate the model ade ‘quacy, the fitness of the model as well as to rectify the K? value for sample size and forthe st of terms in the model. A high valle of ad Justed R? (99.96%) implies a high model significance. ‘The optimum condition of independent variables selected for the best inhibition efficiency (%) of the model as predicted statistically by Minitab 17 softwage was corrosion inhibition of 86.2% at variable ‘conditions of 0.79M concentration of brine, 2.1 mg/ L concentration ‘of binary inhibitor and ata time period of 14 days. 3.2.2. ANN modelling Several neural network architectures, topologies, and transfer funetions have been identified andl analyzed for prediction of inhibicion Monts Toey Communica 21 (2019) 100691 efficiency ofa binary mixture of sesame andl castor oll in brine soluton (G8). The selection of suitable network architecture, its transfer func- ‘ion, ad its topology is essential forthe successful application of ANN as a ype of transfer funtion used to control the learning rate of @ neural network and its performance. Te number of neurons was in- cited by a heuristic approach involving testing different amounts of neurons and topology until the mean square error (MSE) of the output data was reduced by varying the topology (4:n:3) and the number of neurons (n) used from 7 to 2 [ANN architecture was trained using 1000 as a stopping criterion, ‘The mean square error (MSE) and the coefficient of determination were evaluated to determine the neural network predictive ability. The mean square error (MSE) andthe coefficient of determination were calculated using the equations below. (eae ° 1g 3 : Where, 21 isthe caleulated values, zn is the experimental values, acum I the average experimental values, Nj isthe predicted values, Zj {s the experimental value, n is the number of runs performed in the experiment. MSE value was established as 0.072984 and the coefficient of determination (R*) was 0.99999, The predicted values by ANN was in good accordance with the experimental values. The coeficient of determination (R) for inhibition efficiency is 0.99998, The weights and bias forthe output and hidden layer ae illustrated in Table 4 3.2.3. Experimental analysis of results '3D response surface and contour plots (Figs. 4-6) are the graphi iMustration of the quadratic equations to demonstrate the effect of variables on the responses. From Fig. 4, itean be observed that at low concentration of binary inhibitor and low concentration of rine the Inhibition efficiency achieved was highest. Increasing the concentration Of brine at low concentration of binary inhibitor decreases the inhib tion efficiency, these could be attributed to high concentration brine providing a more corrosive environment for the mild steel and less Concentration of inhibitor is available to slow down the rate of the corrosion process. A high concentration of inhibitor of 23:mg/ L in a brine concentration of 1.1M provides inhibition efficiency of 65% ‘compared tothe same concentration of inhibitor of 23 mg/ Lin lower brine concentration of 07 M with 84% inhibition eficeney. This ean be attributed to the fact that a higher concentration of brine has a more corroding effect than a lower brine concentration, Fig. 5 adduced that at alow reaction time and low concentration of brine the inhibition efficiency was at maximum, Increasing the concentration of brine at ‘table 3 ANOVA table for inhibition efficiency of the binary inhibitor Sree ‘Sum of ares De ‘ontbon | ngs hams Pvalue Pane Mote 05590 7 vo “05590 sro ‘09 52 000 » v8 1 oe 3a 308 3058 om a mast 1 199 26106 2.86 arses 000 Pore ror 000 5 oo 00 000 Ro 99.57%, Adjusted KE — 9.96%, Predicted RE — 99.85%, Where DE = Degree of freedom, Aj 8 = Adjusced stm of wares, Aaj MS = Adjusted Mean squares. Dr Opens a ‘Table + Weights for hidden and outa layer ayer Wels ofthe Biden ayer ‘utp (=) low reaction time reduces the inhibition effcieney, this can be as @ result of @ more corrosive environment provided by a higher con: ‘centration of brine. Although, as time increases and the concentration ‘of brine reduces the inhibition efficiency is at maximum. This ean be ‘attributed to a reduction in the concentration of brine resulting in a less ‘corrosive environment for the mild stel. From Fig. 6, itean be observed that at low reaction ime and lovr concentration of binary inhibitor the inhibition efficiency reduces, and as the time increases and the con- ‘cenration of inhibitor increases te inhibition efficiency increases. For fnstance, inreasing the concentration of inhibitor from 23 t0 20 mg/ I. ‘a a constant period of time 3 days, the inhibition efficiency increases by 22%, This further proves the efficiency of the binary inhibitor in Cite Brae * Concentation of Binary inhi Monts Toey Communica 21 (2019) 100691 Inhibiting the corrosion rate of mild stel in brine solutions “The relative significance ofthe three input variables was evaluated using Neural Power version 2. (CPC-X software, USA) shown in Fig. 7. As displayed in Fig. 7, al of the three variables (Concentration of brine, Concentration of binary inhibitor, and Time witha relative importance 0 58.26%, 29.02%, and 17.72% respectively) have strong effects on the Inhibition efficiency. Therefore, all the variations studied in this work could not be negated inthis current analysis. The degree of potency of Variables was found in the following order of Concentration of brine > Concentration of inhibitor > Time binary 3.2.4. Comparison of ANN anid RSM models ‘The predictive exceution of the ANN and RSM models were com: pared on the bass of coefficient of determination (R°), Relative Percent Deviation (RPD) and Root Mean Square Error (RMSE) given by the ‘model. The general model capability can be validated inthe prediction accuracy forthe authentication ofa data set. The values of R?, RPD, and [RMSE for the ANN model were greater than those for the RSM model as shown in Table 5, This result implies thatthe ANN model for prediction has higher accuracy than the RSM model used for prediction. This also indicates that the experimental data has a high level of accuracy fitted using the ANN model. Be tor (mg/t) * ig. 4. Surtae and contour pots strain the ef of brine concentration (M) and Concent ‘Boncettaton 8 Binary Inhithtor (g/L) om of binary inhibitor (rg/ 1) tnibition efficiency (0, Dr Opens a Bes e238 z : § Monts Toey Communica 21 (2019) 100691 ‘Time (Days) Fig. 8. Surface and contour pts ilstrating the effects of brine concentration (M) and Time (ay) on Inhibition efficeney (8). In the validation of the experimental dataset, R, RPD and RMSE values for the RSM and ANN models are displayed in Table 5. Table 5 showed that both models gave a very high performance of the data set, validation, but ANN performed invariably better than RSM, The per: formance prediction of the ANN model for dataset validation affirms 2 ‘greater capacity for the generalization on the given ease over RSM, Also, Fig. 8 illustrates the predicted and experimental values for each ‘experimental run to obtain the inhibition efficiene. From Fig. 8, iti ‘obvious thatthe trained neural network has an effective approximated ‘experimental values. The higher predictive capability of the ANN can be ascribed fo its potential to generally approximate the system non- linearity, én contrast to RSM which is limited to second-order degree polynomial [29]. Nevertheless, when the ANN technique is usd it must bbe noted that its predictions are limited tothe range of process factors applied in the training process (29,32) 3.3. Effect of immersion time ‘The variation of inhibition efficiency with immersion time is ilu strated in Figs. 5 and 6. Fig. 5 Mlustrates the changes observed in the experimental inhibition efficiency against immersion time and con- centration of brine. At low brine concentration, the inhibition efiiency Increases with time this was in agreement with previous studies [28,34]. Also, at a high concentration of brine, the inhibition efficiency increases as time increases. A similar occurrence was also experienced in Fig 6. Te inhibition efficiency inereases asthe immersion time was increased at a low concentration of binary inhibitor. Similarly, in creasing the concentration of binary inhibitor to maximum, the in hibition efficiency inereased as the immersion time increased but after 10 days of immersion the inhibition efficiency decreases indicating that the inhibition potential of inhibitors decreased as a function of theit exposure in aid [35]. Overall, for this study, the relative importance of jimmersion time is 17.72% which wns lower than the concentration of brine and concentration of binary inhibitor of 53.26% and 29.02% re spectively (ig. 2) 3.4. Porentiodynamic polarization ‘The effects of a binary mixture of sesame and castor oil on the be: havior of anodic and cathodic polarization of mild steel in brine Dr Open, ea Monts Toey Communica 21 (2019) 100691 é ; Experimental inhibition | aBRBRSESRSRSAS = i i: ze [ “pus ae ae See “3 us os ur «ea Fig. 6. Surface and contour plats ilstrating the fet of Concentration of Binary inhibitor (ng/L) on Tie (Day) Cone. of Brine (M) Cone. of Binary Inhibitor (may) Time (Days) Importance (6) Fig. 7. Relative importance ofthe variables on the response Dr Opens a ‘Table S Validation data for RSM and ANN mode! developed sing PD and RMSE. ‘once of tiation) Relative percent devistion RFD) oot mea square ene BSE) RSM R?=09997 ¢ *. EXPERIMENTAL NKBION EFICIENCY (4) [EXPERIMENTAL INHIBITION EFFICIENCY (x) 0 30 7 0 30 9 «7 90 EFICENCY Fig. 8. Comparing experimental and predicted values fr RSM (lf) and ANN igho. solution have been investigated by polarization measurements and re corded as displayed on the Tafel plots in Fig. 9a and b for 0.3M and 4.1 M concentration of brine respectively. I is ilustrated from the data presented in Table 6 that both cathodic hydrogen evolution and anodic ‘metal dissolution of mild steel reaction were inhibited with the addition of the binary mixture of sesame and castor oil to 0.3M and 1.1 M of brine solution. The inibitory natuce of thie reaction was mare visible withthe further addition of the binary inhibitor. Fig, Sa and b illustrate the effect ofthe binary inhibitor (equal vo umes of sesame and castor ols) on the kinetics of mild steel corrosion inhibition proces in 0.3 and 1.3 M of brine solution. Inhibitor added in brine solutions influence the corrosion reactions and could be observed by the changes inthe corrosion potentials as well as alterations in both ‘cathodic and anodic polarization curves. The equilibrium corrosion Potentials (Epa) Is altered towards more negative values in the pre- ‘sence of inhibitor. Also, dereasing the catholic polarization curves was ‘above the anodie polarization curves. Both observations imply thatthe fhibitor reduces the cathodic corrosion reactions successfully than ‘anodic corrosion reaction [261. Further analysis showed that the po larization curves were slightly altered when different amounts of i hibtor were added fn 0.3 M brine solution (Fig. Sa). This reveals that the amount of inhibitor added has litle effect on the inhibition me- chanism, ‘The technical terms associated with the kinetis of corrasion pro: ‘ess, such as, equilibrium cotrosion potential (Eps), corrosion eurtent density (hear, anole Tafel slope (b.) and cathodic Tafel curve slope (was calculated from the polarization curves at different inhibitor ‘concentrations and an overview was stated in Table 2. Observing the values of Evy showed thatthe addition of more ofthe binary inhibitor ‘caused the corrosion potential of the mild steel to shift towards a greater negative potential with reference 10 Fen in 0.3 and 1-1 M of Monts Toey Communica 21 (2019) 100691 Lg (Nem i “ }| ih "8 Snot E.vagagco b 2. e 2 eS 7 — arn wom E,V gage Fig 9. The log of current agsnst potential for 02M Brine. b. The log of curtent against potent for 1.1 M Brine brine solutions. According to Riggs and others [37], 2 compound is classified as anodie or eathodic inhibitor based on the displacement of Epon f the displacement is > 85 mv in reference 10 Eyer the inhibiCor can be reported as a cathodic or anodic inhibitor, otherwise, it wil be classified as a mixed type inhibitor. ln this study, the maximum change observed in the corrosion potentials was 60.42 mV and 117.81 mV in 10.3M and 1.1 M brine solution respectively. The binary inhibitor used in this study can be classified as a mixed corrosion inhibitor in mod cxately low concentration range for brine solutions this was corrobo- rated by Popoola 122] for sesame oils. Thus, this type of binary in- hibitor simultaneously aets as anodic and eathodic inhibitorsin a higher concentration of brine solutions The values of anodic and cathode (by and b,) slopes presented in Table 6 were changed significantly as the concentration of inhibitor, this suggests that inhibitor adsorption on the ‘metal surface probably affects the mechanism at which mild ste! dis solves in brine solutions, “Table 6 illustrate that Further addition of inhibitor reduces the corrosion curent in concordance with the concentration of inhibitor ‘sed, The values of corrosion current density were obtained by extra: polation of cathodic Tafel (b,) lines to the respective free corrosion Potential (Ea) [27]. The lowest js Yale was deduced for both concentrations (0.3 and 1.1M brine) at 300mg /1. concentration of binary inhibitor. However, using 0.3M of brine the corrosion current value was lower than in 1.1 M of brine, showing the binary mixture was ‘more effective corrosion inhibitor in 0.3 M of brine solution than 1.1 M brine solution. The bass for mild steel corrosion inhibition of mild stoel {in brine solutions ean be as a result of the resistance values of linear polarization (Rt). The resistance of the mild steel polarization in brine solutions was significantly inereased in the presence of the inhibitor (Table 6) Consequently, corrosion inhibition was realized. On the other hhand,i¢ can be dectared that the adsorption of inhibitor molecules on the mild steel surface atthe active eathodie and anodie sites made the Dr Opens a Monts Toey Communica 21 (2019) 100691 ‘Table 6 Kinetic parameters obtained from Tafel los of mild steel immersed in 03M and 1.1 M Brine containing the binary inhibitor. Tie Sain Coco inary nie m/) Ena (OV IRAVARC ewe WAR!) Ry GV dee) (aW/ ded) Rp Gem) metal (mild steel) more protected from corrosion (polarization) re- Table? sulting into corrosion inhibition It can be summarized thatthe binary inhibitor (sesame and eastor ‘il preferentially adsorbed on the active sites of the mild steel surface resisting corrosion reaction by improving the electrode insulation po: tential (low corrosion current). Inhibitor concentration improves the inhibition efficiency accompanied by a low change in corrosive solu- tions. Larger surface coverage and low rate of adsorption of inhibitor molecules can be adjured by the stated fact. ‘3.8, Electrochemical impedance spectroscopy (EIS) Electrochemical impedance spectroscopy was used to obtain more {information about the kinetics of mild stel corosion in the presence of the binary inhibitor mixture (sesame and castor oil). The electo- ‘chemical process takes place at the open-circuit potential and was ex ‘amined by the electrochemical impedance spectroscopy. EIS measure: ments of the mild stel electrode at its open circuit potential were ‘observed after immersing in brine solution alone for 15 min and in the presence of different concentrations of the binary mixture. These ex: periments were performed over a 10k to 10m frequency range. Fig. 10 illustrates the Nyquist plots for mild stool dissolution in unin hibited and inhibited brine solution with different concentrations ofthe binary inhibitor. lt was observed that the width of the Nyquist plots Increased as the concentration of the binary inhibitor increased, in- icating the strengthening of the inhibitive film. Furthermore, the Ny ‘quist plots present imperfect semicircle which accure due tothe rough electrode surface and surface homogeneity [38]. The EIS spectrum re- ‘corded fr mild steel in 1 M brine solution at 27 °C (Fig. 10) showed one ‘depressed capacitive loop. This same ttend was also observed in mild steel immersed in 1M brine containing binary inhibitor (15-30 mg/1). ‘The depressed semicirele with the center under the real axis is a char ‘acteristic behavior ofa solid electrode attributed to the roughness ofthe surface and inhomogeneity ofthe metal cletrodes [7] EIS experiment was performed for mild steel immersed in 1M Brine solution uninhibited and inhibited with binary inhibitor at 27 °C as shown in Fig. 10. The respective kinetic parameters are displayed in co er) 100180 Zp (Ohms cm’) Fig. 10. Nyquist plot for mild tel in LM brine uninhibited and inhibited with ferent concentration of the binary inhibitor. 20280 EIS parameters obtained for mild sel in 1 M Brine uninhibited and inhibited with binary inhibitor Teter Gaeieig/h) Reem) Ry Cem!) Cine ) IE Where Re ste solution esistance,K ste charge anser resistance and Gy {the double layer capacitance ‘Table 7. The EIS plots of mild ste! in brine solution uninhibited and Inhibited with the Binary inhibitor at 27 °C showed an inductive Toop in low frequency (LF) region while it displays a depressing capacitive oop at high-frequency region (HE). Thus indicates the occurrence of a Far aie process on free electrode sites, From ES illustrated in Table 7, the charge transfer resistance (Ry) values increases as the concentration of binary inhibitor increase ths indiating the formation intermediates responsible for the anodie controling process from the metal dissol tion and subsequently inhibiting corrosion. twas also observed tha the Re values increase whereas the double-layer capacitance (Gx) values decrease in the presence of the binary inhibitor at diferent con centrations. The inerease in the Re, and decrease in Gy, values in in hibited solutions are due toa rection in local dielectric constant and/ ‘or to an increase inthe thickness of the double layer [39]. These rests imply that the binary inhibitor in this study inhibits mild steel corrosion by adsorbing on the metal/electrode interface [38]. The adsorption of the binary inhibitor on the electrode reduces the electrical capacity because it displaces water molecules and other fons intially adsorbed ‘on the surface. The depletion of this capacity with inereasing binary Inhibitor concentration may be due to the formation of a protective layer at the surface (7). The maximum value of Ra attained was 207.62 em for binary inhibitor at 27 °C. The reslls obtained in this stdy shouted thatthe binary mixture of sesame and castor cil acts as a good coerosion inhibitor at 27 °C in brine solutions. 6, Surface characterization ‘To ascertain whether corrosion inhibition occurs asa result of the formation of a protective lm by the binary inhibitor on the metal surface, SEM images were taken. Fig. 11 illustrates the surface mor phology ofthe SEM analysis of the examined samples. The igure shows the mierogeaph of mild steel examined after exposure 100.5 and 1.1 M of brine solution. The figures reveal the SEM images after immersion, severe damage on the surface occurs asa result of material dissolution starting on grain boundaries located between perlite and ferite. Grain boundaries aro well known to be active sites swhere lattice and dis location defeets accumulate, therefore the deposition of from oxide loxyhydroxide is often initiated on the grain boundaries to be slowly spread evenly on the whole metal surface, The rate at which mild steel degrades depends on the presence of a good protective film on its surface Opt eat. Monts Toey Communica 21 (2019) 100691 ig, 1. SEM images of coded mild tel i bine solution, 4. Conclusion ‘The results presented in this study (weight loss tests, polarization tests, and SEM micrography) demonstrates the use of a binary mitre ‘of sesame and eastor oll a8 a good corrosion inhibitor. Using Minitab Statistical software to determine the weight lose measurements, op ‘imum condition of the independent variable selected by the software for best inhibition efficiency was corrosion inhibition of 86.2% at 0.79% brine concentration, 22.1 mg/L concentration of binary in hibit and at a ime period of 14 days. The polarization studies de- ‘onstrate that binary mixture of sesame and castor oil act as a mixed type inhibitor in moderately low concentration of brine solutions while ‘at high brine concentration it simultaneously aets as anadie and ‘cathodic inhibitors. The study suggests that adsorption of inhibitors on the metal surface probably affects the mechanism at which mild stecl ‘dissolves in brine solutions Declaration of Competing Interest ‘The authors declare that they have no known competing financial interests or personal relationships that could have appeared to itl ‘ence the work reported in this paper. References [n} Rani Yous H.Yozane,Sytbeis, aacteeaton and ssesent of Daly (etaneco-pttsdrved meat la acolo cong or aida ste op. Org. Cost 75 (10) (20) 1454-1454 (Vom, EE Eien MAA Quran ge eon nd satan cor ‘von nies for mead los eerie, 1 Ma Lg 248 2007) in comrte 1 Min Mer Chara Eay 2 (2019) 271-276 Foc antcrsin protrton, rg. Ong. Coat 63 (208) 307515, {nhs Ila endaren pln eect archaea ston, Cree Ss com aaae-2n, fect of mite, Maer Chem. Py 99 (208) 4 ‘tin corn irs, Cou 8 30 08) 1428 cy 3 a fo io ts no on oy oa, os 06 on os om 01 ey a fa ea es 8) wm es coy 01 wn 1 Opie rahaton of te init fect of ae pln exact on the aid ferro of mld sted, Co. Se 0 (2008) 2953 tes, plaszton, ES, F2C, EDX and SEM sta, Eeceot. Ata 52 C207) iQ. thang, QR. Ca LX Go, KY. Le, lt fei, tiene al tunic eld an the corr copper in rite hyo cl Cos St 30 (208) {ans Se 52 (2005) 164, fae al extract eee, a Chem. S14 (1) (216) 105-127 SIA Catandeye A.0. leat, Ui Et gan, Ivetton of he we of IM bala zany Kalo, ¥ mad, Nal ls as conasin in ‘tos for tne tel Cham, Technol, Fe Ol 482) O13) 40-52 Samu nce 2M HCL / SON inert, in Heceacbem Se 7 (2012) BY. Ajodel, CX: Cheng. Modeling en opinion of syns production eam Siagemeration nng epose me met and atl eel eter ‘Sition:s novel pprsth or ea” inh aston, NCS Omeg 8208) {Sern o 1018 cath Mel ne cd slo, Coron So 48) 206), ‘rom sesame (Sesrnum indicum) ol using barium hydrosde ss heterogems (Gots) a1 28, Bene BO, eal Btn dan Gm yee el . Ogunade, 0. Rotini, A Mojsle, A Slomoa, G,Angye, Experimental dataset of ten ne palmer and ening optician, Data Beef 208) Dr Open, ea {52} Kat Rave JM, Avaric PS. le, OS Stankovic ¥.Velhov ‘ptinieaon of utasound-asisted base

    You might also like