Optimization of process parameters for the green synthesis of silver nanoparticles using Plackett-Burman and 3-level Box–Behnken Design

The present work describes the synthesis of silver nanoparticles (AgNps) from leaf extracts of Piper betle and Jatropha curcas using a green approach. Green synthesis of nanoparticle is superior over other methods of synthesis as it is ecofriendly and cost effective. The phytochemical components in the leaf extract play a vital role in reducing the AgNO 3 and hence synthesizing the silver nanoparticles. During this reduction activity, the several factors which affect the synthesis of nanoparticles were investigated and optimized according to the yield of nanoparticles. The experimental conditions investigated were pH, temperature, pressure, time of reaction, microwave radiation, UV radiation and concentration of plant extract, silver nitrate & sunlight. Satisfactory yields of silver nanoparticles synthesis from plant extract were obtained by using optimum conditions when compared to conventional synthesis of nanoparticles. The optimization parameter was later A mathematical model was formulated to correlate the interactive inuence of the parameters and the signicant reduction. Plackett-Burman design (PBD) indicated that concentration of plant extract, concentration of silver nitrate and sunlight were the major parameters affecting the synthesis of silver nanoparticle. The mutual interactions of these variables are mapped in the design by 3 Box-Behnken design (BBD). The signicant factors and their interactions in the green synthesis were examined by analysis of Variance (ANOVA). The result indicated that the BBD was a good predictive model for the experimental results. Though the plant extracts are different, the characterization of the synthesized nanoparticle after optimization of parameters showed uniform size and shape i.e. spherical shape and size of 41 nm. of any other independent variables. P-value suggest us the signicance of the parameter for the synthesis of nanoparticle


Results And Discussion
Green synthesis of Silver nanoparticle: The reduction activity of Ag + to Ag 0 is shown by the color change from light green to brown color. The concentration is calculated by a calibration graph measuring the optical density at 430 nm. Figure 1 a) shows the plant extract of yellow color and b) shows the change of color to brown showing the reduction activity of the plant extract to silver nanoparticles. The phytochemical components present in the leaf extracts enable a quicker reduction of ions when compared to other biological methods. Flavones, organic acids and quinines present in the leaf extract are water soluble phytochemicals responsible for immediate reduction of Silver ions. These were con rmed through the phytochemical analysis and was found that in Piper betle Flavones were 32.5 mg/g, organic acids 8%(w/w) and quinines about 43 mg/g. Similarly, the phytochemical components in Jatropha curcas Flavones were 37 mg/g, organic acids 12% (w/w) and quinines about 39 mg/g.
Optimization of Silver nanoparticle: The optimization parameters in uencing the silver nanoparticle synthesis were analyzed using Plackett burman design (PBD) and Box Behnken Design (BBD).
The results obtained through the preliminary experiments were analyzed using statistica software.

Plackett-Burman design (PBD):
Plackett-Burman experimental design is most simple, less time consuming and well-established system for screening different variables and selection of most signi cant variable and conditions for better results. In our study, a total of 9 independent variables were screened using Plackett-Burman experimental design ( Table 1).
The experiment was conducted in 12 runs to study the effect of the selected variables on the synthesis of silver nanoparticles. (Table 2). In present study, Plackett-Burman experimental design showed a markedly wide variation in yield of silver nanoparticles :19.57-41.53µg/mL in Piper betle extract and 20.46-41.51µg/mL in Jatropha extract. This variation is mainly due to the effect of different optimization parameters on the synthesis of Silver nanoparticles and it has re ected the importance of optimization of variables to increase the yield of silver nanoparticles. The maximum yield of 41.53 µg/mL from Piper Betel extract and 41.51µg/mL from Jatropha extract was achieved in the run number 1, while the minimum yield of 20.46 µg/mL from Piper betle extract and 19.57 µg/mL from Jatropha extract were observed in the run number 12. The Run1 has X 1 , X 3 , X 7 , and X 8 as the high value parameters and X 2 , X 4 , X 6 and X 10 as low value parameters. The results show that pH, Pressure, concentration of plant extract and silver nitrate had a greater impact on nanoparticle synthesis. The plant extract has the reducing agent and the silver nitrate has the silver ion which is a major component for the silver nanoparticle synthesis. Pressure is also a contributing factor as increase in pressure in uence the dynamic movement of the molecules resulting in more interaction and thereby synthesizing more nanoparticles. Change in pH contributes the H + ions there by changing the conformation of the reducing agent and hence allowing increasing the yield of nanoparticles. Also the Run 12 has the minimum yield on silver nanoparticle and all the parameters were kept at low level (-1) at this trial. Table 3 show the effect of each independent variable on the silver nanoparticles synthesis. The values shows the determination of the effect of each variable and a large numerical value either positive or negative indicates that a factor has a large impact on synthesis; while an effect close to 0 means that a variable has little or no effect. With respect to the table we can see that seven variables out of nine namely, concentration of AgNO 3 , pH, temperature, time, microwave radiation, sun light exposure have positive affect whereas other two variables namely, concentration of plant extract and UV radiation exposure. Regression coe cients estimate the unknown parameters and describe the relationship between a predictor variable and response. The sign of each coe cient indicates the direction of the relationship between a predictor variable and the response variable. Standard error is a measure of the statistical accuracy of an estimate, equal to the standard deviation of the theoretical distribution of a large population of such estimates. Main effect is the effect of one independent variable on the dependent variable. It ignores the effects of any other independent variables. P-value suggest us the signi cance of the parameter for the synthesis of nanoparticle The Pareto chart shows the order of effects of nanoparticles synthesis ( Figure 4). The length of each factor is proportional to the absolute values of the estimated effects. Therefore these terms are supporting hierarchy. The nal decision on the most effective parameters requires the assessment of half-normal probability plot. Furthermore the positive and negative signs showed that the response is improved from the low to high level or not.
The Pareto chart shows 7 signi cant and 2 non-signi cant variables (Table 3) affecting the synthesis of silver nanoparticles. The leaf extract was the most signi cant variable with p value = 0.04 (Piper betle) and 0.0002 (Jatropha curcas) as leaf extract has the major reducing agent in it, followed by silver nitrate with p-value = 0.04( Piper betle) and 0.0003(Jatropha curcas) as silver nitrate contribute the silver ions for nanoparticle synthesis, sunlight exposure with pvalue = 0.07 (Piper betle) and 0.001(Jatropha) as sunlight serves as a photocatalyst for the reduction of silver ions ,pH with p-value = 0.015 (Betel) and 0.011(Jatropha curcas) as pH changes the structure of the reducing agent , Pressure with p-value = 0.19(Piper betle) and 0.003(Jatropha curcas) and it in uences the rate of , UV radiation with p-value = 0.20(Piper betle) and0.0005(Jatropha curcas) , MW radiation with p-value 0.47(Piper betle) and 0.01(Jatropha curcas) as it can proceed with the reduction activity with minimum time, Temperature with p-value = 0.57(Piper betle and 0.502(Jatropha curcas) as temperature increases the rate of interaction but the reducing agent may degrade at higher temperatures, and time with p-value =0.86(Piper betle) and 0.0095(Jatropha), as time increases the synthesis of silver nanoparticles increases and after a while attains saturation due to non availability of silver ions or the reducing agent. Table 4 is on ANOVA analysis and F-test. Analysis of variance (ANOVA) is used to determine whether there are any statistically signi cant differences between the means of three or more independent variables. ANOVA uses F-tests to statistically test the equality of means. The degrees of freedom associated with SSR will always be 1 for the simple linear regression model. The values from the ANOVA table shows that they are statistically signi cant. Table 5 shows the observed and predicted values of Silver Nanoparticles synthesis from the two leaf extracts Piper Betle and Jatropha curcas. The half normal probability plot was also used to identify the signi cant parameters. In this plot, The factors having effects near the straight line through zero are not signi cant while those deviating from the straight line are signi cant. [28] Figure 3 shows the half normal plot, can be used to which the algorithms of Length and Dong are applied to identify signi cant effects and to estimate standard deviation of the effects. Signi cant effects in half normal plots are detected through visual inspection (29). The slope of the line through the effects assumed to be non-signi cant gives an estimate of the standard deviation (σ) of the error. From Figure 2 it can be concluded that plant extract, Concentration of Silver Nitrate solution and sunlight has greater impact on AgNp synthesis. The silver nitrate solution serves as the major donor of the silver ion contributing for the synthesis of silver nanoparticles. Sunlight act as a photo catalyst in the reduction of silver ion. The plant extract has protein, nitrate reductase which acts as an reducing agent in the reduction process [30]. Therefore, at high and low pH, less accumulation of nanoparticles may occur due to the aggregation of the protein structure [31,32] The three parameters (plant extract, Concentration of Silver Nitrate solution and sunlight) which have a greater in uence on the silver nanoparticle synthesis were selected for Box Behnken design.

Box Behnken Design (BBD):
A Box-Behnken experimental design with 3 independent variables at 3 different levels, low (-1), medium (0) and high (+1) was used to study the effects on dependent variables. A Box-Behnken experimental design has the advantage of requiring less experiments (15 batches) than would a full factorial design (27 batches). Table 6 shows the Box-Behnken experimental design and yield of silver nanoparticle obtained through experimentation.
The results of ANOVA analysis are depicted in Table 7. The F value, p-value, MS value indicate that the relation between the response and the selected parameters is statistically signi cant [33]. The adequacy of the model developed was evaluated based on the correlation coe cient R 2 and standard deviation value [34]. The closer the R 2 value of unity and smaller the standard deviation implying more accurate response and repeatability that could be predicted by the model. The R 2 and R 2 -(adj.) for Piper betle were obtained as 1 and 1 whereas for Jatropha curcas were obtained as 1 and 0.9999 respectively. The compatibility of R 2 to R 2 means a good adaptation of the theoretical values for the experimental data of the model [35]. The model can be considered as a practical model for the prediction of the factors within the tested ranges.
By applying multiple regression analysis on the experimental data, the following second order polynomial equation was found to explain the silver nanoparticle synthesize from piper betel (3) and Jatropha curcas (4) plant extract by only considering the signi cant terms and is shown below: The value of the correlation coe cient (R 2 ) of Equation (3) and (4) was found to be 0.99 and 1 respectively, indicating a very good t. The results clearly indicate that the Y (Yield of Silver nanoparticles) value is strongly affected by the variables selected for the study. This is also re ected by the wide range of values for coe cients of the terms of Equation 3 and 4. The main effects of X 1 , X 2 , and X 3 represent the average result of changing 1 variable at a time from its low level to its high level. The interaction terms (X 1 X 2 , X 1 X 3 , X 2 X 3 , X 1 2 , X 2 2 , and X 3 2 ) show how Y changes when 2 variables are simultaneously changed.
The negative coe cients for all 3 independent variables indicate an unfavorable effect on Y, while the positive coe cients for the interactions between 2 variables (X 1 ,X 2 and X 3 2 ) indicate a favorable effect on Y. Among the 3 independent variables, the lowest coe cient value is for X 2 X 3 in equation 3 and 4, indicating that this variable is insigni cant in prediction of Y.
The relationship between the dependent and independent variables was further elucidated by constructing contour plots. Figure 4 A shows the Response surface plot and contour plot of silver nanoparticle synthesis from Piper betle extract and (i) shows the effect of Silver nitrate (mM) and plant extract on Silver Nanoparticle synthesis. It was obvious that when silver nitrate was at a low level, the effect of plant extract was negligible.
When the silver nitrate concentration was at a higher level, silver nanoparticle synthesis steadily increased with increasing Y as 55 µg/mL.   [36], ii) the plant extract which is the biological reducing agent due to the presence of the nitrate reductase in synthesizing the silver nanoparticles [37] and iii) sunlight which act as a photo catalyst in the synthesis of silver nanoparticle [38,39]. This method of synthesis of silver nanoparticles through a green approach proves to be much effective and cheap method as it requires the natural plant extract and sunlight as a catalyst.
The observed and the predicted yield of silver nanoparticle synthesis is shown in Figure 5 from a)Piper betle and b) Jatropha curcas plant extract versus the predicted yield of silver nanoparticles and found that the BBD predictions are more accurate and near to the regression line suggesting the superiority of BBD in predicting silver nanoparticle synthesis. The clustering and concordance of points around the diagonal line con rms the compatibility of the model to predict the experiments.

Characterization of Silver nanoparticles by SEM:
SEM analysis gives the size and shape of nanoparticles produced. Average particle size of silver nano-particles from Jatropha curcas was found to be 42 nm whereas from Piper betel, it was found to be 41 nm. The nanoparticles produced were of uniform size and spherical in shape ( Figure 6). Thus through optimization of the critical factors stable uniform sized and shaped particles were generated.

Conclusion
In the present study, stable silver nanoprticles were green synthesized via biological reduction method with the help of Piper betle and Jatropha curcas extract. The major factors affecting the sythesize of nanoparticles were optimized using statistical tools. Statistical optimization of factors affecting the synthesize of silver nanoparticlea using Plackett-Burman and Box-Behnken design appears to be a valuable tool for the syntheis of silver nanoparticle from Piper betle and Jatropha curcas. This study on varying nine different parameters on the synthesis of silver nanoparticle by Plackett Burman shows the optimum paramters to be chosen while synthesing the silver nanoparticle. It is very clear from the statistical model that the parameters like plant extract, silver nitrate concentration and sunlight has greater in uence on the synthesis of the silver nanoparticle. This was further analyzed by Box Behnken design. The plant extract serve as a source of reducing agent, silver nitrate solution serve as a donor for silver ions and sunlight acts as a catalyst enhancing the activation energy of the molecules and thereby giving more yeild for the nanoparticle synthesis. The yeild of silver nanoprtucles can be increased by chosing the proper parameters and maintaining the optimal conditions. The BBD shows an obvious interaction between the parameters and hence proving the signi cance of the model.
More stable and uniform sized nanoparticle were synthesized by optimizing the critical factors for the green synthesis of silver nanoparticle. Thus proving the green synthesis of silver nanoparticle is a cost effective, economic and stable method of synthesis.

Materials And Methods
Green synthesis of Silver nanoparticle: The synthesis of silver nanoparticles was carried out from two leaf sources, Piper betle and Jatropha curcas under the optimized conditions. The leaves were washed thoroughly and air dried to remove the moisture. Phytochemical analysis were carried out to nd the components in the leaf extract. 10 gm of the dried leaves were then crushed in mortar and pestle by adding adequate water making a solution of 100ml. This solution was then ltered using Whatman No.1 lter paper. The synthesized nanoparticles were collected by centrifugation and washed twice with distilled water to remove any contaminants.
Optimization of Silver nanoparticle: The optimization of variables for the synthesis of silver nanoparticles was carried out by Plackett Burman design. Nine factors (pH, temperature, pressure, time of reaction, microwave radiation, UV radiation, concentration of plant extract, silver nitrate and sunlight) were used for analysis. All factors were tested in triplicates (S-1 to S-24). The maximum and minimum limits of experimental factors were determined in preliminary experiments. Aqueous solution of silver nitrate (AgNO 3 ) was prepared in distilled water. The leaf extract of Piper betel and Jatropha curcas was added to AgNO 3 solution and stirred continuously.
Color change of the leaf extract was observed and optical density was noted at 430 nm using UV-Visible Spectrophotometer. After stabilization of reaction, a pellet was prepared by taking 20 ml of silver nanoparticles solution in falcon tubes and centrifuged at 10000 rpm for 15 min. Pellet was thoroughly puri ed by washing repeatedly with distilled H 2 O and centrifuging for 10 min at 10000 rpm. Pellet was air dried for about 24 hrs. The yield for all the samples were calculated from the standard graph using the standard equation: y=0.33x+0.068, where y is the optical density and x is the concentration of silver nanoparticle.
Plackett Burman design: PBD was applied to detect the best three signi cant variables and thereby improve the sequential optimization of process variables (8).PBD is much used for screening experiments, nding out the much impacted factors and heavily confounded with 2 factor interactions. It is a two factorial design and can identify the critical parameters required for elevated synthesis of silver nanoparticles by screening "n" variables through n+1experiments. The experimental design screening of the variables is presented in Table 1.
PBD assumes that each parameter is independent and can be described by the rst-order model: where Y is the predicted target response (extraction yields of silver nanoparticles), β 0: the model intercept, β i : the regression coe cient, X i : an independent parameter.
All the variables selected for the study were denoted as numerical factors and investigated at two level intervals designated as -1 for low level and +1 for high level. A nine factor Placket Burman design was used to study the effect of variables on synthesis of silver nanoparticles. Variables chosen were pH X 1 (2, 10), temperature, X 2 (20°C, 100°C), Pressure X 3 (5psi,15psi), Time of reaction X 4 (2mins,90mins), , Microwave radiation X 5 (10secs,40sec), UV radiation X 6 (10mins, 40mins), concentration of plant extract, X 7 (0.1 g/ml, 0.4 g/ml), silver nitrate X 8 (0.001 M, 0.01 M) and sunlight X 9 (10mins, 90mins). X 10 and X 11 are considered as dummy variables. The values represented are lower (-) and upper limit (+) of each variable. The maximum and minimum limits for analysis were decided after performing a set of 14 preliminary experiments. Analysis was done using Statistica software which generated a 12 run table for 9 factors. All experiments were performed in triplicates to minimize experimental errors. Dummy variables (X 10 and X 11 ) are used to estimate experimental errors in data analysis. Triplicates of all experiments were done and mean of AgNps biosynthesized was taken as the response. Based on the biosynthesized AgNps, the factorial experiment was analyzed using regression analysis and ANOVA.
Box Behnken Design: The Box-Behnken design is a response surface methodology (RSM) design that requires only three levels (low, middle and high) to run an experiment. BBD was used to further optimize the synthesize conditions and was based on the parameters that were already screened as being signi cant factors by PBD. Plant extract (X 1 ), concentration of silver nitrate (X 2 ), and sunlight exposure (X 3 ) were selected for further optimization, with uctuation ranges of 0.1-0.4 g/mL, 0.1 -1 mM and 30-90 min respectively. Each experiment was performed in triplicate. Analysis of variance (ANOVA) was used to analyze the experimental data from 17 experiments and the results were tted by the response surface regression: where Y is the predicted response, XiXj represents coded independent variables, β 0 is the offset term, β i is the i th linear coe cient, β ii is the i th quadratic coe cient, and β ij is the ij th interaction coe cient. Statictica software was employed for experimental design as well as for regression and graphical analysis of the experimental data.

Characterization of Silver nanoparticle by SEM analysis:
After nding out the optimized parameters having major effect on the synthesis of nanoparticle, silver nanoparticles were synthesized at those optimized conditions. The size and shape of the nanoparticles synthesized were studied using SEM analysis. The characterization of the silver nanoparticles was carried out by Carl Zeiss ultra 55 SEM. Table 2: Plackett Burman design for evaluation of 9 independent variables (X 1 -X 9 ) for synthesis of silver nanoparticles from a) Jatropha curcas extract b)