a. Catalyst characterization
Natural CaCO3 has microcrystalline forms such as calcite, aragonite, dolomite or vaterite .Among them, aragonite structure has biocompatibility properties. This microcrystalline form can be found naturally in the crust of all clams and often in bivalves (pH=10.32). Compared to natural CaCO3 particles in the microcrystalline form of calcite (pH=9.91), seashells are layered. Layers are composed of two parts, an inorganic section inclusive CaCO3 and an organic section containing chitin, protein or polysaccharides cahins which make them a natural nano-sized biocomposite. This biocomposite can show its unique catalytic activity for organic transformations when its highly ordered and compact layers are crushed by milling to afford a material demonstrating more surface area. In this study, seashells were collected from the southern coast of Caspian sea, Babolsar, Iran. The shells were thoroughly washed with distilled water, refluxed in EtOH for 30 min and then oven-dried at 50 °C for 1 h. Obtained dried seashells were ball-milled in a stainless steel ball mill vessels at 25 Hz and ambient temperature.
The FTIR spectra of the ball-milled seashells (3), commercial CaCO3 and chitin are shown in (Fig. 2). The absorption bands observed at 1452, 1081, 840 and 710 cm−1 are related to the aragonite CaCO3 nanocrystals43-45. By comparing of the FT-IR spectra of the catalyst and commercial CaCO3, it can be found that the biocomposite catalyst 3 is mainly comprised of CaCO3. Furthermore, results of energy-dispersive X-ray (EDX) analysis showed that the nano-biocomposite seashells (3) includes elements such as calcium, carbon and oxygen (Fig. 3). Moreover, the scanning electron microscopy (SEM) image of the catalyst 3 showed an almost uniform spherical particles distribution by the average of particles size about 28-43 nm (Fig. 4).
Also, the atomic force microscopy analysis of seashell nano-biocomposite shows the topology of the catalyst 3 surface. AFM images confirms that the catalyst particles are at the nanoscale (Fig. 5). On the other hand, the X-ray diffraction (XRD) pattern of the ball-milled seashells nano-biocomposite (3) demonstrated that the crystalline quality of aragonite CaCO3 in the obtained powder is maintained throughout the process (Fig. 6)43-45.
In addition, the thermogravimetric analysis (TGA) of the catalyst 3 demonstrated that significant reducing of the weight seashells was not observed until around 600 °C. This indicates that seashells have high thermal stability more than than commercial CaCO3 (Fig. 7) in the calcite microcrystalline from.
b. Optimization of the esterification of isoamyl alcohol with acetic acid
Our aim in this research was the improvement of the RSM model for finding the best effective relationship between four variables including reaction time, reaction temperature, molar ratio of acid: alcohol, and catalyst loading. Therefore, analysis of variance (ANOVA) was assessed with regard to significance of the effect of operational parameters and their interactions on the yield of esterification of isoamyl alcohol with acetic acid. The raw isoamyl acetate (4) yield data were not well fit to various models (linear, quadratic, and cubic). Accordingly, a convenient data transformation was required40,46,47With this transformation, the experimental data well fit to the quadratic model. The causing model equation, in terms of the variables, is shown in Eq. 1. Generally, it is favourable to fit the lowest order polynomials that sufficiently describe the system. Therefore, a quadratic polynomial model was fitted to obtain the yield of isoamyl acetate (4). The quadratic model has been chosen based on stepwise procedure, and the model terms have been selected according to their p values (> 0.05). The obtained results of ANOVA analysis were shown in Table 1.
The ANOVA analysis, as shown in Table 1, confirmed the adequacy of quadratic model to demonstrate the actual relationship between the response and the significant variables, since the probability value was lower than 0.0001. Moreover, it can be seen from Table 1, the F value of 30.66 indicates that the above mentioned model is significant. On the other hand, the calculated F value of lack of fit of 1.56 indicated that the lack of fit was not significant relative to the pure experimental error and confirms the reliability of the model. The p values from Table 1 were used to check the significance of each of the factors and interaction between the factors. The value of prob F less then 0.05 for a variable implies that its effect is significant at the 95% of confidence interval11,48,49. Values greater than 0.1 display that the variable is not significant. Hence, the effects of terms A, B, C, D, AB, AC, AD, A2, B2 and D2 are significant to explain the model
Table 1
ANOVA table of the quadratic response surface model.
Source
|
SSa
|
Dfb
|
MSC
|
F-Value
|
Prob>F
|
Model
|
9258.76
|
10
|
925.88
|
30.66
|
<0.0001 Significant
|
A-Temp
|
5342.73
|
1
|
5342.73
|
176.91
|
<0.0001
|
B-Time
|
1993.69
|
1
|
1993.69
|
66.02
|
<0.0001
|
C-Cat
|
99.53
|
1
|
99.53
|
3.30
|
0.0853
|
D-Molar ratio of acid:alcohol
|
341.95
|
1
|
341.95
|
11.32
|
0.0033
|
AB
|
191.20
|
1
|
191.20
|
6.33
|
0.0210
|
AC
|
264.88
|
1
|
264.88
|
8.17
|
0.0100
|
AD
|
278.31
|
1
|
278.31
|
9.22
|
0.0068
|
A2
|
308.62
|
1
|
308.62
|
10.22
|
0.0047
|
B2
|
119.34
|
1
|
119.34
|
3.95
|
0.0614
|
D2
|
575.79
|
1
|
575.79
|
19.07
|
0.0003
|
Residual
|
673.34
|
19
|
30.20
|
|
|
Lack of fit
|
566.39
|
14
|
33.35
|
1.56
|
0.3278
not significant
|
Pure Error
|
106.95
|
5
|
21.39
|
|
|
Cor Total
|
983257
|
29
|
|
|
|
a) Stand for sum of square
b) Degree of freedom.
C) Mean of source
|
Table 2
Other statistical parameters of this model.
Std. Dev
|
5.50
|
R – Squared
|
0.9416
|
Mean
|
54.54
|
Adj R- Squared
|
0.9109
|
C.V
|
%10.08
|
Pred R-Squared
|
0.7788
|
PRESS
|
2174.48
|
Adeq Precision
|
20.572
|
The model created was in the coded format and is shown in Eq. 1. Thus, the final Second-order polynomial equation is:
Yield = +48.24+13.69*A + 8.36*B + 1.87*C – 3.46*D +3.46*A*B – 3.93*A*C -4.17*A*D + 2.22*A2 +1.38*B2 +3.03*D2
(Eq. 1)
Other statistical parameters of the model, a coefficient of determination adjusted for the number of parameters in the model relative to the number of points in the design was 0.9416. This value showed that the model was trustworthy in the predicting the response and, at least 94.16 % of the variability in the data, could be defined by the second-order polynomial model. The R2pred and R2adj values were 0.7788 and 0.9109, respectively, that show good fitness of the model because these values are in reasonable agreement with together, as the values should differ by no more than 0.250,51. The adequate precision is a measure of the signal to noise ratio and a quantity greater than 4 is desirable52. At this investigation, adequate precision of 20.572 is a good signal to noise ratio and proves the ability of model to navigate the design space (Table 2)53. These statistical values along with lack of fit tests display that the quadratic model is adequate to predict the response (the yield of isoamyl acetate (4). The relationship between the experimental values and predicted values are denoted in (Fig. 8). It can be seen from this Fig. that the experimental values are very close to predicted using the model equation, showing linear distribution with R2 = 0.9416. This value displayed that this experimental model is acceptable and reproducible.
The best approach to predict the relationships between responses, variables, and interactions is the contour and three-dimensional plots. As shown in (Fig. 9 and 10), three-dimensional (3D) response surfaces and two-dimensional (2D) contour plots illustrate the effects of different variables on the response. In these plots, two variables varied, whereas the other variables are kept constant and they describe the type of interactions between two tested variables and the correlations between response and variables levels. 3D response surface and contour plot in (Fig. 9) illustrate the influence of interaction between the reaction temperature (A) and molar ratio of acid: alcohol (D) on values of time (B) and catalyst loading (C) were set to 165 min and 25.5 mg, respectively. These results clearly show that at temperature of 56.5 °C, enhancing the molar ratio of reactants had no notable effect on the yield of isoamyl acetate (4) whereas at 98 °C lower yields were obtained when high molar ratio of reactants are used11,54. Moreover, (Fig. 10) display relationship between reaction temperature (A) and catalyst 3 loading (C). It can also be seen that by increasing reaction temperature to 98 °C, considerable yields were achieved. Instead, higher loadings of the catalyst had no significant effect on the yield of reaction.
The main purpose of the response surface methodology (RSM) is detecting the optimal conditions to maximize the percentage of the yield of the favorite product (response). For this purpose, four factors were measured in the range of (±1) to optimize the process while the response (the yield of isoamyl acetate (4) was fixed to maximum value55. According to this method, the optimum values of the factors under solvent-free conditions were 15.7 mg for the catalyst 3 loading, 1:3.7 molar ratio of isoamyl alcohol: acetic acid, 98 °C for the reaction temperature, and 219 min for the reaction time. The yield of desired product under optimal experimental conditions (91 %) is in excellent agreement with the predicted value (89 %). Good settings between the experimental and predicted values show the credibility and adequacy of the model to predict the yield of isoamyl acetate (4) by esterification of isoamyl alcohol with acetic acid under solvent-free conditions. Indeed, the crude reaction mixtures were analyzed by gas chromatography (GC). Furthermore, gas chromatography-mass spectroscopy (GC–MS) analysis of the crude reaction mixture was also performed. The results have been shown in the Supplementary Material.
The unique catalytic properties of ball-milled seashells, as a biodegradable, eco-friendly, and recyclable nano-biocomposite, can be related to synergy of aragonite microcrystalline from rather than calcite one in the commercial CaCO3 samples, porosity and hygroscopic properties of the material.
To demonstrate the efficiency of this methodology, Table 3 compares the ability of various heterogeneous catalysts in the esterification reaction of isoamyl alcohol with acetic acid which represents significant excellence of the ball-milled seashells more than others.
Table 3
Comparison of the performance of various catalysts in the esterification of isoamyl alcohol (1) with acetic acid (2).
Entry
|
Catalyst
|
Temp (°C)
|
2: 1 mole ratioa
|
Loading (mg)
|
Time (min)
|
Yield %
|
1
|
NaX
|
120
|
1:1
|
2000
|
240
|
9256
8856
9233
9411
7515
6418
8057
91e (this work)
|
2
|
NaY
|
120
|
1:1
|
2000
|
240
|
3
|
β– MnO2
|
124
|
1:1.8
|
144
|
210
|
4
|
SO4-2/ TiO2
|
130
|
7:1
|
1.9b
|
300
|
5
|
Lipase
|
30
|
10:8
|
12c
|
480
|
6
|
Lipase
|
37
|
2:1
|
60d
|
480
|
7
|
Cs(K)XH3 –X PW12O40TPA
|
180
|
1;1
|
500
|
|
8
|
Ball-milled seashells
|
98
|
3.7:1
|
15.7
|
219
|
a) Molar ratio of acid:alcohol.
b) 3.2 wt % with respect to acetic acid.
C) 12% (w/w) enzyme
d) 60 IU of immobilized lipase
e) Average of two runs at optimal conditions
|