^{1}

^{1}

^{*}

Water in diesel nano-emulsion (WiDNE) due to their nano size, kinetically stable gives its beneficial in commercial and environmental aspects. However, the capability of this fuel strongly depends on the method of preparation, stability and their physic-chemical properties. Central composite design (CCD) method was used to optimize variable interactions in order to obtain maximum stability. Methodology RSM method with six independent variables was selected in order to understand the impacts on droplet size. The response surface and 3D plots of the quadratic polynomial model were created for studying the combination effect on response. Dynamic light scattering DLS technique was used for measuring of droplet sizes. The analysis result by ANOVA was with 95% confidence displaying F value model was 52.82. The results displayed model was fulfilled with the assumptions of ANOVA. This study has relied on Design Expert software to locate the optimum droplet size situations. The measured diameter is 26 nm, with 0.0297 errors between actual conditions and measured value. The optimum blend properties of prepared WiDNE fuel were compared with conventional diesel. Improvements in physical properties were observed in presence of water in WiDNE.

Emulsions consist of two immiscible liquids which one as small droplets in the dispersed liquid, and can be classified as oil-in-water or water-in-oil relying on which phase is constituting the disperse phase [

Nano-emulsion is a class of high stability emulsions with extremely small droplets in the range of 20 - 200 nm with no apparent flocculation or coalescence [

The most common process in the preparation of nanoemulsions is high energy method [

In this study, WiDNE fuel preparation by using high shearing method. The emulsification process includes simulating droplet model for the optimum condition based on the Central Composite Design (CCD) [

The path for implementing experimental design was evaluated for two purposes. The first is to identify a subset of the original processes factors that have substantial main and interaction impacts on the final WiDNE fuel properties achieved. Second, factors have to be significant optimized statistically to determine the best WiDNE properties, according to the assumed Response Surface Method (RSM) model.

Conventional diesel fuel produced from local Daura Refinery was used as continuous phase of WiDNE fuel. The characterization of diesel fuel is shown in

No. | Properties | Value |
---|---|---|

1 | Density at 15%. (Kg/ liter) | 0.8932 |

2 | Total sulfur (%wt) | 2.88 |

3 | Total Nitrogen (ppm wt) | 320 |

4 | Acidity (mg KOH g) | 0.1 |

5 | Viscosity at 60%. (cSt)) | 25.15 |

6 | Pour point (˚C) | 6 |

7 | Cloud point (˚C) | 8 |

8 | Aniline point (˚C) | 71.8 |

The important step for forming stable emulsion was to select the suitable surfactant blend, Span80 and Tween80 that were used in the present work. The Specifications of the surfactants are shown in

WiDNE were generated by dispersing water content into diesel fuel according to

1) Surfactants Tween80 (HLB = 15) and Span80 (HLB = 4.3) were blended to get the desired Tween80/Span80 ratio (HLB = 8).

2) The prepared blend was added into the diesel fuel, and homogenized at 15,000 rpm for 10 min.

3) Distilled water was added gradually, into the mixture of diesel and surfactants, and mixed according to

Consequently CCD is a combination of both mathematical and statistical technique used to design experiments; evaluate the factors of process; obtaining the model; interaction between variables and find optimum condition to analyze the problem [

The sum of square sequential model was used to compare different models. It showed the statistical significance of adding new terms step by step in increasing order. It provided accounts of variation and associated P-values (Prob > F). The model was selected based on the highest order that was significant (P-value small) and not aliased, on lack of fit (P-value > 0.10) and reasonable agreement between adjusted R-squared and predicted R-squared (within 0.2 of each other). The summary table of the sequential model sum of square is shown on

Span 80 | Tween 80 | |
---|---|---|

symbol | S | T |

Appearance | Brown viscous | Amber sticky |

Mwt | 428.61 | 1310 |

Density (g/ml)@20˚C | 0.99 | 1.08 |

HLB | 4.3 | 15 |

NO. | W% | HLB | TS% | TIME(min) | rpm | pH | D (nm) |
---|---|---|---|---|---|---|---|

1 | 18 | 5.7 | 2.6 | 40 | 17,000 | 9.8 | 30 |

2 | 18 | 5.7 | 2.6 | 30 | 17,000 | 11.2 | 29.2 |

3 | 12 | 5.7 | 3.4 | 40 | 17,000 | 9.8 | 21 |

4 | 12 | 5.3 | 3.4 | 30 | 13,000 | 9.8 | 27.3 |

5 | 12 | 5.7 | 3.4 | 30 | 17,000 | 11.2 | 22 |

6 | 18 | 5.3 | 2.6 | 30 | 13,000 | 9.8 | 31.3 |

7 | 18 | 5.3 | 3.4 | 30 | 17,000 | 11.2 | 27 |

8 | 18 | 5.3 | 2.6 | 40 | 13,000 | 11.2 | 33 |

9 | 12 | 5.3 | 2.6 | 40 | 17,000 | 9.8 | 25 |

10 | 12 | 5.3 | 2.6 | 30 | 17,000 | 11.2 | 24.3 |

11 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 27 |

12 | 12 | 5.7 | 2.6 | 40 | 13,000 | 11.2 | 29 |

13 | 18 | 5.7 | 3.4 | 30 | 13,000 | 9.8 | 29 |

14 | 18 | 5.3 | 3.4 | 40 | 17,000 | 9.8 | 26.5 |

15 | 18 | 5.7 | 3.4 | 40 | 13,000 | 11.2 | 27 |

16 | 12 | 5.7 | 2.6 | 30 | 13,000 | 9.8 | 26.1 |

17 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 27.5 |

18 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 27 |

19 | 12 | 5.3 | 3.4 | 40 | 13,000 | 11.2 | 24.2 |

20 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 28 |

21 | 12 | 5.3 | 2.6 | 40 | 13,000 | 9.8 | 24.1 |

22 | 18 | 5.3 | 3.4 | 30 | 13,000 | 11.2 | 32 |

23 | 18 | 5.7 | 3.4 | 30 | 17,000 | 9.8 | 27.3 |

24 | 12 | 5.3 | 2.6 | 30 | 13,000 | 11.2 | 30 |

25 | 12 | 5.7 | 3.4 | 30 | 13,000 | 11.2 | 27 |

26 | 18 | 5.7 | 3.4 | 40 | 17,000 | 11.2 | 22.3 |

27 | 12 | 5.3 | 3.4 | 30 | 17,000 | 9.8 | 25 |

28 | 12 | 5.7 | 3.4 | 40 | 13,000 | 9.8 | 23.5 |

29 | 18 | 5.7 | 2.6 | 30 | 13,000 | 11.2 | 32.3 |

30 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 28 |

31 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 27.9 |

32 | 12 | 5.7 | 2.6 | 30 | 17,000 | 9.8 | 26 |

33 | 18 | 5.3 | 2.6 | 30 | 17,000 | 9.8 | 33 |

34 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 28 |

35 | 18 | 5.3 | 2.6 | 40 | 17,000 | 11.2 | 30.2 |

36 | 12 | 5.7 | 2.6 | 40 | 17,000 | 11.2 | 25.04 |

37 | 18 | 5.3 | 3.4 | 40 | 13,000 | 9.8 | 30 |
---|---|---|---|---|---|---|---|

38 | 18 | 5.7 | 2.6 | 40 | 13,000 | 9.8 | 29 |

39 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 27 |

40 | 12 | 5.3 | 3.4 | 40 | 17,000 | 11.2 | 21 |

41 | 15 | 6.5 | 3 | 35 | 15,000 | 10.5 | 25 |

42 | 15 | 5.5 | 3 | 35 | 25,000 | 10.5 | 21.7 |

43 | 15 | 5.5 | 3 | 35 | 15,000 | 7.5 | 30 |

44 | 15 | 5.5 | 3 | 35 | 15,000 | 13.5 | 28 |

45 | 15 | 5.5 | 5 | 35 | 15,000 | 10.5 | 29.5 |

46 | 0 | 5.5 | 3 | 35 | 15,000 | 10.5 | 2 |

47 | 15 | 4.5 | 3 | 35 | 15,000 | 10.5 | 30 |

48 | 15 | 5.5 | 3 | 35 | 5000 | 10.5 | 35 |

49 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 30.1 |

50 | 15 | 5.5 | 3 | 35 | 15,000 | 10.5 | 28 |

51 | 30 | 5.5 | 3 | 35 | 15,000 | 10.5 | 29 |

52 | 15 | 5.5 | 1.5 | 35 | 15,000 | 10.5 | 36.5 |

53 | 15 | 5.5 | 3 | 60 | 15,000 | 10.5 | 29.5 |

54 | 15 | 5.5 | 3 | 10 | 15,000 | 10.5 | 42 |

Source | Sum of Squares | df | Mean Square | F Value | p-value Prob > F | Model Significant |
---|---|---|---|---|---|---|

Block | 12.88 | 2 | 6.44 | |||

Model | 1401.79 | 27 | 51.92 | 52.82 | <0.0001 | Significant |

A-w% | 505.33 | 1 | 505.33 | 514.07 | <0.0001 | |

B-HLB | 22.72 | 1 | 22.72 | 23.11 | <0.0001 | |

C-TS% | 99.21 | 1 | 99.21 | 100.93 | <0.0001 | |

D-TIME | 99.79 | 1 | 99.79 | 101.52 | <0.0001 | |

E-RPM | 138.22 | 1 | 138.22 | 140.61 | <0.0001 | |

F-pH | 0.74 | 1 | 0.74 | 0.75 | 0.3942 | |

AB | 7.64 | 1 | 7.64 | 7.78 | 0.0102 | |

AC | 2.18 | 1 | 2.18 | 2.22 | 0.1491 | |

AD | 0.097 | 1 | 0.097 | 0.098 | 0.7564 | |

AE | 0.44 | 1 | 0.44 | 0.45 | 0.509 | |

AF | 1.82 | 1 | 1.82 | 1.86 | 0.1858 | |

BC | 2.9 | 1 | 2.9 | 2.95 | 0.0985 | |

BD | 0.46 | 1 | 0.46 | 0.47 | 0.5001 |

BE | 8.00E−04 | 1 | 8.00E−04 | 8.14E−04 | 0.9775 | |
---|---|---|---|---|---|---|

BF | 0.19 | 1 | 0.19 | 0.19 | 0.6674 | |

CD | 6.34 | 1 | 6.34 | 6.45 | 0.018 | |

CE | 7.84 | 1 | 7.84 | 7.98 | 0.0094 | |

CF | 7.64 | 1 | 7.64 | 7.78 | 0.0102 | |

DE | 0.19 | 1 | 0.19 | 0.19 | 0.6674 | |

DF | 0.46 | 1 | 0.46 | 0.47 | 0.5001 | |

EF | 22.71 | 1 | 22.71 | 23.11 | <0.0001 | |

A2 | 212.62 | 1 | 212.62 | 216.3 | <0.0001 | |

B2 | 2.98 | 1 | 2.98 | 3.03 | 0.0947 | |

C2 | 27.26 | 1 | 27.26 | 27.73 | <0.0001 | |

D2 | 50.61 | 1 | 50.61 | 51.48 | <0.0001 | |

E2 | 0.66 | 1 | 0.66 | 0.67 | 0.4195 | |

F2 | 0.017 | 1 | 0.017 | 0.017 | 0.8962 | |

Residual | 23.59 | 24 | 0.98 | |||

Lack of Fit | 19.99 | 17 | 1.18 | 2.29 | 0.1348 | Not Significant |

Pure Error | 3.6 | 7 | 0.51 | |||

Total Corrected for the means | 1438.26 | 53 |

The lack of fit tests was included because extra design points beyond what was needed for the model were involved to provide estimation of pure error. Since it was not desirable, so a small F value and probability greater than 0.1 were desired.

Data obtained in

The appropriate of the models relied on the value of R^{2} coefficient, the model recording R^{2} = 98.34%. Furthermore, a suitable agreement with the adjusted coefficient of determination was discovered.

In this study, the Adj-R^{2} value of 96.48% and Pred. R^{2} 84.40% were obtained. Both values of Adj-R^{2} and R^{2} very adjacent to 1.0 were high and supported a high closing between the predicted and the observed values.

The Pred-R^{2} of 84.40% was in rational agreement with the Adj. R^{2} of 96.48%; the difference was less than 0.2. Value of lake of fit was 0.1348 > 0.05 (not significant) showing that the hypothesis of significance of fitting was rejected for RSM method. The model with p-value of 0.0001 means that there was only 0.01% chance that the F-value of the model. Response surface model equation for regression was in actual units to optimum diameter size by using Design Expert software as illustrated in the Equation (1).

Statistical analysis | Values |
---|---|

R-Squared | 0.9834 |

Adj. R-Squared | 0.9648 |

Pred. R-Squared | 0.8440 |

Adeq. Precision | 51.660 |

D ( nm ) = 27 . 93 + 2 . 48 × A − 0. 53 × B − 1 . 25 × C − 1 . 1 0 × D − 1 . 29 × E − 0. 1 0 × F − 0. 48 × AB − 0. 2 × AC + 0.0 5 × AD + 0. 11 × AE − 0. 23 × AF − 0. 3 0 × BC + 0. 12 × BD − 0.00 5 × BE + 0.0 7 × BF − 0. 44 × CD − 0. 49 × CE − 0. 48 × CF + 0.0 7 × DE + 0. 12 × DF − 0. 84 × EF − 0. 54 × A 2 − 0.0 6 × B 2 + 0. 24 × C 2 + 0. 26 × D 2 − 0.0 3 × E 2 − 0.00 6 × F 2 (1)

Equation (1) is in terms of coded variables that can be used for making expectations toward the responses to each factor at given levels. Negative value for individual factor in equation means in reducing droplets size. By default, the low levels of the factors were coded as −1 and the high levels of the factors interaction were coded as +1. The coded equation was useful for evaluating the relative effluent of the variables by comparing to the coefficients factor, where water W% (A), HLB (B), TS% (C), mixing time (min) (D), mixing speed rpm (E), and pH (F).

It was observed from

The results displayed fulfilled with the model assumptions of ANOVA. It indicates no sign presented that the error terms were correlated with each other.

The most significant step in the preparation of WiDNE is selection of an appropriate variables and the effect of each one on the response. In this study, RSM method with six variables was selected in order to understand the impacts on droplet size. Each of these variables had a different effluent and limited impact except W%, all other factors were associated to reduce droplet size.

Therefore, the 3D plots are the best method for illustrating the interaction between variables and effect of each others as can be shown in

In order to clarify the impact of the independent variables according to the results, response surface and 3D plots of the quadratic polynomial model were

created by varying two of independent variables (within the experimental range) for studying the combination effect on response.

Thus the Figures were generated by varying two variables in x and y axis while holding the other variables at optimum values. In this research, WiDNE fuels were prepared by two steps, the first step was prepared of a WiDNE at low rotating speed. Then, high energy homogenizer process was used to additionally decrease in the droplet size. The resistance of droplets to deformation was determined by the surface tension [

The coalescence rate of droplets is specified by the ability of emulsifiers that are adsorbed on the surface of droplets, which is ruled by the concentration of surfactants and surface activity

According to Equation (1), the coefficient value of HLB was −0.53 for WiDNE preparation. The higher value means this variable is significant. The negative value means that HLB leads to decrease the droplet size but in less effect than positive.

The surfactants role is very central in WiDNE system. It implements two functions. First, it will increase the interaction between the two immiscible systems; diesel fuel and water by decreasing interfacial tension. Second, it can be improved to stabilize the WiDNE system. Stability behavior of emulsions is highly reliant on concentration of surfactant and its nature.

Surfactant molecules organize themselves near interfacial film between water dispersed phase and diesel fuel a continuous phase, in order to stabilize the water dispersed droplets into the diesel phase. High concentration of surfactants avoids the integration of water droplets [

Obtaining the best conditions for the WiDNE fuel preparation process is one of the advantages of using this study. As shown in

Optimum blend of prepared WiDNE fuel are compared with conventional diesel, as shown in

The optimum water ratio was 12%, based on weight percent. The proportion of water inside the emulsified fuel must be kept within the specified limits, because the increase in the proportion of water quantity leads to a decrease in heat of combustion and affect the properties of fuel ignition.

Surfactant plays an important role in the stability of the WiDNE. This role lies in the ability of the surfaces to reduce the interfacial tension between the two liquids and facilitate the process of emulsification and protect droplets from adhesion, aggregations and thus protection from separation.

The exposure time of water droplet within the WiDNE fuel has a direct relationship to the conditions and the rotation speed. Increasing the time led drops to homogeneity and uniform distribution within the emulsion, but this will be limited if exceeded the optimum time, where considered consumed in energy and time. The optimum time was chosen to be 30 min. The rpm leads to the rupture of the droplets to smallest size. Therefore, it is the main factor affecting the size of droplets; the best speed is chosen to be 15,000 rpm. Alkalinity pH does not have a strong effect on droplets size but has interaction with other factors; it is optimized to be 10.5.

Properties | Diesel fuel | WiDNE fuel |
---|---|---|

Calorific value (kJ/kg) | 44,800 | 38,850 |

Flash point (˚C) | 54 | 61 |

Viscosity at 40˚C (cSt) | 3.268 | 4.56 |

Pour point (˚C) | 9 | 7 |

Density g/cm^{3} | 0.87 | 0.882 |

The variation in the calorific value of WiDNE fuel is shown in

Flash point is an indication of the fuel’s volatility. For safe handling of liquid fuel, it should always have high value. The influence of water content on flash point of WiDNE fuel is shown in

The effect of the presence of water on the viscosity of WiDNE fuel is shown in

The impact of water on density of WiDNE fuel is shown in ^{3} to 0.882 g/cm^{3} for WiDNE. It is known that the water is heavier than diesel fuel. The impact of water is obvious in increases the density when blending with diesel. However, more fuel density means less volume will be taken for storing purpose.

Pour point of WiDNE fuel +7˚C were less compared with net diesel +9˚C. Increasing water content on WiDNE fuel caused reduction in pour point as shown in

1) Response surface methodology (RSM) was used to enhance process variables for preparation of WiDNE fuel.

2) Surfactant concentration has very positive effect on emulsion stability.

3) Increasing in water content decreased the emulsion stability.

4) Mixing and time have enhanced the stability significantly up to a certain limit beyond which it remains the same.

Khidhir, A.G. and Hamadi, A.S. (2018) Central Composite Design Method for the Preparation, Stability and Properties of Water-in-Diesel Nano Emulsions. Advances in Chemical Engi- neering and Science, 8, 176-189. https://doi.org/10.4236/aces.2018.83012