JFUE 8688 No.

of Pages 7, Model 5G
26 November 2014

Fuel xxx (2014) xxx–xxx

1

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

4
5

3 Measurement, correlation and prediction of biodiesel blends viscosity

6 Q1 Sibel Geacai a, Olga Iulian b, Irina Nita a,⇑
7 a
Department of Chemistry and Chemical Engineering, ‘‘Ovidius’’ University of Constanta, 124, Bd. Mamaia, 900521 Constanta, Romania
8 b
Department of Applied Physical Chemistry and Electrochemistry, Faculty of Applied Chemistry and Materials Science, ‘‘Politehnica’’ University of Bucharest, 132, Calea
9 Q2 Grivitei, 010737 Bucharest, Romania

10
11

1 3
h i g h l i g h t s
14
15  Viscosity and refractive index experimental data for systems with biodiesel for temperature range 293.15–323.15 K.
16  The accuracy of different models for viscosity calculations is presented.
17  Correlative–predictive equations give better results than the ones predictive.
18  Calculation of viscosity of biodiesel blends from the refractive index of the blends.

19
a r t i c l e i n f o a b s t r a c t
2
3 1
5
22 Article history: The viscosity is one of the most important transport properties of a fuel, influencing especially the injec- 36
23 Received 17 May 2013 tion system, mainly at low temperatures when viscosity increases. Biodiesel is considered an alternative 37
24 Received in revised form 7 November 2014 fuel for diesel engines, but some of its properties such as density and viscosity, have higher values than 38
25 Accepted 13 November 2014
those of diesel fuel. In this paper, the viscosity of binary blends biodiesel + diesel fuel, biodiesel + benzene 39
26 Available online xxxx
and biodiesel + toluene in the temperature range 293.15–323.15 K is presented. Experimental data were 40
used to evaluate the accuracy of viscosity calculations with different models, both for blends of biodiesel 41
27 Keywords:
with pure components (benzene or toluene) and for biodiesel with diesel fuel. The biodiesel + benzene and 42
28 Biodiesel
29 Diesel fuel
biodiesel + toluene blends were used as base of comparison, to better understand the behavior of biodie- 43
30 Blend sel + diesel fuel blends. The Grunberg–Nissan, Wilke and McAllister equations were used to predict the 44
31 Viscosity blends viscosity from the viscosities of pure components and to correlate the experimental data. Also, 45
32 Refractive index the viscosity of the biodiesel blends was calculated from the refractive index of the blends or by the means 46
33 Q4 Prediction models of viscosity–temperature and viscosity–temperature–composition empirical correlative equations. 47
34
Ó 2014 Published by Elsevier Ltd. 48
49

50
51
52 1. Introduction was obtained biodiesel with better qualities than those of diesel 66
fuel, like biodegradability, significant reduction of CO, SOx, 67
53 High consumption of conventional fuels and environmental unburned hydrocarbons, volatile organic compounds and particu- 68
54 concerns had created urgent needs for alternative fuels. Transport late matter emissions. 69
55 sector uses a significant amount of petroleum-derived fuels. Even More than 95% of biodiesel is made today at industrial level 70
56 though new oil reserves were discovered in the recent years, from edible oils [4]. Six raw materials for biodiesel production 71
57 non-renewable fuels could not meet the world energy demand in dominate the entire world: rapeseed, sunflower, palm, cottonseed, 72
58 the long term. Fuels made from petroleum can be replaced with soybean and peanut oils [5–8]. Biodiesel properties differ from one 73
59 fuels from renewable sources (e.g. vegetable oils) because they sample to another, due to esters varying composition determined 74
60 have equivalent features with fossil fuels [1–3]. by the type of feedstock. Biodiesel is a mixture of esters, while die- 75
61 Vegetable oils have a high viscosity and their use in diesel sel fuel is a mixture of hydrocarbons. 76
62 engines causes different problems, such as low atomization, Biodiesel is an alternative fuel for diesel engines, but some of its 77
63 incomplete combustion, injector clogging and low oxidation stabil- properties such as density and viscosity, have higher values than 78
64 ity due to polyunsaturated content [1]. To minimize or to avoid those of diesel fuel. Viscosity is one of the most important transport 79
65 these problems, the vegetable oil were transesterified and thus properties of a fuel, influencing especially the injection system, 80
mainly at low temperatures when viscosity increases. Biodiesel 81

Q3 ⇑ Corresponding author. Tel./fax: +40 241 606 434. and its blends with diesel fuel show similar viscosity–temperature 82

E-mail address: inita@univ-ovidius.ro (I. Nita). dependence [2,9]. Some methods for predicting biodiesel and 83

http://dx.doi.org/10.1016/j.fuel.2014.11.041
0016-2361/Ó 2014 Published by Elsevier Ltd.

Please cite this article in press as: Geacai S et al. Measurement, correlation and prediction of biodiesel blends viscosity. Fuel (2014), http://dx.doi.org/
10.1016/j.fuel.2014.11.041
 JFUE 8688 No. of Pages 7, Model 5G
26 November 2014

2 S. Geacai et al. / Fuel xxx (2014) xxx–xxx

84 diesel fuel + biodiesel blends viscosity are presented in literature into a separating funnel and kept for about 12 h to reach the 147
85 [2,10–16]. Blends of diesel fuel with 5–10% biodiesel are currently two-phase separation, the crude biodiesel and glycerin. Biodiesel 148
86 commercialized as fuel for diesel engines. obtained after glycerol removal was washed with distilled water 149
87 To be used in transport area, biodiesel and its blends with diesel up to 50 °C. Biodiesel was washed with water to remove traces 150
88 fuel must fulfill standard requirements. The ASTM D975 (USA) and of unreacted methanol and catalyst. Biodiesel layer was separated 151
89 EN 590 (Europe) standards are used to evaluate the quality of die- from water. This process was repeated until the pH of washing 152
90 sel fuels and also are used in the case of diesel fuel + biodiesel water was the same as distilled water. At the end of the process, 153
91 blends with up to 5% biodiesel. For blends with 6–20% biodiesel biodiesel was dried on calcium sulfate. 154
92 content, the ASTM D7467 is used and for pure biodiesel, the ASTM The methyl esters composition of the biodiesel samples was 155
93 D6751 (USA) and the EN 14214 (Europe) requirements must be determined by gas chromatography using a Clarus 500 GC chro- 156
94 fulfilled [17]. matograph equipped with a FID detector and the separation of 157
95 As a continuation of our studies on properties of biodiesel esters was realized on a capillary column of high polarity, with 158
96 blends [18,19], the objectives of this work are to present new polysiloxane as stationary phase. Isothermal separation was made 159
97 experimental data and to evaluate the accuracy of some equations at 210 °C using hydrogen as carrier gas flow of 20 ml/min. Fatty 160
98 for viscosity data correlation or prediction from pure components acid composition of the biodiesel is presented in Table 1. 161
99 properties. For this purpose already used equations for biodiesel Based on their chemical composition, the average molar mass of 162
100 blends were tested (Grunberg–Nissan eqs.), also Wilke equation biodiesel samples was calculated using Eq. (1) [23]: 163
101 from petroleum domain and McAllister equation from classical X 164
MBio ¼ xi  M i ð1Þ 166
102 thermodynamics of molecular mixtures, equations that have not
103 yet been used for biodiesel + diesel fuel blends. Grunberg–Nissan where MBio represents the mean molar mass of biodiesel, Mi and xi, 167
104 (one parameter) and McAllister equations are semiempirical or the molar mass and molar fraction of i-th fatty acid methyl ester, 168
105 theoretical-based models with one or two adjustable parameters, respectively. 169
106 more difficult to use for complex blends of biodiesel + diesel fuel. The mean molar mass of diesel fuel was determined using the 170
107 Because the experimental determination of refractive index is sim- cryoscopic method with benzene as solvent. The determinations 171
108 pler than of viscosity, the viscosity evaluation from refractive index were realized on an ice bath. The accuracy of the thermometer 172
109 data can be useful in many practical cases. An equation from was ±0.3 °C. Three measurements of the melting point of diesel 173
110 petroleum domain was tested to correlate viscosity and refractive fuel + benzene solutions were made. The mean value was taking 174
111 index for biodiesel blends. into account for diesel fuel molar mass calculation. 175
112 Given the fact that biodiesel + benzene and biodiesel + toluene The properties of biodiesel and diesel fuel are presented in 176
113 blends are not recommended as fuels for transport due to the tox- Table 2. 177
114 icity of aromatic compounds, their data were used for comparison
115 purposes to better understand the quality of viscosity data correla- 2.2. Kinematic viscosity and refractive index measurement 178
116 tion of biodiesel + diesel blends. The biodiesel + benzene and bio-
117 diesel + toluene blends are simpler regarding their composition Blends of biodiesel + diesel fuel (Bio + D), biodiesel + benzene 179
118 and components structure, compared to biodiesel + diesel fuel (Bio + B) and biodiesel + toluene (Bio + T) were prepared by volume 180
119 blends; they have no polar molecule and are miscible throughout fractions (v1 – volume fraction of biodiesel) of 0.05, 0.10, 0.15, 0.20, 181
120 the composition with biodiesel. 0.25, 0.30, 0.35, 0.40, 0.45, 0.50, 0.55, 0.60, 0.65, 0.70, 0.75, 0.80, 182
121 After our knowledge, properties of biodiesel + diesel fuel blends 0.85, 0.90 and 0.95. All components are completely miscible. The 183
122 of different degrees of blending at different temperatures using bio- experimental uncertainty in volume fractions was estimated to 184
123 diesel obtained from rapeseed oil and other types of vegetable oils be less than ±0.002. 185
124 have been investigated. There are a lot of studies referring to biodie- The kinematic viscosity (g) was measured using an Anton Paar 186
125 sel + diesel fuel blends [18–28], but not all specifies the biodiesel SVM 3000 viscometer calibrated with double distilled water and 187
126 type, its chemical composition and average molecular weight, fact pure solvents. This viscometer is equipped with a Peltier thermo- 188
127 which has been achieved in this study; the biodiesel + benzene stat whose accuracy is ±0.02 °C. The kinematic viscosity was mea- 189
128 and biodiesel + toluene blends have been less studied [29]. sured in the temperature range 293.15–323.15 K at 5 K intervals 190
with an accuracy of ±0.35%. The kinematic viscosity data reported 191
129 2. Material and methods here are means of triplicate determinations. 192
The refractive index was measured for temperature ranging 193
130 2.1. Materials from 298.15 K to 323.15 K. A refractometer Abbé, Atago 3T type 194
coupled to a thermostated bath was used to measure the refractive 195
131 Biodiesel used in this study was synthesized by transesterifica- index. The reproducibility of refractive index data was of 104 and 196
132 tion of commercially available rapeseed oil. Methanol with a purity the thermostated bath had an accuracy of ±0.05 °C. 197
133 of 99.5%, and potassium hydroxide 99% purity as catalyst, both of
134 Czech origin (Lachi-ner), were used in the transesterification 3. Results and discussions 198
135 process. Binary blends with biodiesel were prepared using 99.7%
136 purity benzene from Merck, 99.7% purity toluene from Merck 3.1. Experimental data 199
137 and diesel fuel from a local company.
138 The vegetable oil (rapeseed oil) was converted into fatty acid The experimental kinematic viscosity data (g) for biodiesel bin- 200
139 methyl esters (biodiesel) by transesterification reaction. Molar ary blends with diesel fuel, benzene and toluene, respectively, for 201
140 ratio methanol/oil was 6:1 and the catalyst (KOH) was 1.5% by
141 weight based on oil. The catalyst was dissolved in alcohol and Table 1
Fatty acid compositions of biodiesel.
142 the mixture was added over the oil sample. Transesterification
143 reaction progresses at temperatures of 60 °C at atmospheric pres- Fatty acid mass percent
144 sure, under continuous mixing for 2 h. The reaction was performed 14:0 16:0 18:0 18:1 18:2 20:1
145 in a reaction vessel (500 ml) equipped with condenser, magnetic
Biodiesel 0.21 25.89 3.11 59.80 10.60 0.39
146 stirrer and thermometer. The mixture obtained was introduced

Please cite this article in press as: Geacai S et al. Measurement, correlation and prediction of biodiesel blends viscosity. Fuel (2014), http://dx.doi.org/
10.1016/j.fuel.2014.11.041
 JFUE 8688 No. of Pages 7, Model 5G
26 November 2014

S. Geacai et al. / Fuel xxx (2014) xxx–xxx 3

Table 2
Properties of biodiesel and diesel fuel.

Property Units Method Biodiesel Diesel fuel

3
Density at 20 °C g/cm ASTM D4052 0.8840 0.8414
Kinematic viscosity at 40 °C mm2/s ASTM D445 4.6715 3.5439
Water content ppm SR EN ISO 12937/2003 258.36 130
Sulfur content mg/kg ASTM D5453 0.00 0.14
Flash point °C ASTM D93 142 60
Methylic esters of fatty acids % (m/m) SR EN 14103:2003 96.50 3.0
Molar mass g/mol 288.62 213.89

202 twenty-one different compositions between 293.15 K and literature: 3.97–4.24 mm2/s at 313.15 K [16]. Data from Table 3 213
203 323.15 K, are presented in Table 3. The refractive index experimen- shows the viscosity of biodiesel + diesel fuel blends with no more 214
204 tal values for biodiesel + benzene blends are presented in Table 4, than 60 vol% biodiesel (4.055 mm2/s) is in accordance with standard 215
205 for the biodiesel + diesel fuel and biodiesel + toluene, the data demand for diesel fuel (ASTM D975 recommended values: 216
206 being previously presented [19]. 1.9–4.10 mm2/s at 40 °C). 217
207 As seen in Table 3, the kinematic viscosities of biodiesel vary in The kinematic viscosity of all studied blends of biodiesel with 218
208 the range of 7.381–3.633 mm2/s and are higher than those of diesel fuel, benzene or toluene increases with the increase of 219
209 diesel fuels, 5.331–2.792 mm2/s, in the temperature range biodiesel content. The binary blends demonstrate temperature- 220
210 293.15–323.15 K. The obtained kinematic viscosity data for dependent behavior; their kinematic viscosities decrease nonlin- 221
211 biodiesel correspond with the recommended EN 14214 values: early with temperature. The refractive index of biodiesel + benzene 222
212 3.5–5.0 mm2/s at 313.15 K temperature and with Alptekin data in blends decreases with biodiesel content increasing and with 223

Table 3
Kinematic viscosity of biodiesel blends with diesel fuel, benzene or toluene at different temperatures.

System T (K) v1a g (mm2/s) v1 g (mm2/s) v1 g (mm2/s) v1 g (mm2/s) v1 g (mm2/s) v1 g (mm2/s) v1 g (mm2/s)
Biodiesel + diesel fuel 293.15 0.00 5.331 0.15 5.412 0.30 5.560 0.45 5.826 0.60 6.234 0.75 6.761 0.90 7.381
0.05 5.359 0.20 5.438 0.35 5.616 0.50 5.971 0.65 6.400 0.80 6.956 0.95 7.619
0.10 5.386 0.25 5.499 0.40 5.703 0.55 6.098 0.70 6.576 0.85 7.168 1.00 7.831
298.15 0.00 4.831 0.15 4.968 0.30 5.133 0.45 5.398 0.60 5.729 0.75 6.114 0.90 6.579
0.05 4.878 0.20 5.012 0.35 5.215 0.50 5.488 0.65 5.848 0.80 6.264 0.95 6.751
0.10 4.925 0.25 5.072 0.40 5.298 0.55 5.608 0.70 5.955 0.85 6.429 1.00 6.927
303.15 0.00 4.187 0.15 4.318 0.30 4.522 0.45 4.762 0.60 5.025 0.75 5.370 0.90 5.768
0.05 4.220 0.20 4.383 0.35 4.603 0.50 4.841 0.65 5.162 0.80 5.498 0.95 5.890
0.10 4.254 0.25 4.447 0.40 4.684 0.55 4.930 0.70 5.262 0.85 5.633 1.00 6.022
313.15 0.00 3.544 0.15 3.633 0.30 3.742 0.45 3.885 0.60 4.055 0.75 4.237 0.90 4.457
0.05 3.567 0.20 3.666 0.35 3.787 0.50 3.929 0.65 4.110 0.80 4.321 0.95 4.560
0.10 3.591 0.25 3.704 0.40 3.842 0.55 3.992 0.70 4.164 0.85 4.389 1.00 4.672
323.15 0.00 2.792 0.15 2.860 0.30 2.946 0.45 3.061 0.60 3.193 0.75 3.336 0.90 3.499
0.05 2.812 0.20 2.889 0.35 2.986 0.50 3.104 0.65 3.239 0.80 3.388 0.95 3.553
0.10 2.832 0.25 2.917 0.40 3.027 0.55 3.144 0.70 3.285 0.85 3.448 1.00 3.633
Biodiesel + benzene 293.15 0.00 0.732 0.15 1.009 0.30 1.327 0.45 1.734 0.60 2.419 0.75 3.555 0.90 5.373
0.05 0.860 0.20 1.085 0.35 1.462 0.50 1.872 0.65 2.730 0.80 4.069 0.95 6.528
0.10 0.934 0.25 1.206 0.40 1.597 0.55 2.145 0.70 3.041 0.85 4.721 1.00 7.831
298.15 0.00 0.690 0.15 0.948 0.30 1.181 0.45 1.586 0.60 2.204 0.75 3.216 0.90 4.864
0.05 0.801 0.20 1.031 0.35 1.309 0.50 1.736 0.65 2.528 0.80 3.579 0.95 5.904
0.10 0.865 0.25 1.107 0.40 1.437 0.55 1.970 0.70 2.853 0.85 4.252 1.00 6.927
303.15 0.00 0.647 0.15 0.873 0.30 1.113 0.45 1.461 0.60 2.007 0.75 2.920 0.90 4.356
0.05 0.740 0.20 0.921 0.35 1.214 0.50 1.606 0.65 2.317 0.80 3.222 0.95 5.281
0.10 0.815 0.25 1.017 0.40 1.315 0.55 1.807 0.70 2.628 0.85 3.784 1.00 6.022
313.15 0.00 0.582 0.15 0.792 0.30 1.020 0.45 1.305 0.60 1.759 0.75 2.406 0.90 3.468
0.05 0.669 0.20 0.838 0.35 1.116 0.50 1.397 0.65 1.959 0.80 2.662 0.95 4.020
0.10 0.747 0.25 0.929 0.40 1.213 0.55 1.573 0.70 2.160 0.85 3.065 1.00 4.672
323.15 0.00 0.518 0.15 0.688 0.30 0.909 0.45 1.141 0.60 1.496 0.75 2.060 0.90 2.810
0.05 0.600 0.20 0.747 0.35 0.974 0.50 1.262 0.65 1.662 0.80 2.261 0.95 3.246
0.10 0.632 0.25 0.823 0.40 1.029 0.55 1.369 0.70 1.888 0.85 2.535 1.00 3.633
Biodiesel + toluene 293.15 0.00 0.741 0.15 1.024 0.30 1.308 0.45 1.885 0.60 2.621 0.75 3.699 0.90 5.525
0.05 0.821 0.20 1.147 0.35 1.440 0.50 2.100 0.65 2.955 0.80 4.171 0.95 6.164
0.10 0.901 0.25 1.219 0.40 1.670 0.55 2.311 0.70 3.248 0.85 4.863 1.00 7.831
298.15 0.00 0.737 0.15 0.957 0.30 1.229 0.45 1.722 0.60 2.423 0.75 3.370 0.90 4.918
0.05 0.798 0.20 1.049 0.35 1.366 0.50 1.915 0.65 2.672 0.80 3.837 0.95 5.682
0.10 0.864 0.25 1.139 0.40 1.565 0.55 2.210 0.70 2.986 0.85 4.334 1.00 6.927
303.15 0.00 0.734 0.15 0.891 0.30 1.148 0.45 1.535 0.60 2.164 0.75 3.000 0.90 4.311
0.05 0.774 0.20 0.951 0.35 1.273 0.50 1.709 0.65 2.389 0.80 3.353 0.95 5.002
0.10 0.828 0.25 1.044 0.40 1.410 0.55 1.949 0.70 2.705 0.85 3.806 1.00 6.022
313.15 0.00 0.686 0.15 0.831 0.30 1.052 0.45 1.358 0.60 1.845 0.75 2.487 0.90 3.401
0.05 0.724 0.20 0.887 0.35 1.146 0.50 1.506 0.65 2.015 0.80 2.717 0.95 4.000
0.10 0.766 0.25 0.957 0.40 1.251 0.55 1.665 0.70 2.253 0.85 2.984 1.00 4.672
323.15 0.00 0.657 0.15 0.773 0.30 0.973 0.45 1.230 0.60 1.622 0.75 2.147 0.90 2.910
0.05 0.680 0.20 0.828 0.35 1.052 0.50 1.346 0.65 1.772 0.80 2.337 0.95 3.280
0.10 0.721 0.25 0.889 0.40 1.133 0.55 1.483 0.70 1.941 0.85 2.539 1.00 3.633
a
Biodiesel volume fraction.

Please cite this article in press as: Geacai S et al. Measurement, correlation and prediction of biodiesel blends viscosity. Fuel (2014), http://dx.doi.org/
10.1016/j.fuel.2014.11.041
 JFUE 8688 No. of Pages 7, Model 5G
26 November 2014

4 S. Geacai et al. / Fuel xxx (2014) xxx–xxx

Table 4
Refractive index of biodiesel blends with benzene at different temperatures.

System T (K) v1a nD v1 nD v1 nD v1 nD v1 nD v1 nD v1 nD

Biodiesel + benzene 298.15 0.00 1.4968 0.15 1.4895 0.30 1.4824 0.45 1.4757 0.60 1.4695 0.75 1.4639 0.90 1.4585
0.05 1.4944 0.20 1.4871 0.35 1.4801 0.50 1.4737 0.65 1.4678 0.80 1.4618 0.95 1.4570
0.10 1.4921 0.25 1.4848 0.40 1.4779 0.55 1.4715 0.70 1.4656 0.85 1.4605 1.00 1.4548
303.15 0.00 1.4938 0.15 1.4873 0.30 1.4803 0.45 1.4739 0.60 1.4677 0.75 1.4619 0.90 1.4562
0.05 1.4918 0.20 1.4851 0.35 1.4781 0.50 1.4718 0.65 1.4656 0.80 1.4600 0.95 1.4545
0.10 1.4895 0.25 1.4827 0.40 1.4760 0.55 1.4698 0.70 1.4637 0.85 1.4582 1.00 1.4528
313.15 0.00 1.4876 0.15 1.4816 0.30 1.4752 0.45 1.4691 0.60 1.4634 0.75 1.4577 0.90 1.4524
0.05 1.4857 0.20 1.4795 0.35 1.4730 0.50 1.4673 0.65 1.4615 0.80 1.4557 0.95 1.4508
0.10 1.4837 0.25 1.4774 0.40 1.4711 0.55 1.4654 0.70 1.4595 0.85 1.4541 1.00 1.4495
323.15 0.00 1.4800 0.15 1.4737 0.30 1.4677 0.45 1.4625 0.60 1.4576 0.75 1.4524 0.90 1.4478
0.05 1.4779 0.20 1.4717 0.35 1.4661 0.50 1.4610 0.65 1.4559 0.80 1.4508 0.95 1.4462
0.10 1.4759 0.25 1.4699 0.40 1.4644 0.55 1.4592 0.70 1.4542 0.85 1.4493 1.00 1.4452
a
volume fraction of biodiesel.

271
224 temperature increasing. The same behavior present the biodie- 2M 1 þ M 2
M12 ¼ ð5Þ 273
225 sel + diesel fuel and biodiesel + toluene blends [19]. 3
226 From Tables 3 and 4 it can be observed that viscosity and refrac- 274
227 tive index of biodiesel + diesel fuel blends varies in the same man- M1 þ 2M 2
M21 ¼ ð6Þ 276
228 ner with composition and temperature as for biodiesel + benzene 3
229 and biodiesel + toluene blends. xi is the molar fraction of component i, Mi, the molar mass of com- 277
230 The experimental blends kinematic viscosity data were corre- ponent i and gij are the equation parameters adjustable from exper- 278
231 lated with composition, temperature, and both composition and imental data. 279
232 temperature. Similarly, the possibility to calculate the kinematic Wilke’s equation [32] is: 280
233 viscosity from refractive index was studied. The blends with bio- 281
x1 g 1 x2 g 2
234 diesel and diesel fuel were considered pseudo-binary mixtures g¼ þ ð7Þ 283
x1 þ x2 h12 x2 þ x1 h21
235 because both biodiesel and diesel fuel are themselves mixtures of
236 esters and hydrocarbons, respectively. where 284
 285
 1=2  1=4 2
g1 M2
237 3.2. Viscosity–composition correlation 1þ g2 M1
h21 ¼ n h  io1=2 ð8Þ
238 For the viscosity–composition correlation, the Grunberg–Nissan 8 1þ M
M2
1
287
239 (Eqs. (2) and (3)), McAllister (Eq. (4)), and Wilke (Eq. (7)) equations
288
240 were used. The Grunberg–Nissan (Eq. (2)) and McAllister (Eq. (4)) g M1
241 equations are frequently used for viscosity data correlation of h21 ¼ h12  2  ð9Þ
g1 M 2 290
242 molecular mixtures, being semiempirical or theoretical-based
243 models with one or two adjustable parameters in classical thermo- where h12, h21, are quantities depending on viscosity and molar 291
244 dynamics. The Wilke equation is used for petroleum compounds. mass of pure components. 292
245 The Grunberg–Nissan simplified equation and Wilke equation are For all these models, the accuracy was estimated with the per- 293
246 predictive, and the Grunberg–Nissan and McAllister are correla- cent absolute average deviation (AAD): 294
  295
247 tive–predictive equations. n  
248 The Grunberg–Nissan equation [30] is: 100 X gcalc;i  gexp;i 
249 AAD ¼   ð10Þ
n i¼1  gexp;i  297
251 ln g ¼ v 1 ln g1 þ v 2 ln g2 þ v 1 v 2 G12 ð2Þ
where gcalc, gexp, are the calculated and experimental values of vis- 298
252 where g is the blend viscosity, gi and vi are the viscosity and the vol-
cosity, respectively, and n is the number of experimental data. 299
253 ume fraction of component i, respectively, and G12 is the interaction
The results of calculation: the correlation parameters and AAD 300
254 binary parameter. According to Grunberg and Nissan, G12 is
values for Eqs. 2, 3, 4 and (7) are presented in Table 5. From the 301
255 regarded as measure of the strength of interactions between the
table, we can see that for biodiesel + diesel fuel blends, the predic- 302
256 mixing species. In this study, Eq. (2) was considered as a model with
tive Eqs. (3) and (7) allow the viscosity blends calculation with best 303
257 one parameter, the G12 term being calculated by regression analysis.
results for Eq. (7), the Wilke equation. This equation can be used 304
258 Originally proposed by Allen for kinematic viscosity prediction
for predictive calculation of biodiesel + diesel fuel blends viscosity 305
259 of biodiesel from its fatty acid composition [25], Grunberg–Nissan
with 1.49–4.57% accuracy on the studied temperature domain. For 306
260 simplified equation, calculates the blend viscosity:
261 biodiesel + benzene and biodiesel + toluene systems, the Wilke 307
263 ln g ¼ v 1 ln g1 þ v 2 ln g2 ð3Þ equation (Eq. (7)) gives also better results than the Grunberg–Nis- 308
san simplified equation (Eq. (3)), but the calculated error is quite 309
264 McAllister’s two-parameter equation, based on Eyring’s theory
high, 4.83–13.08% on the same temperature domain. 310
265 of absolute reaction rates, takes into account the interactions both
The correlative–predictive equations give better results than 311
266 of like and unlike molecules by two-dimensional three-body inter-
the ones predictive, as we expected. The correlative–predictive 312
267 actions [31]:
268 equations (Eqs. (2) and (4)) represent very well the experimental 313
ln g ¼ x31 ln g1 þ x32 ln g2 þ 3x21 x2 ln g12 þ 3x22 x1 ln g21 data for biodiesel + diesel fuel systems and well, for biodie- 314

þ x31 ln M 1 þ x32 ln M 2  ln ðx1 M1 þ x2 M 2 Þ þ 3x21 x2 ln M 12 sel + benzene and biodiesel + toluene systems. 315
Also, we can notice that the experimental data for the 316
270 þ 3x1 x22 ln M21 ð4Þ
biodiesel + diesel fuel system can be well represented with both 317

Please cite this article in press as: Geacai S et al. Measurement, correlation and prediction of biodiesel blends viscosity. Fuel (2014), http://dx.doi.org/
10.1016/j.fuel.2014.11.041
 JFUE 8688 No. of Pages 7, Model 5G
26 November 2014

S. Geacai et al. / Fuel xxx (2014) xxx–xxx 5

Table 5
Correlation parameters and AADa values for Eqs. (2)–(4) and (7).

System Temperature (K) Models

Grunberg–Nissan Eq. (2) Grunberg–Nissan Eq. (3) McAllister Eq. (4) Wilke Eq. (7)
G12 AAD (%) AAD (%) g12 g21 AAD (%) AAD (%)
Biodiesel + diesel fuel 293.15 0.3192 0.51 5.13 6.7733 5.3957 0.16 4.57
298.15 0.2036 0.23 3.25 6.0925 5.1560 0.11 2.76
303.15 0.1413 0.30 2.28 5.3981 4.5739 0.16 1.78
313.15 0.1380 0.17 2.27 4.1835 3.8154 0.16 1.97
323.15 0.1074 0.12 1.76 3.3251 2.9990 0.11 1.49
Biodiesel + benzene 293.15 0.8432 4.24 14.48 4.2898 6.0789 3.94 12.39
298.15 0.8337 3.21 14.15 4.0116 5.3133 3.25 11.52
303.15 0.7542 2.57 12.69 3.6699 4.7688 2.91 11.43
313.15 0.5372 3.51 9.28 2.6686 4.3847 3.97 12.24
323.15 0.3345 2.22 5.62 2.2417 3.7690 3.35 13.08
Biodiesel + toluene 293.15 0.6504 2.90 11.34 4.2618 5.2015 1.82 12.52
298.15 0.7126 1.95 12.59 4.1296 4.3427 0.86 9.81
303.15 0.8604 1.66 15.41 3.9279 3.4005 1.13 6.18
313.15 0.7596 1.62 13.50 3.0194 2.9199 0.95 4.83
323.15 0.5773 1.08 9.95 2.7401 2.4321 1.02 5.34
a
Calculated with Eq. (10) over 21 compositions at a given temperature.

318 predictive and correlative equations. For biodiesel + benzene and ura et al. for viscosity calculation of individual saturated fatty acid 349
319 biodiesel + toluene blends, the results are weaker. This could be methyl ester (FAME) and their mixtures [21]. Based on Eyring’s 350
320 explained by greater structural difference between the components theory of absolute reaction rates, the equation can be used to 351
321 of biodiesel + benzene and biodiesel + toluene systems, compared estimate the kinematic viscosities of biodiesel blends at various 352
322 with those of biodiesel + diesel fuel system. The different chemical temperatures and compositions: 353
323 nature of the ‘‘pure’’ components from the pseudo-binary mixtures 354
studied, determines different specific intermolecular interactions c d  v1
324 ln g ¼ a þ b  v 1 þ þ ð12Þ 356
325 between the components, affecting the properties of the mixtures. T T

326 3.3. Viscosity–temperature correlation where a, b, c and d are the correlation parameters depending on 357
molar enthalpy and entropy of viscous flow, v1 is the volume frac- 358

327 The viscosities of biodiesel blends with diesel fuel, benzene or tion of biodiesel. 359

328 toluene were correlated with temperature by the Andrade equa- The correlation parameters were calculated by a two-step 360

329 tion [13]: procedure presented by the authors [21]. Eq. (12) may be written 361
330 as Eq. (13): 362
332 g ¼ a  eb=T ð11Þ 363
0
ln g ¼ a0 þ b v ð13Þ 365
333 where a and b are the correlation parameters, T is the absolute tem-
334 perature. The equation is recommended for temperatures below the
335 boiling point of the liquid. where (a0 ) and (b0 ) are: 366

336 The Andrade correlation parameters for kinematic viscosity and 367
c 0d
337 AAD% values for biodiesel blends (10%, 25%, 50% and 75% v/v) are 0
a ¼ aþ b ¼bþ ð14Þ
T T 369
338 presented in Table 6. The accuracy of viscosity–temperature corre-
339 lation with Eq. (11) for biodiesel + diesel fuel blends varies approx- First, the values of (a0 ) and (b0 ) are the intercepts and slopes, 370
340 imately in the similar range for all compositions, AAD varying from respectively, of the plots ln g(v1) for different blends. Then, a, c 371
341 1.03% to 1.63%. For biodiesel + benzene and biodiesel + toluene and b, d, are the intercepts and slopes of straight lines resulted 372
342 blends, the error in viscosity–temperature correlation with Eq. from the plots of (a0 ) against (1/T) and (b0 ) against (1/T), respec- 373
343 (11) varies from 0.31% to 1.61% in terms of AAD. The Andrade tively. Their values from Eq. (12) and AAD for each biodiesel blend 374
344 equation represents well the viscosity variation with temperature are presented in Table 7. The obtained equations can be used prac- 375
345 for all studied systems. tically to calculate the viscosity of the blends at different composi- 376
tions and temperatures. The equations correlate satisfactorily the 377
346 3.4. Viscosity–composition–temperature correlation experimental data for biodiesel + diesel fuel system, AAD being 378
3.20–5.08%, and unsatisfactorily for biodiesel + benzene and 379
347 Viscosity dependence on temperature and composition for bio- biodiesel + toluene systems with AAD values between 3.10% and 380
348 diesel blends was evaluated with Eq. (12) proposed by Krisnangk- 17.56% for the studied temperature domain. 381

Table 6
Correlation parameters and AAD values for Eq. (11).

v1a Biodiesel + diesel fuel Biodiesel + benzene Biodiesel + toluene

a b AAD (%) a b AAD (%) a b AAD (%)
0.10 0.0056 2018.13 1.63 0.0170 1175.15 1.28 0.0798 709.78 0.31
0.25 0.0061 2000.99 1.58 0.0222 1166.95 1.08 0.0403 994.58 1.46
0.50 0.0051 2079.56 1.39 0.0252 1261.26 0.71 0.0174 1399.15 1.61
0.75 0.0032 2252.80 1.03 0.0093 1741.56 0.39 0.0095 1747.09 0.97
a
Volume fraction of biodiesel.

Please cite this article in press as: Geacai S et al. Measurement, correlation and prediction of biodiesel blends viscosity. Fuel (2014), http://dx.doi.org/
10.1016/j.fuel.2014.11.041
 JFUE 8688 No. of Pages 7, Model 5G
26 November 2014

6 S. Geacai et al. / Fuel xxx (2014) xxx–xxx

Table 7 7.1
Results of correlation with Eq. (12).

System Equations Temperature AAD

(K) (%)
6.6

ln g ¼ 4:8462  1:2728  v 1

Viscosity, mm2/s
Biodiesel + diesel fuel 293.15 3.32
þ 1903:4 þ 458:99v1 298.15 5.03
T T 6.1
303.15 3.88
313.15 5.08
323.15 3.20
5.6
Biodiesel + benzene ln g ¼ 3:3521  1:8728  v 1 293.15 6.61
þ 870:14 þ 1210:8 v1 298.15 7.00
T T
303.15 6.59 5.1
313.15 4.57
323.15 3.10
Biodiesel + toluene ln g ¼ 2:3436  3:5037  v 1 293.15 17.30 4.6
v1 298.15 17.56 1.450 1.455 1.460 1.465 1.470
þ 512:98
T þ 1716:5
T
303.15 15.75 refracve index
313.15 15.80
323.15 17.46 (A)
v1 – volume fraction of biodiesel.
8.0
382 3.5. Viscosity–refractive index correlation

383 Experimental determination of refractive index is simpler than 6.0

384 of viscosity. Because of this, the viscosity calculation from refrac-
385 tive index data is useful in many practical cases. In this work, the Viscosity, mm2/s
386 viscosity of biodiesel blends at different temperatures was esti-
387 mated from blend refractive index with Eq. (15) used to evaluate 4.0
388 the viscosity of liquid hydrocarbons and petroleum fractions [33]:
389
1 b 2.0
¼aþ ð15Þ
391 g nD

392 where nD is the refractive index of the blend, a, b are the correlation
393 parameters. 0.0
1.450 1.460 1.470 1.480 1.490 1.500
394 The values of correlation parameters, a and b from Eq. (15) and
395 AAD for each biodiesel blends are presented in Table 8. The refracve index
396 accuracy of viscosity calculation using Eq. (15) is relatively good (B)
397 for biodiesel + diesel fuel blends. The smallest AAD when using
398 Eq. (15) was 0.39%, and the greatest was 1.18%. The errors in vis-
399 cosity evaluation with Eq. (15) for biodiesel + benzene and, respec-
8.0
400 tively, biodiesel + toluene systems are greater (7.22–12.73%) than
401 that corresponding to biodiesel + diesel fuel systems. Eq. (15) is
402 recommended to calculate the viscosity from refractive index of
403 biodiesel + diesel fuel blends. 6.0
Viscosity, mm2/s

404 The correlation between viscosity and refractive index for bio-
405 diesel + diesel fuel, biodiesel + benzene and biodiesel + toluene
406 blends respectively, is presented in Fig. 1. The points correspond
4.0
407 to measured viscosity values, and the lines represent the calculated
408 viscosity with Eq. (15).

2.0
Table 8
Q6 Correlation parameters and AAD values for Eq. (15).

Temperature (K)
0.0
298.15 303.15 313.15 323.15 1.450 1.460 1.470 1.480 1.490 1.500
Biodiesel + diesel fuel refracve index
a 9.4491 10.9499 10.5771 12.9882
b 13.5316 15.6657 15.0159 18.3691 (C)
AAD (%) 1.18 0.55 0.59 0.39
Biodiesel + benzene Fig. 1. Correlation between viscosity and refractive index for Bio + D (A), Bio + B (B)
a 37.8186 40.6241 46.8127 57.7720 and Bio + T (C); , experimental values; –, calculated values with Eq. (15).
b 54.8417 58.8121 67.5739 83.1327
AAD (%) 8.16 8.19 7.22 7.32 4. Conclusion 409
Biodiesel + toluene
a 37.3353 40.1131 45.6986 49.4029 Experimental viscosities data for binary blends of biodiesel with 410
b 54.1366 58.0744 65.9723 71.0482
AAD (%) 12.73 11.61 8.73 7.55
diesel fuel, benzene or toluene were reported. Also, refractive 411
indices data for biodiesel + benzene system were presented. The 412

Please cite this article in press as: Geacai S et al. Measurement, correlation and prediction of biodiesel blends viscosity. Fuel (2014), http://dx.doi.org/
10.1016/j.fuel.2014.11.041
 JFUE 8688 No. of Pages 7, Model 5G
26 November 2014

S. Geacai et al. / Fuel xxx (2014) xxx–xxx 7

413 accuracy of different models from classical thermodynamics and [8] Seddon RH. Vegetable oils in commercial vehicles. Gas Oil Power 1942:136–46. 463
[9] Monteiro MR, Ambrozin ARP, Lião LM, Ferreira AG. Critical review on analytical 464
414 from practice of petroleum compounds proposed to correlate/pre- 465
methods for biodiesel characterization. Talanta 2008;77:593–605.
415 dict the viscosity was evaluated. [10] Joshi RM, Pegg MJ. Flow properties of biodiesel fuel blends at low 466
416 The correlative–predictive equations (Grunberg–Nissan and temperatures. Fuel 2007;86:143–51. 467
[11] Pinto AC, Guarieiro LN, Rezende MJC, Ribeiro NM, Torres EA, Lopes WA. 468
417 McAllister) give better results than the ones predictive (Grun- Q5 469
Biodiesel: an overview. J Braz Chem Soc 2005;16:1313–30.
418 berg–Nissan simplified equation and Wilke) for all studied [12] Makhija RC, Stairs RA. Viscosity of associated liquids. Can J Chem 470
419 systems. 1970;48:1214–8. 471
[13] Tat ME, Van Gerpen JH. The kinematic viscosity of biodiesel and its blends with 472
420 The experimental data for biodiesel + diesel fuel system can be
diesel fuel. J Am Oil Chem Soc 1999;76:1511–3. 473
421 well represented by both predictive and correlative equations. For [14] Liew KY, Seng CE, Oh LL. Viscosities and densities of the methyl esters of some 474
422 biodiesel + benzene and biodiesel + toluene blends, the results are n-alcanoic acids. J Am Oil Chem Soc 1992;69:155–8. 475
423 weaker, due to greater structural difference between the ‘‘pure’’ [15] Perry RH, Green DW. Perry’s chemical engineers’ handbook. 7th 476
ed. McGrawHill; 1997. 477
424 components of the systems, which determines different specific [16] Alptekin E, Canakci M. Determination of the density and the viscosity of 478
425 intermolecular interactions between the components affecting biodiesel-diesel fuel blends. Renew Energy 2008;33:2623–30. 479
426 the properties of the mixtures. [17] Moser BR, Vaughn SF. Evaluation of alkyl esters from Camelina sativa oil as 480
biodiesel and as blend components in ultra-low-sulfur diesel fuel. Bioresour 481
427 The Grunberg–Nissan simplified equation and Wilke equation Technol 2010;101:646–53. 482
428 can predict the viscosity of blends with best results for biodie- [18] Nita I, Geacai S, Iulian O. Measurements and correlations of physico-chemical 483
429 sel + diesel fuel blends. The Wilke equation gives better results properties to composition of pseudo-binary mixtures with biodiesel. Renew 484
Energy 2011;36:3417–23. 485
430 than the Grunberg–Nissan simplified equation for all systems. 486
[19] Geacai S, Nitßă I, Iulian O, Geacai E. Refractive indices for biodiesel mixtures.
431 The prediction with Grunberg–Nissan modified equation is less U.P.B. Sci. Bull. 2012;74:149–60. 487
432 accurate, probably because this equation is recommended for [20] Benjumea P. Basic properties of palm oil biodiesel–diesel blends. Fuel 488
2008;87:2069–75. 489
433 ‘‘ideal’’ mixtures that is not the case of studied systems. 490
[21] Krisnangkura K, Chanida S, Kornkanok A, Supathra L. An empirical approach
434 Empirical equations to correlate the viscosity with refractive for predicting kinematic viscosities of biodiesel blends. Fuel 2010;89:2775–80. 491
435 index for biodiesel with diesel fuel and biodiesel with benzene or [22] Tesfa B, Gu F, Mishra R, Ball AD. LHV predication models and LHV effect on the 492
performance of CI engine running with biodiesel blends. Energy Convers 493
436 toluene blends were proposed. A relatively good relation between
Manage 2013;71:217–26. 494
437 viscosity and refractive index was obtained for biodiesel with die- [23] Anand K, Sharma RP, Pramod SM. A comprehensive approach for estimating 495
438 sel fuel systems. Therefore, the equations type (15) can be used for thermo-physical properties of biodiesel fuels. Appl Therm Eng 496
439 the prediction of biodiesel + diesel fuel systems viscosity from 2011;31:235–42. 497
[24] Ramirez-Verduzco LF, Garcia-Flores BE, Rodriguez-Rodriguez JE, Jaramillo- 498
440 refractive index measurements. Jacob A del R. Prediction of the density and viscosity in biodiesel blends at 499
441 The viscosity variation with temperature is represented well for various temperatures. Fuel 2011;90:1751–61. 500
442 all studied systems by the Andrade equation. The viscosity depen- [25] Allen CAW, Watts KC, Ackman RG, Pegg MJ. Predicting the viscosity of 501
biodiesel fuels from their fatty acid ester composition. Fuel 1999;78:1319–26. 502
443 dence on temperature and composition is represented satisfacto- [26] Moser BR. Efficacy of specific gravity as a tool for prediction of biodiesel- 503
444 rily by Krisnangkura et al. equation for biodiesel + diesel fuel and petroleum diesel blend ratio. Fuel 2012;99:254–61. 504
445 less well for biodiesel + benzene and biodiesel + toluene systems. [27] Chen YH, Huang BY, Chiang TH, Tang TC. Fuel properties of microalgae 505
(Chlorella protothecoides) oil biodiesel and its blends with petroleum diesel. 506
Fuel 2012;94:270–3. 507
446 References [28] Moser BR. Impact of fatty ester composition on low temperature properties of 508
biodiesel–petroleum diesel blends. Fuel 2014;115:500–6. 509
447 [1] Srivastava A, Prasad R. Triglycerides-based diesel fuels. Renew Sustain Energy [29] Vural US, Durmaz F, Kocyigit O, Kocyigit H, Muradoglu V, Akin B. Excess molar 510
448 Rev 2000;4:111–33. volumes, viscosity, refractive index and Gibbs energy of activation of binary 511
449 [2] Ma F, Hanna MA. Biodiesel production: a review. Bioresour Technol biodiesel + benzene and biodiesel + toluene mixtures at 298.15 and 303.15 K. 512
450 1999;70:1–15. Russ J Phys Chem A 2008;82:2260–8. 513
451 [3] Ramadhas AS, Jayaraj S, Muraleedharan C. Use of vegetable oils as I.C. Engine [30] Grunberg L, Nissan AH. Mixture law for viscosity. Nature 1949;164:799–800. 514
452 fuels – a review. Renew Energy 2004;29:727–42. [31] McAllister RA. AIChEJ 1960;6:427–31. cited by; 515
453 [4] Gui MM, Lee K, Bhatia S. Feasibility of edible oil vs. non-edible oil vs. waste Soliman K, Marschall E. J Chem Eng Data 1990;35:375–81; 516
454 edible oil as biodiesel feedstock. Energy 2008;33:1646–53. Dominguez M, Pardo JI, Gascon I, Royo FM, Urieta JS. Fluid Phase Equilib 517
455 [5] Gunstone FD. Market report. Lipid Technol 2009;21:207. 2000;169:277–92; 518
456 [6] Korbitz W. Biodiesel production in Europe and North America, an encouraging Matos JS, Trenzado JL, Gonzales E, Andrade R. Fluid Phase Equilib 519
457 prospect. Renew Energy 1999;16:1078–83. 2001;186:207–34. 520
458 [7] Bacon DM, Brear F, Moncrieff ID, Walker KL. The use of vegetable oils in [32] Wilke CR. A viscosity equation for gas mixtures. J Chem Phys 1950;18:517–9. 521
459 straight and modified form as diesel engine fuels. In: Fazzolare RA, Smith CR, [33] Kanit K, Tawatchai Y, Ratanachai P. An empirical approach in predicting 522
460 editors. Beyond the energy crisis-opportunity and challenge. Third biodiesel viscosity at various temperatures. Fuel 2006;85:107–13. 523
461 international conference on energy use management, Berlin (West), vol. 524
462 III. Oxford: Pergamon Press; 1981. p. 1525–33.

Please cite this article in press as: Geacai S et al. Measurement, correlation and prediction of biodiesel blends viscosity. Fuel (2014), http://dx.doi.org/
10.1016/j.fuel.2014.11.041