Article pubs.acs.org/IECR
Study of Two Different Types of Minimum Flash-Point Behavior for Ternary Mixtures Horng-Jang Liaw*,†,‡ and Hao-Ying Chen‡ †
Department of Safety, Health and Environmental Engineering, National Kaohsiung First University of Science and Technology, Kaohsiung, Taiwan, ROC ‡ Department of Occupational Safety and Health, China Medical University, Taichung, Taiwan, ROC S Supporting Information *
ABSTRACT: Binary mixtures exhibiting minimum flash point behavior (MinFPB) are more hazardous than their pure, individual components. In real processes, multicomponent solutions are more frequently encountered than binary systems, but analyses of MinFPB for multicomponent mixtures are very rare. We investigated the MinFPB of ternary mixtures, with two minimum flash point binary mixtures, including isopropyl alcohol + ethanol + octane and 2-butanol + ethanol + octane (and their component binary combinations), and one ternary mixture, cyclohexanol + ethanol + octane, with a single binary mixture exhibiting MinFPB, and its component binary mixtures. Two different types of MinFPB were observed for the three ternary solutions, one with two constituent binary mixtures exhibiting MinFPB and one with a single binary mixture exhibiting MinFPB. in order to prevent confusion with maximum flash point behavior.16 In the real world, multicomponent mixtures are much more frequently encountered than binary mixtures; nonetheless, discussions of MinFPB are almost always limited to binary mixtures. In a comprehensive literature review, one study relevant to MinFPB of ternary mixtures was identified, which reported the elimination of MinFPB by addition of a specific third component.12 A flash point prediction model is necessary to describe the MinFPB of mixtures. Several alternative models for predicting the flash points of different types of mixtures are proposed. Models based on the assumption of ideal solutions,6,17−19 which are the most frequent, show limitations when applied to nonideal mixtures,1−3,20 thus, they cannot be applied to minimum flash-point solutions. Models taking into account the nonideality of the solution through liquid-phase activity coefficients have a wider application range and efficiently predict the flash points of miscible mixtures.1−3,13,20−22 Nonideality of the liquid phase is particularly responsible for the occurrence of extreme flash-point behavior such as minimum and maximum flash-point behaviors,10−12,15,16,23 with strong implications for fire and explosion hazard assessment of the mixtures. The extreme behavior is set in parallel with minimum-boiling and maximum-boiling azeotropic behavior in the vapor−liquid equilibrium (VLE). The general model for predicting the flash points of miscible mixtures proposed previously efficiently predicts the experimental data of ternary mixtures3 and is the most frequently used model by many authors.8,10,14,24−27 Such a model was used in this study to describe the MinFPB of ternary mixtures.
1. INTRODUCTION Several accidents, such as the Shengli illegal dumping incident in 2000 in Taiwan,1−3 a series of explosions of essential oils in 2003 in Taiwan,3 and the crash of a gasoline tanker that burst into flames resulting in a bridge melting and collapsing in 2007 in San Francisco4,5 highlight the importance of safe storage, usage, and transportation of flammable and combustible liquids. The fire and explosion hazards of liquids are primarily characterized by their flash points.6 The flash point is defined as the temperature determined experimentally at which a liquid emits sufficient vapor to form a combustible mixture with air.7 In addition, the flash point involves the operability of storage, ignition, and combustion of a fuel.8 The behavior of flash point variation for mixtures is applicable to fuel design. Ellis first pointed out in 1976 that a mixture of butyl alcohol and mineral spirits exhibited a minimum flash point,9 defined as a flash point less than that of each of the pure components.1 Nonetheless, the first successful estimation of the flash point of a mixture, ethanol + octane, exhibiting minimum flash-point behavior (MinFPB) was not reported until 2002.1 A liquid solution demonstrating MinFPB is more hazardous than the individual components of the combination because the flash point of the mixture, over a range of compositions, is lower in comparison to its constituents.1 This special behavior for a specific mixture, such as ethanol and octane, is attributable to the fact that the repulsive interaction between different molecules is much stronger than that between the same molecules, such that more molecules escape into the vapor phase than for an ideal mixture, so the solution’s flash point is substantially lower.10 This behavior has also been demonstrated in mixtures of 2-butanol + octane, isopropyl alcohol (IPA) + octane, methanol + methyl acrylate, isobutanol + toluene, and ethylbenzene + n-propanol;11−15 such products are termed minimum flash-point solutions and their behavior is termed MinFPB.11 In this manuscript, we use the acronym MinFPB rather than MFPB, as used by Vidal et al.10 and Catoire et al.,15 © 2013 American Chemical Society
Received: Revised: Accepted: Published: 7579
February 22, 2013 May 9, 2013 May 16, 2013 May 16, 2013 dx.doi.org/10.1021/ie400585k | Ind. Eng. Chem. Res. 2013, 52, 7579−7585
Industrial & Engineering Chemistry Research
Article
2.2. General Model for Predicting the Flash Point of Miscible Mixtures. The flash point of a miscible mixture can be estimated by the modified Le Chatelier equation, the Antoine equation, and a model for estimating activity coefficients, γi3
Recently, the predictions of the general model for miscible mixtures combined with the original Universal quasichemical Functional Group Activity Coefficients (UNIFAC) and the modified UNIFAC-Dortmund 93 models showed good agreement with the measured flash points.28 Thus, the original UNIFAC model and the modified UNIFAC-Dortmund 93 model were used as well as the nonrandom two-liquid (NRTL) model to calculate the liquid phase activity coefficients for the estimation of flash points. The purpose of this study was to investigate the MinFPB for ternary mixtures as applied in hazard assessment and fuel design. In order to ensure the accuracy of the experimental data, we compared it with its predicted analogues. Recovery of alcohol solvents used in the reprocessing of nuclear fuel was done by extraction with aqueous hydrocarbon solutions.29 The binary mixtures, IPA + octane, 2-butanol + octane, and ethanol + octane are minimum flash-point solutions.1,11,12 We selected three ternary systems including IPA + ethanol + octane, 2butanol + ethanol + octane, and cyclohexanol + ethanol + octane for investigation of MinFPB.
1=
∑ i ≠ kl
xiγiPisat Pisat , fp
log Pisat = Ai −
(1)
Bi T + Ci
(2)
where the summation runs over all flammable components, kl is the nonflammable components of the mixture. Psat i,fp, the vapor pressure of the pure flammable component, i, at its flash point, Ti,fp, can be estimated using the Antoine equation. The activity coefficient, γi, can be estimated using a thermodynamic model, such as the NRTL or UNIFAC-type model, as used in this study. For a ternary solution of flammable solvents, eq 1 can be reduced as follows: 1=
2. MATERIALS AND METHODS 2.1. Experimental Protocol. An HFP 362-Tag Flash Point Analyzer (Walter Herzog GmbH, Germany), which meets the requirements of the American Society for Testing and Materials (ASTM) D56 standard,30 was used to measure the flash points of a variety of ternary minimum flash point solutions (IPA + ethanol + octane, 2-butnaol + ethanol + octane, and cyclohexanol + ethanol + octane) with different compositions. The apparatus consisted of an external cooling system, test cup, heating block, electric igniter, sample thermometer, thermocouple (sensor for fire detection), and indicator/operating display. The apparatus incorporated control devices that programmed the instrument to heat the sample at a specified rate within a temperature range close to the expected flash point. The literature data and the estimated data based on our previously proposed model3 were used as the expected flash points for pure substances and mixtures, respectively. The flash point was automatically tested using an igniter at specified temperature test intervals. If the expected flash point was lower than or equal to the change temperature, heat rate-1 was used, and the igniter was fired at test interval-1. If the expected flash point was higher, heat rate-2 was adopted, and the igniter was fired at test interval-2. Using the standard method, the change temperature was laid down by the standard and could not be changed. The first flash-point test series was initiated at a temperature equivalent to the expected flash point minus the start-test value. If the flash point was not determined when the test temperature exceeded the sum of the expected flash point plus the end-of-test value, the experimental iteration was terminated. The instrument operation was conducted according to the standard ASTM D56 test protocol30 using the following parameters: start test, 5 °C; end of test, 20 °C; heat rate-1, 1 °C/min; heat rate-2, 3 °C/min; change temperature, 60 °C; test interval-1, 0.5 °C; and test interval-2, 1.0 °C. The liquid mole fraction was determined from the mass measured using a Setra digital balance (EL-410D: sensitivity, 0.001 g; maximum load, 100 g). The prepared mixtures were stirred by a magnetic stirrer for 30 min before the flash point test. Isopropyl alcohol (99.8%) was HPLC/Spectro grade reagents (Tedia Co. Inc.; USA). Ethanol (99.8%), octane (99%), 2-butanol (99%), and cyclohexanol (99%) were obtained from Panreac (Spain).
∑ i ≠ kl
xiγiPisat Pisat , fp
=
x1γ1P1sat P1,satfp
+
x 2γ2P2sat P2,satfp
+
x3γ3P3sat P3,satfp
(3)
The temperature from the solution of eqs 2 and 3 and the equations to describe liquid phase activity coefficients are deemed to be the flash point of the ternary mixtures.
3. RESULTS AND DISCUSSION 3.1. Parameters Used. The general flash-point model for miscible mixtures was used to estimate the flash points of three ternary mixtures, IPA + ethanol + octane, 2-butanol + ethanol + octane, and cyclohexanol + ethanol + octane; these data were compared with the corresponding experimental data. Liquidphase activity coefficients were estimated by using the NRTL,31 original UNIFAC,32,33 and UNIFAC-Dortmund 93 models.34,35 Binary interaction NRTL parameters regressed from the VLE data were used along with parameters adopted from the literature (Table 1).29,36−39 The group volume and surface area Table 1. VLE Parameters of the NRTL Equation for the Binary Solutions
*
mixture
A12
A21
α12
reference
IPA (1) + ethanol (2) IPA (1) + octane (2) ethanol (1) + octane (2) 2-butanol (1) + ethanol (2) 2-butanol (1) + octane (3)
−61.71 450.21 651.91 403.90 289.50
60.58 478.35 604.97 −281.74 406.51
0.3040 0.4694 0.47 0.2721 0.47
36 37 38 39 29
Aij = (gij−gjj)/R.
parameters, and the UNIFAC group interaction parameters for the two UNIFAC-type models were obtained from the literature;32−35 the pure component volume parameters and pure component area parameters are listed in Table 2. The pure compound data are listed in Table 3. The pure flash points were measured using the flash point analyzer, and the Antoine coefficients were obtained from the literature.37,38,40 3.2. Comparison of Simulated and Measured Flash Points. Experimental flash-point data for the three ternary mixtures covering their entire composition range are given in the Supporting Information and are displayed in Figures 1−3. The standard deviations of most of these data are less than 1 7580
dx.doi.org/10.1021/ie400585k | Ind. Eng. Chem. Res. 2013, 52, 7579−7585
Industrial & Engineering Chemistry Research
Article
Table 2. Pure Component Volumes and Surface Areas Used in the UNIFAC-Type Models original UNIFAC
UNIFAC Dortmund
compound
ri
qi
ri
qi
2-butanol cyclohexanol ethanol isopropyl alcohol octane
3.9235 4.8189 2.5755 3.2491 5.8486
3.664 4.128 2.588 3.124 4.936
3.5930 4.6310 2.4952 2.9605 5.0600
4.0514 5.1838 2.6616 3.3433 6.3702
Table 3. Antoine Coefficients and Experimental Flash Point for Studied Solution Components Antoine coefficients substance 2-butanola cyclo hexanolb ethanola isopropyl alcohola octanea a
A
B
C
reference
Tfp,exp/°Cc
6.32690 8.35237
1157.363 2258.560
−104.830 −21.376
40 37
22.0 ± 2.4 67.2 ± 3.5
7.24222 7.56634
1595.811 1366.142
−46.702 −75.030
40 38
13.0 ± 0.6 12.9 ± 0.8
6.04394
1351.938
−64.030
38
14.5 ± 1.4
Figure 2. Comparison of predicted and experimental flash points for the ternary mixture of 2-butanol (1) + ethanol (2) + octane (3) (complete data sets are available from the authors upon request). □, experimental flash points; blue •, predicted flash points calculated using the original UNIFAC model; red •, predicted flash points calculated using the UNIFAC Dortmund 93 model; green •, predicted flash points calculated using the NRTL model.
b
log(P/kPa) = A−B/[(T/K)+C]. log(P/mmHg) = A−B/[(T/ K)+C]. cThe uncertainty is represented by the value of double standard deviation.
Figure 1. Comparison of predicted and experimental flash points for the ternary mixture of IPA (1) + ethanol (2) + octane (3) (complete data sets are available from the authors upon request). □, experimental flash points; blue •, predicted flash points calculated using the original UNIFAC model; red •, predicted flash points calculated using the UNIFAC Dortmund 93 model; green •, predicted flash points calculated using the NRTL model.
Figure 3. Comparison of predicted and experimental flash points for the ternary mixture of cyclohexanol (1) + ethanol (2) + octane (3) (complete data sets are available from the authors upon request). □, experimental flash points; blue •, predicted flash points calculated using the original UNIFAC model; red •, predicted flash points calculated using the UNIFAC Dortmund 93 model.
°C, and the greatest standard deviation was that for cyclohexanol, 1.7 °C. The cyclohexanol + ethanol + octane mixture’s constituent binary mixture of ethanol + octane exhibited a minimum flash-point. The IPA + ethanol + octane mixture and the 2-butanol + ethanol + octane mixture had two constituent binary mixtures that exhibited MinFPB. Figures 1−3 also depict the predicted flash points for the ternary mixtures calculated from the general flash point prediction model using the different equations to estimate the
activity coefficients; each is consistent with the experimentally determined data, although there are differences between them. Table 4 shows that there were no significant differences in the predictive capabilities of the NRTL, original UNIFAC, or UNIFAC-Dortmund 93 equations used to estimate the activity 7581
dx.doi.org/10.1021/ie400585k | Ind. Eng. Chem. Res. 2013, 52, 7579−7585
Industrial & Engineering Chemistry Research
Article
range were moderate, and their values were almost between 4.7 and 6.4 °C (Figures 1 and 4). In contrast, the flash points variations of the IPA + ethanol + octane mixture were substantial over a wide composition range around the constituent binary IPA + ethanol mixture and pure octane (Figures 1 and 4). Figure 4 shows that the shapes of flash point variation curves are very similar for each ratio of IPA/ethanol, with flash points approaching each minimum value over a wide composition range. The trend of the experimental data indicated that the minimum flash point increased as the ratio of IPA/ethanol increased, although two measured flash-point values, 4.5 and 4.6 °C, slightly less than the minimum flash point of ethanol + octane, 4.7 °C, were observed. Because of measurement uncertainty, we cannot conclude that the minimum flash point of the IPA + ethanol + octane mixture is less than that of the constituent binary mixture of ethanol + octane. 3.3.2. 2-Butanol + Ethanol + Octane. For the 2-butanol + ethanol + octane mixture, its constituent binary mixtures of ethanol + octane and 2-butanol + octane both exhibited MinFPB (Figure 5, 2-butanol/ethanol = 0 and ∞). The flash
Table 4. Average Temperature Deviation between Calculated and Experimental Flash Points, ΔTfpa, for the Studied Solutions, Comparing Models ΔTfp/K mixture
NRTL
original UNIFAC
UNIFAC Dortmund
IPA + ethanol + octane 2-butanol + ethanol + octane cyclohexanol + ethanol + octane
0.42 0.49 -
0.51 0.40 0.90
0.35 0.38 0.93
a
Deviation of flash point: ΔTfp = ∑N|Tfp,exp − Tf p,pred|/N.
coefficients. The differences in flash point were very small. Since the general flash point prediction model described the ternary mixtures exhibiting MinFPB as shown in Figures 1−3 and Table 4, the predicted flash points provided by such a model were used to assist in the analysis of MinFPB for each ternary mixture. 3.3. MinFPB of Ternary Mixtures. 3.3.1. IPA + Ethanol + Octane. The IPA + ethanol + octane mixture’s constituent binary mixtures of ethanol + octane and IPA + octane exhibited MinFPB (Figure 4, IPA/ethanol = 0 and ∞). The flash points
Figure 5. Two dimensional plot of predicted and experimental flash points determined at a constant ratio of 2-butanol to ethanol for 2butanol (1) + ethanol (2) + octane (3). □, experimental flash points; , predicted flash points calculated using the original UNIFAC model; − − −, predicted flash points calculated using the UNIFAC Dortmund 93 model; − - −, predicted flash points calculated using the NRTL model.
Figure 4. Two dimensional plot of predicted and experimental flash points determined at a constant ratio of IPA to ethanol for IPA (1) + ethanol (2) + octane (3). □, experimental flash points; , predicted flash points calculated using the original UNIFAC model; − − −, predicted flash points calculated using the UNIFAC Dortmund 93 model; − - −, predicted flash points calculated using the NRTL model.
points for 2-butanol + octane over some of the composition range were almost constant around the minimum measured value of 10.7 °C, which is lower than that of 2-butanol and octane at 22.0 and 14.5 °C, respectively. As an aliquot of 2butanol was replaced with ethanol, the flash point of the ternary mixture decreased (Figure 5). At the limit of no 2-butanol, the flash point approached the binary mixture of ethanol + octane at the minimum measured value of 4.7 °C, over a wide composition range, which was the minimum flash point for the 2-butanol + ethanol + octane mixture. Inspection of Figures 1, 2, 4, and 5 shows that the flash point variation trends for the two ternary mixtures (IPA + ethanol + octane and 2-butanol + ethanol + octane) exhibited MinFPB, with two local minimum flash points similar to the flash points
of ethanol, IPA, and octane were 13.0 °C, 12.9 °C, and 14.5 °C, respectively. However, the flash points for the binary mixtures of ethanol + octane and IPA + octane were about 5.0 and 6.5 °C, respectively, over a wide composition range. Thus, the flash points of both binary mixtures were lower than the flash points of the three separate components. In contrast, the IPA + ethanol mixture behaved ideally, and its flash point was almost the same as its constituent flash points. The flash points within a wide composition range for ethanol + octane and IPA + octane approached their minimum constituent measured values of 4.7 and 6.4 °C, respectively, which is reflected in the ternary mixture of IPA + ethanol + octane. The flash point variations within a wide composition 7582
dx.doi.org/10.1021/ie400585k | Ind. Eng. Chem. Res. 2013, 52, 7579−7585
Industrial & Engineering Chemistry Research
Article
addition of a high-boiling point compound (cyclohexanol) to the low-boiling compound octane, except when a substantial quantity of cyclohexanol was added (see Figure 6, cyclohexanol/ethanol = ∞). This behavior was also observed for other highly nonideal solutions, without exhibiting MinFPB, or just barely exhibiting MinFPB, owing to the great difference in flash points for the individual components, such as 1-butanol + octane, methanol + methyl acrylate, and isoamyl alcohol + isoamyl acetate.1,11,28 Since a highly positive deviation mixture, minimum-boiling azeotrope, is encountered considerably more often than a highly negative deviation mixture, maximumboiling azeotrope, in real-world environments,41 the existence of a small quantity of contaminating low-boiling compound always makes the original chemical much more dangerous, and it is usually not possible to reduce the fire and explosion hazard of a chemical significantly with addition of a high-boiling point compound, except for addition of huge quantities of the highboiling point compound. The flash point of cyclohexanol at 67.2 °C is much greater than that of ethanol and octane at 13.0 and 14.5 °C, respectively. Figures 6 and 7 indicate that addition of
of their constituent binary mixtures. The small and relatively greater difference in minimum flash point between their two minimum flash point binary mixtures for IPA + ethanol + octane and 2-butanol + ethanol + octane resulted in moderate and relative significant flash point variation in a similar wide composition range, respectively. 3.3.3. Cyclohexanol + Ethanol + Octane. For cyclohexanol + ethanol + octane, a single minimum flash-point was observed only in the constituent binary mixture of ethanol + octane (Figure 6, cyclohexanol/ethanol = 0). Additionally, the binary
Figure 6. Two dimensional plot of predicted and experimental flash points determined at a constant ratio of cyclohexanol to ethanol for cyclohexanol (1) + ethanol (2) + octane (3). □, experimental flash points; , predicted flash points calculated using the original UNIFAC model; − − −, predicted flash points calculated using the UNIFAC Dortmund 93 model.
mixture of cyclohexanol + ethanol behaved as an ideal solution. As proposed in our previous study,11 the conditions sufficient for a binary mixture to form a minimum flash point solution are
γ1∞P1sat |T2,fp
(5)
Figure 7. Two dimensional plot of predicted and experimental flash points determined at a constant molar fraction of cyclohexanol for the ternary mixture of cyclohexanol (1) + ethanol (2) + octane (3). □, experimental flash points; , predicted flash points calculated using the original UNIFAC model; − − −, predicted flash points calculated using the UNIFAC Dortmund 93 model.
These equations imply that a binary mixture with a small difference in flash point for individual components and highly positive nonideal behavior has a greater possibility of forming a sat sat minimum flash point solution, where Psat 1 |T2,f p/P1,fp and P2 |T1,f p/ ∞ ∞ Psat 2,f p are relevant to the former, and γ1 |T2,f p and γ2 |T1,f p are relevant to the latter. The binary mixtures of ethanol + octane, IPA + octane, and 2-butanol + octane are examples. Since the flash point of cyclohexanol, 67.2 °C, is much greater than that of octane at 14.5 °C, the MinFPB was not observed in the binary mixture of cyclohexanol + octane, although such a mixture is a highly positive nonideal solution. The highly nonideal behavior made the flash point decrease sharply as a small quantity of low-boiling point compound, such as octane, was added to the high-boiling point compound cyclohexanol; only a slight increment in flash point was observed upon
cyclohexanol to the binary minimum flash point solution of ethanol + octane increased the flash point of the solution. Thus, the minimum flash point for cyclohexanol + ethanol + octane was the same as the minimum flash point of the binary mixture of ethanol + octane. As the quantity of cyclohexanol increased, the MinFPB became less conspicuous. If the molar fraction of cyclohexanol was increased to 0.8, the MinFPB disappeared (Figure 7). Elimination of the MinFPB of binary mixtures is possible by addition of a third component; however, the flash point of the mixture did not increase for our previously selected examples.12 In the case of adding cyclohexanol to ethanol + octane, not only was the MinFPB eliminated but also the flash point of the mixture increased. This phenomenon indicated that it is possible to reduce the fire and explosion hazard of
P1,satfp γ2∞P2sat |T1,fp P2,satfp
>1 (4)
>1
7583
dx.doi.org/10.1021/ie400585k | Ind. Eng. Chem. Res. 2013, 52, 7579−7585
Industrial & Engineering Chemistry Research
Article
binary minimum flash point solutions by addition of a third component.
Greek letters
αij = NRTL parameter ΔTfp = deviation of flash point γ = activity coefficient
4. CONCLUSIONS Ternary mixtures with a single or two minimum flash point binary constituent mixtures were investigated. The flash point variation trends for the two ternary mixtures with two constituent binary mixtures exhibiting minimum flash points (IPA + ethanol + octane and 2-butanol + ethanol + octane) were very similar, with substantial flash point variation around the binary mixtures (IPA + ethanol, 2-butanol + ethanol) and pure octane, and relatively moderate flash point variation over a wide range of compositions. For the ternary mixture (cyclohexanol + ethanol + octane) exhibiting a single minimum flash point binary mixture, the flash point increased with addition of the high-boiling point compound (cyclohexanol), and the MinFPB disappeared upon reaching a sufficient amount of cyclohexanol. We observed two different types of MinFPB useful for the classification of minimum flash points of ternary mixtures, which are helpful for hazard assessments of flammable solvent mixtures and fuel design. Nevertheless, more experimental data and exploration of different types of MinFPB for ternary mixtures are required for the classification of minimum flash point solutions for ternary mixtures.
■
Superscripts
∞ = infinite dilution Subscripts
■
REFERENCES
(1) Liaw, H.-J.; Lee, Y.-H.; Tang, C.-L.; Hsu, H.-H.; Liu, J.-H. A Mathematical Model for Predicting the Flash Point of Binary Solutions. J. Loss Prev. Process Ind. 2002, 15, 429−438. (2) Liaw, H.-J.; Chiu, Y.-Y. The Prediction of the Flash Point for Binary Aqueous-Organic Solutions. J. Hazard. Mater. 2003, 101, 83− 106. (3) Liaw, H.-J.; Chiu, Y.-Y. A General Model for Predicting the Flash Point of Miscible Mixture. J. Hazard. Mater. 2006, 137, 38−46. (4) Liaw, H.-J.; Gerbaud, V.; Chiu, C.-Y. Flash Point for Ternary Partially Miscible Mixtures of Flammable Solvents. J. Chem. Eng. Data 2010, 55, 134−146. (5) Liaw, H.-J.; Gerbaud, V.; Chen, C.-C.; Shu, C.-M. Effect of Stirring on the Safety of Flammable Liquid Mixtures. J. Hazard. Mater. 2010, 177, 1093−1101. (6) Crowl, D. A.; Louvar, J. F. Chemical Process Safety: Fundamentals with Applications, 2nd ed.; Prentice Hall PTR: NJ, 2002. (7) CCPS/AIChE, Guidelines for Engineering Design for Process Safety; American Institute of Chemical Engineers: New York, 1993. (8) Guo, Y.; Yang, F.; Xing, Y.; Li, D.; Fang, W.; Lin, R. Study on Volatility and Flash Point of the Pseudo Binary Mixtures of SunflowerBased Biodiesel + Methylcyclohexane. Fluid Phase Equilib. 2009, 276, 127−132. (9) Ellis, W. H. Solvent Flash Points − Expected and Unexpected. J. Coat. Technol. 1976, 48, 44−57. (10) Vidal, M.; Rogers, W. J.; Mannan, M. S. Prediction of Minimum Flash Point Behaviour for Binary Mixtures. Process Saf. Environ. Prot. 2006, 84, 1−9. (11) Liaw, H.-J.; Lee, T.-P.; Tsai, J.-S.; Hsiao, W.-H.; Chen, M.-H.; Hsu, T.-T. Binary Liquid Solutions Exhibiting Minimum Flash-Point Behavior. J. Loss Prev. Process Ind. 2003, 16, 173−186. (12) Liaw, H.-J.; Chen, C.-T.; Cheng, C.-C.; Yang, Y.-T. Elimination of Minimum Flash-Point Behavior by Addition of a Specified Third Component. J. Loss Prev. Process Ind. 2008, 21, 82−100. (13) Gmehling, J.; Rasmussen, P. Flash Points of Flammable Liquid Mixtures Using UNIFAC. Ind. Eng. Chem. Fundam. 1982, 21, 186− 188. (14) Ha, D.-M.; Lee, S.; Back, M.-H. Measurement and Estimation of the Lower Flash Points for the Flammable Binary Systems Using a Tag Open Cup Tester. Korean J. Chem. Eng. 2007, 24, 551−555. (15) Catoire, L.; Paulmier, S.; Naudet, V. Experimental Determination and Estimation of Closed Cup Flash Points of Mixtures of Flammable Solvents. Process Saf. Prog. 2006, 25, 33−39. (16) Liaw, H.-J.; Lin, S.-C. Binary Mixtures Exhibiting Maximum Flash-Point Behavior. J. Hazard. Mater. 2007, 140, 155−164. (17) Affens, W. A.; McLaren, G. W. Flammability Properties of Hydrocarbon Solutions in Air. J. Chem. Eng. Data 1972, 17, 482−488. (18) White, D.; Beyler, C. L.; Fulper, C.; Leonard, J. Flame Spread on Aviation Fuels. Fire Saf. J. 1997, 28, 1−31. (19) Garland, R. W.; Malcolm, M. O. Evaluating Vent Manifold Inerting Requirements: Flash Point Modeling for Organic Acid-Water Mixtures. Process Saf. Prog. 2002, 21, 254−260.
ASSOCIATED CONTENT
S Supporting Information *
Experimental flash-point data for the three ternary mixtures covering their entire composition range. This material is available free of charge via the Internet at http://pubs.acs.org.
■
exp = experimental data f p = flash point i = species i k, m, n = group k, m, n pred = predictive value
AUTHOR INFORMATION
Corresponding Author
*Phone: 886-7-6011000 ext. 7606. E-mail:
[email protected]. tw. Corresponding author address: Department of Safety, Health and Environmental Engineering, National Kaohsiung First University of Science and Technology, 2 Jhuoyue Road, Kaohsiung, Taiwan. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors thank the National Science Council of Taiwan ROC for supporting this study financially under grant #NSC 100-2221-E-039-006.
■
NOMENCLATURE A, B, C = Antoine coefficients Aij = binary parameter (K) g = binary parameters of the NRTL equation (J/mol) kl = nonflammable components N = number of experimental data Psat i = saturated vapor pressure (kPa) Psat i,fp = saturated vapor pressure of component, i, at flash point (kPa) qi = pure component area parameter R = gas constant (8.314 J/mol·K) ri = pure component volume parameter T = temperature (K) Ti,fp = flash point temperature of pure component, i (K) x = liquid-phase composition 7584
dx.doi.org/10.1021/ie400585k | Ind. Eng. Chem. Res. 2013, 52, 7579−7585
Industrial & Engineering Chemistry Research
Article
(20) Liaw, H.-J.; Tang, C.-L.; Lai, J.-S. A Model for Predicting the Flash Point of Ternary Flammable Solutions of Liquid. Combust. Flame 2004, 138, 308−319. (21) Liaw, H.-J.; Wang, T.-A. A Non-Ideal Model for Predicting the Effect of Dissolved Salt on the Flash Point of Solvent Mixtures. J. Hazard. Mater. 2007, 141, 193−201. (22) Lee, S.-J.; Ha, D.-M. The Lower Flash Points of Binary Systems Containing Non-Flammable Component. Korean J. Chem. Eng. 2003, 20, 799−802. (23) Catoire, L.; Paulmier, S.; Naudet, V. Estimation of Closed Cup Flash Points of Combustible Solvent Blends. J. Phys. Chem. Ref. Data 2006, 35, 9−14. (24) Li, D.; Fang, W.; Xing, Y.; Guo, Y.; Lin, R. Effects of Dimethyl or Diethyl Carbonate as an Additive on Volatility and Flash Point of an Aviation Fuel. J. Hazard. Mater. 2009, 161, 1193−1201. (25) Xing, Y.; Shao, D.; Fang, W.; Guo, Y.; Lin, R. Vapor Pressures and Flash Points for Binary Mixtures of Tricyclo [5.2.1.02.6] Decane and Dimethyl Carbonate. Fluid Phase Equilib. 2009, 284, 14−18. (26) Noorollahy, M.; Moghadam, A. Z.; Ghasrodashti, A. A. Calculation of Mixture Equilibrium Binary Interaction Parameters Using Closed Cup Flash Point Measurements. Chem. Eng. Res. Des. 2010, 88, 81−86. (27) Carareto, N. D. D.; Kimura, C. Y. C. S.; Oliveira, E. C.; Costa, M. C.; Meirelles, A. J. A. Flash Points of Mixtures Containing Ethyl Esters or Ethylic Biodiesel and Ethanol. Fuel 2012, 96, 319−326. (28) Liaw, H.-J.; Gerbaud, V.; Li, Y. H. Prediction of Miscible Mixtures Flash-Point from UNIFAC Group Contribution Methods. Fluid Phase Equilib. 2011, 300, 70−82. (29) Hiaki, T.; Taniguchi, A.; Tsuji, T.; Hongo, M. Isothermal Vapor−Liquid Equilibria of Octane with 1-Butanol, 2-Butanol, or 2Methyl-2-propanol. Fluid Phase Equilib. 1998, 144, 145−155. (30) ASTM D 56. Standard Test Method for Flash Point by Tag Closed Tester; American Society for Testing and Materials: West Conshohocken, PA, 2005. (31) Renon, H.; Prausnitz, J. M. Local Compositions in Thermodynamic Excess Functions for Liquid Mixtures. AIChE J. 1968, 14, 135−144. (32) Fredenslund, A.; Jones, R. L.; Prausnitz, J. M. Group− Contribution Estimation of Activity Coefficients in Nonideal Liquid Mixtures. AIChE J. 1975, 21, 1086−1099. (33) Fredenslund, A.; Gmehling, J.; Michelsen, M. L.; Rasmussen, P.; Prausnitz, J. M. Computerized Design of Multicomponent Distillation Columns Using the UNIFAC Group Contribution Method for Calculation of Activity Coefficients. Ind. Eng. Chem. Process Des. Dev. 1977, 16, 450−462. (34) Weidlich, U.; Gmehling, J. A Modified UNIFAC Model. 1. Prediction of VLE, hE, and γ∞. J. Ind. Eng. Chem. Res. 1987, 26, 1372− 1381. (35) Gmehling, J.; Li, J.; Schiller, M. A Modified UNIFAC Model. 2. Present Parameter Matrix and Results for Different Thermodynamic Properties. Ind. Eng. Chem. Res. 1993, 32, 178−193. (36) Gmehling, J.; Onken, U. Vapor-Liquid Equilibrium Data Collection, Vol. 1, Part 2a., DECHEMA: Frankfurt, Germany, 1977. (37) Gmehling, J.; Onken, U.; Arlt, W. Vapor-Liquid Equilibrium Data Collection; DECHEMA: Frankfurt, Germany, 1978; Vol. 1, Part 2b. (38) Hiaki, T.; Tatsuhana, K.; Tsuji, T.; Hongo, M. Isobaric Vapor− Liquid Equilibria for 2-Methoxy-2-methylpropane + Ethanol + Octane and Constituent Binary Systems at 101.3 kPa. J. Chem. Eng. Data 1999, 44, 323−327. (39) Gmehling, J.; Onken, U.; Arlt, W. Vapor-Liquid Equilibrium Data Collection; DECHEMA: Frankfurt, Germany, 1982; Vol. 1, Part 2c. (40) Boublik, T.; Fried, V.; Hala, E. The Vapor Pressures of Pure Substances, 2nd ed.; Elsevier: Amsterdam, Holand, 1984. (41) Balzhiser, R. E.; Samuels, M. R.; Eliassen, J. D. Chemical Engineering Termodynamics: the Study of Energy, Entropy, and Equilibrium; Prentice-Hall: NJ, 1972.
7585
dx.doi.org/10.1021/ie400585k | Ind. Eng. Chem. Res. 2013, 52, 7579−7585
