Flammability Limits: a Review with Emphasis on Ethanol for Aeronautical

Home , Flammability limit

Journal of Hazardous Materials 241–242 (2012) 32–54

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials

j ournal homepage: www.elsevier.com/locate/jhazmat

Review

Flammability limits: A review with emphasis on ethanol for aeronautical

applications and description of the experimental procedure

a b,∗ c c

Christian J.R. Coronado , João A. Carvalho Jr. , José C. Andrade , Ely V. Cortez ,

a c b

Felipe S. Carvalho , José C. Santos , Andrés Z. Mendiburu

Federal University of Itajubá – UNIFEI, Mechanical Engineering Institute – IEM Av BPS 1303, Itajubá, MG CEP 37500903, Brazil

São Paulo State University – UNESP, Campus of Guaratinguetá – FEG Av. Ariberto P. da Cunha 333, Guaratinguetá, SP CEP 12510410, Brazil

National Space Research Institute – INPE, Combustion and Propulsion Laboratory – LCP Rod. Pres. Dutra, km 39, Cachoeira Paulista, SP CEP 12630-000, Brazil

h i g h l i g h t s

Develops a comprehensive literature review on ethanol ﬂammability limits.

Difference in standard procedures lead to different experimental values of the ﬂammability limits.

Methodology for experiments to ﬁnd the FL’s of ethanol for aeronautical applications.

a r t i c l e i n f o a b s t r a c t

Article history: The lower and upper ﬂammability limits of a fuel are key tools for predicting ﬁre, assessing the possibility

Received 22 May 2012

of explosion, and designing protection systems. Knowledge about the risks involved with the explo-

Received in revised form 23 August 2012

sion of both gaseous and vaporized liquid fuel mixtures with air is very important to guarantee safety

Accepted 16 September 2012

in industrial, domestic, and aeronautical applications. Currently, most countries use various standard

Available online 24 September 2012

experimental tests, which lead to different experimental values for these limits. A comprehensive liter-

ature review of the ﬂammability limits of combustible mixtures is developed here in order to organize

Keywords:

the theoretical and practical knowledge of the subject. The main focus of this paper is the review of

Flammability limits

Ethanol the ﬂammability data of ethanol–air mixtures available in the literature. In addition, the description of

methodology for experiments to ﬁnd the upper and lower limits of ﬂammability of ethanol for aero-

Visual criterion

Pressure and temperature dependence nautical applications is discussed. A heated spherical 20 L vessel was used. The mixtures were ignited

with electrode rods placed in the center of the vessel, and the spark gap was 6.4 mm. LFL and the UFL

◦

were determined for ethanol (hydrated ethanol 96% INPM) as functions of temperature for atmospheric

Contents

1. Introduction ...... 33

1.1. Objectives and scope ...... 33

1.2. Flammability limits ...... 33

1.3. State of the art ...... 33

1.3.1. Theoretical methods to determine ﬂammability limits...... 35

2. Standard methodology for ﬂammability determination ...... 36

2.1. Visual criterion ...... 36

2.2. Pressure criterion ...... 37

2.3. Brief description of ﬂammability test methods discussed in this paper...... 38

3. Inﬂuences of temperature, pressure, turbulence and ignition energy on ﬂammability tests ...... 38

3.1. Temperature ...... 38

3.2. Pressure...... 39

∗

Corresponding author at: São Paulo State University – UNESP, Campus of Guaratinguetá – FEG, Av. Ariberto P. da Cunha 333, Guaratingueta, SP 12516-410, Brazil.

Tel.: +55 12 31232838.

E-mail addresses: [email protected], [email protected] (J.A. Carvalho Jr.).

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 33

3.3. Turbulence ...... 40

3.4. Comparison between the limits in air and in oxygen ...... 41

3.5. Ignition energy ...... 41

3.6. Flammability limits of fuel mixtures ...... 42

4. Flammability limits in the aeronautical industry...... 42

4.1. Importance for the aeronautical industry ...... 42

4.2. Ignition ...... 43

4.3. Flammability properties of aviation fuels ...... 44

5. Tests for ﬂammability limits of ethanol: review ...... 45

6. Experimental tests with ethanol ...... 47

6.1. Flammability apparatus ...... 47

6.2. Description of experimental procedure ...... 47

6.3. Mass of ethanol needed to form ﬂammable mixtures with air ...... 50

6.4. Description of the procedure to calculate LFL and UFL from experimental data ...... 50

6.5. Results and discussion...... 51

6.5.1. Behavior of ﬂame propagation ...... 51

6.5.2. Results for atmospheric pressure ...... 52

7. Conclusion...... 52

Acknowledgements ...... 52

References ...... 52

1. Introduction the following considerations: (a) visual inspection of the ﬂame

produced by a spark (spark ignition) in an oxidant-combustible

1.1. Objectives and scope mixture inside a transparent and closed vessel (visual criteria);

(b) pressure or temperature measurement at the moment of igni-

A comprehensive review of the literature about the ﬂammability tion inside a closed vessel (pressure criteria). In the 1950s and the

limits of combustible mixtures is developed here in order to orga- 1960s, the tests were conducted using glass tubes with diame-

nize theoretical and practical knowledge on the subject. The main ters of up to 60 mm and heights up to 300 mm, using the visual

focus of this paper is a review of ﬂammability data of ethanol–air criteria to verify ﬂammability of certain mixtures. Today, the visual

mixtures available in the literature. In addition, the description of criterion is still in use, as is the pressure criterion. Vessels used

an experimental method for determination of the upper and lower for the tests are generally spherical and/or cubic and of different

limits of ﬂammability of ethanol for aeronautical applications is dis- sizes.

cussed. It also shows a diagram of the experimental prototype built There are standards for setting up experimental procedures to

to obtain the ﬂammability limits of ethanol and the main results at determine whether a mixture is ﬂammable or not. In all these

atmospheric pressure. standards, the variables are the same: the size of the combustion

chamber (test vessel), temperature, operation pressure, electrode

discharge time (spark ignition or fuse wire ignition), and ignition

1.2. Flammability limits

energy of the electrodes. Table 1 shows the ﬂammability limits of

some fuels that were selected as part of the scope of this work.

Flammability limits are the main properties that represent

Among the fuels chosen are those used in the aircraft industry, such

ﬂammability characteristics of speciﬁc fuels. These limits are the

as gasoline for commercial and military aviation and kerosene, as

borders that separate the oxidant–combustible mixture regions

well as some vehicle fuels such as gasoline, diesel, natural gas, etc.

in which ﬂame propagation occurs and does not occur. There

It also shows the main alcohols and other alternative fuels of inter-

are two kinds of ﬂammability limits for an oxidant–combustible

est. Most of the data shown on this table were obtained from Refs.

mixture: the minimum concentration of fuel for which ﬂame prop-

[1–8], except those marked with letters. For readers interested in a

agation is possible (lean mixture), known as the lower ﬂammability

particular fuel, Appendix A of the paper published by Zabetakis [2]

limit (LFL), and the maximum concentration of fuel for which

and Table 44 on page 130 of the paper published by Cowards and

ﬂame propagation is possible (rich mixture), known as the upper

Jones [1] are recommended.

ﬂammability limit (UFL). There is another parameter known as LOC

(limiting oxidant concentration) that is widely used along with

the limits of ﬂammability. LOC is the concentration of oxidant in 1.3. State of the art

a fuel–oxidant–diluent mixture below which deﬂagration cannot

occur under specified conditions. This section will present the latest scientific studies of flamma-

Flammability limits have been thoroughly discussed in scientiﬁc bility limits. The following sections emphasize ﬂammability limit

literature. Probably the ﬁrst works on this topic were those devel- s of ethanol primarily for the aircraft industry.

oped by Coward and Jones [1] and by Zabetakis [2], both for the Naegeli and Weatherford [13] studied the ﬂammability haz-

Mines Department of the U.S. Government. Afterwards, Kuchta [3] ard of storing pure alcohol and diesel fuel/alcohol blends in fuel

expanded upon the data obtained by Zabetakis. Nestor [4] and the tanks. The authors measured the ignition limit for the C1 through

Fuel Flammability Task Group [5] worked speciﬁcally with ﬂamma- C4 alcohols and alkanes, ethylene, isooctane and methylal in a

bility limits for the aeronautical industry. Measurements of ignition combustion pump using an automotive-type spark plug as an igni-

energy for aeronautic fuels (Spark Ignition Energy Measurements tion source. However, the UFL of the alcohols were very different

in JET A) were reported by Shepherd et al. [6]. In addition, the from published ﬂammability limit data, though the upper limit for

works published by Ott [7] and Kosvic et al. [8] contribute for the methanol is in the area of reported values. According these authors,

knowledge in the area. methanol’s LFL is 7.9 mol% and UFL is 26 mol%. On the other hand,

Different methods are used to evaluate ﬂammability limits. The for ethanol LFL is 4.4 mol% and UFL is 14.3 mol%. All these data were

success of an attempt can be determined by the combination of for 0.97 atm and at a temperature of 364 K [13].

34 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

Table 1

Flammability limits for some fuels in air, at atmospheric pressure and 298 K, vol% [1–8].

Chemical name Chemical formula LFL UFL

1 Acetylene CH:CH 2.5 100.0

2 Ammonia NH3 15 28

3 Benzol (Benzene) C6H6 1.3 7.9

4 Blast Furnace Gas – 35 74

5 Butane CH3CH2CH2CH3 1.6 8.4

6 Butyl Alcohol CH3(CH2)2CH2OH 1.4 11.2

7 Carbon Monoxide CO 12.5 74

8 Coal Gas – 5.3 32

9 Coke-oven Gas – 4.4 34

10 Cyclepropane (CH2)3 2.4 10.4

11 Cyclohexane C6H12 1.3 8.0

12 Decane CH3(CH2)8CH3 0.8 5.4

13 Ethane CH3CH3 3.0 12.5

14 Ethyl Alcohol C2H5OH 3.3 19.0

15 Ethylene H2C:CH2 2.7 36.0

16 Ethylene Glycol C2H6O2 3.2 21.6

17 Diesel (gas oil) 0.5 5.0

18 Gasoline, Premium Automotive – 1.3–1.4 6.0–7.6

19 Gasoline, Regular Automotive – 1.3–1.4 6.0–7.6

20 Gasoline, Commercial Aviation – 1.0 6.0–7.6

21 Gasoline, Military Aviation 1.0 6.0–7.6

22 Heptane CH3(CH2)5CH3 1.1 6.7

23 Hexane CH3(CH2)4CH3 1.1 7.5

24 Hydrogen H2 4.0 74.2 a

25 LH2 – 4.0 75.0

26 Natural Gas – 3.8–6.5 13–17

27 Natural Gas 5 15.6

28 LNG and CNG – 5 15

29 Kerosene – 0.7 5

30 Methane CH4 5.0 15.0

31 Methyl Alcohol CH3OH 6.0 36.0

32 Methyl Ether (CH3)2O 3.4 27.0

33 Naphtha – 0.8 5

34 Nonane C9H20 0.8 2.9

35 Octane CH3(CH2)6CH3 1.0 6.5

36 p-Dioxane OCH2CH2OCH2CH2 2.0 22.0

37 Pentane CH3(CH2)3CH3 1.5 7.8

38 Propanal CH3CH2CHO 2.6 17.0

39 Propane (LPG) CH3CH2CH3 2.1 9.5

40 Propylene CH2:CHCH3 2.0 11.0

41 Propyl alcohol CH3CH2CH2OH 2.2 13.7

42 Syngas (wood gas) – 16 –

43 Toluene C6H5CH3 1.2 7.1

Ref. [9]. b

Ref. [10].

Ref. [11]. 96.16% CH4 by vol., 2.54% CO2, 1.096% C2H6, approximately 0.189% hydrocarbons higher than C3, and the remainder including nitrogen, hydrogen sulﬁde, and

water is approximately 0.015%. d

Ref. [12].

Shebeko et al. [14] proposed a new analytical method According to Brandes et al. [15], it is well known that explo-

for the calculation of ﬂammability limits in mixtures of sive mixtures can exist at temperatures below the ﬂash point (FP).

combustible–oxidizer–diluent. They revealed new regularities for Experiments show that the difference between ﬂash point and LFL

LFL and the compositions of mixtures near the peak point of the may be up to 15 K and in some special cases even more. There

ﬂammability curves. are some liquids without a ﬂash point (pure substances as well

Van den Schoor and Verplaetsen [9] studied the UFL as mixtures) that are able to form an explosive vapor–air mixture.

of ethane–air, propane–air, n-butane–air, ethylene–air, and According this author, knowledge about LFL lets work take place at

propylene–air mixtures. The UFL were determined experimen- higher temperatures than using ﬂash point and including a safety

tally at initial pressures up to 30 bar (this paper does not specify margin of for example 12 K or more. A good example for this case

the study at reduced pressures and negatives temperatures) and is cleaning processes using ﬂammable liquids.

◦

temperatures up to 250 C. The experiments were performed in Van den Schoor et al. [16] reported four different numerical

a closed spherical vessel with an internal diameter of 200 mm. methods to calculate the UFL of methane–air mixtures at initial

◦

The mixtures were ignited by using electric current to fuse a pressures up to 10 bar and initial temperatures up to 200 C by

coiled tungsten wire that was placed at the center of the vessel. comparison with experimental data. According this author: (i)

They concluded that UFL increases linearly with initial temper- at atmospheric pressure, calculation of UFL by calculating planar

ature; however, the slope of the straight line is not a constant ﬂames with the inclusion of radiation heat loss is satisfactory;

but depends on the initial pressure. Finally, a comparison of data (ii) at elevated pressures, calculated UFL values are signiﬁcantly

for the alkanes showed that the effect of preferential diffusion high and large differences are found between the different reac-

plays an important role in near-upper ﬂammability limit combustion mechanisms. Van den Schoor et al. [17] reported two different

tion. experimental methods to determine ﬂammability limits which

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 35

were compared, evaluated, and exempliﬁed by determining the A general rule for estimating ﬂammability limits is to consider

ﬂammability limits of methane–hydrogen–air mixtures for hydro- the upper ﬂammability limit equal to three times the stoichio-

gen fuel molar fractions of 0, 0.2, 0.4, and 0.6 at atmospheric metric value and the lower ﬂammability limit equal to 50% of the

pressure and room temperature. stoichiometric value [26]. However, as pointed out in the same ref-

Chen at al. [18] predicted the upper/lower ﬂammability lim- erence, there are many exceptions to the rule, as can be seen in

its of hydrocarbons diluted with inert nitrogen gas. These authors Table 1.

reported that there are linear relations between the reciprocal of Some theoretical methods in the literature are discussed in the

the upper/lower ﬂammability limits and the reciprocal of the molar following sections. Ethanol ﬂammability limits will be determined

fraction of hydrocarbon in the hydrocarbon/inert nitrogen mixture. for example and in order to contrast the theoretical and experi-

Also, Gharagheizi [19] studied a quantitative structure–property mental data. According to experimental data, ethanol’s LFL is 3.3%

relationship to predict the UFL of pure compounds. The obtained and UFL is 19% (vol%).

model is a ﬁve-parameter multilinear equation.

Koshiba et al. [20] characterized the explosion properties of mix- 1.3.1.1. Stoichiometric concentration. In order to determine

tures of n-pentane, diethyl ether, diethylamine, or n-butyraldehyde ﬂammability limits using stoichiometric concentration, the

with nitrous oxide and nitrogen using three parameters: explosion following formulas are used:

limit, peak explosion pressure, and time to peak explosion pres-

LFL = 0.55 Cs, [2, 27] (1)

sure. The explosion experiments were performed in a cylindrical

vessel at atmospheric pressure and room temperature. According

UFL = 3 Cs, [26] (2)

this author, measurements showed that explosion ranges of the

UFL = 3.5 Cs, [27, 28] (3)

mixtures containing nitrous oxide were narrower than those of the

mixtures containing oxygen.

in which Cs is stoichiometric concentration.

Shoshin and de Goey [21] reported an experimental study of the

For example, ethanol’s stoichiometric reactions are: 1

LFL of methane–hydrogen–air mixtures in tubes of different diame-

C2H5OH + 3 O2 + 11.28 N2 → 2 CO2 + 3 H2O + 11.28 N2. For this

ters (6.0–50.2 mm). The ﬂames propagated upward from the open

reaction, Cs = 6.54%. Using Eqs. (1) and (2), LFL = 3.59% and

bottom end of the tube to the closed upper end. According this

UFL = 19.62%.

author, LFL value decreased with tube diameter for methane–air

and (90% CH4 + 10% H2)–air mixtures for tubes. This effect has been

1.3.1.2. Combustion enthalpy. A rough estimation of LFL in air may

attributed to the stronger combined effect of preferential diffusion

be obtained using the following rule of thumb from Tareq [27]. The

and ﬂame stretch in narrower tubes for ﬂames which resemble ris-

equation to determine the lower ﬂammability limit is:

ing bubbles. Also, Miao et al. [22] reported both LFL and UFL of

4354

hydrogen-enriched natural gas with hydrogen concentrations of

LFL = − , [27] (4)

20%, 40%, 60% and 80% as well as these of natural gas and hydro- comb

gen. The experimental results showed that the ﬂammability data of

in which Hcomb is combustion enthalpy (kJ/mol) and LFL is in

methane–hydrogen mixtures are applicable to hydrogen-enriched

volume percent.

natural gas mixtures.

For example, ethanol has a combustion enthalpy of

Piqueras et al. [23] reported a method for the prediction of LFL at −

1281 kJ/mol [2]. Applying Eq. (4), LFL is 3.39%.

high pressures for the mixtures of H + O in CO and N , between

2 2 2 2 The two aforementioned methods are generally applied to

1.0 and 300 bar and 288–348 K. According to this author, the use

hydrocarbons and parafﬁns.

of CO2 as a diluent increases the operational margin from 4.5 mol%

H2 at 1 bar up to ca. 7.0–9.0 mol% H2 at 200 bar due to the increase

1.3.1.3. Vapor pressure. The lower ﬂammability limits can be deter-

in heat capacity. On the other hand, the use of nitrogen or air as a

mined using the fuel vapor pressure and standard atmospheric

diluent only increases the margin from 5.2 mol% H2 at 1 bar up to

pressure as follows:

6.0 mol% H2 at 200 bar.

Rowley et al. [24] reported an estimation of the LFL of organic LFL × 100, [23] (5)

compounds as a function of temperature. This author showed that

the LFL of general organic compounds may be accurately esti- in which pf is the fuel vapor pressure at ﬂash point and p0 is standard

mated from structural contributions and the heat of formation atmospheric pressure.

◦

of the fuel. Prediction of LFL and its dependence on temperature The ﬂash point of ethanol is 13 C [29]. At this temperature,

closely agrees with experimental data obtained in an ASHRAE- it has a vapor pressure of 28 mmHg [10]. With p0 = 760 mmHg,

style apparatus.Kondo et al. [25] reported the ﬂammability limits LFL = 3.68%.

of CH2 CFCF3 (HFO-1234 yf), CH2F2 (HFC-32), and methane mea-

sured at pressures from atmospheric pressure to 2500 kPa in a 5 L 1.3.1.4. Algebraic method. The following equation allows LFL to

spherical stainless-steel vessel. In general, the changes to lower be determined from the data of the minimum number of moles

flammability limits are small compared to upper flammability lim- required for a flame propagation reaction: its.

100

Below are some simple methods to calculate the ﬂammability LFL = , [29] (6)

1 + na

limits from stoichiometric concentration, enthalpy of combustion,

and/or the vapor pressure of the fuel. in which na is the number of moles of air per mole of fuel in the

mixture at LFL.

For example, for ethanol, the reactants are 1 C2H5OH + 3

1.3.1. Theoretical methods to determine ﬂammability limits O2 + 11.28 N2. LFL obtained is 3.29%.

The scientiﬁc literature provides different theoretical methods The combustion reaction can be described using LFL as fol-

− → ˛ ˛

to determine ﬂammability limits. In all these methods, the theoret- lows: LFL (CnHmOl) + (100 LFL) (0.21 O2 + 0.79 N2) 1 CO2 + 2

˛ ˛ ˛ × ˛ × ical values for lower ﬂammability limits ﬁt the experimental data H2O(v) + 3 O2 + 4 N2, in which 1 = LFL n, 2 = LFL m/2,

well; on the other hand, the theoretical values for upper ﬂamma- 3 = 21 − LFL (0.21 + n + m/2 − 1/2), and ˛4 = 79 − LFL × 0.79.

bility limits have some differences from experimental data. Therefore, for ethanol, the combustion reaction will be: 3.299

36 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

Table 2

Main characteristics of international standards for measurement of ﬂammability limits [32].

DIN 51649-1 EN 1839 (T) EN 1839 (B) ASTM E 681-01

Ignition vessel Cylindrical glass vessel Cylindrical glass vessel Spherical or cylindrical closed Glass vessel V = 5 L

∅ = 60 mm, H = 300 mm ∅ = 80 mm, H = 300 mm vessel V > 5 L

Ignition source High tension spark, duration High tension spark; duration Fusion wire (explosion), High tension spark, duration

0.5 s; power 10 W 0.2 s; power 10 W E = 10–20 J high tension spark; 0.5 s; power 10 W

duration 0.2 s

Criteria Visual, flame detachment Visual, flame detachment and Pressure ≥ 5% of test initial Visual, flame propagation up to

propagation >100 mm, or absolute pressure. 13 mm of the wall (horizontal

without ﬂame detachment (H) or vertical)

of H > 240 mm

Repetitions 5 4 4 1

Sample size 0.1 mol% (FL ≤ 10 mol%), 10% rel. (FL ≤ 2 mol%), 0.2 mol% 10% rel. (FL ≤ 2 mol%), 0.2 mol% Arbitrary

0.2 mol% (EL > 10 mol%) (FL >2 mol%) (FL > 2 mol%)

Flammability limit Last point without ignition Last point without ignition Last point without ignition Average values between the

ﬁrst ignition point and the last

point without ignition

FL, ﬂammability limits.

→

C2H5OH + 20.3 O2 + 76.39 N2 6.59 CO2 + 9.89 H2O(v) + 10.4 2.1. Visual criterion

O2 + 76.39 N2. This method is generally applied to organic

substances with the general formula CnHmOl [30]. Test vessels must be transparent; they are incapable of resisting

a total deﬂagration charge. A high-speed video camera is recom-

mended.

2. Standard methodology for ﬂammability determination

The German standard utilizes vertical cylindrical tubes with

smaller volumes than those established by the American standard.

Methods of measuring ﬂammability limits for gases have been

For example, DIN standard 51649 requires a cylinder diameter of

known well for a long time, and there have been many attempts

60 mm, with a volume of almost 0.85 L, while European standard

to standardize them. However, no standard method has been esti-

EN 1839 (T) requires a cylinder diameter of 80 mm, with a volume

mated yet [31]. Modern ﬂammability test methods to determine

of 1.5 L, and ﬁnally, American standard ASTM E681 [28] requires

ﬂammability limits and oxygen concentration (LOC) are summa-

spherical vessels with volumes of 5 L or 12 L, with internal diam-

rized in Fig. 1. Tables 2 and 3 show a comparison between standard

eters of 222 mm and 295 mm, respectively. European standard EN

test methods for ﬂammability limits at atmospheric pressure and

1839 offers two different visual criteria: 298 K.

Table 3

Comparison of ﬂammability data obtained using modern standards [32–35].

Substance LFL and UFL (vol%) isobar conditions LFL and UFL (vol%) isochor conditions

EN 1839 T ASTM E681 EN 1839B ASTM WK5917 and E2079

Ethylene

◦

LFL (20 C) 2.5 ± 0.10 2.2 2.5 ± 0.10

◦

UFL (20 C) 32.3 ± 0.40 33.3 30.8 ± 0.70

◦

LOC (20 C) 7.7 7.5 8.1 8.5 (120 L)

Methane

◦

LFL (20 C) 4.4 ± 0.20 3.8 4.6 ± 0.20 5.0 ± 0.1 [26] (8, 20 and 120 L test vessels)

◦

UFL (20 C) 16.6 ± 0.20 16.9 16.8 ± 0.20 15.9 ± 0.10 (20 L) [26]15.8 ± 0.10 (120 L) [26]

◦

LOC (20 C) 10.2 9.9 11.3 11.2 (120 L)

n-Hexane

◦

LFL (20 C) 1.0 ± 0.10 1.0 ± 0.10

◦

UFL (40 C) 8.3 ± 0.10 7.7 ± 0.25

1,3,5-Trimethyl-benzene

◦

LFL (70 C) 0.8 ± 0.08 0.8 ± 0.08

◦

UFL (130 C) 7.1 ± 0.20 6.6 ± 0.10

Hydrogen [24,27]

◦

LFL (20 C) 3.6 3.75 4.2

◦

UFL (20 C) 76.6 75.1 77.0

◦

LOC (20 C) 4.4 4.1 4.7

Ammonia [24,27]

◦ a

LFL (20 C) 14.3 13.3 14.2

◦ a

UFL (20 C) 32.5 32.9 39.4

Schroder and Molnarne [32].

Ural and Zalosh [34].

Fuss et al. [35].

Performed in a 5 L ﬂask instead of a 12 L ﬂask.

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 37

Fig. 1. Flammability test methods [22].

•

The ﬁrst visual criterion is “ﬂame detachment”, which requires the fact that results obtained in larger vessels are more reliable;

an upward movement of the detached ﬂame from the ignition therefore, they do not allow tests in small vessels. European

spark (around 10 cm apart). standard EN 1839 does not allow tests in vessels smaller than

•

The second visual criterion is that the ﬂame reach the top of the 5 L, while the American standard permits tests in vessels down to

cylinder with a minimum distance of 24 cm. 4 L.

A mixture is considered ﬂammable if both conditions are met. European standard EN 1839B permits two different kinds of

American standard ASTM E681 [28] gives two different visual igniters:

criteria as well:

• •

For the 5 L vessel, the ﬂame front must move upward and outward Induction spark: A series of sparks are inducted by electrodes in

from the ignition source to the vessel walls or at least 13 mm up alternate current using a high voltage transformer. The power of

the wall. the spark depends on the mixture for each test and also on the

•

For the 12 L vessel, flame propagation is defined as moving details of the air gap and the circuit configuration. This standard

upward and outward from the ignition source to the vessel walls. suggests a power of 10 W. Therefore, the total energy emitted by

◦

It forms an arc larger than 90 measured from the ignition point the spark is limited to 2 J per 0.2 s of average discharge time.

•

to the vessel walls. Fuse wire: This kind of igniter emits more energy. The energy

fed to this wire must be between 10 and 20 J. Only a part of this

Schroder and Molnarne [32] reported comparative measure- electric energy is available for ignition in the gas phase because

ments (Table 3) using European standard EN 1839T (vol = 1.5 L) and most of the energy is used to heat and to vaporize the fuse wire.

American standard ASTM E681 (vol = 5 L). The mixture was pre-

pared outside the test vessels for both cases; the air–fuel mixture The American standard permits a continuous electric arc and the

was the same for both experiments. The ignition source was also same fuse wire as those required by the European standard. It also

the same, 10 W for 0.2 s. allows coal sparks, discrete electric sparks (almost 10 J of energy

storage in the capacitor bank), and chemical igniters.

2.2. Pressure criterion The European standard uses the criterion of 5% above the ini-

tial ﬁxed pressure in the vessel to determine whether a fuel is

The methods that are used most often are European standard EN ﬂammable or not. The American standard uses a 7% criterion. How-

1839B and American standard ASTM E2019. They both recognize ever, according to Brandes and Ural [33], the following remark must

38 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

be considered: the criterion of 7% may not be appropriate for certain

mixtures. This is the case when vessel volume is small or ignition

energy is substantially larger than 10 J. It is a good idea to perform

an exploratory test in the vicinity of limit mixtures to evaluate the

validity of the selected pressure rise criterion. In fact, users are free

to choose any pressure rise criterion as long as it is backed up by

appropriate baseline tests [33].

The authors state that when collecting data from ﬂammabil-

ity tests, researchers must keep in mind that tests carried out in

small volumes or with high ignition energies tend to result in

wider ﬂammability intervals. Considering the fact that tests car-

ried out with small ignition energies or small vessels may not have

noticeable ﬂammability with certain mixtures, it is important to use

sufﬁcient ignition energy and vessel volumes that are large enough

for ignition.

Table 4 shows a comparison of ﬂammability limits obtained

experimentally using different standard methods (criteria) for the

most important combustible gases.

To sum up, tests carried out in small volumes with high ignition Fig. 2. Temperature effect on ﬂammability limits [1].

energies result in wider intervals of ﬂammability and thus more

conservative results at the expense of higher safety costs. Consid-

is used to establish ﬂammability. Draft WK5917 employs the same

ering the fact that tests carried out with small ignition energies

concept as E2079 and is available for the LFL and UFL determina-

or small vessel volumes might not have a noticeable ﬂammability

tions [33,38]. EN1839B. The ignition procedure is identical to the

for certain combustibles, it is important to use sufﬁcient ignition

previous method. It uses a tube apparatus with an inner closed

energy and large enough volumes. A common practice is to use

spherical vessel with a volume of 14 dm . The mixture is prepared

high-speed video cameras in order to analyze the ﬁlm from differ-

with partial pressure and the ignition source is a fusing (explod-

ent points of view in the same laboratory. According to Takahashi

ing) wire. The criterion for a successful ignition is a pressure rise

et al. [36], if the vessel is large enough, the experimental values

of at least 5% of the starting absolute pressure. The criterion for

of ﬂammability limits may approach the values that would be

ﬂammability is a 5% minimum pressure rise after ignition [31].

obtained in open space.

2.3. Brief description of ﬂammability test methods discussed in 3. Inﬂuences of temperature, pressure, turbulence and

this paper ignition energy on ﬂammability tests

US bureau of mines ﬂammability tubes. This method used a visual 3.1. Temperature

criterion. The test vessel is made of glass (50 mm diameter and

1000 mm height, minimum) with a pressure release at the bot- Heat released by a reaction mixture and heat transfer to the

tom. The criterion for successful ignition (go) is when the ﬂame surroundings are both important in determining the course of the

travels up from the tips of the electrodes to at lest half way combustion process. According to Jones [39], it is not possible to

through the tube [1,2,33]. EN1839T. This method used a visual cri- assign an ignition temperature for a particular mixture, since igni-

terion. It uses an 8 cm wide, 30 cm tall, open top glass cylinder tion depends not only on the reagents but also on heat exchanged

with a spark igniter at the bottom (0.2 s and 10 W). The crite- with the surroundings. However, this parameter is found in the

rion for ﬂammability is propagation of the ﬂame 10 cm vertically combustion literature [40].

above the igniter or 12 cm horizontally at any point on the ﬂame’s A low ignition temperature means that the mixture is poten-

path. [31]. ASTM E681. This test method covers the determina- tially hazardous. The higher the activation energy of the reaction,

tion of the lower and upper concentration ﬂammability limits for the higher the ignition temperature. For mixtures with a high

chemicals with sufﬁcient vapor pressure to form ﬂammable mix- activation energy, a spark is needed to initiate combustion. After

tures in air at atmospheric pressure at the test temperature. This ignition by spark, the heat generated is usually sufﬁcient for the

test method is based on electrical ignition and visual observations reaction to be self sustainable [40].

of ﬂame propagation [28]. ASTM E918. This method is designed If the temperature increases, the lower ﬂammability lim-

to determine ﬂammability limits at elevated temperatures (up its should decrease, while the upper ﬂammability limits should

◦

to 200 C) and initial pressure of up to 200 psia (1.38 MPa). This increase. In other words, ﬂammability interval widths increase

practice is limited to mixtures that have explosion pressures less when initial mixture temperature increases. This can be seen in

than 2000 psia (13.79 MPa). The pressure resistant metal test ves- Fig. 2.

sel is a vertical cylinder, only 1 L in size, which is much smaller As has been discussed, LFL values decrease when temperature

than that used in E681. Flame propagation is deﬁned as a com- increases; consequently, less combustion energy will be necessary

bustion reaction that produces at least a 7% rise of the initial to ignite the ﬂame. Drysdale [41] stated that LFLT and UFLT for any

absolute pressure [33,37]. ASTM E2079 and draft standard WK5917. temperature T at 1 atm pressure can be calculated based on the

These test methods cover the determination of the limiting oxy- limits determined for 298 K using the following equation:

gen (oxidant) concentration of mixtures of oxygen (oxidant) and

LFLT LFL298[1 − 0.00078(T − 298)], [2] (7)

inert gases with ﬂammable gases and vapors at a speciﬁed initial

pressure and initial temperature. They can also be used to deter-

UFLT = UFL298[1 + 0.000721(T − 298)]. [41] (8)

mine the limiting concentration of oxidizers other than oxygen.

This method has been designed to address problems associated The ﬂammability limits of ethanol at 298 K and 1 atm are 3.3% for

with the E681 and E918 methods. The test vessel required is LFL and 19% for UFL [1,2], both in volume concentration, so Fig. 3

larger than 4 L and nearly spherical. A 7% pressure rise criterion can be made from Eqs. (7) and (8) for different temperatures. Later

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 39

Table 4

◦

Flammability limits of hydrogen, ethylene, methane and ammonia in air, T = 20 C [32].

DIN 51649-1 (mol%) EN 1839 (T) (mol%) EN 1839 (B) (mol%) ASTM E681-01 (mol%)

LFL (H2–air) 3.8 3.6 4.2 3.75

LFL (H2–air) 75.8 76.6 77.0 75.1

LFL (H2–40% N2–air) 3.6 3.6 4.4 3.65

UFL (H2–40% N2–air) 38.2 38.4 38.2 37.3

LFL (C2H4–air) 2.3 2.4 2.6 2.15

UFL (C2H4–air) 33.0 32.6 27.4 33.3

LFL (C2H4–40% N2–air) 2.4 2.4 2.6 2.85

UFL (C2H4–40% N2–air) 8.4 8.2 6.9 8.05

LFL (CH4–air) 4.2 4.3 4.9 3.8

UFL (CH4–air) 16.6 16.8 16.9 16.9

LFL (CH4–40% N2–air) 4.3 4.5 5.1 4.15

UFL (CH4–40% N2–air) 6.5 6.4 5.7 6.35

LFL (NH3–air) 14.3 14.3 14.2 13.3

UFL (NH3–air) 31.7 32.5 39.4 32.9

LFL (NH3–20% N2–air) 15.4 15.2 16.2 14.1

UFL (NH3–20% N2–air) 19.8 20.4 21.3 20.9

this same ﬁgure will be compared with experimental data of this In airplanes, the temperature of the fuel and vapor depends on

work for ethanol. the rate of heat transfer to the fuel from the surroundings. Fuel

Kondo et al. [42] reported that ﬂammability limits of gases are temperature is not normally measured in ﬂight but can be esti-

◦

dependent on temperature at temperatures from 5 to 100 C in a mated from a heat transfer model or measured in a special ﬂight

12 L spherical ﬂask basically following ASHRAE method. The mea- test. A heat transfer model must include heat transfer from the

◦

surements were done for methane, propane, isobutane, ethylene, heated cabin (nominally at 21 C), heat transfer from the air packs,

propylene, dimethyl ether, methyl formate, 1,1-diﬂuoroethane, heat losses through the rear spar to the ambient atmosphere in the

ammonia, and carbon monoxide. As a result, the linear tempera- wheel well, and heat transfer through the side-of-body ribs from

ture dependence of LFL was found to be adequately predicted using the adjacent fuel in the wing. The amount of fuel in the tank, the rate

a limiting ﬂame temperature concept that is constant independent at which fuel is removed, and the rate of climb or descent are addi-

of the experimental temperature. In addition, for compounds like tional factors that must be considered in addition to the properties

methane, propane, isobutane, propylene, methyl formate, and 1,1- of the particular batch of fuel [43].

diﬂuoroethane, a limiting ﬂame temperature concept along with

an assumption that geometric mean “G” [37] is independent of

3.2. Pressure

temperature enables one to predict the temperature dependence

of UFL reasonably well, at least in the temperature range exam-

When compared to the effect of fuel concentration, atmospheric

ined. However, this author mentions that the UFL of ethylene,

pressure variations do not have a considerable effect on ﬂamma-

dimethyl ether, and carbon monoxide deviate from this predicted

bility limits. The effect of pressure variation is neither simple nor

temperature dependence. Also, ethylene and dimethyl ether have

uniform, but it is speciﬁc for each mixture. The variation is generally

special burning characteristics at high concentrations. In addition,

unnoticeable for the ﬁrst few hundred mmHg, which decreases due

the ﬂammable range of CO may be too wide to apply the present

to atmospheric pressure; the effect increases until the ﬂame cannot

model to UFL [42].

propagate at low pressure [1].

As the temperature rises, lower ﬂammability limits are gradu-

Increases above atmospheric pressure do not always increase

ally shifted down and upper limits are shifted up.

the interval width of ﬂammability limits. For some mixtures, the

Fig. 3. Temperature effect on ﬂammability limits of ethanol at 1 atm [2,36].

40 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

Fig. 4. Effect of pressure and concentration on ﬂammability limits at 298 K for avi-

Fig. 6. Effect of pressure and temperature on ﬂammability limits for saturated avi-

ation gasoline [1].

ation gasoline vapor–air mixtures [1].

about 0.25 psi) when cruising since the vents are located in a low

interval decreases as pressure increases. So a mixture that has ﬂame

pressure region on the wing tips [43].

propagation at atmospheric pressure may not have it at higher

pressures. From tests conducted at the FAA William J. Hughes Technical

Center, Atlantic City International Airport, Atlantic City, NJ, it was

A moderate pressure increase above atmospheric pressure

determined that at sea level through 10 kft, LOC is approximately

decreases the interval width of ﬂammability intervals for

12% O2, with a linear increase from 12% at 10 kft to approximately

hydrogen–air or carbon monoxide–air mixtures. In the same con-

14.5% at 40 kft. There was little variation in tests with various sparks

ditions, the ﬂammability interval width for parafﬁn gases mixed

and arcs as ignition sources at sea level, with LOC ranging from

with air decreases on the LFL side and increases on the UFL side [1].

12.0% to 12.8%. Also, a heated surface capable of igniting a fuel air

Fig. 4 shows the effect of pressure and concentration on ﬂamma-

mixture proved insufﬁcient for ignition in a tank inerted to just

bility limits at 298 K for aviation gasoline. Fig. 5 presents the effect

14%. Peak pressures resulting from ignition at oxygen concentra-

of pressure and temperature on ﬂammability limits for saturated

tions 1–1.5% above LOC values decreased as the altitude increased

aviation gasoline vapor–air mixtures [1].

to 30 kft, while time to reach peak pressure increased. According

LFL has a small decrease as pressure rises, but UFL has a sig-

to Summer [46], further experiments to examine the trend of peak

niﬁcant increase [2]. Below atmospheric pressure, as pressure

pressure rise as a function of both altitude and oxygen concentra-

decreases, the values of the two limits approach each other and the

tion are needed.

gap between them is narrowed until a certain pressure is reached

According to Nestor [4], for the LFL region, the color of the ﬂame

(for methane, 125 mmHg at 293 K) [1,44]; at low pressures, the

is blue at an altitude between 0 and 20 kft, the ﬂame is blue-green

ﬂame cannot propagate through the mixture; according to Stull

at an altitude between 20 and 40 kft, and the ﬂame is green at an

[45], this is due to the fact that the gas (fuel) concentration is too

altitude between 40 and 60 kft. For the UFL region, the ﬂame is yel-

low to sustain combustion.

low at an altitude between 0 and 20 kft, yellow-green at an altitude

Fuel in an airplane is exposed to a range of conditions during

between 20 and 40 kft, and green at an altitude between 40 and

ﬂight. The pressure and temperature in the atmosphere are a func-

60 kft.

tion of altitude (see Fig. 6). The pressure within the fuel tank is

slightly lower than ambient pressure (the difference in pressure is

3.3. Turbulence

Few observations have been made in relation to the effect of tur-

bulence on ﬂammability limits. However, it has been shown that

the LFL for methane and ethane in air are slightly reduced due to

turbulence generated by a fan or a mixer [1]. Additionally, ﬂamma-

bility limit interval widths are slightly increased due to an increase

in mixture ﬂow rate.

Turbulence has an inﬂuence on mixtures of methane and air.

When a small fan was set at the right speed (fast enough but not

too fast) in a methane–air mixture contained in a 4 L globe, the

LFL of methane was 5.0% compared with 5.6% observed for a still

mixture in the same vessel. A streaming movement of the gas

mixture produces similar effects on LFL. At a speed of 35–65 cm/s,

ﬂame was propagated in a 5.02% methane–air mixture but not at

Fig. 5. Temperature and pressure as functions of altitude, standard atmosphere [43]. any speed in a 5.00% mixture. The same ﬁgure was obtained when

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 41

Table 5

Flammability limits in air and in oxygen [1,32].

Fuel LFL (%) UFL (%)

Air O2 Air O2

Hydrogen 4 4 74 94

Carbon monoxide 12 16 74 94

Ammonia 15 15 28 79

Methane 5 5 15 61

Propane 2 2 10 55

Ethane 3 3 12.5 66

Isobutene 1.8 1.8 8.4 48

Ethylene 2.7 2.9 36 80

Propylene 2 2 11 53

Coal gas 5.3 5.3 32 70

Fig. 7. Minimum ignition energy [41].

movement of the mixture was produced by expansion caused by

its own combustion in experiments on the propagation of ﬂame ASTM E2079-07 standard test methods cover the determina-

from the closed end to the open end of a large vessel [1]. tion of the limiting oxygen (oxidant) concentration of mixtures of

For a mixture of ethane in air, in a 4 L globe, the LF for a still oxygen (oxidant) and inert gases with ﬂammable gases and vapor

mixture was 3.10% ethane. With a fan running at high speed inside at a speciﬁed initial pressure and initial temperature. These test

the vessel, a 3.2% mixture did not ignite, but when the fan was run methods may also be used to determine the limiting concentra-

at a moderate speed, a 3.0% mixture exploded, producing a pressure tion of oxidizers other than oxygen. These test method are not

of 4.3 atm [1]. applicable to mixtures that undergo spontaneous reaction before

Coward and Jones [1] reported that the LFL of methane, ethane ignition is attempted, nor to mixtures that have maximum deﬂa-

and natural gas air mixtures can be reduced by about 12% by a gration pressures below the maximum operating pressure of the

suitable amount of turbulence produced either by a fan or by the test apparatus [51].

movement of the mixture itself. The effect disappears at higher Zlochower and Green [52] reported data on limiting (minimum)

velocities, at which no ﬂame is observed [47]. Sokolik et al. [48] oxygen concentration (LOC) with N2 added for methane, propane,

showed that for each mixture there is a certain maximum turbu- ethylene, carbon monoxide, and hydrogen and some of their binary

lence level for turbulent ﬂame front propagation velocity. A further mixtures. They also address the issue of the ﬂammability limits

increase in turbulence causes a drop in ﬂame velocity and ﬁnally of these pure gases in air. Their results had excellent agreement

the ﬂame goes out. Karpov and Severin [49] observed the same between the 120 L and 12 L results, good agreement with the 20 L

effect for mass burning rate. Lockwood and Megahed [50], in a results, and reasonable agreement with the earlier ﬂammability

study of ﬂame lift above small burners, concluded that ﬂames tube values for LFL. However, they differ from the more conserva-

can be quenched by turbulence when the molecular diffusion rate tive European values. Several studies give examples of LOC data

between the microscale turbulent vortices exceeds the chemical applications [53–56]. Much of the previous literature LOC data

kinetic rate. However, Chomiak and Jarosinski [47] raised some [1–3] were measured in a ﬂammability tube.

objections to this.

3.5. Ignition energy

3.4. Comparison between the limits in air and in oxygen

Explosion hazards are quantiﬁable in terms of the product of

Knowledge about LOC is needed to make some chemical and the probability of existence of a ﬂammable mixture and the sepa-

industrial processes safe. This information may be needed in order rate probability of the presence of an adequate ignition source [57].

to start up or operate a reactor while avoiding the creation of LFL and UFL are thus boundaries deﬁning the extent of a compo-

ﬂammable gas compositions therein or to store or ship materials sition’s domain of explosibility at room temperature and ambient

safely. The NFPA (National Fire Protection Association) has guide- pressure within which explosion is possible in the presence of an

lines for the practical use of LOC data, including the appropriate adequate ignition source, but beyond which explosion is impossible

safety margin to use [51]. regardless of the strength of the ignition source [57].

In general, the lower limit is almost the same in oxygen and in air A potential ignition source in an accidental ﬁre and explosions

for a gas. Until 1960, very few comparative results were available is an electric shock or spark. Other possible sources are hot sur-

to show whether there really was a small difference between the faces or adiabatic compressions of the air–combustible mixture. It

limits in oxygen and air. Nowadays, it is known that the lower limit was found that for a combustible–oxidant (air) mixture, there is a

is practically the same in oxygen and in the air. On the other hand, minimum energy required to cause an ignition when the mixture

the upper limit is much higher in oxygen than in air. This happens is within the ﬂammability interval. This energy is called ignition

because the lower limit is in the oxidant excess zone, and switching energy.

nitrogen for oxygen does not affect the important parameters as Ignition energy depends on the mixture’s composition. The

does ﬂame temperature [40]. minimum ignition energy for hydrocarbons in air was measured

Using central ignition in a globe vessel or with downward prop- for several common substances. Most of the results were around

agation of ﬂame in a tube, the LFL of methane is rather higher in 0.2 mJ, usually for a rich mixture. Some examples of minimum igni-

oxygen than in air, presumably because the mean molecular heat of tion energies are shown in Fig. 7. Based on experimental results,

oxygen is higher than that of nitrogen between room temperature the minimum ignition energy for petroleum derived fuels including

◦

and that of a ﬂame limit, for example 1200–1400 C. For upward or aviation kerosene is expected to be close to 0.2 mJ [6].

horizontal propagation the LFL of methane is slightly lower in oxy- Flammability limits depend on the ignition source. The most

gen than in air. UFL of ammonia, methane, and propane is much common sources are sparks, generally used with the visual crite-

greater in oxygen than in air [1]. Table 5 shows results for ﬂamma- rion, and fuse wires, generally used with the pressure criterion. The

bility limits in air and in oxygen. spark systems are capacitance sparks, which are developed from

42 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

discharged condensers. The duration of these discharges can be as results for mixtures that contain H2 and unsaturated hydrocarbons;

short as 0.01 ␮s or as long as 100 ␮s for larger engines [26]. it is also valid only if the components are chemically similar.

Energy and its distribution in both time and space have an effect Using Le Chatelier’s law, Carvalho and McQuay [40] showed that

on the values obtained for ﬂammability limits. This effect is particu- if there are two different hydrocarbons in individual mixtures with

larly true for difﬁcult to ignite materials that are not very energetic air that are both in the lower ﬂammability limit, when they are

when burning. What is more, if too high an energy ignition source added together, the resulting mixture will also be in the lower

is used, all that can be seen is dissipation of the ignition energy and ﬂammability limit. Similar reasoning is valid for the upper lim-

not the propagation of a ﬂame. The spark is an optimal and accept- its. For the demonstration, consider: (a) a mixture of a moles of

able ignition source [28]. An electric spark was used as an igniter hydrocarbon A and x moles of air, and (b) a mixture of b moles of

mechanism in the experiments in this study. hydrocarbon B and y moles of air, both in the lower ﬂammability

As already mentioned in the previous section, tests conducted limit. Then:

in smaller volumes using larger ignition energies tend to result in 100a

LFL = (%), (11)

wider ﬂammability range and therefore they are more conservative A

a + x

could lead to potentially higher prevention and protection costs.

100b

Considering the fact that low ignition energy or small test volume

LFLB = (%), (12)

b + y

can mask the ﬂammability of certain fuels, the prudent practice is to

rely on ﬂammability data obtained using sufﬁciently large ignition

Using Le Chatelier’s law:

energies in sufﬁciently large test enclosures [33].

100

Minimum ignition energy is a function of electrode spacing. It

LFLA+B = , (13)

C / + C /

becomes asymptotic to a very small spacing below which no igni- ( A LIA) ( B LIB)

tion is possible. Minimum ignition energy decreases as electrode

in which

spacing increases, reaches its lowest value at a certain spacing,

100a

then begins to rise again. For small spacings, the electrode removes C = (%), (14)

A a + b

large amounts of heat from the incipient ﬂame and thus large min-

imum ignition energy is required. As the spacing increases, the 100b

CB = (%). (15)

surface area to volume ratio decreases, and consequently the igni-

a + b

tion energy required decreases [26].

Thus:

Many papers have been published about ignition [58–63].

It should be noted that experiments on ignition energies car-

100

ried out by Pasamehmetoglu and Unal [64] in 11.25 L and 400 L LFLA+B =

(100a/(a + b))((a + x)/100a) + (100b/(a + b))((b + y)/100b)

combustion vessels at initial pressures of 100 kPa and tempera-

100 100(a + b)

tures of 295 K. Ignition energy bounds of methane–nitrous oxide, =

= , (16)

a + b + x + y

((a + x + b + y)/(a + b)) ( ) ( )

ammonia–nitrous oxide, and ammonia–nitrous oxide–nitrogen

mixtures have been determined. Lower and upper ﬂammability

which is the concentration of (a + b) moles of fuel in a mixture with

limits (mixing fan on, turbulent conditions) for ignition energies of

(x + y) moles of air, indicating that the new mixture will be exactly

8 J were: H2–N2O: 4.5–5.0% H2 (LFL), 76–80% H2 (UFL); CH4–N2O:

in the lower ﬂammability limit.

2.5–3.0% CH4 (LFL), 43–50% CH4 (UFL); NH3–N2O: 5.0–5.2% NH3

(LFL), 67.5–68% NH3 (UFL).

4. Flammability limits in the aeronautical industry

3.6. Flammability limits of fuel mixtures

4.1. Importance for the aeronautical industry

LFL and UFL for fuel mixtures can be determined by the Le Chate-

Since the introduction of kerosene as a fuel for civilian air trans-

lier’s law [3]:

port in the 1950s, aeronautical designers have been careful about

100

the ullage in aircraft fuel tanks because they may contain a mixture = ,

LFLmixture (9)

(C /LFL ) + (C /LFL ) + . . . + (Ci/LFLi)

1 1 2 2 of various vapors, fuel, and air which can ignite in the presence of

100 an ignition source when this mixture is within the ﬂammability

= ,

UFLmixture (10) limits of the fuel in use. After the ﬂight TWA 800 accident, which

(C1/UFL1) + (C2/UFL2) + . . . + (Ci/UFLi)

fell near the coast of New York in 1996, concerns about formation

in which C1,C2,. . .,Ci (%, volumetric basis) are the proportions of of vapors in aircraft fuel tanks reached a peak. It was determined

each fuel gas in the fuel mixture. According to Kondo et al. [65], the that the crucial reason for the lost plane was an explosion inside

Le Chatelier’s law accurately predicts LFL with for a wide variety of the Fuel-CWT tank (airplane wing). One signiﬁcant factor for TWA

combustible mixtures. 800 was the very small mass of fuel in the CWT. At low mass-to-

The ﬂammability limits of mixtures with propane were well volume ratios, Jet A vapor pressure decreases due to depletion of the

explained by Le Chatelier’s original formula [66]. Likewise, Kondo more volatile components in the fuel [43]. At the moment of explo-

et al. [67] measured a carbon dioxide dilution effect on ﬂammability sion, the plane had approximately 50 gal of Jet A fuel. The fuel and

limits for various ﬂammable gases. The values obtained were ana- air contained inside the tank were heated by the air conditioning

lyzed using Le Chatelier’s extended formula developed in a previous system (air packs) located directly below the CWT. This increased

study. fuel vaporization in the ullage region, forming a ﬂammable mix-

On the other hand, for most cases, experimental values for UFL ture which later caused the accident. Since then, ﬂammability limit

are lower than those predicted using the law. Von Niepenberg studies for the aeronautical industry have been increasingly taken

et al. [68] used Le Chatelier’s law to predict ﬂammability limits into account in aircraft design [5,6,43].

for mixtures that contained inert gases. Hustad and Sonju [69] had Required conditions for ignition and ﬂame propagation inside

experimental results that were very near those predicted by Le the aircraft fuel tank depend on various parameters that include

Chatelier’s law for fuel mixtures at high temperature and pressures. fuel type, temperature, tank pressure, and oxygen concentration.

According to Glassman [26], the law does not produce accurate It was also noted that the foam and spray of fuel, which may

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 43

4.2. Ignition

There is little data regarding ignition energy for aeronautical

fuels, but minimum ignition energy of 0.2 mJ for hydrocarbons is

suggested in CRC [70] as being suitable for aeronautical fuels. Exper-

iments by Nestor [4] and Ott [7] used sparks with energy ranging

between 4 and 20 J. These values are applied to create standard

ﬁgures involving ﬂammability limits in most aeronautical fuel man-

uals [36]. Results achieved by Nestor [4] for ﬂammability limits

with ignition energy of 20 J are shown in Fig. 8. In this ﬁgure, the

“safe” regions are those in which fuel ignition will not occur and

the medium region indicates ﬂammability. The vertical line indi-

cates the altitude in thousands of feet (e.g., the cruising altitude

in commercial aviation is typically 35,000 ft). Results from other

researchers are shown in Fig. 9. In this ﬁgure, the data are plotted

in the form of tank altitude versus fuel temperature for speciﬁed

spark ignition energy. Table 6 shows a summary of Jet A fuel ignition

tests applied since the 1960s.

In Table 6, Nestor [4] conducted a test in a 4-ft-long vertical

transparent tube with an ignition source at the bottom. Successful

ignitions at various fuel temperatures were observed as upward

ﬂame propagation to the top of the tube [5]. The ignition source

in Ott’s test was a moving arc that traveled up and down a pair

of 1-ft-long electrodes. Ott [7] used the pressure generated during

combustion as a measure of fuel ﬂammability for both static fuel

and sloshing fuel tests. The 1992 Air Force [71] tests were similar to

Nestor’s [4] ﬂame tube static ignition tests, except that the 1992 Air

Force tests involved higher ﬂash point samples and a broader range

of spark energies [5]. Shepherd et al. [6] conducted Jet A ignitability

tests designed to determine the effect of fuel load and spark ignition

energy at various fuel temperatures.

Fig. 8. Flammability limits for Jet A fuel in air [70] based on data by Nestor [4] with

ignition energy of 20 J. According to Shepherd et al. [6], the key factor in determin-

ing ignition energy is the composition and concentration of fuel

vapor. The concentration of fuel vapor in equilibrium with a liq-

uid fuel is proportional to fuel vapor pressure P and inversely

form during refueling or during a ﬂight with fuel oxygen libera- ,fuel

proportional to ambient pressure Pa:

tion, might increase fuel ﬂammability limits. In this sense, when

working with ﬂammability limits of fuels for the aeronautical P T

X ( fuel)

= .

industry, the following details must be taken into account: altitude fuel (17)

changes, temperature changes, ullage changes, tank ventilation,

state changes of fuel (spray) by agitation due to aircraft movement,

The vapor pressure of fuel in turn [43] depends ﬁrst on liquid

and fuel mixtures during refueling operations.

fuel temperature Tfuel and second on the amount of liquid and the

Fig. 9. Jet A ﬂammability data comparison [6].

44 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

Table 6

Ignition tests for Jet A fuel [6].

◦ 3

Author Flash point ( F) Vessel size (L) Fuel concentration (kg/m ) Ignition source Vessel pressure (atm)

Nestor [4] 118, 120, 125 Tubular (9.9) 100 Multiple sparks (5 and 20 J) 1 and 0.58

Nestor [4] Tank (9.5) 100 Spark (16 and 24 J) 1

Ott [7] 188, 123 Tank (303) 100 Electric arc 1 and 0.68

Air Force [5] 130, 140 Tubular (12.4) 100 Spark (3 mJ to 16 J) 1 and 0.69

Shepherd et al. [6] 118 Cylindrical (2) 3 and 200 Spark (1 mJ to 100 J)

Fig. 10. Measured ignition energy for Jet A vapor–air mixtures, 200 kg/m mass load,

and an initial pressure 0.585 bar [6].

Fig. 11. Measured ignition energy for Jet A vapor–air mixtures, 3 kg/m mass load,

and initial pressure 0.585 bar [6].

history of handling. To examine temperature dependence, Shep-

herd et al. [6] studied dependence on fuel load by testing two

situations, one with 200 kg of fuel per m of tank fuel correspond- [3] for mist (spray) of Jet A sprayed directly onto electrodes. Fig. 14

ing to a 25% full tank and the other with a nearly empty tank with shows the relation between Jet A ignition energy and altitude.

3 kg/m corresponding to a thin layer of residual fuel. The tests had

a tank pressure of 0.585 bar, which corresponds to an altitude of 4.3. Flammability properties of aviation fuels

about 14,000 ft, at which the TWA Flight 800 explosion occurred.

In the ﬁrst case, Fig. 10, the experiments produced the following There are at least 23 speciﬁcations in ASTM D1655 for Jet A or

◦

results in the 1.8-L vessel (about 450 mL of liquid fuel): below 30 C, Jet A-1 fuels, which set maximum or minimum limits for stated

the ullage mixture cannot be ignited even with spark energies of up properties or measurements. Compare this to nine speciﬁcations

to 100 J. Ignition energy decreases rapidly as temperature increases. in the ASTM Diesel fuel requirements or six speciﬁcations in the

◦

Above 50 C, the mixture can be ignited with spark energies on automotive gasoline requirements [72]. According to ASTM D1655,

the order of 1–10 mJ. In the second case, Fig. 11, the experiments there are three grades of aviation turbine fuel: the kerosene-based

◦

produced the following results in the 1.8-L vessel (about 7 mL of Jet A and Jet A-1 (approximate boiling point range of 160–300 C)

liquid fuel): dependence on temperature was very similar to the and wide-cut or naphtha-based Jet B (approximate boiling range of

3 ◦

200 kg/m case. The expected mass-load effect is not obvious [6]. 50–300 C). The difference between Jet A and Jet A-1 is that freezing

◦ ◦

Fig. 12 shows the minimum ignition energy for Jet A and Jet B point of Jet A-1 is −47 C and Jet A is −40 C. Jet B is little used

fuel. Fig. 13 shows the values of ignition energy measured by Kuchta today except in some Arctic operations, although it is essentially

Fig. 12. Minimum electric ignition energy for aviation fuels Jet A and Jet B [5].

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 45

Table 7

Summary of ﬂammability limits for ethanol–air mixtures at 298 K and 1 atm [1].

Tube dimensions (cm) Flammability limits (vol%)

Diameter Length LFL UFL

Upward ﬂame propagation

30.6 39 (steel) 3.48 –

15 300 (steel) 4.16 –

10.2 96 (steel) 3.28 –

7.5 150 (glass) 3.56 –

5 150 (glass) 4.24 18.95

5 130 (glass) 3.69 18.00

5 91 (glass) 4.30 –

2.5 150 (glass) 5.02 –

2.5 25 (glass) 5.0 14

Fig. 13. Dependency of ignition energy on fuel temperature for Jet A sprays [3].

Outward ﬂame propagation

15 300 (steel) 4.23 –

7.5 150 (glass) 3.70 –

the same fuel as JP-4, the standard US military fuel in use until

5 150 (glass) 4.32 13.80

about 10 years ago. The current US military aviation fuels are JP-8, 591 (glass) 4.40 –

which is nearly equivalent to Jet A, and JP-5, which has a high ﬂash 2.5 150 (glass) 5.18 –

point speciﬁcation [5].

Downwards ﬂame propagation

Jet A/A-1 and JP-8 fuels are typically mixtures of hundreds of 15 300 (steel) 4.37

compounds, controlled only by the deﬁned boiling point ranges. 7.5 159 (glass) 3.75

6.2 33 (glass) 3.70

Current analyses can identify over 200 chemical species in sample

5 150 (glass) 4.44 11.50

fuels [73]. Jet A type fuels consist of the order of 75–85% paraf-

5 91 (glass) 4.50

ﬁn, both straight chain and cyclic, with the balance almost entirely

5 70 (glass) 3.81

aromatic compounds [74]. 2.5 150 (glass) 5.21

1.9 40 (glass) 3.95 13.65

Regarding ﬂammability properties of aviation fuels, the two

5.0 11.5

most widely used parameters are the fuel’s ﬂash point and ﬂamma-

bility limits given in terms of altitude and temperature data. The

current standardized methods for determining ﬂash point are

5. Tests for ﬂammability limits of ethanol: review

ASTM D56 and ASTM 3828 [46].

The ﬂammability limits of a given fuel provide more information

Coward and Jones [1] also carried out tests using ethanol; they

than the ﬂash-point measurement, especially when presented as

used different vessel sizes and separated the limits by ﬂame prop-

regions of ﬂammability in terms of altitude and fuel temperature.

agation direction. Table 7 summarizes their results. In Table 7, the

At any given altitude (pressure), a fuel will have a lean (or lower)

lower limit of ethyl alcohol in air with upward ﬂame propagation

ﬂammability limit (LFL) and a rich (or upper) ﬂammability limit.

in a 5 cm diameter tube open at the ﬁring end is 4.25 or 4.40%. The

Within these limits, ignition of vapors may occur, while outside

limits in a small vessel have been determined to be 2.6 and 9.0%;

the limits, no combustion phenomenon should be observed. Over 3 3

in a 1400 cm upright closed tube with 600 cm of liquid at the

the years, there have been numerous experiments to determine the

bottom and central ignition, the limits were 3.2% (saturation of the

ﬂammability limits of aviation turbine fuels. A typical plot of these ◦ ◦

atmosphere at 10.60 C) and 18.9% (saturation at 41.2 C) [1].

limits, including ignition energies, for Jet A and Jet B fuels is shown

Zabetakis [2] also carried out tests using ethanol. The ﬂamma-

in Fig. 9, as taken from Fuel Flammability Task Group [5].

bility limits of ethanol with carbon dioxide–air and ethanol with

Both the ﬂash point and ﬂammability limits given above assume

nitrogen–air mixtures at 298 K and 1 atm are shown in Fig. 15.

isothermal and equilibrium conditions, not considering such things ◦

Fig. 16 shows data for ethanol–water vapor mixtures at 100 C and

as possible condensation of fuel vapor and variation of ullage com-

position. In order to provide useful data for fuel tank inerting

requirements, data on the reduction of fuel ﬂammability as a func-

tion of oxygen depletion are needed [46].

Fig. 15. Flammability limits of ethanol mixed with inert substances at 298 K and

Fig. 14. Minimum ignition energy with altitude [5]. 1 atm [2].

46 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

Table 8

Flammability limits for the most important alcohols [2].

Fuel Formula Molecular weight Stoichiometric Combustion enthalpy LFL (vol%) UFL (vol%)

(g/mol) concentration (vol%) (kcal/mol)

Methyl alcohol CH3OH 32.04 12.25 159 6.7 36

Ethyl alcohol C2H5OH 46.07 6.58 306 3.3 19

n-Propyl alcohol C3H7OH 50.09 4.45 448 2.2 14

n-Butyl alcohol C4H9OH 74.12 3.37 596 1.7 12

n-Amyl alcohol C5H11OH 88.15 2.72 742 1.4 10

n-Hexyl alcohol C6H13OH 102.17 2.27 888 1.2 –

atmospheric pressure. Table 8 shows ﬂammability limit results for he reports ﬂammability values for pure ethanol of 3.3 vol% for LFL

◦

the most important alcohols. and 19 vol% for UFL, both at 25 C.

Seaton [75] proposed a mathematical model derived from Le Brooks and Crowl [80] carried out experiments on ﬂamma-

Chatelier’s law to estimate the ﬂammable limits of vapors in air. bility of some alcohols. The experiments were conducted for

Both the lower and the upper ﬂammable limits can be estimated methanol, ethanol, acetonitrile, and toluene. The experimental

from nothing more than the chemical structure of a compound. device consisted of a 20 L volume spherical vessel with a fuse wire

Although there are some outstanding exceptions, the model pre- igniter and ignition energy of 10 J. Fuel was injected using a preci-

dicts LFL values to within 10% of published values. Likewise, the sion syringe. Nitrogen and oxygen were introduced with pressure

model predicts most UFL values to within 20% of published values. gauges. The pressure criterion was applied, considering those above

In the case of LFL for ethanol, he estimated 3.39% (vol) and for the 7% of initial pressure as ﬂammable. The LFL obtained were close to

UFL, he estimated 16.1% (vol). the data in the literature. All the experiments were carried out at

McCormick and Parish [76] prepared a study sponsored by the 298 K and 1 atm.

National Renewable Energy Laboratory (NREL-USA) about the tech- The ﬂammability test device had the following components:

nical barriers to the use of ethanol in diesel fuel. They used ethanol

ﬂammability limits values of 3.3 for LFL and 19% for UFL (vol% at

◦ - A high precision pressure transducer (+/− 0.05% full scale) for low

25 C). In that work, a report prepared for Growmark, Inc. [77]

pressures, 0–1.7 bar abs, utilized for the gas mixture.

and Battelle [78] demonstrated that blends of 10%, 15%, and 20%

- A high pressure transducer, 0–13.6 bar abs, utilized to gauge the

ethanol in conventional diesel have combustion safety characteris-

pressure during the explosion.

tics essentially identical to pure ethanol. These data were acquired

- Thermocouples to gauge the temperature at the center of the

using diesel ethanol blends that contained no emulsiﬁer.

vessel, at the inner wall, and at the external wall.

Waterland et al. [79] prepared a report sponsored by an agency

of the United States government about the safety and performance

assessment of ethanol/diesel blends (E-Diesel). In this document,

The main difference from the other equipment is the liquid

injection port, which is connected to the vessel and protected from

the explosion pressure by a high pressure manual globe valve.

A 1 mL Kloehn #400 precision syringe was used. The liquid was

injected when the vessel pressure was 2.07 kPa (below saturation

pressure of all analyzed fuels). The liquid evaporated as soon as it

entered the vessel due to low pressure inside the vessel. Each test

takes from 30 to 45 min, and they were divided into four stages.

The experimental result from Brooks and Crowl [80] for the LFL

of ethanol was 3.7% (according to Kutcha [3], the LFL is 3.3%). On the

other hand, the experimental result for the OCL (oxygen concentra-

tion limit) was 9.8% (according to Kutcha [3], the OCL is 10.5%). The

◦

saturation pressure of ethanol was taken at 25 C and 593.5 mmHg

(Perry and Green [81]). The results are shown in Fig. 17.

Brooks and Crowl [82] carried out experiments with the

ﬂammability of vapors above aqueous solutions of ethanol and ace-

tonitrile in a 20 L combustion apparatus. They reported that the

lower ﬂammability limits (LFL) of ethanol decreased from 3.7% for

pure vapor to 3.2% for saturated water vapor.

Astbury [83] presented a brief review of the properties and

hazards of some alternative fuels, including ethanol and methanol.

He also reports some ﬂammability values for the aforementioned

alcohols and other alternative fuels of interest.

Ma [84] reported a methodology based on thermal balance

for estimating the ﬂammability limits of a mixture with ethanol.

Melhem [85] presented a general method for the estimation of

ﬂammability envelopes for chemical mixtures based on chemical

equilibrium. The impact of the mixture’s initial temperature, the

presence of diluents, and elevated system pressures are implicitly

accounted for. Hansen et al. [86] published a review which dis-

cussed the properties and speciﬁcations of ethanol blended with

diesel fuel with special emphasis on the factors that were critical

Fig. 16. Flammability limits of ethanol mixed with water vapor–air mixtures at

◦

100 C and 1 atm [2]. to the potential commercial use of these blends. The three previous

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 47

Fig. 17. Experimental data for ethanol ﬂammability limits [80].

authors reported that for ethanol at ambient pressure and room liquid injection. The liquid was injected when the pressure in the

temperature, LFL was 3.3 (vol%) and UFL was 19 (vol%). reaction vessel was 2 kPa (0.3 psia), which is below the saturation

Pidol et al. [87] studied the properties of ethanol blended fuels vapor pressure for ethanol used in this research. The liquid ﬂashed

and evaluated their behavior in conventional diesel combustion to vapor upon injection into the reaction test vessel. The evaporator

and advanced combustion with low temperature combustion (LTC). was used only in the experiments to ﬁnd the UFLs.

These authors reported that the addition of ethanol into Diesel fuel The rubber stopper was attached to four rods connected to four

affects some key properties such as blend stability, cetane num- springs to hold it secure in the vessel port. The 20 L spherical glass

ber or ﬂash point, and the fuel formulation was thus improved. For vessel was equipped with a high precision pressure transducer and

these studies, Pidol et al. used the ethanol ﬂammability limit values two thermocouples, one positioned near the wall (type E), and the

reported by Rakopoulos et al. [88]. other positioned in the inner core of the vessel (type K). The pres-

sure transducer measured initial pressure and pressure variation

during the explosion. The spark gap was 6.4 mm (1/4 in.), and igni-

6. Experimental tests with ethanol

tion energy was 90 J. The ﬂammability apparatus was automated

using a Lab-View program.

6.1. Flammability apparatus

A heating chamber ﬂammability apparatus was specially con- 6.2. Description of experimental procedure

structed on the basis of the American standard ASTM E-681 [28]. A

20-L spherical test vessel was used. It should be noted that the vol- The Visual Criterion would be the most appropriate and was

ume speciﬁed by the American standard was increased because no chosen for determining the ﬂammability limits of ethanol [89]

rule is provided for use at high temperatures and reduced pressures, (pure and/or mixed) with air. As mentioned, American standard

when the situation requires less air and fuel. A drawing of the appa- E681-04 [28] was chosen as a basis for building a prototype heat-

ratus, which was adapted from Brooks and Crowl [80], is shown in ing chamber that was modiﬁed in order to use the visual criterion

Fig. 18. and validate the experimental method for ethanol.

An evaporator was used in the connection line to the vessel. In the experiments it is essential that the vessel have suf-

This guaranteed that the fuel sample entered the vessel as a vapor, ﬁcient ethanol vapor pressure to form ﬂammable mixtures in

even at low temperatures. A 1 mL hypodermic syringe was used for air at atmospheric and negative pressures (below atmospheric

Fig. 18. Flammability apparatus.

Adapted from [80].

48 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

irregular ﬂame propagation or insufﬁcient luminescence in the vis-

ible spectrum).

As there is interest in testing ethanol as a fuel for aircraft, the

tests should be bounded by the initial pressure of the local environ-

ment down to 13 kPa (100 mmHg). Considering a typical cruising

altitude of the aircraft of 40,000 ft (approximately 18.82 kPa),

13 kPa is well below the limit of the American standard. The tem-

◦

perature of the heating chamber can vary between 20 and 200 C.

A uniform mixture of a gas or vapor with air should be ignited in a

closed vessel and propagation of the ﬂame from the ignition source

will be observed. The concentration of ethanol is varied between

each test until the composition that will achieve ﬂame propagation

is determined.

Each run took between 15 and 25 min to complete. First, the

reaction vessel was purged with nitrogen and the pressure was

reduced to less than 2 kPa abs (0.3 psia) to insure removal of all

water and other combustion products from previous runs. Second,

the liquid was injected into the reactor at low pressure to insure its

complete vaporization. Third, the synthetic air was loaded into the

reactor and the concentrations of all the components were deter-

Fig. 19. Visual criterion for ﬂame propagation inside a 20 L vessel [28].

mined from the partial pressures. Fourth, the video camera was

turned on, the mixture was ignited, and the results were collected

pressure) (future experiments) with different temperatures. The and stored.

method is based on electric ignition and the visual observation of All of the experiments were run at atmospheric pressure and

◦

ﬂame propagation in the mixture in the vessel. Users could expe- approximately 25 C to compare with published data for this pres-

rience some problems if the ﬂames are difﬁcult to observe (e.g., sure and validate the experimental procedure.

Fig. 20. Minimum volume of ethanol as function of vessel size for different temperatures at 1 atm.

Fig. 21. Maximum volume of ethanol as function of vessel size for different temperatures at 1 atm.

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 49

◦

Fig. 22. LFL, initial temperature: 42 C, sample volume: 1.70 mL, ﬂammable mixture.

The experimental procedure was carried out according to the k) Activate the electrode (for ignition).

following the sequence: l) Observe whether the ﬂame propagates or not, according to the

standard indications.

m) Start the process again in order to obtain values with and with-

a) Double check the equipment.

out ﬂame propagation for the lower and upper ﬂammability

b) Purge of a glass vessel twice using an inert gas (e.g., N2).

limits.

c) Purge the vessel once using an oxidant (e.g., synthetic air, 80%

N2 and 20% O2).

d) Set a pressure value that guarantees ethanol evaporation dur- Each test was video recorded, and later each video was edited

ing its introduction into the vessel (the ethanol vapor pressure with computer video editing software. This allowed determining

curve should be used as a reference). whether a sample had ﬂame propagation or not at ﬁxed tempera-

e) Set the desired volume of ethanol (a 1 mL scalp 19G hypodermic ture and pressure conditions. Standard indications for determining

syringe could be used; a magnetic stirrer should be operating whether or not a mixture is ﬂammable are of great importance.

when the sample is introduced). A mixture is labeled as ﬂammable if the ﬂame propagates

f) Once the sample has been completely vaporized, slowly allow upward and outward from the vessel walls, forming an arc larger

◦

access of oxidant until the ﬁnal pressure is reached. than 90 measured from the ignition source to the vessel walls

g) Check temperature and pressure inside the vessel at all times. [25,30]. Flame must be continuous along the region of the vessel

h) Allow the magnetic stirrer to operate for another 2 min and then shown in Fig. 19. If ﬂame propagation could not be reproduced,

turn it off. or ﬂame extension was not clear, for example, there was no uni-

i) Turn off the room lamp. form propagation, ﬂame structure was irregular, or the ﬂame did

j) Start video recording and simultaneously record the tempera- not ﬁll at least half the vessel even for the maximum ﬂammable

ture and pressure inside the vessel. concentration in air, an ignition probability of 50% should be used.

50 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

◦

Fig. 23. LFL, initial temperature: 40 C, sample volume: 1.40 mL, non-ﬂammable mixture.

6.3. Mass of ethanol needed to form ﬂammable mixtures with air LFL:

No. of moles of ethanol in the vessel

× =

Minimum and maximum masses of ethanol needed to form a LFLEthanol 100

Total no. of moles of gas in the vessel

ﬂammable mixture in vessels of 5 and 12 L is calculated for 1 atm

◦

and 25 C. A minimum and maximum mass of ethanol needed to LV . 1

= × 100 (19)

form a ﬂammable mixture is calculated in vessels of 5 and 12 L MW (V/22.4)(p/p0)(T0/T)

(considering the dimensions of the American Standard, E681-04

◦ 3

[28]) mixed with air at 1 atm and 25 C. in which LV is the sample volume, cm ; is the sample speciﬁc

From previous discussions in this article, the following mass, g/cm ; MW is the sample substance molar mass, g/mol; p is

ﬂammability limits will be considered: LFLEthanol = 3.3% and the pressure of the test, mmHg; T is the initial temperature of the

UFLEthanol = 19%. For a 5-L vessel at 298 K, the number of moles of test, K; p0 is the standard pressure, 1 atm (760 mmHg = 101.3 kPa);

C2H5OH at the LFL is: and T0 is the standard temperature, 273 K. Thus:

5 273 LV .

nLFL = · 0.033 = 0.006784 (18) . LFL = k, (20)

22 4 298 Ethanol MW

Since the molar mass of ethanol is 46.07 g/mol, the mass of

in which

ethanol to form a ﬂammable mixture with air at the LFL will be

0.006784 46.07 = 0.32 g (=0.40 mL). (22.4)(p0)(100)

k = . (21)

If the vessel volume is 12 L, the minimum mass of ethanol is

(V)(T0)

0.76 g (0.97 mL). For the mixture at the upper limit of ﬂamma-

bility, the calculation procedure is entirely analogous. For the 5-L For example, a test carried out in the 12-L vessel (12,000 cm )

◦

vessel, the maximum mass of ethanol is 1.83 g (2.32 mL), and for at 25 C and 715 mmHg (atmospheric pressure in the town of

the 12-L vessel, the maximum mass of ethanol is 4.39 g (5.56 mL). Guaratinguetá, Brazil) with a sample of ethanol ( = 0.789 g/cm3

Figs. 20 and 21 present the different minimum and maximum vol- and MW = 46.07 g/gmole) produced the ﬁrst ﬂame spread with 1 mL

umes of ethanol, for different sizes of vessel, for tests at different (L1) of ethanol. The last volume that did not produce a ﬂame was

temperatures. This calculation was essential to ﬁnd the minimum 0.9 mL (L2) of ethanol. The sample volume for LFL was then taken

and maximum mass of ethanol for the 20 L vessel that is used in the as the average: 0.95 mL. For this case:

ﬂammability apparatus used in this study.

(22.4)(719)(100)

k = = 519.65

(12)(T )

6.4. Description of the procedure to calculate LFL and UFL from 0

experimental data

The value of k remains unchanged for all tests in a 12-L vessel.

Therefore, the LFL obtained was

The experimental results should be optimized, tested, and

eventually modiﬁed for volumetric concentration formats in order . .

(0 95)(0 789)(298) k ◦

LFL = = 3.52% (at 715 mmHg and 25 C).

Ethanol .

to compare with other experimental data from the scientiﬁc liter- (46 07)(715)

ature. The procedure is shown for LFL. For UFL, it is the same. For

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 51

◦

Fig. 24. LFL, initial temperature: 170 C, sample volume: 1.10 mL, ﬂammable mixture.

◦

6.5. Results and discussion rises to approximately 600–800 C in less than 0.1 s. This behavior

is similar to that speciﬁed by the standard for the case of ﬂame

6.5.1. Behavior of ﬂame propagation propagation. Fig. 23 illustrates the case with no ﬂame propagation.

Fig. 22 shows a sequence of photographs for a test with propaga- When the volume of ethanol increases from the value corre-

tion in a mixture near LFL. Flame propagation was evident and there sponding to the LFL, the ﬂame propagates across the vessel wall,

was an abrupt rise of temperature and pressure. This caused a phe- brightness increases as does the intensity of the detonation noise.

nomenon similar to an explosion. The use of protective glasses and This can be seen in Fig. 24.

headphones is important when carrying out any tests, especially The behavior is similar for UFL, except that the noise is not as

in those to determine the LFL. No vessel, electrode, or other equip- intense as it is for LFL. In most experiments to determine UFL,

ment was lost during the tests. The springs held the cover in place. no detonations were heard, unlike in the experiments for LFL. It

It was noted that for mixtures near the LFL, the ﬂame propagates should be noted that the absence of detonation does not imply

upwards from the ignition spark, reaches the top of the vessel and an absence of ﬂame propagation. The sudden rises in temperature

then extends to the wall and propagates downwards. Temperature registered by the thermocouples and the ﬂame propagation across

52 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

Fig. 25. Hydrated ethanol–air mixtures, LFL, and UFL at 101 kPa.

the vessel wall were the factors that led to catalog the mixture as percentage of water (4%) present in the sample (hydrated ethanol

◦

ﬂammable. 96% INPM). This behavior at the upper limit due to the presence

of inert substances had already been anticipated by Zabetakis [2].

6.5.2. Results for atmospheric pressure

In this stage of the investigation, about 70 experiments were

7. Conclusion

performed; 90% of these tests were video recorded and all of them

were recorded in LabView. Fig. 25 shows ﬂammability results in

A comprehensive literature review of ﬂammability limits

terms of fuel volume as a function of temperature and of tempera-

of combustible mixtures with emphasis on ethanol was con-

ture as a function of the volume percentage of gaseous fuel. Results

ducted. The experimental procedures to determine the limits were

obtained here agreed with data published in the scientiﬁc litera-

thoroughly discussed. Flammability limit results at atmospheric

ture on ﬂammability limits of ethanol for behavior of both LFL and

pressure for ethanol were presented.

UFL and in LFL values [2,3,41,80]. This agreement showed that the

equipment was operated correctly and also that the experimental

procedure adopted was carried out appropriately. Acknowledgements

As mentioned, LFLT and the UFLT for any temperature T at 1 atm

pressure can be calculated based on the limits determined for 298 K Funding for this work was provided by the Fundac¸ ão de Amparo

using Eqs. (7) and (8). Predictions by these equations are also shown à Pesquisa do Estado de São Paulo – FAPESP, Brazil, through projects

in Fig. 25. LFL results agreed well with results from previous inves- 2009/09738-7 (regular sponsorship) and 2009/09008-9 (post doc

tigations. It is specially worth stating that the present work found fellowship to CJRC), and by Empresa Brasileira de Aeronáutica

◦

3.5% for the ethanol LFL at 25 C, while Kuchta [3] reported 3.3% and (EMBRAER).

Brooks and Crowl [80] 3.7%, for LFL at the same temperature.

Fig. 25 also shows that the behavior of UFL as a function of

References

temperature is similar to that predicted in the literature (given by

Eq. (8)). However, UFL values were about 4.5% (absolute) lower in

◦ [1] H.F. Coward, G.W. Jones, Limits of Flammability of Gases and Vapors, US Bureau

all cases. Zabetakis [2] reported 19% for UFL at 25 C. Other articles of Mines Bulletin, United States Government Printing Ofﬁce, Washington, 1952,

[85,87] report the same value, but they referred to Zabetakis’ data. p. 503.

[2] M.G. Zabetakis, Flammability Characteristics of Combustible Gases and Vapors,

A value of 16.1% for UFL at the same temperature was reported

US Bureau of Mines Bulletin, United States Government Printing Ofﬁce, Wash-

[75]. The experiments were repeated three times for this pressure

ington, 1965, p. 627.

and the data were conﬁrmed. Extreme care was taken to avoid [3] J.M. Kuchta, Investigation of Fire and Explosion Accidents in the Chemical, Min-

ing and Fuel-Related Industries – A Manual, US Bureau of Mines Bulletin, United

condensation of the samples when they were injected into the

States Government Printing Ofﬁce, Washington, 1985, p. 680.

vessel. An evaporator was utilized to guarantee sample injection

[4] L.J. Nestor, Investigation of Turbine Fuel Flammability within Aircraft Fuel

in the vapor phase. The difference is probably due to the small Tanks, Federal Aviation Administration, USA, 1967.

C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54 53

[5] Fuel Flammability Task Group, A review of the ﬂammability hazard of jet a fuel [36] A. Takahashi, Y. Urano, K. Tokuhashi, S. Kondo, Effect of vessel size and shape

vapor in civil transport aircraft fuel tanks, Final Report No. DOT/FAA/AR-98/26, on experimental ﬂammability limits of gases, J. Hazard. Mater. A 105 (2003)

Federal Aviation Administration, USA, 1998. 27–37.

[6] J.E. Shepherd, J.C. Krok, J.J. Lee, Spark ignition energy measurements in Jet A, [37] ASTM E918-83, Standard Practice for Determining Limits of Flammability of

Report FM97-9, Explosion Dynamics Laboratory, 2000. Chemicals at Elevated Temperature and Pressure, American Society for Testing

[7] E. Ott, Effects of fuel slosh and vibration on the ﬂammability hazards of and Materials (ASTM) Washington Ofﬁce, 2011.

hydrocarbon turbine fuels within aircraft fuel tanks, Technical Report AFAPL- [38] ASTM-E2079, Standard Test Methods for Limiting Oxygen (Oxidant) Concen-

TR-70-65, Fire Protection Branch of the Fuels and Lubrication Division, tration in Gases and Vapors, American Society for Testing and Materials (ASTM)

Wright-Patterson Air Force Base, Ohio, 1970. Washington Ofﬁce, 2011.

[8] T. Kosvic, L. Zung, M. Gerstein, Analysis of aircraft fuel tank ﬁre and explosion [39] J.C. Jones, Combustion Science: Principles and Practice, Millenium Books, New-

hazards, Technical Report AFAPL-TR-71-7, Air Force Aero Propulsion Labora- town, Australia, 1993.

tory, Wright-Patterson Air Force Base, Ohio, 1971. [40] J.A. Carvalho, M.Q. McQuay, Principles of Applied Combustion, UNESP, São

[9] F. Van den Schoor, F. Verplaetsen, The upper explosion limit of lower alkanes Paulo, Brazil, 2007 (in Portuguese).

and alkenes in air at elevated pressures and temperatures, J. Hazard. Mater. [41] D. Drysdale, An Introduction to Fire Dynamics, 3rd ed., John Wiley & Sons,

128 (2006) 1–9. University of Edinburgh, UK, 2011, ISBN 978-0-470-31903-1.

[10] Physical & thermodynamic properties (methanol, ethanol, propanol, and [42] S. Kondo, K. Takizawa, A. Takahashi, K. Tokuhashi, On the temperature depend-

butanol), Chem. Eng. J. 1 (June) (1976) 10. ence of ﬂammability limits of gases, J. Hazard. Mater. 187 (2011) 585–590.

[11] S.Y. Liao, Q. Cheng, D.M. Jiang, J. Gao, Experimental study of ﬂammability limits [43] J.E. Shepherd, J.C. Krok, J.J. Lee, Jet A Explosion Experiments: Laboratory Testing,

of natural gas–air mixture, J. Hazard. Mater. B 119 (2005) 81–84. Report FM97-5, Explosion Dynamics Laboratory, 1997.

[12] C.J.R. Coronado, Technical-economical analysis of an 100-kg/hr biomass gasiﬁer [44] J. Arnaldos, J. Casal, E. Planas-Cuchi, Prediction of ﬂammability limits at reduced

for powering of an internal combustion engine, Ph.D. Thesis, in Portuguese, pressures, Chem. Eng. Sci. 56 (2001) 3829–3843.

São Paulo State University, Campus of Guaratinguetá, Post Graduate Program [45] D.R. Stull, Fundamentals of ﬁre and explosion, A.I.Ch.E. Monograph Series No.

in Mechanical Engineering, 195 pp., 2007. 10, New York, 1977.

[13] D. Naegeli, W.D. Weatherford Jr., Practical ignition limits for low molecular [46] S.M. Summer, Technical Note Limiting Oxygen Concentration Required to

weight alcohols, Fuel 68 (1989) 45–48. Inert Jet Fuel Vapors Existing at Reduced Fuel Tank Pressures, Report No.

[14] Y. Shebeko, W. Fan, I.A. Bolodian, V.Y. Navzenya, An analytical evaluation of DOT/FAA/ARTN-02/79, U.S. Department of Transportation, Federal Aviation

ﬂammability limits of gaseous mixtures of combustible–oxidizer–diluent, Fire Administration, Ofﬁce of Aviation Research, Washington, DC, 2003.

Saf. J. 37 (2002) 549–568. [47] J. Chomiak, J. Jarosinski, Flame quenching by turbulence, Combust. Flame 48

[15] E. Brandes, M. Mitu, D. Pawel, The lower explosion point: a good measure for (1982) 241–249.

explosion prevention: experiment and calculation for pure compounds and [48] A.S. Sokolik, W.P. Karpov, E.S. Semenov, Phys. Comb. Expl. 3 (61) (1967).

some mixtures, J. Loss Prevent. Process Ind. 20 (2007) 536–540. [49] W.P. Karpov, E.S. Severin, Phys. Comb. Expl. 14 (33) (1978).

[16] F. Van den Schoor, F. Verplaetsen, J. Berghmans, Calculation of the upper [50] F.C. Lockwood, I.E. Megahed, Combust. Sci. Technol. 19 (1978) 77.

ﬂammability limit of methane/air mixtures at elevated pressures and temper- [51] ASTM E2079, Standard test method for Limiting Oxygen (Oxidant) Concentra-

atures, J. Hazard. Mater. 153 (2008) 1301–1307. tion in Gases and Vapors, American Society for Testing and Materials (ASTM)

[17] F. Van den Schoor, R.T.E. Hermanns, J.A. van Oijen, F. Verplaetsen, L.P.H. de Goey, Washington Ofﬁce, 2010.

Comparison and evaluation of methods for the determination of ﬂammability [52] I.A. Zlochower, M.G. Gregory, M. Green, The limiting oxygen concentration and

limits, applied to methane/hydrogen/air mixtures, J. Hazard. Mater. 150 (2008) ﬂammability limits of gases and gas mixtures, J. Loss Prevent. Process Ind. 22

573–581. (2009) 499–505.

[18] C.C. Chen, T. Wang, H. Liaw, H. Chen, Nitrogen dilution effect on the ﬂammability [53] L.G. Britton, Flammability hazards of lower aliphatic aldehydes at elevated

limits for hydrocarbons, J. Hazard. Mater. 166 (2009) 880–890. pressure and temperature, Process Saf. Prog. 17 (1998) 138–148.

[19] F. Gharagheizi, Prediction of upper ﬂammability limit percent of pure [54] L.G. Britton, Operating atmospheric vent collection headers using methane gas

compounds from their molecular structures, J. Hazard. Mater. 167 (2009) enrichment, Process Saf. Prog. 15 (1996).

507–510. [55] F.T. Bodurtha, Industrial Explosion Prevention and Protection, McGraw-Hill,

[20] Y. Koshiba, T. Takigawa, Y. Matsuoka, H. Ohtani, Explosion characteristics of New York, USA, 1980, pp. 13–16.

ﬂammable organic vapors in nitrous oxide atmosphere, J. Hazard. Mater. 183 [56] K.L. Cashdollar, I.A. Zlochower, G.M. Green, R.A. Thomas, M. Hertzberg, Flamma-

(2010) 746–753. bility of methane, propane and hydrogen gases, J. Loss Prevent. Process Ind. 13

[21] Y.L. Shoshin, L.P.H de Goey, Experimental study of lean ﬂammability limits of (2000) 327–340.

methane/hydrogen/air mixtures in tubes of different diameters, Exp. Therm. [57] M. Hertzberg, K. Cashdollar, A.I. Zlochower, Flammability limit measurements

Fluid Sci. 34 (2010) 373–380. for dusts and gases: ignition energy requirements and pressure dependences,

[22] H. Miao, L. Lu, Z. Huang, Flammability limits of hydrogen-enriched natural gas, in: Twenty-ﬁrst Symposium (International) on Combustion/The Combustion

Int. J. Hydrogen Energy 36 (2011) 6937–6947. Institute, 1986, pp. 303–313.

[23] C.M. Piqueras, J. García-Serna, M.J. Cocero, Estimation of lower ﬂammability [58] J. Bond, Sources of ignition: ﬂammability characteristics of chemicals and prod-

limits in high-pressure systems. Application to the direct synthesis of hydro- ucts, J. Hazard. Mater. 31 (1992) 190.

gen peroxide using supercritical and near-critical CO2 and air as diluents, J. [59] F. Norman, F. Van den Schoor, F. Verplaetsen, Auto-ignition and upper explosion

Supercrit. Fluids 56 (2011) 33–40. limit of rich propane–air mixtures at elevated pressures, J. Hazard. Mater. 137

[24] J.R. Rowley, R.L. Rowley, W.V. Wilding, Estimation of the lower ﬂammability (2006) 666–671.

limit of organic compounds as a function of temperature, J. Hazard. Mater. 186 [60] R.K. Eckhoff, M. Ngo, W. Olsen, On the minimum ignition energy (MIE) for

(2011) 551–557. propane/air, J. Hazard. Mater. 175 (2010) 293–297.

[25] S. Kondo, A. Takahashi, K. Takizawa, K. Tokuhashi, On The pressure dependence [61] S. Kondo, A. Takahashi, K. Tokuhashi, Calculation of minimum ignition energy

of ﬂammability limits of CH2 CFCF3, CH2F2 and methane, Fire Saf. J. 46 (2011) of premixed gases, J. Hazard. Mater. 103 (2003) 11–23.

289–293. [62] M. Caron, M. Goethals, G. De Smedt, J. Berghmans, S. Vliegen, E. Van’t Oost,

[26] I. Glassman, Combustion, 2nd ed., Academic Press, San Diego, CA, 1996. A. van den Aarssen, Pressure dependence of the auto-ignition temperature of

[27] A.A. Tareq, Flammability characteristics of pure hydrocarbons, Chem. Eng. Sci. methane/air mixtures, J. Hazard. Mater. 65 (1999) 233–244.

58 (2003) 3629–3641. [63] V. Tihay, P. Gillard, D. Blanc, Ignition study of acetone/air mixtures by using

[28] ASTM E681, Standard Test Method for Concentration Limits of Flammability laser-induced spark, J. Hazard. Mater. 209–210 (2012) 372–378.

of Chemicals (Vapors and Gases), American Society for Testing and Materials [64] U.J. Pfahl, M.C. Ross, J.E. Shepherd, K.O. Pasamehmetoglu, C. Unal, Flamma-

(ASTM), W. Conshohocken, PA, 2010. bility limits, ignition energy, and ﬂame speeds in H2–CH4–NH3–N2O–O2–N2

[29] Safety information sheet of the chemical product – anhydrous ethyl alcohol mixtures, Combust. Flame 123 (2000) 140–158.

fuel, Brazilian Oil Inc., PETROBRAS S.A., No. FISPQ: Pb0004 P, 23.01.09. [65] S. Kondo, K. Takizawa, A. Takahashi, K. Tokuashi, A. Sekiya, A study on ﬂamma-

[30] M. Vidal, W. Wong, W.J. Roger, M.S. Mannan, Evaluation of lower ﬂammability bility limits of fuel mixtures, J. Hazard. Mater. 155 (2008) 440–448.

limits of fuel–air–diluent mixtures using calculated adiabatic ﬂame tempera- [66] S. Kondo, K. Takizawa, A. Takahashi, K. Tokuhashi, A. Sekiya, Flammability limits

tures, J. Hazard. Mater. 130 (2006) 21–27. of isobutane and its mixtures with various gases, J. Hazard. Mater. 148 (2007)

[31] B.S. Fuman Zhao, Experimental measurements and modeling prediction of 640–647.

ﬂammability limits of binary hydrocarbon mixtures, Master of Science Thesis, [67] S. Kondo, K. Takizawa, A. Takahashi, K. Tokuhashi, Extended Le Chatelier’s for-

University of Tianjin, PR China, 108 pp., 2008. mula for carbon dioxide dilution effect on ﬂammability limits, J. Hazard. Mater.

[32] V. Schroder, M. Molnarne, Flammability of gas mixtures. Part 1. Fire potential, A 138 (2006) 1–8.

J. Hazard. Mater. A 121 (2005) 37–44. [68] H.P. Von Niepenberg, R. Oppenberg, F. Rick, Zusammenstellung Sicher-

[33] E. Brandes, E.A. Ural, Towards a global standard for ﬂammability determination, heitstechnischer Daten zur Vermeidung van Explosionen in Feurräumen,

in: Loss Preventions Symposium, American Institute of Chemical Engineers, Industriefeurung-Essen 11 (1978) 17–31.

Spring National Meeting, New Orleans, 2008. [69] J.E. Hustad, K.O. Sonju, Experimental studies of lower ﬂammability limits of

[34] E.A. Ural, R.G. Zalosh, A mathematical model for lean hydrogen air–steam mix- gases and mixtures of gases at elevated temperatures, Combust. Flame 71

ture combustion in closed vessels, in: Twentieth Symposium (International) on (1988) 283–294.

Combustion, The Combustion Institute, 1984, pp. 1727–1734. [70] CRC, Handbook of Aviation Fuel Properties, CRC Report 530, Society of Auto-

[35] O. Fuss, M. Molnarne, A. Schonbucher, V. Schroder, Sauertoffgrenzkonzen- motive Engineers, Warrendale, PA, 1983.

tration von brennbaren Gasen and Dampfen, Chem. Ing. Tech. 75 (2003) [71] Fuel characterization test, Minimum energy for ignition, Wright Patterson Air

101–104. Force Base, Dayton OH, Classiﬁed Air Force report, Attachment D, 1992.

54 C.J.R. Coronado et al. / Journal of Hazardous Materials 241–242 (2012) 32–54

[72] E.M. Goodger, Liquid fuels for transport, Prog. Energy Combust. Sci. 8 (1982) [81] R.H. Perry, D.W. Green, Perry’s Chemical Engineers Handbook, 7th ed., McGraw-

233–260. Hill, NY, 1997.

[73] H.T. Mayﬁeld, JP-8 Composition and Variability, USAF Report AL/EQ-TR-1996- [82] M.R. Brooks, D. Crowl, Vapor ﬂammability above aqueous solutions of

0006 (AD-A317177), May 1996. ﬂammable liquids, J. Loss Prevent. Process Ind. 20 (2007) 477–485.

[74] W.G. Dukek, Aviation and other gas turbine fuels, in: M. Howe-Grant (Ed.), Kirk- [83] G.R. Astbury, A review of the properties and hazards of some alternative fuels,

Othmer Encyclopedia of Chemical Technology, vol. 3, John Wiley & Sons, 1992, Process Saf. Environ. Prot. 86 (2008) 397–414.

pp. 788–812. [84] T. Ma, A thermal theory for estimating the ﬂammability limits of a mixture, Fire

[75] W. Seaton, Group contribution method for predicting the lower and the upper Saf. J. 46 (2011) 558–567.

ﬂammability limits of vapors in air, J. Hazard. Mater. 27 (1991) 169–185. [85] G.A. Melhem, A detailed method for estimating mixture ﬂammability limits

[76] R. Mc Cormick, R. Parish, Advanced Petroleum Based Fuels Program and Renew- using chemical equilibrium, Process Saf. Prog. 16 (1997) 203–218.

able Diesel Program Milestone Report: Technical Barriers to the Use of Ethanol [86] A. Hansen, Q. Zhang, P. Lyne, Ethanol–diesel fuel blends, a review, Bioresour.

in Diesel Fuel, November 2001, NREL/MP-540-32674. Technol. 96 (2005) 277–285.

[77] Growmark, Inc., Flammability Limits for Ethanol/Diesel Blends Final Report [87] L. Pidol, B. Lecointe, L. Starck, N. Jeuland, Ethanol–biodiesel–diesel fuel blends:

from Battelle Agreement No. CP037328R, September 1998. performances and emissions in conventional diesel and advanced low temper-

[78] Battelle, Flammability limits for ethanol/diesel blends, Final Report prepared ature combustions, Fuel 93 (March) (2012) 329–338.

by Battelle, Columbus, OH, 1998. [88] D.C. Rakopoulos, C.D. Rakopoulos, R.G. Papagiannakis, D.C. Kyritsis, Combustion

[79] R. Waterland, S. Venkatesh, S. Unnasch, Safety and performance assessment heat release analysis of ethanol or n-butanol diesel fuel blends in heavy-duty

of ethanol–diesel blends, National Renewable Energy Laboratory, September DI diesel engine, Fuel 90 (2011) 1855–1867.

2003, NREL/SR-540-34817. [89] S. Kumar, N. Singh, R. Prasad, Anhydrous ethanol: a renewable source of energy,

[80] M.R. Brooks, D. Crowl, Flammability envelopes for methanol, ethanol, acetoni- Renew. Sustain. Energy Rev. 14 (2010) 1830–1844.

trile and toluene, J. Loss Prevent. Process Ind. 20 (2007) 144–150.