Combustible Organic Materials: Determination and Prediction of Combustion Properties [2 ed.] 9783110782042

Citation preview

Mohammad Hossein Keshavarz Combustible Organic Materials

Also of Interest Encyclopedia of Oxidizers Eckart Walter Schmidt, 2022 ISBN 978-3-11-075024-9, e-ISBN (PDF) 978-3-11-075029-4, e-ISBN (EPUB) 978-3-11-075039-3

Encyclopedia of Liquid Fuels Eckart Walter Schmidt, 2022 ISBN 978-3-11-075025-6, e-ISBN (PDF) 978-3-11-075028-7, e-ISBN (EPUB) 978-3-11-075038-6

The Properties of Energetic Materials. Sensitivity, Physical and Thermodynamic Properties Mohammad Hossein Keshavarz, Thomas M. Klapötke, 2021 ISBN 978-3-11-074012-7, e-ISBN (PDF) 978-3-11-074015-8, e-ISBN (EPUB) 978-3-11-074024-0 High Explosives, Propellants, Pyrotechnics Ernst-Christian Koch, 2021 ISBN 978-3-11-066052-4, e-ISBN (PDF) 978-3-11-066056-2, e-ISBN (EPUB) 978-3-11-066059-3

Energetic Materials Encyclopedia: Volume 1-3 Thomas M. Klapötke, 2021 ISBN 978-3-11-067465-1

Energetic Compounds. Methods for Prediction of their Performance Mohammad Hossein Keshavarz, Thomas M. Klapötke, 2020 ISBN 978-3-11-067764-5, e-ISBN (PDF) 978-3-11-067765-2, e-ISBN (EPUB) 978-3-11-067775-1

Mohammad Hossein Keshavarz

Combustible Organic Materials Determination and Prediction of Combustion Properties 2nd Edition

Author Prof. Dr. Mohammad Hossein Keshavarz Malek-ashtar University of Technology Department of Chemistry PO Box 83145/115 Shahin-shahr Iran [email protected]

ISBN 978-3-11-078204-2 e-ISBN (PDF) 978-3-11-078213-4 e-ISBN (EPUB) 978-3-11-078225-7 Library of Congress Control Number: 2022938754 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2022 Walter de Gruyter GmbH, Berlin/Boston Cover image: chrispecoraro/E+/Getty Images Typesetting: Integra Software Services Pvt. Ltd. Printing and binding: CPI books GmbH, Leck www.degruyter.com

Preface to the second edition Everything said in the preface to the first edition still holds and essentially does not need any addition or correction. In this revised second edition, the manuscript has been updated and added some recent aspects of combustible organic compounds: (i) Some errors which unfortunately occurred in the first edition have been corrected, and the references have also been updated where appropriate. (ii) Recent works have been reviewed and discussed in each chapter. Moreover, new sections have also been inserted including: (a) Chapter 1 – Two new sections (1.2.1.6 and 1.2.2.3) are introduced for the prediction of flash points of organic compounds using machine learning-developed models, and the improved structural group contributions based on experimental data of a wide range of organic compounds. Moreover, two new sections (1.2.3.8 and 1.2.3.9) are presented for reliable estimation of the flash points of organosilicon compounds and organic compounds containing hazardous peroxide functional groups. Two new sections (1.3.1 and 1.3.2) are also introduced for the prediction of flash point of liquid mixture of organic compounds using mixing rules. (b) Chapter 2 – Two new sections (2.2.4 and 2.2.5) are introduced for the prediction of autoignition temperatures of organic ether compounds and organic hydroxyl compounds. A new section (2.3) is presented for the estimation of the ignition delay of different types of liquid fuels. (c) Chapter 3 – A new section (3.2.4) is introduced to use machine learningdeveloped models for the prediction of the lower flammability limit (LFL) and the upper flammability limit (UFL) of pure organic compounds. (d) Chapter 4 – Two new sections (4.2.3 and 4.2.5) are introduced for the prediction of the net heat of combustion of organic compounds from group contribution-based property models and machine learning-developed models. A new section (4.2.4) is presented for the calculation of the gross heat of combustion of organic compounds as well as salts and ionic liquids of prediction of the net heat of combustion of organic compounds using a generally applicable group- additivity method. A new section (4.2.6) is also introduced for reliable prediction of the net heat of combustion of organosilicon compounds. Mohammad Hossein Keshavarz

https://doi.org/10.1515/9783110782134-202

Contents Preface to the second edition

V

1 Flash Point 1 1.1 Measurement of the FP 2 1.2 Predictive methods of the FP for pure compounds 5 1.2.1 Empirical models 5 1.2.2 The SGC methods 40 1.2.3 QSPR models 48 1.3 Estimation methods of the FP for mixtures 69 1.3.1 Mixing rules 70 1.3.2 Assessment of combination of the mixing rule of Liaw et al. and QSPR models 72 1.3.3 Empirical methods for liquid mixtures 72 1.4 Summary 76 2 Autoignition 77 2.1 Measurement of the AIT 78 2.2 Predictive methods of the AIT for pure compounds 79 2.2.1 The use of SGC by a polynomial of degree 3 for organic compounds 81 2.2.2 A simple QSPR model for various classes of hydrocarbons 84 2.2.3 A new and reliable model for prediction of the AIT of organic compounds containing energetic groups 86 2.2.4 Simple method to assess the AIT of organic ether compounds with high reliability 92 2.2.5 Reliable prediction of autoignition temperature of organic hydroxyl compounds 94 2.3 Autoignition and ignition delay 96 2.4 Summary 101 3 Flammability Limit 103 3.1 Measurement of the LFL and UFL 103 3.2 Predictive methods of the flammability limits 104 3.2.1 The predicted LFL as a function of temperature 105 3.2.2 The use of SGC method for prediction of the LFL and UFL of pure hydrocarbons 106 3.2.3 Extended method for prediction of the UFL of pure compounds 3.2.4 Machine learning-developed models for prediction of LFL and UFL 113

110

VIII

3.3 3.4

Contents

Flammability limit estimation of the hydrocarbon‐inert gas mixture 115 Summary 117

4 Heat of Combustion 119 4.1 Experimental methods for determination of heats of combustion 120 4.2 Different approaches for prediction of the heats of combustion 121 4.2.1 Predicting the standard net heat of combustion for pure hydrocarbons from their molecular structure 123 4.2.2 Prediction of the standard net heat of combustion from molecular structure 126 4.2.3 A comprehensive methodology for prediction of the net heat of combustion from group contribution-based property models 129 4.2.4 A generally applicable group additivity method for the calculation of the gross heat of combustion of organic compounds as well as salts and ionic liquids 136 4.2.5 Machine learning-developed models of prediction of the net heat of combustion of organic compounds 143 4.2.6 Reliable predictions of the net heat of combustion of organosilicon compounds 144 4.2.7 A new method for predicting the gross heat of combustion of polynitro arene, polynitro heteroarene, acyclic and cyclic nitramine, nitrate ester and nitroaliphatic compounds 145 4.3 Summary 149 5 Polymer Flammability 151 5.1 Experimental method based on pyrolysis combustion flow calorimetry 152 5.2 Different approaches for prediction of flammability parameters 153 5.2.1 SGC method of Walters and Lyon for prediction of the heat release capacity 154 5.2.2 SGC method of Lyon et al. for prediction of total heat release (heat of combustion), char yield, and heat release capacity 158 5.2.3 The simplest model for reliable prediction of total heat release (heat of combustion) 162 5.2.4 A simple model for reliable prediction of the specific heat release capacity of polymers 164

Contents

5.2.5

A simple method for the reliable prediction of char yield of polymers 165 Summary 170

5.3 Problems

171

Answers to Problems List of Symbols

177

Appendix A

183

Appendix B

223

Appendix C

241

Appendix D

255

References

259

About the Author Index

277

175

275

IX

1 Flash Point Flash point (FP) of an organic liquid compound is the lowest temperature at which a liquid can form an ignitable mixture in air near the surface of the liquid where its measurement requires an ignition source. The American Society for Testing and Materials (ASTM) defines FP under specified testing conditions and 101.3 kPa pressure [1]. The FP data are widely used to evaluate the fire and explosion hazards of liquids. They have great practical significance in the handling and transporting of such chemicals in bulk quantities. Above the FP temperature, a liquid can produce enough vapor to form a flammable mixture with air. Organic liquids with lower FP can be ignited more easily because the FP of a volatile organic compound is the lowest temperature at which vapors of a fluid will ignite. Fire point is the temperature at which the vapor continues to burn after being ignited. When the ignition source is removed at the FP, which is different from the fire point, the vapor may cease to burn. Since FP and the fire point do not depend on the temperature of the ignition source, a certain concentration of vapor for each flammable liquid in the air is necessary to sustain combustion. Thus, when an ignition source of sufficient strength is applied, the FP of an organic flammable liquid is the lowest temperature at which there will be enough flammable vapor to ignite. The FP of a pure flammable organic compound or mixture is an important parameter in hazard classification of flammable liquids. Since the FP is an approximation of the lower temperature limit, the temperature at which a chemical compound evolves enough vapors to support combustion, it is essential to be updated for safe handling, transportation, and storage of many substances [2]. The values of FP for common organic compounds are widely reported because the knowledge of the combustion potential of an organic compound is crucial when designing safe chemical processes. They can also help firefighters in extinguishing fires because a high FP liquid fire pool can be extinguished by cooling with water mist. Meanwhile, liquids with low values of the FP usually need to be blanketed by dry chemicals or foams [2, 3]. Since cooking oil, for example, has a high FP and its temperature during burning is high (603–733 K), its gasification heat is also high and additional thermal energy is required for fuel evaporation. As the rate of supply of fuel vapor or burning rate is reduced sufficiently enough not to support the flame, the cooking oil fire can be extinguished by water mist mainly through cooling the fuel surface [4]. Due to the presence of large heat capacity of water, it absorbs high quantities of heat and pulls down the temperature of the fire. Thus, water is a good candidate for extinguishing the fires of high FP liquids. Since foam can act as a blanket and not as a heat absorber, it can be used for extinguishing the fires of high or low FP [2]. For industrial processes, the FP is also used to determine the vapor explosion potential [5, 6].

https://doi.org/10.1515/9783110782134-001

2

1 Flash Point

1.1 Measurement of the FP There are five main apparatuses used internationally for measurement of the FP that are given in Table 1.1. Table 1.2 also shows eight different test methods of ASTM, which has been defined based on four of these apparatuses. As seen in Table 1.2, there are two methods for measuring the FP [3]: (1) the closed cup method and (2) the open cup method, which are used for liquids with low and high FP, respectively. The open cup method can measure the FP using the conditions that are met in open vessels and that would be encountered in spills [7]. The vessel is conducted in a container where it is exposed to the air outside. The temperature of the desired substance in the vessel is gradually raised and an ignition source is passed over the top of it. At the FP, the temperature reaches a point at which it “flashes” and ignites. The recorded FP will vary according to the height of the source above the cup, i.e., the distance between the substance and the ignition source. Since low boiling components of the mixture may be lost to the surrounding atmosphere prior to the application of the flame, which gives higher values of the FP [7], this is one disadvantage of the open tester. Closed cup methods may give somewhat lower results than open cup methods. This situation is due to the presence of a physical barrier, which prevents the volatile particles from escaping and causes to approximate an equilibrium between vapor and the air in the Table 1.1: Five main FP devices currently used. Device

Temperature uniformity

Tagliabue (Tag) Pensky-Martens Cleveland Small-Scale (Setaflash) Abel

Liquid Bath Stirred; Metal shell Metal plate across cup base Preheated to fixed temperature Water bath and air gap

Sample volume (mL) 50 75 70 2 or 4 79

Main use Less viscous compounds Viscous compounds Open cup tests Small-scale and flash/no-flash tests European tests

Table 1.2: ASTM standardized method for measuring the FP. Method

Device

Open/ Closed

Heating rate °C·min–1

Range (°C)

D 56 D 92 D 93 D 1310 D 3278 D 3828 D 3941 D 3941

Tag Cleveland Pensky-Martens Tag Small-scale Small-scale Tag or Pensky-Martens Tag or Pensky-Martens

Closed Open Closed Open Closed Closed Closed Closed

1–3 5–6 4–5 1 Fixed temperature Fixed temperature Fixed temperature 0.5

C< =CH2 =CH– =C< =C= =CH =C– –OH –O– >C=O –CHO (aldehyde) –COOH (acid) –COO– (ester) HCOO– (formate) –NH2 –NH– >N– =N–

–2.97 –1.14 –1.64 –0.1 –2.08 –2.21 –2.03 –1.01 –4.36 –0.23 12.63 2.67 5.02 4.92 15.33 5.33 3.86 6.91 2.74 0.91 1.93

–0.65 –1.34 –4.11 –5.92 –0.76 –1.52 –3.06 –3.69 –2.01 –1.73 14.94 3.22 3.8 3.39 23.4 6.95 6.83 7.57 5.38 0.41 3.24

Functional Group –C≡N –NO2 –F –Cl –Br –I –SH –S– –CH2-(ring) –HC< (ring) =CH– (ring) >C< (ring) =C< (ring) –O– (ring) –OH (ring) >C=O –NH– (ring) >N– (ring) =N– (ring) –S– (ring) –CO–O–CO– (anidride)

φi

φj

5.84 7.41 2.84 7.33 7.66 12.17 –1.77 –0.84 –2.49 –0.26 –1.4 –2.52 –0.39 4.13 10.78 5.38 8.14 14.67 6.03 –0.6 10.6

7.31 10.33 3.74 9.65 6.17 11.06 0.79 –0.97 –1.91 –1.57 –1.11 –6.62 –1.13 8.92 12.92 5.73 12.61 0.62 4.71 2.6 11.49

Example 1.3: Use eqs (1.5) and (1.6) and calculate the FP of the following isomers: (a) 2-Formylbenzoic acid; NBP = 561 K [43] COOH CHO

1.2 Predictive methods of the FP for pure compounds

11

(b) 4-Carboxybenzaldehyde; NBP = 702 K [43] COOH

CHO

Answer: According to Table 1.3 and eqs (1.5) and (1.6), these compounds have one –CHO (aldehyde), one –COOH (acid), four =CH– (ring), and two =C< (ring) functional groups, which provide the FP of these compounds as follows: (a) Equation (1.5): FP = 12.14 + 0.73 ðNBPÞ+1 ð−CHOÞ+ 1ð−COOHÞ + 4ð=CH− ðringÞÞ + 2ð=C < ðringÞÞ = 12.14 + 0.73 ð702Þ + 1ð4.92Þ + 1ð15.33Þ + 4ð − 1.4Þ + 2ð − 0.39Þ = 538.47 K Equation (1.6): FP = 11.07 + 0.72 ðNBPÞ + 1ð−CHOÞ+1ð−COOHÞ+4ð=CH− ðringÞÞ+2ð=C < ðringÞÞ = 11.07 + 0.72ð702Þ + 1ð3.39Þ + 1ð23.4Þ + 4ð − 1.11Þ + 2ð − 1.13Þ = 536.6 K (b) Equation (1.5): FP = 12.14 + 0.73 ðNBPÞ + 1ð−CHOÞ + 1ð−COOHÞ + 4ð=CH− ðringÞÞ + 2ð=C < ðringÞÞ = 12.14 + 0.73ð561Þ + 1ð4.92Þ + 1ð15.33Þ + 4ð − 1.4Þ + 2ð − 0.39Þ = 435.54 K Equation (1.6): FP = 11.07 + 0.72 ðNBPÞ + 1ð−CHOÞ + 1ð−COOHÞ + 4ð=CH− ðringÞÞ + 2ð=C < ðringÞÞ = 11.07 + 0.72ð561Þ + 1ð3.39Þ + 1ð23.4Þ + 4ð − 1.11Þ + 2ð − 1.13Þ = 435.08 K The experimental data of the FPs of 4-carboxybenzaldehyde and 2-formylbenzoic acid are 552 and 440 K [43], respectively. For eq. (1.5), the resultant errors for 4-carboxybenzaldehyde and 2-formylbenzoic acid are 12 and 4 K, respectively. For eq. (1.6), the resultant deviations for 4-carboxybenzaldehyde and 2-formylbenzoic acid are 13 and 5 K, respectively.

1.2.1.5 Non-linear model for the prediction of FP using NBP Serat et al. [48] used a large dataset of 1,660 components from different material classes to find the following equation, which has good extrapolation ability and correlative power:

12

1 Flash Point

0

P

P

P

1

B 0.33156 + N i Ci + Mj Dj + Ek Ok C B C i j k B C FP = 2.9513 × NBP × B  C B  C  P P P @  A 1 + 0.33156 + Ni Ci + Mj Dj + Ek Ok    i j k

(1:7)

where the NBP is in K; Ni, Mj, and Ek are the numbers of occurrences of individual group contributions; Ci is the first-order group contribution of type i; Dj is the second-order group contribution of type j, and Ok is the third-order group contributions of the type k. All parameters except NBP are dimensionless. Tables 1.4–1.6 give the values of Ci, Dj, and Ok. Example 1.4: Use eq. (1.7) and calculate the FPs of the following organic compounds: (a) o-Terphenyl; NBP = 609.15 K [43]

(b) Trilactic acid; NBP = 619 K [43] O O

O

O

HO O

OH

(c) N,N’-Di-tert-butylethylenediamine; NBP = 462.15 K [43]

NH

NH

(d) Methylcyclopentadiene Dimer; NBP = 473 K [43]

1.2 Predictive methods of the FP for pure compounds

13

(e) Pentafluoroethyl Trifluorovinyl Ether; NBP = 281.15 K [43] F F

O

F

F

F F

F

F

Answer: The use of eq. (1.7) and Tables 1.4–1.6 can predict the FP for each part as follows: (a)

First-order group contributions Groups

P

Ci

Ni

Ni C i

i

aCH aC except as above

–0.00050 0.00290

14 4

14 × (–0.00050) 4 × (0.00290)

Second-order group contributions Groups

P

Dj

Mj

Mj Dj

j

AROMRINGs1s2

–0.00269

1

1 × (–0.00269)

Third-order group contributions Groups

P

Ok

Ek

Ek Ok

k

–0.00598

aC-aC (different rings) 2 P P P Ni Ci + Mj Dj + Ek Ok = –0.01008 i

j

2 × (–0.00598)

k

0

P

P

P

1

B 0.33156 + N i Ci + Mj Dj + Ek Ok C B C i j k B C FP = 2.9513 × NBP × B  C B  C  P P P @  A 1 + 0.33156 + Ni Ci + Mj Dj + Ek Ok    i j k

= 2.9513 × 609.15 ×

The measured FP is 436 K [43].

0.33156 − 0.01008

!

1 +j0.33156 − 0.01008j

= 437.36 K

14

1 Flash Point

(b) First-order group contributions Groups

P

Ci

Ni

Ni C i

i

CH3 CH OH COOH CHCOO

3 1 1 1 2

0.00403 –0.00751 0.02297 0.02409 –0.00572

3 × (0.00403) 1 × (–0.00751) 1 × (0.02297) 1 × (0.02409) 2 × (–0.00572)

Second-order group contributions Groups

P

Dj

Mj

Mj Dj

j

CHCOOH or CCOOH OH–CHn–COO (n in 0..2)

1

0.00524

1 × (0.00524)

1

0.00133

1 × (0.00133)

Third-order group contributions Groups

P

Ok

Ek

Ek Ok

k

– P

Ni Ci +

i

P j

Mj Dj +

P

– Ek Ok = 0.04679

–

–

k

0

P

P

1

P

N i Ci + Mj Dj + Ek Ok C B 0.33156 + B C i j k B C FP = 2.9513 × NBP × B  C  P P P @  A 1 + 0.33156 + Ni Ci + Mj Dj + Ek Ok    i j k


 0.33156 − 0.04679 = 501.47 K = 2.9513 × 619 × 1 + j0.33156 − 0.04679j The measured FP is 499 K [43]. (c) First-order group contributions Groups

Ni

Ci

P

Ni C i

i

CH3 C CH2NH

6 2 2

0.00403 –0.02247 0.00097

6 × (0.00403) 2 × (–0.02247) 2 × (0.00097)

1.2 Predictive methods of the FP for pure compounds

(continued)

Second-order group contributions Groups

P

Dj

Mj

Mj Dj

j

(CH3)3C

2

0.00501

1×(0.00501)

Third-order group contributions Groups

P

Ok

Ek

Ek Ok

k

– P

Ni Ci +

i

P

Mj Dj +

j

P

– Ek Ok = –0.0088

–

–

k

0

P

P

1

P

N i Ci + Mj Dj + Ek Ok C B 0.33156 + B C i j k B  C FP = 2.9513 × NBP × B  C  P P P @  A 1 + 0.33156 + Ni Ci + Mj Dj + Ek Ok    i j k


 0.33156 − 0.0088 = 332.82 K = 2.9513 × 462.15 × 1 + j0.33156 − 0.0088j The measured FP is 336 K [43]. (d) First-order group contributions Groups

Ni

Ci

P

Ni C i

i

CH3 CH2 (cyclic) CH (cyclic) CH=C (cyclic)

2 2 4 2

0.00403 –0.00068 –0.00150 –0.00334

2 × (0.00403) 2 × (–0.00068) 4 × (–0.00150) 2 × (–0.00334)

Second-order group contributions Groups

Mj

Dj

P

Mj Dj

j

(CHn=C)cyc–CH3 (n in 0..1)

4

–0.00578

4 × (–0.00578)

Third-order group contributions Groups

Ek

Ok

P

Ek Ok

k

CH 4 P P multiringP Ni Ci + Mj Dj + Ek Ok = –0.01026 i

j

k

0.00182

4×(0.00182)

15

16

1 Flash Point

0

P

P

1

P

N i Ci + Mj Dj + Ek Ok C B 0.33156 + B C i j k B C FP = 2.9513 × NBP × B  C  P P P @  A 1 + 0.33156 + Ni Ci + Mj Dj + Ek Ok    i j k
= 2.9513 × 473 ×

 0.33156 − 0.01026 = 339.47 K 1 + j0.33156 − 0.01026j

The measured FP is 326.15 K [43]. (e) First-order group contributions Groups

P

Ci

Ni

Ni C i

i

C=C C–O CF3 –F except as above

–0.03340 –0.02117 0.03262 0.02605

1 1 1 5

1 × (–0.03340) 1 × (–0.02117) 1 × (0.03262) 5 × (0.02605)

Second-order group contributions Groups

P

Dj

Mj

Mj Dj

j

CHm–O–CHn=CHp (m,n,p in 0..3)

1

0.00065

1 × (0.00065)

Third-order group contributions Groups

P

Ok

Ek

Ek Ok

k

–0.00598

aC-aC (different rings) 2 P P P Ni Ci + Mj Dj + Ek Ok = 0.0878 i

j

2 × (–0.00598)

k

0

P

P

P

1

N i Ci + Mj Dj + Ek Ok C B 0.33156 + B C i j k  C FP = 2.9513 × NBP × B B  C  P P P @  A 1 + 0.33156 + Ni Ci + Mj Dj + Ek Ok    i j k
= 2.9513 × 281.15 × The measured FP is 233 K [43].

 0.33156 − 0.0878 = 245.15 K 1 + j0.33156 − 0.0878j

Table 1.4: List of the first-order groups, their contributions to the FP, and their number of occurrences in the molecules. Groups

aC fused with aromatic ring aC fused with nonaromatic ring aC except as above aN in aromatic ring aC–CH3 aC–CH2

Ci

1090 793 236 81 107 67 46 18 8 3 1 13 6 440

Description

Examples

0.00403 –0.00120 –0.00751 –0.02247 0.00226 –0.00210 –0.00593 –0.01332 –0.03340 –0.00174 –0.00525 0.00078 0.00094 –0.00050

–CH3 (methyl group) except as below –CH2– except as below >CH– except as below >C< except as below CH2=CH– except as below –CH=CH– except as below CH2=C< except as below –CH=C< except as below >C=C< except as below CH2=C=CH– CH2=C=C< except as below CH≡C– except as below –C≡C– except as below =CH– aromatic carbon

35

–0.00631

44

–0.00089

42 18 98 94

0.00290 0.00650 0.00105 –0.00360

fused aromatic carbon atom with an aromatic ring fused aromatic carbon atom with a non-aromatic ring aromatic carbon except as above aromatic nitrogen atom aromatic carbon connected to CH3 aromatic carbon connected to CH2

Propane (2); Isobutane(3) n-Butane (2); n-Hexane(4) 2,3-Butanediol (2); 3-Ethylpentane(1) Neopentane (1); 2,2,3,3-Tetramethylhexane(2) 1-Nonene (1); 1,5-Hexadiene(2) trans-2-Eicosene(1); cis,trans-2,4-Hexadiene (2) Isobutane(1); 2,5-DimethyL-1,5-Hexadiene(2) 2-Methyl-2-Butene (1); 2,5-DimethyL-2,4-Hexadiene (2) Hexachloro-1,3-Butadiene (2); Chlorotrifluoroethylene(1) 1,2-Butadiene (1); 1,2-Hexadiene(1) 3-Methyl-1,2-Butadiene (1); 1-Octyne(1); Propargyl chloride(1) 1-Pentene-3-yne (1); pent‐4‐en‐2‐yn‐1‐ol (1) Monomethyl terephthalate (4); 2-(alphaMethylbenzyloxy)-1-Propanol (5) 8-Hydroxyquinoline (2); 2,6-Naphthalenedicarboxylic acid(2) 4,6-Dimethyldibenzothiophene(4); 2-Mercaptobenzothiazole(2) Diphenylacetylene(1); m-Terphenyl(4) Pyrazine(2); Melamine(3) Quinaldine(1); 4-Chloro-o-Xylene(2) Ethylbenzene(1); o-Diethylbenzene(2)

1.2 Predictive methods of the FP for pure compounds

CH3 CH2 CH C CH2=CH CH=CH CH2=C CH=C C=C CH2=C=CH CH2=C=C CH≡C C≡C aCH

Frequency of occurrence

(continued)

17

18

Table 1.4 (continued) Frequency of occurrence

Ci

aC–CH

33

aC–C

Description

Examples

–0.00127

aromatic carbon connected to CH

28

–0.02136

aromatic carbon connected to C

aC–CH=CH2

13

–0.00320

aC–CH=CH aC–C=CH2 aC–C≡CH aC–C≡C OH

7 5 1 1 175

–0.01140 –0.00587 0.00075 –0.01010 0.02297

aC–OH COOH

48 83

0.01841 0.02409

aC-COOH

19

0.01795

CH3CO CH2CO CHCO aC-CO

29 20 3 9

0.01497 –0.00073 0.00229 0.00619

CHO aC–CHO

45 13

0.01328 0.00720

aromatic carbon connected to CH=CH2 aromatic carbon connected to CH=CH aromatic carbon connected to C=CH2 aromatic carbon connected to C≡CH aromatic carbon connected to C≡C –OH (for aliphatic chains) except as below aromatic carbon connected to OH carboxyl group (C(=O)OH) except as below aromatic carbon connected to carboxyl group carbonyl connected to CH3 carbonyl connected to CH2 carbonyl connected to CH carbonyl connected to aromatic carbon aldehyde group except as below aldehyde connected to aromatic carbon

1,3,5-Triisopropylbenzene(3); 1,2,4,5Tetraisopropylbenzene(4) 1,4-Di-tert-Butylbenzene(2); 1,3,5-Tri-tert-Butylbenzene (3) m-Divinylbenzene (2); p-Divinylbenzene(2) cis-1-Propenylbenzene(1); trans-1-Propenylbenzene(1) p-Isopropenylstyrene(1); p-Isopropenyl phenol(1) Ethynylbenzene(1); 1,2‐diethynylbenzene (2) Diphenylacetylene(1); (prop‐1‐yn‐1‐yl)benzene (1) 1,2-Propylene glycol(2); Neopentyl glycol(2) Bisphenol a(2); o-Cresol(1) Pimaric acid(1); Isopimaric acid(1) o-Chlorobenzoic acid(1); Trimellitic acid(3) Acetylacetone(2); Acetone(1) 3-Pentanone(1); Chloroacetyl chloride(1) Diisopropyl Ketone(1); Dichloroacetyl chloride(1) Benzophenone(1); Acetophenone(1) Glutaraldehyde(2); Furfural(1) Terephthaldehyde(2); Benzaldehyde(1)

1 Flash Point

Groups

35 41 11 2

0.01246 0.00160 –0.00572 –0.01985

ester connected to CH3 ester connected to CH2 ester connected to CH ester connected to C

HCOO aC-COO

17 35

0.01206 0.00325

aC-OOC

2

0.05710

COO except as above

36

0.01116

CH3O CH2O CH–O C–O aC–O CH2NH2 CHNH2 CNH2 CH3NH CH2NH CHNH

41 71 7 5 14 39 7 2 3 17 1

0.00739 0.00014 –0.00265 –0.02117 0.00386 0.01022 0.00663 –0.00950 0.01821 0.00097 0.00040

–O– connected to CH3 –O– connected to CH2 –O– connected to CH –O– connected to C –O– connected to C CH2 connected to –NH2 CH connected to –NH2 C connected to –NH2 CH3 connected to >NH CH2 connected to >NH CH connected to >NH

CH3N

8

–0.00177

CH3 connected to >N–

CH2N

9

–0.01328

CH2 connected to >N–

formate group >C=O of ether group connected to an aromatic carbon -O- of ether group connected to an aromatic carbon ester group except as above

Methyl acetate(1); n-Propyl acetate(1) Ethyl n-Butyrate(1); Propionic anhydride(1) n-Propyl isobutyrate(1); Trilactic acid(2) Ethyl trimethyl acetate(1); Hydroxypivalyl hydroxypivalate(1) Benzyl formate(1); Methyl formate(1) Dioctyl phthalate(2); Benzyl benzoate(1) Acetylsalicylic acid (1); Phenyl acetate(1) n-Propyl methacrylate(1); Dimethyl-1,4-cyclohexane dicarboxylate(2) Methyl ethyl ether(1); Methylal(2) Acetal(2); Isopropyl isobutyl ether(1) Diisopropyl ether(1); Bis(difluoromethyl)ether(1) Di-t-butyl peroxide(2); Di-tert-butyl ether(1) trans-3,5-Dimethoxystilbene(2); Anethole(1) n-Propylamine(1); Isobutylamine(1) Isopropylamine(1); L-phenylalanine(1) 2-Methyl-2-aminobutane(1); tert-Butylamine(1) Dimethylamine(1); N-Methylcyclohexylamine(1) Triethylenetetramine(2); Di-n-Butylamine(1) Diisopropylamine(1); N‐[1‐(isopropylamino)ethyl] hydroxylamine (2) Dimethylethanolamine(1); Tetramethylethylenediamine (2) Triethylamine(1); Triallylamine(1)

1.2 Predictive methods of the FP for pure compounds

CH3COO CH2COO CHCOO CCOO

(continued)

19

20

Table 1.4 (continued) Frequency of occurrence

Ci

aC–NH2 aC–NH

24 9

Description

Examples

0.01087 0.00160

aromatic carbon connected to NH2 aromatic carbon connected to NH

4

–0.00839

aromatic carbon connected to >N–

9 24 4 1 2 15 2

0.02780 0.00795 –0.01075 –0.01586 –0.00223 0.03063 0.01557

–NH2 except as above >CH2 connected to –C≡N >CH– connected to –C≡N >C< connected to –C≡N aromatic carbon connected to –C≡N –C≡N except as above –N=C=O connected to >CH2

aC-NCO

6

0.01119

CH2NO2 CHNO2 aC–NO2

3 1 29

0.01495 0.01141 0.00998

1 8 1 1

0.04834 0.03077 0.03288 0.01404

aromatic carbon connected to –N=C=O –NO2 (nitro group) connected to >CH2 –NO2 (nitro group) connected to >CH– –NO2 (nitro group) connected to aromatic carbon –NO2 (nitro group) except as above –CO–NH2 (amide group) –CO–NH– connected to CH3 –CO–N< connected to two CH3

Melamine(3); o-Phenylenediamine(2) N,N’-DIphenyl-p-Phenylenediamine(2); p-Nitrodiphenylamine(1) N,N-Dimethylaniline(1); p-Dimethylaminobenzaldehyde (1) Cyclopentylamine(1); Dicyandiamide(1) Succinonitrile(2); Hydracrylonitrile(1) Isobutyronitrile(1); Acrolein cyanohydrin(1) Acetone cyanohydrin(1); 2,2‐dihydroxypropanenitrile(1) Nicotinonitrile(1); Benzonitrile(1) Cyanogen(2); Cyanogen chloride(1) n-Butyl isocyanate(1); 1,6-Hexamethylene diisocyanate (2) 2,6-Toluene diisocyanate(2); 3,4-Dichlorophenyl isocyanate(1) Nitroethane(1); 1-Nitropropane(1) 2-Nitropropane(1); 2‐nitropentane (1) m-Dinitrobenzene(2); o-Nitrodiphenylamine(1)

aC–N NH2 except as above CH2CN CHCN CCN aC–CN CN except as above CH2NCO

NO2 except as above CONH2 CONHCH3 CON(CH3)2

Nitromethane(1); Hexanamide(1); Trifluoroacetamide(1) N-methylacetamide(1); N,N-Dimethylacetamide(1);

1 Flash Point

Groups

1

0.02264

aC–NH(CO)H

1

0.02115

aC–NHCO

3

–0.01224

NHCONH NH2CONH

1 1

0.02663 0.05138

CH2Cl CHCl CCl CHCl2 CCl3 CH2F CHF2 CF2

41 8 1 5 5 6 6 2

0.01481 0.00357 –0.01124 0.03679 0.05920 0.01619 0.02531 0.00761

CF3

10

0.03262

–C connected to three F

CCl2F CClF2 aC–Cl aC–F aC–I

1 1 32 3 1

0.04787 0.03432 0.01161 0.00561 0.00764

–C connected to both Cl2 and F –C connected to both Cl and F2 aromatic carbon connected to Cl aromatic carbon connected to F aromatic carbon connected to I

Benzamide(1); Formanilide(1); Acetaminophen(1); Acetanilide(1) 1,3-Dimethyl urea(1); Monomethyl urea(1); 1,4-Dichloro-trans-2-Butene(2); Ethyl chloride 2,3-Dichlorobutane(2); 2,3-Dichloro-1-Propanol(1) tert-Butyl chloride(1); Dichloroacetaldehyde(1); 1,1-Dichloroethane(1) Trichloroacetaldehyde(1); 1,1,1-Trichloroethane(1) 1,2-Difluoroethane(2); 1,1,1-Trichlorofluoroethane Difluoromethyl methyl ether(1); 1,1-Difluoroethane(1) Pentafluoroethyl methyl ether(1); 1,1,2,2,3Pentafluoropropane(1) Pentafluoroethyl trifluorovinyl ether(1); Trifluoroacetamide 1,1-Dichloro-1-fluoroethane(1); 1-Chloro-1,1-Difluoroethane(1); o-Dichlorobenzene(2); 3-Chloro-o-xylene(1) Pentafluorophenol(5); Fluorobenzene(1) Iodobenzene(1); (continued)

1.2 Predictive methods of the FP for pure compounds

aromatic carbon connected to –CO– NH2 (amide group) aromatic carbon connected to –NH–C (=O)H aromatic carbon connected to –NH– C=O >C=O connected to two >NH >C=O connected to both >NH and –NH2 –Cl connected to >CH2 –Cl connected to >CH–Cl connected to C –CH connected to two Cl –C connected to three Cl CH2 connected to F –CH connected to two F –C connected to two F

aC–CONH2

21

22

Table 1.4 (continued) Ci

aC–Br –I except as above –Br except as above –F except as above

5 6 15 18

–Cl except as above CHNOH CNOH OCH2CH2OH OCHCH2OH OCH2CHOH –O–OH CH2SH CHSH CSH aC–SH –SH except as above CH3S CH2S CHS CS

Description

Examples

0.00730 0.02396 0.01665 0.02605

aromatic carbon connected to Br iodine except as above bromine except as above fluorine except as above

49 1 1 27 5

0.01862 0.03838 0.02472 0.01419 0.00830

chlorine except as above >C=N–OH aldoxime (oxime) –HC=N–OH ketoxime (oxime) –O–CH2–CH2–OH –O–CH–CH2–OH

19 13 21 2 3 1 4 11 20 2 3

0.00519 0.03754 –0.00236 –0.00586 –0.01505 0.00698 0.02324 0.00430 –0.00592 –0.01849 –0.02504

m-Dibromobenzene(2); p-Bromotoluene(1) Diiodomethane(2); n-Butyl iodide(2) 1,1,2,2-Tetrabromoethane(4); 1,2-Dibromododecane(2) p-Chlorobenzotrifluoride(3); 2-Chloro-1,1Difluoroethylene(2) cis-1,2-Dichloroethylene(2); Methyl chloride(1) Acetaldoxime(1); 2-Butoxime(1); TRIETHYLENE GLYCOL(2); Diethylene glycol(1) Propylene glycol 2-tert-Butyl ether(1); 2-Methoxy propanol-1(1) Tripropylene glycol(2); Dipropylene glycol(1) Peracetic acid(1); Methyl hydroperoxide(1) Methyl-3-Mercaptopropionate(1); 1,2-Ethanedithiol(2) sec-Butyl mercaptan(1); Isopropyl mercaptan(1) tert-Nonyl mercaptan(1); tert-Octyl mercaptan(1) Phenyl mercaptan(1); Methyl mercaptan(1); Cyclohexyl mercaptan(1) Dimethyl disulfide(2); Methyl ethyl sulfide(1) Diethyl disulfide(2); Ethylthioethanol(1) Diisopropyl sulfide(1); Malathion(1) DI-tert-butyl disulfide(2); DI-tert-butyl sulfide(1)

–O–CH2–CH–OH –O–OH –CH2–SH –CH–SH –C–SH aromatic carbon connected to –SH –SH except as above –CH3 connected to –S– (thioether) –CH2 connected to –S– (thioether) –CH connected to –S– (thioether) –C– connected to –S– (thioether)

1 Flash Point

Frequency of occurrence

Groups

2

0.00667

SO SO2 aC-P

1 2 1

0.01760 –0.00096 –0.02074

C2H3O

7

0.01584

C2H2O

1

0.01065

CH2 (cyclic)

206

–0.00068

aromatic carbon connected to –S– (aromatic thioether) –S(=O)– (sulfoxide) –S(=O)2– (sulfone) non-cyclic phosphorus connected to aromatic carbon –CH2–CH–O- cyclic ether with a three-atom ring (epoxide) -CH2-C-O- cyclic ether with a threeatom ring (epoxide) –CH2– in a ring

CH (cyclic)

131

–0.00150

>CH– in a ring

C (cyclic) CH=CH (cyclic) CH=C (cyclic)

31 47 42

–0.00404 –0.00122 –0.00334

>C< in a ring –CH=CH– in a ring –CH=C< in a ring

C=C (cyclic) CH2=C (cyclic)

6 6

–0.00199 0.00158

NH (cyclic) N (cyclic) CH=N (cyclic) C=N (cyclic)

18 14 2 1

0.01256 –0.01117 0.01224 0.01402

>C=C< in a ring carbon in a ring double bonded to a sidechain carbon >NH in a ring >N– in a ring –CH=N– in a ring >C=N– in a ring

Methyl phenyl sulfide(1); Diphenyl disulfide(2) Dimethyl sulfoxide(1); DI-n-Propyl sulfone(1); DI-n-Butyl sulfone(1) Triphenylphosphine(1); alpha-Epichlorohydrin(1); Allyl glycidyl ether(1) 1,2-Epoxy-2-Methylpropane(1); Bicyclohexyl(1); 1-cis-2-trans-3-Trimethylcyclopentane (2) 1,2,3,4-Tetramethylcyclohexane(4); 1-cis-2-trans-4Trimethylcyclopentane(3) Tetramantane(2); 1,1-Dimethylcyclopentane(1) 1,3-Cyclohexadiene(2); Cyclohexene(1) Methylcyclopentadiene dimer(2); 1-Methylcyclopentene (1) Tetrachlorothiophene(2); Palustric acid(1) beta-Pinene(1); Isoagatholal(1) Piperazine(2); Saccharin(1) Triethylenediamine(2); N-Methyl-2-Pyrrolidone(1) Pyrazole(1); Isoxazole(1) 2-Mercaptobenzothiazole(1); (continued)

1.2 Predictive methods of the FP for pure compounds

aC–S–

23

24

Table 1.4 (continued) Ci

O (cyclic) CO (cyclic) S (cyclic) SO2 (cyclic) >NH –O– –S– >CO

46 26 32 3 4 4 1 4

SiHO SiO SiH2 SiH Si Ccyclic=N–

Description

Examples

0.01169 0.01082 0.00670 0.02174 0.01765 –0.04086 0.46297 0.01025

–O– in a ring >C=O in a ring –S– in a ring –S(=O)2– (sulfone) cyclic >NH except as above –O– except as above –S– except as above >C=O except as above

1 27

–0.00787 –0.00924

>SiH– connected to –O– >Si< connected to –O–

3 4 24 1

0.00643 –0.00447 –0.02604 –0.00825

Ccyclic=CH–

1

–0.01681

Ccyclic=C

2

–0.01429

N=N C=NH >C=S

1 1 1

–0.00646 –0.02386 –0.00468

–SiH2– primary silicon group >SiH– secondary silicon group >Si< tertiary silicon group carbon in a ring double bonded to a sidechain nitrogen carbon in a ring double bonded to a sidechain =CH– carbon in a ring double bonded to a sidechain =C< –N=N– in not ring (azo) >C=NH in not ring >C=S in not ring

1,4-Dioxane(2); alpha-Tocopherol(1) 2-Pyrrolidone(1); Anthraquinone(2) 2-Mercaptobenzothiazole(1); Thiophene(1) 3-Methyl sulfolane(1); Sulfolane(1) Hexamethyldisilazane(1); Dicyandiamide(1) Dicumyl peroxide(2); Divinyl ether(1) Dicyclohexyl sulfide(1); 2-Hydroxypropyl acrylate(1); 2-Hydroxyethyl methacrylate(1) Trimethoxysilane(1); BIS[3-(triethoxysilyl)propyl]disulfide(2); Dimethyldimethoxysilane(1) Dichlorosilane(1); Dimethyl silane(1) Trichlorosilane(1); Trimethyl silane(1) Hexamethyldisilazane(2); Hexamethyldisiloxane(1) Cyclohexanone oxime(1); 5-Ethylidene-2-norbornene(1); Neoabietic acid(1); Terpinolene(1) p-Aminoazobenzene(1); Dicyandiamide(1); pentan‐3‐imine (1) N-Methylthiopyrrolidone(1);

1 Flash Point

Frequency of occurrence

Groups

HCONH

2

0.02699

SiH3 CH=C=CH

1 1

0.00735 –0.00248

OP(=S)O

1

–0.01856

–NH–C(=O)H not connected to an aromatic atom (amide) –SiH3 silane group –CH=C=CH– two cumulated double bonds –P(=S)(–O–)–O–

tert-Butylformamide(1); N-Methylformamide(1) Methyl silane (1); propylsilane(1) 2,3-Pentadiene(1); hepta‐3,4‐diene (1) Malathion(1); O,O‐dimethyl ethylsulfanylphosphonothioate (1)

1.2 Predictive methods of the FP for pure compounds

25

Mj

(CH3)2CH (CH3)3C CH(CH3)CH(CH3)

123 59 10

–0.00140 0.00501 –0.00468

CH(CH3)C(CH3)2

5

–0.00209

C(CH3)2C(CH3)2

4

0.00938

13

Description

Examples Isopentane (1); Isobutane(2) Neopentane(2); 2,2-Dimethylbutane(1) 2,3,4-Trimethylpentane(2); 2,3-Dimethylbutane(1) 2,3,3,4-Tetramethylpentane(2); 2,2,3-Trimethylbutane(1)

–0.00017

CH connected to two methyl group C connected to three methyl group CH–CH present in the longest chain of a hydrocarbon bond with two methyl group neighbors, one on each side CH–C present in the longest chain of a hydrocarbon bond with three methyl group neighbors CH–CH present in the longest chain of a hydrocarbon bond with four methyl group neighbors, two on each side Two conjugated double bonds in a chain

93

–0.00040

–CH3 connected to sp2 carbon in a chain

cis-2-Butene(2); Propylene(1)

117

–0.00093

>CH2 connected to sp2 carbon in a chain

Linoleic acid(4); Triallylamine(3)

20

0.00303

Stigmasterol(2); sec-Butenyl glycol ether(1)

10 18

–0.00183 –0.01645

CH3COCH or CH3COC

3

–0.01040

>CH– or >C< connected to sp2 carbon in a chain CH or C connected to HC=O (aldehyde group) –CH2–C(=O)–CH3 (ketone connected to both –CH3 and CH2) ketone connected to both –CH3 and CH or C

CHCOOH or CCOOH

13

0.00524

CHn=CHm–CHp=CHk (k,m,n,p in 0..2) CH3–CHm=CHn (m,n in 0..2) CH2–CHm=CHn (m,n in 0..2) CHp–CHm=CHn (m,n in 0..2; p in 0..1) CHCHO or CCHO CH3COCH2

carboxylic acid connected to CH or C

2,2,3,3-Tetramethylpentane(1); 2,3-Dimethyl-2,3-diphenylbutane(1) Isoprene(1); Chloroprene(1)

2-Methylheptanal(1); 2-Methylpropanal(1) Acetylacetone(2); Methyl ethyl ketone(1) 3-Methyl-2-pentanone (1); Methyl isopropyl ketone(1) Dichloroacetic acid(1); Trichloroacetic acid(1)

1 Flash Point

Frequency of occurrence

Groups

26

Table 1.5: List of the second-order groups, their contributions to the FP, and their number of occurrences in the molecules.

8

0.00045

CH3–C(=O)O–CH or CH3–C(=O)O–C

Ethylidene diacetate(2); Isopropyl acetate(1)

4 40 9 3 4

0.00286 0.00047 0.00942 0.01351 0.00133

Acetic anhydride(1); Isobutyric anhydride(1) Tripropylene glycol(2); 2-Butanol(1) Triacetone alcohol(2); trans-1,8-Terpin(1) Acetone cyanohydrin(1); Lactonitrile(1) Methyl lactate(1); Methyl glycolate(1)

12

–0.00549

6

0.00415

2

0.01149

5

–0.00317

3

0.00875

1

0.00393

–O– connected to two ketone secondary alcohol tertiary alcohol N≡C–CH–OH or N≡C–C–OH OH–CH2–C(=O)O– or OH–CH–C(=O)O– or OH–C–C(=O)O– carbon-carbon bond with two –OH neighbors, one on each side carbon-carbon bond with –OH and nitrogen neighbors, one on each side carbon-carbon bond with two –NH2 neighbors, one on each side carbon-carbon bond with –NH2 and >NH neighbors, one on each side carbon connected to nitrogen and carboxylic acid carbon connected to two carboxyl groups

2

–0.03016

Succinic acid(1); Tartaric acid(1)

5

–0.00421

1

0.01696

carbon-carbon bond with two carboxyl groups, one on each side carbon connected to –OH (hydroxy group) and carboxylic acid CH3–O– connected to carbon-carboxyl group bond

Tartaric acid(1); 1,4-Benzenedicarboxylic acid,bis(2,3-dihydroxypropyl)ester(2) Monoethanolamine(1); Diethanolamine(2) Ethylenediamine(1); 1,2-Propanediamine(1) Tetraethylenepentamine(2); N-Aminoethyl ethanolamine(1) L-Phenylalanine(1); Lysine(1) Malonic acid(1); 2‐ethylpropanedioic acid (1)

Tartaric acid(2); Hydroxycaproic acid(1) Methoxyacetic acid(1);

1.2 Predictive methods of the FP for pure compounds

CH3COOCH or CH3COOC CO–O–CO CHOH COH NCCHOH or NCCOH OH–CHn–COO (n in 0..2) CHm(OH)CHn(OH) (m,n in 0..2) CHm(OH)CHn(NHp) (m,n,p in 0..2) CHm(NH2)CHn(NH2) (m,n in 0..2) CHm(NH)CHn(NH2) (m,n in 1..2) CHm(NHn)–COOH (m,n in 0..2) HOOC–CHn–COOH (n in 1..2) HOOC–CHn–CHm– COOH (n, m in 1..2) HO–CHn–COOH (n in 1..2) CH3–O–CHn–COOH (n in 1..2)

(continued)

27

28

Table 1.5 (continued) Mj

HS–CH–COOH

1

0.02342

HS–CHn–CHm–COOH (n, m in 1..2) NC–CHn–CHm–CN (n, m in 1..2) OH–CHn–CHm–CN (n, m in 1..2) HS–CHn–CHm–SH (n, m in 1..2) COO–CHn–CHm–OOC (n, m in 1..2) OOC–CHn–CHm–COO (n, m in 1..2) NC–CHn–COO (n in 1..,2) COCHnCOO (n in 1..2)

1

–0.02151

1

–0.00638

1

0.01809

1

0.01928

4

–0.00315

2

0.00234

2

0.01505

4

0.01106

5

0.00065

8

–0.00706

1

0.02000

14

0.00217

CHm–O–CHn=CHp (m,n,p in 0..3) CHm=CHn–F (m,n in 0..2) CHm=CHn–Br (m,n in 0..2) CHm=CHn–Cl (m,n in 0..2)

Description

Examples

HS-carbon bond connected to carboxyl group bond HS-carbon bond connected to carboncarboxyl group bond nitrile-carbon bond connected to carbonnitrile bond … .. HO-carbon bond connected to carbon-nitrile bond carbon-carbon bond connected to two –SH, one on each side carbon-carbon bond connected to two carboxyl groups by –O–, one on each side carbon-carbon bond connected to two carboxyl groups by >C=O, one on each side carbon connected to nitrile and >C=O of carboxyl group carbon connected to >C=O (ketone) and >C=O of ester group carbon-oxygen bond connected to carboncarbon double bond carbon-carbon double bond connected to –F

Thioglycolic acid(1);

carbon-carbon double bond connected to –Br carbon-carbon double bond connected to –Cl

3-Mercaptopropionic acid(1); Succinonitrile(1); 2‐ethylbutanedinitrile(1) Hydracrylonitrile(1); 1,2-Ethanedithiol(1); butane‐2,3‐dithiol(1) Glyceryl triacetate(2); 2-Acetoacetoxy ethyl methacrylate(1) Malathion(1); Diethyl succinate(1) Methyl cyanoacetate(1); Ethyl cyanoacetate (1) 2-Acetoacetoxy ethyl methacrylate(1); t-Butyl acetoacetate(1) Methyl vinyl ether(1); Ethyl vinyl ether(1) Tetrafluoroethylene(4); Chlorotrifluoroethylene(3) Vinyl bromide(1); cis-1,2-Dichloroethylene(2); Vinyl chloride(1)

1 Flash Point

Frequency of occurrence

Groups

8

–0.01372

30

–0.00588

6

0.00244

6

0.01061

3

–0.00971

1

–0.01901

10

0.00128

aC–CHn–OH (n in 1..2)

8

–0.00019

aC–CHn–CN (n in 1..2)

1

–0.02201

aC–CHn–CHO (n in 1..2) aC–CHn–SH (n in 1..2)

1

–0.01711

1

–0.02142

2

–0.00343

2

–0.00489

aC–CHn–COOH (n in 1..2) aC–CHn–OOC–H (n in 1..2)

carbon-carbon double bond connected to –C≡N carbon-carbon double bond connected to –C(=O)–O–CH0,1or 2 carbon-carbon double bond connected to aldehyde group carbon-carbon double bond connected to carboxyl group aromatic carbon-carbon bond connected to halogen elements aromatic carbon-carbon bond connected to nitrogen aromatic carbon-carbon bond connected to –O– aromatic carbon-carbon bond connected to –OH aromatic carbon-carbon bond connected to –C≡N aromatic carbon-carbon bond connected to aldehyde aromatic carbon-carbon bond connected to –SH aromatic carbon-carbon bond connected to –COOH aromatic carbon-carbon bond connected to –O–C(H)=O

Fumaronitrile(2); Acrylonitrile(1) Dibutyl maleate(2); Methyl acrylate(1) Methacrolein(1); Acrolein(1) Itaconic acid(1); Acrylic acid(1) Benzotrichloride(3); Benzyl dichloride(2) Benzylamine(1); 1‐[2‐(aminomethyl)phenyl] methanamine (2) BIS(alpha-Methylbenzyl) Ether(2); Benzyl ethyl ether(1) m-Tolualcohol(1); Benzyl alcohol(1) Phenylacetonitrile(1); 2-Phenylpropionaldehyde(1); Benzyl mercaptan(1); [4‐(sulfanylmethyl) phenyl]methanethiol(2) 4-Methoxyphenylacetic acid(1); Ibuprofen(1) alpha-Methylbenzyl alcohol formate(1); Benzyl formate(1)

1.2 Predictive methods of the FP for pure compounds

CHm=CHn–CN (m,n in 0..2) CHn=CHm–COO–CHp (m,n,p in 0..3) CHm=CHn–CHO (m,n in 0..2) CHm=CHn–COOH (m,n in 0..2) aC–CHn–X (n in 1..2) X: Halogen aC–CHn–NHm (n in 1..2; m in 0..2) aC–CHn–O– (n in 1..2)

(continued)

29

30

Table 1.5 (continued) Mj

2

–0.00105

1

–0.00113

16

–0.00859

aC–C(CH3)3

8

0.00461

aC–CF3

5

–0.04778

(CHn=C)(cyc)–CHO (n in 0..1) (CHn=C)cyc–CH3 (n in 0..1) (CHn=C)cyc–CH2 (n in 0..1) (CHn=C)cyc–Cl (n in 0..1) CHcyc–CH3

1

–0.00120

27

CHcyc–CH2

aC–CHn–OOC–H (n in 1..2) aC–CHn–COO (n in 1..2) aC–CH(CH3)2

CHcyc–CH CHcyc–C CHcyc–CH=CHn (n in 1..2)

Description

Examples

Benzyl benzoate(1); Benzyl acetate(1)

–0.00578

aromatic carbon-carbon bond connected to –O–C=O aromatic carbon-carbon bond connected to –C(=O)–O– aromatic carbon-carbon bond connected to two –CH3 aromatic carbon-carbon bond connected to three –CH3 aromatic carbon-carbon bond connected to three –F –CH=C< or >C=C< in a ring connected to –C(H) =O –CH=C< or >C=C< in a ring connected to –CH3

9

–0.00309

–CH=C< or >C=C< in a ring connected to –CH2

Methylcyclopentadiene dimer(2); 2,3Dimethylthiophene(2) Furfuryl alcohol(1); Propenyl cyclohexene(1)

1

–0.01165

–CH=C< or >C=C< in a ring connected to –Cl

Tetrachlorothiophene(4);

44

–0.00451

>CH– in a ring connected to –CH3

25

–0.00448

>CH– in a ring connected to non ring –CH2–

9 5 3

–0.00003 –0.00175 –0.00573

>CH– in a ring connected to non ring >CH– >CH– in a ring connected to non ring >C< >CH– in a ring connected to non ring –CH =CH1or 2

Methylcyclopentane(1); cis-1,2Dimethylcyclopentane(2) trans-1,4-Diethylcyclohexane(2); Ethylcyclopentane(1) Isopropylcyclopentane(1); Sitosterol(1) tert-Butylcyclohexane(1); alpha-Terpineol(1) Vinylnorbornene(1); Vinylcyclohexene(1)

Ethyl phenyl acetate(1); 1,3,5-Triisopropylbenzene(3); 1,2,4,5Tetraisopropylbenzene(4) 1,4-DI-tert-Butylbenzene(2); 1,3,5-TRI-tertButylbenzene(3) Benzotrifluoride(1); 3-Nitrobenzotrifluoride (1) Furfural(1);

1 Flash Point

Frequency of occurrence

Groups

2

0.00059

12 3 2

–0.00677 –0.01616 –0.01302

1 1 3

–0.02513 –0.01969 0.00135

>CH– in a ring connected to non ring >C=CH1or 2 >CH– in a ring connected to –OH >CH– in a ring connected to –NH2 >CH– in a ring connected to non ring NH– carbon bond >CH– in a ring connected to –SH >CH– in a ring connected to –C≡N >CH– in a ring connected to –COOH

CHcyc–CO CHcyc–S– CHcyc–CHO

1 1 2

0.00571 –0.23428 –0.01054

>CH– in a ring connected to –C=O >CH– in a ring connected to –S– >CH– in a ring connected to –C(H)=O

CHcyc–O–

3

–0.00490

>CH– in a ring connected to –O–

1 1 2 25 5

–0.00162 –0.00918 –0.00457 –0.00679 –0.00262

>CH– in a ring connected to –O–C(H)=O >CH– in a ring connected to –C(=O)–O– >CH– in a ring connected to –O–C(H)=O >C< in a ring connected to –CH3 >C< in a ring connected to –CH2–

4 4 8

–0.00278 0.01183 0.00280

>C< in a ring connected to –OH >N– in a ring connected to –CH3 >N– in a ring connected to –CH2–

CHcyc–OOCH CHcyc–COO CHcyc–OOC Ccyc–CH3 Ccyc–CH2 Ccyc–OH >Ncyc–CH3 >Ncyc–CH2

d-Limonene(1); beta-Terpineol(1) Inositol(6); beta-Cholesterol(1) Cyclopentylamine(1); Cyclopropylamine(1) Dicyclohexylamine(2); N-Methylcyclohexylamine(1) Cyclohexyl mercaptan(1); Cyclopropyl cyanide(1); 1,4-Cyclohexanedicarboxylic acid(2); Cyclopropane carboxylic acid(1) Cyclopropanecarboxamide(1); Dicyclohexyl sulfide(2); 1,2,3,6-Tetrahydrobenzaldehyde(1); Cyclohexanecarboxaldehyde(1) Cyclohexyl hydroperoxide(1); Methoxydihydropyran(1) Cyclohexyl formate(1); Dimethyl-1,4-Cyclohexanedicarboxylate(2); Acetomethoxane(1); Cyclohexyl acetate(1) Isophorone diisocyanate(3); Camphor(2) 1-Methyl-1-Ethylcyclopentane(1); 1,1Diethylcyclohexane(2) beta-Terpineol(1); 1-Methylcyclohexanol(1) N-Methylpyrrolidine(1); N-Methylpyrrole(1) 4-(2-Aminoethyl)Morpholine(1); NEthylmorpholine(1)

1.2 Predictive methods of the FP for pure compounds

CHcyc–C=CHn (n in 1..2) CHcyc–OH CHcyc–NH2 CHcyc–NH–CHn (n in 0..3) CHcyc–SH CHcyc–CN CHcyc–COOH

(continued)

31

32

Table 1.5 (continued) Mj

AROMRINGs1s2 AROMRINGs1s3 AROMRINGs1s4

58 33 73

AROMRINGs1s2s3 AROMRINGs1s2s4 AROMRINGs1s3s5 AROMRINGs1s2s3s4

Description

Examples

–0.00269 –0.00494 0.00014

ortho substitution in benzene meta substitution in benzene para substitution in benzene

13 33 11 3

–0.00650 –0.01117 –0.00532 –0.01075

AROMRINGs1s2s3s5

4

–0.01713

AROMRINGs1s2s4s5

4

–0.00700

PYRIDINEs2

1

–0.00042

substitution in positions 1-2-3 (for benzene) substitution in positions 1-2-4 (for benzene) substitution in positions 1-3-5 (for benzene) substitution in positions 1-2-3-4 (for benzene) substitution in positions 1-2-3-5 (for benzene) substitution in positions 1-2-4-5 (for benzene) substitution in position 2 (for pyridine)

PYRIDINEs3 PYRIDINEs4

3 1

0.01991 0.00116

PYRIDINEs2s6 AROMRINGs1s2s3s4s5

1 2

–0.00766 –0.01108

o-Terphenyl(1); o-Cymene(1) m-Ethyltoluene(1); m-Cresol(1) 1-(4-Ethylphenyl)-2-(4-Ethylphenyl)Ethane (2); p-Xylene(1) 1,2,3-Trichlorobenzene(1); 2,3-Xylenol(1) Trioctyl trimellitate(1); 4-Chloro-o-Xylene(1) 1,3,5-Triisopropylbenzene(1); Mesitylene(1) 1,2,4-Trimethyl-3-Ethylbenzene(1); 1,2,3Trimethyl-4-Ethylbenzene(1) 4,6-Dinitro-o-sec-Butylphenol(1); 1,2,3,5Tetramethylbenzene(1) 1,2,4,5-Tetramethylbenzene(1); Pyromellitic acid(1) 2-Methylpyridine(1); 2‐[(pyridin‐2‐yl)methyl] pyridine (2) Niacin(1); Nicotinonitrile(1) 4-Methylpyridine(1); 4‐[(pyridin‐4‐yl)methyl] pyridine (2) 2,6-Dimethylpyridine(1); Pentaethylbenzene(1); Pentamethylbenzene (1)

substitution in position 3 (for pyridine) substitution in position 4 (for pyridine) substitution in positions 2-6 (for pyridine) substitution in positions 1-2-3-4-5 (for benzene)

1 Flash Point

Frequency of occurrence

Groups

Table 1.6: List of the third-order groups, their contributions to the FP, and their number of occurrences in the molecules. Ok

HOOC–(CHn)m–COOH (m>2, n in 0..2)

5

–0.02955

NH2–(CHn)m–OH (m>2, n in 0..2)

1

0.01590

OH–(CHn)m–OH (m>2, n in 0..2)

5

–0.00607

NH2–(CHn)m–NH2 (m>2; n in 0..2)

3

–0.00342

NC–(CHn)m–CN (m>2)

2

–0.00005

aC–(CHn=CHm)cyc (fused rings) (n,m in 0,,1)

17

–0.00021

aC–aC (different rings)

14

–0.00598

aC–CHncyc (different rings) (n in 0..1) aC–CHncyc (fused rings) (n in 0..1)

2

–0.00076

20

–0.00386

Description

Examples

two carboxyl groups connected to each other through linear alkane chains with more than two C –NH2 group connected to –OH group through linear alkane chains with more than two C 2 –OH groups connected to each other through linear alkane chains with more than two C 2 –NH2 groups connected to each other through linear alkane chains with more than two C 2 cyano-groups connected to each other through linear alkane chains with more than two C aromatic carbon connected to non-aromatic CHn=CHm (alkenyl group) in a same ring 2 aromatic carbons connected outside the rings aromatic carbon connected to a carbon in a different ring aromatic carbon connected to a nonaromatic carbon in a same ring

Adipic acid (1); Azelaic acid(1)

3-Amino-1-propanol (1)

1,3-Propylene glycol(1); 1,4-Butanediol(1) Hexamethylenediamine(1); 1,3-Propanediamine(1) Glutaronitrile(1); Adiponitrile(1)

Acenaphthalene(2); Indene(1)

Biphenyl(1); p-Terphenyl(2) 1-Phenylindene(1); Cyclohexylbenzene(1) 1,2,3,4-Tetrahydronaphthalene(2); alpha-Tocopherol(1)

1.2 Predictive methods of the FP for pure compounds

Frequency of occurrence

Groups

(continued)

33

34

Table 1.6 (continued) Ok

aC–(CHn)m–aC (different rings) (m>1; n in 0..2)

3

0.00052

CHcyc–CHcyc (different rings) CH multiring

3

–0.00943

33

0.00182

C multiring

16

–0.00541

aC–CHm–aC (different rings) (m in 0..2)

11

0.00265

aC–(CHm=CHn)–aC (different rings) (m,n in 0..2) aC–CO–aC (different rings)

5

0.00400

1

0.07690

aC–CHm–CO–aC (different rings) (m in 0..2) aC–COcyc (fused rings)

1

–0.08448

5

–0.00881

16

–0.00559

aC–Scyc (fused rings)

Description

Examples

2 aromatic carbons in different rings connected to each other through linear alkane chains with more than two C two carbons in different rings, connected outside the rings >CH– in multi rings (>CH– belongs to more than one ring) >C< in multi rings (>C< belongs to more than one ring) 2 aromatic carbons in different rings connected to each other through non ring carbon 2 aromatic carbons in different rings connected to each other through non ring carbon-carbon double bond 2 aromatic carbons in different rings connected to each other through keton (>C=O) 2 aromatic carbons in different rings connected to each other through keton-carbon bond aromatic carbon connected to a non-aromatic >C=O in a same ring aromatic carbon connected to a non-aromatic –S– in a same ring

4‐(2‐phenylethyl)pyridine (1); 1,2-Diphenylethane (1) 2-Cyclohexyl cyclohexanone(1); Bicyclohexyl(1) Stigmasterol (3); 1,3-Dimethyladamantane(2) Stigmasterol (2); Isoagatholal(1) Triphenylmethane (3); Tetraphenylmethane(6) Triphenylethylene (2); Tetraphenylethylene(4) Benzophenone (1)

2-hydroxy-1,2-diphenylethanone (1);

Anthraquinone(4); Phthalic anhydride(2) 4,6-Dimethyldibenzothiophene(2); Benzothiophene(1)

1 Flash Point

Frequency of occurrence

Groups

–0.00112

6

0.00036

aC–(N=CHn)cyc (fused rings) (n in 0..1)

1

–0.00477

aC–O–aC (different rings)

1

–0.00436

aC–CHn–O–CHm–aC (different rings) (n,m in 0..2) aC–Ocyc (fused rings)

1

0.04622

4

–0.00924

AROM,FUSED[2]

58

0.00512

AROM,FUSED[2]s1

18

0.00255

aromatic carbon connected to a nonaromatic -NH– (or >N–) in a same ring 2 aromatic carbons in different rings connected to each other through nitrogen aromatic carbon connected to a nonaromatic –N=CH- (or –N=CN–

1

0.00840

N=C=O

2

0.03915

Ccyc–N=C=O

1

–0.02272

pyridine ring containing successively 1 aromatic carbon atom (aCH), 1 aromatic nitrogen atom (aN), 2 aCH and 2 fused aromatic carbon (faC) (aCH—aN—aCH— aCH—faC—faC) pyridine ring containing successively 2 fused aromatic carbon (faC), 1 aromatic nitrogen atom (aN), 2 faC and 1 aromatic carbon atom (aCH). (faC—faC—aN—faC— faC—aCH) >N– in multi rings (>N– belongs to more than one ring) >N– except as above –N=C=O (isocyanate) connected to non ring or non aromatic atom –N=C=O (isocyanate) connected to ring atom

Isoquinoline(1); 5,6,7,8‐tetrahydroisoquinoline (1)

Acridine (1); 5,7‐diazapentacene (2)

Triethylenediamine (2); octahydro‐ 1H‐quinolizine (1) 2,2ʹ,2ʹ’-Nitrilotris-acetonitrile (1); tribenzylamine (1) Methyl isocyanate (1); Isophorone diisocyanate(1) Cyclohexyl isocyanate(1); Isophorone diisocyanate(1)

1.2 Predictive methods of the FP for pure compounds

PYRIDINE,FUSED[2-iso]

37

38

1 Flash Point

1.2.1.6 Machine learning-developed models of prediction of flash point Park et al. [49] developed easy-to-apply machine learning-developed models for predicting the FP of pure organic compounds. They used readily available variables (i.e., the numbers of atomic elements, molecular weights, and normal boiling points) as default variables. They utilized two steps as follows: 1) Building multiple linear regression (MLR) models by incorporating default input variables. This step identifies an optimal subset of predictors by constructing an MLR model via the sequential floating backward selection (SFBS) algorithm. 2) Building MLR models by incorporating interaction and transformed variables to improve the predictions from the models in step 1. This step incorporates nonlinearity and interaction terms to construct MLR models via the sequential floating forward selection (SFFS) algorithm by selecting the optimal subset of default variables. Park et al. [49] constructed the model for predicting the FP using step 2 as follows: pffiffiffiffiffiffiffiffiffi FP ¼ 259:0  7:532nCl  17:04nSi þ 49:33 NBP  84:17 log nC þ 39:47 log MW  366:7 log NBP  4:569 expðnBr Þ  27:75n2I þ NBPð0:024nSi  0:1217nS Þ (1:8) þ 5:733ðnC × nS Þ þ nCl ð2:083nC  1:38nH þ 5:421nO  11:17nN Þ where nC , nH , nO , nN , nCl , nSi , nBr , nI , and nS are the number of carbon, hydrogen, oxygen, nitrogen, chlorine, silicon, bromine, iodine, and sulfur atoms, respectively; MW is the molecular weight of an organic compound of interest in g/mol. Example 1.5: Use eq. (1.8) and calculate FPs for the following organic compounds: (a) Trifluoromethyl trifluorovinyl ether; NBP = 247.15 K [43] F F

O F

F F

F

Chemical Formula: C3F6O Molecular Weight: 166.02

(b) Trimethoxysilane; NBP=357.5 K [43] O SiH O

O

Chemical Formula: C3H10O3Si Molecular Weight: 122.20

1.2 Predictive methods of the FP for pure compounds

39

(c) Trimethylchlorosilane; NBP=330.75 K [43] Cl

Si

Chemical Formula: C3H9ClSi Molecular Weight: 108.64

Answer: (a) pffiffiffiffiffiffiffiffiffi FP ¼ 259:0  7:532nCl  17:04nSi þ 49:33 NBP  84:17 log nC þ 39:47 log MW  366:7 log NBP  4:569 expðnBr Þ  27:75n2I þ NBPð0:024nSi  0:1217nS Þ þ 5:733ðnC × nS Þ þ nCl ð2:083nC  1:38nH þ 5:421nO  11:17nN Þ pffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 259:0 7:532 × 0 17:04 × 0 þ 49:33 247:15  84:17 logð3Þ þ 39:47 logð166:02Þ  366:7 logð247:15Þ  4:569 expð0Þ  27:75ð0Þ2 þ 247:15ð0:024 × 0  0:1217 × 0Þ þ 5:733ð3 × 0Þ þ 0 × ð2:083 × 3  1:38 × 0 þ 5:421 × 1  11:17 × 0Þ ¼ 204:49 K The measured FP is 209 K [43]. (b) pffiffiffiffiffiffiffiffiffi FP ¼ 259:0  7:532nCl  17:04nSi þ 49:33 NBP  84:17 log nC þ 39:47 log MW  366:7 log NBP  4:569 expðnBr Þ  27:75n2I þ NBPð0:024nSi  0:1217nS Þ þ 5:733ðnC × nS Þ þ nCl ð2:083nC  1:38nH þ 5:421nO  11:17nN Þ pffiffiffiffiffiffiffiffiffiffiffi ¼ 259:0  7:532 × 0  17:04 × 1 þ 49:33 357:5  84:17 logð3Þ þ 39:47 logð122:20Þ  366:7 logð357:5Þ  4:569 expð0Þ  27:75ð0Þ2 þ 357:5ð0:024 × 1  0:1217 × 0Þ þ 5:733ð3 × 0Þ þ 0 × ð2:083 × 3  1:38 × 10 þ 5:421 × 3  11:17 × 0Þ ¼ 289:19 K The measured FP is 269.15 K [43]. (c) pffiffiffiffiffiffiffiffiffi FP ¼ 259:0  7:532nCl  17:04nSi þ 49:33 NBP  84:17 log nC þ 39:47 log MW  366:7 log NBP  4:569 expðnBr Þ  27:75n2I þ NBPð0:024nSi  0:1217nS Þ þ 5:733ðnC × nS Þ þ nCl ð2:083nC  1:38nH þ 5:421nO  11:17nN Þ pffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ 259:0  7:532× 1  17:04 × 1 þ 49:33 330:75  84:17 logð3Þ þ 39:47 logð108:64Þ  366:7 logð330:75Þ  4:569 expð0Þ  27:75ð0Þ2 þ 330:75ð0:024 × 1  0:1217 × 0Þ þ 5:733ð3 × 0Þ þ 1 × ð2:083 × 3  1:38 × 9 þ 5:421 × 0  11:17 × 0Þ ¼ 249:64 K The measured FP is 245 K [43].

40

1 Flash Point

1.2.2 The SGC methods The second category for predicting the FP contains models, which are developed based on the SGC methods. According to the SGC methods, properties are calculated as a function of the number and the type of predefined functional groups constituting the compound. Linear SGC method is the simplest form, which offers the following linear equation to predict the property: X (1:9) ϕ¼cþ ni ϕi where ni and ϕi are the number and contribution of functional group i, respectively, and c is a constant. The SGC methods [42, 50] are widely used for the prediction of different properties, e.g., solid phase heats of formation [51], gas phase heats of formation [52], heats of fusion and fusion temperature [53, 54], and heats of sublimation [55]. For predicting the FP for some classes of organic compounds, the SGC methods are widely used. Application of the SGC methods for calculation of the FP property of pure compounds was done on the basis of different approaches. Albahri [56] presented a structural SGC method for predicting the flammability characteristics of pure hydrocarbon fluids that have a significant contribution to the overall flammability characteristics and arrive at the sets of groups. The investigated flammability characteristics include the FP, the autoignition temperature (AIT), and the upper and lower flammability limits of about 500 different substances. The calculated values of the SGC method can predict the said flammability properties of pure components from the knowledge of only the molecular structure. The proposed method can predict FP with average percentage error (APE) of 1.8%. Pan et al. [57] constructed relationships between structure and the FP of 92 alkanes by means of an artificial neural network (ANN) using the SGC method. For these alkanes, the average absolute deviation of the predicted FP is 4.8 K, and the root mean square (rms) error being 6.86. Gharagheizi et al. [58] proposed a collection of 79 functional groups to correlate the FP of pure components. This approach constructs a neural network−group contribution correlation to estimate FP of pure components. It was used for predicting the FP of 1,378 pure components of various chemical families. The SGC method lacks accuracy for large datasets of compounds from diverse families. Another drawback of the SGC method is its inability to distinguish between the properties of isomers. Although the SGC methods result in good calculation/prediction of the FP of some classes of organic compounds, their applications are generally limited to a particular group of materials where their values of groups are specified. Some of the best available SGC methods are reviewed here.

1.2.2.1 Organosilicon compounds Wang et al. [59] proposed a predictive model of the FP for organosilicon compounds via the SGC method. They built up their method by a training set of 184 organosilicon

41

1.2 Predictive methods of the FP for pure compounds

compounds with the average error of 8.91 K. The predictive capability of the proposed model has been demonstrated on a testing set of 46 organosilicon compounds with the average error of 11.15 K. The proposed equation to predict the FPs for organosilicon compounds has the following form: X

FP = 224.54 +

! νi f i

− 8.5904 × 10−4

X

i

!2 νi f i

(1:10)

i

where νi is the number of group i in a molecule and fi is the group contribution for the ith contributed group. The values of fi are given in Table 1.7. Example 1.6: Calculate the FP of allyldimethylchlorosilane with the following molecular structure: Si

Cl

Answer: (a) The molecular structure for allyldimethylchlorosilane consists of two (–CH3), one (=CH2), one (=CH–), one (>CH2), one (>SiCH2 >CH– >C< =CH2 =CH– =C< ≡CH ≡C– >CH2 (ring) >CH– (ring) =CH– (ring) =C< (ring) –F –Cl –Br –OH –O– (non ring) –O– (ring) >C=O (nonring) ketone

–4.7543 15.6399 29.7614 41.8024 –0.9360 11.8316 28.2627 25.8011 10.14026 12.0741 26.3594 11.6540 28.2310 –17.3884 21.5543 29.2475 49.4208 8.7550 17.4888 36.9982

Serial No. 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39

Group

fi

–NH2 >NH (non-ring) >NH (ring) >N– (non-ring) –N= (ring) –CN –S– –SiH3 >SiH2 >SiH– >Si< >SiH– (ring) >Si< (ring) >N– (ring) –Cl (attached to Si) –I (attached to Si) –N=C=S –N=C=O –N=C=N–

26.7186 17.0723 20.6411 23.0458 58.1676 46.3254 38.2233 5.1962 7.0559 22.9298 35.2149 8.7554 21.0877 –19.7121 9.4766 46.5773 69.7229 18.3904 57.1338

42

1 Flash Point

P

νi fi = 2ð − 4.7543Þ + 1ð − 0.9360Þ + 1ð11.8316Þ + 1ð15.6399Þ

i

+ 1ð35.2149Þ + 1ð9.4766Þ = 61.7184

2 P P −4 FP = 224.54 + νi f i − 8.5904 × 10 νi f i i

i

= 224.54 + ð61.7184Þ − 8.5904 × 10−4 ð61.7184Þ2 = 282.99 K The experimental value is 278.15 K [28].

1.2.2.2 MNLR and ANN structural group contribution methods Albahri [45] used a structural group contribution method to determine the FP using two techniques: multivariable non-linear regression (MNLR) and ANN. The set of 37 atom-type structural groups was used to represent the FP for about 375 substances. The final FP equation has the following form: ! X ni FPi (1:11) FP = 180.594 + 23.3514 i

where FPi is the atom-type structural group contribution and ni is the number of structural groups in the molecule. The values of FPi are given in Table 1.8. Table 1.8: The structural group contribution values [45]. Serial No.

Group

(FP)i

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

–CH3 >CH2 >CH– >C< =CH2 =CH– =C< =C= ≡CH ≡C– >CH2 (ring) >CH– (ring) >C< (ring) =CH– (ring) =C< (ring) –Cl –OH –O– (non ring) >C=O (non-ring) ketone

0.2823 0.6199 0.8512 1.0179 0.1431 0.5856 0.9265 0.6446 0.3692 0.7478 0.6125 0.6517 0.4666 0.5859 1.0747 0.9295 3.2230 0.1692 2.5228

Serial No. 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37

Group –HC=O (aldehyde) –COOH (acid) =O –NH2 >NH (non-ring) >N– (non-ring) –O– (ring) >C=O (ring) >NH (ring) –N= (ring) >N– (ring) –H >S >SO –SH =S ≡N =S=

(FP)i 2.0626 5.1405 1.8826 1.7216 4.7711 0.4800 0.9161 4.0227 2.2231 3.4118 1.8971 0.0000 1.8503 7.1659 1.2870 2.0413 2.8685 0.1067

1.2 Predictive methods of the FP for pure compounds

43

Example 1.7: Calculate the FP of (a) p-diethyl benzene and (b) methyl diethanolamine with the following molecular structure: HO

N

HO (b)

(a)

Answer: (a) The molecular structure for p-diethyl benzene consists of two (–CH3), two (>CH2), four (=CH (ring)), and two (>C= (ring)).
 P ni FPi FP = 180.594 + 23.3514 i

= 180.594 + 23.3514ð2ð0.2823Þ + 2ð0.6199Þ + 4ð0.5859Þ + 2ð1.0747ÞÞ = 327.7 K The experimental value is 329.3 K [59]. (b) The molecular structure for methyl diethanolamine consists one (–CH3), four (>CH2), two (–OH), and one (>N– (non-ring))
 P ni FPi FP = 180.594 + 23.3514 i

= 180.594 + 23.3514ð1ð0.2823Þ + 4ð0.6199Þ + 2ð3.2230Þ + 1ð0.4800ÞÞ = 406.8 K The experimental value is 400 K [60].

1.2.2.3 The improved SGC approaches based on experimental data of a wide range of organic compounds Albahri and Esmaeil [61] introduced a general QSPR for predicting the FP of 1471 pure compounds. They used ANN and MLR along with the SGC approach to calculate FP. They proposed four SGC methods based on MLR, which resulted in almost the same accuracy with an average absolute error (AAE) ranging from 4% to 5% and a correlation coefficient (r) from 0.93 to 0.96. They also implemented the ANN method to enhance the predictions of one of the methods. Among four SGC models, one model is more reliable, which is based on the Benson binary structural group contribution method for the prediction of thermodynamic properties of organic compounds [62]. Benson’s method explains the effect of the next group neighbor on the structural group contribution. For example, n-pentane consists of two methyl groups (−CH3) and three ethyl groups (–CH2). Benson’s group contribution method uses the symbols C–(H)3(C) for −CH3 which means a carbon atom attached to another carbon atom and three hydrogen atoms, and C-(C)2(H)2 for –CH2 which means a carbon atom attached

44

1 Flash Point

to two other carbon atoms and two hydrogen atoms. Albahri and Esmaeil [61] expanded the number of structural groups to 360, which are given in Table 1.9. They used the following equation for the calculation of FP: FP ¼ 214:730 þ

X

! ni FPi

(1:12)

i

where eq. (1.12) has the same parameters as in eq. (1.11). Table 1.9: The structural group contribution values. Serial no.

Group

FPi

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

C-(CH2)(H)3 C-(CH)(H)3 C-(C)(H)3 C-(C)2(H)2 C−(C)3(H) C-(C)4 Cd-(H)(O) Cd-(H)2 Cd-(C)(H) C-(Cd)(C)(H)2 C-(Cd)(H)3 Cd-(C)2 C-(Cd)(C)2(H) Cd-(Cd)(H) C-(Cd)(C)3 C-(Cd)(0)(H)2 C-(O-C)(H)3 C-(O-CO)(H)3 C-(CO)(H)3 C-( C)(CO)(H)2 C-( C)2(CO)(H) C-( C)3(CO) C-(C)(O)(H)2 C-(C)2(O)(H) C-( C)3(O) C-(O)2(H)2 CO-(CH2)(O) CO-(CH)(O) CO-(C)(O) CO-(O)(H) CO-(C)(H) CO-( C)2 C-(C)(Cl)(H)2

1.702268 2.362253 2.905774 11.31439 11.49337 11.01741 –7.72132 –2.94745 1.86281 20.22657 5.699348 4.301482 18.12588 10.44619 16.72775 34.11388 –0.28915 –1.24201 2.446113 16.63454 15.50676 –7.42827 12.4331 10.28266 12.52383 12.9672 26.57928 28.1063 29.50753 18.19533 35.7308 36.64033 20.39277

Serial no. 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213

Group O-(CO)2 O-(CO)(C) O-(CO)(Cb) O-(C)(O) O-(Cb)2 O-(N)(H) O-(C)(B) O-(NI)(H) Nd-Cd N-(Cb)2(H) N-(C)(H)2 N-(CO)(C)(H) N-(CO)2(H) N-(Cb)3 N-(C)2(O) N-(CO)(H)2 N-(CO)(C)2 N-(Cb)(H)2 N-(H)3 N-(C)2(Cb) NO-(Cb)(O) S-(Cd)2 SO2-(C)2 SO2-(O)2 SO-(O)2 SO2-(C)(Cl) S-(Cb)(C) S-(Cb)(H) S-(Cb)(Cd) S-(Cb)2 S-(C) S-(C)2 S-(C)(S)

FPi 43.76638 26.65063 21.85995 13.48235 29.35316 96.23402 12.8913 90.69796 0 17.13775 34.80115 172.3795 82.1507 23.28608 −5.22373 132.459 258.6805 113.8244 −2.28905 48.45599 0.459609 −24.3553 −74.2167 0 1111.469 0 −1.26002 30.51287 31.28435 8.130267 37.87289 6.564962 13.87924

45

1.2 Predictive methods of the FP for pure compounds

Table 1.9 (continued) Serial no.

Group

34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77

C-(C)2(Cl)(H) C-(C)(Cl)2(H) Cb-(H) Cb-(C) C-(Cb)(H)3 C-(Cb)(C)(H)2 C-(Cb)(C)2(H) C-(Cb)(C)3 Cb-(O) Cb-(COOH) Cb-(Cb) C-(S)(H)3 C-(C)(S)(H2) C-(C2)(S)(H) C-(C3)(S) Cb-(N) C-(N)(H)3 C-(C)(N)(H)2 C-(C)2(N)(H) C-(C)(CN)(H)2 C-(C)2(CN)(H) O-(Cb)(H) O-(CH2)(H) O-(CH)(H) O-(C)(H) O-(C)2 O-(CO)(CH3) O-(CO)(CH2) O-(CO)(CH) O-(CO)(H) N-(CH2)(H)2 N-(CH)(H)2 N-(cyclopentyl)(H)2 N-(C)2(H) N-(C)3 N-(CB)(H)2 N-(CB)(C)(H) N-(CB)(C)2 NI-(CB)2 S-(C)(H) S-(C)2(H) ClC-(Cb)2(C)2 C-(C)(Pb)(H2)

FPi 27.08847 18.2391 12.52023 11.34031 11.69134 17.7546 22.40979 24.3508 4.279604 2.351915 20.75047 12.83064 34.96559 27.85825 39.73863 29.69458 –10.6428 0.108719 –12.1044 88.92266 3.162605 58.8481 57.80813 56.78384 52.8897 16.09527 32.4068 11.60002 43.15965 69.54379 43.64943 44.31381 –15.0556 50.30669 55.81647 48.83514 38.09461 72.16653 14.43147 8.940247 –4.544 16.2834 12.23922 36.15272

Serial no. 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257

Group S-(Cb)(S) S-(CO)(C) CN-(C) CN-(Cb) P-(Cb)3 PO-(O)3 Pb-(C)4 IFBrB-(O)3 Si-(Cl)2(C)2 Si-(Cl)(C)3 Si-(Cl)2(Cb)2 Si-(Cd)(C)(Cl)2 Si-(Cl)2(C)(H) Si-(Cl)2(Cb)(C) Si-(C)2(O)2 Si-(C)2(H)2 Si-(O)3(Cd) Si-(C)(H)3 Si-(C)4 Si-(Cl)3(H) Si-(Cl)3(Cd) Si-(Cl)3(C) Si-(Cl)3(Cb) Si-(O)3(H) Si-(C)3(H) Si-(O)4 Si-(C)3(O) Si-(O)3(C) As-(H)3 C-(C)(F)2(H) C-(C)(Cl)(F)2 C-(C)(SO2)(H)2 C-(C)2(CO)(O) C-(C)2(F)2 C-(C)2(I)(H) C-(C)2(N)2 C-(C)2(NCO)(H) C-(Cb)(CL)(H)2 C-(Cb)(CL)2(H) C-(Cb)(CL)3 C-(Cb)(F)3

FPi −5.06435 −48.529 61.74718 −11.3261 0 −36.0213 0 28.94304 −6.93501 7.516088 2.303569 39.98964 62.73838 −0.49951 88.34547 −2.09958 23.53402 440.021 20.31119 674.5113 −33.4795 99.55062 −18.5803 47.66145 23.4774 0 514.3225 38.02179 717.96 270.919 53.20077 −51.7301 −9.22231 −9.04923 329.5077 23.51738 73.16727 44.75239 37.69953 51.41807 28.19505 48.04505 20.53551 14.4678 (continued)

46

1 Flash Point

Table 1.9 (continued) Serial no.

Group

78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121

C-(CN)2(H)2 C-(CO)2(H)2 C-(C)2(CN)(O) C-(C)(CN)(O)(H) C-(C)(CO)(O)(H) C-(Cd)(Cl)(H)2 C-(Si)(H)3 C-(O-Cb)(H)3 C-(C)(Ct)(H)2 C-(C)2(Cb)(H) C-(Cd)2(H)2 C-(CO)(O)(H)2 C-(C)(O)2(H) C-(Cb)(Cd)(H)2 C-(C)(I)(H)2 C-(C)3(N) C-(Ct)(H)3 C-(Si)(C)(H)2 C-(Cb)(CN)(H)2 C-(Cb)(N)(H)2 C-(C)2(Cb)(O) C-(Cb)(O)(C)(H) C-(Br)(C)2(H) C-(C)(Cl)3 C-(S_d)2 C-(F)3(H) C-(Br)2(C)(H) C-(C)(Br)(H)2 C-(C)(F)(H)2 C-(F)2(O)(H) C-(SO2)(C)(H)2 C-(C)(I)(H)2 C-(C)(N)(H)2 C-(C)(SO2)(H)2 C-(Ct)(O)(H)2 C-(Cb)2(H)2 C-(CO)(Cl)3 C-(CO)(Cl)(H)2 C-(CO)(CN)(H)2 C-(CO)(Cl)2(H) C-(CO)(S)(H)2 C-(Cb)(O)(H)2 C-(CN)(H)3 C-(CN)(N)(H)2

FPi 46.92557 39.85702 0.971509 14.52666 10.07595 50.61136 –16.2506 –3.91222 8.867255 –68.6781 20.22103 35.4101 1.884557 12.73698 59.93888 –3.8344 –18.0505 –6.267 90.5428 0.164646 32.69482 10.12787 74.12041 31.08138 28.41993 19.49507 89.85733 33.15446 11.06997 –12.4044 123.6583 48.93486 0.37198 123.7643 32.97794 29.04126 23.68892 41.04579 57.907 42.68768 104.7859 21.41973 –11.9863 12.16394

Serial no. 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301

Group C-(Cb)2(C)(H) C-(Cb)3(H) C-(Cb)4 C-(Cd)(C)(O)(H) C-(Cd)(C)3 C-(Cd)(CN)(H)2 C-(Cd)(CN)(O)(H) C-(Cd)(N)(H)2 C-(CH3)(H)3 C-(CN)(O)(H)2 C-(NCO)(C)(H)2 C-(NI)(H)3 C-(NOO)(C)(H)2 C-(NOO)(C)2(H) C-(O-PO)(H)3 C-(O-Si)(H)3 Cb-(Br) Cb-(Ct) Cb-(F) Cb-(I) Cb-(NCO) Cb-(NOO) Cb-Ni Cd-(C)(H) Cd-(Cb)2 Cd-(CI)(H) Cd-(CN)(C) Cd-(N)(H) Cd-(N)(H)2 Cd-(Ni)(H) Ci-(Cd)(H) Ci-(Ni)(H) Co-(Cb)2 CO-(Cd)(C) CO-(Cd)(N) CO-(CL)(C) CO-(N)(H) CO-(O)2 CO-(S_d) Ct-(Cb) N-(Cb)(CO)(H) N-(Cd)(N)(H) N-(Cd)2(H) N-(CO)(C)2

FPi −16.7629 12.44551 14.50403 65.52741 31.13962 82.50458 0 10.30946 −38.29 58.63306 49.87121 26.83137 43.70682 44.6957 11.13537 −15.947 41.52182 0 17.51172 33.72573 59.32379 81.632 26.33332 −52.4901 1.644899 9.082943 74.21739 41.19129 81.43868 22.17519 89.9957 0 14.45082 37.39993 −5.51614 26.84618 139.8446 22.36975 −28.7301 0 110.598 0 0 −9.60354

47

1.2 Predictive methods of the FP for pure compounds

Table 1.9 (continued) Serial no.

Group

122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165

C-(N)(H)3 C-(Ct)(H)3 Ct-(C) Ct-(H) Cb-Cl Cb-P Cb-Cd Cb-Si Cb-(SO2) Cb-(S) Cb-(NO2) Cb-(CN) Cb-(CO) Cd-(Cb)(H) Cd-(S)(H) Cd-(S)(C) Cd-(CO)(H) Cd-(C)(Cl) Cd-Nd Cd-(Si)(H) Cd-(Cd)(C) Cd-(Cd)(Cl) Cd-(Cl)(S) Cd-(CN)(H) Cd-(CO)(H) Cd-(CO)(C) Cd-(H)2 Cd-(O)(H) Cd-(C)(H)2 Cd-(Ct)(H) Ct-(H) Ct-(C) CO-(O)(Cd) CO-(CH3)(O) CO-(O)(Cb) CO-(N)(Cb) CO-(H)(Cb) CO-(O)2 CO-(CI)(Cb) CO-(CO)(C) CO-(CO)(O) CO-(C)(N) CO-(CH)2(H) CO-(S)

FPi –73.2119 –22.6526 –33.8963 26.81877 26.04 16.87215 16.95241 37.96856 53.09815 38.3059 33.10893 78.14488 28.71712 22.49603 31.79635 33.94626 30.72371 5.957604 74.65742 –44.2943 8.070961 2.863679 66.88214 79.21026 33.64071 34.32709 –14.0829 22.26357 10.79239 42.94284 –10.604 19.26734 25.21707 42.47368 16.25749 5.545333 46.04634 2.904668 14.35224 31.90546 36.52457 26.08012 72.62344 40.47271

Serial no. 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345

Group NCO-(C)(H) NCO-(C)2 NCO-(Cb)(H) Ni-(C)2 Ni-(Cd) NI-N NO2-(Cb) NOO-(C) NOO-(Cb) O-(CO)(O) O-(O)(H) C-(O-SO2)(H)3 Sd-(C) Sd-(CO) Si-(C)2(Cl)(H) Si-(Cl)2(H)2 C-(C)(F)(H)2 C-(C)3(Cb) O-(Si)(H) C-(H)3(O) C-(O)(Cl)(H)2 O-(CH3)(H) NCO-(Cb)2 C-(H)3(Br) C-(H)3(Cl) C-(H)2(Br)2 Cd-(F)2 Cd-(O)(F) O-(Cd)(C) C-(C)(O)(F)2 C-(F)3(C) C-(CO)(N)(H)2 N-(Cb)(H)(N) C-(H)2(N) C-(Cb)2(Cd)(H) C-(H)4 C-(H)2(O)2 C-(H)3(I) CO-(Cb)(CH3) Cd-(Br)(H) Cd-(F)(Cl) Cd-(F)(H) CO-(H)2 O-(Cd)(CH2)

FPi −1.57011 1.217665 −4.44028 0 −16.5986 0 47.952 34.69645 0 53.50014 0 16.63537 0 0 46.63777 −29.2969 0 31.61326 −144.897 −3.00656 8.180413 72.2765 0 11.96636 −26.0135 101.6628 −73.8673 34.00857 0 0 0 221.4974 51.83922 3.284967 0 −129.58 4.32885 52.20494 45.86199 10.91382 37.30561 −66.2987 −8.73007 7.618764 (continued)

48

1 Flash Point

Table 1.9 (continued) Serial no.

Group

166 167 168 169 170 171 172 173 174 175 176 177 178 179 180

CO-(S)( C) CO-(O)(Cl) CO-(CO)(H) CO-(S)(O) CO-(C)(F)3 O-(Cb)( C) O-(PO)( C) O-(PO)(Cb) O-(Si)( C) O-(SO2)( C) O-(SO2)( H) O-(SO)( C) O-(Si)2 O-(CO)(Cd) O-(NO)(H)

FPi

Serial no.

104.048 78.5022 23.63497 0 20.3632 46.26479 50.29503 33.51059 –166.103 54.07459 742.2943 –514.16 –381.446 41.23404 40.02891

346 347 348 349 350 351 352 353 354 355 356 357 358 359 360

Group CNO-(Cd)(O) O-(CN)(Cd) NI-(O)(C) CN-(H) NCO-(C) N-(N)(H)2 C-(H)3(NCO) NO2-C C-(H)3(NO2) SO-(C) C-(H)3(SO) SO2-(C)(Cl) C-(H)3(SO2) CN-(CN) Cl-(SO2)

FPi 0 0 0 40.26993 36.26993 96.26993 0 93.41993 0 146.2699 0 168.4199 0 −5.36503 0

Example 1.8: Calculate the FP of 1,2,3-benzenetriol with the following molecular structure: OH OH

OH

Answer: The molecular structure for 1,2,3-benzenetriol consists of the following structural groups for estimating FP: three (Cb-(H)), three (Cb-(O)), and three (-O-(Cb)(H)): ! X ni FPi FP ¼ 214:730 þ i

¼ 214:730 þ ð3 × 12:520Þ þ ð3 × 4:279Þ þ ð3 × 58:848Þ ¼ 441:67 K The experimental value is 458 K [60]. 1.2.3 QSPR models QSPR methods belong to the third category of methods for predicting the FP in which molecular descriptors, which are numerical quantities calculated from 2-D or 3-D structure of compounds, are used in predictive correlations. Most of QSPR methods are procedures that are more sophisticated. They usually include drawing chemical

1.2 Predictive methods of the FP for pure compounds

49

structure of compounds, minimizing energy, calculating molecular descriptors, and selecting the most effective ones among hundreds of available molecular descriptors. Molecular descriptors are molecular-based parameters where they are numeric characteristics of a pure compound directly calculated from its molecular structure with special algorithms. Several molecular descriptors, normally less than 10 molecular descriptors, are selected to correlate the desired property such as the FP of pure compounds. Several QSPR methods have been introduced in the literature to calculate the FP of pure compounds. Tetteh et al. [63] used radial basis function (RBF) neural network models for the simultaneous estimation of the FP and the NBP based on 25 molecular functional groups and their first-order molecular connectivity index (1χ) has been developed. The success of this model depends on a network optimization strategy based on biharmonic spline interpolation for the selection of an optimum number of RBF neurons (n) in the hidden layer and their associated spread parameter (σ). The method of Orthogonal Least Squares (OLS) learning algorithm was used for training of the RBF networks. Tetteh et al. [63] divided the total database of 400 compounds into training (134), validation (133), and testing (133), where the average absolute errors (AAEs) obtained for the validation and testing sets are 10 and 12 K, respectively. Katritzky et al. [64] used geometrical, topological, quantum mechanical, and electronic descriptors calculated by CODESSA PRO software to develop QSPR models for predicting the FP of 758 organic compounds. They reported multilinear regression models and a non-linear model based on an artificial neural network. Gharagheizi and Alamdari [65] introduced a general QSPR model for the prediction of the FP of 1,030 pure compounds. They used Genetic Algorithm-based Multivariate Linear Regression (GAMLR) technique to select four chemical structure-based molecular descriptors from a pool containing 1,664 molecular descriptors. Gharagheizi et al. [34] used experimental NBP of the compound and two chemical structure-based QSPR parameters. They used a comprehensive database of the FPs containing 1,472 pure compounds of various chemical structures for the development of the model. The most important disadvantage of these kinds of QSPR methods is the complex process of calculation for some molecular descriptors from the chemical structure. Molecular descriptors are numerical quantities, which are used in predictive correlations. This kind of QSPR method is less popular than the other methods because of its lower accuracy and more sophisticated procedure. Thus, the selection of appropriate molecular descriptors such as topological indices, quantum chemical parameters, and electrostatic indices, is a key problem. The selected molecular descriptors can be combined with the other methods such as multiple linear regression, partial least squares, and different types of ANNs. These QSPR methods usually require some computer software such as Dragon [66] for obtaining descriptors, which contain unconventional parameters. For all of these QSPR methods, it is important to use a large dataset of different molecular structures because the compounds with similar molecular structure in training set of QSPR procedure should be used as test set. Since the used experimental data for development of the predictive models of the FP are much more than variables, statistical analysis data may

50

1 Flash Point

be used to confirm the reliability of these methods for those new compounds with similar molecular structures that have not been used in the development of methods. The QSPR correlations based on complex molecular descriptors are not generally simple to develop because these QSPR methods require complex computer codes and expert users. There are several simple QSPR models based on elemental composition and structural parameters, which have been used to predict the FP of some classes of organic compounds [67–72]. Since most of QSPR methods require complex molecular descriptors and complicated computer codes, the best simple and reliable methods for predicting the FP are demonstrated here where they need only molecular structural parameters. 1.2.3.1 Saturated alkanes A simple method was introduced for predicting the FP of pure cyclic and acyclic saturated hydrocarbons [72]. It is based on nC as well as two structural parameters, i.e., increasing (ISP) and decreasing (DSP) structural parameters. The values of rms and the average absolute deviations of the predicted FPs are 4.6 and 5.4 K for a dataset of 120 and 59 acyclic and cyclic alkanes, respectively. This model is given as: FP = A + 16.15 nC + 16.68 ISP − 24.71 DSP

(1:13)

where A is a constant that is equal to 146.6 and 154.9 for acyclic and cyclic alkanes, respectively. Two correcting functions ISP and DSP can revise deviations from the predicted results of FPs of saturated hydrocarbons on the basis of nC. They are specified for saturated cyclic and acyclic hydrocarbons according to the following situations: ISP – This parameter can be used only for large cycloalkanes that have more than seven-membered rings where it is equal to 1.0 for these compounds. DSP – This correction parameter may be used for both cyclic and acyclic compounds as: (i) Cycloalkanes with three- or four-membered rings: The value of DSP is 1.0 for those compounds without any substituent or with the attachment of only methyl groups. (ii) Cycloalkanes with five- or six-membered rings: For the attachment of large n-alkyl with more than nine carbon atoms (nC > 9 ), DSP =ðnC > 9 − 9Þ × 0.3. (iii) Acyclic hydrocarbons with isobutyl molecular fragment (i.e., (CH3)3C–R): For alkyl groups containing less than four carbons (nC < 4 ), DSP=0.3nC < 4 . (iv) Small acyclic hydrocarbons: For nC ≤ 4, DSP= 4.25 – nC ≤ 4 .

51

1.2 Predictive methods of the FP for pure compounds

The parameters ISP and DSP are equal to zero if the conditions for giving them various values are not met. Example 1.9: Cyclopentyldodecane has the following molecular structure. Use equation (1.11) and calculate its FP.

Answer: Since condition (ii) is satisfied, DSP =ðnC > 9 − 9Þ × 0.3 = ð12 − 9Þ × 0.3 = 0.9 FP = A + 16.15 nC + 16.68 ISP − 24.71 DSP = 154.9 + 16.15 × 17 + 16.68 × 0 − 24.71 × 0.9 = 407.2 K The measured FP is 409 K, which is taken from the chemical database of the department of chemistry at the University of Akron (USA) [73]. 1.2.3.2 Unsaturated hydrocarbons A simple method was introduced for predicting the FP of different classes of unsaturated hydrocarbons, including alkenes, alkynes, and aromatics. The numbers of carbon and hydrogen atoms are used as a core function that can be revised for some compounds by a correcting function. The rms error is 9 K for a dataset of 173 unsaturated hydrocarbons. The optimized correlation has the following form: FP = 167.1 + ðFPÞcore + ðFPÞcorrecting

(1:14)

ðFPÞcore = 19.68nC − 2.915nH

(1:15)

ðFPÞcorrecting = 16.77FPð + Þ − 32.66FPð − Þ

(1:16)

where

The quantity nH is the number of hydrogen atoms. The parameter (FP)core is a linear function of nC and nH because it depends on molecular weight and degree of unsaturation of the compound. The factors FP(+) and FP(–) are the contributions of structural parameters of unsaturated hydrocarbons for increasing and decreasing of the FP on the basis of (FP)core, respectively, which can be specified according to the following conditions: (i) FP(+) – This parameter can be applied only to polymethyl benzene where the value of FP(+) depends on the number of methyl groups attached to the benzene ring in the ortho position with respect to each other. Thus, it is equal to 0.25nCH3 . (ii) FP(–) – This correcting function can be applied to aromatic, alkene, and alkyne compounds.

52

1 Flash Point

(a) Aromatic compounds – Two different situations can be considered here: 1. For the attachment of isopropyl directly to the aromatic ring and the presence t-butyl in the molecule, the values of FP(–) are 0.25 nisopropyl and 0.5 nt−butyl , respectively. For example, the value of FP(–)=3 × 0.25=0.75 for 1,3,5-triisopropyl benzene. 2. For the attachment of large normal alkyl group (n′ ≥ 10), FP(–) equals 1.0 where the parameter n′ is the number of carbon atoms in the alkyl group. (b) Alkyne – For alkynes containing one triple bond with general formula R–C≡C–H, FP(–) equals 1.25 − 0.25n′ in which n′ < 5, where n′ is the number of carbon atoms in R substituent, e.g., FP(–)=1.25 – 2 × 0.25 = 0.75 for 1-butyne. (c) Alkene – For alkenes with two alkyl groups attached to double bond in form R1–C=C–R2, FP(–) is 2.25 − 0.75n′ where n′ < 3. For example, FP(–) is to 1.5 for propene. (d) Two double bonds – For the existence of two double bonds in form R1 − C = C − C = C − R2 or R1 − C = C = C − R2 , FPð − Þ is 1.0 − 0.5n′, where only one of the alkyl groups (R1 or R2) should be methyl and the other hydrogen atom, e.g., FP(–) = 0.5 in 2-methyl butadiene. Example 1.10: Calculate the FP of tetradecyl benzene with the following molecular structure by using eq. (1.12).

Answer: Since condition 2 of parts (ii) and (a) is satisfied where n′ ≥ 10, FP(–)=1.0 ðFPÞcore = 19.68nC − 2.915nH = 19.68 × 20 − 2.915 × 34 = 294.5K ðFPÞcorrecting = 16.77FPð + Þ − 32.66FPð − Þ = 16.77 × 0 − 32.66 × 1.0 = − 32.66K FP = 167.1 + ðFPÞcore + ðFPÞcorrecting = 167.1 + 294.5 − 32.66 = 428.9K The reported FP is 433 K [73]. The use of two models Albahri [56] and Rowley et al. [40] give 448 K and 468 K, respectively, which have larger deviations. 1.2.3.3 General correlation between saturated and unsaturated hydrocarbons A reliable correlation has been introduced for predicting the FP of various types of saturated and unsaturated hydrocarbons containing cyclic and acyclic paraffin, olefins, alkynes, and aromatic hydrocarbons [68]. Large available experimental data consisting of 441 diverse hydrocarbons were used to derive and test the general correlation. For training set containing 423 of these 441 hydrocarbons, the values of rms and average absolute deviations are 7.7 and 5.7 K, respectively. General correlation has the following form:

1.2 Predictive methods of the FP for pure compounds

FP = 158.7 + 19.86nC − 2.40nH + 51.12Pð+Þ − 49.63Pð − Þ

53

(1:17)

where P(+) and P(–) are two correcting functions, which can be easily specified on the basis of some molecular fragments given in the following sections. 1.2.3.3.1 Structural parameter P(+) (i) Cycloalkanes: For large cycloalkanes containing more than six-membered rings, the value P(+) is equal to 0.4. (ii) Normal alkanes: For normal alkanes with nC ≥14, the value of P(+) is 0.25. (iii) Methyl-substituted aromatics: For polymethyl aromatics, the value of P(+) is 0.1nCH3 where nCH3 is the number of methyl groups attached to the aromatic ring. 1.2.3.3.2 Structural moieties affecting P(–) (i) Small cycloalkanes: For cycloalkanes with three- or four-membered rings without any substituent or with the attachment of only methyl groups, the value of P(–) equals 0.2. (ii) Small acyclic hydrocarbons: If nC ≤ 5, then P(–)=1.40 – (nC –1.3) × 0.35. (iii) The attachment of isopropyl and t-butyl to benzene ring: The values of P(–) equal 0.85 and 1.5 for the attachment of more than two isopropyl or t-butyl directly to the aromatic ring, respectively. (iv) The attachment of normal alkyl group to benzene ring: The values of P(–) equal 0.5 and 1.0 for the attachment of large normal alkyl group with 10 ≤ nʹ ≤ 12 and nʹ > 12, respectively. (v) Alkynes with the general formula R–C≡C–H: The value of P(–) is 1.16 − 0.18nC where nC ≤ 6. (vi) Alkenes with formula R1–C=C–R2: P(–) equals 1.55 – 0.25nC, where nC ≤ 4. (vii)The existence of two double bonds: For the compounds with formula R1–C=C– C=C–R2 and R1–C=C=C–R2, the values of P(–) are 1.75 – 0.3nC and 0.95 – 0.1 nC, respectively, for which only one of alkyl groups (R1 or R2) should be methyl group and the other hydrogen atom. 1.2.3.3.3 Different behavior of mono-alkyl substituted cyclopentane and cyclohexane Both P(+) or P(–) can participate in predicting the FP of mono-alkyl substituted cyclopentane and cyclohexane, which depend on the number of carbon atoms in the alkyl group. For the presence of nʹ = 1, 2 and nʹ = 3 – 9, the values of P(+) are 0.1 and 0.2, respectively. Meanwhile, the value of P(–) is (n′> 12 – 10) × 0.05 + 0.05, where n′> 12 shows that the number of carbon atoms in the alkyl group is greater than twelve.

54

1 Flash Point

The estimated FPs for 18 further hydrocarbons containing complex molecular structures have been compared with the Rowley et al. method [40], which gives much lower values of the rms and average absolute deviations. Example 1.11: Cyclotetradecane has the following molecular structure. Use eq. (1.17) and calculate its FP.

Answer: For this compound, condition (i) in Section 1.2.3.3.1 is satisfied. FP = 158.7 + 19.86nC − 2.40nH + 51.12Pð + Þ − 49.63Pð − Þ = 158.7 + 19.86 × 14 − 2.40 × 28 + 51.12 × 0.4 − 49.63 × 0 = 390.0 K The measured FP is 433 K [73]. The use of Rowley et al. [40] gives 336 K, which has a larger deviation, i.e., 50 K. 1.2.3.4 Kerosene hydrocarbons The kerosene fuels are a distillate fraction of petroleum with boiling point between 150 and 300 °C. They are a mixture of hydrocarbons containing compounds with 10 to 16 carbon atoms in both straight chain and branched formations. Zohari and Qhomi [71] presented two new, reliable, simple correlations for predicting FP of kerosene hydrocarbons. Since the reliability of one of them is higher, this correlation is given as: FP = 187.8 + 18.298nC − 2.570nH − 4.875nR5 + 21.240FP + − 26.699FP −

(1:18)

where nR5 is the number of five member rings in the molecular formula; FP+ and FP– are the several non-additive structural parameters, which are defined as below:

1.2.3.4.1 Definition of FP+ (1) Hydrocarbon fuels containing aromatic rings: For a phenyl ring containing five or six methyl substitutions (e.g., 1,2,3,4,5-pentamethyl benzene), the value of FP+ is 1.0. (2) The compounds containing three or more rings in their structures: If rings adjoin each other by two or more carbon atoms, two situations are expected: (2.1) For six-membered rings and aromatic, FP+ = 1.2. (2.2) If there is any ring, which is not a six-membered ring and aromatic, the value of FP+ is 0.7.

1.2 Predictive methods of the FP for pure compounds

55

1.2.3.4.2 Description of FP–: (1) For polycyclic nonaromatic compounds if there is an unsaturated bond in their structure, FP– = 1.0. (2) For acyclic hydrocarbon compounds containing two or three quaternary carbon atoms in their structure, the value of FP– is 0.7. (3) In the hydrocarbon fuel compounds with aromatic rings, there are several situations: (3.1) For the attachment of one or two isopropyl substitutions to the phenyl ring, the value of FP– is 0.5. Meanwhile, for three isopropyl substitutions, FP– = 1.5. (3.2) For the attachment of tert-butyl substitution to the phenyl ring, the value of FP– is 0.7. Meanwhile, if isobutyl or isobutylene are introduced to a phenyl ring, the value of FP– is 0.5. (3.3) For the attachment of more than two ethyl groups to a phenyl ring, the value of FP– is 0.5. (4) For the presence of a link as –C–C–, –C=C–, –C≡C– or –C(C)– between two phenyl rings, the value of FP– is 0.5. Example 1.12: Find FP of p-diisopropyl benzene with the following molecular structure:

Answer: For this compound, condition (3.1) in Section 1.2.3.4.2 is satisfied. FP = 187.8 + 18.298nC − 2.570nH − 4.875nR5 + 21.240FP + − 26.699 FP − = 187.8 + 18.298 × 12 − 2.570 × 18 − 4.875 × 0 + 21.240 × 0 − 26.699 × 0.5 = 347.8K The measured FP is 349 K [43].

1.2.3.5 Various classes of amines A simple method has been presented for estimating the FP of various types of flammable amines, which include aliphatic amines such as primary, secondary, tertiary, and cyclic amines as well as aromatic amines and heteroarenes containing nitrogen heteroatom [69]. It is based on the contribution of elemental composition and the effects of two correcting functions as:

56

1 Flash Point

FP = 207.2 + 23.43nC − 7.363nH + 49.41nN + 64.79IP − 62.96DP

(1:19)

where IP and DP are increasing and decreasing parameters based on structural parameters of amines. Since FPs of amino derivatives of organic compounds are related to their volatility, the presence of some molecular moieties such as specific polar groups, branching, and the length of substituents attached to amine can affect the values of IP and DP. 1.2.3.5.1 Prediction of IP (a) Hydroxyl, chloro, or acyclic ether groups: For any amino derivative organic compound containing hydroxyl, chloro, or acyclic ether molecular moieties, IP= nOH + 0.4nCl +0.3naO for the presence of hydroxyl (nOH), chloro (nCl), and acyclic ether (naO) groups. (b) Aromatic compounds containing only –NH2 substituents for which nNH2 ≥ 2: The value of IP equals 0.4 for these compounds. (c) Nitro aniline: For only nitro-substituted aniline, IP = 1.0. (d) R–NH2 with nC ≥ 8: The value of IP is 0.5 for these alkyl mines. (e) −Cð=OÞ−N− or −Cð=NHÞ−N−: For the presence of these groups, IP =1.0. 1.2.3.5.2 Prediction of DP (a) Primary, secondary, and tertiary amine (only nN = 1) in which each substituent has nC ≤ 3: The value of DP is 0.3 for these compounds. (b) (Ar)2NH and (Ar)3N: For aryl amines containing more than one aryl substituent, the values of DP are 0.5 and 1.9 for (Ar)2NH and (Ar)3N, respectively. (c) Alkyl derivatives of pyridine or pyrrole: For these compounds, DP = 0.6. Example 1.13: Calculate FP for the following compound:

NH HN

Answer: FP = 207.2 + 23.43nC − 7.363nH + 49.41nN + 64.79IP − 62.96DP = 207.2 + 23.43 × 12 − 7.363 × 12 + 49.41 × 2 + 64.79 × 0 − 62.96 × 0 = 498.8K The predicted result is close to the measured FP, which is 503 K [43].

57

1.2 Predictive methods of the FP for pure compounds

1.2.3.6 Alcohols and phenols A reliable correlation has been developed on the basis of 929 experimental FP values of different alcohols and phenols including acyclic and cyclic alcohols as well as phenols and alcohols with composite aliphatic-aromatic structures [70]. It is based on the elemental composition and some of the structural parameters such as intermolecular hydrogen bonding. Among all 929 data points, absolute percent error of this model is greater than 10% only in 24 alcohols and less than 5% in 714 cases. It has the following form: FP = 233.7 + 11.67nC − 2.028nH + 18.08nO + 17.13ðnCl + nBr + nI Þ + 33.74HBD − 18.56CF

(1:20)

where nO, nCl, nBr, and nI are the number of oxygen, chlorine, bromine, and iodine atoms, respectively; HBD stands for “Hydrogen Bonding Donors” in intermolecular hydrogen bondings that are the number of H atoms attached to O and N atoms, i.e., –OH, –NH2, and >NH functional groups. Final parameter CF stands for “Correction Factor” to consider the effects of specific intramolecular attractions, branches, heavier atoms, and long chain alkyl groups on raising or lowering the FPs. The values of CF are defined as follows: (a) The attachment of –OH, –NH2 or >NH to tertiary carbon: The value of CF equals to the number of hydrogens of these functional groups, which are affected by tertiary carbon steric hindrance. For example, the reported value of the FP of 1-butanol is in the range of 302.0–319.3 K but for tert-butanol, the FP value is between 282.1 and 288.7 K. (b) The attachment of large alkyl groups to –OH: The value of CF equals –1.0 for alcohols containing a long chain saturated alkyl group with more than 11 carbon atoms. For unsaturated aliphatic alcohols with more than 11 carbon atoms, the value of CF equals +3. (c) The presence of C–S–C or –SH groups in aliphatic alcohols: In these cases, CF is equal to –2.0. Example 1.14: Calculate FP for 2-butyn-1-ol with the following molecular structure: HO

Answer: FP = 233.7 + 11.67nC − 2.028nH + 18.08nO + 17.13ðnCl + nBr + nI Þ+ 33.74HBD − 18.56CF = 233.7 + 11.67 × 4 − 2.028 × 6 + 18.08 × 1 + 17.13 × 0 + 33.74 × 1 − 18.56 × 0 = 320.0K The measured value of FP for this compound is 324.8 K [74].

58

1 Flash Point

1.2.3.7 Pure organic chemicals from structural contributions Rowley et al. [40] used experimental data of the FP of more than 1,000 organic compounds to present a suitable method for predicting their FP based solely on structural contributions. The correlation has the following form: Alcohols or phenols: P FP =

ni gi − 208.30

i

2.40 × lnð8nC + 8nSi + 8nS + 2nH − 2nX − 4nO Þ + 1

+ 196.68

(1:21)

+ 235.21

(1:22)

Other organic compounds: P FP =

ni gi − 510.49

i

2.13 × lnð8nC + 8nSi + 8nS + 2nH − 2nX − 4nO Þ + 1

where ni is the number of the gi structural group contribution in the molecule. Equations (1.21) and (1.22) can be used for a wide range of organic compounds but their reliability is lower than previous QSPR models based on elemental composition and structural parameters. Table 1.10 gives structural contributions for eqs (1.21) and (1.22). Example 1.15: Calculate the FP for 2,2-dimethyl-1-propanol, which was selected from Table 1.11, with the following molecular structure: HO

Answer: Since this compound is an alcohol, eq. (1.21) should be used as: Table 1.10: Structural contributions for eqs (1.19) and (1.20). I

Group

gi

Example

1 2 3 4 5 6 7 8 9 10 11 12 13

≡C– (HC) ≡CH (HC) =C< (HC) =CR< (HC) =CH– (HC) =CRH– (HC) =CH2 (HC) >C< (HC) >CR< (HC) >CH– (HC) >CRH– (HC) >CH2 (HC) –CRH2– (HC)

256.43 261.94 483.40 378.53 219.78 124.16 299.53 561.32 98.67 418.55 313.87 191.61 122.22

2-Pentyne 1-Hexyne 2-Methyl-1-octene d-Limonene 1-Pentene Cyclohexene 1-Pentene Neopentane 1,1-Diethylcyclohexane 4-Methylheptane Methylcyclohexane Pentane Cyclopentane

59

1.2 Predictive methods of the FP for pure compounds

Table 1.10 (continued) I

Group

14 15 16 17 18 19 20 21 22 23 24 25 26

–CH3 (HC) >CH– >CRH– >CH2 –CRH2– –CH3 =C< =CR< =C= =CH– =CRH– =CH2 CR–CR= (fused ring) >C
CR< >C=O

36 37 38 39 40 41 42 43 44 45 46 47

>CR=O O=CH– (aldehyde) O=CROR– –COO– (ester, nonring) –COOH (acid) O=CRORCR=O (aliphatic ring) =O –O– –OR– –OH (alcohol) –OH (phenol) >N– >NH >NRH –NH2 –N= –NR= >NR–

48

–N–Ca

32 33

34 35

gi

Example

259.62 119.79 201.98 162.43 149.72 77.80 194.11 236.45 2239.01 148.59 163.28 37.56 259.24

Propane Isobutyl Formate Cyclohexanol Butanol Cyclohexanol Ethanol Chloroprene Phenol Carbon Disulfide Acrolein Maleic Anhydride Methacrolein a-Pinene

108.68

494.20 551.77 437.19

Acetone Cyanohydrin Isophorone Diisocyanate 2-Pentanone Cyclopentanone Propanal

1192.63 529.37

Diketene Ethyl Acrylate

1034.70 1750.35

Thioglycolic Acid Succinic Anhydride

623.68 176.69 128.89 803.82 806.21 153.69 354.79 325.82 362.58 196.10 243.93 369.80

Di-n-propyl Sulfone Methyl Ethyl Ether Tetrahydrofuran Propanol Nonylphenol Tripropylamine Diisopropylamine Ethyleneimine Urea Acetaldoxime Oxazole 4-(2-Aminoethyl) Morpholine Diphenylamine

130.20

797.35

(continued)

60

1 Flash Point

Table 1.10 (continued) I

Group

gi

Example

49 50 51 52

–CN –NC=O O=C=N–Ca NO2–C– (aliphatic) –NO2 –S– –SR– –SH –Br –Cl –F –I –Si– –O–(Si)

640.67 1193.67 697.80 898.17

Ethyl Cyanoacetate n-Methylacetamide Phenyl Isocyanate Nitroethane

525.91 405.65 221.89 469.16 386.51 251.85 255.41 622.38 89.55 96.01

m-Nitrotoluene Diethyl Sulfide Thiophene Propyl Mercaptan Bromoethane 1-Chloropentane Benzotrifluoride Hexyl Iodide Dimethyldichlorosilane Hexamethyldisiloxane

53 54 55 56 57 58 59 60 61 62

HC indicates the group only applies to hydrocarbons, subscript R is an atom belonging to any ring, and Ca is explicitly an aromatic carbon.

Table 1.11: Sample calculations illustrating the estimation of the FP from eqs (1.19) and (1.20). Chemical Terpinolene

Groups

3, 4 (×2), 6, 13 (×3), 14 (×3) m-Diethylbenzene 4 (×2), 6 (×4), 12 (×2), 14 (×2) 2,2-Dimethyl-1-propanol 17, 19 (×3), 27, 39 p-Cresol 19, 21 (×2), 24 (×4), 40 Azelaic acid 17 (×7), 34 (×2) Dimethylethanolamine 17 (×2), 19 (×2), 39, 41 Tetradecamethylhexa 19 (×14), 61 (×6), 62 (×5) siloxane Propylene glycol 15, 17 (×2), 19, 23, 25, 37, monoallyl ether 39 2-Ethyl thiophene 17, 19, 21, 24 (×3), 55 Maleic anhydride 24 (×2), 30 (×2), 38

Group Sum

OH Group FPcalc (K) FPexp (K)

1552.42

329.5

329.15

1517.68

327.0

329.0

298.3 358.6 491.1 315.0 362.6

303.15 359.15 488.15 312.15 375.15

331.5

327.55

302.3 370.2

300.15 375.15

1308.33 2010.03 3206.41 1437.97 2106.55

39 40

1689.11

39

1188.41 1558.99

39

61

1.2 Predictive methods of the FP for pure compounds

P

ni gi − 208.30

i

FP =

2.40 × lnð8nC + 8nSi + 8nS + 2nH − 2nX − 4nO Þ + 1

=

+ 196.68

ð3 × 77.8 + 162.43 + 108.68 + 803.82Þ − 208.30 + 196.68 = 298.3 K 2.40 × lnð8 × 5 + 8 × 0 + 8 × 0 + 2 × 12 − 2 × 0 − 4 × 1Þ + 1

1.2.3.8 Organosilicon compounds It was indicated that the following correlation exists between FP of organosilicon compounds and their molecular structure [75]: FP ¼ 220:18 þ 13:27 ðnC  0:11nH þ 1:86nCl þ 0:54nO þ 0:67nN Þ þ 40:37FPISM 35:14 FPDSM

(1:23)

where FPISM and FPDSM are increasing and decreasing structural moieties of the FP. The values of FPISM and FPDSM are given in Table 1.12. Example 1.16: Calculate FP for the following compound: Cl

Cl Si

C4H8Cl2Si

Answer: FP ¼ 220:18 þ 13:27 ðnC  0:11nH þ 1:86nCl þ 0:54nO þ 0:67nN Þ þ 40:37FPISM  35:14 FPDSM ¼ 220:18 þ 13:27 ð4  0:11 × 8 þ 1:8 × 2 þ 0:54 × 0 þ 0:67 × 0Þ þ 40:37 × 0  35:14 × 0:5 ¼ 293:15 K The measured value is 300.15 K [43].

62

Table 1.12: The values of FPISM and FPDSM for some organosilicon compounds. Molecular structure

Condition

FPISM FPDSM Example

1.0

0

R1, R2, and R3 are alkyl groups with one or two carbon atoms

0.5

0

R1, R2, and R3 are alkyl groups with one or two carbon atoms and n < 2 and X= Br or I

1.0

0

R1, R2, and R3 are alkyl groups with one or two carbon atoms and Z is Chlorinen < 2 and X = Br or I

1.0

0

1.25

0

R1, R2, and R3 are alkyl groups with one or two carbon atoms

1 Flash Point

R1, R2, and R3 are alkyl groups with one or two carbon atoms

R1, R2, and R3 are alkyl groups with less than five carbon atoms

0

or

or 1.0

0

or

in X

X= alkyl group containing three carbon atoms with thiol (–SH) group

1.5

0

1.2 Predictive methods of the FP for pure compounds

X= alkyl group containing more than five carbon atoms with a substituent containing oxygen and more than four carbon atoms containing nitrogen atoms except for the existence of molecular fragment

1.0

(continued)

63

Molecular structure

64

Table 1.12 (continued) Condition

FPISM FPDSM Example 1 Flash Point

or X= alkyl group containing three carbon atoms with a substituent containing sulfur or nitrogen atoms

0.5

0

or

H-(SiH2)n-H

–

1.5

0

n5

0

0.5

0.25 (n-5)

(continued)

1.2 Predictive methods of the FP for pure compounds

X = alkene group with less than three carbon atoms

67

Molecular structure

68

Table 1.12 (continued) Condition

FPISM FPDSM Example 1 Flash Point

n CH2 ðringÞ + ν > CH − ðringÞ f > CH − ðringÞ = 8 × ð−28.4401Þ + 2 × ð−6.7179Þ = −240.957


AIT = 750.3 +

P


2
3 P P νi fi − 8.644 × 10−4 νi fi − 4.5604 × 10−6 νi fi

i

i

i

= 750.3 + ð−240.957Þ − 8.644 × 10−4 ð−240.957Þ2 − 4.5604 × 10−6 ð−240.957Þ3 = 523.0 K

b)

P i

νi f i = ν > CH2 ðringÞ f > CH2 ðringÞ = 10 × ð−28.4401Þ = −284.401


AIT = 750.3 +

P


2
3 P P νi fi − 8.644 × 10−4 νi fi − 4.5604 × 10−6 νi fi

i

i −4

i −6

= 750.3 + ð−284.401Þ − 8.644 × 10 ð−284.401Þ − 4.5604 × 10 ð−284.401Þ3 = 500.9 K

2

83

2.2 Predictive methods of the AIT for pure compounds

Table 2.1: The structural group contribution values for estimation. fi

Group

fi 88.8289 79.4122 53.4870 −8.9378 134.3524 −70.0383 −28.4801 8.1173 57.5044 −138.3186

35 36 37 38 39 40 41 42

–F (ring) –Cl (ring) –Br (ring) –OH (alcohol) –OH (phenol) –O– (non-ring) –O– (ring) >C=O (non-ring) >C=O (ring) –HC=O (aldehyde) –COOH (acid) –COO– –NH2 >NH (non-ring) >NH (ring) >N– (non-ring) >N– (ring) –N= (non-ring)

−9.8344

43

–N= (ring)

31.8743

−24.2759

44

–CN

80.5038

293.5064

45

–NO2

−52.7670

Serial No.

Group

1 2 3 4 5 6 7 8 9 10

–CH3 >CH2 >CH– >C< =CH2 =CH– =C< CH C– >CH2 (ring)

−22.8857 −28.5961 1.3340 49.7423 −21.6668 −46.3286 −32.7605 −81.0169 −64.4957 −28.4401

25 26 27 28 29 30 31 32 33 34

11 12 13 14 15 16 17 18

>CH– (ring) >C< (ring) =CH– (ring) =C< (ring) =CH– >C (fused) >C (non-fused) –CH3 (attached to at least one halogen atom) >CH2 (attached to at least one halogen atom) >CH– (attached to at least one halogen atom) >C< (attached to at least one halogen atom) –F (non-ring) –Cl (non-ring) –Br (non-ring)

−6.7179 −21.7342 19.5293 −49.068 6.2350 5.8332 15.9976 103.2738

19 20 21 22 23 24

c)

P i

Serial No.

−45.0477 33.9332 −27.8628

νi f i = ν > CH2 ðringÞ f > CH2 ðringÞ + ν = CH − ðringÞ f = CH − ðringÞ

= 3 × ð−28.4401Þ + 2 × ð19.5293Þ = −46.2617

2
3 P −4 P −6 P AIT = 750.3 + νi fi − 8.644 × 10 νi fi − 4.5604 × 10 νi fi i

i −4

i −6

= 750.3 + ð−46.2617Þ − 8.644 × 10 ð−46.2617Þ − 4.5604 × 10 ð−46.2617Þ3 = 702.6 K

2

4.0037 35.2011 −17.7579 −1.8223 24.5474 −4.7926 −49.3834 −41.9897

84

d)

2 Autoignition

P

νi f i = ν=CH − ðringÞ f =CH− ðringÞ + ν=C < ðringÞ f =C< ðringÞ = 8 × ð19.5293Þ

i

+ 2 × ð−49.068Þ = 58.0984

2
3 P P P νi fi − 8.644 × 10−4 νi fi − 4.5604 × 10−6 νi fi AIT = 750.3 + i

i

i

−4

2

−6

= 750.3 + ð58.0984Þ − 8.644 × 10 ð58.0984Þ − 4.5604 × 10 ð58.0984Þ3 = 804.6 K

e)

P i

νi f i = ν−Cl ðnon − ringÞ f −Cl ðnon − ringÞ + ν=C < f =C < + ν=CH − f =CH − = 3 × ð33.9332Þ + 1 × ð−32.7605Þ + 1 × ð−46.3286Þ = 22.7105

AIT = 750.3 +


P i


2
3 P P νi fi − 8.644 × 10−4 νi fi − 4.5604 × 10−6 νi fi i

i

= 750.3 + ð22.7105Þ − 8.644 × 10−4 ð22.7105Þ − 4.5604 × 10−6 ð22.7105Þ3 2

= 772.5 K

2.2.2 A simple QSPR model for various classes of hydrocarbons A simple model was introduced for estimating the AITs of different classes of hydrocarbons, including alkanes, alkenes, cycloalkanes, cycloalkenes, alkynes, and aromatics. It contains four variables as follows [134]: AIT = 647 + 33.33nC − 20.79nH + 58.20F SH + 81.03F BH

(2:2)

Two functions FSH and FBH are related to size and branches of different classes of hydrocarbons, respectively. Function FSH can control the rate of changing of the AIT on the basis of nC and nH for molecules in homologous series of hydrocarbons. Function FBH can adjust a high degree of branching in a hydrocarbon molecule. Different molecules with the same nC have a tendency to react by decreasing nH in the following order: aromatics < cyclics < alkenes < alkanes This general structural sequence is consistent with the ease of free radical formation of the mentioned hydrocarbons. This situation confirms foundation of previous works [139, 140, 152]: the AIT mechanism proceeds by a free radical reaction. Two functions FSH and FBH can also affect the ease of free radical formation, and consequently of oxidation. Thus, some of the structural features can affect the mechanism

2.2 Predictive methods of the AIT for pure compounds

85

of autoignition that includes chain length, the addition of methyl groups, unsaturation, branching, and strain. Two functions FSH and FBH can be specified according to the following situations. 2.2.2.1 FSH The value of FSH depends on the size of hydrocarbon and has some conditions for each class: (i) Saturated acyclic hydrocarbons: This condition is applied only for linear alkanes. For n-paraffins with nC < 5, FSH = 5.5 − nC . The value of FSH equals –1.0 for 5 ≤ nC ≤ 10. For 11 < nC ≤ 17, FSH = 0.0. For n-alkanes with nC > 17, AIT is constant and its value is 475 K. (ii) Saturated cyclic hydrocarbons: This state is valid only for cyclic compounds without any substituent. For nC ≤ 5, FSH = 5.5 − nC . The value FSH equals –2.0 for nC ≥ 6. (iii) Alkenes: (a) For the compounds with general formula CH2 = CR1R2, two different cases can be considered: (1) If R1 and R2 are methyls, FSH equals 2.0. (2) If one of the alkyl groups is methyl and the other one is a nonlinear alkyl group, FSH is 1.0. (b) For CH2 = CH(CH2)nCH3, where n = 2 to 8, FSH equals –1.0. For n > 8, AIT is constant and its value is 515 K. (c) For linear alkenes with general formula R1CH = CHR2 except both R1 and R2 are –CH3 groups, FSH equals –1.0. (d) For the existence of two double bonds except 1,3-butadiene, FSH is –1.0. (iv) Alkynes: (a) For the existence of triple bond, FSH is –1.5. (b) The value of FSH equals –2.2 for the presence of both triple and double bonds simultaneously.

2.2.2.2 FBH (i) Saturated acyclic hydrocarbons: The value of FBH depends on the number of alkyl substituents (nalkyl sub) attached to the longest continuous chain of carbon atoms (LCC): (a) If nalkyl sub >2, FBH = 2.0. (b) If nalkyl sub = 2, the values of FBH are 2.0 and 0.8 for LCC 1

1.0

0

n≤1

0

1.5

R'

F F O F R

O

F

O

F

,

–

R

(continued)

94

2 Autoignition

Table 2.5 (continued ) Organic compounds containing etheric groups

AIT AITþ SFG SFG

Cyclic ether

2.5

0

0

1.0

Condition More than two ether functional groups The presence of two –O-CH2CH2-O- groups

þ AIT ¼ 584:07 þ 24:36 nC  15:57 nH þ 115:25 AITSFG  116:02AIT SFG ¼ 584:07 þ 24:36 ð7Þ  15:57ð16Þ þ 115:25 ð2:5Þ  116:02ð0Þ ¼ 793:6 K

The calculated values of AIT by Albahri-George [131] and Chen et al. [132] are 472.3 and 506.0 K, respectively. Since the experimental value of AIT for 1-(1-methoxypropan-2yloxy)propan-2-ol is 805.2 K [160], the predicted value of the AIT by eq. (2.4) is closer to the reported value than the methods of Albahri-George [131] and Chen et al. [132].

2.2.5 Reliable prediction of autoignition temperature of organic hydroxyl compounds A simple model has been introduced for estimating the AIT of different classes of organic hydroxyl compounds containing the other polar groups such as –O–, –S–, –CN, –C(=O)O–, –NH2, and >NH [162]. For the presence of both –O– and –OH groups, eq. (2.4) can also be used. This model is based on the number of hydrogen atoms and two correcting functions under certain conditions, which can be given as follows: AIT ¼ 731:9  9:1 nH þ 138:7AITCorr; In  145:7AITCorr; Dec

(2:5)

where AITCorr; In and AITCorr; Dec are correcting functions for increasing and decreasing the predicted AIT based on nH . Since the coefficient of nH has a negative sign, different molecules of organic hydroxyl compounds with the same number of carbon atoms tend to increase AIT by decreasing nH molecular moieties in the following order: aromatic substituents < cyclic derivatives < alkenes < alkanes This situation confirms that the AIT mechanism proceeds by a free radical reaction [141, 161]. Two parameters “AITCorr; In ” and “AITCorr; Dec ” can also affect the ease of free radical formation because they affect the mechanism of autoignition through molecular moieties

2.2 Predictive methods of the AIT for pure compounds

95

[140, 161]. The values of AITCorr; In and AITCorr; Dec were found based on molecular fragments as follows: (i) R–O-CHxCH2OH and R-OCH2CH2CH(OH)CH3 – Several situations should be considered: (a) R–O-CH2CH2OH – The value of AITCorr; Dec equals 0.8 if R is (1) alkyl group up to three carbon atoms, (2) containing CH3(CH2)n=0,1,2,. . . as well as the presence of oxygen atoms up to two or only –OH group. The AITCorr; Dec equals 0.4 if R is (1) linear alkyl group containing more than three carbon atoms, (2) including CH3(CH2)n>2- as well as the presence of oxygen atoms of more than two. The values of AITCorr; In equal 0.9 and 0.5 if R includes an aromatic ring and the other HO-CH2-CH2-O- molecular fragment, respectively. (b) R-OCH2CH(OH)CH3 – The values of AITCorr; Dec are equal to 1.5 and 0.7, if R includes aromatic and linear alkyl up to two carbon atoms, respectively. If R includes a tertiary carbon atom attached to an oxygen atom such as tertbutoxy, AITCorr; Dec equals 0.9. (c) Cyclic organic hydroxyl compounds – The values of AITCorr; Dec and AITCorr; In are 0.2 and 1.0, respectively, for the attachment of –OH to cyclohexane (with or without methyl substituent) and higher membered rings. The value of AITCorr; Dec is 1.0 for the attachment of –CH2OH to tetrahydrofuran ring. (d) Attachment of –OH or –CH2OH to tertiary carbon or furan cycle – The value of AITCorr; In equals 0.5 except if all substituents of tertiary carbon atoms have less than four carbon atoms. The value of AITCorr; In is 0.9 if alkyl substituents of tertiary carbon atoms have more than or equal to four carbon atoms. (e) Cyanide or >CHCOCH< – The values of AITCorr; In are 2.0 and 0.5 for the presence or absence of –OH attached to a tertiary carbon atom. (f) The existence of –CH2-CH2-NH-CH2-CH2-, -CH(OH)-CH2-CH2-CH(OH)-, H-C≡C-, and –CH(OH)-CH(OH)-CH2-CH(OH) – For the presence of only molecular moieties –CH2-CH2-NH-CH2-CH2- and -CH(OH)-CH2-CH2-CH(OH)- without further functional groups, the values of AITCorr; Dec are 2.1 and 1.1, respectively. Meanwhile, the values of AITCorr; Dec are 2.1 and 1.1, respectively, for the presence of only molecular fragments H-C≡C- and –CH(OH)-CH(OH)-CH2CH(OH)- without further functional groups. Example 2.5: Calculate the AIT of hexane-1,6-diol and compare its result of eq. (2.5) with the predicted results of Albahri-George [131] and Chen et al. [132]. Answer: Hexane-1,6-diol has the following molecular structure: OH

HO Chemical Formula: C6H14O2

96

2 Autoignition

The use of eq. (2.5) gives AIT ¼ 731:9  9:1 nH þ 138:7AITCorr; In  145:7AITCorr; Dec ¼ 731:9  9:1 ð14Þ þ 138:7ð0Þ  145:7ð0Þ ¼ 604:5 K Using methods of Albahri-George [131] and Chen et al. [132] gives 532.1 and 560.8 K, respectively. The predicted value of AIT by eq. (2.5) is closer to the reported value, that is, 593.2 K [162].

2.3 Autoignition and ignition delay The autoignition process affects the performance of several combustion devices, such as internal combustion engines (ICEs) because it governs compression ignition (CI) combustion and limits the compression ratio in spark ignition (SI) engines to avoid knocking. There is an ignition delay time between the start of injection and the start of combustion during which some of the injected fuel evaporates and mixes with the cylinder charge. The ignition delay includes physical and chemical phenomena. Physical phenomena contain atomization and air mixing in the case of CI and gasoline direct injection engines. Meanwhile, chemical phenomena describe oxidation chemistry, which they lead to very high premixing ratios and low combustion temperatures (LCT modes) [163]. Due to high exhaust gas recirculation (EGR) rates and/or diversity of fuel reactivities (blends or dual fuel), they are kinetically controlled and require a reliable determination of the expected ignition timing. Autoignition depends on the engine type and operating conditions, the fuel/air ratio, and the charge composition as well as the fuel chemical structure. For researchers and the automotive industry, the understanding of autoignition and the accurate prediction of the ignition delay time for organic compounds as fuels are interesting. Ignition delay has wide applications for liquid propellants [128–130]. For low-to-medium temperature conditions, hydrogen abstraction (usually by OH•) dominates for the autoignition of alkanes. Hydrogen abstraction depends on the type of carbon to which H is linked in the absence of the other atoms [164]. The paraffin length as well as the degree and type of branching establish the distribution of C–H bonds, which can affect autoignition significantly. The reactivity of paraffin increases and decreases with the length of the carbon chain and the degree of branching, respectively. Moreover, the reactivity also depends on the molecular structure of paraffin. Minetti et al. [165] compared a set of autoignition data of linear (n-butane, n-pentane, and n-heptane) and branched (2,2-dimethylpropane and 2,2,4trimethylpentane) alkanes in a rapid compression machine (RCM). They found that autoignition delay times increase when decreasing the carbon chain length and with the presence of branches. This foundation is also confirmed by the other authors, for example, the results of experiments of Silke et al. [166] and Li et al. [167] confirmed

2.3 Autoignition and ignition delay

97

that the presence of branches reduces the reactivity, delaying the ignition and decreasing the burning rates. Won et al. [168] investigated the reactivity of large paraffinic fuels to the ratio of methylene (CH2) to methyl (CH3) molecular groups and confirmed that the reactivity of such fuels increases with the relation (CH2/CH3)×[CH2+CH3]. Lapuerta et al. [169] demonstrated that the ignition delay increases with the number of methyl branches, and the reactivity of the tested fuels, including n-hexadecane, 2,6,10-trimethyltridecane, and 2,2,4,4,6,8,8,-heptamethyl-nonane, as well as their mixtures, is related to the CH2/CH3 mass ratio. Most of the correlations for the calculation of the ignition delay are based on Arrhenius-type expressions for a specific fuel. They consider the effect of pressure, temperature, and occasionally the oxygen concentration and the relative fuel/air ratio [167, 170–172]. Other correlations or models related the ignition tendency to the fuel chemical structure. Hernández et al. [173] explored the relationship between autoignition and paraffin length and structure, from 10 to 18 carbons at conditions relevant to diesel engine operation. They proposed an Arrhenius-type correlation containing the effect of pressure, temperature, relative fuel/air ratio, and main chemical groups. The assessment of the ignition quality of a wide range of organic compounds is one key challenge in the identification of novel molecular entities qualifying as biofuels or biofuel blend components derived from oxygen-rich lignocellulosic feedstocks. It is difficult to obtain high-quality experimental ignition delay data because not only the type of equipment but also the experimental conditions influence the ignition behavior. Mixture formation, air temperature, and pressure in the combustion chamber should be carefully controlled during the investigation of a fuel candidate. Dahmen and Marquardt [174] summarized the results from a comprehensive experimental screening campaign targeting a diverse set of pure-component organic fuels and their ignition characteristics in an ASTM D6890 Ignition Quality Tester (IQT) [175]. IQT has constant-volume combustion chamber experiment and rapid screening potential. Dahmen and Marquardt [174] introduced a suitable group contribution modeling to unravel relationships between the ignition delay observed in IQT experiments and the fuel’s molecular structure. Their approach covers a wide range of organic compounds, including acyclic and cyclic, branched and straight, saturated and unsaturated hydrocarbons as well as alcohols, ethers, esters, ketones, aldehydes, and aromatic and polyfunctional compounds. For calculation of ignition delay of organic fuel, several steps should be done: (i) Evaluation of vapor pressure of a desired fuel – Dahmen and Marquardt [174] assumed the existence of a correlation between the vapor pressure pS (in the bar at 298 K) and the physical contribution to the IQT ignition delay. Since experimental data on pS may not always be available, they used the estimation of pS based on the Hoffmann–Florin equation [176, 177]. The method of the Hoffmann–Florin can calculate the value of pS (in the bar at 298 K) as follows:

98

2 Autoignition

fNBP ¼

fTc ¼

1  7:9151 × 103 þ 2:6726 × 103 log10 ðNBPÞ  0:8625 × 106 NBP NBP

(2:6)

1  7:9151 × 103 þ 2:6726 × 103 log10 ðTc Þ  0:8625 × 106 Tc Tc

(2:7)


 fNBP 1:01325 α ¼ 11:52608845  ln fNBP  fTc pc β¼

(2:8)

lnð1:01325=pc Þ fNBP  fTc

(2:9)

 pS ¼ 10 exp α þ 1:7962 × 103 β  13:81551056

(2:10)

where NBP, Tc ; and pc are the normal boiling point (in K), the critical temperature (in K), and the critical pressure (in bar), respectively. The method of Joback’s SGC [178] can be used to calculate NBP, Tc ; and pc from the following equations: NBP ¼ 198:2 þ 2

X

NBPi

(2:11)

i

Tc ¼ NBP40:584 þ 0:965

X

Tc;i 

X

i

pc ¼

0:113 þ 0:0032nA 

!2 31 Tc;i 5

(2:12)

i

X

!2 pc;i

(2:13)

i

where nA is the total number of atoms. Table 2.6 provides the values of NBPi , Tc;i , and pc;i based on the contribution of different groups in the desired organic compound. IQT ignition delay model for compound i is given by several model equations, and inputs for calculation of its IQT ignition delay are as follows:  lnðτi Þ ¼ exp carboni þ oxygeni þ oxygenðringÞi þ descriptori þ 0:8341

(2:14)

where τi is IQT ignition delay in ms. The parameter carboni is related to the number of −CH3 (nCH3 ;i ) and nonring −CH2− groups (nCH2 ðnonringÞ;i ) as follows:   carboni ¼ 0:0449nCH3 ;i  0:2389 ln nCH2 ðnonringÞ;i þ 1

(2:15)

For organic compounds containing oxygen atoms, two parameters “oxygeni ” and “oxygenðringÞi ” are related to the number of –OH groups (nOH;i ), O=CH- (nO¼CH;i ),

99

2.3 Autoignition and ignition delay

Table 2.6: The values of NBPi , Tc;i , and pc;i . Group Nonring increment -CH3 -CH2>CH>C< =CH2 =CH=C< =C= ≡CH ≡CRing increments -CH2>CH>C< =CH= =C< Halogen increments -F -Cl -Br -I Oxygen increments -OH (alcohol) -OH (phenol) -O- (nonring) -O- (ring) >C=O (nonring) >C=O (ring) O=CH- (aldehyde) -COOH (acid) -COO- (ester) =O (except as above) Nitrogen increments -NH2 >NH (nonring) >NH (ring) >N- (nonring) -N= (nonring) -N= (ring) =NH -CN -NO2

Tc;i

pc;i

23.58 22.88 21.74 18.25 18.18 24.96 24.14 26.15 9.20 27.38

0.0141 0.0189 0.0164 0.0067 0.0113 0.0129 0.0117 0.0026 0.0027 0.0020

–0.0012 0 0.0020 0.0043 –0/0028 –0.0006 0.0011 0.0028 –0.0008 0.0016

27.15 21.78 21.32 26.73 31.01

0.0100 0.0122 0.0042 0.0082 0.0143

0.0025 0.0004 0.0061 0.0011 0.0008

-0.03 38.13 66.86 93.84

0.0111 0.0105 0.0133 0.0068

–0.0057 –0.0049 0.0057 –0.0034

92.88 76.34 22.42 31.22 76.75 94.97 72.24 169.09 81.10 -10.50

0.0741 0.0240 0.0168 0.0098 0.0380 0.0284 0.0379 0.0791 0.0481 0.0143

0.0112 0.0184 0.0015 0.0048 0.0031 0.0028 0.0030 0.0077 0.0005 0.0101

73.23 50.17 52.82 11.74 74.60 57.55 83.08 125.66 152.54

0.0243 0.0295 0.0130 0.0169 0.0255 0.0085 – 0.0496 0.0437

0.0109 0.0077 0.0114 0.0074 –0.0099 0.0076 – –0.0101 0.0064

NBPi

(continued)

100

2 Autoignition

Table 2.6 (continued ) Group

NBPi

Sulfur increments -SH -S- (nonring) -S- (ring)

63.56 68.78 52.10

Tc;i 0.0031 0.0119 0.0019

pc;i 0.0084 0.0049 0.0051

nonring –O- (nOðnonringÞ;i ), nonring >C=O (n > C ¼ OðnonringÞ;i ), nonring –COO(nCOOðnonringÞ;i ), ring –O- (nOðringÞ;i ), and ring >C=O (n > C ¼ OðringÞ;i ) groups as follows: oxygeni ¼ 0:7286nOH;i  0:0401nO ¼ CH;i  0:2123nOðnonringÞ;i þ 0:3649n > C¼OðnonringÞ;i þ 0:5260nCOOðnonringÞ;i oxygenðringÞi ¼ 0:0426nOðringÞ;i þ 0:4241n > C¼OðringÞ;i

(2:16)

(2:17)

The parameter descriptori is a function of the number of aromatic bonds (nAr bonds;i ), carbon–carbon double bonds (ndouble bonds;i ), quaternary carbon atoms, or carbon atoms that are attached to four other carbon atoms through single bonds where both nonring and ring quaternary carbons are counted (nquat;i ), and the vapor pressure pSi (in the bar at 298 K). It is given as follows: descriptori ¼ 0:0639nAr bonds;i þ 0:1404ndouble bonds;i þ 0:1923nquat;i þ 0:0429pSi (2:18) For calculation of ignition delay of an organic compound as a fuel, the order of the mentioned equations can be used as follows: (i) Equations (2.11)–(2.13) and Table 2.6 for calculation of NBP, Tc , and pc : (ii) Equations (2.6)–(2.10) for calculation of pSi : (iii) Equations (2.15)–(2.18) for evaluation of carboni , oxygeni , oxygenðringÞi , and descriptori : (iv) The use of eq. (2.14) for the calculation of τi : Example 2.6: Calculate the τi of (a) 2,2-dimethylbutane and (b) 3-methoxy-3-methyl-1butanol. Answer: (a) 2,2-Dimethylbutane has the following molecular structure:

2.4 Summary

101

(i) The use of eqs. (2.11)–(2.13) and Table 2.6 gives NBP=333.6 K, Tc =508.3 K, and pc =31.74 bar. (ii) The use of eqs. (2.6)–(2.10) gives pSi =0.2802 bar. (iii) The use of eqs. (2.15)–(2.18) gives carboni =0.01401, oxygeni =0, oxygenðringÞi =0, and descriptori =0.1297. (iv) The use of eq. (2.14) gives τi =14.28 ms. The measured IQT ignition delay is 9.82 ms [174]. (b) 3-Methoxy-3-methyl-1-butanol has the following molecular structure:

HO

O

(i) The use of eqs. (2.11)–(2.13) and Table 2.6 gives NBP=448.2 K, Tc =619.2 K, and pc =34.60 bar. (ii) The use of eqs. (2.6) to (2.10) gives pSi = 0.00084 bar. (iii) The use of eqs. (2.15)–(2.18) gives carboni = –0.08286, oxygeni = 0.7286, oxygenðringÞi = 0, and descriptori = –0.1562. (iv) The use of eq. (2.14) gives τi = 10.99 ms. The measured IQT ignition delay is 45.15 ms [174]. As shown in part (b), the calculated τi is much lower than the measured IQT ignition delay, suggesting that the presence of some specific molecular moieties may provide large deviations.

2.4 Summary Since experimental values of the AIT of different classes of compounds are scarce and expensive, this chapter introduces different approaches for prediction of the AIT of some classes of materials. This chapter introduces two different approaches of SGC and QSPR methods for prediction of the AIT of organic compounds and different types of hydrocarbons. Among available SGC and QSPR methods, the method of Chen et al. [132] and eqs (2.2) and (2.3) were demonstrated because they can be used for a wide range of different categories of pure compounds. Available QSPR methods require complex descriptors, computer codes, and expert users, except for eqs (2.2) and (2.3) that are based on molecular structures of hydrocarbons. Equation (2.2) can be used to predict the AITs of different classes of hydrocarbons containing alkanes, alkenes, cycloalkanes, cycloalkenes, alkynes, and aromatics. An elemental composition, as well as two functions FSH and FBH in eq. (2.2) to (2.5), can be easily found from molecular structures of hydrocarbons. As compared to three of the best available SGC

102

2 Autoignition

methods, that is, Albahri [56], Albahri and George [131], and Chen et al. [132], eq. (2.2) provides more reliable results for different types of hydrocarbons. Equation (2.3) can also be used for prediction of the AIT of organic energetic compounds containing functional groups nitro, nitrate, nitramine, and peroxide. Equation (2.4) assesses the AIT of organic ether compounds with high reliability. Equation (2.5) can also provide a simple model for estimating the AIT of different classes of organic hydroxyl compounds containing the other polar groups such as –O–, –S–, –CN, –C(=O)O–, –NH2, and >NH. As compared to three of the best available SGC methods, that is, Abahri [56], Albahri and George [131], as well as Chen et al. [132], eqs. (2.2)–(2.5) provide more reliable results. Section 2.3 provides a simple method for calculation of the IQT ignition delay model of compound i. This approach can be applied to a wide range of organic compounds, including acyclic and cyclic, branched and straight, saturated and unsaturated hydrocarbons, as well as alcohols, ethers, esters, ketones, aldehydes, and aromatic and polyfunctional compounds.

3 Flammability Limit It is suitable to keep a combustible compound outside of its flammable concentration range. As mentioned in Chapter 1, the flash point can suggest an approximation of the lower temperature limit in which a chemical evolves enough vapors to support combustion. Flammability limits, which are also referred to as the explosive limits, represent the concentrations of fuel in air, which can support flame propagation. They give the range of fuel concentration usually in terms of percentage volume in the air usually at 298 K where a gaseous mixture can ignite and burn. They are better descriptors of a chemical’s flammability in solids, liquids, and gases. A considerable amount of flammability limit data has been published, but the temperature dependence of the limits is nearly always neglected. Available data are frequently reported at 298 K for gases. For liquids and solids, flammability limit data are often reported at a single arbitrary temperature. The upper and lower flammability limits are some of the most important for safety considerations in storage, processing, and handling, which should be considered in assessing the overall flammability hazard potential of a chemical substance. They are defined in terms of the lower flammability limit (LFL) and the upper flammability limit (UFL) as the degree of susceptibility to ignition or release of energy under varying environmental conditions. Thus, the LFL and UFL show the minimum and maximum concentrations of a combustible substance that is capable of propagating a flame in a homogeneous mixture of the combustible and a gaseous oxidizer (air) under the specified conditions of the test. Below the LFL, there is not enough fuel to cause ignition. If fuel concentration is greater than the UFL, there is insufficient oxygen for the reaction to propagate away from the source of ignition. Determination of the LFL and UFL depends on several variables, e.g., the type of the fuel or chemical, the geometry of the apparatus, strength of the ignition source, test pressure, degree of mixing, oxygen concentration, and concentration of diluents [179].

3.1 Measurement of the LFL and UFL Differences in apparatuses and experimental methods can influence the measured flammability limits significantly. It is important to use a suitable procedure for the precise determination of the flammability limits. This requirement can be established by the use of a standardized apparatus and conditions as specified in ASTM standard E681 [180], which is shown in Figure 3.1. The values of the LFL and UFL are determined by igniting a uniform mixture of a gas or vapor with air in a closed vessel. In the ASTM standard test method, the upward and outward propagation of the flame, away from the ignition source, is noted by visual observation. The concentration of the desired compound is varied between trials. The composition is https://doi.org/10.1515/9783110782134-003

104

3 Flammability Limit

Rubber stopper

Contineous flames along at least 90° arc

Glass flask

Electrodes

Magnetic stirrer Figure 3.1: Visual criterion for flask flame propagation (ASTM E 681) [180].

determined where propagation of the flame is just sustained. The LFL and UFL may be used to determine guidelines for the safe handling of volatile chemicals. They assess ventilation requirements for the handling of gases and vapors.

3.2 Predictive methods of the flammability limits Determination of the flammability limits is difficult and not always practical. Thus, a reliable predictive method that is desirably convenient and fast must be used to estimate them. The LFL and UFL are two of the macroscopic properties of compounds, which are related to the molecular structure. The magnitude and predominant types of the intermolecular forces depend on the molecular structure of the desired material. Four different approaches – empirical methods, QSPR based on complex molecular descriptors, ANN-SGC, and SGC – were usually used for estimation of the LFL and UFL [20, 56, 181–218]. The method of SGC suggests that a macroscopic property can be calculated from group contributions. It was indicated that the SGC approach based on non-linear regression and least square techniques suffers from several shortcomings because these limitations are mainly associated by using a simple correlation, which is unable to capture the complex nature of the LFL

3.2 Predictive methods of the flammability limits

105

and UFL properties. Between two properties of the LFL and UFL, the LFL is more important than the UFL. Because of the complex dependency of the LFL on the molecular structure of the compound, it is a difficult property to estimate or correlate. For the coexistence of several functional groups in one molecule, it is difficult to formulate a model that can incorporate the behavior of all the different groups without taking into account the structure of the molecules. For designing safe chemical and petrochemical processes, accurate knowledge of the LFL for a variety of chemicals is needed. The measured LFL and UFL are little for a wide range of chemicals. Moreover, experimental values of the LFL and UFL are rare for most chemicals at non-ambient conditions. For pure chemicals at a single temperature point, usually 298 K, many methods have been developed to estimate their LFL and UFL [20, 56, 181–217]. Vidal et al. [20] reviewed a few of these methods. Available QSPR methods and ANN-SGC are complicated methods, which require computer codes and expert users. In this chapter, the simplest approach for estimating the temperature dependence of the LFL of general organic compounds as well as several simple and reliable SGC models for estimation of the LFL and UFL are discussed.

3.2.1 The predicted LFL as a function of temperature Catoire and Naudet [205] proposed a simple correlation for the accurate estimation of the LFL of CHNO, and monohalogenated organic pure compounds in air at atmospheric pressure in the 25–400 °C temperature range as: × T−0.51536 LFL ðmol% Þ = 519.957 × ð1 + 5nC + 1.25nH − 2.5nO Þ−0.70936 × n−0.197 C

(3:1)

where LFL (mol%) is mole percentage and T is the temperature in K. Example 3.1: Calculate the LFL (mol%) of (a) 2-nitropropane at 25 °C; (b) acetone at 25 °C, and (c) aniline at 140 °C. Answer: (a) The molecular formula of 2-nitropropane is C3H7NO2. Thus, eq. (3.1) gives: × T−0.51536 LFL ðmol% Þ = 519.957 × ð1 + 5nC + 1.25nH − 2.5nO Þ−0.70936 × n−0.197 C = 519.957 × ð1 + 5 × 3 + 1.25 × 7 − 2.5 × 2Þ−0.70936 × ð3Þ−0.197 × ð298.15Þ−0.51536 = 2.70 (b) The molecular formula of acetone is C3H6O. Thus, eq. (3.1) gives:

106

3 Flammability Limit

LFL ðmol% Þ = 519.957 × ð1 + 5nC + 1.25nH − 2.5nO Þ−0.70936 × n−0.197 × T −0.51536 C = 519.957 × ð1 + 5 × 3 + 1.25 × 6 − 2.5 × 1Þ−0.70936 × ð3Þ−0.197 × ð298.15Þ−0.51536 = 2.59 (c) The molecular formula of aniline is C6H7N. Thus, eq. (3.1) gives: LFL ðmol% Þ = 519.957 × ð1 + 5nC + 1.25nH − 2.5nO Þ−0.70936 × n−0.197 × T−0.51536 C = 519.957 × ð1 + 5 × 6 + 1.25 × 7 − 2.5 × 0Þ−0.70936 ×ð6Þ−0.197 ×ð413.15Þ−0.51536 = 1.21 The experimental values of 2-nitropropane, acetone, and aniline at the specified temperatures are 2.5, 2.6, and 1.2, respectively [219].

3.2.2 The use of SGC method for prediction of the LFL and UFL of pure hydrocarbons Albahri [56] presented a suitable SGC method on the basis of 19 structural groups for predicting the LFL and UFL volume %, denoted as the LFL (vol%) and UFL (vol%), respectively, in the air of about 500 pure hydrocarbons. His model has average errors of 0.04 and 1.25 volume % for the LFL (vol%) and UFL (vol%), respectively. He compared the predicted results to that of other methods in the literature and found it to be far more accurate. Table 3.1 shows structural groups for estimating the LFL and UFL volume percent in the air of pure compounds at 298 K. Thus, the LFL (vol%) and UFL (vol%) in the air are calculated by: 2

P




2 3

P

ðLFLÞi + 0.0689 × ðLFLÞi 6 4.174 + 0.8093 × 7 6 7 i LFL ðvol%Þ = 6
3
i 4 7 4 5 P P + 0.00265 × ðLFLÞi + 3.76 × 10−5 × ðLFLÞi i

2 6 6 UFL ðvol%Þ = 6 4

(3:2)

i

P


P

2

3

ðUFLÞi + 0.3587 × ðUFLÞi 18.14 + 3.4135 × 7 7 i i
3
4 7 5 P P −4 + 0.01747 × ðUFLÞi + 3.403 × 10 × ðUFLÞi i

(3:3)

i

Example 3.2: Calculate the LFL (vol%) of (a) 2,2-dimethylbutane (neohexane); (b) trans-1,3-pentadiene; and (c) 1-methyl-2-n-propylbenzene

3.2 Predictive methods of the flammability limits

107

Table 3.1: Group contribution for estimating the LFL (vol%) and UFL (vol%) in air of pure hydrocarbons at 298 K. Serial No.

Group

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

–CH3 (Paraffin) >CH2 (Paraffin) >CH– (Paraffin) >C< (Paraffin) =CH2 (Olefin) =CH– (Olefin) =C< (Olefin) =C= (Olefin) ≡CH (Olefin) ≡C– (Olefin) >CH2 (Cyclic) >CH– (Cyclic) >C< (Cyclic) =CH– (Cyclic) =C< (Cyclic) =CH– (Aromatic) >CH2 (Aromatic) =C< (Aromatic, fused) =C< (Aromatic, non-fused)

(LFL)i

(UFL)i

–1.4407 –0.8736 –0.2925 0.2747 –1.3126 –0.7679 –0.2016 –0.4473 –1.2849 –0.4396 –1.0035 –0.4955 0.1058 –0.8700 –0.5283 –0.8891 –1.0884 –0.3694 –0.2847

–0.8394 –1.1219 –1.2598 –2.1941 0.2479 –0.3016 –0.6524 0.0675 3.8518 1.3924 –0.8386 –0.9648 –2.2754 –0.0821 –0.1252 –1.2966 –1.6166 –1.4722 0.6649

Answer: (a) 2,2-Dimethylbutane has the following molecular structure:

It has four –CH3 (Paraffin), one >CH2 (Paraffin), and one >C< (Paraffin) groups. The use of eq. (3.2) gives: P ðLFLÞi = 4 × ð−CH3 Þ + 1 × ð>CH2 Þ + 1 × ð>CCH2 (Paraffin) groups. The use of eq. (3.2) gives: P i

ðLFLÞi = 2 × ð−CH3 Þ + 3 × ð=CH−Þ + 2 × ð=C< Þ + 2 × ð>CH2 Þ

= 2 × ð−1.4407Þ + 3 × ð−0.8891Þ + 2 × ð−0.2847Þ + 2 × ð−0.8736Þ = −8.7544 2
2 3 P P ðLFLÞi + 0.0689 × ðLFLÞi 6 4.174 + 0.8093 × 7 6 7 i i 6 7 LFL ðvol%Þ = 6
3
4 7 4 5 P P + 0.00265 × ðLFLÞi + 3.76 × 10−5 × ðLFLÞi 2 =4

i

i

4.174 + 0.8093 × ð−8.7544Þ + 0.0689 × ð−8.7544Þ2 3

−5

4

3 5 = 0.81

+ 0.00265 × ð−8.7544Þ + 3.76 × 10 × ð−8.7544Þ

The experimental values of 2,2-dimethylbutane (neohexane), trans-1,3-pentadiene, and 1-methyl-2-n-propylbenzene are 1.2, 1.52, and 0.82, respectively [55]. Example 3.3: Calculate the UFL (vol%) of (a) cis-1,2-dimethylcyclopentane; (b) ethylacetylene (1-butyne); and (c) 1,1-diphenyltetradecane.

3.2 Predictive methods of the flammability limits

109

Answer: (a) cis-1,2-Dimethylcyclopentane has the following molecular structure:

It has two –CH3 (Paraffin), three >CH2 (Cyclic), and two >CH– (Cyclic) groups. The use of eq. (3.3) gives: P i

ðUFLÞi = 2 × ð−CH3 Þ + 3 × ð>CH2 Þ + 2 × ð>CH−Þ

= 2 × ð−0.8394Þ + 3 × ð−0.8386Þ + 2 × ð−0.9648Þ = −6.1242 2
2 3 P P ðUFLÞi + 0.3587 × ðUFLÞi 18.14 + 3.4135 × 6 7 i i 6 7 UFL ðvol%Þ = 6
3
4 7 4 5 P P + 0.01747 × ðUFLÞi + 3.403 × 10−4 × ðUFLÞi i i " # 18.14 + 3.4135 × ð−6.1242Þ + 0.3587 × ð−6.1242Þ2 = = 7.15 + 0.01747 × ð−6.1242Þ3 + 3.403 × 10−4 × ð−6.1242Þ4 (b) Ethylacetylene (1-butyne) has the following molecular structure:

It has one –CH3 (Paraffin), one >CH2 (Paraffin), one ≡CH (Olefin) and one ≡C– (Olefin) groups. The use of eq. (3.3) gives: P i

ðUFLÞi = 1 × ð−CH3 Þ + 1 × ð>CH2 Þ + 1 × ð≡CHÞ + 1 × ð≡C−Þ = 1 × ð−0.8394Þ + 1 × ð−1.1219Þ + 1 × ð3.8518Þ + 1 × ð1.3924Þ = 3.2829 2

6 6 UFL ðvol%Þ = 6 4 " =

18.14 + 3.4135 × + 0.01747 ×


P i

P i

ðUFLÞi + 0.3587 ×

ðUFLÞi


P

3

i

2 ðUFLÞi




+ 3.403 × 10−4 ×

P i

ðUFLÞi #

3 7 7 4 7 5

18.14 + 3.4135 × ð3.2829Þ + 0.3587 × ð3.2829Þ2 = 33.87 + 0.01747 × ð3.2829Þ3 + 3.403 × 10−4 × ð3.2829Þ4

(c) 1,1-Diphenyltetradecane has the following molecular structure:

110

3 Flammability Limit

It has one –CH3 (Paraffin), ten =CH– (Aromatic), two =C< (Aromatic, non-fused), twelve >CH2 (Paraffin), and one >CH– (Paraffin) groups. The use of eq. (3.3) gives: P i

ðUFLÞi = 1 × ð−CH3 Þ + 10 × ð=CH−Þ + 2 × ð=CCH2 −Þ + 1 × ð>CH−Þ = 1 × ð−0.8394Þ + 10 × ð−1.2966Þ + 2 × ð0.6649Þ + 12 × ð−1.1219Þ

+ 1 × ð−1.2598Þ = −27.20 2
2 3 P P ð UFL Þ + 0.3587 × ð UFL Þ 18.14 + 3.4135 × i i 6 7 i i 6 7 UFL ðvol%Þ = 6 7

 3 4 4 5 P P + 0.01747 × ðUFLÞi + 3.403 × 10−4 × ðUFLÞi i i " # 2 18.14 + 3.4135 × ð−27.20Þ + 0.3587 × ð−27.20Þ = 25.37 = + 0.01747 × ð−27.20Þ3 + 3.403 × 10−4 × ð−27.20Þ4 The experimental values of cis-1,2-dimethylcyclopentane, ethylacetylene (1-butyne), and 1,1-diphenyltetradecane are 7.3, 32.93, and 24.13, respectively [56].

3.2.3 Extended method for prediction of the UFL of pure compounds Albahri [220] has extended eq. (3.3) for prediction of the UFL of pure compounds. He introduced 30 atom-type structure groups to represent the UFL for 550 pure substances. His model has a correlation coefficient of 0.9996 and average absolute deviation of 0.17 vol% as: 2 6 6 UFL ðvol%Þ = 6 4


2 3 P P ðUFLÞi + 1.572 × 10−9 × ðUFLÞi 3.563 + 0.5237 × 7 7
i 3
i 4 7 5 P P + 6.375 × 10−8 × ðUFLÞi + 3.266 × 10−5 × ðUFLÞi i

i

(3:4) Table 3.2 shows structural groups for estimating the UFL volume percent in the air of pure compounds at 298 K.

3.2 Predictive methods of the flammability limits

111

Table 3.2: Group contribution for estimating the UFL (vol%) in air of pure compounds at 25°C. Serial No.

Group

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

–CH3 >CH2 >CH– >C< =CH2 =CH– =C< =C= ≡CH ≡C>CH2 (Ring) >CH– (Ring) >C< (Ring) =CH– (Ring) =C< (Ring) –F –Cl –OH (Alcohol) –O– (Non-ring) >C=O (Non-ring) Ketone O=CH– (Aldehyde) –COOH (Acid) =O – NH2 >N– (Non-ring) –O– (Ring) >C=O (Ring) >NH (Ring) –N= (Ring) –H

(UFL)i 1.114692 0.339248 0.138901 –2.50000 2.305876 2.024712 1.346964 6.542315 18.1876 2.243466 0.570317 0.70827 0.403864 0.672134 0.654589 4.231781 2.254908 6.41015 13.04636 5.417596 13.023 0.047503 0.001000 8.603798 1.900000 3.179851 9.414214 6.019325 2.458781 12.51823

Example 3.4: Calculate the UFL (vol%) of (a) cyclopentanone, (b) 1-decanol, and (c) n-decanoic acid. Answer: (a) Cyclopentanone has the following molecular structure: O

It has four >CH2 (Ring) and one >C=O (Ring) groups. The use of eq. (3.4) gives:

112

3 Flammability Limit

P i

ðUFLÞi = 4 × ð>CH2 Þ + 1 × ð>C=OÞ = 4 × ð0.570317Þ + 1 × ð9.414214Þ = 11.69548
2 P P −9 ðUFLÞi + 1.572 × 10 × ðUFLÞi 3.563 + 0.5237 ×

2 6 6 UFL ðvol%Þ = 6 6 4

i

+ 6.375 × 10−8 ×


P i

2

i

3 ðUFLÞi

+ 3.266 × 10−5 ×


P i

ðUFLÞi

−8

+ 6.375 × 10

3

−5

× ð11.69548Þ + 3.266 × 10 × ð11.69548Þ

7 7 7 4 7 5

3

3.563 + 0.5237 × 11.69548 + 1.572 × 10−9 × ð11.69548Þ2

=4

3

4

5 = 10.30

(b) 1-Decanol has the following molecular structure:

HO

It has one –CH3, nine >CH2, and one –OH (Alcohol) groups. The use of eq. (3.4) gives: P i

ðUFLÞi = 1 × ð−CH3 Þ + 9 × ð>CH2 Þ + 1 × ð−OHÞ

= 1 × ð1.114692Þ + 9 × ð0.570317Þ + 1 × ð6.41015Þ = 10.57807
2 3 P P −9 ðUFLÞi + 1.572 × 10 × ðUFLÞi 6 3.563 + 0.5237 × 7 i i 6 7 UFL ðvol%Þ = 6 7

 3 4 4 5 P P −8 −5 + 6.375 × 10 × ðUFLÞi + 3.266 × 10 × ðUFLÞi i i " # −9 3.563 + 0.5237 × 10.57807 + 1.572 × 10 × ð10.57807Þ2 = = 9.51 + 6.375 × 10−8 × ð10.57807Þ3 + 3.266 × 10−5 × ð10.57807Þ4 2

(c) n-Decanoic acid has the following molecular structure: O

OH

It has one –CH3, eight >CH2, and one –COOH (Acid) groups. The use of eq. (3.4) gives: P i

ðUFLÞi = 1 × ð−CH3 Þ + 8 × ð>CH2 Þ + 1 × ð−COOHÞ = 1 × ð1.114692Þ + 8 × ð0.570317Þ + 1 × ð0.047503Þ = 3.876179

3.2 Predictive methods of the flammability limits

2 6 6 UFL ðvol%Þ = 6 4 " =

3.563 + 0.5237 ×

P i

+ 6.375 × 10−8 ×


P


ðUFLÞi + 1.572 × 10−9 × 3 + 3.266 × 10−5 ×

i

2 ðUFLÞi


P

3

7 7 4 7 5

ðUFLÞi # 3.563 + 0.5237 × 3.876179 + 1.572 × 10−9 × ð3.876179Þ2 = 5.60 + 6.375 × 10−8 × ð3.876179Þ3 + 3.266 × 10−5 × ð3.876179Þ4 i

ðUFLÞi

P

113

i

The experimental values of cyclopentanone, 1-decanol, and n-decanoic acid are 10.4, 5.5, and 5.5, respectively [43].

3.2.4 Machine learning-developed models for prediction of LFL and UFL As shown in Chapter 1 (Section 1.2.1.6), Park et al. [49] used an easy-to-apply machine learning-developed models for predicting the FP of pure organic compounds. They developed the same approach for predicting LFL and UFL as follows: LFL ðvol%Þ ¼ 0:1014 þ nCl ð0:3103nO  0:0613nH Þ 1 ð10:3579 þ 1:6783nF þ 2:0032nCl Þ nC  1  þ 2 1:1007nH þ 0:496n2O  4:3578nS  6:896nSi nC þ

ð3:5Þ

UFL ðvol%Þ ¼ 3:636  0:0686nH þ 0:00001MW2 þ 0:2266n2O þ 0:1740n2N þ 0:8261n2S  0:0118ðMW × nO Þ þ 0:1799ðnC × nF Þ þ 0:0746ðnH × nO Þ 1 ð46:62 þ 5:476nO þ 38:47nSi Þ nC  1  þ 2 5:749nH  2:051n2Cl  7:02nBr  0:0009MW2 nC

þ

ð3:6Þ

Example 3.5: Use eqs. (3.5) and (3.6) for the calculation of LFL and UFL of Tetrafluor odimethyl ether.

114

3 Flammability Limit

Answer: Tetrafluorodimethyl ether has the following molecular structure: F F F

O

F

Chemical Formula: C2H2F4O Molecular Weight: 118.03

LFL ðvol%Þ ¼ 0:1014 þ nCl ð0:3103nO  0:0613nH Þ 1 þ ð10:3579 þ 1:6783nF þ 2:0032nCl Þ nC  1  þ 2 1:1007nH þ 0:496n2O  4:3578nS  6:896nSi nC LFL ðvol%Þ ¼ 0:1014 þ 0 × ð0:3103 × 1  0:0613 × 2Þ 1 þ ð10:3579 þ 1:6783 × 4 þ 2:0032 × 0Þ 2 1 þ ð1:1007 × 2 þ 0:496 × 1  4:3578 × 0  6:896 × 0Þ ¼ 8:01 4 The measured LFL (Vol%) is 8.4 [43]. UFL ðvol%Þ ¼ 3:636  0:0686nH þ 0:00001MW2 þ 0:2266n2O þ 0:1740n2N þ 0:8261n2S  0:0118ðMW × nO Þ þ 0:1799ðnC × nF Þ þ 0:0746ðnH × nO Þ 1 þ ð46:62 þ 5:476nO þ 38:47nSi Þ nC  1  þ 2 5:749nH  2:051n2Cl  7:02nBr  0:0009MW2 nC UFL ðvol%Þ ¼ 3:636  0:0686 × 2 þ 0:00001ð118:03Þ2 þ 0:2266 × 1 þ 0:1740 × 0 þ 0:8261 × 0  0:0118ð118:03 × 1Þ þ 0:1799ð2 × 4Þ þ 0:0746ð2 × 1Þ 1 þ ð46:62 þ 5:476 × 1 þ 38:47 × 0Þ 2  1

5:749 × 2  2:051 × 0  7:02 × 0  0:0009ð118:03Þ2 þ 4 ¼ 24:10 The measured LFL (Vol%) is 37.5 [43].

3.3 Flammability limit estimation of the hydrocarbon‐inert gas mixture

115

3.3 Flammability limit estimation of the hydrocarbon‐inert gas mixture There are two approaches to the prediction of the flammability limit of hydrocarbon and inert gas mixture. The first one is empirical methods. For example, Chen et al. [221] used a linear model to explain the carbon dioxide dilution effect on flammability limits for hydrocarbons. The second approach uses thermal theory. For example, Wu et al. [222] proposed a linear model at upper flammability limits between the molar fraction of diluent in an alkane‐CO2 mixture and calculated the adiabatic flame temperature. Hua et al. [223] introduced a simple SGC method for predicting the flammability limit of mixtures containing inert gas. They have proposed two main factors for deriving their method, which includes the effect of inert gas on flammability limits and the proportion of mixtures. It can be assumed that inert gas does not take part in the reaction mechanism of other prediction models [222, 224, 225]. Thus, Hua et al. [223] considered the whole inert gas as a contribution group responsible for flammability limits. The volume concentration ratio of inert gas to the mixture of hydrocarbon and inert gas can be chosen as the parameter to stand for dilute concentration. Since the inert gases in question do not contain nitrogen from the air, the ratio of inert gas to hydrocarbon is given as rin/hy/(1–rin/hy). Hua et al. [223] introduced the following equations for the calculation of LFL and UFL of hydrocarbon‐inert gas mixture: LFL ðvol%Þ ¼ Exp

X

Nj ðLFLÞj þ

X

j

k

  rin=hy Mk ðLFLÞk þ × Cinert;LFL þ ln LFL′const 1  rin=hy

!

(3:7)

UFL ðvol%Þ ¼ Exp

X j

Nj ðUFLÞj þ

X k

  rin=hy Mk ðUFLÞk þ × Cinert;UFL þ ln UFL′const 1  rin=hy

!

(3:8) where (LFL)j and (LFL)k are the first- and the second-order group contribution factors occurring in Nj and Mk, respectively, for prediction of LFL (VOL%); (UFLj) and (UFLk) are the first- and the second-order group contribution factors occurring in Nj and Mk, respectively, for prediction of UFL (Vol%); rin/hy is the ratio of volume concentration of inert gas to the volume concentration of the mixture of hydrocarbon and inert gas; Cinert,LFL is the contribution factor of inert gas in LFL, where the values of Cinert,LFL for CO2 and N2 are 0.0209 and 0.0056, respectively; Cinert,UFL is the contribution factor of inert gas in UFL, where the values of Cinert,UFL for CO2 and N2 are –0.1047 and –0.0865, respectively; and In (LFL′const) and In (UFL′const) are universal constants. The values of In (UFL′const) for CO2 and N2 are 0.2523 and 2.4005, respectively. The values of In (UFL′const) for CO2 and N2 are 0.2700 and 2.4685, respectively.

116

3 Flammability Limit

Table 3.3: Group contribution factors for the estimation of lower and upper flammability limits. Level

Group

First level

–CH3

Second level

LFLi

UFLi

0.2169

–0.1323

>CH2 >CH– >C< CH2=CH– –CH=CH– CH2=C< –CH=C– >C=C< CH2=C=CH– CH2=C=C– –CH=C=CH– CH≡C– –C≡C–

–0.1091 –0.4928 –0.8139 0.3965 0.1678 0.1732 –0.0551 –0.2483 0.3838 0.13 0.13 0.1942 0.02224

–0.0663 0.0324 0 –0.1561 –0.1718 –0.1381 0 0 0.125 0 0.2347 0 0

(CH3)2CH– (CH3)3C< –CH(CH3)–CH(CH3)– –CH(CH3)–C(CH3)2– –C(CH3)2–C(CH3)2– CHn=CHm–CHp=CHk CH3–CHm=CHn –CH2–CHm=CHn CHp–CHm=CHn

0.0363 0.0226 0.0534 0.0561 0.0084 –0.3063 –0.1409 –0.233 –0.2095

–0.0929 –0.0502 –0.0205 0 0 0.14 0.0382 0.0717 0.1949

Example 3.6: Use eq. (3.7) for calculation of LFL of 3-methyl 1-butene and CO2 mixture with r = 0.4. Answer: 3-Methyl 1-butene has the following molecular structure:

X

Nj ðLFLÞj ¼ 2ðCH3 Þ þ ð > CHÞ þ ðCH2 ¼ CHÞ

j

¼ 2 × 0:2169  0:4928 þ 0:3965 ¼ 0:3375 X

Mk ðLFLÞk ¼ ððCH3 Þ2  CHÞ þ ðCHp  CHm ¼ CHn Þ

k

¼ 0:0363  0:2095 ¼ 0:1732

3.4 Summary

X

X

117

  rin=hy LFL ðvol%Þ ¼ Exp Nj ðLFLÞj þ Mk ðLFLÞk þ × Cinert;LFL þ ln LFL′const 1  r in=hy j k
 0:4 ¼ Exp 0:3375  0:1732 þ × 0:0209 þ 0:2523 ¼ 1:54 1  0:4

!

3.4 Summary This chapter introduces different approaches for prediction of the LFL and UFL of pure hydrocarbons as well as some classes of pure compounds. Among three different methods – ANN-SGC, QSPR, and SGC methods – available ANN-SGC and QSPR methods usually need complex molecular descriptors, computer codes, and the expert users. Thus, several simple and reliable SGC methods have been described in this chapter. Equation (3.1) provides a simple path for the accurate estimation of the LFL (mol%) of CHNO, and monohalogenated organic pure compounds in air at atmospheric pressure in the 25–400 °C temperature range. Equations (3.2) and (3.3) give simple correlations for calculating the LFL (vol%) and UFL (vol%) of pure hydrocarbons in the air at 298 K. Eq. (3.4) is an improved correlation, which can be used as pure compounds. Equations (3.5) and (3.6) provide easy-to-apply machine learning-developed models for predicting the LFL and UFL of pure organic compounds. Equations (3.7) and (3.8) give a simple SGC method for calculation of LFL and UFL of a hydrocarbon– inert gas mixture, including CO2 and N2 gases.

4 Heat of Combustion The heat of combustion of a specified substance can be defined as the heat evolved when it is converted to the standard oxidation products by means of molecular oxygen [189]. It can be used for reactive materials to estimate the potential fire hazards of chemicals once they ignite and burn. The fuel can be either liquid or solid, which contains only the elements carbon, hydrogen, nitrogen, and sulfur. For complete combustion of the fuel, the products of combustion, in oxygen, are gaseous carbon dioxide, nitrogen oxides, sulfur dioxide, and liquid or gaseous water. There are two different values of specific heat energy for the same batch of combustible organic compounds, i.e. the gross (or high) heat of combustion and the net (or low) heat of combustion. The net heat of combustion and the gross heat of combustion can also be called the lower heating value (LHV) and the higher heating value (HHV), respectively. The gross heat of combustion is the quantity of energy released when a unit mass of a combustible compound is burned in a constant volume enclosure, with the products being gaseous, other than water that is condensed to the liquid state. The net (or low) heat of combustion is the quantity of energy released when a unit mass of fuel is burned in a constant pressure, with all of the products, including water, being gaseous. The difference between the gross and net heat of combustion is significant, about 8% or 9%. Because engines exhaust water as a gas, the net heat of combustion is the appropriate value to use for comparing fuels. For a diesel fuel, the net heat of combustion corresponds to a heating value in which the water remains a vapor and does not yield its heat of vaporization. Thus, the energy difference between the gross and net heat of combustion is due to the heat of vaporization of water as [226]: ΔHco ðgrossÞ = ΔHco ðnetÞ +

mwater × 40.80 kJ mol−1 mdiesel

(4:1)

where ΔHco ðgrossÞ and ΔHco ðnetÞ are the gross and net heat of combustion, respectively; mwater and mdiesel are the mass of liquid water in the combustion products and the mass diesel fuel, respectively. The net heat of combustion can be defined as the increase in enthalpy when a substance containing carbon, hydrogen, nitrogen, oxygen, fluorine, bromine, iodine, sulfur, phosphorous, and silicon atoms in its standard state (298.15 K and 1 atm) undergoes an oxidization to produce its final combustion products: CO2 (g), F2 (g), Cl2 (g), Br2 (g), I2 (g), SO2 (g), N2 (g), H3PO4 (s), H2O (g), and SiO2 (cristobalite) [43]. It is a measure of the energy available from a fuel, which is used to compare the heating values of fuels and the stability of compounds. It can be used to assess the potential fire hazard of reactive chemicals and predict the performance of explosive and propellant formulations. The knowledge of ΔHco ðnetÞ for chemicals is essential when considering the thermal efficiency of equipment used to produce power or heat. It also provides a good assessment of the environmental impact of any plant at https://doi.org/10.1515/9783110782134-004

120

4 Heat of Combustion

which complete and incomplete combustion are yet to be defined. For pure chemical compounds, the values of ΔHco ðnetÞ are compiled in databases such as AIChE-DIPPR [43] and API-TDB [227]. These databases contain various chemical families of organic compounds including halogenated compounds, acids, ethers, ketones, aldehydes, alcohols, phenols, esters, amines, anhydrides, and sulfur compounds. Experimental determination of ΔHco ðnetÞ is tedious, expensive and sometimes impossible. A fast, easy, and accurate estimation method can be useful if experimental values of ΔHco ðnetÞ are not available and determining them experimentally is inconvenient or not possible. The predicted results from a reliable method of the desired property can be used to predict this property for other compounds that have not been measured or synthesized. The use of appropriate mixing rules can calculate the heat of combustion of transportation fuels, such as naphtha, kerosene, and diesel from the heat of combustion of their constituent compounds when their compositions are known. This type of calculations is also possible for the use of surrogate fuels that simulate the components of the studied fuel [228, 229]. Thus, the knowledge of ΔHco ðnetÞ of the pure chemical compounds can determine the same property for undefined mixtures, such as petroleum fractions.

4.1 Experimental methods for determination of heats of combustion Measurement of ΔHco ðnetÞ is complicated by the existence of several recognized ASTM standard test methods. Available ASTM methods differ on the basis of the characteristics of the studied liquid. For example, ASTM D240-09 [8] can determine the heat of combustion of liquid hydrocarbon fuels and polymers. This test method covers the determination of the heat of combustion of liquid hydrocarbon fuels ranging in volatility from that of light distillates to that of residual fuels. ΔH oc ðnetÞ is determined by burning a weighed sample in an oxygen bomb calorimeter under controlled conditions. The heat of combustion is computed from the temperatures observed before, during, and after combustion while allowing for the appropriate thermochemical and heat transfer corrections. ASTM D4809-13 [230] is used as a more precise method. For aviation fuels, ASTM D3338/D3338M-09 [231], ASTM D1405/D1405M-08 [232], and ASTM D4529-01 [233] are used. ASTM D4868-00 [234] is used for diesel and burner fuels. The gross heat of combustion is normally determined by the oxygen bomb calorimeter method (Figure 4.1), which includes the heat of vaporization given up when the newly formed water vapor produced by oxidation of hydrogen is condensed and cooled to the temperature of the bomb. For nearly all industrial applications, this water vapor escapes as steam in the flue gases. Thus, water vapor is not available for useful work. To compensate for this loss, the net heat of combustion can be calculated by subtracting the latent heat of vaporization from the gross value obtained

4.2 Different approaches for prediction of the heats of combustion

Thermometer

121

High pressure oxygen combustion bomb

Electrical connections to Fe ignition wire Waterbatch stirrer High pressure oxygen combustion bomb

Figure 4.1: Schematic diagram of bomb calorimeter.

from the calorimeter, but this requires knowledge of the hydrogen content of the sample. Therefore, the net heat of combustion is calculated from the gross heat of combustion by assuming the formation of water is in a gaseous state. If the hydrogen content of a sample is known, the net heat of combustion at 298.15 K can be calculated as follows [235]: ΔHc′ ðnetÞ = ΔHc′ ðgrossÞ − 212.2 × H

(4:2)

where ΔH ′c ðnetÞ is the net heat of combustion at constant pressure in kJ kg−1 ; ΔHc′ ðgrossÞ is the gross heat of combustion at constant volume in kJ kg−1 ; H is the weight percent hydrogen in the sample. Due to the difficulty in accurate determination of the hydrogen content of the sample, and the fact that the hydrogen content of most fuels is fairly low, the gross heat of combustion is usually reported in preference to the net value for most applications.

4.2 Different approaches for prediction of the heats of combustion Experimental determination of the heat of combustion of a new compound is too time-consuming because a complete reproducible combustion cannot be easily obtained. Predicting heat of combustion is valuable before expending resources. Various methods have been developed in the literature to estimate the heat of

122

4 Heat of Combustion

combustion [189, 226, 236–250]. Cardozo [236] proposed a group contribution method to estimate the heat of combustion for organic compounds using three simple correlations that depend on the state of the compound: solid, liquid, or gas. The necessary data are the total number of carbon atoms in the compound and the corrections for various structures and phases. The reported errors exceeded 12.5% for some compounds. This method can calculate the heat of combustion of complex organic compounds. Cardozo’s model [236] relates the length of chain and heat of combustion based on 1,168 experimental data. It cannot be applied for polynitroheteroarenes, acyclic and cyclic nitramines as well as energetic compounds with nitrate functional groups because group correction factors have not been specified for the compounds containing N–NO2 and O–NO2 functional groups and molecular fragments in polynitroheteroarenes. Seaton and Harrison [237] assumed some combustion products for pure or mixtures of compounds including 71 elements to estimate the heat of combustion. This approach is based on the Benson method [50] for predicting the heat of formation, which is considered to be too complex for manual calculations. Suitable group contribution and quantum mechanical methods can also be used for calculating the heat of formation or heat of sublimation, which can predict indirectly the heat of combustion [251–254]. Van Krevelen [238] used the heats of formation for the combustion products and reactants to calculate the heat of combustion. For polymeric reactants, the heat of combustion was estimated from the molar contributions of the chemical groups that constitute the monomer or repeat units. Kondo et al. [255] calculated the heats of formation for several flammable gases by the Gaussian-2 (G2) and/or G2MP2 method to obtain their heats of combustion and related constants for evaluating the combustion hazards. Quantum mechanical methods usually require special complex software, high-speed computers, and expert users. Hshieh [189] developed two empirical equations to predict the gross and the net heats of combustion of organosilicon compounds on the basis of the atomic contribution method. Hshieh and coworkers [240] have also introduced two other empirical equations for predicting the gross and net heats of combustion of organic polymers. Hshieh’s methods [189, 240] have the advantage that only elemental compositions of organosilicon compounds and polymers are input parameters but they are only applicable to the organosilicon compounds and polymers. Diallo et al. [246] developed a model for calculating the heat of combustion of 53 ionic liquids using a multivariable linear regression technique. QSPR models have also been used to predict heat of combustion of organic compounds. Gharagheizi [241] used genetic algorithm based GA-MLR to obtain four parameters multi-linear equation. This model is based on four complex molecular descriptors. Cao et al. [242] introduced another suitable QSPR for prediction of the standard net heat of combustion of organic compounds. This model is based on the atom-type electrotopological state (E-state) indices and ANN technique. Cao et al. [242] suggested a QSPR-ANN model to estimate the heat

4.2 Different approaches for prediction of the heats of combustion

123

of combustion for pure organic compounds using atom-type electrotopological state indices with 49 structural groups as inputs. The model of Cao et al. [242] gives large AAEs for compounds containing fluorine and chlorine atoms. For compounds containing fluorine and chlorine, the model of Cao et al. [196] predicts the heat of combustion with an AAE of 615.98 and 299.65 kJ mol–1, respectively [242]. Gharagheizi et al. [245] developed a complex method for calculating the heat of combustion of pure compounds from group contributions using a three-layered Feed-Forward Artificial Neural Network (FFANN) model with 142 intricate structural groups as inputs. Saldana et al. [78] created a consensus QSPR model for predicting the heat of combustion of 1,624 hydrocarbon-based compounds and 1,143 alcohols and esters. Saldana et al. [78] used various QSPR approaches to build models for predicting the heat of combustion ranging from methods leading to multi-linear models such as Genetic Function Approximation (GFA) and PLS to non-linear models, such as FFANN, General Regression Neural Networks (GRNN), SVM, and Graph Machines (GM). These models except the GM model use molecular descriptors and functional group count descriptors as inputs. Since all of the individual models have AAE of less than 2% except for the GRNN based model, which has an AAE of 4%. Saldana et al. [78] developed a consensus model by averaging the values computed with selected individual models to improve the generality and predictive power compared to individual predictive models. The robust consensus model can predict the heat of combustion with AAE = 0.7%. These QSPR models are based on complex descriptors, which require specific computer codes and expert users. Moreover, QSPR models may have some uncertainty and difficulty for complex molecular structures. A simple and reliable QSPR model has also been used for prediction of the heat of combustion of various energetic compounds including polynitro arene, polynitro heteroarene, acyclic and cyclic nitramine, nitrate ester, and nitroaliphatic compounds [243]. The methods of SGC are suitable methods for estimation of the heat of combustion of organic compounds. Sagadeev and coworkers [256, 257] have introduced some group contributions to calculate the heats of combustion For different classes of organic compounds. The method of Sagadeev and coauthors [256, 257] can be used only for certain types of nitroaromatic and nitroaliphatic compounds. Among different available methods for predicting the heat of combustion, several simple methods that provide reliable estimates and are easier to use are discussed in this chapter.

4.2.1 Predicting the standard net heat of combustion for pure hydrocarbons from their molecular structure Albahri [248] introduced a group contribution method to predict the standard net heat of combustion of pure hydrocarbons from their molecular structures. He used a

124

4 Heat of Combustion

multivariable non-linear regression based on the least square method to arrive at a set of 32 atom-type structural groups. He represented the standard net heat of combustion for about 452 pure hydrocarbon substances. His method is very simple and requires no experimental data. This SGC can predict the standard net heat of combustion from the knowledge of the molecular structure alone with an AAE of 0.71%. It can also predict the standard net heat of combustion of hydrocarbon isomers as well. !3 X X   o o o ΔHc i − 0.00719 × ΔHc i ΔHc ðnetÞ = 154.608 + 1,177.8 × i

X

+ 0.0053 ×

ΔHco

i

!4



(4:3)

i

i

P

ΔHco



is the sum of the atom-type group contribution values for   calculation of the standard net heat of combustion. The values of ΔHco i are given in Table 4.1. where

i

i

Example 4.1: Calculate the standard net heat of combustion for p-diethyl benzene. Answer: The molecular structure of p-diethyl benzene is given as:

It consists of two –CH3 (Paraffin), two >CH2 (Paraffin), four =CH– (Aromatic), one =C< P  o ΔHc i is calculated as: (Aromatic), and one =C< (Aromatic, –p). Thus, i

P i

 o

ΔHc

i

= 2 × 0.5389 + 2 × 0.5158 + 4 × 0.4218 + 2 × 0.3720 + 1 × 0.3974 + 1 × 0.3875 = 4.5815

The use of eq. (4.2) gives: ΔHco ðnetÞ = 154.608 + 1,177.8 × + 0.0053 ×


P i

P i

4

 o

ΔHc


  P  o 3 ΔHco i − 0.00719 × ΔHc i i

i

= 154.608 + 1,177.8 × 4.5815 − 0.00719 × ð4.5815Þ3 + 0.0053 × ð4.5815Þ4 = 5,552.34 kJ mol−1 The measured ΔHco ðnetÞ is 5,555.21 kJ mol–1 [43]. Example 4.2: Calculate the standard net heat of combustion for 6-n-propyl-[1,2,3,4tetrahydronaphthalene].

4.2 Different approaches for prediction of the heats of combustion

Table 4.1: Atom-type structural groups and their contribution values for estimation of the standard net heat of combustion of pure hydrocarbons (kJ mol–1) using eq. (4.3). Serial No.

Group

 o ΔHc i

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

–CH3 (Paraffin) >CH2 (Paraffin) >CH– (2–) (Paraffin) >CH– (3–) (Paraffin) >CH– (4–) (Paraffin) >CH– (5–) (Paraffin) >C< (2–) (Paraffin) >C< (3–) (Paraffin) =CH2 (Olefin) =CH– (Olefin) =C< (2–) (Olefin) =C< (3–) (Olefin) =C= (Olefin) ≡CH (Olefin) ≡C– (Olefin) >CH2 (Cyclic) >CH– (Cyclic) >CH– (–2)cis (Cyclic) >CH– (–2)trans (Cyclic) >CH– (–3)cis (Cyclic) >CH– (–3)trans (Cyclic) >CH– (–4)cis (Cyclic) >CH– (–4)trans (Cyclic) >C< (Cyclic) =CH– (Cyclic) =CH– (Aromatic) >CH2 (Aromatic) =C< (Aromatic, fused) =C< (Aromatic) =C< (Aromatic, –o) =C< (Aromatic, –m) =C< (Aromatic, –p)

0.5389 0.5158 0.4905 0.4963 0.4946 0.4963 0.4700 0.4748 0.4923 0.4713 0.4323 0.4566 0.4494 0.4683 0.4238 0.4948 0.4775 0.4831 0.4770 0.4753 0.4758 0.4657 0.4599 0.4165 0.4371 0.4218 0.3651 0.3720 0.3974 0.3898 0.3882 0.3875

Groups 29 to 32 are applied for non-fused. For non-cyclic compounds, the numbers 2–, 3–, 4–, and 5– show the carbon atom position along the hydrocarbon chain in the second, third, fourth, and fifth position, respectively, where they are calculated from the shortest distance from either end of the hydrocarbon chain. For cyclic compounds, the numbers 2–, 3–, and 4– refer to the second, third, and fourth position along the cyclic ring with respect to the reference group 17, respectively, where they are calculated from the shortest distance along the cyclic ring from either direction. For aromatics, the symbols o–, m–, and p– refer to the ortho, meta, and para positions on the aromatic ring.

125

126

4 Heat of Combustion

Answer: The molecular structure of 6-n-propyl-[1,2,3,4-tetrahydronaphthalene] is given as:

It consists of one –CH3 (Paraffin), two >CH2 (Paraffin), three =CH– (Aromatic), four >CH2 (Cyclic), one =C< (Aromatic), one =C< (Aromatic, –o), and one =C< (Aromatic, –p). Thus, P  o ΔHc i is calculated as: i

P i

 ΔHco i = 1 × 0.5389 + 2 × 0.5158 + 3 × 0.4218 + 4 × 0.4948 + 1 × 0.3974 + 1 × 0.3898 + 1 × 0.3898 + 1 × 0.3882 = 5.99

The use of eq. (4.2) gives:


 P  o P  o 3 ΔHco ðnetÞ = 154.608 + 1,177.8 × ΔHc i − 0.00719 × ΔHc i i i
 P  o 4 + 0.0053 × ΔHc i i

= 154.608+1,177.8 × 5.99−0.00719×ð5.99Þ3+0.0053 ×ð5.99Þ4 =7,216 kJ mol−1

4.2.2 Prediction of the standard net heat of combustion from molecular structure Albahri [250] developed a suitable QSPR method to predict ΔHco ðnetÞ of chemical compounds based only in their molecular structures. He used SGC method to estimate ΔHco ðnetÞ through two models: a Multi-Variable Regression (MVR) based on least squares and an ANN. He applied the SGC method to probe the structural groups that have a significant contribution to the overall ΔHco ðnetÞ. He introduced 47 atomtype structural groups can represent the ΔHco ðnetÞ for 586 pure substances. Among two approaches ANN and MVR, ANN was the more accurate. The method ANN can predict ΔHco ðnetÞ with an average relative error of 0.89%. The results of the MVR model is less accurate, but it is also simple and practical and provides reliable estimates. The SGC method of Albahri [250] is very useful and convenient to assess the hazardous risks of chemicals. The MVR model uses a simple linear addition of the structural group contributions as: ΔHco ðnetÞ =

X j

ΔHco

 j

(4:4)

4.2 Different approaches for prediction of the heats of combustion

127

  The values of ΔHco j are given in Table 4.2. Example 4.3: Calculate the standard net heat of combustion for o-cresol by using eq. (4.3). Answer: The molecular structure of p-diethyl benzene is given as follows: OH

This compound consists of one –CH3, four =CH– (Ring), two =C< (Ring), one –OH (Phenol). Thus, the use of eq. (4.4) gives: Table 4.2: The structural groups and their contribution values for estimation of the standard net heat of combustion of pure compounds (kJ mol–1 ) using eq. (4.4). Serial No.

Group

 o ΔHc j

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

–CH3 >CH2 >CH– >C< =CH2 =CH– =C< =C= ≡CH ≡C– –CH2 (Ring) >CH– (Ring) >C< (Ring) =CH– (Ring) =C< (Ring) –F –Cl –Br –OH (Alcohol) –O– (Non-ring) >C=O (Ketone) –CH=O (Aldehyde) –COOH (Acid) –COO– (Ester) =O (Except as above) –NH2

711.40 609.28 503.00 398.87 667.53 544.47 431.42 503.17 621.61 502.60 612.64 519.26 425.69 522.59 416.78 –101.00 –49.50 –51.00 –88.00 –106.00 230.28 371.63 19.76 33.50 0.1 257.91

128

4 Heat of Combustion

Table 4.2 (continued)  o ΔHc j

Serial No.

Group

27 28 29 30 31 32 33 34 35 36 37 38 39

>NH (Non-ring) >N– (Non-ring) –OH (Phenol) –O– (Ring) >NH (Ring) –N= (Ring) >N– (Ring) –H ≡C ≡O =S >S –SH

126.05 71.89 –110.00 –115.00 60.81 2.31 244.67 120.17 228.36 54.50 278.58 326.55 417.24

40

S

219.52

41 42

≡N =S= (Non-ring)

153.34 90.31

N

43

44

–O–

0.02

4.40

45

N H

46

N

47

NH2

220.57

–76.00

187.93

4.2 Different approaches for prediction of the heats of combustion

ΔHco ðnetÞ =

P j

129

 ΔHco j = 1 × 711.40 + 4 × 522.59 + 2 × 416.78 + 1 × ð−110.00Þ

= 3,525.3 kJ mol−1 The measured ΔHoc ðnetÞ is 3,517.45 kJ mol–1 [43]. Example 4.4: Calculate the standard net heat of combustion for n-eicocylcyclohexane by using eq. (4.4). Answer: The molecular structure of n-eicocylcyclohexane is given as follows:

This compound consists of one –CH3, 19 >CH2, one >CH– (Ring), and five –CH2 (Ring). Thus, the use of eq. (4.4) gives: P  o ΔHc j = 1 × 711.40 + 19 × 609.28 + 1 × 519.26 + 5 × 612.64 ΔHco ðnetÞ = j

= 15,870 kJ mol−1 The measured ΔHoc ðnetÞ is 15,970.80 kJ mol–1 [43].

4.2.3 A comprehensive methodology for prediction of the net heat of combustion from group contribution-based property models Frutiger et al. [258] used the Marrero/Gani (MG) method [259] for the prediction of the net heat of combustion of organic compounds. The method of MG combines the contributions of the first-, second-, and third-order parameters, which correspond to specific functional, polyfunctional, and structural groups, respectively. If the description given by the first-order groups is insufficient, the use of the secondand third-order parameters can provide additional structural information about molecular fragments. The model of Frutiger et al. [258] can be given as follows: ΔHc ðnetÞ ¼ 99:67 þ

X i

Ni Ci þ

X j

Mj Dj þ

X

E k Ok

(4:5)

k

where Ci is the contribution of the first-order group of type j that occurs Ni times, Dj is the contribution of the second-order group of type j that occurs Mj times, and Ek is the contribution of the third-order group of type k that has Ok occurrences in an organic compound. Tables 4.3–4.5 give the list of Ci , Dj , and Ek along with the sample assignments and group occurrences.

130

4 Heat of Combustion

Table 4.3: The values of Ci along with the sample assignments and group occurrences (Ni ). Group

Ci

Example (Ni )

CH3 CH2 CH C CH2=CH CH=CH CH2=C CH=C C=C CHC CC aCH aC fused with aromatic ring aC fused with nonaromatic subring aC except as above aN in aromatic ring aC-CH3 aC-CH2 aC-CH aC-C aC-CH=CH2 aC-CH=CH aC-C=CH2 aC-CCH OH aC-OH COOH aC-COOH CH3CO CH2CO CHCO aC-CO CHO aC-CHO CH3COO CH2COO CHCOO CCOO HCOO aC-COO COO except as above CH3O CH2O CH-O

–663.99 –607.99 –549.79 –488.61 –1,147.93 –1,071.84 –1,067.17 –990.95 –926.95 –1,081.60 –992.35 –506.95 –520.99 –451.26 –441.51 –36.38 –1,097.63 –1,048.92 –993.98 –966.09 –1,592.35 –1,492.82 –1,513.65 –1,520.19 123.75 –293.98 –23.19 –480.27 –913.52 –851.15 –791.53 –662.91 –313.74 –761.17 –695.51 –643.99 –593.15 –518.03 –120.21 –490.03 –62.74 –545.11 –465.57 –402.84

n-Tetracontane (2) n-Tetracontane (38) 2-Methylpentane (1) 2,2-Dimethylbutane (1) 1-Hexene (1) 2-Hexene (1) 2-Methyl-1-butene (1) 2-Methyl-2-butene (1) 2,3-Dimethyl-2-butene 1-Pentyne (1) 3-Decyne (1) Benzene (6) Naphthalene (2) Indane (2) Benzophenone (1) Pyridine (1) Toluene (1) Ethylbenzene (1) Cumene (1) tert-Butylbenzene (1) Styrene (1) 1-Prophenylbenzene (1) α-Methylstyrene (1) Phenylacetylene (1) 1,4-Butanediol (2) Phenol (1) 1,2-Pentanedioic acid (2) Benzoic acid (1) 2-Butanone (1) 3-Pentanone (1) 2,4-Dimethyl-3-pentanone (1) Acetophenone (1) 1-Hexanal (1) Benzaldehyde (1) Butyl acetate (1) Methyl butyrate (1) Ethyl isobutyrate (1) Ethyl 2,2-dimethylpropionate (1) Propyl formate (1) Methyl benzoate (1) Ethyl acrylate (1) Methyl butyl ether (1) Di-n-butyl ether (1) sec-Butyl ether (1)

4.2 Different approaches for prediction of the heats of combustion

131

Table 4.3 (continued) Group

Ci

Example (Ni )

C-O aC-O CH2NH2 CHNH2 CNH2 CH3NH CH2NH CH3N CH2N aC-NH2 aC-NH aC-N NH2 except as above C=N CH2CN CHCN aC-CN CN except as above CH2NCO aC-NCO CH2NO2 CHNO2 CNO2 aC-NO2 NO2 except as above CONH2 CON(CH3)2 aC-CONH2 aC-NH(CO)H aC-NHCO NHCONH NH2CONH CH2Cl CHCl CHCl2 CCl3 CHF2 CF3 aC-Cl aC-F aC-Br -I -Br except as above -Cl except as above

–423.92 –305.66 –797.66 –722.22 –663.64 –850.94 –771.79 –797.82 –775.99 –587.13 –607.23 –579.48 –208.75 –292.55 –1,035.16 –984.06 –977.99 –485.98 –889.27 –682.25 –487.89 –421.66 –605.45 –349.68 91.11 –308.44 –1,616.34 –764.37 –815.60 –727.61 –402.35 –372.02 –503.92 –454.53 –384.01 –240.32 18.66 319.30 –342.17 –213.47 –403.87 33.96 70.60 103.68

tert-Butyl ether (1) Methyl phenyl ether (1) Ethylamine (1) sec-Butylamine (1) tert-Butylamine (1) Dimethylamine (1) Dipropylamine (1) Methyldiethylamine (1) Triethylamine (1) Aniline (1) N-Methyl aniline (1) N,N-Dimethylaniline (1) Cyclobutylamine (1) Ketazine (2) Propionitrile (1) Isobutylnitrile (1) Benzonitrile (1) Acrylonitrile (1) Ethyl isocyanate (1) Phenyl isocyanate (1) 1-Nitropropane (1) 2-Nitropropane (1) 2-Methyl-2-nitropropane Nitrobenzene (1) Nitrocyclohexane (1) Butyramide (1) Dimethylacetamide (1) Benzamide (1) N-Phenylformamide (1) N-(2-Methylphenyl)acetamide (1) N,N′-Dimethylurea (1) Methylurea (1) 1-Chlorobutane (1) 2-Chloropropane (1) 1,1-Dichloroethane (1) 1,1,1-Trichloroethane (1) 1,1-Difluoroethane (1) Hexafluoroethane (1) Chlorobenzene (1) Hexafluorobenzene (1) Bromobenzene (1) Iodoethane (1) Bromoethane (1) Ethyl chloroacetate (1)

132

4 Heat of Combustion

Table 4.3 (continued) Group

Ci

Example (Ni )

CHNOH CNOH OCH2CH2OH -O-OH CH2SH CHSH CSH aC-SH -SH except as above CH3S CH2S CHS aC-SPO3 aC-PO4 CO3 (carbonate) C2H3O CH2 (cyclic) CH (cyclic) C (cyclic) CH=CH (cyclic) CH=C (cyclic) CH2=C (cyclic) NH (cyclic) N (cyclic) CH=N (cyclic) O (cyclic) CO (cyclic) S (cyclic) SO2 (cyclic) >NH -OSiO Si Ccyc=CHP=O N=N >C=S CH=C=CH

–556.44 –623.35 –966.58 95.12 –981.49 –914.40 –849.62 –813.00 –388.04 –977.67 –918.86 –871.19 –763.61 –133.37 –233.14 140.45 –1,021.78 –597.82 –519.59 –184.48 –1.061.20 –771.21 –1,122.13 –116.79 –124.84 –517.74 173.26 –181.22 –286.54 –269.97 –192.02 –66.61 –545.70 –836.63 –1,029.04 –504.63 –166.14 –568.83 –1,610.35

Propionaldehyde oxime (1) Diethyl ketoxime (1) 2-Ethoxyethanol (1) tert-Butylhydroperoxide (1) Ethanethiol (1) 2-Propanethiol (1) 2-Methyl-2-propanethiol (1) Benzenethiol (1) Cyclohexanethiol (1) Dimethylsulfide (1) Diethylsulfide (1) Diisopropylsulfide (1) Phenyl methyl sulfide (1) Triethylphosphite (1) Triphenylphosphate (1) Diethylcarbonate (1) Ethyl oxirane (1) Cyclopentane (5) Methylcyclopentane (1) 1,1-Dimethylcyclohexane (1) Cyclobutene (1) 1-Methylcyclopentene (1) Methylene cyclohexane (1) Cyclopentimine (1) N-Methylpyrrolidine (1) Imidazole (1) Tetrahydropyran (1) Cyclobutanone (1) 2-Methyl-thiophene (1) Cyclobutadiene sulfone (1) Dimethylamine (1) Methoxymethane (1) Methoxydimethylsilane Dimethylsilane (1) Ethylidenecyclopentane Trimethylphosphine oxide (1) (Z)-1,2-Dimethyldiazene (1) Propane-2-thione (1) 2,3-Pentadiene (1)

4.2 Different approaches for prediction of the heats of combustion

133

Table 4.4: The values of Dj along with the sample assignments and group occurrences (Mj ). Group

Dj

Example (Mj )

(CH3)2CH (CH3)3C CH(CH3)CH(CH3) CH(CH3)C(CH3)2 C(CH3)2C(CH3)2 CHn=CHm-CHp=CHk (k, m, n, p in 0,…, 2) CH3-CHm=CHn (m, n in 0,…, 2) CH2-CHm=CHn (m, n in 0,…, 2) CHp-CHm=CHn (m, n in 0,…, 2; p in 0,…, 1) CHCHO or CCHO CH3COCH2 CH3COCH or CH3COC CHCOOH or CCOOH CO-O-CO CHOH COH OH-CHn-COO (n in 0,…, 2) CHm(OH)CHn(OH) (m, n in 0, …, 2) CHm(OH)CHn(NHp) (m, n, p in 0,…, 2) CHm(NH2)CHn(NH2) (m, n in 0,…, 2) CHm(NHn)-COOH (m, n in 0,…, 2) HOOC-CHn-CHm-COOH (n, m in 1,…, 2) HO-CHn-COOH (n in 1,…, 2) HS-CHn-CHm-COOH (n, m in 1,…, 2) NC-CHn-CHm-CN (n, m in 1,…, 2) COO-CHn-CHm-OOC (n, m in 1,…, 2) OOC-CHm-CHm-COO (n, m in 1,…, 2) NC-CHn-COO (n in 1,…, 2) COCHnCOO (n in 1,…, 2) CHm=CHn-Cl (m, n in 0,…, 2) CHm=CHn-CN (m, n in 0,…, 2) CHn=CHm-COO-CHp (m, n, p in 0,…, 3) CHm=CHn-CHO (m, n in 0,…, 2) CHm=CHn-COOH (m, n in 0,…, 2) aC-CHn-NHm (n in 1,…, 2; m in 0,…, 2)) aC-CHn-OH (n in 1,…, 2) aC-CHn-CN (n in 1,…, 2) aC-CHn-COO (n in 1,…, 2) aC-CH(CH3)2 aC-C(CH3)3 (CHn=C)(cyclic)-CHO (n in 0,…, 2) (CHn=C)cyc-CH3 (n in 0,…, 2) (CHn=C)cyc-CH2 (n in 0,…, 2) CHcyc-CH3

3.31 1.67 –2.40 –7.19 –10.35 1.99 –9.36 –9.68 –5.44 –4.65 13.68 9.54 –20.83 28.35 5.98 27.78 –9.09 15.95 18.68 3.98 20.14 –13.32 –4.88 –22.15 –27.42 –23.48 –2.88 –16.44 –54.34 –22.26 23.55 –5.81 6.72 –26.46 29.07 –0.44 16.20 29.94 1.89 36.02 –183.83 –202.78 –173.73 –2.11

2-Methylpentane (1) 2,2,4,4-Tetramethylpentane (2) 2,3,4-Trimethylpentane (2) 2,2,3,4,4-Pentalmethylpentane (2) 2,2,3,3,4,4-Hexamethylpentane (2) 1,3-Butadiene (1) 2-Methyl-2-butene (3) 1,4-Pentadiene (2) 2-Methyl-1-butene (1) 2-Methylbutyraldehyde (1) 2-Pentanone (1) 3-Methyl-2-pentanone (1) 2-Methybutanoic acid (1) Propionic anhydride (1) 2-Butanol (1) 2-Methyl-2-butanol (1) Ethyl lactate (1) Ethylene glycol (1) 2-Amino-1-butanol (1) Ethylenediamine (1) L-Alanine (1) Succinic acid (1) 2-Hydroxyisobutyric acid (1) β-Thiolactic acid (1) 1,2-Dicyanoethane (1) Ethylene glycol diacetate (1) Dimethylsuccinate (1) Methylcyanoacetate (1) Methylacetoacetate (1) 1-Chloro-2-methylpropene (1) Acrylonitrile (1) Ethyl acrylate (1) Propenaldehyde (1) Acrylic acid (1) Benzyl amine (1) Benzyl alcohol (1) Benzyl cyanide (1) Methyl phenyl acetate (1) Cumene (1) tert-Butylbenzene (1) Furfural (1) 1,2-Dimethylcyclopentene (1) 2-Ethylfuran (1) Methylcyclopentene (1)

134

4 Heat of Combustion

Table 4.4 (continued) Group

Dj

Example (Mj )

CHcyc-CH2 CHcyc-C CHcyc-C=CHn (n in 1,…, 2) CHcyc-Cl CHcyc-OH CHcyc-NH2 CHcyc-CN CHcyc-COOH CHcyc-OCcyc-CH3 Ccyc-CH2 Ccyc-OH >Ncyc-CH2 AROMRINGs1s2 AROMRINGs1s3 AROMRINGs1s4 AROMRINGs1s2s3 AROMRINGs1s2s4 AROMRINGs1s3s5 AROMRINGs1s2s3s4 AROMRINGs1s2s3s5 AROMRINGs1s2s4s5 PYRIDINEs2 PYRIDINEs3 PYRIDINEs4 PYRIDINEs2s6

–4.12 –73.25 –24.19 20.66 1.76 10.76 –22.13 –15.16 –34.90 –142.42 –131.31 –30.33 7.20 –11.80 –1.90 2.26 –16.75 –17.76 –8.25 –33.54 –41.18 –10.73 –2.64 3.32 –5.24 –3.47

Ethylcyclohexane (1) tert-Butylcyclohexane (1) Limonene (1) Chlorocyclopentane (1) Cyclohexanol (1) Cyclohexylamine (1) Cyanocyclopentane (1) Cyclopropanecarboxylic acid (1) Methoxycyclohexane (1) 1,1-Dimethyl-cyclohexane (2) 1-Ethyl-1-methyl-cyclopentane (1) 1-Methylcyclopentanol (1) N-Methyl-2-pyrrolidone (1) 2-Methyl-phenol (1), 2-Et-toluene (1) 3-Methyl-phenol (1), 3-Et-toluene (1) 2-Methyl-phenol (1), 4-Et-toluene (1)4 1,2,3-Trimethylbenzene 1,2,4-Trihydroxybenzene (1) 3,5-Diethyltoluene (1) 3-Ethyl-1,2,4-trimethylbenzene (1) 1,2,3,5-Tetramethylbenzene (1) 1,2,4,5-Tetramethylbenzene (1) 2-Methylpyridine (1) 3-Methylpyridine (1) 4-Methylpyridine (1) 2,6-Dimethylpyridine (1)

Table 4.5: The values of Ek along with the sample assignments and group occurrences (Ok ). Group

Ek

Example (Ok )

HOOC-(CHn)m-COOH (m > 2, n in 0,…, 2) OH-(CHn)m-OH (m > 2, n in 0,…, 2) NC-(CHn)m-CN (m > 2) aC-(CHn=CHm)cyc (fused rings) (n, m in 0,…, 1) aC-aC (different rings) aC-CHn,cyc (different rings) (n in 0,…, 1) aC-CHn,cyc (fused rings) (n in 0,…, 1) aC-(CHn)m-aC (different rings) (m > 1; n in 0,…, 2) CHcyc-CHcyc (different rings) CH multiring C multiring aC-CHm-aC (different rings) (m in 0,…, 2)

–9.13 16.35 –27.63 –7.78 –40.49 6.77 –22.20 –13.20 7.70 18.85 –314.73 2.55

1,5-Pentanedioic acid (1) 4-Aminobutanol (1) Glutaronitrile (1) Indene (1), acenaphthylene (2) Biphenylene (2), biphenyl (1) Cyclohexylbenzene (1) Tetralin (2), indane (2) Bibenzyl (1) Cyclohexylcyclohexane (1) Hexahydroindan (2), decalin (2) Spiropentane (1) Diphenylmethane (1)

4.2 Different approaches for prediction of the heats of combustion

135

Table 4.5 (continued) Group

Ek

Example (Ok )

aC-(CHm=CHn)-aC (different rings) (m, n in 0,…, 2) aC-CO-aC (different rings) aC-CHm-CO-aC (different rings) (m in 0,…, 2) aC-COcyc (fused rings) aC-Scyc (fused rings) aC-NHn,cyc (fused rings) (n in 0,…, 1) aC-NH-aC (different rings) aC-O-aC (different rings) aC-Ocyc (fused rings) AROM.FUSED[2] AROM.FUSED[2]s1 AROM.FUSED[2]s2 AROM.FUSED[3] AROM.FUSED[4a] AROM.FUSED[4p] PYRIDINE.FUSED[2] PYRIDINE.FUSED[2-iso] PYRIDINE.FUSED[4]

–14.55 –207.33 193.26 –42.38 6.17 –11.35 35.46 22.39 –23.77 107.32 88.13 98.48 146.01 197.71 197.05 101.92 86.03 184.48

1,2-Diphenylethylene (1) Benzophenone (1) Benzyl phenome (1) Phenolphthalein (1) Dibenzothiophene (1) Carbazole (2) Diphenylamine (1) Diphenyl ether (1) Benzoxazole (1) Naphthalene (1) 1-Methylnaphthalene (1) 2,7-Dimethylnaphthalene (1) Phenalene (3), pyrene (2) Anthracene (1) Phenanthrene (1), pyrene (2) Quinoline (1) Isoquinoline (1) Acridine (1)

Example 4.5: Use eq. (4.5) for calculation of the standard net heat of combustion of pyrene. Answer: The molecular structure of pyrene is

First-order groups: 6 × (aC fused with aromatic ring) + 10 × (aCH) Second-order groups: No second-order groups are involved. Third-order groups: 2 × (AROM.FUSED[3]) + 2 × ( AROM.FUSED[4p]) Thus, the use of eq. (4.5) gives ΔHc ðnetÞ ¼ 99:67 þ

X

Ni Ci þ

i

X j

Mj Dj þ

X

Ek Ok

k

¼ 99:67 þ ½6 × ð520:99Þ þ 10 × ð506:95Þ þ 0 þ ½2 × ð146:01Þ þ 2 × ð197:05Þ ¼ 7; 608:99 kJ mol − 1 The measured ΔHc ðnetÞ is 7,630.94 kJ mol–1 [43].

136

4 Heat of Combustion

4.2.4 A generally applicable group additivity method for the calculation of the gross heat of combustion of organic compounds as well as salts and ionic liquids Naef and Acree [260] used a commonly applicable computer algorithm based on the group additivity method to calculate the gross heat of combustion of organic molecules as well as salts and ionic liquids at standard conditions. They defined the atom groups or “backbone atoms” consisting of a central atom and its immediate neighbor atoms. The central atom is bound to at least two other atoms and is characterized by its atom name, its atom type being defined by either its orbital hybridization or bond type or its number of bonds, where required for distinction, and by its charge, if it is not zero. The neighbor atoms are collected in a term which lists all neighbors following the order H > B > C > N > O > S > P > Si > F > Cl > Br > I. This term includes the bond type of its bond with the backbone atom (if not single), its atom name, and its number of occurrences (if >1). The symbol J is used for better readability of a neighbor’s term containing iodine. If the total net charge of the neighbor atoms is nonzero, the charge is appended to the neighbor term by a “(+)” or “(−).” A nitrogen atom with three single bonds is shown by the atom type “N sp3.” Meanwhile, O and S with two single bonds are denoted with atom types “O” and “S2,” respectively, where neighbor atoms are part of a conjugated moiety, and the neighbor term is further supplemented by the terms “(pi),” “(2pi),” or “(3pi),” respectively. Table 4.6 indicates group examples and their meaning in which the backbone atom is pronounced in the “meaning” column in boldface for clarity. Table 4.6: Group examples and their meaning for organic compounds. Atom type

Neighbors

Meaning

Atom type

Neighbors

Meaning

C sp3 C sp3 C sp3 C sp3 C sp3 C sp3 C sp3 C sp3 C sp3 C sp2 C sp2 C sp2 C sp2 C sp2 C sp2 C sp2 C sp2

H3C H3N H2C2 H2CO HC3 HC2Cl HCO2 C3 N C2F2 H2=C HC=C HC=N H=CN HN=O C2=O C=CN =CNO

C–CH3 N–CH3 C–CH2–C C–CH2–O C–CH(C)–C C–CH(Cl)–C C–CH(O)–O C–C(C)2–N C–CF2–C C=CH2 C=CH–C N=CH–C C=CH–N O=CH–N O=C(C)–C C=C(C)–N C=C(N)–O

N sp3 N sp3 N sp3 N sp2 N sp2 N sp2 N(+) sp3 N(+) sp3 N(+) sp2 N aromatic N(+) sp O O O P3 P4 P4

H2C H2C(pi) C2N(2pi) H=C C=N =CO H3C H2C2 CO=O(−) :C2 =N2(−) HC HC(pi) Si2 C3 CO2=O N2O=O

C–NH2 C–N*H2 C–N*(N)–C C=NH N=N–C C=N–O C–NH3+ C–NH2+– C–NH3+ C–NH2+–C O=N+(O−)–C C:N:C N=N+=N(−) C–OH C–O*H Si–O–Si C–P(C)–C O=P(O2)–C O=P(O)(N)–N

4.2 Different approaches for prediction of the heats of combustion

137

Table 4.6 (continued) Atom type 2

C sp C sp2 C aromatic C aromatic C aromatic C sp C sp C sp C sp C sp C aromatic

Neighbors

Meaning

Atom type

Neighbors

Meaning

N=NO NO=O H:C2 a H:C:N :CN:N H#C b C#N #CN =C2 =C=O :CN:N

N=C(N)–O O=C(N)–O C:CH:C C:CH:N C:C(N):N C#CH N#C–C C#C–N C=C=C C=C=O C:C(N):N

S2 S2 S4 S4 Si Si

HC(pi) CS CO=O2 O2=O C2Cl2 OCl3

C–S*H C–S–S C–S(=O)2–O O–S(=O)–O C–SiCl2–C O–SiCl3

a

represents an aromatic bond # represents a triple bond * lone-pair electrons form π-orbital conjugated bonds with neighbor atoms.

b

Naef and Acree [260] have also extended their model by considering specific steric interactions and hydrophobic effects. Table 4.7 gives the definitions of these special groups and their explanation. Table 4.7: Special groups and their meaning. Atom type Neighbors Meaning H

H acceptor Intramolecular H bridge between acidic H (on O, N, or S) and basic acceptor (O, N, or F) at distance