New method estimates the parameters for evaluating process reactivity hazards

June 23, 2003
A new method of evaluating process hazards combines experimental thermal analysis, computational chemistry methods, and thermodynamic-energy relationships. This results in a better understanding of reaction stoichiometry for a given process, which allows better estimates of system thermodynamics and kinetics that are necessary to evaluate hazards of runaway reactions due to chemical reactivities.

A new method of evaluating process hazards combines experimental thermal analysis, computational chemistry methods, and thermodynamic-energy relationships. This results in a better understanding of reaction stoichiometry for a given process, which allows better estimates of system thermodynamics and kinetics that are necessary to evaluate hazards of runaway reactions due to chemical reactivities.

Identifying the major reaction pathways provides the most important information, namely pressure and heat hazards due to a loss of control leading to thermal or physical system failure.

The new method focuses research on the most likely and most hazardous reaction stoichiometry and reduces the need for detailed experimental analysis. More detailed and advanced experimental analysis is still required for more complex systems.

Process hazards

Reactive chemicals can exist in every unit of a chemical process. These chemicals are stable for certain operating conditions, but sometimes a slight deviation of these operating conditions will trigger unwanted reactions.

These reactions can be highly exothermic and can lead to a thermal runaway reaction, explosion, or fire. In another cases, these reactive chemicals lead to unwanted but thermally neutral reactions. These reactions may generate a large amount of gas, which can cause equipment failures.

Characterizing the reactive chemicals is the ultimate goal in a hazard evaluation. The primary difficulty in this characterization is the large number of chemicals and wide variety of conditions under which these chemicals may undergo an uncontrollable hazardous reaction.

To understand the role of process chemistry as a root cause of runaway situations, one must determine the rate and maximum quantities of energy and gas that are generated by the primary and secondary exothermic reactions.

This article discusses the application of thermal analysis techniques, computational chemistry models, and thermodynamic-energy relationships to estimate reactivity evaluation parameters. The combination of these techniques helps minimize experimental work and provides the required parameters for evaluating reactivity hazards and a more comprehensive understanding of process chemistry.

Thermal stability

To increase safety, the chemical industry designs equipment to prevent, control, or withstand runaway reactions. Temperature, concentration, impurities and contaminants, presence of air, confinement, and type of solvent all influence chemical stability.

Thermally unstable chemicals have caused many catastrophic incidents in the past few years. When cooling systems fail to absorb energy generated by unwanted exothermic reactions, a temperature and pressure buildup can lead to a runway reaction and result in thermal explosions or fires.

System contamination is another crucial source of thermal instability. Halle and Vadekar reported one example of iron and iron oxide from rusting equipment that catalyzed an ethylene hydrogenation reaction and caused a runaway scenario (OGJ, June 17, 1991, pp. 33-36).

In most of the reactivity-related incidents, a lack of reactive chemistry knowledge was one of the most frequent root causes.

Strategies for safe and economic processes require accurate information about chemicals and chemical reactions, but sufficient information is often unavailable from traditional sources. Using calorimetric measurements and classical and quantum chemistry models to determine properties of chemicals and analyze chemical systems is an efficient approach.

Molecular simulation and computational methods are practical engineering tools that allow designers to predict thermophysical properties, estimate reaction rates, and understand the molecular-level causes of macroscopic behavior measured in the laboratory.

Molecular simulation models, such as the analysis of unstable reaction systems that cause thermal runaway reactions, can supplement experimental measurements. We discussed this strategy of chemical reactivity evaluation in a previous work.1

Parameters

An effective evaluation of chemical reactivity must be based on process operating conditions, chemistry mechanisms, conditions under which chemical reactive hazards can appear, and parameters for quantifying reactive chemical hazards. One must identify these conditions and parameters to design optimum, safe, and economical processes.

One must evaluate the maximum quantities, and rate of production, of energy and gas generated by primary and secondary exothermic reactions to evaluate hazards of reactive systems. The assessment process must identify reaction stoichiometry.

Click here to enlarge image

Fig. 1 shows that an estimate of thermodynamic, kinetic, and stoichiometric parameters is the first step in quantifying the reactivity evaluation parameters. Those parameters describe the severity of a reactive system's exothermic behavior and include:

  • The initiation temperature of significant exothermic behavior, Tonset.
  • Maximum adiabatic temperature increase, Tad.
  • Heat of reaction with adiabatic conditions, ΔHr,ad.
  • Reaction mixture boiling point, Tbp.

Maximum temperature attained in the reaction runaway, Tmax.

Required time to reach maximum reaction rate with adiabatic conditions, TMRad.

Reaction activation energy, Ea.

Fig. 1 shows that these parameters depend on the availability of stoichiometric, thermodynamic, and kinetic parameters. The accuracy of this reactivity evaluation depends directly on the quality of the estimated parameters.

For simple systems, such as single-reaction systems, this evaluation is acceptable when one uses approximate thermal analysis techniques. For complex, multireaction systems, this evaluation is based on overall thermodynamic and kinetic measurements, which provide poor or no description of the reaction mechanism.

One must understand reaction mechanisms or pathways when evaluating reactivity hazards. This helps focus the attention on the most reactive components (reactants, intermediates, or products), and it helps explain the pressure behavior.

Experimental thermal analysis is the traditional procedure to evaluate reactivity; however, two other sources give helpful estimates: computational chemistry models and thermodynamic-energy relationships.

Thermal analysis

Most safety and thermal reaction risk estimations are based on the exact characterization of a reaction system including reaction stoichiometry, thermodynamic, and kinetic parameters.

A thermal calorimetric analysis is the traditional approach for the most accurate evaluation of chemical reactivity; most research has focused on applying thermal analysis techniques to evaluate reactive chemical hazards.

Because assessment procedures vary from one industry to another, many different thermal analysis techniques and procedures exist for categorizing reactive chemicals.

Applied techniques range from rapid-screening thermal analysis techniques to more accurate advanced techniques. Some of these thermal analysis devices are used for various applications, while others are specifically designed for chemical reactivity hazard evaluation.

The experimental approach is accurate and can simulate a real runaway scenario; however, due to the large number of chemical compounds and reaction scenarios in production processes, experimental evaluation is expensive and time consuming. Thermal analysis testing is also hazardous for highly energetic or toxic materials.

A reactive chemical hazard evaluation requires reliable estimates of the kinetic parameters for the main reaction pathway. Calorimetric analysis provides acceptable parameters for homogeneous and one-dimensional chemical systems.

For complex reactive systems, experimental procedures provide overall system thermodynamic and kinetic data but do not explain reaction stoichiometry. A system analysis is often required in addition to laboratory measurements.2

Thermal analysis techniques will continue to be a major source of reactivity evaluation data; however, more theoretical data can generate thermodynamic and kinetic parameters, and help identify important reaction pathways.

Computational chemistry models

Thermal analysis without additional screening steps is time consuming and expensive due to the large number of possible reaction pathways, even for relatively simple systems. Unexpected exothermic secondary reactions can dictate the magnitude and time scales of heat releases during a runaway reaction. This makes it difficult to interpret the data using experimental techniques.3

Various techniques can generate thermodynamic and kinetic parameters of reactive systems such as molecular contribution methods and computational quantum chemistry. No single theoretical chemistry method yields satisfactory results for all chemical systems. The usefulness of the various methods depends on system size, calculated property, and calculation costs.

Molecular contribution methods are theoretical techniques that use bond and group contributions in known chemical structures to estimate system thermodynamic parameters like Gibbs free energy, heat of formation, and heat of reaction.

Many group contribution methods are available; Benson's method4 used in a software program5 is the most widely accepted method.

These methods are based on correlations based on experimental values of thermodynamic properties for common molecules. Thermodynamic predictions are valuable if the needed molecular groups are available in the database.

Occasionally, these methods cannot differentiate among the various molecular configurations such as isomers, leading to large deviations in calculated enthalpies.

Computational quantum chemistry is based on molecular quantum theory; the motion and distribution of electrons is described in terms of electron probability distributions or molecular orbitals.6

Numerical techniques perform the quantum chemistry calculations. Techniques include the density functional theory, Hartree-Fock, Gaussian 2, and semiempirical parameter techniques.

The fundamental quantum chemistry methods, also called ab-initio methods, are combined with statistical thermodynamics to estimate thermodynamic properties such as enthalpy and entropy of formation of reactants and products, reaction enthalpy and entropy, Gibbs free energy of the ideal gas reaction, and Gibbs free energy of mixing.

Thermodynamic energy relationships

One can understand a reactive system's chemistry by investigating the relation between thermodynamic and kinetic parameters.

Evans and Polanyi78 examined the relationship between the thermodynamics of a reaction and the activation barrier, Ea. They showed empirically that as a reaction becomes more exothermic, its activation barrier generally decreases.

Evans and Polanyi also noted that, in many cases, the activation barrier for a given reaction is related to the heat of reaction:

Ea = Eºa + γP ΔHr

where:

Ea = Activation barrier, kcal/gmole

a = Intrinsic barrier of the reaction, kcal/gmole

ΔHr = Heat of reaction, kcal/gmole

γP = Transfer coefficient, kcal/gmole.

Masel gives a detailed derivation of Equation 1.9 10 Masel9 and Bockris, et al.,11 present a procedure to find γP.

The intrinsic activation barrier is the energy required to distort the reactant orbitals to the transition-state geometry. The intrinsic barriers show that some elementary reactions are more difficult than others. Intrinsic barriers can also show that one reaction pathway is favored over another.12

The Polanyi equation might not work for all cases. For highly exothermic reactions, the Polanyi equation can result in negative activation energy, which is physically unacceptable, or it may start showing deviations from experimental activation energies.13

In this case, the Marcus equation, which is a quadratic extension of the Polanyi equation, is more accurate:1415

Ea = (1+ ΔHr/4Eºa)2 Eºa

This equation has a better agreement to activation barriers for large ΔHr.13 Masel presents a derivation of equation 2.910

Masel and Lee12 concluded that Eºa and γP are constant for the same elementary reaction mechanism and that the type and size of the transferring ligand has a negligible effect on the activation barrier prediction if the heat of reaction is known.

Values for Eºa and γP are available for many reaction mechanisms. The heat of reaction for each elementary reaction step or the overall reaction is measured using thermal analysis or estimated using computational quantum models; therefore, the Polanyi or Marcus equations can calculate the activation energy barriers for each elementary reaction step. Reaction pathways with lower activation energy barriers are the more likely pathways.

If the values of intrinsic barriers and transfer coefficients are not available for a certain reaction mechanism, one can predict them using experimental or theoretical activation energies and heats of reaction values for similar reactive compounds.

One should predict the constants of intrinsic barriers and transfer coefficients for similar groups of compounds to minimize the effects of ligands differences on the predicted energies. This makes the Polanyi and Marcus equations most efficient.

Example calculation

We tested the runaway behavior of styrene-acrylonitrile copolymerization to show the applicability of the new method for evaluating chemical reactivity.

A rapid increase in reaction rate and an accelerating temperature rise characterize thermal runaways in copolymerization reactors. Thermal runaway causes large temperature rises and possible instability as well as a sharp reduction in polymer-copolymer molecular weight and an increased spread in molecular weight distribution.16

We used an adiabatic calorimeter to perform a calorimetric analysis on styrene (99+%) and acrylonitrile (99+%) monomers at styrene:acrylonitrile (S:AN) weight feed ratios of 80:20, 70:30, 60:40, 50:50, 40:60, 30:70, and 20:80.

Click here to enlarge image

Fig. 2 shows temperature profiles at S:AN ratios of 80:20 and 20:80.

Table 1 shows the reaction onset temperature, heat of reaction, and Arrhenius parameters for the styrene-acrylonitrile copolymerization runaway. The monomer feed ratio does not significantly affect the reaction onset temperature, Tonset.

Click here to enlarge image

On the other hand, there is a slight reduction in the heat of reaction as acrylonitrile concentration increases.

The temperature and pressure profiles show that additional reaction activity occurs after the maximum temperature and pressure are reached. This activity is noticeable at acrylonitrile monomer feed compositions greater than 40%. At these acrylonitrile feed rates, there are two maximum temperature and pressure peaks.

The first pressure peak corresponds to the copolymerization reaction. Styrene-acrylonitrile in bulk will copolymerize in the vapor phase.17 Initially, monomers will evaporate causing a pressure increase and then copolymerize to the liquid phase causing the pressure to decrease; this activity forms the first pressure peak.

Due to the high temperature increase caused by a thermal runaway, acrylonitrile monomers start to decompose exothermally. Decomposition products cause the temperature and pressure to further increase, which forms the second maximum peaks. As the feed acrylonitrile monomer increases, the second decomposition temperature and pressure peaks reach higher values.

We conducted a theoretical evaluation to analyze the styrene-acrylonitrile reaction mechanism.

We mainly wanted to predict the relative tendency for homopolymerization and cross-propagation copolymerization reactions. We calculated reaction enthalpies using a semiempirical computational method.18

Because styrene and acrylonitrile are nonsymmetric molecules, there are two reactive ends for each molecule; therefore, we had to consider reactive site orientation in the computations. We used the Polanyi equation to predict the most favorable reaction stoichiometry for the homopolymerization and copolymerization reactions.

Stoichiometries were presented in a previous work.7 Those analyses indicated that, in the presence of two monomers in the mixture, they will probably experience a cross-propagation mechanism.

Comparing these calculations to the experimental results showed that as the concentration of acrylonitrile increases, the copolymerization must slow down because most of styrene monomers are consumed either by the cross-propagation copolymerization reaction or by the homopolymerization reaction.

Because acrylonitrile has a low tendency for homopolymerization in the presence of styrene monomer, most of the uncopolymerized acrylonitrile monomers will decompose at higher temperatures due to the copolymerization runaway; this activity causes another temperature and pressure increase.

Reaction enthalpy is a function of the chain end active site regardless of what is attached to that site on the other side. This is consistent with the random arrangement of monomers in the styrene-acrylonitrile copolymer chain.

We could not have predicted this reaction mechanism based on the calorimetric analysis. The combination of experimental and theoretical methods increased our confidence in the analysis of this system.

Acknowledgment

The authors thank the Mary Kay O'Connor Process Safety Center of the Department of Chemical Engineering at Texas A&M University for sponsoring this research.

References

1. Aldeeb, A., Rogers, W., and Mannan, M., "Theoretical and experimental methods for the evaluation of reactive chemical hazards," Tran. Inst. Chem. Eng., Part B, Vol. 80 (2002), No. 3, pp. 141-49.

2. Maria, G., and Heinzle, E., "Kinetic system identification by using short-cut techniques in early safety assessment of chemical processes," J. Loss Prev. Process Ind., Vol. 11 (1998), pp. 187-206.

3. Bruneton, C., Hoff, C., and Barton, P., "Thermal stability analysis via elucidation of hazardous reaction stoichiometries," Computers Chem. Eng., Vol. 22 (1998), No. 6, pp. 735-45.

4. Benson, S., Thermochemical Kinetics, 2nd edition, John Wiley Inc., New York, 1976.

5. ASTM, "CHETAH: the ASTM Computer Program for Chemical Thermodynamic and Energy Release Evaluation," v. 7.2, American Society for Testing and Materials, West Conshohocken, Pa., 1998.

6. Bruneton, C., Hoff, C., and Barton, P., "Computer aided identification of chemical reaction hazards," Computers Chem. Eng., Vol. 21 (1997), pp. S311-17.

7. Evans, M., and Polanyi, M., "Further considerations on the thermodynamics of chemical equilibria and reaction rates," Trans. Faraday Soc., Vol. 32 (1936), No. 9, pp. 1333-60.

8. Evans, M., and Polanyi, M., "Inertia and driving force of chemical reactions," Trans. Faraday Soc., Vol. 34 (1938), pp. 11-29.

9. Masel, R., Chemical Kinetics and Catalysis, John Wiley & Sons Inc., New York, 2001.

10. Masel, R., Principles of Adsorption and Reaction on Solid Surfaces, New York, John Wiley & Sons Inc., 1996.

11. Bockris, J., Reddy, A., and Gamboa-Aldeco, M., Modern Electrochemistry: Fundamentals of Electrodics, Vol. 2A, 2nd edition, New York, Kluwer Academic/Plenum Publishers, 2000.

12. Masel, R., and Lee, W., "Intrinsic activation energy barriers as a guide to mechanisms of reactions in the gas phase and on solid surfaces," J. Catal., Vol. 165 (1997), pp. 80-90.

13. Hupe, D., and Wu, D., "The effect of solvation on Brønsted b values for proton transfer reactions," J. Am. Chem. Soc., Vol. 99 (1977), No. 23, pp. 7653-59.

14. Marcus, R., "On the theory of oxidation-reduction reactions involving electron transfer: I," J. Chem. Phys., Vol. 24 (1956), No. 5, pp. 966-78.

15. Marcus, R., "Theoretical relations among rate constants, barriers, and Brønsted slopes of chemical reactions," J. Phys. Chem., Vol. 72 (1968), No. 3, pp. 891-99.

16. Sebastian, D., and Biesenberger, J., "Chemical Reaction Engineering, The Fifth International Symposium on Chemical Reaction Engineering," Weekman, V., Jr., and Luss, D., editors, meeting of the American Chemical Society, Houston, 1978, pp. 173-86.

17. American Cyanamid Co., The Chemistry of Acrylonitrile, New York, 1959.

18. Dewar, M., and Thiel, W., J. Am. Chem. Soc., Vol. 99 (1977), p. 4899.

Click here to enlarge image

The authors
Abdulrehman A. Aldeeb is a chemical engineering PhD candidate at Texas A&M University, College Station. He joined the Mary Kay O'Connor Process Safety Center in 2000 as a graduate researcher. Aldeeb previously worked in the pharmaceutical industry as a research and development chemical engineer. He holds a BS in chemical engineering from Jordan University of Science & Technology, Irbid, Jordan, and an MS in environmental engineering from the University of Texas, Arlington.

Click here to enlarge image

William J. Rogers is laboratory director at the Mary Kay O'Connor Process Safety Center and a Texas Engineering Experiment Station research scientist at Texas A&M University. His research interests include thermal behavior of chemical processes, chemical hazard characterization, aerosols, experimental design, and computational chemistry for property estimations. Rogers holds a PhD in physical chemistry from Ohio State University.

Click here to enlarge image

M. Sam Mannan is a professor of chemical engineering and director of the Mary Kay O'Connor Process Safety Center at Texas A&M University. His research interests include hazard assessment and risk analysis, modeling flammable and toxic gas cloud dispersions, inherently safer design, reactive chemicals and runaway reactions, aerosols, and abnormal situation management. Mannan holds a PhD in chemical engineering from the University of Oklahoma.