Energy & Fuels 1998, 12, 547-553
547
Pseudo-Detailed Chemical Kinetic Modeling of Antioxidant Chemistry for Jet Fuel Applications Steven Zabarnick† Aerospace Mechanics Division, University of Dayton Research Institute, 300 College Park, KL-463, Dayton, Ohio 45469-0140 Received September 2, 1997. Revised Manuscript Received February 13, 1998
Chemical kinetic modeling was used to simulate the autoxidation of jet fuel including the chemistry of peroxy radical inhibiting antioxidants and hydroperoxide decomposing species. Recent experimental measurements of oxygen concentration during autoxidation of model hydrocarbon solvents were used to “calibrate” the rate parameters of the mechanism. The model showed good agreement with oxygen profiles of static measurements at 140 °C. At this temperature, the model predicts large increases in oxidation rate upon peroxy radical inhibiting antioxidant consumption to below 1 × 10-5 M. At 185 °C we have shown that peroxy radical inhibiting antioxidants and hydroperoxide decomposers both slow and/or delay oxidation, but the resulting oxygen profiles display different characteristics. We have shown that comparison of these profiles with fuel blending and fuel dilution measurements has the potential to differentiate between the two types of oxidation-slowing species. The modeling predicts that the presence of both types of species in a fuel results in a synergistic behavior.
Introduction Liquid hydrocarbons, and liquid hydrocarbon fuels such as jet fuel in particular, undergo a series of reactions in the presence of oxygen which together are often called “autoxidation.” These reactions occur at moderate temperatures (∼100-300 °C) and produce oxygenated species, such as hydroperoxides, alcohols, ketones, aldehydes, and acids. Subsequent poorly understood reactions can produce particulates, gums, and solid deposits. In general, the production of these and other oxidized species is highly undesirable. For example, in lubricating oils the reaction with oxygen produces acids which are corrosive.1 The production of gums in gasoline can cause injector/carburetor plugging, valve malfunctions, and piston and crankcase fouling.2 In military supersonic aircraft, where jet fuel is used for cooling of lubricating oils and hydraulic fluids, the formation of deposits has been implicated in the fouling of injector nozzles, fuel manifolds, and main engine controls.3 Antioxidants have been used for many decades to slow and/or delay autoxidation. Indeed, the development of antioxidants has been so successful that they are used routinely in many industrial and consumer products, including foods and gasoline. There are many types of antioxidants that have been used in hydrocarbon liquids: hindered phenols, phenylenediamines, peroxide decomposers, and metal deactivators. It is believed that hindered phenols and phenylenediamines slow oxidation † Telephone: (937) 255-3549. E-mail:
[email protected]. (1) Murphy, C. M.; Ravner, H.; Smith, N. L. Ind. Eng. Chem. 1950, 42, 2479-2489. (2) Walters, E. L.; Minor, H. B.; Yabroff, D. L. Ind. Eng. Chem. 1949, 41, 1723-1729. (3) Hazlett, R. N. Thermal Oxidation Stability of Aviation Turbine Fuels; ASTM: Philadelphia, 1991.
by intercepting the alkylperoxy free radicals that carry the autoxidation chain. Peroxide decomposers react with hydroperoxides and prevent their decomposition into free radicals. Metal deactivators complex with metal ions and prevent their catalysis of processes which form free radicals. Thus, each of these antioxidant types is thought to lower the concentration of free radicals by either preventing their formation or rapidly removing them after they are formed. Although antioxidants are widely used and have proven very successful in preventing the formation of detrimental products, chemical kinetic modeling of the chemical mechanism of antioxidant action has received little attention. Fortunately, the primary reactions by which hindered phenol antioxidants intercept alkylperoxy radicals have been studied extensively. Previously, we have performed chemical kinetic modeling of antioxidant action for jet fuel applications.4 This study was limited to relatively short reaction times before significant consumption of the antioxidant species occurred. Also, because of the lack of suitable experimental data, no comparison with laboratory measurements was included. Recent measurements of oxygen concentration during the autoxidation of jet fuel, jet fuel mixtures, and model jet fuel species have provided reasonable data for comparison with modeling.5-8 In the present study we utilize these experimental data (4) Zabarnick, S. Ind. Eng. Chem. Res. 1993, 32, 1012-1017. (5) Jones, E. G.; Balster, L. M.; Balster, W. J. Energy Fuels 1996, 10, 509-515. (6) Zabarnick, S.; Whitacre, S. D. Aspects of Jet Fuel Oxidation. Proceedings of the 1997 ASME Turbo Expo Conference; ASME: New York, 1997; Paper No. 97-GT-219. (7) Zabarnick, S.; Zelesnik, P.; Grinstead, R. R. J. Eng. Gas Turbines Power 1996, 118, 271-277. (8) Balster, L. J.; Balster, W. J.; Jones, E. G. Energy Fuels 1996, 10, 1176-1180.
S0887-0624(97)00157-6 CCC: $15.00 © 1998 American Chemical Society Published on Web 04/08/1998
548 Energy & Fuels, Vol. 12, No. 3, 1998
Zabarnick
Table 1. Reaction Mechanism for Chemical Kinetic Modeling
reaction
Arrhenius A factor (mol, L, s)
activation energy (kcal/mol)
reaction no.
I f R• R• + O2 f RO2• RO2• + RH f RO2H + R• RO2• + RO2• f termination RO2• + AH f RO2H + A• AO2• + RH f AO2H + R• A• + O2 f AO2• AO2• + AH f AO2H + A• AO2• + AO2• f products R• + R• f R2 RO2H f RO• + •OH RO• + RH f ROH + R• RO• f Rprime• + carbonyl •OH + RH f H O + R• 2 RO• + RO• f RO• termination • Rprime + RH f alkane + R• RO2H + SH f products
1 × 10-3 3 × 109 3 × 109 3 × 109 3 × 109 3 × 105 3 × 109 3 × 109 3 × 109 3 × 109 1 × 1015 3 × 109 1 × 1016 3 × 109 3 × 109 3 × 109 3 × 109
0.0 0.0 12.0 0.0 5.0 10.0 0.0 6.0 0.0 0.0 42.0 10.0 15.0 10.0 0.0 10.0 0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
for “calibration” of the kinetic parameters of the model. Also, we compare the predicted and measured oxygen profiles to provide insight into the antioxidant mechanism. In addition, we explore the type of behavior expected by the presence of peroxide-decomposing species. This study will also address relatively long reaction times, where the antioxidant species may become completely consumed. Methodology The methodology used in the chemical kinetic modeling reported here has been reported previously4 and therefore will only be described briefly. The chemical kinetic modeling was performed using the modeling package REACT.9 This is a relatively simple package that integrates the multiple differential equations that result from a detailed chemical kinetic mechanism and yields species concentrations as a function of time. This code does not solve the energy equation and therefore does not include energy production or removal due to chemical reaction. The following is input to the code: reaction mechanism with rate constants, initial concentrations for each species present in the mechanism, reaction time and time intervals for output, and various tolerances for the precision of the computation. The code outputs the individual species concentrations at each time interval. Because jet fuels are mixtures of hundreds of chemical species, it is impractical to model the chemical changes of all components of the mixture. Instead we have chosen to model the bulk fuel as a single compound, RH. RH has the chemical properties of a straight-chain alkane, such as n-dodecane; this is a reasonable simplification as alkanes comprise a large proportion of a jet fuel mixture. This single compound “fuel” can also contain dissolved oxygen (O2), an antioxidant species (AH), and a peroxide decomposer (SH). The reaction mechanism used in the present study is shown in Table 1. We refer to the mechanism as being “pseudodetailed” as it has some of the characteristics of detailed reaction mechanisms, i.e., real chemistry with physically meaningful rate parameters, but is less simplified than the simple empirical modeling using one or two reac(9) Whitbeck, M. Tetrahedron Comput. Methodol. 1990, 3, 497-505.
tions that have been used in jet fuel modeling in the past. The mechanism is a modification of that used previously.4 Modifications involve adjustment of rate parameters to better match experimental measurements and the inclusion of a “peroxide decomposer” species. These modifications will be detailed below. In the previous modeling study,4 we presented a reaction mechanism of jet fuel autoxidation and estimated rate parameters (Arrhenius “A factors” and activation energies) for these pseudo-elementary reactions. These are not true elementary reactions due to the modeling of all fuel species as a single species, RH. The rate constant parameters were estimated using a combination of techniques: comparison with measured rate constants and Benson style “thermochemical kinetics” analysis. Also, some of the rate parameters were adjusted and the effects of these changes were observed in the output. The rate constants, k, were estimated in Arrhenius form, i.e., k ) A exp(Ea/RT); thus, an activation energy, Ea, and an Arrhenius “A factor” were estimated for each reaction. The mechanism used in the current study is a modified version of the previous mechanism. These modifications are based on comparisons of the calculated oxygen concentration time behavior with experimental measurements. The mechanism has been described in detail previously4 and will only be summarized here. In Table 1, R• is a hydrocarbon alkyl radical species, RH represents the bulk fuel as a single hydrocarbon compound, AH is an antioxidant species (i.e., a species with an easily abstractable hydrogen atom, also referred to in the literature as an inhibitor), O2 represents the dissolved oxygen present in the liquid fuel, SH is a peroxide decomposing species, and I is an initiator species. The alkyl radicals required to begin the autoxidation chain are produced by an initiation step, reaction 1. For simplicity we use a unimolecular decomposition of I into R• radicals to represent the poorly understood initiation process. This is an obvious oversimplification of the poorly understood initiation process, but its inclusion is required to begin the chain with a very small production rate of R• radicals. An activation energy of zero was chosen for simplicity, although it is most likely that the real initiation rate increases strongly with temperature. The A factor was chosen along with the initial I concentration to produce an initial R• radical production rate that is large enough to start the autoxidation chain, yet small enough to remain an insignificant source of radicals once the chain has begun. Reaction 2 converts these R• radicals to alkylperoxy radicals, RO2•, by reaction with dissolved O2. This addition reaction is expected to have a near-zero activation energy, and thus proceeds very rapidly. In reaction 3, RO2• abstracts a hydrogen atom from a fuel molecule, generating a hydroperoxide, RO2H, and another R• radical, thus propagating the chain. At sufficiently high temperatures, the hydroperoxide will decompose to yield the additional radicals, RO• + OH (reaction 11). These radicals then undergo additional reactions (reactions 12-16), forming alcohols, carbonyl compounds, water, and other species. If the concentration of RO2• radicals is sufficient, the termination reaction, reaction 4, can occur; this reaction can produce aldehydes, alcohols, and ketones and may regenerate oxygen by a disproportion-
Autoxidation of Jet Fuels
Energy & Fuels, Vol. 12, No. 3, 1998 549
Table 2. Initial Concentrations of Species species
initial concn (M)
I O2 RH AH SH all other species
1.0 × 10-7 1.8 × 10-3 4.4 varied, but typically 2.0 × 10-4 varied 0.0
ation pathway. For simplicity, we have chosen to disregard the subsequent reactions of the products of reaction 4. When the dissolved oxygen concentration becomes low, R• radicals predominate over RO2• and these can self-terminate by reaction 10. When an antioxidant species, AH, is present, reactions 5-9 become important. Also, when a peroxide-decomposing species, SH, is present, reaction 17 may play an important role. The AH and SH reactions will be addressed in detail below. The initial concentrations of each of the species present in the mechanism are also required input to the modeling code. These initial concentrations are listed in Table 2. The RH concentration was estimated from the room-temperature molarity of n-dodecane liquid (molecular weight of 170 amu and density of 0.752 g/cm3). The dissolved oxygen concentration was estimated from the review of oxygen solubility by Battino et al.10 and from a compilation of jet fuel properties.11 From these papers we estimate the dissolved oxygen concentration at room temperature and one atmosphere air in jet fuels and hydrocarbons as ∼79 ppm (by weight). Therefore, we have used 1.8 × 10-3 M for the oxygen concentration. The antioxidant concentration, AH, was varied over a wide range in the calculation, so a typical value of 2.0 × 10-4 M was listed in the table. A very small concentration (1.0 × 10-7 M) of the initiator species, I, was included for the initial production of R• radicals required to get the chain mechanism started. The peroxide-decomposing species concentration was varied. Calibration of the Model A modified version of the previous mechanism was used in the present study.4 To provide agreement with the measured oxidation rates, the rate constant parameters of reactions 1, 3, and 8 were adjusted. The oxygen concentration versus time data of Balster et al.8 was used to “calibrate” the rate parameters in the absence of antioxidant species, AH. Balster et al. measured the oxygen concentration at the outlet of a tubular heat exchanger operated isothermally at 185 °C. These measurements were performed as a function of flow rate to provide various residence times in the heated tube. For the present modeling study, we are most interested in the results of Balster et al. on the oxidation of the paraffinic/cycloparaffinic solvent Exxsol D-80. Exxsol D-80 is a dearomatized, narrow-cut aliphatic solvent with
