5068
Ind. Eng. Chem. Res. 1997, 36, 5068-5073
Safety Charts Simulation of Nitroglycerine/Nitroglycol Spent Acids via Chemical Reaction Kinetics Fotis Rigas,* Ioannis Sebos, and Danae Doulia Safety Health and Environment Unit (SHE1), Chemical Engineering Department, National Technical University of Athens, GR-15780 Athens, Greece
Chemical instability of spent acids from nitration units has caused a great deal of severe accidents in the explosives industry. To cope with this hazard, safety charts have been constructed by various authors. In this work, evaluation of the three most commonly used safety charts found in the literature showed that they are not, in general, consistent with each other, due to different storing temperatures considered. Simulation of these experimental data was accomplished, based on simple reaction kinetics models. By using these models the existing safety charts were investigated and explained properly. In addition, the validated chemical reaction kinetics models can be used to predict the safety performance of both nitroglycerine and nitroglycol spent acids at temperatures in the range 20-70 °C. Introduction The nitration of organic compounds is a potentially dangerous process, since nitrations are exothermic reactions producing under controlled conditions explosive substances. Many accidental explosions of nitration industrial plants have been reported in the literature (Evans et al., 1977; Automatic Process of Handling Spent Acids in the Manufacture of EGDN, 1987; Urbanski, 1965; Van Dolah, 1963; Fritz, 1969). Nitric esters (e.g., nitroglycerine and nitroglycol), unlike other common explosives such as TNT, are prone to turn very unstable on prolonged contact with acids. Thus, many accidents have occurred during the manufacture of nitroglycerine and nitroglycol, particularly while handling or storing spent acid (residual acid after the nitration process) (Automatic Process of Handling Spent Acids in the Manufacture of EGDN, 1987; Urbanski, 1965). The separation of nitric esters from spent acid and their accumulation in unexpected places (e.g., transfer lines, drains, and collecting areas), building up dangerous quantities over a period of time, are the main reasons of accidents taking place at nitric ester production units. Since acidic nitric esters are chemically unstable and sensitive mixtures, all possibilities of unintentional accumulation must be avoided. The accumulation of nitroglycerine (NG) or nitroglycol (EGDN) may be attributed to a variety of reasons (Federation of European Explosives Manufacturers, 1988), namely, the following: (a) A decrease of temperature during handling of spent acid which may cause separation of NG/EGDN in storage tanks. On the contrary, an advanced decrease of temperature in separators may inversely affect the separation of NG/EDDN from spent acid due to an increase of the viscosity. (b) Insufficient dilution of spent acid with water. Dilution with water is a technique used to avoid precipitation of NG/EGDN in drains and collecting tanks. (c) Poor separation of NG/EGDN and spent acid after the reactor, due to impurities in raw materials (glycerine or glycol and mixed acid). * Author to whom correspondence is addressed. Telephone: +30.1.772.3267. Fax: +30.1.772.3163. E-mail: rigasf@ central.ntua.gr. S0888-5885(97)00418-1 CCC: $14.00
(d) Nitric esters leaking from microscopic cracks and pores which may have been developed at welded structural elements. (e) Any kind of obstruction to the flow of nitric esters/ spent acid emulsion or places where pockets of nitric esters might collect. (f) Accumulation of nitroglycerine or nitroglycol at pumps which are not self-draining. These unfortunate situations may combine with human fault (Automatic Process of Handling Spent Acids in the Manufacture of EGDN, 1987), such as (a) inappropriate handling of spent acid; (b) lack of safety mentality of workers handling this byproduct; and (c) lack of frequent inspections of the spent acid installations. In order to handle safely spent acid in the manufacture of nitric esters, safety charts have been constructed indicating regions of composition of this byproduct which can be dangerous in contact with NG or EGDN. In this work some remarks concerning decomposition of acidic nitric esters and evaluation of the three most commonly used safety charts found in the literature are presented. In addition, simulation of the safety charts was achieved based on simple reaction kinetics models, which enables us to interpolate the performance of safety charts at temperatures ranging from 20 to 70 °C. Theoretical Background of Nitric Esters Decomposition Nitric esters are hydrolyzed in an acid solution, producing the corresponding alcohol and nitric acid according to the mechanism
RCH2NO3 + H2O T RCH2OH + HNO3
(1)
Alcohols are more sensitive to wet oxidation than esters. The sensitive point is the alcohol group, which does not exist in a completely esterified molecule. Oxidation follows the reactions
RCH2OH + O f RCHO + H2O
(2)
RCH2NO3 + O f RCHO + HNO3
(3)
According to reaction (2), an increase of OH groups (dinitroglycerine or mononitroglycol content) decreases stability. Reaction (3) is important probably at higher © 1997 American Chemical Society
Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997 5069
temperatures, where decomposition is a combination of both thermal and oxidative effects. In this latter case, it is a question of breaking up radicals in the dinitroglycerine or mononitroglycol molecule. Another reaction equilibrium which may affect decompositions is the following, divided into two parts (Oehman, 1966):
RCH2ONO2 + NO+ + 3H2O T RCH2ONO + NO3- + 2H3O+ (4) RCH2ONO2 + HNO2 T RCH2NO2 + HNO3 (5) Organic nitrites may also be formed by nitrogen dioxide:
RCH2OH + 2NO2 T RCH2NO2 + HNO3
(6)
The organic nitrites are more unstable than the nitric acid esters functioning as follows:
RCH2NO2 + O T RCHO + HNO2
(7)
According to reactions (4)-(7), it is the formation of nitrites that can explain the strong decomposition effect of HNO2 and lower nitrogen oxides, as well as the autocatalytic character of decomposition reactions. Certainly the above reaction systems are not the only reactions in the complicated mechanism of decomposition. Other oxidation reactions follow, but the above is the basic reaction scenario, considering the decomposition as an exclusive molecular process. Considering radicals reactions, the complicated mechanism of decomposition may also be presented via the following radical chain scheme (Camera et al., 1982):
Initiation RCH2OH + HNO3 f RCHO + H2O + HNO2 (8) Propagation HNO2 + HNO3 T N2O4 + H2O
(9)
N2O4 T 2•NO2
(10)
RCH2OH + •NO2 f RC•HOH + HNO2
(11)
RC•HOH + HNO3 f RCHO + H2O + •NO2
(12)
Consequently, the previous reaction scheme denotes again the catalytic effect of nitrous acid, while suggesting that the most powerful oxidizing agent is •NO2, a radical formed by reaction of HNO3 and HNO2. Oxidation reactions must follow, which convert the polyalcohol being nitrated into CO2 and H2O under strongly oxidizing conditions or into ketones or organic acids (e.g., oxalic acid from glycol) under mild conditions. Oehman et al. (1960a,b, 1961) have made a thorough investigation of nitroglycerine, nitroglycol, and spent acid systems. Based on those, the following conclusions result: (a) The formation of HNO2 determines the extent of decomposition. (b) The decomposition takes place much more slowly in the spent acid phase than in the oil phase, and the autocatalytic effect follows much later. This happens because the HNO2 is bound to H2SO4, whose content is
very low in the oil phase, to form NOHSO4. Then, the next hydrolysis reaction occurs:
NOHSO4 + H2O T NO+ + H3O+ + SO42- (13) The decrease of the stabilization effect of H2SO4 by increasing the water content can be explained through equilibrium (13). In fact, NOHSO4 has a relatively high stability. (c) An increase of the water content or the number of hydroxyl groups (mononitroglycol or dinitroglycerine content) results in a decrease of stability. Furthermore, the equilibria (4) and (5) move to the right for the formation of compounds less stable than nitric esters, namely, nitrites. (d) An increase of the water concentration moves equilibrium (4) to the right, reinforcing autocatalysis, due to the formation of nitrites. (e) Nitric acid acts as a stabilizer not only causing a lower degree of hydrolysis but also influencing equilibrium (4). (f) Decomposition increases when the initial HNO2 increases to a certain level. (g) The minimum concentration of HNO2 for initiating a catalytic effect decreases as water concentration increases. (h) The spent acid system is stable between 0.5 and 1.3 mol of H2O/mol of H2SO4, regardless of its HNO3 content. At higher H2O contents the stabilizing effect of HNO3 becomes significant. Presentation of Safety Charts Since the separation from spent acid and the accumulation of nitric esters is the main reason for accidents at nitroglycerine and nitroglycol units, it becomes evident that the safety of nitration plants is strongly dependent on the composition of spent acid, provided the raw materials are in accordance with specifications. The composition and temperature of spent acid are the factors that determine the content of dissolved ester in the acid phase, as well as the content of dissolved HNO3 in the oil phase and the content of hydroxyl groups (mononitroglycol or dinitroglycerine) in the acid phase. Consequently, for the manufacture of nitroglycerine and nitroglycol it is important to know the spent acid composition regions which are safe (or dangerous) in contact with nitric esters. Thus, some practical safety charts have been constructed contributing to this aim. In this paper three of the most commonly used charts are presented in Figures 1-3 (Oehman et al., 1961; PLINKE leaflet; Automatic Process of Handling Spent Acids in the Manufacture of EGDN, 1987). The “high safety area” in Figure 1 is the area where no more than 0.2% HNO2 is formed after heating for 2 h at 70 °C and the “normal safety area” is where 0.21% HNO2 is formed under the same conditions. This is true only when working with pure materials. High HNO2 content systems broaden the danger area. Figure 2 has been constructed for 35 °C and a HNO2 content 0.2 wt % and Figure 3 for normal temperature (20 °C) and the same HNO2 content. The top region of Figure 3 is a stable area, although it is a high water content area, and according to Oehmann’s investigations (Oehman et al., 1960a,b, 1961), it should be characterized as a “dangerous area”.
5070 Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997
Figure 1. Safety chart for spent acid handling from a NG/EGDN plant at 70 °C. Reprinted with permission from BIAZZI S.A.
Figure 3. Safety chart for spent acid handling from a NG/EGDN plant at 20 °C. Reprinted with permission from UNION ESPANOLA DE EXPLOSIVOS, S.A.
Figure 2. Safety chart for spent acid handling from a NG/EGDN plant at 35 °C. Reprinted with permission from BIAZZI S.A.
The explanation here is that the system is so dilute that the oxidation power of nitric acid considerably decreases. Nevertheless, concerning their common region, Figure 2 and the corresponding part of Figure 3 satisfy the results of Oehman et al. (1960a,b, 1961) depicted in Figure 1. Comparison of Safety Charts It is obvious that the safety chart of Figure 3 is the most complete, since it covers the whole region of spent acid composition. A graphical comparison of the three safety charts is shown in Figure 4. Figure 4 consists of the couples of boundary lines of Figures 1-3 for their common region of spent acid composition. In this figure the area to the right of the right line of each boundary set is considered as dangerous area for storing or handling according to the relevant chart. Inside each pair of boundary lines the doubtful area is located with a safe storage duration from 1 to 4 weeks. To the left of the left line of each pair of boundary lines is located the safety area for storing. As shown in Figure 4, Figure 1 is the most conservative safety chart, while Figure 2 is more conservative than Figure 3. This was expected since Figure 1 refers to 70 °C, while Figures 2 and 3 refer to 35 °C and normal
Figure 4. Graphical comparison of the three safety charts for their common region of spent acid composition.
temperature, respectively. Furthermore, the authors of Figure 1 demonstrate after experimentation that, even for nitrations producing spent acids with compositions found upon the boundary lines between normal safety and danger area, spent acid has sufficient stability (Oehman et al., 1960a,b, 1961). Another point of interest is that the top boundaries of Figures 1 and 3 dividing the safe from the doubtful area have similar trends, while the corresponding lines of Figure 2 have a completely different trend, as shown in Figure 4. Thus, for the simulation of safety charts, Figure 2 was not taken into account. It must be stressed that Figures 1-3 have been prepared for pure substances. The most important impurity is HNO2, which is considered by all authors as the most significant instability factor. Chemical Reaction Kinetics from Safety Charts The top boundary lines of safety charts, which divide the safe from the doubtful area of spent acid composition, correspond to a capability of storing spent acid up
Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997 5071 Table 1. Values for A and B for ∆Gf° Estimation B×
compound
A
EGDN EGMN NG DNG H2O HNO3
-59.97 -76.74 -85.46 -102.23 -58.08 -50.56
Table 3. Rate Constants for EGDN Decomposition 102
T (°C)
10.56 8.99 16.32 14.75 1.15 4.98
20 70
k1K
K
k1 (L‚mol-1‚week-1) k2 (L‚mol-1‚week-1)
0.00265 94 0.00050 9.3
2.82 × 10-5 5.40 × 10-5
0.0082 0.0117
Table 4. Rate Constants for NG Decomposition
Table 2. Values of ∆Gr° and K for 20 and 70 °C compound
T (°C)
∆Gr° (kcal/mol)
K
EGDN EGDN NG NG
20 70 20 70
-2.64 -1.52 -2.65 -1.52
94 9.3 94 9.3
to 4 weeks or a maximum acceptable concentration of HNO2 up to 0.2% by weight, according to Oehman (Oehman et al., 1960a,b, 1961). Intending to simulate them, a simple kinetic model was developed, based on one hydrolysis and two oxidation reactions:
T (°C)
k1K
K
k1
k2
20 70
0.003 15 0.000 24
94 9.3
3.35 × 10-5 4.30 × 10-5
0.000 03 0.021 72
The rate of oxidation r1 of EGMN (DNG) is according to reaction (15):
r1 ) k1[EGMN][HNO3]
(21)
The rate of oxidation r2 of EGDN (NG) is according to reaction (16):
r2 ) k2[EGDN][HNO3]
(22)
The total rate of disappearance r of the nitric ester is
K
EGDN (NG) + H2O 9 7 8 EGMN (DNG) + HNO3 (14)
r ) r1 + r2
(23)
k1
EGMN (DNG) + HNO3 98 RCHO + HNO2 + H2O (15)
By combining eqs 17, 21, and 22, eq 23 becomes
r ) k1K[EGDN][H2O] + k2[EGDN][HNO3]
(24)
k2
EGDN (NG) + HNO3 98 RCHO + HNO2 (16) where EGMN is mononitroglycol and DNG is dinitroglycerine. The equilibrium constant K of reaction (14) is
K)
[EGMN][HNO3] [EGDN][H2O]
[H2O] ) [H2O]0 + x [HNO3] ) [HNO3]0 + x
(18)
where constants A and B may be determined as additive functions of groups. Values of A and B are tabulated in Table 1. The standard Gibbs energy of hydrolysis reaction ∆Gr° may be calculated from the ∆Gf° values of Table 1 by using eq 19 Then the equilibrium constant is given
∆Gr° ) ∆Gf°(EGMN) + ∆Gf°(HNO3) ∆Gf°(EGDN) - ∆Gf°(H2O) (19) by eq 20. Values of ∆Gr° and K for 20 and 70 °C are
K ) e-∆Gr°/RT
[EGDN] ) [EGDN]0 - x
(17)
In order to calculate the equilibrium constant, K, the standard Gibbs energy of formation ∆Gf° is first calculated for all the compounds of equilibrium (14) by using the van Krevelen and Chermin group contribution method (Reid et al., 1977). According to this method, ∆Gf° can be correlated with temperature as shown in eq 18
∆Gf° ) A + BT
The concentrations [EGDN], [H2O], and [HNO3] are given by
(20)
given in Table 2 for EGDN and NG. The values of ∆Gr° and K are identical for both EGDN and NG, if the contribution group method is used for their estimation, because these compounds belong to the same homologous series of the completely nitrated esters of a paraffinic polyhydric alcohol.
(25)
where the subscript 0 denotes the initial concentration before decomposition starts and the quantity x is the extent of the oxidation reactions. For simulation of the top boundary lines x takes the value of 0.2% by weight (0.023 mol/L) and rate r the critical value of 0.023 (mol/ L/week). For the initial concentration of [EGDN]0 the solubility of EGDN in spent acid at a certain temperature is considered. Since there is no contact of nitric esters with spent acid after their separation, this is the maximum initial concentration which diminishes gradually due to the reactions (14)-(16). Solubility data of EGDN and NG for various temperatures can be found in FEEM (1988). Values for [H2O]0 and [HNO3]0 were taken from the top boundary lines of the safety charts in weight percent units. In order to use them in eqs 24 and 25, they were first converted to mol per liter, taking into account that the specific gravity of spent acid in the region under consideration ranges from 1.74 to 1.76 (Urbanski, 1965). By fitting data from the top boundary lines for the cases of 20 and 70 °C by eq 24 and assuming a reaction rate r equal to 0.023 (mol/L)/week (equal to the critical concentration of HNO2, above which loss of control of exothermic reactions drives to an accidental explosion), values for constants k1 and k2 can be derived. The values obtained are tabulated in Tables 3 and 4 for the case of EGDN and NG, respectively. As shown in Tables 3 and 4, the principal reaction with regard to the production of the destabilizing agent HNO2 is eq 15 at 20 °C and eq 16 at 70 °C. Thus, the
5072 Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997
upon which it was based were not especially prepared for chemical kinetics estimations. Yet, it offers a first approximation of the reaction constants of principal reactions, and via the simulation obtained, it is suitable for predicting the safety boundary lines of spent acids in intermediate (between 20 and 70 °C) temperatures. Conclusions
Figure 5. Boundaries between the safe and doubtful areas for various temperatures (in °C) for EGDN.
Evaluation of the three most commonly used safety charts found in the literature, demonstrating spent acid compositions which are safe in contact with NG or EGDN, showed that they are not, in general, consistent with each other, due to different storing temperatures considered. Nevertheless, two of them have similar trends, while the third one has a completely different trend. Simulation of the two experimental safety charts with similar trends, based on simple reaction kinetics models, resulted in the estimation of reaction rate constants for 20 and 70 °C. The dependence of the safety boundary lines on temperature was accomplished by making use of the Arrhenius equation, and it was used for the prediction of intermediate boundary safety lines in the temperature range 20-70 °C. The reaction kinetics simulation obtained has only an approximate character, taking into account that the data used were not properly prepared for chemical kinetics estimations. Nomenclature
Figure 6. Boundaries between the safe and doubtful areas for various temperatures (in °C) for NG.
partially nitrated glycol/glycerine promotes mainly production of HNO2 at 20 °C, whereas the fully nitrated glycol/glycerine is the main precursor of the production of HNO2 at 70 °C. This can be explained by the observation that the equilibrium constant K increases by an order of magnitude, when moving from 20 to 70 °C (Tables 3 and 4). Consequently, reaction rate r1 could be used alone at temperature 20 °C and reaction rate r2 at 70 °C, instead of their sum. The Arrhenius equation (eq 26) can be fitted to data from Tables 3 and 4, in order to define the safety boundary lines with temperature.
ki ) ko,ie-E/RT,
where i ) 1 or 2
(26)
Thus, using eq 26, the safety performance of a spent acid mixture in temperatures between 20 and 70 °C can be predicted. Figures 5 and 6 show the top boundary line (the line between the safe and the doubtful area) for 35 and 50 °C for the case of EGDN and NG, respectively. In Figure 5 the experimental results of the safety chart found in PLINKE’s technical leaflet are also plotted for comparison. Thus, it is obvious once again the PLINKE chart does not fit well with the results obtained by Oehman et al. (1961) and SAFEX (1987), which give a quite different trend. It should be stressed, finally, that this work is not indeed a chemical kinetics analysis, since the results
x: extent of reaction DNG: dinitroglycerine EGDN: nitroglycol EGMN: mononitroglycol K: equilibrium constant k: rate constant NG: nitroglycerine r: reaction rate r1: rate of oxidation of EGMN (DNG) r2: rate of oxidation of EGDN (NG) T: temperature ∆G°: free energy Subscripts f: i: 0: r:
formation index taking the values of 1 and 2 initial situation reaction
Literature Cited Automatic Process of Handling Spent Acids in the Manufacture of EGDN. SAFEX 9th International Congress, Casablanca, Morocco, June 1987; pp 131-137. Camera, E.; Modena, G.; Zotti, B. On the Behaviour of Nitrate Esters in Acid Solution. II. Hydrolysis and Oxidation of Nitroglycol and Nitroglycerin. Propellants, Explos., Pyrotech. 1982, 7, 66. Evans, F. W.; Meyer, P.; Oppliger, W. Safety Consideration in the Development and Design of an Industrial Nitration Plant. Proceedings of the International Symposium on Loss Prevention and Safety Promotion in the Process Industry, Heidelberg Germany, Sept 1977; Vol. IV, p 191. Published by Dtsch Ges. fuer Chem. Apparatwesen (EFCE-Eur. Fed. Chem. Eng.) Publ. Ser. n1, 1978. Federation of European Explosives Manufacturers (FEEM). Code of Good Practice: Nitroglycerine/Nitroglycol Handling of Spent Acid Effluents. Publication No. 11, 1988. Fritz, E. J. Anatomy of a Nitration Explosion. Loss Prev. 1969, 3, 41.
Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997 5073 Oehman, V. Redox Measurements Show Safe Nitration Conditions. Sven. Kem. Tidskr. 1966, 78 (1), 20. Oehman, V.; Camera, E.; Cotti, L. The Stability of Acidic Nitroglycerine and Spent Acid. Part I. Explosivstoffe 1960a, 6, 120. Oehman, V.; Camera, E.; Cotti, L. The Stability of Acidic Nitroglycerine and Spent Acid. Part II. Explosivstoffe 1960b, 7, 148. Oehman, V.; Camera, E.; Cotti, L. The Stability of Acidic Nitroglycol System. Explosivstoffe 1961, 5, 95. PLINKE GmbH & Co. Chemieanlagen KG (spent acid concentration company) leaflet. Reid, R. C.; Prausnitz, J. M.; Sherwood, T. K. The Properties of Gases and Liquids, 3rd ed.; McGraw-Hill: New York, 1977.
Urbanski, T. Chemistry and Technology of Explosives; Pergamon Press: Oxford, U.K., 1965; Vol. 2. Van Dolah, R. W. Detonation Potential of Nitric Acid Systems. Loss Prev. 1963, 3, 32.
Received for review June 9, 1997 Revised manuscript received September 8, 1997 Accepted September 17, 1997X IE970418E
X Abstract published in Advance ACS Abstracts, November 1, 1997.
