Acrylonitrile (AN) was first discovered in 1893 and copolymerized with butadiene during WW II to form oil resistant rubber for the military. When acrylonitrile polymerizes, the CH2=CH breaks releasing 17.3 kcal/gmol, making the reaction difficult to control. The polymer is not soluble in the monomer. Polymerization also occurs in the vapor phase. A conservative polymerization heat of reaction of -20 kcal/gmol (-35,977 BTU/lbmol) is predicted by breaking the C= bond to -C- (-377 cal/g or -678 BTU/lb).
This article by Georges Melhem, Ph.D., FAIChE, and Harold G. Fisher, FAIChE, presents methods for assessing existing relief systems — specifically when sodium hydroxide and ammonia chemicals are present — using calorimetry data. Design recommendations for risk management are also included. We conducted a total of 21 calorimetry tests, using the ARC® and APTAC™, to define the chemical reaction hazards. The study obtained a working kinetic model for AN polymerization in the presence of HCN, sodium hydroxide, ammonia, and water.
The Accelerating Rate Calorimeter (ARC®) is an instrument that provides adiabatic pressure and temperature vs. time data for reactive systems. The ARC can be used to obtain information about the thermal behavior of reactions and exothermic onset temperatures. The ARC is primarily used for liquid-phase reactive systems. It is also used for the evaluation of explosives and propellants. Exotherm rates as low as 0.02°C/min can be detected. Developed by Townsend and Tou, this instrument (see Figure 1), provides thermokinetic data that is applicable to the design and safety/performance evaluation of reactors and storage vessels.
An Automatic Pressure Tracking Adiabatic Calorimeter (APTAC™) measures adiabatic pressure and temperature versus time response of a reactive system. Due to its unique pressure balancing system and larger sample size, it has a low thermal inertia (~1.10), similar to a full-scale reactor or storage vessel. During an exothermic event, the system will achieve more real-life adiabatic and temperature rise rates, better imitating the actual vessel being evaluated. Data can be directly applied to larger vessels with little scale-up correction.
Figure 1: Sampling of calorimetry test results
Four Accelerating Rate Calorimeter (ARC) tests were conducted to study the thermal polymerization of AN.
The AN polymerization reaction yields a heat of reaction of -17.3 +/- 0.5 kcal/gmol of AN or -326 cal/g or -586 BTU/lb. The polymerization occurs in the vapor phase as well as in the liquid phase. A conservative polymerization heat of reaction of -20 kcal/gmol (-35,977 BTU/lbmol) is predicted by breaking the C= bond to -C- (-377 cal/g or -678 BTU/lb). This is consistent with the measured heat of polymerization value.
PAN begins to decompose at 250°C exothermically (-373 BTU/lb of PAN) to make HCN, AN, NH3, acetonitrile, methacrylonitrile and heavies (14% lights and 86% heavies by weight). The PAN decomposition reaction will yield hydrogen starting at 350°C. A pure AN polymerization model must consider the decomposition reaction.
Figure 2: AN Dimerization ARC test data summary
Figure 3: AN Dimerization ARC test data summary
Figure 4: AN Dimerization ARC test data summary
Figure 5: Model comparisons with pure, inhibited AN ARC data dT/dt
Figure 6: Model comparisons with pure, inhibited AN ARC data dP/dt
Figure 7: Model comparisons with pure, inhibited AN ARC data P vs. T
Figure 8: Model predictions illustrating the impact of thermal inertia on P vs. T behavior in ARC test cell
Figure 9: Impact of sodium hydroxide on acrylonitrile polymerization
Figure 10: Impact of sodium hydroxide on acrylonitrile polymerization
Figure 11: Impact of sodium hydroxide on acrylonitrile polymerization
Five Accelerating Rate Calorimeter (ARC) tests were conducted to study the reactions of the AN/HCN system at different AN/HCN rations. The reaction of AN with HCN forms succinonitrile and is the most energetic reaction yielding -652 BTU/lb succinonitrile formed. If sufficient HCN is present, the succinonitrile formation reaction will jump start the thermal polymerization and decomposition of AN. The excellent predictions of all five tests were obtained using the same model parameters.
Figure 12: Comparison of model predictions with ARC test containing 17.2% HCN dT/dt
Figure 13: Comparison of model predictions with ARC test containing 17.2% HCN dP/dt
Figure 14: Comparison of model predictions with ARC test containing 17.2% HCN P vs. T
Two Accelerating Rate Calorimeter (ARC) tests were conducted to study the impact of caustic on reaction rates. The reaction rates increased in proportion to the weight fraction (ppmw) of sodium hydroxide in the solution. The weight fraction of sodium hydroxide in both tests was 5.0% and 5.1% by weight respectively. The thermally induced polymerization of AN increased by a factor of 50,000. The polymerization reaction of HCN increased by a factor of 2,850. The decomposition of PAN was not affected by the presence of sodium hydroxide.
Figure 15: Comparison of model predictions with test data A90007 dT/dt
Figure 16: Comparison of model predictions with test data A90007 dP/dt
Figure 17: Comparison of model predictions with test data A90007 P vs. T
Ammonia was added to AN/HCN mixtures in concentrations ranging from 0.3 - 0.42% by weight. The formation reaction of succinonitrile and the polymerization reaction of HCN were greatly accelerated by the presence of ammonia. In addition, the activation energies were found to be less than the uncontaminated values.
Figure 18: Model comparison with ARC test B11188, 12% HCN, 0.3% NH3 dT/dt
Figure 19: Model comparison with ARC test B11188, 12% HCN, 0.3% NH3 dP/dt
Figure 20: Model comparison with ARC test B11188, 12% HCN, 0.3% NH3 dT/dt
Figure 21: Model comparison with ARC test B11188, 12% HCN, 0.3% NH3 dP/dt
Figure 22: Model comparison with ARC test B11188, 12% HCN, 0.3% NH3 P vs. T
Figure 22: Model comparison with ARC test B11188, 12% HCN, 0.3% NH3 T vs. time
The reaction rates increased in proportion to the weight fraction (ppmw) of ammonia in the solution. The weight fraction of ammonia in all seven tests ranged from 0.3 - 0.42% by weight. The reaction of AN with HCN to produce succinonitrile increased by a factor of approximately 3,000. The polymerization reaction of HCN increased by a factor of approximately 3,000 t. The decomposition rate of PAN increased by a factor of 3.5. The thermally initiated AN polymerization reaction increased by a factor of 3.5.
We conducted a total of 21 calorimetry tests, using the ARC® and APTAC™, and isolated the individual acrylonitrile reactions of interest. The thermal polymerization of AN occurs at 190°C to form polyacrylonitrile (PAN). The reaction is energetic and yields -586 BTU/lb of AN reacted. The polymerization occurs in both the liquid and vapor phases. PAN starts to decompose exothermically at 250°C yielding a heat reaction of -373 BTU/lb PAN decomposed. The decomposition reaction starts producing hydrogen at 350°C. AN will react with HCN to form succinonitrile starting at 65°C. This reaction yields -52 BTU/lb succinonitrile produced. HCN will start to polymerize at 65°C to yield an HCN dimer and then an HCN polymer. The reaction yields -494 BTU/lb of HCN polymerized. All reactions, except the decomposition reaction of PAN, are greatly accelerated by the presence of small amounts of ammonia and/or sodium hydroxide.
An example of a crude AN storage tank contaminated with an NH3 solution:
Figure 23: Relief requirements for process induced runaway reaction (ideal flow area)
It is not possible to design an adequate venting system for low design rating AN storage tanks, especially for the following scenarios:
Redundant temperature sensing and alarms triggering sufficient cooling to maintain temperature during runaway at or below 50°C.
The preferred cooling route is reflux condensing.
Cooling requirement should be designed for succinonitrile reaction (652 BTU/lb) at a rate of 2.2 BTU/lb stored/min at or below 50°C.
Consider installing fixed water sprays to protect the tanks from external fire exposure.
Focus on aggressive prevention of contamination with acid or base.
The presence of sodium hydroxide and ammonia will greatly accelerate the reaction of AN with HCN, AN polymerization, and HCN polymerization. even small amounts of sodium hydroxide and ammonia present a significant safety hazard.
Here we have presented a working model, based on 21 calorimetry data sets, that can be used to assess the effectiveness of existing relief systems in case of contamination and/or fire exposure to vessels containing mixtures of AN/HCN.
Safe design and operation of chemical processes requires a thorough understanding of chemical reactivity and thermokinetics. ioKinetic offers a 3 step approach, applicable for relief system design, process optimization, and chemical reactivity hazard evaluation. Request your free quote today or call us at 1-844-ioKinetic.
Sign Up Now