Microsoft Word - 1.docx CHEMICAL ENGINEERING TRANSACTIONS VOL. 77, 2019 A publication of The Italian Association of Chemical Engineering Online at www.aidic.it/cet Guest Editors: Genserik Reniers, Bruno Fabiano Copyright © 2019, AIDIC Servizi S.r.l. ISBN 978-88-95608-74-7; ISSN 2283-9216 Comparison of Test Results from Typical Explosive Properties Screenings, such as the Closed Pressure Vessel Test (CPVT) and the Glass Cylinder Deflagration Test Christian Aeby, Delphine Berset, Christian Kubainsky, Mischa Schwaninger* TÜV SÜD Process Safety, Basle, Switzerland mischa.schwaninger@tuev-sued.ch A fast and reliable screening for explosive properties of organic substances is a crucial and ongoing topic for the fine chemicals industry. However, many test systems require large amounts of material, especially for the evaluation of detonative properties, e.g. the BAM 50/60 Steel Tube Test, which may not be available in early stages of development. Therefore, a CPVT with high-resolution pressure measurement (1 kHz) was implemented, based on Adolf Kühner’s mini-autoclave system as described in Whitmore et al. (1999). To evaluate the violence of the decomposition reactions, the CPVT was calibrated according to a procedure as suggested in Knorr et al. (2007), which is based on comparing the maximum pressure rise rate and peak temperature data of standard materials with known explosive properties. Variations of this system are already implemented in some companies and proved its usefulness (Bodman & Chervin, 2004). The use of this type of stability testing of energetic, organic materials was accompanied by parallel testing of the sample materials with several other screening tests, such as the Glass Cylinder Deflagration Test (VDI 2263-1, 1190), the UN Time/Pressure Test (UNDG MTC, 2015) and also by more simple screenings such as dynamic runs in DSC and in the Lütolf Oven (VDI 2263-1, 1990). This work compares and discusses the different test results, i.e. detonation and deflagration behaviour. 1. Introduction In chemical process development there is a strong demand for a fast and reliable screening of explosive properties of energetic organic compounds. The handling of material with explosive properties is not just associated with legal challenges but also technical issues, e.g. processing in larger scale may become inefficient and thus game-changing decisions are required as early as possible in the development phase. However, in such early stages only small amounts of sample material are available for safety testing. Typically, the screening criteria as outline in the Orange book (UNDG MTC, 2015) are applied to exclude explosive properties. These criteria are very conservative leading to numerous false positive results. The UN screening scheme refers mainly to the heat of decomposition as pivotal criterium and is based on thermal stability screening tests such as DSC (Differential Scanning Calorimetry). If the screening criteria indicate potential explosive properties, it is prudent to treat the material as such until proven otherwise. The CPVT and its criteria are intended to reduce the number of false positives. The aim of this project was to compare results from different screenings tests and to evaluate the recently implemented Closed Pressure Vessel Test (CPVT) in regard to experimental effectiveness. The materials used for testing were mainly standard lab samples with suspected explosive properties. Only one reference material (Azodicarbonamide) as mentioned in the Orange book (UNDG MTC, 2015) was integrated in the test program. 2. Screening tests and evaluation schemes The following screening calorimeters and tests were used to characterise the thermal stability and explosive properties of samples. DOI: 10.3303/CET1977031 Paper Received: 18 December 2018; Revised: 30 April 2019; Accepted: 26 June 2019 Please cite this article as: Aeby C., Berset D., Kubainsky C., Schwaninger M., 2019, Comparison of Test Results from Typical Explosive Properties Screenings, such as the Closed Pressure Vessel Test (CPVT) and the Glass Cylinder Deflagration Test, Chemical Engineering Transactions, 77, 181-186 DOI:10.3303/CET1977031 181 2.1 Calorimetric screenings (DSC and Lütolf Oven) The initial test performed was a dynamic DSC measurement up to 400°C using gold plated high pressure crucibles and a heating rate of 4 K/min according to ASTM E 537. For the evaluation the typical threshold values as mentioned in Appendix 6 of the Orange book (UNDG MTC, 2015) were used: (i) exclusion of self-reactive properties (UN Division 4.1) when the exothermic decomposition energy is < 300 J/g, (ii) exclusion of explosive properties (UN Class 1) when the exothermic decomposition energy is < 500 J/g, (iii) exclusion of detonation propagation (UN Class 1) when the exothermic decomposition energy is < 800 J/g. Another thermal stability test performed was the dynamic measurement in the Lütolf Oven using open glass test tubes with 2 g samples and a heating rate of 2.5 K/min according to VDI 2263-1 (Figure 1a). The temperature difference (ΔT) between the sample tube and a graphite reference sample is recorded and plotted against the oven temperature (Figure 1b). Despite the rather old-fashioned test setup, energetic decomposition reactions can be followed even visually, which makes it a useful screening tool especially for deflagration reactions. The criterium used to exclude deflagration behaviour, as determined in the Glass Cylinder Deflagration Test according to VDI 2263-1 (see more detailed description below), is: ΔTmax < 10 % of the left temperature limit of the exotherm peak (TOnset) in °C. This criterium is applied within TÜV SÜD Process Safety and is an alternative to the above listed 300 J/g criterium for DSC measurements. Figure 1: Picture of Lütolf Oven (1a) and cross-section with thermogram (1b) 2.2 Glass Cylinder Deflagration Test The Glass Cylinder Deflagration Test according to VDI 2263-1 consists of a vertical glass tube (40 mm in diameter), filled with 200 ml of sample material, which is ignited by a glowing plug (about 800°C) either at the bottom or on top. The propagation of the decomposition front is monitored visually and with 3 thermocouples immersed in the sample at different heights. If a propagating deflagration front is observed, then the sample is considered capable of deflagration. Since deflagration is temperature and pressure dependent the sample is typically pre-heated prior to testing (to process temperature or 100 °C). Many deflagration incidents observed in the Basel chemical industry were initiated by hot tramp material in dryers, mixers, mills, sieves and screw conveyers. Therefore, a classification system was developed by Fink and Zwahlen (1992), in which the deflagration behaviour is classified from GKD 1 to 3 for increasing violence (Table 1). Based on this deflagration risk class, unit operation specific recommendations regarding use of equipment and layout of safety measures, e.g. extinguishing systems, can be derived. 182 Table 1: Deflagration risk class GKD Risk class GKD 1 GKD 2 GKD 3 Operator (combination of the criteria below) and or Decomposition energy < 1000 J/g > 1500 J/g Combustibility index @ 100°C 6 Onset (left limit) in dyn. Lütolf or DSC > 150 °C Ignition time (glowing plug) > 60 s ≤ 20 s Propagation velocity ≤ 1 cm/min > 1 ≤ 3 cm/min > 3 cm/min Gas release < 10 l/kg > 50 l/kg Reaction violence Only smoke, no ejection of material Material mainly ejected 2.3 Closed Pressure Vessel Test (CPVT) The CPVT consists of a g-scale differential thermal analysis system with mini-autoclaves with a free volume of about 6 ml and is connected to a high-speed pressure transducer (kHz data capture). The autoclaves are filled with 1 g of sample material and heated with 2.5 K/min up to a maximum of 400 °C. Besides pressure the temperature difference (ΔT) between sample and reference (containing graphite) is recorded and plotted against the reference temperature. The system setup is illustrated in Fig. 2 and a detailed description is given in Whitmore et al. (1999). Based on the maximum rate of pressure rise (MPR) and the peak temperature (TP), which is defined as the reference temperature at peak maximum, an explosive rank may be assigned. The risk ranking was suggested by Whitmore et al. (2001) and is shown in Table 2. Figure 2: Picture of CPVT oven with single mini-autoclave (2a) and cross-section of CPVT (2b) The boundaries for classification are shown in Fig. 3 and were based on the results of the three reference materials BPO75 (Dibenzoyl peroxide 75% with water), AIBN (2,2’-Azobis(2-methylpropanenitrile)) and MN (Malononitrile) following the calibration procedure as suggested by Knorr et al. (2007). 2.4 UN Tests From the official UN tests the Time/Pressure Test and in two cases also the Koenen Test were used to compare the previously described screening test results. The UN Tests were performed according to the Orange book, Manual of Tests and Criteria (UNDG MTC, 2015). 183 Table 2: Explosive rank determined by the CPVT Explosive rank Severest property according to UN Class 1 tests Correspondence to UN transport classification A Detonates (related to BAM 50/60 Steel Tube Test) Potentially class 1 B Deflagrates rapidly (related to Time/Pressure Test) and/or gives violent effect upon heating under confinement (related to Koenen Test) Potentially class 1, but not detonable C Deflagrates slowly (related to Time/Pressure Test) and/or medium or low effect of heating under confinement (related to Koenen Test) Not class 1 D Does not deflagrate and shows no effect of heating under confinement No explosive properties with respect to transport classification Figure 1: Explosive rank classification boundaries 3. Results and discussion The test results for each test are listed in table 3a and 3b. The samples are listed in the order of increasing exothermic decomposition energy Q’ determined by DSC or, when not available, by increasing ΔTmax determined in the Lütolf Oven. 10 samples were tested negative and 5 positive in the Glass Cylinder Deflagration Test. Only those with Q’ > 1000 J/g (determined by DSC) led to positive results of the Glass Cylinder Deflagration Test. However, all positively tested samples showed very violent deflagrations, rated as GKD = 3, which explains the observed high threshold value of 1000 J/g, being more than 3 times the internal limit value of 300 J/g. From the 10 negative Glass Cylinder Deflagration Tests all were false positive regarding the Q’ > 300 J/g criterium for DSC and 9 false positive regarding the 10% criterium for the Lütolf Oven. False negative results could occur, because only samples were chosen which were suspected to be capable of deflagration, i.e. Q’ > 300J/g determined by DSC or the 10% criterium determined in the Lütolf Oven (ΔTmax > 10 % TOnset). 184 Table 3a: Test results (Calorimetric screenings) Sample DSC Lütolf Oven TOnset TP Q' TOnset ΔTmax 10% criterium [°C] [°C] [J/g] [°C] [°C] 1 106 241 309 180 19 pos. 2 157 224 332 181 19 pos. 3 181 225 531 140 35 pos. 4 116 224 601 200 92 pos. 5 220 251 617 200 33 pos. 6 112 231 618 180 18 pos. 7 276 328 735 211 132 pos. 8 198 265 758 199 69 pos. 9 192 249 778 200 100 pos. 10 294 367 1039 301 122 pos. 11 99 116 1118 98 53 pos. 12 159 210 1315 142 48 pos. 13 233 302 1359 240 24 pos. 14 75 135 1520 15 99 169 1695 16 215 242 2827 230 84 pos. 17 209 14 neg. 18 184 100 pos. 19 212 157 pos. Table 3b: Test results (Glass Cylinder Deflagration Test, Time/Pressure Test, CPVT and Koenen Test) Sample Glass Cylinder Deflagration Test Time/Pressure Test CPVT Koenen Result GKD Pend t6.9-20.7 bar Result MPR TP Expl. rank [bar] [ms] [bar/s] [°C] 1 neg. <1 D 2 neg. <1 D 3 neg. neg. <1 227 D 4 neg. neg. <1 D 5 neg. neg. <1 D 6 neg. neg. <1 D 7 neg. neg. 8 D 8 neg. 6 D 9 neg. 2 229 D 10 pos. 3 neg. 235 332 D 11 21 fast 2274 91 B 12 pos. 3 >20.7 63 slow 2529 179 B limiting Ø = 1.5 mm 13 neg. neg. <1 D 14 24.3 30 slow 2298 190 B 15 182 slow 1558 141 B 16 pos. 3 89 slow 5828 234 A limiting Ø < 2.0 mm 17 neg. <1 230 D 18 pos. 3 neg. <1 197 D 19 pos. 3 429 slow 1179 198 C As expected, all negative results from the Glass Cylinder Deflagration Test were confirmed by the Time/Pressure Test. However, the 5 positive Glass Cylinder Deflagration Tests corresponded to two slow deflagrations in the Time/Pressure Test and 3 negative Time/Pressure Tests. The reason for this are the different criteria applied. The observation period in the Glass Cylinder Deflagration Test is in minutes (propagation velocity of > 3 cm/min for violent deflagrations), but it is milli-seconds for the Time/Pressure Test (pressure increase between 6.9 and 20.7 bar in < 30 ms), which is linked to the intended application (unit operations in production vs. transport classification). 185 The same holds true for the CPVT. All negative results from the Glass Cylinder Deflagration Test were confirmed by the CPVT. The 5 positive Glass Cylinder Deflagration Tests corresponded to one explosive rank A (detonation), one rank B (fast deflagration), one rank C (slow deflagration) and two rank D (no deflagration) in the Time/Pressure Test. The comparable classification outcome is not surprising, as the CPVT ranking criteria are derived from UN-Tests, incl. the Time/Pressure Test. The CPVT cannot be regarded as conservative for the evaluation of unit operations, as 2 rank D results (no deflagration) were obtained. The extrapolation of the criteria derived from UN tests for transport classification to other applications especially at elevated temperature is not possible. Further investigations would be required as it is not clear to what extent the higher temperatures affect the ranking. Anyway, the existing ranking system already uses temperature dependent boundaries. All negative results from the Time/Pressure Test (9) were confirmed by the CPVT. The 5 slow deflagrating substances, as tested in the Time/Pressure Test, corresponded to one rank A (detonation), three rank B (fast deflagration) and only one correct rank C (slow deflagration) in the CPVT. The one fast deflagrating substance, as tested in the Time/Pressure Test, got a correct rank B. Thus, for the limited amount of tested samples the CPVT always gave conservative results. Only two samples could be tested with the Koenen test. Sample 12 showed a medium effect of heating under confinement (limiting Ø = 1.5 mm) and got a rank B in the CPVT (violent effect of heating under confinement) and sample 17 showed a medium effect or less (limiting Ø < 2.0 mm) and got a rank A (detonation). Though for both samples the CPVT gave conservative results. 4. Conclusions In this test series the CPVT was a fast and reliable screening test, consuming small amounts of sample (2 x 1 g per test). It led to conservative assessments regarding explosive properties for transport classification, i.e. explosives of class 1 and self-reactive substances of division 4.1. Due to the limited number of samples this finding cannot be generalized. Further evaluation, preferably by round robin tests, is required to verify the ranking system and to confirm the conservative nature of the screen. For the assessment of deflagration behavior in unit operations, e.g. drying processes, the CPVT test showed clear limitations, which are related to the actual ranking system. References ASTM E 537, 2017, Standard Test Method for The Thermal Stability of Chemicals by Differential Scanning Calorimetry, ASTM International, West Conshohocken, USA Bodman G.T., Chervin S., 2004, Use of ARC in screening for explosive properties, Journal of Hazardous Materials, 115, 101–105 Fink P.J., Zwahlen G., 1992, Klassierung deflagrationsgefährlicher pulverförmiger Stoffe und daraus folgende Massnahmen in der Betriebspraxis, VDI Berichte Nr. 975, 99-119 Knorr A., Koseki H., Lib X.-R., Tamurac M., Wehrstedt K.D., Whitmore M.W., 2007, A closed pressure vessel test (CPVT) screen for explosive properties of energetic organic compounds, Journal of Loss Prevention in the Process Industries, 20, 1–6. UNDG MTC, 2015, Recommendations on the Transport of Dangerous Goods: Manual of Tests and Criteria, 6 th Ed., United Nations Publications, New York, USA VDI 2263-1, 1990, Dust Fires and Dust Explosions: Test methods for the Determination of Safety Characteristics of Dusts, Beuth Verlag, Berlin, DE. Whitmore M.W., Baker G.P., 1999, Investigation of the use of a closed pressure vessel test for estimating condensed phase explosive properties of organic compounds, Journal of Loss Prevention in the Process Industries, 12, 207–216. Whitmore M.W., Baker G.P., 2001, A closed pressure vessel test screen for condensed-phase explosive properties in organic materials, Journal of Loss Prevention in the Process Industries, 14, 223–227. 186