DOI: 10.3303/CET23104010 Paper Received: 21 May 2023; Revised: 11 June 2023; Accepted: 9 August 2023 Please cite this article as: Serrano J., Nowicki A., Perrin L., Dufaud O., 2023, Effect of Particle Size Distribution and Inerting Mechanism on Explosion Severity of Organic/mineral Mixtures, Chemical Engineering Transactions, 104, 55-60 DOI:10.3303/CET23104010 CHEMICAL ENGINEERING TRANSACTIONS VOL. 104, 2023 A publication of The Italian Association of Chemical Engineering Online at www.cetjournal.it Guest Editors: Sabrina Copelli, Ernesto Salzano Copyright © 2023, AIDIC Servizi S.r.l. ISBN 979-12-81206-06-9; ISSN 2283-9216 Effect of Particle Size Distribution and Inerting Mechanism on Explosion Severity of Organic/Mineral Mixtures José Serranoa,b, Alexandre Nowickib, Laurent Perrina, Olivier Dufauda,* a Université de Lorraine, CNRS, LRGP, F-54000 Nancy, France bTecaliman, 44323 Nantes, France olivier.dufaud@univ-lorraine.fr The animal feed industry mixes cereals, vitamins, amino acids, and mineral powders to produce a solid mixture called “Premix”. The mitigation of premix explosions is challenging due to the diversity of composition, particle size, and nature of the mixed products. Therefore, determining their explosion safety parameters requires many standardised tests and a time-consuming process. However, it is possible to reduce the extensive use of experimental characterisation by better understanding the physicochemical mechanisms involved. In this context, this project aims to study the influence of Particle Size Distribution (PSD) and the mineral chemical nature on the explosion severity of organic and mineral powder mixtures commonly used for premix manufacturing. Cornflour was mixed with four minerals (sodium chloride, sodium bicarbonate, calcium carbonate, and magnesium oxide) chosen based on their industrial applications and inerting mechanisms (scavenging of radicals, inert gas generation, and heat sink). The powders were sieved to obtain samples with distinct particle size ranges. PSD was analysed ex-situ and in-situ to study the fragmentation behaviour of the products. The explosion tests method was based on the standard ISO/IEC 80079-20-2 using the 20L sphere. The results indicated that due to physical and chemical effects, NaHCO3 is the most efficient inerting agent. Moreover, its initial PSD did not affect the inhibition performance due to its brittleness and the explosibility test pressure gradient, leading to possible inerting overestimation. NaCl reduced the deflagration index (KSt) less efficiently due to the incomplete decomposition into scavenging agents of free radicals, essential for flame propagation. The unsuitable addition of purely thermal inhibitors (CaCO3, MgO) could increase the mixture’s KSt due to a dispersibility improvement, dust cloud PSD reduction and radiation effects. The mineral nature selection during product design could then significantly impact the inherent safety in the premix industry. 1. Introduction A “premix” is a compound feed not intended for feeding animals directly. Premix manufacturers blend cereals, vitamins, amino acids and mineral products to cover nutritional needs by managing complex formulations. Therefore, applying process safety measures is challenging because of the mixed powders’ multiple physical and chemical properties. Nevertheless, it is possible to assume, in a first approach, that the frequent incorporation of inert materials allows the inhibition of dust explosion by applying the moderation principle of inherent safety (Amyotte et al., 2007). However, the inhibition efficiency depends on many variables, such as the fuel composition and non-combustible nature. For instance, Reding & Shiflett (2019) characterised the inhibition efficiency of five suppressant agents mixed with organic and metallic fuels using thermal analysis and showed that similar decomposition temperature ranges for the fuel and inert increase the inhibition effect of physical and chemical mechanisms. Y. Liu et al. (2022) proposed a NaHCO3 inerting mechanism and highlighted the influence of sodium-containing species that scavenge the free radicals in oil shale explosions. However, Chen et al. (2022) compared its inerting effects with the corresponding solid product (Na2CO3) and discovered that the physical effects are predominant when the inert material proportion increases. Similarly, J. Liu et al. (2022) investigated the efficiency of adding CaCO3 to inhibit titanium explosions, in which its thermal stability and purely thermal absorption action reduced the inerting capacity. However, few studies have aimed at inerting organic combustible products. One of them was conducted by Yang et al. (2022), in which they compared the inerting efficiency of NaCl and NaHCO3 on flour explosion at different mineral concentrations. 55 On the other hand, it is well known that the dust cloud PSD also modifies the explosion performance. Jiang et al. (2018) determined the Minimum Inerting Concentration (MIC) of aluminium/NaHCO3 mixtures with different PSD. The authors found that the MIC was nearly independent of the NaHCO3 PSD due to the ratio between the reaction time of the aluminium (d50 = 5 µm) according to its combustion regime (Serrano et al., 2021) and NaHCO3 heating time. Moreover, Bu et al. (2021) researched the suppressant-enhanced explosion phenomenon in aluminium dust flame propagation by adding alumina and improving the dispersibility of the dust cloud by reducing the effective dust cloud PSD, as reported by Bagaria et al. (2019). Nevertheless, this research area has focused mainly on metallic powders, lower PSD (<100 µm) and not considered the combined effects of more than two variables, which reduce their application to protection and mitigation safety measures in the premix industry. Therefore, this study addresses the influence of the inert mechanism and organic/mineral products’ PSD on the explosibility of simplified mixtures commonly used in premix manufacturing to contribute to applying the inherent safety approach and suitable protection measures. 2. Materials and methods The mineral powders (NaCl, NaHCO3, CaCO3 and MgO) were selected according to their relevance in the premix’s formulation and their different main inerting mechanisms (scavenging of radicals, CO2 generation and heat sink). Cornflour was chosen as the organic fraction of the binary mixtures due to its wide industrial use and suitable explosion parameters. The PSD of the samples was modified using a vibrating sieve shaker according to relevant industrial PSD ranges. It was measured ex-situ using the Mastersizer 3000 Particle Size Analyser (Malvern Instruments) equipped with the Aero S dry dispersion unit. Table 1 shows the characteristic diameters of the powder samples (PSD0), including the Sauter mean diameter (d3,2) Table 1: Characteristic diameters and temperatures of the powders before dispersion in the 20L sphere Powder d10 [µm] d50 [µm] d90 [µm] d3,2 [µm] Temperature [°C] NaCl 196 280 394 269 800 – 970 (melting) CaCO3 27 241 350 36 700 – 800 (decomposition) NaHCO3-#1 9 30 60 17 NaHCO3-#2 193 291 434 278 120 – 190 (decomposition) NaHCO3-#3 299 422 589 407 MgO-#1 10 50 126 14 MgO-#2 167 252 373 239 2850 (melting) MgO-#3 462 638 871 621 Cornflour-#1 18 65 134 42 Cornflour-#2 165 240 347 229 300 (decomposition) Cornflour-#3 461 635 865 615 2.1 Particle size distribution in-situ measurement The dust cloud generation based on ISO/IEC 80079-20-2 requires a high-pressure injection, leading to significant differences between the samples’ PSD before and after dispersion due to particle fragmentation. The dust cloud PSD (PSDd) was measured in situ using a HELOS/KR (Sympatec GmbH) laser diffraction sensor equipped with R3 and R5 lenses capable of measuring particle sizes between 0.5 – 175 and 4.5 – 875 µm, respectively. The dispersion chamber is similar to the 20L sphere with visualisation windows that promote optical measurements and allow the granulometric analysis under similar conditions to the standard method. The PSDd was assessed each 5 ms and averaged around the standard ignition delay time (60±10 ms). 2.2 Explosion severity According to ISO/IEC 80079-20-2, a standard 20L spherical vessel equipped with a rebound nozzle was used for the explosion tests. The dust container pressure was slightly adjusted for large-volume samples to consider the air volume reduction and to pursue the 1 bar pressure at ignition. The binary mixtures Pmax and (dP/dt)max were obtained between (1000 – 1200 g/m3) based on the Pmax and (dP/dt)max of the pure cornflour samples. The samples’ mixing process was done in a 3D mixer (Turbula), allowing a high mixing efficiency and reproducibility. 3. Results and discussion Firstly, the influence of each mineral product and its inerting mechanism will be discussed, followed by the effect of the mineral particle size on the inerting efficiency. Finally, the combined effects of fuel concentration and particle sizes of organic and mineral products over binary mixture explosibility will be introduced. 56 3.1 Inert mechanisms efficiencies The cornflour sample (Cornflour-#1) was mixed at three mass proportions (40%, 60%, 80%) with four mineral samples (NaCl, CaCO3, NaHCO3-#2, MgO-#2) with comparable PSD to study the influence of the mineral nature on the maximum explosion overpressure Pmax, (Figure 1a) and maximum rate of pressure rise (dP/dt)max, (Figure 1b). The experimental data was fitted to a polynomial regression and projected by a dotted line until the pure Cornflour-#1 values. The Pmax of the mixtures began to decrease significantly after adding 60% of mineral powder. The mixtures with NaCl, CaCO3 and MgO-#2 led to a similar quasi-linear Pmax drop, as shown in Figure 1a, demonstrating that the mineral concentration to inhibit the explosion is much higher than 60%. Moreover, their inerting effect had a similar thermodynamic influence on the explosion performance considering their heat capacities (864, 834 and 918 J/kg*K, respectively) and thermal stability (Table 1) (J. Liu et al., 2022; Yang et al., 2022). On the other hand, the Pmax of mixtures with NaHCO3-#2 significantly decreased after adding more than 40%, in agreement with (Y. Liu et al., 2022). The 60% NaHCO3-#2 mixture achieved 60% of the Pmax obtained with the other mineral products, as seen in Figure 1a. This behaviour was consistent with the low decomposition temperature (Tdecomp) of NaHCO3 (Table 1) (Chen et al., 2022), which strongly suggests that the particles decomposed completely through an endothermic process and efficiently removed the heat from the flame front required for the cornflour combustion. Figure 1: Evolution of explosibility data of cornflour #1 and mineral compounds mixtures at different organic mass concentrations. (a) Maximum explosion overpressure, (b) Maximum explosion pressure rise. In addition, the Cornflour-#1 mixture needed at least the addition of 20% of NaHCO3-#2, 40% of NaCl and more than 60% of CaCO3 or MgO-#2 to significantly decreased its (dP/dt)max, as seen in Figure 1b. The inerting effect by adding NaCl and NaHCO3-#2 followed a quasi-linear trend over the concentration range studied. On the other hand, the mixtures with CaCO3 and MgO-#2 achieved an unusual non-linear behaviour as the organic composition decreased and even a (dP/dt)max promoting effect by adding 20% of mineral. The improved dispersibility of the cornflour caused this effect due to the disrupted inter-particle contacts and decreased agglomeration trend induced by the mineral particles over their inerting mechanisms, which led to a PSDd reduction (Bu et al., 2021). In addition, the MgO and CaCO3 particles remaining at the flame front (delayed thermal degradation – Table 1) might boost the heat radiation transfer afterwards. Furthermore, the other mixtures did not evidence that behaviour due to a complete thermal decomposition (NaHCO3) and free radicals scavenging mechanism (NaHCO3/NaCl) ahead from the flame front. Similarly, the particular inerting mechanism of the mineral products can be seen more clearly in Figure 2 for each binary mixture with an organic composition of 40% and the corresponding Cornflour-#1 curve. As previously discussed, the MgO and CaCO3 samples led to a similar Pmax evolution, demonstrating that the thermal sink corresponds to their primary inerting mechanism. Moreover, the CaCO3 mixture reached a slightly lower Pmax caused by its delayed thermal decomposition (Table 1), in which a small fraction is converted into CO2 (J. Liu et al., 2022). Nevertheless, the KSt parameter of these mixtures was not significantly different from the pure organic product, illustrating that the application of mitigation safety measures should be equivalent to them. Furthermore, the characteristic temperatures of CaCO3 and NaCl being close, the inerting method of NaCl is different. Indeed, NaCl acts mainly through a chemical effect caused by the Na+ and Cl- 57 ions generated from its thermal decomposition, which “flatten” the overpressure curve, as shown in Figure 2. However, the mixture achieved a comparable Pmax to the previous mineral samples, suggesting that thermal absorption was the primary explosion moderation (Reding & Shiflett, 2019) because of a lack of Na+ radicals to inhibit the explosion considering its Tdecomp (Table 1) (Yang et al., 2022). Moreover, the inerting effect of the NaHCO3 sample was the most effective because of its combined effect of the endothermic thermal decomposition at low temperature (Table 1), O2 dilution and the generation of sodium-containing reactive products within the flame front, which can scavenge the radical species of O- and OH- and modify the propagation phenomenon (Yang et al., 2022). Then, it should be stressed that the physical effects of NaHCO3 play a significant role from the early stages of the explosion and correspond to the main difference with the similar chemical inhibition mechanism of NaCl (Chen et al., 2022). Figure 2: Pressure evolution of 40% Cornflour-#1 and minerals compounds mixtures. 3.2 Particle Size Distribution (PSD) effect on the inerting efficiency The mineral product to study the influence of the PSD0 on the explosion inhibition was NaHCO3 due to its inerting efficiency. The mixtures comprised Cornflour-#1 and two NaHCO3 samples (Table 2) with different PSD0 (Table 1). Both mixtures’ explosibility data ((dP/dt)max and Pmax) were identical, as seen in the table added to Figure 3. Thus, NaHCO3 PSD0 did not affect the inerting efficiency of cornflour explosion, suggesting that the process safety mitigation measures could be applied regardless of their PSD0. Nevertheless, Amyotte et al. (2007) found that the mineral PSD affects inhibition performance. Therefore, the PSD of the mixtures was measured in situ after dispersion (PSDd) to highlight a potential PSD modification due to powder fragmentation (Figure 3). Table 2: Binary mixtures of cornflour with different PSD0 of NaHCO3 Binary mixture Organic sample Organic [%] Mineral sample Mixture#1 Cornflour-#1 60 NaHCO3-#3 Mixture#2 NaHCO3-#1 When considering the PSD0 of Mixture#1, both components could be clearly distinguished due to their distinctive particle size range (Table 1): particles with diameters lower than d60 characterise Cornflour#1 sample, whereas the coarsest particles correspond to the NaHCO3-#3 sample (Table 2). Such analysis could not be applied to Mixture#2 because of the similarities between each component (Table 1), setting up a unimodal distribution with considerably smaller particle size. However, Figure 3 shows that the PSDd of both mixtures were indistinguishable, suggesting an intense particle fragmentation with the same threshold regardless of the NaHCO3 PSD0. In this scenario, the high brittleness of NaHCO3 caused its fragmentation under the energetic pressure gradient and shear stress during an explosion severity test (Bagaria et al., 2019). Therefore, these NaHCO3 samples underwent the same complete thermal decomposition and gas-phase chemical interaction with the flame (Chen et al., 2022). Similar results were obtained by Jiang et al. (2018) when testing Al/NaHCO3 mixtures in the 20L sphere, suggesting that although the conditions of the standard 58 explosibility method estimate most traditional scenarios conservatively, it could lead to overestimating the inerting efficiency of a premix with mineral products and thus miscalculating the safety mitigation measures. Figure 3: Particle Size Distribution (PSD) of Cornflour-#1 and NaHCO3 (#1 and #3) mixtures measured ex-situ (PSD0) and in-situ (PSDd). Explosibility data is shown in the table included. 3.3 Combined effect of PSD and organic fraction The study of the combined effects of the PSD0 of both Cornflour and MgO (Table 1) at different fuel proportions (Φ) followed a Box-Behnken experimental design (Figure 4). The mixture explosion was inhibited only with the Cornflour-#3 sample (d50 = 635 µm) and at the lowest fuel concentration (Φ = 40%), corresponding to the expected most favourable scenario for explosion inerting, highlighting the low explosion inhibition efficiency of MgO powder again. Figure 4: Contours of (dP/dt)max evolution of Cornflour/MgO mixtures versus Φ, organic and mineral d50. Furthermore, the fuel proportion and cornflour PSD0 were the most significant factors for the mixture’s explosibility. The reduction of organic PSD0 showed a non-linear effect over the mixtures (dP/dt)max, and its influence increased with the fuel proportion, as illustrated by Figure 4. In addition, the MgO PSD0 did not show a significant effect at low fuel concentration (Φ = 40%) or when the cornflour PSD0 had a d50 ≥ 240 µm (Cornflour-#2 and Cornflour-#3). However, this factor substantially impacted the mixture’s explosibility when the Φ ≥ 60% and the cornflour d50 = 65 µm (Cornflour-#1). In this scenario, MgO particles improved the dispersibility of cornflour more than they inhibited its combustion inhibition, as highlighted in Section 3.1. 59 Therefore, under such conditions, the (dP/dt)max values were even higher than those of the pure cornflour, as presented in Figures 1 and 4, which illustrated the explosion inerting/promotion duality of a mainly “thermal- action” mineral product concerning its mass proportion. However, physical interactions between mineral and organic compounds might follow different trends according to the PSDmineral/PSDorganic ratio (Bu et al., 2021) and effective PSDd (Bagaria et al., 2019). Therefore, according to ISO/IEC 80079-20-2, the cornflour PSD0 was the only parameter significantly influencing the MgO mass required to inert the explosion. 4. Conclusions Estimating the explosion severity and ignition sensitivity of premixtures used in the animal feed industry is challenging due to the variability of the nature, particle size and composition of their components. Using binary versions of those products and modifying selected factors (fuel concentration, PSD0 and inert nature) contributes to a better understanding of the explosion severity of premix compounds. This study established that NaHCO3 is the most efficient inerting agent, with both physical and chemical effects. In addition, the PSD0 of this mineral did not significantly affect the inerting efficiency because of its fragmentation during the explosibility standard test. Therefore, if the standard procedure can be considered a conservative approach when determining pure combustible powders, because of the increase of the particle surface area by fragmentation, it is no longer so when brittle non-combustible powders are added. Indeed, their fragmentation increases the inhibiting action of minerals and is potentially inadequate concerning the industry reality. Furthermore, the inhibition performance of inert materials with a decomposition temperature higher than fuel pyrolysis seems inadequate (NaCl, MgO, CaCO3) due to a scarcity of more efficient inerting agents (Na+, CO2) in the reactive flow. On the other hand, adding minerals without a chemical mechanism (MgO and CaCO3) below 40% could even increase the mixture KSt due to dust cloud dispersibility and radiative transfer improvement. For that reason, the mineral nature selection step of the product design could significantly impact the inherent safety in the premix industry. Further research is aimed at characterising the ignition sensitivity of organic/mineral mixtures. Acknowledgements The authors would like to acknowledge the support of the EMIX Club members. References Amyotte, P. R., Pegg, M. J., Khan, F. I., Nifuku, M., & Yingxin, T. (2007). Moderation of dust explosions. Journal of Loss Prevention in the Process Industries, 20(4–6), 675–687. Bagaria, P., Li, Q., Dastidar, A., & Mashuga, C. (2019). Classification of particle breakage due to dust dispersion. Powder Technology, 342, 204–213. Bu, Y., Amyotte, P., Li, C., Yuan, W., Yuan, C., & Li, G. (2021). Effects of dust dispersibility on the suppressant enhanced explosion parameter (SEEP) in flame propagation of Al dust clouds. Journal of Hazardous Materials, 404, 124119. Chen, X., Lu, K., Xiao, Y., Su, B., Wang, Y., & Zhao, T. (2022). Investigation on the inhibition of aluminium dust explosion by sodium bicarbonate and its solid product sodium carbonate. ACS Omega, 7(1), 617-628. Jiang, H., Bi, M., Gao, W., Gan, B., Zhang, D., & Zhang, Q. (2018). Inhibition of aluminium dust explosion by NaHCO3 with different particle size distributions. Journal of Hazardous Materials, 344, 902–912. Liu, J., Meng, X., Yan, K., Wang, Z., Dai, W., Wang, Z., Li, F., Yang, P., & Liu, Y. (2022). Study on the effect and mechanism of Ca(H2PO4)2 and CaCO3 powders on inhibiting the explosion of titanium powder. Powder Technology, 395, 158–167. Liu, Y., Zhang, Y., Meng, X., Yan, K., Wang, Z., Liu, J., Wang, Z., Yang, P., Dai, W., & Li, F. (2022). Research on flame propagation and explosion overpressure of oil shale dust explosion suppression by NaHCO3. Fuel, 314, 122778. Reding, N. S., & Shiflett, M. B. (2019). Characterisation of thermal stability and heat absorption for suppressant agent/combustible dust mixtures via thermogravimetric analysis/differential scanning calorimetry. Industrial & Engineering Chemistry Research, 58(11), 4674–4687. Serrano, J., Ratkovich, N., Muñoz, F., & Dufaud, O. (2021). Explosion severity behaviour of micro/nano-sized aluminium dust in the 20L sphere: Influence of the particle size distribution (PSD) and nozzle geometry. Process Safety and Environmental Protection, 152, 1–13. Yang, P. P., Meng, X., Zhang, Y., Liu, J. Q., Yan, K., Li, F., Wang, Z., Liu, Y., Dai, W. J., & Wang, Z. (2022). Experimental study and mechanism analysis on the suppression of flour explosion by NaCl and NaHCO3. Combustion Science and Technology, 1–16. 60 8serrano.pdf Effect of Particle Size Distribution and Inerting Mechanism on Explosion Severity of Organic/Mineral Mixtures