Thermal Accident of Methyl Ethyl Ketone Peroxide Plant: Calorimetric Analysis

Overview About Methyl Ethyl Ketone Peroxide (MEKP) in Chemical Plants

Methyl Ethyl Ketone Peroxide (MEKP) is used as a hardener in the manufacture of resins, synthetic rubber and other petrochemical plastics. It is an ingredient of paints, varnishes and paint removers. MEKP is also used in the fiber glass and plastics industry as a curing agent. It is an organic peroxide, which is explosive in its pure form. Hence it is commercially available as a 40-60% solution with stabilizing agents such as dimethyl phthalate, cyclohexane peroxide, or diallyl phthalate (Gooch, 2011; Barbalace, 2009).

MEKP is listed as a highly toxic substance and is categorized into United Nations Hazard class 5.2 (Eller and Cassinelli, 1994). Its colorless nature and minimal odor has led to accidental ingestion among both adults and children (Prez-Martinez et al. 1997, Bates et al. 2001). Several cases of intentional ingestion for self-harm or suicide have been reported (Mittleman et al. 1986). In addition, poisoning by inhalation and spillage to eyes leading to corrosive damage has also been reported. Unprotected workers are victims of chronic exposure and toxicity (Brigham and Landrigan, 1985).

Commercially available preparations of MEKP are strong oxidizing substances. They are known to produce alkylperoxyl radicals upon contact with metal ions, a process accelerated by the presence of iron in the heme molecule in biological systems. Tissue damage is believed to be caused by these free radicals, which denature organic molecules, including the peroxidation of lipids. Further toxicity is caused by the acidity of the chemicals produced (Akaike et al. 1992).

The behavior of thermal explosions or runaway reactions has been widely studied for many years. In a reactor with an exothermic reaction, it is very easy to accumulate energy and temperature, when the heat generation rate exceeds the heat removal rate by Semenov theory (Semenov, 1984). Methyl ethyl ketone peroxide (MEKPO) is usually applied as initiators and cross-linking agents for polymerization reactions.

One reason for accidents involves the peroxy group (-0-0-) of organic peroxides, due to its instability and high sensitivity for thermal sources. Many thermal explosions and runaway reactions have been caused globally by MEKPO resulting in a large number of injuries and even death, as shown in Table 6.1 (Yeh et al„ 2003; Chang et al., 2006; Tseng et al„ 2006; Tseng et al„ 2007; MHIDAS, 2006).

Table 6.1 displays three accidents for MEKPO in Australia and UK from the Major Hazard Incident Data Service (MHIDAS)

Case Selection

The case study in this chapter deals with a thermal explosion and runaway reaction of MEKPO occurred at Taoyuan County (the so-called Yung-Hsin explosion) that killed 10 and injured 47 people in Taiwan in 1996. Figures 6.1 (a) and (b) show the accident damage situation from the Institute of Occupational Safety and Health in Taiwan.

Accident development was investigated by a report from the High Court in Taiwan. Unsafe actions (wrong dosing, dosing too rapidly, errors in operation, cooling failure) caused an exothermic reaction and the first thermal explosion of MEKPO. Simultaneously, a great deal of explosion pressure led to the top of the tank bursting and the hot concrete broken and shot to the 10-ton hydrogen peroxide (H202) storage tank (d = 1, h = 3). Under this circumstance, the 10-ton H20, storage tank incurred the second explosion and conflagration that caused ten people to perish (including employers, fire-fighters) and 47 injuries. Many plants near Yung-Hsin Co. were also affected by the conflagration caused by the H,02 tank. H202, dimethyl phthalate (DMP), and methyl ethyl ketone (MEK) were applied to manufacture the MEKPO product. To prevent any casualties from runaway reaction and thermal explosion events from occurring, the aim of this study was to simulate an emergency response

TABLE 6.1

Thermal Explosion Accidents Caused by MEKPO Globally

Year

Nation

Frequency

Injuries

Fatalities

Worst Case

1953-1978

Japan

14

115

23

  • 114 (Injuries)
  • 19 (Fatalities) in Tokyo

1980-2004

China

14

13

14

  • 8 (Injuries)
  • 5 (Fatalities) in Honan

1984-2001

Taiwan

5

156

55

  • 49 (Injuries)
  • 33 (Fatalities) in Taipei

2000

Korea

1

11

3

  • 11 (Injuries)
  • 3 (Fatalities) in Yosu

1973-1986

Australia

2

0

0

Not applicable

1962

UK

1

0

0

Not applicable

(Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications)

(a) The aftermath of the Yung-Hsin explosion which devastated the entire

FIGURE 6.1 (a) The aftermath of the Yung-Hsin explosion which devastated the entire

plant, including all buildings within 100 m (b) Reactor bursts occurred by thermal runaway (Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications).

process. Differential scanning calorimetry (DSC) was employed to integrate thermal hazard development. The processing of experimental data and kinetics evaluation was implemented by applying the TDPro and For К software developed by CISP Ltd. The production method is described in detail by Kossoy and Akhmetshin (Yuan, Shu & Kossoy, 2005) for the creation of a kinetic model and the algorithms that are utilized. Due to MEKPO decomposing at low' temperature (30^Ю°С) (Yuan, Shu & Kossoy, 2005; Fu et al., 2003) and exploding with exponential development, developing or creating an adequate emergency response procedure is very important for preventing it. The safety parameters, such as temperature of no return (TNR), time to maximum rate (TMR), self-accelerating decomposition temperature (SADT), maximum temperature (Tmax), etc., were necessary and useful for studying emergency response procedure in terms of industrial applications. In view of loss prevention, the emergency response plan is mandatory and necessary for corporations to cope with reactive chemicals under upset scenarios.

Thermal Hazard Analysis of MEKP by DSC

DSC has been employed widely for evaluating thermal hazards (ASTM E537-76, 1976; Ando, Fujimoto & Morisaki, 1991) in various industries. It is easy to operate, gives quantitative results, and provides information on sensitivity (exothermic onset temperature, T0) and severity (heat of decomposition, AHd) at the same time. DSC was applied to detect the fundamental exothermic behavior of 31 mass % MEKPO in DMP that was purchased directly from the Fluka Co. Density was measured and provided directly from the Fluka Co. ca. 1.025 g cm-'. It was, in turn, stored in a refrigerator at 4 °C. DSC, as shown in Figures 6.2(a) and (b), involved two thermocouples, gold of test crucible (100 bar), STAR0 software, and so on. According to Figure 6.2(b), the S side put on sample crucible and the R side detects the blank crucible.

(a) Measurement principle of a heat flux DSC sensor with a single thermocouple

FIGURE 6.2 (a) Measurement principle of a heat flux DSC sensor with a single thermocouple (b) Measurement principle of the FRS5 interchangeable heat flux DSC sensor with 56 thermocoup lest Reproduced from Wu et al. 2008. International Journal of Chemical Science, Sadguru Publications).

Thermal Analysis of MEKP and H 2 O 2 Through DSC Analysis

Thermal Decomposition Analysis of 31 Mass% MEKPO for DSC

Figure 6.3 demonstrates a comparison of thermal curves of decomposition of 31 mass % MEKPO with four types of H (H = 1, 2, 4, and 10°C min ') by DSC. Table

6.2 summarizes the thermodynamic data by the DSC STARC program for the runaway assessment. MEKPO could decompose slowly at 30-32°C, as disclosed by previous researches (Chen et al., 2006). We surveyed MEKPO decomposing at 30°C, shown in Figure 6.3. Various scanning rates by DSC were used to survey the initial decomposition circumstances. Under the scanning rate of 10°C min-1 situation, the T() was measured at about 47°C and AH of the first peak was evaluated at about 96.87 J g1. As a result, a rapid rise of temperature may cause violent initial decomposition

Heat flow vs. temperature of MEKPO 31 mass% under dynamic various scanning rates by DSC (Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications)

FIGURE 6.3 Heat flow vs. temperature of MEKPO 31 mass% under dynamic various scanning rates by DSC (Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications).

(the first peak) of MEKPO under external fire conditions. Table 6.2 shows thermok- inetic parameters, such as T0, ДН, Tmax, of 31 mass% MEKPO by DSC under various scanning rates. The initial decomposition peak usually releases little thermal energy, so it is often disregarded. The T, of mainly decomposition was about 80°C. The total heat of reaction (ДН1оЫ) was about 1,200 J g1. DSC was applied to evaluate the Ea and frequency factor (A). The Ea under DSC dynamic test was about 168 kJ mol1 and A was about 3.5.1019 (s-1).

Thermal Decomposition Analysis of 20 Mass% H 2 O 2 by DSC

Figure 6.4 displays the exothermic reaction of 20 mass % H,02 under 4°C min-1 of (3 by the DSC. Due to H202 being a highly reactive chemical, operators must carefully control flow and temperature. H,02 was controlled at 10°C, when it precipitated the reaction. H202 exothermic decomposition hazards are shown in Table 6.3.

Kinetic Analysis of Thermal Degradation

There are several well-known methods for evaluating simple autocatalytic models, i.e., for estimating the model parameters. One can derive complex multi-stage kinetic models that depict autocatalytic phenomena in more detail, but special numerical optimization methods are required to estimate parameters of such models as discussed by Kossoy and Koludarova (Kossoy & Koludarova, 1995) for a complex

TABLE 6.2

Thermokinetics and Safety Parameters of 31 Mass% MEKPO by DSC Under Various Scanning Rates

Р

(°C min ')

Mass

(mg)

Initial Decomposition

Main Thermal Decomposition

1sl peak

T,

(°C)

0 8'1)

2nd peak

T2

(°C)

  • 4 may
  • (°C)

AHd

(lg-1)

  • 3rd peak T2
  • (oC)

Tm,« (°C)

AHd

(J g1)

AHloUi 0 g-i)

1

4

30

29.16

70

115

382.76

142

180

825.29

1237.51

2

3.72

35

35.85

75

125

324.11

152

187

768.95

1128.91

4

4

42

41.35

83

135

304.05

160

200

768.13

1113.53

10

4.9

47

96.87

100

140

250.82

175

220

584

931.69

(Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications)

Heat flow vs. temperature of 20 mass% H,0, under 4°C mitr scanning rate by DSC (Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications)

FIGURE 6.4 Heat flow vs. temperature of 20 mass% H,0, under 4°C mitr1 scanning rate by DSC (Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications).

TABLE 6.3

Thermokinetics and Safety Parameters of 20 Mass% H202 by DSC Under Various Scanning Rates

Chemical

Mass (mg)

T„ (°C)

Tm„ (°C)

AHd

0 g-'>

ha

2.47

67

105

395

Reference1

2.2

69

99.6

409.7

(Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications)

model of two consecutive reactions where the second stage is autocatalytic. The reaction mechanism of MEKPO could be represented by the following kinetic model:

According to Equations (6.1) and (6.2), it might be noted that two reaction mechanisms without initial catalyst (B) exist in the incipient stage, indicating that the reaction rate was proportional not only to the reactant concentration, but also to the product concentration. Figure 6.5 applies the TDPro and ForK to simulate the TMR versus temperature. We used 31 mass % MEKPO of DSC experimental data (scanning rate is 4°C min1) to fit the thermal curve and to simulate the TMR model. Under 100°C situation, the TMR of MEKPO was calculated as about three minutes.

Safety Parameter Evaluation

We used DSC parameters, including Ea, A. and ДН. to set up the Semenov equation. SADT for a 25 kg package was calculated by using a wetted surface area (S), S = 0.48124 m2, and heat transfer coefficient (U), U = 1.7034.10 * (kJ nv2 °C_I min '). SADT for a 0.51 Dewar vessel was evaluated as S = 0.0303 m2, and U = 8.7402.10 2 (kJ nr2 °C' min1). The SADT for a 5 and 55 gallon drum was evaluated as S = 0.137 m2, S = 1.51 m2, respectively, and U = 11.34 (J nr2 °C I). The SADT of various vessels was determined as demonstrated in Table 6.4.

TMR vs. temperature (kinetic-based curve-fitting) for 31 mass% MEKPO by DSC for 4°C min  of heating rate

FIGURE 6.5 TMR vs. temperature (kinetic-based curve-fitting) for 31 mass% MEKPO by DSC for 4°C min 1 of heating rate (Reproduced from Wu et al. 2008. International Journal of Chemical Science, Sadguru Publications).

TABLE 6.4

TNR and SADT for Safety Storage and Transportation by Various Vessel Situations

Vessel Type

TNR (°C)

SADT ("O

5 gallon drum

78

72

55 gallon drum

97.9

84.5

UN 25 kg package

62.7

54.2

UN 0.51 L Dessel vessel

55

50

Adiabatic test by VSP 2

100

80

(Reproduced from Wu et al. 2008, International Journal of Chemical Science, Sadguru Publications)

Summary

According to the DSC experimental data, MEKPO decomposes at 30-40°C. If the H is high, the initial exothermic temperature could be delayed and ДН would cause its temperature to rise quickly. Under external fire circumstances. MEKPO can decompose quickly and cause a runaway reaction and thermal explosion. During storage and transportation, a low concentration (< 40 mass %) and a small amount of MEKPO should be controlled. Under differential storage and transportation vessels, for the SADT there was a disparity. This chapter, with a view of predicting the SADT of a 55 gallon drum in Taiwan, came up with a value of about 85°C. H,02 was controlled 10°C, when it joined a MEKPO manufacturing reaction. This is very dangerous for the MEKPO manufacturing process, so the reaction was a concern and controlled at less than 20°C in the reactor. Thermokinetics determined by an autocatalytic thermal curve could be used to assess the thermal explosion hazard for organic peroxides and to determine useful parameters such as T0, SADT, temperature of no return (TNR), and adiabatic time to maximum rate (TMRad). In practice, these data are necessary for the proper choice of safe conditions for application, storage, and transportation in terms of chemical products.

References

Akaike, T., Sato, K., Ijiri, S.. Miyamoto, Y.. Kohno, M.. & Ando, M. (1992). Bactericidal activity of alkyl peroxyl radicals generated by heme-iron-catalyzed decomposition of organic peroxides. Arcli. Biocltem. Biophy. 294 (1) 55-63.

Ando. T., Fujimoto. Y., & Morisaki, S. (1991). Analysis of differential scanning calorimetric data for reactive chemicals. J. Hazard Mat. 28 (3), 251-280.

ASTME537-76. (1976). Thermal Stability of Chemicals by Methods of Differential Thermal Analysis.

Barbalace, K. (2009). Chemical Database: Methyl ethyl ketone peroxide. Environ. Client. Com.

Bates. N.. Driver, C.P.. & Bianchi. A. (2001). Methyl ethyl ketone peroxide ingestion: toxicity and outcome in a 6-year-old child. Peel. 108 (2) 473-476.

Brigham. C.R.. & Landrigan. P.J. (1985). Safety and health in boatbuilding and repair. Am. J. hid. Med. 8(3)469-182.

Chang, R. H.. Tseng. J. M.. Jehng, J. M.. Shu. С. M. & Hou. H. Y. (2006). Thermokinetic model simulations for methyl ethvl ketone peroxide contaminated with 4 OR NaOH by DSC and VSP.7. Therm. Anal. Calorim. 83, 57-62.

Chen, K. Y., Lin. С. M., Shu С. M., & Kao. C. S. (2006). An evaluation on thermokinetic parameters for hydrogen peroxide at various concentrations by DSC. J. Therm. Anal. Calorim. 85, 87-89.

Eller, P.M.. & Cassinelli, M.E. (1994). NIOSH manual of analytical methods. DIANE Publishing, Pennsylvania. United States.

Fu, Z. M., Li, X. R., Koseki. H. К.. & Mok. Y. S. 2003. Evaluation on thermal hazard of methyl ethyl ketone peroxide by using adiabatic method. J. Loss Prev. Process Ind. 16 (5), 389-393.

Gooch, J.W. (2011). Methyl Ethyl Ketone Peroxide. Encyclopedic Dictionary of Polymers. 458.

Kossoy, A. A., & Koludarova, E. (1995). Specific features of kinetics evaluation in calorimetric studies of runaway reactions. J. Loss Prev. Process Ind. 8. 229-235.

Maria, G.. & Heinzle, E. (1998). Kinetic system identification by using short-cut techniques in early safety assessment of chemical processes. J. Loss Prev. Process Ind. 11(3), 187-206.

MHIDAS. Mayor Hazard Incident Data Service. (2006) OHS_ROM. Reference Manual.

Mittleman, R.E., Romig, L.A., & Gressmann, E. (1986). Suicide by ingestion of methyl ethyl ketone peroxide. 7. For Sci. 31(1) 312-320.

Prez-Martinez. A.. Gutirrez-Junquera, C., Gonzlvez-Piera, J.. Marco-Macin. A., Rubio- Guijarro. J., & Moya-Marchante. M. (1997). Oesophageal stenosis in a child caused by ingestion of methyl ethyl ketone peroxide. Eur. J. ofPed. 156 (12) 976.

STAR1' Software with Solaris Operating System. Operating Instructions. (2004). Mettler Toledo. Switzerland.

Tseng, J. M., Chang. R. H., Horng, J. J.. Chang. M. K., & Shu, С. M. (2006). Thermal hazard evaluation for methyl ethyl ketone peroxide mixed with inorganic acids. J. Therm. Anal. Calorim. 85. 189-194.

Tseng. J. M., Chang. Y. Y., Su, T. S., & Shu, С. M. (2007). Study of thermal decomposition of methyl ethyl ketone peroxide using DSC and simulation. J. Hazard. Mater., 142 (3), 765-770.

Wu, S.H., Su. С. H., & Shu. С. M. (2008). Thermal accident investigation of methyl ethyl ketone peroxide by calorimetric technique. Int. J. Chem. Sci. 6(2), 487-496.

Yeh. P. Y., Shu. С. M.. & Duh, Y. S. (2003). Thermal hazard analysis of methyl ethyl ketone peroxide. Ind. Eng. Chem. Res., 43 (1), 1-5.

Yuan, M. H.. Shu. С. H., & Kossoy. A. A. 2005. Kinetics and hazards of thermal decomposition of methyl ethyl ketone peroxide by DSC. Thermochimica Acta, 430. 67-71.

 
Source
< Prev   CONTENTS   Source   Next >