Available online at www.sciencedirect.com
Journal of Hazardous Materials 153 (2008) 10711077
Effects of acetone on methylperoxide runaway rea
Tsin So-Kw, Nat6400
niversan, ROber 2
Runaway t issurelease durin try (Dand with ace he kifitting. Through the reproducible tests in each condition, the results show that acetone is not a contaminant, because it could increase the activationenergy (Ea) and onset temperature (To) when combined with MEKPO, which differs from the hazard information of the material safety data sheet(MSDS). 2007 Published by Elsevier B.V.
Chemictiators andethyl ketont-hexyl hydon . Ining in varibeen incurunstable chorder to aloperating pevolution itemperaturreported bying ignition
Corresponfax: +886 5 5
0304-3894/$doi:10.1016/junaway reactions; Methyl ethyl ketone peroxide (MEKPO); Differential scanning calorimetry (DSC); Acetone; Kinetic and safety parameters
al industries widely apply organic peroxides as ini-cross-linkers during polymerization, such as methyle peroxide (MEKPO), di-t-butyl peroxide (DTBP),roperoxide, cumene hydroperoxide (CHP), and sopractice, acetone is readily available for clean-
ous applications. Historically, many accidents havered by MEKPO in Asia (Table 1); therefore, itsaracteristics need to be understood thoroughly in
leviate the degree of hazard in manufacturing androcesses. If a reaction is exothermic and the heat
n the system is greater than the heat loss, then thee will increase until all reactants are consumed, as
Zatka . Kotoyori indicated that incidents involv-or explosion of thermally unstable substances may
ding author. Tel.: +886 5 534 2601x4416/4499;31 2069.dress: email@example.com (C.-M. Shu).
occur due to failure of temperature control in process ves-sels . In a systematic study, we have attempted to elucidatethe runaway reaction phenomena of MEKPO combined withacetone.
In this area of thermal analysis, many related calorimetershave been employed to discover the unstable materials. Yu etal. use C80D to demonstrate the hazardous material of asphalt-salt mixtures (ASM) . Stoessel applies DSC to evaluate thethermal stability of nitro-aromatics, and then obtains the safetyparameter of time to maximum rate (TMR) . Miyake etal. use microcalorimetries for evaluating the thermal hazardof self-reactive substances of CHP in chemical processes .The accelerating rate calorimeter (ARC) also has been appliedto calculate the safety parameters reported by Whitmore andWilberforce . By vent sizing package 2 (VSP2), Tseng etal. demonstrate that MEKPO is a sensitive material, especiallywhen mixed with inorganic acids  or contaminants . Othermeasurement methods have been used in this area, includingisoconversion methods for estimating the activation energy .
In this study, we attempted to elucidate the status of a runawayreaction for MEKPO in the presence of acetone. If acetone has
see front matter 2007 Published by Elsevier B.V..jhazmat.2007.09.099Yan-Fu Lin a, Jo-MingTsung-Chih Wu c, Chi-M
a Department of Chemistry, National Chung Hsing University, 250 Kub Doctoral Program, Graduate School of Engineering Science and Technology
123 University Road, Sec. 3, Douliou, Yunlinc Department of Industrial Safety and Health, Hungkuang U
Taichung County 43302, TaiwReceived 20 March 2007; received in revised form 18 Septem
Available online 29 Septemb
reactions by methyl ethyl ketone peroxide (MEKPO) are an importang upset situations. This study employed differential scanning calorimetone 99 mass% on three types of heating rate of 2, 4, and 10 C/min; tethyl ketonection
eng b,hu b,
ang Road, Taichung 40227, Taiwan, ROCional Yunlin University of Science and Technology,2, Taiwan, ROCity, 34 Chung-Chie Road, Shalu,C
007; accepted 18 September 20077
e in Asia, due to its unstable structure and extensive heatSC) to draw the experimental data for MEKPO 31 mass%
netic and safety parameters were then evaluated via curve
1072 Y.-F. Lin et al. / Journal of Hazardous Materials 153 (2008) 10711077
A frequency factor (M1n/s)CpEaHkm
Table 1Selected seveYear Loc
1979 Tai1996 Tai1964 Jap1978 Jap2000 Kor2001 Chi2003 Chi
been mixedcompared w
MEKPOwhich had99 mass%,at 4 C. Afdata wereAfterwardsand safetyarranged an
2.2. Differential scanning calorimetry (DSC)
ed inby tspecific heat capacity (J/g K)apparent activation energy (kJ/mol)heat of reaction (J/g)rate constant (M1n/s)mass of reactant (mg)reaction order (unitless)caloric capacity of measuring cell from exother-mic substances (J/g)gas constant (8.314 J/mol K)temperature (C)time to conversion limit (day)time to maximum rate (h)exothermic onset temperature (C)adiabatic temperature rise (C)autocatalytic constant (unitless)
formedcell thmatelysoftwacompater the10 C/chosencondit
Allplayednot usraisedttersdegree of conversion (unitless)scanning rate (C/min)degree of conversion for second and third stages(unitless)
re thermal explosion accidents caused by MEKPO in East Asia
ation Injuries Fatalities Hazardwan (Taipei) 49 33 Explosion (storage)wan (Taoyuan) 47 10 Explosion (tank)an (Tokyo) 114 19 Explosionan (Kanagawa) 0 0 Explosionea (Yosu) 11 3 Explosionna (Jiangsu) 2 4 Explosionna (Zhejiang) 3 5 Explosion
with MEKPO, the degree of hazard can be decreasedith pure MEKPO.
ard procedure for preparation of MEKPOand acetone 99mass%
31 mass% was purchased directly from Fluka Co.,been stored in a refrigerator at 4 C. Acetone
used as a diluent, was also stored in a refrigeratorter mixing of these two materials, the experimentaldetermined by DSC three times in each condition., curve fitting was employed to model the kineticparameters. Mass ratio of MEKPO/acetone has beend listed in Table 2.
AccordiMEKPO isOO bondpositions. Tby autocataof the methA complexthird stages
= A12nd stageddt
When M99 mass%all reactionture (T0) in55.6 to 61.3also had bepeak; theseature-programmed screening experiments were per-a DSC (Mettler TA8000 system), and a measuringan withstand relatively high pressure to approxi-bar (DSC 821e) was used for the experiment. STAReas used to obtain thermal curves . For precise
of the results of curve fitting and achieving bet-equilibrium, the scanning rate was set at 2, 4, andrespectively . The range of temperature rise was
30 to 300 C during the excursion of the test in each
he experimental and curve fitting results are dis-Tables 2, 3 and 4 (these mixture conditions were
our previous studies). Aside from the informationhe literature and MSDS, our results indicated thephenomenon of acetone on the MEKPO runaway
ng to the experimental and curve fitting results,a thermally unstable material because of the weak
. There were three exothermic peaks during decom-he first peak belongs to an n-order reaction, followedlytic reactions for the second and third. Advantagesod are demonstrated by Kossoy and Koludarova .model of consecutive reactions where the second andwere autocatalytic can be expressed by Eqs. (1)(3):
eEa1/RT (1 )n1 n-order reaction (1)
eEa2/RT ( )n2 (z + n3 )autocatalytic reaction(2)
eEa3/RT ( )n4 (z + n5 )autocatalytic reaction(3)
EKPO 31 mass% was combined with acetoneat three different heating rates (2, 4, and 10 C/min),s indicated that the first exothermic onset tempera-creased from 35.4 to 38.1 C, 43.6 to 79.2 C, andC, respectively. The apparent activation energy (Ea)
en reduced during the mixture conditions for the firstphenomena can be readily seen in Figs. 1 and 2.
Y.-F. Lin et al. / Journal of Hazardous Materials 153 (2008) 10711077 1073
Table 2Thermokinetic parameters derived from the DSC data on MEKPO 31 mass% and mixed with acetone 99 mass% for the first peak of the reaction
Sample (C/min) T0 (C) Ea (kJ/mol) n1 n2 A (s1) z H (J/g)MEKPO 31 mass% (2.5 mg) 10 55.6 136.5 1.1 * 43.2 * 68.7MEKPO 31 mass% (3.2 mg) 4 43.6 119.3 0.9 * 36.5 * 65.6MEKPO 31 mass% (3.1 mg) 2 35.4 117.8 0.8 * 36.9 * 65.5MEKPO 31 mass% (2.2 mg) + acetone 99 mass% (1.0 mg) 10 61.3 138.0 1.0 * 42.9 * 76.2MEKPO 31 mass% (2.9 mg) + acetone 99 mass% (1.2 mg) 4 79.2 184.1 0.8 * 55.9 * 57.4MEKPO 31 mass% (2.5 mg) + acetone 99 mass% (1.0 mg) 2 38.1 120.2 0.7 * 37.8 * 66.3The first peak of the reaction. Calculated values based on experimental data from DSC tests. Estimated values are shown in bold. (*) Not applicable.
Table 3Calculated thermokinetic parameters derived from the DSC data on MEKPO 31 mass% and mixed with acetone 99 mass% for the second peak of the reaction
Sample ol)MEKPO 31 mMEKPO 31 mMEKPO 31 mMEKPO 31 mMEKPO 31 mMEKPO 31 m
The second p sts. Es
One imporhazardousacetone 99combiningcompared wof 4 C/minpeak. Afterwere classireactions. Tto 202.7 J/g4, and 10 Ccondition,bined withdemonstratconditions.were the fir
Practicaemployed trials that itemperatur
on. Are t
Table 4Calculated the
MEKPO 31 mMEKPO 31 mMEKPO 31 mMEKPO 31 mMEKPO 31 mMEKPO 31 m
The third pea (C/min) Ea (kJ/mass% (2.5 mg) 10 78.4ass% (3.2 mg) 4 59.4ass% (3.1 mg) 2 51.8ass% (2.2 mg) + acetone 99 mass% (1.0 mg) 10 78.2ass% (2.9 mg) + acetone 99 mass% (1.2 mg) 4 114.3ass% (2.5 mg) + acetone 99 mass% (1.0 mg) 2 56.2
eak of the reaction. Calculated values based on experimental data from DSC te
tant parameter that was employed to understand thecharacteristics is the heat of reaction (H). Whenmass% (about 1 mg) was doped into the test cell forwith the MEKPO 31 mass%, the H was reducedith the pure solution. Especially under a heating rate, the H decreased from 65.6 to 57.4 J/g in the firstthe first peak was induced, the second and third peaksfied as autocatalytic reactions, treated as unexpectedhe H also decreased from 320.9 to 284.7 J/g, 312.6, and 198.7 to 129.4 J/g under the heating rates of 2,/min, respectively. In terms of the third peak of each
the H also was reduced while MEKPO was com-acetone, as indicated in Table 4. All of the resultse that an inhibitive reaction can be formed in mixing
and sohas moparamto maxrunaw
Tad =As far as T0, H, and Ea are concerned, these resultsst to be reported on these issues.
lly speaking, many safety parameters could beo classify the degree of hazard for hazardous mate-s adopted, such as self-accelerating decompositione (SADT) , temperature of no return (TNR),
Howeven-order reaing autocaonly by apbeen used i
From Fiwithin the twas decreathan the mi
rmokinetic parameters derived from the DSC data on MEKPO 31 mass% and mixed
(C/min) Ea (kJ/mol)ass% (2.5 mg) 10 123.6ass% (3.2 mg) 4 130.8ass% (3.1 mg) 2 144.9ass% (2.2 mg) + acetone 99 mass% (1.0 mg) 10 103.7ass% (2.9 mg) + acetone 99 mass% (1.2 mg) 4 136.7ass% (2.5 mg) + acetone 99 mass% (1.0 mg) 2 122.1
k of the reaction. Calculated values based on experimental data from DSC tests. Estimn1 n2 A (s1) z H (J/g)1.1 0.2 18.5 0.0186 198.70.9 0.5 12.5 3 103 312.60.9 0.7 10.1 0.0373 320.90.9 0.3 18.4 6.8 103 129.41.3 0.3 28.9 1.5 103 202.70.9 0.3 11.1 0.0142 284.7
timated values are shown in bold. z: Autocatalytic constant .
ccording to the experimental results, since MEKPOhan two exothermic peaks, both of these two safetyare not suitable for use in this study. In essence, timem rate (TMR) is used for determining the degree ofactions if an accident occurs. TMR was proposed byand Tou  in 1980, who derived convenient ana-essions (Eqs. (4) and (5)) for calculation purposes:
(5)r, the formulas are valid only for simple single stagections. In the case of more complex reactions, includ-talytic reactions, TMR can be properly determinedplying kinetics-base curve fitting. This method hasn the present study.gs. 35, when MEKPO was combined with acetoneemperature range of 20100 C, the degree of hazardsed. Accordingly, pure MEKPO was more dangerousxed ones.
with acetone 99 mass% for the third peak of the reaction
n1 n2 A (s1) z H [J/g]0.8 0.5 25.6 0.0124 483.31.1 1.7 29.2 0.0945 359.10.4 0.4 31.9 0.0711 373.31.1 1.1 21.4 0.0385 436.10.5 0.4 29.6 0.0162 234.80.4 0.6 26.1 0.0125 330.4
ated values are shown in bold. z: Autocatalytic constant .
1074 Y.-F. Lin et al. / Journal of Hazardous Materials 153 (2008) 10711077
Fig. 1. Comparison of experimental data on heat production for MEKPO 31 mass% and mixed with acetone 99 mass% with kinetics-based curve fitting results((1) MEKPO + acetone with scanning rate of 10 C/min for experimental data and curve fitting; (2) MEKPO with scanning rate of 10 C/min for experimental dataand curve fitting; (3) MEKPO + acetone with scanning rate of 4 C/min for experimental data and curve fitting; (4) MEKPO with scanning rate of 4 C/min forexperimental data and curve fitting; (5) MEKPO + acetone with scanning rate of 2 C/min for experimental data and curve fitting; (6) MEKPO with scanning rate of2 C/min for experimental data and curve fitting).
For determining thermal stability, which is characterized bythe time necessary to reach a certain level of conversion at certainconstant temperature of time to conversion limit (TCL) ,it was necessary to determine the stability of materials understorage or transportation conditions. From Figs. 68, within the
temperature range of 20100 C, the results revealed that pureMEKPO was more unstable than in mixed conditions. This wascontrary to the normal perception that acetone may exacerbatethe severity of hazards incurred by MEKPO in terms of runawayreactions.
Fig. 2. Comp((1) MEKPOand curve fittexperimental2 C/min for earison of experimental data on heat production rate for MEKPO 31 mass% and mixewith scanning rate of 10 C/min for experimental data and curve fitting; (2) MEKPOing; (3) MEKPO + acetone with scanning rate of 4 C/min for experimental data andata and curve fitting; (5) MEKPO with scanning rate of 2 C/min for experimental dxperimental data and curve fitting).d with acetone 99 mass% with kinetics-based curve fitting results+ acetone with scanning rate of 10 C/min for experimental datad curve fitting; (4) MEKPO with scanning rate of 4 C/min forata and curve fitting; (6) MEKPO + acetone with scanning rate of
Y.-F. Lin et al. / Journal of Hazardous Materials 153 (2008) 10711077 1075
Fig. 3. TMR vs. temperature (kinetics-based curve fitting) for MEKPO31 mass% and mixed with acetone 99 mass% with heating rate of 2 C/min.
Fig. 4. TMR31 mass% and
MEKPObis-hydropinto severathe data ofenol radica
Fig. 5. TMR31 mass% and
Fig. 6. Time to conversion limit (TCL) vs. temperature (kinetics-based curvefitting) for MEKPO 31 mass% and mixed with acetone 99 mass% with heatingrate of 2 C/min.vs. temperature (kinetics-based curve fitting) for MEKPOmixed with acetone 99 mass% with heating rate of 4 C/min.
ion mechanism analysis
has seven types of structures . When MEKPO,eroxide, was heate...