Thermal Hazard Evaluation of Cumene Hydroperoxide-Metal Ion Mixture Using DSC, TAM III, and GC/MS

molecules

Article Thermal Hazard Evaluation of Cumene Hydroperoxide-Metal Ion Mixture Using DSC, TAM III, and GC/MS

Mei-Li You

Department of General Education Center, Chienkuo Technology University, 1 Chieh-Shou N. Rd., Changhua 50094, Taiwan; [email protected]

Academic Editor: Derek J. McPhee Received: 22 February 2016; Accepted: 23 April 2016; Published: 28 April 2016

Abstract: Cumene hydroperoxide (CHP) is widely used in chemical processes, mainly as an initiator for the polymerization of acrylonitrile–butadiene–styrene. It is a typical organic peroxide and an explosive substance. It is susceptible to thermal decomposition and is readily affected by contamination; moreover, it has high thermal sensitivity. The reactor tank, transit storage vessel, and pipeline used for manufacturing and transporting this substance are made of metal. Metal containers used in chemical processes can be damaged through aging, wear, erosion, and corrosion; furthermore, the containers might release metal ions. In a metal pipeline, CHP may cause incompatibility reactions because of catalyzed exothermic reactions. This paper discusses and elucidates the potential thermal hazard of a mixture of CHP and an incompatible material’s metal ions. Differential scanning calorimetry (DSC) and thermal activity monitor III (TAM III) were employed to preliminarily explore and narrate the thermal hazard at the constant temperature environment. The substance was diluted and analyzed by using a gas chromatography spectrometer (GC) and gas chromatography/mass spectrometer (GC/MS) to determine the effect of thermal cracking and metal ions of CHP. The thermokinetic parameter values obtained from the experiments are discussed; the results can be used for designing an inherently safer process. As a result, the paper ﬁnds that the most hazards are in the reaction of CHP with Fe2+. When the metal release is exothermic in advance, the system temperature increases, even leading to uncontrollable levels, and the process may slip out of control.

Keywords: cumene hydroperoxide (CHP); organic peroxide; differential scanning calorimetry (DSC); thermal activity monitor III (TAM III); gas chromatography/mass spectrometer (GC/MS)

1. Introduction Since the 1970s, the chemical industry has been booming in Taiwan, and it has contributed to the industrial development and economic progress of the country. However, we sporadically hear of disastrous accidents in the chemical process industry [1–3]. In fact, any manufacturing process potentially possesses a variety of risks. Therefore, minimizing risks and creating a safer, worry-free, friendly working environment was the main objective of this study and the ultimate goal. In practice, cumene hydroperoxide (CHP) is an initiator for the polymerization of acrylonitrile–butadiene–styrene (ABS), which has wide applications and is ubiquitous in our daily life. CHP is organic peroxide for the following reasons: its structure has a peroxy bond (–O–O–) [4], it has high thermal sensitivity, and it is susceptible to the catalytic properties of contaminants [5–8]. During the manufacturing process, if the process is aging or other uncontrollable factors such as CHP with incompatible chemicals during their exposure to the substance, it may therefore cause unexpected exothermic reaction in advance [9,10]. Out-of-control reaction processes are normally a result of

Molecules 2016, 21, 562; doi:10.3390/molecules21050562 www.mdpi.com/journal/molecules Molecules 2016, 21, 562 2 of 12 Molecules 2016, 21, 562 2 of 12

result of improper control of the manufacturing process, or problems in operating equipment, improperand, controltherefore, of heat the manufacturingis accumulated process,in the syst orem. problems Figure in1 illustrates operating a equipment,process leading and, to therefore, an heat isout-of-control accumulated reaction. in the system. Figure1 illustrates a process leading to an out-of-control reaction.

FigureFigure 1. Illustration 1. Illustration of of a a process process leadingleading to to an an out-of-control out-of-control reaction reaction.

In theIn past, the past, manufacturing manufacturing processes processes used used forfor suchsuch substances substances have have repeatedly repeatedly caused caused casualties casualties or equipmentor equipment damage. damage. Table Table1 lists 1 lists relevant relevant disaster disaster events that that have have occurred occurred worldwide worldwide since since2001. 2001.

Table 1. List of relevant chemical disasters worldwide since 2001. Table 1. List of relevant chemical disasters worldwide since 2001. Date Country Material Cause Hazard Injuries Deaths Date06/09/2001 CountryAustralia Material Cumene Human Cause error Release Hazard Injuries 0 0 Deaths Explosion 09/09/2002 USA CHP System failure 0 0 06/09/2001 Australia Cumene Human error(Reactor) Release 0 0 09/24/2002 USA CHP Mechanical failure ExplosionExplosion 0 0 09/09/2002 USA CHP System failure 0 0 01/05/2003 USA Cumene Human error (Reactor) Release 0 0 Explosion 09/26/2003 Taiwan CHP/DCPO Thermal decomposition 2 0 09/24/2002 USA CHP Mechanical failure(Reactor) Explosion 0 0 01/05/200305/18/2005 USATaiwan Cumene Cumene Accident Human (capsized) error Release – 0 0 0 0 05/24/2006 Taiwan Phenol Human error (splashed) – 0 1 Explosion 09/26/200303/01/2007 Taiwan USA CHP/DCPO Cumene Thermal Barge accident decomposition (leaked) – 0 2 0 0 Explosion(Reactor) 11/09/2007 Taiwan Acetone Fire (leaked) 0 0 05/18/2005 Taiwan Cumene Accident (capsized)(Tank) – 0 0 Explosion 05/24/200601/30/2008 TaiwanTaiwan Phenol DCPO Human Thermal error decomposition (splashed) –0 00 1 (Reactor) 03/01/2007 USA Cumene Barge accident (leaked)Explosion – 0 0 01/08/2010 Taiwan CHP Fire 0 0 (Reactor)Explosion 11/09/2007 Taiwan Acetone Fire (leaked) 0 0 07/19/2013 Taiwan Phenol Accident (capsized) Release(Tank) 1 0 04/04/2014 USA Acetone Thermal accident Fire 0 0 Explosion 01/30/2008Abbreviations Taiwan used in the DCPO table: cumene Thermal hydroperoxide decomposition (CHP), dicumyl peroxide (DCPO).0 0 (Reactor) Explosion 01/08/2010An organic-peroxide-based Taiwan CHP structure has an unstable Fire double bonding group that 0can readily 0 decompose with exothermic characteristics. CHP itself is incompatible,(Reactor) which makes it react with 07/19/2013acids, bases, metal Taiwan ions, or other Phenol substances [9–11]. Accident (capsized) Release 1 0 04/04/2014Corrosion USAresults from the Acetone surface of metal Thermal undergoing accident a chemical or Fire electrochemical 0 reaction 0 with the Abbreviationssubstances present used in in the the table: environment, cumene hydroperoxide which leads (CHP), to dicumylthe gradual peroxide erosion (DCPO). of the metal interior [12,13]. Corrosion releases metallic materials [14,15], and it can be considered to be equivalent An organic-peroxide-based structure has an unstable double bonding group that can readily decompose with exothermic characteristics. CHP itself is incompatible, which makes it react with acids, bases, metal ions, or other substances [9–11]. Molecules 2016, 21, 562 3 of 12

to adding a solution of metal ions. According to the literature, the commonly used materials for fabricating metal pipes are carbon steel (carbon ferrochrome and manganese), steel, chrome, stainless Molecules 2016, 21, 562 3 of 12 steel, and tin (galvanized iron stainless steel). In the past, there have been several cases of damage caused by plague display pipeline corrosion. On 10 July 2011, several fires occurred at a petrochemical industrialCorrosion complex results along from the the west surface coast of metalin Taiw undergoingan because athe chemical corrosion or electrochemicalof the central coastal reaction withpipeline. the substances Subsequently, present the authorities in the environment, ordered the re whichplacement leads of toall thepipelines. gradual The erosion pipes at of the the plant metal interiorwere changed [12,13]. Corrosionto hot-dip galvanized releases metallic steel pipes. materials On 30 [14 July,15 ],2014, and the it can shocking be considered Kaohsiung to bepropylene equivalent togas adding explosion a solution disaster of occurred metal ions. because According the unde torground the literature, pipeline thehad commonlybeen corroded used by materialswater and for gas over several decades. The corrosion caused the pipeline to rupture, leading to the leakage of fabricating metal pipes are carbon steel (carbon ferrochrome and manganese), steel, chrome, stainless enormous amount of propylene. The leakage left 32 people dead and 321 injured. The line connecting steel, and tin (galvanized iron stainless steel). In the past, there have been several cases of damage often requires welding, and substances such as iron, nickel, zinc, copper, manganese, zinc, lead, and caused by plague display pipeline corrosion. On 10 July 2011, several fires occurred at a petrochemical chromium are used for filling. industrial complex along the west coast in Taiwan because the corrosion of the central coastal pipeline. The main objective was to conduct dynamic research to study the stability of the CHP with Subsequently,metal ions, that the generated authorities by orderedcorrosion the phenomena. replacement of all pipelines. The pipes at the plant were changed to hot-dip galvanized steel pipes. On 30 July 2014, the shocking Kaohsiung propylene gas2. explosionResults and disaster Discussion occurred because the underground pipeline had been corroded by water and gas over several decades. The corrosion caused the pipeline to rupture, leading to the leakage of enormous2.1. Thermal amount Analysis of propylene.by DSC The leakage left 32 people dead and 321 injured. The line connecting often requiresIn an experiment, welding, CHP and substanceswas added to such three as incomp iron, nickel,atible metals, zinc, copper, and the manganese, aforementioned zinc, method lead, and chromiumwas used areto achieve used for thermal filling. decomposition, as shown in Figure 2. We can find advanced phenomena on TheT0 for main mixed objective metal ions was of to sample. conduct However, dynamic researchthere is no to appreciable study the stability difference of the in Δ CHPHd and with Tmax metal. ions,After that adding generated CHP to by ZnBr corrosion2, CuBr2, phenomena. or FeBr2, T0 increased to 95.0, 75.0, and 74.0 °C, respectively. For these additions to ZnBr2, CuBr2, and FeBr2, the ΔHd values were 1192.0, 1100.0, and 1246.0 J/g, respectively. 2. Results and Discussion The parameter Tmax was stable at 160.0 ± 5.0 °C, and the relevant values are listed in Table 2. After adding metal ions to CHP, it will let the exothermic onset temperature (T0), in advance of 2.1. Thermal Analysis by DSC appeared from 74.0–95.0 °C, as shown in Table 2. The exothermic onset temperature (T0) of CHP is 105.0In an°C. experiment, Hence, metal CHP ions was can added be said to threeto cause incompatible the reaction metals, that andaffects the CHP, aforementioned let it T0 was method in wasadvance. used to It achieve is noteworthy thermal that, decomposition, apart from the as shownion curves in Figure of Cu2.2+ We and can Fe2+ find, T0 advancedcan have phenomenaadvanced onphenomenon,T0 for mixed and metal its ionspeak of value sample. is far However, from the theresingle is peak. no appreciable However, differenceΔHd is maintained in ∆Hd andwithinTmax . ˝ After1150.0% adding ± 10% CHP J/g, towhere ZnBr 10%2, CuBr is the2, experimental or FeBr2, T0 erroincreasedr. The reason to 95.0, is 75.0,that, in and this 74.0 test, C,only respectively. CHP is used to provide heat, and the metal ion does not generate heat, and therefore, the metal ion does For these additions to ZnBr2, CuBr2, and FeBr2, the ∆Hd values were 1192.0, 1100.0, and 1246.0 J/g, not affect their heat of decomposition. ˝ respectively. The parameter Tmax was stable at 160.0 ˘ 5.0 C, and the relevant values are listed in Table2.

0 Δ 2.5 CHP_T0: 105.0 ( C), Hd: 1,086.0 (J/g)

2.0 2+ 0 Δ CHP + Zn _T0: 95.0 ( C), Hd: 1,192.0 (J/g)

2+ 0 1.5 Δ CHP + Cu _T0: 75.0 ( C), Hd: 1,100.0 (J/g) 2+ 0 Δ CHP + Fe _T0: 74.0 ( C), Hd: 1,246.0 (J/g) 1.0

Heating flow(W/g) 0.5

0.0

0 50 100 150 200 250 300 o Temperature ( C)

Figure 2. Heat flow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2, or FeBr2, Figure 2. Heat ﬂow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2, or FeBr2, tested using differential scanning calorimetry (DSC) analysis at a heating rate of 4.0 °C/min. tested using differential scanning calorimetry (DSC) analysis at a heating rate of 4.0 ˝C/min.

After adding metal ions to CHP, it will let the exothermic onset temperature (T0), in advance of ˝ appeared from 74.0–95.0 C, as shown in Table2. The exothermic onset temperature ( T0) of CHP is ˝ 105.0 C. Hence, metal ions can be said to cause the reaction that affects CHP, let it T0 was in advance. 2+ 2+ It is noteworthy that, apart from the ion curves of Cu and Fe , T0 can have advanced phenomenon, Molecules 2016, 21, 562 4 of 12

and its peak value is far from the single peak. However, ∆Hd is maintained within 1150.0% ˘ 10% J/g, where 10% is the experimental error. The reason is that, in this test, only CHP is used to provide heat, and the metal ion does not generate heat, and therefore, the metal ion does not affect their heat of decomposition. Molecules 2016, 21, 562 4 of 12 Table 2. Nonisothermal data obtained from differential scanning calorimetry (DSC) tests for 80.0 mass% ˝ cumeneTable hydroperoxide2. Nonisothermal (CHP) data mixedobtained with from ZnBr different2, CuBrial2, andscanning FeBr 2calorimetryat 4.0 C. (DSC) tests for 80.0

mass% cumene hydroperoxide (CHP) mixed with ZnBr2, CuBr2, and FeBr2 at 4.0 °C. Organic Peroxide Incompatibility ˝ ˝ Organic Peroxide Incompatibility T0 ( C) Tmax ( C) ∆Hd (J/g) Mass (mg) Substance Mass (mg) T0 (°C) Tmax (°C) ΔHd (J/g) Mass (mg) Substance Mass (mg) – – 105.0 156.0 1086.0 – – 105.0 156.0 1086.0 CHP 80.0 ZnBr2 0.75 95.0 155.0 1192.0 CHP 80.0 2.5% ˘ 10% ZnBr2 0.75 95.0 155.0 1192.0 mass% 2.5% ± 10% CuBr2 0.65 75.0 158.0 1100.0 mass% FeBrCuBr2 2 0.710.65 74.075.0 158.0 163.0 1100.0 1246.0 FeBr2 0.71 74.0 163.0 1246.0 During the production process, CHP and the mixed metal ions are scrambled, and this may cause anDuring advanced the production reaction. A process, substance CHP will and have the out mixed of control metal phenomenonions are scrambled, earlier and than this the may preset cause an advanced reaction. A substance will have out of control phenomenon earlier than the temperature. The second peak of Cu2+ and Fe2+ may also involve detrimental effects. If the protection preset temperature. The second peak of Cu2+ and Fe2+ may also involve detrimental effects. If the system fails to monitor the production process, it will make the process go out of control, resulting in a protection system fails to monitor the production process, it will make the process go out of control, major disaster. resulting in a major disaster. 2.2. Isothermal Hazard Analysis by TAM III 2.2. Isothermal Hazard Analysis by TAM III The experiments involved the pure substance 80.0 mass% CHP, which was maintained under The experiments involved the pure substance 80.0 mass% CHP, which was maintained under 110.0 ˝C in a constant temperature environment in TAM III. TAM III was used to investigate the 110.0 °C in a constant temperature environment in TAM III. TAM III was used to investigate the runaway reaction of CHP mixed with ZnBr , CuBr , or FeBr by recording the associated heat generated runaway reaction of CHP mixed with 2ZnBr2, 2CuBr2, or2 FeBr2 by recording the associated heat ˝ undergenerated isothermal under conditions isothermal atconditions temperatures at temperatur of 80.0, 90.0,es of 100.080.0, 90.0, and, 100.0 110.0 and,C (Figures 110.0 °C3 –(Figures6). In Figure 3–6). 3, theIn CHP Figure thermostatic 3, the CHP peak thermostatic value can peak be seenvalue to can be be 0.0523 seen W/g,to be which0.0523 W/g, corresponds which corresponds to 9.49 h. Metal to ions9.49 were h. Metal then ions added were to then CHP. added In Figure to CHP.3, QInmax Figureclearly 3, Q increasesmax clearly toincreases 0.1152 W/gto 0.1152 and W/gTMR andiso is approximatelyTMRiso is approximately 3.42 h. The 3.42 parameter h. The parameterTMRiso TMRis 0.41iso is 0.41 and and 0.37 0.37 h ath at the theQ Qmaxmax values of of 0.1011 0.1011 and and 0.9320.932 W/g, W/g, respectively. respectively. FromFrom FigureFigure 33,, wewe foundfound that metal ions ions were were added added to to CHP. CHP. The The exothermic exothermic onsetonset temperature temperature occurred occurred earlier, earlier, and and the the heatheat releaserelease was greater than than twice twice that that for for pure pure CHP. CHP.

CHP 80.0 mass% o CHP 80.0 mass% + ZnBr 5.0 mass% 110 C 2 CHP 80.0 mass% + CuBr 5.0 mass% 0.12 2 CHP 80.0 mass% + FeBr 5.0 mass% 2

0.10 2+ CHP+Cu _Q : 0.1152 W/g, TMR : 3.42 h max iso CHP+Zn2+_Q : 0.1011 W/g, TMR : 0.41 h 0.08 max iso CHP+Fe2+_Q : 0.0932 W/g, TMR : 0.37 h max iso 0.06

0.04 CHP_Q : 0.0523 W/g, TMR : 9.49 h max iso Heat flow (W/g) Heat flow

0.02

0.00

0 5 10 15 20 25 Time (h)

Figure 3. Heat flow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2, or FeBr2, Figure 3. Heat ﬂow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2, or FeBr2, tested using TAM III scanning analysis at 110.0 °C. tested using TAM III scanning analysis at 110.0 ˝C.

Figure 4 presents the TAM III pattern parallel to CHP with metal ions; the pattern was analyzed at 100.0 °C. We found that the metal ions forced TMRiso to occur earlier. Among the metal ions, those of FeBr2 can be observed on the leftmost side. TMRiso was 0.34 h, which was earlier than that of pure CHP by a factor of 0.68. ZnBr2 also showed a similar phenomenon. Its TMRiso was 0.42 h, Molecules 2016, 21, 562 5 of 12

almost coinciding with the time of heat release. The TMRiso value of CuBr2 was 8.66 h, which was earlier than that of pure CHP by a factor of only 2.6.

o Molecules 2016, 21, 562 100 C 5 of 12 0.08 CHP 80.0 mass% CHP 80.0 mass% + ZnBr 5.0 mass% 2 0.07 CHP 80.0 mass% + CuBr 5.0 mass% Figure4 presents the TAM III pattern parallel to CHP with metal ions;2 the pattern was analyzed 0.06 CHP 80.0 mass% + FeBr 5.0 mass% ˝ 2 at 100.0 C. We found that the metal ions2+ forced TMRiso to occur earlier. Among the metal ions, those CHP+Fe _Q : 0.0771 W/g, TMR : 0.336 h 0.05 max iso of FeBr can be observed on the leftmost2+ side. TMR was 0.34 h, which was earlier than that of pure 2 CHP+Zn _Q : 0.0302 W/g,iso TMR : 0.416 h Molecules 2016, 21, 5620.04 max iso 5 of 12 CHP by a factor of 0.68. ZnBr2 also showed a similar phenomenon. Its TMRiso was 0.42 h, almost CHP+Cu2+_Q : 0.0307 W/g, TMR : 8.66 h max iso coinciding with the time0.03 of heat release. The TMR value of CuBr was 8.66 h, which was earlier than almost coinciding with the time of heat release.iso The TMRiso value2 of CuBr2 was 8.66 h, which was that of pure CHP by a factor of only 2.6.

earlier than thatHeat flow (W/g) 0.02 of pure CHP by a factor of only 2.6.CHP_ Q : 0.0293 W/g, TMR : 22.88 h max iso

0.01 100 oC 0.08 CHP 80.0 mass% 0.00 CHP 80.0 mass% + ZnBr 5.0 mass% 2 0.07 CHP 80.0 mass% + CuBr 5.0 mass% 0 5 10 15 20 25 30 35 40 45 50 552 60 65 CHP 80.0 mass% + FeBr 5.0 mass% 0.06 Time (h) 2 2+ 0.05 CHP+Fe _Q : 0.0771 W/g, TMR : 0.336 h Figure 4. Heat flow versus temperature maxplot for 80.0 mass% CHPiso mixed with ZnBr2, CuBr2 or FeBr2, CHP+Zn2+_Q : 0.0302 W/g, TMR : 0.416 h tested using 0.04TAM III scanning analysis atmax 100.0 °C. iso CHP+Cu2+_Q : 0.0307 W/g, TMR : 8.66 h max iso Figures 5 and0.03 6 are shown TAM III thermostat tests at 90.0 and 80.0 °C, respectively. The peak

2 of CHP mixedHeat flow (W/g) with0.02 FeBr is tilted to the right, andCHP_ whenQ pure: 0.0293 CHP W/g, is mixed TMR with: 22.88 ferrous h ions, the max iso peak is the same as that at temperatures less than 100.0 and 110.0 °C. Because the heat value is 0.01 considerably higher than that for other materials, it is dangerous and detrimental at temperatures less than 90.0 °C.0.00 The heat release for CHP mixed with CuBr2 and for pure CHP was approximately 0.016 ± 0.001 W/g. However, the peak of zinc ions is not very obvious, amounting to only 0.00504 W/g. Ferrous ions are added,0 it can 5 reach 10 150.10045 20 W/g, 25 30and the 35 heat 40 release 45 50 increases 55 60 appreciably. 65 When relevant disasters occur, the extent of harm maTimey increase (h) sharply. Figure 6 shows a thermostat

experimentFigure 4.conducted Heat flow atversus a temperature temperature plotbelow for 80.080.0 mass%°C. For CHP this mixed condition, with ZnBr only2, CuBrpeaks2 orof FeBr iron2, and Figure 4. Heat ﬂow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2 or FeBr2, zinc ionstested remain. using TAM There III scanning is no heat analysis release at 100.0 for °C.pure CHP and the addition of zinc ions. We can tested using TAM III scanning analysis at 100.0 ˝C. predict that in the thermostat case at 80.0 °C, no reaction will occur. Figures 5 and 6 are shown TAM III thermostat tests at 90.0 and 80.0 °C, respectively. The peak o of CHP mixed with FeBr90 C2 is tilted to the right, and when CHP pure 80.0 CHP mass% is mixed with ferrous ions, the 0.10 CHP 80.0 mass% + ZnBr 5.0 mass% peak is the same as that at temperatures less than 100.0 and 110.0 °C. Because2 the heat value is CHP 80.0 mass% + CuBr 5.0 mass% considerably higher than that for other materials, it is dangerous and detrimental2 at temperatures CHP 80.0 mass% + FeBr 5.0 mass% less than 90.0 °C.0.08 The heat release for CHP mixed with CuBr2 and for pure CHP2 was approximately

0.016 ± 0.001 W/g. However, the peak2+ of zinc ions is not very obvious, amounting to only 0.00504 W/g. CHP+Fe _ Q : 0.10045 W/g, TMR : 0.348 h Ferrous ions are added, it can reach 0.10045max W/g, and the isoheat release increases appreciably. When 0.06 relevant disasters occur, the extent of harm may increase sharply. Figure 6 shows a thermostat experiment conducted at a temperature below 80.0 °C. For this condition, only peaks of iron and 0.04 CHP+Cu2+_ Q : 0.01516 W/g, TMR : 27.423 h zinc ions remain. There is no heat releasemax for pure CHP andiso the addition of zinc ions. We can CHP+Zn2+_ Q : 0.00504 W/g, TMR : 29.418 h predict that in the thermostat case at 80.0 °C, no reactionmax will occur. iso Heat flow (W/g) CHP _ Q : 0.01752 W/g, TMR : 57.406 h 0.02 max iso o 90 C CHP 80.0 mass% 0.10 CHP 80.0 mass% + ZnBr 5.0 mass% 2 0.00 CHP 80.0 mass% + CuBr 5.0 mass% 2 CHP 80.0 mass% + FeBr 5.0 mass% 0.08 0 5 10 15 20 25 30 35 40 45 50 55 60 65 2 CHP+Fe2+_ Q : 0.10045Time (h )W/g, TMR : 0.348 h max iso Figure 5. Heat0.06 flow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2, or FeBr2, Figure 5. Heat ﬂow versus temperature plot for 80.0 mass% CHP mixed with ZnBr , CuBr , or FeBr , tested using TAM III scanning analysis at 90.0 °C. 2 2 2 tested using TAM III scanning analysis at 90.0 ˝C. 0.04 CHP+Cu2+_ Q : 0.01516 W/g, TMR : 27.423 h max iso CHP+Zn2+_ Q : 0.00504 W/g, TMR : 29.418 h max iso ˝ Figures5 and6Heat flow (W/g) are shown TAM III thermostat testsCHP _ at Q 90.0: 0.01752 and 80.0W/g, TMRC,: respectively.57.406 h The peak of 0.02 max iso CHP mixed with FeBr2 is tilted to the right, and when pure CHP is mixed with ferrous ions, the peak is the same as that at temperatures less than 100.0 and 110.0 ˝C. Because the heat value is considerably 0.00 higher than that for other materials, it is dangerous and detrimental at temperatures less than 90.0 ˝C.

The heat release for CHP0 mixed 5 10with 15 CuBr 202 25and 30 for 35 pure 40 CHP 45 50 was 55 approximately 60 65 0.016 ˘ 0.001 W/g. However, the peak of zinc ions is not very obvious,Time (h) amounting to only 0.00504 W/g. Ferrous ions are added, it can reach 0.10045 W/g, and the heat release increases appreciably. When relevant disasters Figure 5. Heat flow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2, or FeBr2, occur, the extenttested using of harm TAM mayIII scanning increase analysis sharply. at 90.0 Figure°C. 6 shows a thermostat experiment conducted at Molecules 2016, 21, 562 6 of 12 a temperature below 80.0 ˝C. For this condition, only peaks of iron and zinc ions remain. There is no heat release for pure CHP and the addition of zinc ions. We can predict that in the thermostat case at ˝ 80.0 C,Molecules no reaction 2016, 21, 562 will occur. 6 of 12

80 oC CHP 80.0 mass% CHP 80.0 mass% + ZnBr 5.0 mass% 0.05 2 CHP 80.0 mass% + CuBr 5.0 mass% 2 CHP 80.0 mass% + FeBr 5.0 mass% 0.04 2

CHP+Fe2+_ Q : 0.04575 W/g, TMR : 0.429 h 0.03 max iso

0.02 CHP+Cu2+_ Q : 0.01045 W/g, TMR : 25.346 h max iso Heat flow(W/g) 0.01

0.00

0 5 10 15 20 25 30 35 40 45 50 55 60 65 Time (h)

Figure 6. Heat flow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2, or FeBr2, Figure 6. Heat ﬂow versus temperature plot for 80.0 mass% CHP mixed with ZnBr2, CuBr2, or FeBr2, testedtested using using TAM TAM III scanning III scanning analysis analysis at at 80.0 80.0˝ °C.C. CHP was used in the isothermal evaluation by using a TAM III scanning analysis at 80.0, 90.0, CHP was used in the isothermal evaluation by using a TAM III scanning analysis at 80.0, 90.0, 100.0, and 110.0 °C. Table 3 presents thermostat experiments involving the addition of three metal ˝ 100.0,ions. and For 110.0 example,C. Table at 110.03 presents °C, Qmax thermostatof CHP is only experiments 0.0523 W/g. After involving the addition the addition of metal ofions, three Qmax metal ˝ ions. Forshows example, an approximately at 110.0 C,two–foldQmax ofincrease CHP isat onlyQmax.0.0523 The time W/g. to maximum After the additionrate at isothermal of metal ions, Qmax showsconditions an (TMR approximatelyiso) for the scanning two–fold analysis increase at 90.0, at 100.0,Qmax and. The 110.0 time °C towere maximum 57.41, 22.88, rate and at 9.49 isothermal h. ˝ conditionsThe TMR (TMRiso isovalues) for thefor scanningcumene hydroperoxide analysis at 90.0, (CHP) 100.0, mixed and with 110.0 FeBrC2 were, tested 57.41, using 22.88, TAM and III 9.49 h. scanning analysis at 80.0, 90.0, 100.0, and 110.0 °C were 0.34–0.43 h. Metal ions affect the reactions The TMRiso values for cumene hydroperoxide (CHP) mixed with FeBr2, tested using TAM III scanning analysisinvolving at 80.0, CHP 90.0, by 100.0, increasing and 110.0 the ˝reactivityC were 0.34–0.43 of CHP, h.and Metal theyions change affect the the heat reactions mechanism involving markedly—especially, CHP added with incompatibility material FeBr2. The process should be CHP by increasing the reactivity of CHP, and they change the heat mechanism markedly—especially, performed very carefully once the system in the incompatible sections is operated in such a real CHP addedhazard. withCHP is incompatibility affected by the addition material of FeBrmetals.2. Storage The process tanks, pipelines, should be and performed reactors used very in the carefully once themanufacture system in of the CHP incompatible are all made of sections metals, isand operated contact of in the such metals a real with hazard. CHP is CHP inevitable. is affected by the addition of metals. Storage tanks, pipelines, and reactors used in the manufacture of CHP are all made of metals,Table and contact3. Heat flow of the for metals80.0 mass% with CHP CHP mixed is inevitable.with ZnBr2, CuBr2, and FeBr2, tested using thermal activity monitor III ( TAM III ) scanning analysis at 80.0, 90.0, 100.0, and 110.0 °C. Table 3. Heat ﬂow for 80.0 mass% CHP mixed with ZnBr , CuBr , and FeBr , tested using thermal 80.0 °C2 90.0 °C2 100.02 °C 110.0 °C activityOrganic monitor III ( TAM III ) scanning analysis at 80.0, 90.0, 100.0, and 110.0 ˝C. Incompatibility Qmax TMRiso Qmax TMRiso Qmax TMRiso Qmax TMRis Peroxide (W/g) (h) (W/g) (h) (W/g) (h) (W/g) o (h) 80.0 ˝C 90.0 ˝C 100.0 ˝C 110.0 ˝C Organic Incompatibility – – – 0.0175 57.41 0.0293 22.88 0.0523 9.49 Qmax TMRiso Qmax TMRiso Qmax TMRiso Qmax TMRiso Peroxide ZnBr2 – – 0.0050 29.42 0.0302 0.42 0.1011 0.41 CHP 80 (W/g) (h) (W/g) (h) (W/g) (h) (W/g) (h) mass% CuBr2 0.0105 23.35 0.0152 27.42 0.0307 8.66 0.1152 3.42 ––– 0.0175 57.41 0.0293 22.88 0.0523 9.49 CHP 80 ZnBr2 FeBr2 –– 0.0458 0.0050 0.43 0.100529.42 0.350.0302 0.07710.42 0.34 0.0932 0.1011 0.37 0.41 mass% CuBr2 0.0105 23.35 0.0152 27.42 0.0307 8.66 0.1152 3.42 2.3. CalculationFeBr of Thermokinetic2 0.0458 Parameters0.43 0.1005 0.35 0.0771 0.34 0.0932 0.37 The heat determined in a reaction kinetics analysis of substances can be used to enhance the 2.3. Calculation of Thermokinetic Parameters safety of the process design. Each stage of the process may be simulated by assuming faulty Theconditions, heat determined and the simulation in a reaction results kinetics can be used analysis for improving of substances the processing can be used phase, to enhance storage, and the safety of thetransport. process design. In these Each experiments, stage of the the TAM process III thermo may bestat simulated obtained the by assumingrelevant parameters faulty conditions, through and the simulationcalculations results involving can bethe usedArrhenius for improving equation and the the processing apparent activation phase, storage, energy. and transport. In these In the thermostat tests performed in this study, 80.0 mass% CHP was considered at 80.0, 90.0, experiments, the TAM III thermostat obtained the relevant parameters through calculations involving 100.0, 110.0 °C and the following equation was used to calculate the apparent activation energy (Ea): the Arrhenius equation and the apparent activation energy. Molecules 2016, 21, 562 7 of 12

In the thermostat tests performed in this study, 80.0 mass% CHP was considered at 80.0, 90.0, ˝ 100.0, 110.0 C and the following equation was used to calculate the apparent activation energy (Ea): ´E k “ A ˆ expp a q (1) 0 RT where k0 is the reaction rate constant, A is the frequency factor, Ea is the apparent activation energy, R is the ideal gas constant, and T is the absolute temperature. Taking the natural logarithm on both sides of Equation (1) gives: E ln k “ ln A ´ p a q (2) 0 RT

In the Arrhenius equation, Ea and A are the only constant values that do not change with the temperature. The greater the temperature, the reactions are faster. Therefore, the following equation can be regarded as a set of dynamic equations:

Ea Ea Ea ln k1 ` p q “ ln k2 ` p q “ ln k3 ` p q “ ¨ ¨ ¨ (3) RT1 RT2 RT3 where T1, T2, T3 ... are the temperatures in the absolute experiment and k1, k2, k3 ... are constants at the temperatures T1, T2, T3 ..., respectively. From the results obtained using the Arrhenius equation and apparent activation energy of the experimental group, we can derive the results, as displayed in Table4. The Ea value of pure CHP is 155.5 kJ/mol. Furthermore, adding a metal layer containing materials incompatible with CHP leads to a decrease in Ea, and the order of Ea values was ZnBr2 (146.8 kJ/mol), CuBr2 (95.6 kJ/mol), and FeBr2 (21.8 kJ/mol). The minimum energy required to kick off a chemical reaction is called apparent activation energy. When molecules collide, the kinetic energy of the molecules can be used to bend, stretch, and ultimately break bonds, leading to chemical reactions. The lowering of apparent activation energy of CHP with Fe2+ means that they are easy to start to react compared with the other ions.

Table 4. Apparent activation energy of CHP decomposition after mixing CHP with ZnBr2, CuBr2, or FeBr2 individually; the apparent activation energy was obtained using the Arrhenius equation.

Material Contaminant Ea (kJ/mol) CHP – 155.46 CHP ZnBr2 146.76 CHP CuBr2 95.55 CHP FeBr2 21.81

2.4. Literature Comparison and Veriﬁcation

To verify the accuracy of this experiment, Table5 displays a comparison of Ea of CHP, obtained in the present study, with the values obtained from the previous studies.

Table 5. Apparent activation energy values of CHP obtained from the literature.

Authors Ea (kJ/mol) Somma, et al. [6] 139.4 ˘ 0.8 Huang, et al. [10] 161.96 Chen, et al. [5] 180.176 Yuan Lu, et al. [11] 138.7 ˘ 5.4 Duh, et al. [7] 122.0, 125.0

2.5. Pyrolysis Products According to CHP Analysis by GC/MS Organic peroxides are widely used in industrial processes because of their high reactivity. Frequent disasters related to the manufacture of organic peroxides may also occur. CHP is prevailingly used for the manufacture of phenol, acetone, α-methylstyrene (AMS), and other common chemicals. Molecules 2016, 21, 562 8 of 12

Molecules 2016, 21, 562 8 of 12 To understand the corrosion of metal containers and release of metal ions during manufacturing process,Moleculesprocess, we 2016 chosewe, 21chose, 562 the the trace trace amount amount of of Zn Zn2+2+,, Cu Cu2+2+, Fe 2+2+ asas experimental experimental factors, factors, and and finally finally evaluated evaluated8 of 12 thermalthermal cracking cracking products products by by GC/MS GC/MS analysisanalysis to determine product product profiles. profiles. process, we chose the trace amount of Zn2+, Cu2+, Fe2+ as experimental factors, and finally evaluated FigureFigure7 shows 7 shows the ion the spectra ion spectra of diluted of diluted 80.0 mass%80.0 mass% CHP, CHP, after theafter completion the completion of GC/MS of GC/MS analysis. thermalanalysis. cracking For the defaultproducts analysis by GC/MS and withinanalysis 35.5 to determinemin, seven product distinct profiles.crests can be observed. First, at For the default analysis and within 35.5 min, seven distinct crests can be observed. First, at 12.243 min, 12.243Figure min, the7 shows first peakthe ionis cumene. spectra Then,of diluted at 16.564 80.0 min,mass% AMS CHP, is isolated, after the and completion it is followed of GC/MS by the the first peak is cumene. Then, at 16.564 min, AMS is isolated, and it is followed by the third product analysis.third product For the acetophenone default analysis (AP) and at 23.337 within min. 35.5 Thmin,ese seven three distinctproducts crests are repeatedcan be observed. in a subsequent First, at acetophenone (AP) at 23.337 min. These three products are repeated in a subsequent atlas, and all of 12.243atlas, and min, all the of themfirst peak are common is cumene. derivatives Then, at or16.564 byproducts min, AMS during is isolated, the CHP and manufacturing it is followed process. by the themthirdAt are 25.274 product common min, acetophenone derivativeswe found that or(AP) byproductsphenol, at 23.337 2,4,5-trimethyl, min. during These the three appeared, CHP products manufacturing and are at repeated27.383 process. min, in a oxazolidine, subsequent At 25.274 min, we foundatlas,3–phenyl, and that allwas phenol, of formed.them 2,4,5-trimethyl, are common derivatives appeared, or and byproducts at 27.383 during min, oxazolidine,the CHP manufacturing 3–phenyl, wasprocess. formed. At 25.274 min, we found that phenol, 2,4,5-trimethyl, appeared, and at 27.383 min, oxazolidine, 3–phenyl, was formed.

Figure 7. Product analysis of CHP decomposition using GC/MS. Figure 7. Product analysis of CHP decomposition using GC/MS. Subsequently, theFigure three 7. Product incompatible analysis ofmetal CHP materialsdecomposition were using added. GC/MS. Figures 8–10 delineate Subsequently, the three incompatible metal materials were added. Figures8–10 delineate the spectra of diluted CHP with ZnBr2, CuBr2, and FeBr2, obtained through GC/MS analysis. Like thethe spectra Subsequently,aforementioned of diluted the CHPatlas three withof incompatiblepure ZnBr substances,2, CuBr metal2, at andmaterials 12.243, FeBr 2were,16.564, obtained added. and through23.337Figures min,8–10 GC/MS thedelineate three analysis. Likethesubstances—cumene, the spectra aforementioned of diluted AMS,CHP atlas withand of AP—appearZnBr pure2, substances,CuBr 2sequential, and FeBr atly. 12.243,2, obtainedCumene 16.564, isthrough presumed and GC/MS 23.337 to be analysis. a min, product the Like of three substances—cumene,thethermal aforementioned volatilization. AMS, atlas In addition, andof pure AP—appear the substances, spectrum sequentially. obtained at 12.243, for the Cumene16.564, addition and is of presumed23.337 zinc ions min, at to 24.914 bethe athree productmin of thermalsubstances—cumene,is that volatilization.of phenylethyl AMS, Inmethyl addition, and ether. AP—appear the Furthermore, spectrum sequential obtained at 30.670ly. Cumene for min, the addition isethane presumeddione, of zinc to(4-methylphenyl) be ions a product at 24.914 of min is thatthermalphenyl, of phenylethyl volatilization.appears. The methyl spectrumIn addition, ether. of Furthermore, thecopper spectrum ions, at αobtained-cumyl 30.670 min,alcoholfor the ethanedione, addition(CA), appears of zinc (4-methylphenyl) at ions 24.914 at 24.914 min, minand phenyl, is that of phenylethyl methyl ether. Furthermore, at 30.670 min, ethanedione, (4-methylphenyl) appears.ethanedione, The spectrum (4-methylphe of coppernyl) ions,phenyl,α-cumyl evolves alcohol at 30.669 (CA), min. appears The spectrum at 24.914 of min,ferrous and ions ethanedione, shows phenyl, appears. The spectrum of copper ions, α-cumyl alcohol (CA), appears at 24.914 min, and (4-methylphenyl)that the fourth crest phenyl, is that evolves of α-ethyl- at 30.669α-methylbenzyl min. The spectrumalcohol at 24.914 of ferrous min. ionsFinally, shows at 30.669 that themin, fourth it ethanedione, (4-methylphenyl) phenyl, evolves at 30.669 min. The spectrum of ferrous ions shows displays theα waveformα of dicumyl peroxide (DCP). crestthat is thatthe fourth of -ethyl- crest is -methylbenzylthat of α-ethyl-α alcohol-methylbenzyl at 24.914 alcohol min. at Finally,24.914 min. at 30.669 Finally, min, at 30.669 it displays min, it the waveformdisplays of the dicumyl waveform peroxide of dicumyl (DCP). peroxide (DCP).

Figure 8. Product analysis of CHP decomposition after mixing CHP with ZnBr2 by GC/MS.

Figure 8. Product analysis of CHP decomposition after mixing CHP with ZnBr2 by GC/MS. Figure 8. Product analysis of CHP decomposition after mixing CHP with ZnBr2 by GC/MS. Molecules 2016, 21, 562 9 of 12 Molecules 2016, 21, 562 9 of 12 Molecules 2016, 21, 562 9 of 12

Figure 9. Product analysis of CHP decomposition after mixing CHP with CuBr2 by GC/MS. Figure 9. Product analysis of CHP decomposition after mixing CHP with CuBr by GC/MS. Figure 9. Product analysis of CHP decomposition after mixing CHP with CuBr2 2by GC/MS.

Figure 10. Product analysis of CHP decomposition after mixing CHP with FeBr2 by GC/MS. Figure 10. Product analysis of CHP decomposition after mixing CHP with FeBr2 by GC/MS. Figure 10. Product analysis of CHP decomposition after mixing CHP with FeBr2 by GC/MS. According to individual normal curves obtained from DSC shown in Figure 2, after the addition According to individual normal curves obtained from DSC shown in Figure 2, after the addition of AccordingCuBr2 and FeBr to individual2 to CHP, the normal heat wave curves peak obtained appears from twice. DSC However, shown in inFigure the GC/MS2, after experiments, the addition of CuBr2 and FeBr2 to CHP, the heat wave peak appears twice. However, in the GC/MS experiments, ofwithin CuBr andthe FeBrfirst threeto CHP, substances, the heat waveonly CA peak repeat appearsed, as twice. shown However, in each in atlas. the GC/MS Furthermore, experiments, the within2 the first 2three substances, only CA repeated, as shown in each atlas. Furthermore, the withinderivatives the ﬁrst CA three and substances, DCP were onlyobtained CA repeated, by adding as Cu shown2+ and in Fe each2+ in atlas.the CHP Furthermore, preparation the process. derivatives It derivatives CA and DCP were obtained by adding Cu2+ and Fe2+ in the CHP preparation process. It CAis andpredicted DCP were that obtainedthis addition by adding should Cugenerate2+ and Femetal2+ in ions the that CHP catalyze preparation the formation process. Itof is the predicted two derivatives.is predicted Moreover,that this addition in the shouldDSC experiment, generate metal the heatingions that process catalyze in the an formationexothermic of reactionthe two thatderivatives. this addition Moreover, should in generate the DSC metal experiment, ions that th catalyzee heating the process formation in an ofexothermic the two derivatives. reaction Moreover,corresponds in the to the DSC second experiment, thermal thepeak. heating process in an exothermic reaction corresponds to the correspondsA comparison to the secondof the hazard thermal posed peak. by the mixture of CHP and incompatible metals in TAM III second thermal peak. and reactionsA comparison involving of the CHP hazard contaminated posed by withthe mixtur impurieties of CHPshow and that incompatible the greatest hazard metals isin posed TAM byIII A comparison of the hazard posed by the mixture of CHP and incompatible metals in TAM III and CHPand reactions mixed with involving Fe2+. The CHP proposed contaminated mechanisms with for impuri metal-catalyzedties show that decompos the greaitiontest [9]hazard are as is follows: posed by reactionsCHP mixed involving with Fe CHP2+. The contaminated proposed mechanisms with impurities for metal-catalyzed show that the decompos greatestition hazard [9] are is posedas follows: by CHP mixed withFe2+ +Fe C62+H.5C(CH The proposed3)2OOH → mechanisms C6H5C(CH3)2 forO‧ metal-catalyzed+ OH− + Fe3+; decomposition [9] are as follows: Fe2+ + C6H5C(CH3)2OOH → C6H5C(CH3)2O‧ + OH− + Fe3+; C6H5C(CH3)2OO‧ + C6H5C(CH3)2H → C6H5C(CH3)2OOH + C6H5C(CH3)2; 2+ C6H5C(CH3)2OO‧ + C6H5C(CH3)2H → C6H5C(CH´3)2OOH3+ + C6H5C(CH3)2; Fe + C6HH55C(CHC(CH3)32)OOH2OOH + HÑ+ →C6 [CH56HC(CH5C(CH3)23)O2O]¨ ++ + OH H2O;+ Fe ; + + C6H5C(CH3)2OOH+ + H → [C6H5C(CH+ 3)2O] + H2O; C6H5C(CH[C6H35)C(CH2OO¨3)+2O] C6 H→5 C(CH[C6H5O(CH3)2H Ñ3)2C]C6;H 5C(CH3)2OOH + C6H5C(CH3)2; [C6H5C(CH3)2O]+ → [C6H5O(CH3)2C]+; 6 5 3 2 + 6 5 3 2 +→ + 6 5 3 2 + 3 2 6 5 C6H5C(CHC H 3C(CH)2OOH) OOH + H +Ñ [C[CH6HO(CH5C(CH) C]3)2O] [C+H HC(CH2O; ) O] + (CH ) CO + C H OH; C6H5C(CH3)2OOH + [C6H5O(CH3)2C]‧+→ [C6H‧ 5C(CH3)2O]+ + (CH3)2CO + C6H5OH; [C H C(CHC6H5C(CH) O]3+)2ÑOOH[C →H CO(CH6H5C(CH) C]3)+2O; + OH ; 6 5 C6H5C(CH3 2 3)2OOH6 →5 C6H5C(CH3 2 3)2O‧ + OH‧; C6H5C(CH3)2O‧ + C6H5C(CH3)2H →+ C6H5C(CH3)2‧ + C6H5+C(CH3)2OH; C6H5C(CHC6H53C(CH)2OOH3)2O +‧[C + C6H6H5O(CH5C(CH33)2HC] →Ñ C6H[C56C(CHH5C(CH3)2‧ +3 )C26O]H5C(CH+ (CH3)23OH;)2CO + C6H5OH; C6H5C(CH3)2‧ + OH‧→ C6H5CH3C= CH2 + H2O C H C(CHC6H5C(CH) OOH3)2‧Ñ + OHC H‧→C(CH C6H5CH) O3C=¨ + CH OH2¨ ;+ H2O 6 5 3 2 6 5 3 2 C6 H5C(CH3)2O¨ + C6H5C(CH3)2H Ñ C6H5C(CH3)2¨ + C6H5C(CH3)2OH; C6H5C(CH3)2¨ + OH¨ Ñ C6H5CH3C = CH2 + H2O Molecules 2016, 21, 562 10 of 12

3. Materials and Methods

3.1. Samples

CHP, CuBr2, and ZnBr2, each with a concentration of 80.0 mass%, were provided by Alfa Aesar (Haverhill, MA, USA), and FeBr2 was purchased from Acros Organics (Taipei, Taiwan). Thermal hazard studies were conducted with 5.0 mass% CuBr2, ZnBr2, and FeBr2.

3.2. Differential Scanning Calorimetry For determining the material energy, the sample and reference material were placed in different experimental containers. However, they were in the same heating furnace and were heated at a constant rate to measure enthalpy changes. The results were continuously recorded as an energy difference function. Dynamic scanning experiments were performed on a Mettler TA8000 system coupled with a DSC 821e measuring test crucible (Mettler ME-26732) (Mettler-Toledo International Inc., Columbus, OH, USA). The test cell was sealed manually by using a kit equipped with Mettler’s differential scanning calorimetry [16,17]. The scanning rate selected for the temperature-programmed ramp was 4.0 ˝C/min, and the range of temperature rise was from 30.0 to 300.0 ˝C for tests.

3.3. Thermal Activity Monitor The thermal activity monitor (TAM III) and its peripherals were manufactured by Thermometric, Inc., (Jarfalla, Sweden), and the operating temperature range was from 15.0 to 150.0 ˝C. The sensitivity level can reach 10.0 nW, and the temperature error is ˘0.01 ˝C. TAM III [18–20] performs measurements in a cylinder, which is located in a 25.0-L constant temperature oil tank to maintain a highly constant temperature environment. There are two measuring cups (one for the sample and the other for the reference products) in each measurement cylinder. Every measuring cup was equipped with Peltier thermopile thermal sensors for measuring the thermal power released by the sample or reference products. The sample and reference products with Peltier detectors were connected in series. The signal principle is based on the sum of voltage signals from the paired detectors. Under isothermal conditions were detected 80.0, 90.0, 100.0 and 110.0 ˝C, we continuously measured the heat generated by substances during decomposition (gradual change in physical properties/resistance), and we titrated the incompatible chemical to measure its thermal stability. Five mass% ZnBr2, CuBr2, FeBr2 was added to 80.0 mass% CHP in TAM III thermostat tests at 80.0, 90, 100, and 110.0 ˝C, respectively.

3.4. Gas Chromatography and Gas Chromatography/Mass Spectrometer (GC/MS) Gas chromatography (GC) [21] and gas chromatography/mass spectrometer were used to analyze a chemical mixture of interest. In GC, the sample is first heated and vaporized by the carrier gas. The vapors of the first sample can then be divided into individual components through a separation process, and volatile compounds contained in the sample can be identified from the peaks in the chromatogram. GC can be used for quantitative analysis as well as for determining the amount of various components. It is estimated that approximately 10%–20% of the materials can be analyzed using GC, but the material must be volatile and thermally stable. The temperature is set to be 5.0–10.0 ˝C which is greater than the boiling point for achieving optimal separation. A mass spectrometer [22,23] is an analytical spectroscopic instrument that is primarily concerned with the separation of molecular (and atomic) species according to their mass. The MS does not actually measure the molecular mass directly, but rather the mass-to-charge ratio of the ions formed from the molecules. With its heating mode with GC/MS, the products of qualitative analysis will use the product of each fragment peaks and GC/MS built-in database for comparison and analysis of possible species. Molecules 2016, 21, 562 11 of 12

Test conditions:Column: Model no: Agilent 19091S–433, ´60.0 ˝C–325.0 ˝C (350.0 ˝C), HP ´5.0 ms, Capillary: 30.0 m ˆ 250.0 µm ˆ 0.25 µm.Carry gas: Helium (1.0 mL/min).Injection port: Split ratio (1:1), 280.0 ˝C, Injection volume: 1.0 µL.Detector: MSD, 280.0 ˝C.

4. Conclusions Organic peroxides have substantial applications. In particular, they are widely used in industrial processes because of their high reactivity. In the DSC experiment, the parameter T0 was obtained as ˝ 105.0 C, and thermal decomposition ∆Hd was 1,086.0 J/g. The addition of metal ions did not lead to ˝ any appreciable changes in ∆Hd, but T0 increased considerably by 10.0–20.0 C. In the TAM III thermostat experiments, metal ions did not affect the CHP exothermic reaction. 2+ ˝ The value of TMRiso of CHP mixed with Fe at 80.0, 90.0, 100.0 and, 110.0 C will appear during 0.34–0.43 h. TAM III test revealed that Fe2+ had the greatest influence on the extent of the CHP reaction. Iron is a major constituent of components used in manufacturing equipment. Therefore, the risk of a thermal disaster also increased substantially. The CHP manufacturing process should be performed very carefully. GC/MS pyrolysis product analysis and CHP analysis indicated the presence of six significant peaks. Among them, the three largest peaks corresponded to cumene, AMS, and AP, which are common products in the CHP manufacturing process. In addition, they also appeared in other patterns, and they were not influenced by metal ions. Investigating the GC/MS pattern of CHP containing added metal ions, we found the presence of anisole, CA, and DCP in the Zn2+, Cu2+, and Fe2+ patterns. Among them, CA and DCP posed a potential thermal hazard. We can predict the cause of the second peak. During the manufacturing process, metal release may cause incompatible reactions and cause reactions to occur earlier, resulting in unforeseen events. If the monitoring system does not respond in a timely manner or adequately, thermal disasters are inevitable. Therefore, we should increase the inspection frequency of these equipment and components and periodically replace them to avoid incompatibility reactions, which can lead to disasters. If a component is corroded, it poses the risk of major accidents. Follow-up studies are required to find ways to prevent corrosions.

Acknowledgments: The author is grateful to the members of Process Safety and Disaster Prevention Laboratory in Taiwan for technical assistance and their valuable comments. The author is especially indebted to Yi-Lin Chen for her technical assistance with the experiment. Conﬂicts of Interest: The author declares no conﬂicts of interest.

References

1. Tseng, J.M.; Kuo, C.Y.; Liu, M.Y.; Shu, C.M. Emergency response plan for boiler explosion with toxic chemical releases at Nan-Kung industrial park in central Taiwan. Process. Saf. Environ. 2008, 86, 415–420. [CrossRef] 2. Chen, C.C.; Wang, Y.C.; Chen, L.Y.; Dia, J.H.; Shu, C.M. Loss prevention in the petrochemical and chemical-process high-tech industries in Taiwan. J. Loss. Prev. Process Ind. 2010, 23, 531–538. [CrossRef] 3. Lin, C.P.; Chang, H.K.; Chang, Y.M.; Chen, S.W.; Shu, C.M. Emergency response study for chemical releases in the high-tech industry in Taiwan—A semiconductor plant example. Process. Saf. Environ. 2009, 87, 353–360. [CrossRef] 4. CAMEO. Chemical Datasheet of Cumene Hydroperoxide. Available online: https://cameochemicals.noaa.gov/ chemical/478 (accessed on 26 April 2016). 5. Chen, K.Y.; Wu, S.H.; Wang, Y.W.; Shu, C.M. Runaway reaction and thermal hazards simulation of cumene hydroperoxide by DSC. J. Loss. Prev. Process Ind. 2008, 21, 101–109. [CrossRef] 6. Somma, I.D.; Marotta, R.; Andreozzi, R.; Caprio, V. Detailed thermal and kinetic modeling of cumene hydroperoxide decomposition in cumene. Process Saf. Environ. 2013, 91, 262–268. [CrossRef] 7. Duh, Y.S.; Kao, C.S.; Lee, C.; Yu, S.W. Runaway hazard assessment of cumene hydroperoxide from the cumene oxidation process. Process Saf. Environ 1997, 75, 73–80. [CrossRef] Molecules 2016, 21, 562 12 of 12

8. Eto, I.; Akiyoshi, M.; Miyake, A.; Ogawa, T.; Matsunaga, T. Hazard evaluation of runaway reaction of hydrogen peroxide—Inﬂuence of contamination of various ions. J. Loss Prev. Process Ind. 2009, 22, 15–20. [CrossRef] 9. Wang, Y.W.; Shu, C.M.; Duh, Y.S.; Kao, C.S. Thermal runaway hazards of cumene hydroperoxide with contaminants. Ind. Eng. Chem. Res. 2001, 40, 1125–1132. [CrossRef] 10. Huang, C.C.; Peng, J.J.; Wu, S.H.; Hou, H.Y.; You, M.L.; Shu, C.M. Effects of cumene hydroperoxide on phenol and acetone manufacturing by DSC and VSP2. J. Therm. Anal. Calorim. 2010, 102, 579–585. [CrossRef] 11. Levin, M.E.; Gonzales, N.O.; Zimmerman, L.W.; Yang, J. Kinetics of acid-catalyzed cleavage of cumene hydroperoxide. J. Hazard. Mater. 2006, 130, 88–106. [CrossRef][PubMed] 12. Jones, D.A. Principles and Prevention of Corrosion; Prentice Hall: Lebanon, IN, USA, 1996. 13. Ho, T.C.; Duh, Y.S.; Chen, J.R. Case studies of incidents in runaway reactions and emergency relief. Process Saf. Prog. 1998, 17, 259–262. [CrossRef] 14. Wesslinga, B.; Posdorfera, J. Corrosion prevention with an organic metal (polyaniline): Corrosion test results. Electrochim Acta. 1999, 44, 2139–2147. [CrossRef] 15. Wranglen, G. An Introduction to Corrrosion and Protection of Metals; Chapman and Hall Limited: London, UK, 1985. 16. You, M.L.; Liu, M.Y.; Wu, S.H.; Chi, J.H.; Shu, C.M. Thermal explosion and runaway reaction simulation of lauroyl peroxide by DSC tests. J. Therm. Anal. Calorim. 2009, 96, 777–782. [CrossRef] 17. You, M.L.; Tseng, J.M.; Liu, M.Y.; Shu, C.M. Runaway reaction of lauroyl peroxide with nitric acid by DSC. J. Therm. Anal. Calorim. 2010, 102, 535–539. [CrossRef] 18. Tseng, J.M.; Shu, C.M.; Gupta, J.P.; Lin, Y.F. Evaluation and modeling runaway reaction of methyl ethyl ketone peroxide mixed with nitric acid. Ind. Eng. Chem. Res. 2007, 46, 8738–8745. [CrossRef] 19. Wei, J.M.; You, M.L.; Chu, Y.C.; Shu, C.M. Evaluation of thermal hazard for lauroyl peroxide by VSP2 and TAM III. J. Therm. Anal. Calorim. 2012, 109, 1237–1243. [CrossRef] 20. Hou, H.Y.; Tsai, T.L.; Shu, C.M. Reaction of cumene hydroperoxide mixed with sodium hydroxide. J. Hazard. Mater. 2008, 152, 1214–1219. [CrossRef][PubMed] 21. Pavlova, A.; Ivanova, R. GC methods for quantitative determination of benzene in gasoline. Acta Chromatogr. 2003, 13, 215–225. 22. Hosaka, A.; Watanabe, C.; Tsuge, S. Rapid determination of decabromodiphenyl ether in polystyrene by thermal desorption-GC/MS. Anal. Sci. 2005, 21, 1145–1147. [CrossRef][PubMed] 23. Liu, J.; Tang, X.; Zhang, Y.; Zhao, W. Determination of the volatile composition in brown millet, milled millet and millet bran by gas chromatography/mass spectrometry. Molecules 2012, 17, 2271–2282. [CrossRef] [PubMed]

Sample Availability: Samples of the compounds are available from the authors.

© 2016 by the author; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).