EXPERIMENTAL DESIGN AND OPTIMIZATION FOR THERMAL DESORPTION OF ACTIVATED CHARCOAL STRIPS CONTAINING IGNITABLE LIQUIDS FROM FIRE DEBRIS SAMPLES

Zusammenfassung

Fire investigation involves the determination of the origin and the cause of the fire. Fire is commonly initiated or accelerated by using ignitable liquids (ILs). In this study, the thermal desorption of ILs from activated charcoal strips (ACS) was optimized prior to headspace gas chromatography mass spectrometry (GC-MS) analysis for the identification and classification of IL in fire debris samples. Fire debris was simulated by following the modified destructive distillation method where gasoline and diesel were used as the ILs. A Box- Behnken Design (BBD) coupled with Response Surface Methodology (RSM) were used to determine the optimized conditions. A total of 17 experiments (each for gasoline and diesel) were conducted including the 5 central points with incubation temperature (80 – 140 ˚C), incubation time (30 – 60 min), and agitationspeed (250 – 750 rpm) as independent variables; and the sum of the differences between the total ion spectrum (TIS) of gasoline and diesel as dependent variables. According to ANOVA, all three studied variables were significant with a p- value lower than 0.05. The quadratic model was significant with a p- value of 0.0061 and a coefficient of determination equals to 0.911. The lack-of-fit was insignificant which means that therewas no significant differences between the actual and the predicted values. RSM revealed that an increase in incubation temperature and agitation speed increases the response. Moreover, decrease in the incubation time results in a favorable response. Based on the BBD and RSM, the optimized conditions were 140 ºC, 30 min, and 750 rpm. Further optimization of the incubation time was studied by analyzing fire debris samples at different times (5 – 30 min) and 5 min was determined as the optimum condition. The final optimized conditions were applied in combination with chemometrics for the characterization of fire debris samples burned with different ILs. The obtained results demonstrate the feasibility of the method in fire debris analysis.