Analysis And Identification Of Characteristic Components In Cigarette And Interferent Ash Residues

Fires caused by smoking occurred frequently over the past few years.Thus,it is of great importance to quickly and accurately determine whether there is cigarette ash among the fire ground residues in the area of origin after a fire for finding out the fire cause.The research on the precise analysis and identification technology for physical evidence of fire scene caused by smoking has always been the focus of domestic and foreign scholars,which is also one of the important problems to be solved urgently in the field of fire investigation.To address this problem,this paper carried out a study on the pretreatment method and analysis technology for cigarette ash,developed a pretreatment device for cigarette ash sample,and established a precise analysis and identification method for physical evidence of fire scene caused by smoking.Meanwhile,an analysis was conducted on the assignment and formation mechanism of characteristic ions in the cigarette ash sample,and a large amount of data obtained therefrom were recognized by AI.Hence,a new technological system integrating"device development","precise identification","mechanism analysis"and"automatic recognition"of the pretreatment on physical evidence of fire scene caused by smoking was set up.It has solved the technical problems of inability to analyze this kind of physical evidence of fire scene for many years,and also provided a scientific and reliable technical support for the precise affirmation of the fire cause conducted by the fire emergency rescue department in our country,showing an extensive application prospect.The specific research work is as follows:1.A new NCI/TOFMS?negative chemical ionization/time-of-flight mass spectrometer?based method was first put forward to examine the cigarette burning ash samples by using acidized emergent gas.Four pretreatment methods,i.e.solid cigarette ash direct sampling,self-built cigarette ash volatile gas in-situ direct sampling,self-built acidized cigarette ash emergent gas sampling,and self-built alkalized cigarette ash emergent gas sampling,were used to process the cigarette ash samples.Through optimization and comparative analysis,the acidized cigarette ash emergent gas sampling method was established to extract the characteristic components in the residues of cigarette ash.The experimental results show that this method is simple and fast,and suitable for the pretreatment of solid samples,such as cigarette ash.Besides,a direct analysis of cigarette ash solid samples was conducted by using the IR and DSC technologies.The results show that IR technology can effectively distinguish the ash of single interfering substance from the cigarette ash,but it cannot effectively identify the mixture of complex interfering substances and cigarette ash.As there are no characteristic minerals in the cigarette ash that can distinguish it from common interfering substances,the DSC technology is not suitable for analyzing cigarette ash samples.Moreover,the GC-MS,PCI/TOFMS,and NCI/TOFMS technologies were used to analyze the cigarette ash direct in-situ volatile gas,the alkalized cigarette ash emergent gas,and the acidized cigarette ash emergent gas respectively.According to the results,no characteristic volatile organic compounds,inorganic salt cations or inorganic salt anions were detected in the ash direct in-situ volatile gas,and no volatile inorganic salt cations were detected in the alkalized ash emergent gas,but non-volatile characteristic inorganic salt anions were detected in the acidized ash emergent gas.As demonstrated by a wide range of experiments,the NCI/TOFMS technology is efficient for examining the acidized emergent gas of cigarette ash and interfering substance ash.2.To satisfy the actual demands for physical evidence of fire scene,two sets of pretreatment devices were developed respectively for small-batch and large-batch cigarette ash,and the volume of ash samples treated by the device all at once was respectively 0?50 g and 0?500 g.Moreover,a water removal system and a heating and stirring system were designed for the device.The structures of four self-developed acidizing systems were optimized,including the bubbling method based on acidizing glass bottle sample injection,the headspace purging-heating method based on needle puncture outfit sample injection,the headspace purging-heating method based on all-in-one bottle plug puncture outfit sample injection,and the headspace purging-heating and stirring method based on all-in-one bottle plug puncture outfit sample injection.Upon the above optimization,the device adopting the headspace purging-heating and stirring method based on all-in-one bottle plug puncture outfit sample injection was ultimately chosen as the acidizing system of the pretreatment device for small-batch cigarette ash.Also,concerning the acidizing system of the pretreatment device for large-batch cigarette ash,the mechanical stirring mode,motor-driven stirring mode,and the strong magnetic stirring mode were respectively studied.Finally,the strong magnetic stirring mode was determined to be the best for the pretreatment of cigarette ash samples.Then,the service conditions of the cigarette ash pretreatment device were optimized,and the pretreatment procedures were determined:using 5%H₃PO₄ as the inorganic acid of acidized cigarette ash;using 1mL 5%H₃PO₄ to analyze 10 mg ash sample;improving the analytical sensitivity by heating and stirring the sample bottle;setting the heating temperature of sample bottle at 40°C;stabilizing the heating temperature of the ionization zone at 50°C;using clean air as the purging carrier gas;setting the flow rate at200 ml/min;setting the flow rate of back purging air at 500 mL/min;setting the spectrum collection time of each sample to 5 min.3.The characteristic ions in cigarette ash were given,and their assignment,source and formation mechanism were analyzed.The NCI/TOFMS technology was used to analyze the ash of 32 brands of cigarettes,the ash of 22 kinds of single interfering substances,the ash of 9 kinds of mixed interfering substances,and the ash of 9 kinds of mixtures of cigarettes and multiple interfering substances.The results show that there were 5 kinds of ions,with m/z=42,60,75,85 and 88 in the ash of different brands of cigarettes.The aforesaid five kinds of ions were not detected at the same time in the single interfering substance ash,the mixed interfering substance ash or the mixture ash of cigarettes and multiple interfering substances,with no effect on the analysis of cigarette ash,indicating that these five ions can be used as a judgment basis for distinguishing cigarette ash from other ashes.Four ions were detected in the ash of single interfering substance foam board,and only four ions were detected in the ash of 2 kinds of mixtures of cigarettes and multiple interfering substances.In practice,faced with this situation,standard substances should be confirmed by considering the on-site conditions.It was demonstrated that the results are consistent with the facts through verification by comparing the ash of 4 kinds of residues from the actual fire scene and the ash of 4 kinds of mixtures of cigarettes and such residues.After the data obtained were processed by PCA,it was found that the data containing cigarette ashes and the data of interfering substances are in different sample sets,indicating that the method can be used to distinguish cigarette ash and interfering substance ash accurately.The accurate quality correction method was used for the assignment of the five characteristic ions.The results show that the assignments of ions with m/z=42,60,75,85 and 88 are NCO^-,H₂O.NCO^-,HNCO.O^2-,HNCO.NCO^-and HCOOH.NCO^-respectively.The NCI/TOFMS analysis of tobacco paper,cigarette filter,cut tobacco and leaf tobacco shows that the characteristic ions come from the burning ash samples of tobacco leaves,and the characteristic ions have higher strength in the ash samples of tobacco leaves with longer growth cycles.Lastly,the generation of characteristic ions was deduced,and the results show that the five characteristic ions are all derived from the cluster ions generated from the reaction of HNCO and O^2-.4.A method was developed based on the deep learning technology for automatically classifying the data of fire ground residues.The Long Short-Term Memory Model?LSTM?for sequence data processing and analysis was established,and the basic structures of Recurrent Neural Network?RNN?and LSTM were provided.Moreover,the performance of the model was optimized,and the influence of training rounds,the number of hidden layers and the number of training data on the performance of the model was investigated.It was concluded that the performance of the model is better when the training rounds are 7,the number of hidden layers is 3,and the larger the training data is.The models and methods established are used to automatically classify more than 1200groups of data from fire ground residues,and it was found that the training accuracy rate is high,up to 87.6%.Thus,the fire ground residue data can be automatically divided into data containing cigarette ash and data without cigarette ash,which greatly reduces the human interference,shortens the time for analysis,and verifies the effectiveness of the LSTM-based classification method of fire ground residues.All these efforts provide methods and theoretical basis for the rapid determination of the presence of cigarette ash contents in residues at the fire scene,which have great application value.