New Quantitative Structure-Fragmentation Relationship Strategy for Chemical Structure Identification Using the Calculated Enthalpy of Formation as a Descriptor for the Fragments Produced in Electron Ionization Mass Spectrometry: A Case Study with Tetrachlorinated Biphenyls

Differential mass spectrometry correlated with quantum chemical calculations (QCC-ΔMS) has been shown to be an efficient tool for the chemical structure identification (CSI) of isomers with similar mass spectra. For this type of analysis, we report here a new strategy based on ordering (ORD), linear correlation (LCOR) algorithms, and their coupling, to filter the most probable structures corresponding to similar mass spectra belonging to a group with dozens of isomers (e.g., tetrachlorinated biphenyls, TeCBs). This strategy quantifies and compares the values of enthalpies of formation (Δ_f<i>H</i>) obtained by QCC for some isobaric ions from the electron ionization (EI)-MS mass spectra, to the corresponding relative intensities. The result of CSI is provided in the form of lists of decreasing probabilities calculated for all the position-isomeric structures using the specialized software package CSI-Diff-MS Analysis 3.1.1. The simulation of CSI with ORD, LCOR, and their coupling of six TeCBs (IUPAC no. 44, 46, 52, 66, 74, and 77) has allowed us to find the best semiempirical molecular-orbital methods for several of their common isobaric fragments. The study of algorithms and strategy for the entire group of TeCBs (42 isomers) was made with one of the optimal variants for the computation of Δ_f<i>H</i> using semiempirical molecular orbital methods of HyperChem: AM1 for M^+• and [M – 4Cl]^+• ions and RM1 for [M – Cl]⁺ and [M – 2Cl]^+•. The analytical performance of ORD, LCOR, and their coupling resulted from the CSI simulation of an analyte of known structure, using a decreasing number of isomeric standards, <i>s</i> = 5, 4, 3, and 2. Compared with the results obtained by a classical library search for TeCB isomers, the novel strategies of assigning structures of isomers with very similar mass spectra based on ORD, LCOR, and their coupling were much more efficient, because they provide the correct structure at the top of the probability list. Databases used in these CSI do not contain mass spectra, as in the case of a library search, but a series of Δ_f<i>H</i> values obtained by QCC. These techniques are capable of relating relative intensities to the chemical structures of analytes via Δ_f<i>H</i> of ions which turns out to be a good quantitative structure-fragmentation relationship (QSFR) descriptor