Molecular formula assignment for dissolved organic matter (DOM) using high-field FT-ICR-MS: chemical perspective and validation of sulphur-rich organic components (CHOS) in pit lake samples.

Molecular formula assignment is one of the key challenges in processing high-field Fourier transform ion cyclotron resonance mass spectrometric (FT-ICR-MS) datasets. The number of potential solutions for an elemental formula increases exponentially with increasing molecular mass, especially when non-oxygen heteroatoms like N, S or P are included. A method was developed from the chemical perspective and validated using a Suwannee River Fulvic Acid (SRFA) dataset which is dominated by components consisting exclusively of C, H and O (78 % CHO). In order to get information on the application range and robustness of this method, we investigated a FT-ICR-MS dataset which was merged from 18 mine pit lake pore waters and 3 river floodplain soil waters. This dataset contained 50 % CHO and 18 % CHOS on average, whereas the former SRFA dataset contained only 1.5 % CHOS. The mass calculator was configured to allow up to five nitrogen atoms and up to one sulphur atom in assigning formulas to mass peaks. More than 50 % multiple-formula assignments were found for peaks with masses > 650 Da. Based on DBE - O frequency diagrams, many CHO, CHOS1, CHON1 and CHON1S1 molecular series were ultimately assigned to many m/z and considered to be reliable solutions. The unequivocal data pool could thus be enlarged by 523 (6.8 %) CHOS1 components. In contrast to the method validation with CHO-rich SRFA, validation with sulphur-rich pit lake samples showed that formulas with a higher number of non-oxygen heteroatoms can be more reliable assignments in many cases. As an example: CHOS molecular series were reliable and the CHO classes were unreliable amongst other molecular classes in many multiple-formula assignments from the sulphur-rich pit lake samples. Graphical abstract An exemplary frequency versus DBE - O diagram. CHOS components but not CHO (and not CHON2 or CHON2S) components were considered here reliable.