2800 s] or RT1 8500 s & RT2 5000 s were considered. The base mass of each peak was then assigned to a molecular formula with the form CcHhNnOoSs based on its exact mass. Boundary parameters were defined as 2 ≤ c ≥ 100, 2 ≤ h ≥ 100, n ≤ 2, o ≤ 2 and s ≤ 1 and an error of 5 ppm was allowed. Furthermore, only compounds with a maximum number of three heteroatoms, an H/C ratio within the interval [0.4, 2.4] and a number of double bond equiva- lents within the interval [0, 50] were allowed for comparison. All peaks were checked against the NIST 02 MS library. The minimum spectra similarity was set to 850. DIP-HRMS data were processed using the ChromaTOF software (Leco, St. Joseph, MI, USA) and home-built Matlab-scripts (R2020b, the Mathworks Inc., Massachusetts, USA). All features with less than 100 counts and PFTBA-specific fragments with an error window of 5 ppm were removed. Sum formula assignment was then performed for all features according to smallest mass difference and low- est heteroatomic content. The used parameters are analogous to those described above. RESULTS AND DISCUSSION The results of the elemental analyses, solvent extractions and thermogravimetric measurements are summarized in Tab. I. The elemental compositions of both samples are similar with carbon and hydrogen as main com- ponents. However, the heteroatom content is slightly higher in P compared to F, which is particularly due to the higher oxygen content. The lower H/C ratio in P indicates a lower de- gree of saturation. This is consistent with the lower solubility in heptane and toluene and the higher residual carbon content that could be ob- served in the thermogravimetric measurements. Together with significant differences in the sof- tening points, the volatile, low molecular weight fraction seems to be larger in F than in P. Tab. I. Elemental composition, weight percent- ages of solvent-separated fractions and coke residue after heating up to 1500 °C P F C in wt % 91.02 91.79 H in wt % 4.07 4.80 N in wt % 0.93 0.84 O* in wt % 3.03 1.48 S in wt % 0.50 0.52 Others in wt % 0.45 0.57 H/C atomic ratio 0.53 0.62 HS in wt % 0.9 8.2 HI/TS in wt % 42.4 40.3 HI/TI in wt % 56.7 51.5 TG residue (1500 °C, Ar) in % 63.3 52.1 * by difference Fig. 1 shows the results of the GC x GC experiments. By hyphenating mass spectrome- try to GC x GC, a mass spectrum of the corre- sponding separated compound is available for each peak in the chromatogram. These mass spectra can then be used for matching against the NIST database on the one hand. On the other hand, a molecular formula can be as- signed to each base peak, which allows to draw conclusions about the structure of the com- pound, such as ring size or elemental composi- tion. A comparison of sum formulas and reten- tion times shows that the ring size of the base peaks increases with increasing retention time in the first dimension (RT1). The maximum number of double bond equivalents (DBE) is reached at DBE = 17 for both samples, corre- sponding to 6-ring aromatics such as benzo[ghi]perylene. However, the molecular mass distribution within this range differs sig- nificantly between the two samples which is consistent with the results from thermogravi- metric analyses and solubility. While P is com- posed entirely of aromatic compounds, a small fraction of saturated hydrocarbons with chain lengths between 10 and 15 can be observed in