molecular weight distribution analysis (Fig. 5). F shows the most intense signal at 9 DBE for the compound classes N, O and S. This is due to the 3-ring aromatics C12H9N, C12H8O and C12H8S, which are very intense heteroatomic compounds in the GC x GC experiments, too. It is also observed that almost no ions are de- tected in the range of 5-ring aromatics within these three compound classes. It is only from 17 DBE that more intense compounds appear again. In the case of P intense signals of the classes O, N, S, and NS appear only for higher molecular weight compounds of 17 DBE and above. This is noticeable because this mass range can no longer be investigated by GC x GC experiments. This proves the poten- tial of the direct inlet measurements especially for samples with a large fraction of semi-vola- tile compounds. In contrast to the substance classes with one heteroatom, the substance classes with several heteroatoms such as NS, NO and NOS behave similarly in both samples. The compounds of these classes are more evenly distributed over a certain mass range than it was the case especially for F. But here, again, the shift into a higher mass range is clearly visible for P in contrast to F. Another interesting observation is the strong orientation on a diagonal line for all compound classes as already seen for the class of hydrocarbons. The slope of this planar limit is an indication for the growth of the aromatic ring system. The linear and nonlinear integra- tion of benzene rings to an existing aromatic system leads to slopes of 0.75 and 1, respec- tively. In contrast, the integration of saturated rings leads to a slope of 0.2510. Zhang showed that the planar limit of hydrocarbons in CTP has a slope of 0.78, indicating that linear and nonlinear addition of benzene rings is the dom- inant mechanism11. For these samples, the slopes for all compound classes range from 0.75 to 0.85, suggesting that this mechanism is also prevalent for heteroaromatic compounds. Fig. 5. 3D-Isoabundance plots of double bond equivalents (DBE) vs. carbon number (#C) for all compound classes with a minimum of one heteroatom detected by DIP-HRMS. Bubble sizes represent normalized peak areas. CONCLUSIONS Different analytical methods were con- ducted to investigate two organic binders in terms of their elemental composition and mo- lecular structure. Significant differences in their molecular weight distribution were ob- served. In this context, it was shown that DIP- HRMS proved to be extremely useful, espe- cially for the higher molecular weight sample P. With respect to the study of heteroatomic compounds, mass spectrometric analysis was shown to be a robust method to separate differ- ent classes of compounds. Thus, the rule of pla- nar limits known for hydrocarbons could be adapted to heteroatomic compound classes in CTP.