MS data of polyphenols is incredibly useful when it is combined with the UV spectral information
If you have been successful in determining the sub-group of your polyphenol by its UV spectrum, then you practically already know a part of its molecular weight, i.e. that part that contributes to the UV spectrum. Alternatively, if you know how plant cells construct different polyphenol groups, you know what types of options your compound may have for its molecular weight. Nevertheless, you are free to use the MS and MS/MS data as elegantly as needed to accurately estimate the polyphenol structure by the available information. In many cases this leads you very close to the correct structure, especially if you are able to utilize the compound’s retention time data for characterization purposes as well.
Typically negative ion mode of mass spectrometry is more generally suitable for the detection of polyphenols by electrospray ionization mass spectrometry (ESI-MS). For some polyphenol groups such as flavonoid glycosides, positive ion mode is just as suitable, but for most of the hydrolysable tannins positive ion mode favors the formation of tannin fragments rather than the molecular ions. Thus for molecular weight determination negative ion mode is a recommended choice together with low enough (cone) voltages that guide molecular ions into the vacuum region of the mass spectrometer, without fragmentation at the atmospheric pressure region.
Once the negative ion mass spectrum is obtained, it readily reveals molecular weights of low-molecular-weight polyphenols (300-1000 Da) even with low resolution instruments such as triple quadrupoles. However, for larger polyphenols high resolution mass spectra (by TOF or Orbitrap type instruments) is typically needed to accurately calculate the charge state of all the produced ions and thus the weight of the original polyphenol. Examples below show how the basic mass spectra of a monomeric pedunculagin (784 Da) and a dimeric oenothein B (1568 Da) are practically similar and differences can only be found when m/z values of the ions are zoomed for the distribution of the isotopic signals.