Although, a very high reproducibility has been re­ported in a specific intralaboratory study (Granata and Argyropoulos, 1995), the reproducibility of quantitative 31P NMR spectroscopy reported in independent interla­boratory studies is much lower (Table 18.4). Moreover, even in studies conducted at the same laboratory, one can observe significant divergences between the results reported earlier (Granata and Argyropoulos, 1995) and more recently (Xia et al., 2001) for the same lignin sam­ples (see SE aspen and poplar lignins in Table 18.3). As expected, the worst reproducibility of 31P NMR analytical methods has been observed for different types of 5-substituted phenolic hydroxyls (S-units and

5- condensed G units) in a hardwood technical lignin (Table 18.4) due to a very poor signal resolution.

Content of Major Lignin Functional Groups Determined by Different Analytical Methods

Therefore, in these cases, we consider erroneous report­ing separately S-units and 5-condensed G-units, it is more reasonable to report them as "5-substituted pheno — lics". The reproducibility of 31P NMR is overall better for major signals such as aliphatic hydroxyls (AliphOH), phenolic hydroxyls (PhOH), and total OH, especially for Indulin AT lignin. Surprisingly, it is still not of very high quality the results reported for Alcell lignin (Table

18.4) . In part, but not completely, this observation might be explained by inconsistencies in the lignin itself. In addition, 31P NMR reports much lower carboxyl numbers than wet chemistry methods and 13C NMR methods (Ta­ble 18.3). In fact, 13C NMR methods usually report the sum of carboxyl and ester structures in general. For instance, the significantly higher numbers reported by 13C NMR for Alcell lignin might be explained by the sig­nificant contribution of ester structures (predominantly ethyl esters) in this lignin. However, it is quite unreason­able to expect significant amounts of esters in kraft lig­nins isolated from high alkaline solutions. Therefore, it might be concluded that 31P NMR also underestimates the amounts of carboxyl groups in lignins.

In summary, it appears that 31P NMR-I does not pro­vide sufficient resolution even between major signals, PhOH and AliphOH, while 31P NMR-II yields signifi­cantly lower values. Moreover, separation of S-units and G-condensed (at C-5) structures is very ambiguous in 31P NMR-II analysis of technical lignins, consequently so is the evaluation of S/G ratio and the degree of condensation. The amount of 5-substituted (S + G— condensed) and 5-unsubstitued phenolic hydroxyl should be reported instead. Comparative data for the analysis of various technical lignins by the 31P-II NMR method are summarized in Table 18.5.

Due to the high degree of variability in the structure of lignins discussed above, it is difficult to make any gen­eral conclusions on the existing structural differences among the lignins originated from various technical pro­cesses (Tables 18.3—18.5). In addition to the variability linked to the feedstock origin and process variables, the numbers for different structural moieties reported can vary due to the particularities of the used analytical methodologies (Table 18.3). There is overall a lack of comprehensive comparison of technical lignins. In addi­tion, the existing comparative studies are limited to a few lignin functionalities only, such as amounts of phenolic, aliphatic hydroxyl groups, or methoxyl groups.