The 13C NMR analysis of lignin has been considered as a very informative, but not very affordable method due to the very long experimental time originally required for quantitative lignin analysis (Chen and Robert, 1988). Recently, we have optimized the experimental time required for 13C NMR analysis and reduced it from 70 to 15—20 h (Capanema et al., 2004, 2005a). Thus, a large amount of valuable structural information (20—30 results on structural moieties per analyzed lignin sample) can be obtained in a reasonably short experimental time which permits considering the 13C NMR method as the most productive one in lignin analysis. Furthermore, we demonstrated that the use of a CryoProbe NMR (Bruker BioSpin MRI GmbH, Germany) allows for 1 h total quan­titative 13C NMR experimental time (Balakshin, Berlin et al., 2013). Therefore, 13C NMR cannot be considered a time consuming lignin analytical technique anymore. In addition, the CryoProbe yields much better signal res­olution, both for 13C and Heteronuclear Single Quantum Coherence (HSQC) NMR methods. However, currently the use of CryoProbe does require tedious and profes­sional optimization of acquisition and processing param­eters to adjust the baseline, specifically for 13C NMR spectra of lignin samples. Therefore, we cannot recom­mend yet the use of this method on a routine basis. Further development in the use of CryoProbe technology and its use for lignin analytical chemistry is expected to mitigate this limitation.

A careful optimization of the acquisition parameters for lignin analysis using a traditional probe yields 13C NMR spectra with a good and reproducible baseline, which can be easily and reliably corrected during spectra processing. This careful adjustment of the acqui­sition and processing parameters has enabled the recording of reproducible results even in different NMR spectrometers with a relative error of ca. 2—3% for the major lignin peaks. Unfortunately, it is not feasible for now to evaluate interlab variability of the 13C NMR method as it has been used much less often than the 31P NMR method and data for the same lignin preparations are very limited. It is also very important to consider some issues when calculating the amount of various lignin moieties using the original 13C NMR spectra as it has been discussed earlier (Capanema et al., 2005b). However, our team has been acquiring over the past few years significant information on different technical lignins, which is hereby summarized in Table 18.6. Since the data were produced and inter­preted by the same analytical methodology, their com­parison is more accurate than the comparisons based on data obtained from various literature reports. The analysis of technical lignins (Table 18.6) clearly showed dramatic changes in lignin structure resulting from the delignification process. In addition to between-process variations, certain within-process lignin structural changes could be documented. One of the most impor­tant factors in these variations is the feedstock origin. Significant differences in the structure of technical lig­nins from different tree species were obvious and they were significantly larger than the differences in the

Phenolic

structure observed for native lignins in these tree spe­cies. For instance, it was shown that various hardwood lignins degraded differently during kraft pulping result­ing in variations of hydroxyl and carboxyl groups, b-O — 4, Ь-Ь, and b-5 linkages as well as in S/G ratio and degree of condensation (Capanema et al., 2005b; Balak­shin et al., 2008). In fact, species-originated variations are similar or even larger than the variations in major lignin functionalities caused by different delignification technologies, such as kraft and OS processes. The only significant differences observed between kraft and ethanol OS lignins (as analyzed by 1-D NMR) are the incorporation of ethoxyl groups and the significantly higher amounts of carbonyl groups in the latter.

Most of the wet chemistry and 31P NMR methods originally yield results in mmol/g (or mass %) units. The 13C NMR method reports results in number of functional groups per aromatic ring (Ar). A conversion factor based on the C9-formulae is typically used to correlate these values (mmol/g and units/Ar), but the ratio is not obvious. The C9-formulae might not be ac­curate (even for high-purity lignins; contaminations would also contribute to this NMR signal) as the lignin side chain is degraded, to a certain extent, during biomass processing. The 13C NMR lignin analysis with Internal Standard (IS) allows for both types of data presentation. A very good correlation between 13C NMR with IS and 31P NMR data for the hydroxyl group content has been reported (Xia et al., 2001). How­ever, in that publication, the authors did not specify if a correction for lignin acetylation was applied to the 13C NMR data or not. The proportion between the values expressed "per 100 Ar" and those in mmol/g (for the same lignin) allow us to calculate the weight of an average C9-unit (or Ar). The numbers obtained are 271 and 243 for Aspen SE and Poplar SE lignin, corre­spondingly (Table 18.3). These values are much higher than those calculated based on the C9-formulae (195 and 193, correspondingly; Table 18.3) and indicate that the amounts of OH groups in the 13C-IS NMR ex­periments have been calculated based on acetylated lignin. Therefore, recalculation as per the original (non-derivatized) lignin would give numbers of ca. 25-40% higher for the 13C NMR (with IS) vs. 31P NMR-II. It should be mentioned that a very good corre­lation between 13C-NMR-IS data (for a non-acetylated lignin) and the methoxyl group wet chemistry analysis, one of the most reliable analytical methods in lignin chemistry, has been reported (Xia et al., 2001). This in­dicates that 13C NMR data should be considered as more realistic and that the 31P-II NMR method pro­duced significantly underestimated numbers (probably due to incomplete derivatization), in agreement with the earlier discussed results.

In summary, 13C NMR with IS is probably the best analytical approach to obtain the most comprehensive and reliable lignin structural information expressed, both in mol% (units/Ar) and in mmol/g. Unfortu­nately, very little has been reported so far on the meth­odology development and its validation of the 13C NMR lignin analysis with IS vs other analytical techniques.

Overall, the lignin scientific community believes, based on a few publications, that there is a good cor­relation between the different methods used for the analysis of the technical lignin chemical structures. However, a comprehensive review of the existing database (especially of independent publications) clearly shows that this is not the case. In fact, the de­viation between reported data using even the same analytical method (such as 31P NMR) for the same lignin preparation is often very significant. Moreover, the deviation between different analytical methods is in the range of the differences observed between different lignin types. This conclusion indicates that significant efforts still should be made to address these deviations and to standardize the analytical methodology for technical lignins analysis. Mean­while, we should remember the general principle that it is naturally more accurate to compare struc­tural data obtained with the same analytical method in the same lab. The use of "reference" lignin samples (well-investigated lignins, such as Alcell and Indulin

AT) would be also very beneficial to ensure at least a reliable relative comparison.