31P NMR is a powerful tool for the structural elucidation of polyphenols. This fast, easy, precise, quantitative, and highly reproducible analytical procedure, that allows for the quantification and differentiation of the different types of hydroxy, phenolic, and carboxylic groups in lignins and tannins has now become a routine analytical tool.
The development of sustainable biorefinery products is confronted, among others, with the challenge of lignin and tannin valorization. These abundant, renewable aromatic biopolymers have not been widely exploited due to their inherent structural complexity and high degrees of variability and species diversity. The lack of a defined primary structure for these polyphenols is further compounded with complex chemical alterations induced during processing, eventually imparting a large variety of structural features of extreme significance for any further utilization efforts.
Consequently, a protocol for the rapid, simple, and unequivocal identification and quantification of the various functional groups present in natural polyphenols, is a fundamental prerequisite for understanding and accordingly tailor their reactivity and eventual utility.
Quantitative 31P NMR offers the opportunity to rapidly and reliably identify unsubstituted, o-mono substituted, and o-disubstituted phenols, aliphatic OHs, and carboxylic acid moieties in lignins and tannins with broad application potential.
The methodology consists of an in situ quantitative lignin or tannin labeling procedure using a suitable 31P containing probe, followed by the acquisition of a quantitative 31P NMR spectrum in the presence of an internal standard. The high natural abundance of the 31P nucleus allows for small amounts of the sample (~30 mg) and short NMR acquisition times (~30-120 min) with well-resolved 31P signals that are highly dependent on the surrounding chemical environment of the labeled OH groups.
This procedure, which was recently published in Nature Protocols1 has been cited over 3,000 times in the archival literature and has become a routine measurement for lignin and tannin characterization since it provides essential, rapid, and reproducible structural information.
Lignin and tannins
When Green Chemistry was introduced by Paul T. Anastas and John C. Werner2,3, it drastically changed the general conception of Chemistry. In particular, the importance of employing sustainable materials instead of fossil feedstocks, such as oil and coal, as a starting point is highlighted as a crucial aspect2,3. Among the different types of biomass, lignin is the most abundant aromatic biopolymer and can be seen as a potential source for industrial commodities and high-added value products4.
Lignin is the second most abundant wood-constituent (with cellulose being first and hemicellulose third). Its content in plants varies depending on the plant-type: for example hardwoods characterized by a lower amount of lignin compared to softwoods (20% ± 4% vs. 28% ± 4%). In addition, lignin distribution within vegetable tissue is not homogeneous: the higher lignin content can be found in the cell wall5,6. Lignin is a polyphenolic material industrially obtained as a by-product of the paper/cellulose industry7. It is recovered from the wood pulping process, in which wood chips are primarily processed in the presence of OH– and/or OH– + HS– ion conditions to separate cellulose from hemicellulose and lignin (Soda and/or Kraft processes)8,9.
The first attempts to study lignin were made by Payen and Schultze, respectively, in 1838 and 186510. In 1977, Adler summarized all the relevant available knowledge of that time11. It is currently recognized that the lignin building blocks are three phenyl-propanoid units: p-coumaryl, coniferyl, and sinapyl alcohols. These monomers, thanks to a free radical polymerization process, give rise to p-hydroxyphenyl, guaiacyl, and sinapyl units that eventually broadly constitute lignin (Figure 1)12. The lack of a primary structure in lignins implies an inherent difficulty for its structural characterization. Accordingly, the evaluation of the distribution of molecular weight has always been somewhat controversial. Milled wood lignin, the lignin isolated under mild conditions that approximate mostly protolignin10, is composed of oligomers13 which highly interact via supramolecular aggregation processes14,15.
Figure 1: A representative model of softwood lignin in which the different types of bonds are highlighted. Please click here to view a larger version of this figure.
Lignins are commonly classified depending on: (a) the type of wood from which they are derived (e.g., hardwood and softwood), (b) the process used to isolate it. The most crucial industrial lignin types are Kraft, Lignosulfonates, and Organosolv.
The structure of lignin is highly dependent upon its origin and processing chemistry. More specifically, when the rather complex and irregular structure of lignin is compounded with its natural diversity and the complex processing chemistries, a material of extreme variability, diversity, and heterogeneity emerges, limiting its use to low-value applications16. While softwood lignins contain mainly guaiacyl units (G) with negligible amounts of p-hydroxyphenyl groups (G lignin), hardwood lignins are composed by guaiacyl and syringyl subunits (GS lignin) in varying ratios and grass lignins are constituted by guaiacyl, syringyl, and p-hydroxyphenyl (GSH lignin) subunits. The extractive approach used for isolation dramatically affects the structure of the emerging lignin17. Figure 2 depicts three lignin structures, differing by the isolation approach employed. Some considerations regarding the effect of the extraction method could be highlighted. Firstly, Kraft lignin is a dealkylated, highly fragmented, and condensed lignin, while Organosolv lignin has a structure similar to milled wood lignin (isolated using the Bjorkman approach)18,19,20. Finally, lignosulfonates are characterized by a high degree of sulfonation, depending on the intensity and the conditions of the extractive sulfonation process.
Figure 2: Representative structures for technical lignins. In this figure, the differences among the different types of lignin can be seen. (A) Softwood Kraft lignin is highly condensed, (B) lignosulfonates are characterized by sulfonic groups on saturated carbons, and (C) organosolv lignin has a structure similar to the one of milled wood lignin. Please click here to view a larger version of this figure.
Similar to lignins, tannins are polyphenolic compounds that are found in plants. A recent and updated review on tannins' extractive approaches and applications was recently released by Das et al.21. The importance of tannins in everyday life can be highlighted considering two examples: they impart taste and color to wines22; furthermore their poly-phenolic structure offers antioxidant characteristics and makes them ideal for application in the tanning industry23. Tannins are divided into two classes: hydrolyzable and non-hydrolyzable. Hydrolyzable tannins can be considered a polymer of gallic, di-gallic, and ellagic acid esters (Figure 3). These esters result from the esterification of the phenolic acids with sugar molecules (e.g., glucose, rhamnose, and arabinose).
Figure 3: Typical hydrolysable tannins: tannic acid, vescalgin. Please click here to view a larger version of this figure.
Non-hydrolyzable tannins, also known as condensed tannins, are polymers and oligomers deriving from flavan-3-ols. Among flavan-3-ols, catechins and gallocatechin are the most frequent. They are colorless crystalline compounds (Figure 4). The polymerization creates a polymer characterized by a helicoidal structure. The aromatic hydroxy groups are directed on the exterior of the helix, while the pyran oxygens are in the interior.
Figure 4: Proantocyanidin structures: R =H, OH, OCH3. Please click here to view a larger version of this figure.
Characterization of lignins and tannins using NMR
Two types of information are crucial in lignin or tannin characterization: (a) chemical structure (e.g., hydroxy group content, nature, and frequency of interunit linkages) and (b) molecular weight and polydispersity. Since the early studies on lignin, different techniques have been employed to achieve these goals, and two classes of methods have emerged: chemical and physical methods.
In lignin chemistry, chemical methods, such as alkaline nitrobenzene oxidation, derivatization followed by reductive cleavage, permanganate oxidation, and thioacidolysis, have been historically widely used24,25,26,27,28,29. However, even if the analytical protocols have been implemented and optimized, they are time-demanding, laborious, and require extensive experimental skills30. Alternatively, from the beginning of the instrumental analysis, physical methods have been used to perform lignin and tannin characterizations31. These techniques allow overcoming the problems of classical methods making it easy to characterize lignin structure.
Nuclear Magnetic Resonance (NMR) allows obtaining information about lignin structure and chemical composition among the instrumental techniques. In particular, data from quantitative monodimensional 1H NMR spectra and quantitative 13C NMR spectra can provide information about different types of lignin interunit bondings32,33,34,35. Unfortunately, monodimensional spectra suffer from signal overlap, which can seriously undermine signal integration efforts. Quantitative versions of HSQC (Heteronuclear Single Quantum Coherence), Q-HSQC (Quantitative – Heteronuclear Single Quantum Coherence), have been used to understand lignin structure better, providing helpful information about internal linkages. However, they cannot be fully utilized to determine the various buildings units13,36,37 quantitatively.
To overcome the issues associated with mono- and two-dimensional NMR, substrate derivatization has been considered. Among the advantages of this approach is that specific labels can be introduced within the complex macromolecule and no spectral interference results from the solvent in which the labeled substrates are dissolved1. Verkade was the pioneer in this field, performing 31P NMR analysis of phosphorous derivatives, coal derivatives, and related compounds38. In its publication, a screening of different phosphorus-containing reagents (phospholanes) was performed, and the chemical shift of other labeled compounds was recorded. Argyropoulos' team first introduced derivatization for the quantitative and qualitative analysis of hydroxy groups in lignin in 1991. After studying the derivatization of lignin model compounds using phosphorus-containing reagents, his group paved the way for one of the most daily-used techniques in lignin chemistry, 31P NMR analysis39,40,41,42,43. Among the different phospholanes examined, Argyropoulos arrived at the use of 2-chloro-4,4,5,5-tetramethyl-1,3-2-dioxaphospholane (TMDP) as being the most suitable one to perform lignin analysis44. TMDP selectively reacts with hydroxy groups causing the quantitative formation of phosphorus-containing derivatives characterized by specific 31P NMR chemical shifts (Figure 5).
Figure 5: Lignin and tannin phosphytilation chemistry. Labeling lignin and tannin labile H groups is accomplished by in situ reaction. The labeled polyphenols are characterized by specific 31P NMR bands corresponding to the different type of hydroxy groups. Please click here to view a larger version of this figure.
Sample derivatization is performed in a pyridine/chloroform (1.6:1) mixture; this choice results from an accurate evaluation. Pyridine has two advantages. Firstly, selecting a solvent characterized by a Hildebrand parameter of about 22.1 MPa1/2 simplifies and amplifies lignin solubilization45. Consequently, the addition of pyridine as a solvent, whose Hildebrand parameter equals 21.7, is thus optimal. Secondly, the reaction of TMDP with hydroxy groups is accompanied by the formation of hydrochloric acid (HCl) as a by-product with concomitant negative implications toward the facile formation of lignin-phospholane derivatives. For this reason, the resulting HCl needs to be neutralized. When present in significant excess, the basicity of the pyridine, relative to TMDP, allows for the neutralization of the HCl (via the formation of pyridine hydrochloride).
The use of the recommended pyridine/deuterated chloroform binary solvent system is based on three reasons. Firstly, it favors sample dissolution. Secondly, as pyridine hydrochloride is soluble in chloroform, it can prevent precipitation and deterioration of the final spectrum. Thirdly, deuterated chloroform is chosen for its unique singlet signal, allowing locking of the NMR spectrometer during the acquisition process. Sample derivatization is performed in the presence of an internal standard. In this way, when the sample and the standard are derivatized, the comparison of the integrals of the peaks of the sample and the standard allows the quantification of the amount for each type of hydroxy group present. Various compounds have been considered as internal standards. These compounds are characterized by a single hydroxy group per molecule, offering a single sharp signal in the 31P NMR spectrum after derivatization. The selection of the standard must be made carefully. Its signal should not overlap with those of the derivatized sample. Cholesterol was widely used during the early days. However, a partial overlap with signals arising from aliphatic hydroxy group limits its use. For routine analysis, internal standard solutions of N-hydroxy-5-norbornene-2,3-dicarboximide (NHND) are preferred. However, owing to NHND instability, its standard solutions can be stored only for a few days46.
The described method represents the implementation and optimization of the analytical protocol aimed at the qualitative and quantitative characterization of lignins as developed by Argyropoulos37,38,39,40,41,42. Compared to many other techniques available for lignin structural elucidation, the method has been widely accepted as being among the most facile, rapid, and reproducible. The validity of the wet chemical methods (e.g., nitrobenzene, permanganate oxidations, etc.) relies on the good experimental skills of the operator, effectively confining the method to limited operators. Furthermore, it is not uncommon to encounter correction factors in the literature for wet chemical methods to account for several drawbacks. The described 31P NMR protocol does not require advanced experimental skills making this readily applicable, user-friendly, and widely available. Compared to other instrumental analytical methods, 31P NMR is the only technique capable of precisely detecting and quantifying the different hydroxy groups in lignins. For instance, FTIR can be used to identify various hydroxy groups such as 1H NMR. Both techniques, however, suffer since they cannot offer reliable quantitative data due to extensive signal overlap issues. Another widely used technique is UV-Vis spectroscopy, firstly reported by Goldschmid. The approach, however, is limited to a general overall determination of hydroxy groups since it cannot effectively differentiate among aliphatic, aromatic, and carboxylic OHs47.
From an economic point of view, the only limitation of the 31P NMR technique is the price of TMDP, which is a relatively expensive reagent. It costs about 190 USD per gram; consequently, if the cost of analysis would be approximated only to the price of TMDP, excluding those deriving from the pyridine/chloroform mixture and those of the operator time, it would amount to about 24 USD per analysis. To solve this issue, many laboratories resort to synthesizing TMDP, thus, reducing the reagent costs. To do this, pinacol and phosphorous trichloride are reacted in the presence of triethylamine44. Technically, this reaction is relatively easy; however, care in the use of phosphorous trichloride and its work-up, including well-controlled vacuum distillation, is required. More details as to the synthesis of the TMDP can be supplied upon request.
Even though this protocol is among the best in terms of ease, reproducibility, and precision, some critical points need to be highlighted. Firstly, the sample needs to be fully soluble in the identified pyridine/chloroform mixture. This consideration is fundamental because the quantitative phosphitylation reaction of the hydroxyl groups needs to take place under completely homogeneous conditions. If only part of the sample is solubilized, the resulting analysis would be inaccurate. Secondly, the sample to be examined needs to be moisture- and solvent-free since these variables will detrimentally affect the precision and overall success of the analysis. Traces of humidity will react with TMDP giving 2-hydroxy-4,4'-5,5'-tetramethyl-1,3,2-dioxaphospholane. This compound is a pale-yellow flocculating salt, insoluble in the pyridine/chloroform solvent mixture, causing inadequate NMR signal acquisition. Since only a small weight (~30 mg) of a sample is required, it needs to be free of volatiles for its precise weight to be accurately known before the analysis.
Sometimes, sample solvation issues can be promoted (especially for highly oxidized samples) by adding small amounts of a co-solvent (i.e., dimethylformamide), aiding sample dissolution. In principle, every solvent that does not interact with TMDP can be used to help sample dissolution. The election of a co-solvent cannot include co-solvents containing labile hydroxy or amino groups since they react with the reagent, causing misleading final spectra. Notably, dimethylsulfoxide also reacts with TMDP precluding its use as a co-solvent. Pyridine-based ionic liquids, such as 1-allyl-3-butylpyridinium chloride, can be used when solubility issues arise; however, the ionic liquid should be once again dry48. To dissolve lignosulfonates (a lignin type characterized by a high sulfonation degree), a pre-treatment involving the conversion of neutralized groups into their acidic form was demonstrated to be helpful. Lignosulfonates can be conveniently converted into their acidic conditions using acidic exchange resins in aqueous media. The resulting lignosulfonic acids are isolated from the solution by their adsorption on specific resins (e.g., XAD-7) and desorption in ethanol. Evaporation of the ethanolic solutions over reduced pressure at 40 °C allows the isolation of lignosulfonic acids. These lignins can then be characterized by 31P NMR because they are soluble in the pyridine/chloroform mixture proposed by the protocol.
Prolonged vacuum drying at mild temperatures effectively reduces the amount of moisture and other volatiles in each sample. Notably, small quantities of water do not affect the final spectrum because TMDP is added in excess. In addition, in some cases, a small amount of 2-hydroxy-4,4'-5,5'-tetramethyl-1,3,2-dioxaphospholane may result from the humidity present in the NMR tube or the sample vial. In these cases, stirring is sufficient to dissolve the amount of the formed precipitate completely. If a high amount of 2-hydroxy-4,4'-5,5'-tetramethyl-1,3,2-dioxaphospholane is formed, it is suggested to repeat the sample preparation, improving the drying treatment. For instance, before use, all glassware can be briefly heated with a heat gun.
The spectral range used to record the spectrum is broad compared to the region of interest for the signal regarding the different hydroxyl groups. However, this is compulsory to understand whether the sample derivatization occurred successfully. The confirmation of complete sample derivatization is given by the presence of a strong signal around 174 ppm. This sharp peak is due to the unreacted TMDP, and its existence ensures that the reagent was present in excess, and therefore, all hydroxyl groups have been derivatized. If this peak is absent, the two most probable causes are: (1) the amount of TMDP used is insufficient to perform the complete derivatization of the sample, or (2) a high amount of water is present in the sample. In the first case, using a higher amount of TMDP would likely ensure the sample's complete derivatization, and the signal at 174 ppm will appear. In the second case, the sample should be dried more extensively. Once an excess of TMDP is ensured, peak integration can be performed. Before this operation, zoom to a narrower window (150 to 132 ppm) that confines the signals of interest.
The amount of sample (~30 mg) to be analyzed, reported in the above experimental protocol, has been selected to collect good quality spectra for a 300 MHz NMR spectrometer or more. Nevertheless, we have observed that it is possible to reduce the sample amount if a 500 MHz or higher field magnet is used. For instance, in Figure 8D, the NMR spectrum (resulting from a 700 MHz instrument) of a sample prepared with 7.2 mg of lignin is shown. Signal integration of this spectrum offers the same results as those obtained when using higher amounts of lignin. This fact amplifies the applicability of this protocol for all the research in which small quantities of products are available.
Overall, this experimental protocol can be applied to many research and development applications when understanding the origin and fate of the various hydroxy groups present in lignins and tannins is required. In particular, when coupled with GPC and HSQC data, the resulting data offers the opportunity to further elaborate and speculate over the structure of lignin or a tannin. In many instances where chemical modifications are applied to the hydroxy groups of lignin or a tannin, quantitative 31P NMR analyses can be extremely valuable to detect whether these modifications occurred and to which degree. For instance, Figure 9 shows two NMR spectra of the same lignin before and after its oxidation. A simple qualitative evaluation shows the reduction of both aliphatic and aromatic hydroxy groups upon oxidation, thus providing valuable information and guidance.
Figure 9: Quantitative 31P NMR spectra of the same Organosolv lignin derivatized using TMDP (A) Prior and (B) post its oxidation. The spectra were recorded using a 300 NMR spectrometer. Please click here to view a larger version of this figure.
In conclusion, this technique has all the attributes of being among the most essential and powerful tools when inquiries dealing with polyphenolic, OH bearing lignins, and tannins (and even synthetic polymers)49,50,51 need to be made in a variety of fields, ranging from chemistry to engineering, from biology to polymer, and pharmaceutical applications.
The authors have nothing to disclose.
This work over the years has been supported by various financial awards that included organizations such as the Pulp and Paper Research Institute of Canada, McGill University Montreal, the Natural Sciences and Engineering Research Council of Canada, National Science Foundation USA, United States Department of agriculture, and the Solvay company.
100 – 1000 µl Eppendorf micropipette | VWR | 613-0866 | |
20 – 200 µl Eppendorf micropipette | VWR | 613-0865 | |
2-chloro-4,4,5,5-tetramethyl-1,3-2-dioxaphospholane, 95% | Sigma-Aldrich | 447536 | |
Analytical balance (sensibility ± 0.1 mg) | Precisa | LX220 A | |
Binder Vacuum Oven | Binder | VD53 | |
Certified Vial Kit, Low Adsorption (LA), 2 mL, pk of 100 | Sigma-Aldrich | 29651-U | |
Chloroform-d | Sigma-Aldrich | 151823 | |
Cholesterol, Sigma-grade | Sigma-Aldrich | C8667 | |
Molecular sieves, 4A | Sigma-Aldrich | 208604 | |
N-hydroxy-5-norbornene-2,3-dicarboximide, 97% | Sigma-Aldrich | 226378 | |
NMR spectrometer, 300 MHz | Bruker | ||
Norell natural quartz 3 mm NMR tubes | Sigma-Aldrich | NORS33007 | |
Pipette tips, 100-1000 µL UltraFine (blue) | VWR | 613-0342 | |
Pipette tips, 20-200 µL Bevel Point (yellow) | VWR | 613-0239 | |
Pyridine, anhydrous, 99.8% | Sigma-Aldrich | 270970 | |
Stirring bars,micro, 3 mm lenght | VWR | 442-0360 | |
Stirring bars,micro, 6 mm lenght | VWR | 442-0362 | |
Triphenylphospine oxide, 97% | Sigma-Aldrich | T84603 | |
Vials for environmental analysis, WHEATON, 20.00 mL | DWK Life Sciences | WHEAW224609 | |
Weighing paper, grade 531 | VWR | 516-0318P |