Soybeans have since 5000 years been a very important agricultural crop due to their high content of both triglyceride oil and edible proteins [54]. After oil removal the residue can be used for food [55]. When instead, soybean meal is to be used for adhesives, the solid residue is heated at a temperature lower than 70°C to preserve the proteins solubility in alkali [55]. The protein content of oil-free soybean meal is in 35% to 55% range worldwide, while a protein content of 44%–52% range is usual for industrial grades. Carbohydrates and ashes constitute around 30% and 5%–6% respectively of soy meal [56] with around 10% moisture content. Adhesive grade ‘‘untoasted’’ soybean meal is generally the one performing best, but it needs grounding or milling to a really very fine flour [55,57] at least 40%, and preferably 60%–80%, passing through a 46-mm (325-mesh) screen.

Organic gels of tannin and soy protein and its hydrolysates are known and their technology has been published in the food and nutritional industry. They are based on the age old secondary forces interaction between the two materials as used in the manufacture of leather. Their development is aimed at a variety of different nutritional uses but definitely not for adhesives or other solids materials [106–109]. Namely, they remain to the stage of rather soft gels, for example to be used as fat replacers for meat products [110], stabilizing proteins Pickering emulsions with tannins [111], to minimize or remove tannin astringency in certain fruits [112], or even for materials nano-encapsulation using soy protein hydrolysates–tannic acid complexes obtained by photocatalysis [113], and many other similar uses. The focus of this review is instead on solid materials application, thus rather different then the food and nutritional applications described.

Tannin only organic gels can be prepared. Apart the tannin-formaldehyde or tannin-aldehyde gels tannins do also form gels by autocondensation [41–44] and form gels, with formaldehyde or other aldehydes [45–50], their gelling also being catalyzed by reaction with silica and silicates, first found and applied for wood adhesives [41–44] and afterwards also for other applications [46,47,50,114] or by reaction with ammonia [51–53]. These approaches are of interest in some fields, the first approach including also soy-tannins gelling for wood adhesives, or alternatively in combination with another biomaterial, such as lignin [115] or soy flour protein [47]. Soy-tannin-formaldehyde has been proposed as a wood particleboard binder [116]. However, diluted resins may be used for preparing gels. Denatured soy protein reacted with formaldehyde and crosslinked with a condensed tannin gels were prepared at a number of pHs, cured at 85°C, and supercritical CO₂ dried, forming in this manner one of the “greenest” organic aerogels [47]. The authors proposed for these tannin-soy-formaldehyde gel systems a reaction mechanism (Fig. 10) [47], by FTIR, ¹³C-NMR and XPS analyses. Under the particular specific conditions used the method employed was by denaturing first the soy protein to expose the protein skeletal amide groups (N), then reacting it with formaldehyde to form hydroxymethyl groups on the N atoms, followed then by copolymerizing with the tannin (T). The authors indicated that crosslinking occurred by formation of methylene bridges connecting soy and tannin (N–CH₂–T) or soy to soy (N–CH₂–N) and tannin to tannin (T–CH₂–T) (Fig. 7b). The highest Tgel for tannin-soy-formaldehyde gels was at pH 6 and 85°C [47] due to the shift of system reactivity when comparing it to the pH 4–5 and 50°C for tannin-resorcinol-formaldehyde gels [35].

Formaldehyde-free tannin gels are well known and present also an interesting effect. In effect, tannins gel very fast, at room temperature, in really short times. In depth studies on this gelling reaction and its mechanism, and the subsequent hardening, has been carried out [28,42–44] and is caused by the exothermal auto-condensation of tannins catalyzed by certain Lewis acid, such as silica smoke, nano-silica, silicates such as waterglass, boric acid and aluminum trioxide [28,42–44]. The time of gelling varies according to the amount and type of Lewis acid added, from as short as a couple of minutes to as long as 30 min, gelling being also promoted by the strong exothermal reaction generated. Tannins can also produce formaldehyde-free gels by reacting with ammonia at a higher concentration. This causes substitution with amine groups of all or the majority of the tannin hydroxyl groups, followed by crosslinking by spontaneous oligomerization of the aminated tannin through –N = bridges [51–53]. Such hydrogels present a porous texture depending on the relative proportions of the materials used. They can be freeze-dried, dried under normal conditions, or or dried under supercritical conditions. This approach has been used to pyrolyse the hydrogels, leading to carbon xerogels, cryogels or aerogels all particularly rich in N [117,118].

However, the mechanism of formaldehyde based gels explained above [47] is only part of the story. First of all bridges of the type Protein-CH₂-N-CH₂-N-CH₂-Tannin are an incorrect interpretation in Fig. 7, the -N-CH₂-N-bridge of the shoulder on the side of the big 72 ppm peak being a Protein-CH₂-N-CH₂-Protein bridge type only. The peak at 53 ppm is correctly assigned to a Protein-N-CH₂-Tannin. The peaks at 42, 32 and 30 ppm are also correctly assigned to Tannin-CH₂-Tannin-methylene bridges between different tannin sites. A direct tannin-protein linkage without presence of formaldehyde as indicated in the original article of Fig. 10 [47] is not noticeable by ¹³C nuclear magnetic resonance (¹³C NMR), being this masked, if it even occurs, by the bridges formed by the formaldehyde. In effect other pointed studies under different conditions [99] have not been able to observe with ¹³C NMR the presence or not of such direct linkages under different conditions when masked by formaldehyde or other formaldehyde-yielding compounds are used. A different techniques was needed then to be used. In materials where gelling must eventually lead to hardening such as in wood adhesives the mechanism has always been thought to be rather different. A first indication of how different, was obtained from a series of several publications in which soy protein isolate was reacted with both condensed tannins and with hydrolysable tannins [119–121].

In the first two of these [119,120] “tannic acid”, thus a commercial chestnut hydrolysable tannin, was employed for preparing adhesives from soy-flour to bond fiberboards and plywood. Differential scanning calorimetry (DSC) in the 25°C–220°C range, thermogravimetric analysis (TGA) and its temperature dependent derivate (DTG) for different soy adhesives were measured between 40°C and 300°C. Fourier Transform Infrared (FTIR) analysis indicated that tannic acid addition to a soy-adhesive decreased both its viscosity and pH. Thus, tannic acid appeared to improve the adhesion performance of soy flour resins. The DSC analysis indicated that both soy binders’ glass transition temperature and temperature of denaturation decreased by tannic acid addition. TGA and DTG indicated that soy flour thermal degradation appeared to start above 146°C. By FTIR spectroscopy it was determined that tannic acid and the amino acid of soy protein in soy flour appeared to react well (Fig. 11). Moreover, the plywood bonded with the tannic acid modification of soy adhesives showed good water resistance by both delamination and shear strength tests. Fiberboards properties such as Modulus of Rupture (MOR), Modulus of Elasticity (MOE), Internal Bond (IB) strength, and water resistance were improved, by the tannic acid modification of soy adhesives.

In the approach taken in this work urea was also used to decrease and control the viscosity of the soy-tannin combination to allow its application as a wood adhesive. The spectra of soy flour-urea and soy flour-urea-tannic acid are shown in Figs. 11 and 12. These show peak intensity changes due to the reaction of the tannic acid with the soy protein amino acids and that the addition to soy flour-urea of tannic acid increases the intensity of the 3465, 2959, 2926, 1243 and 1054 cm⁻¹ peaks. The C = O stretching peak at 1641 cm⁻¹ of the amide II also decreases in intensity by the addition of tannic acid. Such a reaction looks like occurring according to two mechanisms: (i) by the tannic acid phenolic –OH groups reacting with the side chains amine groups of amino acids possessing a side chain amine group, even with amino groups that are converted eventually in amide groups once they are included in the skeletal structure of the protein. Soy protein contains a remarkably high proportion of amino acids possessing side chain amine groups. Fig. 12 shows that chemical bonds form between tannic acid and amino acids, but again at this stage cross-liking was dominated by the methylene groups formed by formaldehyde with all the other three materials, namely soy, tannin and urea. (ii) By numerous tannin-protein secondary forces interaction: this is the same mass of interactions in tanning proteins to produce leather and why proteins precipitate from solution by tannins addition. The tannin and the protein are almost irreversibly linked by millions of these secondary forces. A 3-stage model of the tannins and proteins interaction has been proposed by Charlton et al. in 2002 [122]: In its initial phase, the tannin aromatic ring planar surface and hydrophobic sites of proteins such as pyrrolidine rings of prolyl residues interact to form hydrophobic associations. At the same time, the hydroxyl groups of the tannin complexing the H-acceptor sites (carbonyl and –NH₂ groups) of proteins are helped in the complexes stabilization by the hydrogen bonds between the two (Fig. 13). Then afterwards, the self-association of the protein-tannin complexes occurs by further hydrogen bonding. This results in larger but soluble tannin-protein complexes then aggregating. In the end the aggregated complexes become sufficiently large to form insoluble precipitates.

The urea addition to the soy resin adhesive is one reason for the lowering of the MOE at higher tannin proportions. Urea is known to contribute to partially block some flavonoid tannins rearrangements and condensation reactions [19], mimosa tannin included. The urea effect on a flavonoid tannin self-condensation reaction is shown in Fig. 14. Urea has shown in previous work to be a blocking agent of tannin self-condensation, since with tannin model compounds it has successfully decreased self-condensation, such as also m-phenylenediamine and phloroglucinol are able to do [19].

The use of tannins in combination with soy flour to make wood particleboard has also been investigated. Thus, the feasibility of using three types of commercial tannins, namely two flavonoid condensed types (quebracho, mimosa) and one hydrolysable type (chestnut tannins) as cross-linking materials for soy adhesives has been investigated. The chemical bond formation and adhesion behaviors of tannin-modified soy adhesives were also analyzed by Matrix Assisted Laser Desorption Ionization Time-of-Flight (MALDI-ToF) mass spectrometry and thermomechanical analysis (TMA), indeed indicating the formation of tannin with amino acids chemical bonds. The TMA analysis showed that The MOE of soy adhesive increased by addition of tannins to its formulation. Also the presence of covalent bonds formation was confirmed by MALDI-ToF mass spectrometry. The improvement of the mechanical resistance of soy bonded joints by addition of tannin had already been found one year before, but due to the lack of the adequate analytical technique the formation of these bonds and their potentiality was not noticed, being drowned in the formaldehyde-based cross-linking, the improvement being logically ascribed to the presence of formaldehyde cross-linking between the two biomaterials [99].

The results indicated that new compounds did form by reactions of different hydrolysable and condensed polyflavonoid tannin extracts with soy protein amino acids. The interaction between tannin constituents and amino acids occurs by reaction of the carboxyl group (−COOH) of gallic acid of chestnut tannin with the amino groups of amino acids side chains but not with the amide groups in the peptide skeletal chains of the protein. It must be pointed out that only single, unbound amino acid amino groups react but that these amino group once transformed in amide groups in the protein peptide skeletal chain cannot react anymore, at least not easily. Thus, the reaction is then limited only to amino acids in the chains that possess a side chain with an unbound amino or carboxylic acid group. The hydroxyl groups (−OH) on the phenolic rings of quebracho and mimosa tannins also react with amino acid functional groups. As the analysis were carried out at ambient temperature two types of linkages were revealed, namely ionic linkages, but also covalent linkages between the two materials. Fig. 15 shows also cross-linking between peptide chains by ionic and covalent bond formation between tannin extracts and amino acids. The difference between the two forms of each compound were determined by MALDI ToF indicating that in this particular case this analysis technique allowed to distinguish fine differences in the structure of the simpler reaction products [120].

It must be pointed out that while this reactions clearly appeared to occur between soy protein hydrolysates and tannin these were anyhow partially masked by the domination exercised in hardening by the effect of the reactions of hardeners such as formaldehyde and hexamine. As these were considered essential to obtain a good adhesive. Thus, species where an ionic bond is formed between lysine and gallic acid from the hydrolysable tannin such as species I.

In MALDI-ToF analysis a peak at 169 + 147 = 316 Da is observed, and the equivalent peak if one –OH or –NH₂ is protonated is found at 317 or 318 Da. The same effect if Na+ is linked to the species. The second possible reaction: is leading to the formation of a covalent bond. Thus a compound of formula II:

is obtained, that would give 146 + 170 + 23 − 18 = 321 Da if Na+ is considered or would give 298 Da if the compounds has not Na+. These are also been found.

Again in these approaches gelling and hardening was predominated through reactions of the two chemical species with formaldehyde or hexamine, this partially masking the contribution of the direct reaction of tannin and soy.

Finally in a third phase [121] hexamethylenetetramine and glyoxal, two different hardener types, were used with two different commercial tannins, chestnut (a hydrolysable tannin) and mimosa (a condensed flavonoid tannin). The two different hardeners, namely hexamine (with tannins a non-formaldehyde yielding compound) [122–125] and glyoxal, were used to eliminate any formaldehyde while still hardening and upgrading soy resins for particleboard bonding. A number of soy resin formulations have been tested by DSC (differential scanning calorimetry) in the 25°C to 200°C range, and by TMA in the 25°C to 250°C range. The physical properties of different adhesive formulations such as viscosity, solid contents, acidity and density were also measured. One-layer laboratory particleboards bonded with the experimental adhesive formulations were prepared and tested. Based on the results obtained, tannins appear to decrease effectively soy-based adhesives viscosity. The DSC indicated that tannin appears to change the thermal properties of soy flour, decreasing its glass transition temperature Tg and its temperature of denaturation Td. The TMA also indicated that chestnut tannin extract reacts well with soy flour and improves its adhesion properties, and that both types of tannin decreased both the gel time and gel temperature of the adhesive. Furthermore, the IB strength of particleboards bonded with soy adhesives modified with a tannin are higher than for particleboard bonded with unmodified soy adhesive. Equally, marked decreases in water absorption and thickness swelling were recorded for the particleboards bonded with soy adhesives modified by an added tannin, hence a further improvement. In a subsequent paper [126], different amounts of soy flour (SF) and soy protein isolate (SPI) were used with the three tannins. While in this paper the contribution of the direct reaction of a tannin with soy protein hydrolysate or the protein in soy flour or both was demonstrated, the use of an added hexamine as hardener still did not demonstrate that such a reaction could be used alone to bond successfully wood without the addition of an added hardener.

At this stage a breakthrough on the pure reaction of tannins with soy proteins occurred [127], coming from a different direction, namely the work on the reaction of a flavonoid tannin with collagen, to explain some behavior of vegetable tannin in leather tanning and its durability or cracking at higher temperature. Collagen powder hydrolysates were reacted with a solution of commercial mimosa bark tannin extract. The mixture was prepared at ambient temperature and at 80°C to determine what reactions, if any, did occur between the collagen protein through its amino acids and the polyphenolic condensed tannin. The reaction products obtained were analyzed by MALDI ToF mass spectrometry. Reactions between the two materials did appear to occur, with the formation of a relatively small proportion of covalent and ionic linkages at ambient temperature but a considerable proportion of covalent linkages tannin-protein amino acids and the disappearance of ionic bonds. The linkages between the two materials appeared to be by esterification by the –COOH groups of glutamic and aspartic acid of the aliphatic alcohol-OH on the C3 site of the heterocycle of the tannin flavonoid units (III) and by amination of the phenolic –OHs of the tannin by the amino groups of the non-skeletal side chains of arginine (IV). The proportion of covalent linkages increases markedly and predominate with increasing temperatures. This tightening of the tannin-protein covalent network formed was investigated to understand if it was an additional contributing factor both to leather wear resistance and performance as well to leather shrinking when this is subjected to excessive temperatures.

A sequence of aspartic acid-gallocatechin dimer-arginine-leucin-aspartic acid-glycin-glutammic acid (V).

and/or

A sequence of aspartic acid-catechin-arginine-leucine-aspartic acid-glycine-glutamic acid-catechin (VI).

The residual groups in the above amino acids sequence can also react, forming species of higher molecular weight indicating that extensive covalent cross-linking can in reality occur between tannin and protein as for example in structure VII assigned to the peak at 1805 Da, this being an example of how the body of the protein is cross-linked by the tannin

and/or structure VIII

All the above indicated that in the tanning of hides with vegetal tannins a certain number of covalent and ionic bonds do occur between tannin and protein contributing to leather solidity and wear resistance [127]. The proportion of these linkages increases with increasing temperatures of tanning, and in particular the proportion of tannin-protein covalent linkages does increase until these become the only one presents [127]. While such covalent cross-linking may well be a contributing factor to the stability and performance of leather, when this becomes excessive as when leather is subjected to higher temperatures, may well be also a strong additional, contributing factor to leather shrinking. This work however brought the realization that such mechanism may well be active for other proteins too, and after all for soy protein for wood adhesives [127].

For this reason a study of concept was established on the initial application for totally bio adhesives for wood based only on the covalent reaction between SPI and a commercial flavonoid tannin, namely quebracho tannin, without any added hardener [128]. The adhesive was composed exclusively of the two vegetable biomaterials mentioned, thus is totally environment friendly and non-toxic as tannin has been classified as being not at all toxic by the European Commission REACH program [129]. The pre-reaction between the two yielded the best plywood bonding results when limited to a temperature of 40°C, final cross-linking being achieved during the plywood higher temperature hot pressing procedure, as for any other thermosetting adhesive (Fig. 16) [128]. Pre-reaction at higher temperature, as used in the leather work [127] namely 60°C and 80°C, did not work successfully for wood adhesives because its use achieved extensive premature cross-linking that made it lose any activity to cross-link further when hot pressed for bonding plywood (Fig. 16). The reaction was followed by thermomechanical analysis, by MALDI ToF mass spectrometry, and by plywood shear strength testing dry, after 24 h cold water soak and 1 h in boiling water. The adhesive of this approach lends itself to be further reinforced by the multitude of different original approaches on soy resins and adhesives already developed by several other research groups [60–105].

The MALDI ToF analysis showed on top of similar species already identified for the tannin-collagen work [127] also tannin-soy protein isolate species by direct reaction of the flavonoid tannin with short peptide sequences of the protein [128]. Examples of these are the species at 1465 Da of a sequence of glutamic acid-glycine-aspartic acid-robinetenidin-catechin-arginine-leucin-aspartic acid-glycin-glutamic acid, where the peptide sequence is linked to a flavonoid dimer composed of robinetinidin and catechin (IX)

And a species at 1481 Da of a sequence of glutamic acid-glycine-aspartic acid-robinetenidin-gallocatechin--arginine-leucin-aspartic acid-glycin-glutamic acid, where the peptide sequence is linked to a flavonoid dimer composed of robinetinidin and gallocatechin (X).

Transferring what found from wood adhesives and leather under the conditions exposed above on the covalent reaction of tannin and soy protein also the general reaction of a tannin as exposed at beginning of this review will apply. Thus, for gels of several different types, like for example aerogels, not only it is possible as said to obtain a tannin-gel by addition of silica smoke or silicates, but also to obtain the same type of gels with a soy-tannin coreacted material, thus a soy protein-tannin gel by addition of silica smoke, silicates, boric acid and aluminum trioxide [28,41–44] without any formaldehyde. Equally, such a result possibly opens the way to the use of other natural polyphenolic materials, such as lignosulphonates, alkali lignin and organosolv lignin, to form gels of different types, and thermosetting resins and adhesives by reaction with soy protein and soy flour. But this is left to the innumerable research groups concentrating on soy adhesives.

Gels of soy protein and soy flour with tannin are an interesting area of investigation. Soy is a food crop extremely abundant, non-toxic and not expensive while tannin is classed as non-toxic by the European Commision Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) ruling. Numerous works in this area exist and are all based on a number of different approaches. One can distinguish (i) gels by condensation of tannin and soy through methylene linkages generated by reaction of the two biosourced, renewable materials with formaldehyde or other aldehyde. (ii) Gels obtained by autocondensation of the tannin once this is linked by secondary forces to the protein, and (iii) and gels formed by the covalent condensation without added hardeners of the tannin directly with the protein non-skeletal amino et carboxylic acid groups not belonging to the peptide links of the protein. It is this last reaction, already successful in producing plywood adhesives without any other material, but only by tannin reacted covalently with either SPI or Soy flour, that shows potential to even further improve tannin-soy adhesives and resins for a number of application.