Phytic acid oligomers as bio-based crosslinkers for epoxy and polyol resins

The ¹H-, ¹³C-, and ³¹P-NMR spectra were recorded with a Bruker AV NEO 400 NMR (400 MHz) and the samples were dissolved in D₆-DMSO.

IR spectra were measured using a Bruker Vertex 70 (4000-375 cm⁻¹) with ATR and a resolution of 4 cm⁻¹.

The Modular Compact Rheometer MCR 102 from Anton-Paar was utilized to measure relative values of storage and loss modulus and to determine the glass transition temperature after network formation. A single time plate-plate system with a gap of 1 mm and a diameter of 25 mm was used.

The samples were prepared for the rheological measurements using the DAC 150 1.FV Speedmixer from Hausschild GmbH.

The equimolar mass of the synthesized products was determined using conductivity titration and pH-titration with phenolphthalein. The Touch 900 with the 856-conductivity module from Metrohm was used.

About 0.2 g of the product were dissolved in a mixture of water and ethanol (30:70) and titrated with 0.5 mol/L KOH in ethanol. For the pH titration, three drops of 0.1% phenolphthalein in ethanol was added.

In accordance with DIN EN ISO 1522, DIN EN ISO 11664, DIN EN ISO 2813, DIN EN 13523-11, and DIN EN ISO 2409, the application properties of the coatings were measured with the following equipment: BYK spectro guide sphere gloss; BYK micro-TRI-gloss; BYK crosshatch kit (glass) and the Koenigspendel from Erichsen, respectively.

Phytic acid (PA) was obtained from Acros Organics B.V.B.A 50% in water. It was then concentrated to about 80% using a rotary evaporator (40–60°C; 70–30 mBar). For IR analysis, it was further concentrated to 84% (60°C; 30–23 mBar). Dodecan-1-ol ≥98 % was purchased by Roth GmbH. Octane-1,8-diol,n-butyl glycidyl ether, phenolphthalein (1% in ethanol) and the different solvents (toluene, methanol, and butanone) were purchased from Sigma-Aldrich and used without further purification. Also, 0.5 mol/L KOH in ethanol, dibutylphosphate, tributylphosphate and butyl phosphate (mixture of mono and diesters) were purchased from Merck. 3-Phenyl-1-proanol 99% was purchased from Thermo Fischer Scientific. p-Phenylenedimethanol >99% was purchased from Tokyo Chemical Industries. From Allnex GmbH, the commercially available hydroxy functional acrylic resin (Macrynal® SM 500/60X) and the two epoxy systems Beckopox® 387w/52WA and the Beckopox® 301/75X were used. Maprenal® MF800/55IB and MF920/75W was purchased from Prefere Resins Holding GmbH. For the pigment paste experiments, Bayferrox 130 M from LANXESS was used. Glass plates (10 × 20 × 0.3 cm³) were purchased from Rocholl GmbH.

Table 1 shows the theoretical compositions of the diols.

To check the thermal stability, phytic acid was treated for 12 h under reaction conditions analogous to the PA-oligomer 1 synthesis. Furthermore, it was examined whether changes in the phosphorus NMR are visible when concentrating the phytic acid prior to heat treatment.

All formulations were mixed in a 1:1 molar ratio. The formulations are prepared using the Speedmixer. For the rheological measurements, the coatings are applied with a spatula on the single time measurement plate. The temperature was increased by 2°C per min from 20 to 160°C and then held for 30 min. The shear deformation was constant by 0.1% and the frequency by 10 rad/s. All samples except the melamine formulations were within the LVE-range.¹⁸ To measure the glass transition temperature the cured samples are cooled by 2°C/min to 20°C and then reheated to 100–180°C also by 2°C/min. The deformation and stress were the same as by the curing experiments. To determine the room temperature curing of the epoxy resins, the samples were measured at room temperature for 10 to 48 h. In Fig. 3, theoretical reactions with epoxy or hydroxy-functional resins are shown.²¹

For the application testing, the formulations were applied with a 4-sided draw bar 120 µm gap on glass plates. After 15 minutes at 80°C, the samples were then cured at 160°C for 30 min. The application tests were then carried out in accordance with the relevant DIN norms.

Pigment pastes were prepared with Bayferrox 130 M and PA oligomers 3 and 4. In each case, an aqueous pigment paste and one based on butyl acetate were prepared. For this purpose, the oligomer was first dispersed in the aqueous phase or dissolved in the butyl acetate. The pigment was then incorporated in a speed mixer (3000 rpm; 1 min) to form a concentration of 66% pigment to 16% PA-oligomer. The pigment pastes were then prepared with the respective epoxy binders and crosslinked either at ambient temperature or by the influence of heat.

The purchased butyl phosphates were analyzed with IR spectroscopy and ³¹P-NMR spectroscopy.

Figure 4 shows the phosphorus NMR of phytic acid 50% and 81% after 12h under reaction conditions. The ppm shift from near zero to ~− 0.72 ppm can be explained by the difference in pH. While 50% phytic acid has a pH value of 0.4, 81% has a pH of − 0.2. The pH after the reaction could not be measured because the resulting phytic acid was a slightly yellowish, highly hygroscopic, crystalline substance. The shift from higher to lower ppm due to increase in acidity is a well-known phenomenon in phosphorus NMR.²² In addition, the spectra do not show different P species. Conductivity titration of phytic acid after 12 hours under reaction conditions shows that a concentration of >90% phytic acid can be achieved by this method. After the separation of the phases, we found about 0.5% of unknown chemicals in the toluene phase. These molecules could be a natural pollutant from the extraction process of phytic acid or generated products through the thermal destruction of the phytic acid molecule. Therefore, we can conclude that phytic acid is mostly thermally stable under these conditions. The IR spectra of both phases are shown in the supplemental material (Fig. S1). The solid product corresponds to the concentrated phytic acid spectra.

In this work a simple method to create phytic acid-oligomers through esterification is presented. The structure cannot be 100% determined, because of the various possibilities to form linear or cyclic structures, with different amount of mono-, di- and tri-phosphate esters. But our results prove that a successful esterification with this method is possible.

In the following IR spectra (Fig. 5), a significant broadening of the sensitive P=O vibration (1150–1257 cm⁻¹) is clearly visible. The shift and broadening are an indication that the transformation from mono to di-/triester is successful. Furthermore, if the IR spectra of dibutyl- and tributylphosphate (Fig. S2) are compared, the same shift is visible. The reduced intensity of the three different POH board vibrations is also an indication that the amount of free POH is greatly reduced. Typical hydroxy vibrations from unreacted dodecanol and hydrolyzed phytic acid are not visible for the oligomer synthesis, which is a good indicator that a high conversion is achieved. The same effects are also observed with the other unmodified oligomers (Fig. S3). However, the IR spectrum of the PA-ester 1 still shows a high proportion of free alcohol, which indicates poor conversion.

The comparison of the different ³¹P-NMR spectra (Fig. 6) of the products shows that we can identify different species of phosphorous atoms. The shift of the signals could be caused by a slightly different pH value. The samples were weighed without a neutralization step and therefore contain different numbers of free POH groups. What is clearly recognizable, however, is that we detect only one signal for phytic acid, two signals for the theoretical PA ester, and four signals for the oligomer. This is in line with our expectations, since in the case of phytic acid only monoesters should be present in the case of PA-ester 1 mono and diester and in the case of PA-oligomer mono-, di- and theoretical triester. The additional signal around − 13.3 ppm illustrates that a small number of poly- or pyrophosphates are formed through the reaction. PA-oligomer 3 and 4 display the same effect in the figure, but only oligomer 2 shows a small amount of unreacted the polyphosphate (Fig. S4). The relatively small signal spacing between the different esters is also reflected in the NMR measurements of the different butyl phosphate esters (Fig. S5).

The ¹H-NMR spectrum (Fig. 7) of dodecanol (top) shows that the signals 3 and 5 are significantly influenced through the hydroxy functionality. Therefore, both could be used to check if a reaction occurs. After the reaction, signal 3 shows a shift of +0.13 ppm and signal 5 of +0.42 ppm. The signal shift to higher ppm is typical for an esterification.¹² We hypothesize that the broadening and multiplet formation of the newly generated signal 6 is produced by the formation of a multitude of slightly different esters. Additionally, the phytic acid C–H signals have nearly the same shift, which results in a poor evaluability of this signal group.

To calculate the conversation rate of the alcohols signals 3 and 4 were compared. The conversion of the alcohols was significantly higher (93%) for the oligomer than for the PA-ester 1 (~24%).

If we compare the spectra of PA-oligomers 3 and 4 (Fig. S6), we can see that the line broadening is much stronger than in the syntheses with dodecanol. One possible reason for this is that the alcohols used have a significantly shorter carbon chain, and the repetition of the phenyl component produces a line broadening effect similar to that of the repeating units of polymers.²³ The conversion was determined from the signal of the second methylene group between the unreacted and the converted phenyl propanol. In the case of PA-oligomer 4, no more free alcohol could be detected, so that a conversion close to 100% is assumed.

To further prove that this was a true esterification and not a transesterification between dodecanol and phytic acid, a conductometric titration was performed. The resulting equimolar weight should be increased in comparison to the equimolar weight of pure phytic acid. PA-oligomer 1 has an equivalent weight of 252 g/eq (Fig. 8), which is about 97% of the theoretically calculated value of 259 g/eq. To prove that a further reaction with epoxy was also successful, the equimolar weight of PA-oligomer 2 was also determined with a value of 707.4 g/eq. The amount of free unreacted alcohol was included in the calculation. An equivalent mass of 86.25 g/eq was determined for the dry product of PA-ester 1. This corresponds to approximately 4 units of lauryl alcohol per phytic acid.

We assume that the reaction proceeds in two repetitive steps (Fig. 9). First, an intra- or intermolecular phosphorus anhydride is formed by the elimination of water. This step is probably reversible and the newly formed anhydride is decomposed if water is present in the environment. The next step ocurrs at the toluene, phytic acid surface. The alcohol functionality attacks the phosphorus atom, leading to multiple electron transfers and finally to the formation of the new ester. The newly formed PA ester has a typical surfactant structure and the weak to strong foam formation during the reaction is further evidence that the propsed molecule is formed. In the next stage of the reaction it is important, if difunctional alcohol is avaiable or not. We suggest that the high difference of the conversion rates between PA-ester and PA-oligomer could be explained through the formation of micelles. Unreacted phytic acid is trapped in the micelles and cannot react further. Coalescence of the micelles and evaporation of the water results in various phytic acid crystal formations (Fig. S7). During this process, hollow structures can form which are filled with toluene and unreacted alcohol. The crystals are extremely hydrophilic and only for a few minutes stable, if exposed to air. On the other side, if difunctional alcohol is present, the foam formation is reduced over time and instead of a significant crystal phase a clear toluene/product phase is generated. A possible explanation for this effect could be the prevention of micelle formation by the diol components. The reaction with the diol and further reaction with a second phytic acid leads to a significant enlargement of the molecule, which is presumably sterically no longer able to form effective compact micelles and is thus unable to protect free phytic acid from unreacted alcohol.

The curing of the various samples was observed using oscillation measurements (Figs. 10, 11, 12; Fig. S8, 9). All samples show a strong and stable increase of the storage modulus which is classical for network formations. Since a plate-plate system was used, the solvent of the samples could not evaporate easily. This results in a unusual progression at the beginning of the different network formation. The gel point is defined by the build-up of higher structures in the polymer matrices and its value is determined by the point at which the storage modulus exceeds the loss modulus. It is also the minimum temperature required for curing. A second indicator for the successful curing is the glass transition temperature. The glass transition temperature is determined using the maximum of the loss modulus. All measurable samples show typical behavior of polymer networks before and after the glass temperature (Fig. S10). Only the sample with Macrynal® SM500/60X shows a small rubbery elastic region shortly after the glass transition temperature. The melamine samples could not be measured without destruction of the internal structures (sample exceeds LVE-range) therefore, a glass transition temperature could not be determined. Nevertheless, a successful network formation can be assumed since the storage modulus rises significantly above the loss modulus before exceeding the LVE.²⁴ To further demonstrate network formation, extraction experiments were performed with water, methanol, and butanone, in which phytic acid could not be extracted. This proves that the oligomers not only act as a catalyst but are also a part of the newly created network. Additionally, network formation experiments by ambient temperature with bisphenol A digylcidyl ether and PA-oligomer 3 and 4 were successfully conducted (Fig. 13; Fig. S.11).

For application testing the products were equimolar mixed and with a 120 µm 4-sided applicator frame applied to glass plates and subsequently cured. To improve the surface of the water-based systems the PA-oligomer 1 was dispersed in 50% water and then mixed with the water-based systems. In Table 1 the application data are summarized. The theoretically high network density due to the high functionality of the phytic acid oligomers is reflected in the high pendulum hardness. The strong difference between PA-oligomer 1, 3, and 4 could be explained through the different building blocks. The glass transition temperature across the different oligomers is relatively similar, which could be explained by the relatively large binder systems compared to the different PA oligomers, which thus have a higher influence on the glass transition temperature (Table 2). The relatively low glass transition temperature of Beckopox 387w/52WA and Macrynal can be explained by the water present or formed in the respective samples. Due to the plate-plate geometry of the rheometer, this water cannot evaporate and thus interferes with the crosslinking reaction. This leads to an apparently lower Tg than with a sample where the water can evaporate freely.²⁵ The gel point in this test is strongly influenced by the solvent since outgassing at the rheometer is hindered. The actual gel point of the epoxies is therefore probably significantly higher, especially for the water-based systems. The slightly poorer adhesion of oligomer 1 could be explained by the dodecanol carbon chains in the epoxies, which provide lower interaction forces with the polar surface compared to the shorter aromatic alcohols of oligomers 2 and 3.²⁶ In addition, the steric of PA oligomer 1 and the reaction mechanism of crosslinking in the case of OH acrylate could also have an influence on the adhesion strength. The MEK stability tests show that the epoxy systems crosslinked with the oligomers exhibit excellent solvent stability. In contrast, the crosslinked OH acrylates exhibit significantly poorer to virtually no solvent stability. This indicates a very low degree of crosslinking, which can be explained on the one hand by the steric nature of oligomer 1 and on the other hand by the crosslinking reaction. Crosslinking by esterification has the disadvantage that the water formed cannot always escape completely from the film. This can lead to a lower conversion or even to the cleavage of the ester bonds. Overall, the data show that the oligomers are suitable for the crosslinking of industrial OH acrylate and epoxy resins and, especially in the case of epoxides, for the preparation of coatings with good application properties.

No optimization of the pigment paste has yet been carried out in the tests, which are merely intended to serve as proof-of-concept as to whether the PA-oligomer can fulfill the functions of a hardener and dispersant at the same time. The pigment pastes are stable for at least 3 weeks.

The color measurements of the dried pigment pastes in Table 3 show that the difference between both colors is relatively small with a ΔE value of 1.7 and 1.2, respectively; only trained eyes will notice a difference. However, this is only a proof of concept since all pastes were dispersed under the same relatively simple conditions.

Table 4 shows measurement results of the cured coatings. The water-based systems have a lower pendulum hardness in both room temperature drying and thermal curing, which could indicate that a reaction between water and a part of the epoxy binder occurs. The gloss value of the air-dried coatings is in most samples higher than the cured samples, which is probably because no solvent mixture was used to optimize the curing process. As a result, the solvent escapes too quickly during curing, resulting in a rougher surface. The similar L*a*b* values of the experiments indicate that the water- and solvent-based pigment pastes are similar in their dispersion properties. Overall, the creation of a pigment paste and the crosslinking could be successfully carried out with the PA-oligomers.