Sustainable Soil Additives for Water and Micronutrient Supply: Swelling and Chelating Properties of Polyaspartic Acid Hydrogels Utilizing Newly Developed Crosslinkers

Besides the diamines HMD and LYS, bifunctional glycine derivatives of EG and DEG were successfully synthesized by classical-acid-catalyzed, azeotropic esterification using p-toluenesulfonic acid as a catalyst, resulting in the compounds di-p-toluenesulfonic acid bis-(glycine)-diethylene glycol ester (TosGLY₂DEG) and di-p-toluenesulfonic acid bis-(glycine)-ethylene glycol ester (TosGLY₂EG). Scheme 2 shows the reaction scheme between glycine and (1) DEG and (2) EG.

¹H-NMR spectra in Figure 1 show a successful synthesis of each glycine-based crosslinker in its tosylate form. The crosslinker was used without further purification for the next reaction step to form PASP-based hydrogels. In the subsequent crosslinking reaction, triethylamine (TEA) was added to form a salt with p-toluenesulfonic acid and thereby provide the reactive free amine groups of glycine for crosslinking of PSI chains.

PSI was successfully synthesized according to two different synthesis methods. PSI-1 was synthesized from maleic anhydride (MA) and urea in the presence of a mixture of sulfuric (VI) acid (H₂SO₄) and orthophosphoric (V) acid (H₃PO₄) [27]; PSI-2 was prepared in the presence of H₃PO₄ [28]. Both syntheses were carried out at 180 °C under reduced pressure. The yield of PSI samples from the two different syntheses was determined. Samples were characterized by GPC and FTIR for a structural analysis. Table 1 and Figure 2 show the results.

The molecular weight obtained from using MA and urea as monomers led to much lower values. The FTIR spectra of both synthesized PSI samples are shown in Figure 2b. The absorption bands at 1709 cm⁻¹ and 1794 cm⁻¹ correspond to ν(C=O) of succinimide. The appearance of both bands is explained by the two carbonyl groups in a close neighborhood in the PSI structure [29], which can be easily recognized in the PSI synthesized from ASP. The PSI sample synthesized from MA and urea shows these bands as a shoulder of the absorption at 1680 cm⁻¹, which corresponds to the C=O-stretching vibration of an amide bond (amide I) together with the amide II peak at about 1600 cm⁻¹. This indicates the presence of branched or ring-open(ed) structures. In a first step, aspartic acid is formed from the reaction between maleic anhydride/maleic acid and urea [30]. The main difference in the subsequent polymerization reaction compared to the procedure from ASP with H₃PO₄ is in this case the lower amount of acid during polycondensation: a ratio of acids to ASP is 1:2 (“ASP”) or 1:4 (“MA + urea”), respectively. This can cause ASP units to more preferably react with succinimide units undergoing aminolysis and ring-opening, leading to a branched structure. GPC data show that this reaction procedure also leads to shorter polymer chains. An advantage is generally a more homogeneous mixture due to the mixing of all reactants in the aqueous phase. The synthesis from ASP in combination with a high amount of phosphoric acid results in linear PSI chains with a high molecular weight as reported in study [28]. PSI synthesized from MA and urea was not able to form hydrogels after crosslinking. The short polymer chains and reduced reactive moieties due to a branched structure with less succinimide units prevent this material from being a feasible polymer to prepare PASP-based hydrogels. Therefore, PSI-2 was used in all hydrogel syntheses, and discussed hydrogels are based on this polymer.

The FTIR spectra of linear PASP and the different crosslinked clPASP gels are shown in Figure 3. The disappearance of the absorption band at about 1710 cm⁻¹ corresponding to the C=O-stretching vibration of the succinimide structure and appearance of the C=O-stretching vibration of the carboxylate salt (COO⁻Na⁺ units from hydrolyzed polymer) at about 1580 cm⁻¹ alongside an observable shoulder at about 1650 cm⁻¹ (amide) show successful hydrolysis to PASP. Furthermore, the increase in the amide bound at about 1650 cm⁻¹ compared to the one of the carboxylate salt shows an increase in amide groups. This indicates an increasing crosslinker amount in the hydrogels [31]. Hydrolysis of PSI units leads to the presence of one amide group and one unit of carboxylic acid salt, while crosslinking leads to one amide present in the PASP chain and one amide forming the crosslink to another chain. Therefore, the increased absorption correlates with a higher crosslinker amount. The results show that in all cases, an increasing crosslinker amount added in the reaction process leads to an increased amount in the hydrogels. It seems that more HMD is incorporated compared to LYS as the FTIR results indicate. This was also observed in much faster hydrogel formation using HMD compared to LYS. Upon crosslinking with the GLY derivatives, additionally, the C=O absorption of the ester of the crosslinker molecules can be seen in the spectra.

The effect of the crosslinker type and amount on the swelling properties of the PASP hydrogels was determined. Swelling degrees were measured as time-dependent in distilled water and subsequently fitted according to the Voigt model [32]. AUL was measured by applying 0.1 and 0.3 psi while swelling in a self-3D-printed PLA tube.

Figure 4 shows the experimentally determined swelling degrees dependent on swelling time in distilled water of clPASP crosslinked with HMD, LYS, and GLY₂(D)EG as well as the in situ-crosslinked and Cu/Zn-loaded hydrogels, respectively.

Depending on the crosslinker type and amount, generally, swelling degrees between 1000 and 9000% were reached. Hydrogels prepared with in situ Cu/Zn-loading both have a comparable swelling degree of about 5800% (LYS20.0(Cu)) and 6200% (LYS20.0(Zn)). In a comparison between HMD- and LYS-crosslinked gels, the ones with HMD obtained a higher swelling degree despite the presumed higher crosslinking (see Figure 4). It seems that LYS decreases water uptake compared to HMD, while an increase in LYS crosslinking and modification leads to an increase in water uptake. Via fitting regarding the Voigt model, the parameters S_max and t₆₃ were determined as shown in Table 3.

Crosslinked hydrogels with HMD show a decrease in S_max with an increase in crosslinkers alongside a decrease in t₆₃, showing a faster uptake of water, which is expected [33]. Nevertheless, gels crosslinked with LYS and GLY₂(D)EG showed a higher S_max with an increase in crosslinkers. The value of t₆₃ decreases slightly as observed for HMD-crosslinked hydrogels, which indicate a higher crosslinking density nevertheless. An explanation could be an incomplete crosslinking when using these crosslinkers. Leaving one of the two amino groups of the crosslinker pendant on the polymer chains leads to a derivatization of the hydrogels. This could contribute to a higher water uptake. Generally, crosslinker incorporation using GLY₂(D)EG was less efficient compared to HMD and LYS, because no hydrogels were formed at all using a 5% crosslinker. Using LYS, crosslinking times were much higher compared to HMD. Both indicate less reactivity and therefore a derivatization alongside the crosslinking reaction, leaving pendant groups in the polymer network after hydrolysis. In the case of LYS, this provides an additional amine and carboxylic acid group. Their hydrophilicity could explain the increasing water uptake with an increasing addition of LYS while crosslinking.

The swelling equilibrium under load S_AUL was measured after 2 h, applying 0.1 and 0.3 psi weight, respectively, resulting in the values S_{(0.1 psi; 2 h)} and S_{(0.3 psi; 2 h)}. Figure 5 shows the results of the measurements. Additionally to the clPASP hydrogels, the polyacrylate-based hydrogel Stockosorb^® (Evonik, Essen, Germany) was tested in AUL measurements.

Under load, the water uptake decreases drastically, depending on the different hydrogel–crosslinker systems. However, this effect is generally less prominent in the case of PASP hydrogels with different crosslinkers and crosslinker content in comparison to a Stockosorb^® PAA hydrogel. Under 0.3 psi, various PASP hydrogels even had a higher swelling degree than the PAA gel. HMD-crosslinked hydrogels show a linear increase in residual swelling under load compared to free swelling with an increasing amount of HMD. Higher HMD contents lead to less reduction in water uptake under load compared to free swelling. Nevertheless, the absolute S_AUL is still higher with less crosslinkers. This relation is not strictly the case for the other prepared PASP hydrogels. This could be explained by the less reactive crosslinkers LYS and (D)EG derivatives compared to HMD. The crosslinking and functionalization of hydrogels happen in parallel and the hydrogel chemistry is changed additionally to a crosslinking of the polymer chains. This seems to affect AUL and free swelling. Generally, we can conclude that water uptake of the synthesized PASP-based hydrogels is less sensitive under load. The samples LYS10.0/20.0 and GLY₂DEG20.0 especially show good swelling properties under load. A possibility to improve AUL of hydrogels is surface crosslinking of the gel particles. Surface crosslinking and a generally denser polymer network decrease S_max but lead to higher AUL [34]. Ashkani et al., synthesized hydrogels from different crosslinked polymers. A hydrogel based on poly(2-acrylamido-2-methyl-propane sulfonic acid) showed the best AUL compared to other gels based on PAA and polyacrylamide [35]. The presence of the methyl propane sulfonic acid groups in this polymer as bulky side groups is one structural difference between the studied hydrogels and could contribute to this effect. As discussed before, LYS or (D)EG derivatives could be less present as crosslinkers but also be incorporated as pendant groups in the gel network. Therefore, they act as bulky side groups as well, which could explain the improved AUL.

Hydrogels were loaded with the micronutrients copper or zinc, respectively. Two approaches were carried out. As already shown in Section 2.2.2, the two metal ions were included in the crosslinking process with LYS, which was used, on the one hand, as a crosslinker for PSI and, on the other hand, it should act as a chelating agent for Cu(II) and Zn(II). In this way, copper-loaded PSI was synthesized in one step during crosslinking. Subsequent hydrolysis led to the corresponding hydrogels. Swelling properties of such gels are already shown (Figure 4 and Figure 5 and Table 3). A second approach was the immersion of previously synthesized HMD-crosslinked hydrogels as well as linear PASP in the ion solutions of copper or zinc. Table 4 summarizes samples prepared in such a way.

Swelling of hydrogels in these solutions as well as swelling ratios of the resulting loaded hydrogels were measured. Figure 6 shows the HMD5.0 hydrogel and linear PASP mixtures with different concentrated Cu(II) or Zn(II) solutions.

Linear PASP immersed in metal salt solutions had a low water uptake with 477% being the highest value in this series (HMD5.0 in 0.025 mol/L Zn(II) solution) and was not considered further. Among the loaded PASP hydrogels, however, various samples maintained a considerable swelling degree. A higher concentration of metal ions leads to not or slightly swollen solids, while a decrease in metal ion concentration leads to swelling of the hydrogels in the solution. This can be explained by two reasons. Water uptake in higher-concentrated salt solutions is lower due to osmotic effects. Additionally, divalent ions such as Cu(II) and Zn(II) can contribute to a higher ionic crosslinking due to ionic interactions and complexation in PASP [36], which leads to less water uptake. Upon the addition of linear PASP to the ion solutions, a precipitate was observed compared to a clear solution of PASP in water. The insoluble polymer is explained by its dense ionic crosslinking due to high ion concentrations. PASP can act as a complexation agent for divalent ions such as Cu(II) or Zn(II) [36]. All loaded, washed, and dried polymers were immersed in deionized water to determine the swelling degrees after 2 h. Figure 7 shows the results of the swelling degrees of the gels. The linear PASP, precipitated in ion solutions shows no swelling. Loading HMD5.0 in a 0.0125 molar zinc solution leads to a hydrogel with a swelling degree of about 8000%, and the corresponding gel with copper reaches 2621%. Compared to other reported gels based on CMC or κ-carrageenan, a high water uptake was reached when loaded with zinc or copper ions [26,37]. Nevertheless, hydrogels based on carboxymethyl tamarind kernel gum loaded with zinc could be prepared with very high water uptake [38]. PASP, however, is a material with many possibilities of modification especially in its intermediate form PSI via aminolysis [15]. The incorporation of pendant groups to the network might not only tailor water uptake under different conditions, but could also introduce specific functionalities for chelating of (micro-)nutrient ions.

Figure 8 shows the swelling properties of the in situ-loaded LYS20.0(Cu/Zn) in comparison to the solely LYS-crosslinked counterpart as well as the HMD5.0 + Cu/Zn hydrogels with their unloaded counterpart in distilled water. Kinetic parameters are compared in Table 5.

In situ LYS-crosslinked and loaded hydrogels show similar swelling properties compared to the solely LYS-crosslinked gels. Loading with copper ions decreases the water uptake. S_max of subsequent loaded hydrogels differs more from the pure gel. In the case of Zn(II) loading, an increase is observed. Loading with copper (II) leads to a decrease in S_max. This correlates with the data shown in Figure 7. Copper ions have a higher binding affinity to the PASP polymer via complexation [36].

Additionally, UV-VIS spectra were recorded to monitor the copper species inside the hydrogel structure. Cu(II) could be present as precipitated or absorbed sulfate or as a chelate complex with the aspartic acid units, with LYS, or with both. Figure 9 shows the recorded spectra.

The results give no clear indication if copper is bound via complexation with LYS or ASP units. However, the spectra show that copper is not solely immobilized in the gel as dissolved or precipitated sulfate. A shift of the broad absorption of CuSO₄ at 816 nm is observed in all samples. This indicates that copper salts are bound to the hydrogel in different ways, allowing for a broad release profile of micronutrients.

Figure 10 shows the SEM images taken exemplarily from an HMD5.0 hydrogel sample before and after swelling. SEM images of swollen and freeze-dried hydrogel show the presence of pores dispersed throughout the hydrogel. These pores were created when water entered the interior of the hydrogel, which ensure that HMD5.0 has excellent water absorption capacity [39,40].

As an important feature of the hydrogels, biodegradability should be ensured. PASP and its crosslinked gels are reported to be biodegradable. In this study, HMD5.0 was exemplarily investigated regarding its biodegradability in sewage sludge and soil. In comparison, the PAA product Stockosorb^® and sodium alginate as controls were measured. Therefore, the biological oxygen demand (BOD) as an indicator for microbial activity was determined over time for the mentioned polymers in both systems (sewage sludge and soil). To determine the activity, the BOD of a control, e.g., sewage sludge, and soil without hydrogel was measured. Figure 11 shows the various measured BODs (vs. control) dependent on time.

A BOD below zero signifies less activity than the control, and a positive BOD signifies more degradation. Therefore, higher BOD values indicate higher biodegradation. In sewage sludge, all three hydrogels seem to show a lag phase of two days, until the degradation starts. This lag could be due to a change in salinity because of the swelling of the gels, leading to an increase in salt concentration. As shown above, alginate is very easily biodegradable, whereas crosslinked PASP (HMD5.0) has a lower biodegradability. Crosslinked PAA seems to be not biodegradable. Furthermore, the data indicate that PAA inhibits microbial activity in sludge since it has a lower BOD than sludge without hydrogels. Figure 11 also shows the degradation of hydrogels in soil. The graphs indicate that sodium alginate and HMD5.0 are better degradable than PAA. Nevertheless, the data show some degradability of PAA in soil. What is striking is the decline in BOD for HMD5.0 after 9 d. It signifies that other processes generate pressure/gas, other than carbon dioxide. This influences the test system. Nitrification and denitrification processes could cause the release of nitrogen and therefore a rising pressure in the flasks, which affects the measured values.

Since no inhibition chemical for nitrification was added, this could be the reason for the high BOD of PAA. In future experiments, this will be determined and biodegradation tests will be optimized for hydrogels.

Di-p-toluenesulfonate of bis-(glycine)-diol ester (GLY₂DEG or GLY₂EG, respectively) crosslinker: The products were synthesized according to the following protocol. In a round-bottom flask, equipped with a Dean–Stark trap and condenser, glycine (0.028 mol), diol (DEG or EG) (0.13 mol), and p-toluenesulfonic acid monohydrate (0.030 mol) were dissolved in 60 mL of toluene by stirring. The reaction mixture was heated under reflux for 10 h until no more water was separated. The excess of toluene was evaporated under reduced pressure and the resulting products were dissolved in methanol. Unreacted glycine was filtered off and the crude product was recrystallized from an isopropanol/ethyl acetate mixture [41].

¹H-NMR spectra (D₂O, 60 MHz, δ [ppm]):

(1)

7.56–7.21 (d, 8H, 2Ar-H), 4.34 (t, 4H, COOCH₂CH₂), 3.88 (s, 4H, NH₃⁺CH₂), 3.75 (t, 4H, COOCH₂CH₂), and 2.31 (s, 6H, Ar-CH₃).

(2)

7.61–7.22 (d, 8H, 2Ar-H), 4.44 (s, 4H, COOCH₂), 3.89 (s, 4H, NH₃⁺CH₂), and 3.75 and 2.32 (s, 6H, Ar-CH₃).

Synthesis of PSI from MA and urea catalyzed with phosphoric and sulfuric acid mixture (PSI-1): MA (9.8 g, 0.100 mol) was first reacted in water at 80 °C for 40 min to obtain maleic acid. Afterwards, urea (4.2 g, 0.070 mol) and H₂SO₄/H₃PO₄ (0.91 g/0.84 g, 0.009 mol/0.009 mol) were added and the reaction was carried out in a vacuum at 180 °C for two hours. The product was dissolved in DMF (10 mL per 1.0 g) and precipitated in water (30 mL at room temperature) in order to remove acid and unreacted residues. PSI was washed with water until neutral, filtered, and dried at 50 °C in a convection oven for 12 h. FTIR (ATR) measurements and a GPC analysis (N,N-dimethylacetamide with 0.3% LiBr, PMMA calibration) were carried out.

Synthesis of PSI from ASP using phosphoric acid (PSI-2): ASP (10.0 g, 0.075 mol) was mixed with H₃PO₄ (4.30 g, 0.037 mol) at room temperature to form a homogeneous mixture. The mixture was placed in a flask connected to a vacuum pump after being heated to 180 °C. Pressure was about 20 mbar during 3 h of reaction time. Work-up and characterization were performed as described for PSI-1.

Crosslinking with HMD: PSI (2.0 g, 0.021 mmol) was dissolved in DMF. HMD was added at room temperature as a 10 wt.% solution in ethanol. After approx. 15–25 min, an elastic gel was formed. After a total of 2.5 h, the gel was crushed, washed thoroughly with ethanol, and dried in a vacuum at 100 °C for 12 h in order to remove ethanol and any solvent residues of DMF. Table 6 summarizes reactant ratios and obtained product yield.

Crosslinking by LYS: PSI (2.0 g, 0.021 mol) was dissolved in 10 mL of DMSO. LYS was added at room temperature as a 10 wt.% solution in methanol. The solution was heated to 80 °C and stirred for an additional 24 h until an elastic gel was formed. Afterwards, the gel was crushed, washed thoroughly with ethanol, and dried in a vacuum at 100 °C for 12 h. Table 7 summarizes reactant ratios and obtained product yield.

Crosslinking by GLY₂(D)EG: PSI (2.0 g, 0.021 mol) was dissolved in 10 mL of DMF. Bis(glycine)-diol (DEG or EG) was diluted with 4.0 mL of DMF before addition to the PSI solution. Afterwards, TEA was added dropwise at room temperature to form a homogeneous mixture. After a total reaction time of 24 h at 80 °C, an elastic gel was formed. The gel was crushed, washed thoroughly with ethanol, and dried in a vacuum at 100 °C for 12 h. Table 8 summarizes the reactant ratios and obtained product yield.

Subsequent loading of hydrogels with micronutrients Cu/Zn via swelling in ion solutions: An amount of 50 +/− 1 mg of the dry hydrogel sample and linear PASP was placed in 10.0 mL of 0.100, 0.050, 0.025, and 0.0125 mol/L solutions of CuSO₄ or (Zn(Ac)₂). After 2 h, the gels were filtered, washed, and dried at 50 °C. Swelling degrees were determined. From clPASP-HMD5.0 and clPASP-LYS20.0 loaded in the 0.0125 mol/L solution of zinc(II) and copper(II), respectively, swelling kinetics were measured as described below.

In situ loading of hydrogels with micronutrients Cu/Zn: PSI (1.0 g, 0.010 mol) was dissolved in 28.0 mL of DMF, and a solution of LYS (0.64 g, 0.004 mol, in 8.0 mL of water) was added. Afterwards, the mixture was stirred thoroughly at 60 °C for 1 h. TEA (629 μL, 0.004 mol) and CuSO₄ (0.32 g, 0.021 mol) or Zn(Ac)₂ (0.38 g, 0.021 mol), respectively, were mixed in 4.0 mL of water, diluted with 2.0 mL of DMSO, and afterwards added to the PSI solution. After addition, the mixture was stirred thoroughly at 80 °C for 3 h. The work-up was performed as described in Section 4.2.3.

To a 1.0 molar NaOH solution, clPSI samples were added while stirring vigorously. Directly, crosslinked polyaspartic acid (clPASP) was formed, which can be recognized by swelling of the solid and the rapid absorption of the NaOH solution. The mixture was reacted for 2 h before the product was washed with water (200 mL) and ethanol (40 mL) and dried at 50 °C. Table 9 summarizes the used amounts and ratios of clPSI and NaOH and obtained product yield of clPASP hydrogels.