*Article* **Optimising Soy and Pea Protein Gelation to Obtain Hydrogels Intended as Precursors of Food-Grade Dried Porous Materials**

**Lorenzo De Berardinis, Stella Plazzotta and Lara Manzocco \***

Department of Agricultural, Food, Environmental and Animal Sciences, University of Udine, 33100 Udine, Italy **\*** Correspondence: lara.manzocco@uniud.it

**Abstract:** Dried porous materials based on plant proteins are attracting large attention thanks to their potential use as sustainable food ingredients. Nevertheless, plant proteins present lower gelling properties than animal ones. Plant protein gelling could be improved by optimising gelation conditions by acting on protein concentration, pH, and ionic strength. This work aimed to systematically study the effect of these factors on the gelation behaviour of soy and pea protein isolates. Protein suspensions having different concentrations (10, 15, and 20% w/w), pH (3.0, 4.5, 7.0), and ionic strength (IS, 0.0, 0.6, 1.5 M) were heat-treated (95 ◦C for 15 min) and characterised for rheological properties and physical stability. Strong hydrogels having an elastic modulus (G0 ) higher than 10<sup>3</sup> Pa and able to retain more than 90% water were only obtained from suspensions containing at least 15% soy protein, far from the isoelectric point and at an IS above 0.6 M. By contrast, pea protein gelation was achieved only at a high concentration (20%), and always resulted in weak gels, which showed increasing G0 with the increase in pH and IS. Results were rationalised into a map identifying the gelation conditions to modulate the rheological properties of soy and pea protein hydrogels, for their subsequent conversion into xerogels, cryogels, and aerogels.

**Keywords:** plant proteins; heat gelation; gelling behaviour; structure; pH

### **1. Introduction**

Xerogels, cryogels, and aerogels indicate dry porous materials produced by removing the solvent from a gel. Most studies have been carried out on the development of inorganic dried porous materials (e.g., silica and carbon-based) [1–3] to be used in a wide variety of applications, such as catalysis, environmental remediation, energy storage, and insulation [4–7]. Nevertheless, in recent years, growing interest has been focused on the development of biopolymeric-based dried porous templates, due to their biocompatibility, and non-toxic profile. Thanks to these characteristics, their application has been successfully extended to life science fields, including the biomedical and pharmaceutical sectors [8–10]. The potentialities of dried porous materials in the food sector are nowadays attracting large attention, due to their unique physico-chemical properties and techno-functionalities. Both cryogels and aerogels have been suggested as innovative delivery systems to protect bioactives and flavours during processing, storage, and digestion [11–16]. In addition, their capacity to absorb large amounts of food solvents has been identified as a key feature to modulate food structural properties [17,18]. For instance, they have been suggested as templates for oil structuring, leading to fat replacers with improved nutritional properties [16,19–21]. By contrast, as concerns xerogels, to the best of our knowledge, no applications in the food sector have been reported, despite the high potentialities of these materials have been demonstrated in other life science sectors.

To produce food-grade dried porous material, an aqueous gel is first produced by inducing the networking of the selected biopolymer in water, leading to a hydrogel [22]. To obtain a xerogel, subsequently, water is removed from the network by evaporative drying. The latter can also be performed by evaporating ethanol after substituting hydrogel

**Citation:** De Berardinis, L.; Plazzotta, S.; Manzocco, L. Optimising Soy and Pea Protein Gelation to Obtain Hydrogels Intended as Precursors of Food-Grade Dried Porous Materials. *Gels* **2023**, *9*, 62. https://doi.org/ 10.3390/gels9010062

Academic Editors: Francesco Caridi, Giuseppe Paladini and Andrea Fiorati

Received: 15 December 2022 Revised: 4 January 2023 Accepted: 10 January 2023 Published: 12 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

water with ethanol [23,24]. The evaporative drying usually induces capillary forces during solvent removal, so that xerogels present low porosity [25]. Cryogels are instead obtained through freeze-drying, and thus by water sublimation [18]. This reduces the capillary forces, leading to materials with large pores and channels left upon the sublimation of water crystals grown during freezing [26]. Finally, aerogels are obtained by replacing the water contained in the starting gel with ethanol, followed by ethanol removal with a flow of CO<sup>2</sup> in the supercritical state [27]. This technique preserves the structure of the initial network, and the dried material is thus characterised by low density and high internal surface area, due to the presence of micro- and macropores [28].

Food-grade xerogels, cryogels, and aerogels can be prepared either from polysaccharides or proteins. As concerns proteins, most literature studies focus on animal ones (e.g., whey, egg white, casein, gelatin) [29,30], while studies on the development of dried porous templates from plant proteins are limited to a few works exploiting silk fibroin, patatins, and soy proteins [31–37]. The interest in plant-based products is constantly growing due to their low environmental impact, low cost, and the possibility of being obtained from food industry wastes, in a circular economy perspective [38–40]. For these reasons, plant proteins are ideal candidates for developing sustainable dried porous materials intended as innovative ingredients for the food sector. However, the production of plant-based xerogels, cryogels, and aerogels is rather challenging. This is mainly due to the poor gelling properties of vegetable proteins compared to their animal counterparts. Protein gelation is commonly induced by heat treatment, during which the protein chains unfold, exposing their reactive groups, which subsequently drive protein reassembling in a 3D network. Although both covalent (i.e., S-S bridges) and weak interactions (i.e., hydrophobic interactions, hydrogen bonds, and electrostatic interactions) play an important role in the formation and stabilisation of protein gels [41], the availability of free -SH groups available for covalent stabilisation is known to lead to stronger gels. The possibility to obtain strong hydrogels is pivotal in determining their suitability in the conversion into dried porous materials, since the stronger the gel, the higher its capacity to structurally withstand the subsequent drying steps. In this regard, plant proteins present a lower number of -SH groups as compared to animal ones [42]. Moreover, the extraction process performed to isolate the protein fraction from the vegetable matrix, where it is intimately embedded in fiber–protein complexes, is known to induce structural modifications in the protein chains, further reducing gelling properties [43]. Nevertheless, several factors, including protein concentration, pH, and ionic strength, can be properly modulated to improve the plant protein gelling capacity. In this regard, the increase in protein concentration usually leads to a denser protein network, accounting for the formation of firmer gels that better maintain the original volume upon water removal [41]. When gelation occurs at a pH approaching the isoelectric point (pI), globular and strongly aggregated protein structures are formed, mostly driven by hydrophobic interactions [44,45]. At a pH far above or below the pI, instead, proteins form a fine-stranded network, as a result of the presence of surface charges which prevent intimate protein aggregation [46]. For example, aerogels derived from gels prepared near protein pI have been shown to present higher structural stability during drying, associated with lower density and higher pore sizes as compared to aerogels prepared far from the pI [28,47]. Gelation properties are also affected by ionic strength (IS). The increase in IS reduces electrostatic repulsive forces among protein chains, favouring the formation of a stronger network. For instance, the elastic modulus of pea protein gels was increased by 12 times by adding 0.3 M NaCl [41]. However, beyond a salt concentration threshold, specific for each protein (usually >2.0 M), a weakening of the hydrogel structure is commonly observed, due to salt-induced stabilisation of the protein structure, which suppresses protein unfolding during gelation [48,49].

This work aimed to systematically study the effect of gelation conditions on the physical properties of plant protein-based hydrogels, with the final aim of identifying the conditions leading to hydrogels suitable for the development of dried porous materials. For this purpose, soy and pea proteins were selected as the protein sources widely used as

alternatives to animal proteins. Aqueous suspensions containing increasing amounts of soy and pea protein isolates (SPI and PPI) at different pH (3.0, 4.5 pI, 7.0) and IS (0.0, 0.6, 1.5 M) were heat-treated to induce gelation. The obtained hydrogels were characterised for rheological properties and physical stability, and the results were rationalised into a gelation map. of soy and pea protein isolates (SPI and PPI) at different pH (3.0, 4.5 pI, 7.0) and IS (0.0, 0.6, 1.5 M) were heat-treated to induce gelation. The obtained hydrogels were characterised for rheological properties and physical stability, and the results were rationalised into a gelation map. different 0.6, 1.5 M) were heat-treated to induce gelation. The obtained hydrogels were characterised for rheological properties and physical stability, and the results were rationalised into a gelation map. **2. Results and Discussion**  of soy and pea protein isolates (SPI and PPI) at different pH (3.0, 4.5 pI, 7.0) and IS (0.0, 0.6, 1.5 M) were heat-treated to induce gelation. The obtained hydrogels were characterised for rheological properties and physical stability, and the results were rationalised into a gelation map. **2. Results and Discussion**  effect final aim identifying for of M) heat-treated and stability, 0.6, 1.5 M) were heat-treated to induce gelation. The obtained hydrogels were of soy and pea protein isolates (SPI and PPI) at different pH (3.0, 4.5 pI, 7.0) and IS (0.0, 0.6, 1.5 M) were heat-treated to induce gelation. The obtained hydrogels were characterised for rheological properties and physical stability, and the results were rationalised into a gelation map. **2. Results and Discussion** 

This work aimed to systematically study the effect of gelation conditions on the physical properties of plant protein-based hydrogels, with the final aim of identifying the conditions leading to hydrogels suitable for the development of dried porous materials. For this purpose, soy and pea proteins were selected as the protein sources widely used as alternatives to animal proteins. Aqueous suspensions containing increasing amounts

This work aimed to systematically study the effect of gelation conditions on the physical properties of plant protein-based hydrogels, with the final aim of identifying the conditions leading to hydrogels suitable for the development of dried porous materials. For this purpose, soy and pea proteins were selected as the protein sources widely used as alternatives to animal proteins. Aqueous suspensions containing increasing amounts

This work aimed to systematically study the effect of gelation conditions on the physical properties of plant protein-based hydrogels, with the final aim of identifying the conditions leading to hydrogels suitable for the development of dried porous materials. For this purpose, soy and pea proteins were selected as the protein sources widely used as alternatives to animal proteins. Aqueous suspensions containing increasing amounts

physical properties of plant protein-based hydrogels, with the final aim of identifying the conditions leading to hydrogels suitable for the development of dried porous materials. For this purpose, soy and pea proteins were selected as the protein sources widely used as alternatives to animal proteins. Aqueous suspensions containing increasing amounts

This work aimed to systematically study the effect of gelation conditions on the physical properties of plant protein-based hydrogels, with the final aim of identifying the

*Gels* **2023**, *9*, x FOR PEER REVIEW 3 of 12

*Gels* **2023**, *9*, x FOR PEER REVIEW 3 of 12

*Gels* **2023**, *9*, x FOR PEER REVIEW 3 of 12

*Gels* **2023**, *9*, x FOR PEER REVIEW 3 of 12

#### **2. Results and Discussion** *2.1. Effect of Protein Type and Concentration 2.1. Effect of Protein Type and Concentration 2.1. Effect of Protein Type and Concentration of Concentration 2.1. Effect of Protein Type and Concentration*

**2. Results and Discussion** 

#### *2.1. Effect of Protein Type and Concentration* SPI and PPI solutions were prepared at increasing concentrations from 10 to 20% SPI and PPI solutions were prepared at increasing concentrations from 10 to 20% SPI and PPI solutions were prepared at increasing concentrations from 10 to 20%

SPI and PPI solutions were prepared at increasing concentrations from 10 to 20% (w/w) at pH 7.0, and thermally treated. Table 1 reports the appearance of the obtained SPI and PPI samples. (w/w) at pH 7.0, and thermally treated. Table 1 reports the appearance of the obtained SPI and PPI samples. (w/w) at pH 7.0, and thermally treated. Table 1 reports the appearance of the obtained SPI and PPI samples. (w/w) at pH 7.0, and thermally treated. Table 1 reports the appearance of the obtained SPI and PPI samples. at thermally obtained elastic of SPI and PPI solutions were prepared at increasing concentrations from 10 to 20% (w/w) at pH 7.0, and thermally treated. Table 1 reports the appearance of the obtained SPI (w/w) at pH 7.0, and thermally treated. Table 1 reports the appearance of the obtained SPI and PPI samples.

**Table 1.** Appearance, elastic (G0 ), loss modulus (G00), loss tangent (tan δ), and water-holding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) systems obtained after heat treatment of protein solutions at 10, 15, and 20% w/w; at pH 7.0; and 0.0 ionic strength. capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) systems obtained after heat treatment of protein solutions at 10, 15, and 20% w/w; at pH 7.0; and 0.0 ionic strength. capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) systems obtained after heat treatment of protein solutions at 10, 15, and 20% w/w; at pH 7.0; and 0.0 ionic strength. capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) systems obtained after heat treatment of protein solutions at 10, 15, and 20% w/w; at pH 7.0; and 0.0 ionic strength. capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) systems obtained after heat treatment of protein solutions at 10, 15, and 20% w/w; at pH 7.0; and 0.0 ionic strength.

**Table 1.** Appearance, elastic (G'), loss modulus (G"), loss tangent (tan δ), and water-holding

**Table 1.** Appearance, elastic (G'), loss modulus (G"), loss tangent (tan δ), and water-holding

**Table 1.** Appearance, elastic (G'), loss modulus (G"), loss tangent (tan δ), and water-holding

**Table** loss


As expected, for both SPI and PPI, the increase in protein concentration resulted in a visible increase in system structuring [50,51]. At a given protein concentration, SPI always led to a more structured system as compared to PPI, so a minimum protein concentration As expected, for both SPI and PPI, the increase in protein concentration resulted in a visible increase in system structuring [50,51]. At a given protein concentration, SPI always led to a more structured system as compared to PPI, so a minimum protein concentration As expected, for both SPI and PPI, the increase in protein concentration resulted in a visible increase in system structuring [50,51]. At a given protein concentration, SPI always led to a more structured system as compared to PPI, so a minimum protein concentration As expected, for both SPI and PPI, the increase in protein concentration resulted in a visible increase in system structuring [50,51]. At a given protein concentration, SPI always led to a more structured system as compared to PPI, so a minimum protein concentration As expected, for both SPI and PPI, the increase in protein concentration resulted in a visible increase in system structuring [50,51]. At a given protein concentration, SPI always led to a more structured system as compared to PPI, so a minimum protein concentration of 15 and 20% (w/w) was required to form a semi-solid system by using SPI and PPI, respectively (Table 1). This difference was also confirmed by the rheological analysis. Supplementary Figure S1 reports the frequency sweep test results for SPI and PPI hydrogels obtained from 20% (w/w) protein solutions.

For both proteins, G0 higher than G00 and parallel to G00 was obtained, indicating the formation of gel systems [52]. The moduli of the PPI gel showed a higher frequency dependence (higher slope) than those of the SPI gels. The latter showed a negligible frequency dependence, indicating that a stronger gel structure was obtained; SPI gels also presented rheological moduli higher than those of the PPI gel, and a lower loss tangent (tan δ) (Table 1). These results confirm the higher gelling ability of SPI as compared to PPI. In agreement with the literature [53,54], this difference between SPI and PPI gelation properties can be attributed to the different compositions of the globulin fraction of the considered proteins. Soybean globulins are mainly represented by glycinin (11S) and β-conglycinin (7S), which present higher solubility than pea ones (legumin 11S and vicilin 7S). As a result, a higher protein fraction would remain homogeneously suspended during the gelation of soy proteins [43,53]. Moreover, soybean globulins have been previously demonstrated to present a threshold gelling concentration lower than pea ones [55]. formation of gel systems [52]. The moduli of the PPI gel showed a higher frequency dependence (higher slope) than those of the SPI gels. The latter showed a negligible frequency dependence, indicating that a stronger gel structure was obtained; SPI gels also presented rheological moduli higher than those of the PPI gel, and a lower loss tangent (tan δ) (Table 1). These results confirm the higher gelling ability of SPI as compared to PPI. In agreement with the literature [53,54], this difference between SPI and PPI gelation properties can be attributed to the different compositions of the globulin fraction of the considered proteins. Soybean globulins are mainly represented by glycinin (11S) and βconglycinin (7S), which present higher solubility than pea ones (legumin 11S and vicilin 7S). As a result, a higher protein fraction would remain homogeneously suspended during the gelation of soy proteins [43,53]. Moreover, soybean globulins have been previously demonstrated to present a threshold gelling concentration lower than pea ones [55]. The higher strength of the gel obtained with SPI rather than PPI was also related to formation of gel systems [52]. The moduli of the PPI gel showed a higher frequency dependence (higher slope) than those of the SPI gels. The latter showed a negligible frequency dependence, indicating that a stronger gel structure was obtained; SPI gels also presented rheological moduli higher than those of the PPI gel, and a lower loss tangent (tan δ) (Table 1). These results confirm the higher gelling ability of SPI as compared to PPI. In agreement with the literature [53,54], this difference between SPI and PPI gelation properties can be attributed to the different compositions of the globulin fraction of the considered proteins. Soybean globulins are mainly represented by glycinin (11S) and βconglycinin (7S), which present higher solubility than pea ones (legumin 11S and vicilin 7S). As a result, a higher protein fraction would remain homogeneously suspended during the gelation of soy proteins [43,53]. Moreover, soybean globulins have been previously demonstrated to present a threshold gelling concentration lower than pea ones [55]. The higher strength of the gel obtained with SPI rather than PPI was also related to formation of gel systems [52]. The moduli of the PPI gel showed a higher frequency dependence (higher slope) than those of the SPI gels. The latter showed a negligible frequency dependence, indicating that a stronger gel structure was obtained; SPI gels also presented rheological moduli higher than those of the PPI gel, and a lower loss tangent (tan δ) (Table 1). These results confirm the higher gelling ability of SPI as compared to PPI. In agreement with the literature [53,54], this difference between SPI and PPI gelation properties can be attributed to the different compositions of the globulin fraction of the considered proteins. Soybean globulins are mainly represented by glycinin (11S) and βconglycinin (7S), which present higher solubility than pea ones (legumin 11S and vicilin 7S). As a result, a higher protein fraction would remain homogeneously suspended during the gelation of soy proteins [43,53]. Moreover, soybean globulins have been previously demonstrated to present a threshold gelling concentration lower than pea ones [55]. The higher strength of the gel obtained with SPI rather than PPI was also related to systems the frequency dependence (higher slope) than those of the SPI gels. The latter showed a negligible frequency dependence, indicating that a stronger gel structure was obtained; SPI gels also presented rheological moduli higher than those of the PPI gel, and a lower loss tangent (tan δ) (Table 1). These results confirm the higher gelling ability of SPI as compared to PPI. In agreement with the literature [53,54], this difference between SPI and PPI gelation properties can be attributed to the different compositions of the globulin fraction of the considered proteins. Soybean globulins are mainly represented by glycinin (11S) and βconglycinin (7S), which present higher solubility than pea ones (legumin 11S and vicilin 7S). As a result, a higher protein fraction would remain homogeneously suspended during the gelation of soy proteins [43,53]. Moreover, soybean globulins have been previously present than pea [55]. The higher strength of the gel obtained with SPI rather than PPI was also related to an improvement in gel stability, as shown by the higher WHC values (Table 1). The

*Gels* **2023**, *9*, x FOR PEER REVIEW 4 of 12

*Gels* **2023**, *9*, x FOR PEER REVIEW 4 of 12

*Gels* **2023**, *9*, x FOR PEER REVIEW 4 of 12

*Gels* **2023**, *9*, x FOR PEER REVIEW 4 of 12

hydrogels obtained from 20% (w/w) protein solutions.

hydrogels obtained from 20% (w/w) protein solutions.

hydrogels obtained from 20% (w/w) protein solutions.

hydrogels obtained from 20% (w/w) protein solutions.

of 15 and 20% (w/w) was required to form a semi-solid system by using SPI and PPI, respectively (Table 1). This difference was also confirmed by the rheological analysis. Supplementary Figure S1 reports the frequency sweep test results for SPI and PPI

of 15 and 20% (w/w) was required to form a semi-solid system by using SPI and PPI, respectively (Table 1). This difference was also confirmed by the rheological analysis. Supplementary Figure S1 reports the frequency sweep test results for SPI and PPI

of 15 and 20% (w/w) was required to form a semi-solid system by using SPI and PPI, respectively (Table 1). This difference was also confirmed by the rheological analysis. Supplementary Figure S1 reports the frequency sweep test results for SPI and PPI

of 15 and 20% (w/w) was required to form a semi-solid system by using SPI and PPI, respectively (Table 1). This difference was also confirmed by the rheological analysis. Supplementary Figure S1 reports the frequency sweep test results for SPI and PPI

For both proteins, G' higher than G" and parallel to G'' was obtained, indicating the

For both proteins, G' higher than G" and parallel to G'' was obtained, indicating the

For both proteins, G' higher than G" and parallel to G'' was obtained, indicating the

For both proteins, G' higher than G" and parallel to G'' was obtained, indicating the

The higher strength of the gel obtained with SPI rather than PPI was also related to an improvement in gel stability, as shown by the higher WHC values (Table 1). The increased density network obtained by increasing protein concentration was actually able to retain more water, due to the better distribution of the solvent in the 3D structure, as well as to the higher number of protein residues available for the interaction with water [56]. an improvement in gel stability, as shown by the higher WHC values (Table 1). The increased density network obtained by increasing protein concentration was actually able to retain more water, due to the better distribution of the solvent in the 3D structure, as well as to the higher number of protein residues available for the interaction with water [56]. an improvement in gel stability, as shown by the higher WHC values (Table 1). The increased density network obtained by increasing protein concentration was actually able to retain more water, due to the better distribution of the solvent in the 3D structure, as well as to the higher number of protein residues available for the interaction with water [56]. an improvement in gel stability, as shown by the higher WHC values (Table 1). The increased density network obtained by increasing protein concentration was actually able to retain more water, due to the better distribution of the solvent in the 3D structure, as well as to the higher number of protein residues available for the interaction with water [56]. increased density network obtained by increasing protein concentration was actually able to retain more water, due to the better distribution of the solvent in the 3D structure, as well as to the higher number of protein residues available for the interaction with water [56].

#### *2.2. Effect of pH 2.2. Effect of pH 2.2. Effect of pH 2.2. Effect of pH 2.2. Effect of pH*

The precursor protein solutions were adjusted to pH 3.0, 4.5, and 7.0 and thermally treated. Independently of the pH, self-standing gelled systems were only obtained at 15 and 20% (w/w) SPI concentrations and at a 20% (w/w) PPI concentration. As representative examples, Table 2 reports the appearance and the rheological parameters of the hydrogels obtained from the SPI and PPI solutions at 20% (w/w) protein concentration and adjusted at the different pH values. The precursor protein solutions were adjusted to pH 3.0, 4.5, and 7.0 and thermally treated. Independently of the pH, self-standing gelled systems were only obtained at 15 and 20% (w/w) SPI concentrations and at a 20% (w/w) PPI concentration. As representative examples, Table 2 reports the appearance and the rheological parameters of the hydrogels obtained from the SPI and PPI solutions at 20% (w/w) protein concentration and adjusted at the different pH values. The precursor protein solutions were adjusted to pH 3.0, 4.5, and 7.0 and thermally treated. Independently of the pH, self-standing gelled systems were only obtained at 15 and 20% (w/w) SPI concentrations and at a 20% (w/w) PPI concentration. As representative examples, Table 2 reports the appearance and the rheological parameters of the hydrogels obtained from the SPI and PPI solutions at 20% (w/w) protein concentration and adjusted at the different pH values. The precursor protein solutions were adjusted to pH 3.0, 4.5, and 7.0 and thermally treated. Independently of the pH, self-standing gelled systems were only obtained at 15 and 20% (w/w) SPI concentrations and at a 20% (w/w) PPI concentration. As representative examples, Table 2 reports the appearance and the rheological parameters of the hydrogels obtained from the SPI and PPI solutions at 20% (w/w) protein concentration and adjusted at the different pH values. The precursor protein solutions were adjusted to pH 3.0, 4.5, and 7.0 and thermally treated. Independently of the pH, self-standing gelled systems were only obtained at 15 and 20% (w/w) SPI concentrations and at a 20% (w/w) PPI concentration. As representative examples, Table 2 reports the appearance and the rheological parameters of the hydrogels obtained from the SPI and PPI solutions at 20% (w/w) protein concentration and adjusted at the different pH values.

**Table 2.** Appearance, storage modulus (G0 ), loss modulus (G00), loss tangent (tan δ), and waterholding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) hydrogels at 20% protein concentration at pH 3.0 and 4.5, and 0.0 ionic strength. **Table 2.** Appearance, storage modulus (G'), loss modulus (G"), loss tangent (tan δ), and waterholding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) hydrogels at 20% protein concentration at pH 3.0 and 4.5, and 0.0 ionic strength. **Table 2.** Appearance, storage modulus (G'), loss modulus (G"), loss tangent (tan δ), and waterholding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) hydrogels at 20% protein concentration at pH 3.0 and 4.5, and 0.0 ionic strength. **Table 2.** Appearance, storage modulus (G'), loss modulus (G"), loss tangent (tan δ), and waterholding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) hydrogels at 20% protein concentration at pH 3.0 and 4.5, and 0.0 ionic strength. **Table 2.** Appearance, storage modulus (G'), loss modulus (G"), loss tangent (tan δ), and waterholding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) hydrogels at 20% protein concentration at pH 3.0 and 4.5, and 0.0 ionic strength.


a, b, c, d: means indicated by different letters in the same column are significantly different (*p* < 0.05). a, b, c, d: means indicated by different letters in the same column are significantly different (*p* < 0.05). a, b, c, d: means indicated by different letters in the same column are significantly different (*p* < 0.05). a, b, c, d: means indicated by different letters in the same column are significantly different (*p* < 0.05). a, b, c, d: means indicated by different letters in the same column are significantly different (*p* < 0.05).

Similar to data achieved at pH 7.0 (Table 1), also at pH 3.0 and 4.5, SPI led to higher system structuration as compared to PPI. At pH 4.5, which is close to protein pI, a particulate gel, otherwise known as a microgel, was obtained with both proteins [57,58]. Proteins actually show a higher tendency towards aggregation in the isoelectric region, where the net charge is low, and thus protein–protein interactions are promoted with the formation of spherical particles, which, at a high protein concentration, can randomly associate into larger self-supporting hydrogels [58]. By contrast, at pH values away from the pI, where strong electrostatic repulsions are present, the gels present a fine-stranded structure.

For both proteins, the decrease in pH from 7.0 (Table 1) to 3.0 (Table 2) caused a significant decrease in system structuration, as evidenced by the rheological parameters. In fact, not only both moduli showed lower values for gels prepared at pH 3.0 as compared to those obtained at neutral pH, but they also presented a slightly higher frequency dependence. In this regard, Supplementary Figure S2 shows the effect of the pH change on the frequency sweep results of PPI gels prepared at 20% (w/w) protein concentration at pH 3.0 and 7.0. A significant decrease in gel strength was instead observed upon adjusting the protein solution at pH 4.5 (Tables 1 and 2). This can be attributed to the different microstructure of the hydrogels obtained at different pHs. In particular, microgelled systems obtained near the pI are stabilised by weak surface interactions among spherical protein aggregates, which can easily flow one on the other [59]. By contrast, at pHs far from the pI (pH 3.0 and 7.0), stranded gel structures are obtained, stabilised by numerous disulphide bridges and weak-interaction entanglement regions, thus accounting for the higher resistance to mechanical perturbation [58]. Moreover, in the isoelectric region, protein solubility is minimised, resulting in a significant decrease in well-solubilised protein fractions able to efficaciously interlink in a 3D gel network [53].

For both SPI and PPI, pH had a negligible effect on gel stability, as indicated by the comparable WHC values (Tables 1 and 2). This is probably due to the counterbalancing effect of the high protein concentration on the effect of pH. In other words, the effect of the different gel architectures induced by pH would be made negligible in the presence of a high protein concentration, which would increase the network density, thus allowing a high solvent retention [53].

### *2.3. Effect of Ionic Strength*

The precursor protein solutions were added with different NaCl amounts to modulate the ionic strength (IS) of the system. As representative examples of the effect of this parameter at low protein concentrations, Table 3 shows the appearance of systems obtained upon the thermal treatment of 10% (w/w) SPI and 15% (w/w) PPI solutions, at pH 7.0, and having 0.6 and 1.5 M IS.

Although the final system showed an evident phase separation, as compared to the system with no salt added (Table 1), which showed a liquid-like homogeneous structure, the increase in IS resulted in a local gelling effect with the formation of a microgel-like structure. This effect can be traced back to the shielding effect of salt ions of the protein surface charge, favouring protein aggregation [60]. The positive effect of the IS increase on SPI and PPI gelling properties was also observed at a higher protein concentration. In this regard, Table 3 reports the appearance and the rheological parameters of the hydrogels obtained from 20% (w/w) SPI and PPI solutions at pH 7.0, at 0.6 and 1.5 M IS. As compared to the gels obtained without salt addition (Table 1), the increase in IS resulted in particulate gels, well-evident in the case of the PPI-based systems (Table 3). This was due to the changes induced by the increase in IS in the gel microstructure, which shifted from a fine-stranded structure (low IS) to a particulate structure (high IS) [22]. NaCl concentration increase also caused a considerable increase in both SPI and PPI gel strength, as indicated by the increase in G0 values (Tables 1 and 3), as shown in Supplementary Figure S3, which reports the frequency sweep results for PPI gels at 20% (w/w) protein concentration at 0.0 and 1.5 M IS. The presence of Na<sup>+</sup> ions actually promotes protein–protein interactions during gelation, due to the reduction of the repulsive electrostatic interactions between protein chains [51]. Moreover, the increase in IS is known to promote the so-defined "salting-in" effect, i.e., the increase in the solubility of globulins, which are the main protein fraction of both SPI and PPI [61]. A higher IS thus results in higher availability of well-hydrated proteins available for networking during gelation [51,62]. modulate the ionic strength (IS) of the system. As representative examples of the effect of this parameter at low protein concentrations, Table 3 shows the appearance of systems obtained upon the thermal treatment of 10% (w/w) SPI and 15% (w/w) PPI solutions, at pH 7.0, and having 0.6 and 1.5 M IS. modulate the ionic strength (IS) of the system. As representative examples of the effect of this parameter at low protein concentrations, Table 3 shows the appearance of systems obtained upon the thermal treatment of 10% (w/w) SPI and 15% (w/w) PPI solutions, at pH 7.0, and having 0.6 and 1.5 M IS.

The precursor protein solutions were added with different NaCl amounts to

The precursor protein solutions were added with different NaCl amounts to

For both SPI and PPI, pH had a negligible effect on gel stability, as indicated by the comparable WHC values (Tables 1 and 2). This is probably due to the counterbalancing effect of the high protein concentration on the effect of pH. In other words, the effect of the different gel architectures induced by pH would be made negligible in the presence of a high protein concentration, which would increase the network density, thus allowing a

For both SPI and PPI, pH had a negligible effect on gel stability, as indicated by the comparable WHC values (Tables 1 and 2). This is probably due to the counterbalancing effect of the high protein concentration on the effect of pH. In other words, the effect of the different gel architectures induced by pH would be made negligible in the presence of a high protein concentration, which would increase the network density, thus allowing a

**Table 3.** Appearance, storage modulus (G0 ), loss modulus (G00), loss tangent (tan δ), and waterholding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) hydrogels at 10, 15, or 20% (w/w) protein concentrations at 0.6 and 1.5 M ionic strength. **Table 3.** Appearance, storage modulus (G'), loss modulus (G"), loss tangent (tan δ), and waterholding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) hydrogels at 10, 15, or 20% (w/w) protein concentrations at 0.6 and 1.5 M ionic strength. **Table 3.** Appearance, storage modulus (G'), loss modulus (G"), loss tangent (tan δ), and waterholding capacity (WHC) of soy protein isolate (SPI) and pea protein isolate (PPI) hydrogels at 10, 15, or 20% (w/w) protein concentrations at 0.6 and 1.5 M ionic strength.


*Gels* **2023**, *9*, x FOR PEER REVIEW 5 of 12

*Gels* **2023**, *9*, x FOR PEER REVIEW 5 of 12

structure.

structure.

3D gel network [53].

3D gel network [53].

high solvent retention [53].

high solvent retention [53].

*2.3. Effect of Ionic Strength* 

*2.3. Effect of Ionic Strength* 

Similar to data achieved at pH 7.0 (Table 1), also at pH 3.0 and 4.5, SPI led to higher system structuration as compared to PPI. At pH 4.5, which is close to protein pI, a particulate gel, otherwise known as a microgel, was obtained with both proteins [57,58]. Proteins actually show a higher tendency towards aggregation in the isoelectric region, where the net charge is low, and thus protein–protein interactions are promoted with the formation of spherical particles, which, at a high protein concentration, can randomly associate into larger self-supporting hydrogels [58]. By contrast, at pH values away from the pI, where strong electrostatic repulsions are present, the gels present a fine-stranded

Similar to data achieved at pH 7.0 (Table 1), also at pH 3.0 and 4.5, SPI led to higher system structuration as compared to PPI. At pH 4.5, which is close to protein pI, a particulate gel, otherwise known as a microgel, was obtained with both proteins [57,58]. Proteins actually show a higher tendency towards aggregation in the isoelectric region, where the net charge is low, and thus protein–protein interactions are promoted with the formation of spherical particles, which, at a high protein concentration, can randomly associate into larger self-supporting hydrogels [58]. By contrast, at pH values away from the pI, where strong electrostatic repulsions are present, the gels present a fine-stranded

For both proteins, the decrease in pH from 7.0 (Table 1) to 3.0 (Table 2) caused a significant decrease in system structuration, as evidenced by the rheological parameters. In fact, not only both moduli showed lower values for gels prepared at pH 3.0 as compared to those obtained at neutral pH, but they also presented a slightly higher frequency dependence. In this regard, Supplementary Figure S2 shows the effect of the pH change on the frequency sweep results of PPI gels prepared at 20% (w/w) protein concentration at pH 3.0 and 7.0. A significant decrease in gel strength was instead observed upon adjusting the protein solution at pH 4.5 (Tables 1 and 2). This can be attributed to the different microstructure of the hydrogels obtained at different pHs. In particular, microgelled systems obtained near the pI are stabilised by weak surface interactions among spherical protein aggregates, which can easily flow one on the other [59]. By contrast, at pHs far from the pI (pH 3.0 and 7.0), stranded gel structures are obtained, stabilised by numerous disulphide bridges and weak-interaction entanglement regions, thus accounting for the higher resistance to mechanical perturbation [58]. Moreover, in the isoelectric region, protein solubility is minimised, resulting in a significant decrease in well-solubilised protein fractions able to efficaciously interlink in a

For both proteins, the decrease in pH from 7.0 (Table 1) to 3.0 (Table 2) caused a significant decrease in system structuration, as evidenced by the rheological parameters. In fact, not only both moduli showed lower values for gels prepared at pH 3.0 as compared to those obtained at neutral pH, but they also presented a slightly higher frequency dependence. In this regard, Supplementary Figure S2 shows the effect of the pH change on the frequency sweep results of PPI gels prepared at 20% (w/w) protein concentration at pH 3.0 and 7.0. A significant decrease in gel strength was instead observed upon adjusting the protein solution at pH 4.5 (Tables 1 and 2). This can be attributed to the different microstructure of the hydrogels obtained at different pHs. In particular, microgelled systems obtained near the pI are stabilised by weak surface interactions among spherical protein aggregates, which can easily flow one on the other [59]. By contrast, at pHs far from the pI (pH 3.0 and 7.0), stranded gel structures are obtained, stabilised by numerous disulphide bridges and weak-interaction entanglement regions, thus accounting for the higher resistance to mechanical perturbation [58]. Moreover, in the isoelectric region, protein solubility is minimised, resulting in a significant decrease in well-solubilised protein fractions able to efficaciously interlink in a

N.D.: not determined, since the system did not gel. a, b, c: means indicated by different letters in the same column are significantly different (*p* < 0.05). N.D.: not determined, since the system did not gel. a, b, c: means indicated by different letters in the same column are significantly different (*p* < 0.05). N.D.: not determined, since the system did not gel. a, b, c: means indicated by different letters in the same column are significantly different (*p* < 0.05). N.D.: not determined, since the system did not gel. a, b, c: means indicated by different letters in the same column are significantly different (*p* < 0.05). not the system a, b, by different letters same column are significantly different (*p* < 0.05). N.D.: not determined, since the system did not gel. a, b, c: means indicated by different letters in the same column are significantly different (*p* < 0.05).

Although the final system showed an evident phase separation, as compared to the system with no salt added (Table 1), which showed a liquid-like homogeneous structure, the increase in IS resulted in a local gelling effect with the formation of a microgel-like structure. This effect can be traced back to the shielding effect of salt ions of the protein surface charge, favouring protein aggregation [60]. The positive effect of the IS increase on SPI and PPI gelling properties was also observed at a higher protein concentration. In this regard, Table 3 reports the appearance and the rheological parameters of the hydrogels obtained from 20% (w/w) SPI and PPI solutions at pH 7.0, at 0.6 and 1.5 M IS. As compared to the gels obtained without salt addition (Table 1), the increase in IS resulted in particulate gels, well-evident in the case of the PPI-based systems (Table 3). This was due to the changes induced by the increase in IS in the gel microstructure, which shifted from a fine-stranded structure (low IS) to a particulate structure (high IS) [22]. Although the final system showed an evident phase separation, as compared to the system with no salt added (Table 1), which showed a liquid-like homogeneous structure, the increase in IS resulted in a local gelling effect with the formation of a microgel-like structure. This effect can be traced back to the shielding effect of salt ions of the protein surface charge, favouring protein aggregation [60]. The positive effect of the IS increase on SPI and PPI gelling properties was also observed at a higher protein concentration. In this regard, Table 3 reports the appearance and the rheological parameters of the hydrogels obtained from 20% (w/w) SPI and PPI solutions at pH 7.0, at 0.6 and 1.5 M IS. As compared to the gels obtained without salt addition (Table 1), the increase in IS resulted in particulate gels, well-evident in the case of the PPI-based systems (Table 3). This was due to the changes induced by the increase in IS in the gel microstructure, which shifted from a fine-stranded structure (low IS) to a particulate structure (high IS) [22]. Although the final system showed an evident phase separation, as compared to the system with no salt added (Table 1), which showed a liquid-like homogeneous structure, the increase in IS resulted in a local gelling effect with the formation of a microgel-like structure. This effect can be traced back to the shielding effect of salt ions of the protein surface charge, favouring protein aggregation [60]. The positive effect of the IS increase on SPI and PPI gelling properties was also observed at a higher protein concentration. In this regard, Table 3 reports the appearance and the rheological parameters of the hydrogels obtained from 20% (w/w) SPI and PPI solutions at pH 7.0, at 0.6 and 1.5 M IS. As compared to the gels obtained without salt addition (Table 1), the increase in IS resulted in particulate gels, well-evident in the case of the PPI-based systems (Table 3). This was due to the changes induced by the increase in IS in the gel microstructure, which shifted from a fine-stranded structure (low IS) to a particulate structure (high IS) [22]. Although the final system showed an evident phase separation, as compared to the system with no salt added (Table 1), which showed a liquid-like homogeneous structure, the increase in IS resulted in a local gelling effect with the formation of a microgel-like structure. This effect can be traced back to the shielding effect of salt ions of the protein surface charge, favouring protein aggregation [60]. The positive effect of the IS increase on SPI and PPI gelling properties was also observed at a higher protein concentration. In this regard, Table 3 reports the appearance and the rheological parameters of the hydrogels obtained from 20% (w/w) SPI and PPI solutions at pH 7.0, at 0.6 and 1.5 M IS. As compared to the gels obtained without salt addition (Table 1), the increase in IS resulted in particulate gels, well-evident in the case of the PPI-based systems (Table 3). This was due to the changes induced by the increase in IS in the gel microstructure, which shifted from a fine-stranded structure (low IS) to a particulate structure (high IS) [22]. N.D.: not determined, since the system did not gel. a, b, c: means indicated by different letters in the same column are significantly < phase (Table the increase in IS resulted in a local gelling effect with the formation of a microgel-like back the charge, aggregation The on gelling properties was a In regard, reports the of SPI IS. As compared to the gels obtained without salt addition (Table 1), the increase in IS gels, well-evident in case of PPI-based induced the in the gel microstructure, to particulate (high NaCl increase also a considerable increase SPI by the in G' values S3, for also affected WHC IS, (Table 3). Similar from [63,64] and egg white proteins [65–68] and can be attributed to the microstructural induced by this regard, gel an fine-stranded network, presents to at high water tightly Likewise, [70] and Urbonaite, an WHC, larger contrary, 48.48 ± 0.29 b 11.52 ± 0.07 b 0.24 ± 0.01 c Although the final system showed an evident phase separation, as compared to the system with no salt added (Table 1), which showed a liquid-like homogeneous structure, the increase in IS resulted in a local gelling effect with the formation of a microgel-like be to protein surface charge, favouring protein aggregation [60]. The positive effect of the IS increase on SPI and PPI gelling properties was also observed at a higher protein concentration. In this regard, Table 3 reports the appearance and the rheological parameters of the hydrogels obtained from 20% (w/w) SPI and PPI solutions at pH 7.0, at 0.6 and 1.5 M IS. to obtained (Table 1), resulted in particulate gels, well-evident in the case of the PPI-based systems (Table 3). This was due to the changes induced by the increase in IS in the gel microstructure, which shifted from a fine-stranded structure (low IS) to a particulate structure (high IS) [22]. IS also affected gel stability. In the case of the SPI gels, WHC decreased with IS, despite the higher gel strength (Table 3). Similar results were found for gels from both soy [63,64] and egg white proteins [65–68] and can be attributed to the microstructural changes induced by the presence of ions. In this regard, Munialo et al. [69] have demonstrated that a gel with an evenly distributed fine-stranded network, obtained at low IS, generally presents higher WHC as compared to particulate gels, obtained at high IS, where water is less tightly trapped. Likewise, Maltais et al. [70] and Urbonaite, et al. [71,72] reported an inverse correlation between aggregate size and WHC, with larger aggregates resulting in lower WHC. On the contrary, in the case of PPI hydrogels, the increase in IS promoted an increase in the WHC. It can be inferred that, in this case, the increased gel structural properties obtained upon NaCl addition (Tables 1 and 3) prevailed over the microstructural changes induced by the IS increase.

NaCl concentration increase also caused a considerable increase in both SPI and PPI gel strength, as indicated by the increase in G' values (Tables 1 and 3), as shown in Supplementary Figure S3, which reports the frequency sweep results for PPI gels at 20% (w/w) protein concentration at 0.0 and 1.5 M IS. The presence of Na+ ions actually promotes protein–protein interactions during gelation, due to the reduction of the

Supplementary (w/w) protein concentration at 0.0 and 1.5 M IS. The presence of Na+ ions actually promotes protein–protein interactions during gelation, due to the reduction of the

NaCl concentration increase also caused a considerable increase in both SPI and PPI gel strength, as indicated by the increase in G' values (Tables 1 and 3), as shown in Supplementary Figure S3, which reports the frequency sweep results for PPI gels at 20% (w/w) protein concentration at 0.0 and 1.5 M IS. The presence of Na+ ions actually promotes protein–protein interactions during gelation, due to the reduction of the

NaCl concentration increase also caused a considerable increase in both SPI and PPI gel strength, as indicated by the increase in G' values (Tables 1 and 3), as shown in Supplementary Figure S3, which reports the frequency sweep results for PPI gels at 20% (w/w) protein concentration at 0.0 and 1.5 M IS. The presence of Na+ ions actually promotes protein–protein interactions during gelation, due to the reduction of the

NaCl concentration increase also caused a considerable increase in both SPI and PPI gel strength, as indicated by the increase in G' values (Tables 1 and 3), as shown in Supplementary Figure S3, which reports the frequency sweep results for PPI gels at 20% (w/w) protein concentration at 0.0 and 1.5 M IS. The presence of Na+ ions actually promotes protein–protein interactions during gelation, due to the reduction of the

NaCl concentration increase also caused a considerable increase in both SPI and PPI gel strength, as indicated by the increase in G' values (Tables 1 and 3), as shown in

of globulins, which are the main protein fraction of both SPI and PPI [61]. A higher IS thus results in higher availability of well-hydrated proteins available for networking during

results in higher availability of well-hydrated proteins available for networking during gelation [51,62].

demonstrated that a gel with an evenly distributed fine-stranded network, obtained at low IS, generally presents higher WHC as compared to particulate gels, obtained at high IS, where water is less tightly trapped. Likewise, Maltais et al. [70] and Urbonaite, et al. [71,72] reported an inverse correlation between aggregate size and WHC, with larger

of globulins, which are the main protein fraction of both SPI and PPI [61]. A higher IS thus results in higher availability of well-hydrated proteins available for networking during

of globulins, which are the main protein fraction of both SPI and PPI [61]. A higher IS thus results in higher availability of well-hydrated proteins available for networking during

of globulins, which are the main protein fraction of both SPI and PPI [61]. A higher IS thus results in higher availability of well-hydrated proteins available for networking during

IS also affected gel stability. In the case of the SPI gels, WHC decreased with IS, despite the higher gel strength (Table 3). Similar results were found for gels from both soy [63,64] and egg white proteins [65–68] and can be attributed to the microstructural changes induced by the presence of ions. In this regard, Munialo et al. [69] have demonstrated that a gel with an evenly distributed fine-stranded network, obtained at low IS, generally presents higher WHC as compared to particulate gels, obtained at high IS, where water is less tightly trapped. Likewise, Maltais et al. [70] and Urbonaite, et al. [71,72] reported an inverse correlation between aggregate size and WHC, with larger aggregates resulting in lower WHC. On the contrary, in the case of PPI hydrogels, the

IS also affected gel stability. In the case of the SPI gels, WHC decreased with IS, despite the higher gel strength (Table 3). Similar results were found for gels from both soy [63,64] and egg white proteins [65–68] and can be attributed to the microstructural changes induced by the presence of ions. In this regard, Munialo et al. [69] have demonstrated that a gel with an evenly distributed fine-stranded network, obtained at low IS, generally presents higher WHC as compared to particulate gels, obtained at high IS, where water is less tightly trapped. Likewise, Maltais et al. [70] and Urbonaite, et al. [71,72] reported an inverse correlation between aggregate size and WHC, with larger aggregates resulting in lower WHC. On the contrary, in the case of PPI hydrogels, the

are the

to the et [69]

IS also affected gel stability. In the case of the SPI gels, WHC decreased with IS, despite the higher gel strength (Table 3). Similar results were found for gels from both soy [63,64] and egg white proteins [65–68] and can be attributed to the microstructural changes induced by the presence of ions. In this regard, Munialo et al. [69] have demonstrated that a gel with an evenly distributed fine-stranded network, obtained at low IS, generally presents higher WHC as compared to particulate gels, obtained at high IS, where water is less tightly trapped. Likewise, Maltais et al. [70] and Urbonaite, et al. [71,72] reported an inverse correlation between aggregate size and WHC, with larger aggregates resulting in lower WHC. On the contrary, in the case of PPI hydrogels, the

IS also affected gel stability. In the case of the SPI gels, WHC decreased with IS, despite the higher gel strength (Table 3). Similar results were found for gels from both soy [63,64] and egg white proteins [65–68] and can be attributed to the microstructural changes induced by the presence of ions. In this regard, Munialo et al. [69] have demonstrated that a gel with an evenly distributed fine-stranded network, obtained at low IS, generally presents higher WHC as compared to particulate gels, obtained at high IS, where water is less tightly trapped. Likewise, Maltais et al. [70] and Urbonaite, et al. [71,72] reported an inverse correlation between aggregate size and WHC, with larger aggregates resulting in lower WHC. On the contrary, in the case of PPI hydrogels, the

IS also affected gel stability. In the case of the SPI gels, WHC decreased with IS, despite the higher gel strength (Table 3). Similar results were found for gels from both soy

gelation [51,62].

gelation [51,62].

gelation [51,62].

gelation [51,62].

### *2.4. Gelation Map 2.4. Gelation Map*  Collected data were further elaborated and rationalised in order to obtain a gelation

Collected data were further elaborated and rationalised in order to obtain a gelation map (Figure 1), which is useful to have an immediate view of the gelation performances of SPI and PPI under the considered conditions. map (Figure 1), which is useful to have an immediate view of the gelation performances of SPI and PPI under the considered conditions.

increase in IS promoted an increase in the WHC. It can be inferred that, in this case, the increased gel structural properties obtained upon NaCl addition (Tables 1 and 3)

*Gels* **2023**, *9*, x FOR PEER REVIEW 7 of 12

prevailed over the microstructural changes induced by the IS increase.

**Figure 1.** Gelation map of soy (SPI) and pea protein isolate (PPI) at increasing protein concentration (%, w/w), pH, and ionic strength (IS). The mean values of elastic modulus (G' × 103 Pa) of the gelled systems are also reported within cells. **Figure 1.** Gelation map of soy (SPI) and pea protein isolate (PPI) at increasing protein concentration (%, w/w), pH, and ionic strength (IS). The mean values of elastic modulus (G<sup>0</sup> <sup>×</sup> <sup>10</sup><sup>3</sup> Pa) of the gelledsystems are also reported within cells.

The obtained map clearly highlights the complex effect of protein type, pH, IS, and their combination on the sample structure. For example, the higher gelation propensity of SPI as compared to PPI is immediately visible, as well as the higher structuration obtained far away from the protein isoelectric region pH or increasing the IS. This map represents a useful tool to identify optimal conditions leading to SPI and PPI gels presenting the desired physical properties. In particular, the conditions allowing for the preparation of hydrogels presenting a network strong enough to withstand the conversion into xerogels, cryogels, and aerogels can be identified. Moreover, additional considerations can be drawn, with the aim of optimising the production process of these dried porous materials. For example, at pH 3.0 or 7.0, in view of minimising the consumption of SPI, and thus raw material costs, while also maintaining a strong gel structure, the possibility to reduce the SPI concentration from 20 to 15% (w/w) while increasing the ionic strength can be identified. Similarly, in the case of PPI, it is immediately evident how only weak gels can The obtained map clearly highlights the complex effect of protein type, pH, IS, and their combination on the sample structure. For example, the higher gelation propensity of SPI as compared to PPI is immediately visible, as well as the higher structuration obtained far away from the protein isoelectric region pH or increasing the IS. This map represents a useful tool to identify optimal conditions leading to SPI and PPI gels presenting the desired physical properties. In particular, the conditions allowing for the preparation of hydrogels presenting a network strong enough to withstand the conversion into xerogels, cryogels, and aerogels can be identified. Moreover, additional considerations can be drawn, with the aim of optimising the production process of these dried porous materials. For example, at pH 3.0 or 7.0, in view of minimising the consumption of SPI, and thus raw material costs, while also maintaining a strong gel structure, the possibility to reduce the SPI concentration from 20 to 15% (w/w) while increasing the ionic strength can be identified. Similarly, in the case of PPI, it is immediately evident how only weak gels can be obtained at 20% concentration.

#### be obtained at 20% concentration. **3. Conclusions**

**3. Conclusions**  The results collected in this study show that the gelling behaviour of vegetable proteins is highly dependent on both the protein nature and formulation parameters (protein concentration, pH, ionic strength). In particular, hydrogel strength can be enhanced by choosing soy proteins over pea ones, as well as avoiding the isoelectric region and increasing the ionic strength. The obtained gelation map can be considered a useful tool to identify the optimal conditions to produce soy and pea protein hydrogels with physical properties suitable for the subsequent conversion into xerogels, cryogels, and aerogels.

The results obtained in this research, although relevant to soy and pea protein isolates solely, clearly indicate the potential of plant proteins as interesting precursors for the production of food-grade and plant protein-based dried porous materials. Further studies are therefore required to investigate the correlation between the physical and technofunctional properties of the precursor hydrogel and the resulting dried materials. In this regard, different drying processes such as evaporative drying, freeze-drying, and supercritical drying can be applied to convert the obtained hydrogels into xerogels, cryogels, and aerogels, respectively. At the same time, a comprehensive characterisation of the dried templates obtained thereof could be performed. The latter should include the physical characterisation of the materials (e.g., SEM microstructure, BET surface area, porosity) but also their interaction properties with food fluids (oil, water) to obtain a first insight into their applicability as innovative food ingredients.

### **4. Materials and Methods**

### *4.1. Soy and Pea Protein Solution Preparation*

Aqueous solutions presenting different ionic strength (IS), 0.6 and 1.5 M, were prepared by adding NaCl (Sigma Aldrich, Milan, Italy) in deionised water (System advantage A10®, Millipore S.A.S, Molsheim, France). Deionised water without the addition of NaCl was considered to have an IS equal to 0.0 M. Aqueous solutions were added with 10, 15, or 20% (w/w) of soy (SPI) or pea (PPI) protein isolates (Myprotein, Manchester, England). The suspensions were subjected to high shear mixing at 1120× *g* for 1 min (Polytron PT-MR3000, Kinematica AG, Littau, Switzerland), and pH was adjusted to 3.0, 4.5, and 7.0 by adding 1 M NaOH or HCl.

### *4.2. Heat Treatment*

To induce gelation, soy and pea protein suspensions were transferred in 50 mL-sealed falcon tubes and subjected to thermal treatment in a water bath (95 ◦C for 15 min), followed by cooling in an ice bath (0 ◦C for 15 min). The heat-treated samples were then stored at 4 ◦C for 48 h, until analysis.

### *4.3. Image Acquisition*

Images were captured with a digital camera (EOS 550D, Canon, Milano, Italy) in an image acquisition cabinet (Immagini & Computer, Bareggio, Italy). The digital camera was positioned in an adjustable stand positioned at 45 cm from the samples and enlightened by 4 × 100 W frosted photographic floodlights, in a position allowing minimum shadow and glare.

### *4.4. Rheological Properties*

Hydrogel rheological properties were tested using an RS6000 Rheometer (Thermo Scientific RheoStress, Haake, Germany), equipped with a Peltier system for temperature control. The analysis was performed with a parallel plate geometry, with a gap of 2.0 mm at 20 ◦C. Hydrogels were cut into cylinders with 2 mm of height and 20 mm of diameter. The linear viscoelastic region (LVR) was determined using an oscillatory sweep test (0.01 to 1000 Pa at 1 Hz frequency). The frequency sweep tests were carried out increasing the frequency from 0.1 to 20 Hz, at stress values selected in the LVR.

### *4.5. Physical Stability*

The physical stability of hydrogels was evaluated based on their water-holding capacity (*WHC*). Hydrogels were accurately weighed (*W*1) and transferred into 1.5 mL-Eppendorf microcentrifuge tubes, and then centrifugated at 15,000× *g* for 15 min at 4 ◦C (D3024, DLAB, Scientific Europe S.A.S, Schiltigheim, France). The supernatant was then removed, and the samples were weighed again (*W*2). The *WHC* was determined according to Equation (1).

$$\text{WHC} = \frac{W\_1 - (W\_1 - W\_2)}{W\_1} \cdot 100 \tag{1}$$

### *4.6. Data Analysis*

Data are expressed as the mean ± standard deviation of at least three measurements resulting from two replicates. The statistical analysis was performed using the program R version 4.1.2 (The R Foundation for Statistical Computing, Vienna, Austria). The homogeneity of the variance was evaluated with Bartlett tests, a one-way ANOVA was applied, and the difference between the averages was assessed by the post-hoc Tukey test (*p* < 0.05).

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/gels9010062/s1, Figure S1: Elastic (G0 ) and viscous (G00) modulus of soy (SPI) and pea (PPI) hydrogels obtained from 20% (w/w) protein solutions. Figure S2. Elastic (G0 ) and viscous (G00) modulus of pea protein isolate (PPI) hydrogels obtained from 20% (w/w) protein solutions at pH 3.0 and 7.0. Figure S3. Elastic (G0 ) and viscous (G00) modulus of pea protein isolate (PPI) hydrogels obtained from 20% (w/w) protein solutions at 0.0 and 1.5 M ionic strength.

**Author Contributions:** L.D.B.: Investigation, Formal Analysis, Visualisation, Writing—Original Draft; S.P.: Conceptualisation, Methodology, Writing—Review and Editing, Supervision; L.M.: Conceptualisation, Resources, Writing—Review and Editing, Supervision, Funding Acquisition. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data available on request to the corresponding author.

**Acknowledgments:** The authors thank Tijana Bjelogrlic for helping in the sample preparation and analysis.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Article* **Pd-Loaded Cellulose NanoSponge as a Heterogeneous Catalyst for Suzuki–Miyaura Coupling Reactions**

**Laura Riva <sup>1</sup> , Gloria Nicastro <sup>1</sup> , Mingchong Liu <sup>1</sup> , Chiara Battocchio <sup>2</sup> , Carlo Punta 1,3 and Alessandro Sacchetti 1,\***


**Abstract:** The (eco)design and synthesis of durable heterogeneous catalysts starting from renewable sources derived from biomass waste represents an important step for reducing environmental impacts of organic transformations. Herein, we report the efficient loading of Pd(II) ions on an eco-safe cellulose-based organic support (CNS), obtained by thermal cross-linking between TEMPO-oxidized cellulose nanofibers and branched polyethyleneimine in the presence of citric acid. A 22.7% *w*/*w* Pd-loading on CNS was determined by the ICP-OES technique, while the metal distribution on the xerogel was evidenced by SEM–EDS analysis. XPS analysis confirmed the direct chelation of Pd(II) ions by means of the high number of amino groups present in the network, so that further functionalization of the support with specific ligands was not necessary. The new composite turned to be an efficient heterogeneous pre-catalyst for promoting Suzuki–Miyaura coupling reactions between aryl halides and phenyl boronic acid in water, obtaining yields higher than 90% in 30 min, by operating in a microwave reactor at 100 ◦C and with just 2% *w*/*w* of CNS-Pd catalyst with respect to aryl halides (4.5‰ for Pd). At the end of first reaction cycle, Pd(II) ions on the support resulted in being reduced to Pd(0) while maintaining the same catalytic efficiency. In fact, no leaching was observed at the end of reactions, and five cycles of recycling and reusing of CNS-Pd catalyst provided excellent results in terms of yields and selectivity in the desired products.

**Keywords:** nanocellulose; Suzuki–Miyaura coupling; heterogeneous catalysis; sustainable catalyst; nanocellulose-based xerogels; green chemistry

### **1. Introduction**

Pd-catalyzed Suzuki–Miyaura cross-coupling is one of the most investigated C-C bond formation reactions, widely applied for the synthesis of complex molecules, including pharmaceuticals, semiconductors, supramolecular structures, and pesticides [1–4].

Many efforts have been devoted over the years to the design and synthesis of Pd(II) precatalysts by selecting proper ligands such as *N*-heterocyclic carbenes [5–7], phosphine [8,9], palladacycles [10], the PEPPSI (pyridine-enhanced precatalyst preparation) system [11], and allyl-based ligands [12], in order to improve the efficiency of this catalysis under homogeneous conditions.

Ligands guarantee excellent donor abilities, high steric hindrance, and, in most cases, the stabilization of a Pd(0) reduced form, which is considered to be the active species once generated in situ [13].

While most of these approaches allow operation at low catalyst loadings, under mild conditions, and even in green solvents, including water [14,15], they all suffer from the limits related to homogeneous catalysis, which can be summarized in the direct costs for

**Citation:** Riva, L.; Nicastro, G.; Liu, M.; Battocchio, C.; Punta, C.; Sacchetti, A. Pd-Loaded Cellulose NanoSponge as a Heterogeneous Catalyst for Suzuki–Miyaura Coupling Reactions. *Gels* **2022**, *8*, 789. https://doi.org/10.3390/gels8120789

Academic Editor: Pavel Gurikov

Received: 11 November 2022 Accepted: 29 November 2022 Published: 2 December 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

catalysts' synthesis, and the indirect ones for their efficient recovery and reuse, which is also a key issue for the environmental impact of the process.

For these reasons, in recent years many solutions have been proposed for the immobilization of Pd(II)-precatalysts onto heterogeneous networks [16–19], also opening the route to continuous-flow synthetic processes [20,21]. However, in most cases the prefunctionalization of solid supports with proper ligands is necessary to guarantee high efficiency in fixing Pd(II), minimizing leaching phenomena [22,23]. Moreover, the increasing demand for bio-based materials derived from renewable sources, such as biomass waste, in the framework of the circular economy, pushes towards the design of new solutions as heterogenous supports for organometallic catalysis, capable of minimizing synthetic steps and providing sustainable solutions with low environmental impact.

In this context, in recent years we designed and developed a microporous cellulosebased nanosponge (CNS), having TEMPO ((2,2,6,6-Tetramethylpiperidin-1-yl)oxyl)-oxidized cellulose nanofibers and 25 kDa branched polyethylenimine (bPEI) as main components [24]. The high porosity of the system derived from the freeze-drying process followed for converting the original hydrogel-like suspension of the two polymers into the resulting xerogel, with ice crystals acting as pores' templates, while its chemical stability was guaranteed by the thermally induced (~100 ◦C) formation of amide bonds between the carboxyl groups of the oxidized nanocellulose and the amine groups of the polyamine. Further optimization of pristine formulations by addition of citric acid (CA) allowed the nanostructure to have higher mechanical resistance [25] and to better fix bPEI, improving the eco-safety [26–29] and the sustainability [30] of the material. More recently, the nanoporosity of the material was revealed by small-angle neutron scattering (SANS) investigation [31] and by FTIR-ATR analysis of the H-bond network, which evidenced water nanoconfinement in the nanostructure [32,33].

CNS have found ample application in different fields, including wastewater remediation [29,34], sensing [35,36], as drug-delivery systems [25], and as heterogenous catalysts for promoting amino-catalyzed organic reactions [37]. Very recently, we were inspired by the high heavy-metal adsorption efficiency of CNS exploited in wastewater treatment [24,29]. This property had to be ascribed to the strong chelating action of primary, secondary, and tertiary amino groups, provided by the presence of bPEI in the xerogel network. Inspired by this behavior, we envisioned the opportunity to consider these materials as suitable organic heterogeneous supports for transition metal ions, opening the synthesis of a new class of heterogeneous organometallic catalysts for organic reactions. In a first attempt, we confirmed this hypothesis by designing CNS-Cu- and CNS-Zn-loaded catalysts, which were successfully used to promote the synthesis of aromatic acetals [38].

With these premises, herein we propose CNS as easy-to-prepare biobased and sustainable supports for Pd(II) ions. As CNS do not require a pre-functionalization by specific ligands to efficiently trap Pd(II), the environmental and economic impact of these systems is minimized. The new CNS-Pd(II) composite turned out to be an efficient and stable heterogeneous precatalyst towards the Suzuki–Miyaura cross-coupling reaction.

### **2. Results and Discussion**

### *2.1. CNS-Pd Synthesis and Characterization*

CNS-Pd was synthesized as reported in Scheme 1, following a two-step procedure: (i) the production of CNS and (ii) the loading of Pd(II) ions on the resulting xerogel. For the first step, TOCNF and bPEI were mixed in a 1:2 weight ratio in deionized water in the presence of 18% of CA with respect to primary amino groups of bPEI. The resulting hydrogel was transferred in molds and underwent a freeze-drying process, providing a highly porous xerogel. The latter was heated in an oven at ~100 ◦C in order to impart chemical stability and mechanical resistance to the final material and allowing the formation of amidic bonds by dehydration between the carboxylic groups of TOCNF and CA, and the primary amino groups of bPEI [25]. In the second step, the nanosponge was ground in a mortar before use to increase the superficial area and consequently the Pd-sorption

*Gels* **2022**, *8*, x FOR PEER REVIEW 3 of 16

CA, and the primary amino groups of bPEI [25]. In the second step, the nanosponge was ground in a mortar before use to increase the superficial area and consequently the Pdsorption efficiency. Pd loading was performed by soaking the CNS material in a saturated

*Gels* **2022**, *8*, x FOR PEER REVIEW 3 of 16

efficiency. Pd loading was performed by soaking the CNS material in a saturated PdCl<sup>2</sup> solution, running several loading cycles. sorption efficiency. Pd loading was performed by soaking the CNS material in a saturated PdCl2 solution, running several loading cycles.

CA, and the primary amino groups of bPEI [25]. In the second step, the nanosponge was ground in a mortar before use to increase the superficial area and consequently the Pd-

**Scheme 1.** Preparation of CNS-Pd. **Scheme 1.** Preparation of CNS-Pd. reported in the literature. More specifically, evidence of amide bonding was shown by

A complete chemical and morphological characterization of CNS has already been reported in the literature. More specifically, evidence of amide bonding was shown by FTIR [32,34] and 15N CP-MAS solid-state NMR analyses [25]; microporosity of the system was detected by scanning electron microscopy (SEM) images and better-investigated by microcomputed tomography quantitative analysis, resulting in about 70% of the bulk material [25,29]; finally, nanoporosity of the xerogel was revealed by a small-angle neutron scattering (SANS) [31] study and an investigation of water nanoconfinement in the network by ATR-FTIR [32,34]. We thus proceeded with an in-deep characterization of the new Pd-loaded system. Scanning Electron Microscopy/Energy Dispersive Spectroscopy A complete chemical and morphological characterization of CNS has already been reported in the literature. More specifically, evidence of amide bonding was shown by FTIR [32,34] and <sup>15</sup>N CP-MAS solid-state NMR analyses [25]; microporosity of the system was detected by scanning electron microscopy (SEM) images and better-investigated by microcomputed tomography quantitative analysis, resulting in about 70% of the bulk material [25,29]; finally, nanoporosity of the xerogel was revealed by a small-angle neutron scattering (SANS) [31] study and an investigation of water nanoconfinement in the network by ATR-FTIR [32,34]. We thus proceeded with an in-deep characterization of the new Pdloaded system. Scanning Electron Microscopy/Energy Dispersive Spectroscopy (SEM–EDS) analysis revealed a homogeneous distribution of the metal on ground CNS (Figure 1). FTIR [32,34] and 15N CP-MAS solid-state NMR analyses [25]; microporosity of the system was detected by scanning electron microscopy (SEM) images and better-investigated by microcomputed tomography quantitative analysis, resulting in about 70% of the bulk material [25,29]; finally, nanoporosity of the xerogel was revealed by a small-angle neutron scattering (SANS) [31] study and an investigation of water nanoconfinement in the network by ATR-FTIR [32,34]. We thus proceeded with an in-deep characterization of the new Pd-loaded system. Scanning Electron Microscopy/Energy Dispersive Spectroscopy (SEM–EDS) analysis revealed a homogeneous distribution of the metal on ground CNS

**Figure 1.** SEM images of CNS-Pd (**A**) and distribution of Pd on the catalyst's surface, obtained with **Figure 1.** SEM images of CNS-Pd (**A**) and distribution of Pd on the catalyst's surface, obtained with **Figure 1.** SEM images of CNS-Pd (**A**) and distribution of Pd on the catalyst's surface, obtained with EDS (**B**).

An EDS absorption spectrum is also reported in Figure 2, where the Pd and Cl<sup>−</sup> sig-

An EDS absorption spectrum is also reported in Figure 2, where the Pd and Cl<sup>−</sup> sig-

EDS (**B**).

nals can be clearly observed.

(Figure 1).

EDS (**B**).

nals can be clearly observed.

An EDS absorption spectrum is also reported in Figure 2, where the Pd and Cl− signals can be clearly observed. by ICP-OES analysis, obtaining a value of 22.7% *w*/*w*, corresponding to 2.13 mmolPd/gCNS.

A quantification of the amount of Pd(II) actually loaded on CNS-Pd was measured

*Gels* **2022**, *8*, x FOR PEER REVIEW 4 of 16

**Figure 2.** EDS absorption spectrum of CNS-Pd. **Figure 2.** EDS absorption spectrum of CNS-Pd.

To obtain information on the structural features of the CNS-Pd catalyst, X-Ray pho-A quantification of the amount of Pd(II) actually loaded on CNS-Pd was measured by ICP-OES analysis, obtaining a value of 22.7% *w*/*w*, corresponding to 2.13 mmolPd/gCNS.

toelectron spectroscopy (XPS) studies were performed on the pristine catalyst as well as on the catalyst recovered after the first catalytic cycle *(*see Section 2.6 for reaction conditions). The pristine CNS matrix was also investigated to provide useful references for the assignment of Pd3d spectral features. Measurements were carried out at C 1s, N 1s, O 1s, and Pd 3d core levels. A complete collection of core-level binding energy (BE), full width at half-maxima (FWHM) values, and proposed assignments is reported in Table S1 in the Supporting Information; here, the Pd3d spin-orbit components will be discussed with particular attention since they are of major interest for the assessment of the Pd–CNS interaction in the catalyst and for the investigation of its role in the catalytic reaction. C1s and N1s spectral features will also be briefly discussed since the reproducibility of such signals confirms the stability of the catalyst molecular structure upon use. To obtain information on the structural features of the CNS-Pd catalyst, X-ray photoelectron spectroscopy (XPS) studies were performed on the pristine catalyst as well as on the catalyst recovered after the first catalytic cycle (see Section 2.6 for reaction conditions). The pristine CNS matrix was also investigated to provide useful references for the assignment of Pd3d spectral features. Measurements were carried out at C 1s, N 1s, O 1s, and Pd 3d core levels. A complete collection of core-level binding energy (BE), full width at half-maxima (FWHM) values, and proposed assignments is reported in Table S1 in the Supporting Information; here, the Pd3d spin-orbit components will be discussed with particular attention since they are of major interest for the assessment of the Pd–CNS interaction in the catalyst and for the investigation of its role in the catalytic reaction. C1s and N1s spectral features will also be briefly discussed since the reproducibility of such signals confirms the stability of the catalyst molecular structure upon use.

C1s and N1s spectra collected on the pristine CNS and on the CNS-Pd catalyst before and after the catalytic process show analogous features, in excellent accordance with the chemical composition of CNS; in more detail, C1s spectra are composite and at least four spectral components can be individuated by applying a peak fitting procedure, assigned, respectively, to aliphatic C atoms (BE = 285.00 eV), C atoms bonded to N or O in C-N, C-O functional groups (286.3 eV), O-C-O or C=O carbons (287.5 eV), and carboxylic COOH functional groups (288.9 eV) [39–43]. The relative amount of each species is well-reproducible in the three samples, as reported in the column "atomic ratios" in Table S1. C1s spectra of CNS-Pd pristine and recovered from the catalysis reaction are reported in Figure 3A,B. As for the N 1s spectra, a main signal is always found at about 400 eV, as expected for N atoms in the polyamine [43]. At higher BE values, a signal of very low inten-C1s and N1s spectra collected on the pristine CNS and on the CNS-Pd catalyst before and after the catalytic process show analogous features, in excellent accordance with the chemical composition of CNS; in more detail, C1s spectra are composite and at least four spectral components can be individuated by applying a peak fitting procedure, assigned, respectively, to aliphatic C atoms (BE = 285.00 eV), C atoms bonded to N or O in C-N, C-O functional groups (286.3 eV), O-C-O or C=O carbons (287.5 eV), and carboxylic COOH functional groups (288.9 eV) [39–43]. The relative amount of each species is wellreproducible in the three samples, as reported in the column "atomic ratios" in Table S1. C1s spectra of CNS-Pd pristine and recovered from the catalysis reaction are reported in Figure 3A,B. As for the N 1s spectra, a main signal is always found at about 400 eV, as expected for N atoms in the polyamine [43]. At higher BE values, a signal of very low intensity is also observed, probably due to oxidized nitrogen atoms belonging to impurities of the CNS.

sity is also observed, probably due to oxidized nitrogen atoms belonging to impurities of the CNS. Pd3d spectra show the presence of several contributions in different BE positions for the pristine and recovered catalysts, as reported in Figure 3C (pristine CNS-Pd) and D (recovered CNS-Pd).

**Figure 3.** (**A**) XPS C 1s core-level spectra acquired on pristine CNS and (**B**) CNS recovered after catalysis; (**C**) XPS Pd 3d core-level spectra acquired on CNS before and (**D**) after catalysis.

The Pd3d spectrum of pristine CNS-Pd shows the presence of Pd(II) ions only; however, by applying the peak-fitting procedure, it is possible to point out at least three spin-orbit pairs whose BE positions are indicative of different chemical environments. The signal at lower BE (Pd3d5/2 BE = 337.38 eV) can be associated with PdO [43] or, as suggested by some authors, with Pd(II) ions coordinating carboxylic groups [44]. A less-suitable assignment could be atomic Pd(0) interacting with amorphous carbon, as suggested by Bertolini et al. [45]. Moreover, the low-BE component is also less intense (about 7% of all Pd contribution) in the pristine CNS-Pd Pd3d spectrum. The second feature (Pd3d5/2 component at 338.77 eV BE) is indicative of ionic Pd(II) in PdCl<sup>2</sup> [43], as also confirmed by the position of the Cl2p signal (Cl2p3/2 BE = 198.40 eV, see Table S1) [46]. Finally, the most intense signal, at higher BE value (Pd3d5/2 BE = 340.19 eV, 64% of all the palladium in the catalyst), is due to Pd(II) ions interacting with the amine-like nitrogens of the polyamine moieties in coordination compounds, similarly to Pd(NH3)4Cl<sup>2</sup> [43].

In Figure 3C, a Ca2p signal arising from Ca(II) ions contained as impurities in the CNS matrix is observed and partially superimposed to the 3/2 Pd3d spin-orbit components; an analogous signal is also observed in pure-pristine CNS (see Supporting Information, Table S1). The Ca2p signal is not found in the recovered catalyst, probably due to repeated washing with ultrapure water.

After recovering, the Pd signal measured for CNS-Pd catalyst suggests a different composition. The most-intense spectral component (71% of Pd species) is shifted at low BE values (Pd3d5/2 BE = 335.46 eV) and it is now indicative of reduced Pd(0) atoms. At higher BE values, a less intense signal can be observed (Pd3d5/2 BE = 338.20); from the literature, this minor (29%) contribution could be due to Pd(0) atoms interacting with an amorphous carbon matrix [45] or Pd(II) ions coordinating halogenated organic molecules (Br-containing reactant residues, for example, or reaction intermediates and byproducts) [47,48]. Measurements at the Br3d core level reveal a small number of bromine atoms covalently bonded to C (Br3d5/2 BE = 68.80 eV) in the recovered sample, supporting the latter assignment (see Table S1).

#### *2.2. Suzuki–Miyaura Reaction Optimization 2.2. Suzuki–Miyaura Reaction Optimization*

The reaction between 4-Br-anisole (**1a**) and phenylboronic acid (**2**) in water (Scheme 2) was selected as the model for the optimization of the conditions. CNS-Pd precatalyst was used in 2–10% *w*/*w* with KOH (2 eq) as base. In addition, to facilitate the solubilization of the reagents, the phase-transfer agent TBAB (0.6 eq) was used. <sup>1</sup>H-NMR in CDCl<sup>3</sup> with acetonitrile as internal standard was used to calculate the reaction conversion during the reaction optimization process. In all cases, a complete selectivity toward the desired product could be observed, without the formation of any undesired by-products. The reaction between 4-Br-anisole (**1a**) and phenylboronic acid (**2**) in water (Scheme 2) was selected as the model for the optimization of the conditions. CNS-Pd precatalyst was used in 2–10% *w*/*w* with KOH (2 eq) as base. In addition, to facilitate the solubilization of the reagents, the phase-transfer agent TBAB (0.6 eq) was used. 1H-NMR in CDCl<sup>3</sup> with acetonitrile as internal standard was used to calculate the reaction conversion during the reaction optimization process. In all cases, a complete selectivity toward the desired product could be observed, without the formation of any undesired by-products.

**Scheme 2.** Suzuki–Miyaura reference reaction. **Scheme 2.** Suzuki–Miyaura reference reaction.

The influence of reaction time and temperature was first investigated with the use of a microwave reactor, according to the principle of *"*time–temperature equivalence*"*, shortening the reaction time under the condition of increasing temperature. For comparison, conventional heating was also considered. The influence of reaction time and temperature was first investigated with the use of a microwave reactor, according to the principle of "time–temperature equivalence", shortening the reaction time under the condition of increasing temperature. For comparison, conventional heating was also considered.

As reported in Table 1, after working at room temperature (T = 25 °C) with long reaction time (48 h), the reaction yield was not satisfactory (entry 1). As expected, at higher temperatures yields increased and shorter reaction times could be applied, going from 80 °C, 30 min (entry 2), with a yield of 57% to 100 °C, 20 min (entry 3), with a yield of 72%. Finally, applying a temperature of 100 °C for a reaction time of 30 min (entry 4), a satisfying 92% yield was achieved. It has to be noticed that in the same conditions but with the use of conventional heating, a very low 19% yield was obtained (entry 5), thus highlighting the unique role and the high effectiveness of microwave irradiation in this catalytic As reported in Table 1, after working at room temperature (T = 25 ◦C) with long reaction time (48 h), the reaction yield was not satisfactory (entry 1). As expected, at higher temperatures yields increased and shorter reaction times could be applied, going from 80 ◦C, 30 min (entry 2), with a yield of 57% to 100 ◦C, 20 min (entry 3), with a yield of 72%. Finally, applying a temperature of 100 ◦C for a reaction time of 30 min (entry 4), a satisfying 92% yield was achieved. It has to be noticed that in the same conditions but with the use of conventional heating, a very low 19% yield was obtained (entry 5), thus highlighting the unique role and the high effectiveness of microwave irradiation in this catalytic system.


system. **Table 1.** Time and temperature reaction optimization <sup>a</sup> .

4 100 0.5 92 5 <sup>b</sup> 100 0.5 19 <sup>a</sup> Reaction conditions: 0.268 mmol of **1a** (1 eq), 0.161 mmol of TBAB (0.6 eq), 0.563 mmol of KOH (2 <sup>a</sup> Reaction conditions: 0.268 mmol of **1a** (1 eq), 0.161 mmol of TBAB (0.6 eq), 0.563 mmol of KOH (2 eq), and 1 mg (2% *w*/*w*) catalyst in 2.5 mL of water under MW irradiation. <sup>b</sup> Conventional heating instead of MW irradiation was used.

eq), and 1 mg (2% *w*/*w*) catalyst in 2.5 mL of water under MW irradiation. <sup>b</sup> Conventional heating instead of MW irradiation was used. In order to achieve higher yields, the amount of catalyst was increased (Table 2). No significant results were obtained. Indeed, going from 2% (entry 2) to 5% (entry 3) and 10% *w*/*w* (entry 4), the yields did not substantially improve. As expected, in the absence of catalyst (entry 1) and in the presence of a metal-free catalyst (CNS, entry 5), no conversion was observed. According to these results, the 2% *w*/*w* catalyst loading was taken as the optimal condition. It is important to highlight that according to the measured loading of In order to achieve higher yields, the amount of catalyst was increased (Table 2). No significant results were obtained. Indeed, going from 2% (entry 2) to 5% (entry 3) and 10% *w*/*w* (entry 4), the yields did not substantially improve. As expected, in the absence of catalyst (entry 1) and in the presence of a metal-free catalyst (CNS, entry 5), no conversion was observed. According to these results, the 2% *w*/*w* catalyst loading was taken as the optimal condition. It is important to highlight that according to the measured loading of palladium on CNS, the 2% *w*/*w* of catalyst corresponds to a very low 0.45% *w*/*w* amount of Pd, making this system very efficient in terms of metal loading in the reaction.

palladium on CNS, the 2% *w*/*w* of catalyst corresponds to a very low 0.45% *w*/*w* amount of Pd, making this system very efficient in terms of metal loading in the reaction. The role of the base was also considered, and different inorganic bases, namely, NaOAc, Na2CO3, K2CO3, and KOH, were tested in these conditions. Results are reported in Table 3.


**Table 2.** Catalyst loading optimization <sup>a</sup> .

<sup>a</sup> Reaction conditions: 0.268 mmol of **1a** (1 eq), 0.161 mmol of TBAB (0.6 eq), 0.563 mmol of KOH (2 eq), and the listed amount of catalyst in 2.5 mL of water under MW irradiation (T = 100 ◦C, 30 min). The *w*/*w* percentage of catalyst is referred to as **1a**. <sup>b</sup> metal free CNS was used.

**Table 3.** Role of the base in the reaction <sup>a</sup> .


<sup>a</sup> Reaction conditions: 0.268 mmol of **1a** (1 eq), 0.161 mmol of TBAB (0.6 eq), 0.563 mmol (2 eq) of base, and 2% *w*/*w* of catalyst in 2.5 mL of water under MW irradiation (T = 100 ◦C, 30 min).

The results obtained show that, in base-free conditions (entry 1), the pH is probably not alkaline enough to promote the reduction of Pd (II) to Pd (0), thus obtaining poor yields [49]. With the use of NaOAc, Na2CO3, or K2CO<sup>3</sup> (entry 2, 3, and 4), an adequate alkaline environment could not be provided, with a consequent unsatisfactory activation of the catalytic cycle, resulting again in low yields [49]. The only base that gave good results was the strong base KOH, able to give a yield of 92% (entry 5).

Finally, the use of the phase-transfer agent was explored. As the coupling reaction is run in water and under heterogeneous conditions, producing water-insoluble products from organic reagents, tetrabutylammonium bromide (TBAB) was added to the reaction mixture. Optimization of the amount of TBAB in the reaction was performed by gradually reducing it from the starting 0.6 to 0.15 equivalents, obtaining the results shown in Table 4.

**Table 4.** Role of the phase-transfer catalyst in the reaction <sup>a</sup> .


<sup>a</sup> Reaction conditions: 0.268 mmol of 1a (1 eq), TBAB, 0.563 mmol (2 eq) of KOH, and 2% *w*/*w* of catalyst in 2.5 mL of water under MW irradiation (T = 100 ◦C, 30 min). <sup>b</sup> A 10<sup>×</sup> scale-up of the reference reaction.

The results underline that the reaction yield was somewhat higher when using small amounts of TBAB (97% with 0.15 eq, entry 3). It can be also noticed that, without using TBAB, acceptable yields can be obtained (90%, entry 4). The role of TBAB turned out to be crucial when scaling up the reaction. In fact, when performing a ×10 reaction scale-up (entry 5 and 6), the presence of TBAB was essential in increasing the efficiency of the reaction in the presence of large quantities of polar solvent. In this case, without the use of TBAB, a low 60% yield was indeed obtained.

After all this screening, the optimized conditions were defined as follows: 100 ◦C, 30 min, 2% *w*/*w* catalyst, corresponding to 0.45% *w*/*w* of palladium, KOH as base, and 0.15 equivalents of TBAB. For the model Suzuki–Miyaura reaction, in these conditions <sup>a</sup> TON = 1.2 <sup>×</sup> <sup>10</sup><sup>2</sup> and a TOF = 6.5 <sup>×</sup> <sup>10</sup>−<sup>2</sup> s −1 could be calculated. These values are

reasonably good and comparable to many industrial catalytic processes, but are obtained by operating in the presence of an eco-friendly cellulose-derived palladium support. In fact, while we are aware that higher TONs can be obtained under homogeneous conditions, these values are comparable or even better than those of most of the examples reported in the literature for fixing Pd on heterogeneous supports [16–18,20–23]. Moreover, in this case, the choice of the support falls on a bio-based system derived from waste biomass.

To verify the reliability of this catalytic system, a complete mass recovery of the reaction after column chromatography purification was performed. Reaction conditions, work-up, and purification are described in detail in the Section 4 and with this test we were able to verify that, in the optimized conditions after the purification process, a 95% yield could be achieved, calculated by weighting the white crystalline powder obtained after the whole process.

#### *2.3. Suzuki–Miyaura Substrate Scope Gels* **2022**, *8*, x FOR PEER REVIEW 9 of 16

After the definition of the optimized conditions, different substrates were tested in the reaction. All the experiments are reported in Table 5.

**Table 5.** Suzuki–Miyaura cross-couplings of ArX and PhB(OH)<sup>2</sup> catalyzed by CNS-Pd <sup>a</sup> . **Table 5.** Suzuki–Miyaura cross-couplings of ArX and PhB(OH)<sup>2</sup> catalyzed by CNS-Pd.


10 **1h** -CN -H -H -H -Br 99 11 **1i** -H -H -H -H -Cl 55 <sup>a</sup>Reaction conditions: 0.268 mmol of **1a** (1 eq), TBAB (0.15 eq), 0.563 mmol (2 eq) of KOH, and 2% <sup>a</sup> Reaction conditions: 0.268 mmol of **1a** (1 eq), TBAB (0.15 eq), 0.563 mmol (2 eq) of KOH, and 2% *w*/*w* of catalyst in 2.5 mL of water under MW irradiation (T = 100 ◦C, 30 min). <sup>b</sup> A 0.6 eq value of TBAB. <sup>c</sup> A 10% *w*/*w* value of catalyst.

*w*/*w* of catalyst in 2.5 mL of water under MW irradiation (T = 100 °C, 30 min). <sup>b</sup> A 0.6 eq value of TBAB. <sup>c</sup> A 10% *w*/*w* value of catalyst. *2.4. Multiple Cross-Coupling Suzuki–Miyaura Reactions* After studying the effect of substituents, we then performed tests to confirm that CNS-Pd could also be used for multiple couplings. For this study, we selected 1,4-dibromobenzene (Scheme 3) and 1,3,5-tribromobenzene (Scheme 4) as organ halides, increasing the amount of phenylboronic acid (2.5 equivalents for **1j** and 3.5 equivalents for **1k**) and An early analysis on the position of the methoxyl substituent was carried out. As can be seen from the results (entry 1, 2, and 3), with the optimized reaction conditions it was possible to obtain excellent yields with the substituent in para- and meta-position. In the case of the methoxy substituent in the ortho- position (entry 3), we observed a decrease in yield, probably due to the steric effect of the substituent in the position adjacent to the halogen involved in the reaction mechanism. The test with more catalyst (10% *w*/*w*, entry 4) further confirmed the low conversion, supporting the hypothesis of the steric effect of the substituent.

slightly changing the reaction conditions (reaction time increased to 1.5 h for **1j** and to 3 h for **1k**). After the position analysis, a study with structurally different substituents was then performed. In the absence of substituents on the benzene ring of the organ halide (entry 8), a satisfactory yield of more than 90% was obtained. In the presence of electron-donating activating substituents on the benzene ring in para-position, such as -CH<sup>3</sup> (entry 5) and -NH<sup>2</sup> (entry 9), good yields were achieved, in particular a yield of 72% with p-Br-toluene and 76% with p-Br-aniline. In the presence of an electron-withdrawing group, such as

In these reaction conditions, we were able to obtain high selectivity (>95%) towards the product **4j** in the reaction with 1,4-dibromobenzene (Scheme 3), reaching a conversion of 55%. Increasing the phenylboronic acid amount (from 2.5 to 5 equivalents compared to **1j**) and maintaining the same reaction conditions (1.5 h, 30 °C), we were able to reach a

In the reaction with 1,3,5-tribromobenzene (Scheme 4), under the conditions of 3 h, at 100 °C, with 3.5 equivalents of **2** compared with **1k** reagent, we obtained a very high selectivity (>95%) towards product **5k**, with a conversion of 82%, confirming the catalytic

**Scheme 3.** Reaction of double cross-coupling with 1,4-dibromobenzene.

efficiency of CNS-Pd also in multiple cross-coupling reactions.

very high conversion (90%) without losing selectivity toward product **4j**.


**Entry ArX R R' R″ R‴ X Conversion (%) 1a** -H -OCH<sup>3</sup> -H -H -Br 99 **1b** -H -H -OCH<sup>3</sup> -H -Br 98 **1c** -H -H -H -OCH<sup>3</sup> -Br 54

#### *2.4. Multiple Cross-Coupling Suzuki–Miyaura Reactions 2.4. Multiple Cross-Coupling Suzuki–Miyaura Reactions*

*Gels* **2022**, *8*, x FOR PEER REVIEW 9 of 16

**Table 5.** Suzuki–Miyaura cross-couplings of ArX and PhB(OH)<sup>2</sup> catalyzed by CNS-Pd.

After studying the effect of substituents, we then performed tests to confirm that CNS-Pd could also be used for multiple couplings. For this study, we selected 1,4 dibromobenzene (Scheme 3) and 1,3,5-tribromobenzene (Scheme 4) as organ halides, increasing the amount of phenylboronic acid (2.5 equivalents for **1j** and 3.5 equivalents for **1k**) and slightly changing the reaction conditions (reaction time increased to 1.5 h for **1j** and to 3 h for **1k**). After studying the effect of substituents, we then performed tests to confirm that CNS-Pd could also be used for multiple couplings. For this study, we selected 1,4-dibromobenzene (Scheme 3) and 1,3,5-tribromobenzene (Scheme 4) as organ halides, increasing the amount of phenylboronic acid (2.5 equivalents for **1j** and 3.5 equivalents for **1k**) and slightly changing the reaction conditions (reaction time increased to 1.5 h for **1j** and to 3 h for **1k**).

**Scheme 3.** Reaction of double cross-coupling with 1,4-dibromobenzene. **Scheme 3.** Reaction of double cross-coupling with 1,4-dibromobenzene.

**Scheme 4.** Reaction of triple cross-coupling with 1,3,5-tribromobenzene. **Scheme 4.** Reaction of triple cross-coupling with 1,3,5-tribromobenzene.

*2.5. Leaching Tests* Tests were performed to evaluate the leaching of Pd from the CNS-Pd catalyst under the reaction conditions (see Section 4). The amount of Pd in solution before and after the reaction was evaluated by means of ICP-OES analysis. After one reaction cycle, only 7% In these reaction conditions, we were able to obtain high selectivity (>95%) towards the product **4j** in the reaction with 1,4-dibromobenzene (Scheme 3), reaching a conversion of 55%. Increasing the phenylboronic acid amount (from 2.5 to 5 equivalents compared to **1j**) and maintaining the same reaction conditions (1.5 h, 30 ◦C), we were able to reach a very high conversion (90%) without losing selectivity toward product **4j**.

of the initial amount of metal (0.016 mg when 1 mg of CNS-Pd is used) was released in the solution. No significant metal loss was observed after a second cycle, thus confirming that, after stabilization, Pd remains anchored on the heterogeneous support, paving the way for the possible recycling of CNS-Pd. In the reaction with 1,3,5-tribromobenzene (Scheme 4), under the conditions of 3 h, at 100 ◦C, with 3.5 equivalents of **2** compared with **1k** reagent, we obtained a very high selectivity (>95%) towards product **5k**, with a conversion of 82%, confirming the catalytic efficiency of CNS-Pd also in multiple cross-coupling reactions.

#### *2.6. Recyclability Tests 2.5. Leaching Tests*

life.

**3. Conclusions**

To evaluate the possibility of reusing CNS-Pd several times without losing catalytic activity, reusability tests were performed between **1a** and **2** to give **3a** under the optimized conditions. Five consecutive reaction runs were realized and, after each cycle, the reaction Tests were performed to evaluate the leaching of Pd from the CNS-Pd catalyst under the reaction conditions (see Section 4). The amount of Pd in solution before and after the reaction was evaluated by means of ICP-OES analysis. After one reaction cycle, only 7% of

yield was evaluated by 1H-NMR with acetonitrile as internal standard to evaluate the ef-

achieved without losing selectivity towards product **3a**, demonstrating the long catalyst

(a) Reaction conditions: 0.268 mmol of **1a** (1 eq), TBAB (0.15 eq), 0.563 mmol (2 eq) of KOH, and 2%

In this work, cellulose-based nanosponges (CNS) were prepared. TEMPO-oxidized cellulose nanofibers and branched polyethyleneimine were cross-linked together in the presence of citric acid to obtain a micro- and nanoporous structure and this latter was then loaded with Pd(II) to obtain a potential heterogeneous catalyst for Suzuki–Miyaura coupling reactions. The morphology and structure of the material were characterized by various SEM–EDS and ICP-OES analyses. The CNS-Pd system was also thoroughly investigated by XPS analysis to evaluate the behavior of the metal in the catalytic activity. An optimization study on the heterogeneous reaction was conducted, reaching optimal condition for the obtainment of high-rate yields. The catalytic recycling ability of the material

**Entry Cycle Number Yield (%)** 1 I 97 2 II 99 3 III 98 4 IV 99 5 V 96

As can be observed, CNS-Pd can be considered a heterogeneous catalyst with good

**Table 6.** Results from Recyclability tests <sup>a</sup>

ficiency of the reused catalytic system. Results are reported in Table 6.

.

*w*/*w* of catalyst in 2.5 mL of water under MW irradiation (T = 100 °C, 30 min).

the initial amount of metal (0.016 mg when 1 mg of CNS-Pd is used) was released in the solution. No significant metal loss was observed after a second cycle, thus confirming that, after stabilization, Pd remains anchored on the heterogeneous support, paving the way for the possible recycling of CNS-Pd.

### *2.6. Recyclability Tests*

To evaluate the possibility of reusing CNS-Pd several times without losing catalytic activity, reusability tests were performed between **1a** and **2** to give **3a** under the optimized conditions. Five consecutive reaction runs were realized and, after each cycle, the reaction yield was evaluated by <sup>1</sup>H-NMR with acetonitrile as internal standard to evaluate the efficiency of the reused catalytic system. Results are reported in Table 6.

**Table 6.** Results from Recyclability tests <sup>a</sup> .


<sup>a</sup> Reaction conditions: 0.268 mmol of **1a** (1 eq), TBAB (0.15 eq), 0.563 mmol (2 eq) of KOH, and 2% *w*/*w* of catalyst in 2.5 mL of water under MW irradiation (T = 100 ◦C, 30 min).

As can be observed, CNS-Pd can be considered a heterogeneous catalyst with good reusability characteristics, as, after five cycles, more than 96% conversion was still achieved without losing selectivity towards product **3a**, demonstrating the long catalyst life.

### **3. Conclusions**

In this work, cellulose-based nanosponges (CNS) were prepared. TEMPO-oxidized cellulose nanofibers and branched polyethyleneimine were cross-linked together in the presence of citric acid to obtain a micro- and nanoporous structure and this latter was then loaded with Pd(II) to obtain a potential heterogeneous catalyst for Suzuki–Miyaura coupling reactions. The morphology and structure of the material were characterized by various SEM–EDS and ICP-OES analyses. The CNS-Pd system was also thoroughly investigated by XPS analysis to evaluate the behavior of the metal in the catalytic activity. An optimization study on the heterogeneous reaction was conducted, reaching optimal condition for the obtainment of high-rate yields. The catalytic recycling ability of the material and the catalytic effect for different substituents were also examined, confirming the possibility of reusing this sustainable catalyst several times.

### **4. Materials and Methods**

All of the reagents were purchased from Merck (Darmstadt, Germany). Cotton linter was obtained from Bartoli paper factory (Capannori, Lucca, Italy). Deionized water was produced within the laboratories with a Millipore Elix® Deionizer with Progard® S2 ion exchange resins (Merck KGaA, Darmstadt, Germany). All 1H-NMR spectra were recorded on a 400 MHz Brüker (Billerica, MA, USA) NMR spectrometer. Microwave reactions were conducted in a Biotage® Initiator+ (Uppsala, Sweden). Other equipment used in the procedures includes a Branson SFX250 Sonicator (Emerson Electric Co., Ferguson, MI, USA), a SP Scientific BenchTop Pro Lyophilizer (SP INDUSTRIES, 935 Mearns Road, Warminster, UK), a Büchi Rotavapor® R-124 8 (Flawil, Switzerland), and a Thermotest Mazzali laboratory oven (Monza, Italy). Scanning electron microscopy (SEM) was performed using a variable pressure instrument (SEM Cambridge Stereoscan 360) at 100/120 pA with a detector BSD. The operating voltage was 15 kV with an electron beam current intensity of 100 pA. The focal distance was 9 mm. The EDS analysis was performed using a Bruker Quantax 200 6/30 instrument (Billerica, MA, USA). The metal concentrations were measured by

ICPOES atomic emission spectroscopy using a Perkin Elmer Optima 3000 SD spectrometer (Wellesley, MA, USA).

### *4.1. TEMPO-Oxidized Cellulose (TOC) Production and Titration and Synthesis of Cellulose NanoSponges (CNS)*

TEMPO-Oxidized Cellulose (TOC) production and titration were performed according to a procedure previously reported in the literature [34,50,51]. After TOC synthesis and characterization, TEMPO-Oxidized Cellulose Nanofibers (TOCNF) for the production of CNS were produced by suspending 3.5 g of TOC in 0.14 L of deionized water, adding 0.210 g of NaOH pellets. The suspension was ultrasonicated with a Branson SFX250 Sonicator, achieving a homogeneous suspension of TOCNF. This latter was then acidified with 12 N HCl (2 mL), filtered under vacuum on a Büchner funnel, and washed with deionized water (250 mL) reaching neutral pH. Then, two aqueous solutions of anhydrous citric acid (CA) (1.792 g in 10 mL) and 25 kDa branched polyethylenimine (bPEI) (7 g of bPEI in 10 mL) were mixed with the TOCNF solution, while continuously stirring until reaching a white and homogeneous hydrogel and obtaining a final TOCNF concentration of 3% *w*/*w*. The mixture was then placed in well plates used as molds, quickly frozen at −20 ◦C, frozen-dried for 48 h at −52 ◦C, 140 µbar using an SP Scientific BenchTop Pro Lyophilizer, and finally thermally treated in the laboratory oven performing a heating ramp from a temperature of 55 ◦C to a maximum of 98 ◦C. This temperature was kept for 16 h. At the end of the process, CNS was ground with a mortar and then washed with deionized water (500 mL) to remove the excess bPEI. This procedure for the synthesis of CNS was described in our previous works [37,38].

### *4.2. Preparation of the Catalyst*

After washing, 275 mg of CNS was ground into a fine powder. In the meantime, a saturated acidic solution of PdCl<sup>2</sup> (101 mg in 50 mL of 0.1 M HCl) was prepared. Ground CNS was soaked in this solution until complete adsorption of Pd (10 min), repeating the adsorption cycle three times in the same conditions. After every sorption cycle, Pd-loaded CNS was filtered on a Buchner funnel, washed with 100 mL of deionized water and 50 mL of ethanol, and, finally, left to dry in open air, obtaining the catalyst CNS-Pd.

### *4.3. Catalyst Characterization*

Scanning Electron Microscopy (SEM) was performed using a variable-pressure instrument (SEM Cambridge Stereoscan 360) at 100/120 Pa with a detector VPSE. The operating voltage was 20 kV with an electron beam current intensity of 150 pA. The focal distance was 8 mm. The specimens were analyzed in High Vacuum mode after metallization.

EDS analysis was performed using a Bruker Quantax 200 6/30 instrument.

Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) analysis was performed on solid materials to determine the percentage of palladium in CNS-Pd. ICP-OES was conducted using a Perkin Elmer Optima 3000 SD spectrometer and the samples were treated with nitric acid (HNO3) to completely dissolve the organic portion of the material before the determination of the metal concentration.

X-ray photoelectron spectroscopy (XPS) measurements were carried out using a custom designed spectrometer, described in previous studies [52] and equipped with a nonmonochromatized Mg Ka X-ray source (1253.6 eV pass energy 25 eV, step 0.1 eV). For this experiment, photoelectrons emitted by C1s, O1s, N1s, Pd3d, Cl2p, Br3d, and B1s core levels were detected on powder samples of the pristine and recovered CNS-Pd catalyzer. All spectra were energy-referenced to the C 1s signal of aliphatic C atoms at 285.00 eV binding energy (BE) [53]. Atomic ratios were calculated from peak intensities using Scofield's cross-section values [54]. Curve-fitting analysis was performed using Gaussian profiles as fitting functions after subtraction of a polynomial background. For qualitative data, the BE values were mostly referred to from the NIST database [43].

### *4.4. Suzuki–Miyaura Reaction*

In a 5 mL microwave vial, CNS-Pd (1–5 mg, 2–10% *w*/*w*), KOH (0.536 mmol, 2 eq), phenylboronic acid (0.402 mmol, 1.5 eq), and 2.5 mL of water as solvent were added. Then, tetrabutylammonium bromide (TBAB) (0.0402–0.1608 mmol, 0.15–0.6 eq) and reagent **1a**–**k** (0.268 mmol, 1 eq) were put in the vial (all the percentages and equivalents are referred to as reagents **1a**–**k**). The reaction was performed under microwave irradiation in a temperature range of 40–130 ◦C and for reaction times between 10 min and 3 h.

### *4.5. Reaction Work-Up and Product Purification*

At the end of the reaction, 2 mL of ethyl acetate was added into the reaction mixture and stirred for 10 min. The solution was then filtered in a glass straw equipped with cotton to remove the catalyst and the reaction vial was washed three times with 2 mL of water and three times with 2 mL of ethyl acetate. The organic and aqueous phases were transferred in an extractor funnel and 5 mL of 0.1 N HCl solution was added. The aqueous phase was then extracted three times with 15 mL of ethyl acetate and all the organic phases were collected together, and then washed once with 10 mL of NaOH 0.1 N and twice with 10 mL of a saturated solution of NaCl. Then, the organic phase was collected and anhydrified with Na2SO4. Lastly, the final organic solution was then filtered off and the solvent was removed under vacuum to obtain a crude for the NMR analysis. The purification of the product was performed with flash column chromatography using a solvent mixture of hexane and ethyl acetate in a 95:5 ratio.

Yields were calculated by means of <sup>1</sup>H-NMR analysis using acetonitrile (ACN) as internal standard. The analysis was performed on the crude reaction without purification process. For the standard reaction (4-Br-anisole as reagent **A**), the yield calculated by internal standard technique with the <sup>1</sup>H-NMR was further confirmed with the yield obtained with mass recovery after purification with column chromatography, as described in the previous paragraph.

### *4.6. Leaching Test*

In a 20 mL microwave vial, CNS-Pd (12 mg), KOH (360 mg), and 15 mL of water as solvent were added. Then, TBAB (312 mg) was put in the vial. The reaction was performed under microwave irradiation at 100 ◦C for 30 min. Once finished, the reaction was filtered off on a Buchner funnel to remove the solid and the solution was analyzed through ICP-OES analysis to quantify the Pd released in the reaction environment.

### *4.7. Reusability Test*

Re-usability tests were conducted on the reference reaction with *p*-Br-anisole (Scheme 2). In a 5 mL microwave vial, CNS-Pd (10 mg, 10% *w*/*w*), KOH (1.072 mmol, 2 eq), phenylboronic acid (98 mg), and 5 mL of water as solvent were added. Then, TBAB (0.804 mmol, 1.5 eq) and *p*-Br-anisole (0.536 mmol, 1 eq) were put in the vial. The reaction was performed under microwave irradiation at 100 ◦C for 30 min. At the end of the reaction time, the vial was centrifuged, and the reaction solvent was removed. Three cycles of washing of CNS-Pd inside the vial were then carried out as described: a 5 mL measure of AcOEt was added into the reaction vial, and the solution was left stirring for 10 min and subsequently centrifuged. The supernatant was removed, taking care not to remove the catalyst, and the washing was repeated twice more. At the end of the washing passages, CNS-Pd was allowed to dry inside the vial, and this latter was then used for a new reaction cycle. This procedure was repeated five times.

### *4.8. NMR Product Characterization*

**Internal Standard (ACN):** <sup>1</sup>H NMR (400 MHz, CDCl3) δ 2.04 (s, 3H); **Product 3a (4- Methoxybiphenyl):** <sup>1</sup>H NMR (400 MHz, CDCl3) δ 7.55–7.50 (m, 4H), 7.40 (t, *J* = 7.5, 2H), 7.29 (t, *J* = 7.3, 1H), 6.97 (d, *J* = 8.8, 2H), 3.84 (s, 3H) [55]; **Product 3b (3-Methoxybiphenyl):** <sup>1</sup>H NMR (400 MHz, CDCl3) δ 7.61 (d, *J* = 7.5, 2H), 7.45 (t, *J* = 7.6, 2H), 7.39–7.35 (m, 2H), 7.20 (d, *J* = 7.7, 1H), 7.15 (t, *J* = 2.0, 1H), 6.92 (dd, *J* = 8.2, 2.3, 1H), 3.87 (s, 3H) [56]; **Product 3c (2-Methoxybiphenyl):** <sup>1</sup>H NMR (400 MHz, CDCl3) δ 7.54 (d, *J* = 8.3, 2H), 7.41 (t, *J* = 7.6, 2H), 7.34–7.31 (m, 3H), 7.03 (t, *J* = 7.5, 1H), 6.99 (d, *J* = 8.5, 1H), 3.81 (s, 3H) [56]; **Product 3d (4-Methylbiphenyl):** <sup>1</sup>H NMR (400 MHz, CDCl3): δ = 7.64 (d, *J* = 8.0 Hz, 2H), 7.54 (d, *J* = 8.0 Hz, 2H),7.47 (t, *J* = 7.2, 2H), 7.37 (t, *J* = 7.2 Hz, 1H), 7.28 (d, *J* = 8.0 Hz, 2H), 2.44 (s, 3H) [55]; **Product 3e (Biphenyl-4-carbaldehyde):** <sup>1</sup>H NMR (400 MHz, CDCl3): δ = 9.96 (s, 1H), 7.86 (d, *J* = 8.0 Hz, 2H), 7.66 (d, *J* = 8.0 Hz, 2H), 7.55 (d, *J* = 8.4 Hz, 2H), 7.40 (t, *J* = 7.2 Hz, 2H), 7.34 (t, *J* = 7.2 Hz, 1H) [55]; **Product 3f (Biphenyl):** <sup>1</sup>H NMR (400 MHz, CDCl3): δ = 7.54 (d, *J* = 8.0 Hz, 4H), 7.39 (t, *J* = 8.0 Hz, 4H), 7.31–7.27 (m, 2H) [55]; **Product 3g (4-Aminobiphenyl):** <sup>1</sup>H NMR (400 MHz, CDCl3, TMS) δ 7.52 (d, *J* = 7.9 Hz, 2H, CH), 7.41–7.36 (m, 4H, CH), 7.25 (t, *J* = 7.3 Hz, 1H, CH), 6.73 (d, *J* = 8.5 Hz, 2H, CH), 3.68 (s, 2H) [57]; **Product 3h (2-Cynobiphenyl):** <sup>1</sup>H NMR (400 MHz, CDCl3): δ (ppm) 7.76 (dd, *J* = 8.0Hz, 1.2Hz, 1H), 7.64 (td, *J* = 7.6 Hz, 1.2Hz, 1H), 7.58–7.54 (m, 2H), 7.54–7.40 (m, 5H) [58]; **Product 4j (1,1**0 **:4**0 **,1"-terphenyl):** <sup>1</sup>H NMR (400 MHz, Chloroform-d) δ 7.70 (s, 4H), 7.69–7.64 (m, 4H), 7.50–7.45 (m, 4H), 7.41–7.35 (m, 2H) ppm [43]; **Product 5k (1,3,5- Triphenylbenzene):** <sup>1</sup>H NMR (400 MHz, CDCl3): δ = 7.71 (s, 3H), 7.63 (d, *J* = 7.2 Hz, 6H), 7.42 (t, *J* = 8.0 Hz, 6H), 7.31 (t, *J* = 6.8 Hz, 3H) [59].

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/gels8120789/s1. Table S1: XPS data analysis results: core-level binding energy (BE), full width at half-maxima (FWHM) values, internal atomic ratios, and proposed assignments. Reference [60] is cited in the supplementary materials.

**Author Contributions:** Conceptualization, C.P. and A.S.; methodology, L.R., G.N., C.B. and M.L.; validation, L.R., G.N., C.B. and M.L.; formal analysis, investigation, L.R., G.N., C.B. and M.L.; resources, A.S. and C.P.; data curation, A.S.; writing—original draft preparation, L.R.; writing review and editing, C.P., A.S. and C.B.; supervision, A.S. and C.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**

