Eco-friendly Development of New Biodegradable and Renewable Polymers Based on Di(meth)Acrylated and Acrylamidated Monomers Derived from Limonene Dioxide

Abstract

In this work, new biobased polymeric materials were synthesized using interesting limonene-based multifunctional monomers as building blocks: the tetrafunctional 2,8-dihydroxy-1,9-diacrylate and 2,8-dihydroxy-1,9-dimethacrylate, and the trifunctional 2-hydroxy-1-acrylamide. These monomers were prepared by the addition reaction between the diepoxidized limonene and (meth)acrylic acid or N-hydroxyethyl acrylamide. The complete conversion of unsaturations of limonene into epoxides, as well as the formation of monomers, were confirmed by Nuclear Magnetic Resonance analysis of proton and carbon (¹H and ¹³C NMR), Fourier Transform Infrared Spectroscopy (FTIR) and Mass Spectrometry (ESI-MS). The monomeric products were polymerized via miniemulsion polymerization with different initiators and co-stabilizers, resulting in new poly(meth)acrylates and polyamide polymers, which is a hydrogel. The polymers showed a high degree of crosslinking, porosity and good thermal properties characterized by Scanning Electron Microscopy (SEM), FTIR, Differential Scanning Calorimetry (DSC) and Thermogravimetric Analysis (TGA). In addition, the polymeric materials were tested to evaluate the formation of biofilm by the action of Comomanoas sp, whose results indicated that the synthesized new biobased polymers are susceptible to biodegradation.

Graphical Abstract

Introduction

The search for alternatives to replace, even partially, oil-based materials has increased due to the worsening of environmental problems such as visual pollution of accumulated polymeric garbage, atmospheric pollution and, the finite amounts of available petroleum [1,2,3,4]. Renewable polymers stand out, mostly, because they are based on biomass raw material, which is renewed through solar energy, reducing the CO₂ extra emission in the atmosphere [5, 6]. Allied with this, it is possible to implement cleaner synthetic chemical processes in renewable polymer production [1, 2, 4, 5].

Among renewable sources, the use of terpenes has been highlighted [5]. Limonene is a commercially accessible terpene acquired as a by-product of the reused industrial waste from citrus manufacturing worldwide. This low-cost and highly available natural product is a convenient greener raw material to be exploited in scientific research [5, 7, 8]. Currently, there is a scientific focus on the synthesis and polymerization of monomers derived from limonene oxides. This compounds can be prepared from the conversion of the limonene unsaturations into epoxides, carried out under mild experimental conditions, at room temperature (≅ 25 °C), and that do not necessarily involve the use of initiators, catalysts, and hazardous organic solvents [5, 9,10,11]. The presence of the oxirane ring makes the limonene oxide promptly reactive to be directly polymerized or to allow the entry of a variety of nucleophilic compounds in a successive step [1, 5, 11,12,13]. The diepoxide enables the occurrence of polymeric reactions at both oxidized ends of the terpenoid structure, generating stereocomplexed or amorphous polymers, and studies with this terpenoid as a starting material are still in full development and many possibilities can be explored [1, 5, 11, 12, 14,15,16].

(Meth)acrylation is a reaction to add radically polymerizable groups, from an methacrylic acid (MA) or acrylic acid (AA), to the structures of various molecules in order to make them potent (meth)acrylate monomers [17]. Coupling methods of AA and MA in terpenes are already widely seen in the literature, and there are different routes and methodologies [16,17,18]. However, the (meth)acrylation via the epoxide ring opening with acrylic acid or methacrylic acid, aimed at producing a β-hydroxyethyl (meth)acrylate derivative monomer, was reported in recent studies for different terpenes. The opening of the oxirane ring allows the entry of the acrylate functional group at one end while an hydroxyl (OH) group is formed at the other end, constituting at least in a bifunctional monomer, which can lead to the formation of cross-linked bio-based polymers or copolymers [17,18,19,20,21].

Biobased monomers functionalized with acrylamides (ACM) are little explored [17]. The few acrylamidations in biomasses carried out so far are focused on modifications of vegetable oils, lignin, proteins and carbohydrates [17, 18, 22, 23]. N-hydroxyethyl acrylamide (HEAA), for example, is a low-cost and readily available ACM. It has a hydroxyl group in its structure that facilitates coupling to functional groups of other organic molecules [17]. In the available literature, acrylamidations applied directly to terpenes or epoxidized terpenes by ring-opening mechanism were not easily found. However, HEAA has been applied as a chain transfer agent in ring opening reactions in L-lactide/ε-caprolactone to couple in these compounds and form acrylamide macromonomers. Which indicates that HEAA, may have nucleophilic potential to be inserted by ring opening mechanism in other types of compounds, such as epoxidized ones [23,24,25].

The potential formation of multifunctional monomers through the addition of acrylate groups to terpene oxides can increase the entry points for specific and previously thought chemical modifications to construct custom-made polymeric materials and/or Bio-based Chemical Building Blocks (BCBB) from biomass terpenic [6, 26,27,28]. Additionally, this strategy to synthesize multifunctional monomers agrees with the Green Chemistry concepts, since it is based on simplified reaction systems, with reduced number of reagents, avoiding not-renewable organic solvents, and toxic reagents/catalysts. Such characteristics contrast with synthetic strategies already reported in literature for terpene modification, such as hydroboration-oxidation reactions followed by esterification or thiol-ene reactions of click chemistry [1, 5, 16, 29, 30].

Through miniemulsion polymerization, bio-based materials have been produced with more ecological synthetic routes and with advantageous characteristics. Some of them are the drivers of the size and morphology of the newly formed polymeric particle and the possibility of polymerizing highly viscous monomers, commonly observed in monomers that have hydroxyl groups. Furthermore, organosoluble and water-soluble initiators can be used in the polymeric reaction [17, 31,32,33,34]. One of the main advantages offered by the miniemulsion in terms of sustainability is the aqueous phase as a continuous medium instead of organic solvents [32, 35, 36]. Another advantage is the possibility of replacing fossil components used to prepare the miniemulsion with renewable ones. For example, medium-chain esters comprising capric/caprylic acid triglycerides were employed as part of a renewable co-stabilizers consisting of a mixture of free fatty acids, usually derived from vegetable oils, such as coconut or palm oil [37,38,39,40]. Furthermore, despite the aforementioned advantages of the miniemulsion technique, this heterogeneous system is still rarely applied to polymerize monomers derived from terpenes. For this purpose, techniques such as photopolymerization, mass polymerization and solution polymerization can be easily found in recent literature [19, 20, 41].

In order to promote a sustainable cycle, the formation of biofilms from renewable materials becomes an attractive alternative to put forward biodegradability. Biofilms are communities of microorganisms connected to materials by adhesion, which consists of bacteria embedded in a matrix constituted by carbohydrates such as polysaccharides, structural proteins, and, to a lesser extent, extracellular DNA [42]. The biofilms can be formed by a single population or by multiple populations forming a community, which allows for group communication end an architectural infrastructure to promote survival [43]. Biofilm formation is influenced by factors such as species within the community, metabolic network, and quorum-sensing signalling through extracellular DNA and other metabolites [42]. A variety of microorganisms have been described as biofilm formers, both as individuals and as communities that occur on different surfaces of biotic and abiotic materials. Comamonas, a Gram-negative bacterium, is a genus already known to be abundant in biofilms in the environment. Studies have shown that this genus has a broad metabolic capacity and uses various substances as nutrients, such as several organic compounds, and its possible use in bioremediation process has been described [44,45,46]. In a previous study by Peixoto et al. [44], Comamonas sp. was isolated in the Brazilian savanna in deposited plastic waste, and its ability to biodegrade a recalcitrant polymer such as polyethylene was described.

Thus, the present work focuses on the valorization of limonene by its epoxidation, intended for the production of renewable monomers (di(meth)acrylated and acrylamidated), followed by their polymerization. For this, the diepoxidized limonene undergone catalyst-free acrylation (with AA), methacrylation (with MA) and acrylamidation (with HEAA) reactions to form interesting sustainable polymer building blocks. To the best of our knowledge, for the first time, the acquired monomers were homopolymerized via miniemulsion polymerization and under different reaction conditions. The present work also displays as differential the use of in situ and real-time FTIR tools for monitoring the chemical modification reactions to obtain the renewable monomers. Alongside comprehensive structural assignment analysis, this approach allowed for the exploration of the formation mechanisms of the polymeric structures. Three unique polymeric materials were prepared and fully characterized to investigate its physical and chemical properties. These materials were then inoculated with the bacteria Comamonas sp. (isolated from the Brazilian Savanna) in an unprecedented manner, aiming to evaluate the susceptibility of these new polymers to biodegradation.

Experimental Section

Materials

The reagents used in the experimental section were: Sodium Bicarbonate (Sigma-Aldrich Brasil Ltda, SP, Brazil, 99.5%); Acetone (Vetec Quimica Fina, 99.5%); Magnesium Sulfate (Dynamic, 98%); Diethyl Ether (Dynamic, 98%); Acrylic Acid, AA (Sigma-Aldrich, 99%); Methacrylic acid, MA (Sigma-Aldrich, 99%); N-Hydroxyethyl acrylamide, HEAA (Sigma-Aldrich, 97%); Hydroquinone (Vetec Quimica Fina, RJ, Brazil); Hexadecane (Sigma-Aldrich, 99%); Potassium Persulfate, KPS (SigmaAldrich, 99%); 2,2’-Azobis(2-methylpropionitrile), AIBN (Sigma-Aldrich, 98%); Sodium Dodecyl Sulfate, SDS (Sigma-Aldrich, 99%); Distilled water; Anhydrous Methanol (Sigma-Aldrich, 99.8%); Anhydrous Tetrahydrofuran, THF (Sigma-Aldrich, 99.9%); Chloroform (Dinâmica, 99.8%), Crodamol GTCC (Croda do Brasil, SP, Brazil).

Monomers Synthesis

Two successive modification steps of the limonene were performed to produce a bio-based monomer (Fig. 1): (1) limonene diepoxidation to affords the limonene diepoxide (LD) described in the Supplementary Material, and (2) the LD epoxide rings opening by AA, MA and HEAA. Independent to the nucleophile, the epoxide opening reactions were conducted following the same procedure. In a 100 mL three-neck flask connected to a reflux condenser chilled with ice water were added 5 g of the limonene diepoxide, the nucleophile (4.3–10.8 g of acrylic acid-AA or 5.2–18 g of methacrylic acid-MA or 6.9–10.3 g of N-hydroxyethyl acrylamide-HEAA) and hydroquinone as inhibitor (0.8–0.9 w% in relation to the total mass of the reaction components). The flask was then submerged in an oil bath and the reactions were started by increasing the reaction temperature (75–130 °C) under constant stirring at 1000 rpm. Table 1 summarizes the amounts of each component, temperature, and reaction time used to perform the syntheses. After the end of the (meth)acrylation reaction, a purification step was carried out. The products of interest were separated by liquid-liquid extraction with diethyl ether (60 mL) and then washed with a saturated sodium bicarbonate solution (three-four washes with 50 mL each) to remove the excess of unreacted acids. Residual water was removed by addition of anhydrous magnesium sulfate, which was subsequently filtered off. The final products were obtained free of solvent after rotatory evaporation. For the acrylamidation reaction, liquid-liquid extraction was performed with chloroform (60 mL) and distilled water (60 mL). Both phases (organic and aqueous) had the solvent removed and were analyzed separately. The (meth)acrylated and acrylamidated products were analyzed by FTIR, NMR (¹H and ¹³C) and mass spectrometry.

Fig. 1

General scheme of monomer synthesis with their chemical structures and abbreviated nomenclature

Table 1 Reaction information of monomers synthesis

Monomers Synthesis: in-situ and real-time monitoring

The synthesis of all monomers was monitored in real time with the FTIR probe ReactIR-15, according to syntheses no. 6, 6, and 3 referring to M1LD, M2LD and M3LD, respectively, and described in Table 1.

The experimental procedure for the FTIR monitored preparation of the monomers followed the one previously described for these syntheses, but with a sequential addition of the components to allow the individual identification of its FTIR trend. At the beginning of the acquisition, only the LD component was present in the reaction medium. Soon after, hydroquinone was added in one portion. The nucleophile was added in three equal portions (additions 1, 2 and 3) at different times (about 10 to 15 min between one addition and another). The first addition of nucleophile was done at around 40 °C for all experiments, and the temperature was gradually raised until reaching the temperature determined for each type of reaction.

Polymers Synthesis

The monomers M1LD (synthesis no. 6), M2LD (synthesis no. 6) and, M3LD (synthesis no. 3) were polymerized via miniemulsion polymerization reaction. The tests were carried out varying the proportion of organic phase (20–15 w%) and aqueous phase (80–85 w%), the co-stabilizer content in relation to the monomer (4–8 w%), the initiator species (AIBN and KPS) and the reaction temperature (65–80 °C) according to the initiator used. Therefore, the composition of the organic and aqueous phases is summarized below, and Table 2 shows the reaction characteristics of each test, as well as their related nomenclature.

• (i) Organic phase (OF): 1.5 g of monomer + 0.06–0.12 g of n-Hexadecane or Crodamol GTCC co-stabilizers + 0.0075 g of organosoluble initiator, AIBN (used only in tests with this initiator).

• (ii) Aqueous phase (AF): 6–8.5 g of water + 0.018–0.035 g of SDS surfactant (added below the CMC) + 0.0075 g of water-soluble initiator, KPS (used only in tests with this initiator).

The experiment occurred as follows, in a 25 mL beaker was added the organic phase and in another 25 mL beaker was added the aqueous phase, both remained under constant stirring for 10 min. Then, the two phases were mixed and kept under constant stirring for 20 min. Soon after, the organic phase was dispersed into the aqueous phase by using the Ultra-Turrax disperser for 5 min while stirring at 25,000 rpm coupled to an ice bath. After that, the mixture was sonicated for 5 min with 70% amplitude, also in an ice bath. The emulsified mixture went to the polymerization step in a 10 mL one-neck flask closed with a septum and, after purging nitrogen for 10 min inside the flask, the reaction was initiated with an increase in temperature. The total reaction time was 24 h. Finally, the final product was dried in an oven at 60 °C for 24 h. From P(M1LD)-T5, P(M2LD)-T1 and P(M3LD)-T1 samples, biofilm formation was implemented as described in the Support Information file (SI-Biofilm formation).

Table 2 Reaction recipes for the polymer’s synthesis

Characterization Methods

Vibrational spectroscopy data in the mid-infrared region was collected with the Mettler-Toledo ReactIR-15 equipment (Mettler Toledo) to follow in situ the LD and monomers syntheses. The equipment probe was immersed in the reaction medium. Spectral data were acquired in the sampling interval of 60 s (for epoxidation reaction) and 30 s (for acrylation, methacrylation and acrylamidation reactions) in the spectral region from 2500 cm^-1 to 650 cm^-1 at 8 wavenumber resolution by using a MCT Detector equipped with a Happ-Genzel apodization and DiComp (Diamond) probe connected via AgX 9.5 mm x 2 m Fiber (Silver Halide). The iC IR™ software was used for data acquisition and mathematical treatment of the spectral data.

In situ Fourier Transform Infrared Spectroscopy (FTIR)

Vibrational spectroscopy data in the mid-infrared region was collected with the Mettler-Toledo ReactIR-15 equipment (Mettler Toledo) to follow in situ the LD and monomers syntheses. The equipment probe was immersed in the reaction medium. Spectral data were acquired in the sampling interval of 60 s (for epoxidation reaction) and 30 s (for acrylation, methacrylation and acrylamidation reactions) in the spectral region from 2500 cm^-1 to 650 cm^-1 at 8 wavenumber resolution by using a MCT Detector equipped with a Happ-Genzel apodization and DiComp (Diamond) probe connected via AgX 9.5 mm x 2 m Fiber (Silver Halide). The iC IR™ software was used for data acquisition and mathematical treatment of the spectral data.

Nuclear Magnetic Resonance (NMR)

The spectroscopic technique of hydrogen and carbon nuclear magnetic resonance (¹H and ¹³C NMR) was applied in liquid samples (LD and monomer samples), using a Bruker spectrometer model Magneto Ascend 600, operating in a magnetic field of 14T at 600 MHz with a 5 mm probe. The analysis took place with 20 mg of sample dissolved in 0.5 mL of deuterated chloroform (CDCl₃), resulting in spectrograms expressed in parts per million (ppm). For ¹H NMR spectra, TMS tetramethylsilane (0.00 ppm) was used as an internal reference, and for ¹³C NMR spectra, deuterated chloroform (77.0 ppm). The spectra were processed using the MestreNova 6.0 program.

Mass Spectrometry Technique (ESI-MS)

Was carried out on liquid samples (LD and monomer samples) diluted with methanol and was operated by a two-unit device formed by the Q-TOF mass spectrometer equipment from AB Sciex and model TripleTOF 5600 + coupled to an ultra-high performance liquid chromatograph (UHPLC) from the Eksigent Ekspert brand and model 100-XL. This device was used in the positive [M + H]⁺ mode.

Gel Content

The test Was applied in polymeric samples. These were immersed in THF for 24 h at room temperature (≅ 25 °C) and then passed through a PTFE syringe filter (0.45 μm). The sample weight before immersion, initial weight (W_i), and the sample weight retained on the filter after drying, final weight (W_f), were the necessary parameters to calculate the gel content (GC %) by Eq. (1) below.

$$\:GC\left(\text{\%}\right)=100\left[1-\frac{\left({W}_{i}-{W}_{f}\right)}{{W}_{i}}\right]$$

(1)

Swelling Test

Were applied to evaluate the swelling ratio and the resistance of the materials in the presence of water and THF. Macerated PM3LD samples were weighed (Wi) and submerged in THF or water for 48 h at room temperature (≅ 25 °C). Then, the swollen samples were weighed (Wf), after removing excess solvent, and the swelling ratio (S) was calculated according to Eq. (2) below.

$$\:S\left(\text{\%}\right)=100\left[\frac{\left({W}_{f}-{W}_{i}\right)}{{W}_{i}}\right]$$

(2)

Scanning Electron Microscopy (SEM)

The morphology of the polymeric particles and the biofilms formation was evaluated using a Jeol scanning electron microscope, model JSM-7001 F. The samples were previously metallized through the sputtering process, covering them with a thin layer of gold, and the equipment operated with a 15 keV electron beam on them.

Results and Discussion

Limonene Dioxide and its Derived Monomers

Considerations about the monomer’s synthesis. The syntheses of the monomer resulted in yellowish liquid samples for the monomeric products M1LD and M2LD, and a light brown liquid sample for the M3LD product. All monomeric samples showed high viscosity, with the M1LD samples showing the highest apparent viscosity. The products resulting from acrylation and methacrylation are lipophilic, while the acrylamidation products showed an amphiphilic behaviour (product from organic phase of the extraction), and a hydrophilic behaviour (product from aqueous phase of the extraction). For a better understanding throughout the manuscript, the product of the organic phase of the extraction is referred to as M3LD, as it is the main product. The product of the aqueous phase is referred to as M3LD-AP and is only seen in the mass spectral characterization. It was found that M3LD solubilizes in the presence of methanol, ethanol, acetone, chloroform, water, and does not solubilize in diethyl ether and hexadecane.

The criterion used to evaluate the performance of monomeric products’ syntheses described in Table 1 was the conversion of LD into (meth)acrylated or acrylamidated monomers. The extension of the conversions was assessed by monitoring the suppression of ¹³C NMR peaks relating to carbons in the epoxy group. A more detailed discussion about the calculation of the conversion is reported in the following topics (NMR Analysis of Monomers).

With the conversion criterion established, it was observed that syntheses made with greater excesses of AA, MA and HEAA in relation to LD achieved greater conversions. Acrylation was the first type of reaction performed. It was seen that, reaction carried out at temperature of approximately 100 °C (synthesis no. 1) resulted in undesirable polymerization of the reagents. From this data, and based on the experimental procedure of the study prepared by Medeiros et al. [47], with focus on the acrylation of epoxidized fatty acid methyl ester (EFAME), the synthesis carried out at a temperature of 75° C for 3 h. However, the reaction with more than 3 h at 75 °C (synthesis no. 3) is not recommended in order to avoid the formation of extremely viscous fraction that are indicative of polymerization. Therefore, the acrylation synthesis of no. 6 led to obtaining the M1LD sample with a higher conversion of epoxide groups into acrylate groups. To afford the methacrylation, it was necessary to apply higher temperatures and reaction times in order to obtain satisfactory results, with synthesis n^o. 6 providing the M2LD sample with a higher conversion of epoxide groups into methacrylate groups. Finally, acrylamidation syntheses also required higher temperature to deliver better yields. The best conversion of epoxide groups into acrylamidated groups for the M3LD sample was obtained as far as it could be tested by synthesis n^o. 3, with visual polymerization being observed with the reaction temperature above 120 °C (synthesis n^o. 4).

(FTIR) Analysis.Figure S1 (Supporting Information), contains the infrared spectrum obtained for the starting terpenic molecule (limonene), limonene dioxide or diepoxide (LD) and, monomers (M1LD, M2LD, M3LD). Bands referring to the limonene unsaturations (see Supporting Information, Table S2) are not observed for the LD spectrum, indicating the transformation of the starting molecule. Additionally, the bands in 1099 cm^-1 and 1075 cm^-1 indicate the presence of oxirane rings, with the C-O-C ether stretching characteristic of this spectral region [48,49,50]. It is not possible to observe the presence of bands related to the C-O-C stretching in spectrum related to monomers, because of the overlapping with other more intense bands. Although, there are bands that indicate the coupling with nucleophiles (AA, MA, HEAA) in the terpene structure. Broad bands at 3463 cm^-1 for M1LD, 3441 cm^-1 for M2LD and, 3413 cm^-1 for M3LD are typical of OH stretching, indicating the presence of hydroxyls in the monomers. The OH group attached to the molecule is evidenced by the bands observed at 1292 cm^-1, 1298 cm^-1 and 1250 cm^-1 referring to the C-OH stretching of alcohol. These bands are indicative that the opening of the oxirane ring with AA, AM and HEAA may have occurred, originating β-hydroxylated products according to the S_N2 mechanism [51] Furthermore, it is also possible to see the appearance of the bands at 1052 cm^-1, 1039 cm^-1 and 1084 cm^-1, possibly referring to the O-CH₂ and O-CH stretching, which indicate the addition of nucleophiles to LD [34, 50,51,52,53,54,55]. These bands are more evident comparing the infrared spectra of the monomers with their nucleophilic precursors (Figure S2 and Table S3, Supporting Information). FTIR Vibrational bands related to the structural parts coming from the nucleophiles, that differentiate one monomer from the other, are detailed in the FTIR characterization of monomers (see Table S1, Supporting Information).

FTIR real-time monitoring. The LD synthesis was performed with real-time monitoring to better understand the behaviour of the epoxidation reaction. Figure 2a contains a set of mid-infrared spectra, collected every 60 s, and arranged in time, generating a surface that allows one to evaluate the epoxidation throughout the reaction. It can be noted, along the time axis, that the intense band between 1000 cm^-1 and 1200 cm^-1 grows as the reaction progresses. This region was related to DMDO dioxirane and LD epoxy groups, since they appear as the Oxone^® solution is dosed to the medium. At the same time, the bands peaking at 1220 cm^-1, 1356 cm^-1, and 1424 cm^-1, related to the methyl of acetone and dimethyldioxirane, decrease during the course of the reaction [56,57,58]. This reaction profile can also be accompanied by the initial and final 2D spectra shown in Figure S3a.

The profile of the epoxidation reaction through trend lines constructed from the spectral surface was acquired using IC IR^® software (Fig. 2b). When selecting a peak of the spectrum that changes its intensity as a function of time, the software displays a graph with the variation in the height of the selected peak over the course of the reaction. In this way, it is possible to evaluate initial components being suppressed and products being formed, as well as intermediates of the reaction, whose signal has its intensity increasing and decreasing throughout the reaction. Based on the reaction mechanism, the first step is the formation of dimethyldioxirane (DMDO), a spontaneous and instantaneous transformation. As soon as DMDO is formed, it readily reacts with limonene, which leads to a decrease in its concentration in the reaction medium, a fact that was attributed to the trend of the 1360 cm^-1 peak. At the same time, limonene monoepoxide (LM) is formed, followed by limonene diepoxide (LD). The formation of the LM as an intermediate could be followed by the trend line of the 1712 cm^-1 peak, which increases until around 0.25 h and decreases from this moment on. As the reaction progresses, the LD product is formed, a fact that was attributed to the peak trend line at 1062 cm^-1, which continues to increase its height up to 0.90 h, characterizing the end of the reaction. The dashed line in Fig. 2b is a trend based on the deconvolution of the spectral data over time made by the ConcIRT^® tool, which pointed to the formation of a product that is a combination of the DMDO production reaction with the formation of the LM intermediate and coincides with the real trend line of the monoepoxide formation/consumption.

Fig. 2

In situ FTIR analysis of the LD and monomer reactions. a, c, e, and g: 3D FTIR spectrum and b, d, and f: peak height profile

In line with the literature, the formation of a reaction intermediate is expected, given that the endocyclic unsaturation is more reactive than the exocyclic unsaturation of limonene [59]. This is due to two conditions that add up. First, the degree of substitution of the cycloalkane double bond increases the reactivity at the site. This unsaturation, because it is a more substituted structural pattern, helps to stabilize the partial positive charge generated during the chemical modification and, consequently, the reaction is faster there and becomes favourable. In addition, the ring tension caused by the 120° internal angle between the cycloalkane carbons also contributes to the reaction in endocyclic unsaturation. It is known that the ideal bond angle at sp³ carbons is 109.5°, due to their tetrahedral geometry, which makes the stress relief entropic, and the site reactivity tends to increase [21, 60, 61].

The syntheses no. 6 for M1LD (Fig. 2c-d), no. 6 for M2LD (Fig. 2e-f), and no. 3 for M3LD (Fig. 2g-h) were also monitored in situ and in real time. It is important to emphasize that real-time monitoring of the epoxide opening reaction step is a complementary study and was carried out separately. For the three reactions, attention was focused on the reaction behaviour of the LD in the presence of nucleophiles. The three additions of nucleophiles are indicated in the Fig. 2d, f,g by numbers 1, 2 and 3. It was possible to visualize on the 2D FTIR graph (Figure S3b-d) bands in the range of 850–750 cm^-1 belonging to the LD that decreased in the 3D spectrum as the reaction times elapsed. Similarly, the same criterion was used to observe peaks that were not present at the beginning of the reaction and that appear in the final spectra. These peaks, in the range of 1200–1000 cm^-1, can be seen in the 3D spectra, showing an intense growth profile since the first addition of nucleophile and are compatible with the final monomeric product.

From another perspective, these reported behaviours were displayed in the form of trend graphs. For reaction with AA (Fig. 2d), it was possible to observe three different trend lines, one for the consumption of LD, another for AA and one for the formation of a product associated with M1LD. For the AM addition reaction (Fig. 2f), the spectral regions led to two trend lines. One represents the consumption of the initial reagents, being a combination between them (LD + AM) and the other is consistent with the formation of a product, M2LD. In the case of the reaction with HEAA (Fig. 2h), the product formation trend line also shows the consumption of the nucleophile and, given that the spectral regions between them are very similar, the combination of these two events was noted in the same trend. Then, the growth followed by the decay of this trend line was observed at the moments when the additions of HEAA were made, indicating the entry and consumption of this reagent. After the three additions, this trend line goes back up, signaling that the formation of a product continues to progress over time. This interpretation is also observed by the trend line generated by the ConcIRT^® tool. The profile found can be associated with the consumption of HEAA, since it consistently increases in the first addition and the decays follow the same pattern of the trend line related to HEAA + M3LD as the reaction occurs. Therefore, there are three trend lines and one of them, as well as for the other reactions, is the consumption of LD.

Thus, reactions start as soon as the nucleophiles encounter the LD. It can be seen from the trends found based on the selected peaks. The beginning of the formation of a product for each type of reaction is observed when the first amount of the nucleophile is added. This behaviour is also verified with the simultaneous decay of the trend line that represents the initial reactants. Such events occur even without a high temperature increment, since all the first dosages of added nucleophiles were made around 40 °C. The temperature was gradually increased after each subsequent addition of nucleophile (AA, AM or HEAA). This procedure was established to avoid premature polymerization due to the high exothermicity of the three reactions. Thus, the opportunity was taken to monitor the temperature, recording the reaction temperature at different times, until reaching the previously established upper temperature limit (according to Table 1). So, the total period to reach reaction temperatures was approximately 1 h 50 min for acrylation, 2 h 30 min for methacrylation and 3 h 20 min for acrylamidation.

In addition, it is possible to infer that increments in the trend line of product formation over time may be related to the formation of oligomers. Thus, for the acrylation reaction, a product was formed and remained constant until 1 h 15 min (indicated in Fig. 2d) under a temperature around 67 °C and after this period, there is a subtle growth in the trend line of product formation when the temperature reaches around 70 °C, which is possibly the point of oligomers formation. In the same way, it takes around 2 h 15 min for methacrylation (indicated in Fig. 2f) with a temperature of 100 °C. After this time, the increase in the trend line is clear when the temperature is already close to 105 °C. Finally, this same reaction behaviour is observed for acrylamidation. The trend line remained constant until the 3 h reaction time (indicated in Fig. 2h), when the temperature reached 110 °C and, after this period, the trend line started to rise again with the temperature already higher. Therefore, in this first study, it was possible to verify that the production of these monomers, in the used molar proportions of reagents, can be achieved with shorter reaction times and lower reaction temperatures in order to avoid the formation of oligomers as secondary products in subsequent studies. It was also noted that the greater the molar weight of the nucleophiles, the greater the temperature and reaction time to reach the products of interest.

NMR and ESI-MS Analyses of LD.Figure S4 allows the comparison of the ¹H NMR spectrum obtained for limonene with the spectrum obtained for LD. By confirming the structures of these molecules, the highlighted regions indicate the suppression of the peaks related to the hydrogens of endocyclic and exocyclic unsaturations of limonene (5.40 ppm and 4.70 ppm, respectively) and the appearance of peaks related to hydrogens of endocyclic and exocyclic epoxides (3.07–2.96 ppm and 2.63–2.49 ppm, respectively), as well as point out the stereochemical behaviour of the LD molecule. It is found in the literature that the epoxidation of any enantiomer of limonene will result in a cis/trans mixture containing four diepoxidized products (Scheme S2), indicated by the presence of four singlets in the “F” region in Figure S4 [1, 5, 12, 15, 62]. The ¹³C NMR spectrum of LD (Figure S5) shows the sets of peaks 58.6–59.0 ppm and 53.1–52.6 ppm related to de endocyclic and exocyclic carbons of the epoxide groups, respectively. The NMR characterizations made for LD were confirmed by its 2D spectra (HSQC and HMBC), attached in Figure S6. From the ¹H NMR spectra for limonene and LD, a relative area integration method was employed to measure the conversion of limonene unsaturations to epoxides [47]. However, the integrals in the region of vinylic hydrogens in the LD spectrum resulted in values equal to zero, as can be seen in Figure S4, which indicates that the conversions of endocyclic and exocyclic unsaturation into epoxides were close to 100%. Allied with this, the high-resolution mass spectrometry analysis reinforced the transformation of limonene into LD. The signal observed with greater intensity in the spectrum of Figure S7 was the one with m/z 191.1039, that corresponds to the adduct [M + Na]⁺ with molecular formula C₁₀H₁₆NaO₂⁺ and exact mass 191.1043 g.mol^-1.

NMR Analysis of Monomers. The characterizations of the ¹H and ¹³C NMR spectra for the M1LD, M2LD and M3LD monomers (see Figures S8a-b, S9a-b and S10a-b) show that there was transformation of the LD molecule. The ranges of chemical shift regions between 6.5 and 3.25 ppm in the ¹H NMR graph and 60–170 ppm in the ¹³C NMR graph comprises the principal peaks that indicate the addition of AA, AM and HEAA nucleophiles into the LD molecule. Especially the peaks indicated by the letters L, M, Q and R evidence the opening of the oxirane rings of the LD, inserting the (meth)acrylates and acrylamide groups at one end and forming hydroxyl groups at the other end. By convention, the characterization of the ¹H and ¹³C NMR spectra of the precursor nucleophiles are contained in Figures S11 (AA), S12 (MA) and S13 (HEAA) to better visualize their respective peaks in the monomer spectra.

Fig. 3

Comparison between ¹³C NMR of (a) LD, (b) M1LD, (c) M2LD and (d) M3LD

Figure 3 highlights the characteristic region of endocyclic and exocyclic epoxides for the LD molecule and the characteristic region of unsaturations from (meth)acrylates and acrylamide groups. It can be noted for the (meth)acrylated monomers (Fig. 3b-c) there are no carbon signs of epoxides, indicating the formation of disubstituted products, from the total consumption of the epoxy groups of the starting LD molecule. This becomes even clearer by analyzing the HSQC and HMBC 2D NMR spectra for M1LD and M2LD (Figure S14 e S15). While for M3LD (Fig. 3d), there are still signs referring to epoxides, indicating the presence of a monosubstituted product and remaining LD molecule and this could also be seen by the 2D NMR spectra (Figure S16).

Through the NMR analysis, it was also possible to identify the presence of regioisomers for the three monomers. The mixture of diastereoisomers from LD, which reacts with nucleophiles, also leads to diastereoisomeric products. Associated with this, there is the formation of regioisomers with different substitution patterns from the LD molecule. Especially in reactions of opening the oxiranic ring with reaction media constituted by an excess of Brønted-Lowry acid (BL), an S_N2 attack with inverted stereochemical configuration tends to occur on the more substituted carbon instead of on the less substituted carbon. In this case, acid catalysis through protonation of epoxide oxygen favors this type of attack. A partial positive charge is generated on the most substituted carbon of the epoxy group, leaving the C-O bond in this place longer and, consequently, weaker, making oxygen a good leaving group and directing the nucleophilic attack of the carboxylate (from AA and MA) on that more partially positive carbon (see Scheme S4) [21, 60, 63,64,65,66].

The opening of the oxirane ring tends to occur first in the endocyclic group, as this is more reactive than the exocyclic group. The epoxide attached to the ring of the terpene structure forms a bicyclic system, in addition to its natural ring tension due to the C-O bond with forced internal angles of 60° between the tetrahedral carbons and angular oxygen, there is also the cycloalkane of the terpene, which also has ring strain and tends to further increase the total tension of the region where the epoxide is located. Thus, the opening of the endocyclic oxirane ring provides considerable stress relief and entropy increase. This sum of tensional factors makes the molecule more reactive at this location, causing some chemical modification of ring opening to start selectively at the endocyclic epoxide, that is, the reaction is faster at this location. Therefore, this factor contributes to favour the formation of a major monosubstituted product. However, the excess of acid in the medium, associated with higher reaction temperatures helps to induce an increase in exocyclic epoxide reactivity so the reaction proceeds rapidly at this site as well [21, 60, 61, 63, 66, 67].

These reported behaviours happen efficiently in the acrylation and methacrylation reactions. However, MA, is a slightly weaker and bulkier acid when compared to AA. According to the literature, while the pKa for AA is around 4.06, the pKa for MA is around 4.46. Associated with it, the S_N2 reaction is dependent on the volume of nucleophiles, and reactions performed with larger nucleophiles are slower. These factors can be associated to explain why a greater amount of AM in the methacrylation compared to LD, as well as a greater temperature increment were necessary to affords yields similar to that for the acrylation reaction. Optimization of these reaction conditions helped to direct the reaction towards a higher frequency of acid catalysis, providing a greater number of protons from the AM and carboxylate nucleophile for opening both epoxides of the LD [21, 61, 68, 69].

On the other hand, a divergent behaviour could be observed for acrylamidation. One of the reasons for that relies on lower acidity of the O-H of the hydroxyl group (pKa 14–16) when compared to the carboxylic acids, making the alcohol of the HEAA a weaker acid catalyst for the S_N2 reaction on the epoxide. In this scenario, without assiduous acid catalysis, the opening of the exocyclic epoxide becomes less frequent considering its lower reactivity, favoring the production of a monomeric monosubstituted product in relation to a disubstituted product [21, 60, 63, 66, 67, 70].

When considering these factors, higher temperatures were used to improve the addition on both epoxides. However, this approach was not enough to accomplish it, as shown by the calculated endocyclic and exocyclic epoxide conversions for acrylamidation reaction. The opening of the endocylic epoxide showed the highest percentage of conversion, around 77%, while the conversion of the exocyclic one was around 51%. The conversion was calculated from the clearer ¹³C NMR instead of ¹H NMR (see Figure S18 and Equations S1-S2). Considering the conversion value obtained for the exocyclic epoxide, the opening of this oxirane ring is possibly associated with the minor formation of the diacrylamidated product and with the formation of an oligomer (Figure S10).

Figure S18 shows a comparison between the ¹H and ¹³C NMR spectra of the LD with those of the M3LD. When observing the comparisons, it is verified that the ¹H spectrum of the monomer (Figure S18a) does not offer the necessary clarity to determine if the peaks in the region of 2.63–2.50 ppm are related to the exocyclic epoxide hydrogens or if they are related to other hydrogens of the molecule, indicating that the epoxide has been deleted. This doubt makes the application of an integration method unreliable to calculate the conversion of epoxides into acrylate groups using ¹H NMR graphs. The same can be seen for the comparison between LD/M1LD and LD/M2LD (Figure S17). In contrast, the epoxide region in ¹³C NMR is cleaner (Figure S18b). Due to the larger spectral window for ¹³C resonance and the fact that the ¹³C NMR experiment, when acquired decoupled from ¹H, presents resonance signals with simple multiplicity for ¹³C nuclides linked to ¹H, facilitate the visualization of epoxide signals in the spectrum.

Once signs of residual epoxide are detected in a monomer sample, as it happened for the M3LD, it would be possible to quantify it through the relative integration of ¹³C resonance signals. Such an analysis needs to be done with care, as ¹³C nuclides may present differences in their relaxation times that would lead to significant differences between the relative area measured for one signal with respect to another, a basic argument for quantification of the conversion by NMR. When comparing the areas of the signals attributed to the carbons of the oxirane rings of the isomeric mixture of LD with the areas of the methyl F and G of this same sample (see Figure S5), it is verified that the relative integration is close to that expected, 4 carbons of the rings oxiranes to two methyl carbons. In this case, it is plausible to propose the use of the relative integration of these resonance signals from the ¹³C NMR spectra to calculate the conversions of endocyclic and exocyclic epoxides into acrylamide groups, similarly to what is done using the hydrogen spectra. For the M1LD and M2LD, as no peak is observed in the relaxation region of the epoxide carbons (52.63–60.43 ppm) in the ¹³C NMR of these monomers, it can be inferred that the conversions of the oxirane rings of the LD into (meth)acrylate groups were close to 100%.

ESI-MS Analysis of Monomers. For the (meth)acrylated monomers, an ion with m/z 335.1471 was detected and assigned with the sodium adduct of the M1LD molecular ion [M + Na]⁺ (Fig. 4a), and an ion with m/z 363.1785 was assigned to the sodium adduct with the M2LD molecular ion [M + Na]⁺ (Fig. 4b), evidencing the formation of di(meth)acrylated monomeric products, named as 2,8-dihydroxy-1,9-limonene diacrylate (unprecedented in the literature) and 2,8-dihydroxy-1,9-limonene dimethacrylate (already reported by Schimpf et al. [19]).

For the reaction to prepare the acrylamidated monomer, the analysis of the crude showed the mono and the disubstituted products (Figure S19). Two materials were obtained after the liquid-liquid extraction with water and chloroform: one from the organic phase (M3LD) and another from the aqueous phase (M3LD-AP). The ESI-MS analysis of these products showed, Fig. 4c and d, respectively, the diacrylamidated monomer in the aqueous phase together with free HEAA, and its absence in the organic phase. By looking at the spectrum of Fig. 4c, one can see the peak for the monosubstituted monomeric product (ion with m/z 306.1685 assigned to the sodium adduct with the M3LD molecular ion [M + Na]⁺) next to the remaining LD (m/z 191.1047), while the disubstituted (or diacrylamidated) product is found in the spectrum of Fig. 4d (ion with m/z 421.2312 assigned to the sodium adduct with the M3LD-AP molecular ion [M + Na]⁺) next to the free HEAA nucleophile (m/z 138.052).

Fig. 4

Mass spectrum of (a) M1LD, (b) M2LD, (c) M3LD and (d) M3LD-AP

Based on the epoxide conversion values obtained from the ¹³C NMR in association with the result of the mass spectrometry analysis for this acrylamide monomer, it can be inferred that the monosubstituted product is mostly present. However, the acrylamidation product, before being washed, is composed of monosubstituted and disubstituted (or diacrylamidated) monomeric products that can be polymerized, named as 2-hydroxy-1-acrylamide of limonene and 2,8-dihydroxy-1,9-diacrylamide of limonene, respectively, and are unprecedented in the literature. The minor diacrylamidated product remained in a sample composed predominantly of free HEAA and was not polymerized.

The presence of ionization artefacts was considered when analyzing the peak at m/z 263.1261 (Fig. 4a), a value that corresponds to the same m/z of the intermediate formed by the monoaddition of AA in the LD, still with remaining epoxide. However, it was considered that this product is not present in the monomer sample, since the ¹³C NMR spectrum reveals the suppression of all peaks referring to the carbons of the epoxy groups. In view of this, it is possible, then, that a fragmentation by cleavage of [M + Na]⁺ of M1LD has occurred (Scheme S3), forming an ionization artefact [71, 72]. The same behaviour was associated with the peak at m/z 431.24 in Fig. 4a, being an ionization artefact of oligomers 2 and 3. Ionization artefacts were also seen for methacrylated and acrylamidated products, following the same cleavage mechanism shown in Scheme S3.

The presence of oligomers in monomers reactions can be associated with the result of two secondary reaction mechanisms favored by the increase in temperature. One of them is the nucleophilic conjugate addition (1,4 addition), also known as Michael addition reaction, which possibly occurred between AA or MA and the unsaturation of the (meth)acrylate group attached to the aliphatic part of the M1LD/M2LD molecules (see Scheme S4), forming the oligomer 1 in M1LD/M2LD samples. The other, probably, is the opening mechanism of the exocyclic epoxide through the SN2 attack of the molecule (A) (Scheme S4), which is a bridge for the formation of oligomers 2 and 3 in M1LD/M2LD samples and oligomer 1 in M3LD samples.

To summarize, Charts S1, S2 and S3a-b shows the possible chemical structures formed, of the monomer, oligomers and, the possible ionization artefacts found for the (meth)acrylation and acrylamidation products.

Furthermore, it was possible to verify the high purity of the monomeric products through ESI-MS analysis. Peaks that would be related to the presence of the remaining nucleophile (ions with m/z = 95.0109, m/z = 109.0265 and m/z = 138.0520 attributed to the sodium adduct with molecular ion M1LD [M + Na]⁺, M2LD [M + Na]⁺ and M3LD [M + Na]⁺, respectively) in the M1LD, M2LD and M3LD samples were not observed in the graphs of Figure S20a, S20b and S20c (expanded regions of Fig. 4a, b and c where the nucleophile peaks should be detected). These observations confirm the efficiency of the proposed washing step to remove the unreacted nucleophile after the end of each monomeric reaction.

Limone Di(meth)Acrylated and Acrylamidated Polymers

Considerations About the Polymer´s Synthesis

After the end of the miniemulsion polymerization reactions, in all the reaction conditions to synthesize M1LD and M2LD, formation of creamy and solid phases was observed, in addition to the liquid phase intrinsic to miniemulsion polymerization. Furthermore, the creamy part sedimented quickly. Thus, the miniemulsion turned out to function as a macroemulsion [32]. For the reactions carried out with M3LD, which were only carried out with the co-stabilizer Crodamol GTCC due to the insolubility of the acrylamide monomer in hexadecane, it was observed that after the end of the reactions there was only a gelatinous polymeric mass in the middle of the reaction water, indicative of polymer gel, and there was no formation of liquid/creamy polymeric parts. The formation of a hydrogel polymer from M3LD is consistent with what is found in the literature for polymers synthesized with HEAA acrylamide monomer and with other acrylamide species in general [73,74,75,76,77].

The non-stability of the reactions was already verified in the preparation stage of the polymeric synthesis in the tests carried out with 4% of hexadecane co-stabilizer. To increase the mixture stability, delaying the diffusional degradation of the monomer droplets (or Ostwald ripening phenomenon), a higher percentage of hexadecane (8%) was added keeping the pre-reaction mixture visually stable, that is, homogeneous. This being done for polymers with different percentages of organic phase and different initiator species, as shown by the polymeric tests described in Table 2. However, the results were slightly different from reactions carried out with 4% hexadecane. Polymers synthesized with higher percentages of co-stabilizer obtained smaller solid parts during the reaction but did not entirely prevent formation of creamy and bulk phases. Therefore, due to an apparent greater stability in the preparation of miniemulsion with M1LD using 8% co-stabilizer, all the following polymeric reactions, described in Table 2, were carried out with 8% co-stabilizer hexadecane or Crodamol GTCC. The latter promoted a slight improvement in the homogeneity of the miniemulsions, more notably with the most viscous monomer, M1LD. This is because Crodamol GTCC has excellent emollient and spreadability properties that help promote dilution of highly lipophilic compounds [37,38,39,40].

The change in physical aspects between the monomers and their polymers suggests that polymerization has occurred. After the drying step of the P(M1LD) and P(M2LD) to remove water from the miniemulsion, whitish and yellowish solids were obtained with easy and moderate maceration. The maceration of these polymers resulted in the mixture of fine powders (such as talc powder) and moderately rigid powders, as creamy and bulky parts were formed in the samples of these (meth)acrylated polymers, shown in Figure S21a-b of the SI. The maceration of the polymer made with the acrylamidated monomer P(M3LD) proceeded from the dry gelatinous dark yellow mass, which hardened considerably in the drying stage, becoming a more rigid solid, shown in Figure S21c of the SI. By macerating this material, it was possible to obtain smaller particles that aggregate quickly together, indicating the easy absorption of moisture during maceration, which is justifiable due to its hydrogel nature [78, 79].

(FTIR) Analysis

Figure S22a-d shows the FTIR spectrum obtained for all polymeric tests described in Table 2 compared to its respective monomeric precursors. The main regions of change in intensity were highlighted. The increase in intensity of the polymer bands compared to the monomer bands indicates that polymerization has taken place [50, 51, 80]. It can be noted that in the polymer spectrum, the vibration bands out of the plane for unsaturations appear to be softer and the vibration band in the plane (around 1640 cm^-1) gets wider but is still visible. Probably, because the polymerization termination step is terminating the polymeric chains by disproportionation, forming an unsaturation. Added to these factors, the bands referring to -CH₂ and -CH₃ groups appear much more intensely, indicating that there was a growth of the polymeric chain. There is also a C-O stretching band that grew in the FTIR spectra of the polymers, which may be related to aliphatic chain ether groups produced through cross-linking mechanisms [21, 50, 51, 80,81,82]. Knowing that there is LD remaining in the M3LD monomer, as well as the monosubstituted monomeric product, the relative intensity of the OH band increased in relation to the NH band and can be associated with the opening of the remaining epoxy rings, forming hydroxyl groups. In addition, the band referring to carbonyl stretching seems to be shifted to longer wavenumbers in polymers from M3LD, which increases the evidence that there was polymerization of the carbonyl-conjugated double bond in monomeric molecules, impacting the wavenumber at which the C = O stretch is found.

Gel Content and Swelling test

Analyzing Analyzing the data obtained for all P(M1LD) polymeric tests (Table 3), gel content resulted in percentages greater than or close to 90%, which indicates a high degree of cross-linking network, and it justifies the stability of these acrylate polymers in THF, since this solvent is one of the most efficient to solubilize polymers. In addition, this feature indicates that there was a high conversion of monomer into polymer [15] The GC values decreases for the other polymers, being the polymeric tests for P(M3LD) with the lowest gel content. The acquired degree of cross-linked can be explained by the frequency of enolate radical formed to each polymer and by the presence of β-hydroxyl groups in the starting monomer, which induce cross-linking through chain transfer. Furthermore, cross-linking can also occur due to the combination of polymer chains during the polymerization termination step [20, 62, 82].

Table 3 Gel content and swelling for PM1LD, PM2LD and PM3LD polymeric tests

The (meth)acrylated and acrylamidated monomers are tetrareactive and because of it has an enoate group, which can react through the 1,4-addition mechanism [83,84,85]. Taking the acrylated monomer as an example, in its initiation step, unsaturation reacts as an acceptor on the β-carbon and donor on the α-carbon (Scheme S5a). As a result, an enolate radical is generated on the α carbon which, in the chain propagation step, reacts with the acrylate group of another monomeric molecule and so on, building the polymer through the Michael addition mechanism (Scheme S5b). Probably, the chain propagation rate in this step is high, fostering the bulk phase formation that were visualized [86].

The enolate radical of the propagation product, already in the chain transfer step, can react with any other electrophilic group in its proximity to abstract hydrogen, either from a monomeric molecule, for example, and/or abstracting hydrogen from the hydroxyls of polymeric chains (Scheme 1). Consequently, oxygen becomes an alkoxy radical capable of attacking the β-carbon of the acrylate group of the monomer (Oxo-Micheal addition), as well as reacting with another enolate radical, linking the chains by termination, and forming ether groups (Scheme 1). Probably, this is the main source of cross-linking, and the result is an immense network composed of several cross-reactions. Backing in the initiation step, it is important to mention that the AIBN initiator could also react with the hydroxyl oxygen, removing its hydrogen and forming alkoxy radical, which will follow the same reactional tendency just discussed (Scheme S5a). Furthermore, the termination step can combine two live polymeric chain (Scheme S5c) and in the disproportionation mechanism, the unsaturation formed can be an electrophilic group for 1,4 additions by the alkoxy radical (Scheme S5d).

Scheme 1

Chain transfer mechanism with enolate radical and termination mechanism between enolate/alkoxy radical (main source of reticulation)

Cross-linking for P(M2LD) did not occur at the same frequency reported for M1LD, leading to a lower SR value (see Table 3). This is because the methyl group attached to the α-carbon causes steric hindrance in the region, decreasing the reactivity of enolate radical. In view of this, there is a lower ability of this anion to react with electrophilic groups and the cross-linking pathways end up occurring with less constancy.

By looking deeper into the differences between the calculated GCs for the (meth)acrylated polymers, the use of Crodamol GTCC slightly decreases the cross-linking degree of these polymers. This may be associated with a possible more pronounced plasticizing effect of this compound, which may decrease intermolecular interactions, reducing the cross-linking density [87]. Furthermore, contrary to what was being observed, the GC of P(M2LD)-T4 (made with KPS and hexadecane) showed a higher cross-linking value compared to P(M2LD)-T2 (made with AIBN and Hexadecane). This difference was probably due to a higher percentage of KPS initiator added accidentally when preparing the miniemulsion, which may have increased the incidence of cross-linking mechanisms from KPS.

To understand the lowest GC values obtained for P(M3LD) tests, it is necessary to point out the lower acid strength of the acrylamide functional group, which possibly led to a lower degree of crosslinking compared to the GCs calculated for P(M1LD) and P(M2LD) tests. In the P(M3LD) chain propagation mechanism, the enolate radical formed on the α-carbon, adjacent to the carbonyl, is less stable. This occurs because the amide carbonyl has a lower ability to stabilize the enolate radical due to the competition between the delocalization of the carbonyl oxygen unpaired electron and the unpaired electron pair of the -NHR group, which is a pi donor. This also occurs with the ester group, nevertheless, is more pronounced for the amide because nitrogen is less electronegative than oxygen [21, 88, 89]. Besides this circumstance, the M3LD monomer is a monoacrylamide, having one less site to promote Michael addition. Therefore, it is possible that a lower frequency of chain transfer induced by the enolate radical occurs for P(M3LD). In this way, the possibilities of reticulation mechanisms are reduced, and the number of crossed networks becomes smaller.

Digging deeper now into the differences between P(M3LD) tests, a higher degree of cross-linking can be seen for the polymer made with KPS initiator, P(M3LD)-T2. The M3LD monomer is amphiphilic, and the direct miniemulsion technique was employed. Therefore, a part of the monomer with less hydrophilicity tended to form monomeric droplets in the organic phase and another part, with less organophilicity, tended to accommodate better in the aqueous phase. The KPS initiator, as it is water-soluble, releases the initiator radicals in the aqueous environment. These recently released radicals react with the monomeric particles found in the aqueous phase and, also, with the monomeric particles found inside the droplets, in the organic phase. With this, a greater regularity of cross-linking mechanisms may occur when using KPS. Differently from what happens with the AIBN initiator, which is essentially organosoluble and concentrates its release of radicals in the organic phase [32, 33, 82]. Therefore, in the case of this amphiphilic monomer, a lower repetition of cross-links can be associated with AIBN use.

In correlation to this, there is still the heterogeneity of products that the M3LD has. Considering that the remaining LD molecules reacted during polymerization, opening the remaining epoxides, chain propagation and crosslinking mechanisms from this molecule may have been concentrated in the organic phase of the miniemulsion, since LD has a hydrophobic character. When AIBN was used in the P(M3LD)-T1 test, a cross-link density from LD may have been greater than a cross-link density from acrylamide residues presents in monomeric molecules, which tend to be hydrophilic. Knowing that LD solubilizes in aprotic polar solvents, such as THF, and that polymers based on this molecule also tend to do so, it is associated that part of the polymer came to interact with this solvent due to a higher frequency of chains crosslinked by the LD. The collapse of this network may have contributed to a negative swelling rate (− 11%). On the other hand, P(M3LD)-T1 proved to be more resistant to water, supporting the above reasoning and the degree of swelling in water was higher (75%), due to the permanence of reticulated networks [15, 66, 90,91,92]. The polymer P(M3LD)-T2 synthesized with KPS, on the other hand, had a greater resistance in THF and in water, due to the higher frequency of cross-linking from both main products of the M3LD (monomeric and LD molecules) and, possibly, concentrating the cross-linking mechanisms from the acrylamide residues of the monomeric molecules. This can reduce the access of THF to the organophilic part of the polymer, assuming that the organophilic chains will be less exposed than the hydrophilic chains, which with KPS expanded favourably. Added to this, there is less interaction between THF with the amide groups, which also contributes to a greater swelling of these polymers in the presence of water due to the greater interaction between amides and water molecules [91, 92].

Fig. 5

Polymeric structures from M1LD, M2LD and M3LD and their possible sources of crosslinking

Based on the degrees of crosslinking, on the FTIR analysis and on the results obtained for M1LD, M2LD and M3LD it is suggested the formation of three new biobased polymeric products, which were named as poly(2,8-dihydroxy diacrylate of limonene), poly(limonene 2,8-dihydroxy dimethacrylate) and poly(2-hydroxy acrylamide of limonene), these last one being a hydrogel. The possible structures that constitute these polymers are represented in Fig. 5 and the probable sources of cross-linking are also indicated. It is important to mention that the epoxide present in the monosubstituted product of M3LD can react, as well as the LD molecules, opening the oxiranic rings. Thus, most likely, the homopolymerization formed polymeric chains from the epoxide and acrylamide portions of the monosubstituted molecule, as well as molecules from the remaining LD.

SEM Analysis of polymers

From the SEM images for P(M1LD), only made for polymeric tests with 20% OF, in Fig. 6a-b (macerated sample) and in Figures S23-28, the main characteristic observed was the presence of pores in the polymeric particles of all polymers. These pores do not have a homogeneous pattern of size and shape, exhibiting many irregular spherical cavities. Some pores were selected in the images to verify their sizes, relative to the scale of the microscopic image, and were measured by the Image J software. The selected sizes ranged from 60 nm to 21 μm, which characterizes these cavities as macropores for being larger than 50 nm [93, 94]. Porosity, when not acquired by means of porogenic agents, can be induced through a high cross-linking degree of the polymer, originating from multifunctional monomers, which directly affect the formation of cavities in the polymeric structure, which is the case of the PM1LD polymers synthesized in this work [94,95,96]. In addition to the porous characteristic, clusters of larger and smaller polymeric particles were noted. From the analysis of the surfaces of these particles, two types of morphology are suggested. One based on the visualization of a denser structure that appears to have been arranged in layers, which are shaded and indicated in the figures as LM (Layered Morphology). And the other is based on a surface with different reliefs and/or cavities, caused by the grouping of larger particles with smaller ones. This appearance was associated with the physiognomy of corals and because of that it was called CM (Coral-type morphology), which is highlighted and indicated in the SEM Figures.

Furthermore, it was observed that there is no homogeneous format between the particles, but the P(M1LD)-T1 (Fig. 6a-b and Figure S23) and P(M1LD)-T3 (Figure S25) showed some spherical particles grouped together, which were also shaded and indicated in the images as SP (Spherical Particle). However, this characteristic was not observed in the P(M1LD)-T2 (Figure S24) and P(M1LD)-T4 (Figure S28) and the only difference between these polymeric tests is the use of AIBN or KPS initiators in their syntheses. It is known that the control of the number and morphology of polymeric particles in this type of reaction is done through the surfactant and co-stabilizer, but since the same amounts of these components were applied in these tests, it is possible, then, to associate that this morphology may have been acquired due to a better efficiency of polymerization in miniemulsion using the organosoluble initiator AIBN (P(M1LD)-T1 and P(M1LD)-T3 tests), which favors a more homogeneous nucleation of the droplets and added to the effects of emulsifiers, it was possible to see the formation of spherical particles [32, 33].

Entering this perspective and correlating the components of the miniemulsion with the morphology found, the SEM images of the tests carried out with the co-stabilizer Crodamol GTCC are shown in Fig. 6c (film sample) and in Figures S27-28. These images clarify the difference in morphology that Crodamol GTCC, associated with AIBN or KPS, provided. It is clear in Fig. 6c and Figure S27 (P(M1LD)-T5 made with AIBN) the formation of well-defined spherical particles that vary in size, between the micrometric and nanometric scale (9 μm-164 nm). In the SEM images of Figure S28 (P(M1LD)-T6 made with KPS) the spherical particles were not so well-defined, as they are too much agglomerated, although, the sizes varied less (400 nm–275 nm).

This particle size variation may be related to the presence of oligomers in the M1LD samples that polymerized forming larger particles, and to the resistance that this very viscous monomer has in spreading when preparing the miniemulsion. The co-stabiliser Crodamol GTCC proved to be more efficient in acting as an emollient to the monomer, as its caprylic/capric triglycerides confer high lipophilicity and thus substantially influence the size and shape of the polymeric particles [40].

Fig. 6

SEM images for polymers: a-b. P(M1LD)-T1 c. P(M1LD)-T5, d-e. P(M2LD)-T1 and f-h. P(M3LD)-T2

Even with this optimization, the Oswald ripening effect continued to be present, however, in a lesser constancy compared to the hexadecane co-stabilizer. This led to a more uniform polymeric particle shape, but with different sizes. As a result, there are parts of the polymer with particles of similar sizes and with greater surface area, which appear to be arranged in layers (LM), providing a smoother surface. In other parts, consisting of agglomerates of particles with different sizes, a coral-type morphology (CM) was visualized, providing an irregular surface. This explanation aligns with the fact that M1DL contains different regioisomers and oligomeric byproducts, each contributing to various sources of crosslinking. Monomers of the same size tend to organize uniformly, forming lamellar structures, while the presence of oligomers can disrupt this organization, leading to a less uniform, coral-like morphology with bridge formations. In addition, there is not a large amount of visually well-defined pores, which may be linked to the smaller scale of formed polymeric particles. It is plausible, then, to associate that the pores in the structure of these polymers may be on a very small scale, which the equipment did not allow to visualize. However, cavities can be seen between the particles.

The SEM images obtained for P(M2LD) polymers shows in Fig. 6d-e and Figures S29-32 the formation of spherical particles with more regular sizes and with a larger surface area. Compared to PM1LD polymers, there was less size variation (3 μm-109 nm). The polymeric particles, either macerated (Fig. 6d) or in film (Fig. 6e), are well distributed and, due to their small and constant sizes, form a smooth surface and apparently arranged in layers (LM). There is also a coral-type morphology (MC), but not very evident. Well-defined pores were not visualized based on the scale-up used, only small cavities between particles. There were no significant differences in the morphology and size of the P(M2LD) polymers that could be linked to the use of different initiator species and co-stabilizers in the miniemulsion.

However, in the preparation stage of the miniemulsion with M2LD monomer, both with hexadecane and Crodamol GTCC, it was observed that this monomer was easier to dispersed as nanodroplets than the M1LD monomer. Possibly because this methacrylated monomer had a lower viscosity compared to the acrylated monomer. The greater spreadability may have contributed to provide better control of morphology and particle size, using both co-stabilizers associated with both types of initiators. Which leads to inducing that a lower degree of viscosity allowed a greater balance between the osmotic pressure inside the drop and the Laplace pressure around the drop, preventing the monomeric molecules from escaping out of it and thus, the morphology and size of particles tend to be more homogeneous and smaller [97].

Finally, the SEM images for P(M3LD) polymers shown in Fig. 6f-h and Figures S33-34 show macerated (Fig. 6f) or whole polymeric particles (Fig. 6g-h) without a well-defined morphology and format. A mass with a smooth surface is observed, in which the particles seem to be fused to each other. It is possible to see a merged surface (SM) (Fig. 6h), which can be related to the heterogeneity of the polymer composed of polymeric chains with more hydrophilic parts (monosubstituted monomer molecules) and other more organophilic parts (from LD molecules). Thus, the polymeric particles with greater hydrophilicity accommodate better in the water of the miniemulsion and when the polymer is left to dry, the particles with less hydrophilicity deform because they do not accommodate as well in the water of the miniemulsion, providing this apparent merged surface. Some well-defined pores were observed on the surface of the polymer carried out with the KPS initiator (Fig.6f). However, there is no uniformity between the sizes or a good distribution (6 μm -300 nm). The visualization of pores can be associated with the highest degree of cross-linking calculated for the P(M3LD)-T2. It can also be associated with the possibility of a reticulated network with greater spacing between one cross-link and another, which may have favored the better swelling ratio calculated for this polymer immersed in water. The average distance between cross-links directly affects swelling of hydrogels, with very small or very large spacing decreasing the swelling ratio [92].

Thermal analysis

DSC and TGA. For P(M1LD) polymeric tests made with hexadecane co-stabiliser and KPS initiator, the glass transition temperature (T_g) values, obtained from DSC curves (Figure S35), are lower than with the AIBN initiator (Table 4), and it decreases even more for polymers synthesized with 15% of organic phase, without sharp differences between the percentages of hexadecane used (4 or 8%) in the syntheses. This effect also occurred for DSC curves for P(M2LD) polymeric tests (Figure S36), although with a less pronounced difference between values (Table 4). These results may be related to the better efficiency of AIBN in the polymerization reactions of the organophilic (meth)acrylate monomer, favouring nanodroplets nucleation due to the greater amount of initiator radicals within the monomer droplet prompt to undergo reaction. Under this condition, the polymer chains may expand rapidly with smaller sizes and the cross-linking points end up having smaller spacings relative to each other (tight scaffolding - Fig. 7a). This effect shifts the T_g to higher temperatures, as it decreases the mobility between the chains, leaving the material stiffer. The polymers synthesized with KPS may have generated larger polymer chains with cross-linking points more spaced apart, due to the compartmentalized behaviour of miniemulsion polymerization in presence of the water-soluble initiator, affecting the nucleation of the organic monomeric droplets. Thus, the chains have a higher mobility and the T_g decreased (wide scaffolding - Fig. 7b) [32, 33, 82].

Table 4 Glass transition temperatures of polymer tests and thermal events from TG curves in figures S36-35 and figures S37-39, respectively

Fig. 7

Representation of P(M1LD) polymeric chains with (a) higher cross-linking density (b) lower cross-linking density

The P(M1LD) and P(M2LD) polymeric tests made with co-stabiliser Crodamol GTCC did not show significant differences between the glass transition temperatures of the syntheses with AIBN and KPS. However, the T_g values shifted to lower temperatures compared to the values obtained in the tests with hexadecane. This result is justified by the more pronounced plasticising effect that Crodamol may have and is probably related to its emollient property, which proved effective in combination with M1LD and M2LD monomers [98, 99]. This factor improves the elongation capacity of the chains and shifts the T_g to lower temperatures [100, 101]. Thus, considering that the cross-linking mechanisms occurred at similar rates with both initiators in the presence of Crodamol, it was observed the decrease in T_g of the tests made with AIBN and the increase in T_g of the tests made with KPS. In the case of the M2LD monomer, which spread well in both co-stabiliser (hexadecane and Crodamol GTCC), the plasticising effect of the latter stood out compared to hexadecane, decreasing the T_g in the tests with both initiators. Hexadecane also has a plasticising character, but with the T_g results obtained in this work, it is possible to infer that Crodamol GTCC acted better. For the polymeric tests P(M3LD), T_g values were inconclusive, due to the insensitive response of the equipment in indicating the T_g of this cross-linked hydrogel polymer.

By TG/DTG analysis three stages (stage I, II and III) of mass loss were verified for all P(M1LD) (Figure S37), P(M2LD) (Figure S38). and P(M3LD) (Figure S39) polymer tests. The stage I, among the polymers, ranged from 152 °C to 242 °C (Table 4). The polymer with the highest thermal resistance is then P(M1LD) followed by P(M3LD) and finally P(M2LD). Therewith, the temperatures obtained for stage I show moderate thermal stability of the obtained materials and, presumably, is related to the more volatile polymer chains of lower molar mass. Stage II may be related to the larger chains formed from the monomeric molecules. Finally, Stage III, may be related to the chains formed from the oligomers present in the monomers and which came to expand their chains through the polymerization reaction.

It was noted that for the polymers P(M1LD) and P(M2LD), starting from AIBN, the onset of decomposition (stage I) occurred at higher temperatures than the polymers starting from KPS. The higher stiffness is inherent to the tighter scaffolds of the crosslinked network (polymers from AIBN), which may have attributed higher thermal resistance to initiate decomposition of these materials. For P(M3LD) polymeric tests, this same justification is valid, however it occurred with the test made from KPS due to the amphiphilic nature of the acrylamide monomer, which allowed the cross-linking mechanisms to be centred in the aqueous phase of the miniemulsion. No pattern of temperature variation was observed between the tests performed with the two types of co-stabilisers (hexadecane and Crodamol GTCC). The next thermal event, stage II, occurred at similar temperatures among all polymers. And stage III occurred at higher temperatures for P(M2LD) polymers (about 432 °C) and lower temperatures for P(M1LD) and P(M3LD) polymers (about 415 °C). The end of decomposition for all polymers occurred close of 450 °C, and at higher temperatures, the TG curve remains constant and does not reach total mass loss, which is also relative to cross-linked materials [102].

SEM Analysis of Biofilms

The SEM images of the biofilm (Fig. 8a-i and Figure S40-42), suggest all three polymer types are susceptible to biofilm formation. The presence of clusters of Comamonas-sp bacteria in the form of bacilli was highlighted in the images, as well as the formation and attachment of biofilms, arising from the consumption of the polymer matrices by the bacteria and consisting of proteins, polysaccharides and nucleic acids [103]. Compared to the SEM images of the controls, there is biofilm and bacteria randomly spread throughout the length of the polymers. It is possible to suggest that the surface of the polymers appears to be visually altered, presenting new cavities and holes scattered over their surfaces. These appear to be different from the cavities/holes/pores of the control images, images of Fig. 6 and Figures S22-S34, indicating the beginning of their biodegradations in a short time, about 10 days of treatment (assay). The degraded surfaces (DS) were highlighted in the images and, by convention, the smooth (SL) and intact surfaces of the control polymers were also signalled in the images. The degradation of the polymer matrix occurred intensely for the P(M3LD)-T1 polymer, in which its surface and interior are quite deteriorated (Fig. 8h-i), being the polymer that suffered the most from the action of bacteria. The polymers P(M1LD)-T5 and P(M2LD)-T1 also suffered enough attacks to damage their initial structure but proved to be more resistant than P(M3LD). The biofilm originating from P(M2LD)-T1 (Fig. 8e-f) appear to be the ones with the greatest biofouling, with the formation of biofilm mats between the bacteria. Therefore, P(M1LD)-T5 (Fig. 8b-c) appears to be the most resistant polymer to biofilm formation and biodegradation. A greater difficulty in biodegrading acrylic polymers has already been observed in literature, even in association with terpenic molecules. Factors such as topology of polymer surfaces, water solubility, ionic structure and molecular flexibility can favour or disfavour an environment for biofouling and biofilm attachment in acrylic polymers [104,105,106,107].

Fig. 8

SEM images for controls and biofilms from polymers: (a-c) P(M1LD)-T5, (d-f) P(M2LD)-T1, and (g-i) P(M3LD)-T1

However, the results obtained in this work probably occurred due to a sum of favorable factors. Limonene is a biodegradable compound and the association of this molecule through chemical modifications with AA, MA and HEAA, may have been a contributing factor to induce the biodegradability of the polymeric materials from this association [108]. In addition, components used as building blocks in the construction of these polymers may also enable a greater probability of biodegradation. Crodamol GTCC is biodegradable and was used as a co-stabilizer in the synthesis of polymers used in biofilm assays. Finally, the conditions under which a material is set to degrade substantially influences the outcome [44, 103]. The bacteria used in this work has great potential to degrade plastics. Peixoto et al. [44] evidenced the metabolic activity of Comamonas-sp bacteria in the polythene matrix as the only energy source. Thus, it is possible that the microbial assimilation for the polymers P(M1LD)-T5, P(M2LD)-T1 and P(M3LD)-T1 occurred in a similar way. The secretion of possible extracellular degrading enzymes, oxidoreductases, and proteins by digesting Comamonas sp. bacteria, when utilizing these polymers as carbon and energy sources, may have stimulated the cleavage of the polymer chains. This promoted the initiation of biodegradation and depolymerization, modifying the substrate to which these bacteria are inserted [44, 109]. Such a scenario was observed, especially for P(M3LD)-T1. However, the result obtained here can be better investigated by applying analysis that contributes to the understanding of these bacteria in the polymeric matrices shown in this manuscript. In addition, a longer inoculation time may contribute to higher biodegradation rates.

Conclusion

In this work, (meth)acrylation and acrylamidation reactions were applied directly to limonene dioxide without the use of organic solvents or catalysts, resulting in the production of four multifunctional biobased monomers, three of which are novel. In situ and real-time monitoring provided crucial information for optimizing the production of di(meth)acrylated, diacrylamidated (M3LD-FA), and acrylamidated (M3LD) monomers.

Three novel limonene-based renewable polymers were synthesized via miniemulsion polymerization of di(meth)acrylated and acrylamidated monomers using an environmentally friendly reaction mechanism. Two of these are poly(meth)acrylates, and the other is a polyamide hydrogel. The use of different initiators and co-stabilizers in miniemulsion preparation allowed for the comparison of their effects on polymer properties, resulting in moderate to high cross-linking and porosity without the need for additional agents. The P(M1LD) and P(M2LD) polymers formed nano- and micrometric spherical particles, while PM3LD exhibited moderate swelling. The polymers displayed different T_g values, enhancing their application potential, showed good thermal stability, and began biodegrading after ten days of exposure to Comamonas sp. bacteria.

This approach enabled the development of renewable, biodegradable polymers as an alternative to petroleum-based synthetics, helping to address the oil shortage. The properties of these polymers allow for a range of applications, from incorporating copolymers or comonomers into the observed pores to more direct uses, such as films for packaging and cosmetics. Therefore, exploring potential applications can serve as a starting point for future studies on the polymers synthesized in this work.