Molecular Characterization of the Organic Fraction of Municipal Solid Waste and Compositional Evolution during Oxidative Processes Assessed by HR-MAS 13 C NMR Spectroscopy

: The use of high-resolution magic angle spinning (HR-MAS) 13 C NMR spectroscopy is proposed here as an innovative and non-destructive approach to investigate the chemical composition of the organic fraction of municipal solid waste (OFMSW) and for monitoring the evolution of their composition during the oxidative iron-based Fenton treatment to which the initial matrix is subjected. The high quality and the good resolution of the 13 C HR-MAS NMR spectra allowed an accurate assignment and quantiﬁcation of the various types of carbon present in the analyzed organic matrix. Moreover, the HR-MAS has also shown its effectiveness in monitoring the different oxidative processes to which the same initially organic matrix has been subjected. The results obtained from the HR-MAS spectra on the collected samples during the different oxidative experiments, indicate that Fenton treatment is able to modify the percentage of the different types of carbons as a function of the concentrations of both Fenton reactants, H 2 O 2 and Fe 2+ salt, and of the oxidative process time.


Introduction
Solid waste management has become a relevant problem in modern society since the development of urban areas has led to a growth in waste production and consequently to a reduction in the places where all types of materials can be deposited [1]. The reduction in the quantity of waste to be stored in landfills has become a pressing problem and, consequently, different strategies and treatment methods suitable for this purpose are adopted [2,3]. First of all, recycling combined with avoidance, separate collection, and reuse contribute not only to waste reduction but also to energy saving and to the reduction in the amount of polluting and climate-altering emissions [4,5]. All these actions, and in particular separate collection, are very useful in reducing and recycling waste, but are not sufficient to achieve complete abatement, hence the need to develop new waste disposal and treatment methods that can convert waste into new products to be used in different fields so that the potential of waste can be fully exploited [6,7]. In order to choose the right technique, the first step is to classify waste according to its composition [8]. In this study, our attention is focused on the organic fraction of municipal solid waste (OFMSW), which includes food residues or food preparations and assimilable fractions, and which constitutes more than 30% of the total weight of municipal solid waste [9]. This waste fraction is generally converted into biogas [10] through anaerobic digestion [11][12][13] or transformed into soil improvers through a biological process of composting for use in agriculture [8,14,15]. These practices have many disadvantages related both to the time required for bacteria to degrade the various organic substances and to the possible problems of sickening smell in an organic matrix represented by carboxymethyl cellulose (CMC) [37]. In this work, the authors use the integrals obtained from the 13 C HR-MAS spectra to calculate the DS of different CMC samples and, to validate the method, they conduct experiments on specially prepared standards under controlled conditions.
The main purpose of this work was to probe the feasibility of using the HR-MAS technology to characterize the families of organic materials present in a sample of OFMSW and to monitor possible variations in their composition following iron-based oxidative treatments (Fenton reaction). To this end, we used the qualitative and quantitative information extracted from the 13 C-{ 1 H} HR-MAS NMR spectra recorded on a starting sample of OFMSW and on samples collected during, or at the end, of several oxidative treatments to which the same starting material was subjected. Note that, in order to understand if the HR-MAS technique was able to appreciate any variations in the composition of organic families, we followed different Fenton treatments applied to the same starting matrix, varying different parameters, namely the concentrations of the Fenton reagents (FeSO 4 and H 2 O 2 ) and the reaction times.

Sample Collection and Preparation for NMR
The samples were supplied from another department of the project and the experimental details of this process are given outside this work, with only the highlights being reported below. The waste used for the Fenton treatment was obtained through a sampling procedure lasting 12 days. In each of these days, waste was taken from a group of four different families, each day a different one. The waste collected on each sampling day (about 5 kg) was mixed and milled to a millimeter size by means of a steel blender. The resulting minced material was stored at −20 • C for 12 days. After this time, all the daily rates were defrosted and mixed together; the material obtained was divided into bags of 200 g, re-frozen and then used, after heat treatment at 110 • C, as the starting matrix for the oxidative treatment. The oxidation reaction [18] was carried out using a laboratory-scale glass reactor apparatus. The sample of organic material, 200 g, was placed into the glass rector and the Fenton reagents were added: 10 mL of a FeSO 4 solution followed by addition of 10 mL of hydrogen peroxide (H 2 O 2 ) solution. The mixture was kept under stirring conditions, the reactor was placed in a water bath, and the internal temperature was kept at 60 • C for the duration of the reaction. The pH of the mixture, initially at 4.5, was allowed to decline to pH 3.0 and was maintained until the reaction was complete. The samples used for the NMR experiments were collected during the phases of oxidation procedure and were differentiated according to the different percentages of H 2 O 2 /FeSO 4 used. A first sample, called Sample 0, was collected from the batch of organic waste subjected to heat treatment (110 • C for 48 h). Three samples, called Sample 1, Sample 2, and Sample 3, were collected, and waste was treated for 12 h with the same amount of FeSO 4 , 0.05%wt, but different concentrations of H 2 O 2 , i.e., 0.002%wt, 0.0015%wt, and 0.0006%wt respectively. In order to understand the transformations that the waste undergoes during the oxidation process and to investigate the best experimental conditions, samples were collected at regular intervals during two different process (process A and process B  [22], where the authors applied the Fenton reaction to an anaerobic digestate sample. In Figure 1, the experimental steps, the reaction conditions and the NMR samples collected during all oxidative treatments are reported. For the NMR experiments, samples were prepared

NMR Spectroscopy and Data Processing
All HR-MAS NMR spectra were obtained on a Bruker Avance 500 MHz spectrometer (Fällanden, Switzerland), operating at a 13 C resonance frequency of 125.758 MHz, equipped with a BVT-3000 device for temperature control and an HR-MAS 5 mm probe. All measurements were made at a spin rate of 4 KHz and at room temperature, acquiring 32 k points with 5000 scans and using the inverse-gated pulse sequence (zgig) in order to avoid the nuclear Overhauser effect (NOE) and obtain quantitative 13 C information from the 13 C-{ 1 H} NMR spectrum. Data were processed using Bruker TopSpin 3.6 software (Bruker BioSpin, Rheinstetten, Germany) [38]: the time-domain spectra were apodised with exponential function, and then Fourier-transformed, phased, and baseline-corrected automatically. In order to remove the background signal due to the Kel-F rotor cap, to each spectrum has been subtracted the 13 C spectrum obtained from a blank sample consisting in D2O and TMS. The chemical shifts were referred to the TMS signal (δ = 0.00 ppm). In the semi-quantitative procedure, the global chemical shift range of carbon in the spectra was subdivided into sub-regions, as shown in Table 1.

Chemical Shift Range (ppm) Assignment
220 to 190 C=O of ketones and aldehydes 190 to 160 C=O in carboxylic acids, esters and amides 160 to 140 C-O-R and C-N-R aromatics; C-1 in polystyrene 140 to 110 Aromatic C-H; C-2, C-3, C-4, C-5 and C-6 in polystyrene 110 to 90 C-1 in carbohydrates; C-2 and C-6 in the lignin syringes unit 90 to 60 C-2 and C-6 in carbohydrates; side chain in lignin 60 to 50 Methoxy groups, Ca in amino acids 50 to 40 Methylene groups (lipids, proteins, polyethylene, polyisoprene, polyamide). Long chain aliphatic C, also biopolymers 40 to 25 Cα and Cβ in polymers such as polyvinylchloride, polypropylene, polystyrene 25 to 0 Methyl groups (lipids, peptides, polypropylene). Short chain aliphatic C Each signal was integrated using the TopSpin 3.6 software and was expressed as a percentage respect to the total integration area. The integration errors were calculated using the root-mean-square (RMS) method, by which the noise RMS was calculated as:

NMR Spectroscopy and Data Processing
All HR-MAS NMR spectra were obtained on a Bruker Avance 500 MHz spectrometer (Fällanden, Switzerland), operating at a 13 C resonance frequency of 125.758 MHz, equipped with a BVT-3000 device for temperature control and an HR-MAS 5 mm probe. All measurements were made at a spin rate of 4 KHz and at room temperature, acquiring 32 k points with 5000 scans and using the inverse-gated pulse sequence (zgig) in order to avoid the nuclear Overhauser effect (NOE) and obtain quantitative 13 C information from the 13 C-{ 1 H} NMR spectrum. Data were processed using Bruker TopSpin 3.6 software (Bruker BioSpin, Rheinstetten, Germany) [38]: the time-domain spectra were apodised with exponential function, and then Fourier-transformed, phased, and baseline-corrected automatically. In order to remove the background signal due to the Kel-F rotor cap, to each spectrum has been subtracted the 13 C spectrum obtained from a blank sample consisting in D 2 O and TMS. The chemical shifts were referred to the TMS signal (δ = 0.00 ppm). In the semi-quantitative procedure, the global chemical shift range of carbon in the spectra was subdivided into sub-regions, as shown in Table 1. Table 1. Chemical shift regions of 13 C-{ 1 H} NMR spectra of OFMSW [29,30].

Chemical Shift Range (ppm) Assignment
220 to 190 C=O of ketones and aldehydes 190 to 160 C=O in carboxylic acids, esters and amides 160 to 140 C-O-R and C-N-R aromatics; C-1 in polystyrene 140 to 110 Aromatic C-H; C-2, C-3, C-4, C-5 and C-6 in polystyrene 110 to 90 C-1 in carbohydrates; C-2 and C-6 in the lignin syringes unit 90 to 60 C-2 and C-6 in carbohydrates; side chain in lignin 60 to 50 Methoxy groups, Ca in amino acids 50 to 40 Methylene groups (lipids, proteins, polyethylene, polyisoprene, polyamide). Long chain aliphatic C, also biopolymers 40 to 25 Cα and Cβ in polymers such as polyvinylchloride, polypropylene, polystyrene 25 to 0 Methyl groups (lipids, peptides, polypropylene). Short chain aliphatic C Each signal was integrated using the TopSpin 3.6 software and was expressed as a percentage respect to the total integration area. The integration errors were calculated using the root-mean-square (RMS) method, by which the noise RMS was calculated as: where λ is the intensity of the i-point (in ppm) out of the total n-points. The RMS value, obtained in this way, represents the height of the noise. Subsequently, this value was used to estimate the error expressed as a percentage of the intensity of the signal and this percentage has been used to estimate the error bar of the integral of each NMR peak [39].

Results and Discussion
The most important advantage of the HR-MAS technique compared to the solid-state NMR is the possibility to obtain well-resolved NMR spectra [40]. Both 1 H and 13 C NMR spectra can be obtained with a spectral quality similar to the liquid state. However, since the organic waste is a complex system, the 1 H HR-MAS NMR spectra are quite complex, due to the overlapping of the signals of the various compounds, and, therefore, for our investigations, the 13 C spectra have been used. A typical 13 C-{ 1 H} NMR spectrum of the organic waste obtained using the HR-MAS NMR technique is shown in Figure 2a, where the carbon spectrum of Sample 0 (in blue) is compared with an OFMSW spectrum found in literature and obtained using the CP-MAS NMR technique (in black) [26]. Although the two spectra show a similar distribution of the peaks, an improvement in resolution is clearly visible in the HR-MAS spectrum in which the signal lines are much narrower. Figure 2b shows the NMR HR-MAS 13 C-{ 1 H} spectrum of Sample 0 in which the different regions are indicated. The region from 12 to 40 ppm is the classic region of aliphatic carbons signals, probably due to the presence of lipids and fatty acids in the OFMSW mixture that we used. The region from 58 to 100 ppm is due to the carbon atoms signals linked to heteroatoms. Note that this region of the spectrum has the same profile as the typical CP-MAS starch spectrum, as shown in the box of Figure 2b where the comparison of this region is reported [31]. From 125 to 130 ppm, this is the region of unsaturated carbon atoms signals, i.e., signals due to unsaturation in fatty acids and/or to aromatic species; finally, the region from 169 to 183 ppm is related to the signals of carbonyl carbons (acids, esters, ketones, amides, and aldehydes).
A detailed list of the overall chemical shift range of carbons is shown in Table 1 and these chemical shift ranges are used to evaluate the area of the signals. Indeed, unlike the CP-MAS spectrum in which it is possible to integrate only large signal ranges, in the spectrum of Figure 2b, the good resolution allows a better integration on the individual signals. In particular, the area of each signal of the spectrum was integrated and the quantitative information was grouped according to the nature of the atoms (aliphatic carbon atoms: C aliph ; carbon atoms linked to heteroatoms: C X ; unsaturated carbon atoms: C unsat and carbonyl carbon: C carb ). In Table 2, the 13 C NMR integrated area of the different carbon type of Sample 0 are reported. In order to understand if the HR-MAS method is able to appreciate the possible chemical differences among the initial organic substrate and the products after the oxidative process, three samples, Sample 1, Sample 2, and Sample 3, were analyzed. The three samples are related to oxidative processes in which the initial OFMSW matrix was treated with the same amount of FeSO 4 but with different quantities of H 2 O 2 . It is worth noting that the carbon spectra of Sample 1, Sample 2, and Sample 3 have the same spectral profile of Sample 0 and not substantial variations in the chemical shift values were observed in the spectra of the four samples (see Figure S1). The integration process on the various types of carbon was repeated for these samples and the results, reported in Table 2 and illustrated in Figure 3, were compared with the corresponding values obtained for Sample 0. Each sample analyzed shows different contents of C containing groups, and these fluctuations cannot be attributed to the error in the evaluation of the integral area, but, evidently, they derive from the different oxidative processes. The greatest variations occur for the aliphatic zone and for the region concerning the carbons linked to heteroatoms. Indeed, the highest content of aliphatic C is found for Sample 3, which is the sample obtained with the lowest amount of H 2 O 2 (0.0006 %wt), while the highest content of C linked to heteroatoms is obtained for the Sample 1, in which the quantity of H 2 O 2 was the highest (0.002 %wt). These results indicate that: a) the Fenton treatment is able to make changes in the initial OFMSW matrix; b) HR-MAS 13 C NMR spectroscopy is effective in appreciating the variations that individual families of compounds undergo following the different treatments, in terms of reagent concentrations used. A detailed list of the overall chemical shift range of carbons is shown in Table 1 and these chemical shift ranges are used to evaluate the area of the signals. Indeed, unlike the CP-MAS spectrum in which it is possible to integrate only large signal ranges, in the spectrum of Figure  2b, the good resolution allows a better integration on the individual signals. In particular, the area of each signal of the spectrum was integrated and the quantitative information was grouped according to the nature of the atoms (aliphatic carbon atoms: Caliph; carbon atoms linked to heteroatoms: CX; unsaturated carbon atoms: Cunsat and carbonyl carbon: Ccarb). In Table  2, the 13 C NMR integrated area of the different carbon type of Sample 0 are reported. In order to understand if the HR-MAS method is able to appreciate the possible chemical differences among the initial organic substrate and the products after the oxidative process, three samples, Sample 1, Sample 2, and Sample 3, were analyzed. The three samples are related to oxidative processes in which the initial OFMSW matrix was treated with the same amount of FeSO4 but with different quantities of H2O2. It is worth noting that the carbon spectra of Sample 1, Sample  Table 2. Integral values, in percentage, of different carbon type, aliphatic (C aliph ), heteroatoms-linked (C X ), unsaturated (C unsat ) and carbonylic (C carb ), calculated for the Samples 0, 1, 2, 3, and 4 at 12 h.

C aliph (%) C X (%) C unsat (%) C carb (%)
Sample 0 54 ± 1 32 ± 3 9 ± 1 5 ± 1 Sample 1 43 ± 1 44 ± 1 9 ± 1 4 ± 1 Sample 2 53 ± 2 28 ± 5 13 ± 1 6 ± 1 Sample 3 56 ± 1 31 ± 1 8 ± 1 5 ± 1 Sample 4 60 ± 1 26 ± 1 5 ± 1 9 ± 1 Caliph (%) CX (%) Cunsat (%) Ccarb (%) Sample 0 54 ± 1 32 ± 3 9 ± 1 5 ± 1 Sample 1 43 ± 1 44 ± 1 9 ± 1 4 ± 1 Sample 2 53 ± 2 28 ± 5 13 ± 1 6 ± 1 Sample 3 56 ± 1 31 ± 1 8 ± 1 5 ± 1 Sample 4 60 ± 1 26 ± 1 5 ± 1 9 ± 1 The experiments just described show that the NMR HR MAS method provides relevant information on the composition variations of the organic fraction of the waste subjected to the Fenton treatment. In order to better understand what happens to the various families of compounds during the oxidative process and to investigate the best reaction conditions, the HR-MAS NMR technique was applied to the study of two oxidative processes, A and B, carried out in different reaction conditions, as described in Section 2.1. A total of nine samples were collected for the two processes, four for process A and five for process B. For these samples, no alterations were observed in the profile of the 13 C HR-MAS NMR spectra, but there are marked variations in the areas of the values associated with each region of the spectra. The results for process A are reported in Table 3 and illustrated in Figures 4 and S2, while results for process B are reported in Table 4 and illustrated in Figures 5 and S3, for both processes the values of The experiments just described show that the NMR HR MAS method provides relevant information on the composition variations of the organic fraction of the waste subjected to the Fenton treatment. In order to better understand what happens to the various families of compounds during the oxidative process and to investigate the best reaction conditions, the HR-MAS NMR technique was applied to the study of two oxidative processes, A and B, carried out in different reaction conditions, as described in Section 2.1. A total of nine samples were collected for the two processes, four for process A and five for process B. For these samples, no alterations were observed in the profile of the 13 C HR-MAS NMR spectra, but there are marked variations in the areas of the values associated with each region of the spectra. The results for process A are reported in Table 3 and illustrated in Figure 4 and Figure S2, while results for process B are reported in Table 4 and illustrated in Figure 5 and Figure S3, for both processes the values of the integrated areas are compared with the corresponding values of the initial OFMSW matrix (Sample 0). Table 3. Integral values, in percentage, of C aliph , C X , C unsat , and C carb , calculated for Sample of process A and comparison with Sample 0.

Process A C aliph (%) C X (%) C unsat (%) C carb (%)
Sample 0 54 ± 1 32 ± 3 9 ± 1 5 ± 1 Sample 4_3 h 63 ± 2 24 ± 2 6 ± 1 Appl. Sci. 2021, 11, x 8 of 12 the integrated areas are compared with the corresponding values of the initial OFMSW matrix (Sample 0).  • Process A. From the data reported in Table 3 and Figure 4, it can be deduced that the reaction proceeds according to two main steps. The first stage, stage I, goes from 0 up to 9 h: during this time, as the reaction proceeds, a decrease in the content of carbon atoms • Process A. From the data reported in Table 3 and Figure 4, it can be deduced that the reaction proceeds according to two main steps. The first stage, stage I, goes from 0 up to 9 h: during this time, as the reaction proceeds, a decrease in the content of carbon atoms linked to heteroatoms is observed. Since the data present in the literature on organic waste show that the amount of carbons linked to oxygen (C-O) is greater than the amount of any other C-X [23], it is probable that this also occurs for our samples and, therefore, this decrease can be attributable, for example, to the oxidation of alcoholic groups, mostly present in sugars, to carbonyl groups. On the other hand, however, there is no substantial variation in the areas related to the carbon atoms of the carbonyl groups and this suggests that probably, in this period of time, further oxidation of carbonyl groups with loss of CO 2 occurs. It should be emphasized that it cannot be deduced in any way from the recorded spectra as to whether the CO 2 produced comes from the newly formed or pre-existing carbonyl groups. Stage II occurs around 12 h: at this time, there is an evident increase in the area of carbon atoms linked to heteroatoms due, probably, to an oxidation of aliphatic carbons and, to a lesser extent, to the unsaturated ones. As evidence of this, a decrease in the areas due to these carbon groups can be appreciated. Table 4. Integral values, in percentage, of C aliph , C X , C unsat , and C carb , calculated for samples of process B and comparison with Sample 0 and Sample 3.

Process B C aliph (%) C X (%) C unsat (%) C carb (%)
Sample 0 54 ± 1 32 ± 3 9 ± 1 5 ± 1 Sample 3  • Process B. From the data reported in Table 4 and from the relative distribution of Figure  5, it is possible to infer that the process B proceeds in the same way as for process A, but the two steps described above start and finish at different times. Indeed, Stage I of Sample 5 goes from 0 to 6 h, unlike Sample 4, in which it continues up to 9 h, while Stage II occurs at the time of about 8 h compared to 12 h for process A. Moreover, for Sample 5, it is possible to hypothesize a third stage between 9 and 12 h: at this moment, the trend previously observed in Stage I, i.e., an oxidation of the alcoholic groups to carbonyls with simultaneous loss of CO2, occurs. In Table 4 and Figure 5 are also reported the NMR data obtained for the Sample 6_8 h, in which the same concentrations of Sample 5 are used but the reaction proceeds continuously for 8 h. As can be seen from Table 4 and Figure 4, in Sample 6_8 h, there is a slight increase in the carbon atoms bound to the heteroatoms compared to the situation observed at 6 h. A decrease in the aliphatic content at 8 h is not   Table 4 and from the relative distribution of Figure 5, it is possible to infer that the process B proceeds in the same way as for process A, but the two steps described above start and finish at different times. Indeed, Stage I of Sample 5 goes from 0 to 6 h, unlike Sample 4, in which it continues up to 9 h, while Stage II occurs at the time of about 8 h compared to 12 h for process A. Moreover, for Sample 5, it is possible to hypothesize a third stage between 9 and 12 h: at this moment, the trend previously observed in Stage I, i.e., an oxidation of the alcoholic groups to carbonyls with simultaneous loss of CO 2 , occurs. In Table 4 and Figure 5 are also reported the NMR data obtained for the Sample 6_8 h, in which the same concentrations of Sample 5 are used but the reaction proceeds continuously for 8 h.
As can be seen from Table 4 and Figure 4, in Sample 6_8 h, there is a slight increase in the carbon atoms bound to the heteroatoms compared to the situation observed at 6 h. A decrease in the aliphatic content at 8 h is not observed, but a decrease in the unsaturation content is appreciable and suggests that this process (Stage II) begins with the oxidation of the double bonds and that it only subsequently proceeds with the oxidation of the saturated chains, as highlighted in Sample 5_9 h.

Conclusions
In this work, we have successfully applied, for the first time, NMR 13 C HR-MAS NMR spectroscopy, an innovative and non-destructive technology, to organic waste materials in order to investigate their chemical composition and verify possible chemical variations, with respect to the initial substrate, by following several oxidation treatments as a function of both the concentrations of the Fenton reagents (FeSO 4 and H 2 O 2 ) and the reaction times. The better performance of the HR-MAS 13 C NMR approach with respect to the CP-MAS method, the most commonly used technique to analyze the organic fraction of waste, relies upon the possibility of obtaining high-quality and much more resolved spectra, whose analysis contributes to a better understanding of the chemical nature of these waste. In addition, this HR-MAS approach has also proved suitable for monitoring the oxidative processes based on the Fenton reaction to which the initial waste matrix was subjected. Indeed, the results obtained highlight the qualitative/quantitative modifications of the different types of carbons as a function of the concentrations of both Fenton reactants and the duration of the oxidative process. These results are in agreement with the data reported in the literature regarding 13 C NMR spectra, both in solids (CP-MAS) and in high resolution for extracts, in characterizing the organic component of OFMSW and in following the various natural composting processes [30]. Moreover, our results in terms of the variation of organic composition of the initial materials are in agreement with the data from thermogravimetry (TG) and FT-IR analysis applied to extracts of a digestate sample subjected to different Fenton treatment varying the concentration ratios between the reagents FeSO 4 and H 2 O 2 [22].
To our knowledge, this case study represents the first example in which the HR-MAS approach has been successfully applied to the organic fractions of OFMSW. Finally, the advantages in terms of rapidity, simplicity, and diversity of obtainable information makes this technique very attractive for further studies on different matrices of OFMSW and also to monitor other degradation/transformation processes to which organic waste can be subjected.  Data Availability Statement: The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.