2.1. Treatments of SBP and Recoveries of C in Crude Soluble Products
According to the experimental conditions reported in
Section 4.1, the SBP available from previous work [
1] was dissolved in plain water at pH 10. The solution was used as a control against the same solution added with hydrogen peroxide. The purpose was to test the effect of the added hydrogen peroxide oxidant in comparison with water as the only terminal oxidant in the control solution.
Table 1 reports the experimental conditions under which the SBP solution was treated and also the soluble carbon as mol/mol % of the total C in the SBP before the treatments, which was recovered from the control solution, and the solutions with added hydrogen peroxide at the end of the treatments. During the treatments, some insoluble material formed. This was separated from the soluble phase via centrifugation. The recovered insoluble material accounted for 3–9% of the total C in the SBP before the treatments (
Table S1). The formation of insoluble matter in the oxidation of lignocellulose materials has been reported also in the ozonisation of SBP [
7] and of pine kraft lignin in alkaline solution [
29]. It was attributed to the dehydrogenative coupling and cross-linking of ozonised phenolic moieties induced by active oxygen radicals. In the present work, the insoluble co-product was undesirable. Values of the recovered C mol/mol % for the recovered soluble material (
Table 1) and for the insoluble material were calculated from the respective recovered dry mass weights and C contents reported in
Table S1.
The data in
Table 1 for treatments No. 1-D0 and No. 4-L0, respectively performed without and with light irradiation, in the absence of added H
2O
2 show that the pH of the starting SBP solution in water decreases after 14 days from 10 to 9. For treatments No. 2-D2, 3-D3, 5-L2, 6-L3, performed in the presence of added H
2O
2 at a 2–3 H
2O
2/C mole ratio, without and with light irradiation, the pH decreases to about 5. According to previous works on the ozonisation of the as-collected MBW [
2] and SBPs obtained from fermented MBW [
7], the pH decrease reported in
Table 1 is consistent with the oxidation of SBP organic matter and the formation of carboxylic functional groups and/or CO
2. The same treatments were replicated adding KOH to keep the pH 10 of the pristine SBP solution constant during the treatments performed under the same above conditions. The data for treatments No. 7–12 in
Table 1 show that the amount of added KOH needed to maintain pH 10 is equivalent to the formation of 0.24 acid equivalents produced per mole of the starting organic carbon (H
+/C eq/mol) in the case of the No. 7-D0 and No. 10-L0 treatments, of 0.34 H
+/C eq/mol in the case of No. 8-D2 and No. 11-L2, and 0.40 H
+/C eq/mol in the case of the No. 9-D3 and No. 12-L3 treatments. The data confirm that the production of organic functional COOH groups and/or CO
2 occurs also for the SBP solutions containing no added H
2O
2 and that the produced H
+/C eq/mol amount depends on the amount of added H
2O
2.
Parallel to the pH decrease,
Table 1 shows that the recovered soluble C at the end of treatments No.1–6 decreased along with a decrease in pH of the recovered solution. Similar trends were observed for the weight of the recovered soluble matter and for the C/N mol/mol ratio in the dried recovered soluble matter (
Table S1).
An evaluation of the data for the recovered soluble C in
Table 1 should account for the variability of the measured values, due to handling the recovery of the reaction medium, separating the soluble and insoluble phases, drying, weighing and analysing the products for their C and N content (see
Section 4). To this purpose,
Table 2 reports the mean values and standard deviations calculated over the values for treatments No. 1–6 and No. 7–12, separately. Compared to treatments No. 7–12, treatments No. 1–6 are characterised by significantly lower mean values and also much higher standard deviations values. The large differences in mean and standard deviation values between the two groups stem mainly from the production of CO
2 and its fate. While in treatments No. 7–12 at a constant pH of 10, the produced CO
2 remains in the recovered material in the form of a carbonate, in treatments No. 1–6, it is lost in the gas phase. On one hand, the 5.9 standard deviation for treatments No. 7–12 is likely to represent largely variability due to handling, separating and analysing the crude matter at the end of the treatments. On the other hand, a comparison of the data for the two groups of treatments could potentially allow for an assessment of the effect of pH on the soluble organic matter recovered in the different treatments. This poses the issue of understanding how much of the recovered soluble C in treatments No.7–12 is due to the formation of CO
2 and of organic COOH functional groups.
To assess the relative contributions of CO
2 and organic C, samples of the products obtained in
Table 1 treatments were further treated with HCl in order to obtain CO
2-free samples (see
Section 4).
Figure 1 reports the composition of C types and functional groups in the products determined by 13 C NMR solid state spectroscopy.
The content of the C types and functional groups listed in
Figure 1 was estimated from the measured areas of the 13C NMR resonance band covering the following chemical shift (δ, ppm) ranges: 0–53 for aliphatic (Af) C; 53–63 ppm for amine (NR) and methoxy (OMe) C; 63–95 ppm for alkoxy (OR) C; 95–110 ppm for anomeric (OCO) C; 110–140 ppm for aromatic (Ph) C; 140–160 ppm for phenol/phenoxy C (PhOY, Y = H, R); and 160–185 ppm for carboxylate and amide (COX: X = OM, NR; M = metal, R = H, alkyl and/or aryl) C. The total integrated band area over the 0–185 ppm range was assumed to represent the total C moles in the analysed sample (see
Section 4).
For the crude soluble products listed in
Figure 1A,B, the COX resonance signal was generally very broad, covering all the 160–185 ppm resonance range. Some spectra exhibited a high intensity sharp signal overriding the broad band in the 160–185 ppm range. In these cases, the measured band area for the COX resonance accounted for 40–60% of the total integrated band area over the 0–185 ppm range. This occurred particularly for the products obtained in the treatments at controlled pH 10 (
Figure 1B), which were expected to contain potassium carbonate, formed as a consequence of the mineralisation of the SBP organic matter in the aqueous alkaline medium. Potassium carbonate is characterised by a sharp intense 13C resonance signal at 170.3 ppm [
30]. To determine the content of potassium carbonate, some selected samples of the crude soluble products obtained at constant pH 10 were treated with HCl (see
Section 4), and the expected CO
2-free products were analysed via 13C NMR spectroscopy. The same HCl treatment and spectroscopic analysis was performed on selected samples of the crude soluble products obtained at an acidic pH. This allowed us to assess the possible effects of the HCl treatments, other than decarbonation, on the CO
2-free organic matter of the pristine crude soluble products.
2.2. Recoveries of Soluble Organic C and CO2 for Treatments in Absence of Added H2O2
For the scope of the present work, it was of primary importance to compare the crude soluble products obtained in the treatments carried out in absence of added H
2O
2 with their corresponding samples treated with HCl. According to the data in
Figure 1, the 1-D0 and 4-L0 samples do not show any significant composition difference compared to the corresponding 1-D0dec and 4-L0dec samples subjected to the HCl decarbonation treatment. For each C type and functional group,
Table 3 reports the mean and standard deviation calculated for the 1-D0 and 1-D0dec samples and for the 4-L0 and 4-L0dec samples, separately.
It may be observed that the relative standard deviation values range from 2.8% to 11%, except in the case of the PhOY functional group, which shows rather large 19% and 51% values. This poses doubts about the significance of the 13C signals measured in the PhOY 140–160 ppm. In most cases, the broad PhOY resonance band could hardly be distinguished from the background noise.
On the contrary, the data in
Figure 1 show evidence of large significant composition differences for the 7-D0 and 10-L0 samples compared to the corresponding 7-D0dec and 10-L0dec samples subjected to the HCl decarbonation treatment. In these cases, the relative standard deviations were found to range from 16% to 58% of the mean values calculated for all C types and functional groups. The data in
Figure 1 show that the COX content, much lower in the decarbonated samples than in the pristine crude soluble 7-D0 and 10-L0 samples, is primarily responsible for the large composition differences measured for these pairs of samples. These results prove that the measured area of the broad COX resonance band in the 7-D0 and 10-L0 products includes contributions of the resonance signal of CO
2 C (in form of potassium carbonate) and of organic carboxyl C (COXorg), which were calculated according to the following Equations (1) and (2):
where (
Figure 1) A and C are the values of COX and Af, respectively, for 7-D0dec or 10-L0dec; B is the value of Af for 7-D0 or 7-L0; and D is the value of COX for 7-D0 or 7-L0.
Table 4 reports the calculated values of COXorg and CO
2 (mol/mol %) carbon contained in the products obtained in the 1-D0, 4-L0, 7D0 and 10-L0 treatments listed in
Table 1 and in
Figure 1A,B. As expected, the data confirm that in the products obtained in treatments 7-D0 and 10-L0 at constant pH 10, the produced CO
2 is retained in the form of potassium carbonate in the strong alkaline water phase. On the contrary, no potassium carbonate is found in the products obtained in treatments 1-D0 and 4-L0 without pH control. The amount of CO
2 calculated from the 13C spectroscopic data for 7-D0 and 10-L0 products corresponds to the values calculated from the consumption of the alkali to keep pH 10 constant during the treatments (see
Table 1, column H
+/C).
Similar calculations using Equations (1) and (2) were applied to the data reported in
Figure 1B for the crude soluble product obtained in treatments 11-L2 and 11-L2dec at constant pH in the presence of hydrogen peroxide at a 2 H
2O
2/C mol/mol ratio. The calculated amount of CO
2 C in the 11-L2 sample was 53.2 mol/mol % against 8.7 mol/mol % for organic COX C. These results indicated that the mineralisation of organic C in the presence of hydrogen peroxide is more than double than that (
Table 4) calculated for the 7-D0 and 10 L-0 samples of the crude soluble product obtained in the treatments at constant pH 10 in the absence of hydrogen peroxide without and with light irradiation, respectively.
2.3. Molecular Weight Distribution in Pristine SBP and Crude Soluble Products Obtained in Absence of Added H2O2
Further information on the effects of the 1-D0, 4-L0, 7-D0 and 10-L0 treatments of SBP in the absence of added H
2O
2 was obtained by fractionating the recovered crude soluble materials through sequential membrane ultrafiltration. To this end, the pristine SBP and the recovered soluble products were fed to polysulphone membranes with decreasing molecular cutoffs at 750, 150, 100, 50, 20, 5, and 0.2 kDa, and the collected retentates at each step and the final permeate through the 0.2 kDa membrane were collected, weighed and analysed for their C content.
Figure 2 reports the results of the fractionation process.
The data in
Figure 1A show that the R750, R150 and R100 in order of decreasing weight abundance are the major high molecular weight fractions of the pristine SBP solution (0-SBP in
Figure 1) ultra-filtered readily after its preparation at pH 10. However, keeping the SBP in the alkaline aqueous solution for 14 days (treatment No. 1-D0 in
Table 1) yields a product (1-D0 in
Figure 1A) exhibiting a drastic compositional change with the R750 fraction reduced at a 5.2% level and the R150 and P0.2 fraction increased up to 54% and 40.8%, respectively, compared to the composition of the pristine SBP. For the SBP solution irradiated with simulated solar light for the same time, the crude soluble product (4-L0 in
Figure 1A) exhibits a further reduction in the fractions’ molecular weight, with the R750 and R150 accounting for 6.5% and the R20 and P0.2 fractions accounting for over 89.1% of the total recovered material. As shown in
Table 1, during the 14-day treatments, the pH of the two 1-D0 and 4-L0 solutions decreased to 9. For the products obtained in the treatments carried out at constant pH 10 (i.e., 7-D0 and 10-L0 in
Figure 1A), the reductions in molecular weight seemed less compared to the 1-D0 and 4-L0 products in
Figure 1A. For 7-D0 (in
Figure 1A) obtained without irradiation of the pH 10 reaction medium, the R750 fraction accounted for 30% and the R150 for 6.5%, compared to 5% for R750 and 54% for R150 in the 1-D0 product (in
Figure 1A) obtained without pH control. For 10-L0 (in
Figure 1A) obtained from the irradiated reaction medium, the R50 accounted for 14.2% against 0% for the 4-L0 product, while the lower molecular weight fractions R20, R0.2 and P0.2 altogether were 86.2% in the 10-L0 product and lower than 93.3% in the 4-L0 product (
Figure 1A) obtained without pH control of the reaction medium. For the 4-D0 and 10-L0 products, the P0.2 fraction is supposed to contain potassium carbonate, the product of the mineralisation of organic C (see
Section 2.2).
Figure 1B reports the total C distribution over each of the above materials. For the pristine SBP (0-SBP in
Figure 1B), most of the total carbon (69%) is accounted by the R750 fraction. For the crude soluble products obtained in each treatment, the carbon recoveries are given as
w/
w % relative to the carbon in the pristine SBP. The C recovery values are calculated based on the mass data given in
Figure 1A and the C content measured for each fraction. It may be observed that the pattern of C recovery distribution changes significantly depending upon the type of treatment. The carbon recovered with the R750 fractions of the crude soluble products ranges from 0.5% for 10-L0 to 30% for 7-D0, while most of the remaining C (46–81%) is recovered with the lower molecular weight fractions.
Both the mass and C recovery data in
Figure 1 indicate that all treatments cause depolymerisation of the pristine SBP organic matter and that the effect is stronger in the treatments carried out via irradiation and/or without pH control of the reaction medium. These findings offer highly relevant information, which, coupled with the results reported in
Section 2.1 and
Section 2.2, undoubtedly demonstrate the autocatalytic properties of SBP to react with water and lead to the depolymerisation and mineralisation of its own organic matter.
2.4. Products Obtained in the Presence of Added H2O2
Figure 3A reports the C recoveries, as mol/mol % relative to the carbon of the pristine SBP, with each fraction isolated from each crude soluble product obtained in all SBP treatments listed in
Table 1. The C recoveries are calculated from the weight and carbon content of the fraction isolated through the same membranes listed in
Figure 2. The different colours of the histogram columns identify the different retentates and permeates obtained via the ultrafiltration.
The data show the effect of hydrogen peroxide on the molecular weight distribution of the crude soluble products. For the treatments carried out at constant pH 10, without (7-D0, 8-D2, 9-D3) and with irradiation (10-L0, 11-L2, 12-L3), the data show an increase in the lowest molecular weight fractions R0.2 and/or P0.2 occurring in the presence of added H
2O
2 compared to the reaction in the absence of H
2O
2. This is readily evidenced in
Figure 3B, which reports the total C recovered with the retentate (R0.2) and permeate (P0.2) fractions as the % of the total C recovered with all fractions.
It may be observed in
Figure 3B that, in the case of the treatments carried out without light irradiation and with no pH control (1-D0, 2-D2 and 3-D3), the total production of R0.2 and P0.2 increases upon increasing the added content of H
2O
2 from 2 to 3 moles per SBP C mole. In the other cases, no effect or significant trend appears evident by increasing the content of H
2O
2 above 2 moles per SBP C mole. A somewhat similar trend is observed due to light irradiation. Compared to the 2-D2 treatment, the 5-L2 treatment caused a strong increase in the total production of R0.2 and P0.2. In all other cases, no definite effect or trend may be picked out as being caused by light irradiation.
The data in
Figure 3 show that the depolymerisation and/or mineralisation of SBP organic matter is particularly evident in treatments 3-D3, 5-L2, and 6-L3, where the sum of the R0.2 and P0.2 % values in
Figure 3B accounts for 74–93% of the values reported in
Table 1 for the total C recovered with 3-D3, 5-L2, and 6-L3 treatments. Considering the data in
Table 1 and
Figure 1,
Figure 2 and
Figure 3, it appears evident that the SBP treatments in the presence of hydrogen peroxide at the 2 and 3 H
2O
2/C ratio are too drastic, due to the enhancement of depolymerisation and mineralisation via the oxidation of the SBP organic matter. Similar results were obtained for the SBP treatments at controlled pH 10 and the 0.1–0.5 H
2O
2/C mole ratios not reported here. Under these conditions, 0.06–0.15 H
+/C eq/mole were produced. This indicated that the oxidation of the SBP organic C to organic COOH functional groups and/or CO
2 was quite less than it was in treatments No. 7–12 at the 2–3 H
2O
2/C mole ratio reported in
Table 1.
2.5. Chemical Composition and Properties of the Molecular Weight Fractions Obtained in All Treatments
Figure 4 and
Figure 5 report data related to the chemical composition and surface activity properties of the fractions obtained from the crude soluble products obtained in all treatments carried out in the present work. More detailed data are given in
Table S3.
For reactions involving products of complex chemical composition such as SBP, the ratio of the carbon to nitrogen (C/N) content in a product has been used as an index for the effect of a chemical or a biochemical reaction on the chemical nature of the product. For example, for the chemical hydrolysis of a wide variety of municipal biowaste composts, the following Equation (3) has been proposed [
1]:
where Z is the ratio of the measured total aliphatic, aromatic, carboxylic, phenol, phenoxide, methoxy, amine, amide and ketonic carbon over the sum of amine, amide and carboxylate functional groups. A similar significant linear relationship has been confirmed between the C/N values of different composts with the C/N values of the derived SBP products [
27]. The relationship allows one to predict the chemical composition of SBPs obtained from different composts. According to the authors of [
27], it constitutes a valuable tool to assist the industrialisation of the SBP production process.
The data in
Figure 4 show that the C/N values vary over a very wide range, from a minimum 3.3 value for the P0.2 fraction of the crude soluble product obtained in the 3-D3 treatment, listed in
Table 1, to 49 for P0.2 of the crude soluble product obtained in the 15-L0 treatment of the pristine SBP solution carried out at controlled pH 10, under irradiation and in the absence of added H
2O
2. Most of the fractions in
Figure 4 have C/N values different from the 8.7 value for the pristine SBP reported in
Table S1. The variability of the C/N and molecular weight values of the fractions composing the crude soluble products obtained in the different treatments of the pristine SBP reflects the complexity of the supply chain that includes the pristine MBW and the SBP anaerobic digestate, from which the products are derived. In essence, each fraction in
Figure 4, obtained via ultrafiltration of the crude soluble products described in
Section 2.1,
Section 2.2,
Section 2.3 and
Section 2.4, is a mixture of molecules with not only different molecular weights, but also different chemical compositions. With specific references to the P0.2 fractions, the C/N values higher than 8.7 (measured for the pristine SBP) are likely contributed mostly by the content of potassium carbonate CO
2. This occurs for most of the P0.2 fractions isolated from the crude soluble products obtained in the treatments at constant pH 10, as anticipated by the data for the crude soluble products obtained at constant pH 10 (see
Section 2.2). On the contrary, low C/N values, for example, C/N values of 3.3 and 4.4 for P0.2 isolated from the 3-D3 and 6-L3 crude soluble products, indicate the presence of small molecules containing N, as organic oxymes isolated in the ozonisation of SBP [
7].
Table 5 reports mean C/N values calculated from the single C/N values (
Figure 4 and
Table S3) of the fractions isolated as retentates or permeates through membranes with the same cutoff values.
Mean C/N values have relatively high relative standard deviations, which do not allow for the confirmation that the differences between average values are statistically significant. Many factors contribute to high standard deviations. For the complex molecular mixtures dealt with in the present work, the major factors are the different reaction conditions adopted, e.g., the applied H
2O
2/C mole ratio, together with the replicability of the course of the reaction under the same experimental conditions and the change of the solution conformation of the molecular pool constituting the crude soluble products from which the corresponding fractions in
Figure 4 are isolated. The polymeric molecules in the crude soluble products, through their basic and acidic functional groups, may establish intermolecular H-bonds. Depending on their concentration in water, these molecules form aggregates [
24,
25,
26,
27,
28,
29,
30,
31,
32,
33,
34,
35,
36] of different sizes trough the H-bond network established between macromolecules of different compositions. In dilute solutions, the size of these aggregates [
37] is likely to decrease, and therefore the composition of retentates through a membrane with a specific cutoff may change depending on the concentration of the solution fed to the membrane. Yet, within the limitations posed by the relatively high variability of the C/N parameter, the data in
Table 5 show some apparent differences that may help to rationalise the nature of the crude soluble products. For example, the lowest molecular weight P0.2 fractions have the highest C/N average (14.4) and standard deviation (11.0) corresponding to STDr 76%. These results are most likely due to the fact that the P0.2 fractions contain variable amounts of carbonates, as anticipated in
Section 2.2. The same may be true for the R0.2 fractions. By comparison, all other fractions exhibit lower average C/N values in the 6.7–9.5 range and lower STDr in the 20–55% range. Excluding the P0.2 and R0.2 fractions, the apparent order of average C/N values is R20 > R750 > R150 = R50 > R5. The C/N data therefore indicate that, based on Equation (3) and on the molecular weight, the R750 and R5 have, respectively, the lowest and the highest relative content of amino carboxylic and peptide functional groups plausibly organised in protein-like moieties with different molecular weights.
According to previous works [
7,
37,
38], the measurement of surface tension in water solutions is a diagnostic tool, which indicates the behaviour of a product in a solution and its potential performance as surfactant. In the present case, surface tension measurements were carried out as a mean for rating the many different fractions listed in
Figure 3 and
Table 4 on the basis of the surface tension measured for their water solutions and therefore for their potential prospected value in the chemical market. The results of these measurements for selected samples are reported in
Figure 5 and also in
Tables S3 and S4. The graphical representation of the measured surface tension (γ) values in different colours (
Figure 5) allows us to observe readily that the experimental γ data may be divided into three groups, i.e., γ < 50; 50 ≤ γ < 60; and γ ≥ 60.
The data points and the standard deviation bars in
Figure 5, and the statistical analysis given in
Table 6, show that the mean values of the three groups are significantly different from each other. The SBP treatments at H
2O
2/C 0.5–0.1 mole ratios without pH control, and without or with irradiation yield products with the lowest γ < 50 values (
Figure 5). All others have γ > 50 values. All samples obtained by virtue of the autocatalytic properties of SBP in the absence of added H
2O
2 seem to exhibit the highest γ ≥ 60 surface tension values. On the other hand, the treatments in the presence of hydrogen peroxide at 2–3 H
2O
2/C mol/mol seem to yield fractions with some improvement in the surface activity properties, but they produce a high degree of depolymerisation of the pristine SBP organic matter. The treatments at H
2O
2/C 0.5–0.1 mole ratios produce also depolymerisation of the pristine SBP organic matter but yield the fractions exhibiting the best surface activity property.