The electrochemical cell oxidized the benzyl phenyl ether at three different electrode potentials (0.6, 0.8, and 1.0 V vs. Hg/HgO) in the absence and presence of DMSO. The solution obtained after dissolving the lignin model compound in a 1:3 volume ratio of acetonitrile (CH3CN) to sodium hydroxide (NaOH) was split into two parts. One of the resulting solutions was supplemented with DMSO as the radical scavenger, resulting in two indistinguishable solutions except for DMSO's presence in one. Then, each solution was divided into two parts, with one half being sealed and designated as the neat sample and the other fated for oxidation.

Using SPME GC–MS, aromatic compounds such as 2-phenylmethyl phenol, phenoxymethyl benzene, 1,2-benzenedicarboxylic acid, and butyl 2-methylpropyl ester were identified from the oxidation of the model compound.

Based on the findings from the oxidation of the lignin model compound, the first step in depolymerization may be the abstraction of the benzylic H atom followed by cleavage of the β–O–4 bond. The outcomes might point to a positively charged initial oxidation product that undergoes further reactions with OH⁻ generated electrochemically at the cathode during the hydrogen evolution reaction. The highest potential investigated was 1.0 V vs. Hg/HgO.

Table 2 reports the oxidized products obtained from the \(}_}}}\) and \(}_}}}^}}\) samples. Figure 2 shows the TICs of the \(}_}}}^}}\) oxidized sample overlaid on the \(}_}}}\) TICs. The oxidized samples have different intensities, and several additional peaks observed in the \(}_}}}\) samples were absent in the \(}_}}}^}}\) samples. The peaks observed in Fig. 2 but excluded from Table 2 came from the solvent, SPME fiber, or GC column and should be neglected. The large peak at the retention time of 9.56 min corresponds to the benzyl phenyl ether.

TICs for \(}_}}}\) oxidized (red) and \(}_}}}^}}\) oxidized (blue) samples of benzyl phenyl ether at 1.0 V vs. Hg/HgO where W is with DMSO. The two oxidized samples were measured with five replicates

PC score plots in Fig. 3a display associations between the samples and features in the loadings. The first PC accounts for 69% of the total variance, while the second PC accounts for 31%. The ellipses indicate the 95% confidence interval around the mean for each class. The scores significantly differ between samples with and without DMSO. Loadings near − 1 or + 1 have a significant effect, while loadings close to 0 have a marginal effect. For the loadings in Fig. 3b, the \(}_}}}\) samples are the red positive peaks, whereas the blue negative loadings are for the \(}_}}}^}}\) samples. From the mass spectral loadings in Fig. 3c, many ions are observed for \(}_}}}\) samples. In contrast, in the case of \(}_}}}^}}\) samples, comparatively fewer ions are observed, indicating the involvement of •OH radicals in the depolymerization of the model. The prominent blue peaks at m/z 65 and 81 are characteristic of aromatic rings and are most likely due to unreacted benzyl phenyl ether. The scores in Fig. S2 differ for the neat, oxidized with DMSO, and oxidized without DMSO samples. Note that the free radical-mediated oxidation differs most from the neat sample and has the best reproducibility. Some base-catalyzed hydrolysis was observed for the neat samples.

a Factor analysis (FA) to visualize the distributions of the \(}_}}}\) and \(}_}}}^}}\) oxidized samples of benzyl phenyl ether. The percent relative variance is the first value in parentheses on the PC label, whereas the second is the absolute variance. b Retention time loadings to visualize the main difference between \(}_}}}\) and \(}_}}}^}}\) oxidized samples of benzyl phenyl ether. c The mass spectral loadings of \(}_}}}\) and \(}_}}}^}}\) oxidized samples of benzyl phenyl ether

The above results suggest that including the radical scavenger has caused the difference in oxidation at a high potential (1.0 V vs. Hg/HgO). Any substantial variations would not be anticipated in the presence of the radical scavenger if the reaction mechanism were only mediated by direct electrochemical oxidation of the lignin model compound, i.e., scavenging •OH using DMSO would not affect the electrochemical conversion of the model. However, suppose the mechanism is mainly radical-mediated. In that case, introducing the radical scavenger should result in differences in oxidation products because the radical scavenger interferes with the radical-driven oxidation process.

This part of the discussion pertains to the oxidized compounds at 0.8 V vs. Hg/HgO. Table 3 lists the oxidation products of benzyl phenyl ether with and without DMSO at 0.8 V vs. Hg/HgO.

After oxidation, several peaks can be observed in the \(}_}}}\) oxidized sample in Fig. S3. They are oxidized products such as gamma-oxo benzene butanoic acid, benzenemethanol, and cinnamaldehyde. It is observed for both \(}_}}}\) and \(}_}}}^}}\) oxidized samples (Table 3) at 0.8 V but not in the electrochemical oxidation at 1.0 V belongs to the same class of substituted phenolic compounds. They are probably the result of the cleavage of the C–O bond between aromatic units in the model compound.

The scores in Fig. S4(a), of \(}_}}} \) and \(}_}}}^}}\) do not overlap, indicating that the two sets of samples differ significantly. The positive and negative loadings in Figs. S4(b) and S5 contribute to the differentiation in the scores, indicating the depolymerization of the lignin model compound by free radical and direct electrochemical oxidation.

This part of the discussion pertains to the oxidized compounds at 0.6 V vs. Hg/HgO. The TICs in Figs. S6 and S7 reveal differences between the neat and oxidized samples. When comparing the two oxidized samples: \(}_}}}\) (without DMSO) and \(}_}}}^}}\) (with DMSO) in Fig. S8, similar peaks were found at retention times 7.54 and 9.12 min, which are likely attributed to 3-phenyl 2-propenal and 2,4-bis-(1,1-dimethylethyl) phenol, respectively (see Table 4). The prominent peak at 9.56 min is benzyl phenyl ether. Figure S9(b) also has the same benzyl phenyl ether in both the positive and negative loadings caused by subtle drift in retention time.

From the mass spectral loadings (shown in Fig. S10), fewer fragments were formed compared to higher potentials, indicating that the potential was too low to oxidize the model effectively. Even at low potential (0.6 V vs. Hg/HgO), the scores of two oxidized samples with and without DMSO do not overlap (see Fig. S9(a)), indicating the involvement of free hydroxyl radicals in the electrochemical depolymerization process.

Fewer compounds were observed oxidizing at 0.6 V vs. Hg/HgO compared to the higher oxidation potentials, which was expected. Knowing that conversion at low potentials may be minimal, different oxidation potentials of diverse ranges (low, medium, and high, i.e., 0.6, 0.8, and 1.0 V vs. Hg/HgO) were selected to investigate the effect of potential on electrochemical conversion. These results indicate that 0.6 V vs. Hg/HgO is likely too low to oxidize the lignin effectively.

The observed oxidized product, 1,2-benzenedicarboxylic acid dibutyl ester (molecular weight, 278), has a greater molecular weight and a longer retention time than benzyl phenyl ether (molecular weight, 184). This result can be attributed to polymerization or cross-linking reactions among the reactive intermediates or the rearrangement of functional groups during electrochemical oxidation, forming larger molecules.

Building on this study, future research will apply the depolymerization approach used for benzyl phenyl ether to lignin, a more complex biomass. The work will advance our understanding of lignin conversion technologies and contribute to developing more sustainable and cost-effective methods for producing valuable aromatic compounds from lignin.