Oxidative Depolymerization of Kraft Lignin to Aromatics Over Bimetallic V–Cu/ZrO2 Catalysts

Zirconia-supported vanadium–copper catalysts (VCux:yZr) were used for the oxidative depolymerization of softwood LignoBoost Kraft lignin (LB). Various VCux:yZr catalysts were prepared (x:y = 0:1, 1:4, 1:2, 3:4, 1:1, and 1:0) by incipient wetness impregnation, and reactions were performed in alkaline water at 150 °C under an O2 pressure of 5 bar for 10 min. 1H–13C HSQC NMR spectroscopy was used for product identification and quantification. The most promising catalyst was VCu1:2Zr, giving a total monomer yield of 9 wt% and the highest selectivity for vanillin (59%). This catalyst was characterized before and after use by N2 physisorption, XRD, TGA, SEM-EDS, and XPS. Cleavage of the main interunit linkages in LB, including the β-O-4 bonds and recalcitrant C–C bonds, was also observed. The findings of this study demonstrate the potential of the V–Cu/ZrO2 catalyst system in the production of value-added aromatics from technical lignin under relatively mild conditions. This would contribute to the more sustainable use of an underutilized side-stream in forest-based industries, provided catalyst reuse can be successfully demonstrated.


Introduction
Lignin is a complex aromatic biopolymer comprising up to 30 wt% of lignocellulosic biomass, and has significant potential as a renewable feedstock for the production of chemicals, fuels, and functional materials [1,2]. Lignin is also a major byproduct in the pulp and paper and cellulosic ethanol industries, where the bulk raw material is referred to as technical lignin [3,4]. The highly heterogeneous nature of lignin makes its valorization challenging and, as a result, Omar Y. Abdelaziz omar.abdelaziz@chemeng.lth.se an energy-efficient means of lignin conversion, providing targeted functional chemicals, including high-value chemicals such as aromatic aldehydes, aromatic acids, and alkyl carboxylic acids [14,15]. Such functional chemicals can be used in numerous applications, for example, plastics, adhesives, coatings, textiles, and pharmaceutical precursors [16,17]. Metallic catalysts can also be used to reduce energy consumption by lowering the temperature and pressure, and shortening the reaction time, while providing higher selectivity and monomer yields [18]. Lignin oxidation for vanillin production has been generally limited to lignosulfonates, but the implementation of new catalyst systems could provide novel opportunities for the oxidation of other lignin raw materials [19,20].
The use of homogeneous metal-based catalysts including metal ions, oxovanadium complexes, and salen complexes for lignin oxidation has been reported in the literature [21]. An advantage of homogeneous catalytic systems is that a variety of ligands that increase the activity and stability of the catalysts and facilitate the selective cleavage of specific lignin linkages can be used. However, the separation and reuse of homogeneous catalysts is challenging. There is thus a need for new cost-efficient catalysts to overcome these problems [19,22]. The solution may be to engineer inexpensive heterogeneous metallic catalysts. However, the challenges of metal leaching, loss of surface area, poisoning of active sites, and coking on the catalyst surface must be overcome, while ensuring high catalytic activity and selectivity [23,24].
Several heterogeneous metallic catalysts have been reported in the literature for oxidative lignin conversion, including supported noble metals such as Pd and Au, supported transition metals such as Cu and Mn, as well as non-supported mixed oxides such as perovskites containing Mn and Co [25][26][27][28][29]. Although promising results have been obtained, loss of catalytic activity upon the reuse of the catalyst, due to leaching of active metals, has been reported in the majority of these studies. Therefore, a major challenge in the field of lignin oxidation lies in finding an effective heterogeneous catalyst that is stable under the reaction conditions required.
In the present study, zirconia-supported vanadium-copper catalysts (VCux:yZr) were prepared and used for the oxidative conversion of Kraft lignin into vanillin and other high-value aromatic monomers. The active metals used in this study were chosen based on our previous study [30], in which a combination of Cu and V salts was used as a homogeneous catalyst in the oxidative conversion of Kraft lignin into aromatics. ZrO 2 was used as the catalyst support due to its relatively high surface area and redox properties, as well as its mechanical, thermal, and alkaline stability. ZrO 2 facilitates the dispersion of CuO, increases reducibility, and improves the adsorption of reactants [31]. ZrO 2 has also been used as a support material in oxidative lignin depolymerization with Co as the active phase [32].

Chemicals and Materials
The following chemicals were used in the experiments; they were of reagent grade and were used without further purification, unless otherwise stated: vanadyl acetylacetonate (VO(acac) 2

Catalyst Preparation
Catalysts were prepared by incipient wetness impregnation with a metal salt loading of 5 wt% on the support material ZrO 2 at different molar ratios. The theoretical metal loadings of the catalysts were in the range of 1-1.6 wt%. A mixture of appropriate amounts of Cu(OAc) 2 ·H 2 O and VO(acac) 2 was first dissolved in a small volume of deionized water. ZrO 2 powder was then added, and the mixture was stirred. It was important to maintain a suitable water-to-support ratio to obtain a paste with a consistency that would ensure the even distribution of the metals on the surface of the support. The paste was then dried at 110 °C overnight, and was subsequently calcined at 450 °C for 6 h. After calcination, the catalysts were fractionated to a size range of 24-18 mesh (0.71-1 mm). This range was chosen to facilitate the recovery of the catalyst from the reaction mixture.

Catalytic Oxidation
The catalytic oxidation reactions were performed in a 300 mL mechanically stirred autoclave equipped with a Parr 4848 reactor controller (Parr Instrument Company, Moline, IL, USA). In a typical experiment, 0.5 g of LB was dissolved in 50 mL of an aqueous 2 M NaOH solution. The solution was then loaded into the autoclave, together with 0.25 g of the solid catalyst (only catalyzed reactions), resulting in a catalyst-to-lignin ratio of 1:2 w/w. The reactor was purged (three times) and pressurized with molecular O 2 to 5 bar, and heated to 150 °C (± 2 °C) under stirring at 500 rpm. The heating time was approximately 25-30 min. After the desired temperature had been reached, it was maintained for a reaction time of 10 min. The reaction time, temperature, and O 2 pressure were chosen based on the optimized conditions reported in our previous study using a homogeneous V-Cu catalyst system [30]. After reaction, the autoclave was cooled in an ice bath for approximately 75 min until it reached room temperature, and was then depressurized after the final pressure had been noted. The product mixture was then transferred to a Büchner funnel using 4 mL H 2 O, where the spent catalyst was filtered off using a filter paper. The reaction mixture was stored in a freezer at -18 °C before further workup.

Workup Procedure
The product mixture (pH > 13) was thawed and acidified with aqueous HCl to pH ~ 2, to precipitate the residual lignin. The mixture was centrifuged at 3900 rpm for 20 min to separate the liquid phase from the solid fraction. The liquid phase was then transferred to a separation funnel and the solid fraction was stored in the freezer until further treatment. The liquid phase was extracted three times with EtOAc (2 × 10 mL and 1 × 30 mL). The volume of the aqueous fraction was measured before storing it at 5 °C until further analysis. The organic phase was dried over anhydrous Na 2 SO 4 , and after 15 min filtered with filter paper that was washed with 2 × 2.5 mL EtOAc. The EtOAc was removed by evaporation for 40 min in a rotary evaporator with a bump trap (IKA RV 10) at 40 °C and 210 mbar. The bump trap was then emptied, and evaporation continued at 40 °C and 200 mbar for approximately 4 h, until a dark-brown, viscous bio-oil was obtained. Figure 1 shows a block flow diagram of the workup procedure used to extract the bio-oil fraction from the reaction mixture.

Product Analysis
Samples were prepared for NMR analysis by redissolving 50 mg of the bio-oil in 550 µL of DMSO-d 6 and transferring the solution to 5 mm NMR tubes. The samples were quantitatively analyzed to determine the contents of the aromatic monomers vanillin, vanillic acid, acetovanillone, 4-hydroxybenzoic acid, and 4-hydroxybenzaldehyde. The monomer yield (wt%) relative to the initial lignin mass, and the monomer selectivity (%) relative to the total yield of these five monomers were determined. Three standard solutions were prepared, each containing gravimetrically determined amounts of monomers in DMSO-d 6 . All spectra were acquired at 25 °C using an 800 MHz Bruker Avance III instrument equipped with a TCI CryoProbe and a Sample-Jet sample changer. 1 H-13 C heteronuclear single-quantum coherence (HSQC) NMR spectra with 140 ppm spectral width in the 13 C dimension were acquired as data matrices of 1024 ( 1 H) × 256 ( 13 C) complex data points to probe the linkage region and the chemical composition of the substrate. Reaction products were identified and quantified by the acquisition of 1 H-13 C HSQC NMR spectra with 60 ppm spectral width centered in the aromatic spectral region, sampling 2048 ( 1 H) × 512 ( 13 C) complex data points with 2 accumulations per increment, using an interscan recycle delay of 1.2 s and non-uniform sampling of 50% of the data points in the indirect dimension. All NMR spectra were processed with ample zero filling and baseline corrections in in the oxidatively depolymerized samples were vanillin, vanillic acid, acetovanillone, 4-hydroxybenzoic acid, and 4-hydroxybenzaldehyde.
The optimum molar ratio of V and Cu supported on ZrO 2 was determined from screening loadings with pure Cu, V:Cu molar ratios of 1:4, 1:2, 3:4, 1:1, and pure V. The total monomer yields obtained from all the catalyzed reactions were 7.8-9.3 wt%, and were generally higher than the control reaction without an added catalyst. The lowest monomer yield from the catalyzed reactions was obtained with the VCu1:0Zr catalyst; this could be due to the formation of other product species that could not be detected.
The VCu1:2Zr catalyst afforded the highest selectivity for vanillin (59.1%), while the other catalysts had selectivities in the range of 50.6-53.9%. The vanillin selectivity was only 1.7 wt% higher than in the control reaction. However, in comparison with the control reaction, the VCu1:2Zr catalyst gave a considerably higher yield of 5.3 wt% vanillin, and a total monomer yield of 9.0 wt%, which was about 20% higher than in the control reaction. These results indicate that the VCu1:2Zr catalyst was favorable in terms of high vanillin selectivity and high monomer yield from LB. Table 1 presents the total monomer yield in this study, together with the results of other similar studies presented in the literature for comparison.
The maximum vanillin yield obtained from the oxidative depolymerization of LB reported previously by our research group using a homogeneous V-Cu catalyst system under similar reaction conditions was approximately 3.5 wt%, albeit with an initial LB concentration of 25 g/L [30]. The initial LB concentration used in the present study was lower (10 g/L), which may have influenced the yield. However, based on the encouraging results, the VCu1:2Zr catalyst was selected for the investigation of catalyst recovery and reuse in subsequent experiments.
The reusability of the VCu1:2Zr catalyst was investigated by using the spent catalyst in two subsequent oxidation reactions to observe the effect on the monomer yield and vanillin selectivity. As can be seen from Fig. 2, the vanillin yield was 4.3 wt% in both recycling experiments (VCu1:2Zr-RU and VCu1:2Zr-RU II), similar to the yield obtained in the control reaction. The total monomer yields obtained in the two recycling experiments were 7.7 wt% and 7.8 wt%, which is only 0.2-0.3 wt% higher than in the control reaction, and slightly lower than that with pure ZrO 2 . These results indicate that the catalyst might not be stable under the reaction conditions tested; this was further investigated and confirmed through catalyst characterization (see Sect. 3.2).
2D 1 H-13 C HSQC NMR was used to estimate the relative amounts of the interunit linkages present in the LB substrate and in the residual lignin that was recovered after oxidative all dimensions using Bruker Topspin 3.5 pl 7 and were integrated in the same software.

Catalyst Characterization
N 2 physisorption measurements were performed on an ASAP 2020 Micromeritics instrument using liquid nitrogen. Samples were degassed under vacuum at 90 °C for 1 h, and then at 200 °C for 4 h prior to measurements. The specific surface area, pore size, and pore volume of the various catalysts were determined using the Brunauer-Emmett-Teller and Barrett-Joyner-Halenda methods [33,34].
Powder X-ray diffraction (XRD) patterns were obtained with a Huber G670 powder diffractometer in the 2θ range 3-100° in steps of 0.005°, using CuK α1 radiation (λ = 1.54056 Å) emitted from a focusing quartz monochromator, for 10 min. The data were collected in transmission mode from a rotating flat plate sample inclined 45° relative to the primary beam.
Thermogravimetric analysis (TGA) was performed in an atmospheric air environment using a Mettler Toledo TGA/ DSC 1 Star e system. The heating rate was 10 °C/min from room temperature to 600 °C, which was then maintained for 1 h.
Scanning electron microscopy (SEM) was conducted on a high-resolution FEI Quanta FEG 250 ESEM microscope fitted with an Everhart-Thornley detector operated at 10-20 kV, with a spot size of 4, and 500-2000 times magnification. Energy-dispersive spectroscopy (EDS) was coupled with SEM for elemental mapping. SEM-EDS data were obtained when operating the equipment at 20 kV with a spot size of 3.5 and 4000 times magnification with the same microscope, coupled to an Oxford Instruments X-Mas 50 mm 2 EDS analyzer.
X-ray photoelectron spectroscopy (XPS) was carried out with a Thermo Scientific system at room temperature using monochromatic AlKα radiation (1484.6 eV). The base pressure in the analysis chamber was maintained at 2 × 10 − 7 mbar. The number of scans used was 5, 10, 15, and 15 for O, Zr, V, and Cu, respectively, for the fresh catalyst samples, and 5, 10, 20, 20, and 20 for O, Zr, V, Cu, and S, respectively, for the spent catalysts.

Catalyst Performance
All product mixtures (bio-oils) from the oxidative depolymerization reactions were subjected to 1 H-13 C HSQC NMR analysis to identify and quantify the aromatic monomers produced (Fig. 2). The five main aromatics identified between the control and the catalyzed reactions, which may be due to the relatively low metal loading of the catalysts. However, the fact that the control reaction resulted in the cleavage of lignin interunit bonds indicates that the reaction conditions were favorable for oxidative depolymerization. Interestingly, the residual lignin obtained from the reaction catalyzed by the VCu1:2Zr catalyst showed the highest cleavage of all three interunit bonds, implying that this is a promising catalyst for the oxidative depolymerization of technical lignin.
The 2D 1 H-13 C HSQC NMR spectra from the biooils obtained in the reactions employing the VCu1:0Zr, VCu1:2Zr, and VCu0:1Zr catalysts are shown in Fig. 3, together with the results from the control reaction. Each depolymerization reactions catalyzed by the VCu1:0Zr, VCu1:2Zr, VCu0:1Zr catalysts, and the control reaction. The results for the fractions of the main linkages, β-O-4, β-5, and β-β, are given in Table 2. It should be emphasized that such determinations are semiquantitative, through comparison to aromatic signal as internal reference, assuming that virtually all aromatic rings have three protons and that the gamma signals from the linkages are integrated due to their favorable NMR relaxation behavior. The fact that two protons contribute to the linkage signal in the HSQC and three protons to the aromatic signals was accounted for.
As can be seen from Table 2, both the ether bonds and the more robust C-C bonds were cleaved in the oxidative depolymerization reactions. No significant differences were seen the β-O-4 linkages and β-β linkages were low. This further indicates that the combination of V and Cu rendered the catalyst more selective to linkage cleavage.

Catalyst Properties
The catalysts prepared were characterized using various techniques to obtain information on their textural and structural properties. The VCu1:2Zr catalyst was the main focus of characterization, as this afforded the highest yield of vanillin, and was used in the recycling experiments.
The specific surface area and pore volume determined by N 2 physisorption, together with the calculated average pore size are given in Table 3 for all the catalysts prepared. The table also includes data for the spent VCu1:2Zr catalyst after one oxidation reaction with and without subsequent recalcination, and after being reused twice (Entries 5, 6, and 7, respectively).
Pristine zirconia (Entry 1) had a specific surface area of 85.8 m 2 /g, a pore volume of 0.20 cm 3 /g, and an average pore size of 90.4 Å (after suspension in deionized water and calcination). When the support was loaded with the V and Cu metal oxides, all of the three values increased (Entries 2-4 and 8-10), suggesting some agglomeration of the metal oxides with interparticle pores.
After the VCu1:2Zr catalyst had been used once, the reaction mixture removed, and the catalyst washed with water and acetone (Entry 5), all three textural properties were lower than those in the fresh catalyst (Entry 4), possibly due to heavier organic compounds blocking some of the catalyst pores. Moreover, the specific surface area of the catalyst was similar to that of the pristine ZrO 2 after recalcination (Entry 6), suggesting that some of the carbon deposits on the surface were removed during calcination, and that metal oxide clusters on the surface were also lost. After the VCu1:2Zr catalyst had been reused twice, the specific surface area and pore volume barely changed, while the pore size decreased slightly (Entry 7), indicating that the active phase of the catalyst remained largely unmodified between the two consecutive reactions.
The most prominent peaks of the ZrO 2 support (28.1° and 31.4°) corresponded well with the peaks observed for the monoclinic ZrO 2 , and these peaks were also observed in the diffraction patterns of the VCu0:1Zr, VCu1:2Zr, and VCu1:0Zr catalysts (Fig. 4), indicating that the monoclinic spectrum is plotted together with the NMR spectrum obtained from the LB substrate.
It can be seen that most of the signal from the β-O-4 linkages disappeared in the spectrum obtained from the bio-oil using the VCu1:2Zr catalyst, confirming that they were successfully cleaved during this reaction. Furthermore, the signals from the β-β linkages also decreased significantly, compared to the experiments with the other two catalysts. A smaller decrease in these two signals was observed in the control reaction, indicating that less cleavage of the β-O-4 and β-β bonds occurred in the absence of the metal catalyst. The signals from these bonds observed with the VCu1:0Zr catalyst were only slightly lower than those in the control reaction, which indicates that V alone was not very selective for the cleavage of these types of bonds in the lignin substrate. The VCu0:1Zr catalyst induced more cleavage than the VCu1:0Zr catalyst, however, the signals from both  them. However, the peak intensity was unchanged in the diffraction pattern of the VCu0:1Zr catalyst, which was richest in Cu (dark blue line), and in that of the VCu1:0Zr catalyst containing no Cu (light blue line). This observation suggests that Cu was not detected in the XRD pattern. Likewise, no peaks were observed in the diffraction patterns of the analyzed catalysts at the positions corresponding to V 2 O 5 (20.3° and 26.2°), implying that the diffraction patterns obtained arose only from the ZrO 2 support, likely due phase of ZrO 2 was retained after impregnation with the metals. However, an additional small shoulder peak appeared in the diffraction pattern at the most prominent peak (30.2°) of the tetragonal phase of ZrO 2 for all the metal-loaded catalysts. This suggests that metal impregnation induced a minor transformation of the support to the tetragonal crystalline phase of ZrO 2 .
The most prominent peaks in the diffraction pattern of monoclinic CuO (35.6° and 38.8°) are close to peaks observed for monoclinic ZrO 2 , thus making it difficult to distinguish being due to storage of the fresh catalyst for several months under ambient conditions before analysis.
Upon heating to 600 °C, the fresh and reused catalysts showed a total weight loss of ∼1.7 and ∼2.2 wt%, respectively (disregarding the adsorbed water), implying that they were reasonably thermally stable, and thus suggesting that catalyst degradation was not responsible for the reduced catalytic effect of the VCu1:2Zr catalyst observed during to the low V and Cu loadings, in combination with good metal oxide dispersion.
TGA was carried out on the fresh VCu1:2Zr catalyst and the catalyst after being reused twice (VCu1:2Zr-RU II) without preceding calcination (Fig. 5). The two profiles obtained were very different at temperatures below 120 °C, where the fresh catalyst lost more adsorbed water (~ 2 wt%) than the reused catalyst (~ 0.3 wt%). This was identified as  reused catalyst (Fig. 7e and f). This indicates that leaching of the active metals could have occurred during oxidative depolymerization.
The relative amounts of metals and sulfur in the VCu1:2Zr and VCu1:2Zr-RU II catalysts determined by SEM-EDS (bulk) and XPS (surfaces) are given in Table 4. The results show that V and Cu had indeed been leached from the surface of the catalyst by use, which could explain the lower monomer yields obtained in the recycling experiments (Sect. 3.1). In addition, sulfur had accumulated on the catalyst surface after being reused twice, probably due to the presence of sulfur-containing functionalities in pristine LB. This could also cause deactivation of the catalyst.
Since the results of TGA analysis showed that the VCu1:2Zr catalyst was stable at temperatures up to 600 °C in air, the harsh alkaline reaction medium or the combination of elevated temperature and alkaline solution could have caused leaching of the active metals from the catalyst (e.g., the formation of soluble copper complex ions and reuse (Sect. 3.1). In addition, the reused catalyst had lost ∼0.5 wt% at a temperature of about 260 °C, likely corresponding to the release of organic compounds retained in the pores of the catalyst after reaction, which is consistent with our previous observations on catalytic lignin depolymerization under oxidative conditions [27]. Clogging of pores could reduce the activity of the catalyst, demonstrating the need for recalcination during catalyst regeneration.
SEM images were obtained for the fresh VCu1:2Zr and reused VCu1:2Zr-RU II catalysts to examine their morphological features (Fig. 6). When comparing the images of the catalysts on the 20 μm scale, it was clear that the fresh catalyst (Fig. 6a) had a smoother surface than the reused catalyst (Fig. 6b), which had an uneven layer with several protrusions on the support surface.
SEM-EDS elemental mapping of the catalysts (Fig. 7) showed that Cu and V were evenly distributed on the surface of the fresh catalyst ( Fig. 7b and c), whereas the amounts of both metals were considerably decreased in the results. Vanillin, vanillic acid, acetovanillone, 4-hydroxybenzoic acid, and 4-hydroxybenzaldehyde were identified as the main monomers in the bio-oil fraction obtained. V and Cu had a combined effect on the cleavage of various lignin interunit linkages, and the VCu1:2Zr catalyst system exhibited an improvement in the total monomer yield of about 20% compared to the control reaction. Different V-Cu molar ratios were investigated, showing the superiority of VCu1:2Zr, which exhibited the highest selectivity for vanillin. However, the performance of the VCu1:2Zr catalyst decreased with reuse as a result of leaching of V and Cu species from the catalyst surface. Leaching was confirmed by SEM-EDS and XPS analysis and led to the conclusion that the catalyst system was not stable under the reaction conditions investigated. The accumulation of sulfur on the surface of the used catalyst was also observed. Thus, further optimization is required to identify the optimal operating conditions and to ensure the long-term stability of this catalyst system. vanadates). Further investigations are therefore required to elucidate the main cause of metal leaching and to devise strategies for catalyst stabilization.

Conclusions
The oxidative depolymerization of LB into high-value aromatic monomers over heterogeneous V-Cu catalysts with O 2 as oxidant has been demonstrated with promising   catalyst characterization. The NMR spectra were recorded using the 800 MHz spectrometer at the NMR Center DTU, supported by the Villum Foundation.
Funding Open access funding provided by Lund University.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.