Francesco
Brandi
,
Markus
Antonietti
and
Majd
Al-Naji
*
Max Planck Institute of Colloids and Interfaces, Colloid Chemistry Department, Am Mühlenberg 1, 14476 Potsdam, Germany. E-mail: majd.al-naji@mpikg.mpg.de
First published on 10th September 2021
Sodium lignosulfonate (LS) was valorized to low molecular weight (Mw) fractions by combining solvothermal (SF) and catalytic hydrogenolysis/hydrogenation fragmentation (SHF) in a continuous flow system. This was achieved in either alcohol/H2O (EtOH/H2O or MeOH/H2O) or H2O as a solvent and Ni on nitrogen-doped carbon as a catalyst. The tunability according to the temperature of both SF and catalytic SHF of LS has been separately investigated at 150 °C, 200 °C, and 250 °C. In SF, the minimal Mw was 2994 g mol−1 at 250 °C with a dispersity (Đ) of 5.3 using MeOH/H2O. In catalytic SHF using MeOH/H2O, extremely low Mw was found (433 mg gLS−1) with a Đ of 1.2 combined with 34 mg gLS−1. The monomer yield was improved to 42 mg gLS−1 using dual catalytic beds. These results provide direct evidence that lignin is an unstable polymer at elevated temperatures and could be efficiently deconstructed under hydrothermal conditions with and without a catalyst.
Lignin is a complex three-dimensional network of cross-linked phenylpropanoid units connected through different types of ether bonds, i.e., β–O–4, 4–O–5, γ–O–α, and carbon–carbon bonds, e.g., 5–5 and β–β linkages, viz.Fig. 1.15,30–34 Some of the big challenges in lignin valorization are its heterogeneity, recalcitrance against treatments, and the sensitivity of lignin fragments for follow-up condensation reactions.12,30 Currently, four main technologies are used to separate lignin from the other biomass components, i.e., kraft, soda, organosolv and sulfite processes.21,30,35,36 These processes rely on lignin separation from the LCB matrix via solubilization by applying harsh conditions with pH = 1–5 for sulfite and organosolv or pH = 11–13 for soda and kraft, as well as high temperatures (140–220 °C).21,37 The resulting lignin is named technical lignin and is indeed different from the primary plant product.21,37 In the DoE Report, the technical lignin is proposed as the most probable future source of lignin since it is abundant and already available from industrial processes, such as the pulping process.29 Nowadays, the 1st scenario is still the major application of technical lignin and every year, from the principal 250 Mt, about ∼70 Mt of technical lignin is separated, of which an estimated 95% is burned to produce energy.38,39 Of that amount, approximately 2 Mt per y are produced only by the sulfite process.21,38
In the sulfite pulping, debarked wood chips are treated with sulfite and bisulfite salt solution (Na is the major counter-ion to bisulfite) in acidic aqueous media (pH 1–5) at temperatures ranging from 125 to 190 °C.21,40 In this process, the ether bonds in lignin are partially hydrolyzed followed by the sulfitization step (introducing –SO3− functionalities), or oxidative coupling to form parasitic C–C bonds.21,29 Consequently, the resulting technical lignin is named sodium lignosulfonate (LS). It possesses a wide range of molecular weight (Mw) from 1000 to 100000 g mol−1, the obtained Mw values depend on process severity.21 LS contains a large amount of sulfur (3 to 8 wt%) and a reduced number of C–O bonds (Fig. 1) compared to the native lignin in wood.30 Uniquely, LS is highly water-soluble due to the presence of SO3− groups. Sulfonate endows LS with surfactant properties and it is used in current materials applications as a macromolecule (2nd scenario), e.g., as superplasticizer in cements and clays,41 and as dispersing agents for polymeric foams,42,43 dyes44 and colloidal dispersions.45,46 Additionally, LS has been used to create completely new lignin-based materials, e.g., bio-composites for fire-retardant thermal insulators,47,48 or S-doped carbonaceous materials for high-performance electrochemical devices such as batteries and supercapacitors.47,49,50 All the abovementioned applications have been reported to depend strongly on the molecular weight distribution (Mw) of the obtained LS.20,48,51
As LS has a high phenolic content, it also shows great potential for the production of bio-based aromatics via oxidative or reductive catalytic fractionation (O or RCF) and is considered the 3rd scenario.45,52–56 In the last decade, RCF, pioneered by the Sels group, has been reported as one of the methods with the highest potential for lignin valorization towards phenolic single units, mostly starting directly from wood.12,23,28,57–62 Based on this process, an approach for the valorisation of lignin toward both phenolic monomers and phenolic oligomers has been proposed to scale up such a process to industry.23,63 The simultaneous valorization of monomers and dimers in such an approach has the potential to provide a further step toward the complete usage of lignin in the biorefinery.
Despite the potential of RCF, few studies have been reported using LS solutions. To this end, Shu et al.53 reported LS hydrogenolysis using a Pt on a carbon catalyst and CrCl3 homogeneous co-catalyst at 280 °C, 3 MPa of H2, in batch systems. Additionally, the effect of a Ni and Mo bimetallic catalyst on Al2O3 was studied in ethylene glycol and supercritical ethanol at 310 °C and 26 bar of H2 in batch systems.64,65 This approach led to LS fragmentation mostly towards dimers and oligomers resulting in an optimized 88 wt% oil yield, while the minor fraction of monomers was only qualitatively studied.64,65 All these studies were performed in batch systems, which present disadvantages when compared to continuous flow systems, e.g., complex product–catalyst separation, time and energy-consumption in discontinuous steps and cost efficiency.10,12,59,66,67 Horáček et al.52 reported only the LS fragmentation in a continuous flow system at a high reaction temperature (320 °C) over a bimetallic Ni and Mo catalyst supported on Al2O3. In this study and due to over hydrogenation, guaiacol was found as the major product of LS catalytic fragmentation (yield higher than 1.8 wt%).
Herein, we investigate the LS depolymerization in a continuous flow system under mild reaction conditions using the water/alcohol solvent system. In addition, the effect of catalyst-free solvothermal fragmentation (SF) of LS to low Mw fractions and phenolic monomers was explored. Furthermore, the latter step was coupled to hydrogenolysis/hydrogenation fragmentation (SHF) in a continuous flow system using the water/alcohol solvent and 35 wt% Ni supported on NDC as a catalyst.68,69 Finally, the advantage of operating the catalytic depolymerization of LS in a continuous flow system is highlighted with respect to the batch system.
This system consists of: (A) an HPLC pump equipped with a pressure sensor (Knauer Azura P 4.1S Series), (B) a mass-flow controller to supply a specific H2 flow (Model SLA58050 from Brooks), (C) a T-piece for the gas–liquid mixer (Swagelok SS-400-30) to mix the supplied H2 with the reaction solution before reaching the pre-heating unit and the catalyst bed, (D) a two-side open heating unit equipped with a heat controller (Model # 4848 from Parr Instrument Company), and (E) a sampling unit equipped with proportional relief valves also used as a pressure regulator (Swagelok SS-RL4M8F8-EP), cf. Fig. S2 in the ESI.† To ensure efficient heating, a cylindrical aluminium adapter with three boreholes was matched inside the heating unit. One bore is used for the preheating unit for the reactant before it comes in contact with the catalyst, another hole for the thermocouple (Model # A472E5 Parr Instrument Company) and the third one for the tubular reactor (Fig. S3 in the ESI†). The reactions were performed in a stainless steel tubular reactor (inner diameter = 21 mm, outer diameter = 25 mm, length = 280 mm), viz. Fig. S4 in the ESI.† Finally, an experiment using two consecutive fixed bed reactors filled with 10 g each of 35Ni/NDC was conducted. To perform these experiments two fixed bed reactors (D1–D2) were coupled in a setup similar to the one described above (Fig. S5 in the ESI†).
The collected samples were injected into a size exclusion chromatography (SEC) system without further processing, while the product analysis procedure was established to separate the residual LS from the reaction mixture, and therefore analyze the separated mixtures by GC-MS, GC-FID, 2D HSQC NMR, FTIR, elemental analysis and EDS. The established product analysis procedure is described in detail in the ESI in section S5.†
The batch experiment was conducted in a stainless steel autoclave equipped with PTFE liners and a magnetic stirrer from Berghof (Model: BR-100). The catalytic experiments were performed at 250 °C using 30 cm3 of LS solution in a MeOH/H2O solvent mixture (cLS = 2.5 wt%) with 1.0 g of 35Ni/NDC and 7.0 MPa of external H2 pressure for 1 and 3 h. Prior to heating to 250 °C, air was purged out using 1.0 MPa of N2 three times. The analysis of the reaction products followed the same protocol that was established for the continuous flow experiments.
The SEC analysis of an aqueous LS solution (1 wt%) showed a weight average molecular weight (Mw) of 12390 g mol−1 with a dispersity (Đ) of 9.1, viz. Fig. S8 in the ESI.† This high Đ value indicates that LS presents a polydisperse distribution of the molecular weight. The 2D HSQC NMR spectra of the LS showed the absence of syringyl units (S), and only guaiacyl units (G) have been found in the typical aromatic region (δC/δH 100–150/6.2–7.5), viz. Fig. S9 in the ESI.†71,72 This indicates that the utilized LS in this work has been produced from softwood, in which the S/G molar ratio is zero. Generally, lignin rich in G units is favorable for certain applications such as additives for polymers and resins. This is due to the reactive ortho-position of the phenolic rings, which promotes radical-initiated crosslinking.73 2D HSQC NMR of LS showed the presence of sulfo-groups in the α position (α-Sα,β,γ) at δC/δH of 66–68/4.5–4.7, 79–82/4.7–5.1 and 60–62/3.9–4.0, respectively. These are typical for LS, viz. Fig. S9 in the ESI.†71,72 Moreover as expected, FTIR analysis showed the typical O–H stretching bond at around 3400 cm−1, the asymmetric and symmetric OSO stretching at 1150–1200 cm−1 and 1036 cm−1 and the typical signals of the aromatic ring vibration at around 1400 cm−1respectively, viz. Fig. S10 and Table S1 in the ESI.†74
The solvothermal fragmentation reactions (SF) were performed at 150 °C, 200 °C, and 250 °C in the absence of the catalyst. In all cases, i.e., using MeOH/H2O and EtOH/H2O mixtures and H2O as the solvent, the SEC elugrams exhibited the successive disappearance of the peak at low retention times (<10 min), corresponding to the high molecular weight fraction, cf.Fig. 2 and Fig. S11 in the ESI.† For the MeOH/H2O solvent mixture, the derived mass average molecular weight (Mw) was decreased from 12390 g mol−1 for untreated LS to 9132 g mol−1, 6093 g mol−1 and 3856 g mol−1 following the increase of reaction temperature to 150 °C, 200 °C and 250 °C, respectively (Fig. 2, Table S2 – entries 2 to 4 and Fig. S12 in the ESI†). Similarly, the EtOH/H2O solvent mixture showed a decrease of Mw with the increase of reaction temperatures, i.e., from 150 °C to 200 °C and to 250 °C, giving the lowest Mw value of 3000 g mol−1 at 250 °C, viz. Fig. 3, Table S2 – entries 5 to 7 and Fig. S12 in the ESI.† Also using pure H2O as a solvent, the Mw was found to be following the same behaviour, i.e. decreasing with the temperature increase, with the lowest Mw of 5578 g mol−1 at 250 °C (Fig. 3, Table S2 – entries 8 to 10 and Fig. S12 in the ESI†). Analogously to Mw, the number averaged molecular weight distribution (Mn) decreased with the increase of the reaction temperatures, giving the minimum at 250 °C, i.e., Mn = 586 g mol−1, 556 g mol−1 and 957 g mol−1 for MeOH/H2O, EtOH/H2O and H2O respectively, cf. Fig. S12 and Table S2 – entries 4, 7, and 10 in the ESI.†
The significant decrease in the molecular weight in all cases indicates that the cleavage of the weaker linkages of LS (mostly phenyl ether bonds) has already occurred at the lower reaction temperatures, i.e., 150 °C and 200 °C. Rather expectedly, this fragmentation is becoming most effective at the highest applied temperature, i.e., 250 °C. However, the molecular weight dispersity (Đ) was found to be higher than 5.0 at all of the investigated temperatures using all solvent systems (Table S2 – entries 2 to 10 – in the ESI†). This indicates that fragmentation occurred only partially, resulting in a heterogeneous mixture with different sizes of oligomers, mostly tri-, tetra- and pentamers. An additional explanation for such a heterogeneous molar mass distribution could be attributed to in situ re-polymerization (side reactions) of the cleaved intermediates or the occurrence of radical polymerization.79 Nonetheless, the water/alcohol mixtures exhibited lower molar mass distributions over all the investigated temperatures. This result is attributed to the better stabilization of the intermediates against re-condensation by water/alcohols mixtures. Based on this, MeOH/H2O and EtOH/H2O samples were selected to qualitatively analyse the monomers using GC-MS, viz. Fig. S13 in the ESI.† From this chromatogram, a wide range of monomers (16 compounds) has been identified and reported in Table S3 in the ESI.† Moreover, the GC-MS chromatogram exhibited peaks in the high retention time region (>20 min), which can be attributed to dimers. However, these dimers were not identified and recognized by the compound database of the GC device. Among the identified compounds, 4-propyl guaiacol (G1), 4-ethyl guaiacol (G2), homovanillyl alcohol (G3), dihydroconyferyl alcohol (G4), guaiacol (G5) creosol (G6), eugenol (G7) and isoeugenol (G8) were quantified using GC-FID, and cumulative monomer yields of 8.7 mg gLS−1 and 2.9 mg gLS−1 were calculated for MeOH/H2O and EtOH/H2O, respectively (Table 2 and Fig. S14 in the ESI†). These higher cumulative monomer yields are attributed to better monomer stabilization by the MeOH/H2O solvent mixtures. It is noteworthy to be mentioned here that SF in continuous flow systems produced partially depolymerized LS. In this simple approach, the molecular weight distribution has been found to depend directly on the reaction temperature. This finding can be attributed to the thermal elimination of single phenol units from LS. This is a typical behaviour of a thermodynamically unstable polymer that undergoes reversible addition–fragmentation reactions. Moreover, the derived dispersity (Đ = 6.5 at 250 °C using MeOH/H2O) indicates that this equilibrium is perturbed by side reactions such as condensation of the cleaved fragments. This phenomenon provides clear evidence of the presence of a ceiling temperature, above which the polymer is spontaneously depolymerized into smaller units. Interestingly, the presence of a ceiling temperature is a general phenomenon in polymer science but was mostly not reported in the community of lignin biorefinery.80
Sample | c LS/wt% | T/°C | G1/mg gLS−1 | G2/mg gLS−1 | G3/mg gLS−1 | G4/mg gLS−1 | G5/mg gLS−1 | G6/mg gLS−1 | G7/mg gLS−1 | G8/mg gLS−1 | H1/mg gLS−1 | Y CM/mg gLS−1 | STYCM/mg h−1 gNi−1 |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
n.a. = not applicable. Note: The reaction conditions of these experiments can be found in the captions of Fig. 2, 3, 4, 5, 6 and S11 at ESI.† | |||||||||||||
SF-MeOH/H2O | 1 | 250 | 1.2 | 0.48 | 1.5 | 0.94 | 0.80 | 0.04 | 3.4 | 0.35 | n.a. | 8.7 | n.a. |
SF-EtOH/H2O | 1 | 250 | 0.89 | 0.01 | n.a. | 0.94 | 0.93 | 0.05 | n.a. | 0.10. | n.a. | 2.9 | n.a. |
SHF-MeOH/H2O | 1 | 150 | 6.7 | 0.22 | n.a. | 0.72 | 0.80 | 0.46 | n.a. | n.a. | 0.19 | 9.1 | 1.6 |
1 | 200 | 6.6 | 0.35 | 2.9 | 5.1 | 0.89 | 0.04 | n.a. | n.a. | 0.37 | 16 | 2.7 | |
1 | 250 | 16 | 2.7 | 3.1 | 8.2 | 2.2 | 0.22 | n.a. | n.a. | 0.83 | 34 | 5.8 | |
SHF-EtOH/H2O | 1 | 150 | n.a. | n.a. | n.a. | 0.72 | 1.0 | n.a. | n.a. | n.a. | 0.04 | 1.7 | 0.29 |
1 | 200 | 2.4 | 0.12 | 0.6 | 2.6 | 0.61 | 0.01 | n.a. | n.a. | 0.04 | 6.3 | 1.1 | |
1 | 250 | 17 | 0.72 | 2.4 | 9.7 | 0.78 | 0.10 | n.a. | n.a. | 0.53 | 31 | 5.3 | |
Flow 2.5 wt% | 2.5 | 250 | 16 | 4.9 | 0.71 | 5.6 | 3.7 | 1.6 | n.a. | n.a. | 0.1 | 33 | 14 |
Dual column | 2.5 | 250 | 20 | 5.7 | 0.94 | 9.6 | 3.4 | 1.6 | n.a. | n.a. | 0.2 | 42 | 9.0 |
Batch 1h | 2.5 | 250 | 12 | 0.70 | 2.0 | 1.9 | 1.3 | 0.17 | 0.53 | 0.48 | n.a. | 19 | n.a. |
In this part, the influence of coupling a catalytic hydrogenolysis/hydrogenation step with the abovementioned solvothermal treatment on LS depolymerization was investigated. Herein, the presence of a redox catalyst (35Ni/NDC) facilitates LS depolymerization through hydrogenolysis of ether bonds of LS fragments, mostly β–O–4, and 4–O–5 (viz. Fig. S18 in the ESI†). Simultaneously, the unsaturated and unstable fragments formed solvothermally are catalytically hydrogenated, preventing their recondensation (viz. Fig. S18 in the ESI†).59,81 For this purpose, a set of experiments was performed at three different reaction temperatures, i.e., 150 °C, 200 °C and 250 °C, using water/alcohol with a weight ratio of 1:1, i.e., EtOH/H2O and MeOH/H2O, as well as using only H2O as a solvent. These experiments are noted as follows: SHF-MeOH/H2O, SHF-EtOH/H2O, and SHF-H2O.
The dependence of the molecular weight (from SEC) on the reaction temperature in these experiments is shown in Fig. 4 and Fig. S19.† By increasing the reaction temperature from 150 °C to 200 °C and to 250 °C, the SEC elugrams showed a disappearance of the peak at low retention time (RT < 9 min), which corresponds to the high molecular weight fraction, in all solvent systems. Similarly, Mw and Mn exhibited significant decay when the temperature was increased from 150 °C to 200 °C and to 250 °C for SHF-MeOH/H2O, SHF-MeOH/H2O and SHF-H2O, viz.Fig. 3, Fig. S19, and S20 in the ESI.† The Mw was found to have decreased from 12390 g mol−1 (before reaction) to 433 g mol−1, 502 g mol−1 and 1393 g mol−1 at 250 °C for SHF-MeOH/H2O, SHF-EtOH/H2O, and SHF-H2O, respectively (Fig. 4 and Table S2 – entries 13 and 16 in the ESI†). All these values are smaller than those for SF (Table S2 – entries 4, 7 and 10 – in the ESI†).
Nevertheless, using pure H2O as a solvent resulted in a higher molecular weight at all investigated temperatures compared to its water/alcohol counterparts. This observation is in agreement with the SF experiments, wherein using pure water as a solvent lead to the formation of oligomers and dimers.77,82 Moreover, Đ has substantially decreased from 6.5 to 1.2 for SHF-MEOH/H2O and SHF-EtOH/H2O, and to 3.8 for SHF-H2O i.e., from highly dispersed to relatively monodisperse (Table S2 – entries 4 and 13 in the ESI†). This decrease of both Đ and Mw in SHF clearly indicates a controlled and efficient LS depolymerization in the presence of the catalyst with respect to SF. These observations are in good agreement with the proposed stabilization of the formed monomers via hydrogenolysis/hydrogenation due to the presence of both H2 and the redox catalyst. Thus, this step is required to prevent recondensation of the fragmented fraction.59,81
For detailed insights into the structure of the sample taken from SHF-MeOH/H2O at 250 °C, characterization of the freeze dried solid residue (denoted as SHF-LS) was performed via elemental analysis (EA) including analysis of Na, Ca, and Mg via ICP-OES, FTIR and 2D HSQC NMR. The EA showed a decrease of the S content from 7.0 wt% to 5.0 wt%, while the O content decreased slightly from 36 wt% to 34 wt%, cf.Table 1. Differently, the Na, Ca and Mg amounts in the freeze-dried solid residue remained constant before and after the experiment. The difference in the S content is attributed to the elimination of sodium sulfate and sodium sulfide as a result of sulfonate hydrolysis and reduction. We also find a slightly higher acidity of the product solution (pH = 5.5) when compared to the reactant solution (pH = 6.3). It is important to note that at this pH, the water soluble phenolates are not formed without subsequent loss of aromatics in the water phase. The 2D HSQC NMR spectra for SHF-LS showed the disappearance of the sulfo-group signals (α-Sα,β,γ at δC/δH 66–68/4.5–4.7, 79–82/4.7–5.1 and 60–62/3.9–4.0), which is an indication of the sulfonate group removal from LS (Fig. S21 in the ESI†). In addition, low intensity of Ni2S3 with respect to Ni0 was found in the XRD pattern of the spent catalyst indicating a small amount of S2− reacting on the catalyst surface (Fig. S22 in the ESI†), which however does not influence the catalyst performance, as proven by SEC and monomers yield. In addition, FTIR analysis of SHF samples showed a decrease of the OSO stretching bands at 1036 cm−1 and at around 1200 cm−1, when compared to the original FTIR spectrum of LS, viz. Fig. S9 and Table S1 in the ESI.† Moreover, FTIR showed that the C–H stretching modes at 2954 and 2937 cm−1 exhibited a higher peak intensity in SHF samples than in the original LS. This suggests a higher concentration of alkyl chains in the SHF sample due to the LS hydrogenolysis/hydrogenation.
In order to evaluate the depolymerization efficiency at different temperatures, the SHF product has been qualitatively and quantitatively analyzed using GC-MS and GC-FID, respectively. By GC-MS it was possible to identify a wide range of monomers (18 compounds), as reported in Table S5 in the ESI.† Among the identified compounds, 4-propyl guaiacol (G1), 4-ethyl guaiacol (G2), homovanillyl alcohol (G3), dihydroconyferyl alcohol (G4), guaiacol (G5), creosol (G6), and 4-propyl phenol (H1) were the predominant ones and have been quantified from GC-FID chromatograms (Table 2, and Fig. S13 and S22 in the ESI†). Interestingly, the compounds with unsaturated alkene tails, i.e., eugenol (G7) and isoeugenol (G8), viz. Fig. S14 in the ESI,† were found only in traces, indicating the efficient hydrogenation of the double bonds at the alkene tails using 35Ni/NDC. At a reaction temperature of 150 °C, the cumulative monomers yield (YCM) was found to be 9.1 mg gLS−1, 1.7 mg gLS−1 and 0.1 mg gLS−1 with MeOH/H2O, EtOH/H2O and H2O as the solvent, respectively (Fig. 5, Table 2, and Fig. S24 and S25 in the ESI†). These compounds have boiling points between 200 °C and 300 °C, which allows them to be separated from the product mixture for further chemo-catalytic upgrading.28,83 Alternatively, separation based on solvent extraction can be performed.28,63,83,84 Once separated, the monomer can be valorised toward phenols and alkenes via hydroprocessing and dealkylation, while the oligomers can be used for materials applications.23,63
Increasing the temperature from 150 °C to 200 °C and 250 °C corresponded to an increase in the YCM from 9.1 mg gLS−1 to 16 mg gLS−1 and to 34 mg gLS−1 for SHF-MeOH/H2O mixtures and from 1.7 mg gLS−1 to 6.3 mg gLS−1 and to 31 mg gLS−1 for SHF-EtOH/H2O, which correspond to the space–time cumulative monomer yield (STYCM) of 1.6, 2.7, and 5.8 mg h−1 gNi−1 for MeOH/H2O and 0.29, 1.1, and 5.3 mg h−1 gNi−1 for EtOH/H2O. In contrast, SHF-H2O exhibited a minimal YCM increase from 0.1 mg gLS−1 to 0.7 mg gLS−1 and to 1.7 mg gLS−1 at 150 °C, 200 °C, and 250 °C, respectively, viz. Fig. S25 in the ESI.† This lower YCM using pure H2O instead of the water/alcohol mixture is in agreement with the previous study conducted with lignin77,82 In this context, low YCM is attributed to the poor solubility of the formed monomers in water, which lead to their adsorption on the catalyst surface, which subsequently led to low YCM with respect to the water/alcohol solvent mixture.82 Moreover, the hydrogenation of phenols in pure H2O can lead to over-hydrogenation toward hexanols.85 This is confirmed by the presence of only four monomers among the previously identified compounds, i.e.G2, G3G5, and H, viz. Fig. S25 in the ESI.† The low YCM combined with the high Mw of the SHF-H2O experiment indicates the necessity of using the water/alcohol mixture. In all the cases using alcohol/water mixtures, except EtOH/H2O at 150 °C, the monomers were similarly distributed with 4-propyl guaiacol (G1) and dihydroconyferyl alcohol (G4) as the major products, viz.Fig. 5. However, the dominant distribution of phenolic monomers with reduced tail, such as G1–G6 and H1, indicates the effective hydrogenolysis/hydrogenation of LS fragments in the SHF process by the 35Ni/NDC catalyst. Nevertheless, it is important to underline that the selected compounds do not represent the entire amount of phenolic monomers, and for this reason, the cumulative monomer yield remains under-estimated. The values of monomer yield in both water/alcohol solvent systems at low temperature, i.e., at 150 °C and 200 °C, are in-line with the maximum theoretical yield (∼10 mg gLS−1) that Rinaldi et al. reported only for the β–O–4 linkages of LS, indicating the cleavage of these linkage types.30 However, at 250 °C, the cumulative monomer yield was found to be three times higher than the maximum theoretical yield, indicating the cleavage of other linkages in addition to β–O–4. These observations indicate that at 150 °C and 200 °C, depolymerization occurs mostly via hydrogenolysis of ether bonds. Whereas at 250 °C, in addition to β–O–4, a wide number of linkages, e.g., 4–O–5 and 5–5 bonds, are cleaved solvothermally followed by the stabilization of the unsaturated bonds, viz in Fig. S26 in the ESI.† This process is within the typical lignin fragmentation at high temperatures (250–400 °C).79,86 The solvothermal cleavage of bonds produces unstable intermediates, such as radicals and double bonds, which are hydrogenated using 35Ni/NDC, viz. Fig. S26 in the ESI.† This proposed scheme would confirm the protective action against recondensation of the 35Ni/NDC catalyst. For the catalytic hydrogenolysis of β–O–4 and 4–Ov5 ether bonds, we proposed that these steps are following a classical hydrogenation mechanism. This mechanism based on dissociative H2 adsorption on Ni followed by LS adsorption and bond cleavage, is analogous to those of studies in the literature conducted with model compounds, viz Fig. S27.†87–89
In the studied system, the SHF-MeOH/H2O solvent mixture showed a higher monomer yield over all temperature ranges than SHF-EtOH/H2O. The higher activity of MeOH/H2O confirms the results of the LS experiments and is attributed to the high polarity of this mixture that allows better solubility of LS, which led to high diffusion of LS through the 35Ni/NDC catalyst and higher LS hydrogenolysis/hydrogenation rates. These findings are in strong agreement with other studies conducted with MeOH on catalytic lignin depolymerization reported by the Sels group.76,90 This demonstrated the crucial role of a polar solvent in lignin solubilization which could lead to improved process efficiency. Therefore, MeOH/H2O has been selected for further investigations as the solvent mixture, due to its higher YCM when compared to EtOH/H2O or pure H2O.
In terms of weight hourly space velocity (WHSV), increasing the concentration from 1 to 2.5 wt% corresponded to an increase from 0.17 to 0.43 gLS h−1 gNi−1. Herein, Mw and Mn of 430 and 391 g mol−1 with Đ of 1.1, respectively, were found (see Table 3). These values are comparable to the one obtained from the experiment conducted using 1.0 wt% LS solution (Table 3). Also, the cumulative monomer yield was found to be similar in both cases. These results show no significant influence of the higher concentration (from 1.0 wt% to 2.5 wt%) of LS on the catalyst performance. Accordingly, increasing the concentration from 1 to 2.5 wt% corresponded to an increase of STYCM from 5.8 to 14 mg h−1 gNi−1. Because of the higher weight productivity, the 2.5 wt% concentration was selected to investigate the catalyst stability over time on stream (TOS), viz. Fig. S28 in the ESI.† Exposing the 35Ni/NDC catalyst to the reaction conditions for 9 h, did not affect the YCM, which was found to be constant (ca. 30 mg gLS−1). However, the N2 physisorption of 35Ni/NDC-TOS exhibited a loss in specific surface area from 578 to 272 m2 g−1 (Fig. S29 in the ESI†). This decrease in the surface area indicates product deposition onto the catalysts, due to the affinity between aromatics and porous carbon, which is in agreement with a similar study conducted by our group for vanillin hydrogenation.68 The XRD diffraction patterns of the spent catalyst after 9 h of TOS (35Ni/NDC-TOS) showed similar peaks to those of the spent catalyst after 1 h (Fig. S22 in the ESI†), with mainly Ni0 reflections and the presence of Ni(OH)2 and Ni3S2 phases. This remarkable resistance of Ni0 for such a long time under harsh conditions (250 °C and the presence of S), is attributed to the establishment of a Mott–Schottky stabilization effect between the Ni0 and the N of the carbon support.68,91,92
Sample | M w /g mol−1 | M n /g mol−1 | Đ /— | Y CM /mg gLS−1 | STYCMb/mg h−1gNi−1 |
---|---|---|---|---|---|
n.a. = not applicable.a Calculated by SEC.b Calculated from GC-FID results.c Not determined as the Mw is high with respect to 1 h. | |||||
LS | 12390 | 1364 | 9.1 | n.a. | n.a |
SHF-MeOH/H2O | 433 | 360 | 1.2 | 34 | 5.8 |
Flow 2.5 wt% | 430 | 391 | 1.1 | 33 | 14 |
Dual column | 393 | 337 | 1.1 | 42 | 9.0 |
Batch 1 h | 1615 | 491 | 3.3 | 19 | n.a. |
Batch 3 h | 6942 | 684 | 10 | n.d.c | n.a |
One of the advantages of operating in continuous flow systems is the possibility to couple multiple reactors together.67 Herein, two identical packed reactors with 35Ni/NDC have been coupled in series, in order to prolong the contact time between the LS solution and the catalyst and consequently optimize the monomer yield (Fig. 6 and Fig. S5 in the ESI†). Moreover, coupling two flow reactors in series aims to mimic an industrial scale up of such process. Therein, the concentrated 2.5 wt% LS solution has been continuously fed through two coupled reactors packed with 35Ni/NDC. Accordingly, the double amount of the catalyst is used, resulting in two-times lower WHSV (from 0.43 to 0.21 gLS h−1 gNi−1). The molecular weight distributions in this experiment showed a slight decrease when compared with the sole reactor experiment using 2.5 wt% (Table 3). In contrast, the cumulative monomer yield increased from 33 to 42 mg gLS−1 when compared to the single column experiment (Fig. 6 and Table 2), i.e. prolonging the residence time resulted in an improved monomer yield. In addition, this monomer yield (42 mg gLS−1) is four times higher than the maximum theoretical monomer yield (∼10 mg gLS−1) reported by Rinaldi et al. based on the cleavage of β–O–4 linkages in LS.30 However, the benefits of using two columns in terms of YCM are combined with STY decreasing from 14 to 9.0 mg h−1 gNi−1, viz.Table 3. This indicates that an excess amount of the catalyst is used in the dual columns experiment, with a non-optimized WHSV.
Extending the residence time by coupling two reactors yielded an even higher cumulative monomer yield of 42 mg gLS−1 with respect to the experiment in a single tubular reactor. However, the decreased STY of this approach (9.0 mg h−1 gNi−1) pointed out the need for higher weight productivity in view of further possible improvements.
The obtained monomer yield is relatively low, but also expected and well-known for technical lignin.30 This is due to the high sulfonate (and counter-ion) content, as well as the low amount of C–O bonds of LS when compared to native lignin.30 This is the intrinsic limitation of using technical lignin as the starting feedstock for lignin depolymerization. Nevertheless, in our contribution, we reported on a high cumulative monomer yield when it is compared to other work presented in the literature using similar lignin.52,53
In addition to the monomers, the SHF process simultaneously produces lignin oligomers, which can be also applied in a wide range of applications. Based on this, we believe that our approach has the potential to use both the monomer and the oligomers fractions, moreover the weight of the product solutions can be tuned according to the temperature.
We also have to underline that all these data were obtained using an industrial LS waste, i.e., containing 7.0 wt% sulfur and 10 wt% sodium, rather than impurity-free lignin. Certainly, other lignin types could allow high targeted monomer yields. Nevertheless, giving value to industrial waste will increase the sustainability of already established biorefinery processes and accelerate the transition toward a sustainable circular bio-economy.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1gc01714d |
This journal is © The Royal Society of Chemistry 2021 |