Kraft Lignin Electro-Oxidation under Ambient Temperature and Pressure

Abstract

Lignin is the largest natural source of aromatic chemicals. Due to its complex polymeric structure, Kraft lignin is under-utilized and usually combusted for heat generation, thus resulting in CO₂ emissions in the Kraft process. To valorize lignin with renewable electricity and to convert it into value-added aromatic chemicals, efficient electrochemical methods need to be discovered, based not only on its apparent yield of building block chemicals but also on its energy efficiency. In this study, the electro-oxidative performance of six different metals was systematically evaluated. The results showed that the synthesized Ni-based catalyst can increase the vanillin and vanillic acid apparent yield by 50–60% compared to when Ni-based catalyst is absent. We also found that the oxygen evolution reaction (OER) is more than a competing reaction since the presence of oxygen synergistically aids oxidation of the lignin to increase aromatic chemical production by 63% compared to the sum of vanillin generation from both voltage-only and oxygen-only scenarios. With the novel proposed notion of charge efficiency, we showed that utilizing a thinner layer of Ni catalyst balances the OER and the oxidative reaction of lignin, thus improving the charge efficiency for vanillin by 22%

1. Introduction

Due to its ability to reduce industrial dependence on fossil resources, especially petroleum, the production of organic chemicals from biomass via green processes with net-zero emissions has drawn great attention from scientific research communities as well as public policymakers [1,2]. Lignin, one of the most abundant biomass-derived wastes from the pulp and paper industry, is rich in aromatics, making it a good natural source of value-added chemicals [3]. Lignin conversion has been carried out with the traditional approach, where the lignin-containing black liquor is incinerated for power generation in the Kraft process [4], resulting in significant CO₂ emissions [5]. The main objective of lignin valorization, therefore, is to find an alternative to the traditional incineration methods [6] and to lower carbon emissions and energy costs. Thus, the sustainability aspects, such as energy requirement, of the method should be considered the key factors for lignin conversion design [7]. Various techniques have been studied to upgrade lignin, including pyrolysis, enzymatic conversion, catalytic oxidation, and catalytic reduction [8,9,10].

Among all the possible methods, electrocatalysis has demonstrated great potential for greener lignin valorization because it uses less toxic and harmful reagents such as Nitrobenzene [11], which can contribute to lowering costs and improving safety. Furthermore, Canada’s electricity grid is based on renewable sources, e.g., hydro-/solar-/wind- electricity, which enables production of upgraded biomass with a low carbon intensity.

As shown in Figure 1, lignin is composed of three main phenolic units, namely, Syringyl (S), p-Hydroxyphenyl (H), and Guaiacyl (G) units [12]. These units are connected via different chemical linkages such as 5-O-4, β-1, and 5–5 bonds and the most abundant β-O-4 bond which is targeted in this work, as shown in the illustration [13,14]. Valorizing lignin requires depolymerization via the cleavage of these linkages, so the phenolic units would be further converted into the final products such as vanillin, vanillic acid, and syringaldehyde [15].

A number of studies on the topic of electrochemical valorization of lignin have been published using different catalysts or different catalyst synthesis methods with various experimental setups. Nickel-based material has been the most tested catalyst due to its outstanding activity towards biomass oxidation as well as its stability against corrosion in an alkaline environment [18]. Movil et al. deposited Ni and Co nanoparticles onto a Pt disk electrode using the water–ethanol dispersion method and tested for lignin reaction in a typical three-electrode setup, confirming lignin electro-oxidative depolymerization along with facile hydrogen production [19]. Subsequently, Ni-Co catalysts were prepared by electrodeposition on a 1.5 cm $\times$ 1.5 cm titanium foil in a typical three-electrode setup for softwood alkali lignin oxidation, yielding vanillin, 3-methylbenzaldehyde, and acetovanillone [20]. In another three-electrode setup prepared by Yan et al. from Sun’s group, nickel foam was tested for organosolv lignin oxidation, resulting in a combined 17.5% yield of vanillin and syringaldehyde [21]. Several commercially available Ni, Co, Fe, and Ti-based alloys as well as black-liquor-activated Ni foam were tested for Kraft lignin degradation [18]. For higher overall product yield, high-temperature electrolysis was performed by Zirbes et al. from Waldvogel’s group in a sealed three-electrode cell with nickel foam and nickel sheet as the working electrode [11]. Besides Ni-based anodes, PbO₂-based catalysts were also tested. Li’s group reported BHT (butylated hydroxytoluene) production, which is used as an antioxidant in the food and cosmetic industries, from black liquor lignin using Pb/PbO₂ anode [22]. PbO₂-based catalysts were also tested by Li’s group with Aspen lignin for 4-methylanisole production [23] and they discussed different cathode pairing for bamboo lignin [24], rice straw lignin [25], corn stover lignin from paper-making black liquid [26], and cornstalk lignin [27] in a single-cell reactor. To develop a more uniform deposition of PbO₂ nanoparticles and avoid surface defects, a deposition method with multi-wall nanotube on a rotating disk electrode (RDE) was developed by Bateni et al. and tested for lignin depolymerization, yielding vanillin as well as BHT after 24 h and 48 h reaction [28].

Although different papers report the production of various aromatic molecules, one specific product, vanillin, is commonly seen in the reported value-added products and often possesses the highest yield compared to other products. However, due to dissimilar catalyst preparation methods, catalyst geometric sizes, reactor sizes, reaction conditions such as temperature and pressure, reaction time, and lignin types, it is not straightforward to compare the characteristics of the catalysts themselves. In addition, given the complex structure of lignin, under ambient temperature and pressure, the absolute mass yield of the targeted products was generally on the order of milligrams per gram of lignin used, making it more difficult to draw a clear conclusion of the catalyst performance solely based on the absolute yield in a one-pot batch conversion mode. There are two gaps in the electrochemical lignin reaction literature: (1) the analysis of intrinsic electrocatalytic performance and (2) an overall energy requirement comparison by charge transfer efficiency. Key parameters analogous to the surface area–normalized current density and Faradaic efficiency in the established field of water electrolysis [29] should be included.

Moreover, to the best of our knowledge, most studies used undivided cells for lignin electrolysis which therefore have a mixture of products from both oxidation and reduction processes, making it difficult to distinguish the electrocatalytic effects at each half-cell. The mixing of the electrolyte could also lead to the crossing over of the oxidative products to the cathode, which could potentially lower the accuracy of the test results for oxidation reaction when quantifying the oxidation products. Moreover, due to the complex and often unknown degree of polymerization in lignin, the exact reaction mechanism is not known, which prevents accurate monitoring of the charge transfer for each product and calculation of the Faradic efficiency. In this paper, we utilized a divided H-type cell reactor to focus only on the lignin oxidation process under room temperature and ambient pressure with benchmark metal catalysts prepared through the physical vapor deposition method. Instead of optimizing the product absolute mass yield by recycling the reactant flow or having a longer reaction time, we propose a normalization metric for catalyst comparison of their energy efficiency. In addition, we demonstrate that the oxygen evolution reaction (OER) is greater than a competing reaction since the oxygen generated synergistically aids oxidation of the lignin to increase aromatic chemical production.

2. Materials and Methods

2.1. Materials

The Kraft lignin was obtained from Hinton Pulp, a division of West Fraser Mills. Vanillin (ReagentPlus^®, 99%), vanillic acid (97%), guaiacol ($\ge$98%), syringaldehyde ($\ge$98%), and acetovanillone standards were purchased from Sigma-Aldrich Canada (Oakville, ON, Canada). The organic solvent used for extraction, HPLC-grade dichloromethane ($\ge$99%, containing 40–150 ppm amylene as stabilizer), was also purchased from Sigma-Aldrich Canada (Oakville, Ontario, Canada).

2.2. Electrochemical Setup

The electrolysis was conducted in a divided 3-electrode H-type electrolytic cell (from Corrtest Instruments (Wuhan, China)) with a Fumasep anion exchange membrane (Fumatech, Fuel Cell Store, Bryan, TX, USA) separating the cathode and anode chambers, as shown in Figure S1. An Ag/AgCl electrode (3 M KCl with saturated AgCl, Sigma-Aldrich (Oakville, ON, Canada)) was used as the reference electrode. The reference electrode was frequently cleaned and calibrated to avoid organic component fouling. On the cathode side where the catholyte is also 1 M KOH, a 2 cm $\times$ 2 cm platinum foil (from Corrtest Instruments (Wuhan, China)) was used as the counter electrode for all the electrolysis experiments. All the electrochemical experiments were performed with a Metrohm Autolab PGSTAT204 potentiostat.

2.3. Catalyst Preparation

In all sample preparations, YLS-30T^® carbon papers from Futiantian Technology Co., Ltd. (Xiamen, China) were utilized as the catalyst support. Copper samples with 200 nm thickness were prepared by using the physical vapor deposition (PVD) method, using the NexDep^® platform from Angstrom Engineering. An electron beam was utilized to evaporate copper pellets (99.9999% purity provided by Angstrom Engineering (Cambridge, ON, Canada)) at 6 × 10⁻⁶ torr vacuum to sputter a 100 nm layer; then, after a 10 min break, the next 100 nm layer was deposited. The deposition rate was set at 1.0 Ås⁻¹ in both steps. The same procedure was developed to sputter 200 nm samples of titanium, chromium, aluminum, and nickel by using the BJD^® platform from Angstrom Engineering. The same purity of metal targets, the same base pressure, and the same deposition rates were respected. Silver samples were also prepared by e-beam evaporation inside BJD^®, but with a 2.0 Ås⁻¹ deposition rate, as mandated by the provider’s guidelines. Regarding the 100 nm Ni sample, the same procedure was followed with no need of the 10 min break between the sputtering steps.

2.4. Oxidative Electrolysis of Kraft Lignin

In the H-cell described above, 0.4 g of Kraft lignin was dissolved into 40 mL of 1 M KOH, making a final concentration of 10 g/L. The anode chamber was stirred by a magnetic stir bar at a stirring rate of 400 rpm with 1 cm $\times$ 1 cm sized catalysts immersed in the lignin solution.

For product identification, 25 mL of the lignin solution was taken and mixed with an equal volume of 1 M H₂SO₄ as shown in Figure S2. After centrifuging, 20 mL of the clear solution was transferred into a separation funnel and extracted with 20 mL HPLC-grade dichloromethane thrice (3 × 20 mL) as shown in Figure S3. The organic phase from the organic solvent liquid–liquid extraction was then moved to a BUCHI rotary evaporation system with water bath at 47 $℃$ and dried to a dark-brown solid. The solid was then dissolved in 3 mL of HPLC-grade dichloromethane before injection into GC-MS.

Chronoamperometry test were performed at a potential of 1 V vs. Ag/AgCl for 1500 s. After reaction, 2 mL reaction solution samples were mixed with an equal volume of 1 M H₂SO₄ to fully precipitate the unreacted lignin. The mixture was then filtered with a 0.2 $\mathsf{\mu }$m PTFE syringe filter (Sigma-Aldrich) and then analyzed by using HPLC-UV for product quantification.

2.5. Characterization

To quantify the targeted products, HPLC tests were performed on a Thermo Ultimate 3000 HPLC system using a Thermo Acclaim 300 C-18 column with a UV detector. The mobile phase was prepared according to previous studies [30] where two solutions, A and B, were used. Solution A is composed of water, methanol, and acetic acid at a ratio of 89:10:1, and solution B is prepared by mixing water, methanol, and acetic acid at 9:90:1. The HPLC method runs for 60 min. It starts with 100% A and a linear gradient of B is run to increase B from 0 to 40% in the first 30 min. After B reaches 40%, the mobile phase mixture is held for another 30 min. The results are collected with a UV detector at 280 nm for phenolic compound characterization. The GC-MS was performed on an Agilent 6890 N with an Rxi-5 ms column (length: 30 m, inner diameter: 0.25 mm, film: 0.25 $\mathsf{\mu }$m) with an Agilent 5975B MS detector. The GC-MS test runs for 26 min with an initial temperature of 40 $℃$ and the temperature increases at a rate of 15 $℃$/min, which is maintained until the end of the analysis. Helium was used as the carrier gas at a flowrate of 1.5 mL/min. The ion masses were recorded in the range of 40 to 300 (m/z) in the scan mode. The detected compounds were identified according to the NIST library database. To increase the apparent yield for a more concentrated product profile analysis, the GC-MS sample was prepared using 1 cm × 1 cm Ni foam electrode under 1 V and reacted for 10 h.

3. Results and Discussion

3.1. Preliminary Electrochemical Tests

Firstly, cyclic voltammetry (CV) was performed on all the sputtered materials at a scan rate of 5 mV/s from −0.5 V to 1 V vs. Ag/AgCl electrode (saturated AgCl in 3 M KCl) as depicted in Figure 2a. Ag, Ni, and Cu all showed a reversible pattern in the CV window from −0.5 V to 1 V. Ag showed two oxidation peaks, one at 0.30 V and the other at 0.68 V, and two reduction peaks at 0.30 V and −0.06 V, respectively, where the formation and reduction of Ag₂O and AgO occurred [31]. The oxidation peak for Ni was seen at 0.40 V and the reduction peak was seen at 0.25 V. For Cu, a small oxidation peak was seen around 0.33 V and a reduction peak was seen at 0.56 V. In contrast, Cr and Ti only showed their oxidation peak in the first CV cycle and no oxidation or reduction peak was seen in the later CV tests. This indicates that the electrochemical oxidation and reduction process on the surface of Cr and Ti under the prescribed experimental condition is not reversible in the tested potential window. This also reveals the ‘valve properties’ of Ti, where surface TiO₂ shows poor conductivity and protects the base layer of Ti [25]. The surface of the catalysts remains in the oxidized form of CrO₂ [32] and TiO₂ [25]. For Al, the surface of the catalyst is immediately oxidized into Al₂O₃ by being exposed to air after the sputtering process, and it did not show any oxidation peak during any of the CV scans. It is acknowledged that under 1 V, the tested metals are all in their oxidated form on the catalyst surface, which aligned with the main objective of this study to compare these catalysts under the same applied voltage. A chronoamperometry (CA) test was conducted under 1 V vs. Ag/AgCl for 1500 s where all the metals except for Ni almost completely lost their activity as depicted in Figure 2b. The CA test result illustrates the current response for all the different anode materials. Among all the tested materials, 200 nm Ni showed the highest current density for the longest reaction time, making it the most active anode material for lignin oxidation in this scenario. However, all other metals besides Ag and Ni, such as Ti, Al, and Cr, were deactivated very quickly probably due to the lack of catalytic activity for their metal oxides. It is also worth noting that Cu was more active than Ag in the first 120 s, but it was deactivated quickly and lost its activity after around 600 s. Although Ni shows the highest activity among all the tested catalysts, its current density also decreased with time. This phenomenon indicated the deactivation of the catalyst due to the phenolic radical intermediates poisoning the catalyst surface, which was reported in one of the previous studies [33].

3.2. Chemical Identification

Besides the targeted compounds vanillin and vanillic acid, GC-MS reveals some of the products which could not be identified in the HPLC because no standard chemicals were available for quantification. Several chemicals were identified with the NIST library. Although the concentrations cannot be quantified using GC-MS, the relative change of amount was determined by comparing the integrated GC-MS peak area at the same retention time under the same GC-MS operation conditions. To increase the apparent yield for a more concentrated sample for product analysis, the GC-MS sample was prepared using a 1 cm × 1 cm Ni foam electrode under 1 V and reacted for 10 h.

Table 1. Identified chemical species by GC-MS whose amount was decreased after the reaction.

Time (min)	Name	Structure
10.20	2-Methoxy-4-vinylphenol
12.15	Homovanillyl alcohol
13.06	Homovanillic acid
13.32	2,4′-Dihydroxy-3′-methoxyacetophenone

and

Table 2. Identified chemical species by GC-MS whose amount was increased after the reaction.

Time (min)	Name	Structure
7.88	Guaiacol
10.55	4-Acetoxybenzaldehyde OR Benzaldehyde, 4-hydroxy-
11.00	Vanillin
11.76	Acetovanillone
12.30	Vanillic acid

show the chemicals whose amount decreased and increased, respectively. For the full GC-MS chromatogram, refer to Figure S4.

Based on the literature and the GC-MS results, we propose a simplified mechanism for electrochemical lignin oxidation as shown in Scheme 1. The depolymerization of lignin results in the monolignols, G-type, H-type, and S-type monolignols (coniferyl alcohol, p-coumaryl alcohol, and sinapyl alcohol, respectively) [34]. The main product from lignin depolymerization, vanillin, is derived from the oxidative cleavage of the $\beta$-O-4 aryl ether linkage within the lignin polymer (shown with a red rectangle) and electrochemical transformation of the coniferyl alcohol groups inside the lignin [35]. Syringaldehyde and guaiacol were also produced in this study; however, the change of the concentration from electrocatalysis was too minor to draw a clear conclusion regarding the yield of these two products. Therefore, they were not quantified in the latter product analysis section of the experiment.

In the proposed scheme, the G and S monolignols were then further oxidized into smaller molecules such as guaiacol, vanillin, acetovanillone, and syringaldehyde. However, acetovanillone could also act as an intermediate and be oxidized into vanillin [36]. Vanillin could be further oxidized into vanillic acid. It is also possible that vanillic acid can be produced directly from the oxidation of coniferyl alcohol without passing through the intermediate aldehyde state [37].

3.3. The Apparent Yield of the Targeted Chemicals

In this study, vanillin and its oxidized form vanillic acid were the major detectable and quantifiable products. The HPLC was calibrated with vanillin and vanillic acid standards (see Figure S5 in the Supplementary Information) and used to measure the quantity of vanillin and vanillic acid before and after the reaction. The HPLC results proved that vanillin and vanillic acid are present in the Kraft lignin solution before the reaction, with an average concentration of 0.196 ppm and 0.332 ppm, respectively. Meanwhile, a blank sample was stirred in open air without applying any voltage to take the oxygen effect from the air into consideration. (See Figures S6 and S7 for HPLC chromatogram representation.)

Figure 3 shows the increase in product concentration compared to the initial concentration. All electrode materials except Cr and Ti led to the increase in either vanillin or vanillic acid concentration under applied potential compared to simply stirring the solution in open air. Furthermore, when compared to the YLS carbon paper without any sputtered metal, Ni, Ag, and Cu significantly increased the generation of the targeted products, which is consistent with the CA results in Figure 2b, where Ni, Ag, and Cu demonstrated oxidative current density in lignin reaction.

The 200 nm Ni sample has the largest yield in both vanillin and vanillic acid compared to YLS carbon paper with no metal, yielding a 62.4% increase in vanillin and a 54.1% increase in vanillic acid. The catalytic effect of Ag and Cu towards vanillin production was not significantly different than bare carbon paper. Previous studies on Ag catalysts support the view that the AgO catalyst is less selective towards aldehyde production: it was found in a methanol electrooxidation study that formate is the only product from methanol [38]. It was also mentioned in previous studies on methanol oxidation that under 300 K and high surface oxygen coverages, the Ag surface reconstructs into AgO, which tends to suppress the selectivity of formaldehyde formation [39]. While maintaining a similar vanillin yield to that of pure carbon paper, Ag and Cu sputtered catalysts promoted vanillic acid generation, with a 16.8% and 10.8% increase compared to bare carbon paper samples, respectively. In contrast, Al did not show a significantly improved catalytic effect towards either of the targeted chemicals. Cr and Ti maintained the same vanillic acid yield as the carbon paper, while their vanillin yield decreased by 11.5% and 14.4%, respectively, which may indicate that the rate of vanillin oxidation [11] surpassed the rate of vanillin generation. Similarly, the decrease in vanillic acid yield for Cr and Ti can be attributed to over-oxidation of vanillic acid. According to a previous study, the most likely aromatic product that the vanillic acid was converted to is protocatechuic acid [40], though it is also possible that it was oxidized into gas products such as CO₂ [41]. However, due to the low conversion rate, the oxidation product of vanillic acid was not identified or quantified in this study.

3.4. ESCA Normalization

The apparent yield could change considerably depending on the surface area, shape, and morphology of a catalyst; thus, it cannot reflect the intrinsic catalytic characteristic of the electrode materials. A normalization method based on the electrochemical active surface area (ECSA) is used here to normalize the different sizes, shapes, structures, and thicknesses of the catalyst. The method of ECSA measurement through electrochemical double-layer capacitance (C_DL) was adapted from previous studies [29,42]. The ECSA measurements are shown in Figures S8–S10.

$$\mathrm{ECSA}=\frac{{\mathrm{C}}_{\mathrm{DL}}}{{\mathrm{C}}_{\mathrm{s}}}$$

(1)

As the specific capacitance C_s was assumed to be the same for catalysts under the same electrolyte condition, the ECSA normalization was simply conducted by normalizing the apparent yield to the C_DL values.

$$\frac{{\mathrm{ECSA}}_1}{{\mathrm{ECSA}}_2}=\frac{{\mathrm{C}}_{\mathrm{DL}1}}{{\mathrm{C}}_{\mathrm{s}}}\times \frac{{\mathrm{C}}_{\mathrm{s}}}{{\mathrm{C}}_{\mathrm{DL}2}}=\frac{{\mathrm{C}}_{\mathrm{DL}1}}{{\mathrm{C}}_{\mathrm{DL}2}}$$

(2)

To assess the intrinsic characteristic of the deposited 200 nm Ni, which gave the best apparent yield for both vanillin and vanillic acid among all the 200 nm samples, a 1 cm $\times$ 1 cm Ni foam and a thinner layer of 100 nm Ni sample were tested for 1500 s under the same experimental conditions and used as a benchmark for the deposited Ni catalyst. The ECSA normalization results are shown below in

Table 3. ECSA normalization of Ni-based catalysts product yield after 1500 s reaction.

Catalyst Type	Vanillin Apparent Yield from HPLC (ppm)	Vanillic Acid Apparent Yield (ppm)	CDL (F)	ECSA Normalized Vanillin Yield (g/h/F)	ECSA Normalized Vanillic Acid Yield (g/h/F)
Ni foam	4.35	3.49	2 × 10−4 (R2 = 0.9973)	2.088	1.6752
200 nm Ni	0.44	1.19	8 × 10−5 (R2 = 0.9994)	0.528	1.428
100 nm Ni	0.43	1.28	4 × 10−5 (R2 = 0.9933)	1.032	3.072

.

The apparent product yield of Ni foam is 9.78 times more vanillin and 2.86 times more vanillic acid than the 200 nm Ni catalyst. Based on the double-layer capacitance measurement, we found that Ni foam possesses 2.5 times more active area than 200 nm Ni, suggesting that using porous structured materials with a different morphology compared to sputtered particles improves the production per active surface area for both vanillin and vanillic acid. When compared with the 200 nm Ni sample, a 100 nm Ni catalyst maintained a similar apparent yield while having only half of the double-layer capacitance. As a result, the 100 nm Ni demonstrated higher vanillin and vanillic acid yield per active surface area. Surprisingly, its normalized vanillic acid yield is even higher than that of Ni foam, indicating that a thinner layer of 100 nm Ni sample prepared by physical vapor deposition could significantly make the reaction more selective towards vanillic acid and also resulted in higher total product yield per active surface area.

3.5. Effects from OER and Charge Efficiency

Figure 3 shows that only stirring in open air leads to a slight increase in vanillin and vanillic acid, so it is reasonable to postulate that the oxygen in the air could assist the lignin oxidation process. As a result, the oxygen evolution reaction (OER) might be not only a side reaction that is competing with the lignin oxidation for the electron supply but also an aid to the lignin oxidation.

In order to obtain an overview of the effects on the lignin oxidation and the OER with a Ni-based catalyst, we scanned a 200 nm Ni catalyst by cyclic voltammetry (CV) as presented in Figure 4. We found that the current density is higher with lignin present on Ni²⁺ before the Ni oxidation peak at 0.4 V. After the Ni²⁺ was oxidized to Ni³⁺, the current density difference becomes larger, which indicates that the Ni³⁺ in NiOOH has a higher activity towards lignin oxidation [43]. The current density is lower with lignin present in the system in the OER region after 0.6 V, indicating that the lignin reaction suppresses the OER reaction.

To validate this hypothesis, as presented in Figure 5, to exclude the effect from dissolved oxygen in the solution as well as the oxygen in the head space, N₂ was purged for 20 min before applying constant potential. Then, the reaction was conducted under 1 V with the N₂ continuously being purged to decrease the oxygen level in the solution. We found that with 100 nm Ni as the catalyst, N₂ purging led to a 47% decrease of vanillin production compared to the unpurged case. With purged O₂, having a Ni catalyst did not make a significant difference in vanillin production. This validated our hypothesis that increasing the concentration of oxygen drives Kraft lignin oxidation towards vanillin production. The 100 nm Ni under 1 V applied voltage with purged O₂ showed the highest vanillin production, indicating a synergistic effect of oxygen and applied electricity. A similar trend was seen with the concentration increase of vanillic acid from purging oxygen. From these results, we postulate that the presence of oxygen, whether it is from anodic OER reaction or purged external oxygen gas, generates a synergetic effect for lignin electro-oxidation towards vanillin and vanillic acid.

The role of the OER thus can be treated as both a competition reaction that consumes electrons for oxygen generation and a factor that accelerates the conversion of lignin to vanillin and vanillic acid. To balance these two characteristics from the OER, we need a new metric to quantify the electrons that go to the wasted oxygen that escaped the system versus the oxygen that helps the lignin reaction. In traditional electrochemistry, Faradaic efficiency (FE) is usually used for characterizing this phenomenon; however, the FE has not been reported in the field of lignin electro-oxidation as it is challenging to calculate FE because the natural variety of lignin molecular structures causes numerous possible reaction pathways with a different number of electrons transferred for each pathway. Moreover, the electron transfer cannot be monitored for these unidentified and unquantified products. A previous study estimated four transferred electrons per mole of product generated [43]; however, each product might have different reaction pathways and a fixed number is not representative of all products. As an alternative, we propose a new metric for energy efficiency ($\mathsf{\epsilon }$) for lignin product analysis based on the total charge supply, where we treated the n (charge transfer per mol of product) as a generic parameter, neglecting the Faradaic coefficient as it is a constant and does not interfere with the relative comparison. The apparent product concentration increase (c) was used as a substitute for the molar amount for the product due to the relatively low yield rate, as shown in Equations (3) and (4). Charge supply was monitored for both tests, and the vanillin and vanillic acid yield was then normalized to the total charge consumed as a representative measure of their charge efficiency.

$$\mathrm{FE}\left(\mathrm{Faradaic}\mathrm{Efficiency}\right)=\frac{\mathrm{n}\times \mathrm{F}\times \mathrm{mol}}{Q_{\mathrm{total}}}$$

(3)

$$\mathsf{\epsilon }=\frac{c}{Q_{\mathrm{total}}}$$

(4)

As tested in the ECSA normalization section (shown in Figure 6a), a thinner layer of 100 nm Ni had a similar level of vanillin and vanillic acid production rate to the thicker 200 nm Ni catalyst. However, from the CA tests, the charge supply for these two catalysts was different, as depicted in Figure 6b.

Throughout the whole 1500 s reaction time, 200 nm Ni consumed 82 C of charge, while 100 nm Ni consumed 65 C of charge. Moreover, 100 nm Ni had 96.5% of the vanillin yield of the 200 nm Ni sample and 105% of its vanillic acid yield. As a result, as shown in

Table 4. Charge efficiency comparing 100 nm Ni and 200 nm Ni sample.

Catalyst Type	Vanillin Apparent Yield from HPLC (ppm)	Vanillic Acid Apparent Yield (ppm)	Charge (C)	Vanillin Charge Efficiency (ppm/C)	Vanillic Acid Charge Efficiency (ppm/C)
Ni foam	4.35	3.49	75	5.80 × 10−2	4.65 × 10−2
200 nm Ni	0.44	1.19	82	5.41 × 10−3	1.45 × 10−2
100 nm Ni	0.43	1.28	65	6.58 × 10−3	1.95 × 10−2

, a thinner layer of deposited Ni catalyst can maintain the high vanillin and vanillic acid production while saving the supplied electrical charge, resulting in 22% higher charge efficiency for vanillin and 35% higher charge efficiency for vanillic acid.

Considering that lignin suppresses OER at 1 V (as shown in Figure 4), we postulate that due to the bulky polymeric molecular structure of lignin, the mass transfer is intrinsically difficult for lignin molecules, which resulted in moderate current density towards the OER and vanillin production. That is, the OER can occur in the smaller pores of the thin Ni catalyst, which lignin molecules cannot access for oxidation reaction, and therefore, simultaneous OER would lower the lignin product charge efficiencies, but the presence of oxygen can improve the ECSA-normalized yield on vanillic acid (Figure 5). As a result, Ni foam had better charge transfer performance given its 3D structure, which allowed more active surface area for the oxidative conversion of lignin. Moreover, while Ni is active for lignin oxidation in all the catalysts studied, ECSA-normalized vanillic acid production is almost three times greater than vanillin in thin Ni catalysts, whereas vanillin is preferred over vanillic acid on Ni foam. This is consistent with our hypothesis that surface oxygen species facilitate a higher degree of oxidation of lignin species. To further improve reaction selectivity and charge selectivity towards lignin oxidation, we would need to have a thinner layer of active catalyst such as 100 nm Ni, while maintaining a foam-like 3D structure akin to Ni foam to allow macro molecules to access the active surface area.

4. Conclusions

In this study, six different metal catalysts prepared with the vapor deposition method were tested in a H-cell setup at room temperature and pressure for Kraft lignin oxidation. With GC-MS, various aromatic compounds such as vanillin, vanillic acid, syringaldehyde, and guaiacol were identified. Two targeted chemicals, vanillin and vanillic acid, were also quantified with HPLC. Given the best performance of 200 nm Ni catalyst, nickel-based catalysts were selected for further investigations of selectivity optimization. It is also noteworthy to observe the synergetic effects between purged oxygen and applied voltage which boosted vanillin production. To address the lack of specific and accurate metrics to evaluate intrinsic electrocatalyst performance, a key challenge resulting from the complex nature of lignin, we propose that catalyst active surface area normalization and charge efficiency be considered in future studies on lignin electro-oxidation. The active surface area normalization accounts for the different shapes and morphology of the catalyst surface under the same geometric area, while the supplied charge normalization acts as an alternative for the Faradaic efficiency in the lignin electrochemical oxidation field to evaluate the charge transfer efficiency for different catalysts. We found that after the ECSA normalization, Ni foam showed a higher production rate for vanillin, while a smaller increase was seen for vanillic acid generation per active area. Thinner deposited Ni of 100 nm thickness catalyst also showed higher vanillin and vanillic acid generation per active surface area than 200 nm Ni, and its ECSA-normalized vanillic acid production was higher than that of Ni foam. We also found through the supplied charge normalization that the thinner deposited Ni layer can maintain a relatively high yield for targeted products while requiring less electric charge. These two metrics could be applied to future lignin electrocatalytic valorization projects to design thinner, more porous layered catalysts, which would provide more insights into the electrochemistry of these reactions. We also recommend that researchers should apply this approach to different types of lignin, for example, from different plant sources, to determine whether our conclusions apply to other types of lignin than Kraft lignin.