Development of Lignin-Based Mesoporous Carbons for the Adsorption of Humic Acid.

There is an increasing urge to make the transition toward biobased materials. Lignin, originating from lignocellulosic biomass, can be potentially valorized as humic acid (HA) adsorbents via lignin-based mesoporous carbon (MC). In this work, these materials were synthesized for the first time starting from modified lignin as the carbon precursor, using the soft-template methodology. The use of a novel synthetic approach, Claisen rearrangement of propargylated lignin, and a variety of surfactant templates (Pluronic, Kraton, and Solsperse) have been demonstrated to tune the properties of the resulting MCs. The obtained materials showed tunable properties (BET surface area: 95-367 m2/g, pore size: 3.3-36.6 nm, V BJH pore volume: 0.05-0.33 m3/g, and carbon and oxygen content: 55.5-91.1 and 3.0-12.2%, respectively) and good performance in terms of one of the highest HA adsorption capacities reported for carbon adsorbents (up to 175 mg/g).

Introduction

The transition from a petroleum-based economy toward a biobased economy gained a lot of interest in recent years, both from industry and the academic world, due to environmental and sustainability concerns. Lignocellulosic biomass is an especially interesting feedstock because it is not competing with food and feed production.[1] Depending on the biomass source, plant cell wall material is composed of cellulose (35–80%), hemicellulose (5–35%), and lignin (15–35%).[2] There are multiple possibilities for their conversion from both cellulose and hemicellulose to ethanol, furfural, platform sugars, and their derivatives[3−5] and from lignin to fuels, chemicals, and materials.[4,6−8]

Lignin is an amorphous material consisting of highly branched aromatic macromolecules based on three basic phenylpropanoid components (coumaryl, coniferyl, and sinapyl alcohols) bonded together.[9] In nature, lignin provides lignocellulosic materials with inner strength, rigidity, and dimensional stability due to the chemical bonding with cellulose and hemicellulose, water impermeability and its transport within the plant, and it also has antioxidant, antimicrobial, and antifungal properties among others.[10,11] Currently, lignin is considered as a byproduct of the lignocellulose bioethanol production and paper industry, and it has a relatively low value and quality. From the approx. 70–100 million tons of lignin produced annually, only less than 5% is not used as an energy source.[12] To valorize lignin and improve its use, high-value applications are being developed, and lignin biorefineries converting lignin into high-value products are increasingly drawing attention.[13] Possible value-added applications of lignin include its conversion to chemicals, fuels, and carbonous materials, its applications in polymeric materials for adhesives, coatings, and composites, and as antioxidant or flame retardant among others.[7,8,11−15]

Carbonous materials, such as mesoporous carbons (MCs), are widely used in multiple fields, such as adsorption, separation, catalysis, drug delivery, fuel cells, and energy storage and in electronics as a result of their excellent properties: unique morphology, high specific area, pore volume, and excellent mechanical, chemical, and thermal stability.[16−21] They can be obtained via different routes: soft- and hard-templating, physical and chemical activation, sol–gel method, hydrothermal carbonization, or direct carbonization.[16,19,20,22−24] In this research, the soft-templating method of a lignin precursor was applied. In this method, a carbon precursor is cross-linked in the presence of a structure-directing agent (surfactant, block copolymer) after their self-assembly, to form an organized mesophase. Afterward, the surfactant is removed by means of a thermal treatment, followed by the carbonization of the cross-linked resins, as presented in Figure . One of the most common synthetic routes in soft-templating methods is evaporation-induced self-assembly (EISA). The principle of the EISA method is as follows: polymerizable precursors and structure-directing agents form micelles in the solution, and upon solvent evaporation, their concentration is increasing, promoting more organic–organic self-assembly of the components.[25] Effective self-assembly is driven by the concentration gradient in EISA.[26] Tunability of the MC properties is related to the selection of the precursor, synthetic route, surfactant type and concentration, and carbonization temperature.[20,27]

Figure 1

MC synthesis via the soft-templating EISA method. Adapted with permission from ref (28). Copyright 2014 Elsevier.

Commonly used petroleum-based carbon precursors for MCs are phenolic resins, such as phenol, resorcinol, or phloroglucinol–formaldehyde resins, phenolic-functionalized melamine, and (phenol)–urea–formaldehyde resins.[16,20,25,29−32] The synthesis of MCs from biobased precursors is gaining interest in modern materials science research, and it was described in multiple reviews. Precursors, such as plant biomass, lignin, fungus, algae, chitin, chitosan, gelatin, ionic liquids, various sugars, amino acids, and their derivatives, have been demonstrated.[16,19,20,24,33−37] Use of lignin in this application has an advantage of using a non-toxic, renewable, and widely available resource, which is structurally similar to phenol–formaldehyde, one of the well-established precursors. Lignin can be incorporated into MC fabrication using various pathways, such as templating methods and physical and chemical activation.[15,38,39]

Lignin-based MCs fabricated via the soft-templating have been reported in the literature.[40−45] In these reports, lignin–formaldehyde or lignin–phenol/phloroglucinol–formaldehyde resins are synthesized and cured in the presence of a surfactant (Pluronic type), with the disadvantage of the use of toxic phenol and/or formaldehyde. In the next step, carbonization of the resin is performed. Such MCs have a relatively high surface area (81–685 m2/g) and pore volume (0.075–0.62 cm3/g).[40−45] In general, lignin-based MCs find their applications in the fields of supercapacitors, conductive and energy storage materials,[40,41,46−48] catalysis,[44] delivery systems,[28,43] adsorption of organic molecules or heavy metals from water,[49,50] or air pollutant adsorption.[50,51]

In contrast to the state-of-the-art,[41−44,50,52] where the toxic formaldehyde is used to cross-link the system, we developed a formaldehyde-free method to cross-link lignin and for further MC fabrication. Here, propargyl-modified lignin was cross-linked via the Claisen rearrangement, in the presence of a surfactant, using the soft-templating EISA strategy (Figure ). Propargylated lignin has been reported to self-cross-link via the Claisen rearrangement,[53,54] and it can be used as a precursor in single component carbon fiber preparation.[55] This is the first report of the Claisen-rearrangement approach for lignin cross-linking used for MCs precursor synthesis. Four different surfactants were selected as a template: Pluronic F127 and P123, Kraton G1652, and Solsperse M387. The most widely used surfactants are represented by the Pluronic family, consisting of poly(ethylene oxide) (PEO) and poly(propylene oxide) (PPO) block copolymers. The Pluronics were selected because they are up to now the workhorses for the preparation of MCs. While PPO orients toward the core of the micelles, PEO stretches out into the aqueous medium where it undergoes H-bonding with, for example, OH groups in the carbonous precursor resin. In this work, we investigated surfactants with various levels of hydrophobicity. Therefore, more hydrophobic surfactants like Kraton G1652 and Solsperse M387 were tested. Those surfactants have not yet been reported as a template for MCs synthesis. Kraton G1652 is a hydrogenated triblock copolymer with polystyrene end blocks and a poly(ethylene/butadiene) middle block (SEBS), while Solsperse M387 is a basic and weakly cationic and H-bonding dispersant with known affinity for aromatics.[56] They were selected based on their foreseen possibility to interact with the lignin precursor via π–π stacking. Cured resins were carbonized at three different temperatures (600, 800, and 1000 °C), characterized, and applied as adsorbents for humic acid (HA).

Figure 2

Synthetic route of obtaining the lignin-based MC precursors via the soft-templating method. The resin is cured via Claisen rearrangement after lignin propargylation.

HA is one of the components of the surface and ground water. It is a complex, polydisperse biomacromolecule, categorized as dissolved organic carbon.[57] HA dissolved in water not only influences the color, taste, and odor of water but also causes fouling during water purification, and it is a potential hazard for human health during potable water processing.[58] It is a precursor to potentially carcinogenic trihalomethanes and haloacetic acids produced as byproducts during water disinfection by chlorine or chloramines.[58,59] Moreover, HA tends to form organometallic complexes with heavy metals, increasing their bioavailability.[60] There are multiple ways to remove HA from water, including adsorption, membrane filtration, coagulation, flocculation, electrochemical methods, catalytic oxidation, and decomposition.[57,61] Adsorption is an important HA elimination technique. Examples, advantages, and disadvantages of the various adsorbents are a topic of multiple reviews.[57,62,63] Those adsorbents can be categorized in the following groups: carbonaceous, mineral, and polymeric materials and (nano)composites and nanomaterials (nanoparticles and nanotubes). The examples of such materials are mesoporous and activated carbons, clays and zeolites, ion-exchange membranes, carbon nanotubes, and metallic nanoparticles and their composites, respectively.[57,62,63] The MCs synthesized in this article will be benchmarked with state-of-art MC, synthesized using the lignin–formaldehyde approach (soft templated with Pluronic F127),[42] and commercial activated carbon, to evaluate their performance as HA adsorbents.

The main objective of this work is the development of lignin-derived MCs suitable for the adsorption of HA. The strategy is to synthesize carbonous materials from lignin by its cross-linking via the Claisen rearrangement, in the presence of different surfactants, using the soft-templating EISA approach. The investigated surfactants varied in structure and hydrophobicity. The ultimate goal is to correlate the MC properties with their HA adsorbent capacity.

Results and Discussion

MC Precursor Synthesis

The propargylated lignin is expected to interact with the outer part of the micelles, which is formed by the self-assembly of the surfactant as the solvent (2-methyltetrahydrofuran) evaporates. Upon heating above 150 °C, Claisen rearrangement of the propargylated lignin takes place.[53] First, benzopyran structures are formed through [3,3] and [1,5]-sigmatropic and oxa-Diels–Alder reactions.[64] Then, benzopyran moieties present in the modified lignin structure were thermally polymerized at 180 °C for 14 h to ensure maximal conversion (Figure ). The conversion has been determined by means of FT-IR, by observing the disappearance of the alkyne moiety band (3300 cm–1). The conversion was high, in the range of 88–99%, as shown in Figure . Examples of the spectra before and after cross-linking are shown in Figure S2, Supporting Information.

Figure 3

Claisen rearrangement conversion determined by FT-IR of P1000 lignin in the presence of the four surfactants.

MC Synthesis and Characterization

Once the lignin-based resins were cross-linked, they were carbonized at three different temperatures (600, 800, and 1000 °C). As lignin is carbonized, various pyrolyzates are formed at high temperatures: lignin dehydration and decarboxylation takes place, followed by the formation of polycyclic aromatic hydrocarbons (PAHs), coke, tar, and char. Above 400 °C, methoxy groups in lignin are transformed to catecholic and pyrogallolic structures, together with side-chain C–C and C–O bond cleavage, resulting in the start of lignin depolymerization and repolymerization and initial PAHs, char, and coke formation.[65] Around 550 °C, catecholic and pyrogallolic structures tend to be transformed into coke, together with increased PAH formation, especially when the temperature increases above 700 °C. At elevated temperatures, larger-molecular-weight PAHs are being formed. At temperatures above 1000 °C, PAHs are the dominant product of lignin tar decomposition.[66]

It was generally observed that the carbon content, as determined by elemental analysis, increased with increasing carbonization temperature (Table ). For instance, the MC P123, carbonized at 600 °C, had 67.1 wt % carbon, while the same resin carbonized at 800 and 1000 °C had a carbon content of 69.0 and 91.1 wt %, respectively. G1652 and F127 resin sets were an exception, with a similar carbon content for 600 and 800 °C carbonized materials.

Table 1

Characteristics of the MCs and Activated Carbon

material	SBETa[m2/g]	VBJHb[cm3/g]	Vmicroc[cm3/g]	Dp BJHd [nm]	Ce [wt %]	He [wt %]	Ne [wt %]	Oe [wt %]
activated carbon	1126	0.32	0.22	2.7	86.0	0.3	4.5	9.1
G1652_600	136	0.05	0.04	34.4	58.5 ± 0.6	1.6 ± 0.1	<1	3.0 ± 0.1
G1652_800	102	0.08	0.03	22.5	55.5 ± 2.1	<1	1.4	12.2 ± N.A.
G1652_1000	107	0.07	0.03	36.6	91.0 ± 1.2	<1	1.6	8.2 ± 0.3
M387_600	242	0.14	0.06	5.8	58.9 ± 0.8	1.7 ± 0.1	3.8 ± 0.1	6.1 ± 0.5
M387_800	95	0.12	0.02	5.4	63.9 ± 2.2	<1	4.6 ± 0.3	9.6 ± 0.4
M387_1000	142	0.15	0.03	5.4	84.7 ± 0.8	1.2 ± 0.1	3.0 ± 0.1	11.4 ± 0.2
P123_600	371	0.30	0.10	10.2	67.1 ± 0.3	1.7 ± 0.1	<1	5.8 ± 0.2
P123_800	277	0.28	0.07	12.2	69.0 ± 0.3	1.1 ± 0.1	<1	8.5 ± 0.1
P123_1000	291	0.33	0.06	10.7	91.1 ± 1.7	1.1 ± 0.1	<1	9.0 ± 0.1
F127_600	367	0.14	0.09	3.7	63.9 ± 0.5	1.7 ± 0.1	<1	5.5 ± 0.2
F127_800	348	0.12	0.09	3.3	72.2 ± 0.4	<1	<1	3.6 ± 0.2
F127_1000	217	0.12	0.04	3.5	77.4 ± 1.1	<1	<1	7.4 ± 0.1

Determined using the Brunauer–Emmett–Teller (BET) method.

BJH desorption cumulative volume of pores between 17.000 and 3,000.000 Å width. No pore volume was detected above 50 nm in pore size distribution.

Calculated by the BJH model from the desorption branches of the isotherm.

Calculated by the V–t method.

Determined by elemental analysis.

During lignin carbonization, oxygen containing moieties, such as hydroxyl, methoxy, and carbonyl groups, are removed, and the remaining carbon network is transformed to PAH, while the carbon framework remains. This explains the increased carbon content with carbonization temperature.[67] This effect was observed for all the surfactants used in this study. MCs carbonized at 1000 °C exhibited a similar carbon content as the reference activated carbon, except for MC F127_1000.

Characterization of the synthesized MCs by elemental analysis did not yield a complete mass balance of H, O, N, S, and C atoms in most cases. Therefore, additional X-ray photoelectron spectroscopy (XPS) experiments were performed on P1000 lignin and propargylated P1000 lignin, and the P123 MC was set to investigate this issue. Elemental analysis provides information about the content of carbon, hydrogen, nitrogen, sulfur, and oxygen, while XPS can detect all elements except hydrogen and helium. XPS showed that sodium and chlorine are present in the MC samples (but not in P1000 lignin), up to 4.6 and 4.4 wt %, respectively. Since the workup includes use of concentrated sodium chloride solution, those findings are plausible.

Nitrogen sorption measurements (Figure for the materials carbonized at 800 °C and Figures S4 and S5, Supporting Information, for the materials carbonized at 600 and 1000 °C) exhibited isotherms of type IVa for all the synthesized MCs, or isotherm type Ib for the commercial activated carbon, as expected. Materials with isotherm type IV are classified as mesoporous, while those which exhibit isotherm type I are considered to be microporous to mesoporous (with small mesopores, up to 2.5 nm).[68] It can be noticed that for F127 and P123 materials, the hysteresis is not closing. This phenomenon has been reported for calcinated phenol–formaldehyde resins made with F127 template and may be related to swelling of the polymer in nonsolvent (here: condensed nitrogen).[69,70] The carbonized materials exhibited different hysteresis loops, associated with capillary condensation, pores blocking, and/or cavitation-controlled evaporation. A H2a hysteresis loop was observed for F127 carbonized materials, which is typical for the materials exhibiting pores blocking in a narrow range of pore necks or cavitation-induced evaporation.[68] M387 carbonized materials showed a combination of H2a and H2b hysteresis loops, which indicates pores blocking caused by both narrow and wide pore necks, what can be supported by the pore size distribution (PSD).[68] Another type of hysteresis loop, which was found for G1652 and P123 carbonized materials, was type H4, often seen for micro-MC systems.[68] Together with a shift of the hysteresis position to higher p/p0 values, an increase in the pore size was observed (Figure ).

Figure 4

Nitrogen sorption isotherm (top) and PSD (bottom) of the resins carbonized at 800 °C.

The results, as listed in Table , clearly indicate the importance of the selection of the right surfactant in order to tune the pore size and volume. Exceptionally good results in terms of Barrett–Joyner–Halenda (BJH) pore volume were obtained for P123, followed by F127. Possibly, the interaction of those two surfactants, characterized by the different length of the PPO–PEO–PPO units (Pluronic F127 PEO106–PPO70–PEO106 and Pluronic P123 PEO20–PPO70–PEO20), results in different micellization efficiency due to the ratio of PEO to PPO units. Overall, the Pluronic surfactants in Claisen rearrangement-based lignin cross-linking tend to be a better precursor to mesoporous materials as the MCs templated by them exhibited higher BET surface areas and higher pore volumes (Figure ).

Figure 5

Summary of total BET surface areas (top) and pore volumes (bottom) of the synthesized MCs.

The difference in material properties may arise due to different properties of the surfactant, such as chemical structure, chain architecture, chain lengths and ratio, and so forth, which result in different types of interactions (self-assembly of the structure-directing agent and the precursor) influencing the structure of the resulting MCs.[71] For instance, there is hydrogen bonding between the hydroxyl functional groups of lignin (phenolic OH and aliphatic OH groups) and the ethers of the hydrophilic PEO outer blocks of the micelles, with the hydrophobic PPO buried inside to avoid contact with the hydrophilic hydroxyl groups in lignin Pluronics.[27,72] Kraton G1652 [a poly(styrene-ethylene/butadiene-styrene) (SEBS) elastomer] and Solsperse M387 (a basic and weakly cationic and H-bonding dispersant) may interact via weaker hydrophobic interactions and/or π–π stacking with aromatic moieties in the modified lignin. The stronger interactions between lignin and Pluronics can explain the better material properties obtained with those surfactants compared to the hydrophobic surfactants. Especially, interesting MCs were obtained with a Pluronic P123 template. Those MCs had the highest surface area SBET, combined with the highest VBJH pore volume. Another interesting case was the MCs templated with Pluronic F127. Those materials showed similar surface area and slightly lower VBJH pore volume. Interestingly, materials templated with Pluronic F127 exhibited narrow PSD, with a major peak at 4 nm, while Pluronic P123 had a wide peak with a maximum at 8 nm, tailing to 80 nm. The broader PSD with Pluronic P123 as template can be correlated to the shorter hydrophilic PEO segment (PEO20) compared to that of F127 (PEO106), resulting in weaker hydrogen bonding between the template and precursor.[73] Materials prepared with those two surfactants showed a decrease in surface area with increasing carbonization temperature, which can be a result of the pore shrinking and/or collapse during the carbonization process. This may also cause a lower pore volume.

The obtained TEM images (Figure ) confirm the observations from nitrogen sorption data regarding the morphology of the pores. Activated carbon consists of a bulky structure, with micro- and mesopores. In the TEM micrographs of M387_800, the porous domains are visible. Mostly small, uniform mesopores are observed, which is supported by the nitrogen sorption data. P123_800 consists of a more confined structure, with clearly visible mesopores and possibly macropores. A high number of pores are visible, which was also observed in the nitrogen sorption data, represented by the high pore volume. In F127_800, mesoporous domains can be observed, randomly distributed. The number of pores seems to be moderate, but the pores tend to have a relatively narrow size range.

Figure 6

TEM micrographs of activated carbon and lignin-based resins carbonized at 800 °C.

Surprisingly, Kraton G1652 templated MCs exhibited spherical carbon domains along the agglomerated structures. These spherical structures range from 20 nm to 1 μm in diameter. These spheres are observed only in the G1652 templated carbons. Due to the dense structure of the spheres, visualization of distinct pores is hindered, which could also be explained by the low pore volume observed in nitrogen sorption experiments. The carbon (nano)spheres morphology could originate from the interaction with the surfactant, which may enable lignin to rearrange in these spherical structures during the EISA step. Kraton G1652 consists of SEBS units and is consequently hydrophobic. During the resin synthesis, the surfactant is interacting with the hydrophobic and aromatic domains of the lignin precursor. The hydrophilic domains, for example, remaining hydroxyl groups in lignin, could have been repelled by the surfactant and agglomerated together via hydrogen bonding. This self-assembly could have resulted in the spherical structures, which are observed for the carbonized materials. Li et al. reported the synthesis of MC spheres, with 1320 m2/g specific surface area and 3.5 cm3/g pore volume with a hierarchical pore size of 3.5–60 nm via a dual templating method, soft and hard templating.[74] The carbon spheres exhibited a more defined pore structure compared to the material obtained by Kraton G1652, and the spheres they obtained were in the range of 3–5 μm.[74] The synthesized novel material could open an alternative way to synthesize porous carbon nanospheres.

The synthesized carbons can be benchmarked with literature results where a soft-templating approach with lignin as a precursor and Pluronic F127 as a template was used (Table ). The resin cross-linking uses very harmful formaldehyde or less harmful glyoxal, while the lignin precursor is different in all the reported cases. Here, we propose lignin cross-linking via the Claisen rearrangement as an alternative. The surface area of the MCs obtained in our study is similar to the phenol–formaldehyde references, except from the one reported by Chen et al.[40] In that study, a high amount of KOH was used, which can increase the porosity by chemical activation in the MC synthesis combined with the soft-templating methodology. The total pore volume of our materials is comparable with that reported by Saha et al.(43) and lower than that reported by Chen et al.(40) and Qin et al.(42) but better than the ones obtained by Herou et al.(41) When the pore volume of the mesopores is compared, the best-performing materials from this study are superior to all the references, except for that reported by Saha et al.(43) and Chen et al.,[40] showing similar or slightly better characteristics. Pore sizes of our MCs vary between 3.3 and 36.6 nm. PSD is related to the surfactant used as a template. All the literature references using the lignin–formaldehyde cross-linking approach used Pluronic F127 as a template, and the pore size was in the range of 3.4–3.8 nm, which is in the same range as our materials templated with Pluronic F127 (3.3–3.7 nm). The use of another template enables to tune this property, as presented in Table . In general, materials presented in this study exhibit good performance compared with what has been reported already in the field of porous lignin materials, with the advantage that our materials are prepared via a formaldehyde-free synthesis route.

Table 2

Lignin-Based MCs Obtained via Soft-Templating with Pluronic F127 Method Summary

lignin precursor	cross-linking method, carbonization temperature	SBET [m2/g]	Vtot [cm3/g]	VBJH [cm3/g]	Vmicro [cm3/g]	Dp BJH [nm]	ref.
P1000 lignin	Claisen rearrangement, 600 °C	367a	0.23b	0.14c	0.09d	3.7e	best result from this article
lignosulfonate (Sigma-Aldrich)	lignin with formaldehyde, 600 °C	685a	0.44b	0.21g	0.23d	3.7g	(40)
Kraft lignin (Sigma-Aldrich)	lignin–phenol–formaldehyde (lignin/phenol 1:3), 900 °C	260a	NA	0.08g	NA	3.4g	(44)
softwood lignin (Nanjing Forest University)	lignin with formaldehyde, 600 °C	236–466a	0.48–0.62g	NA	NA	3.8g	(42)
methanol soluble fraction obtained from Kraft lignin	lignin with formaldehyde, 1000 °C	205–418a	0.19–0.50h	0.11–0.34h	0.02–0.16h	NA	(43)
Masson pine alkali lignin (Nanping Paper-making Co. Ltd)	lignin with formaldehyde, 900 °C	345a	NA	0.03f	NA	3.4e	(45)
Organosolv Lignin (Fraunhofer Institute Iena)	lignin with glyoxal, 900 °C	81a	0.11i	0.09g	NA	3.6i	(41)

Determined using the BET method.

Calculated as the amount of nitrogen adsorbed at a relative pressure of 0.95.

BJH desorption cumulative volume of pores between 17.000 and 3,000.000 Å width.

Calculated by the V–t method.

Calculated by the BJH model from the desorption branches of the isotherm.

Not defined in the experimental section how this value was obtained.

Derived from the adsorption branches of the isotherms by the BJH method.

Calculated by employing nonlocal DFT.

Calculated from the adsorption line using a DFT model.

Unfortunately, none of our synthesized MCs exhibited an ordered porous structure, which has been reported by Qin et al.(42) and Wang et al.[45] The formation of ordered pores arises from the high hydroxyl group content of the used low-molecular-weight lignin (1–1.5 kDa used by Qin et al.,[42] and Mw of lignin used by Wang et al.(45) was not reported), which enables conductive self-assembly of the micelles.[75] Therefore, the observed non-ordered pore structure in our case could arise from a lower amount of hydroxyl groups (0.79 mmol/g, unmodified P1000 5.51 mmol/g) after the functionalization of the lignin precursor with propargyl units and its relative higher molecular weight (4.3 kDa) compared to the lignin used by Qin et al.(42) By consequence, the H-bond interaction between our propargylated lignin precursors and the surfactant is weaker (there are less hydrogen bonds), resulting in less ordered porous materials because the interactions are already destroyed in the heating process.

HA Adsorption Studies

The main objective of this work was the development of the MCs suitable for the adsorption of HA and to correlate their properties with the adsorbent capacity. The synthesized MCs are compared with activated carbon by benchmarking their HA adsorption capacities. The results are summarized in Figure .

Figure 7

Comparison of HA adsorption of activated carbon and mesoporous materials synthesized in this study.

Activated carbon showed a HA adsorption capacity of 153 mg/g. Other reported commercial activated carbons exhibited lower HA adsorption capacities (up to 56 mg/g) than the activated carbon used in this study, probably because of lower specific surface area, (VBJH) pore volume, and pore size of those activated carbons.[76−80] The HA adsorption on activated carbon is hindered by the size exclusion effect due to the small pore size.[81] The adsorption capacity of activated carbon is in the same range as the G1652_800, G1652_1000 M387_800, M387_1000, P123_800, P123_1000, and F127_1000 MCs. HA adsorption for materials carbonized at 800 and 1000 °C is comparable for most of the materials, except for the F127 set, where adsorption of the material carbonized at 1000 °C is significantly increased compared to the 800 °C one. Materials carbonized at 600 °C showed the lowest HA adsorption in each set, with pronounced difference in the performance compared to materials carbonized at higher temperatures and reference activated carbon. This could be attributed to the fact that the materials carbonized at 600 °C showed the lowest carbon content (Table ).

Libbrecht et al. reported that an increase in the pore size and carbon content promote higher HA adsorption due to the hydrophobic π–π stacking adsorption mechanism.[67] Unfortunately, we do not observe it for our materials, which have a lower carbon content than the ones reported by Libbrecht et al. which have 77–95% carbon. The lower carbon content in the reported materials would be expected to yield in lower adsorption since carbon enhances the hydrophobic character of the material and therefore the van der Waals and hydrophobic interactions.[67] Upon a pH increase, the carboxyl groups of the HA deprotonate first (pKa = 5), and at higher pH values, the phenolic hydroxyl groups (pKa = 10) deprotonate as well. This makes HA more hydrophilic. The adsorption experiments were performed at a pH of 7; therefore, the HA will be charged and become more hydrophilic. It was observed that the MCs, which had an oxygen content above 7 wt %, showed good HA adsorption if they exhibited moderate to good material properties, such as surface area and mesoporous volume. A possible explanation might be the presence of non-covalent interactions between HA and the adsorbent, for example, hydrogen bonding. The hydrogen bonding could enhance the interaction strength between the adsorbent and the adsorbate and increase the adsorption capacity. Therefore, the increased oxygen content can be beneficial for HA adsorption. Additionally, the large pore size enables the adsorption of HA, whose reported hydrodynamic radius is 7.1 nm.[27] For materials with a pore size lower than 7.1 nm, the size exclusion effect may prevent the adsorption of large amounts of HA.

Unfortunately, it is difficult to draw general conclusions regarding the relationship between synthesis route (used approach, surfactant, and experimental conditions), materials properties, and HA adsorption. For each case, the surfactant had a different effect on the properties of the materials, and therefore, the combination of different factors (BET surface area, pore volume, diameter, and elemental composition) influenced the adsorption properties. As multiple properties influence HA adsorption, it is very difficult to understand the interplay between them and the HA adsorption. The G1652 MCs exhibited good to very good HA adsorption, despite the low pore volume. Their performance can be connected with the high oxygen content in those MCs, combined with big pore sizes. M387 materials, on the other hand, exhibited low pore sizes, but their good performance can be attributed to the high oxygen content and moderate mesoporous volume. P123 MCs have the best performance from all the synthesized materials, and the excellent performance can be the result of the combination of the highest mesoporous volume, high BET surface area, pore size above 10 nm, and high oxygen content. F127 materials exhibited moderate to good performance, which can be correlated with the high BET surface area and moderate mesoporous volume.

When compared with other carbon materials used for HA adsorption (Table ), it turns out that the performance of the materials synthesized in this study is very good and is outperformed only by the materials synthesized from resorcinol–formaldehyde resin by Libbrecht et al. and coal-based mesoporous activated carbons prepared by Lorenc-Grabowska et al.(67,82) Other materials have significantly lower HA adsorption capacity than our materials, despite their usually higher surface areas and pore volumes.[81,83−86] In comparison to commercial activated carbons, their performance is drastically lower than the MCs presented in this study.[76−80]

Table 3

Comparison of Carbon Materials Used as HA Adsorbents

material	SBETa[m2/g]	Dp BJH [nm]	Vtot [cm3/g]	HA adsorption [mg/g]	conditions (pH, temp [°C])	ref.
commercial filtrasorb 200 activated carbon	NA	NA	NA	24	5.0, 25	(76)
commercial adsorbent JZN	149	NA	0.63b	3	7.0, 25	(83)
commercial, treated coconut shell-based activated carbons	NA	NA	NA	8–70	NA, 30	(80)
commercial activated carbon treated with sulfuric and phosphoric acids	659–724	NA	NA	19–26	NA, RT	(77)
commercial cellulose and powdered activated carbon	102–834	NA	NA	89–30	2.0, 37	(79)
chitosan-treated granular activated carbon (filtrasorb 400)	NA	NA	NA	56–71	7.0, 25	(78)
chitosan-encapsulated activated carbon	316b	3.28b	0.26b	72	6.4, 30	(85)
ammonia and hydrogen-treated mesoporous activated carbons	415–518	NA	0.51–0.53c	90–137	NA, RT	(84)
coal-based mesoporous activated carbons	367–850	NA	0.24–0.47b	170–200	12.0–12.5, 25	(82)
soft-templated MCs based on resorcinol–formaldehyde resins	422–670	3.5–27d	0.74–1.56c	155–352	7.0, 25	(67)
hard-templated (mesoporous SiO2) ordered MC	988	NA	1.33c	51	6.0, 25	(81)
carbonized rice containing nanosized magnetite	NA	NA	NA	80	7.61, RT	(87)
magnetic porous carbon based on pyrolyzed biomass, containing γ-Fe2O3	63	NA	NA	96	5.0, NA	(88)
dual-pore carbon shells synthesized by templating against silica nanospheres	750–1350	2.4–9.5b	0.4–1.1b	12–100	NA, 25	(86)
P123_1000	290	10.7d	0.37c	175	7.0, 25	this work
P123_800	277	12.2d	0.28c	165	7.0, 25	this work
F127_1000	156	3.5d	0.12c	165	7.0, 25	this work
G1652_800	102	22.5d	0.07c	161	7.0, 25	this work
M387_800	95	5.4d	0.13c	158	7.0, 25	this work
AC	1026	2.3d	0.50c	153	7.0, 25	this work

Determined using the BET method.

Not defined in the experimental section how this value was obtained.

Calculated as the amount of nitrogen adsorbed at a relative pressure of 0.95.

Calculated by the BJH model from the desorption branches of the isotherm.

Conclusions

A series of novel, lignin-based MCs were successfully prepared using a Claisen rearrangement route. The soft-templating method using EISA was applied to obtain mesoporous materials after carbonization. We presented the material characteristics obtained through the synthesis using various surfactants used as a template and at different carbonization temperatures. All the materials are benchmarked as HA adsorbents. The ultimate goal of this study was to correlate carbonous material properties with the humic adsorption capacity. High HA adsorption can be influenced by various properties of MCs, such as carbon and oxygen content, pore size, volume, and BET surface area. These properties can be steered by changing the surfactant type and concentration, resin preparation methodology, and carbonization temperature; however, it is not easy to understand the complex interplay between all these aspects in the material preparation.

The MCs synthesized in this research exhibited excellent performance as HA adsorbents compared to other carbonous materials in terms of adsorption capacity. The best results are obtained with Pluronic P123 surfactant and carbonized at 1000 °C because this material possesses high surface area and pore volume with average pore size above 7 nm, with one of the highest carbon content obtained in this study. This work is one of the examples of lignin use in value-added applications, such as adsorbents for water purification, which is an interesting highlight adding to lignin valorization. This is a proof-of-principle application, but we believe that it is possible to use these materials for other applications as well, such as supercapacitors and energy storage materials.[16,20,89,90]

Experimental Section

Chemicals

All chemicals were used as received, unless stated otherwise. Protobind P1000 Lignin, soda lignin extracted from wheat straw [Mw = 4.3 kDa, D̵ = 2.6, measured on a gel permeation chromatograph (GPC) with refractive index (RI) detection, calibrated with polystyrene standards, using tetrahydrofuran as the eluent, after acetylation of the lignin,[91] was provided by Green Value S.A. It contains less than 4 wt % carbohydrates and less than 2 wt % ash. Elemental analysis of this lignin yields the following composition: 59% C, 6% H, 26% O, 1% N, and 0.1% S.[92] Propargyl bromide solution (80 wt % in toluene solution, with 0.3% magnesium oxide as the stabilizer), potassium iodide (≥99.5%), potassium carbonate (≥99.0%), Pluronic F127 [poly(ethylene oxide)106–poly(propylene oxide)70–poly(ethylene oxide)106, Mw = 12,600 g mol–1], Pluronic P123 [poly(ethylene oxide)20–poly(propylene oxide)70–poly(ethylene oxide)20, Mw = 5,800 g mol–1], formalin solution (37% in H2O, with 10–15% methanol as the stabilizer), sodium hydroxide (≥98%), nitric acid (70%), 2-methyltetrahydrofuran (2-Me-THF) (anhydrous >99.0%), cyclohexanol (99%), pyridine (dried over molecular sieves, 99.8%), chromium(III) acetylacetonate (99.99%), HA (>99%), and hydrochloric acid (37%) were purchased from Sigma-Aldrich. Acetone, ethyl acetate, and tetrahydrofuran were purchased from Biosolve. Molecular sieves (3 Å) were purchased from Merck. Chloroform-d (dried over molecular sieves, D, 99.96%) was purchased from Cambridge Isotope Laboratories. The linear triblock copolymer based on SEBS with a styrene/rubber ratio of 30/70 Kraton G1652 (Kraton Cooperation) was kindly provided by Kraton Polymers Belgium. The Solsperse M387 dispersant (Lubrizol Corporation) was kindly provided by Integrated Chemicals Specialties BV. Activated carbon, Organosorb 200–1, was kindly provided by Desotec.

Synthesis

Lignin Propargylation

The propargylation of P1000 lignin (Figure ) was performed in a 4 L reactor, equipped with a reflux condenser and mechanical stirrer, under nitrogen flow. First, 200 g (1.29 mol, 1 equiv of the sum of phenolic, aliphatic OH and COOH groups) of P1000 lignin was dissolved in 2 L of acetone, then 941.2 g (6.88 mol, 5 equiv) of potassium carbonate and 113.1 g (0.68 mol, 0.5 equiv) of potassium iodide were dispersed in the lignin solution. Subsequently, the temperature was increased to 60 °C, and 759 mL (6.88 mol, 5 equiv) of 80 wt % propargyl bromide solution in toluene was added dropwise into the reaction mixture. The reaction mixture was then refluxed for 24 h. Afterward, the reaction mixture was cooled, and the product was filtered. The solid residue from the reaction mixture was washed with acetone, the filtrates were combined, and the solvent was evaporated. The solid residue was dried at 80 °C in a vacuum oven overnight. The crude product was dissolved in 500 mL of ethyl acetate and washed three times with brine. The solvent was evaporated, and the product was dried in a vacuum oven at 80 °C overnight. 141.5 g of the propargylated lignin was obtained as a brown powder (yield 64%). The propargyl moiety content was determined by measuring the conversion of the phenolic OH, aliphatic OH, and COOH by 31P NMR spectroscopy. The conversion of phenolic OH, aliphatic OH, and COOH was 96, 64, and 96%, respectively, which equals to 4.38 mmol/g of propargyl moieties. The spectra before and after modification are presented in Figure S1.

Preparation of Lignin-Based Resins and Curing via Claisen Rearrangement

Propargylated lignin can be cross-linked via Claisen rearrangement, above 150 °C.[53] Here, propargylated P1000 lignin was cured in bulk, in the presence of the corresponding surfactant to obtain the MC precursor.

3.75 g of propargylated P1000 lignin and 11.13 g of the surfactant (Pluronic P123, Pluronic F127, Kraton G1652, or Solsperse M387) in a 1:3 weight ratio of lignin/surfactant were dissolved in 60 mL of the green solvent 2-methyltetrahydrofuran to ensure proper mixing of the components. After dissolution, the solvent was evaporated, and the material was transferred to a Binder oven. The Claisen rearrangement reaction (Figure ) was carried out at 180 °C for 14 h. The samples were named by the surfactant name. The conversion was determined via FT-IR analysis (Figure S2).

Preparation of MCs

Carbonization of the cured resins was performed in a Carbolite tube furnace type TFZ 12/65/550, according to the procedure described by Libbrecht et al.(67) 2.5 g of resin was weighed into two ceramic pans, and the oven was flushed with nitrogen for 30 min. Then, there were two heating steps: the first to remove the surfactant template and the second to fully carbonize the material. During the first heating step, the sample was heated up to the template decomposition temperature, at a heating rate of 1 °C/min. When the desired temperature was reached, it was kept isothermal for 2 h. A second heating step with a heating rate of 2 °C/min was performed until the desired carbonization temperature was reached (600, 800, or 1000 °C). Then, the temperature was maintained for 3 h to ensure full carbonization of the sample. After carbonization, the material was cooled to ambient temperature at 5 °C/min. Flushing, heating, and cooling steps were performed under a constant nitrogen flow of 0.4 NL/min. The obtained material was weighed, grinded, and stored under dry conditions. The different carbons were named by surfactant name_carbonization temperature.

The decomposition temperature of the surfactants was investigated using TGA analysis. Thermal stability of the samples was determined on a TA TGA Q500 instrument in a nitrogen atmosphere, from 25 to 700 °C at a heating rate of 10 °C/min. The thermographs of the surfactants are presented in Figure S3.

Characterization

Lignin and Modified Lignin Characterization

The aromatic and aliphatic hydroxyl groups as well as the carboxylic acids (before and after modification) of lignin were determined by 31P NMR spectroscopy after sample derivatization, according to the method described by Korntner et al.(93) 10 mg of dried (in a vacuum oven at 80 °C, overnight), lignin-based material was dissolved in a 500 μL of anhydrous pyridine and deuterated chloroform mixture (1.6:1, v/v). Then, 100 μL of internal standard solution, cholesterol (19.7 mg/mL in anhydrous pyridine and deuterated chloroform mixture, 1.6/1, v/v, 0.0051 mmol), 50 μL of relaxation agent chromium (III) acetylacetonate (10 mg/mL, in anhydrous pyridine and deuterated chloroform mixture, 1.6:1, v/v), and 50 μL of derivatizing agent, 2-chloro-1,3,2-dioxanephospolane, were added to the solution. The solution was stirred for 10 min, transferred into a dry 5 mm NMR tube, and measured. The measurement was performed on a Bruker AVANCE III HD Nanobay 300 MHz apparatus (121.49 MHz for 31P NMR experiments) using the standard phosphorus pulse program, at ambient temperature, with 512 scans, relaxation delay (5 s), acquisition time (2 s), transmitter excitation frequency (140 ppm), and spectral width (396 ppm). The chemical shifts were reported in parts per million, and the raw data were processed in MestReNova software. Chemical shifts were referenced from the sharp signal of the reaction product between residual water and 2-chloro-4,4,5,5-tetramethyl-1,3,2-dioxaphospholane at 132.2 ppm.

Molecular weight of lignin and its dispersity were obtained from GPC analysis. It was performed at 30 °C using a Waters GPC equipped with a Waters 2414 RI detector. Tetrahydrofuran was used as the eluent at a flow rate of 1 mL/min. Three linear columns were used (Styragel HR1, Styragel HR4, and Styragel HR5). Lignin samples were dissolved in 1.5 mL of tetrahydrofuran, and samples were filtered with a PTFE syringe filter (pore size, 0.2 μm) prior to measurements. Molecular weights were obtained relative to polystyrene standards. Lignin was acetylated prior to the measurement, according the procedure described by Gosselink et al.(91)

Resin Characterization

IR spectra of the resins before and after curing were recorded on a PerkinElmer Frontier spectrometer, equipped with Pike technologies GladiATR accessory in the range of 600–4000 cm–1, with a resolution of 2 cm–1 using 32 accumulative scans. The conversion was determined by following the disappearance of the alkyne C≡C–H peak at 3320 cm–1, while the peak of the aromatic C=C at 1600 cm–1 was used as an internal standard (Figure S2). The conversion was determined based on the integration of the representative peaks with SpectraGryph 1.2 software.

MCs Characterization

Elemental analysis was performed to investigate the chemical composition of the carbon. The measurement was performed with a UNICUBE (Elementar Analysensysteme GmbH). For the carbon, hydrogen, nitrogen, and sulfur content, the elemental composition of the synthesized materials was determined by the analysis of the combustion products (CO2, H2O, SO2, and NO2). The oxygen content was determined by pyrolysis of the sample in a pyrolysis chamber in a separate experiment. The calibration of the instrument was performed with sulfanilamide as the model compound. Data were analyzed with EAS UNICUBE software.

The nitrogen sorption studies were performed on a Micromeritics Tristar II 3020 surface area and porosimeter analyzer at 77 K. Samples were pretreated at 300 °C to remove moisture under nitrogen flow for 6 h. The relative pressure range was measured from 2.5 × 10–3 to 0.99. The total pore volume Vtot was calculated as the amount of nitrogen adsorbed at a relative pressure of 0.95. The pore volumes of micropores Vmicro was calculated from the V–t plot. The pore volumes of mesopores and macropores, VBJH, were calculated using BJH desorption cumulative volume of pores between 17.000 and 3,000.000 Å width. The total surface area SBET was determined using the BET method. The average pore size (Dp BJH) and the PSD were calculated from the desorption branch using the BJH method. The reported data including SBET, Vtot, Vmicro, VBJH, and PSD were calculated by software MicroActive from Micromeritics.

The morphology of the synthesized materials was investigated using transmission electron microscopy (TEM). The TEM images were acquired on a JOEL JEM 2200-FS TEM with a high tension of 200 kV and 0.1 s exposure time. For the measurement, the powdered sample was placed on a copper grid (Lacey F/C 200 mesh Cu, Ted Pella Inc.).

XPS measurements were carried out using an Ultra AxisTM spectrometer (Kratos Analytical). The samples were irradiated with monoenergetic Al K1,2 radiation (1486.6 eV), and the spectra were taken at a power of 144 W (12 kV × 12 mA). The aliphatic carbon (C–C and C–H) at a binding energy of 285 eV (C 1s photoline) was used to determine the charging. The spectral resolution— the full width at half-maximum of the ester carbon from poly(ethylene terephthalate) (PET)—was better than 0.68 eV for the elemental spectra. The elemental concentration is given in weight %, but it should be considered that this method can detect all elements except hydrogen and helium. The information depth is about 10 nm.

HA Adsorption

The HA adsorption capacity of the synthesized carbons was determined using the methodology described by Libbrecht et al.(67) First, a 500 mg/L HA solution was prepared by dispersing 500 mg of HA in 1 L of deionized water. HA was dissolved by adding 0.1 M sodium hydroxide solution until it reached pH 12, and ultrasonic treatment was carried out for 30 min. After dissolution, the solution was neutralized to pH 7 with 0.1 M hydrochloric acid solution. Next, 10 mg of the adsorbent was added to 50 mL of the 500 mg/L HA solution. These mixtures were incubated in a thermostatic shaker (Infors HT Multitron Standard) at 25 °C and 200 rpm for 7 days to ensure equilibrium. After incubation, the solution was filtered through a 0.45 μm PET syringe filter, and its HA concentration was measured by UV–vis at 254 nm using a Thermo Scientific Evolution 60 UV–vis spectrometer with VISIONlight 4 software. The HA concentration of the 500 mg/L solution was determined by the same procedure (filtration and UV–vis analysis). All adsorption experiments were performed in triplicate. The adsorption capacity was calculated using the equationwhere qe (mg/g) is the HA adsorption capacity at equilibrium, C0 and Ce (mg/L) are the initial and equilibrium concentration of HA, m (mg) is the mass of the adsorbent, and V (L) is the solution volume.