Esterification of levulinic acid with ethanol catalyzed by sulfonated carbon catalysts: Promotional effects of additional functional groups

Esterification of levulinic acid with ethanol catalyzed by sulfonated carbon catalysts: Promotional effects of additional functional groups

Catalysis Today xxx (xxxx) xxx–xxx Contents lists available at ScienceDirect Catalysis Today journal homepage: www.elsevier.com/locate/cattod Ester...

844KB Sizes 18 Downloads 178 Views

Catalysis Today xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Catalysis Today journal homepage: www.elsevier.com/locate/cattod

Esterification of levulinic acid with ethanol catalyzed by sulfonated carbon catalysts: Promotional effects of additional functional groups ⁎

Isao Ogino , Yukei Suzuki, Shin R. Mukai Division of Applied Chemistry, Graduate School of Engineering, Hokkaido University, N13W8, Kita-ku, Sapporo, Hokkaido 060-8628, Japan

A R T I C L E I N F O

A B S T R A C T

Keywords: Esterification Acid catalysis Sulfonic acid Neighboring group effect Levulinic acid Biomass

Analysis of literature data on acid-catalyzed esterification reaction of levulinic acid (LA) with ethanol (EtOH) has suggested that some sulfonated carbon catalysts exhibit higher active-site performance than other solid acid catalysts such as macroreticular resins bearing sulfonic acid groups, zeolites, mesoporous silica functionalized with alkyl- and arene-sulfonic acid groups. To elucidate factors that enable the higher performance of sulfonated carbon catalysts, a series of sulfonated carbon catalysts was synthesized by sulfonating various carbon materials whose concentrations of surface oxygen-containing functional groups, porous structure, and swelling ability differ significantly. The catalysts were tested not only in the liquid-phase esterification reaction of LA with EtOH but also in the reaction of acetic acid (AcA) with EtOH because the latter reaction serves as a test reaction to probe the performance of –SO3H sites with minimal influence by mass transfer limitation and to provide an insight into a role of γ-keto group of LA in catalysis. The results show that all catalysts exhibit nearly the same turnover frequency per –SO3H site in the esterification reaction of AcA with EtOH despite widely different structural properties. In contrast, the data indicate that neighboring functional groups such as –COOH and –OH facilitate the reaction of LA with EtOH presumably through hydrogen-bonding interaction between these surface functional groups and γ-keto group of LA. These results suggest a general design strategy to improve the performance of solid acid catalysts further by precisely tuning the distance between –SO3H sites and neighboring functional groups.

1. Introduction Levulinic acid (LA) has been identified as one of the most important value-added chemicals derived from biomass [1]. Two functional groups (ketone and carboxylic acid) in LA render it as important building blocks for the production of various biomass-derived commodities [2,3]. In addition, LA has been produced in large quantity for years [2], and new processes have been recently developed [4], expanding its large-scale production. Esters of LA have been used as ingredients for flavour and fragrance and tested as an additive for transportation fuels [5–8]. Ethyl levulinate, for example, can be produced by esterification of LA with ethanol (EtOH) or by ethanolysis of furfuryl alcohol [9]. Like many other liquidphase esterification reactions [10–13], esterification of LA with shortchain alcohols is catalyzed by both soluble and solid acid catalysts [5,6,8]. Mineral acids such as H2SO4 are cost-effective catalysts to produce ethyl levulinate via esterification reaction. However, solid acid catalysts provide advantages of ease of separation from reaction products and lack of corrosion of reactor materials. Therefore, various solid



acid catalysts including zeolites [14], resin catalysts [15], supported heteropoly acids [16], sulfonated ZrO2 [17] or SnO2 [14], organosilica or mesoporous silica incorporating –SO3H groups [18,19], and sulfonated carbons [20–22] have been investigated as a potential catalyst for this conversion. Analysis of literature data on acid-catalyzed esterification of LA with short-chain alcohols has suggested that some sulfonated carbon catalysts [20,21] exhibit higher performance than other solid acid catalysts such as Amberlyst-15 (ion-exchange resin consisting of macroreticular polystyrene with strongly acidic sulfonic group) [14,18], SAC-13 (SiO2-supported Nafion) [18], mesoporous silica (SBA-15) functionalized with alkyl sulfonic acid groups [18], zeolite Beta [14,23], and HY zeolite [14] when initial rates were calculated and compared per strong acid site quantified by NH3 temperature-programmed desorption (TPD), base titration using NaOH aq. or other methods reported in the literature and summarized in Table S1 in the Supplementary information. Only limited types of acid catalysts (soluble p-toluenesulfonic acid (p-TSA) [8] and organosilica nanotubes functionalized with both heteropoly acid and ZrO2 [16]) appear to

Corresponding author. E-mail address: [email protected] (I. Ogino).

http://dx.doi.org/10.1016/j.cattod.2017.10.001 Received 1 August 2017; Received in revised form 28 September 2017; Accepted 2 October 2017 0920-5861/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Ogino, I., Catalysis Today (2017), http://dx.doi.org/10.1016/j.cattod.2017.10.001

Catalysis Today xxx (xxxx) xxx–xxx

I. Ogino et al.

2.2. Syntheses of catalysts

exhibit activity comparable to the sulfonated carbon catalysts. This analysis has raised questions about why some sulfonated carbon catalysts seem to perform better than others, and about whether there is a specific carbon catalyst performs better than others. In this work, we chose to investigate carbon-based catalysts for this reaction, aimed to address the second question, and thereby obtain a design principle to improve catalyst performance for this reaction further. Carbon-based catalysts potentially provide advantages because their porous structure and surface properties can be tailored by choosing the type of carbon materials and synthesis conditions of carbon precursors. In our previous work with phenolic-resin-derived carbon catalysts [24], we demonstrated that surface hydrophobicity and mesoporosity of catalysts can be tailored by synthesis conditions of precursor resins and their pyrolyzation temperature. In addition, we showed how these factors influence the performance of catalysts in three different acid-catalyzed reactions. In the current work, the type of carbon materials has been expanded to traditional carbon catalysts based on activated carbon and carbon nanotubes as well as graphite oxides and carbohydrate-based catalysts, which have become more common recently [13,20,25]. Challenges to achieve this goal include complex interplays of effects of porous structures, surface hydrophobicity/hydrophilicity, and swelling ability of some catalysts [25] that is also common in ion-exchange resin catalysts but challenging to evaluate its effect on overall performance [26]. Thus, we sought a way to resolve these effects by synthesizing a set of sulfonated carbon catalysts from various carbon sources, and conducting a test esterification reaction of acetic acid (AcA) with EtOH in addition to that of LA with EtOH (Scheme 1). We chose the reaction of AcA with EtOH because it can be considered as prototypical for esterification reaction, and the reactants are small enough to minimize effects of mass transfer limitations (kinetic diameter of AcA is reported to be 0.436 nm [27]). In addition, by comparing the data for the reaction of LA with that of AcA, we anticipated that we can investigate the effect of γ-keto group of LA on catalysis. We compare catalyst performance by apparent initial turnover frequency (TOF) rather than rate per catalyst mass to compare active-site performance and to minimize effects of product water on –SO3H groups [28,29]. We report our findings that additional surface functional groups such as –COOH and –OH groups in some carbon catalysts (carbonaceous catalysts) facilitate the reaction of LA with EtOH.

Catalysts were synthesized by sulfonating carbon materials using sulfuric acids under N2 atmosphere. Graphite oxides (GO) were synthesized by oxidizing natural graphite powder (z-5F) by Hummers method [30]. Pyrolyzation of glucose and cellulose was conducted by heating 1.0 g of D-glucose and cellulose powder, respectively, in 100 mL min−1 N2 to 673 K at a ramp rate of 5 K min−1 and holding the temperature at 673 K for 15 h. The resultant materials are designated as Glu and Cel, respectively. Hydrothermal treatment of glucose was conducted by heating a D-glucose solution (1 g/mL) in a sealed 23-mL Teflon-lined autoclave at 453 K for 24 h under static conditions. The resultant material is designated as HTGlu. Four different carbon gels (CG) were synthesized by pyrolyzing two different resorcinol-formaldehyde resins at different temperatures [24,31]. The resins were synthesized at a molar ratio of resorcinol (R) and sodium carbonate (C) of 50 and 200, and pyrolyzed at 673 or 1273 K. The resultant materials are designated as CG-x-y where x and y represent a R/C ratio and pyrolyzation temperature, respectively. CGs were sulfonated as reported previously [24]. Activated carbon (AC), multi-walled carbon nanotubes (CNT), GO, Glu, Cel, and HTGlu were sulfonated by heating them in concentrated H2SO4 at 353 K for 15 h, followed by deep washing with distilled water. Sulfonated materials are designated by adding S at the beginning of each name. For example, sulfonated AC and CNT are designated as SAC and SCNT, respectively. 2.3. Characterization of catalysts Surface functional groups of the carbon catalysts were characterized by IR spectroscopy. Measurements were performed in a transmission mode under dynamic vacuum using a JASCO FT/IR-6100 Fourier transform spectrometer with a spectral resolution of 4 cm−1. Acidic functional groups of the carbon catalysts were quantified by Boehm titration [32]. The standardization of NaOH solutions was performed using potassium hydrogen phthalate as the primary standard and phenolphthalein as the indicator. The carbon and oxygen concentrations in the catalysts were determined through CHN analysis (CHN: MICRO CORDER JM10, J-SCIENCE LAB Co.) while their sulfur concentrations were determined by ion chromatography analysis (Dionex ICS1600, Thermo Fisher Scientific Inc.) both at the Global Facility Center of the Creative Research Institution at Hokkaido University. The carbon and oxygen concentrations obtained from the elemental analysis were corrected after the moisture content of carbon catalysts was determined by thermogravimetric analysis (TGA), which assumed that the weight loss below 373 K arises from desorption of physisorbed water. Powder X-ray diffraction of carbon materials were measured on a Rigaku RINT Ultima IV with Cu Kα radiation (λ = 1.5418 Å) and a D/teX Ultra detector. Data were collected in a continuous mode over 5 ≤ 2θ ≤ 60° in 0.05° step width with a scan speed of 0.16° s−1. TG measurements were conducted on a Shimadzu TGA-50 thermogravimetric analyzer by heating approximately 10 mg of a catalyst in a platinum crucible to 1073 K at 10 K min−1 in a 20 cm3 min−1 N2 flow. The textural property of catalysts was characterized through nitrogen adsorption measurements. The adsorption isotherms were collected at 77 K on an adsorption apparatus BELSORP-mini II (MicrotracBEL Co.). Prior to analysis, samples were heated at 523 K in a 30-mL min−1 N2 flow for 4 h. The surface areas of the catalysts were calculated by the Brunauer-EmmettTeller (BET) method [33] in a relative pressure range of 0.05–0.3. The micropore volume (Vmicro) of catalysts was calculated from N2 uptake at P/P0 = 0.15. The mesopore volume (Vmeso) of catalysts was calculated by subtracting Vmicro from the total volume calculated from N2 uptake at P/P0 = 0.98. Mesopore size distributions were determined by applying the Dollimore-Heal method [34] to the adsorption branch of isotherms. To evaluate hydrophilicity of catalyst surface, water vapor adsorption isotherms were measured at 298 K on an adsorption apparatus

2. Experimental 2.1. Materials Levulinic acid (97%), acetic acid (99.7%), ethanol (99.5%), D-glucose, cellulose, 0.01 N NaOH aq., and 0.01 N HCl aq. were purchased from Wako Pure Chemical Industries Ltd. Resorcinol (99.0%), formaldehyde (36.0 wt% aq.), sodium carbonate (99.8%), 1 N HCl aq., tert-butyl alcohol (99.0%), and sulfuric acid (98%) were purchased from Tokyo Chemical Industry Co. Ltd. Activated carbon (Norit GAC 1240W), Amberlyst-15 (hydrogen form), and multi-walled carbon nanotubes (SMW210, Sigma-Aldrich, BET surface area = 350 m2 g−1) were purchased from Sigma-Aldrich Japan. Natural graphite powder (z5F) was received from Ito Graphite Industry Co. Ltd.

Scheme 1. Esterification reactions tested in this work.

2

Catalysis Today xxx (xxxx) xxx–xxx

I. Ogino et al.

calculated O/C atomic ratio of the synthesized catalysts ranges widely from 0.045 to 0.53 (Table 1). The data show relatively low values for SAC and SCNT while the data show high values for SGlu, SCel, and SHTGlu. As-prepared graphite oxides typically have an O/C ratio ∼0.42 [35]. However, sulfonation resulted in a reduction of the O/C ratio to 0.19. Although sulfonation introduces sulfonic acid groups and consequently oxygen atoms in GO, it also causes acid-catalyzed dehydration reactions, which reduce the oxygen atom content. Overall, the O/C ratio decreased after the sulfonation treatment. A similar result has been reported in the literature [36]. The data show that SCG catalysts have moderate concentrations of oxygen atoms with O/C atomic ratios ranging from 0.15 to 0.25. SCG catalysts obtained through high temperature pyrolyzation (SCG-50-1273 and SCG-200-1273) show lower O/C ratios than those obtained through low temperature pyrolyzation (SCG-50-673 and SCG-200-673) because higher fractions of oxygencontaining functional groups had been removed via higher temperature pyrolyzation.

BELSORP-Max (MicrotracBEL Co.). Prior to analysis, samples were heated in vacuum at 523 K for 4 h. To compare the results for carbon catalysts possessing significantly different porous structure, the volume of water vapor uptake at a relative pressure of 0.15 (VH2O) was normalized with the volume of nitrogen uptake at the same relative pressure (VN2). The normalized value is designated as x0.15.

x 0.15 =

VH2O × 100 [%] VN2

(1)

Microstructure of the carbon catalysts was characterized by Raman spectroscopy. The data were collected on an inVia Reflex Raman microscope (Renishaw) using 532 nm radiation at the Open Facility of the Creative Research Institution at Hokkaido University. 2.4. Reaction experiments All liquid-phase esterification reactions were carried out in a 3.5-mL glass vial immersed in a water bath on an EYELA RCH-20L hot plate stirrer equipped with an ETS-D5 temperature controller. In a typical reaction, the reactor was charged with an equimolar mixture of AcA and EtOH or a 1:3 ratio mixture of LA and EtOH, an internal standard, sealed and heated to 333 K while stirring at 600 rpm. Shortly after the temperature of the mixture reached the desired value, the specified amount of a catalyst was added to the reaction mixture and the timer was started. The mass loading of catalyst was adjusted to run reactions with identical molar ratio of −SO3H sites and carboxylic acid reactant (3 × 10−3 mol-(–SO3H)/mol-AcA or LA). Before use in reactions, all catalysts were ground (particle size < 50 μm) and dried at 363 K for 2 h. Sample aliquots (30 μL) were withdrawn from the reaction mixture using a micropipette and analyzed using a Shimadzu GC-17A gas chromatograph equipped with a capillary column (Inert WAX-HT, GL Sciences, 30 m × 0. 25 mm) and an FID detector. Under these reaction conditions, the external mass transfer resistance is negligible. Apparent initial turnover frequency (per −SO3H group) was calculated from the initial conversion data.

3.1.2. Surface functional groups characterized by IR spectroscopy and Boehm titration To identify oxygen-containing functional groups in the synthesized catalysts, IR spectroscopy characterization was conducted. The data characterizing SGO, SCel, SHTGlu, SCG-50-673 and SCG-200-673 (Fig. 1A) show the presence of –COOH, –OH and –SO3H groups: ∼1730 cm−1 for ν(C=O) and ν(COOH) [37], ∼1550–1650 cm−1 for ν(C=C) [37] presumably overlapped with a band for the bending mode of physisorbed water, ∼1230 cm−1 for ν(C–O) of C–OH groups [37] or epoxide groups, and 1030 cm−1 for ν(SO3−) [20]. The presence of –SO3H groups was confirmed further by XPS analysis for SCG catalysts [24]. SAC, SCNT, SCG-50-1273 and SCG-200-1273 absorbed too much incoming IR light to allow for meaningful analysis presumably because of their aromaticity. To quantify the acidic functional groups, Boehm titration method was used [32,38]. The technique uses several bases with different basicity (NaOH, Na2CO3, and NaHCO3) to titrate acidic functional groups with different pKa values. The titration data are complemented by elemental analysis of sulfur to quantify −SO3H in addition to –COOH (and lactone), and phenolic –OH groups. The results show that SGO, SCel, SGlu, and SHTGlu contain high concentrations of –COOH groups (≥1.3 mmol g−1) (Table 1). In addition, SHTGlu and SGO have high concentrations of –OH groups. The total concentration of oxygen atoms determined by elemental analysis is compared with that arising from acidic functional groups (–SO3H, –COOH, and –OH) that had been determined by titration and

3. Results and discussion 3.1. Characterization of synthesized catalysts 3.1.1. Oxygen/Carbon atomic ratio To quantify the total concentrations of oxygen atoms in the synthesized catalysts, elemental analysis was conducted and the results were corrected for water content by TGA experiments (Table 1). The Table 1 Physicochemical properties of sulfonated carbon catalysts. catalyst

SAC SCNT SGO SGlu SCel SHTGlu SCG-50-673f SCG-200-673f SCG-50-1273f SCG-200-1273f

O/C ratioa [–]

0.12 0.045 0.19 0.37 0.27 0.53 0.23 0.25 0.15 0.17

concentration of acid groupsb[mmol g−1] –SO3H

–COOH

–OH

0.50 0.14 0.24 0.53 0.50 0.45 0.97 0.88 0.79 0.62

0.30 0.03 1.4 1.3 1.5 1.4 0.49 0.74 0.20 0.48

0.08 0.11 1.1 0.44 0.11 1.4 0.82 0.31 0.90 0.64

a

SBET [m2 g−1]

Vmicroc [cm3 g−1]

Vmesod [cm3 g−1]

x0.15e [%]

947 251 77 0.81 2.6 7.9 618 658 596 570

0.41 0.10 0.010 < 0.001 0.001 0.001 0.25 0.27 0.25 0.23

0.090 1.97 0.056 < 0.001 0.008 0.005 0.20 0.46 0.31 1.5

8.2 12 2500 32000 14000 20000 58 55 18 30

Determined by CHN analysis. Determined by Boehm titration and ion chromatography analysis of sulfur. c Micropore volume determined by N2 uptake at P/P0 = 0.15. d Mesopore volume determined by subtracting Vmicro from the pore volume determined by N2 uptake at P/P0 = 0.98. e Ratio of water vapor uptake at (298 K) and nitrogen uptake (77 K) at P/P0 = 0.15. f Ref. [24]. b

3

Catalysis Today xxx (xxxx) xxx–xxx

I. Ogino et al.

Fig. 1. (A) FT-IR spectra characterizing sulfonated carbon catalysts. The light red, green, blue, and purple color rectangles indicate frequency regions corresponding to ν(C=O, COOH), ν(C=C) + the bending mode of physisorbed water, ν(C–O) of C–OH groups or epoxide groups, and ν(SO3−), respectively. (B) Concentrations of acidic functional groups and total oxygen atoms. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.1.3. Microstructure characterized by Raman spectroscopy Because most of the synthesized catalysts except SAC and SCNT are amorphous in nature as characterized by PXRD (Fig. S1 in the Supplementary information), the microstructure (defects and size of cluster of aromatic rings) of the synthesized catalysts was characterized by Raman spectroscopy. The data for all catalysts show two broad peaks at ≈1350 cm−1 (D-peak) and ≈1580 cm−1 (G-peak) with a shoulder at ≈1620 cm−1 (D’ band) (Fig. 2). The D-band arises from aromatic sixmembered rings with its intensity being proportional to the number of aromatic rings [39]. However, it is only active in the presence of defects such as functional groups and lattice defects. The D’ band is also associated with the presence of defects [39]. On the other hand, the Gband arises from any sp2 carbon pairs. SCG-200-673, SGlu, SCel, and SHTGlu show lower ID and broader G-peaks than other catalysts, indicating less degrees of clustering of aromatic rings in them with highly disordered microstructure. SCG-200-1273 shows a higher ID relative to IG than SCG-200-673 probably because the higher temperature pyrolyzation resulted in a greater degree of clustering of aromatic rings. Fig. 2. Raman spectra characterizing the selected sulfonated carbon catalysts.

3.1.4. Textural property and surface hydrophilicity/hydrophobicity characterized by N2 and H2O vapor adsorption Textural property and surface hydrophobicity of catalysts were characterized by N2 at 77 K and H2O vapor adsorption at 298 K, respectively (Fig. 3 and Table 1). SAC is a microporous material as shown by its type I isotherm, and has a high BET surface area. On the other hand, SCNT is a micro-mesoporous material with a high Vmeso presumably associated with void spaces between aggregated nanotubes. SGO exhibits a low BET surface area because of partial restacking of oxidized graphene sheets upon pyrolyzation. Catalysts derived from carbohydrates (SGlu, SHTGlu, and SCel) through the current approach are essentially non-porous in dry form like some catalysts reported in the literature [20,40,41]. All of the SCG catalysts are micro-mesoporous

elemental analysis (Fig. 1B). In the calculation of the concentration of oxygen atoms arising from the acidic functional groups. the number of oxygen atoms per functional group (3 for –SO3H and 2 for –COOH) was taken into consideration. The results show that the concentration of total oxygen atoms is about 5 times higher than the concentration of oxygen atoms arising from acidic functional groups (Fig. 1B), indicating the presence of non-acidic oxygen-containing functional groups. In particular, SHTGlu and SCel contain significant concentrations of such groups (> 20 wt%).

Fig. 3. (A) N2 adsorption isotherms and (B) H2O vapor adsorption isotherms characterizing sulfonated carbon catalysts: ( , ) SAC; ( , ) SCNT; ( , ) SGO; ( , ) SGlu; ( , ) SCel; ( , ) SHTGlu; ( , ) SCG-50-673; ( , ) SCG-200-673; ( , ) SCG-501273; ( , ) SCG-200-1273. The open and closed symbols represent adsorption and desorption branches, respectively.

4

Catalysis Today xxx (xxxx) xxx–xxx

I. Ogino et al.

Fig. 4. (A) O/C atomic ratios and (B) x0.15 of sulfonated carbon catalysts. (C) apparent TOF for the esterification of AcA with EtOH. Conditions: AcA (25 mmol), EtOH (25 mmol), catalyst −SO3H sites (75 μmol); reaction temperature = 333 K. (D) apparent TOF for the esterification of LA with EtOH. Conditions: LA (25 mmol), EtOH (75 mmol), catalyst −SO3H sites (75 μmol); reaction temperature = 333 K. SCG-1, -2, -3, and -4 represent SCG-50-673, -200-673, -50-1273, -2001273, respectively.

having moderate BET surface areas ≈600 m2 g−1 and Vmicro ≈0.25 cm3 g−1, but show a variation of Vmeso, ranging from 0.20 to 1.5 cm3 g−1 because different R/C ratios in the resin synthesis result in different sizes of RF resin particles and void spaces between them

[24,42]. SCG catalysts prepared with a higher R/C ratio in the precursor resin synthesis but the same pyrolyzation temperature show a larger Vmeso value. The H2O vapor adsorption data show that SAC as well as resin-based 5

Catalysis Today xxx (xxxx) xxx–xxx

I. Ogino et al.

Fig. 5. Arrhenius plots for selected catalysts: ( ) SCG-200-673, (

resin-based catalysts pyrolyzed at 673 K (SCG-50-673 and -200-673) show type II isotherms, indicating favorable interaction between water molecules and their surfaces; the surfaces of these catalysts are relatively hydrophilic, consistent with the high concentrations of acidic functional groups in these catalysts (Fig. 1 and Table 1). To compare the degree of surface hydrophobicity of catalysts having different porosity, the water adsorption data were normalized by the nitrogen adsorption data. The percentage of the uptake of water vapor at 298 K with respect to the uptake of nitrogen at 77 K at their relative pressures of 0.15 (denoted as x0.15 [%]) serves as a good metric to determine the degree of surface hydrophobicity [24]; a low x0.15 value indicates that the surface is hydrophobic. Such methods have been used to characterize surface hydrophobicity of crystalline microporous materials that have well-defined microporous structures [43], but found to be applicable to sulfonated carbon catalysts [24]. SAC and SCNT show very low x0.15 values (Fig. 4), indicating high surface hydrophobicity presumably because of low concentrations of oxygen-containing functional groups (Fig. 1 and Table 1). On the other hand, SCG catalysts show a variation of x0.15 values, depending primarily on their pyrolyzation temperature, and the catalysts derived from 1273 K pyrolyzation (SCG-50-1273 and -200-1273) have relatively hydrophobic surfaces. All of these catalysts exhibit x0.15 values less than 100%, indicating partial filling of their micropores with water molecules. In sharp contrast, the catalysts derived from SGO, SGlu, SCel, and SHTGlu show x0.15 values exceeding 2000%. These results indicate that the catalysts swell significantly in the presence of water vapor. Carbohydrate-based catalysts like SGlu, SCel, and SHTGlu are known to swell [40], but despite the relative low O/C ratio (Figs. 1 and 4), SGO was also found to swell to a similar degree in the current work. The x0.15 value mostly shows the same trend as the concentration oxygen atoms in catalysts represented by O/C ratio (Fig. 4). However, catalysts with a similar O/C ratio (i.e. SGO vs. SCG-200-1273 (SCG-4), SCel vs. SCG-200-673 (SCG-2)) show significantly different degrees of swelling, suggesting other factors like a degree of crosslinking between aromatic domains might influence their swelling ability as is the case for ion-exchange resins [44,45]

) SGlu, ( ) SCNT, and

( ) SAC. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.2. Liquid-phase esterification reactions The results from the reaction of AcA with EtOH show that the catalysts exhibit nearly the same apparent TOF (≈0.02 s−1) per −SO3H groups despite their different properties (Fig. 4C). In addition, these TOF values are similar to those reported for several sulfonic-acid containing carbon catalysts (Table S1 in the Supplementary information) [46–48], considering the difference in the reaction temperature used in these works. Our previous work shows that when various fractions of –SO3H groups in SCG-50-673, which has smallest mesopore volume and average size in all SCG catalysts, were poisoned with Na+ and tested in the esterification reaction, the rate decreased linearly as the degree of poisoning was increased [24]. In addition, the rate became nearly zero when the data were extrapolated to 100% degree of poisoning [24]. Therefore, we inferred that −SO3H groups function as a single-site and all of them can be accessed by reactants with negligible mass transfer limitations. Although SGO, SGlu, SCel, and SHTGlu show low surface areas in dry form (Table 1), they exhibit similar activity to SCG catalysts. Therefore, we infer that their high swelling ability in the highly polar reaction mixture renders access of reactants to their −SO3H groups inside particles easily. The nearly the same TOF values of all catalysts indicate that −SO3H groups in them function similarly toward the esterification reaction under the present conditions, and neighboring functional groups such as –COOH and –OH groups barely affect the performance of −SO3H groups. Esterification of carboxylic acids by soluble acid catalysts generally proceeds via protonation of carboxylic acids, followed by nucleophilic attack by alcohols. It is often invoked that esterification on solid acid catalysts bearing Brønsted acid sites proceeds via a similar mechanism

Fig. 6. Effects of concentrations of additional oxygen-containing functional groups on rate of the esterification of levulinic acid with ethanol: ( ) SAC; ( ) SCNT; ( ) SGO; ( ) SGlu ( ) SCel; ( ( ) SCG-200-1273.

) SHTGlu; ( ) SCG-50-673; ( ) SCG-200-673; ( ) SCG-50-1273;

Scheme 2. A proposed mechanism for the rate enhancement by proximity effect enabled by interaction of γ-keto group of levulinic acid with the surface functional group.

catalysts pyrolyzed at 1273 K (SCG-50-1273 and -200-1273) show a type III isotherm (Fig. 3B), and on the other hand, SCNT shows low uptake of water vapor, indicating that their surfaces are relatively hydrophobic. On the other hand, SGO SGlu, SCel, and SHTGlu as well as 6

Catalysis Today xxx (xxxx) xxx–xxx

I. Ogino et al.

Acknowledgments

[16,49,50]. Therefore, the esterification of LA with EtOH was anticipated to proceed via the same mechanism as that of AcA with EtOH. However, the results show a variation of apparent TOFs (Fig. 4D). Despite the high internal and external porosity of SAC and SCNT, respectively, they exhibit significantly lower TOF values than SGO, SGlu, SCel, and SHTGlu. Although SCG-50-673 (denoted as SCG-1 in Fig. 4D) show less mesopore volume than SCG-50-1273 (SCG-3), SCG-50-673 shows substantially higher TOF than SCG-50-1273. In addition, SCG200-673 (SCG-2) shows slightly higher TOF than SCG-200-1273. These results suggest that high concentrations of surface functional groups cause promotional effects on rate of the reaction. To investigate the cause of the promotional effect on rate further, the intraparticle mass transfer limitations were evaluated using the Weisz-Prater criterion [51] for SCG-50-1273 (SCG-3), which showed the lowest TOF (Fig. 4D). For this analysis, first, the binary diffusion coefficients of LA in EtOH was estimated using the Wilke-Chang equation [52] as described in the Supplementary Information. Then, the effective diffusivity of LA was estimated in using the equation reported by Beck and Schulz [53] with the hydrated molecular dimension of LA (0.57 nm [54]). Finally, the Weisz-Prater criterion was evaluated. The result confirmed the absence of intraparticle mass transfer limitations. Although SAC, SCNT, SGlu, and SCG-200-673 (SCG-2) show different TOFs, apparent activation energies for these catalysts are all ≈47 kJ mol−1 (Fig. 5), which compares well with that reported for the esterification of LA with n-butanol catalyzed by sulfuric acid (54 kJ mol−1) [55]. Thus, we infer that the difference in the catalytic activity is caused by the difference in the pre-exponential factor. When the apparent TOF values are plotted against a sum of concentrations of –COOH and –OH groups (Fig. 6), the plot shows a positive trend between them, indicating that the esterification rate is enhanced by the presence of additional functional groups. Such rate enhancement was not observed for the esterification of AcA with EtOH. Therefore, considering all results, we propose that interaction between γ-keto group of levulinic acid with the additional surface functional groups (–COOH and –OH) increases the sticking coefficient of reacting molecules and thereby facilitates the reaction (Scheme 2).

The authors thank Shunpei Takahashi, Daiki Andoh, and Dr. Shinichiroh Iwamura for their technical assistance in catalyst synthesis and reaction experiments. The authors thank the staff members at the Joint-use facilities at Hokkaido university supported by “Nanotechnology Platform” Program of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. The authors thank the staff members at the Open Facility of Hokkaido University for their supports in characterizing the materials. The authors thank Ito Graphite Industry for providing the graphite sample. This work was supported in part by JSPS KAKENHI Grant number 26420774. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cattod.2017.10.001. References [1] J.J. Bozell, G.R. Petersen, Green Chem. 12 (2010) 539–554. [2] J.J. Bozell, L. Moens, D.C. Elliott, Y. Wang, G.G. Neuenscwander, S.W. Fitzpatrick, R.J. Bilski, J.L. Jarnefeld, Resour. Conserv. Recycl. 28 (2000) 227–239. [3] F.D. Pileidis, M.M. Titirici, ChemSusChem 9 (2016) 562–582. [4] A. Scott, Chem. Eng. News 94 (2016) 18–19. [5] A. Démolis, N. Essayem, F. Rataboul, ACS Sustain. Chem. Eng. 2 (2014) 1338–1352. [6] E. Ahmad, M.I. Alam, K.K. Pant, M.A. Haider, Green Chem. 18 (2016) 4804–4823. [7] D.J. Hayes, Catal. Today 145 (2009) 138–151. [8] F.G. Cirujano, A. Corma, F.X. Llabrés i Xamena, Chem. Eng. Sci. 124 (2015) 52–60. [9] P. Neves, S. Lima, M. Pillinger, S.M. Rocha, J. Rocha, A.A. Valente, Catal. Today 218–219 (2013) 76–84. [10] J.C. Manayil, V.C. dos Santos, F.C. Jentoft, M. Granollers Mesa, A.F. Lee, K. Wilson, ChemCatChem 9 (2017) 1–9. [11] A. Osatiashtiani, B. Puértolas, C.C.S. Oliveira, J.C. Manayil, B. Barbero, M. Isaacs, C. Michailof, E. Heracleous, J. Pérez-Ramírez, A.F. Lee, K. Wilson, Biomass Convers. Biorefin. (2017) 1–12. [12] C. Pirez, A.F. Lee, C. Jones, K. Wilson, Catal. Today 234 (2014) 167–173. [13] L. Peng, A. Philippaerts, X. Ke, J. Van Noyen, F. De Clippel, G. Van Tendeloo, P.A. Jacobs, B.F. Sels, Catal. Today 150 (2010) 140–146. [14] D.R. Fernandes, A.S. Rocha, E.F. Mai, C.J.A. Mota, V. Teixeira da Silva, Appl. Catal. A 425–426 (2012) 199–204. [15] M.A. Tejero, E. Ramírez, C. Fité, J. Tejero, F. Cunill, Appl. Catal. A 517 (2016) 56–66. [16] D. Song, S. An, Y. Sun, Y. Guo, J. Catal. 333 (2016) 184–199. [17] Y. Kuwahara, W. Kaburagi, K. Nemoto, T. Fujitani, Appl. Catal. A 476 (2014) 186–196. [18] J.A. Melero, G. Morales, J. Iglesias, M. Paniagua, B. Hernández, S. Penedo, Appl. Catal. A 466 (2013) 116–122. [19] S. An, D. Song, B. Lu, X. Yang, Y.H. Guo, Chem.—Eur. J. 21 (2015) 10786–10798. [20] F.D. Pileidis, M. Tabassum, S. Coutts, M.-M. Titirici, Chin. J. Catal. 35 (2014) 929–936. [21] B.L. Oliveira, V. Teixeira da Silva, Catal. Today 234 (2014) 257–263. [22] V.L. Budarin, J.H. Clark, R. Luque, D.J. Macquarrie, Chem. Commun. (2007) 634–636. [23] C.R. Patil, P.S. Niphadkar, V.V. Bokade, P.N. Joshi, Catal. Commun. 43 (2014) 188–191. [24] I. Ogino, Y. Suzuki, S.R. Mukai, ACS Catal. 5 (2015) 4951–4958. [25] S. Suganuma, K. Nakajima, M. Kitano, D. Yamaguchi, H. Kato, S. Hayashi, M. Hara, J. Am. Chem. Soc. 130 (2008) 12787–12793. [26] H.W. Heath, B.C. Gates, AlChE J. 18 (1972) 321–326. [27] T.C. Bowen, R.D. Noble, J.L. Falconer, J. Membr. Sci. 245 (2004) 1–33. [28] B.C. Gates, W. Rodrigue, J. Catal. 31 (1973) 27–31. [29] Y. Liu, E. Lotero, J.G. Goodwin, J. Mol. Catal. A: Chem. 245 (2006) 132–140. [30] W.S. Hummers, R.E. Offeman, J. Am. Chem. Soc. 80 (1958) 1339. [31] R.W. Pekala, J.C. Farmer, C.T. Alviso, T.D. Tran, S.T. Mayer, J.M. Miller, B. Dunn, J. Non-Cryst. Solids 225 (1998) 74–80. [32] H.P. Boehm, Carbon 32 (1994) 759–769. [33] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319. [34] D. Dollimore, G.R. Heal, J. Colloid Interface Sci. 33 (1970) 508–519. [35] S. Gambhir, R. Jalili, D.L. Officer, G.G. Wallace, NPG Asia Mater. 7 (2015) e186. [36] W. Gao, L.B. Alemany, L. Ci, P.M. Ajayan, Nat. Chem. 1 (2009) 403–408. [37] A. Stein, Z.Y. Wang, M.A. Fierke, Adv. Mater. 21 (2009) 265–293. [38] H.P. Boehm, Carbon 40 (2002) 145–149. [39] A.C. Ferrari, J. Robertson, Phys. Rev. B 61 (2000) 14095–14107. [40] M. Okamura, A. Takagaki, M. Toda, J.N. Kondo, K. Domen, T. Tatsumi, M. Hara, S. Hayashi, Chem. Mater. 18 (2006) 3039–3045. [41] X. Mo, D.E. Lopez, K. Suwannakarn, Y. Liu, E. Lotero, J.G. Goodwin, C.Q. Lu, J. Catal. 254 (2008) 332–338. [42] S.A. Al-Muhtaseb, J.A. Ritter, Adv. Mater. 15 (2003) 101–114. [43] R. Gounder, M.E. Davis, AlChE J. 59 (2013) 3349–3358.

4. Conclusions The results shown here highlight the importance of choosing and conducting a test reaction (esterification of AcA with EtOH) that disentangles various factors affecting the total performance of catalysts especially when catalyst structure is complex and its porous structure changes in reaction media. Characterization of synthesized sulfonated carbon catalysts confirm that their porous structure, concentrations of surface functional groups, and swelling ability differ significantly. Catalysts derived from GO and carbohydrates show low porosity in dry form but they swell significantly in the presence of water vapor. The high swelling ability of these catalysts apparently allows facile mass transport of polar and small reactants to sulfonic acid groups initially embedded inside catalyst particles. All catalysts exhibited similar TOF in the reaction of AcA with EtOH under the present conditions. These results provide the basis for analyzing and interpreting the data for more technologically important reaction, esterification of LA with EtOH, and show the rate-enhancement effect by additional functional groups (–COOH and –OH groups) presumably through favourable interaction of LA with these groups. These results suggest an explanation about why some sulfonated carbon catalysts with high concentrations of oxygen-containing functional groups perform better than other solid acid catalysts bearing –SO3H groups. Thus, these results suggest a design strategy by tuning the distance between these additional functional groups and –SO3H groups more precisely to achieve further improvement of the performance of catalysts for this conversion. We anticipate that application of such design strategy is not limited to sulfonated carbon catalysts but applicable to solid acid catalysts in general. 7

Catalysis Today xxx (xxxx) xxx–xxx

I. Ogino et al. [44] A. Chakrabarti, M.M. Sharma, React. Polym. 20 (1993) 1–45. [45] B.C. Gates, E. Gerhard, K. Helmut, S. Ferdi, W. Jens (Eds.), Handbook of Heterogeneous Catalysis, Wiley-VCH, Weinheim, 2008p. 278. [46] M. Hara, T. Yoshida, A. Takagaki, T. Takata, J.N. Kondo, S. Hayashi, K. Domen, Angew. Chem. Int. Ed. 43 (2004) 2955–2958. [47] H. Yu, Y. Jin, Z. Li, F. Peng, H. Wang, J. Solid State Chem. 181 (2008) 432–438. [48] R.P. Rocha, M.F.R. Pereira, J.L. Figueiredo, Catal. Today 218–219 (2013) 51–56. [49] E. Lotero, Y. Liu, D.E. Lopez, K. Suwannakarn, D.A. Bruce, J.G. Goodwin, Ind. Eng.

[50] [51] [52] [53] [54] [55]

8

Chem. Res. 44 (2005) 5353–5363. S. Miao, B.H. Shanks, J. Catal. 279 (2011) 136–143. P.B. Weisz, C.D. Prater, Adv. Catal. 6 (1954) 143–196. C.R. Wilke, P. Chang, AlChE J. 1 (1955) 264–270. R.E. Beck, J.S. Schultz, Science 170 (1970) 1302–1305. K. Lourvanij, G.L. Rorrer, J. Chem. Technol. Biotechnol. 69 (1997) 35–44. H.J. Bart, J. Reidetschlager, K. Schatka, A. Lehmann, Ind. Eng. Chem. Res. 33 (1994) 21–25.