Hydrolytic Cleavage of β-O-4 Ether Bonds of Lignin Model

6 downloads 0 Views 542KB Size Report
Dec 15, 2010 - C-O bonds of R- and β-aryl alkyl ethers.16 The β-O-4 ether linkage is found to be ... acid is needed in their system.22 Recently, an acidic ionic liquid, ..... GG was completely converted in the presence of AlCl3; however, GG.
Ind. Eng. Chem. Res. 2011, 50, 849–855

849

Hydrolytic Cleavage of β-O-4 Ether Bonds of Lignin Model Compounds in an Ionic Liquid with Metal Chlorides Songyan Jia,†,‡ Blair J. Cox,‡ Xinwen Guo,† Z. Conrad Zhang,§ and John G. Ekerdt*,‡ State Key Laboratory of Fine Chemicals, Department of Catalysis Chemistry and Engineering, Dalian UniVersity of Technology, Dalian, Liaoning 116024, People’s Republic of China, Department of Chemical Engineering, The UniVersity of Texas at Austin, Austin, Texas 78712, United States, and KiOR Inc., 13001 Bay Park Road, Pasadena, Texas 77507, United States

The hydrolytic cleavage of β-O-4 ether bonds in lignin model compounds, guaiacylglycerol-β-guaiacyl ether (GG) and veratrylglycerol-β-guaiacyl ether (VG), was studied in 1-butyl-3-methylimidazolium chloride ([BMIM]Cl) with metal chlorides and water. FeCl3, CuCl2, and AlCl3 were found to be effective and functioned catalytically in cleaving the β-O-4 bond of GG, although a number of other metal chlorides are considerably less active. AlCl3 functioned more effectively in cleaving the β-O-4 bond of VG than did FeCl3 and CuCl2. After 120 min at 150 °C, GG conversion reached 100%, and about 70% of the β-O-4 bonds of GG were hydrolyzed, liberating guaiacol, in the presence of FeCl3 and CuCl2, while about 80% of the β-O-4 bonds of GG were hydrolyzed in the presence of AlCl3 with 100% GG conversion. About 75% of the β-O-4 bonds of VG were hydrolyzed in the presence of AlCl3 after 240 min at 150 °C. The catalytic activity is associated with the hydrochloric acid, working as the acid catalyst, formed in situ by the hydrolysis of the metal chlorides. A dehydration product and dimer products from GG were detected and proposed as the possible intermediate products in the GG reaction. One possible acid-catalyzed pathway accounting for the guaiacol production from GG is presented. 1. Introduction Biomass, a renewable feedstock, is an alternative feedstock to the depleting petroleum-based sources, while having a great potential to reduce the carbon footprint.1 Some chemical compounds and new generation biofuels, such as a platform chemical, 5-hydroxymethylfurfural,2,3 and a potential fuel, 2,5dimethylfuran,4 are based on cellulosic feedstocks.5,6 However, the cellulose microfibrils are encapsulated in a matrix consisting of lignin covalently bound to hemicelluloses.7 It is generally accepted that lignin degradation is a rate-limiting step for the degradation of lignocellulose. Therefore, the lignin must be solubilized and/or depolymerized to release the cellulosic fraction.8 Lignin is the third most abundant biomass component after cellulose and hemicellulose, accounting for 18-40 wt % of dry wood.9,10 It is an amorphous polymer with complex chemical structure and is rich in aromatics;11 therefore, lignin is a potential feedstock for some aromatic compounds, such as vanillin.11-13 Moreover, lignin has the potential to yield monomers or fuel.14,15 However, lignin has received less attention relative to cellulose and hemicellulose, mainly because of the complex structure and resistance to degradation. Previous work has revealed that phenylpropane monomers represent the major units of typical lignin structures, which link together primarily through C-O bonds of R- and β-aryl alkyl ethers.16 The β-O-4 ether linkage is found to be dominant (representing ∼50%) among all the linkages of lignin.17,18 Therefore, a lignin depolymerization strategy may target cleaving the β-O-4 ether linkage while preserving the aromatic character of the fragments. Base- and acid-catalyzed lignin depolymerization processes have been explored. The traditional kraft pulping process * To whom correspondence should be addressed. Fax: (512) 4717060. E-mail: [email protected]. † Dalian University of Technology. ‡ The University of Texas at Austin. § KiOR Inc.

employs a base, such as NaOH, and an extra nucleophile, such as NaHS or anthraquinone, to cleave the β-O-4 ether linkage of lignin.16,19 Organic N-bases have also been attempted to break down the β-O-4 bonds in the lignin model compounds, guaiacylglycerol-β-guaiacyl ether (GG) and veratrylglycerolβ-guaiacyl ether (VG), in which 1,5,7-triazabicyclo[4.4.0]dec5-ene was found to be effective for cleaving the β-O-4 bond in a phenolic lignin model compound, GG.20 Acid-catalyzed hydrolysis is another pathway to break the β-O-4 bonds of lignin, and lignin model compound research has been carried out with hydrochloric acid or AlCl3 in dioxane-water or ethanol-water.21,22 For example, Sarkanen et al. used HCl and AlCl3 with a concentration of 0.2 and 0.1 M, respectively, to get enough hydrolysis of the same lignin model compounds used herein with a concentration of 0.015 M, which may imply excess acid is needed in their system.22 Recently, an acidic ionic liquid, 1-H-3-methylimidazolium chloride, has been shown to be effective for this task, in which ∼75% of the β-O-4 bonds of both phenolic and nonphenolic lignin model compounds could be cleaved.23 Generally, strongly acidic conditions are necessary for β-ether linkage cleavage.24 Ionic liquids (ILs) have been used in many research fields, because they are nonflammable, nonvolatile, and recyclable.25,26 ILs have been shown to dissolve wood, cellulose, and lignin,27-33 which opens new opportunities in biomass conversion. Research on sugar and cellulose in ILs has revealed means to produce valuable chemical compounds and fuels.3,34-38 However, relative to sugars and cellulose, fewer studies of lignin reactivity are reported in ILs.20,23,39,40 Metal ions and salts have been applied as catalysts in the biorefining of biomass, such as cellulose hydrolysis and xylotriose degradation.41,42 Metal chlorides have been attempted as catalysts for the dealkylation of alkylphenol lignin model compounds by hydrogenation in a variety of ionic liquids.40 The tested metal chlorides showed no reactivity for the dealkylation reaction. Metal complexes, such as a trisodium meso-tetra-4-

10.1021/ie101884h  2011 American Chemical Society Published on Web 12/15/2010

850

Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

sulfonatophenylporphinerhodium(III) (Rh(TSPP)),43 have also been studied for degrading lignin model compounds. Recently, paired-metal chlorides, such as CuCl2/CrCl2, were reported to hydrolyze the glycosidic C-O bonds of cellulose effectively in 1-ethyl-3-methylimidazolium chloride.38,44 Herein, we report initial results on the hydrolytic cleavage of β-O-4 bonds in both phenolic (GG) and nonphenolic (VG) lignin model compounds in an ionic liquid, 1-butyl-3-methylimidazolium chloride ([BMIM]Cl), with different metal chlorides.

Scheme 1. β-O-4 Bond Cleavage of Lignin Model Compounds

2. Experimental Section

3. Results and Discussion

2.1. Materials. Guaiacylglycerol-β-guaiacyl ether (GG 99%) and 3,4,5-trimethoxybenzaldehyde (TMBA 98%) were purchased from Tokyo Chemical Industry (Japan). Veratrylglycerolβ-guaiacyl ether (VG 97%) was purchased from Astatech (U.S.). LiCl (99%), FeCl3 (98%), FeCl2 · 4H2O (99%), MnCl2 · 4H2O (99%), AlCl3 (98.5%), CrCl3 · 6H2O (98%), CoCl2 (97%), anthrone (AN 98%), p-cresol (99%), and 4-hydroxyacetophenone (98%) were purchased from Acros Organics (Belgium). CuCl (97%), CuCl2 (97%), PdCl2 (99%), NiCl2 (98%), VCl3 (97%), 1-butyl-3-methylimidazolium chloride ([BMIM]Cl 98%), guaiacol (98%), and 4-(1-hydroxyethyl)-2-methoxyphenol (97%) were purchased from Sigma-Aldrich (U.S.). Diethyl ether (99.9%) and hydrochloric acid (HCl 36.9 wt %) were purchased from Fisher Scientific (U.S.). 2.2. General Reaction Procedure. [BMIM]Cl was washed with diethyl ether and vacuum-dried before use. 100 mg of [BMIM]Cl with metal chlorides (e.g., FeCl3, 5 mol % to GG) and 2.25 µL of H2O (4:1 to GG by mol) were added into each vial (0.3 mL) with a magnetic stir bar. The vials were sealed, inserted into a heating and stirring module (Thermo Scientific, U.S.), and stirred at 400 rpm for 2 min at the reaction temperature. The mixtures containing [BMIM]Cl and metal chlorides appeared homogeneous. Next, GG (10 mg, 0.03125 mmol) was added into each vial. The vials were reinserted into the module and stirred at 400 rpm at the reaction temperature. After the reaction was quenched in iced water, the mixtures were diluted by 2 mL of H2O/acetonitrile (1:9 by volume) and analyzed by high pressure liquid chromatography (HPLC). Similar procedures were also followed for the VG reaction. 2.3. Analysis Method. HPLC was performed on a Dionex Ultimate 3000 series with a UV detector (at 280 nm) and a Phenomenex Gemini C6-phenyl column (4.6 × 50 mm). Column temperature was room temperature. H2O/acetonitrile was used as the mobile phase. TMBA was added as an internal standard for quantitative calculations (AN was added as a second internal standard for the reactions with FeCl3, CuCl2, and AlCl3 because unknown products and TMBA could not be separated completely by the HPLC method). Liquid chromatography-mass spectroscopy (LC-MS) was performed on an Agilent 6130 single quadrupole mass spectrometer interfaced to an Agilent 1200 Series HPLC with a diode array detector and a Gemini C18 column (50 × 2.1 mm), and H2O/acetonitrile was used as the mobile phase for LC. Nuclear magnetic resonance (NMR) spectroscopy was performed on a Varian INOVA 500 MHz series system. Fourier transform infrared (FT-IR) spectra were obtained on a Nicolet Avatar 330 FTIR spectrometer with a Smart Performer ATR attachment. Conversion is defined as the molar ratio of GG (or VG) consumed to GG (or VG) added initially. Yield is defined as the molar ratio of compounds produced to the ideal amount of products from GG (or VG) added initially.

GG, a phenolic lignin model compound, was tested first because it is considered to be more reactive. [BMIM]Cl, a common ionic liquid, was used as the solvent. Guaiacol yield was monitored to track β-O-4 bond cleavage, because guaiacol is liberated after the β-O-4 bond of GG is hydrolyzed (as shown in Scheme 1). Water is ubiquitous in a hydroscopic system and necessary for a hydrolysis reaction, such as the hydrolysis of an ether or ester bond, so a controlled amount of water was added at a level that led to ∼2 wt % H2O. GG (10 mg) was reacted in [BMIM]Cl (100 mg) containing metal chlorides (5 mol % to GG) and H2O (2.25 µL) at 130 °C for 120 min. A blank experiment without metal chlorides was tested under the same conditions as above. Figure 1 presents the results for reacting GG in the presence of different metal chlorides and the blank test. The blank test revealed a low level of GG conversion (∼5%), and the only product detected (∼4% yield) was an enol ether (EE), 3-(4hydroxy-3-methoxyphenyl)-2-(2-methoxyphenoxy)-2-propenol, verified by NMR spectroscopy. The formation of EE is consistent with GG dehydration into EE reported in imidazolium-based ILs.20,23,39 LiCl, CoCl2, NiCl2, MnCl2, FeCl2, and CuCl exhibited virtually no activity for GG conversion, because the GG conversions and EE yields were comparable to those found from the blank test. When PdCl2, CrCl3, and VCl3 were used, higher GG conversion (27-54%) and EE yield (∼20%) were observed, as well as some guaiacol formed. The guaiacol yield was 2%, 10%, and 12%, for PdCl2, CrCl3, and VCl3, respectively. In addition, we found some unknown dark precipitate from the experiments with CrCl3 and VCl3. When

Figure 1. GG conversion and product yields of a blank test and experiments with metal chlorides in [BMIM]Cl.

Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

Figure 2. FT-IR spectra of GG and guaiacol.

FeCl3, CuCl2, and AlCl3 were used, GG conversion was 91%, 92%, and 100%, and guaiacol yields of 20%, 22%, and 34% were observed, respectively. FeCl3 and CuCl2 showed a similar performance, while AlCl3 was the most effective. These three metal chlorides were selected as catalysts for additional study. The stability of guaiacol in the reaction medium was tested. After pure guaiacol was heated in the presence of FeCl3, CuCl2, and AlCl3 under the same conditions as those in Figure 1, the recovery of guaiacol was 96%, 97%, and 99%, respectively, which implies the guaiacol formed in the presence of FeCl3, CuCl2, and AlCl3 is not degraded. However, guaiacol recovery was reduced in similar tests with VCl3 and CrCl3, 89.5% and 62.5%, respectively, consistent with the observed dark precipitate for GG conversion. The nature of the precipitate is not identified and will be the subject of future study. Control experiments verify that guaiacol is produced through the β-O-4 bond cleavage of GG. 4-(1-Hydroxyethyl)-2-methoxyphenol, as a lignin model compound, was tested under the same conditions as above. No guaiacol was detected in the presence of FeCl3, CuCl2, or AlCl3, which implies guaiacol was not formed by cleaving the C-C bond from the A-ring of GG but through β-O-4 bond cleavage of the B-ring (as shown in Scheme 1). Figure 2 presents the FT-IR spectra of GG and guaiacol. The absorbance around 3400 cm-1 is associated with the hydroxyl group vibrational stretching modes for the two compounds; symmetric and asymmetric C-H modes produce the absorbances between ∼2800-3000 cm-1; absorbance bands around 1593 and 1500 cm-1 are characteristic of benzene rings; and absorbances around 1253 and 1027 cm-1 are the C-O vibrational stretching bands. As illustrated in Figure 3, the FT-IR spectra of the product mixtures from all three systems extracted by diethyl ether retain many of the same absorbance features as Figure 2 except for the absorbances around 1729 and 1661 cm-1. The absorbance around 1729 cm-1 is the characteristic stretching mode for the CdO bond and implies a ketone or aldehyde was produced. These results are consistent with previous work.23 Hibbert’s ketones would be produced by acidcatalyzed hydrolysis of GG,23,24,45,46 which could account for the CdO bond absorbance. The absorbance around 1661 cm-1 may result from the conjugate effect between the CdO bond and the A-ring of GG after the isomerization of Hibbert’s ketones.23,24,46 The product mixtures were also monitored by LC-MS, and we found two LC peaks that had a (M+H)+/z of 197 in the MS. The LC-MS species are associated with Hibbert’s

851

Figure 3. FT-IR spectra of a product mixture from GG (a) reacted with AlCl3, (b) reacted with CuCl2, and (c) reacted with FeCl3.

Figure 4. Effect of water on GG conversion and guaiacol yield associated with β-O-4 bond cleavage for FeCl3 (blue, g), CuCl2 (red, 3), and AlCl3 (green, 4); and the GG reaction treated by HCl; 15 mol % (to GG) HCl (0, conversion; 9, yield), 10 mol % (to GG) HCl (], conversion; [, yield), and 5 mol % (to GG) HCl (O, conversion; b, yield). Ten mg GG was heated in 100 mg of [BMIM]Cl with metal chlorides (5 mol % to GG) or HCl and H2O at 130 °C for 120 min.

ketones,23,24,45,46 which are companion products from GG after elimination of guaiacol. The Hibbert’s ketones were not established quantitatively herein. Because our results are consistent with previous work,23,24,45,46 in which acids, such as HCl, were used as catalysts for β-O-4 bond hydrolysis, we speculate acid-catalyzed hydrolysis could account for the cleavage reaction in our system and HCl could be the actual catalyst, which was produced in situ by the hydrolysis of AlCl3, CuCl2, and FeCl3 as reported for WCl6 in [BMIM]Cl.47 The available water present in the system affected the GG conversion and guaiacol yields as illustrated in Figure 4. When no water was added, the guaiacol yield dropped to 4%, 5%, and 14% for FeCl3, CuCl2, and AlCl3, respectively. The Lewis acidity of FeCl3, CuCl2, and AlCl3 might lead to coupling condensation or dehydration of GG, which could account for the GG conversion (74% for FeCl3 and CuCl2, respectively; 98% for AlCl3). Once the dehydration of GG happens, the water that is also formed would be available for ether bond hydrolysis. In fact, we found 21%, 10%, and 4% yield of EE (the dehydration product of GG), respectively, in the presence of FeCl3, CuCl2, and AlCl3 after the reaction without added water, which likely explains how guaiacol formed. With the increasing amount of water, more guaiacol was observed, which shows a phenomenon

852

Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

Table 1. Effect of Reaction Temperature on the β-O-4 Bond Cleavage of GG in the Presence of FeCl3, CuCl2, and AlCl3a temperature (°C) 110 130 150 a

FeCl3

CuCl2

AlCl3

conv. (%)

guaiacol yield (%)

EE yield (%)

conv. (%)

guaiacol yield (%)

EE yield (%)

conv. (%)

guaiacol yield (%)

EE yield (%)

54 91 99

8 20 31

18 6 0

46 92 100

7 22 31

7 4 0

89 100 100

25 34 49

6 0 0

10 mg of GG was heated in 100 mg of [BMIM]Cl with catalyst (5 mol % to GG) and 2.25 µL of H2O (4:1 to GG by mol) for 120 min.

similar to that in previous work.23 The rate of increase in the guaiacol yield decreased after the H2O/GG molar ratio reached 16, and the guaiacol yield appeared to approach an asymptotic limit, 59% (FeCl3), 60% (CuCl2), and 73% (AlCl3), at a H2O/ GG molar ratio of 32. More added water increased the rate for the hydrolysis step,23 leading to higher guaiacol yields, because water itself is a reagent in the hydrolysis reaction. GG was completely converted in the presence of AlCl3; however, GG conversion over FeCl3 and CuCl2 decreased slightly beyond an 8:1 H2O/GG molar ratio, which is possibly because the system was diluted by more added water and the acidity decreased. Separate experiments verified that adding water does not affect the solubility of GG in the ionic liquid. In fact, GG is soluble in water at 100 °C. Because of the presence of water, the metal chlorides could hydrolyze to form HCl. To understand if any HCl formed and acted as catalyst, we attempted to measure proton concentration in the system. We first tried to measure the acidity of the [BMIM]Cl/metal chloride/H2O mixtures following the method reported by Thomazeau et al.48 Yet the UV absorption of the metal chloride solutions, such as FeCl3 and CuCl2, masked the absorption of the acid indicators, preventing measurement of the acidity. So a variety of different concentrations of HCl at the same H2O/GG molar ratios as tested above were used. Because hydrochloric acid exists as an aqueous solution, the experiment without water was not tested. As illustrated in Figure 4, GG conversion approached 100% at all H2O/GG ratios when 10 or 15 mol % (to GG) HCl was added. However, GG conversion decreased beyond a H2O/GG molar ratio of 8 when 5 mol % (to GG) HCl was added, which is consistent with the results in the presence of CuCl2 and FeCl3. Comparing the results with metal chlorides and HCl, the guaiacol yield in the presence of AlCl3 was between the experimental runs with 10 and 15 mol % (to GG) HCl, while the guaiacol yields in the presence of CuCl2 and FeCl3 essentially mimicked the behavior in the presence of 5 mol % (to GG) HCl. On the basis of these results, we propose that water present in the system, either added with the reagents or formed in situ through dehydration reactions, hydrolyzed the metal chlorides to form HCl, which functioned as the catalyst. The results also indicate the acidity is higher in the [BMIM]Cl/AlCl3 system. The effect of reaction temperature on GG conversion and guaiacol yield was tested (Table 1). At 110 °C, for [BMIM]Cl/ FeCl3 and [BMIM]Cl/CuCl2 systems, the guaiacol yield is close to the threshold of 5% for catalytic activity. The [BMIM]Cl/ AlCl3 system was catalytic at 110 °C. We also found some EE in the three systems. As reported, EE is a possible intermediate in acid-catalyzed hydrolysis of GG;23,24,45,46 however, EE is not stable under acidic condition and could undergo a subsequent hydrolysis reaction, leading to the β-O-4 bond cleavage of GG to liberate guaiacol. With increasing temperature, GG conversion and guaiacol yield increased, and no EE was observed at 150 °C. In general, lignin consists of more etherified phenylpropane units, and the rate-limiting step is the hydrolysis of the

Table 2. VG Conversion and Guaiacol Yield in the Presence of FeCl3, CuCl2, and AlCl3 at 130 °Ca catalyst FeCl3 CuCl2 AlCl3

VG conv. (%) b

8/24 9/26b 33/61b

guaiacol yield (%) 4/13b 5/14b 19/37b

a 10.4 mg VG was heated in 100 mg [BMIM]Cl with catalyst (5 mol % to VG) and 18 µL H2O (32:1 to VG by mol) for 120 min at 130 °C. b 10 mol % (to VG) catalyst was added.

nonphenolic β-O-4 bonds,49 so we used VG as a nonphenolic lignin model compound. As can be seen in Table 2, when VG was tested at 130 °C, only ∼8% VG conversion and ∼4% guaiacol yield were observed in the presence of FeCl3 and CuCl2. The VG conversion and guaiacol yield were higher in the presence of AlCl3; however, the results appeared to differ from those with GG, which is possibly because GG is more reactive than VG.22 The results of detailed investigation of VG hydrolysis are discussed in greater depth below. Highly acidic conditions may be needed for the β-ether cleavage of VG. More metal chlorides were added to increase the acidity of the systems, and we found both VG conversion and guaiacol yield were improved in the three systems as expected, which implies more highly acidic conditions are necessary for the β-O-4 bond cleavage of VG. GG has a phenolic hydroxide group on its A-ring, while the respective group of VG is etherified by a methyl group (as shown in Scheme 1). The phenolic hydroxide group may also act as a proton donor, forming HCl in situ by the interaction with metal chlorides,47 leading to higher GG conversion and guaiacol yield. To understand the effect of a phenolic hydroxide group, experiments were performed, and the results are presented in Table 3. When no water was added, the VG conversion exhibited the same pattern as Table 2, but produced little or no guaiacol. Water is necessary for the hydrolysis; therefore, little or no guaiacol was formed when water was absent. The VG conversion is possible because of the condensation reaction involving the R- and γ- hydroxyl groups in VG.23 AlCl3, a strong Lewis acid, could lead to more condensation, and H2O formed as a companion product from the condensation, which could facilitate HCl generation in situ; both the H2O and the HCl would enable hydrolysis and lead to the small amount of guaiacol. p-Cresol and 4-hydroxyacetophenone were added as phenolic additives. The two additives were inactive with VG when no metal chlorides were added. However, the VG conversion increased in the presence of the metal chlorides with the phenolic additives, which implies the additives may interact with the metal chlorides to liberate HCl. Acid-catalyzed condensation could happen between the additives and VG,50,51 and between VG and VG.23 Higher VG conversion and guaiacol yield were observed when 4-hydroxyacetophenone was added. The carbonyl group in 4-hydroxyacetophenone has an electronwithdrawing effect, which may increase the reactivity of the phenolic group toward the metal chlorides. When both H2O and additives were used, the guaiacol yield increased, because H2O is critical for the hydrolytic cleavage of the β-O-4 bond. Finally,

Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

853

Table 3. Effect of Additives with a Phenolic Hydroxide Group on the VG Conversion and Guaiacol Yield in the Presence of FeCl3, CuCl2, and AlCl3 at 130 °Ca catalyst

additive

VG conversion (%)

guaiacol yield (%)

FeCl3 CuCl2 AlCl3 none FeCl3 CuCl2 AlCl3 FeCl3 CuCl2 AlCl3 FeCl3 CuCl2 AlCl3

none none none p-cresol/4-hydroxyacetophenone p-cresol p-cresol p-cresol 4-hydroxyacetophenone 4-hydroxyacetophenone 4-hydroxyacetophenone p-cresol/4-hydroxyacetophenone p-cresol/4-hydroxyacetophenone p-cresol/4-hydroxyacetophenone

7 5 33 1.4b/0c 14b/18d 14b/24d 39b/51d 25c 25c 45c 19e/18f 25e/19f 40e/33f

0 0 2 0b/0c 0b/1d 0b/1.5d 2.5b/5d 1c 1c 3c 8e/8f 12e/8f 23e/20f

a 10.4 mg of VG was heated in 100 mg of [BMIM]Cl with catalyst (5 mol % to VG) for 120 min at 130 °C. b 27 mg of p-cresol (8:1 to VG by mol) was added. c 34 mg of 4-hydroxyacetophenone (8:1 to VG by mol) was added. d 54 mg of p-cresol (16:1 to VG by mol) was added. e 54 mg of p-cresol (16:1 to VG by mol) and 18 µL of H2O (32:1 to VG by mol) were added. f 34 mg of 4-hydroxyacetophenone (8:1 to VG by mol) and 18 µL of H2O (32:1 to VG by mol) were added.

Figure 5. GG recovery and product yields for GG reaction in the presence of AlCl3 at 130 °C (the response factors of GG dimers, (M+H)+/z 498 and 303 products are assumed to equal that of GG).

consistent with a GG dimer coupled by an ether linkage produced from the condensation reaction. Herein, the LC peaks with (M+H)+/z of 623 are collectively referred to as GG dimers. Additional products with (M+H)+/z values of 498 and 303 were also detected, which are consistent with previous work involving the 1-H-3-methylimidazolium chloride acidic system.23 The UV absorbance cross sections for the GG dimers and those for the (M+H)+/z 498 and 303 peaks were assumed equal to GG. As illustrated in Figure 5, EE was the main product detected at the onset of the reaction and showed a maximum after 8 min. The GG dimer products display a maximum at 30 min. The subsequent decrease of GG dimers may result from β-O-4 bond cleavage of the dimers to liberate guaiacol. The unknown product with a (M+H)+/z of 498 could be the product formed after one guaiacol is liberated from a GG dimer. In addition, the unknown product with a (M+H)+/z of 303 was also formed in the reaction, which was stable under the conditions. There are some additional unidentified coupling products, and further study is needed to reveal their structural details. On the basis of the results in Figure 5, we speculate EE and GG dimers are possible intermediates in the reaction of GG, leading to the β-O-4 bond cleavage, and we propose one possible acid-catalyzed pathway for the hydrolysis of β-O-4 bond of GG via the possible EE and GG dimer intermediates. In the proposed pathway, parallel reactions, dehydration and condensation of GG, occur first, and then the available water in the system could attack the β-carbon of EE, leading to the β-O-4 bond cleavage to form guaiacol and Hibbert’s ketones, which could undergo tautomerization in the system.24 The GG dimers may also be attacked by water and undergo subsequent deep hydrolysis to form guaiacol and Hibbert’s ketones. The pathway has also been

comparing the last three rows of Table 3 with Table 2, the phenolic additives had a significant impact on both VG conversion and guaiacol yield for FeCl3 and CuCl2, whereas the AlCl3 results were similar. The reasons for these differences were not explored. On the basis of the results above, FeCl3, CuCl2, and AlCl3 are effective for the hydrolytic cleavage of the β-O-4 bond in GG, and AlCl3 is more effective than FeCl3 and CuCl2 for the β-O-4 bond cleavage in VG. To increase the β-O-4 bond cleavage, the reaction was carried out at higher temperature (150 °C). As can be seen in Table 4, most of the β-O-4 bond in GG could be cleaved within 60 min, and the cleavage approached the maximum after 120 min. VG is less reactive than GG, and only ∼50% of the β-O-4 bonds in VG could be cleaved in the presence of AlCl3 (5 mol % to VG) even after 240 min. As more AlCl3 (10 mol % to VG) was added, β-O-4 bond cleavage in VG increased from ∼50% to ∼75%. Sarkanen et al. have reported that the hydrolysis of GG was 4 times faster than that of VG catalyzed by HCl,22 which implies GG is more reactive than VG, so a highly acidic condition is necessary for the VG cleavage reaction. To understand the nature of additional intermediate products and account for the discrepancy between 100% GG conversion and much less than 100% guaiacol yield, studies were conducted at 130 °C with GG in [BMIM]Cl with AlCl3 and water to determine additional products that formed between 0 and 60 min. One expects a condensation reaction involving the hydroxyl groups in GG.23 The products were monitored by LC-MS; some LC peaks with (M+H)+/z of 623 were detected, which is

Table 4. GG and VG Conversion and Guaiacol Yield in the Presence of FeCl3, CuCl2, and AlCl3 at 150 °C for Different Times (nt ) Not Tested)a CuCl2

FeCl3 reaction time (min) 15 60 120 180 240 a

GG conv. (%)

guaiacol yield (%)

61 95 100 100 nt

33 61 69 69 nt

AlCl3

GG conv. (%)

guaiacol yield (%)

63 95 100 100 nt

37 67 70 68 nt

AlCl3

GG conv. (%)

guaiacol yield (%)

VG conv. (%)

guaiacol yield (%)

91 100 100 100 nt

63 76 80 80 nt

nt 57 67 75 79/96b/99c

nt 31 42 50 52/74b/73c

10 mg of GG (or 10.4 mg of VG) was heated in 100 mg of [BMIM]Cl with catalyst (5 mol % to GG or VG) and 18 µL of H2O (32:1 to GG or VG by mol) at 150 °C. b 10 mol % (to VG) AlCl3 was added. c 15 mol % (to VG) AlCl3 was added.

854

Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

reported in previous work.23 More water in the reaction systems increases the probability that water attacks the intermediates and improves the rate for the hydrolysis step, which accounts for the higher guaiacol yield (shown in Figure 4). The hydrolytic cleavage of the β-O-4 bond in VG could also follow the proposed pathway, and further study is needed to establish it. Metal (or metal salts) catalysts have been widely used for lignocellulose transformation in ILs.3,52-54 The work reported herein employs an IL/metal chloride system that likely generates the acid HCl in situ to cleave the β-O-4 bond common in lignin, which is a method to depolymerize lignin and release cellulosic fractions. Moreover, ILs have a capacity for dissolving biomass feedstocks, such as wood, cellulose, and lignin,27-33 which is superior to some traditional organic solvents. In addition, the products in the IL phase could be distilled or extracted by organic solvents, and then the IL is recyclable for successive production, which could reduce the cost. Therefore, it could be that some combination of lignin depolymerization might be feasible in parallel with processes that involve hydrolysis and depolymerization on cellulose that employs the appropriate IL and metal chloride salts. 4. Conclusions We demonstrate a method for the β-O-4 bond cleavage of lignin model compounds in an ionic liquid, [BMIM]Cl, with metal chlorides. FeCl3, CuCl2, and AlCl3 are found to be effective for the β-O-4 bond cleavage of a phenolic lignin model compound, GG. The HCl formed in situ by the hydrolysis of metal chlorides catalyzes the β-O-4 bond cleavage. An increase in available water can lead to more β-O-4 bond cleavage of GG. About 70-80% of the GG β-O-4 bond reacted with water to produce guaiacol at 150 °C after 120 min in the presence of FeCl3, CuCl2, and AlCl3 (each 5 mol % to GG). The AlCl3/ [BMIM]Cl system is also effective for the β-O-4 bond cleavage of a nonphenolic lignin model compound, VG. About 75% of the VG β-O-4 bond reacted with water to produce guaiacol at 150 °C after 240 min in the presence of AlCl3 (10 mol % to VG). A phenolic hydroxide group may interact with metal chlorides to form HCl to increase the acidity in the system. The pathway by which GG reacted was studied, and EE and GG dimers were formed, some of which further reacted to form guaiacol and Hibbert’s ketones. The method described herein may have potential to degrade real lignin (or lignocellulose) in an integrated process. Acknowledgment This work was supported by the National Science Foundation Grant CBET 0849342 and a fellowship to support S.J. from the China Scholarship Council for postgraduates and the Programme of Introducing Talents of Discipline to Universities. Literature Cited (1) Ragauskas, A. J.; Williams, C. K.; Davison, B. H.; Britovsek, G.; Cairney, J.; Eckert, C. A.; Frederick, W. J., Jr.; Hallett, J. P.; Leak, D. J.; Liotta, C. L.; Mielenz, J. R.; Murphy, R.; Templer, R.; Tschaplinski, T. The path forward for biofuels and biomaterials. Science 2006, 311, 484. (2) Roma´n-Leshkov, Y.; Chheda, J. N.; Dumesic, J. A. Phase modifiers promote efficient production of hydroxymethylfurfural from fructose. Science 2006, 312, 1933. (3) Zhao, H. B.; Holladay, J. E.; Brown, H.; Zhang, Z. C. Metal chlorides in ionic liquid solvents convert sugars to 5-hydroxymethylfurfural. Science 2007, 316, 1597. (4) Roma´n-Leshkov, Y.; Barrett, C. J.; Liu, Z. Y.; Dumesic, J. A. Production of dimethylfuran for liquid fuels from biomass-derived carbohydrates. Nature 2007, 447, 982.

(5) Perego, C.; Bianchi, D. Biomass upgrading through acid-base catalysis. Chem. Eng. J. 2010, 161, 314. (6) Nigam, P. S.; Singh, A. Production of liquid biofuels from renewable resources. Prog. Energy Combust. Sci. 2011, 37, 52. (7) Davis, J. R.; Sello, J. K. Regulation of genes in Streptomyces bacteria required for catabolism of lignin-derived aromatic compounds. Appl. Microbiol. Biotechnol. 2010, 86, 921. (8) Loureiro, P. E. G.; Fernandes, A. J. S.; Carvalho, M. G. V. S.; Evtuguin, D. V. The assessment of chromophores in bleached cellulosic pulps employing UV-Raman spectroscopy. Carbohydr. Res. 2010, 345, 1442. (9) Amen-Chen, C.; Pakdel, H.; Roy, C. Production of monomeric phenols by thermochemical conversion of biomass: a review. Bioresour. Technol. 2001, 79, 277. (10) Zakzeski, J.; Bruijnincx, P. C. A.; Jongerius, A. L.; Weckhuysen, B. M. The catalytic valorization of lignin for the production of renewable chemicals. Chem. ReV. 2010, 110, 3552. (11) Bozell, J. J.; Holladay, J. E.; Johnson, D.; White, J. F. Top Value Added Chemicals from Biomass, Vol. II: Results of screening for potential candidates from biorefinery lignin; Pacific Northwest National Laboratory: Richland, WA, 2007. (12) Voitl, T.; Rohr, P. R. V. Oxidation of lignin using aqueous polyoxometalates in the presence of alcohols. ChemSusChem 2008, 1, 763. (13) Voitl, T.; Rohr, P. R. V. Demonstration of a process for the conversion of kraft lignin into vanillin and methyl vanillate by acidic oxidation in aqueous methanol. Ind. Eng. Chem. Res. 2010, 49, 520. (14) Okuda, K.; Man, X.; Umetsu, M.; Takami, S.; Adschiri, T. Efficient conversion of lignin into single chemical species by solvothermal reaction in water-p-cresol solvent. J. Phys.: Condens. Matter 2004, 16, S1325. (15) Yan, N.; Zhao, C.; Dyson, P. J.; Wang, C.; Liu, L.; Kou, Y. Selective degradation of wood lignin over noble-metal catalysts in a twostep process. ChemSusChem 2008, 1, 626. (16) Gierer, J. Chemical aspects of kraft pulping. Wood Sci. Technol. 1980, 14, 241. (17) Adler, E. Lignin chemistry-past, present and future. Wood Sci. Technol. 1977, 11, 169. (18) Chakar, F. S.; Ragauskas, A. J. Review of current and future softwood kraft lignin process chemistry. Ind. Crops Prod. 2004, 20, 131. (19) Venica, A. D.; Chen, C.; Gratzl, J. S. Soda-AQ delignification of poplar wood. Part 1: Reaction mechanism and pulp properties. Holzforschung 2008, 62, 627. (20) Jia, S.; Cox, B. J.; Guo, X.; Zhang, Z. C.; Ekerdt, J. G. Decomposition of a phenolic lignin model compound over organic N-bases in an ionic liquid. Holzforschung 2010, 64, 577. (21) Adler, E.; Pepper, J. M.; Eriksoo, E. Action of mineral acid on lignin and model substances of guaiacylglycerol-beta-aryl ether type. Ind. Eng. Chem. 1957, 49, 1391. (22) Sarkanen, K. V.; Hoo, L. H. Kinetics of hydrolysis of erythroguaiacylglycerol β-(2-methoxyphenyl) ether and its veratryl analogue using HCl and aluminum chloride as catalysts. J. Wood Chem. Technol. 1981, 1, 11. (23) Jia, S.; Cox, B. J.; Guo, X.; Zhang, Z. C.; Ekerdt, J. G. Cleaving the β-O-4 bonds of lignin model compounds in an acidic ionic liquid, 1-H3-methylimidazolium chloride: an optional strategy for the degradation of lignin. ChemSusChem 2010, 3, 1078. (24) McDonough, T. J. The chemistry of organosolv delignification. TAPPI Solvent Pulping Seminar, Boston, MA, November 6-7, 1992. (25) Zhang, Z. C. Catalysis in ionic liquids. AdV. Catal. 2006, 49, 153. (26) Olivier-Bourbigou, H.; Magna, L.; Morvan, D. Ionic liquids and catalysis: Recent progress from knowledge to applications. Appl. Catal., A 2010, 373, 1. (27) Remsing, R. C.; Swatloski, R. P.; Rogers, R. D.; Moyna, G. Mechanism of cellulose dissolution in the ionic liquid 1-n-butyl-3methylimidazolium chloride: a 13C and 35/37Cl NMR relaxation study on model systems. Chem. Commun. 2006, 1271. (28) Kilpela¨inen, I.; Xie, H.; King, A.; Granstrom, M.; Heikkinen, S.; Argyropoulos, D. S. Dissolution of wood in ionic liquids. J. Agric. Food Chem. 2007, 55, 9142. (29) Pu, Y.; Jiang, N.; Ragauskas, A. J. Ionic liquid as a green solvent for lignin. J. Wood Chem. Technol. 2007, 27, 23. (30) Fort, D. A.; Remsing, R. C.; Swatloski, R. P.; Moyna, P.; Moyna, G.; Rogers, R. D. Can ionic liquids dissolve wood? Processing and analysis of lignocellulosic materials with 1-n-butyl-3-methylimidazolium chloride. Green Chem. 2007, 9, 63. (31) Zavrel, M.; Bross, D.; Funke, M.; Bu¨chs, J.; Spiess, A. C. Highthroughput screening for ionic liquids dissolving (ligno-)cellulose. Bioresour. Technol. 2009, 100, 2580. (32) King, A. W. T.; Zoia, L.; Filpponen, I.; Olszewska, A.; Xie, H.; Kilpela¨inen, I.; Argyropoulos, D. S. In situ determination of lignin phenolics

Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011 and wood solubility in imidazolium chlorides using 31P NMR. J. Agric. Food Chem. 2009, 57, 8236. (33) Li, B.; Filpponen, I.; Argyropoulos, D. S. Acidolysis of wood in ionic liquids. Ind. Eng. Chem. Res. 2010, 49, 3126. (34) Li, C.; Zhao, Z. K. Efficient acid-catalyzed hydrolysis of cellulose in ionic liquid. AdV. Synth. Catal. 2007, 349, 1847. (35) Yong, G.; Zhang, Y.; Ying, J. Y. Efficient catalytic system for the selective production of 5-hydroxymethylfurfural from glucose and fructose. Angew. Chem., Int. Ed. 2008, 47, 9345. (36) Sievers, C.; Valenzuela-Olarte, M. B.; Marzialetti, T.; Musin, I.; Agrawal, P. K.; Jones, C. W. Ionic-liquid-phase hydrolysis of pine wood. Ind. Eng. Chem. Res. 2009, 48, 1277. (37) Lima, S.; Neves, P.; Antunes, M. M.; Pillinger, M.; Ignatyev, N.; Valente, A. A. Conversion of mono/di/polysaccharides into furan compounds using 1-alkyl-3-methylimidazolium ionic liquids. Appl. Catal., A 2009, 363, 93. (38) Su, Y.; Brown, H. M.; Huang, X.; Zhou, X.; Amonette, J. E.; Zhang, Z. C. Single-step conversion of cellulose to 5-hydroxymethylfurfural (HMF), a versatile platform chemical. Appl. Catal., A 2009, 361, 117. (39) Kubo, S.; Hashida, K.; Yamada, T.; Hishiyama, S.; Magara, K.; Kishino, M.; Ohno, H.; Hosoya, S. A characteristic reaction of lignin in ionic liquids; Glycelol type enol-ether as the primary decomposition product of β-O-4 model compound. J. Wood Chem. Technol. 2008, 28, 84. (40) Binder, J. B.; Gray, M. J.; White, J. F.; Zhang, Z. C.; Holladay, J. E. Reactions of lignin model compounds in ionic liquids. Biomass Bioenergy 2009, 33, 1122. (41) Shimada, N.; Kawamoto, H.; Saka, S. Solid-state hydrolysis of cellulose and methyl R- and β-D-glucopyranosides in presence of magnesium chloride. Carbohydr. Res. 2007, 342, 1373. (42) Liu, C.; Wyman, C. E. The enhancement of xylose monomer and xylotriose degradation by inorganic salts in aqueous solutions at 180 °C. Carbohydr. Res. 2006, 341, 2550. (43) Watson, P. A.; Wright, L. J.; Fullerton, T. J. Reactions of metal ion complexes with lignin model compounds, Part III. Rh(TSPP) catalyzed formation of guaiacol from β-aryl ethers in exceptionally high yield. J. Wood Chem. Technol. 1993, 13, 411. (44) Su, Y.; Brown, H. M.; Li, G.; Zhou, X.; Amonette, J. E.; Fulton, J. L.; Camaioni, D. M.; Zhang, Z. C. Accelerated cellulose depolymerization

855

catalyzed by paired metal chlorides in ionic liquid solvent. Appl. Catal., A 2010, doi: 10.1016/j.apcata.2010.09.021. (45) Hoo, L. H.; Sarkanen, K. V.; Anderson, C. D. Formation of C6C2enol ethers in the acid-catalyzed hydrolysis of erythro-veratrylglycerol-β(2-methoxyphenyl) ether. J. Wood Chem. Technol. 1983, 3, 223. (46) Hage, R. E.; Brosse, N.; Sannigrahi, P.; Ragauskas, A. Effects of process severity on the chemical structure of Miscanthus ethanol organosolv lignin. Polym. Degrad. Stab. 2010, 95, 997. (47) Lai, L.; Zhang, Y. The effect of imidazolium ionic liquid on the dehydration of fructose to 5-hydroxymethylfurfural, and a room temperature catalytic system. ChemSusChem 2010, 3, 1257. (48) Thomazeau, C.; Olivier-Bourbigou, H.; Magna, L.; Luts, S.; Gilbert, B. Determination of an acidic scale in room temperature ionic liquids. J. Am. Chem. Soc. 2003, 125, 5264. (49) Schultz, T. P.; Fisher, T. H. Alkaline hydrolysis of nonphenolic β-O-4 lignin models: Substituent effect of the A-ring on the rate. Holzforschung 2002, 56, 592. (50) Lin, L.; Yao, Y.; Shiraishi, N. Liquefaction mechanism of β-O-4 lignin model compound in the presence of phenol under acid catalysis. Part 1. Identification of the reaction products. Holzforschung 2001, 55, 617. (51) Lin, L.; Nakagame, S.; Yao, Y.; Yoshioka, M.; Shiraishi, N. Liquefaction mechanism of β-O-4 lignin model compound in the presence of phenol under acid catalysis. Part 2. Reaction behavior and pathways. Holzforschung 2001, 55, 625. (52) Yan, N.; Zhao, C.; Luo, C.; Dyson, P. J.; Liu, H.; Kou, Y. Onestep conversion of cellobiose to C6-alcohols using a ruthenium nanocluster catalyst. J. Am. Chem. Soc. 2006, 128, 8714. (53) Hu, S.; Zhang, Z.; Song, J.; Zhou, Y.; Han, B. Efficient conversion of glucose into 5-hydroxymethylfurfural catalyzed by a common Lewis acid SnCl4 in an ionic liquid. Green Chem. 2009, 11, 1746. (54) Chan, J. Y. G.; Zhang, Y. Selective conversion of fructose to 5-hydroxymethylfurfural catalyzed by tungsten salts at low temperatures. ChemSusChem 2009, 2, 731.

ReceiVed for reView September 10, 2010 ReVised manuscript receiVed October 27, 2010 Accepted November 28, 2010 IE101884H