Applied Energy 154 (2015) 520–527
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Preparation of methyl levulinate from fractionation of direct liquefied bamboo biomass Junfeng Feng a, Jianchun Jiang a,b,⇑, Junming Xu a,b, Zhongzhi Yang a, Kui Wang a, Qian Guan a Shuigen Chen a a Institute of Chemical Industry of Forest Products, Chinese Academy of Forestry (CAF), National Engineering Laboratory for Biomass Chemical Utilization, Key and Open Laboratory on Forest Chemical Engineering, State Forestry Administration (SFA), Key Laboratory of Biomass Energy and Material, Nanjing 210042, China b Institute of New Technology of Forestry, Chinese Academy of Forestry (CAF), Beijing 100091, China
h i g h l i g h t s Methyl levulinate was produced from bamboo by direct liquefied conversion. The yield of methyl levulinate can reach to 30.75 wt% using bamboo methanolysis. Models compounds were used to describe the reaction pathway of direct liquefaction.
a r t i c l e
i n f o
Article history: Received 15 November 2014 Received in revised form 23 April 2015 Accepted 28 April 2015
Keywords: Lignocellulosic biomass Liquefaction Methyl levulinate
a b s t r a c t One-step preparation of methyl levulinate from biomass was investigated. The process used was direct liquefaction under pressure in methanol using a 1 L autoclave. Bamboo, a lignocellulosic biomass, was liquefied using sulfuric acid in subcritical methanol. When sulfuric acid was used as the catalyst, a 30.75 wt% methyl levulinate yield could be obtained from bamboo at 200 °C after a reaction time of 120 min when the catalyst loading was 2.5 wt% per 60 g bamboo. In addition, microcrystalline cellulose, corn starch, methyl glucoside and glucose were selected as model compounds for the liquefaction reaction so that the biomass to methyl levulinate reaction pathway could be investigated. The results suggested that lignocellulosic biomass is a renewable material that can be used to produce a high value-added fuel additive (methyl levulinate) by the direct liquefaction under pressure reaction process. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction The development of renewable biomass energy has attracted extensive attention because of the gradual depletion of conventional non-renewable resources and the greater focus on environmental protection [1]. In recent years, increasing numbers of researchers have become interested in studying various methods of converting biomass into high value-added fuel additives [2–4]. In China, Moso bamboo is the most common lignocellulosic biomass energy crop, and is planted in a large number of areas. They are widely distributed across southeast and southwest China and approximately 15 108 poles moso bamboo are available annually in China. The Moso bamboo has a fast growth rate, is high yielding and has a number of different uses. It has been ⇑ Corresponding author at: Institute of New Technology of Forestry, Chinese Academy of Forestry (CAF), Beijing 100091, China. Tel.: +86 13770663924; fax: +86 25 85482487. E-mail address:
[email protected] (J. Jiang). http://dx.doi.org/10.1016/j.apenergy.2015.04.115 0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.
studied as a raw material for industrial products for a long time. However, a large amount of bamboo waste (accounting for 30–40% of the whole bamboo) is not fully utilized during the production of down-stream products, such as furniture. The lignocellulosic biomass waste is one of the most abundant biomass by-products available worldwide, and mainly consists of cellulose, hemicelluloses and lignin. These components can be converted into high value-added chemicals and fuels [5–7], and can replace non-renewable petroleum resources. One such product is methyl levulinate (MLA), which can be produced from lignocellulosic biomass by liquefaction [8]. Pacific Northwest National Laboratory (PNNL) and the U.S. Department of Energy [9] state that levulinate esters are important chemicals, and are one of the more recognized chemical building blocks that can be produced from carbohydrates. They are short chain fatty esters with numerous potential applications [10], such as a fuel additive to improve petroleum and diesel properties, such as stability, low-temperature fluidity and the flash point [11,12]. In addition, levulinate esters are the preferred substrates for chemical
J. Feng et al. / Applied Energy 154 (2015) 520–527
conversion to many other useful chemicals via different kinds of condensation and addition reactions at the ester and keto groups [13]. The traditional method of producing MLA from lignocellulosic biomass has three steps: (1) produce glucose or 5-hydroxymethyl furfural (HMF) by cellulose hydrolysis in water [8,14]; (2) produce levulinic acid (LA) through dehydration and rearrangement of glucose [15]; and (3) generate MLA by esterification of LA with alcohols. This process has many drawbacks, such as a long reaction route [16–19] and the considerable amount of energy needed to remove the large volumes of water during LA purification. Furthermore, levulinate ester yields are only 16–17 wt% [20]. Currently, catalytic and chemical processes that produce chemicals from biomass are attracting more and more concern. This has meant that the latest studies have focused on the direct conversion of carbohydrates derived from biomass to levulinate esters because the synthetic technique is simple, the products are purified easily by distillation and wastewater is minimized [21]. Some studies have reported alternative ways of converting mono-sugars (simple model compounds) into levulinate esters, such as fructose, glucose and cellulose. Liu [21] reported converting fructose to MLA with a 73.7% yield at 130 °C using a Fe-HPW-1 catalyst. Tominaga et al. [22] investigated an efficient catalyst for MLA synthesis from glucose by combining two different kinds of acids, and achieved a MLA yield of 75%. Wu et al. [23] alcoholyzed cellulose in subcritical methanol and obtained a 55% MLA yield. The yields of levulinate esters have therefore improved, but the cost of the raw materials has also increased. Most processes use non-commercial catalysts (such as Fe-HPW-1) and these catalysts can easily be deactivated. The development of efficient methods that directly convert bamboo into high value-added fuel additive, under mild conditions, still remains a significant challenge [24,25]. The move from model compounds (cellulose, glucose and fructose) to ‘‘real’’ lignocellulosic biomass stocks requires modification of the components during liquefaction so that the liquefied products can easily undergo fractionation, which provides specific products for further refining reactions. Bamboo has a complex lignocellulosic structure mainly consist of crystallized holocellulose and phenolic lignin, which means that the liquefaction and separation process is more difficult than that of cellulose and starch. To overcome these problems, this study aims to investigate the direct conversion of bamboo to MLA by an acid-catalyzed reaction in subcritical methanol. The use of alcohols as solvents for carbohydrate conversion has some advantages [16]. For example, alcohols can suppress humins formation, minimize wastewater discharge and allows higher-grade products to be easily isolated by fractionation. Process development research into converting biomass into biofuels and high added-value chemicals in supercritical and subcritical alcohols has attracted more attention recently [23]. The direct pressurization and liquefaction of bamboo to create MLA in a subcritical methanol medium is a new and alternative method that can effectively utilize an abundant form of lignocellulosic biomass (bamboo). It was also found that high yields of product were generated from liquefaction of bamboo using low concentration sulfuric acid. The effects of different process parameters, including different kinds of catalysts, catalyst amount, biomass to solvent mass ratio, reaction temperature and time were investigated so that we could optimize the degradation of bamboo in a methanol medium at low sulfuric acid concentrations, and obtain the highest possible MLA yield. An advantage of the low concentration acid catalyst system is that little, undesired dimethyl ether is formed from the side
521
reaction that leads to the dehydration of methanol. Only a small amount of spent acid needed to be treated after the reaction, and equipment corrosion is minimal. Overall, this catalytic strategy is an efficient and economical design for converting biomass into high value-added chemicals and fuels. The chemical components and content of MLA in the liquefaction product were analyzed by gas chromatography–mass spectrometry (GC–MS) and gas chromatography (GC). 2. Materials and methods 2.1. Chemicals The bamboo was collected from a local farm in Sichuan Province, China. The bamboo consisted (absolute content) of 46.21 wt% cellulose, 23.69 wt% hemicellulose, 25.14 wt% lignin, 1.45 wt% ash. The corn starch (DP = 600), microcrystalline cellulose (DP = 400), glucose, methyl glucoside (MLG) and MLA were obtained from Aladdin Reagents (99.5% purity); and the methanol and acid catalysts (analytical grade) were from Sinopharm Chemical Reagents (Beijing, China). The materials (corn starch, microcrystalline cellulose and bamboo) were oven-dried at 105 °C over 24 h until they were completely dry, milled and then screened for particles within the size range 0.3–0.425 mm (40–50 mesh). All the other chemicals in the study were analytical grade, commercially available and used without further purification. 2.2. Analytical methods The MLA in the liquefaction product was quantitatively analyzed by gas chromatography (GC) using a flame ionization detector. The MLA was separated on a capillary column (30 m 0.05 lm 0.32 nm) with a programmed temperature range of 50–230 °C using nitrogen as the carrier gas. The internal standard was n-octanol. The quantitative analysis of MLG was carried out using a HPLC instrument (Shimadzu LC–10ATVP) with an Aminex HPX–87H column and a RID–20A detector. The mobile phase was 0.005 mmol sulfuric acid (sonication, deaeration) in water with a flow rate of 0.6 mL/min, and the column temperature was maintained at 50 °C. Gas chromatography mass spectrometry (GC–MS) (Agilent 5975C VL MSD) analysis was undertaken using a capillary column with a 0.05 lm film thickness. Helium was used as the carrier gas at a flow rate of 1.6 mL/min, and the following temperature program was applied: starting temperature of 45 °C for 5 min, heating at a rate of 5 °C/min to a final temperature 250 °C, followed by 20 min at this temperature. The MS detector was operated in the electron ionization mode (70 eV) with an ionization temperature of 220 °C. Typically, 0.2 lL of sample was used. The mass spectra were recorded in electron ionization mode for m/z 50–550. The MLA chemical structure was determined by GC–MS. A GC was used to investigate the absolute content of MLA in the liquefaction product by comparison with an authentic n-octanol reference sample. The corresponding MLA and n-octanol peak area ratio reflected the content ratio according to the standard curve (y = 2.06543x 0.02107, coefficient of correlation (R2) = 0.99998). As the tests involved repeated data points, the average values of the data were obtained by repeating the experiments a number of times and used to calculate the standard deviation of the standard curve to indicate the experimental error range. Eq. (1) was used to calculate the liquefaction product yield from bamboo (on a weight basis), and Eq. (2) was used to calculate the MLA yield from bamboo. Eqs. (3) and (4) were used to measure the main by-product
522
J. Feng et al. / Applied Energy 154 (2015) 520–527
(5-methoxymethyl furfural: MMF and methyl glucoside: MLG) yields and the methanol recovery yield.
reach as much as 27.73 wt%, and the next highest peak appeared at 25.07 wt% for toluenesulfonic acid. This indicated that sulfuric
m ðliquefaction productÞ 100% m ðbambooÞ
ð1Þ
MLA yield ðwt%Þ ¼ m ðliquefaction productÞ mass yield of MLA ðmeasured by GCÞ=m ðbambooÞ100%
ð2Þ
Liquefaction product yield ðwt%Þ ¼
By-product yield ðwt%Þ ¼ 100% m ðliquefaction productÞ mass yield of ðMMF þ MLGÞ ðmeasured by GC and HPLCÞ=m ðbambooÞ ð3Þ Methanol recovery yield ðwt%Þ ¼
m ðrecovery methanolÞ 100% m ðinitial methanolÞ
The reaction liquid mixture was taken out from the autoclave and filtered through a membrane filter with a pore size of 0.8 lm. The liquefaction product was removed from the reaction liquid mixture and evaporated under vacuum at 50 °C to remove methanol.
ð4Þ
acid and toluenesulfonic acid were strong enough to degrade glycoside bonds in subcritical methanol. Phosphoric acid was weaker and the MLA yield was much lower. Formic acid produced the lowest MLA yield of all the acid catalysts in our experiments. 3.2. Effects of various parameters on bamboo liquefaction
2.3. Liquefaction of bamboo by pressurization
3.1. Effect of various catalysts on bamboo liquefaction The reaction was performed in a number of acid catalysts mixed with subcritical methanol. The performance of the different catalysts on MLA yield had an overall trend (Table 1) as follows: sulfuric acid > toluenesulfonic acid > sulfamic acid > phosphoric acid > formic acid > blank (without catalyst). Liquefaction product yield to MLA ratio in subcritical methanol was influenced by the strength of the acid catalyst. The stronger the acid catalyst, the higher the yield of MLA. With sulfuric acid, the yield of MLA can
12
100
10
80
8 Liquefaction product yield Methanol recovery yield By-products yield Methyl levulinate yield Pressure
60 40
6 4
20
Pressure (MPa)
3. Results and discussion
3.2.1. Catalyst loading Sulfuric acid was selected to investigate the effect of catalyst loading on liquefaction product yield because it produced the highest MLA yield. The bamboo was reacted in methanol in the presence of sulfuric acid with loadings from 1.0 to 4.0 wt% (per 60 g bamboo) (Fig. 1). The MLA yield reached a maximum of 27.73 wt% when sulfuric acid was increased from 1.0 to 2.5 wt%. In contrast, the MLA yield fell by approximately 10 wt% when sulfuric acid increased from 2.5 to 4.0 wt%. Catalyst loading determines the availability of acidic reaction sites and also affects the yield of desirable products and undesirable by-products [26]. The acid promoted the formation of MLA as the catalyst loading rose to 2.5 wt%. Other undesired products, such as methyl formate, furan and MLG were also detected by
Yield (wt%)
Materials were introduced into a solution containing an acidic catalyst and methanol, and the mixture was heated at high-pressure (0–30 MPa) in a 1 L autoclave at the specified temperature (0–300 °C), with a stirring speed of 600 rpm, for a designated time period. The heating rate was 3 °C/min. The reaction time was the period during which the highest constant temperature was maintained. After the reaction, the reactor was cooled rapidly in a water bath to room temperature. The gas was collected, weighed and subjected to further analysis. The reaction liquid mixture was removed from the autoclave and filtered through a membrane filter with a pore size of 0.8 lm. The filter cake (residue) was repeatedly washed with methanol to extract the residual as completely as possible. The filtrate was neutralized to pH 7 using NaOH solution, then evaporated under vacuum at 50 °C to remove and recycle the solvents, which left a liquefaction product that did not contain methanol. The residue was dried at 105 °C to give the char yield.
2 0
0 1.0
1.5
2.0
2.5
3.0
3.5
4.0
Catalyst to total materials (‰) Fig. 1. Effect of different catalyst loads. Reaction conditions: bamboo to methanol ratio: 1:7, 200 °C, 60 min.
Table 1 Effect of acid catalysts on the liquefaction of bamboo to MLAa,c. Results
Color of liquefied product Liquefaction product yield (wt%) MLAc yield (wt%) By-productsd yield (wt%) Methanol recovery yield (wt%) a b c d
Catalysts Blankb
Sulfuric acid
Sulfamic acid
Toluenesulfonic acid
Formic acid
Phosphoric acid
Pale brown 7.41 2.87 0.32 90.15
Brown 85.23 27.73 3.09 95.23
Pale brown 50.61 11.24 2.73 94.28
Brown 71.37 25.07 7.13 96.02
Pale brown 15.40 5.07 3.46 89.33
Pale brown 25.28 8.43 5.02 90.19
Reaction conditions: bamboo to methanol ratio of 1:7, 2.5 wt% catalyst, 200 °C, 60 min, stirring speed 600 rpm, average of two runs. Blank represents the non-catalyst treatment. MLA denote methyl levulinate. By-product include 5-methoxymethyl furfural and methyl glucoside.
523
J. Feng et al. / Applied Energy 154 (2015) 520–527
Liquefaction product yield Methyl levulinate yield By-products yield Methanol recovery yield Pressure
140 120
Yield (%)
100
14 12 10
80
8
60
6
40
4
20
2 0
0
3.2.2. Methanol amount In this study, methanol was used as the liquefaction solvent. The subcritical methanol could provide a high autogenetic pressure and good solubility for the products during the reaction, both of which promote the conversion of bamboo to methyl levulinate. The reaction was performed for 60 min, using different amounts of methanol, so that the effect of solvent amount on bamboo liquefaction could be investigated. When the amount of methanol added increased, the MLA yield also rose (Fig. 2), which indicated that bamboo can be degraded into small molecules in subcritical methanol. The main role of subcritical methanol, as a low molecular-weight liquefying agent in bamboo liquefaction experiments, is to dissolve and disperse the liquefied product. The larger methanol amounts improved the conversion of bamboo and slightly increased MLA yield. However, a large rise in the amount of methanol used would greatly increase
120 100
Yield (%)
14
Liquefaction product yield Methyl levulinate yield By-products yield Methanol recovery yield Pressure
12 10
80
8
60
6
40
4
20
2
0 250
300
350
400
450
Pressure (MPa)
140
120
Yield (%)
100
Peaks
12
8
60
6
40
4
20
2
0 20
40
60
80
100
120
140
0 160
Time (min) Fig. 3. Effect of different reaction times. Reaction conditions: bamboo to methanol ratio of 1:7, 2.5 wt% catalyst, 200 °C.
220
Table 2 The GC–MS analysis of bamboo liquefaction productsa.
14
80
200
3.2.3. Reaction time A typical time profile for the conversion of bamboo using sulfuric acid is shown in Fig. 3. As the reaction time rose from 30 to 60 min, the MLA yield increased from 7.23 to 24.46 wt%. As the reaction time increases, more bamboo is converted into MLA and a maximum MLA yield of 30.75 wt% was obtained after 120 min. MLA yield decreased to 21.07 wt% when the reaction time was further extended to 150 min. This trend indicated that the optimum reaction time was 120 min. Furthermore, after 120 min, the formation of dimethyl ether between two methanol molecules had taken place via a dehydration reaction. This phenomenon was not observed when the reaction time was below 120 min. It was also observed that as the reaction time rose above 60 min, the
0 500
10
180
the production costs. Therefore, there is an economic optimal amount of methanol needed to achieve high MLA yields at a reasonable economic cost. In this study, the optimum methanol amount was a bamboo to methanol ratio of 1:7 and this generated a high MLA yield of 27.73 wt%.
Pressure (MPa)
120
160
Fig. 4. Effect of different reaction temperatures. Reaction conditions: bamboo to methanol ratio of 1:7, 2.5 wt% catalyst, 120 min.
Fig. 2. Effect of different methanol amounts. Reaction conditions: 2.5 wt% catalyst per 60 g bamboo, 200 °C, 60 min.
Liquefaction product yield Methyl levulinate yield By-products yield Methanol recovery yield Pressure
140
Tempetature ( 䉝)
Methanol amount (g)
140
Pressure (MPa)
GC–MS in low amounts, as well as 5-methoxymethyl furfural (MMF) at <5 wt%, which may explain the fast conversion to methyl levulinate. At higher acid concentrations, the main by-products (MLG and MMF) increased by about 4 wt% and the MLA yield fell by approximately 10 wt%. This indicated that by-products undergo side reactions at higher acid concentrations, which causes increases in by-product contents in the liquefied product. The higher acid concentrations can seriously corrode reactor equipment, and more spent acid needs to be neutralized after the reaction. Furthermore, the addition of excess sulfuric acid has negative effect on the reaction. In general, the best catalyst loading for obtaining high MLA yields was 2.5 wt%.
a
Time (min)
Compounds name
Area (%)
Similarity (%)
1 2 3 4 5
/ 2.12 2.76 8.94 18.66 20.10
Esters Methyl formate Dimethyl ether Methyl levulinate Pentanoate 4-Methylpentanoate
70.95 2.03 2.36 65.21 0.23 1.12
/ 89 76 94 35 76
6 7 8
/ 7.33 12.32 15.45
Furans Furfural 2-Furanmethanol 5-Methoxymethyl furfural
5.36 0.61 0.12 4.63
/ 90 67 80
9 10 11 12
/ 17.87 23.04 25.37 27.38
Glycosides Methyl-a-D-xylofuranoside Methyl-a-D-glucoside Methyl-a-D-glucopyranoside 4-Methoxy-a-Dglucopyranoside
7.65 1.71 4.03 1.18 0.73
/ 78 92 79 80
13 14
/ 14.23 15.83
Phenols 2-Methoxyphenol 4-Methyl guaiacol
14.09 6.62 7.38
/ 87 76
15
/ 6.75
1.95 0.81
/ 68
16
29.48
Others Tripropylene glycol methyl ether 2,3-Dihydroxy acid
1.14
56
Reaction conditions: bamboo to methanol ratio of 1:7, 2.5 wt% catalyst, 200 °C, 120 min.
524
J. Feng et al. / Applied Energy 154 (2015) 520–527
liquefaction product yield approached 80 wt% (Fig. 3). However, the residue yield decreased from 20 to 15 wt% when the reaction time was extended from 30 to 60 min. Furthermore, as the reaction time got longer, the conversion of bamboo gradually declined over time. Therefore, extending the reaction time may also cause the re-condensation of the liquefied products [7]. 3.2.4. Reaction temperature The effect of reaction temperature on MLA yield was investigated in order to identify the optimum conditions for obtaining the highest possible MLA yield in subcritical methanol. It can be seen from Fig. 4 that reaction temperature played an important role in the course of the experiment. Usually, elevated temperatures enhance the reaction rate and conversion efficiency. The MLA yield grew significantly when the reaction temperature increased from 120 to 200 °C (the reaction time was the same for all temperatures). Then it began to fall as the temperature rose above 200 °C. However, it should be pointed out that MLA will decompose into byproducts at higher temperatures. Additionally, more side-reactions may occur during the reaction. Therefore, temperature elevation does not necessarily improve the MLA formation rate. 3.3. Mass balance and analysis of high yield experiment The best liquefaction product yield and MLA yield conditions were a sulfuric acid catalyst at a loading of 2.5 wt%, a reaction temperature of 200 °C, a biomass to solvent ratio of 1:7 and a reaction time of 120 min. The composition of the bamboo methanol
liquefied product is shown in Table 2. The main liquefaction products were esters (mainly MLA), furans (mainly MMF), glycosides (mainly MLG) and phenols. The phenols are the degradation products from lignin. The others are hemicellulose and cellulose that have formed due to the degradation of the original lignocellulosic biomass. The mass balances of high liquefaction product yields are shown in Table 3. The total feed subjected to liquefaction was 481.5 g, which consisted of 60 g bamboo, 420 g methanol and 1.5 g sulfuric acid. Although the reaction mixture was taken out from the autoclave as quickly as possible some methanol, losses were unavoidable, but were all less than 3 wt% by weight of bamboo (60 g). The liquefaction product yield was more than 75 wt% when averaged over three runs. NaOH was used to neutralize the spent acid to prevent its negative effect on the reaction. About 1.2 g NaOH was used to neutralize the liquid sample. As the tests involved repeated data points, average values for the data were obtained by repeating the experiments three times. 3.4. Purification of MLA by extraction It is difficult to separate MLA from liquefied products effectively using distillation as their boiling points are close. Extraction is a potential way to obtain the end product as the polarities of chemicals are distinct. The liquefied product was a brown liquid obtained by rotary evaporation. The liquefied product can be separated into two phases when a certain amount of polar water is added. The insoluble phase products were mostly composed of MLA and phenols (decomposed from lignin). The MLA can be
Table 3 The balances for the liquefaction of bamboo using an acid catalysta. Entry
1 2 3 a b c
Outputb (wt%)
Input (g)
Weight lossb (wt%)
Bamboo
Methanol
H2SO4
Liquefaction product yield
60.03 60.05 60.02
420.07 420.03 420.05
1.50 1.51 1.50
80.14 78.68 77.12
Liquefied product from bamboo MLA
Glycosides
Phenols
30.75 29.85 30.53
10.65 11.10 9.92
38.74 37.73 36.67
Methanol recoveryc
Gas
Residue
90.11 90.67 92.29
2.15 2.85 2.77
16.06 17.15 17.90
1.65 1.32 2.21
Reaction conditions: bamboo to methanol ratio of 1:7, 2.5 wt% catalyst, 200 °C,120 min. By weight of 60 g bamboo. By weight of 420 g methanol.
Liquefy Bamboo
Filter Bio-oil
Residue
Liquid samples
Evaporate Neutralize
Liquid product Water extraction
Soluble phase (Glycosides, Sugars) Methyl levulinate
Methyl levulinate CH2Cl2 phase Phenols Paste-like phase
CH2Cl2 extraction
Insoluble phase (Methyl Phenols)
levulinate,
Scheme 1. The production and separation of methyl levulinate via its extraction from a liquefied product derived from bamboo.
525
Abundance
J. Feng et al. / Applied Energy 154 (2015) 520–527
1.6x10
6
1.4x10
6
1.2x10
6
1.0x10
6
8.0x10
5
6.0x10
5
4.0x10
5
2.0x10
5
3.5. Proposed reaction pathway for the conversion of bamboo
Methyl levulinate
To explore the reaction pathway for the acid-catalyzed conversion of bamboo to MLA in subcritical methanol media, a series of carbohydrates: microcrystalline cellulose, corn starch, glucose and methyl glucoside, were selected as the substrates for the liquefaction reaction under their optimal reaction conditions. The various carbohydrates generated different yields of MLA (Table 4). The MLA yield from microcrystalline cellulose (16.75 wt%) was dramatically lower than that from glucose (53.47 wt%) under the same reaction conditions. It is well known that glucose units in cellulose molecules are strongly bound between cellulose chains and are relatively difficult to break down compared to other carbohydrates. In addition, cellulose is insoluble in the subcritical methanol medium and this causes poor substrate accessibility to the catalyst. Therefore, the cellulose conversion is hard to produce and less MLA was produced in the reaction. The yield of MLA from corn starch (48.73 wt%) was almost the same as the yield from glucose (53.47 wt%). This result was probably due to the solubility state of corn starch in subcritical methanol, and during the methanolysis it is easily transferred to methyl glucosides [27] as the intermediate product in the acid catalytic process. In addition, with MLG as the model compound, a substantially higher yield (61.04 wt%) of MLA was found compared
n-octanol
0.0
0
5
10
15
20
25
Retention time (min) Fig. 5. The GC result for the extracted MLA. The MLA content was calculated using the internal standard method by comparison with an authentic n-octanol reference sample.
further separated via extraction with CH2Cl2 (Scheme 1). As a result, the content of the extracted MLA was up to 86.3 wt% (measured by GC) with a yield of 26 wt% per 60 g bamboo. The GC result for the MLA extracted with n-octanol (internal standard substance of MLA) is shown in Fig. 5.
Table 4 The product yields for various substrates after liquefactiona. Materials
a b
Reaction temperature (°C)
Yieldb (wt%)
Liquefaction product yield (wt%)
Methanol recovery yield (wt%)
MLA
MMF
MLG
Glucose Methyl glucoside
180 180
99.17 99.53
53.47 61.04
10.14 3.37
6.32 7.59
85.90 92.33
Corn starch
180 120
92.23 99.78
48.73 0.17
5.38 2.31
13.07 95.01
90.21 91.57
Microcrystalline cellulose
200 180
86.93 85.63
23.86 16.75
8.47 8.72
15.31 51.57
89.32 87.04
Reaction conditions: material to methanol ratio of 1:7, 2.5 wt% catalyst, 120 min. MLG, MLA and MMF denote methyl glucoside, methyl levulinate and 5-methoxymethyl furfural, respectively.
CH2OH
CH2OH OH
H
CH2OH H
OH
H
H OH H
H OH H
O
O
O
H
H
OH n
OH
O
H
H OH
H
H O
HO OH O
OH
+
+H , CH3OH
OH HO
Celluclose
+
+H
-H2O
Soluble oligomers
OHC
OH
+
O
+H ,CH3OH
OCH3
-H2O
O
Methyl formate +
O
+H , -H2O
OH O
H OH H H
H
OH
O
O
OH H OH
OH H H H
O
O H
OHC
OH
HO
O
H
H
H
O
O
O
Furfural OH
OH H
OH H CH2OH H
H H
O H
H OH H H H O H H O OH OH H H OH OH O O H H O O
H OH
CH2OH H
H OH
+
H3C
CH3 O
5-Methoxymethyl furfural OH
H
O
+H , CH3OH H
CH2OH
OH H
O
+
OH OH H
O
O
O
By products
Hemicelluclose
Fig. 6a. Possible liquefaction pathway that converts biomass to methyl levulinate in methanol.
Methyl levulinate
526
J. Feng et al. / Applied Energy 154 (2015) 520–527
a
O O
H 3C
CH 3
O
o
OHC
OCH 3 O OH
H
O
O
HO HO
O
OH
CH 3
b
O O
H 3C
CH3
O
O H
O
OH O
HO HO
O
OH
CH 3
c
O O
H 3C
CH 3
O
OH
O H
O
HO HO
O
OH
O
CH3
O O
H 3C
d
CH3
O
OHC
o OCH 3
OH
O H
O
HO HO
OH
O
5
10
15
20
25
O
CH3
30
Fig. 6b. Comparison of the GC–MS chromatograms of products from different materialsa (a) glucose, (b) MLG, (c) corn starch, (d) microcrystalline cellulose. aReaction conditions: material to methanol ratios of 1:7, 2.5 wt% catalyst, 120 min.
with glucose (53.47 wt%). It can be concluded that the reaction pathway from microcrystalline cellulose, corn starch and glucose to MLA involves the intermediate formation of methyl glucosides. Previous studies [28–30] have shown that cellulose (a large polymer with long chains of glucose units) tends to decompose to monosaccharide in water, which can be further decomposed to form HMF and LA. In this paper, we propose a new reaction pathway that uses methanol as the reaction solvent (Fig. 6a). The possible reaction pathway for bamboo liquefaction using methanol as a solvent is based on the GC–MS analysis results (Table 2). The structures for cellulose and lignin are merely pictorial and do not imply a particular sequence. Fig. 6b shows that cellulose was first converted to MLG and then further dehydrated to MMF and methyl levulinate. A high methyl glucoside yield (about 95.01 wt%) was achieved from corn starch after using the acid catalysis at low temperatures (120 °C). Under the same conditions, a high methyl glucoside yield 51.57 wt% was also achieved from
microcrystalline cellulose at low temperatures (180 °C) (Table 4). This result indicated that the biomass was easily converted into the intermediate MLG [31] and then converted into methyl levulinate. During the reaction, the lignocellulosic biomass is first converted in a serial mode to methyl glucoside, dehydrated to MMF, and then further degraded to methyl formate and MLA [32].
4. Conclusions Methyl levulinate (MLA) can be produced from biomass by direct liquefaction under pressure in subcritical methanol. To obtain high yields of MLA from bamboo, various parameters need to be optimized, including the catalyst used, acid catalyst loading, methanol amounts, conversion temperature and reaction time. The sulfuric acid catalyst improved the liquefaction process the most and a 30.75 wt% MLA yield could be obtained from bamboo at
J. Feng et al. / Applied Energy 154 (2015) 520–527
200 °C after a reaction time of 120 min when the catalyst loading was 2.5 wt%. Different model carbohydrates were used to investigate the biomass to MLA reaction pathway. This investigation showed that biomass was easily converted to intermediate methyl glucosides and then converted to MLA. Generally, bamboo is a potential material for MLA production by direct pressurized liquefaction in subcritical methanol. Acknowledgements The authors gratefully acknowledge the financial support provided by the National Natural Science Foundation of China (31422013) for this investigation. The authors thank the Research Institute of New Technology and the Special Fund for Fundamental Research (CAFYBB2014ZD003 and CAFINT2013C05). We also appreciate the help given by Mrs. Li, Department of Chemical and Process Engineering, University of Canterbury, New Zealand, in modifying the manuscript. References [1] Balat M, Balat H. Progress in biodiesel processing. Appl Energy 2010;87(6):1815–35. [2] Naik SN, Goud VV, Rout PK, Dalai AK. Production of first and second generation biofuels: a comprehensive review. Renew Sust Energy Rev 2010;14(2):578–97. [3] Nandiwalea KY, Sonar SK, Niphadkar PS. Catalytic upgrading of renewable levulinic acid to ethyl levulinate biodiesel using dodecatung-stophosphoric acid supported on desilicated H-ZSM-5 as catalyst. Appl Catal A: Gen 2013;460:90–8. [4] Singh R, Bhaskar T, Dora S, Balagurumurthy B. Catalytic hydrothermal upgradation of wheat husk. Bioresour Technol 2013;149:446–51. [5] Demirbas MF. Biorefineries for biofuel upgrading: a critical review. Appl Energy 2009;86:S151–61. [6] Li Y, Zhang XD, Sun L, Zhang J, Xu HP. Fatty acid methyl ester synthesis 3+ catalyzed by solid superacid catalyst SO2 4 /ZrO2–TiO2/La . Appl Energy 2010;87:156–9. [7] Xu J, Jiang J, Hse C, Shupe TF. Effect of methanol on the liquefaction reaction of biomass in hot compressed water under microwave energy. Energy Fuels 2013;27:4791–5. [8] Chang C, Xu G, Jiang X. Production of ethyl levulinate by direct conversion of wheat straw in ethanol media. Bioresour Technol 2012;121:93–9. [9] Werpy T, Petersen G. Top value added chemicals from biomass volume Iresults of screening for potential candidates from sugars and synthesis gas. 2004: p. 45–8. [10] Lee A, Chaibakhsh N, Rahman MBA, Basri M, Tejo BA. Optimized enzymatic synthesis of levulinate ester in a solvent-free system. Ind Crops Prod 2010;32(3):246–51. [11] Digman B, Joo HS, Kim DS. Recent progress in gasification/pyrolysis technologies for biomass conversion to energy. Environ Prog Sust Energy 2009;28(1):47–51.
527
[12] Digman B, Joo HS, Kim DS. Valeric biofuels: a platform of cellulosic transportation fuels. Angew Chem Int Ed 2010;49(26):4479–83. [13] Gürbüz EI, Alonso DM, Bond JQ, Dumesic JA. Reactive extraction of levulinate esters and conversion to c-valerolactone for production of liquid fuels. Chem Sust Chem 2011;4(3):357–61. [14] Hayes DJ. An examination of biorefining processes, catalysts and challenges. Catal Today 2009;145(1):138–51. [15] Chen B, Li F, Huang Z. Integrated catalytic process to directly convert furfural to levulinate ester with high selectivity. ChemSusChem 2014;7(1):202–9. [16] Lin CP, Lin L, Li H, Yang Q. Conversion of carbohydrates biomass into levulinate esters using heterogeneous catalysts. Appl Energy 2011;88:4590–6. [17] Mascal M, Nikitin EB. Comment on processes for the direct conversion of cellulose or cellulosic biomass into levulinate esters. ChemSusChem 2010;12:1349–51. [18] Rowley RL, Wilding WV, Oscarson JL, Zundel NA, Marshall TL, Daubert TE, et al. Dipper data compilation of pure compound properties. New York: Design Institute for Physical Properties American Institute of Chemical Engineers, AIChE; 2004. [19] Windom BC, Lovestead TM, Mascal M. Advanced distillation curve analysis on ethyl levulinate as a diesel fuel oxygenate and a hybrid biodiesel fuel. Energy Fuels 2011;4:1878–90. [20] Mark M, Nikitin EB. High-yield conversion of plant biomass into the key valueadded feedstocks 5-(hydroxymethyl) furfural, levulinic acid, and levulinic esters via 5-(chloromethyl) furfural. Green Chem 2010;12:370–3. [21] Liu Y, Liu CL, Wu HZ. An efficient catalyst for the conversion of fructose into methyl levulinate. Catal Lett 2013;12:1346–53. [22] Tominaga K, Mori A, Fukushima Y, Shimada S, Sato K. Mixed-acid systems for the catalytic synthesis of methyl levulinate from cellulose. Green Chem 2011;4:810–2. [23] Wu X, Fu J, Lu XY. One-pot preparation of methyl levulinate from catalytic alcoholysis of cellulose in near-critical methanol. Carbohyd Res 2012;358: 37–9. [24] Jeremy SL, Jacqueline MR, David MA, Jeehoon H, Youngquist JT, Christos TM, et al. Nonenzymatic sugar production from biomass using biomass-derived cvalerolactone. Science 2014;17:277–80. [25] Zhou LP, Zou HJ, Nan JX. Conversion of carbohydrate biomass to methyl levulinate with Al2(SO4)3 as a simple, cheap and efficient catalyst. Catal Commun 2014;50:13–6. [26] Joshi H, Moser BR, Toler J. Ethyl levulinate: a potential bio-based diluent for biodiesel which improves cold flow properties. Biomass Bioenergy 2011;7: 3262–6. [27] Yabushita M, Kobayashi H, Fukuoka A. Catalytic transformation of cellulose into platform chemicals. Appl Catal B: Environ 2014;145:1–9. [28] Girisuta B, Heeres HJ, Janssen LPBM. A kinetic study on the decomposition of 5hydroxymethylfurfural into levulinic acid. Green Chem 2006;8:701–9. [29] Jung CD, Yu JH, Eom IY, Hong KS. Sugar yields from sunflower stalks treated by hydrothermolysis and subsequent enzymatic hydrolysis. Bioresour Technol 2013;138:1–7. [30] Meine N, Rinaldi R, Schüth F. Solvent-free catalytic depolymerization of cellulose to water-soluble oligosaccharides. Chem Sust Chem 2012;5(8): 1449–54. [31] Deng WP, Liu M, Zhang QH, et al. Acid-catalysed direct transformation of cellulose into methyl glucosides in methanol at moderate temperatures. Chem Commun 2010;46:2668–70. [32] Yin SD, Tan ZC. Hydrothermal liquefaction of cellulose to bio-oil under acidic, neutral and alkaline conditions. Appl Energy 2012;92:234–9.