Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil

Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil

G Model ARTICLE IN PRESS CATTOD-10018; No. of Pages 9 Catalysis Today xxx (2016) xxx–xxx Contents lists available at ScienceDirect Catalysis Toda...

1MB Sizes 0 Downloads 27 Views

G Model

ARTICLE IN PRESS

CATTOD-10018; No. of Pages 9

Catalysis Today xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

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

Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil Peigao Duan a,∗ , Caicai Zhang a , Feng Wang a,∗ , Jie Fu b , Xiuyang Lü b , Yuping Xu a , Xianlei Shi a a College of Physics and Chemistry, Department of Applied Chemistry, Henan Polytechnic University, No. 2001, Century Avenue, Jiaozuo, Henan 454003, PR China b Key Laboratory of Biomass Chemical Engineering of Ministry of Education, Zhejiang University, Hangzhou 310027, China

a r t i c l e

i n f o

Article history: Received 25 October 2015 Received in revised form 28 December 2015 Accepted 19 January 2016 Available online xxx Keywords: Activated carbon Duckweed Catalytic upgrading Upgraded bio-oil Supercritical water

a b s t r a c t This study examined a two-stage (noncatalytic pretreatment followed by catalytic upgrading) hydrothermal processing of crude bio-oil produced from the hydrothermal liquefaction of duckweed. The activities of six activated carbons (ACs)-pine wood AC, coconut shell AC, bamboo stem AC, apricot pit AC, peach pit AC, and coal AC-toward the deoxygenation and denitrogenation of the pretreated duckweed bio-oil were determined in supercritical water at 400 ◦ C for 1 h with the addition of 6 MPa of H2 and 10 wt% AC. All of the ACs exhibited activity similar to Ru/C toward the denitrogenation and deoxygenation of the pretreated duckweed bio-oil. Of the ACs tested, bamboo stem AC produced an upgraded bio-oil with the highest yield (76.3 wt%), the highest fraction (90.13%) of material boiling below 350 ◦ C, and the highest energy density (44.1 MJ/kg). Decreased ash and acidic groups in the pre-treated AC disfavored the production of upgraded bio-oil but aided denitrogenation and desulfurization. The ACs are suspected to leach ions and weak acids into the reaction solution, which would catalyze denitrogenation and desulfurization. The gases mainly consisted of unreacted H2 , CO2 and CH4 together with small amounts of Cx Hy (x ≤ 5, y ≤ 12) hydrocarbon gases produced from the cracking of the upgraded bio-oil. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Bio-oils, which are produced from the thermochemical conversion of biomass, have been recognized as a potential alternative renewable energy source to petroleum-based transportation fuels. Developing bio-oil from renewable biomass would provide environmental and social benefits. As a novel feedstock with high biomass yield and high protein and starch content, duckweed has recently gained increasing attention from researchers and governments for biofuel production [1]. Thermochemical conversion (e.g., pyrolysis and hydrothermal liquefaction (HTL)) is one the main avenues for converting duckweed into a tar-like material (crude bio-oil), which has a considerably higher energy density than that of duckweed biomass [2–6]. However, this type of crude bio-oil is typically too viscous to flow at room temperature, is chemically unstable, and contains quantities of N, O, and S that are significantly higher than the American Society for Testing and Materials

∗ Corresponding authors. Fax: +86 391 3987811. E-mail addresses: [email protected] (P. Duan), [email protected] (F. Wang).

(ASTM) requirement [7]. Thus, this type of crude bio-oil cannot be directly used as a transportation fuel. Further treatments regarding the reduction of viscosity and the removal of heteroatoms from the crude duckweed bio-oil to form hydrocarbons are necessary. To date, there have been few reports on the heterogeneous catalytic upgrading of crude duckweed bio-oil. Only two recent articles from the authors’ lab have been reported [8,9]. In these previous studies, the effects of different commercially available materials, such as Ru/C, Pd/C, Pt/C, Pt/␥-Al2 O3 , Pt/C-sulfide, Rh/␥-Al2 O3 , activated carbon (AC), MoS2 , Mo2 C, Co-Mo/␥-Al2 O3 , and zeolite, on the product fraction yields and the deoxygenation, denitrogenation, and desulfurization of crude duckweed bio-oil were examined in subcritical water. All of the materials, including AC, displayed catalytic activity for deoxygenation, denitrogenation and desulfurization of the crude duckweed bio-oil (relative to noncatalytic treatment). These results motivated the present investigation of whether AC alone could be an effective catalyst for the deoxygenation, denitrogenation, and desulfurization of crude duckweed bio-oil. Water above its boiling point is known as subcritical or supercritical (above the critical point, SCW) water, which typically exhibits many desirable properties, such as a low dielectric

http://dx.doi.org/10.1016/j.cattod.2016.01.046 0920-5861/© 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046

G Model

ARTICLE IN PRESS

CATTOD-10018; No. of Pages 9

P. Duan et al. / Catalysis Today xxx (2016) xxx–xxx

2

Table 1a Proximate and ultimate analyses (wt%, as-received basis) of duckweed. Proximate and ultimate analyses M

A

V

FC

L

P

CH

C

H

N

O

S

13.2

18.0

71.5

10.5

9.0

23.1

26.2

35.4

4.8

3.7

32.3

1.0

M: moisture; A: ash; V: volatiles; FC: fixed carbon; L: lipid; P: protein; CH: carbohydrate.

Table 1b Inorganic composition (wt%, as-received basis) of duckweed. Na

Mg

Al

Si

P

Cl

K

Mn

Fe

Ni

Cu

Zn

Br

Sr

Ca

Ti

Cr

Mo

Ru

0.45

0.61

0.28

1.89

1.17

3.91

6.12

0.16

0.71

0.01

0.01

0.02

0.02

0.08

8.24

0.04

0.01

0.03

0.03

constant, high diffusivity, and adjustable solvency, making it a potential medium for the processing of crude bio-oil [10]. There are reports of the use of charcoals and ACs as catalysts for the deoxygenation of fatty acids [11] in SCW and for upgrading heavy oils without solvent or in supercritical toluene [12–14]. The results of these studies demonstrated that ACs are effective and suitable for upgrading heavy oil, especially heavy oils such as Maya that contain a large amount of heavy metals. ACs have an affinity for heavy hydrocarbon compounds, an adsorption selectivity for asphaltenes, and an excellent ability to restrict coke formation during the upgrading of heavy oils [12–14]. ACs are also used for the adsorptive removal of N and S from diesel oils [15–17]. However, to the best of the authors’ knowledge, there have been no reports on the use of ACs to upgrade crude bio-oils in hydrothermal media, apart from the authors’ single experiment [9]. If sufficiently active, these materials could be attractive as low-cost catalysts for this transformation. In this study, several different commercial ACs were used for the catalytic upgrading of crude bio-oil derived from the HTL of duckweed to fuel-range hydrocarbons in SCW with added H2 .

2. Experimental 2.1. Materials Duckweed (Lemna minor) was collected from rice fields in Hebei Province in northern China and was grown in freshwater in summer for an average of four weeks. After four weeks of growth, the duckweed was harvested, washed with deionized water to remove impurities, and sun-dried. It was dried again in an oven at 110 ◦ C for 12 h prior to use and then crushed in a multi-functional pulverizer to a #100-mesh particle size. The proximate and ultimate analyses of duckweed are listed in Tables 1a and 1b. The inorganic composition of the duckweed was measured by X-ray fluorescence (XRF) using a Bruker S8 TIGER XRF spectrometer, and the results are provided in Tables 1a and 1b. Five biomass-derived ACs and one coal-derived AC were used in the present study. The ACs were commercially available from a local tap water purification company. The biomass-derived ACs were prepared from pine wood, coconut shells, apricot pits, bamboo stems, and peach pits. High-pressure and corrosion-resistant batch reactors (164, 200, and 37 mL) were used to conduct the HTL, non-catalytic hydrotreatment (pretreatment), and catalytic upgrading experiments, respectively. The reactor was un-mixed. Prior to their use in the experiments, these reactors were loaded with water and seasoned at 400 ◦ C for 4 h to remove any residual organic material from the reactors and expose the fresh metal walls to SCW. These reactors were heated using a molten-salt bath that consisted of potassium nitrate and sodium nitrate at a mass ratio of 5:4.

Freshly deionized water was used throughout the experiments. Dichloromethane was obtained from Aladdin, China, at a purity of 99.9% and was used as received.

2.2. Characterization of ACs All of the ACs were dried in an oven at 110 ◦ C for 12 h and crushed to #500 mesh prior to use. No further activation was performed because the authors sought to identify catalysts that would be active in their as-received form. The ash content is determined by ignition of a known weight of the food at 600 ◦ C until all carbon is removed. The residue is the ash, which represents the inorganic constituents of the biomass. The fixed carbon content was determined by subtracting the summation of the volatile matter and ash contents from the total sample mass. The total BET surface area, pore volume, and pore size distribution of all of the ACs were measured by N2 isothermal (77 K) adsorption (Micromeritics, ASAP 2020). The surface area was calculated according to the Brunauer–Emmett–Teller (BET) equation. The pore size distribution was obtained from the Horvath–Kawazoe differential pore volume plot, and the pore size was calculated by the Horvath–Kawazoe method (<2 nm) and the Barrett–Joyner–Halenda (BJH) method (>2 nm). The surface structures of the ACs were observed by scanning electron microscopy (SEM). A thin layer of sample was mounted on a copper sample holder using double-sided carbon tape and was coated with gold (10 nm thickness) to make the samples conductive. SEM studies were carried out using a scanning electron microscope (JEOL JSM-6390LV) at an acceleration voltage of 15 kV. The inorganic compositions of the ACs were determined by Xray fluorescence (XRF) using a Bruker S8 TIGER XRF spectrometer. C, H, O, N, and S were detected using a Flash2000 (CHNS/O, Thermo Fisher Scientific, USA) elemental analyzer. Fourier transform infrared spectroscopic analysis (FT-IR) was performed on a Vertex 70 FT-IR spectrometer (Bruker Optics Corporation) to determine the chemical functionality of the AC. A few milligrams of sample were mixed with approximately 200 mg of KBr powder. The pellet preparation involved grinding the sample with KBr using an agate mortar and pestle and using a hydraulic press and die to create a thin and transparent disk. FT-IR spectra (resolution: 4 cm−1 , scan: 254, range: 4000–400 cm−1 ) were recorded at a controlled ambient temperature (25 ◦ C). A background spectrum was also collected under identical conditions. The XPS analysis was carried out on an ESCALAB 250Xi (Thermo Scientific, USA) spectrometer using monochromated Al K␣X-rays with an energy of 1486.6 eV. The source operated at 12 kV and 20 mA. The measurements were performed under a vacuum at a pressure that was lower than 5 × 10−10 Pa. The survey scans were collected from 0 to 1400 eV. The high-resolution scans were performed over 280–295 eV (C 1s), 525–540 eV (O 1s), 392–408 eV (N

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046

G Model CATTOD-10018; No. of Pages 9

ARTICLE IN PRESS P. Duan et al. / Catalysis Today xxx (2016) xxx–xxx

1s) and 158–170 eV (S 2p) with a pass energy of 20 eV. For the calibration, the C 1s electron bond energy corresponding to graphitic carbon was referenced as equal to 284.8 eV. A Shirley-type background was subtracted prior to quantification. After the baseline was subtracted, curve-fitting was performed using an asymmetric Gaussian–Lorentzian sum function fitting program under an optimized peak shape. This peak-fitting procedure was iterated until an acceptable fit was obtained. pH was measured using a modified ASTM method [18]. Onegram samples of carbon were added to 10 mL of deionized water and stirred. The pH of the resulting solution was measured after 4 h using a pH electrode. The oxygenated surface groups were determined according to the Boehm titration method [19]. One gram of carbon was placed in 50 mL of the following 0.05 M solutions: sodium hydroxide, sodium carbonate, and hydrochloric acid. The vials were sealed and shaken for 24 h; then, 5 mL of each filtrate was pipetted, and the excess base and acid were titrated with HCl and NaOH, respectively. The number of acid sites of various types was calculated under the assumption that NaOH neutralizes carboxylic, phenolic, and lactonic groups, whereas Na2 CO3 neutralizes carboxylic and lactonic groups. The number of surface base sites was calculated from the amount of hydrochloric acid that reacted with the carbon. 2.3. Duckweed liquefaction In total, 400 g of duckweed powder was liquefied in a 164 mL batch reactor in ten independent runs, which produced approximately 80 g of crude duckweed bio-oil. The operating procedure was nearly identical to that reported previously [9]. In each run, 40 g of dry duckweed powder and 50 mL of freshly deionized water were loaded into the reactor, which was then tightly sealed. The reactor was heated to 350 ◦ C for 30 min. The preheating time was approximately 20 min. The crude duckweed bio-oil was extracted with dichloromethane. The dichloromethane in the extract was vaporized using a rotary evaporator. The remaining material was the crude duckweed bio-oil. More details on the extraction process are available in a previous publication [4]. 2.4. Pretreatment The crude duckweed bio-oil underwent hydrothermally noncatalytic hydrotreatment in the 200 mL autoclave reactor. The crude duckweed bio-oil (30 g) and freshly deionized water (21 mL) were loaded into the reactor. The air inside the reactor was displaced with 6 MPa of H2 . The reaction was conducted at 350 ◦ C for 4 h. The preheating time was approximately 25 min. After 4 h of reaction duration, the reactor was removed from the molten salt tank and immersed in an ice water bath for approximately 15 min to stop the reaction. The subsequent procedure was the same as the method for crude duckweed bio-oil recovery from the HTL described in Ref. [4]. The pretreatment experiments were repeated in ten runs under identical conditions, and approximately 230 g of pretreated bio-oil was obtained, which resulted in a 76.7 wt% yield. The mass loss was attributed to the formation of char, gas, and water-soluble products (WSPs). Other causes of mass loss included the inevitable multiple sample transfers required for the separation of the product fractions and the use of evaporation to remove the solvents from the bio-oils. 2.5. Catalytic upgrading The catalytic upgrading experiments were carried out in a 37 mL autoclave reactor. The procedure was nearly identical to that described in our previous publication [9]. The only difference in the present study was that pretreated bio-oil was used instead of

3

crude duckweed bio-oil. In a typical run, 3.0 g of pretreated bio-oil, 0.3 g of AC, 1.5 mL of freshly deionized water and 6 MPa of H2 were loaded into the reactor. The reactions were performed at 400 ◦ C for 1 h. For comparison, one additional experiment with Ru/C was also performed under identical conditions as those for AC. Pretreated bio-oil upgrading reactions were carried out by placing the reactor in a molten salt tank preheated to 450 ◦ C. After approximately 10 min, the reactor reached 400 ◦ C, at which point the reaction time was set to zero. The temperature was controlled at 400 ± 5 ◦ C using an Omega temperature controller connected to a thermocouple residing in a thermowell of the reactor. When the reaction had proceeded for 1 h, the reactor was removed from the molten salt tank and immersed in an ice water bath for approximately 15 min to stop the reaction. The reactor was weighed before and after venting the gas to estimate the gas yield. The precision of the balance used to measure the reactor’s weight was 0.01 g. The gas production was evaluated by subtracting the initial loaded weight (0.19 g) of H2 from the total weight of gas after the reaction. The reactor was opened, and 40 mL of dichloromethane was added to recover the upgraded oil fraction. The reactor was washed two more times with additional 40 mL aliquots of dichloromethane. All of the dichloromethane extracts were filtered before they were transferred to a separatory funnel for organic and aqueous phase separation. After filtration, the filter paper together with the solid residue were dried in an oven at 110 ◦ C for 12 h and then weighed. The non-catalyst solid residue was defined as coke, the amount of which was evaluated by subtracting the weight of the initially loaded AC from that of the solid residue. The separated organic layer was dried for approximately 30 min using 5.0 g of anhydrous magnesium sulfate to absorb any residual water, and the magnesium sulfate was then removed by filtration. The dichloromethane in the dried organic layer was evaporated. The remaining material was upgraded bio-oil, and its amount was determined gravimetrically. The separated aqueous phase was vaporized at 60 ◦ C to recover the non-volatile WSPs. The yield of each product fraction was calculated as its mass divided by the mass (dry basis) of pretreated bio-oil loaded into the reactor. Duplicate independent runs were conducted under nominally identical conditions to quantify the uncertainties in the experimental results. The results reported herein represent the mean values of two independent trials. Uncertainties (when provided) are reported as the experimentally determined standard deviations.

2.6. Product analysis Elemental analysis, including analysis of C, H, and O, was performed on a Flash2000 elemental analyzer (Thermo scientific, USA), which utilized a technique based on a modification of the classic Pregl and Dumas method [20]. The total S and N contents were analyzed according to ASTM D5453-12 [21] and D4629-12 [22], respectively. The analyses were performed in duplicate, and the mean values for the two independent analyses are provided. The Dulong formula (i.e., higher heating value (HHV), (MJ/kg) = 0.338C + 1.428(H-O/8) + 0.095S) was used to estimate the HHV of the bio-oils based on their elemental compositions. An Agilent Technologies 7890A GC equipped with an autosampler, auto-injector and mass spectrometric detector (5975C) was used to analyze the bio-oil. An Agilent J&W DB-5HT non-polar capillary column (30 m length, 0.25 mm i.d., 0.10 ␮m film thickness) separated the constituents. The column was initially held at 40 ◦ C for 4 min. The temperature was ramped up to 300 ◦ C at 4 ◦ C/min and isothermally held for 4 min, yielding a total runtime of approximately 73 min. Helium flowing at 3 mL/min served as the carrier gas. A Wiley mass spectral library was used for compound identification.

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046

G Model

ARTICLE IN PRESS

CATTOD-10018; No. of Pages 9

P. Duan et al. / Catalysis Today xxx (2016) xxx–xxx

4 Table 2 Catalysts characterization. Catalyst

BET area m2 g−1

Micropore volume cm3 g−1

Pore size distribution (%) <2 nm

2–50 nm

>50 nm

Pine-AC Coconut-AC Bamboo-AC Coal-AC Peach-AC Apricot-AC Ru/C

453 604 31 280 123 276 966

0.153 0.254 0.012 0.116 0.048 0.114 0.400

40.81 73.66 21.16 38.56 39.47 57.64 44.15

52.09 23.80 66.89 55.60 54.53 39.90 52.42

7.10 2.54 11.95 5.84 6.00 2.46 3.43

Table 3 pH measurement and Boehm titration (mmol/g). Catalyst type

pH

Lactonic groups

Phenolic groups

Carboxylic groups

Basic groups

Acidity

Total

Pine-AC Coconut-AC Bamboo-AC Coal-AC Peach-AC Apricot-AC Bamboo-AC (pre-treated) Ru/C

12.16 10.09 10.25 12.32 9.38 9.24 8.47 8.96

0.55 0.44 0.33 0.11 0.14 0.41 0.41 0.50

0 0.01 0 0.39 0.13 0.19 0 0

0.96 0.89 0.78 0.77 0.80 0.84 0.84 0.65

3.94 3.44 3.33 3.90 3.33 3.46 3.20 3.60

1.51 1.34 1.11 1.27 1.07 1.44 1.25 1.15

5.45 4.78 4.44 5.17 4.40 4.90 4.45 4.75

Table 4 XPS elemental analysis (% surface atomic composition).

Pine-AC Coconut-AC Bamboo-AC Coal-AC Peach-AC Apricot-AC

C

N

S

O

88.30 89.30 85.01 90.50 82.82 82.61

0.58 0.47 0.94 0.63 0.94 0.78

0.23 0.19 0.28 0.39 0.27 0.25

10.85 10.90 13.66 8.48 15.97 16.36

Thermogravimetric analysis (TGA) of the bio-oils was performed on an SDT Q600 simultaneous DSC-TGA instrument in a nitrogen atmosphere. Bio-oil samples were heated from room temperature to 780 ◦ C at a heating rate of 10 ◦ C/min. The gas flow rate was 20 mL/min. FT-IR and total acid number (TAN) analyses were performed as described previously [4,23]. Moisture was detected by volumetric Karl Fischer titration [24] and was quantified based on the volume of Karl Fischer reagent consumed. 3. Results and discussion 3.1. Characterization of ACs The BET area, pore size distribution, and pore volume of all of the ACs are summarized in Table 2. The corresponding properties of the commercially available Ru/C material are also included in Table 2. The coconut shell AC had the largest surface area (604 m2 g−1 ) and pore volume (0.254 cm3 g−1 ), whereas the bamboo stem AC had the lowest surface area and pore volume (31 m2 g−1 and 0.012 cm3 g−1 , respectively). In contrast, the BET area and pore volume of Ru/C were significantly higher than those of the coconut shell AC due to pretreatment of the support during catalyst preparation [25]. Table 2 shows that micropores and mesopores accounted for 90% of the ACs. Micropores accounted for 74% of the coconut shell AC, whereas mesopores accounted for 67% of the bamboo stem AC. The mesopores of AC reportedly play an important role in the effective conversion of heavy hydrocarbons in the upgrading of heavy oils [12]. The Ru/C material exhibited a similar pore size distribution as the ACs. The surface morphologies of the ACs were characterized by

scanning electron microscopy. Fig. S1 presents representative SEM micrographs. All of the ACs were highly porous, which is consistent with their high surface areas. The organic and mineral compositions of all of the ACs are listed in Table S1. The most abundant element in all of the ACs was carbon, which comprised more than 80% of the coconut shell and bamboo stem ACs. Si, Ca, Al, Fe, Cl, and K were enriched in all of the ACs. The ACs also contained small amounts of transition metals, which might play catalytic roles during the upgrading of bio-oil. Table 3 shows the results from the pH measurement and Boehm titration of all of the ACs. The pH of activated carbon is a measure of its acidity or basicity. Boehm titration is a chemical method to identify the oxygen surface groups on carbon materials. The ACs usually contain a certain amount of inorganic salts and metal oxides, which would dissolve in water during the extraction process, contributing basic ions to the aqueous solution. Essentially, the higher the ash content of the AC, the higher the contact pH. Pine and coal ACs had the highest pH, while peach and apricot pits had the lowest pH. This leaching pH might influence the activity of the AC in the upgrading of bio-oil. According to Boehm, the groups were classified as carboxylic, phenolic, lactonic, and basic. Both acidic and basic groups were present on the surface of all the ACs. However, the content of the total basic groups was higher than that of the total acidic groups. Because the ACs were not acid-washed before the Boehm titration, the number of basic groups detected using HCl may have been overestimated. Carboxylic groups were the dominant acidic groups in all the ACs. Only coal, peach pit and apricot pit ACs contained a certain amount of phenolic groups. Fig. 1 shows the FT-IR spectra of all the ACs, all of which, especially the coconut pit AC, show a wide transmittance band at

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046

G Model

ARTICLE IN PRESS

CATTOD-10018; No. of Pages 9

P. Duan et al. / Catalysis Today xxx (2016) xxx–xxx

C O (carbonyl or quinone), and O C O (carboxyl or ester), respectively. XPS provided results that are consistent with those of the Boehm titration.

1

Transimittance/%

5

0.9 0.8 0.7

3.2. Effect of catalysts on the product fraction yields

0.6 0.5

Pine

Coconut

Bamboo

Coal

Peach

Apricot

0.4 3900

3400

2900

2400

1900

1400

900

400

Wavenumber/cm-1 Fig. 1. FT-IR spectra of different ACs.

100000

Counts / s

80000

Bamboo AC

Apricot AC

Peach AC

Coconut AC

Coal AC

Pine AC

60000 40000 20000 0 300

60000

Bamboo AC Peach AC Coal AC

50000 Counts / s

295

290 Binding Energy (eV)

285

280

531

526

Apricot AC Coconut AC Pine AC

40000 30000 20000 10000 0 546

541

536

Binding Energy (eV) Fig. 2. X-ray photoelectronspectra of the activated carbon: (a) C(1s) and (b) O(1s) spectra.

3200–3600 cm−1 with a maximum at approximately 3400 cm−1 . This band can be ascribed to the O H stretching mode of hydroxyl groups, which is consistent with the results of the Boehm titration. The small band at approximately 1700 cm−1 is assigned to the C O stretching vibrations of ketones, aldehydes, lactones or carboxyl groups. A broad band at 1000–1300 cm−1 is usually observed with oxidized carbons and is assigned to C O stretching in acids, alcohols, phenols, and esters. More detailed information on the surface chemistry of the ACs is derived from the XPS analysis. The contents of elements (in at.%) detected on the surface are listed in Table 4. Compared with the C, N, S, and O values detected by elemental analysis, XPS detected higher C, N, and O contents and a lower S content. A high O ratio on the surface would increase the polar (hydrophilic) character of the AC. Representative C(1s) and O(1s) spectra are shown in Fig. 2. Fig. 2 shows a main narrow C(1s) peak accompanied by a broad, high binding energy component, indicating the presence of more than one type of carbon. The binding energy shoulder at 284.8 eV is ascribed to an sp3 carbon. The binding energies at 286.3, 287.4, and 288.7 eV correspond to C-(O, H) (phenolic, alcoholic, or etheric),

The pretreated bio-oil was converted to four product fractionsoil, solid, gas, and WSP-after it was treated at 400 ◦ C for 1 h with an additional 6 MPa of H2 in SCW. The WSPs were always present in the lowest amount (less than 0.3 wt%) in all of the experiments. Thus, their yields are not provided in the present study. Fig. 3 shows these results. The mass balance closure varied between 99.2 ± 1.8% and 104.7 ± 0.2% depending on the AC catalyst type. Mass balance closures below 100% could arise when the catalyst had a low carbon content. The other two reasons for the low mass balance were attributed to the multiple sample transfers and the solvent evaporation process for the upgraded oil recovery. Mass balance closures exceeding 100% might have been due to the partial incorporation of H2 O and AC into the reaction products. Participation of H2 O was also observed during the upgrading of duckweed bio-oil in subcritical water [8,9]. The upgraded bio-oils obtained by upgrading pretreated bio-oils contained a large proportion of a light-oil fraction, which would be lost during the solvent removal process. Thus, the yields were reported at the low bound. The yield of upgraded oil benefited from the addition of ACs because it increased from 68.2 wt% for noncatalytic upgrading to greater than 70 wt% for most of the ACs. The upgraded oil yield (76.3 wt%) produced with bamboo stem AC was even higher than the yield (74.7 wt%) produced with Ru/C. The presence of AC or Ru/C possibly promoted the hydrogenation of the pretreated bio-oils and thus inhibited the polymerization or condensation of the upgraded oil. However, it was difficult to determine the relationship between the yield of the upgraded oil and the BET area, pore size distribution and pore volume of the ACs. One additional upgrading experiment was performed with bamboo AC that had been pre-treated at 400 ◦ C for 1 h in SCW. The characterization of the pre-treated bamboo AC is listed in Table 3. The pre-treated bamboo AC contained a lower amount of ash and acidic groups than its original form. This pre-treated bamboo AC produced a lower upgraded oil (69.0 wt%) yield than that of its original form (76.3 wt%). Therefore, it is suspected that the decreased ash and acidic groups in the pre-treated AC disfavored the production of upgraded bio-oil. The conditions for producing a lower bio-oil yield could be desirable if they also produced an upgraded oil of higher quality (e.g., if catalytic upgrading produced a lower yield of upgraded oil because it had a lower heteroatom content and higher volatile content). Furthermore, the upgraded oil produced without an AC catalyst had a higher visual viscosity than the oil produced with ACs. The solid yield from the non-catalytic reaction (30.1 wt%) was significantly higher than that from the reactions catalyzed by the ACs. The solid mainly consisted of char, which was formed by several related mechanisms, including the dehydrogenation, polymerization, and condensation of asphaltenes and heavy aromatic compounds contained in the bio-oil [26]. The coke formation was dependent on the pore size of the ACs. A larger proportion of mesopores in the ACs resulted in a lower coke yield, which is consistent with a previous study [12]. The mesopores of AC likely play an important role in the effective conversion of heavy compounds into lighter fractions, restricting coke formation. Moreover, the solvation and dilution characteristics of SCW also suppress coke formation [27]. Further reduced coke formation was observed with Ru/C. The presence of the noble metal may have promoted the hydrogenation of the coke precursor, thus retarding the polymer-

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046

G Model

ARTICLE IN PRESS

CATTOD-10018; No. of Pages 9

P. Duan et al. / Catalysis Today xxx (2016) xxx–xxx

6

Table 5 Elemental composition (wt.%, dry basis) of duckweed, crude bio-oil, pretreated bio-oil and upgraded bio-oils (UBO) produced with different AC catalysts.

Duckweed Crude bio-oil Pretreated oil UBO-no cat. UBO-pine AC UBO-coconut AC UBO-bamboo AC UBO-coal AC UBO-peach AC UBO-apricot AC UBO-Ru/C

TAN

C

H

N

S(ppm)

O

HHV (MJ/kg)

S-ER%

O-ER%

– 40.51 28.13 26.57 24.20 19.32 21.90 21.92 25.56 17.84 9.81

35.4 75.7 79.8 85.1 85.8 86.2 85.5 85.7 85.5 84.9 85.5

4.80 8.7 8.9 10.2 10.4 10.7 10.9 10.7 10.5 10.8 10.8

3.70 2.20 1.81 1.79 1.42 1.24 1.32 1.29 1.20 1.46 1.34

1020.1 724.5 101.7 173.2 109.0 113.1 91.5 83.3 141.8 94.9 152.0

32.3 6.5 3.5 2.2 2.4 1.9 2.2 2.1 2.0 2.1 2.1

14.08 36.89 39.06 42.94 43.43 44.08 44.08 43.88 43.55 43.75 43.96

55.80 82.07 74.98 77.39 80.02 86.11 80.10 73.36 79.64 84.05

55.80 45.80 34.34 35.44 36.65 39.43 36.68 33.60 36.47 38.49

ization and condensation reactions. Therefore, the coke control mechanisms for the AC and noble metal were different. Thermal cracking allows for the breakdown of large hydrocarbon molecules into smaller, more useful hydrocarbon molecules and their further breakdown into gases. Therefore, a certain amount of gaseous products were also derived during the upgrading process. Compared with the other ACs, bamboo stem AC and coal AC produced higher gas yields due to their high performance in the conversion of heavy compounds into light volatiles, as mentioned above. Catalysis resulting from mineral metals in these two ACs cannot be ruled out. It appears possible (perhaps likely) that the activated carbon leaches ions into solution that act as homogeneous catalysts because the gas yield from the catalyzed reaction was larger than that from the uncatalyzed reaction. The gases mainly consisted of unreacted H2 , CO2 and CH4 , together with small amounts of Cx Hy (x ≤ 5, y ≤ 12) hydrocarbon gases produced from the cracking of the upgraded oil. The CO2 could be formed from reactions such as steam reforming and water-gas shifts. The CO2 could also be derived from the hydrothermal decarboxylation of fatty acids, and the Cx Hy (x ≤ 5, y ≤ 12) hydrocarbon gases could be formed via cracking reactions.

3.3. Elemental analysis Table 5 compares the TAN, elemental compositions, and estimated HHVs of the upgraded oils produced from treating the pretreated bio-oil at 400 ◦ C for 1 h with added H2 in the presence or absence of different ACs. The step-wise energy recovery refers to the percentage of energy in the crude bio-oils, pretreated bio-oils, and upgraded bio-oils relative to those quantities in the duckweed, crude bio-oils, and pretreated bio-oils, respectively. The total energy recovery (T-ER) refers to the percentage of energy in the upgraded bio-oils relative to those quantities in the initial duckweed. HTL produced a crude bio-oil that had higher C and H contents and lower O and S contents than the original duckweed biomass feedstock. Pretreatment further increased the C and H contents and decreased the O, N, and S contents in the pretreated bio-oil, which would reduce the risk of catalytic poisoning of the catalyst in the subsequent upgrading process. Pretreatment also decreased the viscosity of the crude bio-oil, decreasing the possibility of coke formation. As listed in Table 5, treating the pretreated bio-oil in the absence of AC increased the H content and decreased the N and O contents of the upgraded oil. However, the S content of the pretreated biooil was higher than that of the pretreated oil, suggesting that the sulfur-containing compounds are stable under severe hydrothermal conditions. The S contents are provided in ppm because they are very low in the upgraded bio-oil. In contrast, the upgraded oil produced with ACs had lower N, S, and O contents than the upgraded oil produced from the uncatalyzed reaction, suggesting that all of the ACs exhibited catalytic activity with respect to den-

itrogenation, desulfurization, and deoxygenation of the pretreated bio-oil. Interestingly, all of the ACs exhibited higher activities than Ru/C in desulfurization because the S content in the upgraded biooil produced with AC was always lower than the oil produced with Ru/C (152.0 ppm), and coal AC exhibited the highest activity toward desulfurization. All of the ACs except pine wood and apricot pit facilitated greater N removal than Ru/C. Of the ACs examined, peach pit AC exhibited the highest activities toward denitrogenation, as indicated by the fact that their resulting upgraded bio-oils contained the lowest N content (1.20 wt%). It is suspected that the ACs leach ions into the reaction solution that would catalyze the denitrogenation and desulfurization reactions. In addition, pretreating the bamboo AC under hydrothermal conditions increased the acidic groups on the surface, and this pretreated AC also performed better in denitrogenation and desulfurization than the un-pretreated AC. Therefore, the weak acids leaching from the AC surface would also contribute to the catalysis of denitrogenation and desulfurization. Of course, adsorptive denitrogenation [16] and catalytic activities from the mineral materials cannot be ruled out. Elemental analysis suggested that the solid residues contained 2.60–2.95 wt% N and 0.10–0.29 wt% S (see Table S2). Although the pine wood AC contained the highest mineral metal content, it showed moderate activity toward denitrogenation, implying that the activities of the different mineral metals in the ACs were different from each other. The pine wood AC also showed the lowest activity toward deoxygenation, as indicated by the fact that its resulting material contained the highest O content. The presence of ACs, such as Ru/C, also prompted hydrogenation, as shown by the higher H contents of the upgraded bio-oils compared to the oils produced without catalyst. Thus, ACs can increase the hydrocarbon content in the oil, thus improving its combustion performance. The addition of ACs also improved the densification of the upgraded oils. The HHV (44.1 MJ/kg) of the upgraded bio-oil produced with bamboo stem AC was the highest, even larger than the oil produced with Ru/C and close to that of petroleum diesel (44.8 MJ/kg) [28]. The initial HTL step produced a crude bio-oil that retained 55.80% of the chemical energy in the duckweed biomass. The subsequent pretreatment step produced a pretreated bio-oil that had 82.07% of the chemical energy of the crude bio-oil. Upgrading the pretreated bio-oil with the various ACs led to step-wise energy recoveries between 73.36 and 86.11%. The energy recovery produced with all of the ACs except peach pit AC was higher than the energy recovery produced without AC. The ACs that gave the highest yields and produced upgraded oils with the highest HHVs also gave the highest energy yields. Bamboo stem AC gave the highest energy yield of 86.11%, which was larger than the energy yield (84.05%) produced with Ru/C. The overall energy recoveries for all of the upgraded bio-oils were significantly lower than their corresponding step-wise energy recoveries. All of the upgraded oils produced in the presence of AC had lower TAN values than the upgraded oils produced in the uncatalyzed

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046

G Model

ARTICLE IN PRESS

CATTOD-10018; No. of Pages 9

P. Duan et al. / Catalysis Today xxx (2016) xxx–xxx

7

Fig. 3. Total ion chromatograms for upgraded bio-oils produced with different catalysts, (a) pine-AC; (b). coconut-AC; (c) bamboo-AC; (d) coal AC; e) peach-AC; (f) apricotAC; (g) Ru/C.

Table 6 Composition of upgraded oils (UPO) (% of total peak area by GC–MS) produced with different catalysts. Experimental conditions

Fatty acids

Fatty acid amides

Sat. hydro carbons

Unsat. hydro carbons

Aromatics

N-comp.

O-comp.

UPO-pine-AC UPO-coconut-AC UPO-bamboo-AC UPO-coal AC UPO-peach-AC UPO-apricot-AC UPO-Ru/C

1.57

1.30 1.12 1.44 1.01 0.17 0.64 0.89

41.60 50.46 46.81 48.96 54.00 49.00 46.57

18.75 20.45 19.29 22.15 22.48 21.17 15.48

29.02 24.07 24.48 26.12 24.85 20.95 27.38

8.00 4.65 6.29 2.89 1.94 4.25 5.92

14.53 15.06 18.27 10.84 11.86 16.17 16.26

reaction, indicating that the presence of AC led to the removal of acidic components (e.g., fatty acids) from the pretreated bio-oil. Oils with a low TAN are considered safe for storage and transportation, whereas those with TANs above the ASTM specification (0.8 mg KOH/g) [7] may lead to operational problems and cause corrosion during storage. All of the AC catalysts showed similar activities toward the removal of acidic components in H2 under hydrothermal conditions. Apricot pit AC produced upgraded oil with the lowest TAN value. However, the noble metal was more effective in removing the acidic components than the ACs. 3.4. GC–MS analysis All of the upgraded bio-oils were analyzed by GC–MS to characterize their molecular composition. The inlet temperature of the GC was 300 ◦ C. At this temperature, approximately 70% of the material from all of the upgraded bio-oils was volatilized; the remaining material had a higher boiling point (the detailed boiling point distribution is presented in a subsequent section). Thus, the data provided are only for the volatile fraction, specifically the crude bio-oils, and are not necessarily representative of the total oil. Fig. 3

compares the total ion chromatograms (TICs) of the upgraded biooils dissolved in dichloromethane. The TICs of all of the upgraded oils produced with ACs were similar to each other and contained many regularly spaced peaks that corresponded to a series of nalkanes starting at C9 . The TIC of the upgraded oil produced with Ru/C contained some differences compared with those of the oils produced with ACs, suggesting that the noble metal played a different role during the upgrading compared with that of AC. A mass spectral library and computer matching analysis were used to facilitate compound identification. For comparison, the identified compounds were categorized into the following groups, and their relative amounts are given in Table 6: fatty acids, fatty acid amides, saturated hydrocarbons, unsaturated hydrocarbons, aromatics, N-containing compounds, and O-containing compounds. Their relative contents were catalyst dependent. Table 6 shows that the upgraded oils produced with ACs contained mainly hydrocarbons (saturated and aromatic). The peach pit AC produced an upgraded bio-oil with the highest hydrocarbon content, whereas the pine wood AC had the lowest hydrocarbon content. The saturated hydrocarbon contents of most of the upgraded bio-oils were similar to the upgraded bio-oil produced with Ru/C. In contrast,

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046

G Model

ARTICLE IN PRESS

CATTOD-10018; No. of Pages 9

P. Duan et al. / Catalysis Today xxx (2016) xxx–xxx

8

Table 7 Boiling point distribution of upgraded bio-oils, (a) pine wood-based AC; (b) coconut shell-based AC; (c) bamboo-based AC;(d) coal-based AC; (e) peach shell-based AC; (f) apricot shell-based AC; (g) Ru/C. Distillate range (◦ C)

Duckweed

Crude oil

Pretreated oil

(a)

(b)

(c)

(d)

(e)

(f)

(g)

35–150 150–200 200–250 250–300 300–350 <350 350–400 400–450 450–500 >500

8.37 2.58 5.32 13.78 15.48 45.53 6.47 4.73 3.23 39.04

14.64 14.16 14.31 11.00 7.78 61.89 6.89 5.41 1.91 23.90

19.98 13.23 14.22 11.22 8.02 66.67 5.83 4.78 2.10 20.62

17.64 24.53 21.35 15.17 8.64 87.33 4.58 2.41 0.73 4.95

22.53 26.40 20.62 13.41 6.15 89.11 2.74 1.12 0.47 6.56

24.81 26.18 19.65 12.91 6.58 90.13 3.28 1.54 0.57 4.48

21.23 25.19 20.84 14.30 7.63 89.19 3.61 1.54 0.42 5.24

25.10 26.14 18.47 10.49 5.15 85.35 2.04 1.20 1.05 10.36

19.58 25.11 20.63 15.07 8.26 88.65 4.29 1.91 0.53 4.62

19.79 21.38 17.24 12.91 8.77 80.09 5.52 4.47 1.23 8.69

3.6. FT-IR analysis

Transimittance%

0.95

0.75 Duckweed Crude oil Pretreated oil Upgraded oil-No cat. Upgraded oil-AC Upgraded oil-Ru/C

0.55

0.35 3600

2800

2000

1200

400

Wave number/cm-1 Fig. 4. FT IR spectra for duckweed, crude, pretreated and upgraded oils produced without and with coconut AC and Ru/C.

Ru/C produced an upgraded bio-oil with the lowest unsaturated hydrocarbon content and highest aromatic content due to its high performance in the hydrogenation reactions of bio-oil. Nevertheless, the ACs exhibited similar activities to that of Ru/C in affecting the molecular composition of the upgraded bio-oils.

3.5. Boiling point distribution The boiling point distributions of the duckweed, crude bio-oils, pretreated bio-oils, and upgraded bio-oils were estimated from the TGA of nitrogen. The application of TGA in simulated distillation is regarded as a “miniature distillation”. The samples were heated from room temperature to 800 ◦ C, and some thermal degradation was likely at higher temperatures [23]. Table 7 lists the boiling point distributions of the bio-oils determined by TGA. HTL increased the proportion of the fraction below 350 ◦ C from 45.53 to 61.89%. The pretreated bio-oil contained slightly more of this low-boiling-point fraction. In contrast, the upgraded oils from the AC-catalyzed reactions had a significantly larger fraction that boiled below 350 ◦ C compared to the pretreated bio-oil. The AC catalysts appeared to have promoted cracking reactions during the upgrading process, which shifted the molecular distribution to more volatile compounds. Treatment with Ru/C produced an upgraded oil with 80.09% of the material boiling below 350 ◦ C. These results agree with the GC–MS chromatograms in Fig. 3, which showed a greater abundance of lighter compounds in the upgraded bio-oils.

FT-IR analysis was performed to determine the functional groups in the duckweed biomass, crude bio-oil, pretreated bio-oil, and upgraded bio-oils. The results are listed in Fig. 4. The relative differences between the band heights are an indication of the relative differences in the concentrations of the corresponding functional groups between the materials. The FT-IR spectrum of duckweed biomass contained strong, broad H-O absorptions at 3400–3650 cm−1 , indicating a rich presence of carbohydrates and proteins in the raw material, which is consistent with the oxygen content in the duckweed. The strong, broad band from 3400 to 3650 cm−1 might also be due to the presence of water in the sample. In contrast, a weak absorbance was observed for the crude bio-oil, suggesting that the carbohydrates and proteins were converted during the HTL process. Pretreating the crude bio-oil further reduced the absorbance at 3400–3650 cm−1 . Small peaks were observed for all of the upgraded bio-oils. All of the bio-oils displayed a strong absorbance from 2850 to 3000 cm−1 , indicating a high content of methylene groups. This result is consistent with the presence of hydrocarbons in the TIC. The absorbance peaks between 1650 and 1760 cm−1 represent the C O stretching vibrations in carboxylic acids, amides and ketones. A strong absorbance was observed for the duckweed and crude bio-oil. All of the upgraded bio-oils produced with AC showed a proportionately lower absorbance over this wavenumber range, which is consistent with the reduced contents of carboxylic acids, amides and ketones. The appearance of distinct bands between 710 and 860 cm−1 in the upgraded bio-oils was most likely due to the presence of substituted benzenes. The spectra of the upgraded bio-oils produced with ACs were similar, suggesting that they contained the same functional groups.

4. Conclusions This study demonstrated that ACs can convert pretreated bio-oil into a liquid fuel that has similar properties to those of hydrocarbon fuels derived from fossil fuel resources. The ACs showed activities similar to that of Ru/C and thus are promising inexpensive catalytic materials for upgrading bio-oils. The net effects of the catalytic supercritical water upgrading process on the molecular composition of the bio-oils included increased amounts of alkanes and aromatics and decreased amounts oxygenated- and nitrogen-containing compounds. The mesopores in the ACs favored the conversion of heavy compounds, restricting coke formation. The upgraded oil produced without catalyst had a higher visual viscosity than the oils produced with AC catalysts. Given these encouraging results, ACs can be considered effective catalysts to improve the properties of crude bio-oil from the HTL of duckweed.

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046

G Model CATTOD-10018; No. of Pages 9

ARTICLE IN PRESS P. Duan et al. / Catalysis Today xxx (2016) xxx–xxx

Acknowledgments The authors gratefully acknowledge the financial support from the Program for New Century Excellent Talents in University (NCET13-775), the National Science Foundation of China (21106034) and Key Laboratory of Biomass Chemical Engineering of Ministry of Education (Zhejiang University) (2015BCE001). 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.2016.01. 046. References [1] W. Cui, J. Cheng, Plant Biology 17 (Suppl. 1) (2015) 16–23. [2] W. Catallo, T. Shup, T. Eberhardt, Biomass Bioenergy 32 (2008) 140–145. [3] S. Xiu, A. Shahbazi, J. Croonenberghs, L. Wang, Energy Sources A 32 (2010) 1293–1300. [4] P. Duan, Z. Chang, Y. Xu, X. Bai, F. Wang, L. Zhang, Bioresour. Technol. 135 (2013) 710–719. [5] R. Baliban, J. Elia, C. Floudas, X. Xiao, Z. Zhang, J. Li, H. Cao, J. Ma, Y. Qiao, X. Hu, Ind. Eng. Chem. Res. 52 (2013) 11436–11450. [6] N. Muradov, M. Taha, A. Miranda, K. Kadali, A. Gujar, S. Rochfort, T. Stevenson, A. Ball, A. Mouradov, Biotechnol. Biofuels 7 (2014) 30. [7] American Society for Testing and Materials, ASTM Int. D6751-07a, 2007, 1–8.

9

[8] P. Duan, Y. Xu, X. Bai, Energy Fuels 27 (2013) 4729–4738. [9] C. Zhang, P. Duan, Y. Xu, B. Wang, F. Wang, L. Zhang, Bioresour. Technol. 166 (2014) 37–44. [10] Z. Fang, C. Xu, Near-Critical and Supercritical Water and their Applications for Biorefineries, Biofuels and Biorefineries. Springer Book Series–Biofuels and Biorefineries, Springer-Verlag, Heidelberg Berlin, 2014. [11] J. Fu, F. Shi, L. Thompson, X. Lu, P. Savage, ACS Catal. 1 (2011) 227–231. [12] H. Fukuyama, S. Terai, M. Uchida, L. Cano, J. Ancheyta, Catal. Today 98 (2004) 207–215. [13] T. Viet, J. Lee, F. Ma, G. Kim, I. Ahn, C. Lee, Fuel 103 (2013) 553–561. [14] C. Xu, S. Hamilton, M. Ghosh, Fuel 88 (2009) 2097–2105. [15] Y. Sano, K. Choi, Y. Korai, I. Mochida, Energy Fuels 18 (2004) 644–651. [16] J. Bu, G. Loh, C. Gwie, S. Dewiyanti, M. Tasrif, A. Borgna, Chem. Eng. J. 166 (2011) 207–217. [17] C. Yu, J. Qiu, Y. Sun, X. Li, G. Chen, Z. Zhao, J. Porous Mater. 15 (2008) 151–157. [18] American Society for Testing and Materials, ASTM Int. D6851, 2011, 1–2. [19] D. Lennon, D.T. Lundie, S.D. Jackson, G.J. Kelly, S.F. Parker, Langmuir 18 (2002) 4667–4673. [20] R. Patterson, Anal. Chem. 45 (1973) 605–609. [21] American Society for Testing and Materials, ASTM Int. D4629, 2012, 1–11. [22] American Society for Testing and Materials, ASTM Int. D4629, 2012, 1–5. [23] P. Duan, B.Wang, Y. Xu, Bioresour Technol. 186 (2015) 58–66. [24] X. Bai, P. Duan, Y. Xu, A. Zhang, P. Savage, Fuel 120 (2014) 141–149. [25] O. Ioannidou, A. Zabaniotou, Renewable Sustainable Energy Rev. 11 (2007) 1966–2005. [26] M. Sert, L. Ballice, M. Yuksel, M. Saglam, Oil Shale 26 (2009) 463–653. [27] T. Sato, T. Tomita, P. Trung, N. Itoh, S. Sato, T. Takanohashi, Energy Fuels 27 (2013) 646–653. [28] A. Demirbas, Energy Policy 35 (2007) 4661–4670.

Please cite this article in press as: P. Duan, et al., Activated carbons for the hydrothermal upgrading of crude duckweed bio-oil, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.01.046