Bioresource Technology 174 (2014) 95–102

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Pressurized pyrolysis of rice husk in an inert gas sweeping fixed-bed reactor with a focus on bio-oil deoxygenation Yangyang Qian, Jie Zhang, Jie Wang ⇑ Department of Chemical Engineering for Energy, Key Laboratory of Coal Gasification and Energy Chemical Engineering of Ministry of Education, East China University of Science and Technology, 130#, Meilong Road, Shanghai 200237, PR China

h i g h l i g h t s  Pressurized pyrolysis of rice husk was conducted in a fixed-bed reactor.  Pressure accelerated the dehydration and decarboxylation of bio-oil.  The bio-oil produced under pressure had less oxygen and higher calorific value.  Pressure increased the yields of acetic acid, PCX and guaiacols.

a r t i c l e

i n f o

Article history: Received 26 August 2014 Received in revised form 1 October 2014 Accepted 4 October 2014

Keywords: Rice husk Pyrolysis Pressure Bio-oil Deoxygenation

a b s t r a c t The pyrolysis of rice husk was conducted in a fixed-bed reactor with a sweeping nitrogen gas to investigate the effects of pressure on the pyrolytic behaviors. The release rates of main gases during the pyrolysis, the distributions of four products (char, bio-oil, water and gas), the elemental compositions of char, bio-oil and gas, and the typical compounds in bio-oil were determined. It was found that the elevation of pressure from 0.1 MPa to 5.0 MPa facilitated the dehydration and decarboxylation of bio-oil, and the biooils obtained under the elevated pressures had significantly less oxygen and higher calorific value than those obtained under atmospheric pressure. The former bio-oils embraced more acetic acid, phenols and guaiacols. The elevation of pressure increased the formation of CH4 partially via the gas-phase reactions. An attempt is made in this study to clarify ‘‘the pure pressure effect’’ and ‘‘the combined effect with residence time’’. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Biomass has been receiving considerable interest as a renewable and carbon-neutral energy resource in recent years due to the worldwide increased concerns on the shortage of fossil fuels, climate change and environmental pollution. Rice husk is a byproduct in mill factories with the productivity of about twenty percent from rice grain. It is available as a biomass feedstock in many agrarian countries primarily in Asia. The annual output of rice husk in China is estimated to be about 70 million metric tons and accounts for about half of the world output (Yoon et al., 2012). Rice husk is concentrated with lignocellulosic components (i.e., cellulose, hemicellulose and lignin) with a half as much calorific value as that of coal, but it has a very low bulk density so that its distant transport is economically limited. In China, rice husk is

⇑ Corresponding author. Tel./fax: +86 21 64252853. E-mail address: [email protected] (J. Wang). http://dx.doi.org/10.1016/j.biortech.2014.10.012 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.

mainly used in burning for rural households and small boilers in local areas. Such energetic uses of rice husk has the demerit of low efficiency and high smoke emission. Pyrolysis offers a promising way to convert rice husk to more energy-intensive and valueadded products such as bio-oil and bio-char. Biomass pyrolysis has been investigated extensively from varying viewpoints. Numerous recent studies have focused on the biooil production by fast pyrolysis of various biomass species, such as cassava stalk and rhizome (Pattiya, 2011), forestry residues (Amutio et al., 2013), jute stick (Asadullah et al., 2008), maize stalk (Zheng, 2008), mallee leaves (He et al., 2012), microalgae remnants (Wang et al., 2013), Miscanthus sinensis (Heo et al., 2010), red oak (Ellens and Brown, 2012) and waste plywood (Jung et al., 2012). The fluidized-bed, conical spouted bed and fall free reactors are employed for fast pyrolysis. Rapid heating rate with short vapor residence time is favorable to maximize the yield of bio-oil as a consequence of the suppressed secondary cracking reactions (Bridgwater, 1999). Traditionally slow pyrolysis is applied for the production of charcoal and gas. However, the slow pyrolysis with

96

Y. Qian et al. / Bioresource Technology 174 (2014) 95–102

gas sweeping over the heated biomass sample permits the volatile matter to leave promptly from the hot zone and can consequently avoid its violent cracking to gases. Gercel (2002) used a fixed-bed reactor to pyrolyze a sunflower-pressed bagasse with the heating rate of 5 °C/min in a flow of nitrogen. A bio-oil yield of 52% was obtained at 550 °C. The average compositional formula of bio-oil was CH1.68O0.165N0.059. Wang et al. (2009) carried out a similar slow pyrolysis of pine wood in a stream of argon, and observed that the compounds in the bio-oil reflected the initial or mild degradation from the lignocellulosic components. The slow pyrolysis of a microalgae (Tetraselmis chui) by passing helium through the packed biomass sample (Grierson et al., 2011) demonstrated that the bio-oil consisted of a wide variety of compounds including fatty acids, alkanes, alkenes, amides, aldehydes, terpenes, phenols, etc. The biomass pyrolysis by any measures to mitigate the volatile matter cracking is inclined to achieve a good harvest of bio-oil. In general, however, the resultant bio-oil contains abundant oxygen. This is deemed to be responsible for its drawback of low heating value, lipophilicity and instability, restricting its application. Several approaches have hence emerged to upgrade the bio-oil before, during and after pyrolysis (Stefanidis et al., 2011; Veses et al., 2014). A common tactics adopted in most approaches is to eliminate the oxygen from bio-oil. The adjustment of operating parameters such as temperature, heating rate and atmosphere during biomass pyrolysis is known as a straightforward way to change the pyrolysis characteristics and the formation of bio-oil (Ben and Ragauskas, 2013; Chhiti et al., 2012; Zhang et al., 2011). However, it can be often difficult to determine an optimal balance of the yield of bio-oil against the quality. There are a number of publications pertaining to the pyrolysis of rice husk. Mansaray and Ghaly (1999) studied the decomposition of four rice husks using thermogravimetric analysis (TGA). Worasuwannarak et al. (2007) investigated the pyrolysis behaviors of rice husk together with rice straw and corncob by means of TG–MS. Zheng (2007) reported the fast pyrolysis of rice husk on a fluidized-bed reactor in which the highest bio-oil yield of 56% (including water) was obtained at 465 °C. Tsaia et al. (2007) used a fixed-bed reactor to pyrolyze rice husk with a rapid heating rate (500–800 °C/min) and obtained a maximal bio-oil yield of 40% (including water). Their GC–MS analysis showed that the bio-oil consisted of copious oxygen-containing compounds. Chen et al. (2011) characterized the bio-oil produced from the pyrolysis of rice husk in a fluidized-bed reactor. Nevertheless, the average bio-oil yield of 41.7% obtained in their experiment at 500 °C appeared to be lower than normal, whereas the average water content of 41.5% in the bio-oil appeared to be higher. Zhou et al. (2013) investigated the catalytic pyrolysis of rice husk mixed with zinc oxide in a fixed-bed reactor towards the bio-oil production, and found that addition of the catalyst reduced the yield of bio-oil and changed its rheological properties. Alvarez et al. (2014) performed the pyrolysis of rice husk using a conical spouted bed reactor. The bio-oil yields (excluding water) of 44–47% and the water yields of around 23% were obtained at 450–600 °C. The yields of acetic acid ranged in 2.35–3.16%. In this work, we examine the pyrolysis characteristics of rice husk in a fixed-bed reactor purged with a pressurized nitrogen gas. To the best of our knowledge, there are few studies in the literature concerning the influence of pressure on biomass pyrolysis. Whitty et al. (2008) investigated the influence of pressure on black liquor using a single-particle reactor and a grid heater. Their interest was centered on the char and gas production. Mercader et al. (2010) proposed the high-pressure thermal treatment for upgrading pyrolytic oil. They found that the treatment under a condition of 300–340 °C and 20 MPa was effective to liberate gas (mainly CO2) and water from the pyrolytic oil. In the present study, we have found that the pressurized pyrolysis has a similar effect on

preferential removal of oxygen from bio-oil by the facilitated dehydration and decarboxylation. 2. Experimental 2.1. Rice husk sample The rice husk used in this work was kindly supplied by a mill factory in Zhejiang province, China. The sample was ground and sieved to the particle size of about 0.15–0.45 mm. The proximate analysis showed that the dried rice husk had 1.55% moisture, 72.58% volatile matter, 16.01% fixed carbon and 9.86% ash. The ultimate analysis showed that the rice husk (daf. basis) had 51.20% carbon, 6.17% hydrogen, 42.16% oxygen (by difference), 0.36% nitrogen and 0.10% sulfur. The componential analysis showed that the rice husk (dry basis) contained 42.94% cellulose, 14.85% hemicellulose, 25.79% lignin and 7.60% neutral detergent solute. Cellulose, lignin and hemicellulose comprised more than 92% of the organic entity. The content of either cellulose or lignin in the rice husk was higher than that reported by Li et al. (2012) and Chen et al. (2011) but close to those reported by Bakar and Titiloye (2013). 2.2. Pyrolysis apparatus and procedures Pyrolysis was carried out in a vertical tubular fixed-bed reactor made of inconel steel. The schematic diagram of the apparatus was illustrated elsewhere (Zhang et al.,2014). In each run, a 4 g sample of rice husk was wrapped by a stainless steel wire mesh and then placed on the ceramic filler in the flat-temperature zone of the reactor. The dehumidified nitrogen gas was flowed from the inlet at the bottom of reactor to the outlet at the top, sweeping through the biomass sample. The gas flow rates were controlled by a mass flow meter from 10 mL/min to 500 mL/min. The gas pressures were controlled by a counterbalance valve from 0.1 MPa to 5.0 MPa. After no gas leakage was guaranteed and the air enclosed inside the reactor was driven out, the reactor was electrically heated at a predetermined heating rate of 15 °C/min from room temperature to 700 °C for a holding time of 60 min. In order to detect the pyrolysis temperature, a thermocouple was inserted into the reactor and its tip was forced to tightly contact the sample mesh. The volatile matter was purged out and condensed in a stainless steel trap immersed in the cooled salt water which maintained the temperatures from 12 °C to 6 °C during the experiment. The tube lines connecting the reactor to the trap was heated at 280 °C with a heating tape to prevent the condensation of the volatile matter. The incondensable gases were collected in the gas bags at a temperature interval of 50 °C in the heat-up stage and at a time interval of 10 min in the temperature-holding stage. The flow rate of total gas including the swept gas and gaseous product was measured by another mass flow meter after depressurization. Char product was recovered after the reactor was dismounted. 2.3. Analytical methods The liquid yield was determined by weighing the trap before and after experiment. The liquid product was then carefully washed out with acetone (AR grade) into a glass bottle, which was then sealed up for later analyses of water and organic compounds in it. The water content in the liquid product was determined using a Coulometry trace moisture analyzer (Mettler Toledo, C20). The compounds in the liquid product were characterized by gas chromatography mass spectrometry (GC–MS, PerkinElmer clams 500). The quantitative determination of typical compounds was carried out on a GC-FID analyzer (Haixin

Y. Qian et al. / Bioresource Technology 174 (2014) 95–102

GC-950) equipped with a HP-5 capillary column. In this analysis, 15 reagents including benzene, toluene, m-xylene, acetic acid, 2-furfural, 2-furanmethanol, phenol, p-cresol, 2,4-xylenol, guaiacol, naphthalene, anthracene, phenanthrene, fluorene and pyrrole were used to prepare their standard solutions by adding to the acetone reagent. All these reagents were of GC grade, which were bought from Aladdin Chemical Co., except that 2,4-xylenol was of AR grade from Shanghai Linfeng Chemical Reagent Co. Ltd. Pyrrole was used as an internal calibration compound. It was confirmed that no peaks of compounds in the liquid product superimposed on that of pyrrole. The four major gases (H2, CO, CO2 and CH4) were quantitatively measured by a gas chromatograph (Agilent 6820) equipped with a thermal conduct detector (TCD). The total gas yield refers to the cumulative yields of these four gases. The char yield was determined by weighing the solid sample before and after experiment. The char sample was stored in a desiccator for elemental analysis. Element analysis (C, H, N and S) was implemented on an elemental analyzer (Vario Macro Cube). The contents of lignocellulosic components in the rice husk sample were determined by the national standard method of China (GBT-20805-2006). The detailed procedures of this analysis were described elsewhere (Shi et al., 2012).

3. Results and discussion 3.1. Gas release

0.35 700 0.30 600

0.25 0.20

500

(

0.15

Temperature

400 )

-1 -1 Release rate ( mmol.min .g -dry biomass)

To help understand the pressure influences on the release rates of gases presented soon after, the release profiles of four main gases during the pyrolysis of rice husk under atmospheric pressure are shown in Fig. 1, where the temperature–time history curve points to the right longitudinal axis. It could be seen that CO2 and CO were the first two gases to evolve at a temperature of 275 °C, followed by CH4 and H2 at 325 °C. This result was consistent with that obtained in our previous work by the pyrolysis of pine wood in a similar pyrolysis experiment carried out under atmospheric pressure (Shi and Wang, 2014). It was evident there that the initial formation of CO2 and CO at low temperature was

0.10 300

0.05 0.00 20

30

40

50

60 70 80 Time (min)

90

200 100 110

Fig. 1. The release rates of main gases during the pyrolysis of rice husk under atmospheric pressure with the flow rate of 500 mL/min. s, CO2; 4, CO; h, CH4; 5H2.

97

mainly due to the decarboxylation and decarbonylation of hemicellulose and extractives, whereas the formation of these gases from cellulose required a temperature higher than 375 °C. Compared to the temperature of 425 °C where the release rates of both CO2 and CO from pine wood reached a maximum (Shi and Wang, 2014), however, the maximal release rates of CO2 and CO from rice husk shifted to a relatively low temperature of 375 °C. It was indicated that the decarboxylation and decarbonylation occurred more rapidly for rice husk than did for pine wood, probably because the mineral matter enriched in rice husk had a catalytic effect, like that in rice straw (Shi et al., 2012). In that work (Shi and Wang, 2014), the evolution of CH4 during the pyrolysis of pine wood exhibited a sharp peak at 450 °C with a broad shoulder tilted from 650 °C. The first peak was primarily attributed to the decomposition of lignin while the later shoulder to the subsequent cracking of the intermediates derived from cellulose and hemicellulose. In contrast, the pyrolysis of rice husk in the present work showed only a broad release rate peak attenuated from 575 °C. It appeared that the dementhanation from cellulose and hemicellulose also occurred more rapidly for rice husk than did for pine wood. Fig. 2 shows the release profiles of four main gases during the pyrolysis of rice husk under varying pressures from 0.5 MPa to 5.0 MPa at the constant linear velocity of gas flow. It should be remarked that the profile of gas release rate could be substantially changed with the linear velocity of gas flow. In the case of pressurized pyrolysis, the linear velocity of gas flow could become very slow even with the maximal volume flow rate of 500 mL/min used in this study. It was estimated that under the condition used in Fig. 2, the linear velocity of gas flow was about 4.0 cm/min, and it took as long as 5.6 min to sweep the gas product out from the reactor system in the temperature range of 200–400 °C, whereas it took only 0.2 min under the condition used in Fig. 1. Consequently, all profiles of gas release rate with time were subjected to a significant temperature delay of about 85 °C in the temperature range of 200–400 °C in Fig. 2. Nevertheless, such a delay could be overlooked when the profiles obtained under different pressures were compared with each other, because of the constant linear velocity of gas flow. An important observation revealed from Fig. 2 was that despite the significant gas release delay, the release rates of CO2 showed a protruding peak with the peak top at the delayed temperature of 425 °C under 3.0 and 5.0 MPa. This indicated that the high pressure strongly accelerated the release of CO2, primarily due to the promoted decarboxylation of hemicellulose and cellulose. In contrast to those obtained under atmospheric pressure (Fig. 1), the release profiles of CO in Fig. 2 exhibited two distinct peaks. It was plausible that the facilitated decarbonylation of hemicellulose and cellulose under pressure enabled the deconvolution of the broad peak of CO release rate. However, both the first and the second peaks of CO declined with elevating pressure. Along with the considerable formation of CO2 and CO during the period of 32–42 min, the release rate of CH4 showed a noticeable peak under 3.0 and 5.0 MPa in this period, ahead of a broad peak occurring. Similarly, a small but distinct peak of H2 appeared during the period of 32–42 min under pressure, and this peak increased with elevating pressure. Table 1 shows the cumulative yields of four main gases generated over the different pyrolysis experiments with varying pressures at the constant linear velocity of gas flow or the constant volume flow rate. The data obtained with the constant linear velocity correspond to those shown in Fig. 2, but in the case of 0.1 MPa, no data of the gas release rate were obtained owing to the extremely low gas flow rate. The errors of the gas yields determined by some repeated pyrolysis experiments are shown in Table 1. The reproducibility in determining the gas yields was quite good, regardless of the pressures boosted.

98

Y. Qian et al. / Bioresource Technology 174 (2014) 95–102

0.4

0.25 700

CO2

700

CO 0.20

600 0.3

500

0.2

400

0.1 0.0 20

0.15

500

0.10

400

300 0.05

30

40

50

60

70

80

90

200 0.00 100 110 20

0.14 0.12

600

300

30

40

50

60

70

80

90

200 100 110

0.08 700

700

H2

CH4

Temperature

Release rate ( mmol.min-1.g-1-dry biomass)

0.5

0.10

600 0.06

600

0.08

500

500

0.04

0.06 400 0.04 300

0.02 0.00 20

400 0.02

30

40

50

60

70

80

90

300

200 0.00 100 110 20

30

40

50

60

70

80

90

200 100 110

Time (min) Fig. 2. The release rates of main gases during the pyrolysis of rice husk under varying pressures at the constant linear velocity of gas flow. 4, 0.5 MPa with 50 mL/min; h, 1.0 MPa with 100 mL/min; 5, 3.0 MPa with 300 mL/min; e, 5.0 MPa with 500 mL/min.

Table 1 The cumulative yields of four main gases under different pressures (mmol/g-dry biomass). Pressure (MPa)

Constant linear velocity

Constant volume flow rate

CO2

CO

CH4

H2

CO2

CO

CH4

H2

0.1

2.02

1.84

0.98

1.35

0.5

2.31 (±0.05) 2.47 2.93

1.80 (±0.03) 1.67 1.31

1.32 (±0.02) 1.36 1.44

1.71 (±0.05) 1.64 1.51

2.15 (±0.03) 2.57

1.82 (±0.02) 1.84

0.76 (±0.01) 1.06

1.79 (±0.02) 1.92

3.31 (±0.04)

1.27 (±0.02)

1.58 (±0.02)

1.87 (±0.02)

2.86 3.09 (±0.04) 3.31 (±0.04)

1.69 1.37 (±0.03) 1.27 (±0.02)

1.13 1.36 (±0.02) 1.58 (±0.02)

1.94 1.84 (±0.02) 1.87 (±0.02)

1.0 3.0 5.0

In the case of the constant linear velocity of gas flow, the yield of CO2 was gradually enhanced with increasing pressure. The yield of CO2 obtained under 5.0 MPa was 1.6 times as much as that obtained under 0.1 MPa. The pressure elevation also formed more CH4. The yield of CO decreased with increasing pressure. In total, however, more oxygen was released as gaseous product under the elevated pressures. The change in the yield of H2 with pressure was not monotonic. The yield of H2 increased significantly from 0.1 MPa to 0.5 MPa, but it was changing somewhat up and down from 0.5 MPa to 5.0 MP. The pressure elevation with the constant volume flow rate had similar effects on the yields of four gases. It should be noted that in two series of experiments with the constant volume flow rate or with the constant linear velocity of flow rate, the pyrolysis experiment carried out at 5.0 MPa employed the same volume flow rate of 500 mL/min. Therefore, to set the constant volume flow rate meant a shorter residence time of vapor in the reactor as the pressures lowered from 5.0 MPa to 0.1 MPa. In this instance, the lower the pressure, the shorter the residence time. It could be accordingly observed from comparing two series of experiments conducted under the same pressures of 0.1, 0.5, 1.0 or 3.0 MPa that all the yields of CO2 and H2 decreased with prolonged residence times,

whereas the yields of CH4 increased. The changes in the yields of CO were not so significant. This result seemed to be strange because the prolonged residence time would generally intensify the decomposition of the volatile matter, resulting in an increase in the yield of each gas. In elucidation of the above contradiction, the gas-phase reaction of CO2 with H2 to form CH4 and H2 is considered to have an influence on the yields of gases. The forward direction movement of this reaction is thermodynamically favorable under high pressure. It was likely that the increased yield of CH4 with increasing pressure arose partially from this reaction. On the other hand, the reaction might not arrive at a thermodynamic equilibrium so that the prolonged residence time increased the conversion of CO2 and H2 to CH4. Likewise, one may have expected the formation of CH4 and H2O via the gasphase reaction of CO with H2. Indeed, this reaction could account for why CO decreased slightly with increasing pressure (Table 1), despite the facilitated decarbonylation of rice husk under the high pressure (Fig. 2). However, it appeared that the vapor residence time had little impact on reaction (2), dissimilar to reaction (1). The concomitant evolution of four gases in the period of 32–42 min (Fig. 2) also underpinned the above postulation on the gas-phase reactions.

99

Y. Qian et al. / Bioresource Technology 174 (2014) 95–102

3.2. Overall distributions of products Fig. 3 shows the overall distributions of char, bio-oil, water and gas products from pyrolysis of rise husk obtained under different pressures, where the total yields of these products are presented pointing to the right longitudinal axis. All yields are represented in mass percent on a dry biomass basis. The yield of bio-oil is obtained in terms of a difference between the total liquid yield and the water yield. The statistical analysis of the data obtained by some repeated experiments showed that the absolute errors were ±0.36%, ±0.46%, ±0.40% and ±0.27%, respectively, for the yields of char, bio-oil, water and gas. The mass closures of the char, biooil, and liquid products were satisfactorily better than 98% in all experiments. Fig. 3a shows the yields of four products obtained under different pressures with the constant linear velocity of gas flow. The yields of char, bio-oil, water and gas products obtained under atmospheric pressure were 33.92%, 27.50%, 21.60% and 16.10%, respectively. This bio-oil yield was smaller than that of about 44% obtained by the pyrolysis of rice husk in a conical spouted bed reactor at 600 °C under atmospheric pressure (Alvarez et al., 2014) but larger than that obtained by the pyrolysis of rice husk on a TG–MS apparatus at 600 °C (Worasuwannarak et al., 2007). The yields of char, water and gas increased with elevating pressure especially from 0.1 MPa to 1.0 MPa at the expense of a reduction in the bio-oil yield. It was implied that the pressure elevation facilitated the polycondensation, dehydration and cracking of the volatile matter to form, respectively, more char deposit, water and gas. However, it was interesting that these effects became insignificant when the pressure was elevated from 1.0 MPa to 5.0 MPa, while an increase in the yield of CO2 remained remarkable (Table 2). Fig. 3b shows the yields of four products obtained under different pressures with the constant volume flow rate. Compared to the results in Fig. 3a, the yields of char and gas obtained under atmospheric pressure were not so much different, whereas the yield of

a

100

40

char

80

30

water

60

gas

20

bio-oil

10 0

40

0

1

2

3

4

20

50 40

char

30

water

0

1

2

3

4

100 80

Peak Retention Compounds no. time (min)

a

1 2 3 4 5 6 7 8 9 10 11

3.01 3.37 4.23 4.83 5.23 5.30 5.94 6.20 6.39 6.47 6.85

12 13 14 15 16 17

7.04 7.23 7.42 7.67 7.98 8.06

18 19 20 21 22 23 24 25

8.25 8.51 8.77 8.85 9.13 9.57 9.62 10.36

26

10.50

Relative area (%) 0.1 MPa 3.0 MPa

Acetic acid Methyl acetate Pyrrolea Furfural 2-Furanmethanol 2-Propanone,1-(acetyloxy)2-Cyclopenten-1-one,2-hydroxy2-Furancarboxaldehyde,5-methylFuran,tetrahydro-2-methyl2H-pyran,3,4-dihydro2-Cyclopenten-1-one,2-hydroxy3-methylPhenol Phenol,2-methoxyPhenol-2-methylPhenol,4-methylPhenol,2-methoxy-4-methylAcetic acid,2-(N-methyl-Nphosphomethyl)aminoPhenol-4-ethyl Phenol,4-ethyl-2-methoxyBebzofuran,2,3-dihydroPhenol,2-Methoxy-4-vinylPhenol,2,6-dimethoxyPhenol,2-methoxy-4-(1-propenyl)Hydroquinone 2-Propanone,1-(4-hydroxy-3methoxyphenyl)2-H-bezopyran-2-one,3,4-dihydro6-hydroxy-

17.39 9.73 / 1.28 4.69 1.39 2.41 – 1.55 1.36 3.49

28.07 – / 4.78 – 1.67 – 2.31 – 2.09 2.34

3.23 4.62 0.74 3.22 3.85 14.24

7.04 10.10 3.00 6.95 9.89 –

2.13 2.17 4.72 3.53 2.23 2.03 1.62 1.63

8.33 7.79 1.93 – 1.67 1.89 – 0.16

1.11

–

27

10.86

D-gluco-heptulosan

4.67

–

28

11.53

2-Propenal,3-(4-hydroxy-3methoxyphenyl)-

0.98

–

Internal standard compound ;–, not detected.

bio-oil was significantly high, up to 36.2%, with a low water yield of only 14.0%. This indicated that the vapor residence time played a more important role under atmospheric pressure in promoting the dehydration of bio-oil. The changes in the yields of char, biooil, water and gas with increasing pressure exhibited trends similar to the results in Fig. 3a. However, in the case of the constant volume flow rate, the changes in the yields of bio-oil and water were more significant because the pressure elevation also prolonged the vapor residence time. In other words, the pressure effect was incorporated with the effect of the vapor residence time in this instance. The increased yield of water with increasing pressure might also be partially due to the prolonged gas-phase reaction between CO2 and H2. The yields of bio-oil and water obtained under 5 MP were 14.1% and 27.0%, respectively. It was apparent from Fig. 2 and Table 1 that the pressurized pyrolysis significantly promoted the dehydration and decarboxylation of bio-oil.

60 3.3. Elemental compositions of char and bio-oil

40

bio-oil

10 0

b

gas

20

0

5

Total yield (wt%-dry biomass)

Product yield (wt%-dry biomass)

50

Table 2 The GC–MS detected compounds in two bio-oils obtained by rice husk pyrolysis under different pressures.

20

5

0

Pressure (MPa) Fig. 3. The overall distributions of char, bio-oil, water and gas products from the pyrolysis of rise husk under different pressures with the constant linear velocity of gas flow (a) or with the constant volume flow rate (b). ., I, total yields; 4, N, char; h, j, bio-oil; 5, ., water; e, r, gas.

Fig. 4 shows the compositions of major elements (C, H and O) in the char products and the bio-oil products obtained under different pressures with the constant linear velocity of gas flow or the constant volume flow rate, where the elemental compositions of the rice husk are depicted for contrast. The fraction of each element is defined as its mole number divided by the total mole number of C, H and O in the raw sample or products. The elemental composition of char was directly determined using an elemental analyzer, while that of bio-oil was estimated in terms of the mass balances of each element in the char, bio-oil, water and gas products. It could

100

Y. Qian et al. / Bioresource Technology 174 (2014) 95–102

0 100

rac tio n

40

60

bio-oil rice husk

60

) (%

Of

80 n tio rac Hf

(% )

20

40

80 100 0

0 20

40

) (% tio n rac Of

(% ) Of rac tio n

3.0 15.0 82.0 82.5 83.0 83.5 84.0 84.5 85.0

C fraction (%)

65

15

60

bio-oil

20

55

25

) (%

15.5

70

10

on

16.0

2.5

0 75 ti rac Hf

16.5

) (%

char

100

5

17.0

1.5 2.0

80

n tio rac Hf

17.5

1.0

60

C fraction (%)

0.0 18.0 0.5

20

char

50

30

45

35 25

40 30

35

40

45

50

55

60

C fraction (%)

Fig. 4. The compositions of major elements (C, H and O) in char and bio-oil products obtained under different pressures with the constant volume flow rate (white marks) or with the constant linear velocity of flow rate (black marks and e).s,d, 0.1 MPa; 4, N, 0.5 MPa; h, j, 1.0 MPa; 5, ., 3.0 MPa; e, 5.0 MPa.

be clearly seen from the top triangle coordinate sub-figure that all char products were much carbon-condensed with very little oxygen remaining as compared to rice husk; the compositions of bio-oil were more disseminated than those of char but all the bio-oil products contained less oxygen than rice husk. The influences of pressure on the compositions of bio-oil and char are illustrated in the scale-magnified sub-figures. It was interesting to observe that the fraction of oxygen in bio-oil decreased appreciably with increasing pressure whether the constant linear velocity of gas flow or the constant volume flow rate was fixed, whereas the fraction of carbon in bio-oil increased appreciably with increasing pressure. In the case of the constant volume flow rate, the fraction of oxygen in bio-oil decreased from 17.6% (mol.) at 0.1 MPa to 6.4% (mol.) at 5.0 MPa, corresponding to the increased in the fraction of carbon in bio-oil from 27.9% (mol.) to 45.3% (mol.). This strongly indicated that the pressurized pyrolysis was beneficial to the bio-oil deoxygenation and could thus improve the quality of bio-oil. More importantly, the pyrolysis under the high pressures of 3.0 MPa and 5.0 MPa behaved in the preferential release of CO2 from bio-oil without causing a more rigorous polycondensation of bio-oil to form char and a more violent decomposition of bio-oil to gases (3). The reason for this result might be the strengthened heat transfer by the high-pressure gas, which intensified the preferential decarboxylation of bio-oil at a low temperature. As for the influence of pressure on the composition of char, the fraction of oxygen in char was slightly reduced with increasing pressure. In a general form, the compositional formula the bio-oil obtained under atmospheric pressure with the volume flow rate of 500 mL/min was CH1.95O0.63, while the compositional formula of the bio-oils obtained under 0.5 MPa and 5.0 MPa with the same flow rate were CH1.79O0.47 and CH1.07O0.14, respectively. Compared to the compositional formula (CH2.84O0.56) of the bio-oil obtained

by the atmospheric pressure pyrolysis of rice husk in a fixed-bed reactor without catalyst by Zhou et al. (2013), our compositional formula of the bio-oil obtained even under atmospheric pressure consisted of less hydrogen. The discrepancy might be because their composition of bio-oil was determined without excluding water. Moreover, the high heating values (HHV) of bio-oil were estimated according to Dulong formula. It was found that the HHV increased gradually from 20.96 MJ/kg to 33.84 MJ/kg with elevating pressure from 0.1 MPa to 5.0 MPa with the constant volume flow rate of 500 mL/min. 3.4. Compounds in bio-oil The bio-oils obtained by the pyrolysis of rice husk under different pressures were detected with the use of GC–MS. It was noticed from the chromatograms that there were more peaks detected for the bio-oil obtained under atmospheric pressure than that obtained under pressure (not shown), implying that the pressurized pyrolysis resulted in the decomposition of some thermally unstable compounds. Table 2 shows the main compounds identified in two typical bio-oils obtained at 0.1 MPa and 3.0 MPa. As expected, almost all the compounds assigned in both bio-oils were the oxygen-containing compounds. No benzene, toluene and xylene (BTX) were detected in both bio-oils, also in the bio-oils obtained under 0.5 MPa, 1.0 MPa and 5.0 MPa (not shown). This suggested that the pyrolysis condition even with the high pressure was not rigorous enough to decompose the oxygen-containing compounds to form such stable compounds as BTX. The comparison between two bio-oils showed that acetic acid and its derivatives were the prominent compounds in two bio-oils. However, the acetic acid derivatives such as peak nos. 2 and 17 present in the bio-oil obtained under atmospheric pressure disappeared in the bio-oil obtained under 3.0 MPa. In addition, some saccharides

101

Y. Qian et al. / Bioresource Technology 174 (2014) 95–102

5.0

0.1 MPa 1.0 MPa 5.0 MPa

4.0 3.0

0.5 MPa 3.0 MPa

a

Yield (wt%-dry biomass)

0.9 0.6 0.3 0.0 5.0

b

4.0 3.0 0.9 0.6 0.3 0.0 Acetic acid

Furural Furfuryl alcohol PCX

Guaiacol

Naphthalene

Fig. 5. The yields of 8 compounds in the bio-oils produced from the pyrolysis of rice husk under varying pressures with the constant linear velocity of gas flow (a) or with the constant volume flow rate (b).

such as D-gluco-heptulosan present in the former bio-oil absent in the latter bio-oil. In the light of the compounds detected by GC–MS, the changes in the yields of 8 compounds with pressure were quantitatively examined. The results are shown in Fig. 5, where the yields of phenol, cresol and xylenol (PCX) are summed together. It was observed that the yield of acetic acid was by far higher than the yields of any other compounds, with a maximal yield of 4.6% being obtained under 3.0 MPa with the constant linear velocity of flow rate. A high yield of acetic acid in the bio-oil may allow for the efficient production of acetic acid in bio-oil refinery. In the case of the constant linear velocity of flow gas (Fig. 5a), the yield of acetic acid increased with increasing pressure from 0.1 MPa to 3.0 MPa, while it dropped slightly from 3.0 MPa to 5.0 MPa. The yields of both furfural and furfuryl alcohol (i.e., 2-furanmethanol) decreased with increasing pressure, whereas the yield of PCX increased with increasing pressure. The pressurized pyrolysis also enhanced the yields of guaiacol and naphthalene. Comparison between Fig. 5a and b pressure demonstrated that a longer residence time substantially increased the yields of acetic acid and PCX, irrespective of the pressures of 0.1, 0.5, 1.0, and 3.0 MPa used. The longer residence time enhanced the yield of furfural but overall reduced the yields of furfuryl alcohol.

4. Conclusions The pressurized pyrolysis was favorable for the preferential deoxygenation of bio-oil as a result of the intensified dehydration and decarboxylation. The bio-oil obtained under 5.0 MPa with a gas flow rate of 500 mL/min was estimated to have a compositional formula of CH1.07O0.14, in contrast to the formula of CH1.95O0.63 for the bio-oil obtained under atmospheric pressure with the same flow rate. Moreover, the high pressure did not cause more polycondensation of bio-oil to char and extensive cracking of bio-oil to gas. Acetic acid, furans, PCX and guaiacols were the typical compounds in the bio-oil obtained under the elevated pressure.

Acknowledgement This work is partially due to the Natural Science Foundation of China (Grand No. 21376080). References Alvarez, J., Lopez, G., Amutio, M., Bilbao, J., Olazar, M., 2014. Bio-oil production from rice husk fast pyrolysis in a conical spouted bed reactor. Fuel 128, 162–169. Amutio, M., Lopez, G., Alvarez, J., Moreira, R., Duarte, G., Nunes, J., Olazar, M., Bilbao, J., 2013. Pyrolysis kinetics of forestry residues from the Portuguese Central Inland Region. Chem. Eng. Res. Des. 91, 2682–2690. Asadullah, M., Rahman, M.A., Ali, M.M., Motin, M.A., Sultan, M.B., Alam, M.R., Rahman, M.S., 2008. Jute stick pyrolysis for bio-oil production in fluidized bed reactor. Bioresour. Technol. 99, 44–50. Bakar, M.S.A., Titiloye, J.O., 2013. Catalytic pyrolysis of rice husk for bio-oil production. J. Anal. Appl. Pyrol. 103, 362–368. Ben, H., Ragauskas, A.J., 2013. Comparison for the compositions of fast and slow pyrolysis oils by NMR characterization. Bioresour. Technol. 147, 577–584. Bridgwater, A.V., 1999. Principles and practice of biomass fast pyrolysis processes for liquids. J. Anal. Appl. Pyrol. 51, 3–22. Chen, T., Wu, C., Liu, R., Fei, R., Liu, S., 2011. Effect of hot vapor filtration on the characterization of bio-oil from rice husks with fast pyrolysis in a fluidized-bed reactor. Bioresour. Technol. 102, 6178–6185. Chhiti, Y., Salvador, S., Commandré, J.M., Broust, F., 2012. Thermal decomposition of bio-oil: focus on the products yields under different pyrolysis conditions. Fuel 102, 274–281. Ellens, C.J., Brown, R.C., 2012. Optimization of a free-fall reactor for the production of fast pyrolysis bio-oil. Bioresour. Technol. 103, 374–380. Gercel, H.F., 2002. Production and characterization of pyrolysis liquids from sunflower-pressed bagasse. Bioresour. Technol. 85, 113–117. Grierson, S., Strezov, V., Shah, P., 2011. Properties of oil and char derived from slow pyrolysis of Tetraselmis chui. Bioresour. Technol. 102, 8232–8240. He, M., Mourant, D., Gunawan, R., Lievens, C., Wang, X.S., Ling, K., Bartle, J., Li, C.Z., 2012. Yield and properties of bio-oil from the pyrolysis of mallee leaves in a fluidised-bed reactor. Fuel 102, 506–513. Heo, H.S., Park, H.J., Yimb, J.H., Sohn, J.M., Park, J., Kime, S.S., Ryu, C., Jeon, J.K., Park, Y.K., 2010. Influence of operation variables on fast pyrolysis of Miscanthus sinensis var. purpurascens. Bioresour. Technol. 101, 3672–3677. Jung, S.H., Kim, S.J., Kim, J.S., 2012. Characteristics of products from fast pyrolysis of fractions of waste square timber and ordinary plywood using a fluidized bed reactor. Bioresour. Technol. 114, 670–676. Li, R., Zhong, Z.P., Jin, B.S., Zheng, A.J., 2012. Application of mineral bed materials during fast pyrolysis of rice husk to improve water-soluble organics production. Bioresour. Technol. 119, 324–330. Mansaray, K.G., Ghaly, A.E., 1999. Determination of kinetic parameters of rice husks in oxygen using thermogravimetric analysis. Biomass Bioenergy 17, 19–31.

102

Y. Qian et al. / Bioresource Technology 174 (2014) 95–102

Mercader, F.M., Groeneveld, M.J., Kersten, S.R.A., Venderbosch, R.H., Hogendoorn, J.A., 2010. Pyrolysis oil upgrading by high pressure thermal treatment. Fuel 89, 2829–2837. Pattiya, A., 2011. Bio-oil production via fast pyrolysis of biomass residues from cassava plants in a fluidised-bed reactor. Bioresour. Technol. 102, 1959–1967. Shi, X., Wang, J., 2014. A comparative investigation into the formation behaviors of char, liquids and gases during pyrolysis of pinewood and lignocellulosic components. Bioresour. Technol. 170, 262–269. Shi, L., Yu, S., Wang, F.C., Wang, J., 2012. Pyrolytic characteristics of rice straw and its constituents catalyzed by internal alkali and alkali earth metals. Fuel 96, 586– 594. Stefanidis, S.D., Kalogiannis, K.G., Iliopoulou, E.F., Lappas, A.A., Pilavachi, P.A., 2011. In-situ upgrading of biomass pyrolysis vapors: catalyst screening on a fixed bed reactor. Bioresour. Technol. 102, 8261–8267. Tsaia, W.T., Lee, M.K., Chang, Y.M., 2007. Fast pyrolysis of rice husk: product yields and compositions. Bioresour. Technol. 98, 22–28. Veses, A., Aznar, M., Martínez, I., Martínez, J.D., López, J.M., Navarro, M.V., Callén, M.S., Murillo, R., García, T., 2014. Catalytic pyrolysis of wood biomass in an auger reactor using calcium-based catalysts. Bioresour. Technol. 162, 250–258. Wang, Z., Cao, J., Wang, J., 2009. Pyrolytic characteristics of pine wood in a slowly heating and gas sweeping fixed-bed reactor. J. Anal. Appl. Pyrol. 84, 179–184. Wang, S., Wang, Q., Jiang, X., Han, X., Ji, H., 2013. Compositional analysis of bio-oil derived from pyrolysis of seaweed. Energy Convers. Manage. 68, 273–280.

Whitty, K., Kullberg, M., Sorvari, V., Backman, R., Hupa, M., 2008. Influence of pressure on pyrolysis of black liquor: 2. Char yields and component release. Bioresour. Technol. 99, 671–679. Worasuwannarak, N., Sonobe, T., Tanthapanichakoon, W., 2007. Pyrolysis behaviors of rice straw, rice husk and corncob by TG–MS technique. J. Anal. Appl. Pyrol. 78, 265–271. Yoon, S.J., Son, Y.L., Kim, Y.K., Lee, J.G., 2012. Gasification and power generation characteristics of rice husk and rice husk pellet using a downdraft fixed-bed gasifier. Renewable Energy 42, 163–167. Zhang, J., Wang, X., Wang, F., Wang, J., 2014. Investigation of hydrogasification of low-rank coals to produce methane and light aromatics in a fixed-bed reactor. Fuel Process. Technol. 127, 124–132. Zhang, H., Xiao, R., Wang, D., He, G., Shao, S., Zhang, J., Zhong, Z., 2011. Biomass fast pyrolysis in a fluidized bed reactor under N2, CO2, CO, CH4 and H2. Bioresour. Technol. 102, 4258–4264. Zheng, J.L., 2007. Bio-oil from fast pyrolysis of rice husk: yields and related properties and improvement of the pyrolysis system. J. Anal. Appl. Pyrol. 80, 30–35. Zheng, J.L., 2008. Pyrolysis oil from fast pyrolysis of maize stalk. J. Anal. Appl. Pyrol. 83, 205–212. Zhou, L., Yang, H., Wu, H., Wang, M., Cheng, D., 2013. Catalytic pyrolysis of rice husk by mixing with zinc oxide: characterization of bio-oil and its rheological behavior. Fuel Process. Technol. 106, 385–391.