Bio-oil and bio-char from low temperature pyrolysis of spent grains using activated alumina

Bioresource Technology 102 (2011) 10695–10703

Contents lists available at SciVerse ScienceDirect

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

Bio-oil and bio-char from low temperature pyrolysis of spent grains using activated alumina Aimaro Sanna a,⇑, Sujing Li a, Rob Linforth b, Katherine A. Smart b, John M. Andrésen a a b

Energy and Sustainability Research Division, Department of Chemical and Environmental Engineering, University of Nottingham, University Park, Nottingham NG7 2RD, UK Division of Food Sciences, School of Biosciences, University of Nottingham, Sutton Bonington LE12 5RD, UK

a r t i c l e

i n f o

Article history: Received 1 June 2011 Received in revised form 18 August 2011 Accepted 21 August 2011 Available online 30 August 2011 Keywords: Biomass Pyrolysis Bio-oil Spent grains Bio-refinery

a b s t r a c t The pyrolysis of wheat and barley spent grains resulting from bio-ethanol and beer production respectively was investigated at temperatures between 460 and 540 °C using an activated alumina bed. The results showed that the bio-oil yield and quality depend principally on the applied temperature where pyrolysis at 460 °C leaves a bio-oil with lower nitrogen content in comparison with the original spent grains and low oxygen content. The viscosity profile of the spent grains indicated that activated alumina could promote liquefaction and prevent charring of the structure between 400 and 460 °C. The biochar contains about 10–12% of original carbon and 13–20% of starting nitrogen resulting very attractive as a soil amendment and for carbon sequestration. Overall, value can be added to the spent grains opening a new market in bio-fuel production without the needs of external energy. The bio-oil from spent grains could meet about 9% of the renewable obligation in the UK. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Bio-ethanol is the main bio-based transportation fuel with a global production of around 67 billion liters in 2008 (REN21, 2009). Its current market share of above 85% is forecast to remain strong for the near and mid-term future since it is the bio-fuel of choice to meet various governmental targets world-wide (Hahn-Hägerdal et al., 2006). Bio-ethanol production is mainly derived from feedstocks available in large quantities locally. In Europe the current bio-ethanol feedstock is cereals, predominantly wheat. Wheat spent grains (WSG) is the main by-product of the wheat deconstruction process with about 300–400 kg of WSG generated from each ton of wheat processed (Sadhukhan et al., 2008; Taheripour et al., 2010). Furthermore, if the WSG is not sold as animal feed (OECD-FAO, 2009) the impact on profit from the process can be seriously impacted. It is forecasted that about 32 Mt of spent grains will be available by 2015 in the US alone (Taheripour et al., 2010), and accordingly, the bio-ethanol industry is currently exploiting new ways to maximize its margin by further recovery of either energy or co-products from WSG. In addition to WSG other industries such as brewing and distilling produce significant quantities of malted barley spent grains, known as brewers spent grains (BSG) and distillers spent grains (DSG). Several uses of spent grains to produce energy have been proposed (Mussatto et al., 2006). The route deemed closest to market involves combined heat and power which can utilize the spent grains

⇑ Corresponding author. Tel.: +44 0 1159514198. E-mail address: [email protected] (A. Sanna). 0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2011.08.092

(SG) to generate energy near bio-ethanol or brewing production sites (Appleyard, 2009). A major drawback with utilizing SG in combined heat and power is the high nitrogen, phosphorus and alkaline metal, such as Na and K, content of the SG which resulting in biofouling and the generation of pollutants such as nitrogen oxide. An alternative use of SG involves bio-gas generation from anaerobic digestion, and although this route has significant political backing, there are constraints with respect to the generation of solid wastes and process water or effluent (Eskicioglu et al., 2010). A more recent proposition involves the use of SG digests as feedstocks for bio-ethanol fermentation, and whilst this has potential the yields of ethanol are poor because neither the release of sugars from the SG nor the simultaneous fermentation of hexose and pentose sugars have been suitably optimized (White et al., 2008). Here we describe an alternative means of releasing energy from SG using thermo-chemical conversion of SG. Using this approach, catalytic pyrolysis can convert SG into a bio-oil suitable for upgrading into transportation fuels and a bio-char where the nitrogen, phosphorus and alkaline metals are concentrated in sequestered carbon (Becidan et al., 2007). Traditional pyrolysis of biomass has been carried out since the early 1970s and is a well established technique that thermo-chemically converts organic matter in an oxygen free atmosphere into a bio-oil, a bio-char and a bio-gas (Bridgwater et al., 1999). The bio-oil obtained is usually a very complex mixture of oxygenated aliphatic and aromatic hydrocarbons that when compared with conventional petroleum fuel, exhibits high viscosity, high oxygen content and low stability (Demirbas, 2007; Bridgwater et al., 1999). A key to large scale use of pyrolysis oil is therefore the removal of oxygen to facilitate its further conversion into industrial

10696

A. Sanna et al. / Bioresource Technology 102 (2011) 10695–10703

commodity chemical feedstocks such as alcohols and olefins (Vispute et al., 2010). In addition to bio-oil, bio-char is attracting growing attention as a valuable by-product since it is able to sequestrate carbon and may be used as a soil amendment (Mathews John, 2008). Bio-char could then be an important tool in CO2 reduction and improve agricultural productivity. A major challenge is the generation of a bio-chemical stable char with high nitrogen loading. Recent Thermo Gravimetric Analysis (TGA) work on the kinetics of spent grains pyrolysis indicate a retention of nitrogen in chars from SG at low temperatures (Boateng et al., 2007; Becidan et al., 2007; Giuntoli et al., 2009). The use of catalysis during pyrolysis, or pyro-catalysis, may result in a high bio-oil conversion at low temperatures whilst still generating chars with high nitrogen content. Recently, calcined alumina has been used for pyro-catalytic conversion of corncobs and mischantus at 600 and 550 °C, respectively (Atesß and Isßikdag˘, 2009; Yorgun and Emre Sßimsßek, 2008). Alumina is a synthetic white oxide of aluminum (Al2O3) that shows Lewis acid catalytic activity when calcined at temperature above 450 °C when surface protons and aluminum cations become mobile and start to change their position in the lattice, and it has shown to be very active towards reducing tar and coke formation during gasification (Corma and Garcia, 2003; Samolada et al., 2000). The number of acid sites is related to the abundance of Al2O3 in the sand and is maximized at Al2O3 content close to 100% (Yorgun and Emre S ß imsßek, 2008). A very interesting aspect of calcined alumina is that it appears to maintain liquid yield while reducing the oxygen content of the oil. The bio-oil yield was increased of about 10-20% with the use of activated alumina compared to the use of silica sand alone (Yorgun and Emre S ß imsßek, 2008; Atesß and Isßikdag˘, 2009). Alumina acid sites promote breaking and formation of carbon–hydrogen bond on or close to double bonds, cracking of carbon–carbon bonds, decarboxylation, decarbonylation and dehydration reactions and could also counteract the base catalyzed polymerization reactions that result from the SG alkaline metals during pyrolysis thereby decreasing the yield of bio-oil and increasing the yield of char (Corma and Garcia, 2003; Nowakowski et al., 2007). Accordingly, this study focuses on lowtemperature pyro-catlytic conversion of SG into a low-oxygen containing bio-oil and high nitrogen containing bio-char using an alumina catalyst. 2. Methods 2.1. Biomass sample and catalyst Wheat spent grains (WSG) and brewers spent grain (BSG) were obtained as by-products from a pilot scale brewer at the University of Nottingham. Alumina sand 40–60 grit with Al2O3 content of 91% was procured from Ken Clark International, Maidenhead, Berckshire, UK and its particle size was distributed between 250 lm and 355 lm, with a mean diameter of 270 lm and was classified as ‘Geldart B’ particles. Alumina is very hard and wear resistant, with high compressive strength even against extreme temperatures and corrosive environments resulting ideal for fluidised bed applications. Also, the thermal conductivity of alumina (30 W/ m K) is higher compared to that of silica sand (1.3 W/m K) enhancing the capacity to transfer the heat to the biomass particles in to the fluidised bed reactor. The particles had a density of 3.8 g/cm3 and a max working temperature of 1750 °C. Alumina was calcined at 460 °C for 80 min before each experiment, total pore volume of 0.04 cm3/g and a surface area of 5 m2/g. 2.2. Pyro-catalytic set-up The experimental devices used in this work principally consisted of a pressurized injection system, a sample chamber, a fluidized bed reactor, an electrical heater and a tar trap as de-

scribed by Sanna et al. (2009) ‘‘see Supplementary Figs. e and f in the online version of this article’’. The tar trap was comprised of 3 Dreshel bottles 500, 250 and 150 ml in series. All Dreshel bottles were placed in an ice trap with ice and water to condense the condensable gases into bio-oil. Thermocouples connected to the reaction chamber were used to check the temperature during the experiments. Approximately 114 g of alumina with a mean particle diameter of 270 lm was placed into the fluidized bed column and fluidized at the desired temperature by a nitrogen flow set at 5.5 L/ min. After stabilization of the system, the sample was continuously injected using a nitrogen flow of 3.5 L/min for 10 min to ensure all the 5 g were injected. The reaction chamber was a schedule 40, SS 316 stainless steel column 65 cm in height and with internal diameter of 4.1 cm, a volume of 858.5 cm3. The time for the thermochemical conversion was of 6 s (retention time based on flow rate of 8.5 L/min and column volume). For each replicate a total of 5 g sample was applied. Three replicates of each analysis were conducted and for each replicate a total of 5 g sample was reinserted into the sample chamber and continuously re-injected for 10 min . In order to establish the effect of the temperature on the yield and properties of bio-oil, pyrolysis experiments were conducted at the following temperatures: 460–490–520 and 540 °C. The bio-oil mass balance was calculated, considering the weigh difference of the Dreshel bottles before and after reactions. The bio-char weight was calculated considering the difference between the alumina bed plus bio-char (after reaction) minus the initial alumina bed weight. Finally the gas yield was obtained by the initial weight minus the weight of the bio-oil and bio-char. After each run the bio-oil was collected in a vial and stored in a refrigerator. The bio-gas produced in each experiment was collected in gas-bags and stored in a refrigerator. The gas sampling was carried out since the injection of the samples into the reactor for 15 seconds. 2.3. Spent grains and pyrolysis products characterization The High Heating Values (HHV) of each sample was evaluated using an IKA 5000 Series Bomb Calorimeter. Calibration was carried out with benzoic acid tablets from Digital Data System Ltd. The composition of the bio-oils was analyzed using a Hewlett Packard HP6890 GC gas chromatograph–mass spectrometer (GC– MS) attached to a Hewlett Packard HP5973 Mass Selective Detector (MSD) with ionizing energy 70 eV, source temperature 280 °C and transfer line 300 °C (Agilent Technologies Inc., Santa Clara, USA). 1.5 mg of the bio-oil was dissolved in 1 mL of dichloromethane and 1 lL injected into the GC (injection temperature 250 °C). Separation was achieved on a VF-1MS-low bleed 100% dimethylpolysiloxane column (30 m length, 0.25 mm internal diameter and 0.25 lm film thickness) with helium as a carrier gas (100 kPa). The column oven temperature programmed was- initial temperature 50 °C (hold 3 min), ramp to 100 °C (20 °C min 1; hold 3 min), ramp to 200 °C (20 °C min 1; hold 10 min). The GC–MS calibration was carried out using a standard mixture of aliphatic compounds with carbon number ranging from 5 to 30. Complex viscosity measurements were performed using a Rheometrics RDA-III high-torque controlled strain rheometer. About 1.5 g of biomass with particle sizes between 350 and 512 lm were pressed under 5 tons of pressure in a 25-mm die to form disks with a thickness of approximately 2.5 mm. The operational parameters of rheometer were a strain of 0.1%, a ramp temperature from 40 to 550 °C at 3–6–10–15–20 °C/min and an auto-tension of 20 g. The high-resolution solid-state 50 MHz 13C Nuclear magnetic resounance (NMR) analysis was conducted using a cross polarization pulse sequence in conjunction with magic angle spinning in a Bruker Avance 200 spectrometer. Tetrakis(trimethylsilyl)silane was added to the samples as an internal standard. Liquid 1H NMR analysis was performed using a Bruker AV(III)500 equipped 60

A. Sanna et al. / Bioresource Technology 102 (2011) 10695–10703

position autosampler, with a 5 mm dual 1H/13C cryoprobe with z gradients, automated gradient shimming capability. The spent grains composition % (cellulose, hemicellulose, lignin, proteins and fatty acids) was obtained by integration and interpretation of the deconvoluted 13C NMR spectra (Bardet et al., 1997; Gil and Neto, 1999; Freitas et al., 2001). The elemental analyses were performed using a Flash EA 1112 series elemental analyzer set to the detection of N, C, S and H. the oxygen content was determined by the following formula: O% = 100 (N% + C% + S% + H%). Dry ash free (daf) calculations were carried out according with ASTM D-3180. DL-Metionina standard from Thermo Electron Corporation (P/N 33840006) was used. Vanadium Pentoxide (BN/164616) from Elemental Microanalysis was added to the samples to detect sulfur. The weight loss experiments to determine the proximate analysis (moisture, volatile matter (VM), fixed carbon (FC) and ash contents) were carried out using a Perkin Elmer Pyris1 thermo-gravimetric analyzer. FC is the material which is left after volatile materials are driven off. Before each run, the crucibles were tared and then about 10 mg of sample were spread in the crucible. High purity of nitrogen was used at 100 ml/min to avoid oxidation reactions. The gases collected in gas bags after each experiment were analyzed using a Hiden analytical HPR20 MS with 200 amu mass range capability to evaluate the content of CO2, H2 and CH4 and equipped with a HAL-RC quadrupole mass spectrometer. The Flow Control Inlet ensured stable mass spectrometer operation through a sample pressure range from 200 mbar to 2 bar. A Quartz Inlet Capillary (QIC) fast sampling column (direct inlet with < 500 ms response to gases/vapors, 2 m length, 0.3 mm internal diameter, 0.02 m orifice diameter, 200 °C inlet temperature) was used to sampling 1cm3/ min (in a continuous way) the gas from the gas-bag at 1 bar. The scan range used was between 1 and 200 amu and the scan speed was of 10,000 amu/s. The MS was run until the signal from the MS was stable. Pure carbon dioxide and nitrogen were used to calibrate the instrument. The condensable compounds that entered into the gas bags such as pentane, hexane, hexene and toluene were analyzed using the same HP GC–MS used for the bio-oils but with a different column. An Agilent Factor Four™ capillary column (VF – 1 ms, 50 m, 0.25 mm, 0.25 lm, part number CP8914) was used with 1 mL/min helium as a carrier gas. 10 lL of bio-gas was injected (injector 250 °C) for each bio-gas using a syringe. The column oven temperature programmed was initial temperature 40 °C (hold 4 min), ramp to 150 °C (10 °C min 1; hold 3 min), ramp to 200 °C (10 °C min 1; hold 10 min). Ten microliters of carbon dioxide was used as a calibration gas. Finally, the energy efficiency used in the sensitivity analysis in Section 3.6 was calculated by the equation Effgeneral = (Etotout/ Etotin) ⁄ 100 where Etotin represents the energy content of the spent grains and Etotout was calculated by removing from the products energy (Echar + Ebio-oil + Egas) the energy required to dry the spent grains (Edrying), the energy to reduce their particle size (Ecrushing), the energy to run the pyrolysis (Epyro) and the energy required to produce the ice to condense the condensable gases into bio-oil (Eice tar-trap). Edrying was calculated considering the heat of evaporation of water, the sensible heat of heating water from 20 to 100 °C and the sensible heat for heating the residual 10% of water not removed by the drying step from 100 to 520 °C.

3. Results and discussion 3.1. Samples characterization Table 1 compares the proximate and ultimate analysis of the wheat and barley spent grains with bituminous coal. WSG and

10697

BSG have a carbon content of 43.2 and 49.8 weight% (wt%), respectively, which is low compared to that of bituminous coal which has a carbon content of 81.5 wt%. Elemental analysis illustrates the challenging characteristics of biomasses containing low carbon content and high oxygen content in comparison with fossil fuels such as coal (Sanna et al., 2009). Table 1 also shows that the calorific values are only 18.35 MJ/kg and 18.55 MJ/kg for WSG and BSG, respectively, compared to 33.5 MJ/kg for coal. Furthermore BSG exhibits lower levels of volatile matter of 61.4 wt% and an oxygen content of 39.4 wt% compared with WSG’s 75.2 wt% and 45.8 wt%, respectively. The nitrogen content was high for both samples with 4.5 and 4.1 wt% detected for WSG and BSG, respectively. It is suggested that the higher nitrogen content reflects the presence of proteins in the samples and this hypothesis was confirmed by analysis to be within the range 5-7 wt%. The solid state 13C NMR spectra of WSG, BSG ‘‘see Supplementary Figs. e and f in the online version of this article’’ were compared to those of lignin, cellulose and hemicellulose. The peak at 21.8 ppm is related to acetyl methyl groups of hemicellulose and the peak around 180–175 ppm is linked with carboxyl carbon of long chain carboxylic acids such as palmitic or oleic fatty acids. The region between 63.5 and 105.8 ppm contains peaks originating from cellulose and hemi-cellulose sugar carbons. In the lignin spectrum, the region between 111 and 150 ppm represents aromatic carbon while the region at 55 ppm represents methoxy groups. The area between 50 and 20 ppm is associated with carbonyl from hemicellulose or aliphatic carbons of saturated fatty acids and proteins in the spent grains spectra (Bardet et al., 1997; Gil and Neto, 1999; Freitas et al., 2001). The WSG curve matches very well with that expected for hemi-cellulose. In contrast, the BSG curve shows more similarity with those of cellulose and lignin indicating a higher presence of these two bio-polymers in the BGS structure. The chemical composition of the samples from 13C NMR correlates well with their chemical history. Wheat and barley grains are mainly composed from starch (60–65 wt%) and proteins (10–15 wt%) (Kanauchi et al., 2001). Saccharification of the grains leave a residue suitable for a thermo-conversion process because of its lignin-cellulosic content has increased dramatically to 75–85 wt% in comparison with the original material (Annetts and Audsley, 2003). Accordingly, 13C NMR shows that the hemicellulose and cellulose represent 42 and 34 wt% for WSG and 38 and 31 wt% for BSG, compared to only about 15 wt% of cellulose and hemicellulose in the starting grains. The lignin increases from about 1 wt% in wheat and barley grains to 4.8 wt% for WSG and 6.8 wt% for BSG and this is reflected on the spent grains HHV (Table 2). In fact, lignin has a high calorific value of 29 MJ/kg while cellulose generally present a lower value of 17–18 MJ/kg (Fernando et al., 2006; Sadhukhan et al., 2008). The presence of proteins, 4.9 wt% in WSG and 6.6 wt% in BSG, could represent a limitation towards fuel use due to the possible emission of nitrogen oxides. Furthermore, the spent grains are rich in fatty acids showing 13 and 17 wt% for WSG and BSG, respectively. The viscosity of wheat spent grains as a function of temperature is shown in Fig. 1a. The viscosity curves can be used to describe the first stages of the pyrolysis reactions by following the transition between the solid and pseudo-liquid state of the cellulose and hemicellulose structures. The influence of five different heating rates was investigated. The solid to pseudo-liquid state transition begins at 260 °C for the lower heating rate of 3 °C/min associated with the breakdown of cellulose. This temperature is shifted from 260 °C to 275 °C when the heating rate is increased to 20 °C/min. Similarly, the temperature of lowest viscosity is shifted to higher temperature from 275 °C at 3 °C/min to 300 °C at 20 °C/min. The minimum viscosities decreased from 27,584 Pa s at 3 °C/min to 7876 Pa s at 20 °C/min indicating that the polymeric components are in a soften-like solid-state that precedes the release of volatiles.

10698

A. Sanna et al. / Bioresource Technology 102 (2011) 10695–10703

Table 1 The characteristics of wheat spent grains (WSG), and brewers spent grains (BSG) compared to bituminous coal are illustrated by the proximate and elemental analysis. Standard deviations (SD) were obtained from triplicate. VM: volatile matter, FC: fixed carbon, HHV: high heating value. wt% dry ash free (daf) Elemental Analysis

WSG BSG Bit. coala a

Proximate Analysis

C

SD

H

SD

N

SD

O

SD

S

SD

Moisture

SD

Ash

SD

VM

SD

FC

SD

HHV(MJ/kg)

SD

43.20 49.80 81.50

0.2 0.3 0.4

6.50 6.38 5.30

0.8 0.4 0.1

4.50 4.14 2.10

0.39 0.15 0.07

45.80 39.36 11.10

0.24 0.76 0.39

0.10 0.10 0.10

na na na

6.60 4.60 0.78

0.21 0.06 0.14

2.20 6.50 6.20

0.25 0.13 1.35

75.20 61.40 36.80

1.21 1.65 1.35

16.00 17.00 63.20

0.40 0.50 0.46

18.35 18.55 33.50

0.16 0.08 0.18

Bituminous coal data from Sanna et al. (2009).

Table 2 Proximate and elemental analysis of WSG and BSG bio-oils obtained at the different experimental temperatures. SDs from triplicates are reported. wt% dry ash free (daf) WSG

Moisture Ash Volatile matter Fixed carbon Carbon Hydrogen Nitrogen Oxygen Sulfur HHV (MJ/kg)

BSG

460

SD

490

SD

520

SD

540

SD

460

SD

490

SD

520

SD

540

SD

11.10 0.00 78.98 21.22 52.65 6.22 6.90 34.22 <0.1 24.52

0.30 na 1.56 0.76 1.18 0.07 0.09 0.76 na 0.18

12.00 0.00 78.30 21.70 51.02 6.13 6.20 36.64 <0.1 23.30

0.56 na 1.14 0.45 0.63 0.18 0.08 0.63 na 0.93

15.72 0.00 79.42 20.58 50.86 5.82 6.06 37.26 <0.1 22.80

0.07 na 0.79 0.67 1.23 0.06 0.37 1.23 na 0.13

10.38 0.00 74.78 25.01 52.25 5.83 7.43 34.48 <0.1 23.30

0.44 na 1.39 0.46 1.67 0.09 0.17 1.67 na 0.40

17.38 0.10 79.36 20.64 55.16 6.01 8.00 30.73 <0.1 25.75

0.36 0.10 0.76 0.34 1.66 0.13 0.35 1.66 na 0.32

17.30 0.00 77.80 22.20 51.30 5.98 7.50 35.12 <0.1 23.00

0.25 na 0.84 0.93 2.92 0.04 0.71 2.92 na 0.03

20.88 0.10 76.97 23.03 52.85 5.95 8.50 32.69 <0.1 24.07

0.13 0.03 0.57 0.50 1.06 0.14 0.04 1.06 na 0.09

18.60 0.00 78.24 21.76 52.47 5.13 7.21 35.09 <0.1 22.11

0.24 na 1.21 0.65 0.92 0.72 0.08 0.92 na 0.13

Therefore, it is possible to maximize the pseudo-liquefaction state using a heating rate about 20 °C/min or higher. After the viscosity reached a minimum, a rapid increase was observed due to the rapid volatilization of smaller compounds and polymerization of the structures into a char with higher viscosity. Fig. 1b shows the predicted temperature at which it is possible to obtain the minimum complex viscosity that indicates the temperature at which the polymeric structure soften (pseudo-liquid state) in relation to the heating rate used in the pyrolysis experiments (about 150 °C/s). The logarithmic trend suggests that between 400 and 460 °C catalytic activities can promote liquefaction of the volatilized gases and retain the aliphatic carbon structure of the original biomass. The derivative thermo-gravimetric (DTG) curves of WSG and BSG ‘‘see Supplementary Fig. c in the online version of this article’’ using a heating rate of 15 °C/min indicate that around 100 °C both samples show a dehydration peak due to water removal. Decomposition of the hemicellulose starts around 250 °C for both the spent grains. The hemicellulose and cellulose decomposition profiles were not clearly distinguishable for the WSG, but very well separated for BSG. The hemi-cellulose decomposition had a maximum between 310 °C and 340 °C for BSG and WSG. The devolatilization of cellulose took place around 380 °C for both samples. The decomposition peaks of WSG fit very well with the viscosity plot previously shown, with the volatilization starting around 300 °C followed by a re-solidification stage associated with the devolatilization peak of WSG. Lignin does not presents a distinct peak and its decomposition was distributed at a range of temperatures that goes from ambient to 900 °C as reported in literature (Yang et al., 2007). The differential weight difference associated with hemi-cellulose is significantly different between the WSG and BSG even if the WGS and BSG have similar hemi-cellulose contents (42% vs. 38). This might be due to the sum of the decomposition of the other components (proteins and lipids) in the same range of temperature of hemi-cellulose decomposition (Giuntoli et al., 2009). In fact, the sum of hemi-cellulose, cellulose, protein and lipids is about 12% higher in WSG than BSG.

3.2. Bio-oil generation and characterization The effect of temperature on the bio-oil yield for WSG and BSG is shown in Fig. 2a and b, respectively. The two samples presented a similar yield trend due to quite similar composition of original materials (Table 1), where the bio-oil curve assumed an inverted V shape. The maximum yield was obtained at 520 °C with 53 and 49 wt% for WSG and BSG, respectively. At 460 °C the bio-char yield is about 20 wt%, while at 540 °C this yield decreases to about 15 wt%. Pyrolysis reactions in presence of alumina are shifted at low temperature so that at low and moderate temperature it maximizes the yield of bio-oil compared to that of non-catalytic reactions (Yorgun and Emre Sßimsßek, 2008; Atesß and Isßikdag˘, 2009). The char yield decreases with increasing temperature due to secondary decomposition reactions of char residue enhanced by the presence of acid sites in the alumina sand that at high temperature maximize the gaseous yield due to acid cracking principally of C-C bonds due to their low bond energy (Atesß and Isßikdag˘, 2009; Chen et al., 2011) and also due to secondary reactions including thermal cracking, re-polymerization and re-condensation of the char residues (Demirbas, 2007). Table 2 lists the proximate and elemental analyses for the biooils produced. The moisture content was quite high for all oils, ranging from 11 to 16 wt% for the oil from WSG and from 17% to 21% for the BSG oil. The volatile matter and fixed carbon were found to be similar for all the bio-oils ranging from 75 to 80 wt% and 21 to 25 wt%, respectively. The volatile matter showed a slight decrease with increasing temperature whereas the fixed carbon tended to increase. The sulfur content was very low for all the bio-oils and the ash content was virtually absent. Hot vapor filtration can also be successfully used to minimize ash and bio-char in the bio-oil (Chen et al., 2011). The carbon and oxygen contents ranged from 50% to 53% and from 31% to 37%, respectively. Also, 5–6% of hydrogen and 6-9% of nitrogen was present in the bio-oils. The energy content of the bio-oils varied between 22 and 26 MJ/kg. The energy content of the WSG and BSG bio-oils was higher at

A. Sanna et al. / Bioresource Technology 102 (2011) 10695–10703

a

10699

The bio-oil yields were highest at 520 °C with 45 and 53 wt% for BSG and WSG, respectively (see Fig. 2). Interestingly, both oils showed their higher energy content at 460 °C with 24.5 MJ/kg and 26 MJ/kg, for BSG and WSG, respectively. The energy content of the original WSG was 18.35 MJ/kg and increased 35% to 24.52 MJ/kg for the oil. Similar enhancement for the BSG bio-oil was 40%. Taking into account the overall conversion rate of WSG into bio-oil of 40% and 53% this indicates that about 55% and 66% of the starting energy of the WSG was retained in the oils at 460 and 520 °C, respectively. Similarly, about 57% and 63% of the initial BSG energy was kept in the bio-oils at 460 and 520 °C. The oxygen content of the oils was significantly lower than the original feedstock, indicating that the pyrolysis process has partially deoxygenated the biomass (Tables 1 and 2). The O/C ratio at 460 °C was found to be the lowest for both samples investigated as shown in Fig. 3. This suggests that lowering the experimental temperature could enhance deoxygenation. The H/C molar ratios of the bio-oils (Fig. 3) indicate that the hydrogen level decreases with increasing temperature. High temperature seems to favor a strong dehydrogenation probably due to increased cracking reactions. Of great interest is the part of the nitrogen contained in the starting spent grains (Table 1) and Table 3 compares its distribution into the reactions products. The SG present high nitrogen yield compared to traditional fossil fuels and this high content is reflected in the bio-oils as can be seen in Table 2. Overall, about 55–80 wt% of the intrinsic SG nitrogen is retained in the oils, while 18-30wt% is transferred to the gas and 13-19 wt% is retained in the char. The nitrogen concentration in the bio-char tends to increase with decreasing temperature. Also, the amount of nitrogen in the bio-oil is lower at 460 °C compared to 490 °C and 520 °C. Therefore, pyrolysis at low temperatures can be considered for its effectiveness on bio-oil quality improvement in terms of nitrogen reduction. Moreover, bio-chars rich in nitrogen might be used as soil amendment and possibly for carbon sequestration. As a result, pyrolysis at low temperature could be used to reduce the nitrogen level of spent grains producing bio-chars and bio-oils with enhanced quality.

b

Fig. 1. Effect of the heating rate on the relative viscosity of WSG plotted versus temperature (a, b) and derivative thermo-gravimetric profile of WSG and BSG (c). (a) Relative viscosities at different measured heating rates, (b) Logarithmic trend of speculated temperature at which the lowest viscosity can be reach at heating rate close to the experimental one (10,000 °C/min). The points represent the detected lowest viscosity at 5 different heating rates (3-a) used to speculate the temperature of max liquid state at heating rate close to 150 °C/s.

460 °C compared to 520 °C. On the other hand, the oxygen content was lower at 460 °C than 520 °C.

a

3.3. Chemical composition of Bio-oils 1

H NMR analysis was carried out to evaluate the overall composition of the bio-oils since it detects virtually all compounds in the oil. The 1H NMR spectra of the WSG bio-oils obtained at 460 and 520 °C were compared. Overall, the bio-oils generated from WSG were complex mixtures of compounds with a wide range of functional groups derived from the cracking of holo-cellulose, lignin, protein and lipid components. The interpretation of the 1H NMR chemical shift ‘‘see Supplementary Fig. d in the online version of

b

Fig. 2. Pyrolysis products distribution. (a) Products from pyrolysis of WSG at different temperatures (460–490–520–540 °C). The yield of bio-oil and bio-char is reported on the left and the gas yield on the right. (b) Products from pyrolysis of BSG at different temperatures (460–490–520–540 °C). The yield of bio-oil and bio-char is reported on the left and the gas yield on the right. Standard deviations on triplicates are reported in the error bars. Standard deviations on triplicates are reported in the error bars.

10700

A. Sanna et al. / Bioresource Technology 102 (2011) 10695–10703

Fig. 3. Comparison of the oxygen/carbon and hydrogen/carbon ratios of the WSG and BSG bio-oils obtained at the 4 different experimental temperatures. Error bars are showing the highest standard deviation of the elemental analysis used to calculate the elements ratios.

Table 3 Effect of temperature on total nitrogen distribution in the starting spent grains, into bio-oils and bio-chars for WSG and BSG. Nitrogen wt% was calculated considering the WSG and BSG oil and char mass balance and the nitrogen content from elemental analysis (standard deviation of each value < 0.7) of starting WSG, BSG and bio-oil and bio-char produced at the different temperatures. wt %

WSG (original) WSG bio-oil WSG bio-char BSG (original) BSG bio-oil BSG bio-char

460 °C

490 °C

520 °C

540 °C

100.0 61.3 19.2 100.0 67.6 18.0

100.0 62.0 18.5 100.0 81.2 18.7

100.0 71.5 16.5 100.0 82.1 15.7

100.0 54.5 16.0 100.0 57.5 13.3

sion of this article’’ include individual aromatics, aldehydes, ketones, esters, ethers, furan derivatives, fatty acids, alcohols, sterols and alkanes. These functionalities represent about 40% of detectable bio-oil compounds. GC–MS normally detects below 40% of the bio-oil (Garcia Peres et al., 2007) and it is not recommended for quantitative analysis. The peaks with higher volatilization rate include C4–C9 and represent compounds mainly derived from depolymerization, dehydration and decarboxylation-decarbonylation of lignin and cellulose. The main reaction pathway can be summarized by the depolymerization of the complex cellulose-hemicellulose structure with the rearrangement into 5-carbon units including furanone, cyclopentanedione, furfural, 5-HMF and pyrane, and 6-carbon units such as phenols and phenols substitutes including vinyl-methoxyphenol. The presence of methoxy groups indicates that demethylation took place during the pyrolysis reactions. The peaks with medium volatility from 20 to 30 min retention time represent saturated fatty acids, mainly palmitic and oleic acids, derived from the lipid residue of spent grains. Finally, the low volatiles above 30 min retention time include C28–C30 and indicate the presence of vegetal steroids such as stigmanstendiene and a-sitosterol, typically found in plants. Also, heterocyclic compounds possibly derived from spent grains proteins, such as oxazolidine and indole were detected. When comparing the GC– MS chromatogram with the 1H NMR some differences can be observed. The 1H NMR quantification indicates that about 14– 17 wt% of the bio-oils is made of aromatic compounds where as 1/3 of the GC/MS peaks are derived from phenolic compound. GC–MS shows an increase of 5-carbon units molecules like 2-(5H)-furanone and the pyrane substitutes (maltol) mainly originate from hemi-cellulose that already contains in its structure pentose sugars like arabinose, xylose and ribose and formed by depolymerization and sequent rearrangement of cellulose in 5/6carbon units including close and open rings compounds. 3.4. Bio-chars characterization

this article’’ was based on published work by other groups and the main groups were quantified (Onay and Koçkar, 2004; Junming et al., 2008). The temperature influences the conversion chemistry. The abundance of aliphatic, aldehydes and olefinic protons decreased and the aromatics increased with increasing temperature. This is in agreement with previous work indicating the shift at low temperature of reactions using alumina and the increase of aliphatic compared to non-catalytic bio-oils. Also, the production of single ing or more than 1-ring aromatic compounds (PAH) is lower when low pyrolysis temperature is used (Atesß and Isßikdag˘, 2009). Also, the pyrolysis at lower temperature reduces the oxygenated protons. The WSG bio-oil at 460 °C had 52 wt% of protons adjacent to oxygen (1.8–5 ppm) compared to 56 wt% for the oil at 520 °C. This is in accordance with O/H molar ratios showing lower oxygen content at lower temperature. The protons adjacent to aliphatic (0.5–1.8 ppm) and aromatic-phenolic or olefinic compounds (5– 6.5 ppm) were 27.4 and 27.1 wt% and 12 and 8 wt% at 460 °C and 520 °C, respectively. Also, 5.5 and 6.3 wt% of protons came from aromatic compounds (6.5–8 ppm) and finally, 2.7 and 2.1 wt% came from aldehydes (8–9 ppm) at 460 °C and 520 °C, correspondingly. Protons adjacent to nitrogen were not considered due to the overlap with oxygenated compounds that are present in higher concentration. The bio-oil generated from spent grains has low aromaticity and contains about 1/3 of aliphatics. Low aromatic content can be explained by the ‘intermediate’ pyrolysis conditions due to a residence time of 6 seconds that allowed extended breakage of the lignin structure. However, even if the oxygen has been reduced during the thermo-conversion, oxygenated compounds represent about half of the bio-oil. Compared to 1H NMR, the main functionalities that emerge from GC–MS analysis ‘‘see Supplementary Fig. d in the online ver-

The elemental analysis and proximate analysis of the chars produced during the pyrolysis are shown in Table 4. The calorific values for the bio-chars ranged from 20 to 25 MJ/kg where the HHV were found to be 25.3 MJ/kg and 25 MJ/kg at 520 °C for WSG and BSG chars, in that order. The carbon contents ranged from about 55% to 67% and its proportion increased with temperature. The highest carbon contents were 67% (SD 2.76) for WSG and 63% for BSG (SD 1.97) at 540 °C and 520 °C, respectively. The sulfur level was below 0.1% for all the bio-chars. A high level of 4–5 wt% nitrogen and 4–5 wt% hydrogen were detected. Nitrogen content decreased with increased pyrolysis temperature but there was little systematic variation in hydrogen and oxygen. The nitrogen trend was found similar in bio-char produced from other biomasses such as algae, straws and poultry litter (Yuan et al., 2011; Bird et al., 2011). The volatile matter of the chars decreased from 38% at 460 °C to 20% at 540 °C and the fixed carbon increased accordingly from 48% to 66%, as expected from the thermal-cracking of the biopolymer structure (Bonelli et al., 2001). The high ash content levels between 10 and 11 wt% and the nitrogen content up to 5 wt% makes the chars poor candidate for co-fire with coal particularly due to the low melting point of alkali and alkaline earth metals that tend to sinter causing ash slagging and fouling, corrosion and loss of fluidization in power generation boilers (Marsh et al., 2008). Also, spent grains release NOx precursors such as NH3, HCN and HNCO between 300 and 900 °C under nitrogen atmosphere (Giuntoli et al., 2009). The O/C molar ratios of the bio-chars give interesting information about the chemistry involved during the pyrolytic biomass deoxygenation. The O/C ratio of the bio-oils were lowest at 460 °C, while the O/C ratio of the bio-chars were highest at 460 (Table 4) indicating

10701

A. Sanna et al. / Bioresource Technology 102 (2011) 10695–10703 Table 4 Proximate analysis and elemental analysis of bio-chars obtained at the different experimental temperatures. Average values of triplicates are reported . wt% daf

Moisture Ash Volatile Matter Fixed Carbon Carbon Hydrogen Nitrogen Oxygen Sulfur H/C ratio O/C ratio HHV (MJ/kg)

W SG

B SG

WSG

BSG

W SG

B SG

W SG

B SG

460

SD

460

SD

490

SD

490

SD

520

SD

520

SD

540

SD

540

SD

3.90 10.00 38.50 47.60 59.50 4.99 5.27 30.24 <0.1 1.01 0.38 23.49

0.42 0.70 0.98 1.05 2.72 0.90 0.08 2.72 na na na 0.18

4.50 9.90 27.00 58.60 54.03 5.33 3.81 36.83 <0.1 1.18 0.51 20.08

0.26 0.52 0.74 0.84 2.29 0.17 0.50 2.29 na na na 0.24

3.50 10.10 35.00 51.40 58.45 4.93 4.81 31.80 <0.1 1.01 0.41 22.06

0.34 0.17 0.29 0.79 1.80 0.25 0.12 1.80 na na na 0.09

2.84 9.80 24.70 61.50 55.17 5.34 4.44 35.04 <0.1 1.16 0.48 21.90

0.31 0.24 0.41 0.88 0.76 0.45 0.08 0.76 na na na 0.31

3.90 10.50 25.70 59.90 65.23 5.11 5.43 24.22 <0.1 0.94 0.28 25.27

0.28 0.56 1.02 0.96 2.78 0.11 0.12 2.78 na na na 0.12

4.80 10.00 24.20 61.00 62.88 4.54 5.23 27.35 <0.1 0.87 0.33 24.98

0.14 0.26 0.37 0.41 1.97 0.14 0.13 1.97 na na na 0.16

5.00 11.00 23.50 60.50 67.26 4.58 4.15 24.01 <0.1 0.82 0.27 22.65

0.21 0.08 0.12 0.39 2.76 0.44 1.15 2.76 na na na 0.07

4.00 11.20 20.00 66.00 62.51 4.13 4.61 28.74 <0.1 0.79 0.34 24.57

0.10 0.27 0.15 0.74 1.20 0.11 0.15 1.20 na na na 0.11

that the chars produced at this temperature are richer in oxygen and might have retained oxygen from the bio-oils. Overall, the high carbon and nitrogen content of bio-char can provide nutrients to the soil and crop productivity could make it attractive to be used as a soil amendment particularly in acid soils leading to an overall process with carbon negative emissions (Bird et al., 2011). 3.5. Gas characterization The composition of gases obtained from WSG and BSG pyrolyzed at 460 and 520 °C was investigated ‘‘see Supplementary Table a in the online version of this article’’. Carbon dioxide is the main gas products accounting for more than 90 mol% (SD < 0.3). As small amount of methane, ethane and hydrogen were detected. The gas yield increased with temperature and the compounds distribution was changed. The carbon dioxide yield decreased from 92.1 at 460 °C to 87.5 mol% at 520 °C and from 94.8 to 92.8 mol% for BSG and WSG, respectively. The lighter hydrocarbons increased in yields according with the increased temperature, where methane raised from 2 to 5 mol% (SD < 0.02), ethane from 0.2 to 0.5 mol% (SD < 0.02) and hydrocarbons C3 to C4 from 0.1 to 0.3 mol%, for BSG. Similar trend was found for the WSG with methane, ethane and hydrocarbons C3 to C4 increasing from 1.5 to 2 mol% (SD < 0.02), 0.2 to 0.5 mol% (SD < 0.02) and 0.1 to 0.2 mol% (SD < 0.02), respectively. There were a wide number of condensable compounds ranging from C2 to C7 that remained in the gas phase, indicating that the tar trap used in this work was not able to condense all the condensable fractions. Compounds such as acetaldehyde, pentane, furan-2-methyl, hexane, hexene, heptene and toluene are liquid at temperature below zero and this was considered during the energy calculation performed in Section 3.6. The high presence of carbon dioxide in the gas phase suggests that decarboxylation reactions were involved during the pyrolysis process, together with dehydration and thermo-cracking reactions. 3.6. Sensitivity analysis and mass and energy balance A sensitivity analysis of the energy used in the pyrolysis process in function of the spent grains moisture content is presented in Table 5 where five different scenarios were considered to assess

the possible mass and energy balances and the general efficiency of the conversion process of 1 kg of spent grains. The spent grains present an initial moisture content of 70–75 wt% and it can be decreased by mechanical belt pressing to 50–60 wt% and further reduced to less than 10% by direct rotary-drum dryers, a procedure that is energy consuming (S.I. Mussatto et al., 2006). The energy to reduce the spent grains to a particle size ranging from 0.2 to 0.6 mm has been establish to be 0.07 MJ/kg (Mani et al., 2004). The energy needed to dry the wet spent grains has been assessed to be 2.8 MJ/kg H2O considering the heat of vaporization of water (2.3 MJ/kg), the specific heat of water at 25 °C (4.2 J/g/°K) and steam at 100 °C (4.2 J/g/°K), the sensible heat for heating water from 20 to 100 °C and the sensible heat for heating the fraction of water not removed by the drying step (10%) from 100 to 520 °C. The energy needed to heat the pyrolysis unit including radiation and exhaust gas losses was based on dry weight and was calculated considering 2.5 MJ/kg of bio-oil produced (Mohan et al., 2006) and ranging from 0.5 to 1.3 MJ. Finally, the energy to produce the ice required for the tar-trap was assessed in 0.36 MJ/ kg (DOE, 2011). The efficiency of the pyrolysis process decreased constantly as expected with the increase in moisture of the feedstock. Moisture content of the starting feedstock represents the most important parameter affecting the energy balance of the pyrolysis process, where the efficiency of the process decreases from 67% in absence of moisture to 55.6% and 41.6% in presence of 40 wt% and 60 wt% of moisture, respectively. Practically, moisture content about 40 wt% has to be expected in the case of spent grains thermo-conversion and this has been considered for the energy and mass flow calculations. The 40% moisture considered in the energy and mass balance can be achieved by the use of membrane filter presses or pressure filters able to leave a cake with moisture as lower as 30% and 10%, respectively (El Shafey et al., 2005; Mujunder, 2007). The energy balance of a hypothetical biorefinery where 1000 kg/day of WSG are processed at 520 °C was calculated taking into account 40% moisture content. The calculations show that 600 kg/day of dry matter of the initial 1000 kg can be inject into the reactor, generating 254 kg/day of dry bio-oil, 96 kg/day of bio-char and 186 kg/day of gas. The energy contained in the gas stream is 1.4 MJ/kg and can be used to cover 1/3rd of the energy needed for the thermo-conversion needed for

Table 5 Sensitivity analysis of the energy used in the pyrolysis process at 520 °C using WSG with different moisture content based on 1 kg of feedstock. Moisture (wt%)

Ein wet (MJ)

Edrying (MJ)

Ecrashing (MJ)

Epyro + losses (MJ)

Eice tar-trap (MJ)

Ebio-oil (dry) (MJ)

Ebio-char (MJ)

Ebio-gas (MJ)

Eoutdry (MJ)

Effgen. (%)

0 10 20 40 60

18.50 16.69 14.83 11.12 7.42

0.00 0.27 0.55 1.09 1.64

0.07 0.07 0.07 0.07 0.07

1.32 1.19 1.06 0.79 0.53

0.36 0.36 0.36 0.36 0.36

9.67 8.70 7.70 5.80 3.90

4.05 3.64 3.24 2.43 1.62

0.43 0.39 0.35 0.26 0.17

12.40 10.84 9.25 6.18 3.09

67.0 64.9 62.4 55.6 41.6

10702

A. Sanna et al. / Bioresource Technology 102 (2011) 10695–10703

Table 6 Potential of spent grains for bio-oil production assuming 30% of the RO will come from bio-ethanol in the UK.

a b c

Year

ROa (%)

Bio-ethanolb (MJ)

SG (MJ)

Bio-oilc (MJ)

Bio-oil RO share (%)

2010 2012 2014 2016 2018 2020

3.5 4.0 5.0 7.5 10.0 15.0

3.9 5.0 5.6 8.4 11.2 16.8

3.1 4.0 4.5 6.7 9.0 13.4

1.3 1.6 1.8 2.7 3.6 5.4

8.9 9.5 8.8 8.8 8.8 8.8

Based on the UK primary energy consumption in 2008 of 10 EJ (DECC, 2010). As summing 30% of the RO coming from bio-ethanol. bio-oil yield and CV at 460 °C.

the pyrolysis (795 MJ). However, the gas stream might be usable to produce chemicals or might be sent to a carbon capture unit where sorbents could further reduce the CO2 emissions considering the low energy content of the gas and its high content of CO2 (>90%), Captured carbon feedstock because of its low toxicity, highly abundant and economical but only few large scale industrial processes use CO2 as a raw material. CO2 can be recycled into solvent, in the food industry, in agriculture as a photosynthesis enhancer or can be used to produce fuels, fine chemical (e.g. urea, organic carbonates such as propylene carbonates) (MacDowell et al., 2010). If considering the overall energy needed for the process (2141 MJ), including the energy needed for grinding, drying, the pyrolysis itself and to produce the ice for the tar-trap, the process would require about 77% of bio-char energy content (1880 MJ) plus all the energy coming from the gas stream (260 MJ). Therefore, under this scenario the process does not need external energy. Summarizing, the total energy put IN was 11,124 MJ/day while the total energy OUT was 5800 MJ/day in terms of bio-oil (6377 MJ including the fraction of bio-char not used to cover the energy required by the process), with a general efficiency of 56%. 3.7. Potential of spent grains from bio-ethanol refineries in the UK Bio-ethanol bio-refineries are recently developing in the UK to meet the Renewable Obligation (RO) and Renewable Transport Fuel Obligation (RTFO). Table 6 shows the bio-ethanol production (Mt) that would be needed in the near future to give a 30% contribute to the 2020 RO set in the UK until 2020, when 15% of the primary energy must be produced from renewables (DECC, 2009). About 7 and 13 Mt of spent grains could be available in 2016 and 2020, respectively. Table 6 also indicates that from 2.7 to 5.4 Mt of bio-oil could be generated from the pyrolysis of bio-ethanol by-products in 2016 and 2020, respectively. This bio-oil could contribute to meet about 9% of the Renewable Obligation or used to produce valuable chemical commodities after further upgrading. The general chemical transformation of the original BSG with chemical formula C100H154O60N7 by pyrolysis indicates that the higher oxygen removal activity is at 460 °C, with bio-oil chemical formula C100H131O42N12 (bio-oil at 520 °C: C100H135O46N14). Also, nitrogen is preferentially moved to the bio-char at 460 °C (C100H120O73N8) than 520 °C (C100H88O53N7). The chemical formulas indicate that low temperature pyrolysis using an active bed material, such as activated alumina can produce a better quality bio-oil. 3.8. Conclusions Overall, the thermo-conversion partially removes oxygen showing higher removal activity at 460 °C than 520 °C leaving a bio-oil with less aromatic compounds. The highest bio-oil yields were 53% and 49% at 520 °C for WSG and BSG, in that order. About 55% and 65% of the starting energy of the spent grains is retained in the

WSG and BSG bio-oils at 520 °C, respectively. Despite the high starting nitrogen content, the process left a bio-oil with less nitrogen of 60–68% at 460 °C compared to 74–82% at 520 °C for WSG and BSG, respectively. In contrast, a higher proportion of nitrogen remains in the bio-char at 460 °C of 18–19% indicating that the bio-char can be used as a soil amendment and carbon sequestration agent. The high O/C ratio of the bio-chars at 460 indicate that the chars produced at this temperature might have retained oxygen from the bio-oils. Despite the higher energy content retained in the bio-oils generated at 520 °C the bio-oils obtained at lower temperature (460 °C) presented better quality in terms of low oxygen, nitrogen content, lower aromatics and should be preferred for this pyrolysis process. Overall, value can be added to the spent grains opening a new bio-fuel market. Acknowledgements This study was carried out with funds provided by: Regione Autonoma della Sardegna, Italy. ‘‘Master & Back Program’’ (sponsor) and Greenolysis Limited, Mansfield, Nottinghamshire, UK. The authors are also grateful to Ignasi Salvado-Estivill, The Environmental Technology Centre (ETC), Z.Z. Faizal and Dr. Miguel Castro. Professor Katherine Smart is grateful to SABMiller for the Sponsorship of her Chair. The research reported here was supported (in part) by the Biotechnology and Biological Sciences Research Council (BBSRC) Sustainable Bioenergy Centre (BSBEC), under the programme for ‘Lignocellulosic Conversion to Ethanol’ (LACE) [Grant Ref.: BB/G01616X/1]. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.biortech.2011.08.092. References Annetts, J.E., Audsley, E., 2003. Modelling the value of a rural biorefinery—part II: analysis and implications. Agric. Syst. 76 (1), 61–76. Appleyard, D., 2009. Biomass and Biogas advances in the UK. Renew. Energy World Mag. 12. Atesß, F., Isßikdag˘, M.A., 2009. Influence of temperature and alumina catalyst on pyrolysis of corncob. Fuel 88, 1991–1997. Bardet, M., Emsley, L., Vincendon, M., 1997. Two-dimensional spin-exchange solidstate NMR studies of C-13-enriched wood. Solid State Nucl. Magn. Reson. 8 (1), 25–32. Michaël, Becidan., Øyvind, Skreiberg., Hustand Johan, E., 2007. Products distribution and gas release in pyrolysis of thermally thick biomass residues samples. J. Anal. Appl. Pyrolysis 78, 207–213. Bird, M.I., Wurster, C.M., de Paula Silva, P.H., Bass, A.M., de Nys, R., 2011. Algal biochar–production and properties. Bioresource Technol. 102, 1886–1891. Boateng, A.A., Hicks, K.B., Flores, R.A., Gustol, A., 2007. Pyrolysis of hull-enriched byproducts from the scarification of hulled barley (Hordeum vulgare L.). J. Appl. Pyrolysis 78, 95–103. Bonelli, P.R., Della Rocca, P.A., Cerrella, E.G., Cukierman, A.L., 2001. Effect of pyrolysis temperature on composition, surface properties and thermal degradation rates of Brazil Nut shells. Bioresource Technol. 76, 15–22.

A. Sanna et al. / Bioresource Technology 102 (2011) 10695–10703 Bridgwater, A.V., Meier, D., Radlein, D., 1999. An overview of fast pyrolysis of biomass. Org. Geochem. 30, 1479–1493. Chen, T., Wu, C., Liu, R., Fei, W., 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. Bioresource Technol. 102, 6178–6185. Corma, A., Garcia, H., 2003. Lewis acids: from conventional homogeneous to green homogeneous and heterogeneous catalysis. Chem. Rev. 103, 4307–4365. DECC, 2009. The UK renewable energy strategy, HM government, Crown Copyright 2009. ISBN: 9780101768627. . DECC, 2010. Energy Statistics, Reference 2010/133, http://www.decc.gov.uk/assets/ decc/Statistics/publications/trends/1083-pn10-133.pdf. Demirbas, A., 2007. The influence of temperature on the yields of compounds existing in bio-oils obtained from biomass samples via pyrolysis. Fuel Process. Technol. 88 (6), 591–597. DOE, 2011. Water-Cooled Ice Machines, Energy Efficiency and Renewable Energy, FEMP, DOE/GO-102011-3203. El Shafey, E.I., Correia, P.F.M., de Carvalho, J.M.R., 2005. Annery waste treatment: leaching, filtration and cake dewatering using a membrane filter press (a pilot plant study). Sep. Sci. Technol. 40 (11), 2297–2323. Eskicioglu, C. et al., 2010. Anaerobic digestion of whole stillage from dry-grind corn ethanol plant under mesophilic and thermophilic conditions. Bioresource Technol.. doi:10.1016/j.biortech.2010.08.061. Fernando, S., Adhikari, S., Chandrapal, C., Murali, N., 2006. Biorefineries: current status, challenges, and future direction. Energy Fuels 20, 1727–1737. Freitas, J.C.C., Bonagambac, T.J., Emmericha, F.G., 2001. Investigation of biomassand polymer-based carbon materials using 13C high-resolution solid-state NMR. Carbon 39, 535–545. Garcia Peres, M., Chaala, A., Pakdel, H., Kretschmer, D., Roy, C., 2007. Characterization of bio-oils in chemical families. Biomass Bioenergy 31, 222– 242. Gil, A.M., Neto, C.P., 1999. Solid-state NMR studies of wood and other lignocellulosic materials. Ann. Rep. NMR spectrosc. 37, 75–117. Giuntoli, J., de Jong, W., Arvelakis, S., Spliethoff, H., Verkooijen, A.H.M., 2009. Quantitative and kinetic TG-FTIR study of biomass residue pyrolysis: dry distiller’s grains with solubles (DDGS) and chicken manure. J. Anal. Appl. Pyrolysis 85, 301–312. Hahn-Hägerdal, B., Galbe, M., Gorwa-Grauslund, M., Lidén, G., Zacchi, G., 2006. Bioethanol – the fuel of tomorrow from the residues of today. Trends Biotechnol. 24 (12), 549–556. Junming, Xu, Jianchun, J., Yunjuan, S., Yanju, L., 2008. Bio-oil upgrading by means of ethyl ester production in reactive distillation to remove water and to improve storage and fuel characteristics. Biomass Bioenergy 32, 1056–1061. Kanauchi, O., Mitsuyama, K., Araki, Y., 2001. Development of a functional germinated barley foodstuff from brewer’s spent grain for the treatment of ulcerative colitis. J. Am. Soc. Brewing Chemists 59, 59–62. Mathews John, A., 2008. Carbon-negative biofuels. Energy Policy 36, 940–945.

10703

Mohan, D., Pittman Jr., C.U., Steele, P.H., 2006. Pyrolysis of wood/biomass for bio-oil: a critical review. Energy Fuels 20, 848–889. Mani, S., Tabil, L.G., Sokhansanj, S., 2004. Grinding performance and physical properties of wheat and barley straws, corn stover and switchgrass. Biomass Bioenergy 27, 339–352. MacDowell, N., Florin, N., Buchard, A., Hallett, J., Galindo, A., Jackson, G., Adjiman, C.S., Williams, C.K., Shah, N., Fennell, P., 2010. An overview of CO2 capture technologies. Energy Environ. Sci. 3, 1645–1669. Marsh, R., Steer, J., Fesenko, E., Cleary, V., Rahman, A., Griffiths, T., Williams, K., 2008. Biomass and waste co-firing in large-scale combustion systems. Energy 161 (3), 115–126. Mussatto, S.I., Dragone, G., Roberto, I.C., 2006. Brewers’ spent grains: generation, characteristics and potential applications. J. Cereal Sci. 43, 1–14. Mujunder, A.S., 2007. Handbook of Industrial Drying, third ed. CRC Press, 1280 pp.. Nowakowski, D.J., Jones, J.M., Brydson, R.M.D., Ross, A.B., 2007. Potassium catalysis in the pyrolysis behaviour of short rotation willow coppice. Fuel 86, 2389– 2402. OECD-FAO Agricultural Outlook 2009–2018, 2009. ISBN 978-92-64-01144-1. . Onay, O., Koçkar, O.M., 2004. Fixed-bed pyrolysis of rapeseed (Brassica napus L.). Biomass Bioenergy 26, 289–299. REN21 (Renewable Energy Policy Network for the 21st Century), 2009. Renewables Global status Report: 2009 Update. . Sadhukhan, J., Mustafa, M.A., Misailidis, N., Mateos-Salvador, F., Du, C., Campbell, G.M., 2008. Value analysis tool for feasibility studies of biorefineries integrated with value added production. Chem. Eng. Sci. 63, 503–519. Samolada, M.C., Papafotica, A., Vasaloa, I.A., 2000. Catalyst evaluation for catalytic biomass pyrolysis. Energy Fuels 14, 1161–1167. Sanna, A., Ogbuneke, K., Andrésen, J.M., 2009. Bio-coke from upgrading of pyrolysis bio-oil for co-firing. Fuel 88 (12), 2340–2347. Taheripour, F., Hertel, T.W., Tyner, E.W.E., Beckman, J.F., DBirut, D.K., 2010. Biofuels and their by-products: Global economic and environmental implications. Biomass Bioenergy 34 (3), 278–289. Vispute, T.P., Zhang, H., Sanna, A., Xiao, R., Huber, G.W., 2010. Renewable chemical commody feedstock from integrated catalytic processing of pyrolysis oils. Science 330 (6008), 1222–1227. White, J.S., Yohannan, B.K., Walker, G.M., 2008. Bioconversion of brewer’s spent grains to bioethanol. FEMS Yeast Res. 8, 1175–1184. Yang, H., Yan, R., Chen, H., Lee, D.O., Zheng, C., 2007. Characteristics of hemicellulose, cellulose and lignin pyrolysis. Fuel 86, 1781–1788. Yorgun, S., Emre Sßimsßek, Y., 2008. Catalytic pyrolysis of Mischantus x giganteus over activated alumina. Bioresource Technol. 99, 8095–8100. Yuan, J.-H., Xu, R.-K., Zhang, H., 2011. The forms of alkalis in the biochar produced from crop residues at different temperatures. Bioresource Technol. 102, 3488– 3497.