Bio-oil Production From Fast Pyrolysis of Waste Furniture Sawdust in a Fluidized Bed

Bio-oil production from fast pyrolysis of waste furniture sawdust in a fluidized bed Hyeon Su Heo a , Hyun Ju Park a , Young-Kwon Park a, * , Changkook Ryu b, * , Dong Jin Suh c , Young-Woong Suh c , Jin-Heong Yim d , Seung-Soo Kim e a Faculty of Environmental Engineering, University of Seoul, 90 Jeonnong-Dong, Seoul 130-743, Republic of Korea b School of Mechanical Engineering, Sungkyunkwan University, Suwon 440-734, Republic of Korea c Clean Energy Research Center, Korea Institute of Science a
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Bio-oil production from fast pyrolysis of waste furniture sawdust in a fluidized bed Hyeon Su Heo a , Hyun Ju Park a , Young-Kwon Park a, * , Changkook Ryu b, * , Dong Jin Suh c ,Young-Woong Suh c , Jin-Heong Yim d , Seung-Soo Kim e a Faculty of Environmental Engineering, University of Seoul, 90 Jeonnong-Dong, Seoul 130-743, Republic of Korea b School of Mechanical Engineering, Sungkyunkwan University, Suwon 440-734, Republic of Korea c Clean Energy Research Center, Korea Institute of Science and Technology, 39-1 Hawolgok-Dong, Seoul 136-791, Republic of Korea d Division of Advanced Materials Engineering, Kongju National University, Gongju 314-701, Republic of Korea e Department of Chemical Engineering, Kangwon National University, Samcheok 245-711, Republic of Korea a r t i c l e i n f o  Article history: Received 31 October 2008Received in revised form 30 May 2009Accepted 3 June 2009Available online 27 June 2009 Keywords: Bio-oilWaste furnitureFast pyrolysisFluidized bed a b s t r a c t TheamountofwastefurnituregeneratedinKoreawasover2.4milliontonsinthepast3years,whichcanbe used for renewable energy or fuel feedstock production. Fast pyrolysis is available for thermo-chem-ical conversion of the waste wood mostly into bio-oil. In this work, fast pyrolysis of waste furniture saw-dust was investigated under various reaction conditions (pyrolysis temperature, particle size, feed rateand flow rate of fluidizing medium) in a fluidized-bed reactor. The optimal pyrolysis temperature forincreased yields of bio-oil was 450  C. Excessively smaller or larger feed size negatively affected the pro-ductionof bio-oil. Higher flowandfeedingrates weremoreeffective fortheproductionof bio-oil, butdidnotgreatlyaffectthebio-oil yieldswithinthetestedranges. Theuseof productgasasthefluidizingmed-ium had a potential for increased bio-oil yields.   2009 Elsevier Ltd. All rights reserved. 1. Introduction The biodegradable fractions in wastes like paper, wood, andfood residue are important sources of biomass for renewable-en-ergy production through thermal or biological conversion. Whiledirect combustion is the dominant method applied to mixedwastes, specificstreams of industrial wastes withhighenergycon-tent could be converted into fuel feedstock using advanced tech-niques, such as pyrolysis and gasification. One example of suchmaterials is waste furniture, of which over 2.4 million tons weregenerated in Korea in the past 3 years (Yoo, 2008). Although thewood in waste furniture could have been treated with paint, sur-face coating, or pesticides, unlike fresh wood or forestry residues,it usually contains less moisture and is available for pyrolysisand gasification after size reduction.Fast pyrolysis is an attractive technology for biomass, fromwhich bio-oil is the preferred product having a great potentialfor useas fuel oil inindustry, or astransport fuel. Fast pyrolysisre-fers to pyrolysis at temperatures of about 500  C, with very highheating rates (>10 3  C/s) and a short vapor residence time (<2s),which can maximize the conversion of biomass into liquid (bio-oil) products. Many researchers have investigated the fast pyroly-sis of different biomass materials in different reactor systems.The yield of bio-oil is as high as 80wt.% of the biomass input(Bridgwater, 1999), and its heating value is lower, ranging from14to 18MJ/kg (Lu et al., 2009). Chiaramonti et al. (2007) reviewed the use of fast pyrolysis oil for power generation in gas turbines,diesel engines, and large-scale power plants by cofiring. Althoughthere are difficulties due to the nature of the oil, such as inhomo-geneity,highviscosity,andcorrosiveness, nomajortechnicalprob-lems have been identified, especially for cofiring at power plants.The biomass type for typical pyrolysis studies has been woodymaterials, but recent studies have reported fast pyrolysis for vari-ous agricultural wastes, such as corn, sunflower, olive, straw, andrice husk (Yanik et al., 2007; Zabaniotou et al., 2008; Zheng,2007; Tsai et al., 2007; Lu et al., 2008). Few studies, however, have reported on the fast pyrolysis of waste wood, such as post-con-sumed furniture, while many have reported on the slow pyrolysis(Helsen et al., 1998; Phan et al., 2008) and thermogravimetricanalysis (Reina et al., 1998).In the present study, fast pyrolysis of waste furniture sawdustwas carried out in a fluidized bed to evaluate the technical feasi-bility of applying this technology to waste furniture. The keypyrolysis parameters, such as the reaction temperature, feed size,flow rate, feeding rate, and fluidizing medium, were varied, andthe product yields and the properties of the bio-oils wereinvestigated. 0960-8524/$ - see front matter    2009 Elsevier Ltd. All rights reserved.doi:10.1016/j.biortech.2009.06.003 *  Corresponding authors. Tel.: +82 2 2210 5623; fax: +82 2 2244 2245 (Y.-K.Park), tel.: +82 31 299 4841; fax: +82 31 290 5889 (C. Ryu). E-mail addresses: (Y.-K. Park), (C. Ryu). Bioresource Technology 101 (2010) S91–S96 Contents lists available at ScienceDirect Bioresource Technology journal homepage:  2. Experimental methods  2.1. Feedstock A waste furniture sawdust sample was supplied by a wastewood treatment facility in Korea. Table 1 presents the results of the ultimate and proximate analyses of the sawdust sample. Ulti-mate analysis was carried out using an automatic elemental ana-lyzer (Flash EA 1112 Series CHNS-O analyzer, CE Instrument) andproximate analysis was performed according to Korean StandardMethodsofWasteQuality.Eachtestcasewasrepeatedintriplicateand the average results were taken. The proximate analysis resultsare shown on an ash-free basis because the waste furniture saw-dust sample was contaminated with a significant amount of soiland sand. Before the experiments, the waste furniture sawdustwas dried in an oven at 110  C for 24h to reduce the amount of waterintheoilproduct.Afterthedryingprocedure,thewatercon-tent in thesawdust samplewas foundto be<1wt.%. Therefore, theresults are presented in this paper on a dry basis.  2.2. Pyrolysis apparatus The fast pyrolysis of waste furnituresawdust was carried out ina fluidized-bed reactor. Fig. 1 shows a schematic diagram of thepyrolysis apparatus that was used. The reactor was made froman SUS 306 stainless-steel pipe, and its internal diameter andheight were 80 and 300mm, respectively. The main reactor andfeed gas were heated electrically. The temperature of the experi-mental system was adjusted using a PID temperature controllerand was monitored using a K-type thermocouple. The errors inthe average reaction temperature were within ±5  C. An electricheating tape that was capable of heating up to 450  C was usedto avoid vapor condensation in the product gas stream tube. Thecondensable phase (bio-oil) of the pyrolysis vapor was collectedusing a series of glass condensers that maintained a low tempera-ture of    25  C, using a circulator (RW-2025G, JEIO TECH), withethyl alcohol used as the cooling solvent. The noncondensable va-pors leaving the condensation system were passed through anelectrostatic precipitator before extraction, to recover the tar. Thevapors were also sampled into samplingbags every20min to ana-lyze their composition.  2.3. Pyrolysis conditions In the fluidized-bed reactor, 1000g of Emery [Al 2 O 3 , (NANKOABRASIVES, Japan)] with a mean particle size of 40 l m was usedas the bed material. The experiments were carried out with thegas velocities above the minimum-bubbling fluidizing velocity( U  mb ), whichwas estimatedto be 0.6cm/s usingGeldart and Abra-hamsen’s equation (Abrahamsen and Geldart, 1980). The reactorsystem was purged with inert nitrogen gas, which was also usedas the fluidizing medium, for approximately 3h before the exper-iments were started. To decrease the heat loss during the experi-ments, the fluidizing gas was preheated to 350  C before beingintroducedtothereactor. Oncethereactorreachedthetargettem-perature, the screw feeder was turned on to continuously feedsawdust into the bed, with fixed mass flow rates ranging from1.5 to 2.5g/min. The total amount of sawdust fed in each test casewas approximately 150g. Each test case lasted for about 75–100min.Table 2 lists the different pyrolysis conditions for each experi-ment. The particle sizes were the average taken from the meshsizes of the sieves that were used; 0.3mm (0.2–0.425mm),0.7mm(0.425–1.0mm),and1.3mm(1.0–1.6mm).Runs1–4wereconducted to investigate the effect of the pyrolysis temperature,which ranged from 400 to 550  C. The fluidizing gas (N 2  at 5l/min), feed rate (2.5g/min), and sawdust particle size (0.7mm)were fixed during the tests. In Runs 5 and 6, the particle size wasvaried to 0.3 and 1.3mm, respectively, while the temperaturewas fixed at 450  C. The gas flow rate was lowered to 3 and 4l/min in Runs 7 and 8, respectively, mainly to increase the residencetime of the pyrolysis vapors. The feed rate was decreased to 1.5g/min in Run 9. In Run 10, the product gas consisting mainly of CO 2 and CO at the downstream of the electrostatic precipitator wasrecirculated into the bed as fluidizing gas, completely replacingthe nitrogen gas. Each test case was repeated in triplicate, andthe variations in the product yields were less than 3wt.% whichwas considered satisfactory for this study. Therefore, the averagevalues were taken for the product yields.  2.4. Product analysis Aftereachtest, thecharwasseparatedfromthebedandthecy-clonetodetermineitsyield.Thepyrolysisgasesatthedownstreamof the electrostatic precipitator were analyzed using a GC-TCDandGC-FID(ACME6000,YoungLinInstrumentsCo.,Ltd.),withtheCar-boxen 1000 (15ft  1/8in.) and HP-plot Al 2 O 3 /KCl columns,respectively. Table 3 lists the analysis methods for the GC. Usingthe accumulated flowrate andthe composition of the gas, its massyield was determined.The bio-oil was collected from the condensers and the electro-staticprecipitator. Sincetheviscousorganicfractionintheelectro-static precipitator was very difficult to fully separate, the oil yieldwas calculated by difference, using the char and gas yields. Thebio-oilthatwasproducedwasinaheterogeneousstate,withaque-ous and heavy organic fractions. Both the chemical and physicalanalyses of the bio-oil, however, were required to be in the homo-geneous state. To acquire representative samples, two bio-oil  Table 1 Ultimate and proximate analyses of the waste furniture sawdust. Ultimate analysis a Proximate analysis  a Component Content (wt.%) Component Content (wt.%)C 49.1 Moisture 9.1H 6.2N 3.0 Combustibles 91.9O b 41.7S – Ash – a On ash-free basis. b Calculated by difference. Fig. 1.  Schematic diagram of fast pyrolysis apparatus.S92  H.S. Heo et al./Bioresource Technology 101 (2010) S91–S96   samples were taken from the 10–20 vol.% parts of the top andbottom sections after sufficient stirring, respectively. Subsequentanalytical work was done in triplicate for each sample, and themean values of the individual results were taken. Since accuratequantitative analyses of bio-oil were difficult to conduct usingthe existing gas chromatography equipment, the area% of theGC–MS chromatogram was considered a good approximation be-cause it indicates the amount of the various chemical compoundsin the bio-oil (Samolada et al., 2000; Zheng, 2007). In this study,both quantitative and qualitative analyses of the bio-oil wereperformed using GC–MS (HP 5973 inert) with an HP-5MS(30m  0.25mm  0.25 l m) capillary column and with heliumas the carrier gas. Table 3 also presents the analysis methods forthe GC–MS. The mass spectra that were obtained by GC–MS wereinterpretedthroughanautomaticlibrarysearch.Thewatercontentof the bio-oil was measured using the ASTM E 203 method. A KarlFischer titrator (Metrohm 787 KF Titrino) was used, and HYDRAN-AL Composite 5K (Riedel-de Haen) and HYDRANAL Working Med-ium K (Riedel-de Haen) were used as the titration reagent andtitration solvent, respectively. The errors of the above analyseswere less than 1%. 3. Results and discussion  3.1. Effects of the pyrolysis conditions on the product distribution Fig. 2 shows the product distribution as a function of the pyro-lysis temperature for Runs 1–4. As the pyrolysis progressed whenthe temperature was increased, the char yield decreased fromabout 35.8% at 400  C to 21.3% at 550  C, releasing more pyrolysisvapors. The bio-oil yield, which was the condensable phase of thepyrolysis vapors, maximized to 58.1% at 450  C and decreased atthe higher temperatures. This was due to the secondary reactionsoftheheavy-molecular-weightcompoundsinthepyrolysisvapors,which is known to become active at temperatures over 500  C(Evans and Milne, 1987). The temperature at the peak of the bio-oil yield (450  C) was slightly lower than the typical ranges of 500–520  Cfor woodbiomass(Scott andPiskorz,1984;Bridgwater,1999). The maximum bio-oil yield of 57.0% was comparable toother studies, suchas 67% for waste wood witha moisture contentof 7.5% and a residence time of about 2.5s (Horne and Williams,1996, Fuel). Note that the moisture content of the sawdust in thisstudy was less than 1%. Tests on maple, which has a much shorterresidence time (0.5s), reported over 80% of the oil yield at 500  C(Scott et al., 1988).Fig. 3 presents the gas composition at different pyrolysis tem-peratures. CO 2  was the dominant gas product at 400  C, but thesecondary reactions of the pyrolysis vapor as well as the furtherprogress of the pyrolysis in the solid led to a rapid increase inthe gas yield, especially for the CO and light hydrocarbons, at tem-peraturesover500  C.Thegascompositionat450  Cwas28.0%CO,62.3% CO 2 , and 9.7% light hydrocarbons (C 1 –C 4 ).When testing the effect of the gas flow rate on the product dis-tribution, the bio-oil yield decreased from 57.0% at 5l/min (Run 2)to 53.0% at 3l/min (Run 7) while the char yield remained unaf-fected.Theresidencetimeofthepyrolysisvaporsisanotherimpor-tant factor that determines the time available for vapor phasereactions, affecting the gas and bio-oil yields. At lower gas flow  Table 2 Pyrolysis conditions. Run 1 Run 2 Run 3 Run 4 Run 5 Run 6 Run 7 Run 8 Run 9 Run 10Reaction temperature (  C) 400 450 500 550 450 450 450 450 450 450Fluidizing medium a N N N N N N N N N PFeed size (mm) 0.7 0.7 0.7 0.7 0.3 1.3 0.7 0.7 0.7 0.7Gas flow rate (L/min) 5 5 5 5 5 5 3 4 5 5Feed rate (g/min) 2.5 2.5 2.5 2.5 2.5 2.5 2.5 2.5 1.5 2.5Input (g) 150 150 150 150 150 150 150 150 150 150 a N: nitrogen, P: product gas.  Table 3 Analysis conditions of GC–MS, GC-TCD and GC-FID. Analysis Test conditionsGC–MS Column type: capillary column (HP-5MS, 5% phenyl-methylpolysiloxane), 30  0.25mm IDInjector temperature: 280  COven temperature: 40  C held for 5min – 40–295  C at 5  C/10min – 295  C held for 20minGC-TCD Column: Carboxen 1000, 15ft  1/8in.Injector and detector temperature: 150  COven temperature: 150  C held for 20min – 150–225  C at 20  C/min – 225  C held for 10minGC-FID Column: HP-plot Al 2 O 3 /KCL, 50m  0.322mm  8.0 l mInjector temperature: 200  C, detector temperature: 250  COven temperature: 40  C for 4min – 40–160  C at 4  C/min – 160–200  C at 2  C/min – 200  C held for 20min Temperature ( 0 C) 350 400 450 500 550 600    Y   i  e   l   d   (  w   t   %   ) 020406080 Char OilGas Fig. 2.  Product distribution as a function of pyrolysis temperature: at a feed size of 0.7mm, flow rate of 5L/min, feeding rate of 2.5g/min and in a nitrogenatmosphere. H.S. Heo et al./Bioresource Technology 101 (2010) S91–S96   S93  rates, the residence time of the vapors in the reactor increases,whichsupportsthe possibility of secondaryreactions suchas ther-mal cracking, repolymerization, and recondensation, leading to adecrease in the bio-oil yield. In the tests, the vapors consist of flu-idizing gas (N 2 ) and pyrolysis vapors. At the sawdust feeding rateof 2.5g/min, the release rate of the condensable/noncondensablepyrolysis vapors became about 0.03g/s, which is much lower thanthat of nitrogen (0.057–0.094g/s). Therefore, the influence of thenitrogen flow rate is dominant for the vapor residence time. Notethat the pyrolysis vapors contain very heavy molecules, and thatthecorrespondingaveragedensitywouldthusbemuchlargerthanthat of nitrogen. At 450  C, the vapor residence time was expectedto increase fromabout 4.3 to 6.6s when the nitrogen flowrate de-creased from 5 to 3l/min.Particle size is another parameter that affects product distribu-tion. When the particle size of sawdust was increased to 1.3mm,the bio-oil yield decreased by about 5% to 53% while the char yieldincreased by 3%. This is related mainly to the heating rate of a par-ticle. A larger particle heats up more slowly and therefore releasesless volatile matter, with more char produced. With a particle sizeof 0.3mm, both the bio-oil yield and the char yield decreased byabout3%whilethegas yieldbecame20.5%from13.3%at aparticlesize of 0.7mm. This can be attributed to the overheating of thesmaller particles, followed by the conversion of the vapors intogas (Islam et al., 1999).Intheproductyieldsattwodifferentfuelfeedrates(Runs2and9), the feed rate of 1.5g/min (Run 9) yielded about 3wt.% less bio-oil while the char yield remained unaffected. This can be inter-preted as the effect of the residence time. The reduced volume of thepyrolysisvaporsduetotheloweredfeedrateincreasedtheres-idence time of the vapors in the reactor, which led to more vaporphase reactions of the condensable compounds yielding lightergas molecules.Fig.4comparestheproductdistributionasafunctionoftheflu-idizingmedium. Interestingly, thebio-oil yieldincreasedtoa max-imum of 66wt.% when the product gas that evolved duringpyrolysis was utilized as the fluidizing medium. The product gasconsisted of 29.2 wt.% CO, 60.7 wt.% CO 2 , and 10.1 wt.% hydrocar-bons (C 1 –C 4 ), which was similar to those for Run 2 at the sametemperature on a nitrogen-free basis. At this point, the influenceof the composition of the fluidizing medium on the pyrolysisbehavior is not fully understood. When using the product gas asa fluidizing agent, however, a reaction between the componentsof the product gas and the pyrolysis vapor in the reactor could oc-cur. There are some reports that using product gas as a fluidizingagent also resulted in higher bio-oil yields (Park et al., 2008,2009; Jung et al., 2008). In Scott and Piskorz’s study (1984), in- creased oil yield (by about 5%) were noticed in the product yieldsof poplar in pilot tests using product gas recycle, compared tothe results of bench scale tests using nitrogen as fluidizing med-ium. However, it is hard to conclude that the increased oil yieldis related to the product gas recycle due to the differences in theparticle sizes tested and process configuration between the benchscale and pilot tests. Further investigations are required on the ef-fect of product gas recycle.  3.2. Characteristics of bio-oil The pyrolysis temperature was the most important parameteraffecting the characteristics of the bio-oil. The water content wassignificantly affected by the pyrolysis temperature, as shown inTable 4. The obtained water content in the bio-oils varied from40 to 60wt.% as the temperature increased, while the yield of water varied from 21 to 28wt.% of the original sawdust. Sincewaste furniture has various additives, such as adhesive, coatingmaterial,anddyes, theresultingbio-oil compositionmaybesome-what different from that of fresh wood. For the pyrolysis liquidsobtained from three different sewage sludge samples, each watercontent (27.2, 44.5, and 46.6wt.%) depended on the organic com-position of the sewage sludge samples (Fonts et al., 2009). More-over, for the bio-oil produced from Quercus Acutissima, thewater contents were within the range of 31.6–62.4% (Lee et al.,2008), which was similar to the results obtained in this study.Luo et al. (2004) also reported that water content of 53.5wt.%was obtained for rice straw. Therefore, the high water content inthis study may have come from the various organic additives of waste furniture. The water content of the bio-oil that was pro-duced at 450  C, with changes in the other operating parameters(Runs 5–10), remained almost constant at 43–46wt.%.Fig. 5 shows the four main categories of the major oil com-pounds identified for Runs 1–4 (temperature: 400–550  C) andRun 10, based on the area of the GC–MS chromatograms. There is Temperature ( o C) 350 400 450 500 550 600    G  a  s   C  o  m  p  o  s   i   t   i  o  n   (   %   d  r  y   ) 020406080100 CO 2 CO C1-C4 gases Fig. 3.  Product gas composition for different pyrolysis temperatures (Runs 1–4). Fluidizing medium Nitrogen Product gas    Y   i  e   l   d   (  w   t   %   ) 020406080 Char OilGas Fig. 4.  Product distribution as a function of fluidizing medium: at a pyrolysistemperature of 450  C, a feed size of 0.7mm, feeding rate of 2.5g/min, flow rate of 5L/min.  Table 4 Water content in bio-oils obtained under different pyrolysis temperatures. Pyrolysis temperature (  C) 400 450 500 550Water content in bio-oil (wt.%) 40.2 45.2 51.0 60.0Yield of water in product (wt.%) 20.9 26.3 27.9 26.2S94  H.S. Heo et al./Bioresource Technology 101 (2010) S91–S96 
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks