Production of Liquid Feedstock from Biomass via ... - ACS Publications

Fast Pyrolysis of Stem Wood in a Pilot-Scale Cyclone Reactor. Henrik Wiinikka ... Jasmine Hertzog , Vincent Carré , Anthony Dufour , Frédéric Aubri...
0 downloads 0 Views 610KB Size
Article pubs.acs.org/EF

Production of Liquid Feedstock from Biomass via Steam Pyrolysis in a Fluidized Bed Reactor Efthymios Kantarelis,* Weihong Yang, and Wlodzimierz Blasiak Royal Institute of Technology (KTH), School of Industrial Engineering and Management, Division of Energy and Furnace Technology, Brinellvägen 23, 100 44, Stockholm, Sweden S Supporting Information *

ABSTRACT: The nature of liquids derived from biomass fast pyrolysis is far from typical oil, and thus, different approaches for bio-oil production and upgrading are needed. In this paper the steam pyrolysis of a pine and spruce wood mixture in a bubbling fluidized bed is investigated. Particularly, the effect of steam to biomass ratio and temperature in relation to products yields and composition has been studied. Products analyses indicate that steam presence affects the yields and composition of all the products (gas, char, liquid) and promotes oxygen removal from the liquid. Increased liquid yields with significantly lower amount of carboxylic acids and higher effective hydrogen index (EHI) were obtained, which makes them more suitable for further upgrading. The levoglucosan (LGA) concentration in the produced liquid is higher compared with conventional N2 pyrolysis, which suggests that steam pyrolylsis can be regarded as an alternative for production of fermentable sugars. Polycondensation reactions are hindered by steam presence while steam seems to act as a hydrogen donor; however, increased water content is a problem that has to be considered as well.



INTRODUCTION The depletion of crude oil reserves, the political instability of oil producing countries, and the environmental burden due to fossil fuels use have incentivized the quest for alternative routes for production of liquid feedstock that can be used for fuels and chemicals production. Biomass is the only carbon containing renewable source, which makes it suitable for fuels and other chemicals production.1 Especially the abundant and nonedible lignocellulosic feedstock has attracted significant interest in recent years. The transformation of biomass to fuels and chemicals is of utmost importance because it is the only viable renewable source for the production of aromatic compounds.2 A promising and rapidly advancing technology for producing a plethora of complex compounds (which otherwise could not be readily available via conventional routes) from biomass is fast pyrolysis, producing a mobile liquid feedstock known as bio-oil. Bio-oil is a complex multicomponent mixture of molecules derived from depolymerization and fragmentation of biomass building blocks, namely, cellulose, hemicellulose, and lignin. It can be classified into the following generalized categories: hydroxyaldehydes, hydroxyketones, sugars and dehydrosugars, carboxylic acids, and phenolic compounds.3 However, unlike oil, its high content of oxygenated compounds makes it difficult to be processed in traditional petroleum refineries. These significant amounts of oxygen are related to undesirable properties, such as high viscosity, corrosiveness, and instability.4 These properties can be quite problematic with respect to equipment, operations, catalysts, and the overall quality of the final product. Hence, even though that bio-oil can be potentially used as a renewable feedstock for the production of transportation fuels and/or chemicals, there are still challenges for direct drop-in into standard refineries and/or petrochemical complexes. The © 2013 American Chemical Society

main challenge is to remove, or reduce, in an efficient manner the oxygen content of the bio-oil and convert it to a more suitable feedstock for exploitation in oil refineries. Contrary to pyrolysis in inert gas atmosphere, there have been a handful of studies investigating the product distribution and composition in the presence of other gases such as CO, CO2, CH4, and H2 as well as steam. Zhang et al.5 have studied the effect of different atmospheres on the yield and composition of liquids. In their study, it is reported that different atmospheres strongly affect product yields and composition, with CO and CO2 atmospheres increasing the acid and ketones content of the bio-oil with parallel decrease of the methoxy containing compounds (which are regarded as polymerization precursors) in favor of monofunctional phenols. The pyrolysis of biomass in the presence of steam is a promising method for the simultaneous production of liquid fuels and/or chemicals as well as activated carbon.6 Steam pyrolysis has been investigated in a number of studies, mainly in fixed bed reactors, also for activated carbon and biochar production.7−9 However, the potential of fixed bed technology to give high liquid yields, especially at commercial scale, is limited.10 Sagehashi et al.11 have studied the pyrolysis of Japanese cedar and its building blocks (cellulose, hemicellulose, and lignin) in the presence of superheated steam at a temperature range of 150−400 °C as an alternative route for carbonization of biomass.11 Biochar is another useful product and can be permanently sequestered in soil because its carbon is highly stable in soil environments and may be sequestered for thousands of years.12 Received: April 2, 2013 Revised: June 25, 2013 Published: June 25, 2013 4748

dx.doi.org/10.1021/ef400580x | Energy Fuels 2013, 27, 4748−4759

Energy & Fuels

Article

Figure 1. Experimental facility. H-401 steam boiler; H-402 and H-403 gas preheaters; R-201 fluidized bed reactor; V-101 biomass hopper; SF-101 metering screw feeder; SF-102 injection feeder; FG-201 cyclone; V-301 Venturi scrubber; P-301 pump; E-301 heat exchanger. A high efficiency cyclone with a nominal cut point of ∼2.5 μm is used for char-vapor separation, while for vapors quenching and liquid collection a venturi scrubber is used. The condensed liquid is recycled and injected through the venturi at the top of the scrubber, acting as cooling medium. During recirculation, condensed liquid is further cooled by means of a water cooled shell and tube heat exchanger. The gases exit the quenching unit at temperatures less than 30 °C. N2 gas is injected to the injection feeder during the experimental runs in order to maintain a slight overpressure and to prevent steam from entering the biomass feeding line. N2 is also used as a tracer gas for determination of the gas yield under steady state conditions. The biomass feeding rate varied from 1.5 to 2.65 kg h−1 while the steam feeding rate was set at 1 kg h−1 for S/B cases of 0.5 and 0.67 and at 0.91 kg h−1 for the S/B = 0.34 case. Those flow rates correspond to nominal gas velocities of 0.24 and 0.22 m s−1. For the nitrogen case the biomass feeding rate was set at 2 kg h−1 and a nitrogen flow of 20.8 Nl min−1, which corresponds to a nominal velocity of ∼0.24 m s−1. For each of the experimental runs, the total operation time was 5 h. Gas Analysis. For gas analysis and process monitoring, a flue gas analyzer (O2, CO, CO2) and a microGC (Agilent 490 microGC quad) were used. The flue gas analyzer consists of an online Non-Dispersive Infrared (NDIR) Maihak MULTOR610 gas analyzer for CO and CO2. O2 concentration was measured using an M&C PMA 25 analyzer equipped with a paramagnetic detector. Online measurement of gas composition ensured oxygen free atmosphere, monitoring of process stability, and validation for the results obtained from the microGC. The microGC is equipped with four columns and thermal conductivity detectors able to detect H2, He, O2, N2, CH4, CO, CO2, H2S, and hydrocarbons up to C6’s. The first channel is equipped with a Molsieve 5 Å column for detection of permanent gases (H2, He, O2, N2, CH4, CO). Channel 2 is equipped with a CP PoraPlotU column, used for separation and analysis of CO2, H2S, and C2 hydrocarbons; the third channel, with an Al2O3 column, analyzes the C3 and C4 hydrocarbons, while a CP-Sil 5 CB column is used at channel 4 for detection of higher hydrocarbons. Ar (channels 1&2) and He

Several studies indicate that the presence of steam atmosphere during biomass pyrolysis influences the products distribution, increasing the yield of the liquids.6−13 Similar results have been reported for other feedstocks, such as coal, oil shale, and lignite.18−21 Characterization of the obtained liquids indicates that liquids exhibit lower oxygen content as compared with liquids obtained from pyrolysis in inert atmosphere and can be utilized as transport fuels and/or for chemicals production.6,13,22,23 In spite of the substantial quantity of studies directed to steam pyrolysis of biomass, few conclusions of the effects of steam during fast pyrolysis can be drawn. To the authors knowledge, no study on biomass fast pyrolysis behavior in the presence of steam has been reported. Study of the steam pyrolysis in a fluidized bed reactor can also provide insights about the initial stages of steam gasification. In this study the effect of steam presence and temperature on the products yield and composition are investigated and discussed.



EXPERIMENTAL SECTION

Experimental Setup. The experimental setup (Figure 1) consists of a SS316L fluidized bed reactor. The height and diameter of the reactor are 95 and 7.2 cm, respectively, with a perforated plate serving as gas distributor (a photograph of the facility can be found as Supporting Information). The gas distributor has holes of a diameter of 1 mm in a triangular arrangement with a pitch of 7.1 mm. The heat supply to the reactor is achieved by gas preheating and by heating elements that surround the reactor body up to a height of 60 cm. Six K-type thermocouples are used along the reactor to monitor the temperature while fluidization is monitored by continuous pressure recording. Biomass is fed using a two stages feeder, with the first stage being the dosing feeder and the second a high revolution speed (900 rpm) injection feeder. 4749

dx.doi.org/10.1021/ef400580x | Energy Fuels 2013, 27, 4748−4759

Energy & Fuels

Article

(channels 3&4) were used as carrier gases. Samples of produced gases were analyzed every 3 min. All the experiments were performed twice in order to verify the observed trends and process stability. In all the cases the produced gas composition was within the standard deviation observed during the initial runs, and thus, the results are considered reproducible. Liquid Analysis. Liquid products were analyzed at Johann Heinrich von Thünen-Institut (vTI), Germany. The water content of the liquids was determined by Karl Fisher titration and the elemental composition using an ICP elemental analyzer. Obtained liquids were characterized by means of GC/FID-MS (for the GC detectable compounds), using an HP 6890 gas chromatograph with a microflow splitter and an HP 5972 mass selective detector under constant He flow. The GC was equipped with a Varian DB 1701(60 m × 0.25 mm) column. Sample was injected using a CTC Analytics (Combi Pal) autosampler. The split ratio was set at 15:1 while the injector and FID temperatures were 250 and 280 °C, respectively. The ion source was at 140 °C, and the ionization energy of the mass detector was 70 eV. Initially the oven temperature was held constant at 45 °C for 4 min followed by a 3 °C/min increase up to 280 °C, where it was held constant for 20 min. Acquired data was evaluated using the MassFinder 4 software. The nonvolatile fraction of the obtained liquids was characterized by means of size exclusion chromatography (SEC). Prior to SEC, the sample was freeze-dried. Using this method only the nonvolatile high molecular residue remains. A Varian Polargel-L column and a UV detector at 254 nm were used. The exclusion limit was 35000 Da using DMSO with 0.1% LiBr as solvent at a flow rate of 1 mL/min. Calibration was performed using PEG standards. Materials. A mixture of pine and spruce sawdust supplied by SCA BioNorr was used with its particle size ranging between 1 and 1.4 mm. Its proximate and ultimate analyses are shown in Table 1.

Table 2. Silica Sand Chemical Analysis (% wt) SiO2 Al2O3 Fe2O3 loss on ignition sintering point

The X-ray diffraction pattern of silica sand, as obtained by a Bruker D2 Phaser X-ray diffractometer using Cu Kα radiation (λ = 1.54 Å), is shown in Figure 2 .

Figure 2. XRD pattern of silica sand.



Table 1. Proximate and Ultimate Analyses of Biomass

RESULTS AND DISCUSSION All the experiments were conducted at Royal’s Institute of Technology (KTH)Division of Energy and Furnace Technologylaboratories in Stockholm, Sweden. In all the experimental runs the residence time of the vapors was similar and did not exceed 3 s until quenching. Effect of Steam on Products Yields and Composition. The effect of steam was studied at 500 °C, which is a typical temperature for maximum liquid production.4 Three different steam to biomass (as received) weight ratios (S/B) were used, namely 0.34, 0.5, and 0.67, and compared with conventional N2 pyrolysis (S/B = 0). The corresponding molar steam to carbon ratios (S/C) (including moisture from the biomass feed) are 0.64, 0.87, and 1.11 respectively. As shown in Figure 3 steam presence affects the product yields, indicating different decomposition mechanisms as compared with N2 pyrolysis. Char yield is slightly and monotonically decreased with increase of S/B. However, for gas and liquid yields different behavior is observed. Here it should be noted that organic liquid refers to the liquid organic matter on a dry basis. Decreased char formation can be partly attributed to enhanced heating rates due to steam presence; however, this effect should not be overrated since heat and mass transfer rates in fluidized bed systems are already high. A maximum organic liquid yield of 41.6% wt, on a dry basis, was obtained at S/B equal to 0.5 with an increasing trend from 0 (N2 pyrolysis) to 0.5. Further increase in S/B seems to favor gas production. The almost constant gas yield from S/B 0 to 0.5 and the char reduction disproportional to the liquid increase indicate steam−vapor interactions.

Proximate Analysis (% wt) moisture volatile matterdb ashdb fixed carbondb HHV (MJ/kg)a Ultimate Analysis (% wt db) C H Ob N S

9.80 83.00 0.31 16.60 20.46

50.70 6.10 42.71 0.18