Energy 162 (2018) 35e44 Contents lists available at ScienceDirect Energy journal homepage: www.elsevier.com/locate/energy The impact of bed material cycle rate on in-situ CO2 removal for sorption enhanced reforming of different fuel types Josef Fuchs a, *, Johannes Christian Schmid a, Florian Benedikt a, Stefan Müller a, Hermann Hofbauer a, Hugo Stocker b, Nina Kieberger c, Thomas Bürgler c a b c TU Wien, Institute of Chemical, Environmental and Bioscience Engineering (ICEBE), Getreidemarkt 9/166, 1060 Wien, Austria Voestalpine Stahl Donawitz GmbH, Kerpelystraße 199, 8700 Leoben, Austria Voestalpine Stahl GmbH, Voestalpine-Straße 3, 4020 Linz, Austria a r t i c l e i n f o a b s t r a c t Article history: Received 7 December 2017 Received in revised form 12 June 2018 Accepted 29 July 2018 Available online 30 July 2018 A dual ﬂuidized bed reactor system produces a nitrogen-free product gas by using steam as gasiﬁcation agent. Additionally, the usage of limestone as bed material allows for the in-situ removal of carbon dioxide out of the product gas. Hence, a hydrogen-rich product gas with a high reduction potential for the steel industry can be generated. This so-called “sorption enhanced reforming” process has already been proven applicable for wood as fuel, but since the costs for biomass like wood have increased signiﬁcantly during the last years, cheaper fuels are of interest. The experimental results of three different biogenic fuel types (soft wood, rice husk and bark) and one fossil fuel type (lignite) are discussed in detail. In the past, research mainly focused on temperature dependency of the process since it seemed to be the main factor for carbon dioxide sorption in the gasiﬁcation reactor and therefore product gas composition. Within this work, it is shown that the bed material cycle rate should not be disregarded and is a key factor within the process. The presented ﬁndings allow a detailed understanding of “sorption enhanced reforming”, the inﬂuence of bed material cycle rate, and the inﬂuence of volatile matter in the fuel. © 2018 Elsevier Ltd. All rights reserved. Keywords: Dual ﬂuidized bed Sorption enhanced reforming Bed material cycle rate Volatile matter Limestone/calcium carbonate Biogenic fuels/biomass 1. Introduction The substitution of fossil fuels by renewables is one of the major challenges for reducing greenhouse gas emissions. Biomass as a renewable source releases the same amount of carbon dioxide as it aggregates during its growth. Therefore, the gasiﬁcation of biomass is a reliable technology for the reduction of greenhouse gas emissions and for gaining a valuable product gas (also called producer gas) for the production of heat, electric power, syngas and also for the usage as a reducing agent in the iron and steel industry. The utilization of cheap alternative biomass sources becomes important, since the increasing use of high-grade biomass (like wood chips) leads to increasing fuel costs as well. Due to challenging characteristics of alternative fuels (accompanying undesired * Corresponding author. E-mail addresses: email@example.com (J. Fuchs), johannes.schmid@tuwien. ac.at (J.C. Schmid), ﬂorian.firstname.lastname@example.org (F. Benedikt), stefan.mueller@ tuwien.ac.at (S. Müller), email@example.com (H. Hofbauer), hugo. firstname.lastname@example.org (H. Stocker), Nina.Kieberger@voestalpine.com (N. Kieberger), email@example.com (T. Bürgler). https://doi.org/10.1016/j.energy.2018.07.199 0360-5442/© 2018 Elsevier Ltd. All rights reserved. chemical substances, high ash content, low ash deformation temperature, high tar and dust contents in the product gas, etc.), further development of existing technologies towards biogenic residues is of great relevance. Further, for the utilization of the product gas in steel industry, a high reduction potential of the product gas is required. To meet the mentioned requirements, the dual ﬂuidized bed (DFB) steam gasiﬁcation is a suitable process for the thermochemical conversion of biomass into a nitrogen-free product gas, which mainly consists of hydrogen (H2), carbon monoxide (CO), methane (CH4), and carbon dioxide (CO2) [1e4]. A steam-blown gasiﬁcation reactor and an air-blown combustion reactor are the main parts. The combustion reactor provides the necessary heat for the overall endothermic steam gasiﬁcation via combustion of residual char from gasiﬁcation. If the residual char is not sufﬁcient for the heat balance, additional fuel can be fed to the combustion reactor as well. The produced heat is transferred into the gasiﬁcation reactor via the bed material, which is typical for ﬂuidized beds. Usually silica sand or olivine [5e7] are used for conventional gasiﬁcation applications as bed material. Besides these well-known bed materials, beneﬁcial effects like enhanced tar reforming  and the selective in-situ removal of CO2 in the gasiﬁcation reactor can 36 J. Fuchs et al. / Energy 162 (2018) 35e44 List of abbreviations Symbols m_ CO2;captured massﬂow of CO2 captured by the bed material in the gasiﬁcation reactor m_ CaO;CR/GR massﬂow of CaO transported from combustion reactor to the gasiﬁcation reactor via the upper loop seal m_ char;GR/CR massﬂow of char which leaves the gasiﬁcation reactor via the lower loop seal massﬂow of fuel fed into the gasiﬁcation reactor m_ fuel m_ steam massﬂow of steam fed to the gasiﬁcation reactor mCaO;initial initial bed material inventory as CaO x mass fraction Abbreviations add additional BeM bed material C carbon char if not designated differently: char from gasiﬁcation reactor to combustion reactor via lower loop seal CR combustion reactor be achieved by a calcium based bed material (calcium oxide (CaO)/ calcium carbonate (CaCO3)) [8e15]. The basic principle of the gasiﬁcation process with selective in-situ removal of CO2, the socalled sorption enhanced reforming (SER) process, is displayed in Fig. 1. The circulating bed material plays a crucial role for the basic operation of the DFB gasiﬁcation system as transport medium of char from the gasiﬁcation reactor into the combustion reactor and as transport medium for heat from the combustion reactor into the gasiﬁcation reactor. Furthermore, the bed material can also act as a catalyst and, therefore, contributes to improvement of the product gas quality. Another possible function of the bed material is the capture of gaseous components from the product gas: The sorption enhanced reforming (SER) process uses limestone/calcium carbonate (CaCO3) as bed material. By operating both reactors in a suitable temperature range, in-situ CO2 capture in the gasiﬁcation reactor according to Equation (1) and its release in the combustion reactor are possible (Equation (2)). The decreased CO2 concentration leads to a more intensive reaction of steam (H2O) with CO. Bed material cooling Fluegas CO2 - enriched Product gas H2 - enriched Heat CaO Fuel Add. fuel CaCO3/CaO Char Steam Air Fig. 1. Principle of gasiﬁcation with CO2 capture (SER). CxHy DFB gas GCMS GR max n.m. Nm3 PG PGY S/C S/F SER temp. TGA fuel add. fuel sum of ethene, ethane and propane dual ﬂuidized bed gasiﬁcation gas chromatograph mass spectroscopy gasiﬁcation reactor maximum not measured cubic meter at 0 C and 101.325 kPa product gas product gas yield steam to carbon ratio steam to fuel(daf) ratio sorption enhanced reforming temperature thermogravimetric analyzer fuel into gasiﬁcation reactor additional fuel into combustion reactor Subscripts th daf db fuel thermal dry and ash-free dry basis fuel into gasiﬁcation reactor Thus, the water-gas shift reaction (Equation (3)) enhances the production of H2 in the gasiﬁcation reactor. Müller et al.  showed that the SER process produces a hydrogen-based reducing agent with a higher reduction potential than the gas produced from the conventional gasiﬁcation process with olivine as bed material. CaO þ CO2 /CaCO3 DHR650 ¼ 170kJ=mol (1) CaCO3 /CaO þ CO2 DHR900 ¼ 166kJ=mol (2) CO þ H2 O 4CO2 þ H2 DHR650 ¼ 36kJ=mol (3) The suitable temperature range for gasiﬁcation and combustion during SER depends on the equilibrium partial pressure of CO2 in Equation (1). The equilibrium curve according to  is reproduced in Fig. 2, where typical operational conditions for both reactors are outlined as well. Typical temperatures in the gasiﬁcation reactor for SER are between 600 and 700 C. In this temperature range, CO2 capture is possible, whereas in the combustion reactor, the bed material is heated up above 810 C. There, a suitable residence time ensures calcination to calcium oxide (CaO). Since the correct temperature assures the driving force for CO2 capture in the gasiﬁcation reactor, it has been identiﬁed as main factor to inﬂuence the product gas composition of the process. Several authors published product gas compositions depending on gasiﬁcation temperature [8,9,17e22]. The data from Ref.  was obtained from SER of soft wood with the advanced 100 kWth test plant at TU Wien. This test plant was used as well for the generation of all data presented within this work. The aim of this work is to point out that other inﬂuencing factors besides the gasiﬁcation temperature have been neglected in past research activities. Therefore, SER experiments of four different fuel types are investigated in detail. Especially the bed material cycle rate is suspected to play a major role for the product gas composition. J. Fuchs et al. / Energy 162 (2018) 35e44 37 Partial pressure CO2 [bar] 0.4 0.35 0.3 CaCO CaCO3 3 CR 0.25 0.2 0.15 0.1 CaO + CO2 GR 0.05 CaO + CO2 0 500 600 700 800 900 1000 1100 Temperature [°C] Temperature [°C] Fig. 2. Equilibrium partial pressure of CO2 for the System CaCO3/CaO. 2. Materials and methods 2.1. Advanced 100 kWth DFB test plant TU Wien has designed and constructed an advanced DFB test plant for the gasiﬁcation of various fuels. A sketch of the plant is shown in Fig. 3. The advanced reactor design enhances the gas-solid contact by a new countercurrent-column, which is placed subsequent to the lower bubbling bed of the gasiﬁcation reactor. The geometrical modiﬁcations in this upper part of the gasiﬁcation reactor lead to an improved bed material hold-up  and enlarge the range of applicable fuels because of higher tar and char conversion rates compared to other DFB systems. Further, gravity separators with gentle separation characteristics instead of cyclones support the usage of soft bed materials such as limestone. The new system prohibits high velocities of gas and particles and minimizes attrition effects. Additionally, a bed material cooling in the upper loop seal enables the deﬁned setting of temperature differences between the gasiﬁcation and combustion reactor for SER. Further, a staged air input into the combustion reactor (primary, secondary, tertiary air in Fig. 3) allows an effective control of the bed material cycle rate without changing the total amount of air. 2.2. Experimental Three of the investigated fuels originate from renewable organic sources - soft wood, rice husk, and bark. Lignite, the fourth presented fuel, is used for reasons of comparison and to point out the fuel ﬂexibility of the system. The detailed chemical analysis of the mentioned fuels is given in Table 1 and was performed by the accredited Test Laboratory for Combustion Systems at TU Wien. All fuels were gasiﬁed with the same type of bed material (limestone, Table 2), which mainly forms CaO during the calcination process with high temperature in the combustion reactor. During the test runs, the contents of the main product gas components like H2, CO, CO2 and CH4 were analyzed: Therefore, a small amount of the produced product gas was cooled down to condense the H2O. Fig. 3. Advanced 100 kWth test plant at TU Wien. Afterwards the gas passed four impinger bottles ﬁlled with rapeseed methyl ester to wash out the tar of the product gas and protect the measurement devices. Online measurement of the main product gas components is done by Rosemount NGA2000 measurement equipment. Higher gaseous hydrocarbons (CxHy) and N2 were analyzed by a gas chromatograph (Perkin Elmer ARNEL e Clarus 500). The main product gas composition in this work is published on N2-free basis to allow a better comparison between the results. Typically, N2 in the product gas can be traced back to the ﬂushing of the measurement equipment and the fuel feeding system and is in the range of about 2 vol.-%db A large number of temperature and pressure sensors and an extensive measurement equipment for mass and volume ﬂows of the input and output streams guarantee an effective process control. A standardized arrangement of sampling equipment according to  was used to analyze the content of dust, water, char and tar in the product gas stream. Instead of using Isopropyl alcohol as a 38 J. Fuchs et al. / Energy 162 (2018) 35e44 Table 1 Ultimate analysis and fuel characteristics of the investigated fuels. Parameter Ultimate analysis Water content Ash content Carbon (C) Hydrogen (H) Nitrogen (N) Oxygen (O) Sulphur (S) Chlorine (Cl) Additional fuel characteristics Volatiles LHV (wet basis) Ash deformation temp. (A) Ash ﬂow temp. (D) Standard/Equipment Unit Soft wood Rice husk Bark Lignite DIN 51701 DIN 14775 Perkin Elmer 2400 CHN Elemental Analyzer wt.-% wt.-%db wt.-%daf wt.-%daf wt.-%daf wt.-%daf wt.-%daf wt.-%daf 7.2 0.2 50.8 5.9 0.21 43.1 0.005 0.005 5.8 15.2 51.2 6.1 0.55 42 0.071 0.106 7.7 7.0 52.3 6.0 0.34 41.3 0.053 0.053 13.0 4.2 68.4 3.9 0.88 26.3 0.397 0.052 wt.-%daf kJ/kg C C 86 17060 1330 1440 81 14740 1350 >1500 78 16480 1150 1210 54 20320 1340 >1500 Sum to 100 wt.-%daf CEN/TS 15289 CEN/TS 15289 DIN 51720 DIN 51900 T2 DIN 51730/DIN 51719 Relative measurement uncertainty: Water content 4.3%; Ash content 9.15e12.65%; Sulphur 7.5%; Chlorine 7.1%; Volatiles 0.45%; LHV 1.0%; C,H,N not available, typically around 5%. Table 2 Chemical composition of bed material particles (limestone). Parameter/name Unit Value CaCO3 MgCO3 SiO2 Al2O3 Fe2O3 Loss of mass after calcination Sauter mean diameter (dsv) wt.-% wt.-% wt.-% wt.-% wt.-% wt.-% mm 95e97 1.5e4.0 0.4e0.6 0.2e0.4 0.1e0.3 ~44 480/290 solvent for tar (as proposed in Ref. ), Toluene is used. This procedure allows the determination of the water content in the product gas via phase separation and volumetric determination. In general all online measured values like temperatures, gas composition, fuel composition and input, or steam and air input were taken over a period of about 30 min steady state operation for every operation point presented in this work. Furthermore, all measured and averaged values are validated with IPSEpro, which is a software system for calculating energy and mass balances also for overdetermined processes. This approach guarantees accuracy and reliability of the presented values, because the errors in measurement are balanced. Since presented values are validated by the balanced and closed mass and energy balance of the whole test plant, no raw measurement data is presented in the results section. Therefore, standard deviation of the raw measurement data is not provided in this work. 2.3. Performance indicating key ﬁgures The calculation of mass and energy balances for different operation points enables a validation of measurement data and the calculation of performance indicating key ﬁgures. Equations (4)e(10) show the applied deﬁnitions for the discussion of the results. Equation (4) shows the product gas yield, which indicates how much product gas can be produced per fuel fed to the gasiﬁcation reactor. The steam to fuel ratio and the steam to carbon ratio (Equations (5) and (6)) are a measure of how much steam is used for the gasiﬁcation of the fuel and respectively of the carbon fed to the gasiﬁcation reactor. Further, the bed material cycle rate in Equation (7) describes the turnover of the total bed material mass based on CaO inventory in times per hour. The CO2 load of the bed material (Equation (8)) indicates how much of the sorbent CaO in the gasiﬁcation reactor has reacted with CO2. The maximum value would be 1, which means that the bed material is fully carbonated when it leaves the gasiﬁcation reactor. Equation (9) shows the amount of carbon captured via the bed material in relation to the total amount of carbon fed into the gasiﬁcation reactor as fuel. Equation (10) is similar to the previous key ﬁgure, but provides information about the carbon which leaves the gasiﬁcation reactor as char to the combustion reactor in relation to the total amount of carbon fed into the gasiﬁcation reactor as fuel. Product gas yield . PGY ¼ V_ PG;db m_ fuel;daf Steamto fuel ratio . 4SF ¼ m_ steam þ m_ fuel xH2O;fuel m_ fuel;daf . i ½Nm3 kg (4) (5) ½kg=kg Steam to carbon ratio . 4SC ¼ m_ steam þ m_ fuel xH2O;fuel m_ fuel xC;fuel ½kg=kg (6) Bed material cycle rate dCaO ¼ m_ CaO;CR/GR mCaO;initial ½h1 i (7) CO2 load of bed material . . XCO2 ¼ m_ CO2;captured 44 m_ CaO;CR/GR 56 ½mol=mol (8) Carbon transport via bed material . . . tCO2 ¼ m_ CO2;captured 44 m_ fuel xC;fuel 12 ½mol=mol (9) Carbon transport via char . . tchar ¼ m_ char;GR/CR xC;char 44 m_ fuel xC;fuel 12 ½mol=mol (10) 3. Results and discussion 3.1. Overview of experimental test runs The results of all test runs conﬁrm that the necessary conditions for the SER process in both reactors (gasiﬁcation and combustion) J. Fuchs et al. / Energy 162 (2018) 35e44 39 Table 3 Operational parameter and results of gasiﬁcation experiments. Initial bed material as CaCO3 (CaO) Sauter mean diameter Gasiﬁcation temp. (lower GR) Max. temperature in CR Fuel to GR Additional fuel to CR Chem. energy in PG H2 CO CO2 CH4 CxHy PG yield CO2 in ﬂue gas Gravimetric tar GCMS tar S/F S/C Bed material cycle rateb Pressure gradient upper CR kg mm C C kW kW kW vol.-%dba vol.-%dba vol.-%dba vol.-%dba vol.-%dba Nm3db/kgfuel,daf vol.-%db g/Nm3db g/Nm3db kgH2O/kgfuel,daf kgH2O/kgC,fuel (molH2O/molC,fuel) h1 mbar/m Soft wood I Soft wood II Soft wood III Rice husk Bark Lignite I Lignite II Lignite III 70 (39) 480 630 840 110 4 81 69.5 8.6 5.6 14.0 2.3 0.9 28.9 n.m n.m 0.8 1.6 (1.1) 4.3 0.25 80 (45) 290 680 820 100 0 59 58.5 7.2 18.8 12.8 2.7 0.8 22.4 1.1 6.0 0.9 1.8 (1.2) 14.6 1.03 80 (45) 290 690 820 102 0 58 56.4 8.0 21.5 11.7 2.4 0.8 21.9 0.9 4.8 0.9 1.7 (1.1) 20.4 1.52 80 (45) 480 640 810 105 29 76 55.5 13.5 11.6 15.8 3.6 0.9 23.7 17.8 29.4 1.2 2.3 (1.5) 12.5 0.8 80 (45) 480 630 840 102 6 77 67.5 6.4 8.8 14.3 3.0 1.0 27.8 n.m. n.m. 1.1 2.0 (1.3) 4.0 0.31 72 (40) 480 660 920 100 0 45 75.7 4.4 9.8 8.4 1.7 0.9 25.6 n.m. n.m. 1.1 1.6 (1.1) 8.8 0.39 72 (40) 480 660 880 100 0 42 72.0 4.8 14.3 7.0 1.9 0.9 23.1 n.m. n.m. 1.6 2.3 (1.5) 12.9 0.67 72 (40) 480 650 920 100 0 42 76.1 4.3 11.1 7.0 1.5 0.9 23.7 n.m. n.m. 1.5 2.2 (1.5) 9.5 0.37 n. m.: not measured. a PG composition on N2-free basis and without minor impurities; N2 content from ﬂushing of measurement instruments for all presented test runs between 1.7 and 2.8 vol.%db. b Pure CaO in both reactors assumed were fulﬁlled (see also Table 3). In general, the gasiﬁcation reactor has to work on the left side of the equilibrium curve of the system CaCO3/CaO to capture CO2 and form CaCO3. Vice versa, the combustion reactor must be operated on the right side of the equilibrium curve to release CO2 and calcine CaCO3 to CaO. However, the most important operating parameter of the process is the gasiﬁcation temperature to guarantee an optimized CO2-sorption parallel to the thermochemical conversion of the fuel. It has been shown in previous works that these two requirements are fulﬁlled at a temperature around 650 C [18,20]. Therefore, the most important control strategy is to keep this temperature in the lower part of the gasiﬁcation reactor (close to the fuel input). Further, the temperature level in the combustion reactor is strongly dependent on the chemical composition of the fuel. For SER of lignite a signiﬁcantly higher maximum combustion temperature was achieved. The reason for this behavior is the low content of volatile matter in the fuel lignite (Table 1). This means that a high amount of char (residual from gasiﬁcation of the fuel) is transported to the combustion reactor. This assumption is supported by the temperature proﬁles of the reactors in Fig. 4. During the gasiﬁcation of lignite, the temperature proﬁle in the combustion reactor increased with its height. This results from the high amount of char, which burns along the reactor height with the ﬂuidizing air. The temperature of the bed material at the exit of the combustion reactor leads to higher temperatures in the countercurrent-column for the gasiﬁcation of lignite, which is advantageous for the reduction of the tar content in the product gas . A comparably smaller amount of char reaches the combustion reactor in case of the gasiﬁcation of bark - the char is already combusted in the lower part of the combustion reactor. The temperature decreases slightly with the height of the reactor. Additionally, a slight fuel input directly into the combustion reactor supports the calcination of bed material during the SER experiments with wood, rice husk and bark. The described differences for lignite and bark are illustrative results for fuels with low volatile matter and high volatile matter in general. Comparing the product gas composition for the SER process of the investigated fuels in Table 3 shows that for soft wood, bark and lignite, H2 contents of about 70 vol.-%db were reached. Only the gasiﬁcation of rice husk led to a signiﬁcantly lower H2 content in the product gas. Several reasons can be identiﬁed for this behavior: The comparably low maximum temperature in the combustion reactor leads to an operation close to the equilibrium curve of the chemical system CaCO3/CaO. It is assumed that complete calcination of the bed material was hindered, which may have impeded the effective CO2 sorption in the gasiﬁcation reactor. And last but not least, the bed material cycle rate was comparably high, which is unfavorable for a high CO2 sorption (this circumstance will be discussed later on in this work). Consequently, a comparatively low H2 content in the product gas was observed for SER with rice husk. The reason for the lower system temperatures is the behavior of the rice husk ash: Fluidization problems were encountered during conventional gasiﬁcation with gasiﬁcation temperatures about 750 Ce800 C and high combustion temperatures of 1050 C . To avoid limitations resulting from agglomeration effects of bed material particles, SER system temperatures were kept as low as possible. Thus, a more smooth operation regarding ﬂuid dynamics of the dual ﬂuidized bed was found for SER of rice husk. Especially, the gasiﬁcation of wood, bark and lignite led to preferable operational conditions without any limitations. Thus, only these three fuels will be considered for further investigation in this work. 3.2. Inﬂuence of bed material cycle rate To demonstrate the inﬂuence of the bed material cycle rate on the product gas composition, the two experiments Lignite II and Lignite III were carried out and the results were compared. The operating parameters like S/F ratio are very similar (see Table 3). Further, the temperature proﬁles of the reactors in Fig. 5 are very similar as well. This is not straightforward since the bed material cycle rate inﬂuences the temperature proﬁle heavily . However, to separate the inﬂuence of temperature from the inﬂuence of bed material cycle rate, a similar temperature proﬁle is indispensable. Due to the bed material cooling in the upper loop seal, which is installed at the test plant and can also be seen in the scheme of Fig. 3, an operation with similar temperature proﬁles was possible. Especially the temperature proﬁles of the gasiﬁcation reactor, which are most important, are nearly congruent. Higher temperatures in Height [m] J. Fuchs et al. / Energy 162 (2018) 35e44 Height [m] 40 Lignite I Bark Gasification temperature [°C] Lignite I Bark Combustion temperature [°C] Height [m] Height [m] Fig. 4. Temperature proﬁle of the gasiﬁcation reactor (left) and the combustion reactor (right) for SER of lignite and bark. Gasification temperature [°C] Combustion temperature [°C] Fig. 5. Temperature proﬁle of the gasiﬁcation reactor (left) and the combustion reactor (right) for SER of Lignite II and Lignite III. the upper part of the combustion reactor (Lignite III) are obvious. However, since the deviation to the equilibrium concentration of the system CaO/CaCO3 is high enough for both experiments favorable conditions for calcination can be assumed in both cases. By analyzing the fundamental ﬁgures of CO2 capture (Table 4) calculated by mass and energy balances with IPSEpro, it turns out that with a low cycle rate more CO2 is captured in total (8 kgCO2/h vs. 6.9 kgCO2/h). Hence, a lower CO2 content can be found in the product gas. Due to the water-gas shift reaction (Equation (3)) and Le Chatelier’s principle, a higher H2 content can be obtained. To get a more detailed insight into the process and its dependency on bed material cycle rate additional experiments needed to be analyzed. All additional data were produced with soft wood as fuel. J. Fuchs et al. / Energy 162 (2018) 35e44 25 Table 4 Comparison of CO2 capture performance of experiments Lignite II and Lignite III. Unit Lignite II Lignite III Bed material cycle rate h1 high (12.9) low (9.5) H2 in PG CO2 in PG Total CO2 capture CO2 load of BeM. vol.-%db vol.-%db kgCO2/h molCO2/molCaO 72.0 14.3 6.9 0.017 76.1 11.1 8 0.027 Since the gasiﬁcation temperature is strongly dependent on the bed material cycle rate, a temperature increase with the bed material cycle rate could not be avoided (Fig. 6), even though the installed bed material cooling was used. However, to demonstrate the scale of inﬂuence of gasiﬁcation temperature versus bed material cycle rate on the product gas composition, the Kendall rank correlation is used. The Kendall rank t is equal to 1 (or 1) if datasets are correlated and is 0 for non-correlated datasets. The correlation between gasiﬁcation temperature and CO2 content in product gas versus bed material cycle rate and CO2 content in the product gas is investigated in Fig. 7. The evaluation indicates that the correlation with bed material cycle rate is higher than with gasiﬁcation temperature. Moreover, a quite high p-value for the correlation with the gasiﬁcation temperature is found, which means that a signiﬁcance of the result is not given. Therefore, from the statistical point of view, a correlation between bed material cycle rate and product gas composition is more probable. However, the number of cases for this analysis is low and therefore must not be seen as proof, but as additional indication for the signiﬁcant inﬂuence of the bed material cycle rate on the SER process. 3.3. General ﬁndings regarding bed material cycle rate The general ﬁndings presented in this work deal with the carbon balance of the gasiﬁcation reactor. Carbon introduced by the fuel can either leave the gasiﬁcation reactor (i) in the product gas as CO, CO2, higher hydrocarbons, tar, ﬂy char or it is transported to the combustion reactor as (ii) CO2 in the bed material (CaO þ CO2 / CaCO3) or (iii) char (solid residual of gasiﬁcation process). The Soft wood I Soft wood II Soft wood III Lignite Add. soft III wood 20 CO2 in product gas [vol.-%db] Parameter/name 41 15 Low correlation Kendall’s = 0.52 p-value = 0.14 10 5 0 620 640 660 680 700 720 Gasification temperature [°C] 25 Initial bed material as CaCO3: 70 – 80 kg Fuel Input: 88 – 110 kW 20 15 High correlation Kendall’s = 0.90 p-value = 0.003 10 5 0 0 5 10 15 20 25 Bed material cycle rate [1/h] Fig. 7. Kendall rank correlation between CO2 in product gas and gasiﬁcation temperature/bed material cycle rate. presented investigation focuses on the carbon which is transported to the combustion reactor ((ii) and (iii)). First, the CO2 load of the bed material is displayed as a logarithmic function of the bed material cycle rate in Fig. 8. It can be learned that the CO2 load increases with a decreasing cycle rate. This behavior is expectable, since it has already been shown that a lower cycle rate increases the total amount of CO2 which is captured in the gasiﬁcation reactor and released again in the Gasification temperature [°C] 700 680 660 Soft wood I Soft wood II Soft wood III Rice husk Add. soft wood Bark Lignite I Lignite II Lignite III 640 620 600 0 5 10 15 20 25 Bed material cycle rate [1/h] Fig. 6. Correlation of bed material cycle rate and gasiﬁcation temperature. Fig. 8. Logarithmic CO2 load of bed material as a function of bed material cycle rate. J. Fuchs et al. / Energy 162 (2018) 35e44 combustion reactor. If the cycle rate decreases while the total amount of captured CO2 increases, the CO2 load must increase. Considering the scale of the CO2 load for very low bed material cycle rates (about 0.15 molCO2/molCaO), limitations regarding CO2 sorption for even lower bed material cycle rates are expected: It is a well-known fact that the sorption capacity of limestones decreases with calcination/carbonation cycles [10,27e31] due to changes of the porous structure on its surface. Grasa & Abanades  proposed a residual sorption capacity of about 0.08 molCO2/molCaO for calcination and carbonation in dry atmosphere (CO2/air mixture). However, it is also well-known that steam increases the sorption capacity signiﬁcantly [32e34]. Thermogravimetric analyzer (TGA) experiments of the limestone used for the test runs in the 100 kWth test plant at TU Wien showed a remaining capacity of about 0.2 molCO2/molCaO after 40 cycles of calcination/carbonation in a 30 vol.-% steam atmosphere. Interestingly, the CO2 load seems to be independent from the used fuel type. This means that for a certain bed material cycle rate, the same CO2 load of the bed material is obtained for every investigated fuel. A similar behavior can be found for the carbon transport via carbonation of the bed material (Equation (1)). Fig. 9 shows the carbon transport via bed material with regard to the total carbon introduced into the gasiﬁcation reactor as a function of bed material cycle rate. For low cycle rates it can be shown that nearly 40% of the carbon in the fuel (0.4 molCO2/molC, fuel) are removed from the gasiﬁcation reactor and transported to the combustion reactor sole by CO2 sorption. The carbon transport via bed material decreases to about 15% for a cycle rate of about 20 h1. The decrease can be explained by a shorter residence time in the gasiﬁcation reactor. Again, no signiﬁcant inﬂuence of the fuel type can be found. Considering the carbon transported in the char from the gasiﬁcation reactor to the combustion reactor, a different behavior is obvious (Fig. 10): In case of lignite, about 60% of the carbon in the fuel (0.6 molC, char/molC, fuel) was transported to the combustion reactor as char, whereas for soft wood and bark only between 20 and 50% could be fed to the combustion reactor as char, depending on the cycle rate. The most reasonable explanation for this behavior is Initial bed material as CaCO3: 70 – 80 kg Fuel Input: 88 – 110 kW Gasification temperature: 620 – 700 °C 0.7 0.6 C transport via char [molC,char/molC,fuel] 42 54 wt.-%daf volatiles in fuel 0.5 0.4 0.3 0.2 78 – 86 wt.-%daf volatiles in fuel 0.1 0 0 0.3 0.2 0.1 Initial bed material as CaCO3: 70 – 80 kg Fuel Input: 88 – 110 kW Gasification temperature: 620 – 700 °C 0 0 5 10 15 20 25 Bed material cycle rate [1/h] Fig. 9. Carbon transport from GR to CR via bed material (CaO þ CO2 / CaCO3) as a function of bed material cycle rate. 10 15 20 25 Fig. 10. Carbon transport from GR to CR via char as a function of bed material cycle rate. the volatile matter of the different fuel types. Lignite contains about 54 wt.-%daf volatiles only. Soft wood and bark contain between 78 and 86 wt.-%daf volatiles. Vice versa, the high amount of char in the fuel lignite was not as easily devolatilized and gasiﬁed in the gasiﬁcation reactor. Therefore, more char was transported to the combustion reactor. This behavior was found in earlier publications for conventional DFB gasiﬁcation without CO2 sorption [1,35] as well. However, the increased residence time of the fuel in the gasiﬁcation Sum of C transport (bed material + char) [molC/molC,fuel] C transport via bed material [molCO2/molC,fuel] 0.4 5 Bed material cycle rate [1/h] Initial bed material as CaCO3: 70 – 80 kg Fuel Input: 88 – 110 kW Gasification temperature: 620 – 700 °C 0.5 Soft wood I Soft wood II Soft wood III Rice husk Add. soft wood Bark Lignite I Lignite II Lignite III Soft wood I Soft wood II Soft wood III Rice husk Add. soft wood Bark Lignite I Lignite II Lignite III 1 54 wt.-%daf volatiles in fuel 0.8 0.6 Soft wood I Soft wood II Soft wood III Rice husk Add. soft wood Bark Lignite I Lignite II Lignite III 78 – 86 wt.-%daf volatiles in fuel 0.4 0.2 0 0 5 10 15 20 25 Bed material cycle rate [1/h] Fig. 11. Sum of carbon transport from GR to CR as a function of bed material cycle rate. Main product gas component [vol.-%db] J. Fuchs et al. / Energy 162 (2018) 35e44 CO2 sorption in the gasiﬁcation reactor and therefore in the product gas composition. In this work it is shown that the bed material cycle rate should not be disregarded. Subject of investigation is the carbon transport from the gasiﬁcation reactor to the combustion reactor, which is possible in two ways: CO2 in the bed material (CaO þ CO2 / CaCO3) or char (solid residual of gasiﬁcation process). The following conclusions can be drawn: 75 H2 65 55 Fuel: Soft wood Fuel input: 88 – 110 kW Gas. Temp.: 620 – 700 °C 25 5 0 CO2 CH4 15 CO C xH y 0 5 10 43 15 20 Bed material cycle rate [1/h] Fig. 12. Main product gas composition for soft wood as a function of bed material cycle rate. reactor due to low bed material cycle rates leads to an exceptionally low carbon transport of about 20% (0.2 molC, char/molC, fuel) via char for fuels with high volatile matter (soft wood and bark). It is assumed that a similar behavior could be found for fuels with low volatile matter (lignite in this case), but during the experiments such low bed material cycle rates have not been investigated for lignite. By considering the sum of both (Fig. 11) carbon transport via bed material and via char, a nearly constant amount of carbon was shifted to the combustion reactor e independent from the bed material cycle rate, but dependent on the used fuel type: For lignite about 80% of the total carbon in the fuel was removed from the gasiﬁcation reactor, whereas for the fuels with high volatile matter, only about 63% on average were removed from the gasiﬁcation reactor. The independence from the cycle rate is surprising, since both individual functions vary in a wide range. Finally, the main product gas composition is displayed as a function of the bed material cycle rate in Fig. 12. It is obvious that the lowest CO2 and highest H2 contents in the product gas can be obtained by low bed material cycle rates. This can be explained easily by the ﬁndings of Fig. 9, which shows that a low bed material cycle rate leads to a high CO2 removal in the gasiﬁcation reactor. Therefore, the water-gas shift reaction (Equation (3)) is stimulated to produce H2. 4. Conclusion Dual ﬂuidized bed steam gasiﬁcation of four different fuels, operating the so-called sorption enhanced reforming (SER) process for the production of a H2-rich reduction gas for steel industry, is investigated in this paper. The SER process aims for the production of a H2-rich product gas by in-situ CO2 removal from the gasiﬁcation reactor via the bed material system CaO/CaCO3. Problem-free operation conditions for the SER process with the four different fuel types are demonstrated and it is shown that the amount of volatiles in a fuel exerts a great inﬂuence on the operation conditions of the two interconnected reactors of a dual ﬂuidized bed gasiﬁcation system. In the past, research mainly focused on temperature dependency of the process since it seemed to be the main factor for (i) The carbon transport via bed material is independent from the investigated fuel type. In the experiments, up to 40% of the carbon in the fuel is captured by the bed material at low bed material cycle rate. (ii) On the contrary, the carbon transport via char is dependent on the volatile matter in the fuel. A signiﬁcantly higher share of carbon is transported as char to the combustion reactor for lignite (low volatile matter). Again, the carbon transport via char is highly dependent on the bed material cycle rate. Low cycle rates lead to a high residence time of char in the gasiﬁcation reactor and therefore a low amount of char can be observed. (iii) The sum of carbon transport to the combustion reactor (carbon in bed material and char) is independent from the bed material cycle rate. For soft wood and bark a total carbon transport of 63% on average is observed, whereas for lignite 80% of the carbon is removed from the gasiﬁcation reactor. (iv) According to the ﬁndings (i) to (iii), higher H2 contents in the product gas can be reached by applying low bed material cycle rates. Acknowledgements The present work is part of the research project ERBA II in cooperation with voestalpine Stahl GmbH and voestalpine Stahl Donawitz GmbH. ERBA II receives ﬁnancial support from the research program “Energieforschung” funded by the “Austrian Climate and Energy Fund”. References  Benedikt F, Fuchs J, Schmid JC, Müller S, Hofbauer H. Advanced dual ﬂuidized bed steam gasiﬁcation of wood and lignite with calcite as bed material. Kor J Chem Eng 2017;34:1e11. https://doi.org/10.1007/s11814-017-0141-y.  Hofbauer H, Rauch R, Bosch K, Koch R, Aichernig C. Biomass CHP plant Güssing - a success story. In: Brigwate AV, editor. Pyrolysis Gasif. Biomass waste. CPL Press; 2003. p. 527e36.  Wilk V, Hofbauer H. Analysis of optimization potential in commercial biomass gasiﬁcation plants using process simulation. Fuel Process Technol 2015;141: 138e47. https://doi.org/10.1016/j.fuproc.2015.07.035.  Li YH, Chen Z, Watkinson P, Bi X, Grace J, Lim CJ, et al. A novel dual-bed for steam gasiﬁcation of biomass. Biomass Convers Bioreﬁnery 2017. https:// doi.org/10.1007/s13399-017-0288-0.  Kuba M, Kirnbauer F, Hofbauer H. Inﬂuence of coated olivine on the conversion of intermediate products from decomposition of biomass tars during gasiﬁcation. Biomass Convers Bioreﬁnery 2016. https://doi.org/10.1007/ s13399-016-0204-z.  Kuba M, Havlik F, Kirnbauer F, Hofbauer H. Inﬂuence of bed material coatings on the water-gas-shift reaction and steam reforming of toluene as tar model compound of biomass gasiﬁcation. Biomass Bioenergy 2015:1e10. https:// doi.org/10.1016/j.biombioe.2015.11.029.  Koppatz S, Pfeifer C, Hofbauer H. Comparison of the performance behaviour of silica sand and olivine in a dual ﬂuidised bed reactor system for steam gasiﬁcation of biomass at pilot plant scale. Chem Eng J 2011;175:468e83. https:// doi.org/10.1016/j.cej.2011.09.071.  Schmid JC, Fuchs J, Benedikt F, Mauerhofer AM, Müller S, Hofbauer H, et al. Sorption enhanced reforming with the novel dual ﬂuidized bed test plant at TU wien. Eur Biomass Conf Exhib 2017:421e8. Stockholm.  Müller S, Fuchs J, Schmid JC, Benedikt F, Hofbauer H. Experimental development of sorption enhanced reforming by the use of an advanced gasiﬁcation test plant. Int J Hydrogen Energy 2017;42:29694e707. https://doi.org/10. 1016/j.ijhydene.2017.10.119.  Diem R, Mueller S, Fuchs M, Schmid JC, Hofbauer H. Sorption-enhanced 44             J. Fuchs et al. / Energy 162 (2018) 35e44 reforming with limestone from iron production. Biomass Convers Bioreﬁnery 2014:95e102. https://doi.org/10.1007/s13399-014-0149-z. Koppatz S, Pfeifer C, Rauch R, Hofbauer H, Marquard-Moellenstedt T, Specht M. H2 rich product gas by steam gasiﬁcation of biomass with in situ CO2 absorption in a dual ﬂuidized bed system of 8 MW fuel input. Fuel Process Technol 2009;90:914e21. https://doi.org/10.1016/j.fuproc.2009.03.016. €ll T, Hofbauer H. Hydrogen production based Müller S, Koppatz S, Fuchs M, Pro on conventional dual ﬂuid gasiﬁcation versus sorption enhanced reforming. In: Int. Conf. Polygeneration Strateg.; 2013. Vienna. Brellochs J. Experimentelle Untersuchung und Prozess-Simulation der AERBiomassevergasung zur Erzeugung eines regenerativen Erdgassubstitutes. Dissertation. 2014. https://cuvillier.de/de/shop/publications/6859experimentelle-untersuchung-und-prozess-simulation-der-aer-biomassevergasung-zur-erzeugung-eines-regenerativen-erdgassubstitutes (accessed March 7, 2016). Hawthorne C, Poboss N, Dieter H, Gredinger A, Zieba M, Scheffknecht G. Operation and results of a 200-kWth dual ﬂuidized bed pilot plant gasiﬁer with adsorption-enhanced reforming. Biomass Convers Bioreﬁnery 2012: 217e27. https://doi.org/10.1007/s13399-012-0053-3. Poboss N, Swiecki K, Charitos A, Hawthorne C, Zieba M, Scheffknecht G. Experimental investigation of the absorption enhanced reforming of biomass in a 20 kwth dual ﬂuidized bed system. Int J Therm 2012;15:53e9. https:// doi.org/10.5541/ijot.321. Baker EH. The calcium oxide-carbon dioxide system in the pressure range 1300 atmospheres. J Chem Inf Model 1962:464e70. https://doi.org/10.1017/ CBO9781107415324.004. Wang X. Hydrogen production by coal gasiﬁcation coupled with in situ CO2 capture. Energy Power Eng 2012;04:210e7. https://doi.org/10.4236/ epe.2012.44029. Poboß N. Experimentelle Untersuchung der sorptionsunterstützten Refor€t Stuttgart; 2016. mierung. PhD Thesis. Universita €ftberger E. In-situ CO2-adsorption in a dual ﬂuidised bed biomass steam Ho gasiﬁer to produce a hydrogen rich gas. TU Wien. PhD Thesis. 2005. Soukup G. Der AER e prozess, Weiterentwicklung in einer Technikumsanlage und Demonstration an einer Großanlage. TU Wien. PhD Thesis. 2009. €rl R, Scheffknecht G. Production of tailored syngas Hafner S, Armbrust N, Spo for Dimethyl Ether synthesis by sorption enhanced gasiﬁcation. In: 12th int. Conf. Fluid. Bed technol; 2017. Krakow. Armbrust N, Schweitzer D, Gredinger A, Beirow M, Poboss TBN, Hawthorne C, et al. Gasiﬁcation of biomass with in-situ CO2 capture and separation in a 200              kWth pilot plant ﬂuidized bed gasiﬁcation infrastructure. Gasif Technol 2014: 2014. Washington D.C. €ll T, Kitzler H, Pfeifer C, Hofbauer H. Cold ﬂow model inSchmid J, Pro vestigations of the countercurrent ﬂow of a dual circulating ﬂuidized bed gasiﬁer. Biomass Convers Bioreﬁnery 2012;2:229e44. https://doi.org/ 10.1007/s13399-012-0035-5. CEN/TS 15439. 2006. p. 2006. Wolfesberger-Schwabl U, Aigner I, Hofbauer H. Mechanism of tar generation during ﬂuidized bed gasiﬁcation and low temperature pyrolysis. Ind Eng Chem Res 2012;51:13001e7. https://doi.org/10.1021/ie300827d. Fuchs J, Müller S, Hofbauer H. Ash related limitations of dual ﬂuidized bed steam gasiﬁcation. 13. Minisymp. Innsbruck: für Verfahrenstechnik; 2017. Grasa G, Abanades J. CO2 capture capacity of CaO in long series of carbonation/ calcination cycles. Ind Eng Chem Res 2006;45:8846e8851. Manovic V, Anthony EJ, Grasa G, Abanades JC. CO2 looping cycle performance of a high-purity limestone after thermal activation/doping. Energy Fuels 2008;22:3258e64. https://doi.org/10.1021/ef800316h. Lysikov AI, Salanov AN, Okunev AG. Change of CO2 carrying capacity of CaO in isothermal recarbonation-decomposition cycles. Ind Eng Chem Res 2007;46: 4633e8. https://doi.org/10.1021/ie0702328. Florin N, Harris A. Enhanced hydrogen production from biomass with in situ carbon dioxide capture using calcium oxide sorbents. Chem Eng Sci 2008;63: 287e316. Wu Y, Blamey J, Anthony EJ, Fennell PS. Morphological changes of limestone sorbent particles during carbonation/calcination looping cycles in a thermogravimetric analyzer (TGA) and reactivation with steam. Energy Fuels 2010;24:2768e76. https://doi.org/10.1021/ef9012449. Liu W, An H, Qin C, Yin J, Wang G, Feng B, et al. Performance enhancement of calcium oxide sorbents for cyclic CO2 capture- A review. Energy Fuels 2012;26:2751e67. https://doi.org/10.1021/ef300220x. Manovic V, Anthony EJ. Thermal activation of CaO-based sorbent and selfreactivation during CO2 capture looping cycles. Environ Sci Technol 2008;42:4170e4. https://doi.org/10.1021/es800152s. Donat F, Florin NH, Anthony EJ, Fennell PS. Inﬂuence of high-temperature steam on the reactivity of CaO sorbent for CO2 capture. Environ Sci Technol 2012;46:1262e9. https://doi.org/10.1021/es202679w. Kern S, Pfeifer C, Hofbauer H. Co-gasiﬁcation of wood and lignite in a dual ﬂuidized bed gasiﬁer. Energy Fuels 2013;27:919e31. https://doi.org/10.1021/ ef301761m.