Tar yield and composition from poultry litter gasification in a fluidised bed reactor: effects of equivalence ratio, temperature and limestone addition

Air gasification of poultry litter was experimentally investigated in a laboratory scale bubbling fluidised bed gasifier. Gasification tests were conducted at atmospheric pressure using silica sand as the bed material. This paper examines the effect of the equivalence ratio (ER) in the range of 0.18–0.41, temperature between 700 and 800 °C, and the addition of limestone blended with the poultry litter on the yield and composition of tar. An off-line solid phase adsorption method was employed in order to quantify tar compounds heavier than styrene, whereas lighter species such as benzene and toluene were measured by means of on-line micro gas chromatography. Total tar yields were in the range from 15.7 to 30.7 gtotal tar kgpoultry litter (dry and ash free basis)−1. These values are considered low with respect to the feedstocks with a higher organic fraction. It also needs to be noted that the yields of benzene and toluene were measured by on-line micro gas chromatography, a technique which inherently delivers higher tar values compared to commonly employed off-line techniques. By varying the ER, poultry litter blended with limestone showed a reduction in total tar yield whereas poultry litter on its own showed an increasing tar yield over the ER range tested. In the presence of limestone, polycyclic aromatic hydrocarbons (PAHs), heterocyclic compounds, toluene and benzene showed a tendency to reduce over the ER range tested. Since the ER also plays a crucial role in tar reduction, the reduction in tar cannot be unambiguously attributed to calcined limestone/lime (CaCO3/CaO). Increasing the temperature was shown to be effective for reducing the total tar yield but the amounts of polycyclic aromatic hydrocarbons increased. However, no definitive correlation could be established between limestone/lime catalytic activity for tar reduction and elevated gasification temperature, because there was no possibility to study their effects separately. The chemical composition of the tar arising from poultry litter is distinctive compared with conventional lignocellulosic fuels linked to the fact that poultry litter has a higher nitrogen content (≈6.5% w/w (dry and ash free basis)). Nitrogen-containing hydrocarbons such as pyridine, 2-methylpyridine, 2-methyl-1H-pyrrole and benzonitrile were identified in significant amounts. This study has demonstrated that poultry litter gasified in a bubbling fluidised bed yielded a product gas with relatively low tar content while its composition reflects the chemical nature of the feedstock.


Introduction
According to the AVEC annual report 2016, 1 the European Union is the world's leading supplier of poultry meat with an annual production of 13.6 million tonnes in 2015.The report also predicted a growth rate of about 1% a year.Even though intensive livestock production has become a more economically viable option compared to traditional farming practices, such industrialised production faces issues associated with its environmental impact due to the accumulation of large quantities of waste with estimates of 1.4 billion tonnes per annum 2 of manure in EU states.The increasing popularity of free range and organic farming supported by European Directives 2007/43/ EC and 1999/74/EC requires poultry farmers to comply with minimum animal welfare standards which results in an increased volume of poultry litter due to utilisation of bedding material (i.e.wood shavings, straw and hay).Poultry litter is a heterogeneous fuel composed of bedding material, excreta, waste feed and feathers.Compared to conventional lignocellulosic feedstocks, poultry litter is recognised as a low value fuel due to its relatively high moisture and ash content.It is also a source of nutrients such as nitrogen, phosphorous and potassium. 3he recent European Commission (EU) regulation no.592/ 2014 paves the path for combustion of poultry litter and onfarm utilisation of the energy.Combustion remains the most advanced and widely applied technology with several commercial scale incinerators of poultry litter currently being used for electricity generation and ash recovery in the UK, the USA and The Netherlands. 4 Recently poultry litter has been successfully gasied in a uidised bed gasier 5,6 as well as in a xed/moving bed gasier. 7These studies have concluded that due to the high content of elements such as phosphorous and potassium, poultry litter is prone to provoke sintering and agglomeration when gasied in a uidised bed gasier.To avoid these problems limestone/lime (CaCO 3 /CaO) have been added to the bed during industrial scale uidised bed combustion 8 and laboratory scale uidised bed gasication 6 of poultry litter.
Tar is a by-product of thermal gasication processes dened as a generic (unspecic) term accounting for all organic compounds in the product gas excluding benzene and lighter gaseous hydrocarbons. 9Although CEN/TS 15439, 2006 9 provides a standardised denition of tar, in practice tar is dened by restrictions of sampling, the identication and quantication methods applied as well as the nal application of the product gas or syngas.For example, the solid phase adsorption (SPA) sampling method combined with gas chromatography (GC) detection provides accurate results for phenolic and 2-5 rings polycyclic aromatic tar compounds.However, measurement reliability drops remarkably when either light volatile compounds (i.e.benzene) or 6 + ring polycyclic aromatic tar compounds are to be quantied.Tar is a black viscous material potentially giving rise to system malfunction if condensation occurs.In practice, tar needs to be removed for applications other than direct burning of the product gas. 10Tar from poultry litter gasication in a uidised bed reactor has not been reported to date.Although, it has been established that a large portion of the nitrogen present in poultry litter is converted into NH 3 and HCN during uidised bed gasication 6 it is expected that the high nitrogen content in poultry litter would also deliver a wide variety of nitrogencontaining compounds in the tar.Jaramillo-Arango et al. 11 investigated the composition of pyrolysis oil from uidised bed tests employing nitrogen rich sewage sludge.They found notable amounts of aliphatic acetamide, single aromatic ring species such as pyridine, pyrimidine, pyrrole, aniline and benzonitrile as well as two ring structures such as quinoline and indole.The composition of the tar from poultry litter gasication is expected to reect the high nitrogen and low lignin content in the feedstock.The formation and decomposition of poultry litter tar is further discussed in the Section 3.
It is well known that tar can be decomposed catalytically with limestone/lime which is an inexpensive, abundant and naturally occurring non-toxic material. 12,13Additionally, it has been reported that Ca-based additives enhance pyrolysis conversion rates. 14An activation energy of 46 kJ mol À1 for tar cracking over a calcined dolomite (CaO-MgO) and 77 kJ mol À1 over an inert bed of silica sand was reported by Delgado et al. 15 In regards to tar mitigation, Simell et al. 12 tested the catalytic activity of carbonate rocks by passing model tar compounds over a xed catalytic bed.The authors concluded that calcined CaO shows good catalytic activity towards tar reduction/reforming.In contrast, a much slower reduction rate has been observed for raw CaCO 3 , which is so slow that it is considered not to occur.However, CaO converts into the carbonated form CaCO 3 , when the CO 2 partial pressure is higher than that of the equilibrium pressure at the process temperature.Tests have shown that at 900 C CaO was carbonated to CaCO 3 only if the partial pressure of CO 2 was higher than 100 kPa.On the other hand, Valverde and Medina 16 demonstrated that the calcination reaction triggers at 857 C under the 50% CO 2 atmosphere.Rapid calcination has also been observed at temperatures below 857 C. With respect to gasication the CO 2 yield is unlikely to reach the equilibrium partial pressure that could be able to reverse the calcination process.Bedyk et al. 17 found out that in inert argon atmosphere CO 2 is released from CaCO 3 in the temperature range 500-700 C forming CaO.While tests in a pure CO 2 atmosphere and a temperature range of 600-800 C resulted in CO 2 adsorption forming CaCO 3 .In the same CO 2 atmosphere CO 2 release occurred at the temperatures above 850 C.
Saw and Pang 13 tested the extent of tar reduction with 0%, 50% and 100% lime as the bed material.The total tar concentration (sum of all tar compounds) decreased exponentially from 5.0 to 0.7 g Nm À3 as the lime loading increased from 0% to 100%.A signicant reduction was also observed for all the individual tar compounds studied.Tar reduction with lime loading is most likely due to the steam reforming of tar in the presence of CaO.The steam reforming reactions of the phenol, cresols and toluene are shown in eqn (1)-( 4). (1) Enhanced production of H 2 may have a negative effect on the tar steam reforming rate because H 2 decreases the catalytic activity of CaO due to the adsorption of H 2 onto its active sites, diminishing the available sites for tar adsorption. 13Likewise CaO can react with other gasication products (i.e.][19] In this paper, an experimental study to characterise the yield and composition of tar in the gas during poultry litter gasication in a laboratory scale uidised bed reactor is presented.The objectives of this study are to investigate the effect of (a) equivalence ratio (ER), (b) limestone addition (blended with the poultry litter) and (c) reactor temperature, on tar yield and its composition.Some basic data regarding overall tar yields were published by Pandey et al., 6 in a complementary publication to the present work.However, a detailed tar analysis is presented here, including the change of tar composition with the process variables.An additional key aspect of the present work is a discussion on the evolution of nitrogen-containing tar compounds present in the product gas.

Materials
Detailed description of poultry litter collection, preparation and characterisation including detailed chemical composition of the poultry litter ash can be found elsewhere. 6However, a summary of relevant information is presented here.The bulk density of the partially dried poultry litter was 360 kg m À3 , with a particle size between 0.7 and 2.8 mm.The limestone used in this study was supplied by Rheinkalk GmbH (Brilon, Germany) with a particle size in the range of 0.9-1.2mm.Ultimate and proximate properties, chemical composition as well as heating value of the poultry litter are reported in Table 1.The content of xed carbon was calculated by subtracting the moisture, ash and volatile matter content from 100%.Likewise, the oxygen content in the fuel was calculated by difference.

Experimental facility
The gasication experiments were conducted using a laboratory scale air-blown bubbling uidised bed gasier located at the Energy Research Centre of The Netherlands (ECN).Experiments were performed at different temperatures (700, 750 and 800 C) and at different ERs between 0.18 and 0.41 by adjusting the air and N 2 ow rate while maintaining the total ow at 12 dm 3 min À1 (Table 2).Along with that an effort has been made to maintain a constant feedstock feed rate.The downstream sections of the reactor up to the cold lter were insulated and maintained at 400 C in order to avoid tar condensation.Tar samples were taken through a SPA sampling port located aer the hot lter.Silica sand with a particle size between 0.25 and 0.50 mm (mean particle size of 0.31 mm) and bulk and absolute densities of 1422 and 2620 kg m À3 respectively was used as the bed material.To avoid any inuence of accumulated ash from previous experiments and to reduce the risk of bed agglomeration, 1.2 kg of fresh silica sand was introduced at the beginning of each experimental day.Gasication experiments were conducted in such a way that the uidising regime remained constant throughout the tests.The minimum uidising velocity was around 0.097 m s À1 at 20 C, calculated according to Wen and Yu's correlation. 20Each test undertaken at the specied gasication conditions lasted about an hour.Steady state was usually reached within the rst 30 min aer commencing fuel feeding.The nal 30 min were dedicated to the tar sampling and analysis of permanent gases using an on-line micro gas chromatograph (mGC) (Varian, CP-4900).Relevant information comprising schematic diagram of gasier, technical data and operating conditions of the experimental setup was previously presented by Pandey et al. 6 and are also concisely outlined in Table 2.It is worth mentioning that tests 1, 2 and 3 were carried out using only poultry litter, while tests from 5 to 14 were conducted with poultry litter blended with 8% w/w of limestone.

Tar measurement methods
The detailed description of the SPA tar sampling method, extraction and chromatographic analysis of tar is provided elsewhere. 21Briey, SPA cartridges were assembled by packing 500 mg of aminopropyl silica sorbent.The sampling volume was adjusted to 100 ml of dry product gas.For each experimental condition duplicate SPA samples were taken.Aer sampling the cartridges were shipped to the University of Limerick -Ireland where the tar compounds were extracted from the sorbent with 3 Â 600 ml of dichloromethane.tert-Butylcyclohexane and 4-ethoxy phenol were added as internal standards to the extracted tar solutions.An Agilent 7890A GC coupled with a triple-axis MSD 5975C was used for identication of the most abundant tar compounds.A Thermo Scientic Trace 1310 GC with a ame ionisation detector (GC-FID) was used for tar quantication.Calibration curves naphthalene/tertbutylcyclohexane and phenol/4-ethoxy phenol were applied to quantify the aromatic and phenolic tar, respectively.The FID detector was deemed more accurate than the MSD for tar quantication based on a statistical comparison of both methods. 21Benzene and toluene concentrations were measured along with the other permanent gases using on-line mGC (Varian, CP-4900).Their yields are presented as an average of four successive measurements conducted at three-minute intervals.Tar yields are expressed on a mass basis as g tar kg poultry litter (dry and ash free-d.a.f.) À1 in order to eliminate any dilution effect of the product gas when slight deviations of biomass feed rate occur, 22 or when the oxygen to nitrogen ratio is reduced to adjust for lower ER. 23Tar yields are presented graphically but are also tabulated in the ESI †.Tabulated yields presented as g tar kg poultry litter (d.a.f.) À1 enable fundamental studies of tar with respect to the feedstock, while yields presented as g tar m dry product gas À3 are more useful for developers and operators of downstream applications such as tar mitigation technologies or using the product gas in internal combustion engine.Total tar in this paper refers to GC detectable tar sampled by SPA inclusive of tar compounds from styrene (M z 104 g mol À1 ) to benz[a]anthracene (M z 228 g mol À1 ) as well as benzene and toluene measured by on-line micro GC instrument.The reason why total tar does not include light tar compounds (i.e.so called BTEX -Benzene, Toluene, Ethylbenzene, Xylene) sampled by SPA is due to delay between sampling and extraction of the SPA cartridges.As reported previously by Horvat et al. 24 a signicant portion of the volatile compounds such as benzene, toluene and xylene are lost during transport resulting in their quantitative underestimation as well as poor measurement repeatability.Instead, benzene and toluene measured by on-line mGC is combined with SPA sampled tar to sum up for total tar.Although, according to CEN/TS 15439, 2006 9 benzene is excluded from the denition of tar, in the present work it is presented as a tar compound since its aromatic chemical structure is more characteristic of tar species than permanent gases.The results of poultry litter tar sampled by SPA are presented in duplicate for each gasication condition to show the repeatability of the measurements and the random errors associated with uctuations in the feeding rate.It is evident from Table 3 that the measurement repeatability in this study is mediocre.One possible reason could be the relatively low tar content in the product gas (i.e.under 10 g tar kg poultry litter (d.a.f.) À1 summing up compounds from styrene to benz[a]anthracene) along with previously mentioned losses during the analysis delay.

Results and discussion
The identied tar compounds are presented in Table 4 in the order in which they eluted.The composition of tar from poultry litter gasication is distinctively different from tar composition from conventional lignocellulosic fuels, specically in terms of nitrogen-containing hydrocarbons.Most of the nitrogen in the poultry litter derives from the animal feed, excreta and feathers rather than from the bedding material.This nitrogen is chemically incorporated into protein molecules and urea. 28,29It is believed that the presence of signicant amounts of pyridine, 2methylpyridine, 2-methyl-1H-pyrrole and benzonitrile in the tar is due to the high level of nitrogen in the fuel (poultry litter).However, the question remains as to whether nitrogen-containing hydrocarbons arising from proteins retain their monomer structure or derive from reforming reactions between permanent gases (i.e.NO x , NH 3 , HCN) and the condensable fraction (i.e.tar) in the product gas.Nitrogen-containing compounds are normally not reported in the relevant gasication literature since insignicant amounts are generated from conventional lignocellulosic feedstock.Fig. 1 shows the structural formulae of the nitrogen-containing compounds identi-ed in this study.
The formation of nitrogen-containing hydrocarbons in the pyrolysis process has been investigated by Dignac et al. 30 In the pyrolysates from fresh vegetables pyridine, pyrrole, benzonitrile and indole derivatives were detected among the other nitrogencontaining hydrocarbons.The authors attributed the pyridine derivatives to the pyrolysis of alanine-containing proteins and peptides, with the benzonitrile derivatives probably formed from pyrolysis of phenylalanine-containing proteins.Pyrrole and its derivatives were formed by cyclisation during pyrolysis of proteins containing the amino acids proline, hydroxyproline, glycine and glutamic acid, but could also be pyrolysis products of pigments such as chlorophyll.The proteins in poultry litter originate from waste feed and feathers, while the chlorophyll derivatives originate from bedding material and waste feed.Poultry excreta also contains nitrogen that possibly plays a role in the formation of nitrogen-containing hydrocarbons as indicated by Inoue et al. 31 who analysed the products of liquefaction of ammonia and cellulose.Brebu and Spiridon 32 investigated the thermal degradation of sheep wool, human hair and chicken feathers containing keratin proteins and attributed the formation of aromatic pyrroles and pyridines to the amino acids in the protein of keratin.The majority of the nitrogencontaining hydrocarbons were found in the aqueous phase of the pyrolysis condensate which needs to be taken into account in the development of tar cleaning and waste water treatment technologies.
Eight individual tar compounds sampled by the SPA method (designated by * in Table 4) are presented quantitatively in Fig. 2-6.Pyridine and benzonitrile represent nitrogencontaining hydrocarbons while phenolic hydrocarbons mainly contain phenol and cresols.It should be noted that two isomers of cresol are summed and presented as a single quantity.Indene, naphthalene, acenaphthylene and phenanthrene are representatives of PAHs.This journal is © The Royal Society of Chemistry 2019 Table 5 includes the yields of principal permanent gases expressed in g gas kg poultry litter (d.a.f.) À1 .Although the yields of permanent gases are not the main focus of the present work, these data aim to support discussion on tar yield and composition.
The scale on the y-axis is kept the same in all graphs in order to simplify comparison of tar yields.The results indicate that tar yields from poultry litter gasication are lower than from feedstocks with a higher organic fraction.This was corroborated by only mild coloration of the white aminopropyl silica sorbent which typically turns dark yellow when product gas with high tar content is sampled.Low total tar yields can be attributed to the very specic composition of poultry litter which has a high ash content and low organic fraction, in particular low lignin content (Table 1).Lignin is known to be a tar precursor giving rise to higher total GC detectable tar and PAHs than cellulose and hemicellulose. 34,35However, smaller quantities of phenols and PAHs can also be formed from cellulose and hemicellulose. 36An ash content of 17.55 wt% (dry basis) in poultry litter is regarded as high but its composition and in particular the concentration of elements such as Ca, Mg, Al, Fe, Zn, Mn 6 which exhibit catalytic tar reduction activity could have played a role in the total tar reduction. 37The total tar yields presented in Fig. 2-6 (15.7 to 30.7 g total tar kg poultry litter (d.a.f.) À1 ) include benzene and toluene measured by on-line micro GC.It is worth mentioning that if benzene and toluene yields measured by on-line micro GC are subtracted from the total tar from poultry litter, the yield of total tar would drop to between 3.1 to 10.3 g total tar kg poultry litter (d.a.f.) À1 .It has been reported previously 24 that on-line mGC measurements give considerable higher benzene and toluene quantities comparing to SPA.Additionally Brage et al. 38 showed that SPA sampling is far superior for quantication of BTEX compounds compared to traditional cold trapping.Therefore, comparing the total tar yield from the relevant literature is complicated due to differences in dening tar, sampling conditions, analytical instrument calibration and reported units.Kinoshita et al. 23 reported total tar yields in the range of 40-45 g total tar kg dry wood sawdust

À1
while conducting the tests under similar ER conditions to those reported here.The tar sampling set up employed was a combination of dry and wet cold trapping Horvat et al. 25 measured total tar between 14-34 g total tar kg biomass (d.a.f.) À1 from raw and torreed Miscanthus x giganteus respectively using the same experimental reactor as being used for this study.In that case tar compounds in the molecular weight range from benzene to benzo[k]uoranthene were sampled by means of the SPA method.Compared to the poultry litter, both raw and torreed Miscanthus x giganteus have lower ash contents of 2.8 and 4.2 wt% and higher lignin content of 21 and 43 wt%, respectively.

Effect of equivalence ratio on tar yield and compositionwithout limestone addition
Fig. 2 includes total tar yields and composition over the ER range between 0.18 and 0.30 at 700 C, without addition of the limestone to the poultry litter.It is evident that total tar as well as nitrogen and oxygen containing tar compounds slightly increase with the ER.Such an observation is in disagreement to the results presented by Kinoshita et al. 23 and Hanping et al. 39 employing wood sawdust, peanut shell and wheat straw as a fuel.They observed a decrease of total tar and oxygen containing tar compounds with increasing ER while keeping the temperature at 700 and 800 C, respectively.However, more recently Horvat et al. 25 found that at constant temperature, the ER has relatively little impact on the yield or composition of tar from a grassy biomass.These three studies 23,25,39 have been considered for comparison due to their ability to study the effect of ER separately from the temperature effect on tar evolution.The concentration of PAH compounds in this study increases with increasing ER.The yields for benzene and toluene increase slightly and then level-off at an ER of 0.3.Similar to the total tar, the yield of product gases increases with ER.At higher ER more oxygen is available for char conversion into product gas as well as reacting with permanent gases and tar.Therefore, as expected an increase in product gas yield and carbon conversion (from 49.1% at ER of 0.18 to over 70% at ER of 0.22 and 0.3) was observed.However, this was accompanied by an increase in tar yield.Poultry litter comprise a very high fraction of extractives (40% on a dry basis, Table 1) but their exact composition is not known.Literature information on the extractive fraction in poultry litter is scarce.It is not known to what extend these extractives contribute to either the product gas or tar yields.

Effect of equivalence ratio on tar yield and compositionwith limestone addition
Fig. 3 presents tar yields for the experiments undertaken between an ER of 0.29 and 0.41, at a gasication temperature of 700 C using poultry litter blended with limestone (8% w/w).Fig. 2 and 3 show data for the same temperature, but the ERs correspond to two different ranges (0.18-0.30 vs. 0.29-0.41).Differences in the ER derive from daily variations of fuel feeding rate, despite efforts being made to maintain a constant feeding rate throughout the experimental campaign.Due to the agglomeration issue associated with the raw poultry litter, it was decided to add limestone (CaCO 3 ), however no adjustment was made to the calibration of the feeding system.Since the range of ER differs for both the limestone blended and raw poultry litter, it is not possible to draw unambiguous conclusions regarding whether the difference in tar yields is due solely to the effect of limestone.The only gasication conditions allowing direct evaluation of the effect of limestone are ER 0.30; T 700 C; without limestone addition and ER 0.29; T 700 C; with limestone addition, respectively.There was 5% less of total tar from the test without limestone addition.The difference between the yields of individual tar compounds vary between -9 and 21.3% taking the test without limestone addition as a reference value.These ndings show no signicant effect of limestone addition on the tar yield and its composition at the lowest gasication temperature.Permanent gases measurements actually reveal higher yields of H 2 , CH 4 , CO in the test without limestone addition, while the concentration of CO 2 was similar for both scenarios.Correlation exist between the yields of permanent gases and carbon conversion efficiency which appears to be 81.8% in the test without limestone addition and 70.8% in the test with limestone addition.Although it has been found that alkaline earth metal oxides (CaO/MgO) employed in steam gasication increase the yields of permanent gases (H 2 , CO 2 ) by promoting the decomposition reactions of tar and light hydrocarbon 13,40,41 it seems this phenomena did not occur in gasication test at 700 C.There could be two potential reasons for low catalytic activity of limestone: limited calcination at low temperature and adsorption of sulphur.In order to understand the transformation of limestone during gasication tests additional characterisation of limestone grains separated from the bed aer tests was carried out.These results are discussed in detail in Section 3.4 and they show that at 700 C limestone was only partially calcined, therefore inactive towards tar reduction.
According to Florin and Harris 42 calcination of limestone forming catalytically active lime does take place in an inert atmosphere at 700 C. In an atmosphere containing CO 2 the carbonation/calcination reactions shi to higher temperature comparing to inert atmosphere because of reversible nature of the calcination reaction at an equilibrium condition. 17Lime alteration can also occur in the presence of gaseous H 2 S (Table 5) forming CaS. 43Direct evidence of catalytic deactivation of lime by H 2 S has not been found in the literature, but catalyst poisoning caused by sulphur has been reported previously. 44,45he sulphur content in the recovered limestone/lime aer the gasication at 700 C increased 3 times compared to the unused limestone (Table 6) proving chemical interaction between H 2 S and the limestone/lime.In Fig. 3 a reduction in total tar is observed over the tested ER range when poultry litter was blended with the limestone.Similar trends are observed in Fig. 4 and 5 showing decreasing total tar over the range of ER tested at gasication temperatures of 750 and 800 C, respectively.It is worth emphasizing that the total tar and yields of individual tar species show the same trend.Benzene and toluene also follow a decreasing trend as the other SPA sampled tar compounds do, although toluene does not seem to be notably affected over the ER range tested.There are at least two factors which could have caused decrease of tar: higher content of reactive oxygen (higher ER) and presence of limestone whose calcination degree increased with temperature.From the data available in Table 5 an increase in ER results in a reduction of both H 2 and CO concentration and an increase in CO 2 in the product gas due to combustion of the volatiles and char.Despite its higher concentration, it seems that the CO 2 did not impact on the catalytic ability of the lime due to carbonization into limestone over the timeframe of the experiments at 750 and 800 C. Thermal gravimetric analysis of limestone/lime aer gasication revealed that the degree of in situ calcination was signicantly higher at 750 and 800 C.Moreover, over 10folds higher sulphur content was observed in the recovered limestone/lime from the gasication experiments compared to the unused limestone.Despite high sulphur adsorption, no clear evidence of reduced catalytic tar reduction capacity of limestone/lime was found during the gasication.
Delgado et al. 40 and Simell et al. 12 reported rapid catalytic deactivation of limestone/lime as a result of coke deposition on the surface of active sites.The authors also stated that both wet (steam) and dry (CO 2 ) gasication eliminate coke from the surface which could explain the increased catalytic activity with increasing ER.Wet and dry gasication reactions are strongly endothermic.Therefore, they are more likely to play a crucial role in the tests at 750 and 800 C than at 700 C.Moreover, at higher ER more oxygen is available to oxidise any deposited coke.It is not clear how the oxygen itself affects the redox equilibrium of limestone/lime.Of relevance to the present work, Campoy et al. 46 conducted gasication tests in an air blown bubbling uidised bed reactor using wood pellets as a fuel.Ote, a silicate subvolcanic rock was compared to calcined limestone over an ER range between 0.23 and 0.36.The yields of gravimetric tar decreased only at an ER above 0.3 while employing silicate rock.On the other hand, the addition of calcined limestone resulted in a slight decrease of gravimetric tar over the entire ER range tested.However, notable variations have been found between silicate rock tests and the tests with added calcined limestone.Gravimetric tar quantities between 40 and 50 g tar kg wood (d.a.f.) À1 resulted from silicate rock, while between 25 and 30 g tar kg wood (d.a.f.) À1 were measured aer the addition of calcined limestone.In a nutshell, there are indications that catalytic tar reduction took place when the gasication tests were performed at 750 and 800 C.However, tar reduction cannot be unambiguously attributed to the calcined limestone/lime because oxygen content, ER, could play a crucial role in tar reduction as well.

Effect of temperature on tar yield and composition-with limestone addition
In Fig. 2-4 the yields of phenols (from 0.11 to 1.22 g tar kg poultry litter (d.a.f.) À1 ) and benzonitrile (from 0.10 to 0.61 g tar kg poultry litter (d.a.f.) À1 ) are relatively high at low gasication temperatures between 700 and 750 C.However, at 800 C only between 0.02 and 0.27 g tar kg poultry litter (d.a.f.) À1 , of phenols and benzonitrile are observed due to their conversion via demethylation, dehydration 47 and denitrication reactions. 48Reforming mechanisms using model compounds such as pyridine, pyrrole and indole have been studied in the context of thermochemical conversion of coal. 48,49Liu et al. 48measured NH 3 and HCN as the main gaseous products from conversion of nitrogen-containing hydrocarbons.Gasication of indole was carried out in supercritical water and the authors concluded that one portion of indole converted directly into aromatic compounds without nitrogen by releasing ammonia, while another portion of indole was converted into nitrogen-containing aromatic compounds such as aniline, o-toluidine and 9-nitroso-9H-carbazole.Zhao et al. 49 pyrolysed pyridine and pyrrole at 600-1200 C in a ow reactor.H 2 and HCN were measured in order to determine the thermal stability of pyridine and pyrrole.The results showed that the thermal stability of pyridine is greater since signicant production of HCN was observed at 825 C while pyrrole generated notable amounts of HCN at 775 C. A thermal degradation (i.e.ring-opening) mechanism was proposed for both nitrogen-containing hydrocarbons studied.The pyridine ring undergoes a series of free radical reactions resulting in H 2 and an aliphatic $R-CN.On the other hand, it is assumed that pyrrole undergoes direct ring opening, therefore reforming into an aliphatic R-CN without passing through free radical reactions.Fig. 6 presents the total tar yields and quantities of ten individual tar species with respect to gasication temperature at an ER of 0.29 AE 0.01.The yield of total tar over the temperature range tested remains steady which is considered as an atypical observation with respect to earlier literature ndings.Depending on the range of temperature tested and the tar sampling method employed the yields of total tar either decreased 22,50,51 or exhibited the peak yield at around 750 C followed by a decrease. 25,33,52Steady total tar yields are attributed to the prevailing effect of high benzene concentrations which increase with temperature over less abundant and diminishing species such as benzonitrile, phenols and indene.It should be noted that if benzene and toluene are excluded, the total SPA tar summed from styrene to benz[a]anthracene shows a signicant reduction with increasing temperature.
According to Delgado et al. 40 higher reaction temperature favours catalytic activity of lime for tar destruction in the temperature range of 780-880 C in a uidised bed biomass gasier.Although, the authors also observed catalyst deactivation due to coke formation and adsorption on the active sites regeneration of lime by coke removal was effectively achieved by steam and dry (CO 2 ) gasication.Fig. 6 indicates that catalytic activity of limestone/lime promoted by elevated temperature could have reduced heterocyclic tar (i.e.phenols, benzonitrile), toluene and indene, while no reduction effect is observed for benzene and PAHs.Similar quantitative curves of individual tar compounds presented in Fig. 6 were previously attributed solely to the temperature effect. 23,25Saw and Pang 13 investigated the inuence of lime loading on the tar yields in a dual uidised bed steam gasier operated at 710-750 C.They observed catalytically driven reduction of tar including tar species in classes from C2 to C5 according to ECN classication. 52owever, benzene was not included in their study.Simell et al. 12 tested calcined carbonated rock for its catalytic reforming potency using benzene, toluene and naphthalene as model tar compounds.The order of reforming toluene > naphthalene [ In summary, perhaps there is a trade-off between phenomena including limestone calcination and lime carbonisation, 12,17 coke deposition, coke gasication (i.e.coke removal), 12,40 and sulphur poisoning. 44,45inuencing its catalytic activity.In any case, in the present work there is no incontrovertible evidence of limestone/lime catalytic activity with respect to temperature and associated tar reduction.
Indene has its peak production at 750 C while the PAH yields gradually increases with temperature.The nitrogencontaining hydrocarbons show different behaviour with respect to increasing temperature.Benzonitrile yield decreases while that of pyridine remains relatively high at elevated temperatures indicating its high thermal stability.Pyridine has a non-branching aromatic chemical structure while the benzonitrile substituent makes it more thermally sensitive.This observation was conrmed by Zhao et al. 49 who reported that pyridine undergoes thermal degradation at temperatures above 825 C. Thermal decomposition of nitrogen-containing tar compounds suggests higher yields of NH 3 which was monitored throughout experimental campaign, 6 but due to the rapid thermal decomposition of NH 3 its concentration decreased sharply with the temperature. 53At this point it is worth mentioning that NO x emissions might be elevated upon combustion of the N-containing product gas.An increase in benzene yield with the temperature correlates with reforming of compounds such as phenols, toluene and benzonitrile. 47

Limestone/lime properties before and aer gasication
In order to get better insight into limestone/lime transformation during poultry litter gasication additional tests were performed.Fig. 7 presents differential thermo-gravimetric Fig. 6 Temperature profile for the tar yields at an equivalence ratio of 0.29 AE 0.01 with limestone addition.(DTG) proles of unused limestone together with the limestone/ lime used in the gasication tests at 700, 750 and 800 C, respectively.Proles (Fig. 7a) correspond to inert nitrogen atmosphere, while the proles (Fig. 7b) refer to reactive carbon dioxide atmosphere.DTG prole of unused limestone in nitrogen does not show the presence of Ca(OH) 2 which would dissociate into CaO and H 2 O.The only peak that was observed at temperature between 750-950 C is associated to the calcination reaction CaCO 3 ¼ CaO + CO 2 . 17Limestone/lime samples taken out from gasication reactor show two decomposition regions.The rst peak between 370 and 520 C increases with the gasication temperature (from about 1 to 13% of mass loss).This can be associated to chemical reaction of hydroscopic CaO with the steam 17 and H 2 S. 19,54 The second peak between 750-950 C decreases with the gasication temperature indicating temperature driven conversion of CaCO 3 into CaO.The DTG proles from nitrogen atmosphere demonstrate that limestone from 700 C gasication did not reach notable degree of calcination.The calcination of unused limestone was assumed to be completed when mass reduction of 45.5% was achieved due to CO 2 desorption in nitrogen atmosphere.The difference between the mass loss from samples aer gasication and the mass loss from unused limestone shows a degree of calcination at different temperatures.Smaller is the peak at temperatures between 750-950 C (Fig. 7a), higher is the degree of in situ calcination during gasication.The mass loss of limestone/lime during in situ gasication was about 2% (700 C), 28% (750 C) and 24% (800 C).The DTG proles of carbon dioxide are in agreement with observations above.Unused limestone and limestone from 700 C gasication did not uptake CO 2 since CaCO 3 was predominant chemical form.The region between 900 and 990 C denotes decomposition of CaCO 3 into CaO and CO 2 . 17In contrast, limestone form 750 and 800 C gasication adsorbed CO 2 in the region between 350 and 800 C. According to Bedyk et al. 17 6 reveals that carbon content dropped, while hydrogen and sulphur increased with the gasication temperature.This again is the evidence that CO 2 was released during gasication.At the higher temperature the amount of carbon retained was low.Along with that increasing hydrogen content indicates the presence of Ca(OH) 2 formed from hygroscopic CaO when in contact with steam.Increased sulphur content could be a proof of CaS in the limestone/lime from 700, 750 and 800 C gasication.In combustion atmosphere CaS decomposes into CaO and SO 2 . 55ased on a characterisation tests it can be concluded that the recovered limestone/lime from 750 and 800 C gasication was calcined in situ to some degree, while this cannot be conrmed for the limestone obtained from 700 C gasication test.A fraction of CaO reacted with a gaseous H 2 S forming CaS.Therefore, the catalytic capacity for tar reduction is probably a trade-off between calcinated limestone (i.e.CaO) and CaS.

Conclusions
Yields and composition of tar from the bubbling uidised bed gasication of poultry litter were investigated as a function of temperature, equivalence ratio (ER) and limestone addition to the feedstock.Principally, limestone was added in order to reduce the risk of bed agglomeration.Along with that its capacity for catalytic tar reduction has been investigated.For the range of gasication conditions tested, the following conclusions can be drawn: (1) due to the high content of catalytically active inorganic species and low lignin content, poultry litter generates low yields of total tar (i.e.sum of SPA tar + benzene and toluene measured by micro GC) in the range from 15.7 to 30.7 g tar kg poultry litter (d.a.f.) À1 for the tested temperatures between 700 to 800 C. At this point it needs to be noted that online micro GC measures considerable higher benzene and toluene quantities compared to more commonly employed offline techniques.
(2) The composition of tar from poultry litter gasication is remarkably different from those of conventional lignocellulosic biomass.Nitrogen incorporated in the protein structures of animal feed, excreta and feathers is likely the reason for the signicant amounts of nitrogen containing hydrocarbons detected in tar (3) limestone addition to the poultry litter does not result in a tar reduction effect based on comparing two tests at an ER of 0.30 AE 0.1 and at a temperature 700 C. (4) In situ calcination of limestone does not occur at 700 C, but it does occur at 750 C and 800 C gasication.Limestone/lime adsorbs sulphur contaminant with the increasing temperature.(5) Temperature is an effective measure to reduce heterocyclic tar compounds such as toluene and indene but the amount of PAHs and benzene increases.Atypically constant total tar yield over the temperature range tested is attributed to the prevailing effect of increasing benzene yield with temperature.There is no incontrovertible evidence of limestone/lime catalytic activity with respect to temperature and associated tar reduction.(6) The measurement campaign once again revealed the issue regarding uncertainty of tar data due to the differences in tar denition, sampling conditions, analytical instrumentation and reported units across the scientic community.(7) The ER shows a distinctive effect on tar yield.In the absence of limestone tar yields increase, while the opposite trend was observed in the presence of limestone.However, tar reduction cannot be unambiguously attributed to calcined limestone/lime as ER may play crucial role in tar reduction as well.

Fig. 2
Fig.2Equivalence ratio profile for the tar yields at reactor temperature of 700 C without limestone addition.

Fig. 3
Fig.3Equivalence ratio profile for the tar yields at reactor temperature of 700 C with limestone addition.

Fig. 4
Fig.4Equivalence ratio profile for the tar yields at reactor temperature of 750 C with limestone addition.

Fig. 5
Fig.5Equivalence ratio profile for the tar yields at reactor temperature of 800 C with limestone addition.
in this region Ca(OH) 2 and CaO react with CO 2 giving CaCO 3 which is again calcined at temperature above 830 C.The elemental composition of limestone/lime samples presented in Table

Table 3
27mpares repeatability from three experimental campaigns conducted in 2012 (ECN Pt.I), 2013 (ECN Pt.II) and 2015 (ECN Pt.IIIcurrent study) using the same test rig but different feedstock as well as different post sampling treatment of the SPA samples.SPA tar samples from campaigns ECN Pt.I and ECN Pt.II were extracted on-site immediately aer the sampling, while as mentioned previously, there was a delay before the extraction step of ECN Pt.III samples.SPA tar yields measured during experimental campaigns ECN Pt.I and ECN Pt.II were signicantly higher25compared with those in the current study.Repeatability was estimated using an open access excel spreadsheet 26 model implementing calculations similar to those in the TAPPI standard T 1200, "Interlaboratory Evaluation of Test Methods to Determine TAPPI Repeatability and Reproducibility".27Inshort,coefficient of variance is calculated for SPA replicates of each experimental condition.So derived coefficients of variance are then combined into average covariance which enables expression of repeatability in percentage.Higher values in Table3indicate poor measurement agreement between replicates.

Table 2
6ummary of operating conditions a during the fluidised bed gasification of poultry litter6 a Later in the manuscript these tests are also referred to as campaign ECN Pt.III.

Table 3
Measurement repeatability of total SPA tar, phenol and naphthalene calculated for three experimental campaigns a Limestone samples were analysed by the thermal gravimetric analyser (TGA).The elemental composition of limestone (C, H, N and S) was determined by a Vario EL cube elemental analyser.To reduce the uncertainties involved in measurements, elemental analyses were performed in triplicates.Results are presented on as received basis.Prior to analysis limestone grains were manually separated from the bed content (i.e.char, ash and silica sand).PerkinElmer's Pyris 1 TGA was employed to test the degree of calcination of the limestone.Approximately 20 mg of limestone grain was placed into an alumina pan without a lid and was heated from 30 to 995 C at the rate of 20 C min À1 under a nitrogen purge owrate of 20 cm 3 min À1 .The same temperature program was performed under a carbon dioxide purge owrate of 20 cm 3 min À1 .

Table 4
33entified tar compounds with the retention time and classification according to Milne et al.33 Benzo[a]anthracene 45.85Tertiary-PAH Fig.1Nitrogen-containing compounds found in poultry litter tar.

Table 5
Yields a of permanent gases during fluidised bed gasification of poultry litter a Values calculated based on data from ref.6.

Table 6
Elemental composition of limestone/lime samples expressed in wt% on as received basis