Search

We assessed the feasibility of adjuvant S-1 and oxaliplatin following neoadjuvant chemoradiotherapy (nCRT) and esophagectomy. Patients treated with nCRT (paclitaxel, carboplatin) and esophagectomy received six 21-day cycles with oxaliplatin (130 mg/m2) on day 1 and S-1 (25 mg/m2 twice daily) on days 1–14. The primary endpoint was feasibility, defined as ≥50% completing treatment. We performed exploratory propensity-score matching to compare survival, ERCC1 and Thymidylate Synthase (TS) immunohistochemistry analyses, proteomics biomarker discovery and 5-FU pharmacokinetic analyses. Forty patients were enrolled and 48% completed all adjuvant cycles. Median dose intensity was 98% for S-1 and 62% for oxaliplatin. The main reason for early discontinuation was toxicity (67%). The median recurrence-free and overall survival were 28.3 months and 40.8 months, respectively (median follow-up 29.1 months). Survival was not significantly prolonged compared to a matched cohort (p = 0.09). Patients with ERCC1 negative tumor expression had significantly better survival compared to ERCC1 positivity (p = 0.01). Our protein signature model was predictive of survival [p = 0.04, Area under the curve (AUC) 0.80]. Moreover, 5-FU pharmacokinetics significantly correlated with treatment-related toxicity. To conclude, six cycles adjuvant S-1 and oxaliplatin were not feasible in pretreated esophageal adenocarcinoma. Although the question remains whether additional treatment with chemotherapy should be provided in the adjuvant setting, subgroups such as patients with ERCC1 negativity could potentially benefit from adjuvant SOX based on our exploratory biomarker research.

The phytoremediation potential of three bamboo species, i.e. Bambusa merilliana, Bambusa blumeana, and Dendrocalamus asper, were evaluated for their total Cu uptake ability in hydroponics. Dendrocalamus asper proved to be the most efficient in terms of Cu phytoremediating potential with a constant positive uptake of 80 μM in a contaminated substrate and a bioconcentration factor of 50.57. Copper accumulation was found to concentrate the most in the roots compared to the amount translocated in the shoots. Analysis of the Cu uptake of D. asper roots at varying concentrations of Cu contamination (40, 80 and 120 μM) allowed for the fitting of the kinetics of Cu uptake and removal with existing kinetic models. The rate of copper removal per mass of plant was the best for the 0th order model in the 80 μM solution with an R2 of 0.954 and rate constant of 3.136 mg-kg−1d−1. The accumulation of Cu within the roots on day 7 (7d) followed the Michaelis–Menten model with an R2 of 0.970. The Michaelis–Menten constant (KM) was 4.87 mg/L and maximum accumulation velocity (Vmax) was 66.26 mg Cu-kg−1-day−1. Results suggest that D. asper is a potential hyperaccumulator plant that can be used in clean-up of domestic and industrial wastes present along river systems in the Philippines.

Copper is the third most utilized metal and is a versatile resource with multiple beneficial uses, but it may also become toxic to aquatic life in excess amount. Thus, there is a need to develop methods to reduce the copper contamination in the environment, particularly in bodies of water. Phytoremediation using Dendrocalamus asper may offer an environment-benign and potentially effective method for copper removal though its effectiveness may take several years to materialize for this technology to become cost-effective. By growing D. asper in synthesized contaminated water and analyzing the change in the copper content of the substrate via atomic absorption spectrophotometry, the removal was found to be optimal at 20 ppm Cu and pH 5. The rate of removal was found to have an order of 2.71 and a kinetic constant of 0.0013 ppm−1.71 day−1. With this, it may be possible to estimate the treatment length of phytoremediation given an initial level of copper contamination and a target concentration.

Coal-fired utility boilers are the single largest anthropogenic source of mercury emissions. Mercury is a naturally occurring trace element in coal and, when combusted, may exist in three different forms: Hg0, Hg2+, or Hg particulate. During oxy-fuel combustion, impurity concentrations, such as SOx, NOx, and Hg, can be up to 4 times higher than concentrations in air combustion. An increased mercury concentration is of concern because mercury is known to attack aluminum heat exchangers required in the compression of CO2. As a result of the elevated concentrations during oxy-fuel conditions, interactions of Hg and SOx were investigated in this study to verify if there is any competition between SOx and Hg. The effect of Hg, SOx, H2O, and temperature on the native capture of Hg by fly ash was assessed using a quartz flow reactor packed with fly ash to simulate a bag filter. Doubling Hg in the system from 5 to 10 μg/Nm3 doubled the amount of Hg captured in the fly ash from 1.6 to 2.8% and increased the amount...

Nitrogen doped Carbon Nanotube with divacancy (4ND-CNxNT) that is decorated with Scandium and Titanium as potential hydrogen storage medium using the pseudo potential density functional method was investigated. Highly localized states near the Fermi level, which are derived from the nitrogen defects, contribute to strong Sc and Ti bindings, which prevent metal aggregation and improve the material stability. A detailed Comparison of the Hydrogen adsorption capability with promising system-weight efficiency of Sc over Ti was elucidated when functionalized with 4ND-CNxNT. Finally, the (Sc/4ND)10-CNxCNT composite material has a thermodynamically favorable adsorption and consecutive adsorption energy for ideal reversible adsorption and desorption of hydrogen at room temperature such that it can hold at least 5.8 wt% hydrogen molecules at the LDA and GGA level.

The Callide Oxy-fuel Project is the world's largest operating oxy-fuel plant. This work details an experimental test campaign at the Callide Oxy-fuel Project monitoring mercury and SO3 levels exiting the fabric filter during transitions between air and oxy firing conditions. The measurements were taken using two custom built probes; the first allowing combined collection of SO3 and mercury over short time intervals; the second allowing on-line measurements of Hgtotal and Hg0 with SOx removal. Total mercury emissions in oxy-firing measured a maximum of 6–7 μg/m3 of which 89% was in oxidised form (Hg2+). The use of low NOx burners had an overriding influence on the mercury measurements reducing the total mercury levels to 0.13 and 0.15 μg/m3 (air, oxy respectively) with no Hg2+ being measured. The SO3 concentrations were also lower than expected, estimated at ∼0.5–0.8 ppm (based on a practical estimate of 1% conversion of SO2). Overall mercury capture in either operating mode was estimated at 92–93% for the existing burners and 98–99% with the low NOx burners used (being 2 of the 4 burners operating). Total SOx captured from the flue gas was 16% in oxy-mode and 19% in air firing. These findings suggest that operational conditions have a primary impact on capture of Hg and SOx during transitions with a secondary impact of firing mode (i.e. air or oxy).

Copper is a commonly used metal in construction, engineering, agriculture and water treatment. Consequently, increased copper concentrations resulting in adverse environmental effects is inevitable. Phytoremediation using Dendrocalamus asper or Philippine giant bamboo (PGB) is a viable option for treatment of copper-contaminated media, but their copper uptake potential remains largely unexplored. As such, the copper uptake of PGB was evaluated under varying environmental conditions, namely initial copper concentration, water hardness and pH.Six-month old propagules were planted in artificially contaminated water in order to determine the copper uptake after 16 days of treatment. Using a Box-Behnken design of experiment, it was found that both initial copper concentration and pH have significant and proportional effects on copper uptake. However, due to possible speciation and/or competition, the optimum copper uptake occurred at 20 ppm Cu and pH 5 (as opposed to 7). A mathematical equation, bearing an R2 = 0.7097, was constructed as a possible model for copper uptake of PGB to understand when PGB phytoremediation is most effective. At copper concentrations lower than 3.81 ppm, higher pH is beneficial to copper uptake and vice versa. Overall, phytoremediation using D. asper or PGB is effective especially at low pH and elevated copper concentrations.

The Callide Oxyfuel Project (COP) is the world's largest operating oxy-fuel plant and has successfully demonstrated the retrofitting capability for oxy-fuel to capture CO2. The compression of the raw flue gas at the COP has also demonstrated that NOx and Hg can also be removed as part of the liquid condensate. This work presents field tests conducted on the condensates to determine the stability of captured NOx and Hg species when depressurised. These tests involved sampling liquid condensate directly into a customised aeration vessel and measuring the evolved gases over an 8–12 h period. The low-pressure condensate (∼4 bar) showed that 3–18% of captured NOx species and 0.5–1.2% of the Hg were volatile, while the high-pressure condensate (24 bar) re-emitted 2–68% of captured NOx and 0.05–12.5% of the captured Hg. These tests showed that volatile Hg was related to volatile NOx and that this volatility of condensates changed with time as the compression plant operated from start-up. Equilibrium calculations of HNO2 in the gas and liquid phases supported the volatility measurements, suggesting that the rate of oxidation of HNO2 to HNO3 in the condensed phase is slow. Overall, the conditions which favoured NOx stability in the condensates, namely longer residence and higher pressure also favoured Hg stability. This work has shown that emissions from an oxy-fuel compression plant must include those emanating from depressurised condensates and suggest that the re-emitted species may not the same as in typical combustion flue gas, but the result of higher-pressure conversion.

The reaction of SO2 with fly ash in the presence of O2 and H2O involves a series of reactions that lead to the formation of SO3 and eventually H2SO4. Homogeneous experiments were conducted to evaluate the effects of the procedural variables, i.e., temperature, gas concentrations, and residence time, on the post-combustion conversion of SO2 to SO3. The results were compared to existing global kinetics and found to be dependent upon SO2, O2, residence time, and temperature and independent of H2O content. For a residence time of 1 s, temperatures of about 900 °C are needed to have an observable conversion of SO2 to SO3. Literature suggested that the conversion of SO2 to SO3 is dependent upon the iron oxide content of the fly ash. Experiments using three different fly ash samples from Australian sub-bituminous coals were used to investigate the catalytic effects of fly ash on SO2 conversion to SO3 at a temperature range of 400–1000 °C. It was observed that fly ash acts as a catalyst in the formation of SO3, w...

Ash produced during oxy-fuel combustion is expected to differ from ash produced during air combustion because of the higher CO2 and SO2 atmospheres in which it is generated. For a quantitative understanding of the sulfation behavior of fly ash in oxy-fuel combustion, fly ash from three commercial Australian sub-bituminous coals was tested and decomposed under an inert atmosphere. Thermal evolved gas analysis was completed for ash produced in both air and oxy-fuel environments. Pure salts were also tested under the same conditions to allow for identification of the species in the ash that capture sulfur, along with thermodynamic modeling using FactSage 6.3. Sulfur evolved during the decomposition of air and oxy-fuel fly ash was compared to the total sulfur in the ash to close the sulfur balance. Both total sulfur captured by the ash and sulfur evolved during decomposition were higher for oxy-fuel fly ash than their air counterparts. Correlations of capture with ash chemistry are presented.

This paper presents results of experiments performed at a 20 kW once-through combustion rig of the Institute of Combustion and Power Plant Technology (IFK) of the University of Stuttgart. A methodology to investigate oxy-fuel process configurations was used in which impurities were injected to the oxidant gas of the once-through reactor to simulate different extents of oxy-fuel recycle gas treatment. Three Australian coals, which had previously been tested in the Aioi furnace of IHI in Japan, were used in the experiments. A comprehensive set of total (Hgtot), elemental (Hg0), and oxidized (Hg2+) mercury concentrations was measured for various air and oxy-fuel combustion conditions. These data enable an evaluation of process parameters that influence the Hg emissions of an oxy-fuel combustion process. A theoretical mass balance between Hg fed to the process (fuel and Hg0 injection) and Hg measured before the filter matched well, indicating that no mercury was captured by fly ash at high temperatures. The c...

Tungsten contamination can significantly impact the performance of future fusion reactors via fuel dilution and radiation in the plasma centre. Therefore, understanding tungsten sputtering and divertor retention is essential for optimising the core performance in tokamaks operating with tungsten. This study investigates these issues numerically in JET high-triangularity, type-I ELMy H-mode plasmas with the Monte-Carlo code DIVIMP used on background plasmas dynamically evolved with the multi-fluid code EDGE2D/EIRENE. During the ELM, the simulations show target temperatures in excess of a few 100 eV, causing two orders of magnitude increase in the tungsten core contamination rate. During the ELM recovery, target densities of five times the pre-ELM values are obtained at the low field side, which strongly enhance divertor retention via increased friction with the main ions. Therefore, the core contamination is determined here dominantly by the intra-ELM divertor plasma characteristics.

The edge transport code EDGE2D-EIRENE has been used to model the inter-ELM phase of H-mode plasmas on JET. A reference plasma was simulated and matched to outer mid-plane profiles and to the heat load distribution at the outer target. While the simulated parallel saturation current was within experimental constraints from Langmuir probes, the target electron temperature was slightly overestimated (factor ∼1.2) compared to the upper experimental range. A deuterium scan was carried out from the simulation of this (lowest fuelled) reference plasma and compared to the same experimental scan. In all simulations, carbon radiation was found to be the dominant power loss channel at the outer target.

Wastewater and sludge are potential resource of phosphorus (P) for fertilizer production. One method of recovering phosphorus is via chemical precipitation. In the study, phosphorus was recovered from wastewater and sludge. First, hydrolysis was carried out to release the phosphorus in the sludge by the addition of 1.0M acid (sulfuric acid) or base (sodium hydroxide) solution mixed for three hours at 200 rpm. The hydrolyzed sludge was filtered, and the pH of the solution was adjusted to 9.0. Precipitation for both wastewater and hydrolyzed sludge solution was carried out using magnesium chloride hexahydrate (MgCl2•6H2O) and ammonium chloride (NH4Cl). The mixture was stirred for an hour for crystallization. Precipitates were allowed to settle for 24 hours before it was filtered and dried in an oven at 55-58oC for 24 hours. The dried sample was grinded and characterized using Fourier transform infrared spectroscopy (FTIR), x-ray fluorenscence (XRF), and scanning electron microscope with energy-dispersive x-ray spectroscopy(SEM-EDX).

In high confinement mode, highly shaped plasmas with edge localized modes in JET, and for heating power of 15–17 MW, the edge fluid code EDGE2D-EIRENE predicts transition to detachment assisted by nitrogen at the low field side (LFS) target when more than 50% of the power crossing the separatrix between ELMs is radiated in the divertor chamber, i.e. ~4 MW. This is observed both in the ITER-like wall (JET-ILW) and in the carbon wall (JET-C) configurations and is consistent with experimental observations within their uncertainty. In these conditions, peak heat fluxes below 1 MW m−2 are measured at the LFS target and predicted for both wall configurations. When the JET-C configuration is replaced with the JET-ILW, a factor of two reduction in the divertor radiated power and 25–50% increase in the peak and total power deposited to the LFS divertor plate is predicted by EDGE2D-EIRENE for unseeded plasmas similar to experimental observations. At the detachment threshold, EDGE2D-EIRENE shows that nitrogen radiates more than 80% of the total divertor radiation in JET-ILW with beryllium contributing less than a few %. With JET-C, nitrogen radiates more than 70% with carbon providing less than 20% of the total radiation. Therefore, the lower intrinsic divertor radiation with JET-ILW is compensated by stronger nitrogen radiation contribution in simulations leading to detachment at similar total divertor radiation fractions. 20–100% higher deuterium molecular fraction in the divertor recycling fluxes is predicted with light JET-C materials when compared to heavy tungsten. EDGE2D-EIRENE simulations indicate that the stronger molecular contribution can reduce the divertor peak power deposition in high recycling conditions by 10–20% due to enhanced power dissipation by molecular interaction.

Migration of beryllium into the divertor and deposition on tungsten in the final phase of the first ITER-like-wall campaign of JET are modelled with the 3D Monte Carlo impurity transport code ERO. The simulation covers the inner wall and the inner divertor. To generate the plasma background for Monte Carlo tracing of impurity particles, we use the EDGE2D/EIRENE code set. At the relevant regions of the wall, the estimated plasma conditions vary around Te≈5eV and ne≈2×1017m-3 (far-scrape-off layer; more than 10 cm away from the LCFS). We calculate impurity distributions in the plasma using the main chamber source as a free parameter in modelling and attempt to reproduce inter-ELM spectroscopic Be II line (527 nm) profiles at the divertor. The present model reproduces the level of emission close to the inner wall, but further work is needed to match also the measured emission peak values and ultimately link the modelled poloidal net deposition profiles of beryllium to post mortem data.

Measurements of the plasma conditions at the low field side target plate in JET ITER-like wall ohmic and low confinement mode plasmas show minor differences in divertor plasma configurations with horizontally and vertically inclined targets. Both the reduction of the electron temperature in the vicinity of the strike points and the rollover of the ion current to the plates follow the same functional dependence on the density at the low field side midplane. Configurations with vertically inclined target plates, however, produce twice as high sub-divertor pressures for the same upstream density. Simulations with the EDGE2D-EIRENE code package predict significantly lower plasma temperatures at the low field side target in vertical than in horizontal target configurations. Including cross-field drifts and imposing a pumping by-pass leak at the low-field side plate can still not recover the experimental observations.

The behaviour of tungsten in the core of hybrid scenario plasmas in JET with the ITER-like wall is analysed and modelled with a combination of neoclassical and gyrokinetic codes. In these discharges, good confinement conditions can be maintained only for the first 2?3?s of the high power phase. Later W accumulation is regularly observed, often accompanied by the onset of magneto-hydrodynamical activity, in particular neoclassical tearing modes (NTMs), both of which have detrimental effects on the global energy confinement. The dynamics of the accumulation process is examined, taking into consideration the concurrent evolution of the background plasma profiles, and the possible onset of NTMs. Two time slices of a representative discharge, before and during the accumulation process, are analysed with two independent methods, in order to reconstruct the W density distribution over the poloidal cross-section. The same time slices are modelled, computing both neoclassical and turbulent transport components and consistently including the impact of centrifugal effects, which can be significant in these plasmas, and strongly enhance W neoclassical transport. The modelling closely reproduces the observations and identifies inward neoclassical convection due to the density peaking of the bulk plasma in the central region as the main cause of the accumulation. The change in W neoclassical convection is directly produced by the transient behaviour of the main plasma density profile, which is hollow in the central region in the initial part of the high power phase of the discharge, but which develops a significant density peaking very close to the magnetic axis in the later phase. The analysis of a large set of discharges provides clear indications that this effect is generic in this scenario. The unfavourable impact of the onset of NTMs on the W behaviour, observed in several discharges, is suggested to be a consequence of a detrimental combination of the effects of neoclassical transport and of the appearance of an island.

JET underwent a transformation from a full carbon-dominated tokamak to a fully metallic device with beryllium in the main chamber and a tungsten divertor. This material combination is foreseen for the activated phase of ITER. The ITER-Like Wall (ILW) experiment at JET shall demonstrate the plasma compatibility with metallic walls and the reduction in fuel retention. We report on a set of experiments (Ip = 2.0 MA, Bt = 2.0-2.4 T, δ = 0.2-0.4) in different confinement and plasma conditions with global gas balance analysis demonstrating a strong reduction in the long-term retention rate by more than a factor of 10 with respect to carbon-wall reference discharges. All experiments are executed in a series of identical plasma discharges in order to achieve maximum plasma duration until the analysis limit of the active gas handling system is reached. The composition analysis shows high purity of the recovered gas, typically 99% D. For typical L-mode discharges (Paux = 0.5 MW), type III (Paux = 5.0 MW) and type-I ELMy H-mode plasmas (Paux = 12.0 MW) a drop of the deuterium retention rate normalized to the operational time in divertor configuration is measured from 1.27 × 1021, 1.37 × 1021 and 1.97 × 1021 D s-1 down to 4.8 × 1019, 7.2 × 1019 and 16 × 1019 D s-1, respectively. The dynamic retention increases in the limiter phase in comparison with carbon-fibre composite, but also the outgassing after the discharge has risen in the same manner and overcompensates this transient retention. Overall an upper limit of the long-term retention rate of 1.5 × 1020 D s-1 is obtained with the ILW. The observed reduction by one order of magnitude confirms the expected predictions concerning the plasma-facing material change in ITER and is in line with identification of fuel co-deposition with Be as the main mechanism for the residual long-term retention. The reduction widens the operational space without active cleaning in the DT phase in comparison with a full carbon device. © 2013 IAEA, Vienna.

Experiments on JET with a carbon-fibre composite wall have explored the reduction of steady-state power load in an ELMy H-mode scenario at high Greenwald fraction 0.8, constant power and close to the L to H transition. This paper reports a systematic study of power load reduction due to the effect of fuelling in combination with seeding over a wide range of pedestal density ((48)×10 19m 3) with detailed documentation of divertor, pedestal and main plasma conditions, as well as a comparative study of two extrinsic impurity nitrogen and neon. It also reports the impact of steady-state power load reduction on the overall plasma behaviour, as well as possible control parameters to increase fuel purity. Conditions from attached to fully detached divertor were obtained during this study. These experiments provide reference plasmas for comparison with a future JET Be first wall and an all W divertor where the power load reduction is mandatory for operation.

During detachment experiments of JET in 2007 (Brezinsek et al., 2009 [1]), 1 × 1022 molecules of 12CD4 were injected at a rate of about 5 × 1020 s−1 into L-mode plasma. The injection was done at the outer strike point in the centre of the horizontal target, just before opening the machine. The deposited layers were analyzed for deuterium post mortem after the removal of tiles. The heaviest local D deposition density is found immediately upstream of the gas inlet, but downstream the deposition is found over a larger area. In total, 3.7 × 1020 deuterium atoms were found locally deposited; if the mean D/C in the deposits is 0.4, then about 10% of the injected carbon was locally deposited. Transport and local deposition of the injected carbon was modeled with the 3D Monte Carlo impurity transport code ERO (Kirschner et al., 2000 [2]). The plasma background was generated with the onion-skin solver of the DIVIMP code (Stangeby and Elder, 1992 [3]). Locally deposited fraction is reproduced with ERO but the distribution not satisfactorily. In particular the poloidal transport of carbon is several times weaker than measured.

Impurity seeding of recycling impurities such as neon or argon, or non-recycling impurities such as nitrogen, is an essential element of power exhaust in future fusion reactors, including ITER. Aside from increasing the fraction of power radiated onto the main chamber, thus reducing the fraction carried by the plasma to the divertor, it also provides a viable method of ELM size reduction (mitigation) in ELMy H-mode plasmas. However, impurity seeding raises the risk of impurity accumulation in the core of the plasma, which degrades reactor performance and can lead to thermal collapse of the discharge. It is therefore imperative to understand the mechanisms of impurity transport in the boundary plasma, including the close field line (edge) region and the open field line (scrape-off layer or SOL) regions, and specifically, to quantify impurity influx across the magnetic separatrix and the 'screening' effect of the SOL. In this paper, we employ the two-dimensional transport code EDGE2D/NIMBUS to investigate the influence of (anomalous) parallel SOL flows, observed experimentally on JET and other tokamaks, on the 'screening' of a recycling impurity (neon). The parallel SOL flow is adjusted by external momentum injection, such that the Mach number at the top of the vessel varies from ~0 to ~0.5. Simulations indicate that the deuterium density in the plasma core increases by ~50% over this range, in agreement with experimental findings (comparing normal and reversed field discharges). Notably, this density increase is associated with (caused by) plasma compression on the inboard side of the torus caused by the parallel SOL flow, thus leading to inverted radial density gradient and a recirculating flow pattern (convective cell) linking the inboard and outboard SOL regions. Simulations predict a 5% decrease in impurity (neon) concentration in the plasma core when the deuterium flow at the top of the vessel increases to Mach number, , with little change for higher Mach numbers.

This paper presents results on experiments carried out at a 20kW combustion rig simulating different extents of oxy-fuel recycle gas cleaning by impurities injection to the oxidant gas of the once-through combustion reactor. A comprehensive set of total (Hgtot), elemental (Hg0) and oxidized (Hg2+) mercury as well as SO3 concentrations was obtained before and after the combustion rig's baghouse filter for in total 14 air and oxy-fuel experiments with 3 Australian coals. Based on this data, an assessment in respect to Hg oxidation, SO2/SO3 conversion and Hg and SO3 capture on the test rig's filter was performed. The air and the oxy-fuel experiments with different extents of recycle gas cleaning, revealed differences in the Hg and SO3 formation and capture behavior: The Hg2+/Hgtot ratios in the flue gas are higher during oxy-fuel combustion compared to air-firing. This effect is even more pronounced at the filter outlet, after flue gas has passed through the filter ash. In some experiments, even a net oxidation of Hg0 entering the filter to Hg2+ was observed. The Hg capture by ash in the baghouse filter has been found to reduce the Hg emissions considerably. However, the Hg capture was altered by the different oxy-fuel recycle configurations, leading to decreased Hg capture efficiencies on the filter for one of the coals. A coal-specific trend of increased SO2/SO3 conversion ratios with increased flue gas SO2 levels was observed that could be related to the ash composition of the three different coals. This and the higher SO2 concentrations in the flue gas lead to considerably higher SO3 levels in oxy-fuel combustion with SO2 recycling. During the experiments, also a considerable capture of SO3 in the baghouse filter was observed (up to 80% under air- and up to 66% under oxy-fired conditions). A reduction of the SO3 capture on the filter under oxy-fuel conditions may be related to the higher SO3 levels in this process.