Umeå University's logo

Solid biomass oxy-fuel combustion with external recirculation was demonstrated in a 100-kW atmospheric, down-fired combustor operated close to stoichiometry. The external recirculation setup comprised a particulate filter, a condenser, a fan and O₂ addition before the burner inlet. A theoretical description of the recirculation process is presented and validated. Two fuels, softwood (SW) and forest residues (FR), with similar residence times, were compared. Gaseous species, including potassium (K) compounds (atomic K, KOH, and KCl), and gas temperature were quantified in real-time by tunable diode laser absorption spectroscopy (TDLAS) and photofragmentation TDLAS at two locations in the reactor core. Major species (CO₂, H₂O, O₂, and N₂) were also measured at the exhaust. Flue gas particles collected with a low-pressure impactor at the exhaust were analyzed by X-ray powder diffraction and scanning electron microscopy. The average CO₂ purity (dry) was 90 % for SW and 86 % for FR. The NO concentration was higher for FR due to the larger nitrogen content in the fuel. The gaseous K species concentrations were higher for FR than for SW (factor 2–3), but not as high as expected from the difference in fuel K content (factor 7), likely due to the high content of Si and Al in FR. Gas-phase K was significantly lower than predicted by thermodynamic equilibrium calculations (TEC), probably due to K adsorption by soot particles. The fine and coarse particle concentrations were significantly higher for FR than for SW due to the higher ash content of FR. The FR fine mode particles consisted mainly of K₂SO₄ and KCl, in good quantitative agreement with TEC of gas phase condensation. Apatite, Ca₅(PO₄)₃OH, likely formed from vaporized Ca and P, was found in the fine mode in all recirculation cases.

Alkali plays a critical role in most systems using biogenic fuels, yet quantification is difficult and often method-dependent, due to partitioning between particles and gas, chemical conversion, and strong interactions with surrounding surfaces. Building on the limited prior work combining online alkali diagnostics, we demonstrate a combined diagnostic approach using photofragmentation tunable diode laser absorption spectroscopy (PF-TDLAS; in situ, path-averaged, gaseous potassium chloride (KCl) and potassium hydroxide (KOH)) and surface ionization detection (SID; point-sampled total potassium (K), gas and particles) in a well-controlled heated reactor with potassium aerosol seeding. Reactor temperature, atmosphere (air, N2, CO2), and the amount and chemical form of injected potassium (chloride, hydroxide, carbonate or sulfate) were varied, and the data were interpreted with computational fluid dynamics (CFD) and thermodynamic equilibrium calculations. Together, the diagnostics resolve potassium partitioning and speciation and quantify what reaches an extraction point: at low temperature they indicate particulate K (high SID, negligible PF-TDLAS signals); at higher temperatures, they capture the particle-to-gas transition. In this work, the transition occurred around 700 °C rather than ∼550 °C predicted by equilibrium, since the injected jet remained colder locally (CFD). At 900 °C, both diagnostics respond consistently in terms of absolute K concentrations and as a function of alkali seeding, while revealing speciation differences (KCl(g) for KCl seeding; KOH(g) plus minor KCl(g) for KOH seeding). Atmosphere and seed chemistry further modulate the signals: CO2 reduces delivered K-loading and promotes carbonate formation during KOH seeding, suppressing KOH(g) in PF-TDLAS, while maintaining elevated total-K in the SID. This combined framework shows that alkali measurement is intrinsically multi-factor and that PF-TDLAS and SID should be applied synergistically, not interchangeably.

Novelty and significance statement

Building on the only earlier combined optical/SID alkali study known to us [1], we present a coordinated, quantitative PF-TDLAS/SID deployment in a more well-controlled high-temperature reactor, substantially expanding the scope by varying temperature, atmosphere, and the amount of injected potassium, and by adding KOH, K2CO3, and K2SO4 seeding. PF-TDLAS provides in situ, path-averaged gas-phase KCl(g) and KOH(g), while SID measures extractive total potassium (gas and particles). The novelty lies in integrating these complementary measurements with CFD-resolved flow and temperature fields and thermodynamic equilibrium calculations to develop a mechanistic interpretation framework for alkali quantification under non-uniform reactor conditions. This framework rationalizes how phase-partitioning, transport/surface losses, spatial non-uniformity, sampling bias, and chemistry-dependent transformations jointly govern the PF-TDLAS and SID signals. The result is a transferable framework for robust alkali quantification with repeatable sub-ppm sensitivity, improving cross-comparability for combustion-relevant systems where alkali partitioning and losses govern fouling, corrosion, emissions, and catalytic effects.

Oxy-fuel biomass combustion can facilitate carbon capture in heat and power plants and enable negative carbon dioxide (CO₂) emissions. We demonstrate oxy-fuel combustion (OFC) of softwood powder in a 100-kW atmospheric down-fired pulverized combustor run at a global oxidizer-fuel equivalence ratio of around 1.25. The simulated oxidizer was varied between oxygen (O₂)/CO₂ mixtures of 23/77, 30/70, 40/60 and 54/46, and artificial air. The concentrations of the main gaseous potassium (K) species: atomic K, potassium hydroxide (KOH) and potassium chloride (KCl), were measured at two positions in the reactor core using photofragmentation tunable diode laser absorption spectroscopy (PF-TDLAS). Major species were quantified by TDLAS in the reactor core and with Fourier transform infrared spectroscopy and mass spectrometry at the exhaust. Flue gas particles were collected at the exhaust employing a low-pressure impactor and analyzed by X-ray powder diffraction and scanning electron microscopy. The measured individual K species concentrations in the reactor core agreed with predictions by thermodynamic equilibrium calculations (TEC) within one order of magnitude and the sum of K in the gas phase agreed within a factor of three for all cases. Atomic K was underpredicted, while the dominating KOH and KCl were slightly overpredicted. The ratios of measured to predicted total K were similar in artificial air and OFC, but the distributions of the individual species differed at the upper reactor position. The gaseous K species and fine particle concentrations in the flue gas were directly proportional to the O₂ content in the oxidizer. The crystalline phase compositions of the coarse mode particles were rich in K- and calcium-containing species. The fine mode particles, which contained most of the K, consisted mainly of K₂SO₄ (94%) and K₃Na(SO₄)₂, which is in excellent agreement with TECs of gas phase condensation. As supported by the solid phase analysis, complete sulfation of K species was achieved for all studied cases. A CO₂ purity (dry) of up to 94% was achieved for OFC.

Potassium (K) plays a key role in biomass combustion as its presence in the gas phase may lead to slagging, agglomeration and corrosion. To facilitate feedstock flexibility, a detailed understanding of the K chemistry during solid fuel conversion is needed. Here, we present photofragmentation laser absorption imaging for simultaneous, quantitative wide-field imaging of gaseous atomic K, potassium hydroxide (KOH) and potassium chloride (KCl) using a high-speed camera. Concentration images of the K species, at levels down to below parts-per-million, were acquired in a physical domain of 3.8-by-16.4 mm with a spatial resolution of 0.13 mm and a temporal resolution of 0.5 s after averaging. The technique was applied for imaging of the K species release from solid coffee husk and wheat straw samples combusted in a laboratory flame. Conversion of KCl salt was investigated as a reference. Significant differences in K species concentration and spatial distribution were observed between the fuels and between the devolatilization and char conversion stages. We also demonstrate spatial mapping of the atomic K fragment recombination rates with 1.2 µs time resolution, which constitutes a step towards reaction kinetics imaging.

In this study, a Eulerian–Lagrangian model is used to study biomass gasification and release of potassium species in a 140 kW atmospheric entrained flow gasifier (EFG). Experimental measurements of water concentration and temperature inside the reactor, together with the gas composition at the gasifier outlet, are used to validate the model. For the first time, a detailed K-release model is used to predict the concentrations of gas-phase K species inside the gasifier, and the results are compared with experimental measurements from an optical port in the EFG. The prediction errors for atomic potassium (K), potassium chloride (KCl), potassium hydroxide (KOH), and total potassium are 1.4%, 9.8%, 5.5%, and 5.7%, respectively, which are within the uncertainty limits of the measurements. The numerical model is used to identify and study the main phenomena that occur in different zones of the gasifier. Five zones are identified in which drying, pyrolysis, combustion, recirculation, and gasification are active. The model was then used to study the transformation and release of different K species from biomass particles. It was found that, for the forest residue fuel that was used in the present study, the organic part of K is released at the shortest residence time, followed by the release of inorganic K at higher residence times. The release of inorganic salts starts by evaporation of KCl and continues by dissociation of K₂CO₃ and K₂SO₄, which forms gas-phase KOH. The major fraction of K is released around the combustion zone (around 0.7–1.3 m downstream of the inlet) due to the high H₂O concentration and temperature. These conditions lead to rapid dissociation of K₂CO₃ and K₂SO₄, which increases the total K concentration from 336 to 510 ppm in the combustion zone. The dissociation of the inorganic salts and KOH formation continues in the gasification zone at a lower rate; hence, the total K concentration slowly increases from 510 ppm at 1.3 m to 561 ppm at the outlet.