NOX formation and reduction in fluidized bed combustion of biomass

Abstract

Biomass has received growing interest in heat and power production due to it being a renewable and CO2 neutral alternative to coal. A promising technology for the utilization of biomass is fluidized bed combustion, offering fuel flexibility and high combustion efficiency. However, the nitrogen content in different biomass varies considerably and can reach up to several weight percent, possibly higher than coal. This may lead to increased emissions of nitrogen oxides, i.e. NOX (NO and NO2) and N2O, which have a detrimental impact on the environment, as NOX lead to the formation of acid rain and photochemical smog, while N2O is a greenhouse gas and ozone depleter. In combination with the tightening regulations on NOX emissions from solid fuel combustion, the understanding of the mechanism of NOX and N2O emissions from fluidized bed biomass combustion and the development of effective countermeasures are of essential importance. The formation and reduction of NOX and N2O in fluidized bed combustion of biomass has been investigated in this PhD project. Continuous combustion experiments were performed in a lab scale fluidized bed reactor to elucidate the influence of co-combustion, fuel properties (nitrogen and ash forming element content), and operating conditions (air staging, temperature, and gas phase composition) on NOX emissions. Based on the results, an additive-based technique for the simultaneous reduction of NOX emissions and bed agglomeration was proposed and examined. In addition, the formation and reduction of NO and N2O during biomass/waste char combustion were studied in a lab scale fixed bed reactor at conditions resembling that of fluidized bed combustion. Fluidized bed mono- and co-combustion of biomass were conducted at air staged and un-staged conditions, from which effluent and local gas composition data were obtained. The investigated fuels included pine wood, beech wood, straw, sunflower husk, sewage sludge, and sunflower seed. The results reveal that from fluidized bed mono-combustion of biomass under un-staged conditions, the fuel bound nitrogen (fuel-N) to NO conversion decreased with an increase in the fuel-N content. This may be related to the larger release of nitrogen species such as NH3 with the volatiles thereby facilitating thermal DeNOX reactions. At un-staged conditions, a synergy effect was observed during straw-sewage sludge and straw-sunflower seed co-combustion, while straw-sunflower husk co-combustion was additive. The interaction during straw-sewage sludge co-combustion was elucidated, showing that at a low ash content in the bed (2-25 g), the NH3 released from sewage sludge facilitated thermal DeNOX reactions leading to a low NO emission. As more ash accumulated in the bed, the oxidation of NH3 catalyzed by sewage sludge ash increased the NO emission. The catalytic effect of sewage sludge ash on NO forming reactions during straw combustion increased with a lower ash preparation temperature and a better mixing of the ash with straw. In straw-sunflower seed combustion, the effluent NO concentration was approximated well by the weighted average of the individual fuels, while the NH3 and NO concentrations above the bed were significantly higher than the weighted values. Air staging proved to be an efficient method for minimizing NO emission frommono- and co-combustion. In addition, the influence of sewage sludge ash on NO forming oxidation reactions was less pronounced at air staged conditions. The influence of fuel properties and operating conditions on NOX emissions was investigated during fluidized bed combustion. Batch fluidized bed combustion experiments of pine wood, beech wood, straw, sunflower husk, sewage sludge, and sunflower seed revealed that similar to continuous combustion, the fuel-N to NO conversion decreased with an increase in fuel-N content. The volatile-NO was generally the main contributor to the total NO, while in the case of beech wood, up to 55% of the NO was released during char combustion. Continuous fluidized bed combustion of pine wood, straw, washed straw, and potassium (KCl, K2CO3, and KOH) doped pine wood and washed straw were conducted. Washing of straw resulted in a higher fuel-N to NO conversion and lower CO emission. All three K-compounds increased CO emission when doped to washed straw. KCl and KOH further decreased the fuel-N to NO conversion, while K2CO3 had a negligible influence. For pine wood, KCl doping increased CO emission without affecting fuel-N to NO conversion, while the other K-compounds did not change the measurements. The increased CO and decreased NO emissions by KCl addition were more pronounced at air staged (primary to total air ratio (λ1/λ) of 0.04) compared to un-staged (λ1/λ = 1) conditions. This may be attributed to the radical recombination property of KCl. In addition, the influence of air staging was investigated during continuous straw and sunflower husk combustion, indicating a minimum in NO emission and fuel-N to NO conversion for λ1/λ values between 0.5 and 0.75. Moreover, the fuel-N to NO conversion increased slightly with increasing temperature for straw combustion, while the fuel-N to NO conversion for KCl-doped washed straw combustion increased and levelled off with temperature. This was related to the inverse trend in the effluent CO concentration. Furthermore, the fuel-N to NO conversion and CO emission were unaffected by replacing the bed partially with kaolin and CaO, or completely with CaO at λ1/λ = 1, or by changing the primary gas from N2 to CO2 at λ1/λ = 0 and 0.04 during straw combustion The use of additives for the simultaneous reduction of NOX emissions and bed agglomeration during fluidized bed combustion of straw was investigated. Other parameters, such as fuel type, additive particle size and introduction method, and air staging, were additionally examined. During straw combustion, kaolin, CaO, and MgCO3 prevented defluidization without changing the fuel-N to NO conversion, while urea decreased the fuel-N to NO conversion without affecting the defluidization tendency. To reduce NOX, the NH-functionality was a necessity. NH4MgPO4 and AlNH4(SO4)2 prevented defluidization while reducing the fuel-N to NO conversion by 40% in straw and 30% in sunflower husk combustion. Local gas composition measurements revealed that the NH-based additives released NH3 above the fuel inlet and/or the bubbling bed. Some additives (urea, (NH4)2SO4, NH4MgPO4, and AlNH4(SO4)2) favored the reduction of NO by thermal DeNOX, while additives that contained Fe ((NH4)2Fe(SO4)2, NH4Fe(SO4)2, and (NH4)3[Fe(C2O4)3]) or facilitated bed agglomeration, i.e. induced poor mixing, ((NH4)2HPO4) increased NO emission near the bed and led to negligible differences in the effluent NO measurements relative to raw straw combustion. When premixing additive and fuel, the larger particles of (NH4)2SO4 dropped to the bed, increasing the NH3 release above the fluidized bed and prolonging the time for defluidization. In comparison, no significant differences in the fuel-N to NO conversion and defluidization tendency were observed when batch adding pellets compared to premixing, possibly related to differences in releasemechanism and location. In addition, air staging (λ1/λ = 0.5) resulted in a 40% reduction of the fuel-N to NO conversion during straw combustion, while the use of NH4MgPO4 and AlNH4(SO4)2 under air staged conditions slightly increased the fuel-N to NO conversion. This was related to the increased oxidation of NH3 in the secondary air jet. The accumulation of sunflower husk ash increased NO and decreased NH3 concentrations above the bed. In addition, the temporal variations were dampened at higher ash content in the bed. This was attributed to the incipient defluidization leading to poorer mixing and the catalytic effect of ash on the nitrogen chemistry. The formation of NO and N2O during raw and demineralized biomass (pine wood, straw, waste wood, bran, sunflower seed, and dried distillers grains with solubles) char combustion was investigated in a fixed bed reactor. The results reveal that the conversion of char-N to NO decreased with an increase in char-N content. As this trend did not correlate with the NO reduction reactivity of the chars, a contributing factor may be the higher yield of N2O from high-N chars. The presence of ash forming elements largely dominated the NO reduction reactivity, showing a reasonable correlation between the NO reduction reactivity and the (K+Ca)/C molar ratio for the chars with a low P content. Chars with high char-P did not follow the correlation, possibly due to the formation of less catalytically active species such as KPO3. The NO reduction reactivity of the raw and demineralized chars did not correlate with the initial porosity, surface area, char-N content, or nitrogen and oxygen functionality or content. The surface of the chars were enriched with (-CN) and/or (-CO) species during NO treatment. The accumulated (-CN) species may exhibit a different reactivity towards NO reduction compared to inherent (-CN) sites. The conversion of char-N during sewage sludge, refuse derived fuel (RDF), and straw char combustion, and the reactivity of these chars towards NO were investigated in a fixed bed reactor. The reactivity of the sewage sludge and RDF chars were an order of magnitude and six-fold higher than that of straw, respectively. This was attributed to the larger Ca and Fe content in the waste based chars. The initial reactivity of the employed chars correlated well with the (K+Fe+Ca)/C molar ratio. As fractional order or random pore model expressions did not yield any obvious advantages, a simple concentration-averaged, first order reaction expression was used to describe the char-NO reaction. The simplified expression was used in a one-dimensional, transient, heterogeneous reactor model, which predicted reasonably well the conversion of char-N to NO during combustion. The reactivities of the waste chars exceeded those of selected coal and biomass char found in literature. This work provides an improved understanding of the conversion of nitrogen during fluidized combustion of biomass. The observed trends may be extrapolated to practical fluidized bed combustion systems for improvement of design and operation to minimize NOX emissions