Optical investigation of gas-phase KCl/KOH sulfation in post flame conditions

.


Introduction
Biomass, including energy crops [1], agricultural products [2] and industrial biomass waste [3], is considered as a sustainable and carbon neutral energy source.It is commonly combusted in boilers to provide heat and/or power.However, biomass such as agricultural residues may contain relatively high concentrations of potassium, chlorine and sulfur.During combustion, these elements lead to the formation of potassium chloride (KCl), potassium hydroxide (KOH), hydrogen chloride (HCl) and sulfur dioxide (SO 2 ), causing problems of slagging and fouling, corrosion, and emissions of harmful gases and aerosols [2].The presence of KCl in the flue gas may cause severe ash deposition and corrosion on the super-heater tubes [4].Sulfation of KCl to K 2 SO 4 by reaction with sulfur-containing additives [5][6][7] has been suggested as a method to mitigate deposition and corrosion processes, since K 2 SO 4 is less corrosive and has a higher melting point than KCl.
More recently, Hindiyarti et al. [11] proposed a number of alternative sulfation routes to explain the observed sulfation rates at lower temperatures.One pathway involved the direct reaction between KOH and SO 2 to form KHSO 3 , with oxidation of sulfite to sulfate as the ratelimiting step, instead of the oxidation of SO 2 to SO 3 : https://doi.org/10.1016/j.fuel.2018 E-mail address: zhongshan.li@forbrf.lth.se(Z.Li).(R6) Another route of K 2 SO 4 formation can be initiated by the reaction of atomic K with SO 2 , forming KOSO, which is then oxidized to KOSO 3 and further to KHSO 4 .These suggested pathways have not been confirmed experimentally, even though they are consistent with results reported by Kassman et al. [13] and Ekvall et al. [15,16].
To evaluate the detailed reaction mechanism, Li et al. [14] studied the formation of K 2 SO 4 from gas-phase KCl and SO 2 in a homogeneous high temperature environment produced by a laminar burner where neither wall-based reactions nor other heterogeneous reactions could contribute to the sulfation processes.The study supported the importance of the reaction sequence R1, R3 and R4 for the sulfation of KCl.
Previous efforts have mostly focused on the sulfation of KCl.However, in combustion of low-chlorine woody biomass KOH is typically formed in larger concentrations than KCl.Despite the importance, knowledge on sulfation of KOH is still limited.In the present work, sulfation of both KOH and KCl was studied in a specially designed counter-flow reactor.Gas-phase KOH and KCl were provided by feeding micro droplets of KCl/KOH water solutions to the hot flue gas from a premixed CH 4 /air flame, having an equivalence ratio of 0.9 or 1.1.The combustion products including the gaseous potassium compounds then mixed with an opposed N 2 flow containing SO 2 to form a gas-phase counter-flow configuration.The temperature distribution was measured and the aerosols formed in the flow were detected.The results were interpreted in terms of an updated chemical kinetic model for Cl/S/K interactions to provide a better understanding of the sulfation process.

Burners
A schematic of the counter-flow burner setup is shown in Fig. 1.A uniform laminar N 2 flow seeded with a known amount of SO 2 was supplied through a 60 mm diameter sintered-metal porous plug of a McKenna burner.The outlet flow speed was 10 cm s −1 and the temperature was 298 K.The concentration of SO 2 in the flow was set to 0 ppm, 245 ppm or 980 ppm, respectively.
A uniform laminar hot flue gas was generated on a Perkin Elmer burner, modified to allow for seeding of water solutions [17].A chamber connected to an upper honeycomb matrix with a diameter of 23 mm was used to mix methane and air for the burner.In the present study, a premixed CH 4 /air flow with an unburned gas speed of 15 cm s −1 was used to generate a flat laminar flame at an equivalence ratio of either 0.9 or 1.1 above the burner.An aqueous solution containing potassium chloride or potassium hydroxide was seeded homogeneously into the flame by a nebulizer as shown in Fig. 2. The salt concentration of the aqueous solution was 0.5 mol l −1 , and the concentration of the gas-phase potassium chloride or potassium hydroxide in the hot flue gas was calculated to be around 215 ppm based on the consumption rate.A N 2 co-flow at room temperature was used to shield the flame from the ambient air with an exit flow speed of 15 cm s −1 .The distance between the Perkin Elmer burner and the McKenna burner was 20 mm.A stable counter-flow was formed between the N 2 flow from the McKenna burner and the hot flue gas from the Perkin Elmer burner.The reactions between SO 2 and KCl/KOH occurred in the mixing zone.

Laser measurement system
When the flue gas containing KCl or KOH mixed with the cold N 2 (SO 2 ) flow, the temperature decreased to ∼1000 K and aerosols were formed.The aerosols were detected from the Mie scattering when illuminated by a 355 nm laser beam from the third harmonic of a Nd:YAG laser (Brilliant B, Quantel) operated at 10 Hz with a pulse energy of 50 mJ, formed into a laser sheet using a telescope lens system.The laser sheet passed through the counter-flow region vertically as depicted in Fig. 1.The signal of the Mie scattering from the aerosols was detected by an intensified CCD camera (ICCD, PI-MAX II, Princeton Instruments).No laser breakdown was observed from illumination of the aerosols.
Using the same system, the temperature distribution of the counterflow was measured through Rayleigh scattering thermometry based on the following equation [18]: flow gas air air flow gas flow gas air (1) where T flow gas is the temperature of the counter-flow, T air is the tem- perature of the reference gas, i.e., the ambient air, which had a temperature of 298 K, I flow gas and I air are the Rayleigh scattering signals of the counter-flow gas and the room temperature air, which can be obtained by the laser setup shown in Fig. 2, and flow gas and air are the molecular Rayleigh scattering cross sections of counter-flow gas and air.
For the Rayleigh scattering cross sections, the same value was used during the temperature calculation.The uncertainty of the obtained temperature caused by the cross section difference between air and the counter flow gas was estimated to be less than 25 K for the temperature at 700 K.

Numerical approach
The sulfation reaction in the present setup was simulated using the counter-flow application code in CHEMKIN [19].The composition of the hot flue gas was obtained based on the one-dimensional modeling of the premixed CH 4 /air flame at an equivalence ratio (ϕ) of 0.9 or 1.1.The flue gas in the post flame, 3 mm from the flame front, was used as the inlet gas of the counter-flow model, which had a temperature of 2060 K (ϕ = 0.9) or 2118 K (ϕ = 1.1).Hence, the inlet speed setting was 103 cm s −1 (ϕ = 0.9) or 107 cm s −1 (ϕ = 1.1).The concentration of the gas-phase KCl or KOH in the flue gas was set to 215 ppm.The cool N 2 flow had an inlet temperature and flow speed of 298 K and 10 cm s −1 , respectively.If added, the concentration of SO 2 in the N 2 flow was either 245 ppm or 980 ppm.The distance between the two flow inlets was 20 mm.A potassium sulfation mechanism was adopted for the simulations, based on previous work [10,11,14].The reaction subsets and thermodynamic properties for H 2 oxidation [20], sulfur chemistry [21], and chlorine chemistry [22] were updated following recent developments.In the present work, the alkali chemistry was revised, with a particular emphasis on the sulfation pathways for KOH and KCl.The full mechanism is available as Supplementary Material.
As discussed in the Introduction, Glarborg and Marshall [10] proposed a mechanism for sulfation of KCl.The rate limiting step was the oxidation of SO 2 to SO 3 , followed by the fast sequence KCl + SO 3 (+M) = KSO 3 Cl(+M) (R1), KSO 3 Cl + H 2 O = KHSO 4 + HCl (R3), KHSO 4 + KCl = K 2 SO 4 + HCl (R4).Modeling studies indicate that this mechanism can explain sulfation rates for KCl obtained in laboratory flow reactors, entrained-flow reactors and premixed flames [10,11,14].However, the generation of O/H radicals is limited at low temperatures, and under these conditions the oxidation of SO 2 to SO 3 is too slow to explain observed the sulfation rate.
In an exploratory study, Hindiyarti et al. [11] proposed a number of alternative sulfation pathways that did not involve SO 3 as an intermediate.Below we re-evaluate these pathways, since they may be important for the sulfation of KOH.Hindiyarti et al. often employed upper limit rate constants for the involved reactions and we have re-evaluated some of these values.The sequence KOH + SO 2 (+M) = KHSO 3 (+M) (R6), KHSO 3 + O 2 = KHSO 4 + O (R7) was of particular interest since it did not involve any radicals as reactants and thereby could occur at lower temperatures.Hindiyarti et al. entered the rate limiting step R7 in the reverse direction, KHSO 4 + O = KHSO 3 + O 2 (R7b) and assumed the reaction to be fast.However, the step involves the breaking of an S]O double bond and we anticipate the reaction to be significantly slower, estimating k 7b by analogy to SO 3 + O.
Another proposed sequence involves KOSO 2 , KO + SO 2 (+M) = KOSO 2 (+M) (R8), KOSO 2 + OH (+M) = KHSO 4 (+M) (R9).By analogy to K + SO 2 (+M), reaction R8 is presumably fast and the second step is expected to be rate limiting.We have updated our estimate of k 9 based on the observation that reactions of alkali species with radicals are not very fast and that the reaction is presumably at its low pressure limit.
Alternatively, KHSO 4 could be formed by the sequence The thermodynamic properties for KO 2 and the rate constant for R10 were adopted from Sorvajärvi et al. [23].The rate limiting step in this sequence is expected to be reaction R12.This reaction is now included in the mechanism in the reverse direction, KHSO 4 + OH = KOSO 3 + H 2 O (R12b) with a rate constant assumed to be similar to that of HOSO 2 + OH = SO 3 + H 2 O.

Results and discussion
A photo of the premixed flame at an equivalence ratio of 0.9 is presented in Fig. 3(a).Potassium chloride aqueous solution was seeded homogeneously into the flame as small droplets from the Perkin Elmer burner.The chemiluminescence of the excited K atom at a wavelength of 404 nm can be observed in the light bluish hot flue gas above the flat flame.As the flue gas approached chemical equilibrium and the temperature decreased above the height of ∼12 mm, the chemiluminescence disappeared.The signal of the Rayleigh/Mie scattering in the region marked by the red dash lines in Fig. 3(a) was captured by the ICCD camera.The Rayleigh scattering signal of the counter-flow with pure water seeding is shown in Fig. 3(b).The scattering signal with KCl solution seeding is shown in Fig. 3(c).Comparing to the Rayleigh scattering signal shown in Fig. 3(b), the only difference in Fig. 3(c) is the appearance of a thin strong signal at the contact area between the hot flue gas and the cold nitrogen flow from the McKenna burner and the co-flow.The data at a given height of 6 mm is plotted in Fig. 3(d).The good agreement between the Rayleigh scattering signals of these two measurements indicates that the temperature distribution is not affected by the introduction of KCl.
The temperature distribution of the counter-flow was calculated based on Eq. ( 1).Fig. 4 compares the measured and predicted temperature profiles from the fuel-lean and the fuel-rich flame, respectively.The predicted temperature profile has been shifted by 0.3 mm compared to the measured values to compensate for minor non-idealities in the experiment.The figure also indicates the location of the stagnation plane for the two flames.
The strong signal from the Mie scattering of aerosols shown in Fig. 3(c) indicates that some gas-phase KCl underwent homogeneous nucleation in the low-temperature region.Combining the results of the Mie and Rayleigh scattering, it can be seen that aerosols of KCl in Fig. 3(c) are mainly formed in the region with a temperature between 700 K and 1000 K; this is discussed in more detail below.
The signal of the Mie scattering under different seeding conditions is presented in Fig. 5 for the lean flame (ϕ = 0.9).The signal was obtained after the subtraction of the Rayleigh scattering in the case without salt and SO 2 seeding.The strength of the Mie scattering signals is similar for the flows with only KCl or KOH seeding.The main difference between them is in the positions where aerosols are formed as shown in Fig. 5(c), which indicates different condensation temperatures for gaseous potassium species.The Mie scattering signal enhanced significantly as SO 2 was added through the McKenna burner flow as shown in Fig. 5(d) and (e).Meanwhile, no scattering was observed in the case with only SO 2 seeding (shown in Fig. 5(i)).The hot flue gas containing KOH produces a much stronger scattering signal with SO 2 than the hot flue gas containing KCl.The signal position also shifted down as SO 2 was added, which can be observed in Fig. 5(g).
The variation of the Mie scattering signal with the height in the center of the counter-flow (radial position = 0 mm) is shown in Fig. 6, together with the measured temperature distribution.For the cases with the lean flame, as the hot flue gas containing KCl was cooled by the N 2 flow without SO 2 , aerosols were formed at the position where the temperature was between 700 and 1000 K (shown in Fig. 6(a)).These aerosols were identified as KCl, which has a vapor pressure of 0.2 mbar at 1000 K [12].When SO 2 was induced in the N 2 flow, the scattering signal was enhanced significantly and shifted a bit to the hot flow side at a temperature between 1000 K and 1400 K.The aerosols formed at this temperature were considered to be condensed K 2 SO 4 , consistent with the fact that the vapor pressure of K 2 SO 4 , e.g., 0.04 mbar at 1300 K [12], is much lower than that of KCl [12], and thus K 2 SO 4 can condense at a higher temperature.Gas-phase potassium sulfate was produced through the sulfation reactions between KCl and SO 2 and subsequently underwent homogeneous nucleation.As the seeding of SO 2 was increased by a factor of four, i.e., from 245 ppm to 980 ppm as shown in Fig. 6(a), the peak value of the signal increased almost by the same factor.This indicates that the amount of K 2 SO 4 aerosols formed in the flow is controlled by the concentration of SO 2 .This finding is in agreement with the results of Ekvall et al. [16] who varied the S/K ratio from 1 to 4.
As shown in Fig. 6(b), aerosols were also formed when KOH was seeded into the lean flame, but in a region with a temperature between 900 K and 1100 K.This phenomenon is surprising, considering that KOH has a considerably higher saturated vapor pressure than KCl at the same temperature (e.g. 8 mbar for KOH at 1100 K vs. 1.5 mbar for KCl at 1100 K [24]).Conceivably the aerosols consist of potassium carbonate (K 2 CO 3 ), which, according to equilibrium calculations with Fact-Sage 7.0, is stable in condensed form under these conditions (see Supplementary Material).This might happen as the reverse process of the decomposition of K 2 CO 3 [25], i.e., 2KOH + CO 2 = K 2 CO 3 + H 2 O (R13), K 2 CO 3 = K 2 CO 3 (s) (R14).This sequence is promoted by the presence of ∼8% CO 2 in the flue gas.Formation of potassium carbonate in the gas-phase is not favored thermodynamically, but conceivably a small amount of gaseous K 2 CO 3 can be formed and nucleate homogeneously [26].
Similar to the flame with KCl seeding, K 2 SO 4 aerosols were detected as KOH mixed with the SO 2 from the N 2 flow.However, the intensity of the scattering signal was much stronger than for KCl seeding, indicating that aerosols were formed in larger amounts during sulfation of KOH.Hence, gas-phase sulfation of KOH by SO 2 appears to occur more readily than homogeneous sulfation of KCl.Moreover, as the concentration of SO 2 was increased from 245 ppm to 980 ppm, the peak of the scattering signal in the KOH case in Fig. 6(b) only doubled, indicating that the sulfation reaction was no longer first order in SO 2 .
Results for the fuel-rich flames (ϕ = 1.1) are shown in Fig. 6(c) and (d).Again, the Mie scattering signal was detected as SO 2 was seeded into the flow, but the signal was weaker by more than a factor of 10 comparing to the corresponding lean flame (ϕ = 0.9).The reduction in the signal is attributed to the lack of oxidizer (O 2 ) under the fuel-rich conditions, since oxidation is required in the sulfation process between KCl/KOH and SO 2 .Moreover, the signal intensity became almost independent of the concentration of SO 2 .Besides, under the condition of KCl seeding, the aerosol signals are detected at relatively low temperatures (between 800 K and 1000 K) and it is uncertain whether the aerosols consist of K 2 SO 4 .An assumption can be that the K 2 SO 4 is also formed under this condition, but its lower concentration induces a lower nucleation temperature based on the relationship between its saturated vapor pressure and temperature.
The reaction between KCl (215 ppm)/KOH (215 ppm) and SO 2 (245 ppm and 980 ppm) in counter-flow was modeled with the detailed reaction mechanism.The results for the KCl and KOH flames, respectively, seeded with 245 ppm SO 2 and for equivalence ratios of 0.9 and 1.1 are presented in Fig. 7.The figure shows the region 14-18 mm from the hot gas inlet, where K 2 SO 4 was formed as KCl/KOH was sulfated by SO 2 during the mixing between the hot flue gas and the cold N 2 /SO 2 flow.In this region the predicted temperature and its variation with position agree well with the experiments (see Fig. 4).The predicted concentration profiles of K 2 SO 4 under different conditions in Fig. 7 are consistent with the measurements.The model predicts that K 2 SO 4 has its concentration peak at a temperature around 1200 K, where the condensation of K 2 SO 4 happens and aerosols are formed according to its saturated vapor pressure [12].In agreement with observations, the The accurate intensity reading is to be found in Fig. 6.The SO 3 pathway is most important in sulfation of KCl in the low temperature region (< 1000 K) of the lean flame, where a small amount of SO 3 (∼1 ppm) is generated in the mixing region (Fig. 7(a)).
For both KCl and KOH, the major sulfation pathway is predicted to involve atomic potassium, formed in significant amounts from thermal dissociation of KCl and KOH, respectively (Fig. 7 An analysis of the A-factor sensitivity of the predicted K 2 SO 4 concentration shows that the sulfation rate is always sensitive to the reactions KO 2 + SO 2 (+M) = KOSO 3 (+M) (R11) and K + SO 2 (+M) = KOSO (+M) (R15), except at low temperature (∼900 K at ∼16 mm) for KCl sulfation where it becomes sensitive to KCl + SO 3 (+M) = KSO 3 Cl(+M) (R1).The rate constant for reaction R15 was measured at 863 K by Goumri et al. [27], while the value of k 11 is a rough estimate from Hindiyarti et al. [11].Both rate constants can be expected to be complex functions of temperature and pressure and involve considerable uncertainties.
Comparing the sulfation of KOH and KCl, less K 2 SO 4 was formed from KCl.This can be explained from the analysis of the calculations.As shown in Fig. 7(b) and (c), atomic K was formed in lower concentrations from KCl compared KOH.Thermal dissociation of KCl and KOH does not show up in the sensitivity analysis since both steps are rapidly equilibrated.However, KOH is thermodynamically less stable than KCl, leading to larger levels of atomic K.
The sulfation processes in the rich flames are similar to those occurring in the lean flames, according to our simulations.However, even though sufficient K 2 SO 4 was produced to form aerosols at a temperature around 1100 K, the concentration of K 2 SO 4 is much lower in the rich flames, in agreement with the experimental results.The difference is attributed to the low concentration of oxidants, which is two orders of magnitude smaller than for the lean cases.Thus, the amount of SO 3 formed in the rich flames becomes negligible, especially in the region where most K 2 SO 4 was formed (at ∼15.5 mm).Any SO 3 formed can be consumed rapidly by H atoms.There is almost no contribution from R1, R2 and R3 to the sulfation.The sulfation process was dominated by the pathway through KOSO 3 as discussed for the lean conditions.In the rich flames, a high concentration of K atom was predicted in the hot gas (e.g. more than 10 ppm K atom is predicted in the hot gas at ∼1500 K with KOH seeding).However, due to the lack of oxidants, reaction R16, i.e. the oxidation of KOSO by O 2 , becomes the rate-limiting step.KOSO is generated easily through R15 and becomes a major potassium compound, especially in the low temperature region with sufficient SO 2 (see Fig. 7(e) and (f)).
In the simulation, as the seeding of SO 2 increased from 245 ppm to 980 ppm with the KCl/KOH seeding remaining at 215 ppm, the increment ratio of the K 2 SO 4 concentration is always much smaller than the increment ratio of the SO 2 seeding.This can also be observed in most measurement results, but the sulfation of KCl in the lean flame is an exception, in which the strength of the scattering signal almost had the same increment ratio as the concentration of SO 2 .The simulations might underestimate the importance of the sulfation pathway through R1 and R3 for KCl at low temperature region (< 1000 K) where the formation and condensation of K 2 SO 4 can happen at the same time.If the effect of the condensation of K 2 SO 4 was taken into consideration in the modeling, more aerosols would be formed through this sulfation pathway.

Conclusions
In the present work, sulfation of KCl and KOH by SO 2 in hot flue gases was investigated.Simultaneous measurement of the temperature distribution and detection of aerosol formation in a counter-flow system allowed identification of aerosols as consisting of KCl K 2 SO 4 , and K 2 CO 3 , respectively, depending on the potassium speciation in the inlet and the availability of SO 2 .The experiments showed that KOH was sulphated more readily than KCl, resulting in larger quantities of aerosols.The sulfation process in the counter-flow setup was simulated using a chemical kinetic model including a detailed subset for the Cl/S/ K chemistry.Similar to the experimental results, much more potassium sulfate was predicted when seeding KOH compared to seeding KCl.For both KOH and KCl, sulfation was predicted to occur primarily through the reactions among atomic K, O 2 and SO 2 , forming KHSO 4 and K 2 SO 4 .The higher propensity for sulfation of KOH compared to KCl was mostly attributed to the lower thermal stability of KOH, facilitating formation of atomic K.According to the model, some sulfation also happened Fig. 8. Reaction pathways for the formation of K 2 SO 4 from KCl (215 ppm) and KOH (215 ppm) with SO 2 (245 ppm) at the position having the K 2 SO 4 concentration peak (∼1200 K) (a, c) and a low temperature (∼900 K) (b, d).For KCl, the K 2 SO 4 concentration peak is at 15.5 mm; For KOH, it is 15.2 mm; the position at 16 mm is selected as the low temperature region.The line thickness indicates the relative strength of the reactions.The flame with equivalence ratio of 0.9 is used to provide the hot gas.through SO 3 in the low temperature region (< 1000 K), especially for KCl (KCl → KSO 3 Cl → K 2 SO 4 ).To obtain a better understanding of the sulfation process for different Cl/K and S/K ratios, more quantitative measurements on the concentrations of KOH, KCl, K atom, and SO 2 in the reaction region are needed.The authors' group has recently obtained spectral-resolved quantitative UV absorption cross section of KCl and KOH [28], which will be used for in situ KCl, KOH and SO 2 concentration measurement in a novel multi-jet burner [29] to generate quantitative data for accurate kinetic understanding of the sulfation processes.

Fig. 1 .Fig. 2 .
Fig. 1.Schematic of the counter-flow setup together with a photo of a typical flame adopted in the experiment.The white dash lines depict the flow field.The light green square indicates the region illuminated by the laser sheet.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3 .Fig. 4 .
Fig. 3.The scattering signal measured in the counter-flow with the flame having equivalence ratio of 0.9.(a) Photo of the counter-flow with KCl seeding in the premixed flame.The box with red dash lines indicates the zone where the Rayleigh/Mie signal were captured; (b) The Rayleigh scattering signal of the counter-flow without seeding; (c) The scattering signal with KCl seeding; (d) The signal from (b) and (c) at height of 6 mm marked with the red dash line in (c).(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 5 .
Fig. 5.The Mie scattering signal from aerosols generated in the counter-flow (ϕ = 0.9) under different seeding conditions.(a) with seeding of KCl solution; (b) with seeding of KOH solution; (c) combination of the results in (a) and (b); (d) with seeding of KCl solution and with SO 2 in the N 2 flow from the McKenna burner; (e) with seeding of KOH and with SO 2 ; (f) combination of the results in (d) and (e); (g) combination of the results in (a) and (d); (h) combination of the results in (b) and (e); (i) only with SO 2 in the N 2 flow.Note: The signal over 5000 counts is saturated in the chosen color bar in order to visualize the weak scattering.The accurate intensity reading is to be found in Fig.6.

Fig. 6 .
Fig. 6.The vertical distribution of the measured temperature and the Mie scattering from aerosols along the center of the counter flow at different conditions.(a) Seeding of KCl into the flame at equivalence ratio of 0.9; (b) Seeding of KOH into the flame at equivalence ratio of 0.9; (c) Seeding of KCl into the flame at equivalence ratio of 1.1; (d) Seeding of KOH into the flame at equivalence ratio of 1.1.Dot lines indicate temperature distribution; solid lines indicate the Mie scattering signal; different color indicates different seeding conditions: without seeding salt solution (gray); seeding salt solution without SO 2 (black); seeding salt solution with 245 ppm SO 2 (red); seeding salt solution with 980 ppm SO 2 (blue).(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) (b) and (c)).The concentration of atomic K can be 1 ppm with 10 ppm of KOH at the temperature of ∼1500 K.The K-atom is then converted to KOSO 3 , either through K + SO 2 (+M) = KOSO (+M) (R15), KOSO + O 2 (+M) = KOSO 3 (+M) (R16) or via K + O 2 (+M) = KO 2 (+M) (R10), KO 2 + SO 2 (+M) = KOSO 3 (+M) (R11).These reactions are promoted by the availability of O 2 and of SO 2 , which diffuses from the cool N 2 side with a concentration of ∼100 ppm.Once formed, KOSO 3 may react with KOH to form K 2 SO 4 directly, KOSO 3 + KOH = K 2 SO 4 + OH (R17), or through the sequence KOSO 3 + H 2 O = KHSO 4 + OH (R12), KHSO 4 + KOH = K 2 SO 4 + H 2 O (R5).

Fig. 7 .
Fig. 7.The simulated mole fraction distributions of key species and temperature along the center of the counter flow reactor of different conditions at equivalence ratio of 0.9 and 1.1.The seeding condition: 245 ppm SO 2 without salt seeding (a, d), with 215 ppm KCl (b, e) and 215 ppm KOH (c, f).