Measurement and Prediction of Ash Deposition Rates for Air- and Oxy-Combustion of a Wide Range of Solid Fuels

Paper from the AFRC 2018 conference titled Measurement and Prediction of Ash Deposition Rates for Air- and Oxy-Combustion of a Wide Range of Solid Fuels

This paper presents a synthesis of a large body of experimental data describing rates of ash deposition from solid fuel combustion from nine fuels burned under a range of air- and oxy-combustion conditions (or 30 conditions in all). The experimental data were obtained from a 100kW (rated) down-flow combustor that allowed self-sustained pulverized solid fuel combustion. These conditions simulated practical conditions as far as temperature- time histories, and gas and particle concentrations are concerned, although the flow was necessarily laminar at the deposit probe location. Deposition rates were measured using a wall temperature controlled deposition probe, and the collected deposits can be divided into the tightly bound "inside deposit", closest to the heat transfer surface, and the loosely bound "outside deposit" that is more easily removed by soot blowing. The size segregated ash aerosol data were obtained using Electric Mobility and Light Scattering methods and Low Pressure Impactors. The focus of this paper is on exploring the ash transformation in different fuels and developing universal relationships between ash aerosol characteristics and the deposition rates. The results from all the fuels (fossil and biomass), under both air and oxy-combustion conditions suggest the formation of the inside deposit layer and the outside deposit layer involve different mechanisms, and their deposition rates have different dependencies. Specifically, the deposition rate of the "inside deposit" layer is proportional to the concentration of sub-micron particles (PM1), of any composition, in the flue gas. In contrast to inside deposits, the deposition rate of the "outside deposit" layer is roughly proportional to the total alkali concentration in the flue gas, but not to PM1 concentration. The formation of submicron particles is found to be significantly boosted under high temperature advanced oxy-combustion cases. The practical consequence of these findings is that, in general, increasing ash concentrations and flame temperatures (as in advanced oxy-combustion) will lead to higher deposition rates.

Measurement and prediction of ash deposition rates for air- and oxy-combustion of a wide range of solid fuels Yueming Wang and Jost O.L. Wendt Department of Chemical Engineering and Institute for Clean and Secure Energy University of Utah, Salt Lake City, UT 84112, USA ABSTRACT This paper presents a synthesis of a large body of experimental data describing rates of ash deposition from solid fuel combustion from nine fuels burned under a range of air- and oxy-combustion conditions (or 30 conditions in all). The experimental data were obtained from a 100kW (rated) down-flow combustor that allowed self-sustained pulverized solid fuel combustion. These conditions simulated practical conditions as far as temperature- time histories, and gas and particle concentrations are concerned, although the flow was necessarily laminar at the deposit probe location. Deposition rates were measured using a wall temperature controlled deposition probe, and the collected deposits can be divided into the tightly bound "inside deposit", closest to the heat transfer surface, and the loosely bound "outside deposit" that is more easily removed by soot blowing. The size segregated ash aerosol data were obtained using Electric Mobility and Light Scattering methods and Low Pressure Impactors. The focus of this paper is on exploring the ash transformation in different fuels and developing universal relationships between ash aerosol characteristics and the deposition rates. The results from all the fuels (fossil and biomass), under both air and oxy-combustion conditions suggest the formation of the inside deposit layer and the outside deposit layer involve different mechanisms, and their deposition rates have different dependencies. Specifically, the deposition rate of the "inside deposit" layer is proportional to the concentration of sub-micron particles (PM ), of any composition, in the flue gas. In contrast to inside deposits, the deposition rate of the "outside deposit" layer is roughly proportional to the total alkali concentration in the flue gas, but not to PM concentration. The formation of submicron particles is found to be significantly boosted under high temperature advanced oxy-combustion cases. The practical consequence of these findings is that, in general, increasing ash concentrations and flame temperatures (as in advanced oxy-combustion) will lead to higher deposition rates. 1 1 1. Introduction The effects of coal and biomass composition and combustion conditions on the mechanisms of ash formation have been widely investigated by numerous researches since last century. However, extensive utilization of these coal and biomass resources is still constrained by several challenges associated with emission of SO , NOx and particular matter (PM) and other issues like corrosion, slagging and fouling [1-9]. Among these problems, this paper will focus on the formation of particular matter and fouling ash in air- and oxy-combustion of various solid fuels, including fossil fuels and biomass. The objective of this work is to understand: (1) the effects of fuel properties and flame temperature on the ash transformation that take place during combustion; (2) the internal relationship between the ash aerosol properties and growth rates of ash deposits for a wide range of fuels. Hopefully, these understandings can aid in predicting the formation of ash deposits for various solid fuels and different combustion conditions. 2 Figure 1 shows the relationship between ash aerosol formation mechanisms and deposit formation mechanisms [10]. The organically-bound or volatile minerals can be easily vaporized at the beginning. Although the refractory oxides (SiO or Al O ) are normally hard to be vaporized, their vaporization can be greatly enhanced in reducing 2 2 3 environment near the fuel particle [11]. The submicron ash aerosols can hereafter be formed through various processes such as nucleation, condensation, coagulation and coalescence, while the supermicron ash aerosols can be directly formed through char fragmentation [5, 12]. Therefore, one might be able to predict at least two modes for the particle size distribution of the generated ash aerosols during combustion process, except the fuels without non organically-bound minerals like residual oil [13, 14]. Sometimes, more than one submicron mode can be observed [15], where each mode is formed by the condensing of mineral matter but in different location. Typically, the vaporization mode with particle diameter between 0.1-0.7 µm is normally formed through the vaporization and subsequent nucleation and coagulation near the fuel particles in the flame zone, while another vaporization mode, with particle diameter smaller than 0.1 µm is formed due to the condensation of the uncooled mineral vapors such as alkali sulfates closer to or inside the sample probe. The fuel particles in a pulverized coal combustion (PCC) unit are usually smaller than 100 µm and the fragmentation is usually percolative in nature [16]. The macropores in char provide oxygen with easy access to the internal and external surfaces, then char structure becomes more and more porous and eventually breaks apart into various fragments with different sizes [16]. Hence, the extent of the char fragmentation dependents on its macroporoisity. Figure 1: Relationship between ash aerosol formation mechanisms and ash deposition mechanisms[10] As shown in Figure 1, deposition of the ash particles may occur by a number of processes. Typically, the large particles are usually deposited by inertial impaction and eddy diffusion, the submicron particles are deposited by thermophoresis or Brownian diffusion, and uncooled vapors can condense or adsorb onto wall surface directly [3]. The impacted particle may stick on the wall or rebound, the criteria is usually based on its viscosity, kinetic energy and impaction angle [17]. The formed ash deposits can further capture more incoming particles or be removed due to erosion [18], which depends on the particle and surface characteristics. Thus, the overall capture efficiency is defined as the ratio of the mass of particulate sticking to surface to total mass of particular injected. Although abundant literatures concerning numerical simulations are available to assist in prediction of ash deposition formation for coal and biomass combustion [18-27], most of them are only valid in narrow compositional ranges. Therefore, the objective of this work is to explore a mechanism to predict the growth of deposition for various fuels with a wide range of ash composition. It has been found that ash deposits on heat transfer surfaces can be mainly divided into two types [28]: (1) ‘inner' deposits that tightly adhere on the wall and can only be removed by scraping; (2) ‘outer' deposits that loosely bound on the wall and can be easily removed by soot-blowing. These ‘inner' deposits are mostly composed of small and sticky particles, which impact on the wall through the combination of thermophoresis and inertial impaction. Thus, it is reasonable to presume that there exists a positive correlation between growth rate of ‘inner' deposits and concentration of submicron particle. This hypothesis has been proved by several studies concerning coal and biomass [29-32]. For example, it is reported by Zhan et al. [30] that the ‘inner' deposition rates can be linearly correlated with the PM1 concentration no matter with the composition for several coal combustion cases. The ‘outer' deposits are mostly formed by the inertial impaction of big and coarse particles, and the weight is usually much higher than ‘inner' deposits. For the majority part of the ash deposits, its growth can be simulated through some complicated mathematical models, but this work intends to develop a simplified model for deposition rates predication. 2. Methods and Materials 2.1 Experimental facilities All the experimental work discussed here was done on a 100 kW (rated) down-fired oxy-fuel combustor (OFC). OFC has been extensively used in previous researches for various pulverized solid fuels including coal and biomass [10, 29, 33-37], its configuration is shown in Figure 2. As shown in Figure 2, OFC is mainly composed of an ‘ignition' zone (port 1 to port 3), a ‘radiation' zone (port 4 to port 9) and a ‘convection' zone in horizontal section. It is a self-sustained (without external heating) and systematically controlled reactor. Furthermore, it operates at realistic stoichiometric ratios, with turbulent diffusion flames in ignition zone causing realistic temperature/time profiles, but the flow becomes laminar after port 4. The ash deposits were collected by a temperature controlled probe, which has been described in details elsewhere [28]. Since fouling deposition is defined as the deposit on convection heat surface with relatively cooler flue gas temperature, the ash deposits were usually collected between port 6 and port 8 wherever the flue gas temperature is close to 1300 K. The wall temperature of the probe is usually controlled at 922 K, which is a typical wall temperature of the superheater in utility boilers. The collected ash deposits can then be divided into side, inside and outside deposits as shown in Figure 2, and the side and inside deposits can be referred as ‘inner' deposits while the outside deposits can be referred as ‘outer' deposits. This paper will only focus on the inside and outside deposits. Ash aerosols were sampled by a water-cooled isokinetic probe, the features of which can be found in [15]. The sampled aerosols were instantly quenched by nitrogen at the inlet to avoid the further coagulation inside the probe, and the isokinetic sampling was accomplished by adjustment of nitrogen flow rate. The real-time measurement of particle size distribution was measured by a Scanning Mobility Particle Sizer (SMPS) and Aerodynamic Particle Sizer (APS). The size segregated ash aerosols were also collected by Berner Low Pressure Impactor (BLPI) for further composition analysis. Figure 2: Configuration of oxy-fuel combustor (OFC) and ash deposits division 2.2 Composition analysis of solid fuels Nine different pulverized solid fuels will be discussed in this paper, given that this paper intends to investigate the characteristics of generated deposition for a wide range of solid fuels. The results from some of these fuels have been reported before [10, 29, 38], but they are recapped here in order to determine a universal mechanism of ash deposition formation. Specifically, these fuels includes Powder River Basin coal (PRB), Illinois coal, 40 wt.% PRB/60 wt.% Illinois blend, Utah Sufco coal 1, Utah Sufco coal 2, rice husk, 20 wt.% rice husk/80 wt.% Sufco 1 blend, 13 wt.% rice husk/ 87 wt.% PRB blend and petroleum coke. These fuels have mean particle diameters ranging from 70-100 µm. Their composition and ash content can be found in Table 1 and Table 2. Although Utah Sufco 1 and Sufco 2 are from same mine in Utah, their ash contents are totally different as shown in Table 2 due to different mining time and location. Note that the significant composition difference among these nine fuels. Among these fuels, the experimental results from Sufco 1, rice husks and petroleum coke will be used to discuss the effect of fuel properties on the formation of ash aerosols and deposition. Hereafter, the results of ash deposition rates from all the fuels will be summarized to determine a simplified correlation between the deposition rates and the ash aerosol properties. Table 1: Composition analysis of various solid fuels Fuel ASH C H N S O HO 2 Volatile FC HHV (%) (%) (%) (%) (%) (%) (%) (%) (%) (kJ/kg) PRB 4.94 53.72 3.59 0.78 0.23 13.05 23.69 33.36 38.01 21115 Illinois 9.42 63.47 4.36 1.24 3.12 8.76 9.64 36.04 44.90 26870 PRB+Illinois 7.63 59.57 4.05 1.06 1.96 10.48 15.26 34.97 42.14 24568 Sufco 1 8.36 67.87 4.77 1.09 0.36 11.44 6.11 38.49 47.04 27677 Sufco 2 13.96 62.41 4.52 1.10 0.46 11.04 6.52 37.36 42.16 27319 Rice husks 33.67 28.47 4.15 1.05 0.10 24.42 8.16 48.96 9.22 11551 Rice husks+Sufco 1 13.42 59.99 4.67 1.08 0.31 14.04 6.52 40.58 39.48 24451 Rice husks+PRB 8.67 50.44 3.66 0.82 0.21 14.53 21.67 35.39 34.27 19871 2.99 82.51 6.02 1.71 FC: Fixed Carbon; HHV: Higher Heating Value 5.65 0.49 0.57 10.18 86.26 35720 Petroleum coke 1 1 2 2 Table 2: Ash analysis of various solid fuels Fuel Al O CaO Fe O MgO MnO PO KO 2 SiO Na O SO TiO PRB 14.78 22.19 5.20 5.17 0.01 1.07 0.35 30.46 1.94 8.83 1.30 Illinois 20.18 3.22 16.46 0.89 0.03 0.10 2.10 51.22 1.06 2.79 0.98 PRB+Illinois 18.02 10.81 11.96 2.60 0.02 0.49 1.40 42.92 1.41 5.21 1.11 Sufco 1 8.34 18.21 5.25 2.84 0.05 0.01 0.33 48.85 3.09 5.96 0.64 Sufco 2 12.09 11.90 3.62 3.94 0.03 0.25 1.13 62.48 0.81 1.83 0.68 Rice husks 1.73 1.31 1.10 0.84 0.83 1.81 2.66 88.51 0.31 0.32 0.18 Rice husks+Sufco1 5.03 9.76 3.18 1.84 0.44 0.91 1.50 68.68 1.70 3.14 0.41 Rice husks+PRB 8.26 11.75 3.15 3.01 0.42 1.44 1.51 59.49 1.13 4.58 0.74 19.40 4.22 7.02 0.66 0.06 0.18 1.17 46.70 0.72 Petroleum coke: There also exists 1.26% NiO and 8.24% V O in ash other than the listed content. 3.77 0.63 2 3 2 3 Petroleum coke 1 1 2 2 5 2 2 3 2 5 2.3 Combustion conditions The tested combustion conditions can generally classified under two types depending on the peak flame temperature in the combustor, for all the fuels discussed in this work. First are the lower peak flame temperature cases (shown by open symbols in Figure 9), which means the fuel is burned under air or oxy-combustion with 27% oxygen in oxidant gas (denoted as OXY27); second are the higher peak flame temperature cases (shown by solid symbols T in Figure 9), which occur when the fuel is burned under oxygen enriched atmospheres (advanced oxy- combustion like OXY50, OXY70 or OXY80). For these oxy-combustion cases, some of them use recycled flue gas (RFG) with different treatment while some cases use once through CO to represent fully cleaned recycled flue gas. It is found that RFG's with different cleanup options do not have significant effect on ash transformation [35], hence RFG is only used in the oxy-combustino of PRB, Illinois coal and their blends. 2 3. Results and Discussion 3.1 The effects of fuel properties and flame temperature on the inorganic transformation during combustion As has been mentioned, the experimental results for Sufco 1, rice husks and petroleum coke will be used here to determine the effect of fuel properties on both formation of ash aerosols and on deposition. Figure 3 shows the measured ash aerosol particle size distributions (PSDs) of these fuels under OXY70 and Air combustion. Figure 3a is the volumetric concentration (µg/m ) and Figure 3b is the input ash concentration (µg/g_ash). It is suggested that all PSDs in Figure 3 are classical bi-modal distributions except the PSD of Sufco 1_Air is tri-modal. Again, the submicron modes are formed by the vaporization of mineral matters and their further nucleation and coagulation, while the super micron modes are formed by the fragmentation. Despite the small difference in the concentration among these fragmentation modes, the peak size of these modes are all at about 3 µm. Although the flame temperatures in OXY70 cases are usually 300 K higher than that in Air cases, it seems that the flame temperature does not affect the formation of fragmentation mode as shown in Figure 3b. In contrast to super micron modes, there are more submicron particles generated in OXY70 cases due to enhanced vaporization of mineral matters under higher flame temperature in advanced oxy-combustion. 3 For the Sufco 1 and rice husks, their submicron modes are located at 0.2 µm for OXY70 and at 0.3 µm for Air, having similar concentration. A difference is that Sufco 1_Air has additional ultrafine mode at 0.04 µm. The larger submicron mode is formed within the flame zone [12] and is also known as accumulation mode, while the smaller submicron mode is believed to be formed by the nucleation of the uncooled mineral vapor (like alkali sulfates) in the post flame zone and can be denoted as a nucleation mode. For the petroleum coke, there exists only one accumulation mode (dp<0.1 µm) for both Air and OXY70, and the concentrations of these ultrafine particles are much higher than Sufco 1 and rice husk. According to the coagulation theory [12], the generated particle diameter will be larger if the concentration of nuclei is higher, but the diameter for accumulation mode in petroleum coke is smaller than coal and rice husk. A possible explanation is that coagulation of nuclei occurs within the boundary layer near fuel particles for Sufco 1 and rice husk, but that occurs after nuclei mixing with flue gas for petroleum coke. Therefore, the nuclei concentration is lower for petroleum coke due to much bigger control volume even though the number of nuclei is higher. It should be noted that the PSDs for petroleum coke exhibit a discontinuity at 0.6 µm, where SMPS and APS data are merged. This is because the burnout of petroleum coke is fairly low due to its low char activity, which can generate unburned char particles. Therefore, the detected super micron char concentration by light scattering processes (APS) will be lower than the actual value since the absorbance of char is much higher than ash, but the electric mobility machine is not affected. (a) (b) Figure 3: Ash aerosol particle size distributions of three fuels under OXY70 and Air Other than the ash aerosol PSDs, it is also essential to know their composition. The size segregated aerosols were collected by low pressure impactor with cutoff size ranging from 32.4 nm to 7.33 µm, and their compositions were measured by EDS. Figure 4 shows the elemental distribution with particle size for Na, K, Si and Ca. As shown in Figure 4, the alkali metals (Na and K) are more enriched in Air than OXY70 for submicron particles for all three fuels. The reduction of the alkali contents in OXY70 is caused by the enhanced reaction between alkali vapors and coarse alumino-silicate particles under higher flame temperature [39]. Thus, less alkali contents can be retained in the vapor to form submicron particles in OXY70. Although the silica content in rice husk is very high, the alumina content is very low, then the scavenging effect of rice husk OXY70 is less efficient than Sufco 1 and petroleum coke. In contrast to alkali metals, the Si and Ca contents are more enriched in OXY70 than Air for submicron particles because their vaporization is promoted under higher flame temperature to generate submicron particles in OXY70. However, this is opposite for Si in petroleum coke and Ca in rice husk. It should be noted that Zeng et al. [40] reported that PM1 is mostly composed of K and Cl when rice husk was burned in a drop tube furnace, but the silicon content was surprisingly high for rice husk in this work. This discrepancy of silicon partitioning might be caused by the difference in rice husk ash and the particle temperature history. Figure 4: Size segregated composition of ash aerosols for three fuels under Air and OXY70 The generated ash aerosols can deposit on heat exchangers to cause problematic fouling issues, and so it is important to understand characteristics of the formed ash deposition. In this work, the composition of inside and outside deposits collected after 1 hour were measured by EDS, given that it has been reported by Zhan et al. that the composition of ash deposits do not change much after the sampling time is longer than 1 hour [28]. The measured compositions of ash deposits for OXY70 and Air are shown in Figure 5a and 5b separately. The inside deposits seems to have more Fe but less Si for these three fuels. Both Sufco 1 and rice husk have more S in inside deposits but it is inverse for petroleum coke. This is because the inside deposits are mainly formed by small and sticky particles like alkali sulfates or calcium sulfates, then the sulfur content is usually more enriched in inside deposits. However, the sulfur content in petroleum coke is greatly higher than other two fuels, and the outside deposits are directly exposed to the flue gas, then the sulfation of the outside deposits is more intense than inside deposits. It should be noted that the Fe content in ash deposits for petroleum coke is significantly higher than the raw ash content, especially for OXY70. This is because: (1) the Fe tend to form big hollow particles, which is easy to be captured; (2) the iron bearing minerals can coat on the surface of the deposited particles; (3) the molten phase FeOFeS can be formed at ~1300 K even under oxidizing conditions, which can aggravate the fouling on deposition probe, though these phase will be the oxidized to magnetite or hematite. For these three fuels, the alkali contents are more enriched in Air than OXY70 for both inside and outside deposits. This is similar with ash aerosol results as shown in Figure 4, because the ash aerosol is the precursor of ash deposits. Furthermore, this enrichment of alkali content in Air is more noticeable in inside deposits than outside deposits because the inside deposits are mainly composed of smaller particles with greater alkali enrichment as shown in Figure 4. After this initial layer is formed, other big and coarse particles can be captured in outside deposits, which have little composition difference. This is also why the Si content in inside deposits are lower than outside deposits. (a) (b) Figure 5: Composition of ash deposits for three fuels under Air and OXY70 3.2 The correlation between deposition rates and aerosol concentrations Figure 6 shows the photos of the top and front view of the collected ash deposits with sampling time of 2 hours for Air (left side) and OXY70 (right side) of blend of Sufco 1 and rice husk. These ash deposits were solidified on the wall by epoxy before the removable couple was carefully taken off from the deposition probe. The ‘deposition angle' in OXY70 is clearly larger than Air because the erosion is much more intense in Air than OXY70 due to the higher particle kinetic energy in Air. There are less particles retained in Air, though the impaction efficiency in Air is larger than OXY70. In other words, the deposition rates in OXY70 is higher than Air, this finding is true for all other fuels. It should be noted that the Reynolds numbers for most cases are lower than 1200 at where the deposits were collected, then it can be treated as laminar flow and turbulent eddy diffusion can be neglected. A thoughtful mathematical model is able to simulate the ash deposition formation, but a simplified correlation between ash deposition and ash aerosol will be more practical and useful to predict the growth rate of ash deposition. In order to better understand the deposition mechanism, it is important to investigate the morphology of ash deposits. Figure 7a and 7b are the SEM images of the outside deposits for Sufco 1 in OXY70 with sampling time of 1 hour, while Figure 7c and 7d are for inside deposits. It is indicated that the particle size of the outside deposit is much larger than inside deposits, and Figure 7a suggests apparent sintering between particles for outside deposits. Figure 7d implies that there are few amounts of submicron particles agglomerated on big particles, and a few of previous studies have discovered that the submicron particles are essential for the formation of inside deposits because these small particles can stick on bigger particles and increase the capture efficiency of these big particles [30-32]. The effect of these submicron particles will be further discussed later. Figure 6: Ash deposits images of a removable coupon in Air (left side) and OXY70 (right side) of Sufco 1 and rice husk blend (a) (c) (b) (d) Figure 7: SEM images in OXY70 of Sufoc: A) outside deposits 500x; B) outside deposits 5000x; C) inside deposits 500x; D) inside deposits 5000x In order to determine growth rate of deposition, the inside and outside deposits were collected at various sampling time (10 minutes, 20 minutes, 30 minutes, 1 hour and 2 hours), then the weights of collected deposits are plotted with time. Figure 8 shows the result for rice husk in OXY70 case, which is representative of all other cases. It is implied from Figure 8 that the weight of inside deposits increases rapidly at the beginning and then stops growing after 2 hours, thus the inside deposition rate is defined as the initial growth rate. The weight of outside deposits, on the other hand, increases continuously within 2 hours, thus outside deposition rate is defined as the average growth rate. Furthermore, the weight of inside deposit is significantly less than outside deposit at the same sampling time. Figure 8: Inside and outside deposits weights with sampling time of rice husk in OXY70 combustion In order to determine a univeral deposition mechaism for a wide range of solid fuels, all the deposition results from nine fuels are reported and discussed here, including 30 inside deposition rates and 19 outside deposition rates in total. It has been found that inside deposits are composed of small particles, while some bigger particles can be ‘glued' by submicron particles as shown in Figure 7d. Zhan et al. pointed out that inside deposition rates can be correlated with PM1 concentration for the first time [10], but only with 3 fuels. In Figure 9a, this correlation is extended to all 9 fuels including some reported results [10, 29]. Note the temperature in Figure 9 suggests the peak flame temperature for each case, the peak flame temperature of each combustion condition is modulated by oxygen content in oxidant gas. The inside deposition rates and PM1 concentrations are greatly higher in higher tmeperature cases than lower temperature cases for the same fuel as shown in Figure 9a. It should be noted that the PM1 concentration is measured through the collected ash aerosol in BLPI with cutoff size smaller than 1 µm. Although the PM1 constitution is quite different between high temperature oxy-combustion and low temperature air combustion cases as shown in Figure 4, the inside deposition rates can be well correlated with the PM1 concentration (r-squared value=0.81 in Figure 9a). Indeed, no matter what the composition of submicron particles are fomed, these small particles are able to deposit on the wall through thermophoresis with high capture efficiency [36]. Traditionally, it was believed that alkali contents are crucial for deposition formation [1]. Attempt has been done to correlate inside deposition with PM1 alkali concentration or total alkali concentration in flue gas, but their r-squared values are all much lower than 0.81. Differently with inside deposits, the outside deposits are mainly formed by inertial impaction and the particle diameter is usually larger than 10 µm. After the inner layer is formed, the larger particles are easier to be captured to form the loosely bound outer layer. The existence of alkali contents could increase the particle capture efficiency since they can form stickier layer on the outer surface of large particles. Hence, it is reasonable to presume the growth rate of outer deposits will be higher if the alkali content is more enriched in flue gas. Figure 9b shows the correlation between outside deposition rate with total alkali concentration in flue gas (r-squared value=0.77). The total alkali concentration is calculated by mulitiplying injected ash concentration with mass fraction of alkali content in raw ash. Figure 9b indicates that the outside deposition rates are propotional to total alkali concentration. It should be noted that the mass fraction of alkali contents (Na O+K O) in raw ash are insufficiently different among these fuels, only ranging from 1.89% to 3.42%, thus the outside deposition rates can also be well correlated with total ash concentration. Alkali content additives, such as potassium acetate and soidum acetate, will be added to the fuels in future work in order to significantly change the mass fraction of alkali content in solid fuels. It should be noted that the ash deposit was collected in lamniar flow zone (Re<1200), and this simplified correlation does not directly consider the impaction efficiency, capture efficiency and erosion, but it indeed provides a easier way to predict the growth rate of outside deposits under these certain conditions. 2 2 (a) (b) Figure 9: (a) Correlation of inside deposition rate with PM1 concentration; (b) correlation of outside deposition rate with total ash concentration in flue gas. 4. Conclusions The experimental results of tight inside deposits and loose outside deposits from nine solid fuels, including fossil fuels and biomass, has been investigated in a 100 kW (rated) down-fired laboratory combustor. Despite the variation of ash content amid these solid fuels, some similar trends for ash transformation and deposition can be concluded. Firstly, alkali contents are more enriched on sub-micron particles for Air combustion compared to for OXY70, due to the increased reaction between alkali vapor and alumnio-silicate particles with higher flame temperature. Although the composition difference in ash deposit is consistent with that of the ash aerosol, inside deposits have more alkali but less Si than outside deposits. Secondly, there are more submicron particles generated in OXY70 due to higher flame temperature. These submicron particles can deposit on the wall to form inside deposits, and some super micron particles can be further ‘glued' by these small particles to stick on the wall. Thus the growth rates of inside deposits are proportional to PM1 concentration no irrespective of their composition. Thirdly, the outside deposits are mainly formed by the inertial impaction of large particles and the alkali content can increase their capture efficiency. Hence the growth rates are proportional to total alkali concentration in the flue gas. Although only simplified correlations are provided in this work to predict deposition growth rates, these results can aid in the simulation of deposition formation. Acknowledgement The author would like to acknowledge the finial support from the National Science Foundation (Award 1603249) and Department of Energy (Award DE-FE0025168). This work is also partially supported by the National Natural Science Foundation of China (Grant 51520105008). Reference [1] [2] [3] [4] [5] [6] [7] [8] [9] Bryers RW. Fireside slagging, fouling, and high-temperature corrosion of heat-transfer surface due to impurities in steam-raising fuels. Progress in energy and combustion science 1996;22(1):29-120. Raask E. Mineral impurities in coal combustion: behavior, problems, and remedial measures. New York: Hemisphere Publishing Corporation; 1985. Baxter LL. Ash Deposit Formation and Deposit Properties. A Comprehensive Summary of Research Conducted at Sandia's Combustion Research Facility. Sandia National Labs., Albuquerque, NM (US); Sandia National Labs., Livermore, CA (US); 2000. Lighty JS, Veranth JM, Sarofim AF. Combustion aerosols: factors governing their size and composition and implications to human health. Journal of the Air & Waste Management Association 2000;50(9):1565618. Xu M, Yu D, Yao H, Liu X, Qiao Y. Coal combustion-generated aerosols: Formation and properties. Proceedings of the Combustion Institute 2011;33(1):1681-97. Nussbaumer T. Combustion and co-combustion of biomass: fundamentals, technologies, and primary measures for emission reduction. Energy & fuels 2003;17(6):1510-21. Demirbas A. Potential applications of renewable energy sources, biomass combustion problems in boiler power systems and combustion related environmental issues. Progress in energy and combustion science 2005;31(2):171-92. Koppejan J, Van Loo S. The handbook of biomass combustion and co-firing. Routledge; 2012. Wendt JO, Sternling C, Matovich M. Reduction of sulfur trioxide and nitrogen oxides by secondary fuel injection. Symposium (International) on Combustion. 14. Elsevier; 1973:897-904. [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] Zhan Z, Fry A, Wendt JOL. Relationship between submicron ash aerosol characteristics and ash deposit compositions and formation rates during air-and oxy-coal combustion. Fuel 2016;181:1214-23. Quann R, Sarofim A. Vaporization of refractory oxides during pulverized coal combustion. Symposium (international) on combustion. 19. Elsevier; 1982:1429-40. Liu H, Wang Y, Wendt JOL. Particle size distributions of fly ash arising from vaporized components of coal combustion - a comparison of theory and experiment. Energy & Fuels 2017. Sippula O, Hokkinen J, Puustinen H, Yli-Pirilä P, Jokiniemi J. Comparison of particle emissions from small heavy fuel oil and wood-fired boilers. Atmospheric Environment 2009;43(32):4855-64. Linak WP, Miller CA, Wendt JO. Fine particle emissions from residual fuel oil combustion: Characterization and mechanisms of formation. Proceedings of the combustion Institute 2000;28(2):26518. Zhan Z, Fry A, Zhang Y, Wendt JOL. Ash aerosol formation from oxy-coal combustion and its relation to ash deposit chemistry. Proceedings of the Combustion Institute 2015;35(2):2373-80. Mitchell RE, Akanetuk AJ. The impact of fragmentation on char conversion during pulverized coal combustion. Symposium (International) on Combustion. 26. Elsevier; 1996:3137-44. Kleinhans U, Wieland C, Frandsen FJ, Spliethoff H. Ash formation and deposition in coal and biomass fired combustion systems: Progress and challenges in the field of ash particle sticking and rebound behavior. Progress in Energy and Combustion Science 2018;68:65-168. Liu C, Liu Z, Zhang T, Huang X, Guo J, Zheng C. Numerical Investigation on Development of Initial Ash Deposition Layer for a High-Alkali Coal. Energy & Fuels 2017;31(3):2596-606. Zhou H, Jensen PA, Frandsen FJ. Dynamic mechanistic model of superheater deposit growth and shedding in a biomass fired grate boiler. Fuel 2007;86(10-11):1519-33. Lokare SS, Dunaway JD, Moulton D, Rogers D, Tree DR, Baxter LL. Investigation of ash deposition rates for a suite of biomass fuels and fuel blends. Energy & Fuels 2006;20(3):1008-14. Lee F, Lockwood F. Modelling ash deposition in pulverized coal-fired applications. Progress in Energy and Combustion Science 1999;25(2):117-32. Shao Y, Wang J, Xu CC, Zhu J, Preto F, Tourigny G, et al. An experimental and modeling study of ash deposition behaviour for co-firing peat with lignite. Applied energy 2011;88(8):2635-40. Rushdi A, Gupta R, Sharma A, Holcombe D. Mechanistic prediction of ash deposition in a pilot-scale test facility. Fuel 2005;84(10):1246-58. Fan J, Zha X, Sun P, Cen K. Simulation of ash deposit in a pulverized coal-fired boiler. Fuel 2001;80(5):645-54. Wang H, Harb JN. Modeling of ash deposition in large-scale combustion facilities burning pulverized coal. Progress in Energy and Combustion Science 1997;23(3):267-82. Richards GH, Slater PN, Harb JN. Simulation of ash deposit growth in a pulverized coal-fired pilot scale reactor. Energy & Fuels 1993;7(6):774-81. Wessel RA, Righi J. Generalized correlations for inertial impaction of particles on a circular cylinder. Aerosol Science and Technology 1988;9(1):29-60. Zhan Z, Bool LE, Fry A, Fan W, Xu M, Yu D, et al. Novel temperature-controlled ash deposition probe system and its application to oxy-coal combustion with 50% Inlet O2. Energy & Fuels 2013;28(1):146-54. Wang Y, Li X, Wendt JO. Ash aerosol and deposition formation mechanisms during air/oxy-combustion of rice husks in a 100 kW combustor. Energy & Fuels 2018;32(4):4391-8. Zhan Z, Wendt JOL. Role of Sodium in Coal in Determining Deposition Rates. Energy & Fuels 2017;31(3):2198-202. Gao Q, Li S, Yang M, Biswas P, Yao Q. Measurement and numerical simulation of ultrafine particle size distribution in the early stage of high-sodium lignite combustion. Proceedings of the Combustion Institute 2017;36(2):2083-90. Li G, Li S, Huang Q, Yao Q. Fine particulate formation and ash deposition during pulverized coal combustion of high-sodium lignite in a down-fired furnace. Fuel 2015;143:430-7. Zhan Z, Tian S, Fry A, Wendt JOL. Formation of Ash Aerosols and Ash Deposits of Coal Blends. Clean Coal Technology and Sustainable Development. Springer; 2016, p. 121-31. [34] [35] [36] [37] [38] [39] [40] Zhang J, Kelly KE, Eddings EG, Wendt JOL. CO 2 effects on near field aerodynamic phenomena in 40kW, co-axial, oxy-coal, turbulent diffusion flames. International Journal of Greenhouse Gas Control 2011;5:S47-S57. Zhan Z, Fry A, Yu D, Xu M, Wendt JOL. Ash formation and deposition during oxy-coal combustion in a 100kW laboratory combustor with various flue gas recycle options. Fuel Processing Technology 2016;141:249-57. Zhan Z, Fry A, Wendt JOL. Deposition of coal ash on a vertical surface in a 100kW downflow laboratory combustor: A comparison of theory and experiment. Proceedings of the Combustion Institute 2017;36(2):2091-101. Morris W, Yu D, Wendt JOL. Soot, unburned carbon and ultrafine particle emissions from air-and oxycoal flames. Proceedings of the Combustion Institute 2011;33(2):3415-21. Zhan Z, Tian S, Fry AR, Wendt JO. Formation of Ash Aerosols and Ash Deposits of Coal Blends. Clean Coal Technology and Sustainable Development. Springer; 2016, p. 121-31. Gallagher NB, Peterson TW, Wendt JOL. Sodium partitioning in a pulverzed coal combustion environment. Symposium (International) on Combustion. 26. Elsevier; 1996:3197-204. Zeng X, Yu D, Fan B. Particulate matter formation characteristics during Zhundong coal combustion at different temperatures. Journal of China Coal Society 2015;40(11):2690-5.