Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite

JFUE 9258

No. of Pages 8, Model 5G

25 May 2015 Fuel xxx (2015) xxx–xxx 1

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel 5 6

4

Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite

7

Qi Gao, Shuiqing Li ⇑, Ye Yuan, Yiyang Zhang, Qiang Yao

8

Key Laboratory for Thermal Science and Power Engineering of Ministry of Education, Department of Thermal Engineering, Tsinghua University, Beijing 100084, China

3

9 1 4 1 2 12 13 14 15 16 17 18 19 20 21 22 23

a r t i c l e

i n f o

Article history: Received 3 March 2015 Received in revised form 5 May 2015 Accepted 14 May 2015 Available online xxxx Keywords: Zhundong lignite Ultrafine PM Flame synthesis Thermophoretic sampling Hencken burner

a b s t r a c t In this work, the pulverized coal combustion of Zhundong lignite was conducted in an optically-accessible, downward Hencken flat-flame burner to investigate the incipient formation of ultrafine particulate matter (PM). The ultrafine PM were collected in-flame from distinct positions under 1800 K, 1500 K and 1200 K using a self-designed thermophoretic sampler, together with extensive conventional transmission electron microscopy (TEM), high-resolution transmission electron microscopy (HRTEM) and energy-dispersive X-ray spectroscopy (EDS) examinations. A novel in situ low-intensity phase-selective laser-induced breakdown spectroscopy (PS-LIBS) was further introduced to diagnose the dynamic behavior of particle-phase sodium during the pulverized coal combustion. The primary particles in the collected ultrafine PM have mean diameter of 8–15 nm, with Si and Na as main mineral components based on TEM/HRTEM–EDS results. The formation of ultrafine PM is regarded as a process of multi-component flame syntheses of mineral precursors, in which the competition between the devolatilization and the coalescence of particles are of most significance. Finally, based on an approach of time scale analysis, three characteristic times of minimum mean diameter of primary particles, the initial increment point of particle-phase sodium and the maximum devolatilization are found to be highly correlated. Ó 2015 Published by Elsevier Ltd.

25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

43 44

1. Introduction

45

It was investigated that coal still takes up 65.7% of the primary energy consumption in China in 2013. The persistent utilization of coal has been claimed as one of main contribution to serious particulate matter (PM) pollution in China. As people start to pay more attention to much smaller PM2.5, PM1.0 or PM0.2 because of healthy and environmental concerns, the collection efficiency of fine particles by electrostatic precipitators (ESP) unexpectedly becomes low, with PM2.5 of 95–99% [1,2] and PM1 of 85–95% [3,4], despite high efficiency of bulk ash particles (PM10+) up to 99.9% or even higher. Nevertheless, the increasing use of low-rank coal (e.g. lignite) in developing countries makes the situation much worse. It is because that such kind of coal often contains a certain amount of easily vaporized species, e.g. ion-exchangeable alkali and alkaline earth metallic (AAEM) species [5,6]. For instance, it was recently reported that an AAEM-rich lignite produces 5.1 times larger amounts of ultrafine PM0.2 than high-rank bituminous coal in the

46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

⇑ Corresponding author at: Department of Thermal Engineering, Tsinghua University, Haidian District, Beijing 100084, China. Fax: +86 10 62773384. E-mail address: [email protected] (S. Li).

furnace, whereas their production amounts of PM2.5 (or PM1) approximately keep the same magnitude [7]. In order to further improve the control of fine PM in both combustion and post-combustion stages, the fundamental studies on the formation mechanism of ultrafine particles, from the bottom level, are of importance and essence. Previously, the characteristics of PM10 emitted from coal combustion were intensively investigated. Seames et al. [8,9] reported that the particle size distributions (PSDs) of PM10 are tri-modal, e.g. ultrafine mode (<0.1 lm), fine fragmentation mode (between 1 and 2 lm) and supermicron mode (>1 lm), respectively. It was successively reported the formation of PM1 mainly results from the volatilization, condensation and aggregation of inorganic species [9–11], the inflation, cracking and bursting of char particles (ceno-sphere structure) and shedding of molten ash due to heating and/or volatile gas expansion [8,9,12–15]. Meanwhile, the formation of PM1–10 was related to the coalescence of mineral matters [16,17] and the fragmentation of char and external minerals [9,18–20]. Recently, Wu and co-workers further clarified, even among PM1, the formation mechanisms of PM0.1–1 and PM0.1 are different [21]. As for ultrafine PM0.1, the distribution was subdivided into five modes, attributed to carbonaceous particles as well

http://dx.doi.org/10.1016/j.fuel.2015.05.028 0016-2361/Ó 2015 Published by Elsevier Ltd.

Please cite this article in press as: Gao Q et al. Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite. Fuel (2015), http://dx.doi.org/10.1016/j.fuel.2015.05.028

61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82

JFUE 9258

No. of Pages 8, Model 5G

25 May 2015 2 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133

Q. Gao et al. / Fuel xxx (2015) xxx–xxx

as size-selective nucleation and growth of refractory oxides and metal nano-particles [22]. Since these PMs, either fine or coarse, were mostly collected from the combustor exit with PSDs and morphology characterization by scanning mobility particle sizer (SMPS)/electrical low pressure impactor (ELPI)/Berner low pressure impactor (BLPI)/aerodynamic particle sizer (APS), computer-controlled scanning electron microscopy (CCSEM), etc. [23,24]. Thus, the direct evidence through in-flame observations is needed to further clarify PM formation mechanism. Due to rapid progresses on flame soot or flame synthesis studies [25], in-flame particle sampling methods (e.g. thermophoretic sampler) may play an important role in studying ultrafine PM formation during pulverized coal combustion. The characterization of chemical compositions in different PM modes is also important. For instance, the mineral components in PM10 mainly include major elements such as Si and Al, minor elements such as Na, Ca, Fe, and trace elements of As, Hg, etc. [26]. Several formation pathways were put forward on the transformation of mineral elements to ultrafine PM. In the stage of char combustion, Quann and Sarofim interpreted the PM1 formation with a high-temperature reaction of MOn (s) + CO (g) = CO2 (g) + MOn1 (g), which governs the volatilization of refractory oxides (with M standing for Si, Fe, Ca and Mg) [27,28]. Nevertheless, Zhang et al. proposed another possibility of PM0.1 formation from the contribution of organically bound mineral elements in volatile matters by qualitatively comparing the pyrolysis and combustion experiments of coals [29]. More recently, Li et al. indicated that the significant amounts of AAEM species (e.g. up to 80% for sodium) were volatilized during the pyrolysis of Victorian Brown coal under the fast heating rate and elevated temperature [5,6]. Furthermore, in the experiment on Zhundong lignite (rich in AAEM species) combustion, the enrichment of sodium (with mass fraction as high as 20–25%) was found to be in ultrafine PM [7]. It implies the important role of easily-vapored mineral precursors in ultrafine PM formation and evolution, which often occurs in an early stage of coal combustion. In order to get such a rapid dynamic releasing of volatiles, several in situ non-intrusive optical diagnostics were introduced [30–33]. Among them, a phase-selective laser-induced breakdown spectroscopy (PS-LIBS) was proven to be one of those promising, because of its capability to distinguish gas-to-particle phase transition of the evaporated minerals [34,35]. In this work, we aim to investigate incipient formation of ultrafine PM during pulverized coal combustion of Zhundong lignite in an optically-accessible, downward Hencken flat-flame burner. A self-designed thermophoretic sampler was used for in-flame sampling of ultrafine PM formed in the early stage of coal combustion, after which an extensive TEM/HRTEM–EDS examination was done. Then, in situ PS-LIBS was introduced to diagnostic the dynamic behaviors of Na release during combustion. Finally, on the basis of these data, sodium element in ultrafine PM and the effect of temperature on their formation are particularly discussed.

134

2. Experimental

135

2.1. Downward Hencken burner system

136

The combustion of Zhundong lignite was performed in an optically-accessible, downward Hencken flat-flame burner that consists of hundreds of multi-element, non-premixed flamelets. Fig. 1 schematically illustrates the details of the set-up. Different from our previous studies [35,36], here we instead introduce an down-fired system for a further reduction of the radial dispersion of particles (as well as holding the rigidity of coal particle streams along the centerline). As shown in Fig. 1, the down-fired system is consisted of the aforementioned Hencken burner and a thin

137 138 139 140 141 142 143 144

furnace sheath to avoid the divergence of flow field. A novel coal feeder, based on the principle of de-agglomeration via high-frequency vibration [36], is utilized to offer a stream of well-dispersed coal particles. The Hencken burner, with a structure of hundreds of I.D. 1.5 mm stainless steel tubes embedded in an O.D.64 mm super-alloy honeycomb, can provide a uniform hot-gas environment for coal combustion. The heating rate of the burner is estimated to be as high as 105 K/s. CO stream, added with a small portion of CH4, act as the fuel gas, whereas a mixture of N2 and O2 acts as the oxidizer. Different combustion conditions can be flexibly achieved via controlling the flow rate of the fuel and oxidant gas by mass flow controllers (MFCs). Inside the furnace sheath, four heat-resisting quartz windows are embedded to get an optical diagnostics on the process of coal combustion. On the bottom of the furnace a pump with a tee valve is installed to offset the influence of thermal buoyancy during the sampling. Three typical temperatures, 1800 K, 1500 K and 1200 K, were chosen for the combustion of Zhundong lignite in this paper. The volume concentration of oxygen in the combustion product gas from Hencken burner was set as 20% for studying coal combustion of all cases, via the adjustment of the flow rates of both fuel and oxidizer streams (as seen in Table 1). The well-dispersed coal particles, carried by N2 with a steady flow rate of 0.085 g/min, were injected into the Hencken Burner through an O.D. 2.5 mm stainless steel tube located in the honeycomb center. The temperatures of hot-gas ambiences were measured by a B-type thermocouple and the coal particle down-flow velocity was detected using the Phase Doppler Anemometry (PDA) (BSAP60 from DANTEC dynamics). The residence time of coal particles (calculated from their velocity data) and gas temperature along the burner are shown in Fig. 3. The ambient temperature can be preserved for as long as 30 mm beneath the burner rim in different cases. The overall operating conditions are given in Table 1.

145

2.2. The sampling systems and PS-LIBS diagnostics

178

A thermophoretic sampler, proposed and successfully used in field of flame synthesis [33,34] was particularly designed to collect ultrafine PMs in flame during coal combustion, as shown in Fig. 2. It is mainly composed of a cylinder and a stepping motor. There are two ports at both ends of the cylinder, one is always kept open and the other is adjustable. With the control of an extra magnetic valve, the piston inside the cylinder can complete the sampling through the high-speed linear motion, about 0.5–1 m s1, in our experiments. Thus the total residence time of the thermophoresis micro-grid in high-temperature environment can be as short as 30–60 m s based on the size of the honeycomb. Such fast sampling velocity can protect the thermophoresis grid from the damage. Further, a computational fluid dynamics (CFD) simulation was performed to get the temperature field, through which the thermophoresis velocities of ultrafine particles are predicted. Then, the time for particle thermophoresis is estimated to be 0.08–0.3 m s, which is about two orders of magnitude less than the sampling time. It means that there is sufficient time for ultrafine PMs to be collected during the sampling without size selectivity. The stepping motor is introduced to get a precise control of sampling positions, which were set to be 5 mm, 10 mm, 15 mm and 20 mm beneath the exit of the burner. Basically, traditional LIBS technique is incapable to distinguish the distribution of elements in different phases. However, by carefully selecting the laser fluence between the gas phase and particle phase breakdown thresholds, the novel PS-LIBS technique, using lower laser power, can only show selectivity in particle phase atoms, with no breakdown emission from gas phase. It was first discovered by our co-workers [34] in the study of flame synthesis and its feasibility on the diagnostics of coal combustion has been

179

Please cite this article in press as: Gao Q et al. Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite. Fuel (2015), http://dx.doi.org/10.1016/j.fuel.2015.05.028

146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177

180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208

JFUE 9258

No. of Pages 8, Model 5G

25 May 2015 Q. Gao et al. / Fuel xxx (2015) xxx–xxx

3

Fig. 1. Schematic diagram of the optically-accessible, downward Hencken flat-flame burner.

Table 1 Experimental operating conditions in this work. Temperature (K)

CO (L/min)

Operating conditions 1800 5.73 1500 3.94 1200 2.34

CH4 (L/min)

N2 (L/min)

O2 (L/min)

N2(carrier) (L/min)

0.1 0.1 0.1

21.63 20.99 18.67

10.78 8.94 6.86

0.07 0.06 0.05

217

verified and described in detail in our recent work [35]. An Nd:YAG laser generator offers second harmonic (532 nm) beam at 10 Hz as the excitation source. It is then focused by a 750 mm lens to a waist radius of 200 lm and detected by an intensified charge-coupled device (ICCD) camera (Princeton Instruments PI-MAX3 1024i-Unigen2-P43). The incident laser intensity used in this work was 8.8 mJ/pulse, guaranteeing the average intensity through the optical window in the range of the particle-phase sodium saturation intensity [35].

218

2.3. Coal properties

219

Zhundong lignite was used in this work. All coal particles were screened to 65–74 lm and dried before experiments. The

209 210 211 212 213 214 215 216

220

Fig. 2. Set-up of the thermophoretic sampling system.

proximate (dry basis), ultimate (dry and ash free basis) and low temperature ash (LTA, ash basis) analysis are shown in Table 2.

221

3. Results and discussion

223

3.1. Characterization and evolution of the ultrafine PM

224

The in-flame collection of ultrafine particles from coal combustion was done along the vertical positions, e.g. 5, 10, 15 and 20 mm from burner rim under different ambient temperatures of 1800, 1500 and 1200 K. Fig. 4 illustrates all TEM images of these collected ultrafine particles. It indicates that ultrafine particles, formed in an early stage of coal combustion, are mainly spherical in shape. Also, a few amounts of agglomerates consisting of several primary particles are found in the samples. For every case, by using 4–12 HRTEM images, more than 150 particles were counted to obtain the mean diameter of primary particles. Here the mean diameter is defined as a value dividing a sum of diameters of all primary particles by particle counts. The evolutions of primary particle mean diameter along the residence time of coal particles under different ambient temperatures are shown in Fig. 5, Fig. 6 and Fig. 7, respectively, in which the corresponding HRTEM images and the EDS spectra of ultrafine particles are embedded as insets. The results suggest that, for the ultrafine particles incipiently formed during

225

Fig. 3. The profiles of both ambient temperature and residence time of coal particles along the distance from burn.

Please cite this article in press as: Gao Q et al. Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite. Fuel (2015), http://dx.doi.org/10.1016/j.fuel.2015.05.028

222

226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241

JFUE 9258

No. of Pages 8, Model 5G

25 May 2015 4

Q. Gao et al. / Fuel xxx (2015) xxx–xxx

Table 2 Proximate analysis, ultimate analysis and LTA analysis of Zhundong lignite used in this work. Proximate analysis (wt.%, dry basis) Fixed carbon Volatile matter Ash HHV (MJ/kg)

63.54 30.58 5.88 21.89

Ultimate analysis (wt.%, dry and ash free basis) C H O N Stotal Cl

71.6 3.16 23.85 0.78 0.52 0.09

LTA analysis (wt.%, ash basis) SiO2 Al2O3 Fe2O3 CaO MgO TiO2 SO3 K2O Na2O P2O5

14.5 3.96 4.85 39.65 3.48 0.36 25.97 0.69 7.48 N/A

Zhundong lignite combustion, the mean diameter of primary particles ranges around 8–15 nm. For all three ambient temperatures, at identical residence time of 12 m s for coal particles, the mean diameters of collected primary particles are 11.83 nm (1800 K), 11.19 nm (1500 K) and 9.83 nm (1200 K), respectively, exhibiting a small difference among them. For the evolution of particle growths, it is noted that the mean diameters of primary particles decrease firstly and then increase again in the collected zone of this experiment for all three ambiences. The characteristic residence time when the mean diameter approaches the minimum are approximately 6 m s (1800 K), 9 m s (1500 K) and 20 m s (1200 K), respectively. Just at these minimum points, it is discovered that the fractions of smaller primary particles are relatively large, as directly seen in TEM images in Fig. 4. For a further characterization, we particularly calculate the population growth rate, (DN/Dt)/N, of primary particles between two successively adjacent time intervals Dt, where N is the average number of primary particles in HRTEM images (with the same magnification). For instance, in 1800 K ambience these rates are 0.46 m s1 (3.1– 6.3 m s), 0.25 m s1 (6.3–9.3 m s) and 0.04 m s1 (9.3–12.4 m s), respectively. It implies that there is a fast population growth of primary particles around 6 m s. Similarly in much lower 1200 K ambience, the rates are 6.0  103 m s1 (6.5–12.9 m s), 0.16 m s1 (12.9–19.7 m s) and 0.07 m s1 (19.7–26.4 m s), respectively, identifying a fast population growth of primary particles around 20 m s.

1800 K/5 mm/3.1 ms

1500 K/5 mm/4.3 ms

1200 K/5 mm/6.5 ms

1800 K/10 mm/6.3 ms

1500 K/10 mm/8.6 ms

1200 K/10 mm/12.9 ms

1800 K/15 mm/9.3 ms

1500 K/15 mm/12.7 ms

1200 K/15 mm/19.7 ms

1800 K/20 mm/12.4 ms

1500 K/20 mm/16.8 ms

1200 K/20 mm/26.4 ms

Fig. 4. TEM images of the incipient ultrafine particles from Zhundong lignite combustion along the distance from burner (i.e. 5, 10, 15 and 20 mm) under ambient temperatures of 1200 K, 1500 K and 1800 K. (temperature/distance/residence time).

Please cite this article in press as: Gao Q et al. Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite. Fuel (2015), http://dx.doi.org/10.1016/j.fuel.2015.05.028

242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266

JFUE 9258

No. of Pages 8, Model 5G

25 May 2015 Q. Gao et al. / Fuel xxx (2015) xxx–xxx

Fig. 5. The morphology, components and evolution of the incipient ultrafine particles formed under 1800 K.

267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290

On the basis of aforementioned results, we may further hypothesize the formation and growth of primary particles in the early coal combustion stage as a process of multi-component flame syntheses of mineral precursors coupled with carbonaceous soot formation. With the release of volatiles during coal pyrolysis, the multi-component, organically-bonded metal precursors (e.g. low-vapor-pressure species) easily co-evaporate and then perform a series of reactions (e.g. oxidation, chloridization, sulfation, etc.) to form product vapors. The product vapors promptly converts to condensed ultrafine particles through a homogenous (or spontaneously heterogeneous) nucleation, which further grow by a successive collision–coalescence mechanism among newly formed particles or a heterogeneous nucleation of vapors on particle surfaces. This is generally termed as a gas-to-particle conversion process in the aerosol community [33,34,37]. The results shown in Figs. 5–7 are well consistent with this hypothesis. That is, the minimum of mean diameters of primary particles at a certain time can be attributed to the balance among the fast nucleation rates due to temperature-dependent devolatilization, the temperatureinsensitive collision rates, and temperature-dependent coalescence rates [38]. For all cases, the positive population growth rates of primary particles indicate the high overlapping of the devolatilization time scale and the coalescence time scale, since the nucleation are thermodynamically approved and ultrafast without energy barrier.

Fig. 6. The morphology, components and evolution of the incipient ultrafine particles formed under 1500 K.

5

An exception of 0.04 m s1 (9.3–12.4 m s) in 1800 K ambience suggests the fast coalescences among primary particles.

291

3.2. Sodium in the incipient ultrafine PM

293

In order to examine inorganic components of ultrafine particles, the energy dispersive X-ray spectra (EDS) were performed for the related samples in three ambient temperatures, as given in the insets of Figs. 5–7. It is noted that Mo and Cu elements are the components of thermophoresis grids and will not be of interest here. It demonstrates that the incipient ultrafine PM, formed in the early combustion stage of Zhundong lignite, mainly contain the refractory element such as Si, accompanied by the volatile Na, whereas the amounts of Al and Mg are pretty limited. For instance, at the residence time of 12.7 m s under 1500 K, Si and Na contributes 69.5% and 23.0% to the total mass of inorganic components in incipient ultrafine PM, while the sum of Mg and Al only account for 7.5%. Under all cases, Si and Na were detected after the sampling positions of 10 mm beneath the exit of the burner. Nevertheless, as for the samples collected at 5 mm, there are so few ultrafine particles that their characteristic X-ray peaks of inorganic elements are covered by the background, e.g. Mo and Cu, as shown in the inset of Fig. 5. It was reported that Na exists in low-rank coal with two forms, the ion-exchangeable carboxylates and the water-soluble salts, e.g. NaCl [5]. The volatile Na, K, Cl, P, etc. commonly contribute to PM1, in particular PM0.1, via the nucleation–coagulation mechanism as well as series of surface reactions during coal combustion [10,21,24]. Especially for high-sodium lignite, a large amount of Na is believed to be released in the early stage of coal combustion. EDS spectral in Figs. 5–7 provide direct experimental evidence that Na exists in the incipient ultrafine particles, even under high ambient temperature. Nevertheless, the saturated vapor pressure of some sodium compounds, such as NaCl and Na2SO4, can be high under elevated temperature, causing it difficult for their vapor to experience the gas-to-particle conversion process. Thermodynamic calculations indicated that Na2SO4 is stable only below 1300–1400 K in coal-fired flue gas atmosphere while the formation of sodium silicate, such as Na2Si2O5, is favored in the range of 1300–1850 K [39]. It was further predicted that the sodium silicate is the main sodium compound in a non-reducing atmosphere of 1300–2000 K when the molar ratio of Na:Cl is larger than unity, which well coincides with the cases in this work since Zhundong lignite possesses a high Na:Cl ratio around 16. In addition, another possible reaction scheme

294

Fig. 7. The morphology, components and evolution of the incipient ultrafine particles formed under 1200 K.

Please cite this article in press as: Gao Q et al. Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite. Fuel (2015), http://dx.doi.org/10.1016/j.fuel.2015.05.028

292

295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332

JFUE 9258

No. of Pages 8, Model 5G

25 May 2015 6

Q. Gao et al. / Fuel xxx (2015) xxx–xxx

334

from NaCl to Na2SiO3 in coal flame in literature was also proposed as below [40]:

335 337

2NaCl þ SiO2 þ H2 O Na2 SiO3 þ 2HCl

333

372

Here the reactant SiO2 may come from the oxidation of organically-bonded silicon precursors [29] or silicon monoxide [23] released in coal combustion. Based on the EDS spectral and above analysis, compounds like Na2OnSiO2 seem to be the main form of sodium existence in incipient ultrafine PM. In order to further reveal the relationship between the phase transition of Na and the formation of incipient ultrafine PM, we particularly employed a novel optical technique PS-LIBS to diagnose the behavior of Na in the early combustion stage of Zhundong lignite. Different from the traditional LIBS technique, the PS-LIBS technique is capable to distinguish atoms in different phases, showing selectivity-only atoms in particle phase with no breakdown emission from gas phase [33–35]. As shown in the inset of Fig. 8, two distinct peaks in the spectra, fit with the characteristic emission lines of atomic 1Na (red line) referring to the NIST database [41], were successfully detected during Zhundong lignite combustion. Fig. 8 also illustrates PS-LIBS signals of Na along the vertical distance beneath the burner in the ambiences of 1200 K, 1500 K and 1800 K, respectively. The normalized intensity of PS-LIBS signal was obtained by integrating the emission intensity from 450 nm to 600 nm. It demonstrates that under 1800 K and 1500 K the normalized intensity, related to the volume fraction of particle phase Na, first decrease slightly and then increase obviously (particularly under 1800 K) with the increase of the residence time. While under 1200 K, there is no apparent changes until a slow increasing rate of the signal intensity after the residence time of 20 m s. The characteristic residence time when the amount of particle-phase Na starts to increase are about 7 m s, 10 m s and 21 m s under 1800 K, 1500 K and 1200 K, respectively. They are remarkably consistent with the time scales that are associated with the minimum mean diameter of primary particles. It states that the appearance of smaller particles is possible to be closely related to the gas-to-particle conversion of sodium precursors. This can be further verified by the increment of Na fraction in the ultrafine PM as a function of the residence time, as tabulated in the inset of Fig. 8, based on the EDS spectral in Figs. 5–7.

373

3.3. Effect of ambient temperature on the ultrafine PM formation

374

It was reported that the ambient temperature plays an important role in several aspects of pulverized coal combustion such as ignition delays, pyrolysis properties and combustion kinetics [36,42–44]. It is known that the ambient temperature also actively affect both the intrinsic reaction rate of carbon oxidation and the diffusion of oxygen to the coal particle surface [45,46]. More recently, in the early stage, a transition from the heterogeneous ignition mode to the hetero-homogeneous ignition mode was found as the ambient temperature increases from 1200 K to 1800 K due to the competition between the heating time scale and the devolatilization one [36]. Here, a similarly theoretical approach on the basis of time-scale analysis is introduced to interpret the role of ambient temperatures in the ultrafine PM formation in the early combustion stage of Zhundong lignite. As shown in Figs. 5–7, it is discovered that the higher the ambient temperature is, the earlier the counted primary particles approaches their minimum value of mean diameters. As discussed above in Section 3.1, this minimum is attributed to a competition between the fast devolatilization rates and coalescence rates, which are both sensitively temperature dependent. Here, a simple

338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371

375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393

1

For interpretation of color in Fig. 8, the reader is referred to the web version of this article.

Fig. 8. Dynamic behavior of the PS-LIBS signals of Na along the distance from burner under ambient temperatures of 1200 K, 1500 K and 1800 K.

one-step first-order kinetic model is adopted to describe the dynamic release of volatiles [47],

394

  dV E ðV 1  VÞ ¼ kðV 1  VÞ ¼ K exp  dt RT p

396

ð1Þ

Here, V and V1 stand for the released and total mass fraction of volatiles, respectively. For Zhundong lignite, the total mass fraction of volatiles is about 0.3058, as presented in Table 2. The kinetic parameters, K and E, are selected from the literature [47]. Tp represents the surface temperature of coal particles, which can be predicted based on the heat balance equation during the pyrolysis,

 
dT qcp V p p ¼ hS T g  T p þ erS T 4w  T 4p dt

395

398 399 400 401 402 403 404

405

ð2Þ

where Vp, S, h, cp, q and e are the volume, surface area, heat transfer coefficient, heat capacity, density and emissivity of coal particle, respectively. The wall temperature Tw and the gas (ambient) temperature Tg are estimated based on the experimental data shown in Fig. 3. In addition, the change of the coal particle density in the devolatilization process is also considered (q0 refers to the initial density of coal particles),

407 408 409 410 411 412 413 414

415

dV dðq=q0 Þ ¼ dt dt

ð3Þ

The predicted results are illustrated in the inset of Fig. 9. The solid line stands for the amounts (mass fraction) of the volatiles released and the dash line stands for the devolatilization rate. It takes only 2.5 m s, more or less, for the release of volatiles to finish under 1800 K while this process lasts for about 20 m s under 1500 K and 210 m s under 1200 K, based on the calculation. Besides, the devolatilization rate in different ambiences goes up firstly, then down, with a maximum approached at 5.0 m s under 1800 K, 11.0 m s under 1500 K and 19.8 m s under 1200 K, respectively. The maximum devolatilization rate under 1800 K is highest, which is approximately 8.5 times and 55.0 times than those under 1500 K and 1200 K. Fig. 9 further illustrates the comparison of three characteristic time scales for all cases, which, respectively, correspond to the minimum mean diameter of primary particles, the initial increment point of particle-phase sodium, and the maximum devolatilization rate. For a convenience, these three characteristic time scales are termed as smpp, sp-Na and spyr, as marked in Fig. 9. As expected, three time scales are well consistent, implying the strong correlations physically. Since the collision rates between ultrafine particles are temperature-insensitive, it can be inferred

Please cite this article in press as: Gao Q et al. Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite. Fuel (2015), http://dx.doi.org/10.1016/j.fuel.2015.05.028

417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437

JFUE 9258

No. of Pages 8, Model 5G

25 May 2015 Q. Gao et al. / Fuel xxx (2015) xxx–xxx

Fig. 9. Calculated released volatile and time scaling analysis under different ambient temperatures.

Table 3 Comparisons of the saturation partial pressure under different ambient temperatures.

a b c d

Substances

Types

P s1800K =P s1500K

Naa Na2Ob NaClc NaOHd

Metals Oxides Salts Alkalies

1.1 5.1 10.8 10.4

Thermochemistry Thermochemistry Thermochemistry Thermochemistry

data data data data

in in in in

the the the the

reference reference reference reference

of of of of

P s1800K =P s1200K 1.4 58.7 419.7 346.1

[48]. [48]. [49]. [50].

465

that the faster devolatilization rates under higher ambient temperature prevail over the coalescence rates, resulting in an initial increment of smaller Na-rich PMs via nucleation mechanism. In addition, as ambient temperature is high, the amounts of particle-phase Na formed in the early combustion stage of Zhundong lignite is relatively large, as shown in Fig. 8. Generally, a higher ambient temperature results in a higher saturation vapor pressure, which somewhat inhibits the gas-to-particle conversion. However, this effect is practically offset by a rapid release of Na element in the fast devolatilization process under elevated temperature. So we compared the saturation ratios of Na element under different ambient temperatures. The saturation ratio is defined as P/Ps, where P refers to the actual partial pressure and Ps refers to the saturation vapor pressure. Considering the complexity of sodium existence in the early combustion stage of Zhundong lignite, we select sodium and its three compounds, reported stable thermodynamically above 1200 K [39], to substitute, as given in Table 3. For instance, when the pyrolysis in 1800 K ambience ends, the actual partial pressure of Na element under 1800 K is predicted to be about 100 times than 1500 K and 900 times than 1200 K based on the calculations, assuming that the release rate of Na element is linear to the devolatilization rate. Whereas, for the sodium-related substances shown in Table 3, their saturation vapor pressure under 1800 K is 1.1–10 times than 1500 K and 1.4–400 times than 1200 K, respectively. In consequence, the saturation ratio of Na element under high temperature ambience is inferred to relatively high in the early combustion stage, which, since then, causes the large amounts of particle-phase Na as clearly shown in Fig. 8.

466

4. Conclusion

467

The formation and evolution of ultrafine PM in the early combustion stage of Zhundong lignite was studied by using an

438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464

468

7

optically-accessible, downward Hencken flat-flame burner. A theoretical time-scale analysis was employed to examine the effect of ambient temperatures on the incipient ultrafine PM formation. The main conclusions are drawn as below:

469

(1) The formation and growth of ultrafine PM in the early combustion stage of Zhundong lignite is regarded as a process of multi-component flame syntheses of mineral precursors. Among all time scales, the competition between the devolatilization and the coalescence process determines the evolution of ultrafine particles, resulting in the minimum mean diameter of primary particles at a certain time. (2) For sodium-rich Zhundong lignite, Si and Na are found to be main inorganic elements in the incipient ultrafine PM in all ambiences, and the as-synthesized Na2OnSiO2 is further inferred on the basis of EDS results. (3) The initial increment point of particle-phase Na, detected by in situ PS-LIBS technique, occurs at a residence time of about 7 m s, 10 m s and 21 m s under 1800 K, 1500 K and 1200 K, respectively. It is found that three time scales of minimum mean diameter of primary particles, the initial increment point of particle-phase sodium and the maximum devolatilization rate are well consistent and correlated. The earlier increment of smaller Na-rich ultrafine particles in higher 1800 K ambience can be attributed to a fact that the faster devolatilization prevails over the coalescence of particles.

473

470 471 472

474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495

Acknowledgements

496

This work is mainly funded by National Natural Science Foundation of China (51390491, U1361201) and by National Key Basic Research and Development Program of China (2013CB228501). Special thanks are due to Prof. Stefan D. Tse at Rutgers University, Prof. Marshall B. Long at Yale University and Prof. Richard L. Axelbaum at Washington University in St. Louis for useful discussion.

497

References

504

[1] Yao Q, Li S, Xu H, Zhuo J, Song Q. Studies on formation and control of combustion particulate matter in China: a review. Energy 2009;34(9):1296–309. [2] Zhuo J, Li S, Yao Q, Song Q. The progressive formation of submicron particulate matter in a quasi one-dimensional pulverized coal combustor. Proc Combust Inst 2009;32(2):2059–66. [3] Helble JJ. A model for the air emissions of trace metallic elements from coal combustors equipped with electrostatic precipitators. Fuel Process Technol 2000;63(2):125–47. [4] Jin X. Research on the capture technology of fine particles in electrostatic precipitator. Master Thesis. Tsinghua University; 2013. [5] Quyn DM, Wu H, Li C-Z. Volatilisation and catalytic effects of alkali and alkaline earth metallic species during the pyrolysis and gasification of Victorian brown coal. Part I. Volatilisation of Na and Cl from a set of NaClloaded samples. Fuel 2002;81(2):143–9. [6] Li C-Z, Sathe C, Kershaw JR, Pang Y. Fates and roles of alkali and alkaline earth metals during the pyrolysis of a Victorian brown coal. Fuel 2000;79(3):427–38. [7] Li GD, Li SQ, 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. [8] Seames WS. An initial study of the fine fragmentation fly ash particle mode generated during pulverized coal combustion. Fuel Process Technol 2003;81(2):109–25. [9] Linak WP, Miller CA, Seames WS, Wendt JO, Ishinomori T, Endo Y, et al. On trimodal particle size distributions in fly ash from pulverized-coal combustion. Proc Combust Inst 2002;29(1):441–7. [10] Neville M, Quann RJ, Haynes BS, Sarofim AF. Vaporization and condensation of mineral matter during pulverized coal combustion. In: Symposium (international) on combustion, vol. 18(1); 1981. p. 1267–74. [11] Clarke LB. The fate of trace elements during coal combustion and gasification: an overview. Fuel 1993;72(6):731–6.

505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535

Please cite this article in press as: Gao Q et al. Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite. Fuel (2015), http://dx.doi.org/10.1016/j.fuel.2015.05.028

498 499 500 501 502 503

JFUE 9258

No. of Pages 8, Model 5G

25 May 2015 8 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590

Q. Gao et al. / Fuel xxx (2015) xxx–xxx

[12] Zhuo J, Li S, Song Q, Yao Q. Analysis of individual PM1 from pulverized-coal combustion. In: Symposium (international) on coal combustion, vol. 6; 2007. p. 837–42. [13] Srinivasachar S, Helble JJ, Boni AA. Mineral behavior during coal combustion 1. Pyrite transformations. Prog Energy Combust Sci 1990;16(4):281–92. [14] Fix G, Seames W, Mann M, Benson S, Miller D. The effect of combustion temperature on coal ash fine-fragmentation mode formation mechanisms. Fuel 2013;113:140–7. [15] Linak WP, Miller CA, Wendt JO. Comparison of particle size distributions and elemental partitioning from the combustion of pulverized coal and residual fuel oil. J Air Waste Manag Assoc 2000;50(8):1532–44. [16] Wang Q, Zhang L, Sato A, Ninomiya Y, Yamashita T. Effects of coal blending on the reduction of PM 10 during high-temperature combustion 1. Mineral transformations. Fuel 2008;87(13):2997–3005. [17] Yan L, Gupta RP, Wall TF. The implication of mineral coalescence behaviour on ash formation and ash deposition during pulverised coal combustion. Fuel 2001;80(9):1333–40. [18] Helble JJ, Sarofim AF. Influence of char fragmentation on ash particle size distributions. Combust Flame 1989;76(2):183–96. [19] Yoshiie R, Tsuzuki T, Ueki Y, Nunome Y, Naruse I, Sato N, et al. Effects of coal types on ash fragmentation and coalescence behaviors in pulverized coal combustion. Proc Combust Inst 2013;34(2):2895–902. [20] Shah KV, Cieplik MK, Betrand CI, Van de Kamp WL, Vuthaluru HB. A kineticempirical model for particle size distribution evolution during pulverised fuel combustion. Fuel 2010;89(9):2438–47. [21] Gao X, Li Y, Garcia-Perez M, Wu H. Roles of Inherent fine included mineral particles in the emission of PM10 during pulverized coal combustion. Energy Fuels 2012;26(11):6783–91. [22] Carbone F, Beretta F, D’Anna A. Factors influencing ultrafine particulate matter (PM0.1) formation under pulverized coal combustion and oxyfiring conditions. Energy Fuels 2010;24(12):6248–56. [23] Li G, Li S, Dong M, Yao Q, Guo CY, Axelbaum RL. Comparison of particulate formation and ash deposition under oxy-fuel and conventional pulverized coal combustions. Fuel 2013;106:544–51. [24] Jia Y, Lighty JS. Ash particulate formation from pulverized coal under oxy-fuel combustion conditions. Environ Sci Technol 2012;46(9):5214–21. [25] Zhao B, Yang Z, Johnston MV, Wang H, Wexler AS, Balthasar M, et al. Measurement and numerical simulation of soot particle size distribution functions in a laminar premixed ethylene–oxygen–argon flame. Combust Flame 2003;133(1):173–88. [26] Quann RJ. Ash vaporization under simulated pulverized coal combustion conditions. Ph.D Thesis, Massachusetts Institute of Technology; 1982. [27] Lee CM, Davis KA, Heap MP, Eddings E, Sarofim A. Modeling the vaporization of ash constituents in a coal-fired boiler. Proc Combust Inst 2000;28(2):2375–82. [28] Senior CL, Panagiotou T, Sarofim AF, Helble JJ. Formation of ultra-fine particulate matter from pulverized coal combustion. Abstracts of papers of the American Chemical Society, vol. 219. American Chemical Society; 2000. p. 671–2. [29] Zhang L, Ninomiya Y, Yamashita T. Occurrence of inorganic elements in condensed volatile matter emitted from coal pyrolysis and their contributions to the formation of ultrafine particulates during coal combustion. Energy Fuels 2006;20(4):1482–9. [30] Van Eyk PJ, Ashman PJ, Alwahabi ZT, Nathan GJ. Quantitative measurement of atomic sodium in the plume of a single burning coal particle. Combust Flame 2008;155(3):529–37.

[31] Häyrinen V, Hernberg R, Aho M. Demonstration of plasma excited atomic resonance line spectroscopy for on-line measurement of alkali metals in a 20 kW bubbling fluidized bed. Fuel 2004;83(7):791–7. [32] Hsu LJ, Alwahabi ZT, Nathan GJ, Li Y, Li ZS, Aldén M. Sodium and potassium released from burning particles of brown coal and pine wood in a laminar premixed methane flame using quantitative laser-induced breakdown spectroscopy. Appl Spectrosc 2011;65(6):684–91. [33] Zhang Y, Li S, Ren Y, Yao Q, Law CK. Two-dimensional imaging of gas-toparticle transition in flames by laser-induced nanoplasmas. Appl Phys Lett 2014;104(2):023115. [34] Zhang Y, Xiong G, Li S, Dong Z, Buckley SG, Tse SD. Novel low-intensity phaseselective laser-induced breakdown spectroscopy of TiO2 nanoparticle aerosols during flame synthesis. Combust Flame 2013;160(3):725–33. [35] Yuan Y, Li S, Yao Q. Dynamic behavior of sodium release from pulverized coal combustion by phase-selective laser-induced breakdown spectroscopy. Proc Combust Inst 2015;35(2):2339–46. [36] Yuan Y, Li S, Li G, Wu N, Yao Q. The transition of heterogeneous–homogeneous ignitions of dispersed coal particle streams. Combust Flame 2014;161(9):2458–68. [37] Tolmachoff ED, Abid AD, Phares DJ, Campbell CS, Wang H. Synthesis of nanophase TiO2 crystalline films over premixed stagnation flames. Proc Combust Inst 2009;32(2):1839–45. [38] Xing YC, Köylü ÜÖ, Rosner DE. Synthesis and restructuring of inorganic nanoparticles in counterflow diffusion flames. Combust Flame 1996;107(1):85–102. [39] Wibberley LJ, Wall TF. Alkali-ash reactions and deposit formation in pulverized-coal-fired boilers: the thermodynamic aspects involving silica, sodium, sulphur and chlorine. Fuel 1982;61(1):87–92. [40] Raask E. Mineral impurities in coal combustion: behavior, problems, and remedial measures. Taylor & Francis; 1985. pp. 92–3. [41] Ralchenko Y, Kramida AE, Reader J . [42] Jamil K, Hayashi J, Li CZ. Pyrolysis of a Victorian brown coal and gasification of nascent char in CO2 atmosphere in a wire-mesh reactor. Fuel 2004;83(7):833–43. [43] Kobayashi H, Howard JB, Sarofim AF. Coal devolatilization at high temperatures. In: Symposium (international) on combustion, vol. 16(1); 1977. p. 411–25. [44] Coda B, Tognotti L. The prediction of char combustion kinetics at high temperature. Exp Thermal Fluid Sci 2000;21(1):79–86. [45] Hurt R, Sun JK, Lunden M. A kinetic model of carbon burnout in pulverized coal combustion. Combust Flame 1998;113(1):181–97. [46] Hurt RH, Calo JM. Semi-global intrinsic kinetics for char combustion modeling. Combust Flame 2001;125(3):1138–49. [47] Fu WB, Zhang YP, Han HQ, Wang DF. A general model of coal particle devolatilization. Sci China (Ser A) 1988;32(7):887–96. [48] Lamoreaux RH, Hildenbrand DL. High temperature vaporization behavior of oxides. I. Alkali metal binary oxides. J Phys Chem Ref Data 1984;13(1):151–73. [49] Stull DR. Vapor pressure of pure substances. Organic and inorganic compounds. Ind Eng Chem 1947;39(4):517–40. [50] Chase MW. NIST-JANAF T T. 4th ed. Monograph 9. J. Phys. Chem. Ref. Data 1998: 1–1951.

Please cite this article in press as: Gao Q et al. Ultrafine particulate matter formation in the early stage of pulverized coal combustion of high-sodium lignite. Fuel (2015), http://dx.doi.org/10.1016/j.fuel.2015.05.028

591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642