ML20106A354
ML20106A354 | |
Person / Time | |
---|---|
Site: | Harris |
Issue date: | 06/15/1984 |
From: | Fisher G, Natusch F CALIFORNIA, UNIV. OF, DAVIS, CA, COLORADO STATE UNIV., FORT COLLINS, CO |
To: | |
References | |
I-EDDL-1, OL, OL-I-EDDL-1, NUDOCS 8408170190 | |
Download: ML20106A354 (49) | |
Text
{{#Wiki_filter:. _.. ,;'n h~ h b
- d h
l 3+ _7 g g.g y [: 6.289 cm y g ^?; ' GGloir UCD 472-502 ,h DOE Research and Development Report in the m,tt,, W ~imita -11 UC-48, Bicmedical and Environmental Research ~ 44%g p, ; Stall j) ,, g ~ ,6 {f:Z ")@fpad ~ a o,,, ~ ~_ __ _m~ ~c ~, ~lf[- dlf[ M .) Othes .a ~ \\ m; iC f dii l.w*
- -j SIZE DEPENDENCE OF H
THE PHYSICAL AND CHEMICAL L?) PROPETITIES OF FLY ASH fg p,1 cund i;jj pa f~cA 'N G.L. Fisher-m ..,a,., a " - Eiid .- 6. u < /.. .d D.F.S. Natusch ,..a.- m. a m.. m g
- . y k
, 71..@ L,<
- ..b g-l
'o ; l ty Rac,iobiology l "; l} Laboratory ~ ~ 1 ., = t.u
- a ' \\ -,,
,1,.. a v& ~ r.,.,.. I y 5 * ' '4 i !?... - I i ' e ' '. I ch 8408170190 840615 , f t,- ii ic
- 3.. h PDR ADOCK 05000400
!* -,L G PDR - - - - - - ~ ~ ~ ~ ~ ~ ~ ~ ^ ~ ~ ~ ~ ~ ~
1 ..). UCD 472 - 502 / a.. -v(. h SIZE DEPENDENCE OF THE PHYSICAL AND CHEMICAL PROPERTIES OF COAL FLY ASH h G. L. FISHER .' 4 y Radiobiology Laboratory u . University of California y Davis, Califirnia 95616 ,1 y
- E D. F. 5. NATUSCH
- t
- 3l Department of Chemistry Colorado State University
~w .O Fort Collins, Colorado 80523
- e;N
- .y O
}ls ' 3 1 5N1 .C M.a c T - ~. : L( Spring,1979 ),$.T C ,.yd .~.;. i., !J This report reproduces a chapter J.., in: Analytical Methods for Coal .b and Coal Products, Vol'ume III (C. Karr, Jr., editor), Academic Press, pp. 489 - 541, 1979. e.,1 '&:it ' kr '.9 .t .I This research was supported by the US Department of Energy. i l a 1
~" **. I. INTRODUCTION "/ In order to assiss the environmIntal significance and potential health hazards associated with exposure to environmental pollutants, detailed studies of physical and f, cheatcal properties are required. It is these properties that determine the route and biological consequences of exposure. The aerodynamic behavior of aerosols released during coal combustion will determine the potential for atmospheric transport and subsequent human exposure. Large particles (>10 um) escaping the power plant's control technology will fall out near the plant, which may ultimately result in general population exposure by ingestion of agricultural products or water. Thus, exposure of agricultural products V by soil or foliar deposition or contamination of water sources in the environs of the power plant will reflect, for the most part, the chemical composition of the larger .u particles. Long-range transport and general population exposure will be associated with .[ the more stable aerosols. These f.ine carticles (<10 um) are of special interest because .,h they are 1.ess efficiently coll _ecteiby existing control ter.tnolonies-hava a ra1Mvfv-i 'j .lonb atnoscheric residence time, and upon inhalation, are efficient 1v dem 4*ad =A e'e ly renoved from the pulmonary region of the respiratory tract. N In a review of pafticulat'e abatemenTr.ecnnologies, Vandegrift et'a1. (1973) described ~ p collection efficiency as a function of particl.e size for a variety of control technologies g including electrostatic precipitators, fabric filters, wet scrubbers, and cyclones. y Average collection efficiencies for a medium-efficienev electrostatic precioitator I se were'90, 70,J DIl 3 M " " 1 0 0.1. and 0.01 um pantirlae eg:n.,4
- nia csDhgly, jj the Venturi wet scrubber (WS) was more efficient (99.5%) for 1.0 um particles and less efficient (<11) for 0.01 um particles. A crossover in the ESP-and WS-efficiency curves U.
was observed at 0.35 us. Respiratory tract deposition of inhaled particles is determir.ert by the physics and chemistry of aerosals, the anatomy of the respiratory tract, and the airficw patter ns in a the lung airways (Yeh et al.,1976). The most important physical factors affecting lung 2,~ deposition of inhaled particles are the aerodynamic properties of the aerosol and the "l chenical' reactivity in the airways. Lung deposition is generally described in terms of frac't'ional particulat'e deposition by mass or number in the three major regions of the ~ Nl4 respiratory tract: the nasopharyngeal, tracheobronchial, and pulmonary regiens (Task Group y,
- on Lung Dynamics,1966). The nasopharyngeal region is composed of the nose and throat, d
extending to the larynx; the tracheobronchial region consists of the trachea ar:d bronchial ..f tree, includine the terminal bronchioles; and the pulmonary region censists of the p resotratory bronchioles and the alveolar structures. Farticles gretter than 10 um are 4 effectively collected in the nasopharyngeal region; tracheobronchial and pulmonary depo-I 1 sition generally increase with decreasing particle size. Fractionai deposition in the , f. v ;' pulmonary region ranges from 30 to 60t of the inhaled aerosol fer particles rangir.g in size from 1.0 to 0.01 um (Task Group on Lung Dynamics,1966). Similarly, tracheo-4 bronchial deposition ranges from 5 to 30% for inhaled aerosols f rom 1.0 to 0.01 um, respectively. Respiratory tract deposition profiles have been calculated for f ron, lead, and t.enzo(a) pyrene in urban aerosols (Natusch and Waliace,1974). The hygroscopicity or J I 1 L J
./..: l d% t ';Qrp jy &%#d. ir:::N _,4 :, l@ y % N D.; y : :ch..jT 2 ? f. e y 9 [ '7 @ T reactivity of an aerosol in the airways may dramatically alter the particle size and 8 ) the regional deposition. Parks et al. (1977) have shown that, upon inhalation, ammont E sulfate aerosols with initial aerodynamic diaraters of 0.8 um and 8% relative humidity ~ may rapidly grow to 2.3 um in the water vapor saturated atmosphere of the respiratory -S'j tract. The rapid growth of the aerosols resulted in deposition predominantly in the J nasopharyngeal region and lower than expected deposition in the tracheobronchial and d pulmonary regions. 3 The rate of clearance of deposited particulate matter from the respiratory tract '3 will be determined, in part, by the chemical behavior in the lung's unique micreenviron-3 ment in the vicinity of the particle. Hygroscopic particlu deposited in the respiratory tract will be rapidly cleared by dissolution and subsequent passage into the bloodstream (,"j for ultimate exposure of internal organs. Less soluble particles deposited on the * .,6 ' mucocilliery escalator of the tracheobronchial region and on the ciliated epithelium of
- i the nasopharyngeal region will be rapidly cleared with half-times on the order of one
') day and a few minutes, respectively (Task Group on Lung Dynamics,1966). Relatively in-soluble particles deposited in the pulmsnary region will be phagocytized by the ]- pulmonary alveolar macrophages (PAM). These particles will be slowly removed by either ] dissolution within PAM or_ transport within PAtt to the mucocilliary escalator. The biological half-time of material in the pulmonary region is very much a, function of particulate chemical composition;. half-times of hundreds of days have been reported for f.{j insoluble particles. -r; It should be emphasized, however, that dissolution of surface-associated chemical components need not be a requisite for their interaction with the biological system. a 7 example, inhaled particles may be phagocytized by macrophages there direct particle 0 ace-cell interaction will take place. A reasonable comparison of " insoluble" .cle interaction may be made with asbestos. [.' In this chapter, the size dependence of physical and chemical properties of coal th is reviewed. Because the size dependence of it.any of +he,,.hemical properties 1 ' rom surface-associated chemic:1 phenomena, a detailed description of surface is provided. An understanding of the bioenvironmental significance of ambient requires a detailed understanding of its checical reactivity and biol'gical '1 T
- tions with fly ash surfaces. This chapter reproduces the material fou.1 in a
-i
- 4 Li. report published through NTIS (Fisher and Natusch,1979).
-2 i II. MORPHOLOCY AND FORMATION OF C0AL FLY ASH f A. Morphological Tnalysis .R Horphological studies by light and electron microscopy have described the hetero-5.{ geneity and structural complexity of coal fly ash. Based on morphological appearance, much can be inferred concerning origin, formation, and chemical composition. McCrone
- j and Delly (1973) indicate that particulate matter derived from combustion products is i
,} readily identified under the light microscope. The fused glassy spheres in coal fly ash 1 are the result of exposure to boiler tenperatures >l200 C. Aside from the water-white 6d glassy spheres, McCrone and Delly (1973) also describe the presence of opaque " magnetite" f spheres and spheres containing trapped gas bubbles, l 2 p
classes us m,. T.Y t.ight microscopy his been used to define 11 major mr.,rpnological 5 l ush particles (Fig.1) in stack-collected, size-fractionated material (Fisher et a., The characteristics employed in morphological characterization were particle 4 The 11 classes include (a) amorphous, nonopaque particles, d 1978). shape ano degree of opacity. (b) amoro$ous,' opaoue. particles, (c) amorphor;s, mixed opaque and nonopa (d) roubed, vesfcular, nonopaque particles, (e) rounded, vesicular, mi,:ed op llow spheres), i k nonopaque particles, (f) angular, lacy, opaque particles. (g) cenosphe-es (ho lid spneres, (j) . (h) pierospheres (sphere filled with other spheres), (1)' nonopaque, so A' morphogenesis ~j opaqn sheres, and (k) spheres with either surface or interna) crystals.' scheme (?!g. 2) has been developed relating the 11 morphological classes to extent ] duratwi of exposure to combustion zone temperatures and probable matrix composition. I Opque amorpheus particles and aragular, lacy, opaque particles were tentative Subseque'nt cs u. oxidized carbonaceous material or f ron oxides (Fisher et al.,1978). SEM-x-ray analysis (Fisher et al.,1979a) indicated that these opaque particles were com- [j Furthermore, cale.ulation ofithe effective atomic posed of low atomic number matricesi number of class b particles based upon Bremstrahlung producticn indicated that this '.;.In 4 class is predominantly composed of elemental carbon (Fisher et al.,1979b). The opaque [.[ spheres (class j) appear to be predominantly magnetite and may be: identified f,Q netic separation or passing a magnet near a liquid count of the sample under a micro-The amorphous and x? scope and (2) by observation of soll clusters of these particles. rounded-vesicular, nonopaque particles (classes a and d) appear to be aluminosilicate -( Roundirq snd vesicularity reflect increased exposure to bniler conditions. i particl es. ~ Further heatirg of these' pa'rticles will give rise to nonopaque spheres tnat are either j Similarly, the m'fxed opaque, nonopaque, \\N# solid, hollo,, or packed with other particles. n,3 amorphous, ar rounded classes will give rise to spherical particles upon increased e The nonopaque, solid spaeres ranged in , g.f - 3 posure to cplun tion conditions in the boiler. Analysis of *1gle particles 40, color from Aate white to yellow to orange and deep red. .o }yg in this class jy SEM-x-ray techniques indicated that the variction in color was ass Cenorphere and plerosphere formation will be 4,} ~ with iron cor tent (Fisher et al., ?979b). Crystals within glassy spheres (as i tyl' ditcussed in detail in the following sections.
- . G..-$
determined by light microscopy) are probably formed by heterogeneous nucleation at t n.m,: In this regard, i surface of the molten silicate droplet (Fisher et al.,1979a). . N. J Gibbon (1978) has demonstrated the presence of mullite crystals within and on the sur- ,l Crystal formation within glassy spheres wys demon- .'. N face of fly. ash particles (Fig. 3). y strated by transmissien electron microscopy (TDi) of hydrofluoric acid-etched replica I'. p gy In this process the f)riginal glassy material is dissolved, but tne ins 7 uble mul'lte 1 - "( ccnfi~d by elect: on-diffracticn analysis. remains. P.ullite str4cture c:9 ,,.3 3 .Ni$ u s
- 4?W[
\\s ?
- ,4 O
i L.~& . ; 1" '._
- ?. '. _ *. j 3
r
- "w e
MIMeier eW 7%Teh.D'r A. A M@ f DY[gMM 1 O e7 ' - ~. - ~ '{
- g,
- ..y
.i .y e 1 B ,20pm 20pm A .::j ..y yn ly .=.4 s , [A)'4 m \\ w %m \\ f 4,, e. .~; W ...'? O -1 _=
==~. -4 W. f M . 4' y 2. 1;;l ft l V ut i s- .+ t F E ',. l ..% ~%
- e. <
I Q;.. ,y,.; .*p, 8 ~ .t . ~ ^ ay#3WPI...d' N
- p
.g t ..v. 'j ,f. *b' .0 .~ y.- s-30pm 20pm .e .;.i n .:-w ,.,s , -/p j$ [. f- ~c ..g 1 = ? $,h h,,
- 11.,
,gy Er.* V /<4 ~ 'c g W ~" ' ~ . 3. '), ,.Opm aM. 20 pm 5 pm- '.:ppj h-J ,' Q.::yt s kn ";..%..:.n* I s
- ,s i t.;' '
at s ? ...i A *y y. " 9,.h! '. i, 4 e [
m .p,g ( y . ~. h,y ! p W-1 ~ .x
- h;
' ' i,, 20 pm. ,+ s s. .o. C '~ g' q';%e ) .hs' O* ^
- ew nn.-
l u "hSO ( 5pm . 4 3 s s '. -. A... [Opn 4 a w m's ._.___.-..J.-. ,.,q 's 'A i. e 'Oy / g/ s iy -(, 30pm es Q Ji9 g_Y ' ", ~ m.$ F ~ ~ ^ 4.. + -i G fr i \\ . ~.' u;J; i ~
- , ~,
tc, ~ - ~ L. n r 15pm 20pm g- _. _ _ _ _ D,..-- - Y@,a vh -- Fig. 1. Light photomicrograph demonstrating ~~ the eleven major morphological s J.S.@ g classes of coal fly ash: (A) amor- ?R d phous, ' ton-opaque particles, 'L (B) amorprous, opaque particles. l Tj (C) ancrphous, mixed opaque anc ip- .gj J non-opaque particles, (D) rounced, l,, vesicular, non-opaque particles,
- %', # ' a,..
-(E) rounded, vesicular, mixed d. i opaque and non-opaque particles. l
- ej
'F) wgultr, lacy, opaque particles, .i ) (G.1 cencspheres, (H) pleroscheres, t.I) non-opaqJe, solid soberes, l g j.- (J) cpaqu? spheres, and (K) spheres i with eitner surface or internal .;j crystals. (Reprinted with permis-15 m si n frot G. L. Fisher, B. A. E , if 2 Prentice, D. Silberman, J. M. Ondov. '.h A. H. Biermann, A. C. Ragaini, and A. R. McFariand, Envircnmental .[ Science and Technoloay J_2,, ~4h~(1978).
- i Ccpyrig'it Tf 78 by tne American Ctemical Society '
4 1 i 5 k - - - -. - - - - - - - - ~ ~ c ~ ~ - - - - - " - ' * " ' ' ' ' ~ ~ ~ * ~ ^ ~ ' ~ ' ~ ' ~ ~ ~ ~ ~ ~ " ~ " ~ ~
a z. y o i 2 i ] opacity l us s t o OPaevt ~ EEIN s =0s.eeacut i.5.T.'f.i - ?'%'N'h *i ttatJ s ';,,,,,,,,, Fig. 2. Fly ash morphogenesis F...,. scheme illustrating ! "ouaw s. a ] probable relationship of l' T s opacity to particle li composition, and relationship of particle l,,6.c, 11 8 i shape to exposure in combustion chamber d? (Reprintedwith ? permission from G. L. 1 P"' -' " "*"'~'"2- - - - d-88 4 Fisher, B. A. Prentice, """) ~ 3
- 0. Silberman, J. M.
Ondov, A. G. Biermann, R. C. Ragaini, and A. R. McFarland, Environmental Science and Technoloay 12, 450 (1978). Copyright 1978 by the American Chemical Society.T- _J 'a 1 'lj. i \\' 'f., .i / '1 ,ri. k.~ g{s ~ Fig. 3. Transmission electron micrograph of a ,.1 Y replica of a fly ash sphere showing - D~ - J '! abundant mullite needles. The M'. '."II C ' h' association of crystals-within-sphere Jd' ,l '/ is retained by the replica; the s j , \\ f original glassy material is dissolved ,/ during the replication process, but 'sk' be e' /s.) nullita is insoluble in HF. (Photos 1 ce: *a?y of G. A. Waits, D. S. McKay, 2pm ' ane 0. t. Gibbon, tyneen s. aonnson Space Center, Houston, Texas.) 1 .il b .D I 4 6
,f Fisher et al. (19781 have quantified the relative abundances of the 11 light- / $ sicroscopically defined morpholegical classes in four size-classified, stack-collected fly ash fractions (McFarland et al.,1977). The four fra'ctions had volume median [ diameters (VMD's) of 2.2, 3.2, 6.3, and 20 pm with associated geometric standard devia-l[ tions (o ) of approximately 1.8 for all fractions. The data in Table I demonstrate that g the relative abundances of all particle classes are size dependent. In particular, only -
- ., J the nonopaque solid spheres increased in abundance with decreasing particle size; all
[,','f other morphological classes appeared to increase in frequency with increasing particle size. Anorphous and vesicular particles (classes a, b, c, d, e, and g) predominated in 1 the coarsest fraction (661 by nisnber), while solid, nonopaque spheres predominated in the finest fraction (871 by nianber). .1 Table I. Relative Abundance (%) of Morphologic Particle Classes in Four Fly Ash Fractionsa A .d Fraction .n b Vm = VMD = VM0 = VM = 7 I Particle class 20 um - 6.3 um 3.2 um 2.2 iss xe (A) Amorphous, nonopaque 7.25 2.13 0.79 0.33 -y '(8) Ancrphous, opaque 0.42 0.18 I% (C) Amorphous, mixed opaque 0.77 0.09 .y and nonopaque iy (D) Kounded, vesicular, nonopaque 12.39 6.67 2.91 2.99
- iy?
(E) Rounded, vesicular, mixed 2.27 0.24 0.03 opaque and nonopaque ( (F) Angtelar, lecy, opaque 1.34 0.57 0.E7 0.33 (G) Nonopaque, cenosphere 41.11 25.22 13.20 7.91 (H) Nenopaque, plerosphere 0.51 0.21 5 (I) Nonopaque, solid sphere 25.56 56.01 79.16 86.99 ( h (J) Opaque sphere 1.56 0.90 0.33 0.24 Ifj (K) Ncnopaque sphere with ~6.80 6.79 3.18 0.95 t ~ crystals q m , 2.*.i " From Fisher et al. (1978c). b Volume median diameter.- i 1 ~~ o l 7 e ~ --,,,.,,......,,,-,,,-,-,--,------------.,,,---r,. ,.,,-,,,,,,,---n,-.--..n, r_---,-,-
\\ 9 r ~' s , 9 s _B.-Morphooenesis, y. T.. 1. Cenosphere Formation' T J 6e mechanism of formatidnlof cenospheres, i.e.. hollow spheres, has been the sub-- ject of a.. number of reports. Raask-(1966) demonstrated that: sphere formation may re. ' ult from melting of mineral inclusions in coal on a nonwetting surface, namely carbon. s He also deconstrated that gas ' generation inside the molten droplet resulte'd in ceno ' sphere formation. He reported two stages of gas evolution. -In the first stage, directly Lafter melting coal-ash slag,-502 and N2 were released. ?The 502 was thought to' result' ~ . o from sulfate decomposition and N2 from air trapped in the melt. Further heating re-- s culted in CD evolution that was catalyzed by addition of fron or f ron oxide to the melt. ~ The author hypothesized that iron carbide was formed at the slag-carbon interfacer and ' ], then reacted with silica resulting in C0 evolution: a 2Fe C + S10 Fe Si + 3Fe + 2C0 W 3 g+ 3 In a subsequent report, Raask (1968) describes-the physical and chemical' properties of cenospheres in pulverized fuel ash collected by the electrostatic precipitators'at.10 power plants. The analysis of major elements indicated that the mass of the cenospheres; consisted of 75-90% aluminosilicate,' 7-10% f ron oxide, and 0.2-0.6% calcium oxide. The mass median diameter of the sieved cenospheres from four power plants rangtJ from 80' to 110 um. Raask (1968).aralyzed the gas content of the cenospheres after breaktng
- the' particles in a hydrogen atmof phere. Approximately 0.2 aan (200C) of gas composed
. of CO2 and N2 was calculcted to.be present in each of the four ashes studied. In con-L trast to hbprevious work '(Raask,1966), no' detectable CO was present. Raask su'ggested - f / a? q 7 ' [ that _the source of the CO2 was the oxidation of carbon by fron oxide: f. 2Fe 023+C We0 + CO (2). 2- ~'? This hypothesis was supported by the observation of a higher Fe0:Fe2 3 ratio in the 0 cenospheres than in the denser. ash. He also speculated that cenosphere nitrogen may. . result from decomposition of silicon nitride: Sf N3 4_+ 6Fe 023 3Sjc + 12Fe04 2N (} + 2 2 -It is also possible that the observed CO ' evolution was due to carbonate mineral de-2 composition. Assuming a diameter of average volume of 100 um, a de'nsity of 0.5 g/cm3 t, and 0.5% Ca0, only 20% of the calcium. need be assor.f ated with carbonate mineral ~to pro-- t vide sufficient CO..In this regard, Fisher et al. (1976) have postulated that CO2 re-2 ~ . leased by crushing fly ash under vacuum (after thorough degassing) was the result of 3 M' carbonate mineral decomposition, which occurred during coal ' combustion. In those studies, CO2 and H O were thought to be due to clay mineral decomposition. In particular. 2 p based on-tha stoichiometry of major elements, Fisher et al. (1976) suggested that'the major clay mineral in the parent coal (western United States) was kaolinite. In a de-tailed -tudy of the transformation of mineral matter in pulverized coal, Saro'im et al. ~ '(1977) demonstrated that the.three major inorganic components in a bituminous coal and lignite were kaolinite, a mi.vture of calcium carbonate and sulfate, and pyrites. These 5- } 8 6 't L a %y y.p 3 ..--+e.r-ee-tem ***
- 'N*~
""'"'"*t-
m 4* 4 authors estimated'that the mineral matter in the Ituriinous coal was 50; kaolinite, 40t ^ pyrite, and 10% calcium sulfate and carbonate, and that in Ifgnite it was 50: kaolinite, 10%' pyrite, and 40% calcium sulfate and carbonate. The mass median, diameter of' the J-mineral matter was 2 um for both coals. The optimal' temperature for cenosphere formation based on ash density was demonstrated to be 1500*K. Thir stimum'was rationalized by calculating the time for sphere formation. At higher te- .atures gas evolution is too rapid and gas will escape from the molten ash; while at lower temperatures sphere ~ formation is too slow relative to the duratfors of the molten state within the furnaces. Padia et al. (1976) have summarized the principal. reactions that take place ir. mineral matter during coal combustion: A1 SI 0 50H)4 + A1 0 '25102 + 2H 0, W 2 25 23 2 + . Al 02 3 + 2510 ' 2 2FeS2 + 5.50 *' I'2 3 + 450 ' UI 0 2 2 CaSO4 + Ca0 + S0, (sa) 3 MgSO + Mg0 + S'0, (6b) 4 3 Fe ISO I + Fe 02 3 + 3S03 (6c) 2 43 CACO 3,+ Ca0 + CO ' (I* I 2 CaMg(CO )2 + Ca0 + Mg0 + 200 (7b) 3 2 ...( These reactions. Eqs. 4-7, all generate gaseous decomposition or nxidation pro-ducts. Kaolinite decomposition '[Eq. (4)), pyrite oxidation [Eq. (5)), calciam and + magnesium sulfate decomposition (Eqs. (6a)-(6b)), a id calcium carbonate (Eq. ' i)] and dolomite decomposition [Eq. (7b)) may all occur at _1000'C or less, and thus raay readily provide gas pressure for cenosphere formation. 2. _Plerosphere Formation Light and electron microscopic studies have identified a morphological class of r spherical particles containing encapsulated smaller spheres (Matthews and Kemp,1971; Natusch et al.,1975; Fisher et al.,1976) (Fig. 4). These encapsulating spheres or plerospheres (Fisher et al.,1976).are similar to cenospheres in that they are composed of an aluminosilicate shell but are filled with individual particles' rather than gas (Fig. 4a-4b). Matthews and Kemp (1971) and Natusch et al. (1975) have established that plerosphere formation is truly the result df encapsulation during particle formation rather than filling of a ruptured cenosphere. For these studies, either the electron - beam of a scanning electron microscope (Matthews and Kemp,1971) or an argon-ion mill-ing apparatus (Natusch et al.,1975) were used to etch through the individual particle surfaces. Subsequent examination of the etched particles indicated the presence of numerous smaller particles within'the plerosphere, thus confirming that encapsulation occurred during particle formation. ~~ 9 ,,..=8 ,-r-e w 3, m w-
p' F L ., :~. : / N ~ m t.' m *-: y. ; y...:. : " 3., a ~j7~.~~mm.y. y ..s g
- wm4.....n.fec w a
i q. {: ~ -,7- ~:p g Lq 3, ..* q.%,a)3.*f.
- g m.,
n t. e-s. u s .. :s r 's . 7f..:r,.(t -'t -,~. .Q* s.. g-,,,. J"
- g. /
1.- 4 D.Q';yf. ( Q..wei.,,- E>c-4 .g y ;7.f \\ -?; '
- t. :
1 ~ s.r c r j N N(n .3y.. T+r d4. s r. n n /.. Q.I.r:; 44 U:.'] ~ @ p*l \\q; g. t 's s : ~..,': 7 ..m. y.,, .y ... a w-V .e i,. w. ~ M'w.s.}./ .;'Y
- ~
e [ tid ~. ,.i.n. tu. i .m + .s< -\\- t..m
- .l r -
.e . ~.
- u.. s f
1, t v.~1 .y ; 2m,.; <,.! . t t-r.n.'. ', t: lu,.% ; A w .s. ' %%,i r ..b .a -.. i, ).,M.'
- s n
..r g,, q[ r-f.-
- g. * "
.g. 1' .l ..t.
- 4.. -
. y,.e . /.... - t. ....,id,' ', -. j, p? K [ .9 ~ q, ,5 -,. . +., - .t. 4 . ;i., ;-.y ,3-1.,,- .-a,. m. .. m., s - v. ,, ? '. 4 ~g' .o., ...w A ..-~n . B >';,ze : ce.e:.~via:w,:w;,cu.- a ~. s: s s.wp W..... ~- .e u. . p):
- ,t~
,'A >.% s .'fy.W W r.%..:'t:,,.. G.2i%w.,,k , ' l ,a-e.s L
- s 9
- .,s.,
.c k._f. '. g,$p, js,sv N.y. 4,5. '~ ~ % ( '7: ~-6 ,,. + -
- y %.r,,.
r' w.5 @.. V./. u ,a Lt.uc k;t 4 / h. V.1"+ h.c e f 3 3 >t pk f: ?.i. : ? ,,/ e, A T, h . N:,g.h,?. s..., > y-F f,. ' .c p.' p....Y, b. s - -., s..J. .f.,. t.'. s.r.. MT.h,i.M r * .. * -:?.,.;. C. c Q g* 'vf,.'",f '4 P t&#
- i f
.e s b, @4,...gf,. : a:,*p *M ; + .,..a/ 'i. i y C...n .m,. / r-, u. t e.. 'e b*s.9+wk I. i .e e f S. ' N..
~ ,gp y 'J: ?
- , n
- Plercsphere formatica is hypothesized t'o result from a process similar to cenosp5ere ,.,; formation (Fisher'et al.,(1976). As the aluminosilicate particle I) progressively heated. : . molten. surface layer develops around the solid core (Fig. 4c). Mineral decomposition - ' with evol'ution of CO and H 0. then results in formation of a bubble around the core. - 2 2 Jwhich remains attached to the molten shell. Furcher heating leads to additional gas , formation causing the core to boil away from the shell. This process may result in con- ~ . comitant formation.ot fine particles. The process dan be rapeited until the.plerosphere. 'is' full of other'plerosphere$ or solid particles or untti the, particle freezes. The i cha'racteristic t'ime for formation of a 50%m-d'an' plerosphere was calculated t'o ' e about - b ,31 msec..Similarly, Raask (1968) cal' ulated the formation time of a 50pm 'cenos'phere' to-c be about 0.3, msec. '3. ' Surface Crystal Formation ~ ~ Surface crystals '(Fig. 5) identified 'by SEM have been, explained by reaction of sulfuric acid with metal oxidas (Fisher et ?1.,1976). This crystal formation process is relatively slow compared to the time required for particle formation. Fisher et al. - (1976) have hypothesized that surface crystal-formation results from 502 hydration and subaquent oxidation on fly ash surfaces to form H SO, which then reacts with meta) 2 4 L osides, predominantly Ca0, or with ambient NH 3 to form either CaSO4 or (NH )250. Such ~ 4 4 a mechanism could also result in formation of soluble compounds from relatively in.
- soluble oxides, e.g.,~ conversion of Pb0 to pbSO. Fine particulate matter has also I("
(been observed by SEM on fly ash surfaces by a nerNer of investigators (Cheng et al., 4 1976; Fisher et al.,1976,1978; Matthews and Xer.,p,1971; natesch et al.,1975;. .Sarofim et al.,1977; Small,1976). Small (1976) has identified four surface morphologies, ~ based on SEM analysis (Fig. 6). Spheres with smooth surfaces comprise <i the most commonly encountered particle morphology (Fig. 6a). Spheres w!th small surface particles -(Fig. 6c) or relatively large surface-associated drcalets, resulting from condensation and solidification (Ff;. 6b), were also observed. Elemental analysis of these two particle classes indicated the surface was predominantly Si and Al and the underlying ' particle was mainly Fe. A. fourth class of. spherical particles with high Fe concentra. tions was found to display an unusual patterr of coarse surfaces (Fig. 6e). Saroffm et al. (1977) reported the presence of submicron sillca particles on laboratory-generated ' fly ash. As an extension of the vaporization-condensation rec %anism of. Davison et al. ' (1974) for trace elements. Sarofim et al. (1977) suggest that stitca deposition resulted from formation of fine silica particles that agglocerate on fly asn surfaces. The formation of.submicron silica particles was thcucht to be because of nucleation of 510 resulting from reaction of SiO2 with car, bon. .4.
- pherule Formation from Natural Processes
- it is interesting to note the natural (nonanthropogenic) occurrence of glassy
~ spheres.with morphological appearance similar to fly ash. Glassy spherules have been 1. reported to be present on shatter cone surfaces (Gay.1976). These spherules were pre. % sumed. to be. the result of meteorite in. act. Similarly, glassy spheres have been s 4 11 t 4 \\ g C
y .a.:. ?*'. 4 x e
- e w
1 s Y' - E 2'. y+,%. *. .'.s. ~ "9 .a* S.2F " 75* % A*W. ' = )l;;... i - - T.-".'; & . q.' ?-... .u *g, 'M .'..6.. i s \\ 5 ? e.. 1 "r'" f Y.~. ..fW, ~*, .. s. Ni,g, .. s.: -3 4 %4.g '- r , +. l ,4 .,e.. . _,i'.<,Y,Q s. - (* ',. M. ) ..g 'e. eg, ~2 9e M), 4 c- ,,,-s . s...ti;' .. 4.,.. g. N. ,..4 ^$UM f. "*?.* n :* . y ' T..y. f. ,t .s ,sw 5 -~ p.w.
- /c
- ~~ m j',
y '~' e or. r. m e x v. .,h *..,. , r. ~ .' # e'f p ' e f gg 4*a*.'* / Y Q r
- Ag
- s*.' 4 j, s. . s 4 y .f* .. v <f,; -..,-5 . e,...,",,p 4... , I.f,
- [24 /
) e e - o. ,,4g gl - J 1. se. y n..,. ' l .e- ....e. (.. ' t .e s.= - k i e, *%.% %. y = e* ,.e.. i f + t = .A l l + .7 '.m 4 f 6 ,1 l * ,i i.. e l 4 1 f / l.
l + l I -. -,T - ' -'-@ \\( %),..$*ie e. ,i f Q 4 .x .3 ' o. p. ,s i t ..,a 1 3 ...z f s..S* f>& ..r' t
- p. a.g 6_%m i,
g j ., (.,, ,1 .y ig .c.;. +%.,a i a.- x..,., .. =.. a,, r.,, tv-- > : w,..a a:.. sj ,..w j Q m..... A.. _.c.:.1-h '"'n i!
- .z..
'o s g m. 4 b' yge -- 9 rm r, m" jfI 3, h, p).,,,y, D,
- kf*.
l' y > ^ ..a r 't. - . W~*. ? A 1 y; E ,. hv.. Ei E.L,.vy bP M4d h','.?.-j;?*&5p,%,:..*& ~.h. ' ly 'y ...r.., y- . #w -Q .,+ r ...a r n b 2 R&3 .'N. ?;:..: ~. :. i...u+:3t< f. g y 47* u-e s.a *%.3,/.; .n i f
- . V=,.;:.: =-
' .,;f J ..%e N..-.. ' l{* '?' f .}. J ^ a .:..:.;,,._';c y.,;( u s pt. %)> f, - .pt p k
- - ):,; : _ =... '.
bu w' m ,_i a,.,,, % gj h' Q, D " 2s von a T7 5 .i'%
- s..
s . d ;, V.,:.***.- y ' ' :.. ;.
- . '- f, y
. 4,%.,3f. w3 4.. / s$- ...s%e Y*
- O.T 1
.:u /..h.~* k4h.. A *f5M' f .. v3.,..',. h* Y g-- j. .-y,, w,,,- 3 1' 47 . i %,e. w,, or ..-.,. w.. 3 v r.- p.. y ,s,ta, Q w L.,5 s , -,7 - w ,f,~' '. 4.[s.f'-A:l/- 4 9 - : 'f W ', 3,,,N g, $ .,, ~ .Qrj$'.l,:3. z # "l'&$ &; D # k 14,*k %a., f.7
- J' 2
, k. : ~. ' d .,sw ''*W
- a* *M.ph.,a
- w....
.pr.,Me. t:.g' ;.. ;t.... ':c. i.Y.> d .4 5. > &4 M O vs.- Q. /e F ss-v 1.. ,. 'h. ,gy-q..pf,-v **.w.. .3 "v .+,..p ..3-m 1 .r -C....., <,. r- . **N.ma.d.2
- [ 4ttf W **'A**i,P 1.,*/,r-. ~.... ~ :.,. :
- }y : ~ .u %m .- o r t.?_Nef;.. .,9 L 1 ~.' p n e l' fj -. a 1 pen F i.J. - ...p s.g .I e
- D -
gl JP, 3 * *, q et-6
- e.
Q ".r+
- q.,.,
e ,m- ..,---3 .,, - - -..,.. ~ - -.,, -..,..,.,, -..
e ~ # + p-n' identified in lunar dnt..anples. 'iagy et al. (1970) recorte:* th'e presence of fine glass -h l beads th.st were either spherical, nearly. spherical, or dueDbell shaped in Apollo !!' hmar y. samples. 'Some broken glass-beads were hollow and vesicular, sintlar to cenospheres;
- particle wrfaces generally were coated with ' fine partidulate matter.
.CH. Co anc CO2 3 .were found to be entrapped in tu glass heads. Carter and' MacGregor (1970) reported - 'that glass. spheres ranged in ccicr from colorless through green, brown, wine-red, to opaque. In discussing the fort ation of the lunar spherules. Gibbon (1975) pointed ou-thJt the r'orphoingy is related tc either impaction or volcanism? The loose lunar soll. is primarily, composed of spherical particles, indicating a high-temperature' origin. Waits et al.. (1973) have performed detailed morpnological studies of lunar samples f' rom the Apolle> 17 and the Russian Luna ilssions (Fig. 7). Smooth vesicular-spheres (Fig. 7a) and spheres with internal vesicles have'been identified. Fine particulate' matter on' sphere - surfaces (Fig. 7b) is thought, to reflect the in-fliou aggregation of fine particles on r)olten sp*iere surfaces. Knobby spteres.(Fig. 7c) that are predominantly crystalline are thought to be feldspar. Detailed surface analysis of some lunar spherules. indicate sub. - micron crystals with lath-shaped habit (Fig. 7d). Pierospheres have also been observed (Figs. 7e and 7f) although it is 'not know'r whether par:icles in the plerospheres of Figs. 7e and f filled a fracture af ter vesicle formation or were formed inside the sphere. Thus it appears that studies of coal fly ash' formation and lunar spherule . formation are complementary and should provide mutual support in understanding the physical and chemical processes involved in the high-temperature morphogenesis of C particulate matter.' !!!. PHYSICAL PR0pERTIES OF COAL FLY ASH: pARTIC'LE SIZE DEPENDENCE In addition to particle morphology, a iumber of other pnysical properties have been J ' investigated in attempts to elucidate the formation and behavioral characteristics of fly ash. Such properties include the mass distributichs rf particle size and the - ~ particle density, specific surface area, electric.1 resistivity, and-ferromagnetic susceptibility. Unfortuna tely the available data are sparse and apply to fly ash collected after a control device or from the device itscif. Consequently, it is-not. ~ lresently possible t a relate physical properties to parameters such as plant operating ennditions or the type of coal fron which the fly ash was derived. Nevertheless, it is qualitatively apparent that the physical properties of fly ask depend upon both of these - parameters. A. Mass Distribution Measurements of the distribution of fly ash particle mass with size are of two distinct types. The first involves determination of the aerodynamic particle size dis. \\ tribution, which normally involves isokinetic collection of fly ash directly from a st'ack gas stream. Several samplin2 devices are available, but 'the most corinon involve - the principle of inertial impaction and enable collection and size classification of fly ash in situ. The principles and nethodology of inertial sampling have recently been re-viewed by Raabe (1976). Newton et al. (1977) and Natusch et al. (1978). Alternatively. 14 -m. g 0 g
. rr F (J .., _,;. f. -. g. 3.,.,,._ / -t.
- 3
- -....,...,..
=e.. ,a[, *..y. g%.. q .y t i c N
- 1. L ;r j
. q '.ty ', S 1 .:s s
- o. --
1 ....c.a....... .~ yz-
- . p
.)
- t4 b,
~. -a
- 7'T '..
y($. . ?. kk ' ' u ';'t i .ae .\\ ..e, a.t,v; e,e.., fi 6. s 4 ps... . t.,,q. y,. - r.. : ~ e .~,,. ; r-5 . 1 2..,,. 7,.t. A r j A(: g n n n.- . ~ A u.i.. a."n%ut .... eCE-iEri..d [ 8 &'5 1.. - ' 3l. a- . ~.. a - - ~- m ; m, m y:,- m n'. h R ~y n ' , p. .., 1.. a.$.~. $. [ e %. t,.f;,
- '*-8* r..y-A
/ m u i,. ,/-1).. % P % ( y..i f,y:., J.;?.# % t. ...a. .. #.d. M..O... i ... ~... .,e.- n......:......,, o. :. ;..v ..; 3.,r. ..., *. u. -. s.: y }* s, s , (s - - Q,w.... ;- Q.'.si,*A. y** . }; p.N V g l ..o : y ( *-*' i*. )',',* . gI*.* b.- ,,, ! ' i [) &.:.. ff; ,0 (.' Y ~ &.g >..,k. -\\, ?..e l ;3,, . f as- >-s 4/1 - r, ~ \\ i ty;. A y.- s. c + ,q. y- / ,[
- k. n..
2 f,,.g , 3, .y g ,. s.*'adAs s.
- t '. ".. d.
b 7,,;s/. - {., e, ..,,s .e* f 3 o N(. ) L ....,_.%.,. h e $ W.L. w.M l N. $I s 6',- Jg 3-- -a D M:.4., :, ? ' i..~ G1 -~u--e-- p-
- ; g::: m*L 3,7 cq._*.7".'('~,*yps. y..y.-
- s
.. = :.s:M. % s, ey f n e+ .k -] 6 h, .~.g.. L, % '.;ye -2 6Qtg e,u n ::x ?.s ..o o v- . n, g ? l\\.s [- r d Q. s*e. : r.,.,,.j 4 t. ,y ,:./ o 1 +* ' @. y...~.,.d ? g 4 'y .. a:J. 9. V, c ~.D?& .v /., <...w [. ~ af
- c. p g
..),.,i' -' (* MI'5' - /" .g, flg. :" .i- -
- le f r.w. :". ';. 5
.;;>o : 1-17 and the f.. '. I " . e r,:.r:e e..., c., 5.; f r.. wt th J lje eget to .e'.te, ele *J';
- n..! i d '. i' ; %b.'" t C PC rl.
l,**- ti.. + .,41 t re. e. i...., rs;,:r ..r is; Aere) S truc ture. 's,. .r** nt** '. I ',' d'PI '!. L. C I O*.'OM. Ljtid0ri E. J0hrt$0rt e .; ji. .e*
- .A.
\\ f L
c ...W 4 ~ uv:,,, bulk, fly ash, such as might be collected from an electrostatic precipitator or. bag house - ~ I can be differentiated;f nto ' physical ~ s12e fraCtf ans by mechanical sieving. 'In either Case. .it is necessary tn separate the various size fractions prior to weigning or microscopic 1 . counting for site estimation. ' These two types of measurement are quite dist11ct. aerodynamic size determnation -enables prediction of the atmospheric and airstream V5avior of each size fractior., such ras :Is' important' for establishing particulate collection efficiency, atmospheric residence. 7 and inhalatf 0n Characteristics (White,1963; Butcher and Charlson. -1972; Natuset. ad : ' Wallace.' 1974; ~ Yeh et al. 1976). : On the other hand, physical size determination pro-vides a straightforward measurement of the physical dimensions of the particles and can ' be directly related to particle number. !!nterconversion between aerodynamic and physical
- sizes can be accomplished using the relatie iship (Koltrappa and 1.ight.1972)
I D,, = D C(D ID /C(D,)o (8) r r r j .D, is the aerodynamic particle diameter. Dr. is the physical particle diameter, cr is'the particle density, r.ae is 1 g/cm by definition and C(Dae) and CL0 ) are-the Cunningham r slip correct
- 3n factors for the diameters D and D. Raabe (1976) has compared the ae r
commonly used conventions for description of the aerodynamic size of respirable aerosols. including a detailed description of the slip correction ar1 dynamic shape factors. Aero-dynamic size distributions are often presented as log-normal probability functions of ~ Tthe grouped mass or num5er data derived from inertial impaction lastrumentation. The general matheieetical approach to fitting size distributions to aerosol data has rece'ntly-b.f j [bedi described by Raabe-(1978). L 'q 9 .Despite the comparative, simplicity of determining both the aerodynamic and physical: size distributions of fly ash mass, the number at available measurements cf fly ash prior - ~ to collection by control equipment is relatively sparse. It has been established, how- ?- = ever (Southern Research Institute,1975), that the $12e and morphology of fly ash depend not only on the nature of the mineral inclusions in' coal, as discussed in the previous section, but also on the manner in which the coal is burned. This latter dependence is illustrated in Fig. 8 for fly ash derived from coal burned in a chain grate stoking unit, , a pulverized coal fed unit, Land a cyclone, fired _ unit (Southern Research Institute,1975). In each' case the fly ash was sampled upstream from control equipment so it is representa. tive of that generated by combustion. -it is apparent from Fig. 8 that the fly ash mass approximates a log normal distri. bution over the aerodynamic size range considered. Furthermcre, although the geometric standard deviations of the three distributions are.similar, their mass median diameters differ. considerably (i.e., Cyclone = 6 um; Pulverized =18 6m; Stoker e42 cm). These observations are in accord with the general principles of particle formatica outlined in the previous section. I. p l t j m { ' l'6 f O
It 100.0 Pulverized b-Cool-Fired -- g ~ 1, Cyclone 0 Stoker Cool-Fired / 44 _ 10.0 E-cool-Fired E d Fig. 8. Size distributions for boiler / particulate emissions from coal .j a: M combustion in a chain grate stoker, /'- W a pulverized coal fed unit and a E // / cyclone fired unit. (Recreduced g,o ./ 5 E /j/ by permission of Southern Researen II Institute,1975, and Electric power w .r /// Research Institute.) s M >= { / /
- 1 0.1
.1 ~ i -) 0.01 'j ODI 0.1 1, 10 50 90 99 99.99 1 WElGHT % LESS THEN STATED SIZE i ? 4d ' l, .J .- l From a practical standpoint, one is primarily interested in the aerodynamic size distribution of the fly ash that is actually emitted from a coal-fired power plant. 1his
- )
is, of course, largely determined by the collection efficiency of the particle control
- .5 equf'pment. Specifically, the size distribution 'of the emitted fly ash is determined by
'l the product of the functions describing the size dependence of fly ash mass entering i a control device and describing the dependence of collection efficiency of this device j on particle size. Examples of the aerodynamic size distribution of fly ash mass emitted from a coal-fired power plant equipped with different control devices are pre- '] sented in Fig. 9. 4 4 J a 1 17 3 w .v. .--,,=.,,v, - - ,--,.--,.---r,--, -nww
c---
-., - =, - - -.. -, - - -
.9 .T....... ' 5 5 "M 10 10 SCRUBBER A:'d ESP 4 4 m 10 e 10 N ..r$o,1 w w n a w z 3 m 3 ~ o 10 M 3 10 z z %....g w w n 2 2 2 10 ->- 10 4 ..a. J a um w w 2 10 oc 10 j w 1 1 1 t il l 0.02 0.10.20.5i 9 I I i 1 %[ l0.01 0.1 12 10 y, ,.s , st DIAMETER (pm) DIAMETER (pm) e sgw Size distributions for particulate emissions from similar production units Fig. 9. ";(j with either electrostatic precipitator (ESP) or a Venturi wet scrubber at the same power plant (Modified with permission from Ondov, et al.1976). w n% s.j .i
- 1.d,
. u, L E.h l B. Density and Magnetic Distributions .. re 7-$b(c; Determination of the density of coal fly ash as a function of particle size is largely of interest in obtaining an understanding of aerodynamic behavior and of the .c&. -- g factors responsible for the intrinsic heterogeneity of coal fly ash. Thus, det'ermination IN:V of the densities of different fly ashes, and subfractions thereof, provides a means of
- p4 interconverting aerodynamic and physical sizes according to Eq. (8). In addition, some
- . ;
- g j;
differentiation between distinct morphological and compositional characteristics can be .4,v ". achieved. For example, cenospheres can readily be distinguished from solid particles on t
- %{}
the basis of density as can predominantly carbonaceous particles from aluminosilicates. W Jgq Determination of fly ash particle density is most simply achieved by means of the ,;)] traditional " float-sink" method that employs a series of liquids of different densities 'Y; to separate particles of greater and lesser density than the liquid (Ruch et al.,1974; Olsen and Skogertoe,.1975). Altern:tivcly, se;: :thn can be achieved by placing the .. 4 particles in a liquid in which a density gradient has been established. ,Q While determination of particle density is of considerable interest in its own right, more definitive insights are obtained if density separations are carried out in 504 r.g.e conjunction with sequential size separations and with differentiation between ferro-y !d magnetic and nonferromagnetic particles. Such a three-dimensional fractionation scheme 'h, w[ has been presented by Natusch et al. (1975), and resulting mass distributions are pre-PA.e, sented in Tables II and III for' fly ashes derived from typical midwestern United States M wm b 18 -M d}
_ bli.dMEL r_ ti $ *D2 IlYObMb bbIAENb9d/'.Ii.IMEEE Mck. L h 3 N N 2l~ Table II. Mass distributton of size-classified, magnette and nonnegnetic fractions of a midwestern bitiminous coal fly ash (%) a Size Mnnmagnetic Magnetic (um) <1.6 1.6-2.0 2.0-2.3 2.3-2.7 2.7-3.0 >3.0 <2.1 2.1-2.5 2.5-2.9 2.9-3.4 3.4-3.6 >3.6 <20 b b 0.2 28.0 b b b 0.6 0.4 1.3 14.9 0.5 20-60' 1.4 1.3 12.1 12.9 0.1 b 0.2 0.6 1.8 11.5 3.1 0.1 60-90 0.7 1.0 0.6 1.1 0.6 0.1 0.5 0.8 1.0 0.2 b b >90 0.1 0.1 0.1 0.6 0.5 0.2 0.1 0.2 0.3 0.2 0.1 b
- From Natusch (1978c). unpublished results.
b Less than 0.051. U$ Table 111. Mass distribution of size-classified, magnetic and nonmagnetic fractions of a western sub-bituminous coal fly ash (1)a 3 Density (g/cm ) Size Nonmagnetic Magnetic (um) <1.6 1.6-2.0 2.0-2.3 2.3-2.7 2.7-3.0 >3.0 <1.6
- 1. 6-2. 0 2.0-2.3 2.3 2.7 2.7-3.0
>3.0 l 20 b b b 0.7 b b b b b b b b ] 20-44 0.2 0.4 0.5 21.3 0.3 0.2 b b b 1.0 b b 44-74 0.5 0.8 1.0 45.6 0.6 0.5 0.1 0.1 0.2 6.8 0.1 0.1 >74 0.2 0.3 0.4 16.1 0.2 0.2 b b b 1.5 b b i
- From Natusch (1978c) unpublished results.
b Less than 0.05%. i t
...N ~" ~f [i.h ..) . bituminous and western sub-bituminous coals. Thes2 data trere obtained by separatio, (. (k.M ' bulk fly ash ints savtral' physical size fractions by sieving.
- {rp $ f})
then subdivided into a number of density fractions that were, in turnEach magnetic and nonmagnetic fractions according to whether the p , separated into 9M4 net or not. 9,$.u The designation of magnetic and nonmagnetic is entirely operat nature. g [ "Wy, &r.M A number of characteristics of coal fly ash can be distinguish J M. M sented in Tables II and III. It is apparent that both fly ashes are compositionally 'g extremely heterogeneous, although there are very considerable differe tsf .,y pass distributions for these two fly ashes. N;f As discussed in the previous section, much of the variation in densities observed is attributable to morphological ra ,v. J
- J -;d compositional characteristics.
This is rather well illustrated by the data in Fig.10, { y,[Q{ g where density distributions have been determined as a function of part before and after crushing the fly ash. indicates the presence of vesicular particles and cenosph EQ.,j fractions, as discussed previously. S. ~g.
- y.MZ'.
Interestingly, determination of the x-ray powder diffraction patterns of the subfractions presented in Tables [ j] f II and III reveals no convincing differences in matrix composition that depend upon either size or density (Natusch et yA-g This finding further supports the contention that the density ' distribu ,.f r.. are largely determined by morphology rather the. by composition. .. p.y} There are, however. very distinct differences between the amount of magnetite (Fe3 4) pre 0 , g netic and nonmagnetic fractions (Fig.11). ...a p responsible for the ferrumagnetic susceptibfif ty of :oal fly ash.Th QM
- SQ.d C.
L'Ryd. Electrical Resistivity Ofstribution
- f.(
The electrical resistivity of coal fly ash is an important physical prope mfd the standpoint of control. Thus, it has been established (Bickelhaupt. 1974,1975) that f[D[A; the collection efficiency of electrostatic precipitators increases with de ty-M-y( ash resistivity. Bickelhaupt (1974,1975) has further shown that both the surface and y volume resistivities of fly ash, at precipitator operating temperatures j ' $'[M[ proportional to the specific concentrations of alkali metals, which are thought as charge carriers. These studies have shown that considerable differences in g#. %, electrical resistivity occur between different fly ashes, and correlations are 5 ,q. between fly ash resistivity and alkali metal content, but no measuremen 'i f relating resistfvity directly to particle site, N Some insight into the dependence c' re:istivity on particle size can b [C. j .,,y considerft.g the data presented in Table IV. This table lists concentrations of potassium measured in fly ash that has been fractionated sequentially acco %W
- $kf density, and ferromagnetism as described previously.
It can be seen that in the non-magnetic fractions (that accounc for 64% of this fly ash) there is a pronounc -h in the concentration of potassium (a'nd also of sodium), both with decreas
- }1.l.K size and with decreasing density.
This suggests that, for this fly ash sample, hi{ w ~.4 n
- e gg b
t
.Py, W .yy, - 10 r 4 {4. tp Y o' .A. PortideDonerer, B. iwticle0iameter. g' O <20pal 20-44pa ./ ,i
- 5 s
n 03 s e s l o J Fig.10. The effect of crushing on the mass 's ci distribution of size-classified fly E / 5 ash fractior.s. The shift to h'gher densities indicates the presence of 0 y -ID Particiociameter-hailc.; or ve:i: lar par;icies (figure . C. by courtesy of D. F. S. Natusch). 44-74pm > 74pai ,,.i j ,1.
- l i
.( g .y e y OS u. o s ,n \\ o s \\ ~ 4 j ,/ ' s, ~ g q 6 y ./ l. ?:6 0 2.12.52.93.3 2.1 232.93.3 ~
- C:.,
3 4 DENSITY (g/cm )
- r,
.ra o 4.-4 M1 NONMAGNETIC L,Mu H-hematite c Q-quortz .e,y Mu-muliite .l H J H Mu . s.' ~x.
- yp:
e t Fig. 11. X. ray powder diffrac-g 7,'1 tion patterns
- ,'.'.'.h demonstrating the
! J.3j compositional 3 MAGNETIC differences betw<en .9 H-hematite l 1 magnetic and non-gj! magnetic fly ash Q-quartz j fractions (Vigure M-magnefite c i l. d.? hy courtesy cf D. E. 5. Natusch). \\ H H a ' - a.; 1 M y gh ".J H H M H H
- p.. >. '
...e ! i i i i l 60* 50' 40* 30* 20" [ 29 Angle ! i i 1 $.1 4 e
- 4
.:, - -,.,...~. - :., c.~ ma;<. _ -. -,;- z..y57,,,;: '
u.: .g).. %ll .2 ~ !.' a Tablo IV._ Concintratitn (%) cf Potassium in Fly Ash Separated S:quentially by Size, Density, and Ferromagnetisma 9 3 Density (g/cm ) .g Q Particle size N Particle type (um) <2.1 2.1-2.5 2.5-2.9 >2.9 N M Nonmagnetic <20 2.69 2.34 2.22 1.73 3 20-44 2.63 2.28 1.33 1.09 j 44-74 1.39 1.63 1.05 0.45 .] >74 1.79 1,48 1.06 0.13~ D Magnetic <20 0.76 0.70 -j 20-44 1.92 1.48 0.73 .C 44-74 1.78 1.60 1.27 0.85 a;3 >74 1.37 1.62 1.49 0.83 .4 a ,( From Natusch et al. (1975). b i[ No meaningful data. d resistivity decreases with particle size and with density. Similar, though less pro-j nounced, density dependencies are observut in the magnetic fractions, but size depend-M encies, if any, are obscure. - Since both decreasing density and decreasing physical size l.N contribute to decreasing aerodynamic size, it is apparent that the efficiency of,
- r M
electrostatic precipitation per unit mass of these size-classified fly ashes increases j[1, with decreasir.g aerodynamic particle size. This is an extremely desirabie characteris- .?. d tic. It should be pointed out, however, that these studies require extensfor, to
- .4J respirable particle sizes.
- y id D.
Surface Area Distribution t') The specific surface area of fly ash particles is an important parameter in de-N temining a number of the behavioral characteristics of coal fly ash. It is the sur- "7) face area of a particle that determines the number of electrostatic charges that can be s- ,?.i placed on that particle in an electrostatic precipitator (White,1963; Bickelhaupt,1974, ll-1975); it is the sur' ice ar,1 of a particle.that determines the extent of condensation .q7 cr adsorption of s;'ccies from the gas ph:!c ('h'.h:t et al., I??4, Natusch and Tonkins, 1977); and it is the surface area of fly ash that deter. nines the rate and extent of its .fi aqueous leaching (Natusch et al.,1975; Matusiewicz and Natusch,1979). To a reasonable approximation, one would expect the specific surface area (square ,(,g meters per gram) of fly ash to increase lir.ea*1y with decreasing particle diame:cr .[f_} since the particles are predominantly spherical. Similar trends would also be expected g for nonspherical particles having similar shape factors (Butcher and Charlson,1972). .:hNe.a -k ',j 22
/J-In fact, th2 Expected trend is cbserved; however, two important points are notcd. ] " First.-th2 surface areas that; are measured for spherical fly ash particles are con-sid:rably greater than those calculated froa E:asured particle diameters. Even taking 1 \\/ - into account the assumptions inherent in surface area measurements, it appears that coal 'p fly ash has a ~ significant " internal" surface area. This is probably in the form of pores i f or cracks or a porous surface layer, although, as previously described, surface crystal N ? formation may contribute significar.tly to the measured surface area. However, several fly ashes show no significant dependence of surface area on particle diameter (Table V), especially for small particles. These data indicate the existence of substantial internal surface area that is effectively proportional to particle volume rather than j external surface area. In this regard, it has recently been suggested (Natusch,197Ea) that collisionally efficient condensation processes may result in deposition of material from the gas phase predomf'antly onto the external particle surface, whereas much less efficient adsorption processes (Natusch and Tomkins,1977) can deposit gases and vapors on both the internal and external surfaces of a particle. 1 1-Table V. Comparison of Measured and Calculated Specific Surface Areas of Size-Classified Fly Ash Fractionsa M Physical d ze Measured Calculated .[h (E) (mjg) {e,2fg) 2 q.s s Q <45 2.02 >267 E.[ 45-63 3.55 191-267 63-90 2.55 133-191 90-125 2.43 96-133 .j 125-180 1.20 67-96 s >180 3.11 <67 ,9
- C.1 a
77,, gg,,nd Natusch (1978). Unpublished results. .. ? f. . y' , l' $ i1
- h 3
i-l l 6 lry , f...l .i .1 l l t I 23 I 1 ~ ~.. ~,.~. ~.., s n -, _.
- a, -m-e = -
... ~ r~v ~ ~ v * * ~ ~ ~~ ~ e."
IV. ELEMENTAL COMPOSITION OF COAL FLY ASH: PARTICLE SIZE DEPENDENCE [; m.. ?[ggsj Studies of the size dependence of the elemental concentration of fly ash can be E5 classified into two categories. The first category consists of those studies that re. l$.N. 9 late the elemental concentration to the particle size of size-classified material. For these studies, sufficient mass of size-classified material is collected to allow ~ g gravimetric determination prior to elemental analysis. The second category are the many
- l-{g.f studies that have employed inertial cascade impactor systems for aerodynamic size classification.
Aerorol sampling is perfomed isokinetically to avoid anomalous alter- .q M ation of the particle size distribution. Because impactor stages are often coated with 'h. sticky adhesive to prevent particle bounce off and reentrainment effects and b:cause only f,( ' taall masses of material may be collected on the stages, accurate gravimetric determin-C."-v ation of sample mass is difficult. To obyfate this complication, specific elemental .j, masses of deposited particles on each stage are often raticed to the mass of an element ,6 that does not demonstrate a markeo concentration dependence with particle size. In '95 this regard, Ondov et al. (1977a) have analyzed four size-classified, stack-collected fly k{gl ash samples ranging in particle size (VMD) from 2.2 to 20 um (McFarland et al.,1977). The elements A1, Si, Ca K. Ce, La, Rb, Nd, Hf, Sm, and Cs varicd in concentration by I-W 1ess than 20% among all fractions and should therefore be suitable for mass estimation. A second approach that has been used in the analysis of impactor data, reports the size 4.g distribution for the mass of each element analyzed, thus avoiding the compounded errors in data derived from elemental ratios. Impactor studies also report elemental con-4 %j eentrations in terms of mass per unit volume of aerosol sampled. Thus, because of the MC limitation of gravimetric detemination, results from impactor studies are often re-ported as ratios of elemental masses or mass-to-volume ratios rather than speci,fic con-g. N.M) centrations. Y'.N Many studies employ the enrichment factor (EF) of Gordon.and Zoller (1973). ) EF is defined as the ratio of an elemental concentration in the fly ash sample to the j The elemental concentration in the coal. To provide normalization relative to total mineral 'Jh content, EF's are often calculated from the ratios of specific elemental contents in the S:Y fly ash samples and coal, respectively, to those of mineral matrix elements in the fly ."M i ash samples and coal, respectively. Thus, the EF may be calculated from: .4 M EF = ([X]3/IM},)/([X]c/N ), (9) T:. I g [g.h where [X], and [X]c represent the mass of element X in the sample and coal, respective C and [M}3 and [M}g represent the content of the matrix element in the sample and coal, a l respectively. A number of " matrix" elements have been used in the EF calculation: Al I (Gordon et al.,1974), Fe (Ragaini and Ondov,1977), Sc (Ondov et al.,1977c), Ce (Coles ,. / 40 et al.,1979), and K (Coles et al.,1978). In the following section, studies of fly ) A.) ash analyses using gravimetrically determined masses will be discussed separately from studies employing smaller masses. .4 '{ c e ef*." 24 h
- +e 3
- a g
A number cf analytical techniqu;s.have b en employId in the determination of the 24 e elemental composition of coal fly ash. For complete analysis of the major, sinor, and
- /y
- r,, '
trace elements, a combinatirn cf anlytical techniques is usually employed. The physical 4 ~4 and chemical heterogeneity in terms of particle size, chemical distribution within and g ,/ among individual particles, and the fused aluminosilicate matrix provide a unique com- %g bination of difficulties for the analyst. The techniques employed for elemental analysis may be divided into two categories: (1) single element techniques that generally require - . f'#- matrix dissolution and (2) multielement techniques that generally are performed on the 4 e undissolved ash. A detailed and extensive review of the elemental analysis of particulate / %g matter has recently been published by Natusch et al. (1978). See also Chapters 11 through /~ 14 in Volume I, and Chapter 45 i- "- " '" - / A. Studies of Soecific Concentraticns Davison et al. (1974) published the first detailed elemental analysis of coal fly ash as a function of particle size. The ash was collected from a power plant using I southern Indiana coal. Two types of fly ash samples were analyzed: (1) fly ash collected by the plant's cyclonic precipitator and (2) stack-collected material. The precipitator ).g ash was size separated by sieving the larger particles and aerodynamically separating the remaining mass. The stack-collected fly ash was aerodynamically classified using .'.,l ~ an Anderson impactor. These authors presented the elemental concentrations in three categories based or! the degree of concentration dependence on particle size. The / ,0 elements' showing " pronounced" concentration trends of increased concentration with de- N
- [.$
creasing particle size were Pb, TI, % Cd. Se, As, Ni, Cr, 2n, and S. Elements N classified as showing limited concentration trends were Fe, Mn, V. Si, Mg, C, Be, and A1. Iron concentrations decreased with particle size for the precipitator ash, while no trend was observed ir. the stack-callected samples. The elr:nents descM bed as showing *no concentration trends were Bf. Sn, Cu, Co, Ti, Ca, and K. The mechanism of concentration enhancement has been postulated to be volatilization of the elettent (or compound) at 0 combustion temperatures (14000-1600 C) followed by condensation on particle surfaces r
- l (Natusch et al.,1974; Davison et al.,1974). Thus, fine particles with their large
] ratio of surface area to mass will preferertf ally concentrate volatilo inorganic species. In particular, those elements displaying the greatest concentration dependence with b particle size generally are associated with elemental forms that boil or sublime at coal combustion temperatures. j Fisher et al. (1977) and Fisher and Chrisp (1978) have 'fescribed the size dapendence of the elemental concentrations in coal fly ash collected from the stack of a power j plant burning low-sulfur, high-ash, western United States ccal. The fly ash was size f. lf classified in situ, downstream from the esp, using a specially designe.d instrument em-I j ploying two cyclone separators in series followed by a 25 Jc". cer.tripeter (McFarland et.
- "[.)
al.,1977 ). Elements were classified into two categories: elemental concentrations j (1) dependent on particle size and (2) independent of particle size. Concentration ce-pendence with particle size was determined qualitativelf with the criterion that constant j concentration trends beyond experimental uncertair.ty were observed for each of the four i .d 25 e r., ,,n, e,.
e, E :s m. M'd 'h, fracti:ns analyzed. In ceder of decreasing depindence on particle size, tha elemeng ,h,[ Zn. As. Sb, W. Mo Ga P'. V. U. Cr, Ba, Cu, Be, and Mn displayed increased concentrat 9 2E.7 with decreasing particle size. Silicon was the only element to decrease in concentratg f with decreasing particle size. -5 s
- Gj The elements not displaying clear-cut concentration dependence on particle size d'b4 for all fractions analyzed were A1, Fe, Ca. Na, K. Ti, Mg, Sr. Ce, La, Rb, Nd, Th; Ni,
.&g g% Sc. Hf, Co, Sm. Dy, Yb, Cs. Ta, Eu, and Tb. Of these elements Na, Sr. Ni, and Co .+. M[j g displayed marked enhancement in the finest fraction relative to the coarsest fraction. A Coles et al. (1979) have described the elemental behavior in the four size-classified J.f fractions in tems of elemental enrichment factors relative to the parent coal. ..f.$ The elements were grouped into three classes: group I elements displayed little or [;. no enrichr.ent in fine particles and were lithophilic; group II elements displayed marked . yG enrichment and were chalcophilic (sulfur associated); and group III consisted of elements "f-with behaviors intermediate to groups I ar.d II. Group I elements included A1, Ca, Cs,
- C Fe, Hf, K, Mg, Mn, Na, Rb Sc, Ta, Th. T1, Ce, Dy, Eu, La, Nd, Sm. Tb, and Yb; group II
+h elements were As, Cd, Ga, Mo, Pb, Sb, Se, W, and Zn; and group III' consisted of Ba, Be, -%..j(tK Co Cr, Cu, Ni, Sr. U, and V. $$4 In a separate report, Coles et al. (1978) described enric' ment factors for 228 h Th. $k 228 a, 210Pb, 226Ra, 238, and 235 40 R 0 0, relative to K in the four size-classified fractions ..s of stack fly ash. Although the EF's fer all radionuclides appeared to increase with { decreasing particle size, 210Fb, the most volatile radionuclide, showed the greatest size .;.gy. dependence. The authors proposed the U is present as either a carbonate [Na2UO(CO)2or 2 3 jg Na4UO (CC )3) that upon heating in an oxidative atmosphere may give rise to either 2 3 3'.2 volatile UO from oxidation of uranite (UO ) or the silicate-soluble, nonvoTatile 3 2 mineral, coffinite IU(Sf 0 )1-x(OH)4,}. Thus, U behavior would be expected to display 4 , e-l,9 an intemediate behavior depending on the relative concentrations of uranite and ~ }h'ij coffinite. The behavior of Th was rationalized to be due to coexistence in submicron ' {hk l'. zircon grains in the coal. The authors suggested that ^ 226 a enrichment may have been R W,') due to 238, whfie no explanation of 228 a enrichment was presented. 0 R , T: Campbell et al. (1978) have studied the elements) distribution of size-classified l3fM2 ESP-collected coal fly ash from a western United States power plant. Reaerosolized [.k ESP fly ash was separated into nine size fractions ranging in size (VMD) f om 0.5 to i ff.., M.n 50 vn. The authcrs describe fine particle enhancment for elements " volatilized during i '.'.S combustion," 1.e., As, Co. Cr, Ga, Pb, Se, and Zn. Their data also demonstrate that K, l k..N3 A1, Mn, Mg, Na, Ba, S. Ni, V, Cu, Cs, Ab, Sb, Br, Mo, and Sn display an inverse con- '"( centration dependence on particle size. Silicon aad pessibly Zr were reported to in-h a,,] crease in concentration with increasing particle size. The concentrations of Ca and Sr demonstrated a maximum at approximately 5 i.m. A similar, concentration pattern was re- ] I'j ported for Ce, Eu, and Yb. .. M. . s ;~ay These studies are in basic agreement with the hypothesis of Natusch et al. (1974) +.C %. j in that.the most volatile elements (or their oxides) Cd In, Se, As, Sb, W. Mo, Ga Pb, , Sh and V, displayed the greatest size dependence. Furthermore, the least volatile elements [97 . :.M Q"4 .d 26 'f.
- 1
~ a
~-~N sy W e :. %e N,f,,did'not display a strong particlo size dependence. e, f With rGgard to enhancement of Ba and e e U, Coles ct al. (1979) postulated that Ba may fcrm the valatile species Ba(OH)2 and U may be volatilized in part as UO. Fisher et al. (1977) have proposed that the presence of 3 4*4 Cr in the organic fraction of coal, Mn and Sr as carbonate minerals, and Cu as sulfides, /
- 4 may explain the behavior of these relatively refractory elements. Campbell et al. (1978) speculated that the concentration profiles exhibiting maximum particle sizes of approxi-1, mately 5 um for Ca, Sr. and the rare earth elements we're because of the presence of 1
these elements in apatite. .' i B. Studies of Relative Concentrations I, Most studies of the chemical properties of size-classified fly ash have empicyed [.~ cascade impactors for stack or plume sampling. Zoller et al. (1974) reported enrich-ment factors relative to Al for stack-collected fly ash. The ash studied was collected downstream from the ESP at a power plant burning pulverized coal containing 10% ash and 15 S.. In agreement with the previously described studies, enhancement of the volattie elements. Sb, Se, As, Pb, Zn, Ni, and I, was observed in the stack fly ash relative to i 'T their concentrations in the coal. Bromine was depleted in the stack ash relative to the
- f coal. The authors point out that the EF's for Se, I, and Br are underestimates because
- ,f portions of these elements were probably in the vapor phase. Elements not displaying
.- Q enrichments included Ti, Sc, Th. Ta, Na, K Rb, Ng, Sr. Ca, Ea, V, Cr, Mn, Fe, Co, and --} six rare earth eleinents. It should be pointed out that although the stack sample was not size classified, a relatively fine particle distribution (i.e., MMD 5-10 um) may be
- e. ~
. Jj presumed for this post-ESP material. In a subsequent report (Gladney et al.,1976), the .a 4 j d research team described the size dependency of the EF's in the stack fly ash. Y Three patterns of elenental behavior were described. The elements Na, K. Rb, Mgi Ca, Sr. Ba, Sc, Ti, V Mn, Co. Zr, Tn, Hf, Ta, ano all rare earths except Ce displued .i an EF distribution that was not size dependent. Interestingly, the authors also report ~ that the relatively volatile elements Cr, Zn, Ni, and Ga, also exhibited little size . ;i dependence. A definite increase in EF of fine particles was observed for Pb, As, and [ [@j Sb. The velatile elements, Se, Br. I, and, to a lesser extent, Hg, displayed bimodal l,k acti'vity. An enrichment minimum was observed from 0.7 to 5.0 um. Iron and Ce displayed ! 3 EF's that decreased with decreasing particle size. l*1 Klein et al. (1975) described the pathways of 37 trace elements through a cyclone-l fed power plant burning ccai of 3% S and 111 ash. Concentration ratios for ESP out'et ( .versus inlet ash indicateo enhancement of As, Cd Cr, Pb, Sb, Se, Y, and Zn in the finer l fly ash fraction. The authors point out that the ESF efficiency was 96.5% during their l first sampling trip, as compared to 09.5% during their second sampling trip. Interest-ingly, the rencval~of the major elements was more complete during the second trip, al. l ( though no change in capture efficiency was observed for Cd, Pd, anc Zn because of ' f,' J association with fine particles. The authors estimate that 60-90% of the Hg was re-l.*[,] leased free the stack as a vapor. In a subsequent study Andren and Klein (1975) pre-sented ex' tensive datt, on the mass balance and chemical fo-m of selenium emissions from the same power plant. The auth3rs concluded that 68% of the Se was incorporated into l fly ash. Based on an ESP efficiency of 99.6%, the authors also concluded that 93% of the ~ a 27 1 a
V kN. = -D . <.1 ...h - Se released ta the environs;nt-is in th2 ' vapor phase. The oxidatisn stats cf Se was og - ../t t:rmined to be Se0 basedLupon inefficient extraction in hcl and complete elemental ex- ! traction in Br/Br~-redox buffer,16M HNO,18M H 50, or 1:1 HNO :HC10. A 3 2 4 3 4 Q Mercury emissions from coal-fired power plants have been described in detail. ],*{.f
- .l ',
Billings and Matson (1972) and Billings et al. (1973) studied mercury emission from a power plant burning low sulfur (<l%), high ash (215) pulverized coal. The authors' con- ..: N . cluded that 90% of the Hg was released from the stack as a vapor and that fly ash parti- ~ - % cles represented less than 11 of the Hg emissions. The annual release of Hg from all 3 1 coal-fired United States power plants was estimated to be l'03,,tric tons in 1971.. 1Similarly. Diehl ~ et al'. (1972) studied Hg emissions from a 100-g/hr pulverized coal com-bustor and a 500-lb/hr pulverized coal. combustor. Although. these authors experienced ~,7 difficulties in their collection of Hg from the flue gas, 35 and 60% of the total Hg was 2% found in the fly ashes generated from. combustion of coals having ash contents of 21.6 and dpf 6.9%, respectively, and sulfur contents of 5.2 and 1.25, respectively. Subsequent studies f?j'C -in the larger combustor using coal with 10.1% ash and 2.1% S. resulted in fly ash contain-Nj ing 12% of the total Hg. The authors present calculations for two Illinois power plants, f . indicating that the Hg content of ash contained'within the plants accounted for 7 and 19% 19 of the total Hg in the coal. Thus, in agreement with B"illings' work, most of the Hg in
- fh coal is volatilized and released as a vapor to the atmosphere. Similarly, Kalb (1975) has reported that the major portion of Hg in coal is volatilized during combustion and 4
released to the atmosphere. Approximately 10% of the volatilized Hg was found to be ad- ,g sorbed onto fly ash; organomercury compounds were not observed. The author points out g that Hg anissions'could be reduced by coal cleaning, which results in removal of higher y density minerals, including pyrite that is relatively high in Hg contents. M. In a review of trace elenent studies relateri to low sulfur, high ash coal combustion in Four Corners, New Mexico, Wangen and Wienki (1976) described enrichment factors for .h electrostatic precipitator ash relative to bottom ash. Enhancement in the precipitator-
- o..-
Wi .- ash was observed for the following elements in order of decreasing magnitude: Se, As, F, u... /.s Sb, Zn, T1, Ng, Mo, Ga, 8 Pb, Y, aad Cr. Enrichment factors near unity were observed s. g. for the other 22 elenents studied. 4l.o Kaakinen et al. (1975) studied the behavior of 17 elements in the inlets and outlets of a power plant burning pulverized coal containing 0.6% S and 61 ash. Although particle ~ ' [.- size was not reported, the author described the specific surface area of his samples. The [,(( surface areas ressured by nitrogen adsorption for the bottom ash, mechanical collector ..",h.1 ' hopper ash, electrostatic-precipitator hopper _ ash, and electrostatic-precipitator-cutlet ^b 2 fly _ ash were 0.38, 1.27, 3.06, and 4.76 q /;. rec cctively. Enhancement in trace .
- element concentration relative to Al was observed for Pb, Mo, As, Zn, $b, and Cu.
g The magnitude of the EF's correlated with relative distance of each outlet downstream - ff from the boiler and the specific surface area of the ashes. The authors point out that [.'[ As enrichment depends on the Ca contant of the coal; As2 3 is associated with low Ca coals 0 while As2 5 is associated with high Ca. Zirconium was the only element displaying a de-0 -(.,( crease was thought to be because of the occurrence of Zr as zircon, a relatively high den- % -[ sith mineral that may be more efficiently captured by the mechanical collector. Contrary to this observation, little or no enrichment was reported for Nb, Sr. Fe, Ab, and Y. '.q 28 3g
- i
'mW ' Ondov ct al. (1977b,c) have performId extensiv2 analyses of element enrichments in l 'e, 1g kly ash as a function of particle size. In th2 study of two large western powIr plants burning high ash, low sulfur, pulnrized coal. Ond:v et al. (1977c) reported consid:rable enrichment of W, U, Ba, 2n, V, In, Ga, Br, As, Se, Sb, and No in fine particles for the plant with an ESP rated at 99.5% efficiency. In the second plant, with a 97% efficient
- c ESP, EF distribution tended to be bimodal for these elements, with a broad maximum of og 4
2-10 us. The authors also point out that Br, Se, Cr, Mn, Ta, Co, and Zn displayed 'en-tf' 'Y richment in both the fine and the large particles, i.e., an EF minimum was observed from approximately 1 to 8 um. The authors indicate that the biphasic distributions may be the result of artifacts in collectior because the larger particles will be collacted on the f first impactor stages, through which vaoor containing volatile elements is initially drawn. l /ll-Fisher et al. (1979d) also reported data supporting bimodal elemental distributions. Filtration studies with neutron-activated, stack-collected fly ash (VMD = 2.2 um; as" 1.8) were performed by dispersing ash samples in buffer at pH 7.4 and filtering thrcugh e a membrane with pore size of 5, 2, 0.8. 0.4, 0.2, 0.1, 3.05, or 0.03 um. The elements were classified into four groups based on their behavior: (1) Na, Ca, Co, Se, Mo, and Ba were partially soluble and did not display filtrate concentrations that were pore-size 4 dependent; (2) Sb, As, In, W. Cr, and U displayed a pattern of filtrate concentrations 4 that appeared to be bimodal; (3) 7, Si, Fe, Ce, Sm. Eu, and Th were only detected in filtrater; from membranes >2 um in pore size; and (4) Zr, Cs Nd, Rb, Tb, Yb, Hf, and Ta j were not detected in the filtrates. For those elements displaying bimodal behavior, a i relatively large increase in concentration was observed in filtrates derived from the vs .E 0.4 um membrane. The concentration profile remained constant thereafter. These data
- ,i suggest a concentration maximum for Sb, As, In, W. Cr, and U in fine particles less than O.4 um in diameter.
Ondov et al. (1977c) have compared en:ictenent factors for the two power plants to y those published by Klein et al. (1975), Xaakinen et al. (1975), and G adney et al. (1976). ,f The comparison (Table VI) for EF's for elements ir. stack-collected fly ash incicates d relatively good agreement tietween studies of different power plants with ESP control systems employing a wide variety of coals. In light of the uncertair. ties, only Mo, Se*, and Mn showed significant differences between plants. The volatile elements Sb, As, and f( Pb were clearly enhanced in samples from all power plants; Zn, Se, Cr, and V were en. hanced in stack ash from those plants with tha most efficient ESP's, i.e., those plants presumabl3 releasing the finest ash. Bromine was the only element displaying a sig. nificant fractional EF. Ondov et al. (1977c) point out that the EF's for stack ash collected from a erit with a venturi wet scruot,er (VWS) art ganeral:y much higher than those for plants with ESP's. The authors attribe these findings, in part, to the high efficiency (iS9%) of removal of particles >2 um and the low efficiency (40%) of removal of particles <2 um by the VW1. In another study, Ondqv et al. (1979b) indicated i that the ratio of VWS-tc-ESP fracticral emissions of submicron, supe.rmicron, and total surpended particles were 1:6.11:1, and 10:1, respectively. They also proposed that 1 J l 3 29 I i .. r= c..- _---. ;, ;;., =, :::,...=f-.f. ---,,- p q- --, i. w w . +.
v - [ ...C E' i:.,.i"J
- G Table VI. Enrichment Factors far Elements in Stack Fly Ash from Coal-Fired Power Plag IM och w'
'. / Western Western Western .i ;%)F U.S.bplant U.S. plant 8 Allen Steam U.S.plantg
- r..cf..f A
(ESP)C plantd Chalk point' Valmont (VWS)9 f t 0,i,wtj Sb 7.0 5.3 6.7 4.0 120 ~ '.INj Cd 6.0 W 4.9 70 .W As 6.6 7.9 6 6.3 100 .s ..4 In 5.5 3.7 20 [ Zn 4.3 4.3 7.8 1.5 2.5 19 4Yd Pb 3.8 8.1 3.7 3.1 '.<v , 2., / - Ga 4.1' 3.0 1.2 ..c 1 U 3.3 2.5 . 4.; 13.5 Pw.; Se 3.0 5.3 5.5 5.7
- 1. 7.
400 f.37. Ba 2.5 2.7 0.7 0.92 13 (*,'[.5 Cr 2.3 2.6 3.0 1.1 100 j%'] Co 2.3 1.7 1.4 1.0 4.3 ' 2;. V-2.0 2.5 2.5 0.75 21 ~..1 k 1.8 3.5 3.0 43 '.dl ~ 0.8 0.54 Mg 1.1 2.7 ?.d:" Fe 1.1 0.90 0.84 0.83 1.0 2.0 - n. jf ( Na 1.0 1.1 0.99 ' 3.2 T, Sc 1.0 1.0 1.0 1.0 f. 1.0 F'g.;? K 1.0 0.7 0.95 0.83 0.86 .b.U Th 0.95 0.90 0.76 0.89 m.
- .;.,;.. l A1 0.86 0.75 '
O.44 0.83 0.94 1.3 Y.<; '. Ca 0.76 0.89 0.92 7.6 . J f.11 Mn 0.68 1.1 0.78 ~' 21 , -).: 0 Be 0.6 0.64 . :.s J C.,
- Br 0.2 0.1 0.17 57
?tlR "$1 ~.'5
- Modified from Ondov et al. (1977c)'.
b Plant A employed an ESP with removal efficiency of 99.6t (Ondov et al.,1977c), c -1 c Plant S employed an ESP with efficiency of 97t on one unit and a venturi wet ,4 ti scrubber (VWS) on a second unit (Ondnv et al.,1979). d ,. g Employed an ESP with 99.5t efficiency (Klein et al.,1975). P. /' c'5 /f ' N ~,., ')
- Employed an ESP with 75% efficiency (Gladney et al.,1976).
y +,g., f Employed a mechanical collector and an ESP with 91% efficiency (Kaakinen et al.,
- ~ 5 1975).
st
- ...v;:
- 3:;
"]
~ Thus, although the WS may = -. corrosion may enhance WS emissions of Cr, Co. Cu, and Zn. ' have a higher removal sfficiency of total suspended particulate matter, the ESP may more i efficiently remove respirable particles. j Ondov et al. (1977b) have reported enrichment factors for~ plume samples collected 8t,3 from a power _ plant with five generating units, of which two units were equipped with ESP's and the other three with WS's. Elemental enrichment factors were relatively con-stant as a function of distance from the stack for Sc Na, K. Cu, and the lanthanides. Enrichments for Mo, V, Ba, U, Ga, In, As, W, and Se increased from the stack to the Subsequent plume samples indicated decreased EF's with distance from the point of plume. The only elements displaying increased enrichments with increased distance fl release. from the ' stack were Br, Sb, In, and Co. The increased Er for Br was postulated to be because of mixing of plume aerosols with high background concentrations of Br, possibly [$ because of automotive sources (Ondov et al.,1977b). In a further comparison of the stack fly ash from an ESP unit with that from a WS l unit, Ondov et al. (1979b) reported that the mass median aerodynamic diameters (MMA0's) for the elements As, Ba, Sb, Se, U, V, and W in the ESP ash were approximately tenfold J The authors concluded 4 higher than in the WS ash, which ranged from 0.47 to 0.59.un. j that despite an eleven-fold higher total particulate emission, the ESP unit is far more efficient at removing submicron particles than is the WS unit. Thus, the scrubber unit ? tested appeared to be less effective at reducing potential inhalation hazards than the precipitator unit., i i.J .t; ~ C. Surface Deposition Models A number of investigators have presented mathematical models relating the concen-
- 1. '
j trations of relatively volatile elements to geometric parameters associated with fly ash
- f Assuming a volatilization-condensation mechanism. Davison et al. (1974)
( particles. proposeo a simple mathematical model for elemental concentration as a function of Their model predicts that the elemental concentration of a volatile particle size. j Kaakinen et al. (1975) presented a species will be inversely dependent on particle size. similar mathematical dependence based on the specific surface area (square meters per , s If the specific surface area is proportional to the surface area:vol-gram) of fly ash. une ratio and if particle sphericity is assumed, then elemental concentration is inverse-ly proportional to particle size. Based on mass transfer arguments, Flagan and and Friedlander (1976) indicated that concentration should be inversely dependent on particle size for Knudsen r. umbers >l (i.e.,'for_ ccndensation when tha particle size is greater than the mean free path of the depositing gas'; and inversely dependent on the. square of the particle size for Knudsen numbers cl. - Aoplication cf this model fits Smith et al. (1978) existing data equally as well as the model of Davison et al. (1974). extended the Flagan and Friedlander (1976) and the Davison et al. (1974) models to in-
- 7.
- clude fire psrticles in which the thickness cf the deposited surface layer approached the
) ] diameter of the total particle. This modification resulted in concentrations that The models were demonstrated 4 I asynptotically approached maxima at particle size <1 um. to fit the concentration dependence cn particle size of reaerosolized ESP-collected a 31 1 1 '1 1 m -.. n.- m==
- az :...==.:=:.
- .._,.. am,.p ye--r-.
-- ~ ~
.. d< - 8' tb fly ash. ~ 8iermann and Ondov (1978) have proposed a model with an inverse squa enc 2 and an asymptotic maximum for concentration as a function of surface thickness. 7 d . Their results indicated that the thickness of surface-deposited chemicals is inve proportional to particle size and that total elemental' composition is proportional to 2 I .l'/d, where'l is the thickness of the surface layer and d the diameter of ! 1 Analysis of 12-stage impactor data with increased resolution in the submicron re supported the mathematical model. Further studies are required, however, to extend the. presently available data on concentration as a function of particle size thus al evaluation of the validity of the existing mathematical models. 4 D. Sunnary In summary, most studies of the size dependence of elemental concentrations in } fly ash support the hypothesis of Natusch et al. (1974); the more volatile elements (or chemical forms) are preferentially associated with fine particles. The fine particle mode (<1.0 pm) in the bimodal elemental distributions is generally considered 2.j to be because of coagulation of primary particles (Whitby,1977). Bimodal size distri-butions may also result from the presence of multiple mineral forms, some o i decompose or may be associated with a fine mineral grain size. It should be noted, how. 9 ever, that the bimodal distribution of the very volatile elements (Se, Br, and I) observed in impactor samples may be artifacts due to vapor condensation on t particles collected on the first impsetor stages. Also, the bimodal distribution of - metallurgical elements may be associated with entrainment of corrosion products in j flue gases. Similarly, small particle enhancement of relatively nonvolatile elements j l may be because of a combination of. decomposition, chemical reaction, mineral size, or elemental association in the organic phase of coal. r a V. MATRIX AND SURFACE CG! POSITION OF COAL FLY ASH A. Matrix Comoosition Elemental analyses of coal fly ash show that the major mat-ix elements are A 1 and Fe together with a few percent of Ca, K. Na, and Ti. Fly e hes derived from western United States.sub-bituminous coals generally contain higher levels of calcium than do bituminous coals and lignites. The actual compounds that constitute the fly ash matrix have been identified for a comparatively small fraction of the mass. The techniques that have proved most ). useful for this purpose are x-ray powder diffraction and infrared spectroscopy 1 (Natusch et al.,1975). In addition, selicted area electron diffraction has been em-ployed in the toentification of small crystals often found associated with the surfa of fly ash. X-Ray powder diffraction studies have demonstrated the presence of a-qu mullite (3A1 0 2510 ), 0 2* hematite (Fe2 3), magnetite (Fe30 ), lime (Ca0), and gypsum 23 2 4 (CaSO 2H O) in aged fly ash (Natusch et al.,1975; Miguel,1976). 4 2 However, there is evidence to suggest that crystalline species, associated with aged fly ash may d from those in the freshly collected material (Fisher et al., 1976, 1978) because of f .32 4
1 I4-4,"troh' citiher the prssence er lack of moisture in storage atmospheres. In addition to these /# t- , t,.
- i crystallin 2 species, x-ray powder diffraction patterns indicate the prtsence of a sub-
%/. stantial amount of mat 1 rial that is amorphous to x-rays (Fig.11). Th2 composition of )
- t**t this material has not been established with certainty; however, it is widely accepted f
'I that it consists of an impure aluminosilicate glass and constitutes the bulk of the fly 1 ash matrix (Natusch et al.,1975; Walt and Thorne,1965; Simons and Jeffery,1960). tg Infrared spectroscopic identification of ' norganic compounds present in coal fly ash i y9 has largely been restricted to the tentative identification of residues of evaporated ~ p aqueous leachates (Jaknbsen et al.,1978). Several sulfate species have been identified; however, it is not clear whether these represent the actual compounds that existed prior to removal from the fly ash. In addition, studies have been made of glass melts derived f. from oxides of aluminin, iron, and silicon (Henry et al.,1978). These have provided information that suppo ts the conter. tion that the matrix of coal fly ash is predominantly an aluminosilicate glass. B. Trace Elemental Distributien Further insights into the facters that determine the distribution cf elements in a i bulk fly ash sample have been obtained from multielemental analyses of the 32 subsamples O presented in Table IV. Specific concentrations of the elements A1, As, Ba, Ca, Co, Cr, ] Cs, Dy, Eu, Fe, Ga Hf, K, La, Mg, Mn, Na, S. Sb, Sc Si, Sm, Sr. Ta, Ti, Th. Y, and Zn V were determined. In addition, x-ray powder diffraction patterns and BET surface areas 5 (by nitrogen adsorption)were obtained (Natusch et al.,1975). a ,f As, an aid to the interpretation of the extensive data sets obtained, multivariate h statistical analyses, in the form of both connon-factor analysis and hierarchical aggre- .[ gative cluster analysis (Harmon,1967; Blackith and Reyment,1971) were employed. Cor. con factor analysis makes it possible to determine tre way in which each measured variable;in I the system is related to a set of n factors conncn to the system as a whole. The q important causalities that give rise to the observed data can thus be inferred. By ,5 comparison, cluster analysis permits an objective assessment of the similarity between individual subsamples. y The results obtained indicated that the distributional pattern of trace elements in q fly ash is cuntrolled by five major factors. These factors have been interoreted to in-clude particle size, particle composition, and the geochemical behavior of the elements. Thus, specific distributional patterns are otserved for the chalcophile, lithophile, and siderophile elemer.ts as classified by Goldschmidt's Geochemicsl Series (Bertine and Goldberg,1971; Coles et al.,1979b). It would appear, therefore, that the size factor arises as a result of the volatilization and condensation of certain terce metals as described earlier (Davison et al.,1974). The dependence on particle ccmposition possibly reflects the association of some elements (e.g., As and Ma) with certain types of mineral inclusions. The dependence on geochemical class of the eier.ents, in all probability, reflectr the different chelical characteristics of each of these classes under high temperature combustion conditions. 33 . -.---- r.ars :mvT .c..a -. wgv a,,,-,,-__ g%_, _
-A y M SEM-x-ray analysis has provided further insight into the complexity of the mates, i.I composition of cal fly ash. ElemIntal analysis of morph 2 logically similar fly ash ^
- 3 gig particles from the NBS fly ash reference material indicated extreme matrix heterogenet.,
i (Pawley and Fisher,1977). Particles rich in K. Ti, Fe, S, or Ca were observed. Indeg.- ~nearly all of the Ti in a field of 100 particles could be accounted for by a single \\ r.] Ti-rich particle.. It is interesting to note the extreme matrix heterogeneity of indi. vidual particles in the N85 fly ash, a material that is well documented as being homo-gg geneous by macroscopic analytical techniques. .m <- 4 'a C. -Surface Composition Wr. As pointed out in previous sections, the inverse dependence of trace elemental con.
- ]
cent' ration on. fly ash particle size is generally held to be due to condensation of ,g metallic species onto particle surfaces from the vapor phase (Davison et al.,1974). One E.y would expect, therefore, to find certain votacilizable elements preferentially concen-trated on particle surfaces. This has been observed (Linton et al., 1976, 1977; Keyser et al.,1978). , Y3 The techniques that have been employed, to date, in analyzing the surface regions of 3,1 coa 1 fly ash are electron spectrometry for chemical analysis (ESCA), Auger electron spectrometry (AES), and secondary fon mass spectrometry (SIMS). In addition, some surface ~ ? -J analytical infomation is available using electron microprobe x-ray spectrometry. The operational characteristics of these techniques are summarized briefly as folicws (Czanderna,1975; Kane and Larrabee,1974; Keyser et al.,1978). !y The electron microscope (EM) and microprobe (EP) bombard the sample with a focused 3 beam of electrons to stimulate emission of x-rays characteristic of the elements present. The technique is useful for analyses of individual micrometer-size particles an'd has a lateral and depth resolution of about I um, detemined by the, x-ray emission volume. 'N The electron probe microanalyzer is described in Chapter 48. -M -f Surface analysis capabilities of EH and EP are poor since the depth resolution is 4 very much greater than the thickness of the surface layer nemally of interest. Indeed, M information about elemental surface predominance can be obtained only by varying the energy of the electron beam (depth penetration) or by fon etching of the outer surface , gj and by comparing elemental ratios for inner and outer surfaces. iQ The ESCA technique employs an x-ray source to eject core-level electrons from the
- d sampl e.
Energy analysis of the resulting photoelectrons provides chemical bonding in-l :._ $ famation since the bonding energies of the core electron are sympathetic to changes in .h.~ the electronic structure of the valence level. Elements present at levels greater than a
- 1. at. ". in tne uppermost 20 A are detected. Depth profiling is achieved by etching the
'1 surface with an ion beam between analyses. For details on ESCA (or XPS) see Volume I, ,:l[ Chapter 11. I@g The utility of ESCA for individual particle analysis is limited because of the 'y difficulty of focusing x-rays to a beam dia::.eter smaller than 1 m, although recent ad-4,7 vances indicate that lateral resolutions of 10 um are feasible. Normally, the sensitivity , -y of ESCA is insufficient to enable observation of trace constituents unless considerable j surface enrichrtent is encountered. ' i +r '.Z 34 . (< J
In AES th2 emission of Augtr electrons is stimulated by bombarding the sample with a 'bea2 cf electrons. The entrgy of the sec:ndary ugir electrons is characteristic of the emitting element. Spectra are recorded in the first derivative mode to discriminate I against a background of inelastically scattered electrons. Elemental detection limits i lie in the range 0.1-1.0 at. % within the analytical volume (depth -20 A). Depth pro-f filing is achieved by etching the sample surface with an ion beam (normally Ar+) as in ESCA..Most AES spectrometers possess microprobe capabilities with incident beam diameters of 1-5 um. In SIMS the sample is bombarded with a stretm of ions (most comonly, negative oxygen ions) and surfact material is physically removed. About 1-10t of the sputtered material.is in the fom of se oncary ions that are mars analyzed by a conventional mass spectrometer. The ion microprobe represents a special configuration of SIMS in which the primary ion beam can be focused to a diameter of about 3-S um. Both ir.dividual particle analysis 3 and elemental-mapping capabilities are thus available. Depth prcfiling constitutes an integral part of the process of secondary fon generation. A major advantage of SIMS is its er.tramely high sensivitiy, with elemental detection j limits ranging frcm 10-E 6 at. %, depending on the element and the primary ion used. i . Typically, it is possible to observe as little as I ug/g in the analytical volume, thereby enabling studies of species present at trace levels. Secondary ion mass spectro-metry is, however,* subject to several types of interferences and artifacts. In parti-if cular, spectral interferences from molecular-and multiple-charged ions make the high i resolving power of a double-focusing mass spectrometer cestrable. ' Also, volatilization ^ losses and migration of sample ions under the influence of the primary fon beam can give rise to spurious depth profiles. Such effects a:e eften difficult to identify in SIMS since removal of sur# ace material is an integral part of the detectier process. l Of the above techniques, AES and SIM! are generally most useful for surface analysis i and the depth-profiling studies, owing to their sensitivity and good lateral and depth resolution. Electron spectrometry for chemical analysis, however, has the important ad-wantage of providing infomation about the identity of molecular species present. With all the techniques, difficulties are encountered in establishing even semiquantitative depth scales, which are nomally attempted by calibratir.g the rate of removal of surface I materf 41 against that obtained for a standard having a surface layer of kriown thickness. The main problem, however, lies in matching.the matrix ccmposition of the standard to that of the material being studied, which, in the case of coal fly ash, is not well definea. Surface analysis and depth profiling studies of both indiviudal coal fly ash particles and groups of oarticles have established that a number of trace elements, in. ciuding C, Cr, K. Mn, Na, Pb, S T1, V, and Zn, are substantially surface enriched, whereas the matria and minor elements, A1, Ca Fe, Mg, 51, and Ti, are not (Linton et al., 1977). lhis observation clearly supports the hypothesis that the more volatile elements, or their compounds, are vsporized during cembustion and then condense on the surfaces of coentrained fly ash particles at lower terr.perature. 1 ] 35 = _. _ .,.,.,w.-- .~,..,n
W Depth profiling stud 1Gs of fly ash have als) demonstrated the utility of using in. '.I,'[.. ~ strumental techniqu;s in conjuncti:n with'ssiv:nt leaching to remove soluble surface [j material. An example of this approach is presented in Fig.12 for the eleme
- "s This study demonstrated that extraction of fly ash with water or dimethyl sulfoxide re.
moves the surface layer of both elements. Determination of the amounts of Pb and T1 in solution then enables estimation of the amounts present in the surface layer. Assuming \\ 7,% y.%.% a surface layer thickness of 300 A, one obtains average concentrations of 2700 ug/g for Pb and 920 ug/g for Tl in the surface layer as compared to bulk particulate concentra-4 .sps tions of 620 ug/g and 30 ug/g for these elements. Similar estimates fo'r several other 7*d trace elements are presented in Table VII (Natusch,1978a). Solvent leaching can also provide some insight into the chemical foms of elements w ?.} For example, although AES and SIMS indicate little surface enrichment of iron, present. [...N M.% aqueous leaching rapidly removes this element from the surface region, thereby indicating its presence in a readily soluble form. Similarly, comparison of the leaching and ~ -Qk depth profiles of K, Fe, Na, and S suggests that these elements may be associated with _.] each other in the surface layer, possibly as alkalf-iron sulfates. Fu: ther support of h the existence of simple and/or complex sulfates is provided by ESCA studies that show -c M Nb that the oxidation states of Fe and 5 in the surface region are +3 and +6, respectively (Wallace,1974). Surface analytical results, such as those presented in Fig.12 and Table VII f,L} demonstrate the considerable differences in composition that exist between the interior ' */ of fly ash particles and their external surface. Since it is the particle surface that v.4 ,~ is in contact with the external environment, determination of surface composition is of M considerable importance, as previously discussed. Finally, it should be remarked that
- h. N.q there are no coherent data that relate surface composition to particle size for. fly ash.
Q However, if the volatilization. condensation process is primarily responsible for surface El enrichment of trace elements, then one would not expect surface concentrations to vary v. . f.. ', greatly with particle size. This is because the amount of vcpor deposited is propor- .e i!rk]j tional to surface area, thereby resulting in a constant elemental concentration per unit surface area. Of course, if other mechanisms are responsible for or contribute to [ surface enrichment (e.g., agglomeration of accumulation mode particles with coarse ' 'I,3 particles or thereal diffusion of trace species to the surface of molten particles).
- Q then some dependence on particle size would be anticipated, i
l 2.I D. Solubility and Leachability
- ~. A 9)P.l A number of workers (Shannon and Fine,1974; Theis and Wirth,1977; James et al.,
t.n N' 1977; Dreesen et al.,1977) have reprted tm the buik solubility of coal fly ash in e,j water is very low and rarely exceeds 2-3% by weight. The bulk solubility is clearly a property of both the glassy and crystalline matrix materials identified earlier, and one i,p would expect elements that are either chemically or physically trapped within this matrix .r. Q to exhibit low solubility. On the other hand, at least some of the material present in a. the surface skin is readily soluble in water (Fig.12). Indeed, it is now quite well es-tablished (Linton et al.,1977; Natusch,1978a; Fisher et al.,1978e) that most of the !Q soluble fraction of fly ash is derived from this surface layer and is thus very rich in l,, i trace elements. 36 i r., '
Grh' E.1 wao oeoen(L 4 *..j. Ff 9..12. W I:n cicroprabe depth profiles for o N To (80 Se eco %c y, e % ( 'aoen* i Pb and T1 f.r un-extracted and ex-y 7, tracted fly ash ze k D samples (reprinted ~ "000} 4 with permission j l p from R. W. Linton, P. Williams, C. A. .~ '80 - 3oooo$ e Evans, Jr. and D. T. j S. Natusch, Anal. - 2o000 Chem. 49 15 C ao j TT977)7, Copyright N,. cooo g 1977 by the American ( f Chemical Society). c o i s 30s.n* 2 g x
- a'
- 4000 j ) s jM h \\ R E - 2000. j .I t O 40 ao 12 0 Kso 200 240 Tene(tecs) = =uwiereci e..ouso cair.c,.e ..np g,,, e fw 4:- .c .) Table VII. Estimted Surface Concentrations of Elements in f Coal Fly Asha . :..') Estimated surface
- )
Bulk concentration in Element concentration 300$ layer g "2j (99/9) (99/9) ~
- [;
As 600 1500 .'.?[ Cd 24 700 / Co 65 440 Cr 400 1400 Pb 620 27C0 S 7100 252,C00 tg ~ 330 760 y 4' ~
- From Natusch (1978a).
1 l 37
- ~#
. - -,3%?aMG" dW'PJhb.fN ^7 Ff*ETJ EEW W" " T#^ 8
- * * * * ** *
- semeesumusp ;g ge wg.aret.m _ _
- ~ -,.,.,, - _. _ _..
,g %[ t* ; ' At th] pres;nt time, th1re is consid:rabia confusi:n involved in interpreting ane y,{ understanding results obtained from differ:nt studies of fly ash solubility. Qff quite different results are obtained by workers using apparently s leachir:g techniques and none are readily transferrable to field studies. It is appropri. ate, therefore, to consider the factors that control leachate composition under both i.d,. laboratory and field conditions and to standardize one or more laboratcry leaching ,g techniques whose results bear some relationship to real field conditions.
- 1 Some insight into leaching behavior can be obtained by recognizing that soluble in-3 organic species M A associated with fly ash particles can dissociate into their compon-g9 y
ent cations Hj and anions Aj in aqueous solution, and that both dissolution and de- .,' { position (e.g., precipitation) processes can occur. Furthermore, cations and anions .j present in solution can interact so as to set up multiple equilibria that may involve .M fon pairing, complexation, precipitation, or acid-base behavior (MgAj or MjAg).The
- M. *R result can be expressed, simplistically by the equations ki n
n n ',ih P IM Agg ; P+IM + IA g g .g k,j -r + + %' tf jA) jM j (10) g k-2'+k2 k,3,4k3 ..M. i
- p m
mn en .Mit 'W I M A) IHA g jg fJ IJ . ~, .,h Here P represents the parent particles and k, k-n are the rate constants for forward and n reverse reactions, respectively. It is apparent from Eq. (10) that, when leaching -i.5 studies are conducted under batch conditions, such that the amount of fly ash and solvent "~:< are maintained constant, an equilibrium will be established between particulate and ,'7. solution species. Consequently, only a fraction of the potentially soluble material will .? S enter solution. On the other hand, if conditions are such that soluble material is 7 N.i.. continuously removed by provision of fresh solvent or by providing a large solution sink
- b';
in the fom of complexing ligands or acids, then all potentially soluble material wil 1 ultimately enter solution. j-Matasiewic: and Natusch (1979) have concucteo very extensive studies to demonstrate the validity of Eq. (10). They have established that the rate and extent of leaching de-a pend upon the leaching method, the fly ash: solution ratio, temperature, pH, complexing ,N agents, particle size, and fly ash origin exactly as predicted from Eq. (10) for both ' ' M.. equilibrium and nonequilibrium conditions. When equilibrium is established between particulate and solution species, as in batch leaching, little dependence on particle ,,r l *a 9 .a l 1 38 ..9 '?
%tt size is obs;rved sinco the amount of a givtn elemtnt is detemined by its solution con-89,-
- $/
hentraticn (solubility) that 1.s only weakly related to the amount of solid phase present. /, Exhaustiva (nonequilibrium) Icaching, howev;r, removes all soluble material, th2 amount
- r,
.of which is directly related to particle size because of its presence in fly ash surface A c t4 layer. '9 It is apparent from the foregoing remarks that any leaching process that establishes the solid-solution equilibrium given in Eq. (10) will result in changes in solution ' ' /,A composition with equilibrium position. It is hardly surprising, therefore, that widely h, differing results are obtained by workers who use different leaching conditions. Never-theless, some significant generalizations can be made. First, it is readily apparent that much higher proportions cf most trace elements are soluble than is the cr.se for matrix elements (Table VIII). This is due, in part, to the predominance of trace /i elements in the particle surface layers in quite soluble forms (probably sulfates, oxides, and carbonates). Second, it is clear that, under batch letching conditions, such as are most likely to occur in the field, the amount of each element entering solution is strongly dependent on the dilution (fly ash: water ratio) and the initial pH (Matusiewicz and Natusch,1979; Dreesen et al.,1977; Theis and Wirth,1977). g .3, h Table VIII. Percentage of Elements Leached from a Typical Coal Fly Asha 1 Element 1 Leached A1 0.2 ^ 8 5 h 4 1 Ca 35 Cr 30 K 40 hg 0.2 Mn 0.4 2 Mo 85 Na 10 P 6 l Pb 100 $1 0.1 Sr 6 'S Zn 6 a l,,",.) From Matusiewicz and Natusch (1979). s
- 4 e
i 9 39 ~ ~ -. - <,,v.r.,.. _. - % m,_ . ~ - - ~ _ _ _ _... _ _ _ _ _ _ _.,. _ _ _.. _ _ -. _. _,, _. - _ _ _ _ _ _ _. _ _ _ _ _.. _ _ _ _ _
.. x.1 ..m%o.. E. Orcanic Constituents .M To date, no exhaustive determination of organic species associated with coal fly ash \\,, 'i has been reported. Rather, emphasis has been placed primarily on the determination of polycyclic organic matter (POM) in fly ash, due to the potential carcinogenicity of , a.s several compotnds of this type (Committee of Biological Effects of Atmospheric Pollutants. 1972). For the most part, the several studies of POM in fly ash have indicated either extremely low or undetectable levels (Cocruittee on Biological Effects of Atmospheric ,l-h ' Pollutants,1972). In a survey of fly ashes representing several coals and combustion conditions, '[ Aslund et al., (1978) found no individual species of POM to be present at concentrations t greater than 20 ng/g. A number of other unidentified organic compounds were, however. g4h; observed at somewhat higher (10x) concentrations. It is important to note that all of Y9[ these st'udies have considered fly ash collected in bulk from power plant control devices. f.8 Only a few studies have been made of PCM present in fly ash that was actually MM.4 emitted and collected from the atmosphere (Natusch,1978b; Tomkins,1978; Stahley,1976). ,sfir However, all have indicated concentrations that are very ruch greater than encountered >.u Yf8l in fly ash collected within the plant. This apparent paradox has been explained by Natusch and Tomkins (1977) who postulate that POM (and probably other organic species) are present as gases at the temperatures encountered within a power plant but rapidly and tj, quantitatively adsorb onto surfaces of emitted fly ash particles as the temperature falls J on leaving the stack. Both laboratory (Miguel et al.,1979) and fleid (Miguel,1976; .k Natusch,1978b) studies support this hypothesis. i ( The actual compounds that have been identified in emitted fly ash are listed in 59 ' Tables IX and X, which present the results of two separate studies in which spepific ',[ concentrations inside and outside the plant, and volume concentrations in the plume.
- 5 were detemined. To our knowledge, only cne study has actually measured POM concentra-
- . y'.;
tions as a function of particle size for emittee fly ash (Natusch,1978b). The results a n p., indicated 11ttle convincin2_dADendCD_Cf._of concentration on aerodynamicJarticle.4ae s U..1' over the range <1.1 to >7.0 pm. However, the fly ash in question was derived from a
- .e g,9 small plant that employed a chain grate stoker, and the particles were found to be ex-
. M, tremely irregular in outline. Furthermore, there was very little change in specific surfacQea,nygr the size range collected We do not, therefore, consider these re-3 '{.y suits to be conclusive. 7,g In fact, if the temperature dependent adsorption mechanism proposed by Natusch and ,M Tomkins (1977) is correct, dne would expect the specific concentration of organic species to vary in proportion to the surface area of the fly ash particles. There is some in-direct evidence for this behavior (Chrisp et al.,1978; fisher et al.,1979c), but /M further work is clearly required. It has been established, however, that adsorption of ( PCM (pyrene) onto fly ash under laboratory conditions occurs, to significantly different extents, on different fly ashes and on magnetic and nonmagnetic fractions of a given fly k.e;y ash (Miguel,1976; Korfmacher et al.,1979a). i Nf.a 4y h' 40 . 'N M
.f** Tablo IX. Measurement of Polycyclic Organic Matter Emitted from a Coal-Fired Porr Plant Stack 8 O
- *t)
CF 5pecific concentration (ug/g) Ds, Compound Inside stat.k Outside stack D Fluorene ND Trace Phenanthrene ND 9 Fluroanthene MD 19 i f Pyrene ND 12 8enrofluorene NO 2 ~ ', 1-ne.chylpyrene MD 1 Benzophenanthrene ND 3 Benzo [alpyrene ND 5 Total fluorescence 3.61 x 10-3 units 3.68 units C a From Tomkins (1978). b J Not detectable. .i o Table X. Emission Factors for Polyc.velic Organic, Matter from Coal Fired 4 Furnaces in (pounds / ton of coal) x 10 Pulverized Chain grate .3 Species firing stoker Hand fired 2'} 8enzo[a] pyrene 0.2 0.52 0.3 3520 Pyrene 0.8-1,6 3.5 5260 Benzole} pyrene 0 -2.3 1.1 880 $26 Perylene 0 -0.6 6.0 8800 Fluoranthene Committee on Biological Effects of Atmospheric Pollutants,1972. a l \\.' o 41 'i l l i l
e :., "?.J Finally, it should b] mentioned that POM ass:ciated t:ith fly ash may undergo I' chemical transformation following adsorption and emiss,1on. In this regard, Korfmacher et al., (1979b) have shown that adsorption onto coal fly ash effectively stabilized most POM against photochemical decomposition, but actually promotes rapid (hours to days) non- .s , J; photochemical oxidation of polycyclic aromatic compounds possessing one benzylic carbon atom. Furthermore, Hughes and Natusch (1978) have shown that exposure of PCM adsorbed on fly ash to typical plume concentrations of sulfur dioxide and nitrogen oxides results in very rapid formation of a variety of derivatives having sulfur or nitrogen containing i 1 substituents. It is possible, therefore, that the chemical nature of PCM associated with coal fly ash emitted from a power plant is likely to change dramatically with time + and distance from the plant. s .:.3 'h i
- .?
- ..e x;
.k.I 1 .j .4 ? ~!.$ .1. **; r.'y ,g ,.r i *.N ',. 9 '1 , f.: bY
- .:i i
i t::n j Y ', <
- s. d
-.8 . '; ? Y$ M ?'E
- 8
?*' } ..;.,7 d'*IM i.: 1 '42
- 3
h REFERENCES 'e. ~. - e c Andren. A. W., and Klein. D. H'. (1975). Environ. Sci. Technol. 9. 856-858. ar. j
- t.
Aslund. A.. Miller, M.. Natusch. D. F. S., and Taylor. :). R. (1978). Unpublished results. g Bertine, K. K., and Goldberg, E. D. (1971). Science 173, 233-235. Bickelhaupt. R. E. (1974). J. Air Pollut. Control Assoc. 24, 251-255. t Bickelhaupt R. E. (1975). J. Air Pollut. Control Assoc. 25, 148-152. Biemann. A. M., and Ondow. J. M. (1978). " Application of Surface-Deposition Models to Size-Fractionated Coal Fly Ash " Prepr. UCRL-81361. Lawrence Livemore Lab., Livennore. California. Billings, C. E., and Matson. W. R. (1972). Science 176, 1232-1233. Billings. C. E., Sacco, A. M., Matson, W. R., Grif fin, R. M., Cenigito. W. R., and Harley, R. A. (1973). J. Air Pollut. Control Assoc. 23. 773-777. l Blackith, R. E., and Reyment. R. A. (1971). " Multivariate Morphometrics." Academic Press New York. Butcher, S. S., and Charlson, R. J. (1972). "An Introduction to Air Chemistry." Academic Press New York. .u I..I Campbell, J. A., Laul, J. C.. Nielson, K. K., and Smith, R. D. (1978). Anal. Chem. 50, 1032-1040. Carter J. L., and MacGregor. I. D. (1970). Science 167, 661-663. 's: Cheng, R. J., Mohnen. V. A., Shen. T. T., Current, M., and Hudson. J. 8. (1976). J. Air ~ Pollut. Control Assoc. 26, 727-824. .\\ Chrisp C. E., Fisher, G. L., and Lamert, J. E. (1978). Science 199. 73-74. } Coles. D. G., Ragaini. R. C., and Or.dov, J. M. (1978). Env$ron. Sci. Technol.17, 442-446. Coles. D. G., Ragaini, R. C., Ondow J. M.. Fisher, G. L.. $11berman, D., and Prentice,
- Q' B. A.
(1979). Environ. Sci. Technol. 13. 455-459. Cossaittee on Biological Effects of Atmospheric Pollutants (1972). " Particulate Polycyclic
- 9 Organic Matter." Natl. Acad. Sci., Washington, D. C.
i ' * '2 Czanderna, A. W.. ed. (1975). " Methods of Surface Analysis." Elsevier, Amsterdam. Davison. R. L., Natusch. D. F. S., Wallace, J. R., and Evans. C. A., Jr. (1974). Environ. Sci. Technol. 8, 1107-1113. Diehl, R. C.. Hattman, E. A., Schultz H., and Hartn, R. J. (1972). " Fate of
- Trace Mercury in the Combustion of Coal." Tech. Prcg. Rep. No. 54. Bur. Mines, Managing Ccal Wastes Pollut. Program, Pittsburgh Energy Res. Cent.. Pittsburgh, Pennsylvania.
Dreesen. D. A.. Gladney. E. S.. Owens J. W., Perkins. 8. L., Wienke C. L., and Wangen. L. E. (1 977). Environ. Sci. Technol. 11, 1017-1015. Fisher, G. L., and Chrisp, C. E. (1978). Symp. Appl. Short-Term Bioassays Fractionation Anal. Complex Environ. Mixtures. (Waters. M. D., Nosnow. S., Faisf agh, J. L., Sandhu, S. $.. and Claxton, L., eds.) EPA 60019-78-027, pp. 4414(2, Williamsburg, Va.. 43 .,i = * - - - - -..... -. ___m,_ /
J .q
- Q y*
) .T Fisher. G. L., and Natusch. D. F. 5. (1979). - U. 5. Dept. Energy, NTIS UCD-472-502. $.N*{ Fisher, G. L., Chang D. P'. Y., and Bruauner, M. (1976). Science 192, 553-555. Fisher. G. L. Prentice, B. A.. Silberman, D., Ondov, J. M., Ragainf. At C. Biermann,. h .e 93 A. H.. McFarland A. R., and Pawley J. B. (1977). Jt. Chem. Inst. Can./Am. Chem.
- ., :y
- Soc. Meet., 2nd Div. Fuel Chem. Symp., Properties Coal Ash, Montreal.
- 'W Fisher, G.
L., Prentice B. A.. Silberman, D..' Ondow. J. M., Ragaini, R.' C., Biermann. . 1* : .' D A. H., and McFarland, A. R. (1978c). Environ. Sci. Technol 12, 447-451. v. i C Fisher, G. L., Chrisp, C. E., and Hayes T. L. (1979a). Conf. Carbonaceous Part. ~ Atmos., Berkeley, Calif. in press. Fisher. 4. L.. Hayes, T. L., Prentice. B. A., and Lai, C. '(1979b).. In " Radiobiology ~ Laboratory Annual Report." UCD 472-125 Univ of California Davis. in press. 1 Fisher. G. L., Chrisp, C. E., and Raabe. O. G. (1979c). Science 204, 879-881. Fisher, G. L., $11berman. D., Prentice. 8. A., Heft. R. E., and Ondow. J. M. (1979d). g s. Environ. Sci. Technol. 13, 689-693. Q Flagan, R. C. and Friedlander, S. K. (1976). Symp. Aerosol Sci. Technol., Meet. Am. x Inst. Chem. Eng., 82nd, Atlantic City, N.J. Gay. N. C. (1976). Science 194, 724-725. 1 Gibbon, D. L. (1978). Personal comunication, Guilford Coll., Greensboro North Carolina. 41adney. E. S., small, J. A., Gordon, G. E.. and Zoller, W. H. (1976). Atmos. Environ. . 4 [:.. 10, 1071-1077. Gordon. 6. E., and Zoller W. H. (1973). Proc. Annu. NSF Trace Contaminants Conf.,1st. Dak Ridge Natl. Lab., Oak Ridge, Tenn. CONF-730802, pp. 314-325 l Gj Gordon. G. E., Zoller, W. H., and Gladney, E. 5. (1974). In " Trace' Substances in i[ i a Environmental Health - VII. A Symposium" (O. D. Hemphill, ed.), pp. 161-166. Univ. of Missouri. Columbia. ! k. Harmon, H. H. (1967). %dern Factor Analysis." 2nd Ed. Univ of Chicago Press, if { Chicago. Illinois. ! D Henry. W. M., Mitchel, R.1. and Knapp K. T. (1978). Proc. Workshop Primary Sulfate 5.I' Emiss. Combust. Sources. Southern Pines. North Carolina EPA Rep. No. 600/9-78-020b, t l tg Vol. !!, pp.185-208. l[- Hughes M. M., and Natusch, D. F. 5. (1978). Unpubitshed results. , h< Jakobsen, R. J., Gendreau, R. M., Henry, W. M., and Knapp. K. T. (1978). Proc. Workshop 4 !! rG Primary Sulfate Emiss. Cembutt. Sources, Southern Pines, North Carolina EPA Rep. No. 600/ 9-75-0200, Vol.1, pp. 253-274. James. W.' O., Janghorbani. M., and Baxter T. (1977). Anal. Chem. 49, 1994-1997. i.%j Kaakinen. J. W., Jorden, R. M., Lawasani, M. H., and West. R. E. (1975). Environ. Sci. .r Technol. 9, 862-69. ! (b ! dC Kalb, W. G. (1975). In " Trace Elements in Fuel" (5. P. Babu, ed.). Advances in i Chemistry Series. No.141. pp.154-174. Am. Chem. Soc., Washington, D. C. ! yy ,fy Xane P. F., and Larrabee. G. B., eds. (1974). " Characterization of Solid Surfaces.*
- .h Plenum. New York.
3 t..a s }'., 4*.; 44 '**1
'.y. 9
- d'.1 ex1-o
, j Newtch G. J., Raabe. O. G., and itkler. B. V. (1977). J. Aeros:1. Sci. 8, 339-347. A h Olsen, K. W., and Skogerboe, R. K. (1975). Environ. Sci. Technol. 9. 227-230. C Ondov. J. M., Ragaini. R. C., and Biemann, A. H. (1976). Prepr. Pap. Nat1. Meet. Div. Environ. Chem.. Am. Chem. Soc., San Francisco, Calif. 200-203. d.4. ' - Ondov. J. M., Ragaini, R. C.. Heft. R. E.. Fisher. G. L., Silberman, D., and Prentice, rl.4
- 8. A.
(1977a). Mater. Res. Symp., 8th, Natl. Bur. Stand., Gaithersburg Md. pp. o 565-572. . i.;f
- a Ondov, J. M., Ragaint. R. C., Biermann. A. H., Choquette, C. E., Gordon, G. E., and
-? Zoller. W. H. (1977b). Prepr. Pap. Natl. Meet., Div. Environ, Chem., Am. Chem. Soc., New Orleans, La. Ondov. J. M., Ragaini, R. C., and Biermann, A. H. (1977c). " Elemental Emissions and Particle-Size Distributions of Minor and Trace Emissions at Two Western Coal-Fired .A. 1p]:f Power Plants Equipped with Cold-Side Electrostatic Precipitators." Prepr. UCRL-80254. Lawrence Livermore Lab., Livermore, California.
- 9 Ondov. J. M., Ragaini, R. C., and Biemann, A. H.
(1979). Environ. Sci. Technol.13, g 598-607. -' 't Padia. A. S., Saroffm. A. F., and Howard, J. B. (1976). Combust. Inst. Cent. States )T Sect. Spring Meet., Columbus, Ohio. ,r. 5' Parks. N. J., Raabe. O. G., Bradley, E. W., Teague. S., and Hopkins, B. (1977). In , a -f " Radiobiology Laboratory Annual Report." pp. 77-82. Univ, of California, Davis. Pawley, J. B., and Fisher, G. L. (1977). J. Microse. (Oxford) 110, 87-101. y,,, Raabe. O. G. (1976). APCA J. 26, 856-860. Raabe. O. G. (1978). Environ. Sci. Technol. 12, 1162-1167. ' ? '" Raask. E. (1966). ASME Trans. J. Eng. Power 88, 40-44.
- .i
,.J Raask. E. (1968). J. Inst. Fuel 43, 339-344. 'b Ragaint. R. C., and Ondov, J. M. (1977). J. Radioanal. Chem. 37, 679-691. t* Ruch, R. R., Giuskoter H. J., and Shimp. J. F. (1974). Ill. State Geol. Sury., Rep. . "if No. 72. ,,j7 '7.C Sarof tm A. F., Howard, J. B., and Padia. A. S. (1977). Combust. Sci. Technol.16, M 187-204. . y, Shannon, D. G., and Fine, L. O. (1974). Environ. Sci. Technol. 8, 1026-1028. Q t+. Simons H. S., and Jef fery, J. W. (1960). J. Appl. Chem. 10, 328-336. i C.. Small, J. A. (1976). Ph.D. Thesis, Univ.' of Maryland, College Park. Smith, R. D., Campbell, J. A.. and Nielson, K. K. (1979). Environ. Sci. Technol.13, 553-558. .If? Southern Research Institute (1975). Rep. No. 259, Electr. Power Res. Inst., Palo Alto, California." Survey of Infomat .Y 4 Stahley, S. (1976). Personal communication. ",3 '(q Task Group on Lung Dynamics. (1966). Health Phys. 12,173-207 1 b.'.iy 1 *lr.d . t.pj '46 7
'Tnels. T. L., and Wirth, J. L. (1977). Environ. Sci. Technol. 11, 1096-1100. Tomkins, B. A. (1978). Ph.D. Thesis, Univ. of Illinois, Urbana. J Vandegraft, A. E., Shannon, L. J., and Gorman, P. G. (1973). Chem. Eng. 80, 107-114. A P Watts G. A., McKay, D. S., and Gibbon, D. L. (1978). Personal communication, SN6, NASA-JSC, Houston, Texas. I'd ~ Wallace, J. R. (1974). Ph.D. Thesis Univ. of Illinois, Urbana. a Walt J. D., and Thorne. D. J. (1965). J. Appl. Chem. 15, 585-604. (1976). "A Review of Trace Element Studies Related to Wangen, L. E., and Wienke, C. L. Coal Combustion in tne Foar Corners Area of i,cw.%f co," Infor.tal Rep. i.A-6401-::5. Los Alamos Sci. Lab., Los Alamos New Mexico. .1 Whitby, K. T. (1977). Mater. Res. Symp., 8th, Nati. Sur. Stand., Gaitherburg, Md. pp. 165-173. White H. J. (1963). " Industrial Electrostatic Precipitation," Addison-Wesley, Reading. [ Massachusetts. fi Yeh, H. C., Phalen, R. F., and Raabe. O. G. (1976). Environ. Health Perspect.15,147-N 156. .y: Zoller, W. H., Gladney E. S., Gordon, G. E., and Bors, J. J. (1974). In " Trace ,J Substances in Environmental Health - VIII. A Symposium" (D.D.Hemphill,ed.), -;, ~ q pp. 167-172. Univ. of Missouri, Columbia. l4 3 i.;1 l.*y ,e
- }~J s
- .' q
, *y .s I,? J n$
- ,.*.]
f
- eN
.1 'i 47 1 L L}}