Ind. Eng. Chem. Prod. Res. Dev. 1904, 23, 252-256
252
product formation
under a hood. The instability of the aziridinium complex should ensure that it is not present in the finished fabric.
CH?
Cell-0-
+'\i -
t iCzh5)zN
C ~ I I - O - ( C H ~ ) ~ N ~ Ct~ HNaCl ~)Z
tost
CHz
(5)
with the competing reaction CHZ
OH- t {CzHgi2;'1
\
"Ow
H O ( C H ~ ) ~ N ( C ~ H ~ ) Z(6)
CH2
The aziridinium complex is a hazardous substance and, although not much is present because it reacts rapidly, precautions were taken to prevent coming in contact with this material. These precautions included handling all reagents with rubber gloves and carrying out all reactions
Acknowledgment
The financial support of this research by the U.S. Army Natick Research and Development Laboratories through Contract No. DAAK60-78C-0017 is gratefully acknowledged. Registry No. Diethylaminoethyl cellulose,9013-34-7;monochloroethyldiethylamine, 100-35-6. L i t e r a t u r e Cited Graham, E. E.; Fook, C. F. AIChEJ. 1982, 28, 245. McKeivey, J. B.; Benerito, R. R. J . Appl. Sci. 1987, I I , 1693. Rousseau, R. W.; Ferreii, J. K.; Macnair, R. N. "Perspiration Poisoning of Protective Clothing Materials. Part IV. Adsorption of Selected Vapors and Production of Modified Undergarment Materials", Technical Report, U S . Army Natick R and D Laboratories, Natick, MA, 1983.
Received for review August 1, 1983 Accepted October 20, 1983
Microanalytical Characterization of North Dakota Fly Ash Steven A. Benson," Dlane
K. Rlndt, George 0. Montgomery, and D. Rlchard Sears
University of North Dakota, Energy Research Center, Grand Forks, North Dakota 58202
Fly ash collected from pilot-scale combustion tests of a Beulah, ND, lignite was examined by scanning electron microscope/microprobe, electron specroscopy for chemical analysis (ESCA), X-ray diffraction, and X-ray fluorescence. The fly ash collection devices used include a five-stage multicyclone and a source assessment sampling system (SASS), which are designed to collect size-fractionated cuts. Muhicyclone samples were analyzed by SEM/microprobe to determine the composition and size of individual particles. ESCA was used to determine the species associated with the surface of the particles. Larger samples collected by SASS were examined by X-ray diffraction and X-ray fluorescence. The results of characterizing the fly ash reveal that certain elements have concentrated either on the surface of the particles or in the very small particulates. This was determined by SEMImicroprobe and ESCA analysis of samples collected from each stage of the multicyclone.
Introduction
The goals of particulate research a t the University of North Dakota Energy Research Center (UNDERC) include detailed characterization of fly ash with respect to those properties which relate to controllability as well as to possible environmental hazards of emissions. The focus is entirely on low-rank western and Gulf Province coals, whose properties are distinctly different from those of eastern coals. Typically the western coals have high moisture, low sulfur, and large variations in the distribution of inorganic constituents. Beulah, a North Dakota lignite, was used in the combustion tests of primary interest to this paper. This lignite is extraordinarily variable in its inorganic constituents. For example, the sodium content can have a tenfold variation within a few hundred meters in a single seam, as reported by Cooley and Ellman (1981). The specific Beulah lignite used was selected for its high sodium content. The fly ash was generated for this research using the UNDERC Particulate Test Combuster (PTC), illustrated in the schematic in Figure 1. This unit is an axially-fired pulverized coal combustor with a nominal consumption of 34 kg of coal/h. The unit is designed to generate ash characteristic of that produced with similar fuel in a full-sized utility boiler. Axial firing maximizes the fly ash/(bottom ash + slag) ratio. The fly ash samples for analysis were collected a t the inlet to the baghouse with
two devices, a five-stage multicyclone, and a source assessment sampling system (SASS). The flues are equipped with heat exchangers permitting delivery of flue gas to the chosen control device at temperatures from -100 to 390 "C. During these tests, the PTC was equipped with an experimental baghouse because of on-going fabric filter research at UNDERC. The PTC instrumentation permits measurement of flue gas conditions such as temperature and concentrations of SO2, NO,, 02, and C02. The coal and fly ash were characterized by several analytical techniques. The inorganics of the coal were examined by an extraction technique which selectively removes the inorganics depending upon their association in the coal as reported by Schobert et al. (1981). Lowtemperature ashing with subsequent X-ray diffraction was used to identify the mineral phases of the coal. The fly ash was examined and analyzed by X-ray fluorescence, X-ray diffraction, scanning electron microscope/microprobe, and electron spectroscopy for chemical analysis (ESCA). Detailed descriptions of the equipment and techniques are given in the next section. E x p e r i m e n t a l Section Equipment and Techniques. A Kevex 0700 System
was used for quantitative elemental analyses of the major oxides. The Kevex system consists of an energy-dispersive X-ray source and a solid state Si(Li) detector capable of
0196-4321/84/1223-0252$01.50/00 1984 American Chemical Society
Ind. Eng. Chem. Prod.
Res. Dev.. Vol. 23,No. 2.
1984
253
Table I. Typical Coal and Coal Ash Analysis of Beulah, ND Lignite ultimate anal., as fired coal ash anal., 5% carbon 46.29% SiO, 22.3 hydrogen
nitrogen sulfur ash
moisture heating value 11 460 J/g (1512Btu/lb)
Figure 1. The 34 kg/h mal combustion unit employed for ash generation. complete automation. The system can use direct, direct-filtered, and secondary excitation. The 0700 subsystem is integrated with a Kevex 7000 Micro-X analyzer that provides data acquisition, spectrum processing, and automation of the analysis scheme. Samples were prepared by making pellets containing 90% samples or ash (750 "C in a standard muffle furnace) and 10% avicel. Each sample was ground for 10 min in a Spex shatterbox and pressed into a 3.8-cm pellet under 206 mPa of pressure. The same procedure was used to make pellets from the standard materials for instrument calibration. Automated powder X-ray diffraction was the major technique utilized to identify the crystalliie phases present in the samples. A Philips APD3600-02 version X-ray diffractometer with a standard goniometer and generator, graphics display devices, a hard copy unit, and a Data General Nova 4 computer was used. Software programs produced lists of peaks, d-spacings, and intensities for each sample. Search and match routines were available to reduce the data as was a program to compare the spectrum with bar graphs of reference compounds available through JCPDS. Samples were pulverized t o -325 mesh and top loaded into a sample holder. Scans were run from 5' to 70' 28 a t 45 kV and 40 MA. Results are reported qualitatively as phases detected and do not indicate quantities present. Scanning electron microscopy characterization was performed by means of a JEOL JSM 35 scanning electron microscope operated at 15 kV with a beam current of 350 PA. The X-rays were detected with a Kevex lithiumdrifted silicon detector and the elemental analysis was performed by means of a Tracor Northern NS-880 X-ray analyzer. Analysis time was 200 s, and the ten elements determined were Na, Mg, AI, Si, P, Si, K, Ca, Ti, and Fe. Particles for SEM characterization were mounted on double-stick tape on a m h o n plug and coated with carbon to provide a conductive surface. The particles analyzed were selected at random by using a grid overlay of points generated by random numbers on the SEM photomicrograph. One hundred particles from each stage were analyzed, determining 10 elements per particle, and sized. All data were entered into a computing system for correlation analysis. Electron Spectroscopy for Chemical Analysis (ESCA) was used t o study the surface properties of the samples. ESCA was chosen because it causes minimal damage to the surface of the samples in comparison to other techniques such as Auger or SIMS (secondary ion mass spec-
5.2996 0.62% 0.99% 8.2% 27.2%
A1,0, Fe,O,
TiO,
CaO MgO Na,O 50,
11.1 9.8 0.9 15.1 5.3 9.9 23.0
troscopy) analysis. Although ESCA causes the least amount of surface damage, it does have the disadvantage of giving only an average analysis of a number of particles because of the large area covered hy the X-ray flux. Samples for analysis were placed on indium foil; the foil was folded over and then pressed between two glass slides. The folded piece was then opened, cut in half, and mounted for analysis. The signal collection time was 30 min for each sample. Coal Characteristics. The inorganics associated with the Beulah lignite consist of a complex mixture of cations hound to organic acid groups, possible chelates or coordination complexes, minerals formed during coalification, and minerals accumulated during deposition. Traditional and nontraditional methods were used to examine the coal characteristics. The traditional analysis of the lignite is summarized in Table I, which lists the ultimate analysis, moisture, heating value (by ASTM (1950) techniques), and major element ash analysis performed by X-ray fluorescence (Benson et al., 1979). One of the two "nontraditional methods" used is lowtemperature ashing in an oxygen plasma at 150 "C (Miller et al., 1979), followed by X-ray diffraction analysis to determine the crystalline phases in the ash. The phases identified include SiOp (quartz), FeSz (pyrite), kaolinite, and CaC03 (calcite). The second method involves a procedure which selectively removes inorganics as reported by Schohert et al. (1981). The ion-exchangeable cations and soluble salts are removed with 1M ammonium acetate. The coal is then extracted with 1M hydrochloric acid to remove chelated species and acid decomposable minerals such as carbonates and oxides. The inorganics remaining in the coal can include FeS2, organically hound sulfur, quartz, and clay minerals. The elements found to he extracted with ammonium acetate are Na (100% of the total Na content), Mg (88%), Ca (41%), and K (57%). The elements removed by HC1 are Mg (20%), A1 (61%), Ca (50%), and Fe (30%). The elements which remain are Si (loo%), Fe (70%), A1 (40%), and S (100%). Ash Sampling. Fly ash is sampled isokinetically in three ways in the P T C (1)by using modified US.EPA (method 5) dust loading filters; (2) by collection in a Southern Research Institute 5-stage multicyclone which simultaneously samples isokinetically and size fractionates the ash as reported by Smith et al. (1979) and Lipscomh (1981); and (3) by collection in a three-stage cyclone module of an Acurex Source Assessment sampling system or "SASS-Train" as reported by Blake (1978) and Smith (1978). Nominal aerodynamic size ranges for the multicyclone and the SASS-Train, at the stated flow rates, appear in Table 11. Nominal size ranges, expressed as aerodynamic diameters in microns, are those corresponding to actual flow rates employed. The figures represent the mass median diameters (Dso)or "cut points" of the size distribution entering or leaving the stage. University of Washington Mark I11 cascade impactors were used t o more precisely determine particle size dis-
254
Ind. Eng. Cham. Rod. Res. h.. Vol. 23. NO. 2. 1984
Table 11. Aerodynamic Size Rang- (Upper and Lower D,, Cut Points) and Sample Flow Rates of Underc's Ash Sampling Equipment instrument flow rate employed, std Llmin stage no. nominal size range (upper and lower D,,), p
30
Southem Research Institute Multicyclone 1 >10.3
2 10.3-5.6
9.36 3 5.6-4.15
4 4.15-1.9
5 1.9-1.55
AcurexIAerotherm SASS train 1
>10
113.2 2 10-3
3 3-1
m
-
25-
so3 Si02
20-
Na20
15-
A'2°3
10-
CaO
Fezof
0-
6-
-
1
2
3
Stage
Figure 2. X-ray fluoreneence analysis of each stege from the SASS train sampler.
tribution as reported by Smith et al. (1978). The bulk of the characterization work was done on the multicyclone and SASS train samples because the masses of particles collected by the impactor are very small and inadquate for analysis.
Results and Discussion Fly ash samples were collected by the SASS train to provide larger masses but only three size c u b for X-ray fluorescence (XRF) and X-ray diffraction (XRD) analysis. The elemental trends determined by XRF analysis show increasing concentration of Na,O and SO3 and decreasing SOz, CaO, Fep03,and MgO with decreasing particle size. Very little change was noted for A1203 or the minor elementa (TiOz, K20,and P,O& These relationships are illustrated in Figure 2. X-ray diffraction was used to identify the crystalline species in each size fraction. The trends, shown in Figure 3, reveal the following relationships: (1) Na2S0, and NaZCa(SO,), are present in the smaller size fractions (stages 2 and 3). CaSO, is present in all stages. (2) SiOz and CaO are noted in larger size fractions (stage 1) and not in smaller size fractions (stage 3). Fe,O,-MgFe,O,-MgO phases decrease from stage 1to stage 3. (3) Poeaible A1$i05 and K$O, phases are noted in stage 3 which is the smallest size fraction. These two bulk particle characterization methods show the trends of the elemental distribution vs. particle size and how the elements are combined. The most interesting trend is the increasing amount of Na,O and SO3 with decreasing particle sizes.
0.814 U- F.,0..F.#AoO.
01. N.*C.180.>, L SC.0
Figure 3. X-ray diffraction spectrographs of the SASS samples from run 220.
Figure 4. Secondary electron image of particles from stage 1 mul-
ticyclone.
Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 2, 1984 255
Table 111. Size, Sodium, and Sulfur Relationships for Each Stage of the Multicyclone
% AI203 401
weight average stage
particle size, pm
Na,O, %
SO,, %
1 2 3 4 5
5.48 1.99 1.29 1.20 1.38
8.60 10.78 13.20 13.98 14.99
10.98 12.53 17.02 19.43 23.05
Na,O/
30
S03a
~.
0.79 0.86 0.78 0.72 0.65
.
. -
10
Na,O/SO, in pure Na,SO, = 0.78.
The scanning electron microscope (SEM) equipped with an energy dispersive X-ray analyzer was used to image and analyze particles down to 1pm in size. A typical SEM photomicrograph is shown in Figure 4. SEM analyses were done on multicyclone samples, stages 1 through 5. The average particle sizes determined from the SEM photomicrographs for each stage are listed in Table 111. Table I11 also lists the Na20 and SO3 concentration averages for all five stages along with the ratio of NazO to SO3. The concentrations of Na20 and SO3 increase as particle size decreases. The ratio of Na20 to SO3 is very close to that of pure Na2S04for stages 1, 3, and 4. The particles collected in stage 5 are not consistent with the other stages in particle size or in the ratio of Na20 to SOB. The compositional relationships can be represented by plots of composition vs. composition for selected constituent pairs. These plots can be used to ascertain the origin of certain species from the original mineral matter of the coal and the interactions that have occurred during combustion. For example, Figure 5 illustrates the compositional plots for several combinations. Figure 5a represents the plot of A 1 2 0 3 vs. Si02which reveals a cluster of points around a slope of approximately 0.8. Minerals identified in Beulah lignite as reported by Schobert et al. (1981) include biotite, kaolinite, muscovite, and illite, where the Al/Si ratio is in the range of 0.67-1.0. From this it can be suggested that the aluminosilicate framework remains intact after combustion. The common clay minerals have a range in composition of 2545% A1203 and 35-50% Si02 Since most of the particles in Figure 5a contain less than these percentages, the aluminosilicates either became combined with additional elements or became coated. In the plots of Si02 vs. SO3 and A1203vs. SO3, Figures 5b and 5c, there is a crude negatke correlation between S O 2 and A1203 with SO3. Despite considerable scatter, many of the points lie in the region along lines connecting clay minerals, at 0% SO3, with Na2S04or glauberite at 0% Si02 This is consistent with clay-like particles becoming coated with Na2S04. The surfaces of the particles were characterized by a sputter-etching technique used to remove surface layers followed by subsequent ESCA analysis. The ESCA analysis technique analyzes to a depth of 20 to 30 A. A typical ESCA scan is shown in Figure 6 for stage 5 multicyclone particles. The major constituents found on the surface include Ba, 0, Ca, Na, C, S, and Si. The binding energies of the elements detected by the detailed ESCA scans show that Na, Ca, and Ba are present as sulfates as cited in the “Handbook of X-ray Photoelectron Spectroscopy” (1979). The Na, Al, and Si may be tied up as silicates, but detailed binding energies for such compounds are unavailable in the literature. It was found that carbon and sulfur were concentrated on the surface of all five stages of the multicyclone samples. The high concentration of carbon can largely be attributed to contamination of the surface of the particles. Sodium was noted
i
a
__
0
i,
.
IO
20
30
40
50
60
7b
80
€30
%Si02
-
% AI203 40
-
Alp03 vs SO3
-h
Muscovlle
% Si02
Si02 vs SO3
I
0
8
10
20
30
40
50
% SO3
Figure 5. Compositional relationships for SEM data combined from all five stages of the multicyclone. 7
1
BINDINE EIIEREY, EV
Figure 6. ESCA scan of stage 5 of the multicyclone.
256
Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 2, 1984
,:PTC-229
-
minosilicates originating from clay minerals of the coal. The latter mechanism is supported by the results of the SEM data in Figures 5b and 5c and the ESCA results. These results support the validity of vaporization-condensation mechanisms of particulate formation. A summary of the possible mechanisms of particulate formation is described by Damie et al. (1982). The work described here is part of an ongoing project at UNDERC to develop a model for ash formation during the combustion of lowrank coals. Acknowledgment The authors wish to thank Robin Roaldson for his assistance in analyzing particles with the scanning electron microscope and Harold Schobert for his helpful comments in preparation of this paper.
SPUTTERED 15 HINS.
STFIGE 5
3-
-282
7
4
-279
-271
-276
-27)
I~IIIIK c*1)Sv.
PTC-229 STQGE 5
,/
rv
-267
-264
-261
f
\ SPUTTERED FOR 38 HINIn 1
,,/
-282
-279
1
= 32228
-276
-271
-278
IIMIK rnrw, rv
-267
-264
-2b1
1
Figure 7. Na peak integration from ESCA scans after 15 and 30 min of sputtering.
on the surface of the larger particles, but in the smaller size fractions (x1.5 pm) the sodium concentration was more uniform throughout the depth analyzed by sputtering shown in Figure 7. This figure shows the integrated area of Na peak after 15 and 30 min of sputtering. Campbell et al. (1979), also showed that Na, S, and C were found on the surface of ash particles. In all cases, except for stage 5, Ca, Al, and Si were found to increase after sputtering, indicating that these elements may occur in the interior of the particle rather than on the surface. Conclusions The most significant trend noted by all techniques was the increase of sodium and sulfur with decreasing particle size. The relatively more volatile sodium salts may sublimate to form very small particles of pure salt or they may condense on the surfaces of other particles such as alu-
Registry No. Fez03, 1309-37-1; MgO, 1309-48-4; NazO, 1313-59-3; SO3,7446-11-9; AZO,, 1344-281; TiOz, 13463-67-7;KzO, 12136-45-7; Pz05, 1314-56-3; NazS04, 7757-82-6; NazCa(S04)z, 15892-81-6; CaS04, 7778-18-9; S O z , 7631-86-9; CaO, 1305-78-8.
Literature Cited ASTM, Part 26, Gaseous Fuels; Coal and Coke; Atmospheric Analysis, American Society for Testhg and Materials, Phlladelphia, 1979. Blake, D. E. Feb 1978, Research Triangle Park, NC, EPA-600/7-78-018. Benson, S. A.; Severson, A. L.; Beckerlng, W. Am. Lab. 1980, 12, 35. Campbell, J. A.; Smith, R. D.; Davis, L. E.; Smith, K. L. Sci. Total Environ. 1979, 12, 75-85. Cooiey, S. A.; Eliman, R. C. Grand Forks Energy Technology Center, Grand Forks, ND, July 1981, U.S. DOE/GFETC/RO-81/2. Damle, A. S.; Ensor, D. S.; Ranade, M. B. Aerosol Sci. Technol. 1982, 1, 119-1 33. Lipscomb, W. 0. U.S. Department of Energy, Research Triangle Park, NC, Feb 1981. DOE/FC/O1193-Tl. Miller, R. M.; Yarzab, R. F.; Qiven, P. H. Fuel 1979, 58, 4. "Handbook of X-ray Photoelectron Spectroscopy"; Perkin-Elmer Corp., Physical Electronlc Division, Eden Prairie, MN, 1979. Schobert, H. H.; Benson, S. A.; Jones, M. L.; Karner, F. R. Proceedings of the International Conference on Coal Science, Dussetdorf, Sept 1981; pp 10-15. Smith, W. B.; Cavanaugh, P. R.; Wilson, R. R., Jr. US. Environmental Protection Agency, Research Triangle Park, NC, March 1978, EPA-600/7-78043. Smith, W. B.; Wilson, R. R.,Jr.; Harris, R. B. Environ. Sci. Technol. 1979, 13, 1387.
Received for review July 20, 1983 Accepted November 17, 1983
This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, mark, manufacturer, or otherwise, does not necessarily constitute or imply ita endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
