close

Вход

Забыли?

вход по аккаунту

?

Integrated Dust Analysis by Physical Methods.

код для вставкиСкачать
corresponding to steps 7 and 8). The alkylmetal compound
can either add CO and thus contribute to chain propagation
(step 14) or, by p-H transfer, give an a-olefin (ethylene at
this stage) and a metal hydride which continues the kinetic
chain (step 15).
Presumably, the metal hydride is also responsible for the
secondary reactions such as hydrogenation and incorporation
of the x-olefins. Coordination and insertion of an a-olefin
into a M-H bond can occur in either of two ways:
R-CH=CH2
+ H-Fe
?
R-CH-CH3
he
R-CH2-CH2-Fe
Step 16 and further growth will lead to monomethylbranched compounds, whereas step 17 gives a linear prolongation of the previously olefinic chain. This explains why methyl
branching is the only significant branching experimentally
observed.
At present, this reaction mechanism cannot be more than
a working hypothesis which explains, at least qualitatively,
most of the experimental findings, without violating known
reaction patterns of transition metal centers.
Several questions remain open; e.g. why is alcohol formation
favored by high pressure and high H2/C0 ratio? Or, why
has it not been possible so far to carry out a Fischer-Tropsch
synthesis in solution, with a soluble transition metal complex?
The answer to the latter question may be related to a general
phenomenon in heterogeneous catalysis : solid reduced transition metal catalysts have, on their surfaces, low valent, coordinatively unsaturated metal species which easily offer the sites
required for catalysis, whereas in solution coordination of
solvent molecules or cluster formation tend to block such
sites.
Received: June 23, 1975 [A 99 IE]
German version: Angew. Chem. 8X. 144 (1976)
F . Fischer: Herstellung flussiger Kraftstoffe a u s Kohle. BorntrCger,
Berlin 1924.
P. Subatler and J . B. Sendrrrns, C. R. Acad. Sci. 134, 514. 689 (1902).
F. Fischrr and H . 7rop,sch, DRP 41 1216 (1922).
a ) H. Pichlrr, Advan. in Catal. 4, 271 (1952): b) H . H. Stordi, N .
Golurnhic, and R. B. Anderron, The Fischer-Tropsch and Related Syntheses. Wiley. New York 1951 c c) H . Kiilbel in K . Winniicker and
L. Kiichlw: Chcmische Technologie. Hanser. Munchen 1959, Val. 3,
p. 492: d ) H . Pichlw and G . Kriiqcr: Herstellung flbssiger Kraf~stoffeilus
Kohle. Gersbach & Sohn, Munchen 1973, and references therein.
F . Fischrr and H . 7ropsch. DRP 484337 (1925).
F . Fischer and H . Pirhler. DRP 731 295 (1936).
I;. Fitcker and H . Pichler, DRP 888240 (1937); DRP 888841 (1937).
H . Pichler and H . B u f l I d ~ ,Brennst. Chem. 2 I . 257, 273, 285 (1940).
a ) W Wenzel.Angew. Chem. Ausg. B 21,225 (1948): b) M . D.Sclifesinger.
H . E. Bensoii, E M u r p h y , and H . H . Stvrch, Ind. Eng. Chem. 46,
1322 (1954); c) DBP 939385 (1951); DBP 974811 (1961) (Ruhrchemie):
d ) Soviet Pat. 386899 (1973); 386900 (1973): e) Yu. Kagiin. A . N . Brrshkirot., L. A . M o r o x x Yu. 5. K r y u k o r . and N. 4. Oriorii, Nefl Khim.
6 , 262 (1966).
H . 7ropsch and H . Kocli, Brennst. Chem. 10. 337 (1929).
H . Koch and H . Hiiherarh. Brennst. Chem. 22, 135, 145 (1941).
R. A . Fririfrf and R. B. Afzrlerson. 1. Am. Chem. SOC.72, 1212 (1950).
H . Pichler, H . Schulz, and F. Hojabri, Brennst. Chem. 45. 215 (1964).
H . Pirhier, H . Schul:, and M . E l s r i i ~ r ,Brennst. Chem. 48, 78 (1967).
H . Pichlcr, H . Schulz, and D. Kiihnr. Brennst. Chem. 4Y. 344 (196x1.
H . Schulr, B. R. Ri m and M . Elsttier, Erdol Kohle 23, 651 (19701.
and literature therein.
G . Henrici-Olic6 and S. O l i l L Adv. Polymer Sci. I S , 1 (1974).
See, r . y.: G . Henriri-OlirCand S. Olird: Polymerisation. Verlag Chemie,
Weinheim 1969.
R. S. Nyholrn in W M . H . Surhrrr, G . C . A . Schuitr, and P. Zwrriving:
Proc. 3rd Internat. Congr Catal., Vol. I. North Holland, Amsterdam
1965, p. 25.
See. e.y.. J. Chart and J. Holperri in F. Basolu and R. L. Burwell.
Catalysis, Progress in Research. Plenum Press. London 1973, Chapter
3.
J. P. Collmrrn and S R. Winter, J. Am. Chem. Soc. 95, 4098 (1973).
I
I
Integrated Dust Analysis by Physical Methods
New analytical
methods (5)
By Hanns M a l i s s a [ * l
1
I
Environmental analyses show that the air which we breathe, and which is so essential
to life, is in general a mixture of gaseous, liquid, and solid components. The solid airborne
particles, whose concentration, homogeneity, chemical composition, size, and shape can vary
over wide ranges, and whose origin may be “natural” or “artificial” are referred to as “dust”.
Dust particles can act, infer alia, as condensation nuclei, catalysts, and directly as hazardous
materials. Unfortunately, we still know far too little about dust. Dust analysis is extremely
difficult and challenging, even for modern analytical chemistry; it is still far from being fully
automated. The simultaneous determination of as many “dust parameters” as possible, and
particularly the synoptic consideration of all available data against a background of physicochemical and technological knowledge on the development, transformation, and effects of dust,
are summarized as “integrated dust analysis”.
1. Introduction
With the concept “integrated dust analysis” an attempt
is made to express the fact that the nature and effects of
[‘I
Prof. Dr H. Malissa
lnstitut fiir Analytische Chemie und Mikrochemie der Technischen Univcrsitiit
Getreidemarkt 9. A-I060 Wien (Austria)
dust can only be understood if the dust sample is examined
as completely as possible in situ, and that one analytical method
is not adequate for this purpose.
The present review outlines the significance and performance
of integrated dust analysis and then goes on to consider a
few general
problems of dust analysis.
Most difficulties in the assessment of analytical results stem
not only from the sample or the methods employed, but
rather from the problem formulated.
The primary question in the case of dust analysis is: “What
is dust and where do the limits of its analysis lie?”
According to Bibbero and Young”] dust is defined as s ~ l i d
particles of “any size”, i.e. airborne particles ranging from
pm (small molecules, large ions) to I O3 pm (particles
already visible to the naked eye). It is in this range of particle
size, which stretches over some seven orders of magnitude
and in which inhomogeneities of chemical composition and
shape arise due to adhesion and agglomeration, where the
main problems of dust analysis lie.
Homogeneity I-log V I
.$
FEMI
Physical Methods
LM
1.
l a r g e Ions
0
Suspended Mailer
f
Chemical Methods
I
F
1
G
Oeposlted Matter
Fig. 1. Particle size in environmcntal analysis. L M =light microscope.
RFA = X-ray fluorescence analysis. EMP=electron microprobe, SEM =scanningelectron microscopy, ELM1 =electron microscopy, FEMl = field electron
microscopy. A: atoms. B: molecules: C : Aitken particles: D: large particles:
E : giant particles; F: fine solids; G : coarse solids.
Figure 1 illustrates the interrelations between particle size,
physical behavior, and homogeneity. The behavior of the particles (such as motion, adsorption and absorption, or chemical
reactions) is determined by the physical and chemical laws
that are valid for finely divided material. As a measurement
of homogeneity”] we use the logarithm of the reciprocal of
the analytical volume (log l / V ) ,which-within a certain standard deviation-yields the same result in measurements on
samples taken from various sites as in repeated measurements
on one and the same sample. The question of homogeneity
is directly related to the degree of resolution of the method
of examination employed. For X-ray fluorescence analysis
(RFA), particles under 100pm are in all events homogeneous,
whereas if electron probe microanalysis or scanning electron
microscopy (SEM) is employed they could be heterogeneous.
fraction Degosited
i
Oh
03
02
0 01 10-8
Fig 2. Deposition of dust according to particle size. 1 : trachcobronchial;
2 : pulmonary: 3: nasopharyngeal (cf. also Fig. 1).
142
This means that the RFA method suffices up to homogeneity
class 5, the SEM method up to 16. The highest homogeneity
would be achieved if every atom in the sample were the
same as the rest (homogeneity class 24).
It is already well known that the harmfulness of dust does
not depend on the chemical composition alone, but also on
the particle size and shape. Recent work by Nutusch and
W ~ [ l a c eshows
[ ~ ~ that small particles are generally more dangerous than large ones. The reason for this lies in the processes
of breathing and deposition (cf. Fig. 2). Usually the dangerous
components such as lead, cadmium, selenium etc. are present
as small particles. It was found for example that prolonged
exposure to coarse grained MnOz was not injurous to health,
whereas a finer dust of the same substance (80% of the particle
diameters < 2 pm) caused pneumonitis. The influence of particle shape has so far not been studied in great detail, but
one can assume that angular particles are more dangerous
than rounded ones. This opens up yet another field for cooperative research (analysts, doctors, etc.). Furthermore, it is also
known that dust is far from being a homogeneous productneither from the point of view of particle size or shape, nor
from the standpoint of composition. This leads to considerable
difficulties in analysis, but can make the investigation of dust
a very interesting task. In the analytical sense dust is an
agglomerate, the characterization of which demands special
consideration of the problems of homogeneity; for its clarification the methods of stereometric analysis are resorted to.
Detailed characterization of dust requires the following analytical steps (where possible, carried out simultaneously):
1. Average analysis of all or individual elements;
2. Identification and quantitative analysis of individual particles of dust;
3. Determination of particle mass (concentration), particle
size, and particle-size distribution;
4. Determination of morphology.
Together with stereometric analysisrz1they form the basis
of integrated dust analysis. Today the analyst has a wealth
of equipment at his disposal for carrying out these determinations, so, in principle, achievement of the aims of such an analysis is not impossible.
2. Sampling
The arsenal of apparatus normally employed for the sampling of airborne dust, which has a particle size of approx. 0.01
to 100 pm, includes filters, dust collectors, electrostatic separators, cyclones, and impactors.
The widely adopted technique of dust sampling with the
aid of filters is highly efficient. By using millipore filters with
a defined pore-size, dust having a particle size down to 0.1 pm
can be trapped with almost 100% efficiency.
Dust collectors are cheap and easy to handle but have
the disadvantage that the dust must be separated from rainwater before the analysis. As a side-effect any SO2 dissolved
in the water can react with the dust and thus lead to a change
in composition. A further disadvantage is the time required
for sampling.
Electrostatic separators are best suited for the collection
of smaller dust particles-down to a particle size of 0.01 p.
Cyclones are optimal for the separation of larger particles
A n g r n . Chon. lnr. Ed. Enyl. i Vol. 15 (1976) No. 3
(down to 5 pm). Impactors offer the advantage of simultaneous
collection and fractionation of particles (working range down
to 0.01 pm).
Next to deciding the correct method of sampling, greatest
importance attaches to the choice of sampling site. This
depends in the first place on the purpose of the investigation :
immission or emission measurements. In each case meteorological factors such as wind direction, wind strength, temperature, moisture etc. must be taken into consideration.
A further important parameter is the duration of sampling.
Each sample collected over a definite period of time yields
average data for that interval of time (cf. Fig. 3), e.g. the
daily average concentration of lead in the atmosphere, if collected over 24h. If, on the other hand, the sampling time
is shortened to 12h, one obtains not only the daily average
value but also the main deviations occurring half-daily. Sampling at three-hourly intervals leads to even more detailed information--for example that the lead concentration is especially
high at peak traffic times.
;
0
0
3
6
9
12
15
18
21
2L
hours
,&1!Zq
Fig. 3. Dependence of analytical Information o n duration of sampling.
Example: lead content of air a t a street crossing in Vienna (after H a r t l ,
Miillnrr. Rcsch, and Wugiier).
Generally speaking, the information content of a sample
is the higher, the shorter is the duration of sampling and
the more complete is the analysis. The shortest possible duration of sampling is determined by the time required to separate
the minimum amount of dust needed for the analysis.
Since the average concentration of dust in towns is about
100 to 50pg/m3 and since dust can be separated from ca.
2&100m3 of air per hour, the amounts of sample available
for analysis lie in the milligram range. This leads to the first
unavoidable claim on the technique of analysis to be employed:
only micro methods are adequately suitable. If one now considers that a series of physiologically important elements are
present in nanogram or microgram quantities in these few
milligrams of dust sample we arrive at the second important
feature required of the analytical technique: The methods
must be highly sensitive, as defined in terms of trace analysis.
Summarizing, this means that micro-trace methods are
required for dust analysis. Also characteristic of such samples
is that dusts are complex multi-element systems and the element to be determined can be present over a wide range
of concentration (of % to ppb). Hence the overall analytical
problem is "simply" that the analyst must have at his disposal
non-destructive multi-element methods for the analysis of microsamples which can cope with a concentration range ofabout
eight orders of magnitude and a particle-size range of about
seven orders of magnitude.
If we check the arsenal of modern analytical chemistry
with regard to the attributes required, we find that physical
micromethods are best suited for dust analysis, since only
these comply with the requirements of the sample and the
previously formulated problem. The problem cannot however
be solved on the basis of a single method.
3. Determination of Average Composition
In integrated dust analysis determination of the average
composition is not only important, but essential-and especially for keeping a check on single particle analysis. The
most important methods at the moment for the average analysis of a dust are X-ray fluorescence analysis (RFA) and atomic
absorption spectroscopy (AAS).
The RFA method can be used for the qualitative and quantitative analysis of all elements having atomic numbers greater
than 11 (Na)[4-61. The practical execution of the analysis
using the appropriate standards is simple- usually the filter
loaded with dust is analyzed directly (without further treatment) (Fig. 4). Since the depth of signal generation in fluorescence excitation is about 50pm, filter material is also excited
along with the thin layer of dust. The characteristic intensity
of the element being measured is directly proportional to
the excited amount of element A in dust and filter. The correct
intensity of an element in the dust is obtained by subtraction
of the intensity of A in a blank filter. Since the surface area
analyzed is usually cu. 3cmZ, then with a signal generation
depth of ca. 50pm, a volume of cu. 0.02cm3 will be excited.
If the density is 2 g/cm3, then approximately 40 mg of sample
of a compact material is analyzed; however, in the case of
analysis of dust on filter paper the value is considerably lower.
--
IA o( MA T A
Sample
Is
Filter
i,.lntensity
of A l c p s l
H,:lmount
of 1 I p g i c m ' l
T ~ Ionization
=
cross-section
w,= fluorestence yield
-A
I
Standard Is,
C,=
MAW
J.iS,-'.<FI
I
-
(IIsT,-(.lFI
c,
-
m~
~ o n c .of A in air ~ p g / m ' i
F = Filter area l c m 2 1
v = Rate o f airflow I m ' / h l
t = Ouration o f sampling Ihl
A
" I
Fig. 4. Quantitative analysis of filter dust with the R F A method.
The quantitative analysis is based on comparison measurements against one or more standards; in most cases these
are filters loaded with a definite amount of the element to
be determinedL7! The standards are often prepared by spraying
a compound on the surface of the filter paper. Gilfrich et
a1.[41have pointed out the errors that can arise if soluble salts
are used as standards instead of insoluble compounds (e.g.,
for Mg and Al, correction factors of 31 and 20"4 respectively,
are necessary).
It is a matter of conjecture whether this type of standardization really is the best, since the element to be analyzed is
often present in different compounds. Thus, e.g., Pb can be
bound to oxygen, halide or sulfate. Depending on the chemical
shift the band maxima of Pb can lie at different wavelengths.
This effect, of course, plays only a subordinate role in RFA,
since the resolution in this method is low. However. the effect
is relevant in the case of electron probe microanalysis (EMP).
143
From the ratio of count rates (cf. Fig. 4)
one obtains the amount of element deposited. The result is
normally given in pg/cmZand converted into the usual concentrations used for hazardous solid materials. The multichannel
analyzers presently being used in many industrial establishments have proved very useful for the analysis of dusts; they
enable the simultaneous determination of 20 elements in
approximately 2 min (in automated operation).
It follows from the direct proportionality between the
intensity of a characteristic X-ray line and the fluorescence
yield, that elements of the 2nd period cannot be determined
with sufficient sensitivity by the RFA method because WA
drops below 1 % at such low atomic numbers.
Excitation with electrons in the macroprobe is useful for
the quantitative analysis of these elements, especially C and
0, since the decrease in fluorescence is compensated by a
corresponding increase in ionization cross-section on account
of the different mechanism of signal generation.
As example, we might mention here the determination of
the C content of street dust with the macroprobe@'. In the
analysis of samples of street dust collected during a survey
in Vienna (Fig. 5), which was concerned with the local and
yearly change of concentration of nine elements, carbon contents of 2.7 to 9.2% were routinely determined. The dust
sample collected in April, with a value of 7.7% (average value
of all 26 collection points), shows the highest proportion of
carbon. This can be ascribed to domestic heating systems
which give rise to considerable emission of soot. In summer
the C-values are around 25 % lower.
j 5 % C
size of the dust. At particle diameters of more than 5pm,
intensity losses of 50 to 90% can occur. The reason for this is
that the absorption of the characteristic radiation in larger
particles is stronger than in smaller ones.
Since the particle size distribution is usually unknown, corrections for this effect can rarely be made. Thus, the accuracy
of dust analyses by the RFA method lies in the region
of ca. &20%. Though this is perfectly satisfactory in the
majority of analyses, in individual cases a higher accuracy-if
possible still with the same sensitivity-is desired. Atomic
absorption spectrometry (AAS) would be suitable for such
anaIysesl'.' '1.
At present, A A S provides good results only for liquid
samples. This means that the dust sample must be dissolvedeither partially (e.g. with HNO,/HCl) or completely (e.g.
with H F / H N 0 3 under pressure in a teflon bomb or with
LiBO2)["]. About 0.5 to 1 mg of sample is required for the
determination of an element. The individual elements can
so far be analyzed only sequentially, but with great accuracy.
Each chemical workup of the dust has the disadvantage
that, on the one hand contamination with impurities can
occur, and, on the other, that the bonding of the elements
can no longer be determined ;moreover, dilution of the system
takes place. Because of this dilution-100mg of dust becomes
100ml of solution, i.e. a dilution of the order of 103-AAS,
though in itself a sensitive method, does not usually show a
higher sensitivity than RFA on evaporation of the sample in a
flame.
Table 1 compares the sensitivities of the determination of
Cu, Zn, and Pb by various methods. The limits of detection
are based on the content of the element in the dust. It can
be seen that the detection limit of AAS, the method usually
employed, is poorer with flame evaporation than that of RFA,
but the precision is better.
Table 1. Sensitivity of the determination of Cu, Zn. and Pb in dust (calculated
according to Refs. [4,6,12]).
Method
Amount of
sample
RFA
AAS (flame)
AAS
(Graphite tube
cuvette)
Optical emission
spectroscopy
1 mg/cm2
1 mg/ml
1 mgiml
50
0.05
10
10
0.01
20 mg
10
10
Limit of detection [a]
CPPml
Cu
Zn
Pb
10
50
200
Precision and
reproducibility
["/:rel.]
920
0.1
:1-5
:
10-30
I
:1&100
[a] Based on concentration in the dust.
Mean values
April 7 . 7 W
rn
July. 5 . 4 W
Oct : 5.3%C
Jan .6.4%C
Fig. 5 . Dust cadastre of Vienna: Carbon content or street dust
[XI
A further disadvantage of RFA (but in this case seemingly
slight disadvantage in comparison to the advantages) is the
dependence of the signal intensity of an element on the particle
144
The sensitivity of AAS can be increased either by chemical
enrichment of the trace metals in the sample solution or by
use of the graphite tube cuvette. In the case of the elements
quoted an increase in sensitivity by a factor of about lo3
is possible with the graphite tube cuvette; however, the accuracy of the determination decreases to ca. +30% rel. (after
Coleman[' 2]). The most recent development of employing socalled plasma torches and incorporating gratings on a Rowland
circle is receiving a great deal of attention, since it will probably
enable the simultaneous flame-photometric determination of
several elements in very low concentrations.
In some cases (e.g. Pb) better limits of detection are achievable with optical emission spectroscopy, since this is also
a method in which the sample is excited directly without
Anynr. Chrm. lot Ed Engl / V d . 15 (3976) N O 3
dilution. However, the results can often be regarded as only
semiq uant itative.
Table 2 . Quantitative investigation of a fly ash by X-ray fluorescence analysis
(RFA) and atomic absorption spectroscopy (AAS) [4].
______________________-----
Element
~
RFA
Cfig,.’cm’I
AAS
[~!cm’l
-_-________________------
S
K
Ca
Fe
NI
cu
Zn
Cd
Pb
3.2
3.5
0.30
0.30
~
0.28
0.008
0.11
0.10
0.80
13
11
0.097
4.8
4.4
Table 2 shows the efficiency of RFA and of AAS in dust
analysis using the results obtained for a sample of fly ash.
The main elements, which are present in the ng or pg range,
can be determined with RFA and/or AAS. Generally speaking
the results can be said to be in good agreement. Large differences can be attributed to lack of sample homogeneity. In
this connection it must be pointed out that precision and
reproducibility in the analysis of dusts are also influenced,
to a not inconsiderable amount, by the statistical error-ie.
by the small number of dust particles to be analyzed. This
statistical error can be conveniently calculated with the aid
o f a formula described by W i l ~ c i n ”which,
~],
to a first approximation, states that the statistical error is 100,l,6‘~, i.e. inversely
proportional to the square root of the number of particles
analyzed. Thus, in the analysis of Cu, one has to reckon
with a statistical error of +30% rel. (RFA) for a Cu content
of lOppm, assuming a particle size of 5pm. Of particular
interest among the numerous new developments in AAS are
the variants with microwave-plasma torches and multichannel
grating spectrometers, since the amount of sample required
is small and an automatic, simultaneous determination of
up to 64 elements is possible.
Other methods used for carrying out average analyses of
highest sensitivity are spark-emission mass spectrometry and
neutron activation analysis. In comparison with the previously
described methods they have considerable disadvantagessuch as highly expensive apparatus, long analysis times, low
accuracy etc.-and consequently find little application.
Fig. 6. Electron microprobe analysls of airborne particles.
with the microprobe. The electron back-scatter picture (center)
not only gives information about the average atomic number,
but also about the shape of the single particle. The energy-dispersion X-ray spectrum (upper right, center) yields semi-quantitative information about the concentration of the elements
having atomic numbers greater than 11 (Na) that are present
in the area scanned. This result might also serve as guide
for any subsequent AAS investigation, in that suitable lamps
and preparative steps can be carefully selected. The distribution
of the elements can be plotted as the element-specific X-ray
signals; the distribution pictures make it possible to give
semi-quantitative information about the elemental composition of individual particles-thus, that the particle on the
upper left, for example, is an aluminum silicate. The next
task consists in quantitatively analyzing a single particle by
point counting. The energy-dispersive X-ray spectrometer is
well suited for the analysis of rough surfaces, since all elements
at the same angle are measured. Crystal spectrometers must
be employed for the determination of elements of the 2nd
period. In binary systems, quantitative analysis is achieved
4. Analysis of Single Dust Particles
The most suitable method for,the identification of dust
components is the elemental analysis of single dust particles.
Since airborne dust has a particle size range of ca. 100 to
0.01 pm the electron microprobe can be used for the analysis
of single particles. The volume of sample excited by the finefocused electron beam in point measurements is only a few
~m~ and thus makes a selecfive analysis of single particles
possible. Sample preparation is simple: the filter dusts
employed in RFA measurements can be analyzed directly
after they have been coated with a thin layer of electrically
conducting carbon by evaporation; loose dusts, such as are
obtained in other dust collecting methods are transferred to
a teflon platelet coated with adhesive carbon conductivity
paste. Figure 6 shows a typical example of a dust analysis
Angru,. Chem. fnr. Ed. Engl.
I Voi. 15
(1976) No. 3
Fig. 7. Scanning electron microscopy of a dust particle collected in the
vicinity of a district heating plant with energy dispersive X-ray analysis.
145
by evaluation of the counting rates and comparison with
the corresponding data for standards containing the elements
under consideration, since this quantity is largely independent
of surface roughness and particle size-i.e. it still applies
to particles that are smaller than the volume excited. Because
of the roughness of the particle surface the achievable accuracy
of the analysis is less than in polished samples, and lies in
favorable cases at about 5-20% rel. This, however, is satisfactory in most cases.
The application of this technique and wavelength dispersion
analysis can afford a wealth of information. Figure 7 shows
electron back-scatter pictures (left) of single dust particles. The
presence of sulfur was determined by recording an energy dispersive X-ray spectrum (center) and its distribution plotted
with the wavelength dispersive spectrometer (right). In order to
obtain more information about the sulfur compounds present,
A
II
UM
1200'
/
/
/'
1000~'
/
5/
c
a temperature differentiated relative conductometric S and C
analysis was carried out (cf. Fig. 8a)"4'. This method yields
specific and typical degassing curves for carbon and for sulfur
compounds.
Figure 8 shows that the degassing or combustion of such
components can be very different, depending on the dust.
In the dust of the first example, which contains 60.6% C
and 10.9%, S, loss of sulfur begins at 270°C. 20%) of the
total content is lost below 500°C, then a further 40% up to
600°C. In the interval to 1000°C the sample slowly loses
another 20%) of sulfur, and finally the rest in the interval
up to 1200°C. In contrast, the volatilization of carbon proceeds
very much differently. It is entirely lost within 200" at relatively
low temperature.
Investigation of a dust, which was collected at the same
time at another site by the Bergerhoff method, and which
also proved to contain similar spherules (Fig. 9) but in smaller
proportion gave total carbon and total sulfur contents of
8.4% and 1.8%, respectively. The degassing curves (Fig. 8b)
are markedly different. 80% of the sulfur in this second example
is volatile only above 1100°C. Loss of carbon is retarded
and shows a significant step at 600°C for a residue of cu.
20%. From volatility curves the presence of "organic" and
"inorganic" carbon as well as sulfur can be concluded. Thermal
decomposition of CaS04 takes place between 1100 and
1300°C. The decomposition of C a C 0 3 begins at 825"C, that
of M g C 0 3 at 325°C.
200
L
20
40
60
80
100%
Fig. 9. SEM picture and energy dispersive X-ray analysis of a dust particle
collected in a city.
5.0;-
Fig. X. Relative conductometric carbon and sulfur analysis with a temperature
program: a ) dust with 60.6%, C and IO.Y% S (cf. Fig. 7); b) dust with
8.4"/, C and I .8 2,S (cf. Fig. 9).
146
,>A
"00
I DO0
iIm-7
d
Fig. 10. IR spectra of. a) soot (cf. Fig. 7). and b) btreet dust (cf. Fig. 9).
Angjrn. Chmt. In!. Ed. Etigl.
Vol. 15 ( 1 9 7 6 ) No. 3
distribution in this case being isotope-specific. Figure 16 shows
as example the ion microprobe analysis ofan approx. 8 pm particle of lampback. The main advantage here is that elements
whose local concentration lies in the 0.001% range-as in
the case of B in this instance-as well as elements of very
low atomic number are still detectable. In the local analysis
of the dust particle illustrated in Figure 16, McHugh and
at
bl
Fig. 14. a ) SEM micrograph of a nickel ore, roasted with NaCI; b) electron
dilfraction diagram of a NaCl particle.
Thus, in investigations on fluorine-containing airborne particles, RadczewskirZz1was able to differentiate between the
naturally occurring cubic K2SiF6 and the hexagonal K2SiF6,
formed in industrial processes. This finding provides an important indication to the origin of the fluoride contamination.
Another technique employed in the identification of small
particles is the evaluation of energy-loss spectra of the transmitted electrons in a scanning transmission electron microscope.
Figure 15 shows a sample of MgO particles and their energyloss spectra. Discrete bands in the continuum correspond
to discrete energy-loss by the electrons in the ionization of
an atom. The MgL2,3 bands arise from the ionization of
the magnesium 2p,,, and
orbitals. A qualitative analysis
would appear to be possible by determination of the ionization
energies. The lateral resolution of this method is ca. 1OOa.
Al
SI
Point analysis:
jmmq
Cl
No,Ca
1 - 10
Mg,AI,K,Fe 0.1 6
0.001 - c l . O l O / O
Fig. 16. Ion microprobe analysis of an oil soot particle [23].
were able to determine semi-quantitatively 21 elements in the concentration range from 10 ppm to more than
10%.
The great importance of local trace analysis is that the
particle’s content of trace elements often depends in a characteristic way on the emitters. Thus, Anderson[241was able to
identify the emitters by determing the Be content of asbestos
fibers (ppm) in dust.
5. Morphology
I
I
t
I
I
I
I
I
0 153045607590eY
1-
100 A
Fig. 15. Energy-loss spectrum of MgO particles on an evaporated A1 layer.
A further limitation to the use of the microprobe arises
from the relative insensitivity of the EMP method: the detection limit lies at a concentration of ca. 0.1 %, with an excited
volume of a few pm3; this means that trace components of a
single dust particle cannot be detected, even though the microprobe permits very low absolute detection limits (ca.
g).
As an extension to electron microprobe analysis, a further
possibility is ion microprobe analysis, which, because of the
higher effective cross-section of signal generation and lack
of background interference, has a sensitivity that is higher
by a factor of 10’ to lo5.
As with the electron microprobe the distribution of the
elements to be found in the dust particles can also be recorded
with the ion microprobe using the scanning technique, the
148
In addition to the average analysis and single particle analysis of a dust, the determination of the morphology of the
particles is frequently also very informative.
The morphology of the dust particles-ie. size, shape, and
surface characteristics-is of great importance from the physiological standpoint. Thus, the possibility of penetration of
dust into the lungs depends on the particle size (cf. Section 1).
Of the physiological effects of shape, it is known that pointed,
rigid fibrous particles (e.g. asbestos fibers) are especially dangerous. The surface properties are decisive for the dust’s ability
to adsorb volatile noxious substances and thus evoke synergistic effects.
The investigation of dust morphology requires the use of
the light microscope (lower limit ca. 1 pm particle
and the scanning electron microscope (lower limit ca. 100A
particle diameters). SEM pictures (cf. Fig. 17) are by far the
most informative[’ l 1 because higher resolutions and much
sharper definitions are possible.
Angew. Chem. Int. Ed. Engi. / Vul. 15 ( 1 9 7 6 ) No. 3
a sample by point, linear, and area analysis with new equipment
(e.g. phase integrator) brought into line with new concepts
and methods of interpretation.
The main analytical points are:
- the chemical composition of the region of sample examined ;
- the qualitative and quantitative analysis of single particles;
- the total number of all selected particles;
- the particle size and particle-size distribution ;
- the morphology.
The synoptic consideration of the results and the chemical
and mathematical evaluation of the data will lead to the
desired information.
Received: July 14. 1975 [A 102 IE]
in revised form: December 1. 1975
German version: Angew. Chem. 88, 168 (1976)
[I] R . J. Bibbero and I . G . Young: Systems Approach to Air Pollution Control. Wiley-Interscience, New York 1974, p. 31.
[2] H . Mrrl~.~sa.
Z . Anal. Chem. 273, 449 (19751.
133 D. F. S. Nurusch and J. R. Wdlucr. Science lX6, 695 (1974).
[4] J . V: Gilfrich, P . G. Burkhalter, and L. S. Birks, Anal. Chem. 45, 2002
[5]
161
171
[8]
[9]
Fig. 17. SEM picture of an oil soot particle (magnification 2000 x )
(1973).
L. S. Birks and J . V: G i p i c h , Anal. Chem. 46, 360 R (1974).
J . R . Rhodes r t a l . . Environ Sci. Technol. 6. 922 (1972).
8. E. Arrz and H . Chessin, Adv. X-Ray Anal. 18, 225 (1974).
E. W a l d m ~ ~ iDissertation,
~i,
Technische Hochschule Wien 1973.
P. W West in A . C . S t e r n - Air Pollution: A Comprehensive Treatise,
Vol. 2, p. 147. Academic Press, New York 1968.
[lo] 5. Wrk Atom-Absorptions-Spektroskopie. Verlag Chemie, Weinheim
'
Special advantages are offered by an energy dispersion spectrometer fitted with a scanning transmission electron microscope, since such an apparatus enables the analysis of single
particles, the determination of crystal structures by recording
electron diffraction diagrams, and the morphological investigation to be carried out in one operation.
6. Conclusions
Integrated dust analysis, a novel analytical strategy in which
physical, chemical, geometrical, and mathematical operations
are combined, provides a wealth of very detailed data. It
is still in the early stages of development, but rapid progress
is expected in this sector over the next few years. The whole
field of dust analysis must be re-examined and the methods
of sampling, the possibilities of stereometric investigation of
Anyew. Chrm. Itit. Ed. Eizyl. 1 Vol. 15 ( I 9 7 6 1 No. 3
1972, p. 159.
[I I] 0.Wriss: Diplomarbeit. Technische Universitat Wien
[12] R. F. Colemarin, Anal. Chem. 46, 989 A (1974).
[I 31 D. Wilsoti, Analyst (London) X9, 53 (1964).
[I41 H . Malissa, E . Pell, and H . Puxbuum, Z. Anal. Chem., in press.
[ I 51 H . Pierrner and A . Sch@ers, Sonderheft der VGB Technische Vereinigung der GroBkraftwerksbetreiber : VGB-Conference on Feed Water
1972, VGB-Conference on Chemistry in the Power Station, Lectures,
Essen 1972.
1161 W. McCrone rt a / . in A. C. Sreni: Air Pollution: A Comprehensive
Treatise, Vol. 2, Academic Press, New York 1968.
[17] K . Sieqhnhn er ul.: ESCA; Atomic, Molecular and Sohd State Structure
Studied by Means of Electron Spectroscopy, Almqvist and Wiksell.
Stockholm 1965; C. Nordlir~g,Angew. Chem. X4, 144 (1972); Angew.
Chem. Int. Ed. Engl. I t , 83 (1972).
[IS] 7: Nucaknc, S. G. Chanq, and A. B Horkrr, Science 186, 259 (19741.
[I91 M . Gtxwerbouer, 2. Anal. Chem. 273. 401 (19751.
[20] A . L. Rickards, Anal. Chem. 44, 1872 (1972).
[ Z l ] M.Bayord in C . A. Ander.wn Microprobe Analysis. Wiley-Interscience,
New York 1973.
[22] D. E. Rotlc:e\r.aki, Siemens-Z. 48, Nr. 3 (1974).
[23] J . A . McHugh and J . F. Srrrriis, Anal. Chem. 44, 2187 (1972).
1241 C . A . Atidersrn, Lecture, 7. Koll. Metallkundl. Analyse, Wien 1974.
'
149
Документ
Категория
Без категории
Просмотров
1
Размер файла
993 Кб
Теги
physical, integrated, dust, method, analysis
1/--страниц
Пожаловаться на содержимое документа