General Reference Materials Relating to the
Measurement of Priority Pollutants
Effluent Guidelines Division
Office of Water and Hazardous Materials
U.S. Environmental Protection Agency
June 1977
-------
Contents
Identification of Organic Compounds in Unbleached Treated Kraft
Paper Mill Wastewaters
Polycyclic Aromatic Compounds In Nature
Asbestos In Your Future
Inductively Coupled Plasma - Optical Emission
Spectroscopy
Simultaneous Multielement Analysis of Liquid Samples
By Inductively Coupled Argon Plasma Atomic-Emission
Spectroscopy
The Jarrell - Ash Plasma AtomGamp
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
subject: General Reference Materials Relating to the
Measurement of Priority Pollutants
DATE:
FROM: William A. Telliard, Chi.
Energy and Mining Branch
JUN 8
to: Robert B. Schaffer, Director
Effluent Guidelines Division
The enclosed information is a set of general references for
utilization by the project officers to further their
understanding of the general application of various analytical
methodologies being employed 1n the BAT review. The articles in
this package apply to the various questions that have developed
over the evolution of the BAT review. One article describes the
use of gas-chromatography and mass spectrometry in evaluating and
measuring the discharge from a pulp mill. This particular
article is a good representation of the activities which many of
the project officers will be Involved with in the immediate
future. In addition, an article is supplied on the occurrence of
polycyclic aromatic compounds in nature. This article is
noteworthy simply because of the analytical methodology employed
and the distribution of the various compounds in the environment.
Another article is provided which addresses itself to the
definition and problems relating to asbestos. This article while
developed by the industry, does provide a rather accurate
chronology of the various problems relating to the definiton of
asbestos and asbestos fibers. This article will provide the
project officers with a basic understanding of the problems
confronting a number of the industrial categories presently under
review.
As you are aware, at present fn the BAT screening procedure we
are employing the use of an inductively coupled argon plasma unit
(ICAP) for the measurement of metals. A rather brief and
concise article is enclosed which describes the operation,
detectability and the range of operation of this particular
instrument. The article also compares the efficiency and
activities of this instrument with atomic absorption, which has
been the classic approach employed by the Agency with regard to
the measurement of metals.
EPA Pjm. 1320-4 (Rev. 6-72)
-------
2
This package also contains a copy of the equivalency study
carried out by the Central Regional Laboratory of the Chicago
Surveillance and Analysis Division. The study was performed in
order to acquire equivalency approval for use of the plasma unit
instead of atomic absorption for the measurement of metals.
I feel this is very important reading, giving the project
officers a understanding of the range and reliability of the
results of the analyzes they are receiving. These references are
supplied simply as a reference base of general information. As
other information becomes available or as additional studies are
provided, we will be pleased to supply them to you and upon your
approval, of course, to the various project officers. This
general information has been circulating throughout the Division
1n some form. This is an attempt to consolidate the various
articles which have been under discussion within the Division.
-------
Identification of Organic Cortpounds in Unbleached
Treated Kraft Paper Mill Wastewaters
-------
CURRENT RESEARCH
Identification of Organic Compounds in Unbleached Treated
Kraft Paper Mill Wastewaters
Lawrence H. Keith
U.S. Environmental Protection Agency, Environmental Research Laboratory, College Station Rd.. Athens, Ga. 30601
¦ Wastewaters from two kraft paper mills in Georgia were
'sampled at various points in the waste treatment systems. Gas
chromatography of the organic extracts and identification of
many of the specific chemical components by gas chroma-
tography-mass spectrometry provided a "chemical profile"
of the effluents. The mills, in different geographical locations,
have very similar raw wastewater compositions out different
wastewater treatments. In spite of these differences, the
treated effluents are qualitatively similar in composition al-
though the quantities of the various components diner. After
two years the raw and treated effluents of both mills were
re sampled. Analyses showed that although concentrations of
the organics varied, the same compounds are still present.
"The need to know what chemicals may escape into the
environment and at what levels they may be harmful leads
rather quickly to a realization that until one can identify these
compounds with certainty and measure their presence in se-
lected compartments of the environment, effective control of
these chemicals is essentially impossible" {!). Knowledge of
the specific chemical composition of treated wastewaters is
basic to the evaluation of the environmental impact of these
wastewaters and to the problem cf analyzing and controlling
their discharge. By tracing the chemicals through the treat-
ment system one can identify which compounds are being
effectively removed and which are resistant to the treatment
in use. Any new chemicals produced during treatment are
readily apparent. Once identifications are made, the ap-
proximate concentration of each compound can be calculated
at each stage of the treatment To our knowledge this study
represents the first attempt to characterize a wastewater
chemicaiiy, trace the dissolved volatile organics through a
treatment system, and correlate this information with the
traditional collective pollution prarameier measurements
(BOD. TOCl. The results detailed in this paper wert gathered
over a six-year period and portions of them have been pre-
sented previously (2—tl.
Paper Mill Treatment Facilities
Two mills, having similar processes but different waste
treatments, were used in this study. The first. Paper Mill "A",
daily produced about 1400 tons of containerboard in March
1972 when the samples were taken. Approximately 13 million
gal of water passed through the treatment system daily. The
treatment system is diagrammed in Figure l.The total BOD
reduction through the whole system is reported to be in excess
of 70%.
The second mill. Interstate Paper Corp.. Riceboro. Ga..
daily produced about 5-10 tons of containerboard in March
1972 when the samples were taken. Approximately 5.5 million
gal of water passed through the treatment system daily. The
treatment system is diagrammed in Figure 2. In thet55G-acre
stabilization lagoon, the highly alkaline effluent G'«
Staple P« 9
iRettntion time 3-6 inonlhil
Figure 2. Waste treatment system diayam of sne interstate Paper Corp
Hill
licit****
CBUt fflft)
tntirflUe Paper Corp.
Ricrtora. a.
clvtflcr
iRittnticn lim fcoul i noun)
Stcondarr CUrtflf
tfctiftton tiiw mcj J toti'il
tine FloctuUiion lini
Attention fine ootf 49 nil
Oritur
(Relntisn timi Aout t-t Muni
(|«mim un tat i ««r>
ftgui* 1. Waste treatment system diagram of Mill "A"
Figure 2. Waste treatment system diayam of sne interstate Paper Corp
mil) at RiceOofO. Ga.
Voli*ne tQ, Number 6, June 1976 SS5
-------
L
immediately frozen. They were stored at —10 °C and thawed
for.use as necessary.
Buth mills were sampled twice. In 1972. grab samples for
chemical characterization were taken at various stages of the
treatment systems and time delays were programmed so that
n "slue" of the effluent could be followed through the treat-
ment labilities. However, quantitative analysis indicated that
the slu): was missed at the outfall of Mill "A". During the
second sampling period, in January 1974. three-day com-
posites were collected from the raw effluent and the outfalls
of both mills b\ automatic sampling devices. Samples were
not taken at intermediate points.
A three-day composite composed of equal volumes from the
Monday. Tuesday, and Wednesday raw wastewater samples
of Mill "A" was prepared. A second three-day composite was
composed of equal volumes from the Wednesday. Thursday,
and Friday treated wastewater samples of Mill "A". Three-day
composities of equal volumes from the Monday. Tuesday, and
Wednesday raw wastewater and of the treated wastewater
samples of the Interstate Mill at Riceboro were prepared.
Because the stabilization pond retention time is at least three
months, no attempt was made to obtain samples of the same
slug of this wastewater. The concentrations in the 1972 grab
samples (Table II i cannot be considered "typical" because the
concentrations of the individual components vary significantly
with time. Comparison of these values with those obtained in
the 1974 sampling period (which are more "typical" because
they represent three-day composites) shows that while some
compounds varied almost by an order of magnitude, others
had very similar values.
Instrumentation
Both Yarian 1400 and Perkin-Elrner 900 gas chromato-
graphs iGC i equipped with flame ionization detectors (FID)
were vused. Generally, a commercially prepared (Perkin-
Elmeri 50-ft support coated open tubular (SCOT) capillary
column coated with Carbowax 20M.'terephthalic acid (K 20
C :
r
Figure 3. Gas chromatograms ol the methyl derivatives from Mill "A"
wastewater using (A) dimethyl sulfate. (6) diazomethane. (C) Me thy 1-8.
and (O) MethElute as methylating reagents
M/TPA) was used for separation. The optimum carrier gas
flow for our GC-MS systems operating under a vacuum and
using a Gohlke jet separator (IG— 18 ml/min) is twice the
amount necessary for optimizing conditions with an auxiliary
GC (operating under atmospheric pressure). From I96S to
1971. a I'erkin-F.lmer/Hitachi RMU-7 double focusing mass
spectrometer (MS) connected to a Perkin-Elmer 900 GC
through a Watson-Bieman separator was used. Now a com-
puterized Finnigan 1015 quadrupole MS connected to an
all-gla>s. single-stage Gohlke jet separator is used. Complete
descriptions of the instrumentation and our techniques for
computer matching the mass spectral data are described
elsewhere [7,8).
Analytical Procedures
Methylation Techniques. Paper mill wastewater extracts
contain two types of extractable volatile compounds: neutrals
(predominately terpenes and their derivatives), and acidic
compounds converted to their methyl derivatives to facilitate
GC separation. The aqueous solution can be methylated di-
rectly with dimethyl sulfate and sodium hydroxide, followed
by extraction of the methyl derivatives with chloroform, or
methyl derivatives of the acid extracts can be made in a sep-
arate step using diazomethane, on-column GC methylation
techniques, or several other common methylation procedures.
Several methods were evaluated—each has its advantages and
disadvantages, but their overall effectiveness is best illustrated
by Figure 3.
A 500-ml portion of the Mill "A" wastewater from sample
point 2 (primary clarifier effluent) was made alkaline to pH
11 with sodium hydroxide and extracted with chloroform to
remove the neutral compounds. Methylation of the aqueous
layer with dimethyl sulfate was followed by re-e.\traction with
chloroform to remove the methylated organics. The procedure
we use (9) is a variation of that described by Bichoet al. (J0..
II i. After concentration in a Kuderna-Danish apparatus to
0.5 ml. 1.2 m1 of the extract was chromatographed on a 50-ft
SCOT column coated with Carbowax 20 M/TPA and pro-
grammed from 100-200 °C at 4°/min with an initial 2-min
hold at 100°. The chromatogram is shown in Figure 3-A.
A 1-1. portion of the same wastewater was extracted with
chloroform at pH 11 to remove neutral compounds and then
made acidic to pH <2 with concentrated hydrochloric acid.
The aqueous layer was re-extracted with four 200-ml portions
of chloroform to remove the acids and phenols. The extracts
were combined and divided into two equal portions, each
representing 500 ml of the waste wafer extract. "
One portion was concentrated to near dryness in a Ku-
dema-Danish apparatus and methylated with diazomethane;
the final volume was adjusted to 0.5 mL The chromatogram
of 1.2 jil of this sample under cond ilions identical to those of
the previous sample is shown in Figure 3-B.
The second portion of the chloroform extract was concen-
trated in a Kuderna-Danish apparatus to near dryness and
MethElute (trimethylanilinum hydroxide in methanol. Pierce
Chemical Co.), which methylates the sample on-column, was
added to bring the volume to 0.5 ml. The chromatogram of 1.2
m1 of this sample under conditions identical to those of the
previous two samples is shown in Figure 3-D.
Another 500-ml portion of the wastewater, after extraction
of neutral compounds at pH 11. was made acidic and extracted
with four 100-ml portions of chloroform. The extracts were
combined, dried, and concentrated to near dr\ nt-ss as before
in a Kuderna-Danish apparatus. F.nough Methyl-8 (DMF
dimethyl arctal in pyridene, l'ierce Chemical Co.) was added
to bring the volume to 0.5 ml, and the solution was heated at
60 ®C for 15 min in a reacti-vial. The chromat<>«ram of 1.2 n\
of this sample, under condition* identical to those of the other
three, is shown in Figure 3-C.
556 Environmental Science & Technology
-------
Comparison of the chromatnuramR in Figure .1 shows that
dimethyl sulfate and MethElute are better reagents for
methyl a ting the paper mill samples than diaznmpthane or
Methyl-8. Phenols, especially guaiacol, were not methylated
well with diazomethane or Methyi-8. Although extraction or
guaiacol and other phenols with chloroform is not 1U0%. in-
complete extraction can be eliminated as a major contributor
to the smallnessof the methylated guaiacol peak (veratrole)
in the diazomethane sample because (he samples for diazo-
methane and MethElute treatment came from the same ex-
tract, divided into two portions; the MethElute sample shows
.a large veratrole peak. The dimethyl sulfate appears to be
equivalent to MethElute with respect to methylation of the
resin and fatty acids. However, the larger phenolic peaks in
the dimethyl sulfate sample and the absence of the MethElute
reagent peak made us decide to use dimethyl sulfate far the
remaining methylations.
GC-MS Techniques. After a sample is injected into the
computer-GC-MS system, the mass spectrometer automat-
ically scans its preset mass range every 4-5 s. When each run
is complete, the computer plots a reconstructed pas chro-
matogram i-tl using the algorithm
of a matching program (Inscribed in i!i» literature !J2). This
rapid-malchirig program was developed jointly by Handle
and the Environmental Research l-anoratory-Atheiis. nnd
utilized a CDC €400 time-shared computer (/.'J).
Results
Identification of Tcrpenes, Table 1 lisis lhe individual
compounds found in the neutral irac'ion of the wa-tewaier
extract?. Most iif the--* are terpens The corretpnndin;: fas
chromaUisram? of Mill "A" extract are shown in Fii'ure -1 wiih
peak rutmlifrt* lhat correspond ». Lheonmpinmd- in Table 1,
Vertical displays of gas chromatosram5. of extra of each step in the treatment process with respect to an
individual compound, to a class, uf compounds, or to all
chromatogTaphable compounds in the wastewater, whether
they are identified o: not. This information in conjunction
with the traditional collective pollution parameters (e.g., BOD
and TOC) provides a much better understanding and evalu-
ation of the effluent composition and its possible environ-
mental effects than the chemical profiles or collective pa-
rameters aiotie.
Two of the neutral extracts from Mill "A" were analyzed
by a Digilab Fourier Transform infrared spectrophotometer
Table I. Neutral Volatile! Identified in Both Kraft Paper Mill Effluents with Approximate Concentrations {13721
Approximate ccmca. in mg/L. Approximate conen. in
Puk
Mill "A" umple pcinu
Jnlerjm* sample points
no.
Compound identified
Confirmed by
1
2
3
4
1
2
3
4
4
Methyl trisulfide
0.003
0.004
0.008
0.001
—
—
—
62
Fenchone
MS, GC
0.007
0.007
0.015
0.015
0.015
O.O02
0.001
<0.001
63
Hexachloroethane
—
—
—
—
—
—
—
<0.001
64
Sabinene
—
—
—
—
—
—
—
0.003
65
Unidentified terpen® ketone
0.055
0.050
0.055
0.04 5
0.040
0.010
0.015
0.004
66
Camphor
MS. GC, IR
0.045
0.045
0.060
0.090
0.090
0.015
0.020
0.035
67
Unidentified terpene ketone
0.045
0.040
0.050
0.045
0.045
0.006
0.020
0.008
68
Unidentified terpene ketone
0.020
0.020
0.025
0.02D
0.020
0.003
0.0C8
0.002
69
Fenchyl alcohol
GC
0.065
0.065
0.035
0.010
0.105
0.040
0.060
—
70
Unidentified terpene ketone
0.025
0.025
0.020
0.004
0.015
0.005
0.009
—
71
Terpene-4-oi
MS, GC
o.oso
0.045
0.040
0.010
0.030
0.010
0.015
<0.001
72
2-Formylthiophene
0.010
—
—
—
—
—
—
—
73
Methyl chavicol
MS, GC
0.045
0.040
0.030
—
0.030
0.010
0.020
—
9
Borneo!
MS. GC. IR
0.275
0.200
0.155
0.090
0.470
0.200
0.260
—
10
O-Terpineol
MS, GC
0.645
0.700
0.625
—
0.490
0.215
0.260
0.080
11
Veratrote
MS. GC
0.020
0.015
0.015
0.008
0.015
0.004
0.008
—
74
2-A«ty rthiophene
GC. IR
0.025
0.025
0.030
0.025
0.012
C.002
0.004
<0.001
75
Myrtenol
0.010
0.010
0.008
—
0.008
—
—
—
76
2-Propiony Ith ioptiene
GC. IR
0.025
0.025
0.025
0.010
0.020
0.005
0.010
<0.001
77
Ane thole
0.007
—
—
—
—
i. —
—
—
78
Benzyl alcohol
0.013
0.012
o.coa
—
0.025
0.004
0.007
—
79
Methyl eugenol
GC
0.002
0.001
—
—
—
—
—
—
SO
Unidentified terpene alcohol
0.006
0.007
0.010
—
0.030
0.015
0.015
0.008
SI
Unidentified aromatic similar to
O.O06
0.006
0.009
—
0.009
o.ocs
o.oos
—
metiiyl isoeugenol (MW« 173)
30
Ethyl palmitate
MS.GC
0.006
82
Unidentified monounsaturated
0.038
C„ fatty acid methyl ester
83
Unidentified Qiunsaturated C„
0.016
fatty acid methyl ester
HA
Unidentified phthalate diester
Total
1.464
1.344
1.223
0.7! I
1.469
0.552
0.760
0.145
Volume 10, Number 6, Juna 1976 S57
-------
«'
66
10
M 74
J7
75|f7»
*70
4 $V,nn
|u"M"
uLvw
¦i: lo
«
\
69 N
10
/
tt , m
»!"'n
. 73
71 76
||l J
11 75 n
A\
66
67
4 «|
l>
1 .»
1
1 1* «
;• i
'11 w
1 f >
¦X J—A
10
15
n n
» * V f
J I I
71 *
II | 75 1 71
"Jv/si-L.
*0.13
n ao
*- -*
X 0.20
«P (I
! I
_J
I » 0.33
C-)n
C
>D .
C-?n
*4
I
JL-
6-Jn
U
I
A
G~4n
U
-A_
Figure 4. Chemical profile of neutral volatile compounds from Mill "A"
extracts of sample points 1-4 (top to bottom)
interfaced with a tfas rhrunuiti^niph (CiC- IR). S[>ertra of the
tinting romixuindsnre obtained "on the fly" as they are with
the (JO-MS system. These miiilyses were helpful in confirming
the functional croups of *evernl unirfentified compounds. For
example. com|xiunds fiT. GK. and 70 were all shown to he ke-
Uines. Computer matching of the mass spectra of lhe-hydroxybenzoic acid)
14
Methyl mandelate (mandelic acid)
—
—
—
0.025
—
—
—
—
IS
Dimethylsulfone
GC
0.240
0.400
0.400
0.055
0.240
0.540
0.400
0.130
16
Unidentified aromatic MW« 166
0.130
0.090
0.050
0.005
0.050
0.030
0.160
—
17
Methyl isomynstate (C„ fatty acid)
GC
—
—
—
0.003
—
—
—
0.001
18
p-Metnoxybenzaldehyde
0.055
0.050
0.020
0.025
0.090
•0.240
0.190
0.005
[p- hydrox ybertzaldehyde)
19
Eugenol
0.025
0.025
0.025
20
Methyl myristate (C,4 fatty acid)
MS. GC
—
—
—
0.020
—
—
—
0.010
21
Umaentified aromatic, MW = 196
—
—
0.045
0.035
—
—
—
—
22
Unidentified nonaromatic. MW = 196
—
—
— •
—
0.170
0.175
0.245
0.065
23
Methyl anteiiopentadecanoate
MS. GC
—
—
0.010
0.020
—
0.005
(C,. fatty acid)
24
Methyl 10-methyltetradecanoate
—
—
—
0.030
—
—
—
0.002
(C„ fatty acid)
25
p-Methoxyacetophenone
GC
0.020
0.025
0.030
—
0.060
0.130
0.080
0.010
(^-hydrovyacetophenone)
S58 Environmental Science & Technology
-------
Table II. (Continued)
Compound identified as methyl derivative Confirmed —
Peak
Approximate concn in mg/l..
Milt "A" sample points
Approximate concn in mg/l..
Inter state sample points
no.
(parent compound)
by
1
2
3
4
1
2
3
4
26
Methyl pentadecanoate
MS. GC
—
—
—
0.020
—
—
—
0.010
(C„ "fatty acid)
27
Methyl isopalmitate (C„ fatty acid)
MS, GC
—
—
0.015
0.030
—
—
—
0.003
28
Methyl palmitate (C„ fatty acid)
MS, GC
0.070
0.190
0.180
0.430
0.140
0.160
0.035
0.017
29
Methyl palmitelaidate
MS, GC
0.005
0.020
0.025
0.125
—
—
—
0.010
(Cu frons-9-unsaturated fatty acid)
30
Ethyl palmitate (C„ fatty acid)
MS. GC
—
—
—
0.030
—
—
—
—
31
Methyl anteisomargarate
MS. GC
0.020
0.025
0.040
0.050
0.030
0.120
0.040
0.001
(C„ fatty acid)
32
Methyl 3,4-dimethoxyphenylacetate
0.050
0.075
0.090
0.080
0.090
0.365
0.150
0.003
{Homovanillic acid)
33
Veratraldehyde (vanillin)
MS, GC
1.500
1.700
0.450
0.410
2.100
4.400
2.600
0.070
34
2-Methylthiobenzothiazole
GC
0.035
0.025
0.030
0.025
—
—
—
—
(2-Mercaptober\zothiazole)
35
Methyl vanillate (vanillic acid)
GC
—
—
0.005
0.005
—
—
—
—
36
Acetoveratrone (acetovanillone)
GC
0.420
0.490
0.450
0.370
0.820
1.600
0.860
0.120
37
Methyl stearate (C„ fatty acid)
MS, GC
0.025
0.065
0.045
0.110
0.100
0.100
0.020
—
38
Methyl oleate
MS, GC
0.470
0.600
0.430
0.400
0.570
0.510
0.120
0.080
(C„ cts-9-unsaturated fatty acid)
39
3,4,5-Trimethoxybenzaldehyde
MS. GC
0.070
0.070
0.070
0.040
0.070
0.100
0.070
0.020
(syringaldehyde)
40
Methyl linoleate (C„ cis,cis-9.
MS. GC
0.350
0.470
0.230
0.100
0.450
0.920
0.160
—
12-diunsaturated fatty acid)
41
3,4-Dimethoxypropiophenone
GC
0.060
0.080
0.025
(3-Methoxy-4-hydroxy-
propiophenone)
42
3,4,5-T rimethoxyacetophenone
GC
0.055
0.050
0.050
0.050
0.090
0.200
0.130
0.005
(acetosyringone)
43
Methyl arachidate
0.030
0.035
0.030
0.025
0.035
0.025
—
—
44
Dihexyl phthatate
—
—
—
—
—
—
—
—
45
Unidentified resin acid "A"
—
—
—
—
0.070
0.075
0.130
0.020
methyl ester
46
Unidentified resin acid "B"
0.030
0.055
0.095
0.100
—
—
—
—
methyl ester
47
Unidentified resin acid "C"
0.100
0.055
0.010
—
—
—
—
—
methyl ester
48
Unidentified resin acid "D"
0.145
methyl ester
49
Unidentified unsaturated fatty acid
—
—
0.005
0.035
—
—
—
—
methyl ester
50
Methyl pimerate (resin acid)
MS, GC
0.245
0.475
0.570
0.800
1.270
0.825
0.610
0.500
51
Methyl sandaracopimerate
MS, GC
0.050
0.060
0.125
0.045
0.340
0.245
0.275
0.110
(resin acid)
52
Unidentified resin acid "E"
0.025
methyl ester
53
Methyl-13-abieten-18-oate
MS. GC
—
0.430
0.800
1.400
0.100
0.035
0.050
1.050
(resin acid)
54
Unidentified unsaturated fatty acid
0.035
—
—
—
0.075
—
—
methyl ester similar to araconidate
55
Methyl isopimerate (resin acid)
MS. GC
0.430
0.660
0.770
0.780
4.400
1.700
1.20C
0.800
56
Methyl abietate (resin acid)
GC
0.370
0.430
0.420
0.050
3.300
1.500
1.900
—
57
Methyl dehydroabietate (resin acid)
MS, GC
1.500
1.300
4.000
1.000
3.300
2.700
3.600
3.900
58
Methyl 6,8,11,13-abietatetraen-
MS, GC
0.065
0.070
0.170
0.095
0.160
0.115
0.280
0.180
lS-oate (resin acid)
59
Methyl neoabietate (resin acid)
MS. GC
0.105
0.125
—
—
1.300
1.200
0.450
—
60
Methyl lignocerate (C14 fatty acid)
MS. GC
—
—
0.055
0.030
—
—
—
—
61
Dioctyl phthalate
—
—
—
—
—
—
—
—
Totals
9.380
10.692
10.246
7.123
21.690
^l.480
17.700
7.243
10-methyltetradecanoic, pentadecanoic, isopalmitic, palmitic,
palmitelaidic, anteisomargaric, and oleic). Six others were
found in the lagoon effluent of one mill but not the other {ethyl
palmitate (as the ester), stearic, linoleic, Iignoceric, arachidic,
and unidentified unsaturated acid #Mj. In both mill waste-
waters a greater number of fatty acids were found in the la-
goon effluent than in the influent. The majority of the new
compounds are saturated low-molecular-weight (C-1-4.C-15,
C-16) branched and straight-chain fntty acids and are prob-
ably metabolites of the lagoon biota. The only fatty acid
identification not verified by comparison of GC retention time
with a standard is the methyl ester of 10-methyltetradeca-
noate. Its identification rests solely on interpretation of its
mass spectrum (9).
Little or no work has been reported on the analyses of fatty
acids in treated paper mill effluents. We have found no pre-
vious report of odd-numbered carbijn and branched fatty
acids in treated wastewaters. The presence of 10-methylte-
tradecanoic acid and ethyl palmitate, although in small
amounts, was unexpected.
Volume 10, Number 6. June 1976 559
-------
Identification of Rosin Acids. A total of 1.1 different resin
artds (Table II) was found in the wastewaters ofltoth mills in
the 1972 samples: eight were identified and confirmed and five
were not identified. The rrsin arid contents of the finished
wastewaters were qualitai ively similar for the two mills. Seven
resin acids were common to the raw wastewater extracts of
both mills and six of these were found in Ixilh finished
wastewater extracts. Differences were primarily found among
the small, unidentified resin acid GC peaks.
In addition to the isopimaric, abietic. dehydroabietic, and
sandaracopimaric acids found in a recent study hy 1-each and
Thakore (/o). we identified pimaric. 6,8.11,13-ahietate-
traen-18-oic, neoabietic. 13-abieten-lS-oic, and several un-
identified resin acids. To our knowledge, neoabietic, 13-abi-
eten-18-oic, and 6,8.11.1.>-abietatetraen-18-oic acids have not
been reported before in kraft pulp mill wastewaters. All of
these resin acids (as their methyl esters) were initially iden-
tified by comparison of their mass spectra with published mass
spectra (76). They were all confirmed by comparison of GC
retention times and GC-MS from our instrument with stan-
dards obtained from Dr. Duane F. Zinkel, USDA Forest
Products Laboratory, Madison, Wis.
In general, most compounds decreased significantly in
concentration as the wastewater passed through the treatment
system. One exception appeared to be the resin acid content.
Some of the peaks appeared to increase in concentration. We
believe this to be the result of large variances in the concen-
tration of the organics in the mill wastewaters with time and
failure to have precisely sampled the "slug" of effluent we tried
to follow through the treatment systems.
Identification of Phenols. Kleven different phenols were
identified in the kraft mill wastewaters (Table II). Based upon
the data obtained in 1974. some correlation ran be made be-
tween the percent removal of the phenols in the two waste
treatment systems and the complexity or degree of substitu-
tion of the phenol molecules. Table III lists the percent re-
moval of these phenols and shows their structures. The ana-
lytical data for the calculations were obtained from 3-day
composite samples quantitated with the aid of the Perkin-
Elmer PEP-1 computer system. The phenols are generally
more resistant to treatment as the complexity of the molecule
increases. This is especially true with Mill "A", which has only
biological treatment, with the single exception of p-hydrox-
vbenzaldehyde. which might be produced in the aerated la-
goons by biological degradation of lignin or vanillin. The
chemical profiles of the acids and phenols from both mills are
shown in Figures 5 and 6. The peak numbers correspond to
the compounds listed in Table II. The compositions of the raw
effluents from the two mills were similar, but the treated ef-
fluent samples from the two mills exhibited more differences.
Computer-Assisted Gas Chromatographic Analyses.
Based on the compound identifications from the 1972 analy-
ses. a computer-assisted GC analysts was used to analyze the
1974 wastewater samples. GC-MS was used to verify corn-
Table III. Percent Removal of Phenols vs. Their Structural Complexity (1974)
Parent compound
Identif iad at
% removed
Mill "A" Interstate
Guaiacol
Veratrole
96
98
p-Hydroxybenzaldehyde
p-Methoxybenzaldehyde
21
75
Vanillin
Veratraldehyde
-OCH
65 . 93
Acetovanillone
Acetoveratrone
,OCHj
77
90
Homovanilhc Acid
Methyl homovanillate
,OCH_
50
Syringaldehyde
3,4,5-Trimethoxybenzaldehyde
62
Acetosyrinqone
3,4,5-T rimethoxyacetophenone
57
67
mn
S60 Environmenial Science A Technology
-------
pound identifications; Once -compound!) were-identified; -
computer-assisted GC analysis was sufficient for subsequent
qualitative and quantitative analyses. Chloroform extracts
of 1 I. of each sample were concentrated to 1.0 ml. A 300-mI
aliquot of each extract was spiked with 30 «j| of a standard
solution (11 mg/ml) of acenaphthene, the internal standard,
bringing its concentration in the extract to 1.0 mg/ml. A
standard solution of acenaphthene, veratrole. methyl palmi-
tate, and methyl dehydroahietate was prepared and chro-
matographed using the PEP-1 computer system. From the
known concentrations and the measured peak areas, response
factors were calculated for each compound relative to that of
acenaphthene (veratrole, 1.49; methyl palmitate. 2.05; methyl
dehydroahietate, 2.32). Concentrations of the various phenols
and fatty and resin acid methyl esters in the samples were then
calculated by computer, we assumed that all compounds in
a given chemical class have similar response factors.
Computer analyses of the three-day composite extracts
from Mill "A" and from the Interstate mill are shown in Fig-
ures 7 and 8, respectively. Peak number designations, notation
of the presence or absence of the compound in the corre-
sponding 1972 sample, and the sum of all the concentration
values were manually added to the data printed out by the
computer.
Collective Pollution Parameters. Comparison of col-
lective pollution parameters (Table IV) indicates that BOD
removal is about 70-90% in both mill wastewaters during
treatment. TOC reduction in the treated wastewaters of both
mills probably varies between 60-90%. Because of large
amounts of suspended solids in the raw effluent of Mill "A",
the TOC values may be erroneous. Reduction in total chro-
matographable acidic materials is about 65-80% in both mills
G-la
4a
¦ ¦ » I 9 | r »
j n ¦x Iff
Mill i(,
The decrease of only 24% in the aeidic material for the 1972
sampling of Mill "A" is probably not representative.
If treatment effectiveness is considered with respect to
classes of compounds, the phenols appear to be the most
susceptible to treatment Reduction in the chromaiopraphable
phenolic content of each mill was very consistent, ranging in
Mill "A" wastewaters from a 73-77% reduction. Chromato-
graphahle phenols were reduced in the Interstate wastewaters
by 94-95%.
Resin acid reductions in the treated wastewaters ranged
from about 50-90%. The apparent increase in resin acid con-
tent of Mill "A" in 1972 is probably due to a nonrepresentative
grab sample.
Fatty acid content of the Interstate wastewaters was de-
creased by about 85-90%, but the overall fatty acid concen-
trations increased in both the 1972 and the 1974 samples of
Mill "A" treated wastewaters. This is probably due to the
production of fatty acids by the microbiota in the aerated la-
goons. An increase in number of branched and odd-carbon
fatty acids was also noted.
To maintain the proper perspective with respect to a mass
balance, the total volatile acidic material as well as the three
major classes of compounds comprising it is presented as a
percentage of the TOC in Table V. The total of all the volatile
acidic components in the wastewater extracts was less than
10% of the total organic carbon content of the wastewaters.
The neutral components were less than 1% of the TOC. The
sum of all the volatile components therefore is still less than
10% of the TOC. Yet this relatively small amount of dissolved
organic material probably represents the bulk of problem-
causing compounds—both toxic and taste- and odor-causing
compounds. The dark brown color of kraft pulp mill waste-
Figure 5. Chemical profile of acids and phenols from MiU "A" extracts
of sample points 1-4 (top to bottom)
Figure 6. Chemical profile ol acids and phenols from Interstate Mill
extracts of sample points 1-4 (top to bottom)
Volume 10. Number 6, June 1976 561
-------
Figure 7. Computer analysis of Mill "A" acids and phenols from (top) raw wastewater and (bottom) fully treated effluent (1974)
HUN
INST
1 - MILL A RAW EFFLUOIT-ACIDS AND PHEN0LS
. HETH0D SO
file
37 3i
TIME
AREA
RRT
RF
C
NAME
10. 14
3.1384
1.000.
1.4900.
.5051.
VERATR0LEI
13.66
1.7237
1.347.
2.0S00.
.3817,
DIMETHYLSULF0NEI
16.70
.2877
1.646.
1.4900.
.0463.
M ETH0 XYBEN IAL DEHYDEI
18.48
.4908
1.822.
1.0000.
. 0S30.
1
19. 10
9.2566
1. 883.
1.0000.
1.0000.
ACDJAPHTHENEl
20.08
.3537
1.980.
1.0000.
.0382.
I
22.10
1.3317
2.184.
2.0500.
.2949.
PALM I TATE I
2.220.
2.0500.
PALM ITELAI DATEl
23.75
.4753
2.350.
1.4900.
• .0765'.
ME H3M0VANILLATEJ
24.73
5.2457
2.450.
1.4900.
.8443.
VERATRALDQfYDEl
26.51
3.0875
2.629.
1.4900.
.4969.
VERATR0NE:
26.82
3.S064
2.661.
2.0500.
.7765.
STEARATE AND BLEATEl
27.47
.9110
2.726.
1.4900.
.1466,
ME SYRINGALDEHYDEl
27.66
" .7588
2.746.
2.0500.
.1680.
LINSLEATEl
27.94
.2840
2.774.
1*4900.
.0457.
3.4-CNPt
28.98
.9289
2.879.
1.4900.
.1495.
3. 4. 5-TMAl
30.59
.5729
3.037.
1.0000.
.0618.
31.13
.5133
3.057.
2.3200.
.1286.
ME RESIN ACIDt
31.82
.4960
3.152.
2.3200.
.1243.
ME RESIN ACIDS B. Ci
32.86
4.0798
3.250.
2.3200.
1.0225.
PIMERATEj
34. 10
6.7065
3. 36 6.
2.3200.
1.6808.
SAND-P AND 13-AB- IS- I
36.20
3.8577
3.563.
2.3200.
.9668.
1S3PIMERATEJ
41.92
26.4614
4. 100.
2.3200.
6.6320.
AS* AND DD1YDR0AB-I
44. 1 1
.3432
4. 305.
2.3200.
.0872,
6.8. 11. 13-AS-t
44.59
.5035
4.350.
2.3200.
. 1262.
NE3ABIETaTEj
45.81
.4653
4.464.
2.0500.
.1030.
L1GN0CERATEI
PEAK
IN *72
11
YES
IS
YES
18
YES
A
--
28
YES
29
YES
32
YES
33
YES
36
YES
37.38
YES.YES
39
YES
40
YES
41
YES
42
YES
- -
- -
—
N0
46.47
YES. YES
50
YES
51.53
YES.N0
55
YES
56. 57
YES.YES
58
YES
59
YES
60
N0
RUN
INST 4
TBTAL> 14.9834 MG/L
y.ZLL A TP. EAT ED EFFLUENT-AC I OS AND PHENOLS
. METH0D
50
file
36
31
TIME
APEA
RRT
RF
C
NAME
PEAX
IN '72
10. 12
.1291
1.000.
1.4900,
.0193.
VERATR0LEI
11
YES
13.64
. 1732
1 . 347,
2.0500,
.0356.
DIHETHYLSITF3NEI
IS
YES
16.70
. 2452
1 .650,
1.4900,
.0366.
METH0XYBENZALDEHYDEl
18
YES
17.42
• 1 129
1.721,
2.0500,
.0232,
MYRISTaTEj
20
YES
19.39
.1112
1.817,
1.4900,
.0166,
ARBMATIC M V 1681
—
--
18.7 1
. 2002
1.348,
2.0500,
.04 12,
ANTEIS0 C-15 1
23
YES
19.11
9.9590
1.388#
1.0000.
1.0000,
ACENAPHTHENEi
A
19.79
. 1403
1.955,
2.0500,
.0238,
PENTADECAN0ATEJ
36
YES
2C. 12
. 1304
1.93S,
1.00GC.
.0181.
1
—
--
21. C3
.nil
2.079,
2.0500.
.0352,
IS0PALM I TATEI
27
YES
22. 11
1.<5S7S
2.187,
2.0500,
.3885,
PALM I TATE«
28
YES
22. 5'
1.1136
2.234,
2.0500.
.2292,
PALM I TELAI DATE I
29
YES
23.66
• 2549
2.343,
1 .4900,
• 033 1,
ME H0M0VANILLATEl
32
YES
24.44
.2631
2.42 1,
2.0500.
• 0541,
MARGARATE:
M
N0
24.72
1 .9430
2.450,
1.4900.
.2914,
VESATRALDEHYDEi
33
YES
26.4?
.7'S 1
2.626,
1.4900,
.1164,
VEP.ATR0NE1
36
YES
26.93
2.2860
2.67 1,
2.0500,
.4705.
STEARATE AND 0LEATEI
37.38
YES
27 .44
. 3729
2.723,
1.4900,
.0S58.
ME SYRlNGALDEHYDEi
39
YES
27.64
. ! 332
2.743,
2.0500,
.0377,
LIN3LEATE:
40
YES
27.92
. 1779
2.77 1,
1.4900,
.0266,
PHTHALATEt
—
28.55
• 3314
2.834,
1.0000,
.0332,
t
—
--
28.98
.4289
2.877,
1.4900,
.0641,
3.4. S-TMAi
42
YES
20.54
. 2222
2.934,
1.0000.
.0223,
t
—
--
30.53
.6436
3.031,
2.0500.
.1335.
ARACHIDATEl
43
YES
31.04
.4631
3.081,
1.4900,
.0692,
D : H EX YL PHTH ALATE1
44
YES
31.78
.2201
3.154,
2.3200,
.0512,
RESIN AND FATTY ACIDj
46.49 .
YES
32.78
. 9428
3.252,
2.3200.
.1963.
PIMERATEj
SO
"yes
33-96
1 .4073
3•368,
2.3200.
.3278.
SANC-P AND 13-AB-18-I
SI.53
YES
3S.48
.2199
3.516,
2.0500,
.0452.
UNSAT FATTY ACIDl
54
N0
36.06
.7912
3.573,
2.3200.
• 1843.
1S0PIMERATE:
S5
YES
41 .43
2. 5728
4.100,
2.3200.
.5993.
AB- AND CEHYDR0AB-1
56. 57
YES
44. 17
• 2192
4.368,
2.3200.
.0510,
NESASIETATEt
59
N0
45.48
.2818
4.496,
2.0500.
.0580,
LIGNOCERATEj
60
YES
49. 59
. 1314
1.0000.
.0131.
1
TBTAL I 3.8! 14 HG/L
N0TEt IO-METHYLTETRADECAN0ATE UNDER ACENAPHTKENE IS ESTIMATED AT
A CONCENTRATION 0F 0.03 MG/L.
FATTY acid METHYL ESTER AT 31.78 MIN. IS UNSATURATED.
562 Environmental Science & Technology
-------
II
Figure 8. Computer analysis of Interstate Mill acids and phenols from (top) raw wastewater and (bottom) fully treated effluent (1974)
flUN INTERSTATE raw effluent-acids and phenols
01 ST
WN
INST
, HETH0D
SO
# file
33
3i
TIME
AREA
RRT
RF
C
NAME
10. 10
8.3184
1.000,
1.4900#
1.3637#
VERATR0LE:
11.89
.4989
1.177,
1.0000#
.0548#
t
13.65
2.7822
1.351.
2.0500#
.6275#
DIM ETHYL SULF0NE:
IS.60
.2228
1•544#
1.0000#
.0245#
i
16.66
.5923
1.649,
1.4900#
.0971#
METH0XYBDJ £ALDEHYDEi
17.39
.3927
1.721,
1.0000#
.0432#
I
17.85
.5308
1.767.
1.0000#
.0584#
t
18.36
1.6764
1.817#
1.0000#
. 1844#
t
18.72
.6656
1.8S3#
1.4900#
.1091#
AR0MATIC MV 182:
I*. 07
9.0883
1.888#
1.0000#
1.0000#
ACENAPHTHENE:
20. OS
.4136
1.985#
1.0000#
.0455.
t
22. OS
.8961
2.186#
2.0500#
.2021#
PALM I TATE:
22.80'
.2664
2.262#
1.0000#
.0293#
I
23.30
.3619
2.312#
2. 0500#
.0816#
ANTEIS0MARGARATE:
23.68
.7918
2.3S1#
1.4900#
.1298#
ME KZM0VAN1LLATEI
24.37
.2071
2.4S0#
2.0500#
.0467#
MARGARATE:
24.66
7.7353
2.450#
1.4900#
1.26S 1#
VERATRALDEHYDEi
26.41
3.8972
2.626#
1.4900#
.6389#
VERATR0NE:
26.71
2.2447
2.656#
2.0S00#
. 5063#
STEARATE AMD 0LEATE:
27.57
1.9567
2.743#
2.0500#
.4413#
MOSTLY LIN0LEATE:
28.88
.4292
2.875#
1.4900#
.0703#
3# 4# 5-TWA t
29.37
.2510
2.925#
1.0000#
.0276#
1
29.96
.2354
2.984#
2.0500#
.0576,
AAA CHI DATE:
30.47
.4545
3.031#
2.3200#
.1160#
HE RESIN ACID:
31.01
.5225
3.081#
2.3200#
.1333#
ME RESIN ACID:
31.73
1.1206
3.148#
2.3200#
.2860#
ME RESIN ACID MV 314i
32.78
7.8243
3.246#
2.3200#
1.9973#
P1MERATE:
33.60
2.1625
3.322#
2.3200#
.5520#
SAN DARAC3 PIMERATE:
33.98
.9214
3.357#
2.3200#
.2352#
13-ABIETEN-18-0ATE:
34.54
.6673
3.409#
1.0000#
.0734#
I
36.23
24.6361
3.566#
2.3200#
6.2338.
IS0PIMERATE:
38.77
.8480
3.802#
1.0000#
.0933#
l
41.98
50.9977
4. 100#
2.3200#
13.0176#
AB- AN D DEHYDR0AB-I
42.80
.7900
4. 176#
1.0000#
.0869#
:
44.62
6.8304
4.345#
2.3200#
1.7436#
6#8# 1 1# 13-AB# RE0AB-I
45.92
.3696
4.465#
2.0500#
.0833#
LIGN0CERATE:
PEAK
IN *72
11
YES
15
YES
18
YES
— "
—
Ne
A
--
28
YES
31
YES
32
YES
M
N0
33
YES
36
YES
37#
38
YES# YES
40
YES
42
YES
43
YES
—
N0
45
YES
—
N0
50
YES
51
YES
53
YES
55
YES
56#
57
YES#YES
58#
59
YES#YES
60
N0
TOTAL J
30.7714 HG/L
INTERSTATE TREATED ETTL-ACJOS AND PHOJ0LS
# HETH8D
SO
file
41
3s
TIME
AREA
RRT
RF
C
NAME
PEAX
IN *72
10.12
.1315
1.000#
1.4900#
.0220#
VERATR0LE:
11
TES
1.350#
2.0500#
DIMETHYLSULF0NE:
15
YES
16.65
. 1423
1.645#
1.4900#
.0239#
HETH0 XYBEJJ ZALDEHYDEt
18
YES
18.41
.1368
1.819#
1.0000#
.0154#
1
—
_ _
19.06
8.8748
1.883#
1.0000#
1.0000#
ACENAPHTHENE:
A
--
20.07
.1581
1.983#
1.0000#
.0178#
I
—
—
22.07
.2320
2. 186#
2.0500#
.0535#
PALMITATEJ
28
YES
22.74
.2390
2.254#
1.0000#
.0269#
:
—
24.41
.2475
2.423#
2.0500#
.0571#
MARGARATE>
H
N0
24.67
.5476
2.450#
1.4900#
.0919#
VERATRALDEHYDEl
33
TES
26.45
• 399S
2.630#
1.4900#
.0670#
VERATR0NE:
36
YES
26.89
.3413
2.675#
2.0500#
.0788#
STEARATE AND 0LEATE1
37.38
N0#YES
2.720#
1.4900#
. HE SYRING ALDEHYDE:
39
YES
28.96
.1394
2.885#
1.4900#
.0234#
1.4# 5-THAI
42
YES
29.31
.2972
2.941#
1.0000#
• 0334#
t
—.
—
30.02
.2021
2.991#
1.0000#
.0227#
t
—
--
31.06
1.6762
3.093#
2.3200.
.4381#
ME RESIN ACID:
45
TES
31.72
1.3443
3.157#
2.3200#
.3514#
ME RESIN ACIDl
48
YES
32.74
2.9344
3.257#
2.3200#
.7670#
PIMERATEj
SO
YES
33.91
3.6459
3.37 1#
2.3200#
.9530#
SAND-P AND 13-AB-18-1
51 # 53
YES.YES
35.98
2.8126
3.572#
2.3200#
.7352#
IS0PIMERATE:
SS
YES
41.39
11.7420
4. 100#
2.3200#
3.0694#
AB- AND OEHYDR0A8-:
56# 57
N0.YES
43.54
.2298
4.309#
2.3200#
.0600#
6.8# 1 1# 1 3-AB# NEB AB- :
58# 59
YES#N0
4.460#
2.0500#
LIOJ0 CERATE:
60
N0
TOTAL I
6.9079 HG/L
Volume 10, Number 6, June 1976 S63
-------
Table IV. Collective Pollution Parameter Measurements and Total Concentrations of Volatile Components in
Acid-Phenol Extracts
Concentrations, mg/l.
Mill
"A" •
Interstate Paper at Riceboro. Ga.
1972
1974
1972
1974
%
%
%
%
Raw
Outfall
Change
Raw
Outfall
Change
Raw
Outfall
Change
Raw
Outfall
Change
BOD.
323
88
-73
320
45
-86
438
70
-84
440
52
-88
TOC
240
230
—4
350
350
0
470
200
-57
490
85
—S3
Total GC organics"
9.38
7.12
-24
14.98
3.81
-75
21.69
7.24
-67
30.77
6.91
-78
Total phenols
4.96
1.15
-77
2.31
0.62
-73
5.53
0.27
-95
3.57
0.23
-94
Total fatty acids
1.01
1.46
+45
1.36
1.59
+17
1.40
0.14
-90
1.34
0.19
-86
Total resin acids
2.04
4.27
+109
10.77
1.39
-87
14.24
6.71
-53
24.37
6.37
—74
" Rented to concentration] o< total gas enromatograonabie organic material.
waters is believed to be due to partially degraded Iignin mol-
ecules. These high-molecular-weight nonvolatile compounds
form a significant portion of the balance of the TOC. Other
contributors to the TOC are the carbohydrates and, to a lesser
extent, tannins and various other highly polar or nonvolatile
compounds.
The increase in the proportion of the TOC represented by
total volatiles in the Interstate wastewaters is a reflection of
the greater decrease in the nonvolatile portion of the organic
content as compared to the volatile content. Lime flocculation
probably removed a large amount of the nonvolatile organics.
In summary, the major difference in the volatile organic
content of wastewaters from the two mills was the fatty acid
content. Whereas the fatty acids decreased significantly
during the Interstate traatment, they increased during
treatment in Mill "A". Reductions in BOD, TOC. total GC
peak areas, and phenolic and resin acid content were similar
in the wastewaters from the two mills.
Acknowledgments
Ail sample preparations and GC separations were done by
Terry Floyd. Mass spectral data were provided by Ann Alford
and Mike Carter. Infrared data were provided by Leo Azar-
raga and Ann McCall.
Charles Davis. Lloyd Chapman, and William J. Verross
provided valuable assistance and wastewater samples from
the Interstate Paper Corp. at Riceboro, Ga. Officials at Mill
"A", which prefers to remain anonymous, were also very-
helpful in providing both samples and information. Without
the cooperation of the staff froni these two mills, this study
would not have been possible.
Duane F. Zinkel (USDA Forest Products Laboratory.
Madison. Wis.) kindly supplied us with reference standards
of the resin acids.
Literature Cited
111 Nat ional Academy of Sciences' report of the Committee for the
Working Conference on Principles of Protocols for Evaluating
Chemicals in the Environment. "Principles for Evaluating Chem-
icals in the Environment", p 299, Washington, D.C., 1975.
l2i Keith. L. H.. "Identification ol Organic Contaminants Remaining
in a Treated Kraft Paper Mill Effluent", 157th National Meeting
of the American Chemical Society, Division of Water, Air, and
Waste Chemistry, pp 76-81, 1969
l.T) Keith L. H . Garrison, A. W„ Walker, M. M„ Alford. A. L.,
Thrusion. A. D., .Jr., "The Role of Nuclear Magnetic Resonance
Spectri**iipv and Mass Spectrometry in Water Pollution Analysis",
15Hth National Met-ting of the American Chemical Society, Division
of Water, Air. and Waste Chemistry, pp 3-6, 1969.
(4) Ciarrison, A. W., Keith, L. H.. Walker, M. M., "The Use of Mass
Spoctrometrv in the Identification of Organic Contaminants in
Water from the Kraft Paper Mill Industry", 18th Annual Confer-
ence on Ma»s Spectrometry and Allied Topics, pp B20.V213, 1970.
Table V. Volatile Acidic Components as Percentages
of TOC
Percentage of TOC
Mill A—
1974
Intarstati
a—1974
Raw
Raw
Component
effluent
Outfall
effluent
Outfall
Total acidic volatiles
4.28
1.09
6.28
8.13
Total phenols
0.66
0.18
0.73
0.27
Total fatty acids
0.39
0.45
0.27
0.22
Total resin acids
3.08
0.40
4.97
7.49
(5) Davis. C. L.. "Color Removal from Kraft Pulping Effluent by
Lime Addition", Environmental Protection Agency, Washington,
D.C., Pub. No. 12040 ENC. p iii, December 1971.
(6) Environmental Protection Agency, "Color Removal from Kraft
Pulping Effluent by Lime Addition", Technology Transfer Capsule
Rep. 2,1972.
(7) McGuire. J. M., Alford. A. L., Carter, M. H., "Organic Pollutant
Identification Utilizing Mass Spectrometry", Environmental
Protection Agencv, Washington, D.C., Pub. No. EPA-R2-73-234,
pp 10-13. July 1973.
(8) Hoyland. J. R., Neher, M. B., "Implementation of a Computer-
Based Information System for Mass Spectral Identification", En-
vironmental Protection Agency, Washington, D.C., Pub. No.
EPA-660/2-74-048, pp 5-33, June 1974.
(9) Keith, L. H., "Analysis of Organic Compounds in Two Kraft Mill
Wastewaters", Environmental Protection Agency, Washington,
D.C., Pub. No. EPA-660/4-75-005. p 99, June 1975.
(10) Bicho, J. G., Zavann. E.. Brink. D. L, "Oxidative Degradation
of Wood II", TAPPI, 49,218-26 (1966).
(11) Brink, D L., Wu, Y. T.. Noveau, H. P., Bicho, J. G.. Merriman,
M. M., "Oxidative Degradation of Wood IV", ibid., 55, 719-21
(1972).
(12) Hertz, H. S„ Hites, R. A., Biemann. K„ AnaL Chem., 43, 681
(1971).
(13) Hoyland, J. R., Neher, M. B., "Implementation of a Com-
puter-Based Information System for Mass Spectral Identification",
Environmental Protection Agency, Washington, D.C., Pub. No.
EPA-660/2-74-048, p 1, June 1974.
(14) Hurtfiord, B. F„ Friberg, T. S„ Wilson, D. F„ Wilson, J.
"Organic Compounds in Pulp Mill Lagoon Discharge", Environ-
mental Protection Agency, Washington, D.C., Pub. No. EPA-
660/2-75-028. p 15. June 1975.
(15) Leach, .1. M„ Thakore. A. N., "Identification of the Constituents
of Kraft Pulping Effluent That are Toxic to Juvenile Coho Salmon
(Oncorkvnchus kisutch)", J. Fish. Res. fird. Canada, 30, 479-83
(1973).
(16) Zinkel, D. F„ Zank. L. C.. Wesolowski, M. F., "Diterpene Resin
Acids , U.S Department of Agriculture Forest Service, Madison,
Wis., Forest Products laboratory, pp C1-C32, 1971.
Received for review October 28. 797.5. Accepted March 12, 1976.
Mention of commercial products is for identification only and dnet
not constitute endor\rnitmt by the Knuirnnmi'ulul frotection Agency
of the U.S. Cwernment.
564 Environmental Science & Technology
-------
Polycyclic Arcmatic Compounds
in Nature
-------
Polycyclic Aromatic Compounds
in Nature
These multiple-ring hydrocarbon molecules have been found
in soils and sediments around the world. They are unusually
stable, and their origins have presented an intriguing puzzle
by Max Blumer
Polycyclic aromatic compounds are
found in great variety in the natural
world; among them are plant and ani-
mal pigments of notable beauty and unusu-
al chemical stability. Alizarin, the red of
military uniforms ui Napoleon's day, be-
longs to this class of compounds; it exhibits
its original brilliance in museum collec-
tions, where other plant pigments have long
since faded. Deep-sea animals, particularly
sea urchins and sea lilies, owe their vivid
coloration to such pigments. Polycyclic ar-
omatic compounds are also noted for their
biological effects. Some of them can induce
cancer or cause mutations, even in very low
concentrations.
Just what are polycyclic aromatic com-
pounds? They are one of the principal
classes of substances where the central mo-
lecular structure is held together by stable
carbon-carbon bonds. The great variety of
carbon compounds—and of life itself, which
is based on such compounds—reflects the
almost limitless possibilities in the spatial
arrangement of carbon atoms in groups,
clusters, chains and rings. Long carbon
chains are common in natural products;
they have been recognized ever since organ-
ic chemistry became a separate discipline.
Branched carbon chains are equally com-
mon and also have a long scientific history.
The recognition in 186S that carbon atoms
can be linked in closed rings was one of the
great conceptual advances of chemistry.
The discovery was made by Friedrich
Kekule, who solved the riddle of the struc-
ture of the benzene molecule after having a
dream in which dancing snakes bit their
own tails.
From Kekule's insight it was only a short
step to the recognition that the naphthalene
molecule has two fused rings and the an-
thracene and phenanthrene molecules have
three. Such multiple-ring, or polycyclic,
compounds are said to be saturated if all the
bonds of the carbon atoms, beyond the min-
imum needed for carbon-carbon bonding,
are linked to hydrogen atoms. They are
called aromatic if some of the carbon atoms
are doubly bonded .to other carbon atoms.
In Kekule's day the rings of aromatic com-
pounds were thought to have single and
double bonds in alternation. That tradition-
al but mistaken view survives in the com-
mon graphical representation of the hexag-
onal benzene ring. According to current
views, the bonds that link the carbon atoms
in the benzene molecule and its polycyclic
analogues all play an equal role and have an
equal value. The theory of chemical bond-
ing helps to explain the exceptional stability
of six-carbon-ring systems.
Polycyclic aromatic hydrocarbons con-
sist of three or more fused benzene rings
in linear, angular or cluster arrangements.
By definition they contain only carbon and
hydrogen. Nitrogen, sulfur and oxygen at-
oms can, however, be readily substituted
for carbon atoms in the rings. The result-
ing "heteroaromatic" compounds are com-
monly grouped with the hydrocarbons,
which they resemble in their properties and
their behavior in analysis.
The identification of polycyclic aromatic
hydrocarbons and the study of their natural
pathways are pursued by geochemists, envi-
ronmental chemists and toxicologists. Since
the quantity of such hydrocarbons in most
samples is low, concentration and prelimi-
nary separation are usually necessary. In
my laboratory at the Woods Hole Oceano-
graphic Institution we use a three-step pro-
cedure consisting of gel filtration, column
chromatography and precipitation. These
methods interact in different ways with the
hydrocarbon mixture; in combination they
isolate the constituents of the mixture rap-
idly. In the first step the polycyclic hydro-
carbons are separated from larger mole-
cules. In the second they are separated from
substances that are held either less or more
strongly in a chromatographic column of
alumina. In the third step a solid precipitate
is formed with a reagent that combines spe-
cifically with aromatic compounds. Foreign
materials are excluded from the precipitate
and can be removed by washing with a sol-
vent. This three-step separation will work
with samples weighed in micrograms.
For very complex mixtures an additional
chromatographic column serves to resolve
the aromatic-hydrocarbon concentrate into
fractions of molecules that have equal num-
bers of rings.
The polycyclic aromatic fractions from
such a separation can be analyzed by vari-
ous techniques; as a rule no single existing
method will provide a complete analysis.
We have found most informative a combi-
nation of three techniques: measurement of
ultraviolet absorption, separation by distil-
lation in a high vacuum and mass spec-
trometry. The ultraviolet spectra of the hy-
drocarbon samples tell us much about the
spatial arrangement of the benzene rings
in them, but they yield little information
on the presence and nature of substituents:
atoms that are extraneous to the rings. Dis-
tillation further resolves complicated mix-
tures; it simplifies mass-spectrometric anal-
ysis and provides structural information,
based on the relations between structure
and volatility. The mass spectra comple-
ment the other data; they measure the over-
all size of the molecules, detect the presence
of substituents and often indicate their
character.
Until recently the detailed composition
of many mixtures of polycyclic aromatic
hydrocarbons found in nature was un-
known, and the sources of the compounds
were mysterious. Can chemical analyses
FOSSILIZED SEA LILY of the genus Milltricrinus was found to contain a series of pig-
ments. The photograph on the opposite page shows a section through a rootstock of the fossil
animaL. which lived about 150 million years ago on the muddy bottom of the Jurassic Sea,
southwest of what is now Basel in Switzerland. The middle of the stalk contains crystallized
pigments, given the name fringelites after Fringeli, the mountain where the fossils were found.
35
-------
provide clues to the processes of formation,
transformation and transportation of or-
ganic compounds in nature? To answer the
question I shall first take up some of the
principles governing the formation of poly-
cyclic aromatic hydrocarbons, after which 1
shall discuss some case histories that reveal
correlations between structure and origin.
Polycyclic aromatic molecular structures
are formed whenever organic substances
are exposed to high temperatures. In this
process, called pyrolysis, energy is released,
and the aromatic products thai are formed
are stabler than their precursors. For exam-
ple, pyrolysis occurs when a match is being
charred by the advancing flame. The char-
coal formed has the structure of graphite:
extended giant molecules consisting of
tightly linked benzene rings. High tempera-
tures and open flames are not required,
however, for the' aromatization and graphi-
tization of organic matter; even the heat
from an electric iron is sufficient to cause
incipient graphiiiiation and scorch fabric.
Given enough time, aromatization proceeds
at even lower temperatures. The aromatic
hydrocarbons of crude oil are formed over
millions of years in sediments that are at
temperatures between 100 and 150 degrees
Celsius.
The composition of the products of ther-
mal aromatization depends on the nature of
the starting material and on the transforma-
tion temperatures. Burning a log in a fire-
place or scorching a fabric with an iron
leads to products that are quite unlike
petroleum, although like petroleum they
contain polycyclic aromatic hydrocarbons.
One feature is particularly dependent on the
formation temperature of the hydrocarbon
mixtures: the abundance and relative distri-
bution of aromatic hydrocarbons that carry
substituent side chains of various lengths,
known collectively as alkyl groups. At very
high temperatures, as in the coking of coal,
the products consist of a relatively simple
mixture of unsubstituted hydrocarbons,
presumably because of the rapid cleavage of
the less stable alkyl bonds. At intermediate
temperatures, as in the smoldering of wood
in a smokehouse, complex mixtures of al-
kylated rings survive. Yet even there long
alkyl chains are not favored. Unsubstituted
hydrocarbons outnumber the substituted
ones, and the percentage of alkylated hy-
drocarbons decreases rapidly as the length
and number of the alkyl chains increase.
A quite different pattern is observed
when the temperature of formation is lower.
That pattern is best ilhistrated by analyses
of crude oils, there alkylated polycyclic aro-
matic hydrocarbons far exceed the unsub-
stituted ones, and the average degree of al-
kylation and the maximum number of car-
bons on the aromatic rings are far higher
than they are in samples produced by high-
temperature pyrolysis. This fact reflects the
conditions of petroleum formation: the time
is adequate to accomplish the energetically
favored aromatization but the temperature
is not high enough to fragment even the
weaker carbon-carbon bonds in the alkyl
chains.
It has been conjectured that polycyclic
aromatic hydrocarbons are being produced
today by thermal processes in soils at envi-
ronmental temperatures. Such mixtures
should be even more heavily alkylated than
petroleum and should therefore possess
characteristics that are readil) detectable,
even when tbey are mixed with hydrocar-
bons from other sources. As far as I know,
no such mixtures have been found.
Are polycyclic aromatic hydrocarbons
also formed by living organisms? If so,
can we distinguish between them and those
from pyrolytic sources? Many beautiful and
stable plant and animal pigments with
structures based on several aromatic rings
are indeed formed by living organisms. By
definition, however, such pigments are not
UNBRANCHED CHAINS
BRANCHED CHAINS
CARBON METHANE BUTANE
2-METHYLBUTANE
SCHEMATIC
STRUCTURES
BUTANE
NONANE
2-METHYLB'JTANE
2-METHYL-6-ETHYLNONANE
VARIETY OF CARBON COMPOUNDS is virtually without limit
With its tour valences, or bonds, the carbon atom readily forms
chains, branched and nnbranched, and ring structures in enormous
variety. In the "ball and stick" models in the top row the large balls
represent carbon atoms and the small balls hydrogen atoms. The
most ubiquitous carbon-containing ring structures in nature are those
based on the benzene ring: CgHfr. Benzene is described as being on-
saturated because not all the available carbon bonds are taken up by
-------
}?
hydrocarbons; they incorporate other ele-
ments, usually oxygen and nitrogen. A vig-
orous discussion continues about sugges-
tions that polycyclic aromatic hydrocar-
bons are also synthesized by living orga-
nisms, directly or by the transformation of
precursors that might resemble the surviv-
ing structures. It is particularly difficult to
resolve the controversy experimentally be-
cause of the ubiquity of aromatic hydrocar-
bons in nature and the ease with which they
are transported and contaminate experi-
ments.
In order to exclude such contamination
Gemot Grimmer of the Biochemical In-
stitute for Environmental Carcinogens in
Hamburg has grown plants under carefully
controlled conditions in greenhouses mon-
itored for the appearance of outside con-
tamination. In spite of extensive air-filter-
ing the most volatile aromatic hydrocar-
bons entered the greenhouse from the out-
side. Less volatile aromatic-ring systems
with four or more benzene rings, however,
were not found in the plants, which appears
to exclude their biosynthesis. The investiga-
tions were necessarily limited to a small
number of plant species, and synthesis may
still occur in other organisms.
There may be an indirect way to resolve
the controversy, or at least to suggest sam-
ples that could be analyzed for biosynthetic
mixtures of polycyclic aromatic hydrocar-
bons. Biosynthesis differs from pyrolytic
and geochemical synthesis in its great se-
lectivity. The living cell synthesizes only a
very limited number of compounds out of
an immense range of theoretically possible
ones; those are the compounds whose par-
ticular properties are required by the orga-
nism. When several members of one chemi-
cal family (for example several straight-
chain hydrocarbons) are formed by the cell,
they are often formed in markedly different
concentrations. In pyrolysis and geochem-
istry few selection rules are at work; the
resulting mixtures are exceedingly complex,
and adjacent members of hydrocarbon fam-
ilies are found in similar concentrations.
The characteristic biosynthetic selection
pattern is observed not only among the
straight- and branched-chain hydrocarbons
in organisms but also among the fatty acids,
amino acids and carbohydrates; indeed, it
is seen among most if not all constituents
of the cell. This appears to be such a fun-
damental biochemical principle that one
expects to observe it also in any mixture
of polycyclic aromatic hydrocarbons that
might be formed by organisms.
We now have the means of recognizing a
thermal or biochemical contribution to an
environmental assemblage of polycyclic ar-
omatic hydrocarbons. For pyrolysates we
can even estimate the formation tempera-
ture. I shall describe some natural polycy-
clic mixtures, give their hydrocarbon com-
position and attempt to deduce how they
were formed. As I proceed I shall note oth-
er correlations between the chemical struc-
tures and the processes of formation of
such hydrocarbon mixtures.
Compounds with an intense fluorescence
in ultraviolet radiation are found in
extracts from soils. Some 30 years ago W.
Kern, then an assistant at the laboratories
of Hoffmann-La Roche in Basel, became
curious about such extracts and studied
them in his spare time. His discovery in
1947 of the four-ring aromatic hydrocarbon
chrysene in a garden soil made him the orig-
inator of a new branch of environmental
chemistry Related compounds were soon
found. The isolation of the strongly carcino-
genic hydrocarbon benz[a]pyrene in rural
soils stimulated worldwide analyses. By the
end of the 1960's it was generally accepted
that from 10 to IS unsubstituted hydrocar-
bons are present in most if not all soils
in similar proportions everywhere. Various
speculations attributed their formation to
soil bacteria, to the decay of plant material
RING COMPOUNDS
CYCLOHEXANE
BENZENE
r^i
NAPHTHALENE
CYCLOHEXANE BENZENE NAPHTHALENE
(SATURATED) (AROMATIC)
other lands of atoms. Chemists use the word aromatic to describe un-
saturated ring compounds. Naphthalene Is the simplest polycyclic, or
multiple-ring, aromatic hydrocarbon. Cyclohexane, C«H [2, is the
saturated analogue of benzene. In the familiar diagram of the ben-
zene ring depicted here single and double bonds are shown as al-
ternating around the ring. Actually all carbon-carbon linkages are
equivalent, as if each pair of carbon atoms had one and a half bonds.
In the illustrations that follow, simple schematic structures sire used.
37
-------
rt
of substituted members containing methyl
groups, alkyl chains and saturated five-
member rings. Sulfur-containing analogues
of the aromatic hydrocarbons are also pres-
ent. These principal structural elements are
encountered in mixed systems in an almost
limitless number of permutations. The
or to fallout from polluted air. The limita-
tions of that picture rapidly became evident
when newer analytical methods with higher
resolving power were applied to the study of
polycyclic-hydrocarbon mixtures in soils
and recently deposited sediments. Within a
short time the number of known or suspect-
ed members of this family had increased by
at least two orders of magnitude.
The following picture of the hydrocarbon
compounds present in soils and recent sedi-
ments has emerged. Unsubstituted aromat-
ic hydrocarbons are the most abundant;
they are accompanied by extended series
ANTHRACENE
TETRACENE
PHENANTHRENE
CHRYSENE
PICENE
PYRENE
ISOMERIC BEMZOPYRENES
PERYLENE
CORONENE
MESO-NAPHTHOOIANTHRENE
POLYCYCLIC AROMATIC HYDROCARBONS consist of two or
more rings fused together in various ways. Rings of six carbon atoms
have the highest stability because they require the least distortion of
the natural bond angles of the carbon atoms. The rings may form lin-
ear arrangements (top row), angular arrangements (second row) or
dusters, as is illustrated by the remaining configurations. The two
benzopyrenes are known as isomers because they differ in geome-
try. Isomeric form at right, benz[a]pyrene, is a cancer-causing agent.
-------
hydrocarbon composition remains remark-
_ ably constant over a wide geographic-range,-
from continental soils to-sea-bottom sedi-
ments and from deposits that vary from
oxidizing to strongly reducing. It is most
unlikely that the great diversity of orga-
nisms associated with such a wide variety of
sites would contribute the same hydrocar-
bon series in the same proportions. The
complexity of composition and the similari-
ty in the concentrations of related com-
pounds also argue against a biochemical
source.
How good, then, is the case for a thermal
origin of these sedimentary hydrocarbon
mixtures? I believe it is excellent. In each
series the unsubstituted hydrocarbon pre-
dominates. Aikyl derivatives with up to 10
substituted carbon atoms are present, but
the concentrations decrease rapidly with in-
creasing alkylation [jfe bottom illustration
on next page) Thai is exactly the pattern
expected for medium-temperature pyrolv-
sis. Are there any pyrolysates of natural
products that exhibit such a pattern? My
colleague William W. Youngblood of Flori-
da Technological University suggested that
we analyze "hickory-smoke flavor," a com-
mercial flavoring obtained by the distilla-
tion of wood at moderately high tempera-
tures.
Remarkably, the alky] distribution pat-
tern of that pyrolytic material matched our
observations in soils and young sediments.
This evidence excluded several thermal
sources as the principal contributors to the
sedimentary hydrocarbons. For example,
the polycyclic aromatic hydrocarbons in
polluted air, derived from incomplete com-
bustion in furnaces and engines at higher
temperatures, contain fewer alkyl deriva-
tives than are found tn the sediment sam-
ples. Crude oil, ar the other end of ihe spec-
trum of formation temperatures, is much
more heavily alkylated. If the polycyclic hy-
drocarbons were formed in decaying plant
material, they should exhibit even more al-
kylation. In any case decaying plant materi-
al must be considered an unlikely contribu-
tor Jo the polycyclic aromatic hydrocarbon
fraction found worldwide in sediments for
the reason I have already noted: the huge
diversity of organisms involved. Thus all
the evidence suggests that the polycyclic
aromatic hydrocarbon fraction is generated
by natural pyrolytic processes. But bow is it
possible for the composition of the fraction
to be so constant over such a wide sampling
range?
I have a provocative suggestion that
would explain the relative uniformity of the
polycyclic aromatic hydrocarbon fraction
in soils and young marine sediments. Large
quantities of pyrolytic products are formed
in forest and prairie fires and are widely
dispersed by prevailing winds. The haze
over the North Atlantic is attributed in part
to such fires. Indeed, Dwight M. Smith,
John J. Griffin and Edward D Goldberg of
the Scripps Institution of Oceanography
have found carbon particles with a reccg-
CH-
,Cri
CH;
ISOMERS OF ALKYLNAPHTHALEN'ES contain by definition the same a umber of carbon
and hydrogen atoms. These are only four of many possible Isomers. The term alkyl refers to the
side chains an the rings. Although isomers are chemically identical, their structural differences
give rise to differences in properties. It is nonetheless difficult to isolate isomers in pure form.
HO.
•OH
PRIMITIVE RED PIGMENT »as extracted by the author from the fossilized rootstalk of the
sea lily Millrricrinus {see illustration on pagt 34). The pigment is one of a family of the frin-
ge lites that differ In the number and position of the hydroxy! (OH) groups on their molecules.
39
-------
C BErZOTHIOF'v-iEME
AZACHRrSENE
0 OH
o
ALIZARIN
ADDITION OF FOREIGN ATOMS lo polycyclic aromatic hydrocarbons greatly increases
the number of structural permutations possible. Alizarin, (be brilliant red pigment of .Napo-
leon's time, has oxygen atoms and bydroxyl groups as peripheral substituenti on its molecule.
It bears a family tekmbUice to the triagelite structure shown at the bottom of preceding page.
POLVCYCUC AROMATIC
HYDROCARBONS
EXAMPLE: TWO ALKYL CARBONS
ON PHENANTHRENE
ALKYL CARBONS
ON AROMATIC
RINGS
(1.10-DIMETHn.PHeNAriTHFENEJ
0 2*68
NUMBER OF ALKYL CARBONS
ON AROMA T3C RINGS
v
• u
z
<
a
z
c
<
0 2 4 6 0 Z 4 60 2 4 6 0 2 4 6
NUMBER OF ALKYL CARBONS ON AROMATIC RINGS
HIGH TEMPERATURE MEDIUM TEMPERATURE LOW TEMPERATURE
{—2,000° C) (-600"-W C ) (-ISCT-IWC.)
NUMBER OF ALKYL CARBONS present as side chains on polycyclic aromatic hydrocar-
bons correlates closely with the temperature at which the compounds were formed. In a typical
mixture (curve dxocarton series in ahem
overlap 10 some degree. They contain at
least several hundred polycyclic aromatic
hydrocarbons, together with their sulfur
and nitrogen analogues and alkyl and cycio
alkyl derivatives in many combinations of
substitution.
Can chemical analysis tell us something
about the origin of these minerals? The al-
ky! distribution of the polycyclic aromatic
hydrocarbons in idnalite and curttsite re-
sembles the distribution in soils and young
sediments. Thus the unsubstituted hydro-
carbons are the most abundant. The alky]
series does not extend as far, however, and
the drop in concentration from one member
of a senes to the next is steeper than it is in
the hickory-wood distillate. That suggests
a pyroiyttc origin at temperatures higher
-------
7& J&&L •¦
tfr-.+ZVJ!™*
%3*>.
^5^t-
~?3^VV..
• * '."S: t
?>'&% * . ,
¦ ^v? --^ j>*S~ + ->.
^TV.*-4 " t* _• .v "•'", f .'"*.• •
>?»i'-r^- ••• •- ¦^ .*¦" -~ • ^;--,-a'"- .'-».. ->-•' . •. .•¦" :—*v.-cv„v- -•
r^ara;
gSTr^';:':'^
fcur -Al —y ' irjlv^
^ * -r ti _jAu*£w. -i t -rtVfcy r 2f£<
HALO OF GRAPHITE forms aronnd growing crystals of ehiastolite mixtures of organic compounds that vary in volatility. The most re-
(aluminum silicate) in certain minerals associated with deeply bur- fractory fraction is graphite, which remains in specimens such as this
ied sediments. The organic matter in such sediments is converted iato one from R. Sa-do of the Woods Bole Oceanographic Institution.
than those in forest and prairie fires but
lower than those in furnaces and engines.
Curiously, the polycyclic assemblages in
idrial.te and curtisite differ from those dis-
cussed so far in the way the benzene rings
are arranged. The ring arrangements can be
divided into three broad categories: linear
(with all the rings in a line), angular (with
the rings in steps) and clustered (with at
least one ring surrounded on three sices).
The various arrangements differ in their sta-
bility. The linear arrangement in anthra-
cene and tetracene is the least stable; tetra-
cene and more complex equ:va.ents to it can
be prepared in the laboratory, but they do
not survive in nature. Clusters of benz sne
rings, as in pyrene, benzopyrtne and pery-
lene, are stabler and are commonly found
in pyrolvsates. The stablest configuration is
the angular arrangement of benzene rings in
phenanthrene, chrysene and picene. Such
stable series abcunc in idrialii; and curtis-
ite, which contain few of the hydrocarbons
of the cluster type tha: are common in soils,
41
-------
sediments, wood tar, tobacco smoke, petro-
leum and automobile-exhaust gases.
Evidently series of the cluster type are
formed in pyrolysis, regardless of the tem-
perature. They survive if the reaction prod-
ucts are rapidly cooled to temperatures
where the decomposition of the less stable
products is arrested, as it is in the case of
wood tar; they also persist in petroleum be-
cause it has never reached temperatures ca-
pable of eliminating or of rearranging the
less stable ring systems. This suggests that
idrialite and curtisite lack the unstable clus-
ter configurations because those configura-
tions have been eliminated by prolonged ex-
posure to elevated temperatures.
With such chemical clues and the geolog-
ical background, one pan interpret the for-
mation and evolution of these hydrocarbon
minerals as follows. Sediments containing
organic compounds are earned by move-
ments of the earth's crust to regions much
deeper than those where petroleum is typi-
cally formed. In those deeper regions pyro-
lytic temperatures probably go as high as
400 to 500 degrees C. The original organic
material is destroyed and its constituents
are rearranged and therm ©dynamically sta-
bilized. One of the products is graphite,
which remains at the original depth. The
stable hydrocarbon gases (in particular
methane) and the polycyclic aromatic hy-
drocarbons (along with their sulfur and ni-
trogen analogues) remain near the original
depth long enough for the least stable hy-
drocarbons to be destroyed. Eventually the
rest of the hydrocarbons move toward the
surface together with the gases, the mineral
waters and the mercury ores, which are rec-
ognized for their geochemical mobility.
Along the migration path the hydrocarbons
are separated by fractional crystallization.
Fractions of higher molecular weight and
higher melting potnt form idrialite and
pendletonite; fractions of lower molecular
weight and lower melting point end up as
curtisite.
Chemical analysis has again given us
clues to the formation of a natural hydro-
carbon assemblage, n suggests the gross
mode of formation and provides evidence
on the temperature of formation and the
duration of thermal exposure. As our inves-
tigations proceeded we visualized an expla-
nation for the formation of hydrocarbons in
such diverse materials as soils, young sedi-
ments, wood tar, tobacco smoke, engine ex-
haust and minerals such as idrialite and
curtisite. We obtained further correlations
between chemical structure and the proc-
esses of hydrocarbon formation from stud-
ies of the polycyclic aromatic hydrocar-
bons in petroleum.
Petroleum may well be the most complex
organic mixture on the earth. It is formed
from the residues of earlier life in buried
sediments by chemical reactions that re-
quire millions of years for completion The
transformation resembles pyrolysis. but the
reactions are exceedingly slow because of
the modest temperatures involved: probab-
ly less than 150 degrees C. Products that
would not be observed in high-temperature
pyrolysates are formed and retained.
The chemical analysis of the polycyclic
aromatic fraction in petroleum is a
challenging task. Adequate analytical reso-
lution requires the combination of many
different techniques, even then pure single
compounds are rarely isolated. Harold J
Coleman and his co-workers at the Bartles-
ville Energy Research Center in Oklahoma
have made extensive analyses of the poly-
cyclic aromatic fractions m crude oils, par-
ticularly those from the region of Prudhoe
Bay in Alaska. Polycyclic aromatic hydro-
carbons are abundant, amounting to about
a sixth of the oil fraction distilling between
370 and 535 degrees C. A similar propor-
CYCLIZATION
AROMA T1ZATION
FORMATION OF FIVE-MEMBER RINGS
CYCLIZATION AND AROMATIZATION of saturated hydrocar-
bon chains produce the complex mixture of polycyclic aromatic com-
pounds found in petroleum. The source material of petroleum, largely
derived from plants, is rich in long carbon chains. At elevated tem-
peratures in buried sediments some of the chains lose a few hydrogen
atoms and are converted into six-member rings (rop). With further
loss of hydrogen the saturated rings are aromatized (middle). Addi-
tional substitution by chains, followed by cyclization and aromatiza-
tion, gradually creates compounds with many rings. Five-member
rings, which also form readily (boiwm). cannot be converted into ar-
omatic rings. If se%eral accumulate around an aromatic nucleus, they
block its further growth to a still larger aromatic system. Aromatic
hydrocarbons surrounded by several five-member rings are abundant
in petroleum. Their production is favored by a reaction time of mil-
lions of years and by the absence of temperatures able to break the
somewhat strained carbon-carbon bonds in the five-member rings.
At
-------
lion is present in some fractions that have a
lower boiling point, and in their higher dis-
tillates and distillation residue. Three com-
positional features are particularly perti-
nent here: the extreme complexity and the
uniformity in the concentration of adjacent
compounds within xhe many hydrocarbon
series, the frequent occurrence of com-
pounds with saturated five-member rings
and the presence of compounds in which
the substituent groups show considerable
steric strain, that is, show bond angles dis-
torted beyond their normal range.
The complexity and the uniformity in the
concentration of adjacent compounds char-
acterize the origin of this crude-oil fraction;
it is not a biochemical pattern but the result
of extensive geochemical scrambling of or-
ganic compounds by nonselective reactions.
In the formation of petroleum some linear
carbon structures are transformed into sat-
urated Sve- and six-member rings. Such
reactions are possible because they stabi-
lize the source material thermodynamically
The six-member saturated rings thus creat-
ed are readily transformed by aromatiza-
tion—the loss of hydrogen—into ihe still
stabler aromatic systems. Those systems
can grow through the accretion and aroma-
tizatioci of new saturated six-member rings
The five-member rings, on the other
hand, are not readily converted into aro-
matic structures, so that with the passage of
time two, three or more such rings may
accrete around an aromatic nucleus, there-
by blocking its further growth into a large
aromatic system. The average number of
five-member rings around an aromatic nu-
cleus is far greater in petroleum than it is in
hydrocarbon series formed al higher tem-
peratures. That fact reflects the much great-
er time available for petroleum formation;
the marshaling of rings can continue, al-
though at low intensity, for millions of
years. In pyroiysis ai higher temperatures
five-member rings in the aromatic struc-
tures are less abundant either because the
reaction time is short, because- the source
material is rapidly converted into stable six-
member and aromatic systems 01 because
the temperature is high enough to break the
carbon-carbon bonds in saturated rings.
Highly strained molecules in which a
single CHj group forms a bridge between
two aromatic carbon atoms, as in 4,5-
methylenephenamhrene [see illustration on
this page], form with different relative abun-
dances in various pyrolysates. They are
most abundant in wood tar, lower by one
order of magnitude in soils and sediments
and by two orders of magnitude in crude
oils. In idriaiite and curtisite such com-
pounds are nearly absent. The distorted
bond angle at the CH 3 bridge implies that
such molecules have a higher energy con-
tent than molecules with less strain and
therefore less stability. We attribute the for-
mation of strained CH2 badges in pvrolysis
to energetic reactions and their survival ei-
ther to rapid quenching (as in tar and
smoke) or to low temperatures fas in petro-
6
6
7
JO
r
*0
4.5-METHYLENEPKENANTHRENE
1.10-ETHVLENEPHENANTHRENE
increasing strain
DECREASING STASfUTY
, INCREASING ABUNDANCE IN
WOOD TAR'
SOILS AND SEDIMENTS«
¦ CRUDE OIL'
•CURTISITE MINERALS
STRAINED FIVE-MEMBER RINGS are formed by pyroiysis (exposure to high tempera-
tures) from alkrl side chains. Only one alkyl carbon is needed (itft) if the parent structure sup-
plies three of the five sides of the ultimate ring. In other cases (nj/ir) two alkyl chains may be
needed. Because the carbon-car boo bonds are under greater strain in fite-member rings than
In six-member ones they have a greater eaergr content and hence a lower ftabiliry Five-
member rings survive most readily if the pyrolytic mixture is quietly quenched, as it is in the
case of smoke from a wood 6re, or if the pyroiysis continues for a long time at low temperature.
leuni formation) that depress the rate ai
which structural rearrangement can pro-
ceed. CH2 bridges have disappeared from
idnalite and similar minerals because they
were exposed 10 high temperatures for long
periods after the initial pyroiysis.
These examples will serve to show the
richness of information that is encoded in
the structure of organic compounds in na-
ture. The parameters I have discussed here
reflect the character of the source materials,
the processes of formation, the tempera-
tures at which the compounds form and the
reaction time. The relations among these
parameters suggest that the compounds
could be used as geological "thermometers"
and "clocks." Their calibration, however,
presents a difficult problem There must be
a complex interdependence among the pa-
rameters, since time and temperature boih
influence the survival of unstable struc-
tures. Yet even at the present level of our
understanding, detailed chemical analysis
has helped us to recognize a wide range of
environmental processes and has guided us
to a unified view of materials that are re-
mote in terms of sources and formation
processes.
Not all natural mixtures of polycyclic
aromatic hydrocarbons have a single
predominant origin Often there is a more
complex situation where more than one hy-
drocarbon source contributes to the sample.
For instance, a recent marine sediment may
contain fossil Fuels From an oil spill in addi-
tion to the hydrocarbons we attribute to the
fallout of soot particles from foresi and
43
-------
c
prairie fires. The different patterns of com-
position should make i: possible to recog-
nize these multiple sources.
With my co-worker Jeremy Sass 1 have
studied one such instance. An oil spill in
Buzzards Bay off Massachusetts has con-
taminated she near-ill or? bottom sediment
ana 4 feel ofl that has a comjaranreiy low
boiling point. Extraction and isolation or
the higher-boiling-point poJycyclic aroma!-
ic fraction and subsequent analyses reveal
the norma) background pattern we observe
not only throughout the bay but also in the
seal of the adjacent land. The unsubstituted
aromatics predominate and the alky! distri-
bution pattern is the one characteristic of
woodsmoke, that is, of the hydrocarbons we
attribute to the fallout of soot. Within the
range of beating points of the compounds in
the spilled fuel, however, alkyl derivatives
are more abundant than the unsubstnuied
hydrocarbons. Thai is the petroleum pat-
tern, and it is remarkable that it can slill
be observed now, almost six years after the
spill.
A particularly interesting example of a
dual soutce comes from our earlier investi-
gation of hydrocarbons in a fossil sea lily
Irora the Jura Mountains of Switzerland.
This fossilized deep-sea animal is -vividly
colored by pigments thai are dispersed
throughout the rock matrix of the fossil.
Present as tiny crystals, the pigment is a
true organic mineral; we named it fringeiite
after Fringeli, the mountain where we dis-
covered it. Since simitar compounds are still
found m living relatives of this animal spe-
cies, we assume that the pigment, or at least
its close precursors, were part of the animal
that lived iit the Jurassic Sea ISO million
years ago.
Fnngelite and its close relatives are aro-
matic compounds, When we attempted to
resolve the fossil pigments by chromatogra-
phy, we noted in the fraction that washed
oat of the column ahead of the pigments
some intensely fluorescing but nearly col-
orless materials. Further study revealed
u continuous pciycycUc series with from
three to seven a-cmauc rings. The lower
members of the series are not particularly
abundant and closely resemble aromatic
mixtures that are found in many other geo-
logical specimens. They may be the survi-
vors of compounds that were synthesized in
the geologic past or, what is more likely,
they may be geological transformation
products, originating in deep sediments that
were not sufficiently rich in organic materi-
al to yield petroleum. At higher molecular
weights we f.nd unusual hydrocarbons, nev-
er before isolated from sediments, thai ex-
hibit precisely those features we had pre-
stpuctural features
STRUCTURAL PREFERENCE
CONDITION Of FORMATION
HIGH FORMATION TEMPERATURE
HIGH CARBON CONTENT in SOURCE
material
LOW DEGREE OF ACKvUTION
OEGREE OP ALKVLATION
LONG EQUtUBRATION TIME
HIGH EQUILIBRATION TEMPERATURE
ANGULAR RING ARRANGEMENT
RING ARRANGEMENT (LINEAR. ANGULAR. CLUSTER)
SHORT EQUILIBRATION TIME
LOW FORMATION TEMPERATURE
HIGH DEGREE Of RING STRAIN
LONG REACTION TIME
LOW FORMATION TEMPERATURE
MULTIPLE FJVE-MEMBER SINGS
NUMBER OP FIVE-MEMBER RINGS
CONFIGURATION OF HYDROCARBON STRUCTURES found
in geochemical mixtures provides net information about the Condi',
tioos andej which tbe structures formed. The type and spatial ar-
rangement of the alkyl chains and of the saturated ani aromatic rings
can vary over a wide range. The presence of preferred structures in
natural mixtures of poly eye lie aromatic hydrocarbons reflects the
composition of the source material, the temperatures of formation
cad trans forma r ion and the duration of chemical processes involved.
-------
10.000 -
1.000 -
z
3
o
3
<
ui
>
100 -
10 -
1 -J
HEXAHYDRC-MESO-NAPHTHOOIANTHRENE
ANTHANTHRENE
PHENANTHRENE
BENZlajPYRENE
hexahydro-meso-anthrodianthrene
y
TWO GROUPS OF HYDROCARBONS are found in the fossil sea lily SiilUricrinui. The less
abundant group consists of fairly small structures of three to six fused benzene rings also widely
found in ancient sediments. The more abundant group consists of much larger polycyclic struc-
tures (color) that have never been found In other deposits. Their structures are so similar to the
fringelite pigments that a common origin seems very likely. This is one of the few cases where
at least part of a natural mixture of polycyclic aromatic hydrocarbons does not originate with
pyrolysis but can be traced back to a substance that was synthesized by a living organism.
dieted for substances of biochemicaJ rather
than geochemical origin. The number of
such compounds is limited, the chemical
structures are select and unusual, and the
concentrations are far higher than the nor-
mal background [see illustration at right].
More detailed structural analysis sup-
ports this interpretation. The exotic hydro-
carbons are the bare aromatic ring systems
on which the fringelite pigments are based.
We are not certain whether those hydro-
carbons were already present in the living
animals and therefore represent genuinely
biochemical products or whether they are
geochemical transformation products of
biochemical precursors closely related to
the fringelite pigments. Either would sup-
port the predictive value of the relations I
have suggested between chemical structure
and biological origin.
Our original view of aromatic hydrocar-
bons in nature has now changed dra-
matically. The picture of a dozen or so sim-
ple compounds distributed over the earth in
similar ratios has given way to the realiza-
tion that the mixture is exceedingly com-
plex and variable and that its origins are
diverse. The mixture is not yet fully ana-
lyzed; indeed, it may not be fully resolvable
at the present state of the analytical art. Yet
even our incomplete analyses reveal the op-
eration of a number of fundamental chemi-
cal processes, and we are learning to read
in the hydrocarbon structures a code that
can be deciphered in terms of sources and
processes of formation, transformation and
mass movement. We have spanned a wide
range of samples, and the evidence for a
thermal origin of polycyclic aromatic hy-
drocarbons in nature is overwhelming. The
biosynthesis of selected aromatic assem-
blages under special circumstances has not
been ruled out, but neither has it been dem-
onstrated convincingly.
Polycyclic aromatic hydrocarbons from
pyrolysis have been present on the earth for
a long time Man has been in contact with
combustion products throughout his histo-
ry, and natural fires and reactions in sedi-
ments formed polycyclic aromatic hydro-
carbons long before the advent of man.
Moreover, we now know that the hydrocar-
bons in smoke, in fallout from the air, in
sediments and in fossil fuels include, in
addition to already recognized compounds
that can give rise to cancer and mutations,
new carcinogens and mutagens. That raises
a new question: Have such materials con-
tributed significantly to the role of mutation
in the evolution of species? They might then
rank among other natural mutagens such as
.ultraviolet radiation and background nucle-
ar radiation.
Polycyclic aromatic hydrocarbons now
enter the environment in larger amounts
than they did in the human and geologic
past, and from some new sources. Among
the dominant sources are the incomplete
combustion of wood, coal and petroleum
and the spillage of raw or refined petroleum.
Attempts have been made to assess the ef-
fects of this new influx of aromatic hydro-
carbons into the environment, yet these at-
tempts predate the recent realization that
the environmental mixture of aromatic hy-
drocarbons is exceedingly complex, that it
has many sources, that it is difficult to ana-
lyze and that many components have never
been tested for their biological activity. It
seems important to reexamine the question
of the environmental effects of the new in-
flux of aromatic substances in the light of
our present knowledge and of our remain-
ing ignorance.
Since Dumas's first study of idrialite and
Kern's discovery of chrysene in soil, our
knowledge of polycyclic aromatic hydro-
carbons in the environment has developed
in step with the analytical art: slowly during
periods of analytical stagnation, rapidly
with the confluence of many modern ana-
lytical methods. I see a similar development
in other areas of environmental organic
chemistry. In each oyr best techniques fail
to approach the complete resolution of
chemical compounds. The question "How
complex is nature?" is an important one and
has many implications. As a geochemist I
see an almost limitless opportunity to inter-
pr;t a complex world in terms of the proc-
esses that have shaped it. As a pan-time
environmental biologist I am frustrated by
the difficulty, if not impossibility, of pre-
dicting the effect of organic chemicals in
nature without a fuller understanding of
their structure.
45
-------
V1
Asbestos in Your Future
-------
Asbestos in Your Future
by C. S. THOMPSON
Manager—Minerals, Ceramics and Paper R&D, R T Vanderbilt Co.
Reprinted from
the December 1976 issue of Mining Congress Journal
-------
Asoestos in Your Future
y C. S. THOMPSON
Manager—Minerals. Ceramics and Paper R&D, R. T. Vanderbilt Co
To many, the title of these comments might infer a ma-
jor discovery of new asbestos deposits, an increase in
production, or new areas of application. However, others
might suspect a more serious and sinister meaning, and,
unfortunately, they are correct.
Relative to world asbestos production and reserves
(with the exception of the huge ore body of short fiber in
the New Idria serpentinite in California), the United
States has always been a small contributor, involving a
very limited number of mining companies. This situation
appears likely to remain the case since nature, apparent-
ly, saw fit to distribute the majority of this unique and
highly useful material elsewhere in the world.
pertise (USBM, USGS, etc.), "asbestos" was defined as
any one of six minerals.
Table 1 lists these six minerals showing the correct ter-
minology for describing the asbestiform varieties of an-
thophyllite. tremolite and actinolite.
The first three minerals listed in the Osha document—
chrysotile, crocidolite and amosite—designate specifical-
ly the fibrous or asbestiform varieties of their vastly more
abundant non-fibrous counterparts. In contrast, the latter
three amphiboles—anthophyllite. tremolite and actino-
lite—as listed erroneously, include all forms of these very
common rock forming minerals which rarely occur in fi-
brous or asbestiform habit.
"Asbestos" minerals created by regulations
However, in the last few years, since excessive expo-
sure to asbestos has been shown to be a serious health
hazard, the United States government, through various
regulatory agencies, has promulgated and proposed nu-
merous regulations that have created "asbestos" miner-
als where they have never existed before, and thereby
caused essentially every mining company in this country
to be an "asbestos" miner and "asbestos" miller. In most
cases, this also makes them guilty of contaminating the
environment with "asbestos" by dumping their tailings
so that the natural weathering processes redistribute
these newly invented "asbestos" minerals into the air.
surface runoff and ground water. In the case of many min-
ers and producers of industrial minerals these new "as-
bestos" materials are distributed into a wide variety of
manufacturing operations as mineral fillers in paint, plas-
tics, rubber, or as basic ingredients in cements, spackling
compounds and ceramic products.
To understand how this startling situation has come
about, let us examine the chronology of the chain of
events which has occurred over the past few years.
Issuance of standard in 1972 started chain of events
It all began with the issuance of the Occupational Safe-
ty and Health Administration (Osha) Asbestos Standard
in July 1972.' based on the National Institute For Occupa-
tional Safety and Health (Niosh) criteria document,2 with
the purpose of controlling human exposure to asbestos.
Without the benefit of readily available mineralogical ex-
Table 1. "Asbestos" minerals as listed in Federal Register
(Listed minerals are italicized)
Asbestiform
Varierv
Chemical Composition
Non-
Asbestiform
Variety
SERPENTINE GROUP
Chrysotile MgstSijOjXOH),
AMPH1BOLE GROUP
Crocidolite Na;Fe3Fe^SigO:2)(OH.F);
Amosite (Mg.FeJ^SijOKKOH).
Antigorite.
lizardite
Riebeckile
Cummington-
ite—gruner-
ite
Anthophyllite
Tremolite
Actinolite
Anthophyllite
asbestos (Mg.FeJ^SigOuXOH.F);
Tremolite
asbestos Ca2Mgj(SieOsj)(OH.F)2
Actinolite
asbestos Caj(Mg.Fe)5(SisO:j)(OH.F)2
Simple error has created confused situation
It was this simple error which has been perpetuated by
other agencies and is now being expanded through pro-
posed regulations by both Osha and the Mining Enforce-
ment and Safety Administration (Mesa) that has created
the confused situation that exists today. This error was
further complicated by defining a "fiber" as any mineral
particle three times longer than it is wide (3:1 aspect ra-
tio), through reference to the Niosh criteria document.
These simplistic definitions of "asbestos" and "fiber"
were eagerly and immediately accepted by essentially all
other agencies as illustrated by the proposed and/or prom-
-------
ulgated regulations listed in table 2. It is obvious from
these documents that the majority of standard producing
Table 2. Initial Reaction. (Primary definitions covering asbestos
exposures)
Agency
Date
Definitions
Osha
USBM
(Coal)
EPA
(Air)
July 7, 1972 6 Minerals (See table 1)
(Revised Oct. Fiber = 3:1 Length: width ratio
18, 1972) = >5 Micrometers in
length
Nov. 7, 1972 Same as Osha
Apr. 6, 1973 Same as Osha
Asbestos material = Asbestos or
any material containing asbestos
Asbestos tailings = Any solid
waste product of asbestos mining
or milling operations which con-
tains asbestos
agencies are reluctant to develop definitions of their own
and are perfectly satisfied to incorporate those promul-
gated by others even to the perpetuation and expansion
of erroneous ones.
Bureau of Mines tried to clarify issue with symposium
The first major breakthrough in correcting the errors in
the original asbestos definitions came as the result of a
concerted effort by the Bureau of Mines to clarify the situ-
ation by holding a Symposium on Talc on May 8. 1973 in
Table 3. Turn about or Sidestep. (Secondary reactions
concerning asbestos definitions)
Agency
Date
Definitions
EPA
(Air)
May 3, 1974
USBM July 1,1974
(Metal-
nonmetal)
Osha Oct. 9, 1974
Nov. 21, 1974
EPA Oct. 25, 1974
(Air)
Mesa Dec. 13, 1974
Same as Osha
Standard revised to include only
"commercial asbestos," not opera-
tions where asbestos was a con-
taminant
Only includes asbestiform varieties
of six minerals. Recognizes exis-
tence of non-fibrous forms. Based
decision on data presented at
symposium on talc3
Agreement with talc producer7
Issued field memorandum8 using
five criteria for fiber identifi-
cation (see table 4)
Set exclusion level of one percent
asbestos content"
Issued memorandum requiring use
of criteria similar to Osha for
counting fibers
Washington, D.C.3 Numerous speakers presented data
on the medical effects of talc and associated minerals on
both humans and animals. Also, methods for the identifi-
cation and classification of these minerals were discussed
in detail.
Subsequent evaluation of all the presentations resulted
in the Bureau of Mines publishing in the Federal Register
of July 1. 1974 Amendments to the Asbestos Dust Stand-
ard.'* Included were mineralogically correct definitions
for asbestos and the recognized asbestos minerals (see
table 3). This document stands today as the turning point
and landmark in governmental regulation of minerals by
the realization that materials to be regulated must be cor-
rectly defined and described by the field of science to
which they belong if meaningful standards are to be devel-
oped. To quote Professor Tibor Zoltai of the Department
of Geology and Geophysics, University of Minnesota, iid
comments to the Minnesota Pollution Control Agency"
regarding the distinction between fibers of natural asbes-
tos and of cleavage fragments:
. Unfortunately, the misuse of some relevant mineral-
ogical concepts and terms are becoming so wide-
spread that it may be extremely difficult to correct
them. However, if that is not done, some mineralogical
concepts and expressions will have double definitions:
one for mineralogists and physical scientists, and one
for use in environmental public health sciences and
practices.
The continuing use of these double definitions would
be most unfortunate as it will undoubtedly lead to addi-
tional misunderstanding and conflict between mineral-
ogists and geologists, on one side, and environmental
and public health personnel on the other. As the con-
cepts and expressions are mineralogical. the logical so-
lution to this problem would be for the personnel in the
second category to restrict themselves to the proper
use of mineralogical terms.
On May 3, 1974, the Environmental Protection Agency
(EPA) issued its "Revised National Standard for Asbes-
tos Air Pollutant"6 in which the asbestos definition was
side-stepped by stating that the standard was meant to
deal only with "commercial asbestos" and that materials
"containing asbestos as a contaminant only are not cov-
ered." They specifically mention talc mines and manu-
facturing operations and the "releases of asbestos fro
taconite milling operations" as being excluded from cov-
erage.
Effort was made to correct error
In the meantime, through the diligent and consistent ef-
forts of a small group of interested parties. Osha became
aware of the errors in definition and the Department of
Labor issued an agreement on Oct. 9, 19747 with one in-
dustrial talc producer stating that "non-fibrous or non-as-
bestiform minerals such as non-asbestiform tremolite are
not within the scope of the standard . . ." Soon there-
after, on Nov. 21, 1974, Osha issued a Field Information
Memorandum8 recognizing the two forms of the amphi-
bole minerals and listing five criteria to be used in deter-
Table 4. Criteria to be used by Osha in counting fibers
A. Particles must appear to be fibrous rather than as crystals or
slivers.
B. The maximum diameter of a fiber to be counted is 3 mi-
crons.
C. The maximum length of a fiber to-be.counted is 30 microns.
D. The length to width ratio must be 5 or more to 1, that is. 5
times or more longer than wide.
E. The separate or individual fibers must contain fibrils or the
"bundle of sticks" effect, unless they are at a nondivtsible
stage. A fibril cannot be subdivided and would be counted,
if it meets the other criteria. The electron microscope ma;§
be used to prove the fibrous nature of the particles. The"
length to width ratio of 5 or more to I is not meant to impiv
that other particles are not hazardous.
mining whether a mineral particle could be considered as
fibrous or asbestiform. These are tabulated in table 4.
-------
Osha's recognition of the problem as evidenced by these
two documents and its repeated verbal assurances that
correct mineralogical terminology and identification
method would be incorporated into the asbestos amend-
ment appeared to signal a real and welcome change in
this agency's attitude toward the scientific correctness of
its standards.
Very soon thereafter, on Dec. 13, 1974, Mesa issued
"Health Division Instruction Memorandum No. 8—Defi-
nition of Asbestos Fiber for Tremolite Occurring in
Talc," which listed criteria for fiber determination similar
to those outlined in the Osha document. The year 1974
ended with a feeling of optimism that science had won
out and that future regulatory documents dealing with
"asbestos" would contain definitions and terminology
that would limit their application to the true asbestos min-
erals upon which all medical data concerning human
health hazards has been obtained.
Plan to rescind Labor Dept. agreement revealed
The first indication that all was not well was a thinly
veiled threat early in 1975 that Osha was considering the
rescinding of the Department of Labor agreement with the
before mentioned talc producer and the withdrawal of
Osha Field Information Memorandum #74-92. This ma-
neuver was based on misinformation supplied by a com-
petitor concerning the mineralogy of the talc products
in question and on continuing pressure from Niosh
t which, to this day, insists on using only the simplistic 3:1
length:width) criterion to define a fiber. A meeting was
*held with the then Assistant Secretary of Labor. John
Stender, with all principals represented, at which Stender
stated that as long as he held office both the agreement
and Field Information Memorandum would remain in
force. Stender left the Department of Labor soon there-
after.
Another document was brought to our attention late in
the summer of 1975 by Bureau of Mines personnel. This
was a study, commissioned by EPA. published by Bat-
telle, Columbus Laboratories entitled "Identification and
Assessment of Asbestos Emissions from Incidental
Sources of Asbestos."10 This publication deals primarily
with the occurrences of amphiboles, all lumped together
under the subtitle of "Asbestiform Minerals of Interest."
All amphiboles are grouped, regardless of crystalline
form, because, the authors state, the excellent prismatic
cleavage characteristic of all members of this mineral se-
ries (some 27 individual species) "causes the mineral to
fragment into fibers." This statement is directly contrary
to the principle, in general acceptance among mineralo-
gists, that mineral fibers grow as fibers and that fibers are
not generated from non-fibrous crystalline material by
fragmentation. This paper goes on to discuss essentially
all the major mining districts in the United States, state by
state, with the consistent inference that the presence of
amphiboles in any rock makes that rock a potential "inci-
dental source of asbestos."
Proposed amendment to standard a backward step
The shoe dropped on Oct. 9, 1975 (Black Thursday)
with the issuance of the Osha proposed amendment to its
Asbestos Standard (see table 5)." Not only did Osha pro-
pose to take the step backward and use the aspect ratio as
the sole criterion in defining a fiber but also to lower the
threshold limit value (TLV) from 5 fibers/cm3 (this TLV
was automatically dropped to 2 fibers/cm3 on July 1,
1976) to 0.5 fiber. In the preamble to this document Osha
continues to show its lack of mineralogical expertise by
Table 5. Back to go and worse. (Most recent reactions
concerning asbestos definitions)
Agency
Date
Definitions
Osha
EPA
(Air)
EPA
(Water)
Mesa
(Metal-
nonmetal)
Oct. 9, 1975 6 Minerals—No distinction as to
crystalline form
Fiber = 3:1 lengthrwidth ratio
= >5 micrometers in length
(negates field memorandum8)
Oct. 14, 1975 No change from original—details
disposal methods
Oct. 16, 1975 No change from original—groups
wollastonite with asbestos for re-
quired effluent controls
June 29, 1975 Removes •"asbestos" terminalogy
and substitutes "mineral fiber"
Mineral fiber = any mineral par-
ticle with a 3:1 aspect ratio or
more
>5 Micrometers in length
<5 Micrometers in width
copying an asbestos definition from a New York Acad-
emy of Science publication dealing with the biological ef-
fects of asbestos rather than consulting a more appropri-
ate mineralogical reference. The definition chosen was in
gross error in that it included all pyroxenes, in addition to
all amphiboles, and classified chrysotile as an example of
the pyroxene group, which it is not. This inclusion of an
additional group of minerals never before considered to
be asbestos indicates the never ending attempts by some
regulatory agencies to expand their jurisdiction indefinite-
ly-
The Osha publication was followed in quick succession
by two documents issued by the EPA on Oct. 14 and 16.
1975, dealing ,with "asbestos" contamination of air and
water, respectively.1213 The former details the method
of disposal of asbestos and any material containing more
than one percent by weight of asbestos. The latter further
modifies the effluent guidelines for mineral mining and
processing numerous minerals including a category en-
titled "Asbestos and Wollastonite." This expansion of
the inferred health hazard to a non-asbestos mineral ap-
parently slipped by unnoticed by the industry since no
counteracting comment has been discovered to date.
Mesa proposal would expand mining regulation
The latest in this long list of regulations ostensibly di-
rected at reducing or eliminating exposure of the worker,
consumer and the general public to "asbestos" is aimed
directly at the mining industry by its own regulatory agen-
cy. The proposed Mesa regulation.14 presented to the
Federal Mine Safety Advisory Committee in Birming-
ham. Al., June 29. 1976. would expand the asbestos regu-
lations to cover essentially all minerals known to man re-
gardless of composition.
Mesa would accomplish this very simply by removing
the term "asbestos," which everyone has such a difficult
-------
'ASBESTOS" MINERALS
Asbestiform Variety
Chrysotiie
¦\€ ar.y-
%f
l
ivv'*.>«>Km*N ;• v. A.-.v- , • - Sjf
Non-Asbestiform Variety
Cummingtonite
Jn-
Photomicrograph — 265 x
r \V ® -r
$..% 1^-
T="
Photomicrograph — 265 x
Amosite
Riebeckite
Photomicrograph ~ 265 x PWrtwrosrap/i - 26.5 x
-------
time defining, from the standard and replacing it with two
words, "mineral fiber." The agency would further simpli-
fy the regulator's task by defining a "fiber" as any "parti-
cle that exceeds 5 microns in length but not 5 microns in
idth, and shall have a length-to-width ratio of at least 3
o 1." While the terminology and definitions are altered,
the requirements for handling these "mineral fibers" re-
main the same as those promulgated for true asbestos.
To my knowledge, there is no mineral assemblage
being mined, milled, processed and either used as such or
discarded as tailings that would not contain significant
amounts of mineral particles which would be classified as
"mineral fibers" using the Mesa definition. The majority
of silicates, including those most common in mineral de-
posits (i.e. amphiboles, pyroxenes and feldspars), and nu-
merous nonsilicates have, by nature, cleavage character-
istics that cause them to elongate during comminution.
Nature itself forms particles meeting Mesa's criteria
through normal weathering processes.
Two points should be made quite clear at this time:
—All of the medical data concerning the nature and ex-
tent of health hazards related to asbestos exposure
has been gained on:
• truly asbestiform varieties of only four minerals,
chrysotile. crocidolite, amosite and anthophyllite
• groups of workers exposed to extremely excessive
amounts of these asbestiform minerals, usually as
mixtures of at least two types.
—No definitive medical data has been presented to in-
dicate that comparable health hazards exist with re-
spect to the non-asbestiform varieties of the true as-
bestos minerals. No references at all regarding expo-
sures to tremolite or actinolite, in any form, were
included in the list of 70 plus references listed in the
Osha Asbestos Standard and the additional 42 con-
taining "new information" in the proposed amend-
ment. The extrapolation of hazard from three or four
extremely rare forms of common minerals to the in-
clusion of cleavage fragments and particles of essen-
tially all other minerals is entirely unwarranted.
Ramifications to mining boggle the mind
There are many ramifications of the present and pro-
posed "asbestos" regulations, which would directly af-
fect the mining industry, that may not be immediately ap-
parent. As mentioned earlier, regulatory agencies show a
great reluctance to develop definition of their own and
eagerly adopt those supplied by others. The EPA is using
the 3:1 aspect ratio exclusively for the six listed "as-
bestos" minerals in its extensive study of water con-
tamination by mines, mills and manufacturing facilities.
At present, the EPA requires that any material containing
one percent or more "asbestos" must be disposed of by
covering with "at least 15 cm (ca. 6 in.) of non-asbestos
C. Sheldon Thompson is R&D Man-
ager-Minerals. Ceramics and Paper De-
partment, R. T. Vanderbilt Co. Thomp-
son began his professional career in the
field of mineralogy in 1955 as a professor
Westminister College in Salt Lake
ty. Following that, Thompson served
as research mineralogist-section leader,
Kennecott Copper Corp; and research
mineralogist-group manager, mineral
sciences. Union Carbide Corp. Mining
and Metals Division. Thompson has a
PhD from the University of Utah.
containing material ... at the end of each operating
day." Try to imagine the situation the mining industry
would be in when the EPA starts to apply the proposed
Mesa "mineral fiber" definition to the hundreds of thou-
sands of tons of mine tailings and overburden dumped
daily. Even worse, try to dream up a soil or cover of any
kind which would not contain any "mineral fibers."
Many state agencies try to improve on their federal coun-
terparts by setting more stringent standards. The Illinois
EPA, for example, requires that any "asbestos" con-
taining material be disposed of in steel drums.
Every mining company would be affected
All the potential problems need not be elaborated on
here. Everyone in the business of mining has been greatly
affected by one or more of the multitude of safety and
health regulations now on the books, but this latest "min-
eral fiber" proposal by Mesa is probably the only stand-
ard that affects every mining company in this nation. To
make it worse, it is based on a lack of medical data, a to-
tally unjustified extrapolation of hazard from fibers to
non-fibers and a complete absence of mineralogical under-
standing and expertise.
In conclusion. I would urge the entire mining industry
to support the efforts now being spearheaded by the
AMC Noncoal Occupational Health Committee to coun-
teract the ever growing umbrella of the "asbestos-miner-
al fiber" regulations. An AMC ad hoc subcommittee on
"mineral fibers" has already made a presentation to the
Federal Mine Safety Advisory Committee in Birming-
ham, resulting in the assignment of that topic to subcom-
mittee for study. Continued efforts will be made to pre-
sent any and all data which will aid all regulatory
agencies, federal and state, to correct past errors in min-
eral definitions and to make use of the vast reservoir of
mineralogical expertise in the U.S. Bureau of Mines and
the U.S. Geological Survey in all future standards dealing
with minerals. To accomplish this, the backing of the en-
tire mining industry will be needed.
Almost everyone knows what asbestos is. but there
seems to be a great many people in numerous regulatory
agencies that do not know what "asbestos" is not. Un-
less these people are enlightened, most everyone who
has not had "asbestos" in his past will find that he has
"asbestos" in his future.
References
'Federal Register. Vol. 37. No. 110. July 7, 1972. p. 11320-1132:
!Cntena For A Recommended Standard . . . Occupational Exposure
To Asbestos. National Institute For Occupational Safety and Health.
1972.
'Proceeding of the Symposium on Talc. Washington. D.C.. May 8,
1973. U.S. Bureau of Mines Information Circular 8639. 1974.
'Federal Register. Vol. 39. No. 127. July 1. 1974. p. 24316.
'Zoltai. Tibor and J. H. Stout. "Comments on Asbestiform and Fi-
brous Mineral Fragments. Relative to Reserve Mining Company Taco-
nite Deposits." Minnesota Pollution Controt Agency. Mar 24. 1976.
•Federal Register. Vol. 39, No. 87. May 3, 1974, p. 15396.
'Letter to H. B. Vanderbilt from John H. Stender. Assistant Secre-
tary of Labor, dated Oct. 9. 1974.
'Field Information Memorandum #74-92. Nov. 21. 1974. "Tremolite
and Talc." Occupational Safety and Health Administration.
•Federal Register. Vol. 39. No. 208. Oct. 25. 1974. p 38064-38073.
'"Kuryvial. R. J.; R. A. Wood, and R. E. Barrett. "Identification
And Assessment of Asbestos Emissions From Incidental Sources Of
Asbestos." EPA-650,'2-74-087. Battelle Columbus Laboratories. Sept
1974.
"Federal Register. Vol. 40. No. 197. Oct. 9. 1975. p. 47652-47665.
"Federal Register. Vol. 40. No. 199. Oct. 14, 1975, p. 48292-48302.
"Federal Register. Vol. 40. No. 201. Oct. 16, 1975. p. 48652-48668.
MAgenda for the 19th Meeting of the Federal Metal and Nonmetal
Mine Safety Advisory Committee, June 29-July 1, 1976. Birmingham.
Al.
-------
?4
Inductively Coupled Plasma -
Optical Emission Spectroscopy
-------
CS
Velmer A. Fassel and
Richard N. Kniseiey
' Ames Laboratory. USAEC and
Department ol Chemistry,
. Iowa State University, Ames, Iowa 50010
With this promising excitation
source for optical emission spec-
troscopy, metals and metalloids can
be determined at the ultratrace level
on or ng samples. Simultaneous
multielement determinations can be
achieved with minimal interelement
effects
Inductively
Coop!sd
Plasma-
Optica! Emission Spectroscopy
Combustion flames provide a re-
markably simple means for converting
inorganic analytes in solution into free
atoms. It is only necessary to intro-
duce an aerosol of the solution into an
appropriate flame, and a fraction or
all of the metallic ions in the aerosol
droplets are eventually converted into
free atoms. Once the free atoms are
formed, they may be detected and de-
termined quantitatively at the trace
level by atomic absorption, emission,
or fluorescence spectroscopic tech-
niques (7).
In spite of the success achieved by
these techniques during the past dec-
ade, many analysts have realized that
flames are a literal nightmare of vio-
lent chemical reactions, not ideally
suited for the purpose. This realiza-
tion led a number of investigators into
devising various electrically generated
"flames" or plasmas possessing higher
gas temperature and less active chemi-
cal environments, with the goal of im-
proving analytical capabilities while
retaining the precise control of sample
introduction and excitation provided
by combustion flames.
Another motivation for exploring
alternative means of generating free
atoms and releasing them into a reser-
voir suitable for spectroscopic obser-
vation was the increasing necessity of
performing simultaneous multiele-
ment determinations nt the nano-
gram/milliliter level for many trace
constituents, often on large numbers
of samples, and occasionally under
conditions for which sample volume
was limited. These requirements could
not be met then, nor are they being
met now by the widely used atomic
absorption spectroscopic (AAS) tech-
niques.
Although strenuous and imagina-
tive efforts to devise viable flame
atomic absorption or fluorescence sys-
tems for simultaneous multielement
determinations are being made by a
number of investigators (2), the task
of fitting these techniques into a mold
that is not natural to them is indeed
difficult. Some progress is being made
in adapting AAS or atomic fluores-
cence spectroscopy (Ar S) to automat-
ed sequential multielement determi-
nations (2), but true simultaneous
multielement analyses on a practical
basis by these techniques pose opera-
tional problems that seem destined to
remain unsolved for quite some time. -
In contrast to AAS or AFS, the obser-
vation of free atoms in emission does
not require an auxiliary primary
source. Simultaneous fnultielement
determinations are therefore possible
in a relatively simple manner.
To answer the question of whether.
electrically generated plasma would be
superior to conventional flames, arcs,
or spark discharges, in terms of pow-
ers of detection, precision, conve-
nience, and freedom from interele-
ment interactions, a number of explor-
atory paths huve l>fcen followed by dif-
1110 A • ANALYTICAL CHEMISTRY. VOL. 46. NO. 13. NOVEMBER 1974
-------
Report
ferenl investigators (3). It is beyond
the scope of this article to take a walk
down these various paths. Rather, at-
tention is focused on inductively cou-
pled plasmas (ICP)—a special type of
plasma that derives its sustaining
power by induction from high-fre-
quency magnetic fields.
The pioneering studies of Reed in
the early 1960's (4,5) and his develop-
ment of ingenious techniques for sta-
bilizing these plasmas set the stage for
a number of subsequent interesting
applications (6). Although Reed rec-
ognized that these plasmas offered at-
tractive possibilities for exciting spec-
tra, the initial analytical studies ap-
parently were undertaken indepen-
dently in two widely separated labora-
tories, one at Albright and Wilson in
Oldburg, England (7-9), and the other
at the Ames Laboratory, USAEC, at
Ames, Iowa (10-14). Later, a succes-
sion of workers in various countries
also explored their potential (J, 75-
20).
The INSTRUMENTATION' feature in
this issue of the JOURNAL provides
further details on the formation, ther-
mal isolation, and stabilization of
these plasmas; the injection of sam-
ples into the plasma: descriptions of
the spectra emitted: and typical ex-
perimental facilities employed for an-
alytical observations. In the remainder
of this REPORT, the performance of
typical analytical systems will be sum-
marized.
Detection Limits
Because of the high dynamic con-
centration range accommodated by
ICP-OES analytical systems, major,
minor, and ultratrace constituents can
be determined with equal facility. For
ultratrace applications, detection lim-
its that are measured experimentally
and reported on a riporous statistical
basis are useful figures of merit be-
cause they reflect the realities of life
with reference to analvte and back-
ground signals and their stability.
Table I shows a compilation of de-
tection limits observed with both
table-model single channel or large
multichannel spectrometers. The
range of values observed on the two
types of spectrometers rarely differed
by more than a factor of five if the
blaze characteristics of the gratings in
the spectrometers were compurnhle.
1CP detection limits reported inde-
pendently by Boumanns and deBoer
(20) and Souilliart and Rubin {IS) are
Table I. Comparison of Experimentally Determined Detection Limits
ICP, »s/ml • FUfnc.'pg/ml
Clamant
AU
6&deB*
S&Rf
AAS
AFS
AC
Ag
0.004
4
0.03
0.005
0.0001
0.008
Al
0.002
0.002
• • •
0.03
0.005
0.005
As
0.04
0.4
• « •
0.1
0.1
50
Au
0.04
• • •
0.04
0.02
0.05
4
B
0.005
0.03
0.03
6
• ••
30
Ba
0.0001
O.OOOC2
...
0.05
0.002
Be
0.0005
0.0004
• • •
0.002
0.01
0.1
8i
0.05
• . .
• • •
0.05
0.05
2
Ca
0.00007
0.00002
• a •
0.001
0.000001
0:0001
Cd
0.002
0.003
...
0.001
0.00001
0.8
Ce
0.007
0.002
0.03
0.5
10
Co
0.003
• • «
• • •
0.005
0.005
0.03
Cr
0.001
0.0003
a • a
0.003
o.ow
0.004
Cu
0.001
0.00Q1
• a •
0.002
0.001
0.01
Dy
0.004
...
0.009
0.2
0.3
0.05
Er
0.001
...
0.01
0.1
0.5
0.04
Eu
0.001
...
0.003
0.04
0.02
0.0005
Fe
0.005
0.0003
a a a
0.005
0.008
0.03
Ca
0.014
0.0006
a « a
0.07
0.01
0.01
Gd
0.007
...
0.01
4
O.OS
2
Ge
0.15
0.004
1
20
0.5
Hf
0.01
,,,
0.04*
8
100
20
Hg
0.2
0.001
• a •
0.5
0.02
40
Ho
0.01
• • •
0.01
0.1
0.1
0.02
In
0.03
...
...
0.05
0.002
0.003
La
0.003
0.0004
0.005
2
2
lu
0.003
. * •
0.01
3
3*"
0.2
Mg
0.0C07
0.00005
0.00005
0.0001
0.001
0.005
Mn
0.C007
0.000G5
a a a
0.002
0.002
0.005
Mo
0.005
0.0002
...
0.03
a.06
0.1
Na
0.6002
0.0003
0.002
0.0001
Mb
0.01
• a •
0.002
1
r"
0.C6
Nd
0.C5
a » •
0.01
2
2
0.2
Ni
O.C06
0.0004
» a a
0.005
0.003
0.02
P
0.04
0.07
• a*
• • •
• ••
Pb
0.008
0.002
0.01
0.01
0.1
Pd
0.007
0.002
• a a
0.03
• •a
0.05
Pr
0.06
• • a
0.03
10
1
0.07
Pt
O.OS
• la
...
0.1
2
Rh
0.003
...
...
0.03
ff.l
0.02
Sb
0.2
• ¦ a
...
0.1
0.05
0.6
Sc
0.003
• ••
• ••
0.1
0.01
0.01
Se
0.03
• • •
• ••
0.1
0.04
100
Sm
0.02
• aa
• • •
2
0.1
0.1
Si
0.01
• •«
0.1
••a
s
Sn
0.3
0.03
• •*
0.02
0.05
0.3
Sr
0.00002
• • •
a a a
0.01
0.01
0.0002
Ta
0.07
• ••
0.03
5
# m
20
Tb
0.2
• • a
0.02
2
0.5
0.03
Ta
0.03
...
...
0.1
O.OS
200
Th
0.003
..
200
Ti
0.003
0.0002
0.001
0.09
O.l"
0.2
Tl
0.2
• • •
a aa
0.03
0.008
0.02
Tin
0.007
• • •
0.01
0.2
0.1
0.02
U
0.03
• ••
• aa
^ a .
10
V
0.006
0.0002
0.001
0.02
0.07
0.01
W
o.oo;
0.001
0.1
3
...
0.5
Y
0.0002
0.00006
0.0005
0.1
..."
0.04
tb
O.C009
0.00004
0.005
0.04
a.oi"
0.002
Zn
0.002
0.016
0.05
0.002
0.00002
50
Zr
0.005
0.0004
0.005
5
• • a
10
• Amps Laboratory, USAEC. Iowa State University. Detection limits 'eor«s*nt concen.
(rations r**ou»rc<> to produce .1 Imo su'n.il twice as i*rc.it .is the jtanoa'on limits jrriort»»fi on t^e s.ime b.isis as
Ai^ic*. Ijbor.itotv ' Kef. f.s. limnsr^prcsrnl ronce"trat*ons required to pro-
duce j lint* sicn.il so times as «^ejt ns tne stjnaard deviation of tne b ich^rouna scatter
(none). * Aft flame detection hints .ire taken from rets, it and it. * No value repotted.
ANALYTICAL CHEMISTRY, VOL. 46. NO. 13. NOVEMBER 1974 > 1111 A
-------
also shown. The statistical criteria on
which these experimental values are
based are noted at the bottom of the
table. For comparison, the most recent
compilations of detection limits ob-
served by flame atomic absorption
(AAS), emission (AE), and fluores-
cence (AFS) are also included.
A definitive appraisal of the collec-
tion of the data in Table I should rec-
ognize that the data for the flame
techniques represent the best values
reported for a variety of primary
sources used in conjunction with a va-
riety of flames, burning at various op-
timal stoichiometrics, and observed at
selected optimal observation sites. In
contrast, the Ames Laboratory detec-
tion limits were observed under iden-
tical experimental conditions, and for
the Boumanns and deBoer data, only
the observation height and argon flow
rate were optimized. In general, pow-
ers of detection can rarely be im-
proved by more than a factor of five
by optimizing height of observation.
The ICP therefore possesses virtu-
ally ideal characteristics for simulta-
neous multielement determinations,
because one set of experimental pa-
rameters provides essentially optimal
conditions for all of the metals and
metalloids. It is unlikely that combus-
tion flames will ever achieve this uni-
versality. Because flame AAS powers
of detection are commonly used
benchmarks, it is interesting to direct-
ly compare the AAS ana ICP-OES
techniques. This comparison is made
in Table II on both a relative and ab-
solute concentration basis, for the
Ames Laboratory ICP values only. It
is evident that there are more ICP
than AAS entries at the low concen-
tration columns, which eventually
leads to no ICP entries at the highest
concentration bracket. Thus, integrat-
ed over the periodic table, the ICP-
OES technique possesses superior
powers of detection on a true simulta-
neous multielement basis.
A valid question, often asked, is
whether detection limits observed
under idealized conditions are trans-
ferable to real-life situations. The data
taken from a more complete study re-
ported elsewhere 123), as summarized
in Table 111. provide documentation
that the ICP-OES values are transfer-
able to a high degree. There is no sig-
nificant deterioration in powers of de-
tection in the presence of an iron ma-
trix. Some typical quantitative results
may also he used to demonstrate the
transferability of the detection limit
-Table II. Comparison Inductively Coupled Plasma-500 ppb
ng>
0.Z-Z ng<
2*20 ng«
2&-100 nyi
> 100 ng*
ICP»
AAS<
ICP
AAS
ICP
AAS
ICP
AAS
ICP AAS
Ba
Ag
Ag
Be
As
Al
Ge
As
B
Be
Mg
Al
Ca
Au
Au
Hg
Dy
Ce
Ca
B
Cd
Bi
Ba
Pt
Er
Gd
Mg
Ce
Co
Ce
Bi
Sb
Hg
Ge
Mn
Cd
Cr
Ga
Er
Si
Ho
Hf
Na
Co
Cu
Hf
Eu
Sn
Pt
La
Sr
Cr
Fe
Ho
Ga
Tb
Sb
Lu
Y
Cu
Mn
In
Ho
TI
Sc
Nb
.Yb
Dy
Na
Mg
In
Se
Nd
Is
Ni
Nb
Mo
Si
P
Eu
Zn
Nd
Pb
Te
Pr
Fe
P
Pd
Tm
Sm
Gd
Pr
Rh
Y
Ta
La
Pt
Sn
Tb
Lu
Se
Sr
Th
Mo
Si
TI
U
Ni
Sm
TI
W
Pb
Ta
Tm
Zr
Pd
Te
V
Rh
U
Yb
Sc
Th
Ti
Tm
V
W
Zn
Zr
* In 200 (0.2 ml).' Ames Laboratory values only.« Best values reported in refs. II and
data. The measured detection limits
for Ti. Ce. Nb. and V are 0.003,0.007,
0.01. and 0.006 ng/m\, respectively
(Ames Laboratory values in Table I).
If the reasonable assumption is made
that a signal at least ten limes greater
than the standard deviation of the
scatter in the background is desirable
for quantitative determination, then
the detection limits should be multi-
plied by a factor of five. For a 0.5 wt %
solution of steel samples, the calculat-
ed determination limits for Ti, Ce, Nb,
and V would then be 0.0003,0.0007,
0.001. and 0.0006 wt %, respectively.
The analytical curves in Figure 1 dem-
onstrate that the calculated predic-
tions are indeed sound and that the
measured detection limits are trans-
ferable. For all four constituents, the
lowest calibrating points either fall
below or near the calculated determi-
nation limit.
Interelement Effects
The relatively high temperature
and long residence times experienced
by the sample species and the inert
environment provided by the plasma
support gas lead to the expectation
that, in comparison to combustion
flames, solute vaporization interferen-
ces should be vastly reduced or nonex-
istent, and the degrees of atomization
should be more complete, if not ap-
proaching 100%, for all metals and me-
talloids. Because the free atoms arc re-
leased in a noble gas environment, de-
Table III. Detection Limits Mea-
sured in Absence and Presence
of Fe Matrix Material
Clement
H:0
solution,
»9/ml
0.S% Fe
solution,
»S/ml
Al
0.G02
0.004
Ce
0.007
0.013
Cr
0.001
0.002
Cu
0.001
0.0003
La
0.003
0.003
Mn
0.003
0.003
Nb
0.01
0.02
Ni
0.05.
0.07
Pb
0.003
0.006
Pr
0.0G3
0.01
W
0.002
0.003
Zr
0.02
0.02
population processes such as metal
monoxide formation should also be
minimized. Thus, the released free
atoms may be viewed as independent
radiating species.
These favorable environmental fac-
tors combined with the prevailing high
temperatures should^ in turn, over-
come most interelement or matrix in-
terference effects found in many
flames, arc. and spark discharges.
These expectations have been con-
firmed by some preliminary observa-
tions on several classical interelement
interference systems and in systematic
studies now in progress. For example,
the analytical studies of Greenfield et
al. (7) ar.d Wendt and Kassel 13, JO)
have shown that the depressing effect
on the free-atom furmatiun of calcium
1116 A • ANALYTICAL CWMISTHY, VOL.
46. NO. 13, NOVEMBER 1974
-------
»—ippbehm ¦¦1 innv
100
•tytj
\
J
.'i
-*
1
\
a
cc
5
o
»-
o
X
a.
IO
Figure 1. Analytical
calibration curves for
determination of Ti,
Ce, Nb, and V in
steels
01"—
QOOOI
0.001 0.01
MPllftlTT CONCENTRATION IN IRON (Wl%)
OJ
in. the presence of increasing concen-
trations of PO.,3- or Al3+ ions was re-
duced to virtually negligible propor-
tions in the plasma. Some interele-
meat effects may occur as a result of
major changes in the concentration of
matrix materials. The degree to which
these interelement effects may occur
depends on a host of experimental
variables, such as height of observa-
tion, argon carrier gas flow rates, and
power input to the plasma {24).
The interference effects that are ob-
served appear to be the result of a
combination of factors, including
changes in the degree of ionization lat-
eral diffusion of the analvre, and to
nebulization efficiency. One of the
pleasing observations made so far is
that observable interelement effects
are Teduced to negligible or tolerable
proportions under the same experi-
mental conditions that lead to excel-
lent powers of detection. For example,
the Ca neutral atom line undergoes
only a 2-3% enhancement, and the Ca
ion line undergoes only a ~lo% de-
pression as the Na concentration in
the solution is increased from 0 to 0.7
wt % (24). Thus, the neutral atom line
of Ca undergoes only a 2-3% change in
signal for the analytical equivalent of
determining Ca in distilled water or a
Na matrix.
Analysis of Microliter
or Microgram Samples
For studies on the biomedical and
nutritional effects of trace metals and
in clinical and forensic analyses, it is
often necessary to determine a num-
ber of elements in microliter or micro-
gram samples. The deta summarized
in Table II have already demonstrated
the exceptional powers of detection
observed on 200-^1 samples. The con-
cept of applying conventional pneu-
matic neoulization o: microliter sam-
ple volumes and introducing the
"plug" of aerosol into the ICP for the
simultaneous determination of many
trace metals was first reported by
Greenfield and Smith {'25), and subse-
quently, Kaiselev et al. (26) applied
this technique to the determination of
trace metals in blood and serum.
Although Kniseley cl al. found that
samples of whole blood could be ex-
aminer! directly, tenfold dilui ions with
dilute HC1 were commonly used.
Serum samples were examined direct-
ly or diluted with dilute HC1. Typical
analytical curves established by the
"standard additions" method are
shown in Figure 2. When 25-jd ali-
quots of human blood diluted 1:9 with
0.1 A/ HC1 are nebulized, the signals
observed for Cu 324." nm during the
passage of the aerosol cloud through
the.plasma are shown in Figure 3.
Analogous recordings of the Mn
403.0-nm line, when 25-^1 samples of
whole blood (undiluted and contain-
ing 50 ng/ml of Mn) were nebulized,
are shown in Figure 4.
One of the -aost sensitive contem-
porary analytical techniques is based
on the thermal atomization of microli-
ter volumes of a sample in gTaphite
furnaces, or from graphite or refracto-
ry filaments. and loops, followed by
the observation of the free atoms
formed in either atomic absorption or
fluorescence. Although exceptional
relative and absolute powers of detec-
tion ana acceptable reproducibility
have been achieved by a multitude of
variations of these atomizatioD sys-
tems, this technique is subject to a
rather extensive list of experimental
constraints which were recently sum-
marized 127). The combination of
thermal vaporization of samples fol-
Flgure 2. Standard addition analytical curves obtained from
25-jjI aliquots of human blood of plasma (tenfold dilution of
whole blood and twofold dilution of plasma)
c
c
120
O
»¦>
o
HO
*
c
100
—¦ ¦ /
2
•
90
"* / O y
>
»•
eo
/
z
60
2
/ ,
US
>
40
<
-J
AC
^ i i f
l t ¦ » i
-Ol 0 jOI 03 J03 045
Vi-.JUUESE Cr>.;£MRATiC% (,.J ,r.)
ANALYTICAL CHEMISTRY, VOL. 46, NO. 13. NOVEMBER 1974 . 1117 A
-------
2^
NORMAL
BLOOD
ABNORMAL
SLOCO
Figure 3. Signals obtained for Cu 324.7
nm for 25-^1 aliguots of human blood di-
luted tenfold. Abnormal sample came
from leukemic patient
fj
vv
J
f
y J w
A /nA
lM
Figure 4. Repetitive recordings of Mn 403.0 nm for 25-j/l samples of undiluted blood.
Blood sample contained 50 ng/ml (a = ±5%)
lowed by ICP atomization and excita-
tion of the vapor offers the twofold
promise of performing uhratrace de-
terminations on a multielement basis
on microliter or microgram-sized sam-
ples and overcoming some of the limi-
tations noted in the AAS or AFS liter-
ature.
First, since the free atoms are actu-
ally generated in the plasma, a single
set of parameters should suffice for
the vaporization of many types of
samples. Second, interelement inter-
ferences arising from recombination or
nucleation of the vapor above the fila-
ment should be minimized because
the plasma subsequently achieves at-
omization of the vapor cloud. Third,
background interferences from the
filament or furnace tube do not exist.
Fourth, analytical curves obtained
from the toroidal-shaped, inductively
coupled plasma are commonly ob-
served to be linear over a concentra-
Whatever you can do with present
xray fluorescence analysis techniques,
you'll do it faster, cheaper, and more accurately
with the EDAX^EXAM* System.
The EDAX/EXAM System is the latest state of
the art technique in xray analysis. Its energy
dispersive (EDAX) operating pnnciple
gives you simultaneous analysis and display of
all elements from Sodium through Uranium.
Qualitative analysis in seconds, quantitative
in minutes. Detailed brochures available
from EDAX International. Inc., PO. Box 135,
Prairie View. 111. 60069.
EDAX
ClRClt 68 ON REAPER SERVICE CARD
111$ A • ANALYTICAL CHEMISTRY. VOL 46. NO. 13. NOVEMGER 1974
-------
)tion range of 4—'i order* of magnitude
I2.t) if appropriate linenr rendout
, capabilities are utilized. In a prelimi-
nary study, detection limits observed
for one set of operating conditions
were highly competitive overall with
the best values observed under a vari-
ety pf experimental conditions for the
nonflame AAS or AFS techniques
127).
Summary
The authors agree with the recent
conclusion of Boumanns and deBoer
{20) that the argon ICP-OES analyti-
cal system "constitutes a most promis-
ing excitation source for simultaneous
multielement analysis of solutions."
Although relatively few laboratories
have taken advantage of the unique
capabilities of this excitation source,
the demonstration that ail of the met-
als and metalloids can be determined
at the ultratrace level, on vl or tig sam-
ples on a simult-'-neous basis with min-
ima] interelement effects, portends in-
creasing application of this excitation
technique to a variety of analytical
problems.
Reference*
(II Review* on thr*e technique* m«v he
fiHind in J. I). Winefuron. Anal. Chem.. 45, 712A
C19T.TH D. G. Mitchell. K. W. .Jackson,
and K. M. Aldmis. ibid., p 121-SA.
(31 V. A. Fassel. "Electrical Hl««ma Spec-
troscopy." XVI Colloquium Spectroscopy
icum Internationale. Adam Hifcer. Lon-
don. England. 1973.
<41 T. B. Heed. J Appl. Phvs.. 32.821,
2534(1961).
(5V T. B. Reed. Int. Set. Technol., <2
(June 1962).
(6) H. L\ Eckert. High Temp. Sci., 6,99
(19741.
(7) S. Greenfield. I. L Jones, and C. T.
Berry, Analyst. 89,713(1964).
(Si S. Greenfield. Proc. Soc. Anal. Chem..
2,111 (19651.
(9) S. Greenfield. P. B. Smith, A. E.
Breeze, and N. M. D. Chilton. Anal.
Chim. Aero. 41, 38o (196S1.
(10) R. H. IVendt and V. A. Fassel, AnaL
Chem.. 37,920(1965).
(11) R. H. Wendt and V. A. FasseL, ibid.,
3&, 337 (1966).
(12) W. Baraett and V. A. Fassel. Spectro-
rhim. Ada. 2.(i», i>4.i l i*«w. —
(13) V. A. Fn-'-cl nml ("i. W. Dickinson,
Antil. Chen . JO. 247 (1%8».
(J4) ('¦. \V. DirVin-nn and V. A. Fassel,
ibid. 41, 1(121 (lWO).
(I.M H. M. IWmill. I). J. Smith. T. S.
West, and S. ('.n-rnfield. Anal. Chim
Acta. S4.W7 (1*171).
(U!) O. K. Kirkliriclil. A. F. Ward, and T.
S. Went.ibid. 62.241 (1972).
(17) (•. F. KirLhriL'ht. A. F. \V ard, and T.
S. West, i/iic . 64, 353 11973).
I If) J. C. Souilliart and J. P. Robin. Anal-
ysts. 1,427 (1972).
(191 P W. J. M Boumanns, F. J. deBoer,
and J. W. Raiter. I'hitips Tech. /fee., 33,
•M> (19731.
(20) P. W. J. M. Boumanns and F. J. de-
Boer, Spccfochim. Acta, 27B, 391
(1973).
(21) J. D. Winefnrdner. V. Svoboda, and I*
J. Cline. "CKC Critical Reviews in Ana-
lytical C hemistry." Aupust 1970.
(22) N. Omeneito. L. M. Fraser, and J. D.
Winefordner. Appl. Speetrosc. Rev., 7
(2). 147 (1973).
(23) R. N. Kniselev. C. C. Butler, and V. A.
Fassel. Anal. Chem., submitted for pub-
lication.
(24) G. Larson. V. A_ Fassel, R. H. Scott,
and R. N". Kniselev, ibid.
(25) S. Greenfield and P. B. Smith, AnaL
Chim. Acta. 59. 341 (1972).
(26) R. X. Knifeley, V. A. Fassel. and C. C.
Butler. Clin. Chem., 19,801 (1973).
(27> D. E. Nison. V. .V FasseL and R. N.
Kniseley, Anal. Chem., 46,210 (1974).
r3
vii .jr
- 31 |
J]
ii
i-i
. !
Velmer A. Fassel is the deputy director of the Ames Laboratory, USAEC and
the Energy & Minerals Resources Research Institute, Iowa State University. He
holds these positions concurrently with his academic position of professor of
chemistry. He earned his BA degree from Southeast Missouri State University
and his PhD (in physical chemistry) from Iowa State L:niversiry. Dr. Fassel is the
author of 136 publications on various aspects of atomic emission and absorption
spectroscopy, molecular spectra and structure, and high-temperature analytical
chemistry. He has received five major awards: the Annual Medal Award of the
Society for Applied Spectroscopy in 1964, the Spectroscopy Society of Pittsburgh
Award in 1969. a special zoic medal presented by the journal Spectrochimica
Acta in appreciation for his sen-ice as coediror for 13 years, the Hasler Award
(1971). and the Anachem Award (19711. He has been a member of the IUPAC
Commission on SpectrochemicaJ and Other Optical Methods of Analysis since
195S and now holds the position of Chairman of this Commission. In 1967 he was
appointed to the Joint Commission on Spectroscopy. International Council of
Scientific Unions. The Optical Society of America and the American Association
for the Advancement of Science have elected Dr. Fassel to Fellow membership.
Richard N. Kniseley received his BA degree from the University of Missouri at
Kansas City and his MS and PhD degrees from Iowa State University. He has
been associated with the Ames Laboratory, USAEC. at Iowa Stale for the past 24
years. His research interests encompass a broad area, including atomic emission
and absorption spectroscopy, electron microprobe analysis, high-temperature
spectroscopic source, and molecular structure. He has authored or coauthored ap-
proximately 60 publications in these area.t and has been an invited lecturer at nu-
merous national and international conferences. He was a National Lecturer for
the Society for Applied Spectroscopy in 1967 and 1974 and an ACS tour speaker
in 1972.
1120 A • ANALYTICAL CHEMISTRY. VOL. 46. NO. 13. NOVEMBER 1974
-------
Hi
Simultaneous Multielement Analysis of Liquid Samples
By
Inductively Coupled Argon Plasma Atamic-Bnission Spectroscopy
-------
u>
SIMULTANEOUS MULTIELEMENT ANALYSIS OF LIQUID SAMPLES
BY
INDUCTIVELY COUPLED ARGON PLASMA ATOMIC-EMISSION SPECTROSCOPY
AT
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION V
CENTRAL REGIONAL LABORATORY
1819 W. PERSHING ROAD
CHICAGO, ILLINOIS 60609
-------
4^
The mention of trade names or commercial products does not imply endorsement
by the Environmental Protection Agency or the Central Regional Laboratory.
-------
44
1.0 Scope and Application
1.1 This method is applicable to fresh water samples including drinking
waters, surface waters, domestic and industrial waste_gfflugntsfor the analyses
of total and dissolved metals of the twenty elements listed in Table 1.
1.2 The detection limits for the 20 elements are listed in Table 2. These
data represent the mean values determined on 5 days, over a 3 month period. Each
individual detection limit was determined by averaging 10 consecutive 10 second
exposures of each element. The numerical values are in ug/l and each is the con-
centration of that element necessary to produce a signal twice the standard devia-
tion of the background noise.
The two sicma detection limit, defined above, is in common use and pro-
duces a figure of merit by which varicus analytical appproaches may be compared,
but, like most detection limits, are almost never at a level that can be accurately
reported. A more useful approach, which is similar to the Lower Optimum Concen-
tration Range (IOCS) reported for atomic absorption use^, is the Lowest Quantitative: ¦'
m
Determinable Concentration (LQ0)~. The LQO is defined as the amount of material
necessary to produce a signal that is 10X the standard deviation of the noise
(i.e. 5X detection limit). The LQD for the 20 elements are also reported in Table Z.
1.3 The relative standard deviation (RSD) of a 1 mg/l standard for all ele-
ments is of the order of 1% over 5-10 minute periods and 2-^* over the course of
an 8 hour day. A typical set of RSDs is presented in Table 3 and demonstrates
variations for 10 consecutive 10 second integration periods.
1.4 The working ranges for all twenty elements are from the LQO tc above
100 ppm. An illustration of these ranges are presented in Table 4, the value re-
ported is the average of four measurements made on the same day over an 3 hour
period. In trie case or calcium two elemental lines are simultaneously employed
and the linear dynamic range is increased to at least 1 COO rag/1. for all elements
-------
the 100 mg/1 upper limit doesn't represent the actual upper limit of linear instru^
mental response for any element. Instead the limiting value indicates to an oper-
ator the need for careful judgement to insure that overlapping effects from other
elements are absent. (See Section 4) But this working range does define sample
concentration values normally reported in day-to-day operation.
1.5 For those procedures and applications described in this manuscript, the
degree of operator skill necessary to perform analysis using an ICAP-AES is similar
to that required for operation of an atomic absorption spectrometer (AAS). However,
the initial set up of an installation, reviewing of data and routine instrumental
problem evaluation would probably still require ready access to an experienced
spectroscopic.
1.5 Approximately 50 chested or otherwise prepared samples and 10 quality
control samples can be run per hour. At this sample run rate a realistic analysis
output of between 5000 and 8000 analysis per man day can be expected. These huge
numbers assume all twenty elements are requested by the sample originator. This
is almost never the case. Furthermore, thousands of analysis per day assume no
sample pretreatment. In fact the majority of sample handling time is associated
with sample digestion. This situation will not change with the implementation of
the ICAP-AES. Lastly, the transfer of -ompleted analyses from the teletype s.neet
to the report form currently being used will be no faster with the plasma system
than it is new and constitutes a significant fraction of the operator's time.
2.0 Summary of the Method
Liquid samoles are aspirated into a high temperature argon -plasma produced
by Inductively coupling radio-frequency electromagnetic radiation to the argon
gas.3 The high temperature4 of the plasma (— 10,000°K) causes desolvation, molecu-
lar breakdown, atomitation, and/or ionization and excitation of the metals in
solution. The resultant radiation produced as the excited atoms relax is passed
-------
through the entrance slit of a dispersive device where it is separated into discrete
wavelengths. The intensity of each of the characteristic wavelengths is associated
with a metal and is measured by a photomultiplier tube. The photocurrent is trans-
formed, by reference to standards, to concentration values which are recorded.
3.0 Sample Handling and Preservation
3.1 General
- , Samples are collected, filtered (for dissolved metals), preserved and
a /•«.*. ,
digested according to approved EPA procedures.1
3.2 Sample Hand! ir.g
Samples are ccV=c~=d in polyethylene containers, preserved with 1 C~ 1
of 5C" HNO^/1 and sealed with plastic caps, containing a polyethylene insert.
. ~AT$
For total metal analysis- samples are digested by the addition q- 3 ml
of conc. HNC3 to 100 ml of representative sample in a Griffin beaker. The beaker
is covered with a ribbed watch glass and carefully heated to dryness cn a hot plate.
After coolino, 3 ml of conc. HN0-; is added and the beaker is heated until the acid
** w
mixture is brought to a gentle boil. At this point, to improve sample disolution,
5 ml of 1:1 HC1 is added to the warm HNO3 mixture and the volumn ^s orougnt to
100 ml with distilled deicnized water.
4.0 Interferences
4.1 Introduction
The ICA? is in many ways an ideal excitation source. The 10,000°K
operating temperature is twice as hot as any chemical flame and therefore produces
significantly larger numbers of excited atoms. Further the plasma cas, a-gon, is
chemically inert and spectroscopics!ly simple. The combination of these character-
istics arrorc the ICAP an interference free existence when compared with flames,
arcs and sparks . Still there are some concerns that must be considered in the
installation ana ODeration of iCAP soectrcmeter system. These are discussed below.
-------
4.2 Chemical Interferences
—————^ ^
The ICA? is virtually free of the classical chemical interferences ' assoc-
iated with both f 1 aire AAS and flame atomic emission spectroscopy (AES). Molecular
formation among solution or gaseous atoms exposed to 10,000°K is not likely and
has not been observed. Carrier gas induced chemical reactions are also very small
because argon is inert.
4.3. Ionization Interferences
Changes in emission intensities due to the presence or absence of easily
ionized material (i.e. 'la, La) has not been reported5'7 nor does this work indi-
cate such problems exist.
4.4 Positional stability of radiation source
The positional stability of tne ?]as~;a is a very important consideration
because very small changes in plasma location can easily cause significant changes
in analytical results. A partial list of variables which would cause the position
or the area of the plasma to chance is as follows:
1. '•'.ovemsnt of plasma torch
2. Movement of ccuol'ng coils
3. Geometry cf annular sample cnannel
4. Variations in either forward or reflected RF power
5. Gas pressure
6. Gas flow
?. Gas purity
8. Chances in inductance couoling
9. Solution viscosity changes
10. Changes in sample uptake race
11. Unknown sources
In the thousands of analyses so far performed, none of these problems
have proved to be significant. However, because these problems can be present,
and can appear at any time, a comprehensive quality assurance audit orcced-re "'s
induced as a part of this metnod. The audits run to dace shew excellent Icr.c
term precision and ca;ly precision of a l.j mg/1 solution for all elements to be
less tnan .vicn no restancarj'tac'cn over a 10 nour period.
*
-------
4 7
4.5 Unwanted Light Interferences _
There are two classes of light interferences which may affect each
channel of the instrument. They are:
1. Direct spectral overlaps.
2. Indirect and/or stray light additions.
In emission spectroscopy the possibility of two or mere elements having
w
overlapping lines is an important consideration. In a direct reader this spectral
type of interference is predictable and can be empirically identified by measuring
the highest reported concentration of a particular element in the absence of a":
other elements. The effect cn all other elements is thus rscorced. treble's
are observed these data can then be factored into tne calculations.
A more uncommon type of direct spectral interference happens when two
_lln.es.. are extremely close but do not normally overlap unless one of the elements
is extremely concentrated. The line broadening or reversal problem is identified
in the same manner as a direct spectral interference and can be controlled by =
careful specification of linear dynamic ranges.
The second class of light interference addresses itself to the observation
that any extraneous lighf3 introduced into the system may create poor results.
The most serious problems are associated with the introduction of unwanted lignt
which is unique to tne sample and not the blank. For example an emission resulting
from a concentrated sample might reflect from some shiny surface within tne direct
reader and cause an error. Any photomultiplier tube which detects any of this
"stray" light will read that intensity as element intensity and therefore concen-
tration. Stray problems of this nature stem from three principle sources:
1. An imperfection in grating.
2. A reflective surface is exposed inside tne spectrometer.
3. A mechanical breakdown or lack of some internal light shielding.
£
-------
In those cases where it has been determined that spectral corrections of the type
stated above are needed, they have been made. Generally speaking these corrections
are small, apply to only a few elements, and are easily verified.
5.0 Apparatus
5.1 Direct Reader
The Jarrell-Ash Plasma Atom Comp 750 equipped with exit slits listed
in Table 1 comprises the basic direct reader. Each slit is observed by a photo-
multiplier (?M) with associated power supply and picoampmeter. The details of
the spectrometer are included in J-A Manual Model 75 Atom Ccmp section A1 and 31.
5.2 Plasma Assembly
- 1C ii
The plasma assembly is composed or a nebulizer , spray chamber , torch
coupling box and power supply. The power supply and coupling box are supplied
by the CCA Corporation and Henry Radio Corporation, respectively, through Jarrell-
Ash Division of Fisher Scientific Company. The nebulizer is a right angle pneumati
aerosol generator constructed of teflon and glass (with external brass fittings).
The plasma torch is all quartz and of standard design which separates coolant argon
plasma argon and accepts nebulized solution. The spray chamber is modeled after
the conical chamber described by Fassell0
5.3 Argon Gas
Argon is supplied from a liquid tank (Linda 5?45], The use of liquid
is preferable because it is cheaper, less labor is involved in moving tanks and
the quality of argon is superior to that from low grade gaseous tar.Rs.
5.0 Reagents
5.1 Water and Acids
Distilled deionized water is used in all cases. All acids are analytical
grade.
-------
6.2 Standards
Standards are prepared monthly (o- as needed) by the addition of Q.SOCml
of 10C0 mg/1 stock solutions of all metals in Table 1 (except Ag which is handled
separately] to a 500 ml volumetric flask. The solutions (except Ag) are acidified
with 25 ml of 1:1 HC1. Silver is prepared daily in 52 HNO3. A blank is prepared
and run with reference to distilled water. Generally, no differences between the
prepared blank and distilled water have been observed and therefore samples are
run by reference to distilled water.
7.0 Procedure
7-1 Turr. on arrc" z-.i
7-2 Turn cn cooling water
7-3 Turn or, R-F power supply (10 rain, warm up) • * * •
7-4 Ignite plasma using Tesla coil
7-5 Adjust forward pov;er to i.GKW
7-5 Adjust reflected power to less than one W.
7-7 Turn on teletype
7-3 Adjust entrance refractor plate such that Hg~=13 reads a maximum on
profile meter.
7-9 Choose a Basic Cata Set to be used and set t">s and date.
7-10 Witn plasma light blocked measure the dark currents of all channels to-
be used.
7-11 With plasma light blocked measure the white light response of syste.fi.
7-12 Uncover plasma, redo 7-5 and record in intensity units blank water be'"r.
aspirated.
7-13 Use the two pcir.t standardiration procedure.
a. List standards needed - coromand L
b. Clear computer memory - comanc J
c. Run each standard ir, duplicate - command EGG
d. Record eacn standard - command r<
-------
Si
Standardize spectrometer for all standards - conmand S
Record gains and offsets of standard curves - command W
Run (EGC) and record the 10C0yg/l standard. This standard
should read 1 COO t 2% for all elements. If any element does
not read within the specified range repeat sequence 7-13.
7-14 Create the operating commands for the days work (i.e. QEGIGIAC will run
a sample twice and print both runs in raw intensity and average both
runs and print the average in concentration units).
7-15 Run the 1000ug/l standard using the operating command for that day.
7-15 Run the blank used. Check against distilled water.
7-17 Run a sample - allow at least a 50 second wash.
7-18 Every 1/2 hour or.20 samples rerun 1QC0 'jg/l standard.
Every 1 nc-r or £0 same;as rerun "CCC .g/" star.ta'-d
7-19 Curing every day of operation collect in addition to the da ;a called
for above:
a. A comolete 10 x 10 sec. evaluation of dark current.
b. A complete 10 x 10 sec. evaluation cf white light.
c. A complete 10 x 10 sec. evaluation of detection limits.
d. Compile anG update the data collected in 7-13.
8.0 Calculations
8.1 Basic Approach
Intensity data for tr.e blank and the mixed 1000uc/l standard are stored
in the computer after being entered by procedures outlined in section 7. A value
of "zero" is assigned to the intensity associated with the blank and a value of
ICOOug/l for the intensity corresponding to that solution. By reference to these
data the computer then assigns concentration values to unknown intensities. The
linear range and validity of this approach can be seen in Table 4. Background
corrections are made by reference to the standard Dynamic Background Correction
package supplied by the vendor. The use cf an Internal Standard;*} is a'so a
possibility but none were usea in these comparisons because no imorove.^ents i?
results have been observed.
e.
f.
g.
8
-------
8.2 Readout of Computation
The photocurrent output is frpnnr.-iH y sannla-i f^n/co^ *nH
by a PDPSM computer. The computer can print intensity (I), ratioed intensity (S)
baseline corrected intensity (3) or concentration (C) using a standard SR33
teletype. The format cf the output and other details are defined by a Basic
Data Set. All pertinent details necessary for the creation and use of the Basic
Data Set are defined in Operation Instructions for 8K Mark II System (C01-08P-3001
08T, C8D, C83).
9.0 Comparab-i : i ty Data
z. t nccur:.cy
The accuracy for 15 of the 20 elements has been verified referer.ee
to "standard" water samples supplied to our laboratory by the U.S. EPA, the U.S.
Geological Survey (U.S.GS) and the National Bureau of Standards (';3S). The U.S.
EPA comparisons for 13 metals in 5 different solutions r.re presented in Tables
5, 6, 7, 8 and 9. No problem^JVcmtheLOOto^"h^jjo^ii^cncentraticns present
were observed. Note that any value below the LCD is presented here only for
comparison anc would not normally be reported. A similar conclusion can be
drawn from the summary of data on two U.S.GS samples for 15 elements presented
in Table 10 and the 6 elements in the NBS mate-ial presented in Table 1. The
U.S.GS material was certified by a round robin evaluation and precision measure-
ments are reported with the mean value. The U.S. ZPA and ,'!3S reference materials
do not have such precision values available at this time. A fc-.'rth material
i
analyzed "eta! s Section of the
CRL to ascertain daily performance. These data are presented in Table 12. Again
the cc.-.parison between referee and proposed analytic:! rr.ethcd results is acce-otac!
for al1 metals analyzed.
1
-------
9.2 Comparability to Reference Method
The reference materials used in trie accuracy studies were prepared in
relatively pure waters and might not properly reflect problems associated with
real samples. Direct comparison ished by two studies called
Study 1 and II. In the first study to compare the results of ICA? to the AAS
reference methods 13 elements were investigated. [No reference method exists
1
for Yttrium and 3oron was referenced to the Curcumin Method.) In this approach
four different water matrices (laboratory, lake, river and effluent water) were
spiked with the 13 metals at two concentration levels. Concentrations were
cr.o sen to fail into the op-Hi mum range for AAS analyses. _r.e settles were esse
to cover a broad soectrum of sample types to approximate samples encountered
in our !!FD£S work. The samples were digested roll owing I?" orocedures* and
spiked at two concentration levels. After standing for ^3 ncurs the samples
were filtered to insure that homogenous aliquots could ::e taken. A set ef 3
aliquots were analyzed by AAS for each element C; 3 different days over apprcxi
T.ately 5 weeks. An equivalent set was a.na-yzed by an ICAP cn 5 different days
over the same period. The results of this study are presented in Taoles 13
through 32. Two total study s .varies are also presented in Taoies 33 and 3-.
The average recovery data in the latter Table would clearly indicate no orcble
with these data. The linear correlation presented in Table 33 is the 'inear si
of a plot of results by AAS (y) vs. ICAP (x). An ideal mate." would result in
a slope of 1.G0 and have an intercept of 0. The data document a satisfactory
comparison of the two methccs.
In the second study of comparabi1ity Z2 pairs of samples which were part c
the normal laboratory AQC program ,-/ere reanalyzed by an ICAP. Each samp'e pair
is a sample ar.d that sample spiked with cne routine AAS spike solution. This
-10-
-------
spike solution contains A1, Ba, Cd, Cr, Cu, Fs, Mn, Ni, Pb and Zn. The results
reported in Tables 35 to 57 represent a comparison of I CAP to AAS for all metals
in the AQC spike requested by the sa~ple originator for that sample. All sample
and spike pairs investigated are included, me total Study II comparison in
terms of percent recovery is given in Table 57. As is the case in the previous
studies, the methods are comparable.
-11-
-------
REFERENCES
Methods for Chemical Analysis of V.'ater end Hastes, Methods Development and
Quality Assurance Research Laboratory, Cincinnati, Ohio 45253, EPA-52516-
74-003.
C. C. Butler, R. N. Kniseley and V. A. Fassel, nnal. Chen. , £7, 825 ( 1975).
V. A. Fassel and R. N. Kniseley, Anal. Chem., &6, 111 OA, ( 1974).
0. J. Kalnicky, R. N. Kniseley and V. A. Fassel, in press (1975).
S. Greenfield, I. L. Jones and C. T. 8erry, Analyst (London), 39, 713 (1954).
G. F. Larson, V. A. Fassel, R. H. Scott and R. i\'. Kniselev, Aral. Chem.,
47, 238 (1375).
G. F. Larscr and 7. A, Fassel, in press ''975).
s. Indiana.
R. N. Kniseley, H. Ai-enson, C. C. 3u11 sr and 7. A. Fassel, AddI. Srectrosc.
25, 235 (19 7-).
R. K. Scout, V. A. Fassel, R. H. Kniselev and 0. E. Nixon, Anal. CheTi., -15,
75 (1974). =
V. A. Fassel, Prcc. lotn Coll. Spectr. In-., Heidelberg, 1971, Adam Hl-ger,
London, 'ir72, p. 63.
G. r. Larson, ¦/. . rassci, r,.
1975 FACES -eetino Ir.dianaocl:
-12-
-------
LIST OF TABLES
Table Mo.
General
1 Element List and Analytical Lines
2 Mean Detection Limits and Lowest Quantitativ
Determinable Concentration
3 Typical Relative Standard Deviations at 1 mg
4 Demonstration of Linear Range
Accuracy
5 U.S. EPA Reference Material 1*71 =1
5 U.S. EPA Reference .Material 1171 = 2
7 U.S. EPA Reference Material 117"; =3
8 U.S. EPA Reference Material ^75 =1
3 U.S. E-; '"ater-a" -75 = 2
10 U.S. GS Sa-'srsr.ce "iterial _?
11 MBS Reference Mater;a: (Prel;t:~ary)
12 In House Check Standa-d Comparison
Study I - Comparison of all Elements ;r> T:g/1 Ranee
13 Low Concentration in Distil led '..;ater
U Precision Study (LC) in D"stilled Water
15 Hich Concentration in Distilled Water
¦6 Precision Study {HC} in Distilled Water
17 Relative Recovery i:; Distilled Water*
18 Low Concentration in Lake Michigan Water
19 Precision Study (LC) in Lake Michigan Water
20 High Concentration in Lake Michigan Water
21 Precision Study in Lake Michigan Water
22 Relative Recovery in Lake Michigan Water
23 Low Concentration in ST?
2a Precision Study (LC) in ST3
25 High Concentration in A STP
26 Precision Study (HC) in A ST?
27 Relative Recovery in A ST?
2S Low Concentration in tne Calumet River
29 Precision Study (LC) in the Calumet River
30 High Concentration in the Calumet River
31 Precision Study (HC) in the Calumet River
32 Relative Recovery in the Calumet River
33 Total Stucy I Comparison - Linear Correlatfc
34 Total Study I Comparison of Relative Recovet
-------
Table No.
Study II - Samples and Spiked Samples
35 S TP ,= 1070
36 STP =3392
37 STP #01142
38 ST? =01172
39 STP £1298
40 STP =13757
41 STP =21361
42 STP =21324
43 STP =3443
44 Electric Power Generating =1227
45 Electric Power Generating t1243
45 Dirty River =21507
4?- - Military Arsenal =1345
48 Crinkin; '.-atsr =1*;0
<13 Geneva". industrial =12?C
50 General industrial =*.32-^
51*" Automotive Industry =7^3^
52 General industrial 1 2—9
53*» Tire Company =3445
54—• Paper industry =7007
55 Communications Industry =0013
56 Communications Incus try =2-94
57 Summary of "> Recovery in Study
-------
Name
X in r,m
flame
X i n nni
Ag
Si 1ver
328.1
A1
A1 uriiinu.n
396.2
S
Boron
249. 7
Ba
Barium
233.5
Ca(l)
Calcium
393.4
Ca(2)
Calcium
354.4
Cd
Cadmium
225.5
Co
Cobal t
238.9
Cr
Chromium
267. 7
Cu
Copper
324.3
Fe
Iron
259.5
Mg
Mn
Mo
Ni
Pb
Sn
Ti
V
v
2n
Magnesium
j Manganese
j Molybdenum
j Nickel
i Lead
Tin
T i za r, i urn
Vanad i uTC
Yttri-j-
Zi nc
279.6
257.6
203.3
341.5
220.3
190.0
334. 7
*>no ~
213.9
ELEMENT LIST ;;;D A MA!
;al l::;es
TASL
A "list cf the elements cjrrent'y. analyzed iy zne 3:>L ICAP-AES i-stru^:
the emission line chosen for each element.
-------
D.L. 100
ug/1 ug/1
D.L. LQO
ug/1 ^g/T
kg
4 20
Ma
<0.5 1
A1
7 35
Kn
1 5
B
3 15
Mo
5 - 25
Ba
1 5
Ni
15 75
Ca
^0.5 1
Pb
12 60
Cd
2 10
Sn
12 60
Co
4 20
Ti
1 • 5
Cr
1 5
V
1 5
Cu
1 5
Y
—1 5
Fe
2 10
In
1 5
~Five Runs over Three ''ctths
meam"detect:o;i l:":ts
AND LOV.EST QUALITATIVELY CETEv'INASLE C3';CE:rRAT"::;i (LJD)
TABLE Z
Ths dstacti D!i n'n-.i't 'D.L.; is tre airo-jnt of rr:£ tere 7 t rat .-riV. pro cues
d signal chat is twice as large as the standard d~."'it:or of tr.5 no:ss.
The lowest quantinative catsmiriile concentration {L'Z) '3 5 t'mss t.'.e
D.l. and is the lowest concentration one can expect' to rapcrt.
-------
6D
9f
to
RSO
r?d
Ag
1.8
Mg
1.1
A1
0.8
Mn
1.1
B
0.3
Mo
1.0
Ba
0.9
Ni
0.5
Ca
0.5
Pb
0.4
Cd
0.9
Sn
1.2
Co
1.0
Ti
1.1
Cr
0.3
M
4
1.1
Cu
1.2
Y
1.0
Fe
1.0
Zn
0.8
i
TYPICAL RELATIVE STA.'iDARC DEVIATION
AT 1 ng/1
TA3LE 3
These RSDs are typ.cal variation for 10 consecutive intargrcii^n
periods. Data ot this type is recorded at the beginning and end
of each day's operation.
-------
£ /
100 mg/1
i
mo mg/1
Ag
96 ± 1
*
105 t 2
A1
95 ± 1
Mn
00 : 3
B
100 r 1
Mo
mr = 2
Ba
106 ± 3
Ni
inn i i
CaO)
97 i 1
Pb
10A - 1
Ca(2)
989 = 2 *
Cd
105 = 4
Sn
105 ± 4
Co
97 = 3
Ti
inn t i
Cr
9P.2 - 0.4
V
?Q ± 5
Cu
95 t 7
V
ion r 2
Fe
94 = 2
Zn
96 i 2
* C = (2) at ICCCTc/i
Average ;f - ~jr,$ over 5 .-curs
Overall Ave~;c= ICQ : - 3cm
DEMONSTRATION OF LINEAR \ANGE
TA3LE 4
-------
t-x
USEPA
I CAP
USEPA
ICAP
yg/1
US/1
•jo/1
•-g/1
Ag
Mg
A1
1100
1030
Mn
4^9
465
B
Mo
Ba
Mi
Ca
Pb
350
350
Cd
73
75
Sn
Co
Ti
Cr
406
400
V
Cu
314
313
re
769
Zn
3C7
371
U.S. EPA RE'ERE.'ICE .''ATERI.AL 1171 - 2
All rsisrsncs niacsr'si run is a single biirvj 2x-p=ri;n5nc.
-------
(e J"
ICA?
ug/1
A.A.S.
Reaion V
Cg/1
US ZPA
•-c/1
- ICAP
! -g/l
A.A.S.
Reoion V
-Ig/1
US EPA
-.c/i
«g
AT
3
3a
Ca
Cd
Co
i Cr
I C-j
» r a
645
25.^
5*
522
157
550
37.5'
49
5"0
X f C
250
5' 0
700
36.7
SO
500
150
250
5C0
Mg
Mn
Mo
Ml
Pb
Sn
Ti
7
12.3-
363
273
254
/ J I
205
350
260
260
350
2S0
250
V?0
2::
U.S. i?~ Rlri?.z::zz MATES IA L
ICAP ; ¦ US : , ICAP ¦ A.A.S
-5/1
: ,S/"l
i
; --?/i
=eci-:n V
. -3/1
;3.'l
i-
. <200
*»r*
¦¦i
»
2. 7*
2.5*
2.5
* T
75
60
¦Xn
£ /
<20
3
VlQ
8c
;
."4 i
52
<50 :
30
r a
7.3*
7.2"
7.2
j
Pb
25
<30
2-
r-»
w
i <20
i 2.5
Sn
i
r*^
19
¦ <40
I 20
Ti
r>
V*
U
1 ' ¦")
<-0
'• t "!
k ) Sj
V
55
. 1 J~.0
70 ;
C*.
<20
'¦ n
r
" ^
- !
<50
20
! Zr.
.. "3
<20
- -
-32 3-"-?-
3; •i.'I.'iCE rATIRIAL ^"5-2
i-3i.; i
' lutri en
Cone. are
Sample ^
fn '"i-/]
r-r"-r|-- ~a*2r*a! p^n as a sire'* :linc zxcer'~sr,z.
-------
USGS
I CAP
USns
I CAP
ug/1
yg/l
ug/l
ug/1
A g
5.4 t 0.9
5.7
Kg
18.3 - 0.9+
13.7
A1
71 = 35
66 A
Mn
15° - H
150
B
92 i 2°*
113
Mo
56.6 : d.fi
62.5
Ba
Ni
9.3 ± 6
<15
Ca
69.5^2.5"
65.7
Pb
23 = 11
15.5
Cd
P»
II-
r*
5.9
Sn
Co
5.1 - 0.5
<4
Ti
Cr
16.5 ± 6
19.7
V
Cu
391 ± Zl
379
Fe
37 r 15
30
2n
!
347 = 2S
3^?
USGS RE."E3Er:CE VA'iSIAL -5 cr.c £7
TABL£ 1"
The 'j'SGS va'us represeits :ne average and standard deviation of all lacerate
who participate! in the Ana:y:ical Evaluation ?rc;r=rr. ~zr Scancarc .^a-'sre'vea
Water Samples 4? and 47 through May, 1 975.
All reference -ataria 1 run 2S a singl3 blind experiment.
-------
:;gs
ICA?
N3S
I CAP
V9/1
vg/1
ug/l
-9/1
Ag
Ha
A1
Mn
42
44
3
Mo
3a
Ni
Ca
Pb
*7
20
Cd
14
12
Sn
Co
Ti
Cr
47
49
V
Cu
8G7
£
«jtO
F.
Zn
-770
i
N33 REFEr.E'iCE
TAcL£ 11
The MSS values are pr=1, iainir;.- 2nd rsprsser.t ^ncert¦ ":£i trsaoriv al vc'-je^
SUDpl'Sd bv :N£S fcr th5"!
¦• -••• »
. f w
"ixsd hez r
;Te:21'5
-------
6 ^
ICA?
Ave.
•jg/1
A AS
Ave.
.0/1
I CAP
Ave.
-1/1
AAS
Ave.
-n/1
Ag
Kg
A1
2257
2510
Mn
50?
rt?4
B
Mo
Ba
12=7
1210
Ni
1229
7 2*2
Ca
Pb
1298
1231
Cd
5C5
492
Sn
Co
7i
Cr
V
Cu
462
470
Fe
1312
" 2r5
Zn
•>¦33
A3*
*+ Av°!"2C2 1 y-a2'"
*+*A'/2r29e 2 rnc-nt
IN-HHU^E CHECK 3TA'!D;i°C C°:'?AP:s~M
of AA5V" .;:::¦ I
-------
STUDY I
COMPARISON OF ICAP AND AAS AT
mg/1 CONCENTRATIONS
IN
OISTILLED ".-/ATER (Digested as Any Sample)
LAKE MICHIGAN (A Clear, Lake)
Res icier,:i = l '..'astes;
CALUMET RIVER (A Oirzy River)
FOR
A SIX ;:eek period to show precision
-------
I CAP
mg/1
AAS
mg/1
I CAP
mg/1
AAS
ma/]
Ag
0.260
0.254
Kg
Q.9
°.8
A1
10.0
10.0
Mn
1! 02
0.9°
B
0.50
0.51*
Mo
25.7
26
Ba
9.4
9.1
N i
5.1
4.96
Ca
25.8
25.4
Pb
5.25
5.0
Cd
0.252
0.249
Sn
10.0
10.
Co
0.97
1.03
Ti
in.o
10.fi
Cr
1.04
1.02
V
o.9
•9.5
Cu
1.00
1.02
Fe
1.18
1.05
Zn
0.55
P. 50
-
LOW CONCENTRATION IN DISTILLED WATER
TABLE 13
I CAP
i o mg/1
AAS
= j mg/1
I CAP
r a mg/1
' .AAS
= z mg/l
Ag
0.007
0.002
Mg
0.2
0.2
A1
0.3
0.1
Mn
0.C2
0.02
B
0.02
0.02*
Mo
0.8
1
Ba
0.2
0.3
Ni
0.1
0.03
Ca
0.5
0.4
Pb
o.os
0.2
Cd
0.005
0.002
Sn
0.2
1
Co
0.02
0.02
Ti
0.3
0.8
Cr
0.01
0.01
V
0.2
0.2
Cu
0.02
o;go
Fe
0.03
0.02
Zn
0.01
0.01
n
6
3
n
PRECISION STUDY - LOW CONCENTRATION
IN DIS.ILLiD '.iAicR * Curcumin Method
TABLE 14
-------
£<1
ICAP
AAS
ICAP
AAS
mg/1
mg/1
mg/1
mg/1
Ag
0.52
0.52
Mg
15.0
14.8
A1
15.3
14.4
Mn
2.05
1.95
B
1.00
0.99 *
Mo
52.1
52
Ba
14.1
13.7
m
10.3
9.9
Ca
49.8
52.9
Pb
10.4
9.9
Cd
0.512
0.50
Sn
1«.7
20.6
Co
1.94
2.04
Ti
19.6
20.1
Cr
2.07
2.07
V
19.5
18.5
Cu
2.00
2.05
Fe
2.25
2.02
Zn
1.06
0.99
1
HIGH CONCENTRATION IN DISTILLED WATER
TABLE 15
ICAP
1
AAS
ICAP
AAS
± a mg/1
t7i7ig/l
i a mg/1
± c mg/I
Aa
0.02
0.004
Mg
0.3
0.2
A1
0.2
0.1
. Mn
0.02
0.03
8
0.04
0.01* .
Mo
0.7
2
Ba
0.1
0.3
Ni
0.2
0.1
Ca
0.4
0.3
Pb
0.2
0.1
Cd
0.006
0.01
Sn
0.5
0.2
Co
0.02
0.05
Ti
0.2
0.1
Cr
0.05
Q.OS
V
0.05
n.u
Cu
0.01
0.11
¦
Fe
0.04
0.02 .
"2n
0.01
C. 02
n
0
3
5
3
PRECISION STUDY - HIGH CONCENTRATION
IN DISiiLuc.0 A'Aic.* *C'Jrcumin tfethcd
-------
JV
ICAP
AAS
ICAP
AAS
tf
V
&
9f
P
JO
A
»
Ag
105
108
Mg
102
ion
AT
106
88
Mn
103
B
TOO
56 *
Ho
105
103
¦Ba
Q4.6
93
Ni
102
98
Ca
96
106
Pb
10?
97
Cd
104
100
Sn
97
113
Co
97
101
Ti
96
95
Cr
103
105
V
?7
90
Cu
100
103
Fe
108
97
In
102
98
* Carc-imir. Method
RELATIVE RECOVERY STUDY
IN
DISTILLED WATER
TABLE 17
Relative Reeavprv = Hich Ccnc. - Low Cone.
Spike Cone. x 100
-------
7 /
I CAP
AAS
I CAP
AAS
mg/1
mg/1
mg/1
mg/1
Ag
0.262
0.258
Mg
21.2
20.9
A1
10.6
10.3
Mn
1.02
0.98
B
0.54
0.55*
Mo
26.3
21
Ba
9.5
9.1
m
5.2
5.02
Ca
62.3
64.1
Pb
5.40
5.1
Cd
0.258
0.252
Sn
TO. 3
10.4
Co
0.97
1.02
Ti
9.9
10.6
Cr
1.05
1.03
V
10.1
9.5
Cu
1.02
1.04
re
1.37
1.20
Zn
0.56
0.50
LOW CONCENTRATION fN LAKE MICHIGAN WATER
TABLE 18
I CAP
AAS I
ICAP
AAS j
±ortg/l
ic.no/1
" !
±cir.g/l
±^mg/l j
Ac
0.009
0.005
Mg
1
0.4
n.3
A1
0.3
0.1
Mn
0.03
0.02
B
0.04
0.04 *
Mo
0.9
1
Ba
0.2
0.3
Ni
0.1
n.07
Ca
0.4
0.4
Pb
0.09
0.1
Cd
0.008
0.003
Sn
0.3
1
Co
0.03
0.02
Ti
0.3
n.8
Cr
0.02
0.01
V
0.3
0.2
Cu
0.03
0.09
Fe
0.02
0.01
Zn
0.01
o.n?
n
6
3
n
6
1 3..
PRECISION STUOr - LOW CONCENTRATION - LAKE MICHIGAN
TABLE 19 * Curcumin Method
-------
ICAP
AAS
ICAP
AAS
mg/1
mg/1
mg/1
mg/1
Ag
0.52
0.52
Mg
25.6
25.7
A1
15.6
14.4
Mn
2.02
1.95
B
1.07
1.03*
Ho
51
52
Ba
13.9
13.7
Ni
10.2
9.83
Ca
83.3
83.8
Pb
10.3
9.9
Cd
0.51
0.51
Sn
19.6
21.8
Co
1.90
2.02
Ti
19.5
20.1
Cr
2.06
2.01
V
19.4
18.5
Cu
1.98
2.05
Fe
2.23
2.07
Zn
1.05
0.98
HIGH CONCENTRATION IN LAKE MICHIGAN WATER
TA3LE 20
ICAP
AAS
ICAP
AAS
±ff!ng/1
f^g/i
± e ng/1
=-mg/1
Ag
0.02
0.01
Mg
0.4
0.2
A1
0.3
0.1
Mn
0.04
0.02
B
0.07
0.03*
Mo
1
2
Ba
0.2
0.3
Ni
0.2
n.ns
Ca
0.5
0.5
Pb
0.2
0.2
Cd
0.01
0.01
Sn
0.6
0.2
Co
0.04
0.05
Ti
n.4
0.1
Cr
0.05
0.07
V
0.4
0.4
Cu
0.05
0.08
Fe
0.06
0.03
Zn
0.02
n.o2
n
5
3
n
6
3
(over 5 wee*
PRECISION STUDY - HIGH CONCENTRATION - LAKE MICHIGAN
TA3LE 21
*Curcunn'n Method
-------
7>
I CAP
AAS
I CAP
AAS
V
&
0/
*
%
*
/9
Ag
103
'104
Mg
88
A1
TOO
88
Mn
100
97
B
94
96*
Mo
101
103
8a
90
93
Ni
99
?6
Ca
104
98
Pb
98
96
Zd
100
104
Sn
Q3
113
Co
93
100
Ti
96
95
Cr
101
98
V
93
90
Cu
96
101
Fe
91
87
2n
98
* Curcumin "2"hod
RELATIVE RECOVERY
IN
LAKE MICHIGAN WATER
TASLE 22
Relative Recovery = ^'9^ Cone. - Low Cone.
Spue Cone.
-------
I CAP
AAS
I CAP
AAS
mg/1
mg/1
mg/1
mg/1
Ag
0.26
0.27*1
Mg
41 .6
42.5
A1
73.5
12.6
Mn
1.18
1.15
B
1.15
1.2 *
Mo
25.7
25
Bfi
9.3
9.2
Ni
5.30
4.99
Ca
80.0
82.6
Pb
5.46
5.2
Cd
0.284
0.279
Sn
9.9
11.2
Co
0.28
1.04
Ti
9.8
10.5
Cr
1.30
1.29
V
10.1
9.5
Cu
1.18
1.22
Fe
7.55
7.1
Zn
1.19
1 JA
LOW CONCENTRATION IN A STP
(Industrial & Residential Wastes)
TABLE 23
I CAP
l
AAS
I CAP
AAS !
±3mg/1
t'-,c/l
±cmg/l
tcmg-/1 j
Ag
0.01
0.007
Mg
0.8
o.«
A1
0.1
0.2
Mn
0.01
0.02
8
0.06
0.1 *
Mo
0.4
1
Ba
0.1
O.a
Ni
0.06
0.02
Ca
0.4
O.o
Pb
0.07
0.2
Cd
0.004
0.001
Sn
0.2
0.8
Co
0.01
0.02
Ti
0.1
0.4
Cr
0.04
0.01
V
0.1
0.2
Cu
0.01
0.08
Fe
0.05
0.1
Zn
0.05
0.05
n
6
3
i n
6
3
PRECISION STUDY - LOW CONCENTRATION - ST?
TABLE 21
* Curcumin Methca
-------
I CAP
AAS
I CAP
AAS
mg/1
mg/1
mc/1
mg/1
Ag
0.52
0.529
Mg
46
47.6
A1
17.6
16.2
Mn
2.13
2.10
B
1.55
1.73*
Mo
49.6
51
Ba
13.7
12.7
Ni
10.2
9.6P
Ca
84.7
83.7
Pb
10.0
9.7
Cd
0.52
0.53
$n
18.9
22
Co
1.90
2.04
Ti
18.9
19.1
Cr
2.27
2.29
V
23.5
27.9
Cu
2.09
2.19
Fe
8.9
8.5
Zn
1.57
1.64
HIGH CONCENTRATIONS IM A STP
{ Industrial & Residential Wastes )
TABLE 25
ICAP
AAS
ICAP
AAS
iamg/1
i.CiiC/1
"
~rag/I
ra^ig/l
Ag
0.02
0.004
Mg
2
0.4
AT
0.4
0.4
Mn
0.04
n.ru
3
0.09
C.05 +
Mo
no
d
Ba
0.3
0.7
Ni
0.3
0.08
Ca
0.5
0.4
Pb
0.2
n. *
Cd
0.01
0.02
Sn
n.4
l
Co
0.04
0.06
Ti
0.4
0.6
Cr
0.08
0..0S
V
0.6
0.8
Cu
0.04
Q.C8
Fe
0.1
C.2
Zn
0.H2
n.rw
n
5
3
m
6
3
PRECISION STUDY - HIGH CCNCE.N75A7I0N - STP
TABLE 26
* Curcumin Mstnod
-------
7t
ICAP
AAS
ICAP
AAS
*
A
Of
-------
I CAP
AAS
I CAP
AAS
mg/1
mg/1
mg/1
mg/1
Ag
0.257
0.259
Mg
23.7
24.0
A1
11.0
9.9
Mn
1.15
l.U
B
0.56
0.55*
Mo
25.7
2f
Ba
9.2
9.0
Ni
5.24
5.04
Ca
122
125
Pb
5.37
5.1
Cd
0.258
0.2*9
Sn
10.1
10.?
Co
1.90
2.02
Ti
9.7
10.fi
Cr
1.03
0.99
V
9.9
9.5
Cu
1.02
1.04
Fe
2.59
2.37
Zn
0.500
0. t5
LOW CONCENTRATION IN THE CALUMET RIVER
TA3LE 23
(The CaTurret River is very do1,luted, its waters resemble discharges from
industrial steel mill pler.ts.)
ICAP
AAS
ICAP
AAS j
±cir,g/1
±s*g/l
ipmg/1
±omg/l j
Ag
0.009
0.003
Mg
0.3
0.7
A1
0.2
0.06
Mn
n. 12
0.02
B
0.03
0.04*
Mo
n. 7
2
Ba
0.2
0.3
Ni
0.09
0.08
Ca
0.9
1
Pb
n.os
0.3
Cd
0.005
0.004
Sn
0.2
n.A
Co
0.04
0.02
Ti
0.3
0.8
Cr
0.03
0.01
V
0.3
0.3
Cu
0.02
0.03
Fe
0.04
0.03
Zn
0.009
0.03
n
6
3
n
6
3
PRECISION STUOr - LOW CONCENTRATION - CAL'JMET RIYER
* Curcumin Matho
TABLE 29
-------
*7 ^
I CAP
AAS
I CAP
AAS ¦
mg/1
mg/1
mg/1
nig/1
Ag
0.49
0.50
Mg
37
37.9
A1
15.2
13.9
Mn
2.0*
2.02
B
1.02
1.10 *
Mo
5n
51
Ba
13.8
14.0
Ni
9.7
9. A3
Ca
119.1
119.6
Pb
P. 7
9.5
Cd
0.49
0.50
Sn
18.7
21.4
Co
2.8
3.0
Ti
18.4
19.3
Cr
1.96
1.98
V
18.6
18.3
Cu
1 .90
2.01
Fe
3.5
3.24
Zn
1.02
n.37
HIGH CONCENTRATION IN THE CALUMET RIVER
TABLE 30
I CAP
AAS
I CAP
i
AAS
±omg/l
±jmg/l
i
ij.^g/l
±7,nq/l
Ag
0.01
o.nos
Mg
0
n.
A1
0.4
0.2
Mn
6.09
n.nnF
B
0.05
0.03*
Mo
2
3
8a
n.5
0.7
Ni
6.3
O.ng
Ca
n.5
0.6
Pb
0.2
0.2
Cd
0.02
0.01
Sn
0.6
n.6
Co
0.1
0.2
Ti
0.8
O.P
Cr
0.06
0.09
V
0.9
O.fi
Cu
0#C8
0.1
0.09
0.3
n.i
Fe
0.08
Zn
n
6
3
n
6
3
PRECISION STUDY - HIGH CONCENTRATION - CALUMET RIVER
*Curcu.Tiin '••ethcd
TABLE 31
-------
? 1
ICAP
*
A
AAS
V
Jo
ICAP
*
*
AAS
Bf
Ag
93
9P
Mg
89
92
A1
84
80
Mn
91
88
B
92
no *
Mo
98
inn
Ba
92
100
Ni
8?
Ca
Pb
87
a*
Cd
93
100
Sn
86
inn
Co
90
98
Ti
87
87
Cr
93
99
V
88
87
Cu
88
97
Fe
91
87
Zn
86
OA
* Cyrcumln Method
RELATIVE RECOVERY
IN THE
CALUMET RIVER
TA3LE 32
Relative Recovery = Hich Cone. - Low Core, *p«.
ipike Ccnc.
-------
Slope
INTCP
yg/1
Slope
INTCP
ug/1
Ag
1.02
*0.003
Mg
1.05
-0.O25
A1
0.87
+0.300
Mn
0. °7
+O.On3
B
1.07
-0.035
k
Mo
1.02
-0.479
Ba
1.03
-0.453
Ni
0.96
+0.003
Ca
1.00
1.3?
Pb
0,°8
+n.l?Q
Cd
1.03
-0.014
Sn
1J1
-0.905
Co
1.08
-0.019
Ti
n.Q8
+0.958
Cr
1.02
-0.048
V
0.99
+0.308
Cu
1.07
-0.039
Fe
0.97
-0.126
Zn
1.01
-0.067
* Cure win Me"hod
sl°>8 ¦ STOP
TOTAL STUDY I COMPARISON
LINEAR CORRELATION
TABLE 33
This data is the least squares slope and intercept of all data collected
in Study I. The AAS Values are plotted on the y axis v.s. the ICAP values
on the x axis.
-------
I CAP
of
AAS
V
10
ICAP
<¦0
AAS
<3/
S3
Ag
101 ±4
10* ; 3
Mg
93 ± 8
?fi : 4
A1
93 ±12
83 ; 7
Mn
C7 ± 5
91 ± 4
B
97 ±4
101 ± 7
*
Mo
ion ± 4
ifi2 ± 1
Ba
91 ± 3
94 ± 4
Ni
97 ± 6
94 = 4
Ca
100 ±6
102 ± 6
Pb
95 ± 7
93 ± d
Cd
93 ±5
101 t Z
Sn
95 ±- 3
103 ± 7
Co
92 ±5
100 r 1
Ti
92 ±4
- °0 ± 5
Cr
99 =5
inn ± 3
V
93 ± 4
90 = 2
Cu
94 * 5
100 ± 2
Fe
106 ±17
102 ±2^
Zn
96 = 7
-T^>
II
VI
* Curcumin •••ethcd
TOTAL STUDY I CO?'PAR:SO.'i
OF
RELATIVE RECOVERY
TABLE 34
The recovery presented is the averageUnd 3) for the observed recovery in
each of the four water types in Study I.
-------
STUDY II
COMPARISON OF ICA? AMD AAS
FOR
SAMPLE AND SPIKED SAMPLE PAIRS
NORMAL LABORATORY AQC PROGRAM
WITH
ALL SAMPLES CHOSEN' AT RANDOM
-------
S3
1
Sample
Spike"1 Sample
Recovery
I CAP
AAS
I CAP
AAS
I CAP
AAS
Element
ug/1
ug/1
vg/i
ug/1
*
*
0/
n
A1
<14
132
672
9^n
P-4
lnl
Ba
Cd
*T
Id
2Uo
2^
102
95
Cr
<1|
40
2Zd
23°
im
oo _ 5
Cu
23
32
211
232
94
mo
Fe
609
579
1715
153*
110
im
Mn
62
65
271
267
104
lni
Ni
<30
16
181
256
91
120
Pb
22
2 i ri
23 i
! 00
1 ' "
| 1-14
Zn
202
219
4(14
^•22
! ino
! 102
! 1
i i
I !
TABLE 35
i
j Sample
i '
Spiked Sample
Recovery i
- \
1 1
ICAP
AAS
ICAP
AAS
ICAP j
AAS
jElemer.t'
•jg n
ug/1
ug/1
ug/1
o* ;
!
«/
A1
9*
328
986
lim
i
m
102
Ba
119
104
100*
032
m
10d
! Cd
-------
Sample
Spikec
Samp 1 ->
Recovery
ICAP
AAS
ICAP
- AAS
ICAP
• AAS
Element
ug/1
ug/1
ug/1
ug/1
9/
,©
of
JO
AT
730
1020
1549
1823
102
101
Ba
Cd
< ^
/
19?
213
l no
ln3
Cr
9
10
201
204
96
ion
Cu
10
12
139
206
90
op
Fe
1 56
lfll
1240
lli'i
108
100
Mn
54
54
25?
261
ins
103
Mi
<30
22
173
232
37
ms
Pb
15
26
209
228
97
i 101
In
! 36
40
225
232
95
! On
; i ; !
I !
STP # 01142
TABLE 37 ¦
Sample I Spiked Sample " ' ; "Recovery
I CAP j AAS J ICAP i AAS j I CAP | AAS
'¦Element! -9/"! j ug/1 j ug/1 ug/1 i % \ %
i i 1 , I I i
A1
Ba
1
t
1
1
Ca
Cr
-------
Sample
Spiked Sample
Recovery
I CAP
AAS
I CAP
AAS
I CAP
AAS
El ement
vg/1
ug/1
yg/1
wg/l
w
' 49
c/
A1
706
30^
KZ5
1528
90
°1
Ba
Cd
K 4
u
1
^0
MM
Cr
40
35
221
222
qfi
c&
Cu
19
25
20A
223
93
99
Fe
1650
1744
27*4
27 "S
109
102
Mn
162
170
367
38£
103
1 P4
Ni
<30
25
! 181
222
°1
1 Of?
Pd
45
57
i 2*3
264
97
GC
Zn
! 193
212
i 393
5
ion
K'l
1 . ¦ ,
i !
SIP =1298
TABLE 39
Samp!e
Spiked Sample ;
Reccvsrv
I CAP
AAS
I CAP
AAS !
I CAP i
AAS
1 Element
uc/1
yg/l
yg/1
ug/i !
~ i
<* 1
•3 !
of
n
A1
i
|
i
i
Ba
Cd
1
C**
/
21 2
icc
106
Cr
-i
0
5
213
ISO
1 nZ
30 I
Cu
£3
223
/¦» ^
,
-------
^ (*
Sample
Spiked Sarcole
Recovery
3 CAP
AAS
I CAP
AAS
I CAP
AAS
Element
ug/1
vg/1
yg/l
ug/1
9/
to
0/
iO
A1
830
800
1552
1F3^
05
1^3
Ba
117
120
S52
872
°2
QC
Cd
1/
20
209
203
°3
°n
Cr
275
261
468
440
87
04
Cu
164
21 fi
339
3*o
88
Q?
Fe
7132
7100
8515
SAnn
Mn
241
255
<144
458
102
in?
Hi
223
2*3
£15
£.38
9£
OR
15 5
150
331
3^
88
OH
Zn
fS7
70S
895
88 a
1TA
an
SIP #21351
TA3LE 41
j I Sample | Spiked Sdinple = Recovery
,
ICA?
^ A C 1
nnJ
I CAP
AAS
I CAP
j AAS
[Element
ua/1
'
,3/1 1
-yc/1
-- - -tro/1 ¦ i
9/
.9
0*
1 J
A1
1074
1152
1897
I960
103
J
i
{ 1 r*i
3a
112
<100
852
84S
93
1 n,?." "
Cd
3
-------
Semple
Spiked Sample j Recovery
Element
I CAP
ug/l
AAS
ug/l
1 CAP
yg/1
AAS
ug/l
I CAP
ct
JO
AAS
19
AT
Ba
Cd
Cr
Cu J
Fe ! 153
173
107?
11*2
93
Q7
Mn
fli
Pb I j i
Zn ! 1 !
ST? #3443
TA3LE 43
Sample
Spiked Sample
Recov
ery
I CAP
AAS
ICA?
AAS
I CAP
AAS
El emer.t
ua/l
'
us/1
'
ug/l
ug/l
J
a/
•9
A,
223
444
959
1268
92
1 rs3
Be
! Cd
< ^
< i0
i
u O.
1 Cr
10
10
202
191
91
Cu
10
0
179
2C4
37
Fe
33?
333
1 3°7
1304
in*
97
Mn
27
22F
234
ini
in* •
Ni
<30
21
196
210
93 •
00
Pb
<2-
22
216
•id
i 00
i'1-'
Zn
17
n
2 n?
193
of,
00
1
1
ELECTRIC POWER GENERATING FLA,'IT =1227
TA3LE 44
-------
Sample
Spiked Samp,e
Recovery
I CAP
AAS
I CAP
AAS
I CAP
AAS
Element
vg/l
vg n
ug/i
^g/i
Of
.«
0/
jO
A1
549
515
1476
nno
116
98
8a
98
SO
551
51*
Cd
< 4
11
216
?U
107
inz
Cr
16
u
225
?in
ina
OR
Cu
n
19
.206
222
CO
i 1 r>?.
Fe
22-
2*o
1 i1 6
1254
120
! mi
Mn
15
16
j 237
223
1 HI
107
Ni
<30
35
I 208
237
1 in/t
101
?b
34
| 239
237
! 1C2
I in?
7 ">
19
23
1 232
22*
1 Tlr
i i p-j
<
1 153
KiCO
2195
i 2272
i in?
i 112
mil . U»y !
-' J i
I CAP
AAS
:cA?
.AAS
I CAP j
AA^ i
Elerr.cr.v
uc/l
*9/1 I
•jg/i
B9/i
*
|
1
A1
06
Cci
<10
2n6
95
IP2
Cr
13
<20
1 99
In?
23
77
Cli
14
; /
192
212
89
Fg
1877
1SOO
2751
2fOn
I'n
i
<¦30
34
181
23?
90.5
1 pn
Pb
w —
*40
227
23-
0<1 '
1 r.n
Zn
70
""i ¦»
21H
J29
c:
' n2
1
POLLUTED RIVES VZ~507
TASLE 46
-------
Samplc
Sp"il;cd Sample
Recovery
Element
I CAP
•jg/l
AAS 1 I CAP
vjq/1 j . yg/l
AAS
ug/1
I CAP j AAS
e' I ©/
>9 j to
A1
Ba
Cd 1
Cr 3
<20
17?
198
i
80 i c5
Cu 1 l
Fe j 903 I 852
160
1563
204
IP 50
SO 1 98
7« t co
Kn 1 7 j <10
Ni ' I
1 83
212
88 j in?
?b ; <2- i <-G
Zn 1 22 1 5o
i / i.
230
2jG I co ; l'"l
2^3 ! 10A ! ir.d
MILITARY ARSS.\'AL =;3^5
TABLE 47
i |
Sample ! Sp"*!:ed Sample | Recovery
j j I CAP I AAS j I CAP j AAS
i£1ernen:! -S/"1 j -9/< | yg/- J -9/1
1
A1 | <20
Ba
Lj
622
I - 1
/? 3 | 78 ?ii
i 1
Co j
Cr i a
4 j 208
i
lOJ ! 1 JC
I
cr. :
i
Cu i :
F e ! .1
— _________ ,
« i <
i I " 1
Mn 1
Ni !
! I
Si ! !
i
i
1
t
1
i
1
1 ! 1
1 : !
-------
Sample
i
Spiked Sample
Recovery
I CAP
AAS
1CAP
MS
I CAP
AAS
Element
ug/1
ug/1
ug/1
ug/l
of
* v
1
i
Element
I CAP
yg/l
AAS .
ug/1
I CAP
ug/1
AAS .
-yg/l
. -I CAP j
«¦ i
.0 j
r.hw
V
.-0
s
1
I
AT
3a
<14
si 00
591 •
904 -
.
*
!
Cd
Cr
< a
3
< 10
< 20
181
183
197
1Q?
90
an
23
2.1
Cu
Fe
3
77
iO
64
164 '
1024
201
1085
30
^5
1,0?
Kn
Hi
7
23
8
*¦ 30
• 194
202
217
infi
-3
cn
K-4
m?
?b
In
<24
] a
< 40
17
".37
190
22f
?l 5
Cd
OS
98
on
!
4
!
GENERAL INDUSTRIAL =1224
TABLE 50
-------
Sample | Spiked Sample
Recovery
Element
I CAP
vg/1
AAS i 5 CAP
yg/1 | wo/1
A^S
vg/1
ICAP j AAS
** 1 •/
.9 | *0
A1
8a
132 J 132
7S2
?24
33 j 9<3
Cd j
Cr 1 10
<20
22
-------
2L
Sample
Spiked Simp if.
•Recovery
Element
I CAP
tg/1
AAS
ug/1
I CAP
ug/1
AAS
ug/i
I CAP j AAS
fl/ 1 0/
1 n
AT
3a
<14
36
539
840
im
Cd
Cr
i
Cu j 35
fe 1
3o
206
o «*> ~
86
Q5
Mrs
Ni
Pb | <24
Zn j
30 j 220
i
205 i no j mi
i i
i >
TIRE CCMPAMV fS--5
TABLE 53
Samol ¦
Spiked S-r.-ple
Recov-2rv
£1 cme^.t
I CAP
-5/1
AAS
¦jg/i
I CAP
ug/i
;as
ug/i
1 :cf i
f/
3
A1 j
£a !
1
i
1
!
1 !
Cg
Cr
20
< iO
< 20
% ¦"> A
1
2no
201
; -9
i cr)
af
-tnz
Cu
Fe
1 CD
1 2?5
2 CO
1380
2'S?
¦rj
2-'3C
i it
; cq
J i
1
i-'n
Ni
59
75
23n
277
i
1 po
irsi
Pb
Zn
zo
zcs
40
7 to
«— w *
237
3<3d
232
*13
i ^
f or
¦> ^
''4
1
! 1
1 i
Q * -
¦ rtf*
£3 ;^nr:v
»' • * > * W W . t > •
T.-®L£
• J 0 /
-------
Sample
Spikec< Sample
Recovery
I CAP
AAS
I CAP
MS
I CAP
AAS
Element
yg/1
ug/1
ug/i
pg/l
of
-------
S4
ICAP ¦ ¦ AAS
U
7T
ICAP AAS
4
:r
% ± a % ± z
Samples
\t a % ± o
Samples
Ag
Ma
A1
11
O
o
C\J
4 1
o
14
Mn
1D3 ± 7 103 ± 2
14
B
Mo
6a
98 = 11 93 = 5
3
Mi
94 = 6 103 = 6
16
Ca
Pb
99 = 8 101 ± 3
17
Cd
99 ± 6 98 ± £
16
Sn
Co
Ti
Cr
98 io 93 r 5
18
V
102 112
1
Cu
91 ± 6 97 ± 3
18
Fs
105 = 13 101 ±9
16
Zn
100 ± 7 98 ± 5
15
^0' <' • • j ¦ -"j • V4»L«j v c...
iff
STUDY II
TA5LE 57
-------
95
Ihe Jarrell-Ash Plasma AtarConp
-------
ft
The Jarrell-Ash Plasma AtomComp
Frederick Brech Richard L. Crawford
Jarrell-Ash Division
setts 02154
1. Description
The Jarrell-Ash AtomComp with Inductively Coupled Argon Plasma exci-
tation (ICAP) forms an analytical system for simultaneous multi-element
determinations of trace metals at the sub-ppm level in solutions. The
basis of the method is atomic emission promoted by coupling the sample,
nebulized to form an aerosol with high temperature argon gas produced by
passage of argon through a powerful radio-frequency zone.
The system shown in Figure 1 comprises three essential modules. The
radio-frequency generator is the stand-alone package on the right-hand
side which delivers up to 2 kw of energy at 27.13 MHz through a water-
cooled one and one-half turn coil connected to a pi network with adjust-
ments to match the output impedance of the power supply to that of the
argon gas discharge. The R.F. coil and impedance matching box are con-
tained by the attachment on the right-hand end of the AtomComp Spectrom-
eter. Another essential module is the plasma torch and associated sample
nebulizer system. These are shown in Figure 2, in which the quartz torch
itself is shown as surrounded by the R.F. coil. The sample nebulizer
couples to the bottom of the torch and terminates in the aspirator, shown
in Figure 1, as drawing sample from a beaker. Flow controllers are pro-
vided for argon gas ionized by the radio frequency energy and separately
for the argon that aspirates the sample into the nebulizing chamber. The
third module is the AtomComp Spectrometer and readout system that is
operated under control of a dedicated mini-computer. Radiation from the
plasma, defined by an entrance slit to the spectrometer, is dispersed by
a grating, and selected frequencies that fingerprint elements of interest
are transduced by photomultinliers, a separate one for each element. The
electronic energy thus produced is treated by the computer and reported
on the teletypewriter directly as concentration of each element.
Jarr«mAah Division
Fisher Scientific Company
590 Lincoln Street
Waltham. Massachusetts 02154
-------
*? 1
-2-
Atomlc emission through the use of an Inductively Coupled Argon
Plasma is not new in concept and indeed Greenfield^ introduced such an
analytical system in the mid-1960's. The promise of high analytical
success, however, did not mature because of a particular difficulty in
coupling the sample aerosol to the argon plasma itself. The aerosol,
in fact, skirted the external surface of the plasma discharge with a
resultant poor energy transfer and inefficient atomic emission. Velmer
(2)
Fassel and co-workers solved this problem by modifying the shape of
the gaseous discharge such that it carried an axial zone cooler than
the surroundings. When viewed in plan from the top, the discharge shows
a central hole, or expressed in other words, is doughnut-shaped, and the
sample aerosol is projected into the axial zone resulting in markedly
improved coupling to the ionized argon, and efficient atomic emission
from the sample aerosol is promoted.
Figure 3a displays the ionized argon discharge with pure deionized
distilled water aspirated into it. The discharge view is an elevation,
and shows the cooler core as a darker zone. The gaseous discharge is
formed by a laminar flow of argon contained in a quartz torch, passing
through an intense magnetic field by applying radio-frequency energy
to a one and a half turn coil surrounding the torch and shown in the
figure as two horizontal bands. Inmediately below the lower band, the
tip of the axial sample introduction nozzle may be seen. This conducts
the nebulized sample into the gaseous discharge. In Figure 3b the pure
water sample has been replaced by a solution containing yttrium, and the
plasma discharge is shown to exhibit a structure. The axial red tip
emerging from the plasma continuum is rich in excited Y atoms, and above
this the blue-greenish zone is rich in ionized Y, and still higher the
plasma cools to yield excited species once more.
The temperature functions of the plasma itself and of the sample
coupled to it are dependent, inter alia, on the separate flow rates of
argon for the plasma and that for sample nebulization, pnd a high quality
(1) S. Greenfield et al., Analyst 89_y 713 (1964).
(2) V. A. Fassel, "Electrical Plasma Spectroscopy," XVI Colloquium Spec-
troscopicum Internationale, Adam Hilger, London, England, 1973.
-------
-3-
control for these is required. Figure 3b represents optimum condition
whereas Figure 3c is at flow conditions that allow overheating to en-
danger the life of the torch, and Figure 3d at other conditions resulting
in too cool a plasma for efficient emission. Control of experimental
conditions are of marked importance to successful and long-term operation
of the plasma, and these are fully catered to in the Jarrell-Ash system.
2. Functional Performance
Inductively Coupled Argon Plasma excitation has a number of qualities
valuable to spectrochemical analysis, and one of importance is that the
discharge is optically thin. Little self-absorption occurs. Another
quality is that the discharge is analytically sensitive and the resul-
tant combination of these properties is that the discharge has a linear
dynamic concentration range exceeding five orders of magnitude for most
elements. It is to be recognized, however, that this is the property of
the discharge itself and full use of this quality for analytical purposes
can be realized only if the dynamic range of the readout system matches
that of the discharge. This is the case for the Jarrell-Ash AtomComp.
The spectrometer is operated under the control of a dedicated mini-
computer. The appropriate spectrum line for each of the elements of
interest is isolated by a separate exit slit for each element and the
radiation thereby selected is transduced by a photomultiplier, the energy
output of which is converted to a voltage stored in a capacitor - one for
each channel. The computer addresses each channel of the analytical
system in turn very many times per second and the voltage on each capac-
itor, when above a pre-selected value, is converted to digits and stored
in a buffer. The capacitor is discharged and thereby ready to acquire
another voltage, and the digital conversion of this is added to the
previous value in core for the same channel. This procedure is continued
for each channel for the duration of the integration period, which most
commonly is set for 10 seconds. By these means the AtomComp readout
system allows more than 2,000,000 counts to be stored for each channel
over a 10-second integration period. Clearly this dynamic range matches
that of the argon discharge itself. This performance is displayed in
-------
-------
' -5-
receives a large volume wash between samples resulting in the analyti-
cally valuable short memory constant.
It has been demonstrated so far that the ionized plasma discharge is
under control for optimized performance; that the wide dynamic concentra-
tion range of the plasma is matched by the readout system; and that these
may be put to effective analytical use by assuring that the concentration
memory constant is 30 seconds or less. Before, however, the final ana-
lytical performance of the total system can be assessed, there are other
modules that also contribute to performance characteristics. Two of
importance are the quality of the spectrometer system as a light measuring
photometer, and the degree of control of energy of the argon plasma itself,
and the coupling of sample aerosol to it.
Every Jarrell-Ash AtomComp is routinely supplied with a means for
measuring the functional performance of the instrument as a photometer.
A white light is provided within the spectrometer, the radiation from
which is not dispersed by the grating but is seen by each of the photo-
multipliers of the system. Accordingly, an exposure to white light for
10 seconds will result in a summed number of digits for each channel.
The summation will be different from channel to channel depending upon
the gain factor of the photomultiplier, and other components of the
channel. Nevertheless, repeat runs of the 10-second integration period
to white light should result in close repeats of the integral in each
respective channel. With 10 separate observations of these values,
statistical treatment can be practiced. In the interests of clarity of
presentation, one set of evaluated data only are presented in Figure 6a.
For a particular channel the average of 10 consecutive runs of 10-second
exposure periods to white light gave 18,535 counts with a standard devi-
ation of 5.27 counts, thus to give a relative standard deviation of
0.0287,. Accordingly, the AtomComp, as a photometer, exhibits a measuring
precision in this instance of better than 1 part in 3000. The allowable
worst case for AtomComps operated under standard conditions is better
than 1 part in 1000.
The repeatability of the plasma as a light emitter could readily be
achieved by measuring the continuum or argon lines at one or several
-------
| 6 3.
-6-
wavelenths. The real interest, however, is not only the quality of the
argon plasma per se, but also the ability to couple a sample element to
it and the reproducibility that can be achieved in that process. The
difficulty, however, is that in the normal use the sample is nebulized
into the argon discharge, and measurements made by this means couple the
properties of such nebulization to that of plasma control without the
ability to separate the repeatability characteristics of each. This
problem was resolved by placing within the spray chamber of the nebu-
lizing system a small pool of Hg, the natural vapor from which was
carrier-borne by argon to the plasma, thereby excluding the aspiration
procedures. Precautions were taken to establish good mixing with the
argon carrier, and to assure that steady state conditions were approached.
The Hg thereby carried to the argon discharge coupled with it to promote
Hg emission and the Hg 2536 was used for the experiment. As with the
white light tests, ten consecutive exposures each of 10-seconds duration
were made and the data statistically evaluated. As displayed in Figure
6b, an average count of 129,309 was obtained and the standard deviation
calculated to be 61.95 counts or an R.S.D. of 0.048%. A second sequence
of exposures yielded an R.S.D. of 0.014%. It is clearly evident that the
supplies of electrical power and argon gas that promote the Inductively
Coupled Argon Plasma are under control to better than 1 part in 2,000.
Nevertheless, when real-life samples are nebulized into the plasma,
reproducibilities poorer than this are achieved. Indeed, the data pre-
sented in Figure 7 display that a concentration of 1 ppm can be deter-
mined to a precision of better than + 1% of that amount for most elements
and for some to better than + 0.5%. Although this is in practice a
highly satisfactory performance, it is poorer than the functional capa-
bility of the system by an order of magnitude. This circumstance indi-
cates that nebulization procedures are under less rigid control than
other aspects of the analytical system. Nevertheless, care should be
taken not to correlate closely the statistical data from the Hg experi-
ment with those from the 1 ppm im'tti-element solution, since the Hg
concentration in the plasma was some 60 times that of the other elements
in solution and improved statistics result therefrom. This, however,
-------
1° I
-7-
does not invalidate the broad conclusion that improved control of
nebulization is desirable.
3. Analytical Performance
It is evident from Figure 3 that the plasma discharge exhibits
structure with emission of atomic and ion spectrum lines spacially
separated. The use of a long entrance slit to admit all line emission
is unacceptable since unused atom or ion species enhance spectral
interference possibilities without compensating advantages. Thus a
choice has to be made for the segments of the discharge to be sighted
by the photomultipliers. In the final analysis the optimum is to have
independent selection of an atom or an ion line as a function of ele-
ment dependent upon the relative intensity, and spectral interferences.
To achieve this, multi-planar level selection of a stigmatic image of
the plasma would become necessary, but in practice this is an uneconomic
approach and a single compromise position proves in fact to be highly
satisfactory. Studies of the distribution of line species in the dis-
charge are best made with photographic records of the complete spectrum
emitted using the stigmatic properties of the Jarrell-Ash 3.4 Meter
Ebert Spectrograph. A short section of such a record displaying spatial
separation of the different line species is shown in Figure 8, and from
such studies an optimum plasma site for a given set of elements can be
selected. There is, however, another mode for achieving like information
which, although not able to display the detailed data yielded by Ebert
spectra, has an aspect of pictorial elegance as shown in Figure 9.
The figures were produced by photographing the plasma discharge
through a transmission grating to form a low dispersion slitless spectro-
graph. The upper turn of the radio-frequency coil shows as a horizontal
dark bar, and it will also be noted that this carries a Hg spectrum that
may be used as a wavelength marker. The spectrum was obtained by attach-
ing a small mirror to the upper turn of the R.F. coil and positioning a
Hg source so that its image was reflected along the camera axis and again
through the transmission £ rating. The spectral features of each compo-
nent of the figure are identified. Figure 9a is that of the Inductively
Coupled Argon Plasma with pu~e water nebulized and displays the continuum
-------
-8-
in the visible. The increased intensity in the deep red region is the
hydrogen alpha emission. When Na is nebulized (Figure 9b), the emission
of the D lines become apparent and it will be noted that such emission
is spatially separated from the continuum. The nebulization of the Ba
produces a more complex spectral structure, but one in which the spatial
separation of atom and ion lines are clearly delineated as shown in Figure
9c. Bnission from the yellow and blue Ba ion lines, 5853.7 II and 4934 II
respectively, occurs at a site well separated from continuum, whereas this
is not the case for the most intense segment of emission from the green
Ba atom line 5535 I. Weaker atomic emission at 5535 also occurs spatially
still higher in the plasma. There are no hard and fast rules on line
selection. When optimum detection limits are desired, as for example,
elements in pure water quality standards, the most intense line for the
element is selected providing it is free from interference from another
line and continuum. Under other conditions, however, there are other
considerations.
To be retained in solution, some elements require to be in an acidi-
fied matrix, sometimes with strengths up to or exceeding 107. acid, and
the question arises as to whether the acidity affects the linearity of
concentration response. Tests were conducted using solutions with the
same concentrations of many elements in each, but the solutions differed
in acidity from 0% to 107. HNO^. The resultant data are presented in
Figure 10 from which it will be noted that ion lines from most elements
retain linearity and are unaffected by acidity and, indeed, of the four-
teen elements in the current test program for which the ion lines were
selected, those of Ba and Pb alone depart from this desirable property.
The non-linearity that does occur is in the form of response suppression
at acid strengths above 2%. On the other hand, all elements determined
by the use of atom lines exhibit non-linearity by response expansion at
acid concentrations above 2%. Furthermore, the effect is more marked
than is the case for the ion lines that display non-linear characteristics
Spectrum line selection is dependent upon the application at hand,
that is to say the elements and ;he matrix in which the determinations
are to be made. For an aqueous matrix in which optimum detection limits
are required, Figure 11 provides a listing for 20 elements. The majority
-------
-9-
and in the ratio of 14 to 6, are ion lines. It may be predicted that for
some time to come there will be much discussion on line selection. In
Atomic Absorption there are few choices. In atomic emission, however,
there are many.
With essential modules of the analytical system under good opera-
tional control; with the best compromise of site selection in the plasma;
with optimized selection of spectrum lines for the application in hand;
it is prudent to determine the detection limits even though the nebuli-
zation procedure, as good as it is for most practical purposes, remains
as the least controlled of the system modules. Using the same procedures
(3)
reported in the paper of Fassel and co-worker , the present Jarrell-Ash
detection limits are presented in Figure 12 for 19 elements. They are
all in the small fraction of a ppm with Pb, Mo and Sn worse than other
elements by an order of magnitude. Detection limits are useful criteria
for the performance of a system, but should not be regarded as the only,
or even the most important of criteria. A system optimized for detection
limit performance would be unsuited for the analyses of impure materials,
and this particularly in regard to the design of the nebulizer system.
The system discussed above is not in that category. It is a system for
wide-ranged specimen applicability. Indeed the system as set up for
detection limit tests was also employed without alteration or adjustment
to the analysis of opal glass, and the data obtained are presented in
Figure 13.
A point of great analytical significance is that the calibration
curve for the pure water standards was also used for element determina-
tions in NBS #91 opal glass standard. The glass was digested in HF,
taken to dryness and the residue taken into a 17. HC1 solution. The agree-
ment between measured and published values are considered acceptable for
Al, Ca and Ti, and marginal for Fe, and unacceptable for Mg and Pb. The
causes for the poor performance of the last two elements, although not
determined, are thought to be associated with sample preparations1
procedures.
At this juncture it is desirable to redirect attention to Figure 7
which displays the precision achievable at element concentrations of
(3) V. A. Fassel and R. N. Kniseley, Anal. Chem. 49 (13), 1110A (1974)
-------
-10-
1 ppm and the analytical performance so displayed is highly satisfactory.
When lower concentration values are subjected to a like statistical
treatment, it is self evident that the precision worsens in proportion
to the approach of the concentration to the detection limit for the ele-
ment. This is revealed in Figure 14 which displays precision data for
0.1 ppm concentrations.
The Inductively Coupled Argon Plasma spectrochemical system is a
newly acquired analytical capability that is not only an extension of
Atomic-Absorption, but also of classical arc and spark spectroscopy.
The method enjoys detection limit capabilities equal to those of Atomic
Absorption, but is void of element restriction well known to that method
when the recombination time constant from atomic to molecular species is
too short for determinations to be made. Almost every element of the
periodic table is amenable to Plasma Spectroscopy and, importantly, the
method has simultaneous multi-element capability. The new method com-
plements classical arc-spark spectroscopy primarily by reason of the
reduction of matrix effects when solids are converted to weak solutions.
These are acceptable to the Inductively Coupled Argon Plasma by reason
of the innate one hundred to one thousandfold sensitivity improvement
of the method compared to arc and spark excitation.
4. Summary
The new method has much to commend it and the more important features
and the practical usefulness of them are enumerated hereunder.
1. Simultaneous multi-element determinations - up to 50 in number.
This endows an analytical system with calculable cost effective-
ness .
2. Linear dynamic concentration range of greater than 100,000
without adjustments, operator intervention or sample manipula-
tion. Thus, for example, the integration of from 3 to say.
2,100,000 counts in a 10-second exposure is made to represent
a concentration range of .001 ppm to 700 ppm, respectively.
This capability is valuable for very many applications, but
-------
-11-
especially so for geochemical samples, where an element may be
a major constituent in one mineral and a minor or trace in
another.
3. Dedicated computer control allows automatic element selection
for different matrices. Specimens of entirely different classes
can be run sequentially as, for example, a water effluent speci-
men for E.P.A. control, followed by a processed food specimen for
Production Quality Assurance purposes.
4. Memory of 5000 ppm element concentrations reduced to the 1 ppm
level in 30 wash periods. This enables the qualities enumerated
in 2 and 3 above to be put to practical utilization.
5. Sensitivity in the sub-ppm range. This enables 17. solutions to
be used even when an element is present in the solid at 0.0017.,
and often at 0.00017.. The volume of solution aspirated is less
than 1 ml and thus no more than 10 ml of solution is required,
and for this a mass of original sample of 0.1 g only is required.
Clearly, even for refractory materials, acid digestion of solids
and the subsequent preparation of solutions should not consti-
tute a major difficulty. Indeed, the limiting consideration is
the analytical validity of small sampling.
6. Nebulization system adapted from Jarrell-Ash Atomic Absorption
instruments. Aspirator is common to both Atomic Absorption and
Inductively Coupled Argon Plasma. Optimization of this compo-
nent has occurred over many years and is a continuing procedure.
7. Dynamic Background Correction is a standard provision in the
Jarrell-Ash system and is selectably usable for matrices that
produce background radiation. This feature is especially use-
ful when organic solvent extraction has been employed in sample
preparation.
8. Hard copy output c*ata in ppm values for high concentrations.
9. Concentration results on 30 elements in less than one minute.
10. Compatible with au'jmated sample presentation.
-------
I 0 2
-12-
Jarre11-Ash Plasma AtomComp System
Figure 1
Jarrell-Ash Plasma Torch
Figure 2
-------
!«i< »,.! :y'|Oil,.'
IteMgiisd is
Figure 3
:• :A; • ¦ jv-.- ;.•$, .
,11, -l
' ¦«
®ri ovet-Ag
4>
it a 1 pha
II alpha
Pure 1^0 Spectrum
5890T
58951
Sodium Spectrum
6'. 97 IT
5854IT 64991
4934IT
4554II 5535T 614211
Figure 9
Barium Spectrum
-------
-14-
1000
JAD PLASMA AT0MC0MP
LINEARITY OF Mn 2576
100
10
.1
.01
.1
.01
100
10
1000
PPM STANDARDS
Figure 4
-------
I (
-15-
1000
I
I
I
PPM
JAD PLASMA AT0MC0MP
MEMORY OF
Mn AT 1000ppm
.0063
.0232
.0084
30 60
120 180 240 300
TIME-SECONDS
Figure :>
-------
I I -X
-16-
ICAP ~ AtomComp Functional Performance Tests
10 runs of 10 second integration
(6a) WHITE LIGHT
Average (counts) 18535
S.D. 5.27
R.S.D. .028%
(6b) Test #1 Test #2
Hg Vapor in Plasma - 2 516 A
Average (counts) 129309 129242
S.D. 61.95 17.45
R.S'.D. .0487. .0147.
Figures 6a & 6b
Jarre11-Ash Plasma AtomComp
Precision at 1 ppm
(10 runs of 10 sec integration)
Element
RSD 7.
Element
RSD 7.
A1
0.45
MS
0.51
B
1.5
Mn
0.55
Ba
1.0
Mo
1.7
Ca
0.32
Ni
0.64
Cd
1.5
Pb
2.5
Co
0.94
Sn
3.1
Cr
0.50
Ti
0.32
Cu
0.28
V
0.40
Fe
0.37
Y
1.3
7. n
0.67
Figure 7
-------
II >
-17-
1^0
100 ppm Fe
Fe DC arc
Si 2631.28 I-
2714.40 II
2727.50 II-
2719.02
2720.90 I-
OH 2811.:
-------
10%
Ion Lines - Ba,Pb
Max. Effect 10%
0
a>
cr»
1 15%
o
a>
>
o
a>
a:
Ion Lines Unaffected-Ca.Cd.Co.Cr.Fe,Mg.Mn,Mo,Sn.Ti,V,Y
Atom Lines — Ag, Al, B, Cu,Ni,Zn
Max. Effect 15%
0
.1%
1%
HNO:
1
J
5% 10%
Figure 10
-------
Ili
-19-
Jarrell-Ash Plasma AtomComp
Analytical Lines
Element
X
Element
X
Ag
I
3280.68
Mg
II
2795.53
A1
I
3961.53
Mn
II
2576.10
B
I
2496.78
Mo
II
2038.44
Ba
II
2335.27
Ni
I
3414.76
Ca
II
3933.67
Pb
II
2203.50
Cd
II
2265.02
Sn
II
1899.90
Co
II
2388.91
Ti
II
3347.11
Cr
II
2677.16
V
II
3093.11
Cu
I
3247.54
Y
II
4177.54
Fe
II
2599.39
Zn
I
2138.56
Figure 11
Jarrell-Ash Plasma AtomComp
Detection Limits
(Line Signal 2 (S.D.) of background scatter)
Element
UK/ml
Element
UK/ml
Ag
.0024
Mg
< .0002
A1
.0022
Mn
.0052
Ba
.0060
Mo
.0340
Ca
.0015
Ni
.0070
Cd
.0150
Pb
.0380
Co
.0050
Sn
.•0120
Cr
.0108
Ti
.0012
Cu
.0053
V
.0018
Fe
.0036
y
.0074
Zn
.0090
Figure 12
-------
20
Jarrell-Ash. Plasma AtomComp
NBS #91 Opal Glass Standard
preparation: RF digestion - dried - take up in 1% HC1
A1 Ca Fe Mg Ti Fb
Solution %
average 2 aums 2.40 5.75 .035 .0105 .0085 .025
Calculated for
%
NBS Values 7.
A12°3
5.75
6.01
CaO
10.17
10.48
Fe2°3
.063
.081
MgO
.022
.008
Ti02
.018
.019
PbO
.034
.097
Figure 13
Jarrell-Ash Plasma AtomComp
Precision at 0.1 ppm
<10 runs of 10 second integration)
Element
R'SD 7,
Element
RSD '
A1
1.7
Mn
2.8
Ba
5.2
Mo
37.
Ca
0.78
Hi
4.0
Cd
10.
Pfa
32.
Co
1.7
Sn
21.
Cr
7.2
Ti
0.50
Cu
1.4
V
0.52
Fe
3.4
Y
5.1
Mg
1.4
Zn
3.6
Figure 14
------- |