EPA-600/4-77-032
June 1977
Environmental Monitoring Series
EVALUATION OF AN INDUCTIVELY
COUPLED PLASMA, MULTICHANNEL
SPECTROMETRIC ANALYSIS SYSTEM
Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Athens, Georgia 30601
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related tields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/4-77-032
June 1977
EVALUATION OF AN INDUCTIVELY COUPLED PLASMA,
MULTICHANNEL SPECTROMETRIC
ANALYSIS SYSTEM
by
R. K. Winge, J. M. Katzenberger, and R. N. Kniseley
Ames Laboratory, U.S. Energy Research
and Development Administration
Iowa State University
Ames, Iowa 50011
Contract No. EPA-IAG-D6-0417
Project Officer
Charles H. Anderson
Analytical Chemistry Branch
Environmental Research Laboratory
Athens, Georgia 30605
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GEORGIA 30605
-------
DISCLAIMER
This report has been reviewed by the Athens Environmental
Research Laboratory, U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S.
Environmental Protection Agency9 nor does mention of trade names
or commercial products constitute endorsement or recommendation
for use.
ii
-------
FOREWORD
Nearly every phase of environmental protection depends on
a capability to identify and measure chemical pollutants in the
environment. The Analytical Chemistry Branch of the Athens
Environmental Research Laboratory develops techniques for
identifying and measuring chemical pollutants in water and soil,,
A relatively new instrumental technique for the rapid,
quantitative analysis of water for approximately two dozen trace
elements utilizes inductively coupled plasma emission spectros-
copy0 This report presents an initial evaluation of one
commercial instrument that embodies the technique„
David W. Duttweiler
Director
Environmental Research
Laboratory
Athens, Georgia
iii
-------
ABSTRACT
An inductively coupled plasma, multielement atomic emission
spectrometric (ICP-MAES) analytical facility was evaluated with
respect to the requirements of the Environmental Protection
Agency for the determination of trace concentrations of the
elements in water. Data are presented on detection limits,
curve fits and reproducibility of analytical curves, photo-
multiplier performance, dynamic range, reproducibility of
analytical results, and stray light properties of the instrument.
The instrument as a whole performed well, but stray light
arising from calcium and magnesium in typical hard waters caused
significant errors at low concentration levels for a number of
critical elements. Suggestions for reduction of the stray light
and for empirical correction of its effects are discussed.
This report was submitted in fulfillment of Contract No.
EPA-IAG-D6-0417 by the Ames Laboratory under the sponsorship of
the U.S. Environmental Protection Agency. This work covers the
period June 24, 1974, to January 28, 1976, and work was completed
June 17, 1976.
iv
-------
CONTENTS
Foreword . . . . iii
ADS traCTI ooooooooooooeooo IV
1- 1-^UL C S eooooooooooooooo VI
-LclDiCS eeaooeoeeoooeoooo VH
Abbreviations ............. viii
Acknowledgment . viii
1. Introduction .000 1
2, Conclusions 2
3o Recommendations . „ 3
4. Experimental Equipment, Software, and
Procedures ............. 5
Apparatus .,,.,.00.0., 5
Software 7
Operating procedure .,00.000 9
Warm-up time of ICP ........ 10
-J o IVw b U. .L L. O oooo-oooooooooo \-£*
Detection limits . 12
Photomultiplier performance ...... 12
Analytical curves ......... 15
Dynamic range .......... 19
Reproducibility of calibration 22
Short term reproducibility 22
Stray light" 25
Solution of stray light problems 35
6. Discussion 40
Dynamic range 40
Detection limits .. 0 ....... 40
Software •„ 40
Stray light 41
7o Publications 43
•KGiC .LwTlCG S ooooooooooooooo^"^"
V
-------
FIGURES
Number PaSe
1 Examples of stabilization curves for deionized
water background on ICPQ 11
2 Examples of wavelength profiles that show the
effects of stray light caused by concentrations
of calcium and magnesium typical in hard water.
0 deionized water, A analyte mixture containing
50 jjtg/^ Cr and 100 ng/£ Sb, + hard water matrix
(150 mg/£ Ca, 40 mg/-l Mg) , X analyte mixture as
above in hard water matrix 26
3 Wavelength profiles of the Cd 226.50 nm line.
The profiles were obtained from four surface
water samples containing different concentrations
of calcium and magnesium (see Table 10).
0 deionized water, + Scott Pond, X Site 10,
0 Site 21, * Site 53, A 200 |ag/t Cd reference
standard ^'
4 Examples of stray light arising in the secondary
optics region of the QA 137 polychromator. All
exit slits were blocked except that for the Mg
279.55 nm line.
0 deionized water, A 1000 mg/£ Mg,
+ 1000 mg/£ Ca 29
5 The secondary optics region of the QA. 137 poly-
chromator showing the exit slit frame, the secon-
dary mirrors and photomultipliers 31
6 Examples of general stray light arising from
general scattering within the polychromator
enclosure. All exit slits were blocked except
that for the Mg 279.55 nm line. Curve symbols
same as on Figure 4 32
7 Geometrical arrangement of secondary optics in
the region of the Mg 279.55 nm line 36
8 Stray light correction coefficients as a function
of calcium concentration.
X As 193.76 nm, A Se 196.09 nm,
0 Pb 220.35 nm, D Sb 206.84 nm J8
VI
-------
TABLES
Number
1 QA 137 Spectral Line Array ........ 6
2 Detection Limits (ng/O ..... .... 13
3 Photomultiplier Dark Current Data ...... 16
4 Examples of Analytical Curve Coefficients and
Error Data ............. 17
5 Examples of Information Given for the Individual
Calibration Points for Each Analytical Curve . . 18
6 Examples of Data From Analytical Curves Covering
a Four Decade Range in Concentration ..... 20
7 Reproducibility of Calibrations for the Zn 213.86
nm Channel Before and After Delivery of the ICPQ
System to the Ames Laboratory ....... 23
8 Reproducibility of Analytical Data From a
"Reference Standard" Run at Regular Intervals
During a Normal Analytical Sequence ..... 24
9 Measured Calcium and Magnesium Concentrations From
Four Surface Water Samples ........ 28
10 Stray Light Concentration Equivalents
From Calcium ............. 33
11 Stray Light Concentration Equivalents
From Magnesium ............ 34
12 Classification of Spectral Lines With Respect to
Stray Light Problems as Observed on the QA 137
Polychromator ............ 42
vii
-------
LIST OF ABBREVIATIONS
ARL -- Applied Research Laboratories
cm -- centimeter
CRT -- cathode ray tube
ICP -- inductively coupled plasma
ICPQ -- inductively coupled plasma quantometer
ICP-AES -- inductively coupled plasma, atomic emission
spectrometry
kW -- kilowatt
t/min -- liters/minute
MHz -- megahertz
Ug/£ ~~ microgram/liter
|jt -- microliter
mg/t -- milligram/liter
mm -- millimeter
mV -- millivolt
riA -- nanoampere
nm -- nanometer
PVC -- polyvinyl chloride
RF -- radio frequency
RMS -- root mean square
SD -- standard deviation
sec -- second
V -- volt
W -- watt
ACKNOWLEDGMENT
The authors wish to express their appreciation to
Dr. William Haas for helpful discussions and for the development
of the computer plotting routines which materially aided the
elucidation of the stray light properties of the QA 137 spec-
trometer.
VL1L
-------
SECTION 1
INTRODUCTION
To fulfill its environmental monitoring responsibilities,
the EPA needs methods for the determination of elemental con-
centrations in a wide variety of natural and polluted waters.
The nature of the water samples, the number of analyses
required, and the need for reliable analytical data, often at
very low concentration levels (1), dictate stringent require-
ments for an ideal analytical method. 1) The analytical method
must have high powers of detection. 2) Potential sources of
contamination (e.g. from sample pretreatment or concentration
procedures) must be minimized. 3) Because 20 or more elements
may need to be monitored in large numbers of samples on a
routine basis, the method should be rapid and be able to deter-
mine the required elements simultaneously. 4) Because of the
wide range of sample compositions that may be encountered in
natural and polluted waters, the analytical method should
possess minimal interelement or other matrix effects.
The ICP-AES method meets more of these requirements than
any other technique (2, 3). The general properties of ICP
methods have been described previously (3-8). The freedom from
interelement effects of ICP-AES methods have been described by
Larson, et al, (9).
This report summarizes the work performed in our laboratory
on an ICP-AES system obtained for the Athens Environmental
Research Laboratory, United States Environmental Protection
Agency, Athens, Georgia. A detailed set of specifications (10)
for and ICP-AES instrument, appropriate for the needs of the EPA,
was prepared and along with a request for quotation was sent to
all manufacturers (vendors) of such instrumentation. Two
vendors responded with quotations. The analytical performances
of the instruments from these vendors were then evaluated in
the vendor's laboratories. On the basis of the quotations and
the initial evaluations, a choice was made to purchase the
Applied Research Laboratories (ARL) Inductively Coupled Plasma
Quantometer (ICPQ). This instrument was tested at the Ames
Laboratory for a period of about three months and was then
transferred to the EPA Laboratory at Athens.
-------
SECTION 2
CONCLUSIONS
An inductively coupled plasma, multielement atomic emission
spectrometric analysis system has been evaluated with respect to
the Environmental Protection Agency's need for a rapid method
for determination of trace elemental concentrations in water.
The data acquisition system of the ICPQ instrument in general
performed well. The detection limit capabilities, the analyti-
cal curve fitting, the analytical curve reproducibility, and the
photomultiplier performance approached the state-of-the-art for
the types of equipment involved. A signal measuring system with
a wider dynamic range capable of exploiting the wide range
characteristics of the ICP would be desirable. A reduction in
warm-up time of the system would also be desirable.
The most serious deficiency in the total system, however,
is the polychromator in which high stray light levels may cause
significant analytical errors, especially for such critical
elements as arsenic, lead, and selenium.
With major reductions in stray light levels, the develop-
ment of empirical methods for correction of the remaining
matrix-related background changes, and improvements in nebulizer
performance, all of which are presently feasible, the ICP-MAES
technique is still the most practical approach for rapid multi-
element determinations of trace concentrations of the elements.
-------
SECTION 3
RECOMMENDATIONS
The instrument under test for this evaluation required more
than % hour warm-up time. A significant reduction of warm-up
time would be desirable in future generations of these instru-
ments, especially for short analytical sequences requiring only
a fraction of an hour to several hours time.
Printout of the data is often the most time consuming part
of the analytical procedure. For large sample loads a faster
printer and disc virtual memory facilities should be considered.
A stepper-motor drive on the entrance slit profile screw of
the polychromator would facilitate wavelength profiling on the
instrument. This wavelength drive motor should be controlled by
the system computer.
A CRT display with graphic capabilities would be helpful
for rapid interpretation of wavelength profiles and other data.
Methods for improvement of the stray light properties of
spectrometers need to be evaluated. Such methods include:
1. improved gratings (e.g. holographically recorded
gratings),
2. better engineering of spectrometers including baffles,
light traps, non-reflective coatings, and general
optical design,
3. interference filters including narrow bandpass trans-
mission and band rejection types.
The relative merits of stray light correction techniques
need to be evaluated. These techniques include:
1. wavelength modulation,
2. wavelength profiles,
3. application of empirically determined interference
correction factors or correction curves.
-------
A reduction in clean-out time of the nebulization system
would be useful especially when wide ranges in elemental con-
centrations may be encountered. Rapid methods for flushing of
the sample uptake tube and nebulization chamber need to be
considered.
A scanning monochromator would be a useful addition to the
multichannel analysis system. The monochromator could be used
as an alternate channel of variable wavelength and would be
most useful if its analytical signals could be processed through
one of the measurement channels of the multichannel system.
The monochromator should also be usable as a scanning instrument
for wavelength recordings.
-------
SECTION 4
EXPERIMENTAL EQUIPMENT, SOFTWARE, AND PROCEDURES
APPARATUS
The main features of the analytical system consist of:
1. The ARL-ICPQ polychromator, equipped with 31 exit
slits for the elements and wavelengths listed in
Table 1. The grating is blazed for ^ 300 run. The
polychromator covers the wavelength range
189.5 - 461.5 nm. Entrance slit, 12 pm; exit slits,
50
2. Plasma radio frequency (RF) power supply: Standard
ARL unit, 27 MHz, 3 kW rating. The specifications
required that the torch assembly be adjustable over
its full vertical and horizontal range without change
in the impedence match between the coupling box and
the load coil. ARL accomplished this by combining
the load coil and coupling box as a single movable
unit.
The considerable heat from the power amplifier
section of the RF generator was exhausted through
a fume duct in accord with ARL recommendations.
Heat and gaseous products from the plasma source
were also vented into the same hood system.
3. Sample generation and excitation hardware: The
plasma torch, nebulizer, nebulization chamber
(similar to Scott return flow type (5)), and gas
flow control system were all standard items supplied
by ARL. Nebulizer uptake tube: 60 cm of 0.75 mm i.d.
polyethylene with a 50 M,£ capillary pipet attached to
the intake end by means of a short section of heavy-
walled PVC tubing. Liquid argon, supplied in cylin-
ders containing 200 kg net argon, was employed for
the plasma and nebulization systems.
-------
Table 1. EPA QA 137 LINE ARRAY
Element Wavelengths (nm)
Ag I 328.07
Al I 308.22 I 396.15
As I 193.76
B I 249.68
Ba II 455.40
Be II 313.04
C I 247.86
Ca II 315.89
Cd II 226.50
Co II 238.89
Cr II 267.72
Cu I 324.75
Fe II 259.94
Hg I 253.65
Mg II 279.55
Mn II 257.61 I . 403.08
Mo II 287.15
Ni II 231.60
Pb II 220.35 I 405.78
Sb I 206.84
Se I 196.09
Sn I 303.41
Sr ii 407.77
Ti II 334.94
V II 292.40
Y II 242.22 II 371.03
Zn I 213.86
I Spectral line originates from neutral atom state.
II Spectral line originates from singly ionized state,
Number of elements 27
Number of wavelengths 31
Atom lines 14
Ion lines 17
-------
4. Data acquisition system:
Photomultipliers Hamamatsu R300 series
Signal measurement Capacitive time-integration
technique of photocurrents
Minicomputer DEC PDF 11/05
Data output Teletype ASR 33 printed
hardcopy with or without
punched paper tape output.
SOFTWARE
Because of the copious data that can be generated by a
rapid multielement analytical scheme such as the ICP-AES method,
the computer software becomes a very important factor in the
total system. Not only must the software produce correct
analytical results from appropriate data, but it must supply
cogent information about such things as system stability and
analytical curves so that the analyst can quickly judge the
quality of the data.
At the time the specifications were written for the EPA
instrument, software packages available with commercial direct
reading spectrometers were oriented largely for the needs of
the iron, steel, and aluminum industries and, in our opinion,
were not adequate for trace element analytical problems. More
appropriate software capabilities were therefore specified for
the EPA system. The specifications requested the following:
1. Ability to correct the reference standard and sample
data for the blank contribution so that valid
analytical curves could be established and accurate
analyses could be performed. Merits of the various
blank correction techniques have been discussed by
Winge, et al.(2).
2. Option of printing blank corrected intensities with
the concentration data in the analytical mode of
operation.
3. Provision for weighting of calibration points to
balance their relative influence on the determination
of the analytical curve. This provision is particu-
larly important for analytical calibration curves
that cover greater than two orders of magnitude in
concentrat ion.
4. Ability to print analytical curve parameters, standard
deviations of the curves, and residuals of the individ-
ual calibration points with respect to the curves.
-------
5. Ability to gather "n" repetitive intensity measure-
ments and to calculate average values and standard
deviations from these data.
6. Ability to correct analytical data for interference
contributions.
7. Ability to ratio analyte intensities to internal
reference element intensities.
A second order, least squares curve fitting routine is
employed for calculation of the analytical curves represented
by the following equation
Y = AQ + AjX + A2X2 (1)
where Y = calculated concentration,
X = net measured intensity,
AQ = intercept on the concentration or Y axis,
AI = slope of the curve at X = 0,
A2 = coefficient of the second-order term.
Examples of analytical curve information from the ICPQ
system will be explained later in Section 5. Theory of the
analytical curve will be briefly described here.
Ordinarily an analytical curve should intersect the
coordinate system very close to its origin. Intercept values
(concentration axis) significantly larger than detection limits,
either positive or negative, may be caused by improper blank
corrections. Residual concentrations of elements in the
reference standards cause negative intercepts, while contamina-
tion of the reference blank yields positive intercepts. Drifts
in instrument sensitivity during the calibration procedure also
may produce positive or negative deviations in an intercept,
particularly if the calibration is based on a single initial
reference blank determination.
The coefficient of the first order term, A^, in Equation 1
corresponds to the slope at zero intensity and because it
corresponds to concentration per unit intensity the instrumen-
tal sensitivity for an element increases as the slope decreases,
If the analytical curve passes through or near to the origin
of the coordinate system, then A^ is effectively the slope at
infinite dilution.
It is appropriate to note here that a weighting scheme was
included in the curve fitting routine of the ICPQ software at
our request. The weighting scheme, by minimizing relative con-
centration errors, produces analytical curves more appropriate
for the wide dynamic range characteristics of ICP-AES methods.
8
-------
Without weighting factors a wide-range analytical curve is
determined largely by a few points at the high concentration end
and large relative errors may result at the low concentration
end of the curve.
The linearity of the analytical curves produced by the ICPQ
system is characterized by the quantity L defined by
L = 1 - A2X2/y = AQ/y + A^/y. (2)
The latter equality in the above equation can be obtained
directly from Equation 1. For a linear relation of the vari-
ables, A~ = 0 and L = 1. For a nonlinear relation, L gives the
fraction of the calculated concentration due to the linear
elements (zero and first order terms) of the calibration curve
equation.
OPERATING PROCEDURE
In general ARL recommendations were followed with regard to
the starting and operating procedures.
Unless noted otherwise the following operating conditions
were used:
RF forward power 1100 W indicated
Reflected power <2 W indicated
Plasma Ar 16 £/min
Auxiliary plasma Ar 1.4 £/m
Aerosol carrier Ar 1.0 £/m
Integration period 10 sec
Observation zone in plasma limited to region approxi-
mately 13 to 17 mm above the load coil by a 4 mm vertical
aperture at the entrance slit.
Photomultiplier voltage: Attenuator position 8-5 (— 750 V)
was normally used for all channels unless high elemental
concentrations in a sample series necessitated voltage
reductions for specific channels.
Because a second order curve fitting routine is used, at
least four reference standards should be used for calibration
of the system. Four to six reference standards were normally
used for studies involved in this evaluation. Also, because of
the second order curve fit, analytical results outside of the
calibrated range for any element should be checked carefully.
-------
Cross contamination of solutions was minimized by immersion
of the sample uptake tube in a dilute acid wash solution for a
few seconds between samples or between sample and reference
blank solutions.
WARM-UP TIME OF ICP
A significant bias in analytical results may be produced
by insufficient'warm-up time. On several occasions, even after
the RF generator had been on for 40 to 60 minutes, relatively
small changes in background levels during calibration caused
significant errors at low concentration levels. On these
occasions, however, the system had stabilized by the time
analyses were begun, because periodic analyses of a reference
standard over a several hour period yielded precision values
approximating detection limit levels.
The stabilization process is shown by intensity-time
profiles for three channels in Figure 1. These profiles and
those from other channels not shown here suggest that the warm-
up of the power supply and of the components in the plasma
enclosure dominate the first portion of the curves, but that
individual differences in the channels begin to appear in the
latter portions of the curves. It should be noted that a com-
puter smoothing process was employed for the plots in Figure 1
and that the ordinate scaling factors differ from plot to plot.
10
-------
As (I) 193.76 nm
Cd (II) 226.50 nm
Cr (II) 267.72 nm
1
1.80
CO
1.70
UJ
cr
1.60
0 10 20 30 '40
0 10 20 30 40
TIME AFTER START , min
18.0
7.0
I6.0r
15.0.
0 10
30 40
Figure 10 Examples of stabilization curves for deionized water background on
ICPQ.
-------
SECTION 5
RESULTS
DETECTION LIMITS
Detection limits obtained periodically from the QA 137
system under several sets of operating conditions are shown in
Table 2, The detection limit is defined as the analyte concen-
tration equivalent to two times the standard deviation of the
background beneath the analyte line. In general, if data from
the same wavelengths are compared under the standard operating
conditions (Condition Set V, Table 2), detection limits from
the ICPQ system compare well with those obtained from other ICP
systems in the Ames Laboratory (3). The significantly poorer
detection limits obtained with Condition Set IV, without
auxiliary plasma gas flow, are surprising because experience in
our laboratory indicates this gas flow is usually not benefi-
cial other than during the startup of the ICP.
PHOTOMULTIPLIER PERFORMANCE
Photomultiplier performance is a critical factor in the
capabilities of a spectrometric analytical system such as the
ICPQ,
Power Supply Stability
A commonly accepted criterion is that a photomultiplier
high voltage supply should be a factor of ten more stable than
the stability required of the photocurrent signal (11). There-
fore, if 0.170 stability is expected in the photocurrent, the
power supply should be stable to within 0.0170. The ICPQ photo-
multiplier power supply voltage was measured nine times (at one
hour intervals) with a five-digit digital voltmeter. The
results: The average value equaled 997.96 V + 0000370o
Dark Current
Photomultiplier dark current and the noise on the dark
current may become significant in the measurement of very low
signal levels. The specifications called for photomultiplier
dark currents not to exceed 0,1 nA and relative standard
12
-------
CO
Table 2. DETECTION LIMITS (yg/1), ICP-QA 137 SYSTEM
12345678
Ag
Al
Al
As
B
Ba
Be
Ca
Cd
-Co
Cr
Cu
Fe
Hg
Mg
Mn
Mn
Mo
Ni
Pb
Pb
Sb
Se
Sn
Sr
Ti
V
Y
Y
Zn
328.
308.
396.
193.
249.
455.
313.
315.
226.
238.
267.
324.
259.
253.
279.
257.
403.
287.
231.
220.
405.
206.
196.
303.
407.
334.
292.
242.
371.
213.
07
22
15
76
68
40
04
89
50
89
72
75
94
65
55
61
08
15
60
35
78
84
09
41
77
90
40
22
03
86
2
—
7
200
4
3
3
6
1
2
20
.
10
6
9
30
60
10
40
70
10
5
5
8
3
2
1
5
09
8
4
2
80
8
100
.5
.2
.2
2
2
4
6
1
2
20
.1
.6
8
7
10
50
50
40
60
80
.04
10
1
5
.2
6
60
7
50
4
.2
.06
1
1
2
2
2
1
20
.6
.3
6
7
4
20
70
20
20
80
.1
6
5
4
5b
3
._
40
5
40
4
.2
.4
8
3
3
3
.5
1
10
1
.2
1
6
7
20
30
20
60
50
.03
3
2
10L
3b
4
__
120
10
180
10
.6
.6
10
10
7
1
2
10
30
7
.5
2
10
50
230
120
150
380
120
.02
10
3
30
.3
30
._
20
3
60
5
.7
.2
6
.5
4
5
.7
2
20
.9
.9
9
10
4
40
40
40
50
80
.9
6
1
30
.9
4
w —
20
4
60
5
.5
.4
20
4
4
2
1
210a '
20
530a
1
5
10
8
40
30
40
50
60
.5
5
1
10
.6
2
2
20
10
70
4
.1
.4
10
3
2
4
1
3
20
.1
.3
3
10
8
30
40
40
40
60
.05
5
2
20
.4
3
-------
Table 2. Continued
Condition set I
ARL Ames
Sunland Lab
CA 9/10/75
7/21/75
II
Ames
Lab
10/16/75
III
IV
V
V
V
Ames Ames Ames Ames EPA
Lab Lab Lab Lab Athens
10/16/75"10/16/75 12/4/75 12/31/75 1/26/76
aVery high concentrations of Fe and Mg had been run previous to the background
measurements for this data set. The acid blank apparently scavenged deposits of
these elements from the nebulizer, nebulization chamber or sample injection tube
of the torch.
bFor unknown reasons the standard deviation of the background for the Y 371.03 nm
channel was unusually high for these measurement sequences.
'Explanation of operating conditions
I
HF power (W)
Plasma Ar flowrate (£/min)
Auxiliary plasma Ar
flowrate (&/min)
Aerosol Ar flowrate (2,/min)
Observation height (mm)
II
Condition Set
III
IV
V
1600
11
1.4
1
20
1600
16
1.4
1
18
1100
16
1.4
1
18
1100
16
none
1
18
1100
16
1.4
1
15
-------
deviations not to exceed 3% of the dark current for an integra-
tion period of 10 seconds (photomultiplier voltage at midrange
or 750 V). The data in Table 3 show that none of the dark
currents exceeded 0.1 nA (0.1 nA « 1 mV) but that all of the
standard deviations exceeded 3%. The latter problem is due
predominantly to the uncertainties introduced by the measurement
electronics rather than to noise on the dark current (see
Section 5, Dynamic Range).
Detection Limits vs. Photomultiplier Voltage
Photomultiplier performance was measured in terms of
detection limits as a function of applied voltage. Most of the
photomultipliers yielded their best detection limits within the
700-800 volt range. The poorest detection limits obtained with-
in the 630-950 volt range were often poorer by a factor of four
or greater than those at the optimum voltage. These data were
collected after the ICP had warmed up for one hour. The system
was allowed to stabilize for 10 minutes after each voltage
change before data were collected.
ANALYTICAL CURVES
Computer listings of data collected on a typical set of
analytical curves generated by the ICPQ system are shown in
Tables 4 and 5. The curve coefficients and root mean square
(RMS) error data can be retrieved from storage at any time by
a simple keyboard command as can the error table data. The
former were usually retrieved immediately after a calibration
was completed. The individual element error tables were usually
not printed until the end of a run or the end of a day because
approximately 30 minutes printing time was required for 30
channels.
As noted in Section 4, under ordinary circumstances the
intercept values for the analytical curves should be very small.
The data in Table 4 show, for example, that the intercepts are
usually at detection limit levels or below. The AO column
yields base solution concentrations directly when samples are
analyzed by the standard addition technique, that is, if valid
blank corrections have been applied.
The RMS error data in Table 4^are calculated from the con-
centration and intensity residuals* and pertain only to the
*An intensity residual is the difference between a measured
intensity and the intensity derived from the analytical curve
for a given concentration. A concentration residual is the dif-
ference between the concentration calculated from the analytical
curve and the concentration assigned to the reference standard.
15
-------
Table 3. PHOTOMULTIPLIER DARK CURRENT DATA'
Rfb
Al
Ba
Ca
Cr
Hg
Mn
Pb
Se
Ti
Y
Ave .
Value
(mV)
5199.
0.4190
0.1920
0.3020
0.1259
0.0130
0.1890
0.1160
0.1740
0.3489
0.2689
% SD
0.0106
11.33
14.06
9.208
18.40
162.3
14.20
23.41
27.38
10.32
10.28
Ag
As
Be
Cd
Cu
Mg
Mo
Pb
Sn
V
Zn
Ave .
Value
(mV)
0.1250
0.1940
0.0629
0.0470
0.2009
0.1660
0.1989
0.3359
0.0270
0.3040
0.3979
% SD
29.75
12.90
42.36
34.81
. 12.94
18.00
11.47
12.24
100.3
15.05
5.530
Al
B
C
Co
Fe
Mn
Ni
Sb
Sr
Y
Ave .
Value
(mV)
0.2030
0.0290
0.4179
0.3220
0.1249
0.6369
0.0279
0.0200
0.0599
0.0699
% SD
22.15
61.79
6.841
6.835
25.36
10.09
116.3
102.7
35.13
46.16
Average values and precision data obtained from ten consecutive ten-second
integrations of the dark current.
•^Reference channel (Rf) for which the input signal is derived directly from the photO'
multiplier high voltage supply.
-------
Table 4. EXAMPLES OF ANALYTICAL CURVE COEFFICIENTS AND ERROR DATA
Channel
As
B
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
Hg
Mg
Mn
Mn
Mo
Ni
Pb
Pb
Sb
Se
RMS Error
Concentration Intensity
mg/£ mV
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
051360 •
001892
003180
000698
004866
002188
039864
021292
008283
049013
011414
018167
011354
020617
023022
108347
034863
019660
019597
015647
0
0
0
0
0
.
2
1
1
1
1
1
t
078687
038986
.23229
.56750
112320
056116
.50954
.81339
.68516
.81559
064180
13.2167
0
0
0
0
0
4
1
1
.
t
.33669
.66851
584106
.70932
086066
077925
038085
034128
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
Curve Coefficients
A2 A-L (slope) AQ(intercept)
mg/2,(mV)"2 mg/A(mv) mg/£
16543269E
19538216E
10950316E
32466687E
19574538E
15552156E
70346869E
68437657E
17270213E
230-30443E
59943646E
27351234E
37298378E
14189282E
66141365E
56251884E
52168965E
18452108E
32751262E
48753619E
-3
-4
-7
-8
-4
-4
-6
-6
-6
-5
-4
-7
-7
-5
-6
-5
-3
-3
-3
-2
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
662933
050018
001440
000455
045164
039947
027913
012182
005290
028390
192077
001420
002777
013123
040216
066741
458038
268478
524383
474166
-.024166
-.002081
-.000884
-.000026
-.000957
-.001562
-.003232
-.001634
-.000992
-.005492
-.017805
-.001267
-.000193
-.003429
-.008370
0.001007
-.025443
-.013218
-.008967
-.006567
-------
Table 5. EXAMPLES OF INFORMATION GIVEN FOR THE INDIVIDUAL
CALIBRATION POINTS FOR EACH ANALYTICAL CURVE
CO
Be
Error
Table
(Be 313.04 nm)
Mo
Error
Table
(Mo 287.15 nm)
Pb
Error
Table
(Pb 220.35 nm)
Std.
No.
01
02
03
04
05
06
01
02
03
04
05
06
01
02
03
04
'05
06
Cone .
mg/Jt
0.003999
0.012000
0.040000
0.119998
0.399993
2.00000
0.099998
0.300003
1.00000
3.00000
10.0000
50.0000
0.099998
0.300003
1.00000
3.00000
10.0000
50. .0000
Int.
mV
8.72314
26.9501
85.9335
264.859
873.187
4255.50
2.59338
7.66674
25.5000
74.5117
246.875
1219.87
0.286666
0.713348
2.18994
6.54321
21.5068
98.1406
Cone. Res.
mg/s.
0.000048
-.000264
0.000821
-.000964
-.000581
0.000954.
0.004067
0.000005
-.017579
0.008101
0.039623
-.034912
-.005905
-.001560
0.019864
0.006065
-.066816
0.048578
Int. Res.
mV
0.106238
-.579433
1.80482
-2.12298
-1.29175
2.22762
0.101154
0.000129
-.437478
0.201948
0.993333
-.904414
-.012901
-.003411
0.043585
0.013442
-.1533.90
0.136593
L
0.999938
0.999803
0.999400
0.998102
0.993811
0.970602
0.999955
0.999870
0.999569
0.998775
0.995968
0.980315
0.999571
0.999115
0.997498
0.992554
0.975869
0.899506
-------
calibration curves. Sample analyses performed after calibration
may be subject to additional errors which cannot be estimated
from the analytical curve data. Because the RMS error values in
Table 4 do not apply to specific concentrations or intensities,
we feel that relative residuals provide a better basis for cal-
culation of the uncertainty in a calibration curve. A relative
residual is the absolute residual divided by the measured or
assigned value for each calibration point. Expressed as a per-
centage the resulting relative RMS or standard error would then
provide an estimate of the error associated with any point on
the calibration curve.
The concentration and intensity residuals in Table 5 show
that the calculated curves fit the calibration points very
closely. If a concentration error has been made in a reference
solution it will produce relatively large residuals. The L
values for the beryllium and molybdenum calibrations in Table 5
show that the deviations of these points from a linear relation
with concentration are hardly significant. The lead data show
a more significant deviation from linearity. If self-absorption
is the cause of the non-linearity, then the L values will
naturally decrease with increasing concentration. In all of the
calibrations performed with the ICPQ system, very few calibra-
tion points had L values less than 0.85-0.90. Most had L
values in the 0.97-1.0 range.
DYNAMIC RANGE
Utilization of the wide dynamic range capabilities of the
ICP excitation source (5, 12) naturally requires wide dynamic
range in the signal measurement system. The ICPQ readout
system has an- absolute dynamic range of 0.01-10,000 mV. The
practical dynamic range is limited, however, by the noise
characteristics of the signal processing electronics. For
example, with a 10 mV signal, the electronics contributed an
uncertainty of about + 0,5% while at 1 mV the uncertainty was
+ 3-470. If + 17o uncertainty is accepted as a reasonable objec-
tive for analytical results, then the practical dynamic range
of the ICPQ is limited to 3-3^ orders of magnitude. The con-
centration interval to which this practical dynamic range
applies can be shifted moderately, of course, by changes in
photomultiplier voltage.
Examples of analytical curve data for a four decade range
test are shown in Table 60 In general, these data indicate
that acceptable curve fits, intercepts, and linearity factors
have been achieved. The linearity factors must be interpreted
with respect to the highest reference standard concentrations
which for the data in Table 6 range up to a factor of 10"
above the detection limits.
19
-------
Table 6. EXAMPLES OF DATA FROM ANALYTICAL CURVES COVERING
A FOUR DECADE RANGE IN CONCENTRATION3
N>
O
Channel
Mn
Cu
Al
Mn
Mn
257.61
nm
Cu
324.75
nm
0
0
0
0
Std.
No.
01
02
03
04
05
01
02
03
04
05
RMS Error
Cone.
mg/J,
.144782 7.
.036428 3.
.141207 2.
.063146 2.
Cone.
mg/£
0.024999
0.250000
2.50000
25.0000
250.000
0.010000
0.099998
1.00000
10.0000
100.000
Cur
Int. A2
mV mg/£(tnV)~2
21684 2.
44151 7.
45126 2.
24193 1.
Int.
mV
0.959991
10.9599
108.988
1070. 18,
9930. 00b
0.660003
8.91015
89.3710
883.375
8514.00
3842E-7
6710E-8
2554E-7
3114E-7
Cone. Res.
mg/£
0.000300
-.003664
-.000645
0.230663
-.227136
0.000622
-.001718
-.002743
0.060111
-.054871
ve Coefficien
A^(slope)
mg/UmV)'1
0.022893
0.011191
0.058717
0.028729
Int . Res .
mV
0.013141
-.160112
-.028245
10.2988
-12.4226
0.055632
-.153587
-.245385
5.42731
-5.44771
ts
AQ(intercept)
mg/£
0.002720
0.001990
0.008106
0.001168
L
1.000002
0.999885
0.998867
0.989078
0.905962
0.999997
0.999939
0.999387
0.994014
0.944394
-------
Table 6. Continued
Al
396.15
nm
Mn
403.08
nm
Std.
No.
01
02
03
04
05
01
02
03
04
05
Cone .
rng/£
0.049999
0.500000
5.00000
50.0000
500.000
0.024999
0.250000
2.50000
25.0000
250.000
Int.
mV
0.609985
8.62011
85.0390
846.062
8374.00
0.649993
8.87988
87.4218
864.750
8537.00
Cone . Res .
mg/£
0.006075
-.014269
-.002284
0.226162
-.219779
0.005157
-.006287
-.013255
0.105510
-.092536
Int. Res.
mV
0.103470
-.243025
-.038924
3.86519
-3.87718
0.179520
-.218857
-.461588
3.68725
-3.35296
L
0.999998
0.999966
0.999674
0.996771
0.968368
0.999998
0.999959
0.999599
0.996077
0.961770
Operating conditions as in Set I, Table 2.
'This intensity value is very close to the upper limit of the digital voltmeter
and, in fact, may have reached the actual readout limit.
-------
REPRODUCIBILITY OF CALIBRATION
Reproducibility of calibration curve data produced by the
ICPQ system was determined before and after delivery to the Ames
Laboratory. The data shown in Table 7 were obtained from nine
independent calibrations of the Zn channel, one each hour for
an eight hour period. Five reference standards (0.03, 0.1, 0.3,
1.0, and 3.0 mg/£ Zn) were used for each calibration procedure.
The data show that the day-long stability of the Zn channel was
much better before than after delivery. The poorer stability
after delivery was due to a rather constant drift in the curve
coefficients rather than to random variation. This drift has
not been explained but subsequent analytical measurements on
the zinc channel did not show drift to be a serious problem.
The 10007o relative standard deviation for the intercept, before
delivery, is not considered significant because this value
corresponds to a concentration equivalent of only about two
times the detection limit for zinc.
SHORT TERM REPRODUCIBILITY
The authors consider the short term stability (over one-
half to one day periods) more important than day-to-day repeat-
ability because complete recalibration with four to six
reference standards can be accomplished in little more time than
is required for triplicate analyses of a single standard for
^standardization. For this reason we did not request that a
restandardization provision be included in the software.
Table 8 shows an example of short term stability obtained
from a reference standard that was run 10 times at approxi-
mately equal time intervals throughout a normal analytical
sequence. During this sequence 12 blank determinations were
made and 38 ground and surface water samples from an Iowa Coal
Project mine site were analyzed. The reference standard was
one of the series of standards used in the initial calibration.
The reference standard elemental concentrations included the
approximate ranges found in the mine site samples. The total
elapsed time from the beginning of the calibration to comple-
tion of the sample analyses was about three hours. No adjust-
ments of system parameters were made during this period.
j^
The averarre values and standard deviation data for most of
the elements in Table 8 show adequate stability for most trace
element analyti :t.l work. Also the results of the individual
analyses showed that most of the variation was random in nature
without evidence of drift. The average values for lead and
antimony, however, show greater errors than expected. These
errors may have been caused by errors in reference standard
concentrations or by lack of stability in these channels during
22
-------
ro
CO
Table 7 REPRODUCIBILITY OF CALIBRATIONS FOR THE Zn 213.86 nm
CHANNEL BEFORE AND AFTER DELIVERY OF THE ICPQ SYSTEM
TO THE AMES LABORATORY^
Before Delivery After Delivery
Ave. Valuea" % RSD Ave . Value3 % *SD
A2 (mg/JZ. per mV2) -0.5x10"
AL (Slope), (mg/fc per mV) 0.00785
AQ (Intercept), (rag/ JO -0.00043
Bkdg (mV) 85.7
+ 46% -.1372xlO"5
+ 0.70% 0.009830
+ 1000% 0.0193
+ 1.07% 81.2
+ 51.2%
+ 3.57%
+ 47.9%
+ 1.74%
Average of nine individual measurements.
^Operating conditions were as Set I, Table 2
-------
Table 8. REPRODUCIBILITY. OF ANALYTICAL DATA FROM A
"REFERENCE STANDARD" RUN AT REGULAR INTERVALS
DURING A NORMAL ANALYTICAL SEQUENCE
Al
Al
As
B
Ba
Be
Co
Cr
Cu
Fe
Hg
Mn
Mn
Ni
Pb
Pb
Sb
Sn
Sr
V
Y
Zn
308.22
396.15
193.76
249.68
455.40
313.04
238.89
267.72
324.75
259.94
253.65
257.61
403.08
231.60
220.35
405.78
206.84
303.41
407.77
292.40
371.03
213.86
Reference
standard
(vg/*)
1000
1000
200
200
200
2
100
20
20
4000
200
200
200
100
200
200
200
200
2000
20
20
1000
Measured
average
(US/*)
984-
1000
193
199
201
2.1
101
19
20
3950
197
196
196
105
283
232
241
229
2030
23
21
1020
Standard
deviation
(yg/O
+ 17
66
44
6.8
1.4
.09
2.9
2.5
.6
58
13
2.5
6.2
5.4
31
10
22
73
11
1.2
.4
16
24
-------
the calibration process (see discussion of warm-up time,
Section IV).
STRAY LIGHT
With respect to the determination of trace elemental con-
centrations the stray light characteristics of the ICPQ poly-
chromator constitute its most serious deficiency. Calcium and
magnesium, because of their very intense ion lines emitted in
the ICP, are the most common elemental sources of stray light,
although other elements may contribute stray light if present
at high concentrations. The general subject of the effects of
stray light on the spectroscopic determination of ultratrace
concentrations of the elements has been discussed by Larson,
et al (13).
Three main instrumental sources of stray light have been
identified in the ICPQ polychromator: 1) grating scatter,
2) reflections and scatter in the secondary optics, i.e. the
region between the exit slits and the photomultipliers, and
3) general scatter from reflections by internal surfaces of the
polychromator.
Examples of stray light, due predominantly to grating
scatter, are shown by the wavelength profiles in Figures 2 and
3. These profiles were obtained by stepwise movement of the
entrance slit along the Rowland circle thereby changing the
angle of incidence of the source radiation on the grating. A
small wavelength interval 0=0.1 nm for the profiles in Figures
2 and 3) thus sweeps past each exit slit. The chromium profile
in Figure 2 shows that a typical hard water matrix (150 mg/£ Ca
and 40 mg/t Mg) elevates the background beneath the line an
amount equivalent to about 11 ng/£ chromium. Similarly, the
antimony profile in Figure 2 shows a background elevation
equivalent to 200 |ag/
-------
Cr (II) 267.72 run
Sb (I) 206.84 nm
34.0 —
no
>-"
H
en
z
UJ
UJ
tr
30.0 —
26.0 —
22.01
3.60
2.80
2.00
500 490 480
470 510
SLIT DIAL POSITION
500 490 480 470
Figure 2. Examples of wavelength profiles that show the effects of stray light
caused by concentrations of calcium and magnesium typical in hard
water. 0 deionized water, A analyte mixture containing 50 ug/£ Cr
and 100 \±g/l Sb, + hard water matrix (150 mg/£ Ca, 40 mg/t Mg), X
analyte mixture as above in hard water matrix.
-------
6.0
5.0
4.0
UJ
3.0
2.0
1
1
1
500 490 480 470
SLIT DIAL POSITION
460
Figure 3.
Wavelength profiles of the Cd 226.50 nm line. The
profiles were obtained from four surface water
samples containing different concentrations of
calcium and magnesium (see Table 10). 0 deionized
water, + Scott Pond, X Site 10, O Site 21,
-------
Table 9. MEASURED CALCIUM AND MAGNESIUM CONCENTRATIONS
FROM FOUR SURFACE WATER SAMPLES FOR WHICH WAVELENGTH
PROFILES OF THE Cd 226.50 nm REGION ARE SHOWN
IN FIGURE 3
~IConcentration(rag/£)
Sample
identification Ca Mg
Scott Pond 10 4
Site 10 100 26
Site 21 290 140
Site 53 220 90
28
-------
Pb(II) 220.35
V(II) 292.40
2.40
1.60
en
2
UJ
t-
ro
0.80
0.00
480-
320 —
464 472 480 488 496
496
SLIT DIAL POSITION
Figure 40
Examples of stray light arising in the secondary optics region of the
QA. 137 polychromator. All exit slits were blocked except that for
the Mg 279.55 nra line. 0 deionized water, A 1000 mg/£ Mg,
+ 1000 mA Ca.
-------
calcium and magnesium solutions. All slits except the Mg 279,55
nm slit were blocked with opaque flat black paper. The psuedo
line profiles that occur as the magnesium line traverses its
slit must therefore result from scattering of the magnesium
radiation to the detectors of the other elements. Secondary
optics scattering of the magnesium wavelength was observed for
14 of the 19 spectral lines between and including the Pb 220.35
and Ti 334.90 nm lines. In general, the secondary optics stray
light becomes more severe near the magnesium wavelength, but as
Figure 4 shows, it is still significant as far away as the
Pb 220.35 nm wavelength. Geometrical relations between the
secondary mirrors and photomultipliers determine which channels
pick up the most stray light. The geometrical complexity of
the secondary optics region of the QA 137 polychromator (with
the spectral line array as listed in Table 1) is shown in
Figure 5.
Sources of stray light other than the grating and second-
ary optics are also present in the QA 137 polychromator.
Examples of such stray light are shown by the profiles in
Figure 6 which were obtained simultaneously with the profiles
in Figure 4 (all slits blocked except Mg 279.55 nm). The
barium and selenium wavelengths are near the extreme ends of
the wavelength range of the QA 137 polychromator and do not
show evidence of secondary optics stray light (Peaks corres-
ponding to the traverse of the magnesium line across its slit
are not apparent)„ Instead, the calcium and magnesium solu-
tions cause a general elevation of the background due to
scattering from various internal surfaces of the polychromater.
This general stray light from 1000 mg/-t calcium corresponds to
a few times the detection limits for barium and selenium (see
analytical curve slope data in Table 4).
Stray light levels have been quantified for 28 wavelengths
on the QA. 137 spectrometer as shown in Tables 10 and 11 „ To
obtain these data the system was calibrated for all elements
listed in the table. The several calcium and magnesium concen-
trations were run as samples with a 1% HNO^-deionized water
solution as the reference blank. The net sample signal was
then converted to an equivalent concentration for each analyte
element. The same measurement sequence was performed at the
analyte line wavelengths and at 0.04 nm intervals above and
below these wavelengths (new reference blank values were
established for each of the offset wavelengths). In general
the data show a high degree of internal consistency. If the
concentration equivalents are significant with respect to the
detection limits and if the three values within a set are
reasonably consistent then stray light is most probably
involved. If the concentration equivalents within the -0.04, 0,
and +0.04 nm set differ by more than a few times the detection
limit, three possible causes are suggested:
30
-------
Figure 5. The secondary optics region of the QA 137
polychromator showing the exit slit frame,
the secondary mirrors and photomultipliers
31
-------
Ba (II) 455.40 nm
Se (I) 196.09 ran
1.40
i
1.00
CO
ro
o.so
0.20
LOO
0.80
0.60
472 480
0.40
484 472
SLIT DIAL POSITION
Figure 6. Examples of general stray light arising from general scattering
within the polychromator enclosure. All exit slits were blocked
except that for the Mg 279.55 nm line. Curve symbols same as
on Figure 4»
-------
Table 10. STRAY LIGHT CONCENTRATION EQUIVALENTS (mg/A) FROM CALCIUM'
Al
Al
As
B
Ba
Be
Cd
Co
Cr
Cu
Fe
Hg
Mn
Mn
Mo
Ni
Pb
Pb
Sb
Se
Sn
Sr
Ti
V
Y
Y
Zn
X
308
396
193
249
455
313
50 mg/1 Ca 200 mg/1 Ca 500 mg/1 Ca
Wavelength relative Wavelength relative Wavelength relative
to analyte line to analyte line to analyte line
(nm) -.04 nm 0 +.04 nm -.04 nm 0 +.04 nm -.04 nm 0 +.04 nm'
22
15
76
68
.40
.04
226.50
238
267
324
259
253
257
403
287
231
220
405
206
196
303
407
334
292
242
371
213
.89
.72
.75
.94
.65
.61
.08
.15
.60
.35
.78
.84
.09
.41
.77
.94
.40
.22
.03
.86
000000
00 1397
0. 5502
00 0036
•0o 0007
-o 0000
0.0055
00009 1
0.0056
-o 0003
0,0019
0. 0047
0. 000 1
000037
0o 0074
000162
0. 1248
0.0660
0. 089 1
0.2938
0,0939
00 0000
00 0010
0.2006
0,0224
00 0000
0. 0067
-00023
0o 1562
0.5781
-00005
0o 0006
00 0000
0o 0088
0»0045
000015
-o 000 1
-.0018
0»0I58
0. 0001
000041
0o 0080
000171
000936
-o 0070
0o 1032
0o 2308
00 0000
0o 0060
0o 0018
0o 0002
000158
0o 0002
0.0078
000035
0o 1709
0o3638
-00190
0.4960
K5179
000061 000128
0o 0007
-o 0000
0o 0022
000057
0. 0022
-o 0003
0, 0016
fi0 030!
00 000 1
0o 0048
-00077
0o 0122
0o 1248
000636
0o 1 126
0.2453
0.0841
000000
0o P081
0. 0003
000145
000001
000040
0.0025
00 0001
0.0232
000262
0«0135
-.0003
000047
0.07 14
0o 0006
0.0156
0.0317
0o 0476
0o 3436
0.2625
0o 3018
0.8189
0. 1882
0o 0004
000061
000043
0. 0635
0. 0008
0.0168
000178
0o 5641
lo 6108
0o 0082
0. 0025
0o 0002
0.0213
0» 0223
0.0115
-00002
0o2038
0.0555
000003
000180
0o 0299.
000516
0o 3436
0.2388
0.2923
0o 7842
0. 173-3
0. 0232
.000068
0.0035
0. 0675
0. 001 1
000224
0.0250
0.61 18
lo 3971
000190
000027
00 0000
0.0180
0o 0257
000191
0.0010
0.0079
0.0761
0. 0007
000212
0.0389
0.0520
0.4023
0o2839
0=3828
0o8189
0.2877
0o 0004
000549
0. 0047
.0.0661
00 0010
'0.0181
0.0381
Io0062
2.7608
000309
0o 0048
0o 0002
2.0389
0.0534
000341
0.0017
0o01 12
0o 1 158
0o 0013
0. 0417
0.0740
0 o 0.9 1 4
0. 6256
0.6596
0. 5313
1.4534
0o4774
0. 001 1
0. 0275
0.01 14
0. 1 182
0. 0025
0.0296
0o 0404
1. 1459
2.9367
0.0242
0o 0047
0. 0004
0o 0406
0.0426
000292
0. 0027
0. 0094
0. 1380
0. 0012
0. 0441
0.07 19
0.0933
0o 5629
0. 6355
0o 5217
lo 321 1
0.6478
0.0564
0o 039 1
0= 0094
0. 1329
0. 0022
0o 0320
0.051 1
1.2387
2.4181
000366
0o 0050
0.0003
0o 0386
0.0532
0.0306
0o 0030
000150
0o 157 1
0.0013
0. 0443
000749
000992
0.7277
0o 6789.
0o7539
lo 3973
0o 6830
0. 0010
0. 1369
000121
0. 1342
0.0026
0.0307
When the gross (sample) signal level is so low that its uncertainty overlaps that
the blank signal, negative net intensities may occur. Under these conditions the
negative values should be of negligible magnitude.
of
-------
Table 11. STRAY LIGHT CONCENTRATION EQUIVALENTS (mg/A) FROM MAGNESIUM0
u>
10 mg/1 Mg 40 mg/1 Mg 100 Mg/1 Mg
Wavelength relative Wavelength relative Wavelength relative
to analyte line to analyte line to analyte line
X(nm) -.04 nm 0 +.04 nm -.04 nm 0 +.04 nm -.04 nm 0 +.04 nm
Al
Al
As
B
Ba
Be
Ca
Cd
Co
Cr
Cu
Fe
Hg
Mn
Mn
Mo
Ni
Pb
Pb
Sb
Se
Sn
Sr
Ti
V
Y
Y
Zn
308.22
396.15
193.76
249.68
455.40
313.04
315.89
226.50
238.89
267. 72
324. 75
259.94
253.65
257.61
403.08
287.15
231.60
220.35
405. 78
206.84
196. 09
303.41
407. 77
334.94
292.40
242.22
371.03
213.86
-00202 -00059 000166
-00067 -00075 0.0027
000933 0.0933 -.0280
-00020 -00036 000005
-00000 000010 000001
-00000 000002 000000
0o 0015 -o 0043 0o 001 1
000000 000029 -00045
-00043 000045 -00007
-00016 -00037 0,0031
-00013 000008 000004
000001 -00012 0.0009
-o 0015 0o 0 142 0, 0031
-00000 000009 000002
-00036 -00023 000031
-o 0029 -00017 0.0113
-00004 0.0044 000024
000428 000311 000038
-00094 -00305 000471
000234 000187 -00093
000383 000622 000143
-o 0098 -o 1036 0o 0593
-00000 000010 000000
-.0047 -00003 000050
-00010 -00016 000010
000039 000039 -00026
-00005 000004 000003
000020 000065 -00026
000017 -00142 000089 =00029 000047 000172
000010 -.0067 000005 -00062 -00010 -00060
0o2239 0o 1493 000933 0o 3265 0.3544 0.2798
000036 000025 000087 000030 00 0025 000077
000001 .000005 000003 000005 0.0006 000004
-00000 000001 -00000 000000 0.0002 -00000
000023 000000 0.0019 0.0035 -.0062 -.0015
000022 000036 -00003 000068 0.0052 0.0006
000019 0oB012 000040 0C0038 0.0052 0.0057
000047 -00015 0.0031 000065 0,0015 0*0051
000002 -00007 000001 -00000 -00003 -00Z-05
0.0019 -00009 000016 000023 0.0041 .0.0023
000142 000047 000301 0.0349 0.0365 0.0380
0.0002 000005 000001 000004 0.0006 0.0004
-00018 -00028 000009 -00013 -.0031 -.0018
000245 000029 0.0233 ,000383 0.0386 0.0401
0.0054 000014 000019 000132 0«0167 000103
000545 000233 000311 0.0858 0.0506 0.0701
0o 1771 000329 0.0471 0.4078 0o2910 000542
000468 -.0093 000327 000750 0.0703 0.0280
0o 1247 000862 0ol391 0.2308 0.1777 0.2115
0.0445 -00888 0.0841 0.0742 -.0197 0.0940
-.0000 000004 000000 -.0000 0.0001 -00000
-.6003 -00061 000005 -.0016 -00029 -.0005
0.0015 -.0007 000018 000014 0.0010 0.0023
000118 000079 000065 000237 0.0131 0.0118
-.0000 0.0004 0.0001 -.0000 0.0010 -.0031
000038 000058 000013 0.0029. 000139 0.0015
a.,
See explanatory note for Table 10
-------
1. secondary optics stray light,
2. a residual impurity of the analyte element in the
. calcium or magnesium test solution,
3. a spectral line interference from calcium or
magnesium.,
The first cause is not probable because the exit slit for the
Mg 279.55 nm line was blocked for this experiment and no exit
slits were present for the strongest calcium lines. An example
of a residual impurity is shown by the results for strontium
from the 50, 200, and 500 mg/£ calcium solutions (Table 10). An
example of a spectral line interference is shown by the results
for the Pb 405.78 nm line with the magnesium solutions
(Table 11). A magnesium line is located 0.019 nm below the
lead wavelength.
Spectral line broadening (in the excitation source and from
instrumental causes) and shifts in the true spectral background
are being investigated on other ICP systems in the Ames Labora-
tory. The background shifts from these causes are generally
small but may become significant as other sources of stray
light are-reduced or eliminated,
SOLUTION OF STRAY LIGHT PROBLEMS
After the various sources of stray light in the ICPQ poly-
chromator had been identified attempts were made to reduce or
eliminate the stray light level or to correct empirically for
its effects.
The largest secondary optics stray light signals from
magnesium were obtained from the Cr 267.72 and V 292.40 nm
channels. The geometrical relationships of these channels with
respect to the Mg 279.55 channel are shown in Figure 7. The
split peak exhibited.by the vanadium profile (Figure 4) suggests
that the scattering is greatest when the solid angle of the
magnesium radiation is at either edge of the magnesium mirror.
Baffles were placed, one at a time, in appropriate locations to
block scattering of the magnesium radiation from one component
to another. Baffle 1, located as shown in Figure 7, had
essentially no effect on the stray light received by the
chromium and vanadium channels. Baffle 2, placed to block
reflections from the glossy manganese photomultiplier shield,
eliminated most of the stray light received by the vanadium
channel but had no effect on the chromium channel. Baffle 3
eliminated most of the stray light received by both chromium
and vanadium. These experiments indicated that the reflective
surface of the magnesium mirror actually transmitted part of the
magnesium radiation which then scattered from the back and edge
35
-------
ROWLAND CIRCLE
(FOCAL CURVE)
Cr 267.72
BAFFLE 3
BAFFLE 2
279.55
_Mq_2_87.l_5_
V 292.40
BAFFLE I
Figure 7. Geometrical arrangement of secondary optics in the region of the
Mg 279.55 nm line.
-------
surfaces of the mirror blank to nearby photomultiplters. The
reduction in stray light to the vanadium channel by Baffle 2
was due mainly to its blockage of the path to the magnesium
mirror rather than to its blockage of reflections from the
manganese photomultiplier shield.
A zero-order light trap is located on the low wavelength
side of the polychromator case. A flat black paper baffle was
added to this trap to prevent a direct reflection of zero-order
radiation from passing through a narrow open joint in the trap
and reflecting directly to the focal curve. Also, cadmium
plated socket-head cap screws on the face of the grating mask
were painted with optically flat black paint. Shiny objects at
each of the lower corners of the grating (apparently part of
the mount) reflected the primary radiation to the grating at an
inappropriate angle and the resulting diffracted light appeared
as diagonal streaks emanating from the grating. According to
advice from ARL a second mask was superimposed on the original
grating mask to block this potential source of stray light.
Because this mask can be no closer to the grating than about
0.5 cm it may vignette the diffracted rays from the grating
for certain wavelengths.
Empirical corrections for stray light were first approached
through the interference correction method as provided for in
the ICPQ software. In this method a concentration correction
factor K is obtained from the product of an interference
coefficient B and interferent concentration C.
K = BCL (2)
where B is the analyte concentration error (caused by the
interferent) divided by the interferent concentration and is
obtained from a solution containing only the interferent. The
real analyte concentration is then obtained by subtraction of
the error components (one for each interferent) from the gross
calculated concentration. In practice this correction tech-
nique over corrected for interferent concentrations greater than
and under corrected for interferent concentrations less than
that used for the calculation of the interference coefficient.
The reason for this failure is illustrated in Figure 8 in which
interference coefficients for arsenic, selenium, lead, and
antimony are shown as functions of calcium concentration. The
curvature correlates quite well with the self-absorption
exhibited by the Ca II 393.37 and 396.85 nm lines. Correction
coefficients based on intensity rather than concentration
should solve the problem. These intense calcium lines exhibit
different degrees of self-absorption, however, hence neither
line provides the proper base intensity for correction of the
background shifts caused by the pair.
37
-------
CO
00
0.0150 —
o
LL.
O
O
O
UJ
Q:
cr
o
o
0.0100
0.0050 —
cr
H
CO
0
100
200 300
Co CONCENT RAT ION,
400
500
Figure 8. Stray light correction coefficients as a function of calcium
concentration,
X As 193.76 nm, A Se 196.09 nm,
0 Pb 220o35 nm, Q Sb 206.84 nm.
-------
Wavelength profiles, e.g. Figures 2, 3, 4, and 6, have been
used extensively for evaluation of stray light problems in real
samples and for correction for the effects of background shifts
whatever their cause. Often the background level at an analyte
wavelength can be established from these profiles, by a line
joining the backgrounds at either side of the line.
39
-------
SECTION 6
DISCUSSION
In general, the basic hardware and software were relatively
trouble free in routine operation during the evaluation period.
Also, the basic operating routines were easy to learn by
personnel who had previous experience with ICP-AES systems. The
malfunctions that did occur included: 1) occasional fouling of
the nebulizer tip from high salt concentration solutions,
2) slippage of the digital counter assembly on the entrance slit
adjustment (profile) screw (the digital counter assembly was
also subject to being struck by the outer cabinet upon opening
for access to the exit slit-photomultiplier region of the
spectrometer), 3) intermittant failure of capacitors to dis-
charge properly on several channels between integrations.
The following section contains general and specific
comments about the operation and performance of the system.
DYNAMIC RANGE
The dynamic range of the system is presently limited by the
signal measurement system (3-3^ orders of magnitude with stan-
dard deviations approximating 170 or less) rather than by the
sample excitation system.
DETECTION LIMITS
Recent design improvements have made ultrasonic transducers
much more resistant to degradation by corrosive solutions„ In
general, the ultrasonic technique produces an order of magni-
tude or more improvement in comparison to detection limits
obtained by pneumatic nebulization (14). Similar improvements
should be realized on the ICPQ as well as other ICP systems.
SOFTWARE
The PDF 11/05 computer was supplied with 8K of memory. The
ARL software used all but about 200 words. Additional routines
would require an increase in computer memory. The software as
supplied, however, was written to be compatible with all ARL
40
-------
ICPQ systems having up to 48 channels. The EPA-Athens instru-
ment had 32 channels. Therefore, approximately one-third of
the space allocated for data storage was wasted (~1000 words).
This space could be used for calculation of detection limits or
interference correction coefficients, or possibly both routines
or others could be fitted into this space. For example, com-
puter control of a stepper motor and of the integration sequence
for wavelength profiling would be very useful. The economics of
changes in the existing software versus purchase of addition
memory must be considered, of course.
STRAY LIGHT
As mentioned earlier the stray light problem constitutes
the most serious deficiency of the ICPQ system. Stray light
levels have been assessed with respect to the more stringent of
the EPA criteria levels for public water supplies and continuous
use irrigation water (1, 2) and the results are shown in
Table 12. Ultrasonic nebulization, with its factor of 10 or
better improvement in detection limits, will be of significant
help with regard to detectability for all of the elements
listed in Table 12 but its greatest value will be for those
elements in column III for which pneumatic nebulization
presently provides insufficient powers of detection. Ultra-
sonic, nebulization is not expected to influence the relative
stray light levels significantly, therefore improvements in the
stray light properties of spectrometers is still a high priority
objective.
41
-------
Table 12. CLASSIFICATION OF SPECTRAL LINES WITH RESPECT
TO STRAY LIGHT PROBLEMS AS OBSERVED ON
QA 137 POLYCHROMATOR
I
Stray light: is a
small fraction of
signal from EPA
criteria level.
concentration.
II
Spectral sensi-
tivity is sufficient,
but stray light from
hard waters could
cause significant
errors at EPA
criteria level
concentrations.
Al
B
Ba
Be
Cu
Fe
Mn
Zn
308
249,
455,
313,
324,
259.
257.
213.
22
68
40
04
75
94
61
86
Al
Co
Cr
Mn
Ni
V
396
238
267
403
231
292
15
89
72
08
60
40
III
Poor sen s :i !.. 1 v i I: y
of the spectral
line (with
pneumatic aebuli.-
zation) and
stray light are
both significant
problems at the
EPA criteria
level concentra-
tions .
As
Cd
Hg
Mo
Pb
Pb
Se
19.3.
226.
253.
287.
22.0.
405.
196.
76
50
65
15
35
78
09
42
-------
SECTION 7
PUBLICATIONS
1. Larson, G. F., V. A. Fassel, R. K. Winge, and
R. N. Kniseley., "Ultratrace Analyses by Optical Emission
Spectroscopy: The Stray Light Problem." Appl. Spectrosc,
30, 384-391, July-August 1976.
2. Winge, R. K., V. A. Fassel, R. N. Kniseley, E. L. DeKalb
and W. J. Haas, Jr. "Inductively Coupled Plasma-
Multielement Atomic Emission Spectrometry: Determination of
Trace Elements in Soft, Hard, and Saline Waters9" submitted
for publication in Environ. Sci. Technol. 1976.
43
-------
REFERENCES
1. "Water Quality Criteria 1972." National Academy of
Sciences and National Academy of Engineering. U.S.
Government Printing Office, Washington, B.C.
EPA-R3-73-033. The Environmental Protection Agency. 1974.
594 pp.
2. Winge, R. K., V. A. Fassel, R. N. Kniseley, E. L. DeKalb,
and W. J. Haas, Jr., Submitted for publication in Environ.
Sci. & Tech. 1976.
3. Fassel, V. A. and R. N. Kniseley. "Inductively Coupled
Plasma-Optical Emission Spectroscopy." Anal. Chem. 46:
1110A-1120A, November 1974.
4. Fassel, V. A. and R. N. Kniseley. "Inductively Coupled
Plasmas." Anal. Chem. 46; 1155A-1164A, November 1974.
5. Scott, R. H., V. A. Fassel, R. N. Kniseley, and D. E.
Nixon. "Inductively Coupled Plasma-Optical Emission .
Analytical Spectrometry." Anal. Chem. 46: 75-80, January
1974.
6. Fassel, V. A. "Electrical 'Flame' Spectroscopy." in
Proceedings of the XVI Colloquium Spectroscopicum Inter-
nationale, Heidleberg 1971." London, Adam Hilger Ltd.,
1972. pp. 63-93.
7. Boumans, P. W. J. M. and F. J. de Boer. "Studies of an
Inductively Coupled High Frequency Argon Plasma for Optical
Emission Spectrometry -III. Compromise Conditions for
simultaneous Multi-element Analyses." Spectrochim. Acta
30B; 309-334, 1975.
8. Greenfield, S., H. McD. McGeachin, and P. B. Smith.
"Plasma Emission Sources in Analytical Spectroscopy-III.
'Talanta 23: 1-14, 1976.
9. Larson, G. F., V. A. Fassel, R. H. Scott, and R. N.
Kniseley. "Inductively Coupled Plasma-Optical Emission
Analytical Spectrometry. A Study of Some Interelement
Effects." Anal. Chem. 47: 238-243, February 1975.
44
-------
10. Winge, R. K. and V. A. Fassel. "Development and Applica-
tion of an Inductively Coupled Plasma Analytical System for
the Simultaneous Multielement Determination of Trace
Elemental Pollutants in Water." Ames Laboratory, U.S.
Energy Research and Development Administration. Ames,
Iowa. Annual Report submitted to Athens Environmental
Research Laboratory, U.S. Environmental Protection Agency,
Athens, Georgia. Report period March 1, 1974 - February 28,
1975. Part II, pp. 1-19.
11. "RCA Photomultiplier Manual." RCA Corporation. Harrison,
New Jersey, Technical Series PT-61. 1971. p. 111.
12. Butler, C. C., R. N. Kniseley, and V. A. Fassel. "Induc-
tively Coupled Plasma-Optical Emission Spectrometry:
Application to the Determination of Alloying and Impurity
Elements in Low and High Alloy Steels." Anal. Chem. 47:
825-829, May 1975.
13. Larson, G. F., V. A. Fassel, R. K. Winge, and R. N.
Kniseley. "Ultratrace Analyses by Optical Emission
Spectroscopy: The Stray Light Problem." Appl. Spectrosc.
30, 384-391, July-August 1976.
14. Olson, K. W., W. J. Haas, Jr., and V. A. Fassel. "Ultra-
sonic Versus Pneumatic Nebulization in Inductively Coupled
Plasma-Atomic Emission Spectroscopy: An Improved Ultra-
sonic Nebulization Facility for Ultratrace Multielement
Analysis." Submitted for publication in Anal. Chem. 1976.
45
-------
TECHNICAL REPORT DATA
fPlease read Instructions on the reverse before completing)
. REPORT NO.
EPA-60Q/4-77-032
3. RECIPIENT'S ACCESSION NO.
4. TITLE ANDSUBTITLE
Evaluation of an Inductively Coupled Plasma,
Multichannel Spectrometric Analysis System
5. REPORT DATE
June 1977
issuing date
6. PERFORMING ORGANIZATION! CODE
7. AUTHOR(S)
R. K. Winge, J. M. Katzenberger, and
R. N. Kniseley
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Ames Laboratory, U.S. Energy Research and
Development Administration, Iowa State Univ.,
Ames, Iowa 50011
10. PROGRAM ELEMENT NO.
1BD612, 1CC614
11. CONTRACT/GRANT NO.
EPA-IAG-D6-0417
. SPONSORING AGENCY NAME AND ADDRESS
Environmental Research Laboratory -
Office of Research and Development
U.S. Environmental Protection Agency
Athens, Georgia 30605
13. TYPE OF REPORT AND PERIOD COVERED
3/1/73-1
15. SUPPLEMENTARY NOTES
14. SPONSORING AGENCY CODE
EPA/600/01
An inductively coupled plasma, multielement atomic emission
spectrometric analysis system has been evaluated with respect to the
Environmental Protection Agency's need for a rapid method for
determination of trace elemental concentrations in water. Data are
presented on detection limits, curve fits and reproducibility of
analytical curves, photomultiplier performance, dynamic range,
reproducibility of analytical results, and stray light properties of
the instrument.
The instrument as a whole performed well, but stray light arising
from calcium and magnesium in typical hard waters caused significant
errors at low concentration levels for a number of critical elements.
Suggestions for reduction of the stray light and for empirical
correction of its effects are discussed.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATl Field/Group
Analytical chemistry,
Spectrochemical analysis,
Water analysis,
Optical equipment,
Spectrometers
Pollutant
Water supplies
Spectroscopic
excitation
QA-137
identification
07D
. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS f This Report)
Unclassif-i
21. NO. OF PAGES
54
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA. Form 2220-1 (9-73)
46
•ft U.S. GOVERNMENT PRINTING OFFICE: 1977- 757-057/6438
------- |