Trace Metals in Waters by GFAAS, in Accordance with U.S. EPA and Health Canada Requirements

a p p l i c at i o n N o t e
Atomic Absorption
Author:
Randy L. Hergenreder
PerkinElmer, Inc.
Shelton, CT 06484 USA
Trace Metals in
Waters by GFAAS,
in Accordance with
U.S. EPA and Health
Canada Requirements
Introduction
Several trace elements are recognized as toxic or
carcinogenic and are regulated in drinking water
by various environmental agencies worldwide.
The U.S. Environmental Protection Agency’s (EPA)
Safe Drinking Water Act includes maximum levels
permitted in drinking water for the elements arsenic
(As), cadmium (Cd), lead (Pb), selenium (Se) and
thallium (Tl). The World Health Organization
(WHO) and Health Canada also have limits on
these elements in drinking water (Table 1). Water
contamination sources can range from naturally occurring deposits exposed from erosion,
to agriculture and industrial discharges. There can also be direct contamination from: Pb
used in plumbing fixtures, Cd found in galvanized pipes, and electronics manufacturing
discharges for Tl.
Precise and accurate measurements at the regulated levels are an important factor
for assuring safe drinking water. U.S. EPA Method 200.91 is the method cited by EPA,
Health Canada, and the WHO for the use of graphite furnace atomic absorption spectroscopy
(GFAAS). In evaluating a GFAAS system for determination of these elements, it must
provide good sensitivity, low noise, limited drift, and accuracy in matrices with high salt
content (hard water) that might be found in drinking waters. In this work, the PinAAcle™
900T, with a unique optical system, is evaluated for the use of EPA Method 200.9 for
As, Cd, Pb, Se, and Tl in drinking waters.
Table 1. Drinking water limits for regulatory agencies
(µg/L).
Organization
As
Cd
Pb
Se
Tl
U.S. EPA
10
5
15
50
2
Health Canada
10
5
10
10
–
WHO
10
3
10
10
–
Experimental
Instrumentation
A PerkinElmer® PinAAcle 900T flame and longitudinal
Zeeman atomic absorption spectrometer was used for all
measurements. Samples were automatically pipetted into
standard transversely heated graphite atomizer (THGA)
tubes (Part No. B0504033) using an AS 900 autosampler.
PerkinElmer Lumina™ single-element hollow cathode lamps
(HCLs) were used as the light source for lead (Part No.
N3050157), cadmium (Part No. N3050115), and thallium
(Part No. N3050183); while single-element electrodeless
discharge lamps (EDLs) were used for arsenic (Part No.
N3050605) and selenium (Part No. N3050672).
Figure 2. AS 900 autosampler depositing a droplet
of water in the THGA tube as seen using the TubeView
furnace camera.
Table 2. Instrument settings used for U.S. EPA method
200.9 on the PinAAcle 900T.
Wave-
Lamp
Read Read
Element length Slit
Lamp Current Delay Time
(nm)
(nm) Type (mA)
(s)
(s)
As
193.7
0.7
EDL
380
1.0
2.5
Cd
228.8
0.7
HCL
4
0.5
2.5
Pb
283.3
0.7
HCL
10
0.5
3.5
Se
196.0
2.0
EDL
280
1.0
3.0
Tl
276.8
0.7
HCL
6
0.5
3.0
*Baseline Offset Correction (BOC) was 5 seconds for all analytes.
Figure 1. PinAAcle 900T atomic absorption spectrometer with AS 900 furnace
autosampler.
The PinAAcle 900T instrument settings are listed in Table 2
and the furnace program is listed in Table 3. The pyrolysis
and atomization temperatures were optimized using a
spiked drinking water sample. The Method Development
feature of the software (WinLab32™ for AA) automated
these experiments. The TubeView™ furnace camera (Figure 2)
on the PinAAcle 900T was used to adjust the pipette tip
to the most appropriate depth and to watch for potential
matrix buildup in the tube. The camera was also used during
method development to verify the drying steps, ensuring
that no sample boiling or spattering occurred.
The furnace programs for individual elements are listed in
Tables 3-7. Argon was the normal gas type. The special gas
type was 95% Ar / 5% H2 and was used in steps 1-3. Step
4 removed the H2 gas from the tube before atomization and
the read step was Step 5 for all elements. All elements used
0.005 mg Pd (Part No. B0190635) + 0.003 mg Mg(NO3)2
(Part No. B0190634) as a matrix modifier. All samples and
standards used an injection volume of 20 µL sample plus
5 µL of matrix modifier. Each analyte was run in duplicate
(2 replicates).
Table 3. Furnace parameters for arsenic (As).
Step
Temp (˚C) Ramp
Time
Hold
Time
Internal
Flow
1
110
1
30
250
Special
2
130
15
30
250
Special
3
900
10
20
250
Special
4
900
1
15
250
Normal
5*
2300
0
5
0
Normal
6
2450
1
3
250
Normal
*Read step
2
Gas
Type
Results
Table 4. Furnace parameters for cadmium (Cd).
Step
Temp
(˚C) Ramp
Time
Hold
Time
Internal
Flow
Gas
Type
1
110
1
30
250
Special
2
130
15
30
250
Special
3
700
10
20
250
Special
4
700
1
15
250
Normal
5*
1700
0
5
0
Normal
6
2450
1
3
250
Normal
*Read step
Table 5. Furnace parameters for lead (Pb).
Step
Temp
(˚C) Ramp
Time
Hold
Time
Internal
Flow
Gas
Type
1
110
1
30
250
Special
2
130
15
30
250
Special
3
850
10
20
250
Special
4
850
1
15
0
Normal
5*
2000
0
5
0
Normal
6
2450
1
3
250
Normal
*Read step
Table 6. Furnace parameters for selenium (Se).
Step
Temp
(˚C) Ramp
Time
Hold
Time
Internal
Flow
Gas
Type
1
110
1
30
250
Special
2
130
15
30
250
Special
3
800
10
30
250
Special
4
800
1
15
250
Normal
5*
2150
0
4
0
Normal
6
2450
1
3
250
Normal
*Read step
Table 7. Furnace parameters for thallium (Tl).
Step
Temp
(˚C) Ramp
Time
Hold
Time
Internal
Flow
Gas
Type
1
110
1
30
250
Special
2
130
15
30
250
Special
3
500
10
20
250
Special
4
500
1
15
250
Normal
5*
1600
0
5
0
Normal
6
2450
1
3
250
Normal
Calibration Statistics
All elements showed calibration correlation coefficients
better than or equal to an r2 value of 0.999. The method
detection limit (MDL) was determined by running seven
separate samples of a 0.5 µg/L standard for As, Pb, Se and
Tl and a 0.17 µg/L standard for Cd (Table 8). The resulting
standard deviation was then multiplied by the Student’s
t-value (ρ = 0.01). The instrument detection limit (IDL)
was calculated by analyzing 10 replicates of a blank and
multiplying the standard deviation by 3. Linear ranges are
the highest standard that was within 5% of its true value.
Table 8. Linear ranges, MDL and IDLs using the PinAAcle
900T and U.S. EPA Method 200.9.
IDL
IDL
MDL
Element (µg/L)
(pg)
(µg/L)
Linear
Range
(µg/L)
As
0.30
6.0
0.17
100
Cd
0.056
1.1
0.027
2.50
Pb
0.33
6.6
0.16
100
Se
0.28
5.6
0.30
200
Tl
0.37
7.4
0.18
100
The sensitivity, as indicated by the characteristic mass, was
better than the expected value for As, Pb and Tl, while the
Se and Cd values were slightly elevated. This could be due
to the difference in modifier and lamp type from those
detailed in the Atomic Absorption Spectroscopy Analytical
Methods “cookbook” guide.
Practical Quantification Limits
In the Arsenic Rule,2 a standard of 3 µg/L must be recovered
within ±30%. This is performed at the practical quantification
limit (PQL) for As. Other elements have similar limits3 and
these are listed in Table 9, along with the values determined
in this study. A lower standard was used here to demonstrate
the ease of this test using the PinAAcle 900T. The low
standard was read as five separate samples and the average
value with percent relative deviation is given.
*Read step
3
Table 9. Results of the PQL study in accordance with U.S. EPA standards.
Element
PQL
Level
(µg/L)
Acceptance
Limit
(± in %)
Standard
Concentration
Used (µg/L)
Average
Found
(µg/L)
%RSD
As
3
30
1
0.876
4.45
88.0
Cd
2
20
1
1.03
1.33
103
Pb
5
*
5
4.99
3.30
99.8
Se
10
20
1
1.06
13.5
106
Tl
2
30
1
0.810
14.3
81.0
%Recovery
*No PQL acceptance limit was found. Preparing for Proposed Revision 3.3
In the proposed U.S. EPA Method 200.9 Revision 3.3, there is a requirement to run an Initial
Precision and Recovery (IPR) standard. The proposed IPR currently consists of four aliquots of
reagent water spiked at 1 to 5 times the lowest standard and run as samples. The percent
recovery and standard deviation are calculated for those four samples (Table 10). There are
currently no criteria for this test.
Table 10. Data for the proposed initial precision and recovery standard.
Average
Found (µg/L)
Std. Concentration
Used (µg/L)
Standard
Deviation
%Recovery
As
4.83
5.0
0.15
96.6
Cd
0.293
0.25
0.0096
117
Pb
5.04
5.0
0.10
101
Se
4.94
5.0
0.13
98.7
Tl
5.00
5.0
0.18
99.9
Quality Control
Two quality control (QC) checks were analyzed to determine the accuracy of the method.
NIST® 1643e Trace Elements in Water is a fresh water reference material. Data and recoveries
are listed in Table 11. The ERA PotableWatR™ Inorganic Metals material was used as the
independent QC check during the analytical runs. This ERA reference material is also used in
round robin U.S. EPA performance testing. The recoveries and data for this experiment and
the U.S. EPA round robin experiment are listed in Table 12. The recovered concentrations
are within the proficiency testing (PT) performance acceptance limits (PALs) which closely
approximate a 95% confidence interval based on historical data.
Table 11. Recoveries for NIST® 1643e natural waters reference material using PinAAcle 900T.
Element
4
NIST® 1643e
(µg/L)
Limits
(±, µg/L)
NIST® Found
(µg/L)
Std. Dev.
(n=2)
%Recovery
As
60.45
0.70
58.2
0.20
96.2
Cd
6.568
0.073
6.58
0.00080
100
Pb
19.63
0.21
19.5
0.15
99.1
Se
11.97
0.14
10.8
0.12
90.1
Tl
7.445
0.096
6.48
0.010
87.0
Table 12. Recoveries for the determination of metals in ERA PotableWatR™ Inorganic Metals.
Certified Value Uncertainty Round Robin %Recovery ERA Found %Recovery
(µg/L)
(%)
Mean (µg/L) Round Robin
(µg/L)
Found
As
25.7
2.5
25.2
98.1
26.8
104
Cd
40.5
0.20
40.1
99.0
37.3
92.0
Pb
24.4
1.2
24.6
101
25.5
105
Se
22.6
1.2
22.5
99.6
22.1
97.6
Tl
7.84
1.4
7.75
98.8
7.35
93.8
Matrix Effects
A laboratory fortified blank (LFB) showed good recoveries. One drinking water sample was
analyzed with a spike and spike duplicate to check for matrix interferences. The spike recoveries
and sample/blank data are shown in Table 13.
Table 13. Sample and LFB recoveries and percent difference of duplicate spikes.
Spike
Lab
%
Sample
Level Blank
LFB Recovery Result
(µg/L) (µg/L) (µg/L)
LFB
(µg/L)
Sample
Sample
Spike
%
Spike Dup
%
(µg/L) Recovery (µg/L) Recovery
As
30
<MDL
29.3
97.5
0.26
32.3
108
31.5
105
Cd
1.5
<MDL
1.42
94.5
0.031
1.49
99.1
1.54
103
Pb
30
<MDL
31.1
104
0.35
28.7
95.8
29.7
98.8
Se
30
<MDL
29.3
97.6
1.5
28.2
89.0
28.2
89.1
Tl
30
0.20
30.5
102
<MDL
30.2
101
33.4
111
Conclusion
The PerkinElmer PinAAcle 900T atomic absorption spectrometer is capable of meeting the
accuracy and precision requirements for the U.S. EPA and Health Canada regulations. The
method detection limit (MDL) for all elements was determined to be well below the practical
quantification limit (PQL) required by the method. This detection limit could be lowered even
further by using a larger sample volume or through the use of end-capped THGA tubes
(Part No. B3000655). Recoveries for two standard reference materials, spiked samples and
lab fortified blanks were all within the method’s acceptable limits.
References
1.U.S. EPA Method 200.9, Revision 2.2 and 3.0.
2.Environmental Protection Agency, "National Primary Drinking Water Regulations: Arsenic
and Clarifications to Compliance and New Source Contaminants Monitoring, Final Rule,"
Federal Register, vol. 66, no. 14 (January 22, 2001), p. 6981.
3.Environmental Protection Agency, “National Primary Drinking Water Regulations:
Inorganic Chemical sampling and analytical requirements,” 40 CFR 141.23.
PerkinElmer, Inc.
940 Winter Street
Waltham, MA 02451 USA
P: (800) 762-4000 or
(+1) 203-925-4602
www.perkinelmer.com
For a complete listing of our global offices, visit www.perkinelmer.com/ContactUs
Copyright ©2011, PerkinElmer, Inc. All rights reserved. PerkinElmer® is a registered trademark of PerkinElmer, Inc. All other trademarks are the property of their respective owners.
009634_01