Report on the use of Graphite furnace

1 Pages 1-10

▲back to top

1.1 Page 1

▲back to top
INSTITUTE OF APPLIED SCIENCES
THE UNIVERSITY OF THE SOUTH PACIFIC
REPORT ON THE USE OF GRAPHITE
rURNACE ATOMIC ABSORPTION
SPECTROSCOPY IN TRACE
METAL ANALYSIS
IAS TECHNICAL REPORT NO. 94/ 04
by
Richard K Coll
Niels O Kristensen1
Rolf Traberg 1
(1Exchange students from the
Engineering Acadamy of Denmark)
April, 1994

1.2 Page 2

▲back to top
REPORT ON THE USE OF GRAPHITE FURNACE ATOMIC ABSORPTION
SPECTROSCOPY IN TRACE METAL ANALYSIS
1NTRODUCTION
The trace level analysis of the metals Zn, Pb, Cr, Cd and Fe using graphite
furnace atomic absorption spectroscopy (GFAAS) has proven difficult in the
laboratory at the Institute of Applied Sciences (IAS). Typical problems
encountered include: considerable variation in replicates, high blank readings,
lack of linearity in calibration plots and lack of agreement between
concentrations determined and those cited for standard reference materials.
In this report we describe the investigation of these problems for these metals.
As well as detailing the most suitable instrumental parameters for the
determination of these elements, we describe the necessary procedures for the
handling of the glassware and the preparation of concentration standards. The
importance of these latter steps cannot be over-emphasised. High precision in
replicates whilst desirable, is no guarantee of the accuracy of the analyte
concentration determined, especially if the sample matrix is complex.
GENERAL INSTRUMENTAL PROCEDURE
The general instrumental procedure employed for the analyses was based on the
techniques outlined in the Perkin Elmer Manual for the GFAAS instrument.
The general procedure is the so-called STPF (Stabilised Temperature Platform
Furnace) concept and is a collection of instrumental and analysis procedures.
The details of the STPF procedure are given on the following page.
1

1.3 Page 3

▲back to top
The stabilised temperature platform furnace (STPF) concept requires the
following:
use of high quality pyrolytic graphite tubes
use of L'vov platform
use of maximum heating power to reach the atomisation temperature
interruption of the internal gas flow during the atomisation step
the measurement of absorbance using peak area rather than peak height
mandatory use of a matrix modifier
use of pretreatment temperatures of no lower than 1000°C below that of
the atomisation temperature
use of a suitable background correction procedure
Perkin Elmer recommend the use of Baseline Offset correction (BOC) prior to the
atomisation step. In this procedure the baseline is measured just prior to
atomisation. This takes into account any drift in baseline during analysis. In
order to get the best results for absorbance using peak area, it is necessary to
employ BOC. However, the GFAAS instrument at IAS does not have the facility
to do BOC. Because of this, in some instances it is better to use peak height
rather than peak area for the measurement of the absorbance. As a general rule
the absorbance should always be measured using peak area, especially if the
sample matrix is complex. The measurement of absorbance using peak height
is permissible if the element is volatile: e.g. zinc, cadmium, etc.
In our studies we employed the STPF concept and experimented in order to
optimise instrumental parameters . In each case the starting point for the
optimisation procedure was the recommended settings in the Perkin Elmer
Manual for the GFAAS instrument.
From our studies it seems that the use of the matrix modifier is one of the key
parts of the procedure. The use of the matrix modifier allows for the use of
higher pretreatment temperatures without significant loss of analyte. This
2

1.4 Page 4

▲back to top
enables atomisation of the sample to be similar to the atomisation of the
standards, with consequent reduction of possible interference effects. It is
important to use the matrix modifier routinely even if there is no obvious change
in precision or sensitivity for the concentration standards and reference
materials. In the analysis of the samples it is assumed that the samples behave
the same as the more simple matrix of the concentration standards or reference
standards . Clearly, if th.e matrix of the samples is unknown, it is difficult to
make this assumption with confidence. In our studies we inevitably saw an
increase in sensitivity upon the use of the matrix modifier. This occurs since
loss of analyse is minimised during the pre-treatment step.
No very pure samples of the two matrix modifiers used [Mg(N03 ) 2.6H20 and
NH 4H2P0 4] were available. As a result of this the absorbance of the blanks
obtained were higher than those seen without the use of the matrix modifier.
Although the blank readings were higher, they were more stable and
reproducible. New ultra-high purity samples of the two matrix modifiers should
be ordered.
Because of the high blanks readings obtained as a result of the use of the
impure matrix modifiers, the detection limits cited are estimates only. These
should be redetermined upon the arrival of the pure matrix modifiers.
THE ANALYSIS OF ZINC
The main problems encountered with the analysis of zinc by GFAAS were
considerable variation of results from one run to another and difficulty in
obtaining a linear calibration plot.
Zinc is a very sensitive element (characteristic mass 0.1 pg/0.0044 As) and is
one of the most common contaminants present in dust, water and acids .
In our studies high blank readings were obtained and we saw considerable
variation in the readings for the blank, the concentration standards and the
3

1.5 Page 5

▲back to top
samples. A further potential difficulty is interference with iron. There is a
spectral overlap between the most sensitive absorbance wavelength of zinc
(213.9 nm) with an iron emission line at 213.8 nm. This is one of the few
known instances of spectral interference in AAS. Clearly this can adversely
effect the results for the determination of zinc at low levels.
For these reasons it was decided to carry out the analysis of zinc using an
alternative wavelength, 307. 6 nm. This wavelength is less sensitive (by about
3500X) than the wavelength of 213.9 nm. Although the detection limit is
improved by the use of the matrix modifier (NH4H2P0 4 ) to ca. 250 ppb, the
detection limit is obviously higher than that obtained using the more sensitive
wavelength. However, there was good agreement between replicates for the
blank, standards and samples. The use of higher concentration standards
means that contamination effects are less qf a problem. The calibration plots
obtained were linear, and the sensitivity is adequate for the determination of the
concentration of this element. The WHO limit for drinking water is 5 ppm, and
so even using the less sensitive wavelength this is easily measurable.
In conclusion the loss of some sensitivity is offset by the increase in precision,
but more importantly the increase in accuracy and reliability of the analytical
results.
A Perkin Elmer consultant informed us that zinc is not normally determined by
GFAAS due to contamination problems. The use of the less sensitive
wavelength seems to have overcome this problem.
The only reference standard we have available for zinc was ca. 70 ppb, too low
for detection at the less sensitive · wavelength. In the future a more
concentrated reference standard should be obtained and used to check the
accuracy of the method.
4

1.6 Page 6

▲back to top
The recommended parameters for the analysis of zinc on the GFAAS are given
in Table 1.
THE ANALYSIS OF LEAD
The main problem encountered for the analysis of lead was the lack of
consistency from one run to another·.
Using the recommended instrumental parameters and the matrix modifier good
consistent results were obtained. The results were readily repeatable and two
reference samples (10.0 ppb and 34.9 ppb) regularly gave results that agreed
within experimental error.
An instance of contamination that occurred during analysis of the 34.9 ppb
reference standard is instructive. An initial high reading was obtained (ca. 40
ppb). This occurred when the reference solution was poured first into a beaker
and then into the auto injector sample vial. When the reference sample was
placed directly into a auto injector sample vial the correct value was obtained.
It is also important that the reference sample is made up in a similar matrix to
the standards, i.e., the same 0.2% AR grade nitric acid. This is necessary since
the AR grade nitric acid used in this laboratory also contains some analyte. This
can be achieved by adding the appropriate amount of the same concentrated
nitric acid used to make up the standards to the solution of the reference
standard (see appendix A).
The recommended parameters for the analysis of lead on the GFAAS are given
in Table 2.
THE ANALYSIS OF CHROMIUM
The analysis of chromium has been characterised by high and variable blanks,
poor calibration plots, and variation for replicates.
5

1.7 Page 7

▲back to top
Chromium is an involatile element (atomisation temperature 2500°), and is
subject to memory effects . However, good results were obtained using the
STPF system and recommended parameters. The memory effect is a particular
problem when analysing low concentration standards. During our investigation
we used three replicates. In every instance in which we measured the
absorbance for a high concentration standard (e.g. with absorbances of ca.
0.30%) the first reading of the following sample was high. This memory effect
is related to the lack of volatility of the element . We investigated a number of
methods of removing this memory effect, e.g. altering the clean-up step,
including an extra clean-up step etc., but were unable to completely eliminate
the memory effect.
It is our recommendation that there should be a blank inserted after the third
standard, and between each sample. The blank here is acting to clean the
furnace before the next analysis . It might seem that this is an extreme measure
as it will require more sample runs, but ultimately will involve less analyses than
repeating spurious results.
It is also recommended that for chromium (and other involatile elements) that
three replicates rather than two be measured . This will help to monitor the
presence of any memory effect.
The recommended parameters for the analysis of chromium on the GFAAS are
given in Table 3.
THE ANALYSIS OF CADMIUM
Cadmium is a volatile and sensitive element (atomisation temperature 1600°C,
characteristic mass 0.35 pg/0.0044 As). Previous analyses have been
characterised by lack of linearity in calibration plots, and inconsistency between
replicates.
The use of the standard settings STPF gave consistent reproducible results.
6

1.8 Page 8

▲back to top
The recommended parameters for the analysis of cadmium on the GFAAS are
given in Table 4 .
THE ANALYSIS OF IRON
Iron is a non-volatile element that is highly subject to contamination . A severe
memory effect was seen and there was great variety in the results. We tried a
- number of variations to the procedure to improve the analysis including many
different wavelengths. However, it proved impossible to obtain linear
calibration plots or consistent results. It is our recommendation that the
analysis of iron be carried out on the flame AAS. Like zinc, most other
analytical laboratories use flame AAS for analysis.
SUMMARY
The analysis of the elements zinc, cadmium, chromium and lead can now be
performed with confidence using the GFAAS. The use of a matrix modifier is
important in reducing the likelihood of any interference effect and in improving
the detection limits. The cleanliness of the glassware and care in making up of
standard solutions is of paramount importance . Great care must also be taken
in handling the auto injector vials. The analysis of iron is best carried out using
flame AAS .
7

1.9 Page 9

▲back to top
APPENDIX A
Preparation of Glassware
When carrying out analyses at the ppb level the cleanliness of glassware is of
great importance. The following procedures were used for all analyses carried
out in this work.
1) The same pipettes, volumetric flasks and caps for the volumetric flasks
were retained throughout the analysis . Standards were made up daily in
100 ml volumetric flasks.
· 2)
New volumetric flasks and pipettes were soaked overnight in 10% AR
grade HN03 • If a volumetric flask was not used for more than two days it
was again soaked in 10% HN03 •
3) Each volumetric flask and pipette was rinsed 3 times with deionised
water, followed by 3 times with the same 0.2% HN0 3 solution used in
preparing the blank and in dilution of the samples or reference standards.
4) The autosampler vials were stored in 10% AR grade HN03 solution prior
to use . Auto-injector vials were handled using plastic gloves.
Preparation of Concentration Standards
All standard solutions were prepared using glassware treated as outlined above .
In addition, the following procedures were employed to reduce the errors in the
concentrations.
1) To reduce percentage errors in the concentrations of the standards,
pipettes of a minimum volume of 5 ml were used.
8

1.10 Page 10

▲back to top
2) The nitric acid diluent (0.2% AR grade HN03 ) was prepared by placing 3
ml of 70% AR grade HN03 into a 1L volumetric flask. This flask was
retained for the preparation of the acid diluent.
3) Stock solutions were prepared as outlined in the Perkin Elmer Manual.
4) The standards were prepared using 0 .2% AR grade HN03 as a diluent.
5) Samples and references were acidified by the addition of 0.3 ml of Ar
grade HN03 into a 100 ml flask and filling to the mark with the sample
solution. Any necessary dilutions of the samples were carried out using
0.2% AR grade HN03 as a diluent .
Procedure for the Handling of Auto Injector Sample Vials
1) Vials were removed from the 10% AR grade HN03 solution and rinsed
three times with 0.2% AR grade HN03 •
2) The vials were then rinsed 3 times with deionised water, and three times
with the analyte solution.
3) Finally the vials were filled with the analyte solution and placed into the
autosampler tray.
9

2 Pages 11-20

▲back to top

2.1 Page 11

▲back to top
APPENDIX 8
Calculation of Matrix Modifier Concentration
The amount of matrix modifier required for each element is given in the Perkin
Elmer Manual. The mass given in the manual is the mass in mg of the
compound required for each injection, e.g. for cadmium the mass required 0.2
mg per injection. The matrix modifier is added using the alternative volume
switch on the autosampler, and typically a volume of 10 µl is delivered. Matrix
modifiers were made up in 100 ml solutions. The calculation of the
concentration of matrix modifier for 100 ml of solution is illustrated in Box 1.
10

2.2 Page 12

▲back to top
APPENDIX C
Trouble-Shooting Guide
1) Variation in replicates
adequate warm-up time allowed for hollow cathode lamp?
check auto-injector tip for alignment
check auto-injector tip for delivering drop onto platform
check tube age, less than ""300 firings
check age of standards: standards made up fresh daily
2) Lack of linearity in calibration plots
absorbance less than 0.4% for most elements and less than 0.2%
for zinc, tellurium and phosphorus
check age of standards: standards made up fresh daily
check for contamination for sensitive elements
3) Poor agreement with reference standards
check for contamination of auto-injector vials
check reference standard acidified using same acid as
concentration standards
11

2.3 Page 13

▲back to top
APPENDIX D
WHO Recommended Limits for Metal Concentration in Drinking Water
The WHO limits for the metal concentration for the metals studied in this work,
along with the detection limits determined are given in Box 2.
12

2.4 Page 14

▲back to top
Table 1 GFAAS Parameters for the Analysis of Zinc
Element
Zinc
Standards (ppb)
#1
500
#2
1000
#3
5000
Matrix modifier
NH 4 H2P0 4
Concentration (g 100 ml.1)
0.104
Lamp current (mA)
10
Wavelength (nm)
307.6
Absorbance mode
peak height
Slit width (mm)
0.7
Sample volume (µL)
20
Alternative volume (µL)
10
Number of replicates
2
Maximum power setting
39
Furnace Program
Step
Temperature
Ramp
Hold
1
90
1
30
2
120
20
30
3
700
20
30
4
1800
0
4, read, stop flow
5
2600
1
3
6
20
1
15
Comments:
Detection limit : ca. 250 ppb
Non-linear for absorbance greater than 0 .2%. Stock solution needs to be less than one month old.
13

2.5 Page 15

▲back to top
Table 2 : GFAAS Parameters for the Analysis of Lead
Element
Lead
Standards (ppb)
Matrix modifier
Lamp current (mA)
#1
10
NH 4H2PO 4 + Mg(NO 3 ).6H 2O
4
#2
20
Concentration (g 100 mL-1)
Wavelength (nm)
#3
40
2g and 1. 7g respectively
283.3
Absorbance mode
Peak height
Slit width (nm)
0.7
Sample volume (µL)
20
Alternative volume (µL)
10
Number of replicates
2
Maximum power setting
40
Furnace Program
Step
Temperature
Ramp
Hold
1
90
20
30
2
120
10
30
3
650
20
30
4
1800
0
4, read, stop flow
5
2600
1
3
6
20
1
15
Comments:
Detection limit : ca. 5.0 ppb. For more complex matrices absorbance measurement using peak area would be
more suitable using following settings: pretreatment temperature 650, and maximum power setting 35.
14

2.6 Page 16

▲back to top
Table 3 : GFAAS Parameters for the Analysis of Chromium
Element
Chromium
Standards (ppb)
#1
5
#2
10
#3
15
Matrix modifier
Mg(NO3) 2 .6H 20
Concentration (g 100 mL·1)
0.866
Lamp current (mA)
12
Wavelength (nm)
357.9
Absorbance mode
Peak height
Slit width (nm)
0.7
Sample volume (µl)
20
Alternative volume (µL)
10
Number of replicates
3
Maximum power setting
40
Furnace Program
Step
Temperature
Ramp
Hold
1
90
20
30
2
120
10
30
3
1650
20
30
4
2500
0
4, read, stop flow
5
2650
.1
3
6
20
1
15
Comments:
Detection limit : ca. 1.0 ppb. Considerable memory effect seen for this element.
15

2.7 Page 17

▲back to top
Table 4 : GFAAS Parameters for the Analysis of Cadmium
Element
Chromium
Standards (ppb)
#1
1
#2
2
#3
3
Matrix modifier
Lamp current (mA)
NH4H2P0 4
6
Concentration (g 100 ml.1)
Wavelength (nm)
2.00
228.8
Absorbance mode
Peak height
Slit width (nm)
0.7
Sample volume (µL)
20
Alternative volume (µL)
10
Number of replicates
2
Maximum power setting
30
Furnace Program
Step
Temperature
Ramp
Hold
1
90
20
30
2
120
10
30
3
700
20
30
4
1600
0
4, read, stop flow
5
2600
1
3
6
20
1
15
Comments:
Detection limit : ca. 0.5 ppb. Contamination likely for this element.
16