The document summarizes research on directly determining bismuth, indium, and lead in sea water samples using electrothermal atomic absorption spectrometry with Zeeman background correction and chemical modifiers. Molybdenum-containing chemical modifiers combined with tartaric acid were found to increase the maximum pyrolysis temperatures for the analytes, allowing more of the sea water matrix to be removed without loss of analyte. The combination of molybdenum, palladium, and tartaric acid or molybdenum, platinum, and tartaric acid produced the best results, yielding recoveries of 94-103% for spiked analytes in synthetic and real sea water, compared to 49-61% without

1. Talanta 49 (1999) 135–142
Direct determination of bismuth, indium and lead in sea
water by Zeeman ETAAS using molybdenum containing
chemical modifiers
Orhan Acar a
, A. Rehber Tu¨rker b,
*, Ziya Kılıc¸ c
a
TAEA, Ankara Nukleer Aras¸tırma 6e Eg˘itim Merkezi, 06983, Ankara, Turkey
b
Gazi U8 ni6ersitesi, Fen Edebiyat Faku¨ltesi, 06500, Ankara, Turkey
c
Gazi U8 ni6ersitesi, Gazi Eg˘itim Faku¨ltesi, 06500, Ankara, Turkey
Received 9 July 1998; received in revised form 20 October 1998; accepted 12 November 1998
Abstract
Direct determination of Bi, In and Pb in sea water samples has been carried out by ETAAS with Zeeman
background correction using molybdenum containing chemical modifiers and tartaric acid as a reducing agent.
Maximum pyrolysis temperatures and the effect of mass ratios of the mixed modifier components on analytes have
been investigated. Mo+Pd+TA or Mo+Pt+TA mixture was found to be powerful for the determination of 50 mg
l−1
of Bi, In and Pb spiked into synthetic and real sea waters. The accuracy and precision of the determination were
thereby enhanced. The recoveries of analytes spiked were 94–103% with Mo+Pd+TA or Mo+Pt+TA and they
are only 49–61% without modifier. © 1999 Elsevier Science B.V. All rights reserved.
Keywords: ETAAS–Zeeman background correction; Bismuth; Indium; Lead; Sea water; Chemical modifiers
1. Introduction
In recent years, determination of volatile ele-
ments such as bismuth, indium and lead in sea
water samples has become increasingly important
in order to monitor the pollution of this environ-
ment. Owing to its high sensitivity and specificity,
electrothermal atomic absorption spectrometry
(ETAAS) with Zeeman background correction is
one of the most promising methods for the direct
determination of trace elements in water samples
[1–3]. However, the high salt contents of sea
water matrix and very low concentrations of
volatile elements cause considerable difficulty in
the direct analysis. During the determination of
volatile elements by ETAAS, chemical interfer-
ences and significant background absorbance of-
ten occur in graphite furnace due to
covolatilization of the analyte with the matrix
[2,3].
To overcome most of the matrix interferences
and to increase the accuracy of determination in
* Corresponding author. Tel.: +90-312-212-2900; fax: +
90-312-212-2279.
E-mail address: rturker@quark.fef.gazi.edu.tr (A.R. Tu¨rker)
0039-9140/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved.
PII: S0039-9140(98)00358-0

2. O. Acar et al. / Talanta 49 (1999) 135–142136
ETAAS, the application of chemical modification
for the stabilization of volatile elements during
pyrolysis stage is one of the methods that has met
with reasonable success for analyte elements [1].
For this purpose many metal salts, their mixtures
and some organic substances have been recom-
mended and used as chemical modifiers [2–11].
The aim of matrix modification is to permit the
pyrolysis temperature to be high enough to re-
move the balk of the interfering substances during
the pyrolysis stage without loss of analyte before
atomization stage. In our previous works [10,11],
the effects of tungsten, palladium and molybde-
num containing chemical modifiers as well as
tartaric acid (TA) for the thermal stabilization of
Bi, In and Pb have been investigated systemati-
cally and applied to the determination of Bi and
Pb in geological samples. Several matrix modifiers
and organic substances have been recommended
and used to determine lead in water samples [7,8].
The aim of presented work is to determine Bi,
In and Pb in synthetic and real sea water samples
by using molybdenum containing chemical
modifiers. Tartaric acid (TA) was added as chem-
ical reductant to the sample solutions together
with the mixed modifiers to provide higher ther-
mal pre-treatment temperatures of analytes, to
modify the form of modifier and help to reduce
chemical interferences. Tartaric acid produces re-
ductants such as C, CO and H2 by thermal de-
composition, similar to ascorbic acid or oxalic
acid used by Byrne et al. [12]. Addition of tartaric
acid should assure an early reduction of modifiers
and analytes to highly dispersed metallic forms in
the temperature programme and elimination of
chlorides as HCl (g). Efficient reductants pro-
duced from TA at the surface of the tube might
react with the chlorides in the aqueous matrices to
form HCl (g) which is removed at low tempera-
tures [13], thus preventing the loss of analyte as
chloride. Therefore, Mo+Pd+TA and Mo+
Pt+TA mixed modifiers that provided the higher
pre-treatment temperatures for the analytes [11]
were applied to the direct determination of spiked
Bi, In and Pb in synthetic and real sea water
samples by ETAAS with Zeeman background
correction system.
2. Experimental
2.1. Instrumentation
A Hitachi Model 180/80 flame and graphite
furnace atomic absorption spectrometer equipped
with a Zeeman effect background corrector, au-
tosampler (P/N-170/126) and an automatic data
processor was used for all measurements. Details
about the AAS equipment, thermal stabilizing
studies and optimized graphite furnace tempera-
ture programme for wall atomization is given in
previous works [10,11]. The operating parameters
were set as recommended by the manufacturer
except that a spectral band pass of 1.3 nm and
alternate wavelength was used for Bi (306.8 nm).
The light sources are single-element hollow
cathode lamps. A Varian model 9176 recorder
was connected to AAS to obtain atomization
profiles of analytes in a 20 mV/FS spans. Hitachi
pyrolytically coated graphite tubes (P/N-180/
7444) were used for atomization throughout. All
absorbance values were obtained by using inte-
grated mode. Argon served as a carrier gas in 300
ml min−1
flow rate.
2.2. Reagents
All solutions were prepared by dissolving ana-
lytical grade reagents in deionized water distilled
in an Elgastat type C114 distillation unit. Stock
standard solutions (2000 mg ml−1
) of Pd and Pt
were prepared according to [9]. 0.4% (m/v) of Mo
(VI) was prepared from H2MoO4 (Merck) dis-
solved in 1% (v/v) ammonia solution. All modifier
solutions were diluted as required. 4% (m/v) tar-
taric acid (TA) solution was prepared daily. 1000
mg ml−1
of Bi, In and Pb standard solutions
(BDH chemicals) were used. Working standard
solutions were freshly prepared by successive dilu-
tion to the desired concentrations using 0.2%
HNO3.
2.3. Procedure
Synthetic sea water was prepared according to
Cantle [14]. 2.67 g of NaCl, 0.54 g of MgCl2, 0.11
g of CaCl2 and 0.08 g of KCl salts were dissolved

3. O. Acar et al. / Talanta 49 (1999) 135–142 137
in a teflon beaker with deionized water. 1 ml of
each Bi, In and Pb standard solutions (20 mg
ml−1
) were spiked into it and this solution was
transferred into a 100 ml glass calibrated flask.
The Teflon beaker was washed two times with
0.2% (v/v) nitric acid. 600 ml of conc. nitric acid
was also added to final solution to avoid the
absorption of analytes onto glass walls and di-
luted to the mark with deionized water. Sea wa-
ter sample was collected from the surface water
of the Marmara sea coast in Tu¨rkiye in a 2 I
ployethylene bottle. A sample portion was trans-
ferred into 100 ml volumetric flasks and imme-
diately acidified with 1 ml of conc. nitric acid. 1
ml of each Bi, In and Pb standard solutions (20
mg ml−1
) were spiked into these flasks because
of natural contents of the analytes are below the
detection limits. All of the flasks were filled with
sample (sea water) to the mark.
Synthetic or Marmara sea water sample con-
taining 200 mg l−1
analytes was diluted by a
factor 4 (0.5 ml of sample +1.5 ml of deion-
ized water) in order to decrease interferences in
the absence of modifier. In the presence of
modifier, 0.5 ml of sample solution and 0.5 ml
of water were added to 1 ml of modifier solu-
tion (20 000 mg ml−1
of TA, 2000 mg ml−1
of
Mo, 400 mg ml−1
of Pd or Pt, 2000 mg ml−1
of
Mo+20 000 mg ml−1
of TA, 400 mg ml−1
of Pd or Pt). 0.5 ml of sample solution and
0.5 ml of TA (4% (m/v)) stock solution instead
of water were added to 1 ml of modifier
solution (2000 mg ml−1
of Mo+400 mg ml−1
of Pd or Pt) for the triple modifier mixtures
such as Mo+Pd+TA. 20 ml of the sample so-
lution (analyte concentration 50 mg l−1
) pre-
pared in the presence or absence of single or
mixed modifiers were injected into the graphite
tube.
The optimized graphite furnace temperature
programme for wall atomization is given in Ref.
[11]. The optimum modifier mass and mass
ratio of the mixed modifier components were
found to be 20 mg of Mo, 4 mg of Pd and Pt,
and 20/4 (mg/mg) of Mo/Pd and Mo/Pt. 200 mg
of TA was used together with the mixed
modifiers [11].
3. Results and discussion
3.1. Stabilizing effects of modifiers on analytes
In order to remove the high contents of salts in
sea water, higher pyrolysis temperatures for
analytes may be preferable. A chemical modifier
has been recommended for a long time to increase
maximum permissible temperature and to mini-
mize interferences in the determination of volatile
elements in ETAAS [2]. Maximum pyrolysis
temperatures (Tmax) of the analytes studied and
the optimum mass ratios of the mixed modifier
components were found in our previous work [11].
Obtained pyrolysis temperatures are approxi-
mately in agreement with Tsalev’s results [13,15].
When Mo+Pd+TA was used, higher pyrolysis
temperatures were obtained than the other
modifier mixtures studied by Havezov et. al. [16].
In this study, thermal pre-treatment curves for
analytes with TA, Pd+TA and Pt+TA modifiers
were also studied and lower pyrolysis tempera-
tures were obtained. The addition of TA to Pd
and Pt did not increase the pyrolysis temperature,
but it increased the formation of analyte atoms at
lower temperatures [12]. The effect of tartaric acid
on atomic absorption signal of analyte can be
explained primarily by reaction that occur
between gaseous analyte-containing molecules
such as PbO (g) and the reducing gases produced
by pyrolysis of tartaric acid during the atomi-
zation temperature ramp [12].
When the Mo+Pd+TA or Mo+Pt+TA was
added to the solution, the pyrolysis temperatures
could be increased up to 1350–1400°C for Bi and
In and 1250–1300°C for Pb to remove much of
sample matrix without the risk of analyte loss [11].
The mechanism of increased thermal stability can
be explained from the reaction between analyte
and modifier elements. Shan and Wang from
X-ray photoelectron spectra, suggested the
formation of Pb–Pd and Bi–Pd intermetallic
bonds on graphite furnace [17].
3.2. Effect of modifiers on the atomization
profiles of analytes
One advantage of chemical modification is that

4. O. Acar et al. / Talanta 49 (1999) 135–142138
the atomization signals become fairly symmetrical
and shifted to higher pre-treatment temperatures
and appearance times [18]. Welz et al. [4] demon-
strated this behaviour by comparing the shapes of
atomization signals of various elements in both
the absence and presence of a Pd–Mg modifier.
In order to demonstrate how the Mo, Mo+Pd+
TA and other Mo-based modifiers affect the at-
omization and background profiles of analytes, a
comparative study was conducted. Fig. 1 shows
that the analyte and background atomization
profiles of sea water spiked with 50 mg l−1
of lead
as an example. The atomization and background
absorption profiles of spiked Pb for sea water and
for aqueous Pb standard solution obtained with
and without modifiers were compared. As can be
seen in Fig. 1, although similar symmetrical atom-
ization signal shapes were obtained for Pb in sea
water, both in the absence and presence of chemi-
cal modifiers, the atomic signals appeared at an
earlier time in the absence of a modifier mixture
than in its presence. The appearance times of the
peaks were identical for sea water and for
aqueous standard. However, the maximum peak
times of Pb in sample were slightly later than
those in the aqueous Pb standard. When Mo+
Pd+TA mixture has been used, the appearance
time of atomization signal for Pb in sample and in
aqueous standard was shifted to a later time while
increasing pyrolysis temperature and no reduction
in atomic absorption signal was observed up to a
maximum pyrolysis temperature above 1200°C. It
may be due to early reduction of analyte and
modifiers to reactive metallic forms. Thermal sta-
bility of these compounds is higher and the corre-
sponding inter-metallic compounds (or alloys
between them) [17] are formed when Mo+Pd+
TA is used. With no modifier was used, small
analyte signal and higher background absorbance
were obtained for Pb in sample even when the
solution was diluted by a factor of 1+3. When
Mo+Pd+TA was used, it was observed that
analyte signal increased while the background
absorption decreased. As can be seen in Fig. 1,
higher signal/noise ratios of Pb was obtained in
the presence of Mo+Pd+TA than this obtained
in the absence of modifier. The reason for this is
attributed to the behavior of mixed chemical
modifier. Tartaric acid may reduce the modifiers
and analytes to their free reactive metals at pyrol-
ysis temperatures less than 800°C [18] and there-
fore the stabilizing effect of modifier is increased.
It can be seen in Fig. 1 that under these condi-
tions the background absorbance from the matrix
is perfectly corrected using Zeeman corrector
which allows direct determination of sea water
into furnace for the determination of analytes.
Fig. 1. Atomization profiles of Pb in aqueous standard and in
sea water sample; where S1, standard (50 mg l−1
); S2, sample
(50 mg l−1
); B1 and B2, background absorption profiles of Pb
in standard and sample respectively.

5. O. Acar et al. / Talanta 49 (1999) 135–142 139
Table 1
Detection limits (LOD, 3s-criterion) and characteristic masses (m0) for Bi, In and Pb determination using some modifiers
Analytesa
Matrix modifiers
PbInBi
m0, pgLOD, mg l−1
m0, pgLOD, mg l−1
LOD, mg l−1
m0, pg
111.4 6.26No modifier 22.69 175.0 15.79 43.0
5.4494.4 37.2TA 13.2118.43 168.4
10.15 81.1 5.07 31.1Mo 17.03 161.9
78.2 4.83Mo+TA 15.57 131.6 8.74 30.2
4.7776.4 28.9Pd+TA 7.9914.64 120.6
5.44 65.9 2.25Mo+Pd 27.110.26 105.3
3.68 33.2 0.97Mo+Pd+TA 7.03 23.551.6
a
Analytical range: 20–400 mg l−1
for Bi and In; 5–80 mg l−1
for Pb with or without of single or mixed modifier mixtures. Sample
volume: 20 ml.
3.3. Analytical conditions and calibration
The determination of Bi, In and Pb in sample
solutions were performed with and without Mo-
based modifiers by using the calibration on the
base of single element solutions in nitric acid for
comparison. Calibration against standard solu-
tions in the presence or absence of modifiers was
performed for analytes by using optimum
parameters such as furnace temperature pro-
gramme, modifier mass and mass ratios of the
mixed modifiers as described in [11]. To obtain
calibration graph, working standard solutions of
analytes in analytical ranges given in Table 1 were
added to the modifier solutions as described in
Section 2.3. All calibration graphs were linear and
correlation coefficients were about 0.999.
Detection limits (LOD, 3s-criterion) and char-
acteristic mass (m0) of Bi, In and Pb spiked to sea
water (20 mg l−1
of each analyte) were determined
in the presence or absence of some modifiers and
given in Table 1. As can be seen in Table 1, the
lowest detection limits and characteristic masses
of analytes were obtained by using Mo+Pd+
TA mixed modifier.
3.4. Application
For new modifiers, they are particularly impor-
tant that their stabilizing powers for analyte ele-
ments persist also in the presence of complex
matrix and of high salt concentrations. Therefore,
Mo+Pd+TA and Mo+Pt+TA mixed
modifiers providing higher pyrolysis temperatures
for the analytes and their components providing
lower pyrolysis temperatures were applied to the
determination of Bi, In and Pb in synthetic and
real sea water samples. Results were given in
Tables 2 and 3. The determination of analytes in
sample solutions were performed with and with-
out modifier by using calibration graph method
and seven parallel sample solutions. As can be
seen in Tables 2 and 3, a good accuracy was
obtained by using mixed modifiers containing TA.
When higher chloride concentrations are to be
expected in samples such as sea water, the chlo-
ride interferences are to be important to deter-
mine volatile elements. As can be seen in Tables 2
and 3, Mo+Pd+TA or Mo+Pt+TA mixed
modifiers could be used for the determination of
Bi, In and Pb in sea water with about 5% relative
error. Tartaric acid acts on the chloride interfer-
ence in the matrix. It can be plausibly concluded
that tartaric acid is a valid modifier for reducing
the chloride ions in samples [13]. The products of
TA can be convert chloride ions to HCl (g) and
high contents of chloride ions vaporize at low
temperatures in the pyrolysis stage. Especially
Mo+Pd+TA and Mo+Pt+TA mixtures were
efficient to allow the use of pyrolysis temperatures
in the range of 1350–1400°C for Bi and In and
1250–1300°C for Pb. These temperatures could be

6. O. Acar et al. / Talanta 49 (1999) 135–142140
Table 2
Recovery tests performed for bismuth, indium and lead determination in synthetic sea water
Pyrolysis temperature (°C) ConcentrationsElement determined Modifier Recovery (%)
Founda
Spiked (mg l−1
)
800 50.0 24.494.6 49Bi No modifier
5829.193.750.0750TA
50.0 36.192.4Mo 1000 72
50.0 41.292.5Pd 1200 82
42.592.150.0 851150Pd+TA
1100 50.0 39.792.2 79Pt
8441.992.050.01100Pt+TA
50.0 40.791.9Mo+Pd 1300 81
50.0 47.291.8Mo+Pd+TA 1400 94
8843.991.950.01300Mo+Pt
1350 50.0 48.491.5 97Mo+Pt+TA
28.992.550.0 581000In No modifier
50.0 31.592.3TA 63850
50.0 36.492.1Mo 1100 73
7939.692.150.01250Pd
41.991.9 84Pd+TA 1150 50.0
7738.392.450.01200Pt
50.0 39.392.0Pt+TA 1150 79
50.0 42.591.8Mo+Pd 1400 85
9748.391.550.01350Mo+Pd+TA
1300 50.0 43.491.9 87Mo+Pt
9648.291.51350 50.0Mo+Pt+TA
50.0 24.392.6Pb No modifier 49800
50.0 31.992.5TA 750 64
35.692.250.0 711000Mo
7738.492.1Pd 1100 50.0
8040.191.850.01150Pd+TA
50.0 37.991.9Pt 1050 76
50.0 39.292.0Pt+TA 1100 78
9145.392.150.01250Mo+Pd
9547.591.6Mo+Pd+TA 1300 50.0
50.0 42.991.9Mo+Pt 1200 86
10351.491.51250 50.0Mo+Pt+TA
a
Mean of seven replicate measurements with 95% confidence level.
sufficient to volatilize many matrix constituents
without loss of analyte elements. When no
modifier is used, the error is very high because of
matrix interferences. Such effects are negligible
when Mo+Pd+TA or Mo+Pt+TA is present
and accuracy and precision are satisfactory. It can
be expected, therefore, that the recommended
chemical modifiers are applicable to the determi-
nation of Bi, In and Pb in sea water samples.
4. Conclusion
Direct determination of volatile elements such
as Bi, In and Pb in sea water by ETAAS using
Mo-based modifiers were examined. In general
terms, the analytical problems arising from the
sample matrix can be controlled only with dilu-
tion and using Mo+Pd+TA or Mo+Pt+TA
mixed modifier and Zeeman background correc-

7. O. Acar et al. / Talanta 49 (1999) 135–142 141
Table 3
Recovery tests performed for bismuth, indium and lead determination in Marmara sea water
Pyrolysis temperature (°C) ConcentrationsElement determined Recovery (%)Modifier
Founda
Spiked (mg l−1
)
30.694.450 61800Bi No modifier
50 32.793.4TA 65750
50 36.492.6Mo 1000 73
37.692.550 751050Mo+TA
50 40.592.3Pd+TA 1150 81
7839.292.4501100Pt+TA
50 43.892.1Mo+Pd 1300 88
50 48.891.6Mo+Pd+TA 1400 98
8441.992.3501300Mo+Pt
1350 50 47.191.8 94Mo+Pt+TA
30.794.250 611000In No modifier
50 32.593.7TA 65850
50 35.492.9Mo 1100 71
7437.292.2501150Mo+TA
50 41.292.6Pd+TA 1150 82
39.892.3 80Pt+TA 1150 50
50 42.892.0Mo+Pd 1350 86
50 49.091.4 981400Mo+Pd+TA
50 45.392.3Mo+Pt 1300 91
9547.591.61350 50Mo+Pt+TA
50 28.394.1Pb No modifier 57800
50 32.893.7TA 750 66
35.293.250 701000Mo
37.492.6 75Mo+TA 1050 50
7839.192.4501150Pd+TA
50 38.392.5Pt+TA 1100 77
50 41.592.3Mo+Pd 1250 83
10150.391.4501300Mo+Pd+TA
50 41.192.4Mo+Pt 1200 82
50 47.591.7Mo+Pt+TA 951250
a
The mean of seven replicate measurements with 95% confidence level.
tion. Recoveries of analytes spiked to water sam-
ples were 94–103% with Mo+Pd+TA or Mo+
Pt+TA. The recoveries of analytes are only
49–61% without modifier.
Acknowledgements
The supports of Turkish Atomic Energy Au-
thority and Gazi University Research Found are
gratefully acknowledged.
References
[1] A.L. Garcia, E.B. Gonzalez, A. Sanz-Medel, Microchim.
Acta 112 (1993) 19.
[2] Y.-Z. Liang, M. Li, Z. Rao, Fresenius J. Anal. Chem. 357
(1997) 112.
[3] H. Shan-Da, L. Wen-Rong, S. Kun-Yauh, Spectrochim.
Acta 50 (1995) 1237.
[4] B. Welz, G. Schlemmer, J.R. Mudakavi, J. Anal. At.
Spectrom. 7 (1992) 1257.
[5] B. He, Z.-M. Ni, J. Anal. At. Spectrom. 11 (1996) 165.
[6] D.L. Tsalev, V.I. Slaveykova, J. Anal. At. Spectrom. 7
(1992) 147.

8. O. Acar et al. / Talanta 49 (1999) 135–142142
[7] S. Sachsenberg, T. Klenke, W.E. Krumbein, H.J. Schelln-
huber, E. Zeeck, Anal. Chim. Acta 279 (1993) 241.
[8] I. Sekerka, J.F. Lechner, Anal. Chim. Acta 254 (1991) 99.
[9] S. Imai, N. Hasegawa, Y. Nishiyama, Y. Hayashi, K.
Saito, J. Anal. At. Spectrom. 11 (1996) 601.
[10] O. Acar, A.R. Tu¨rker, Z Kılıc¸, Fresenius J. Anal. Chem.
357 (1997) 656.
[11] O. Acar, A.R. Tu¨rker, Z Kılıc¸, Fresenius J. Anal. Chem.
360 (1998) 645.
[12] J.P. Byrne, C.L. Chakrabarti, G.F.R. Gilchrist, M.M.
Lamoureux, P. Bertels, Anal. Chem. 65 (1993) 1267.
[13] D.L. Tsalev, V.I. Slaveykova, Spectrosc. Lett. 25 (1992)
221.
[14] J.E. Cantle, Atomic Absorption Spectrometry, Elsevier,
New York, 1982.
[15] V.I. Slaveykova, D.L. Tsalev, in: D.L. Tsalev (Ed.), XXVI
CSI, July 2–9 Selected papers, vol. 7, 1989, p. 85.
[16] I. Havezov, A. Detcheva, J. Rendl, Mikrochim. Acta 119
(1995) 147.
[17] X.-Q. Shan, D.X. Wang, Anal. Chim. Acta 173 (1985) 315.
[18] X.-Q. Shan, B. Wen, J. Anal. At. Spectrom. 10 (1995)
791.
.