Clinical Biochemistry 46 (2013) 123127

Contents lists available at SciVerse ScienceDirect

Clinical Biochemistry
journal homepage: www.elsevier.com/locate/clinbiochem

Direct determination of tin in whole blood and urine by GF AAS

S.V. De Azevedo a,, F.R. Moreira a, 1, R.C. Campos b, 2
a
Laboratory of Toxicology, Center of Studies of Worker's Health and Human Ecology, National School of Public Health, Oswaldo Cruz Foundation, Rio de Janeiro, Brazil
b
Department of Chemistry, Pontical Catholic University of Rio de Janeiro (PUC), Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Objective: The aim of this work was to develop a procedure for the determination of tin in whole blood
Received 29 May 2012 and urine by GF AAS with a minimum sample pre-treatment, using Pd/Mg as chemical modier.
Received in revised form 1 September 2012 Design and methods: The analyses of tin were conducted using an atomic absorption spectrometer with
Accepted 19 September 2012 Zeeman background correction. The laboratory staff volunteered blood and urine samples for the experimental
Available online 29 September 2012
studies and application of the methodology.
Results: Samples were just diluted with 0.2% v/v Triton X-100, and pyrolysis and atomization temperatures
Keywords:
Tin
of 1300 and 2200 C were used. External calibration was performed with matrix matched calibration solutions.
Blood Limits of detection of 2.7 and 0.8 g L1 were reached for blood and urine, respectively. The method was
Urine applied to the determination of Sn in blood and urine of eleven subjects not occupationally exposed, working
Toxicology in a laboratory of toxicology in a large Brazilian city, and the results ranged from 7.4 to 11.2 g L1 and 0.8
Atomic absorption spectrometry to 2.2 g L1, for blood and urine, respectively. Accuracy was assessed by analysis of standard reference mate-
rials for tin in blood (Contox I, lot TM144-1097, Kaulson Laboratories, USA) and urine (Seronorm, lot 0511545,
Sero AS, Norway).
Conclusions: Results showed good agreement between experimental and reference values according to the
Student's t test at 95% of condence.
2012 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.

Introduction as a protective coating in food, beverage and aerosol cans. It is also

Tin (Sn) is a naturally occurring metal, obtained from ores such as
cassiterite (SnO2). In nature, tin can combine with other elements to
form inorganic as well as organotin compounds. Tin and its compounds
can be found in air, water and soil, near the rocks, mines and industries
where it is present or used [13].
Tin is widely used in industry due to features such as low melting
point, afnity to form alloys, and corrosion resistance. Inorganic com-
pounds of Sn (iSn) are used in toothpaste, perfume, soap, food addi-
tives and dyes [1,4]. The main commercial applications of organic
compounds (oSn) are as stabilizers in PVC, pesticides for agricultural
use, plastic stabilizers, preservatives (wood, cotton and paper), in the
glass industry and as marine antifouling agent [13]. Tin metal is used
Abbreviations: GF AAS, Graphite furnace atomic absorption spectrometry; AAS,
Atomic absorption spectrometry; ICP MS, Inductively Coupled Plasma Mass Spectrom-
etry; Pd, palladium; Mg, magnesium; Pd(NO3)2, palladium nitrate; Mg(NO3)2, magne-
sium nitrate modier; Sn, tin; SnO2, tin dioxide; iSn, inorganic compounds of Sn; oSn,
organic compounds of Sn; PVC, Polyvinyl Chloride; HNO3, nitric acid; ATSDR, Agency
for Toxic Substances and Disease Registry.
Corresponding author at: Laboratory of Toxicology, Center of Studies of Worker's
Health and Human Ecology (CESTEH), National School of Public Health (ENSP),
Oswaldo Cruz Foundation (FIOCRUZ), 1480 Leopoldo Bulhes St., Manguinhos, Rio de
Janeiro, RJ, CEP 21041-210, Brazil. Fax: +55 21 2270 3219.
E-mail address: biosva@yahoo.com.br (S.V. De Azevedo).
1
Fax: +55 21 2270 3219.
2
Deceased author.
present in alloys such as brass, bronze and pewter, and some welding
materials [1,4]. However, it is one of the elements least studied in re-
gard to human health, especially in relation to its presence in biolog-
ical indicators such as blood and urine [5].
Tin as well as other trace elements occurs naturally in soil at low
concentrations but in forms which are not readily available to humans
mostly [1,2]. However, some activities such as mining make the metal
available to the environment, contributing to the contamination of the
surrounding areas [6].
Exposure to Sn and its compounds can produce several effects
such as neurological, hematological and immunological. Inhalation
of iSn can induce to pneumoconiosis and ingestion may lead to gas-
trointestinal effects. Exposure to oSn inhibits the synthesis of heme
oxygenase and may be genotoxic, while its skin contact may cause se-
vere irritation and burning. Other effects include kidney and liver
damage [1,3]. Studies on human health related to tin are still incipient
in the literature, in part due to the scarcity of experiments in biological
uids of interest such as blood and urine [5].
Graphite furnace atomic absorption spectrometry (GF AAS) has
been widely used for the determination of trace elements in biological
uids due to its low limit of detection, reduced sample volume and min-
imum sample pre-treatment, saving time and reducing risks of loss or
contamination. In addition, other advantageous characteristics such as
selectivity, accuracy, precision, and accessibility make the technique
very attractive for such determinations [713].

0009-9120/$ see front matter 2012 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.clinbiochem.2012.09.020
124 S.V. De Azevedo et al. / Clinical Biochemistry 46 (2013) 123127

Other sensitive and accurate spectrometric techniques such as ICP
MS are turning more popular in the analysis of clinical uids, but due
to its lower costs, especially if a few elements are to be determined,
GF AAS is still largely used in the clinical laboratory [10]. Levels of
tin in blood and urine reported in the literature are scarce and
conicting [1], and the availability of simple, robust, reliable and ac-
cessible procedures may help to change this picture. Thus, the aim
of the present study was the development of a sensitive and accurate
method for the determination of total tin in blood and urine with a
minimum of sample pre-treatment by GF AAS.
Experimental
Instrumentation
A Perkin Elmer (Norwalk, CT., USA) AAnalyst 800 atomic absorp-
tion spectrometer equipped with a longitudinal Zeeman-effect back-
ground correction system and an AS-800 autosampler was used for
all measurements. Integrated absorbance was evaluated exclusively
and each measurement was the average of at least three replicates.
Instrumental operating parameters of the AAS apparatus are shown
in Table 1.
Laboratoryware
All plastic and glassware were decontaminated by immersion in
5% (v/v) Extran (Merck, Rio de Janeiro, Brazil) overnight and abun-
dantly rinsed with tap water. Afterwards, the material was rinsed
with deionized water, immersed in 10% (v/v) HNO3 (Merck) for at least
48 h, copiously rinsed with ultrapure water, obtained from a Milli-Q sys-
tem (Millipore, Bedford, USA) and dried at 60 C before use, avoiding any
contact with dust and metallic surfaces.
HNO3 in order to obtain 6 g of Mg(NO3)2 and 10 g of Pd as Pd(NO3)2
in 10 L of the modier solution.
Blood and urine samples
The laboratory staff volunteered blood and urine samples for the
experimental studies and application of the methodology. Eleven
whole blood and urine samples were obtained from male and female
adults not occupationally exposed and selected randomly.
The accuracy of the procedures was checked using reference samples.
In the absence of whole blood reference materials, a serum sample was
used as control (Contox I Serum, lot TM144-1097, Kaulson Laboratories,
USA), with a tin concentration of 32 g L1. On the other hand, a ref-
erence sample of urine (Seronorm Urine, lot 0511545, Sero AS, Norway)
at a concentration of 54.62.7 g L1 was used for controlling the urine.
Reference samples were reconstituted according to the manufacturer's
instructions and pre-treated as described in Sample preparation section.
Sampling and storage
Biological samples were collected after the study participants have
signed an informed consent form. Whole blood was collected by veni-
puncture with disposable and sterile needles using vacuum tubes
(Vacuette Ltd., SP, Brazil) with heparine as anti-coagulant, specic
for the determination of trace elements. Samples of urine were collected
in previously decontaminated containers of polyethylene. In addition,
subjects were instructed to follow some procedures for proper collection
of urine specimens such as washing hands before sampling, not touching
inside the lid or container, and close the sampling vessels immediately
after collection. Samples were stored at 20 C until analysis.
Sample preparation

Reagents, modier and standard solutions
All the reagents used were of analytical reagent grade. Aqueous
standard solutions were daily prepared in 0.2% (v/v) HNO3 (Merck)
by appropriate dilutions of a 1000 g mL 1 Sn standard solution
(Merck, New York, USA). Matrix matched calibration curves were pre-
pared using blood and urine samples with Sn content below the limit of
detection (LOD), spiked with adequate micro-volumes of the standard
solutions and diluted in 0.1% (v/v) Triton X-100 (Merck, Darmstadt,
Germany), as described below in Sample preparation section. Solutions
of Pd (NO3)2 (Merck, Darmstadt, Germany) and Mg(NO3)2 (Perkin
Elmer, Part NB019-0634), 10 g L1 each, were diluted in 0.2% (v/v)
Table 1
Instrumental operating parameters of the AAS apparatus.
Operating conditions
Blood and urine samples were diluted in 0.1% (v/v) Triton X-100.
The dilution ratios were 1 + 6 and 1 + 4 for blood and urine, respec-
tively. An aliquot of 200 L of blood or urine was added to 1200 or
800 L of 0.1% (v/v) Triton X-100, respectively.
Calibration curves were daily prepared using four standard inter-
mediate solutions (50, 100, 200 and 500 g L 1) from a Sn stock
1000 g L 1 standard in a 0.2% (v/v) HNO3 solution. Then 200 L of
either samples were mixed with 1130 or 750 L of 0.1% (v/v) Triton
X-100 for blood or urine, respectively, and 70 (blood) or 50 (urine)
L of each one of those standard intermediate solutions in a polyethylene
tube. After homogenization by mechanical shaking (vortex), an appro-
priate volume was transferred to the autosampler cup of the GF AAS in-
strument. All of these operations were performed in a laminar ow
hood to avoid contamination.
Results and discussion

Primary source
Lamp current
Analytical wavelength
Background correction system
Slit width
Mode
Graphite furnace operation
Atomization tube
Sheath/Purge gas
Injection volumes (modier/sample, L)
Modier (deposited mass, g)
Diluent (concentration, v/v)
Dilution ratio (urine / blood)
Sn Electrodeless Discharge Lamp (EDL) The statistical calculations (uncertainties associated to the calibra-
(Perkin Elmer Part n N3050675)
tion curves parameters and comparisons between characteristic mass
280 mA
286.3 nm values and sensitivity ratio to the unit, etc.) were performed according
Zeeman effect based (Longitudinal) to Gardiner [14], using a 95% condence level. All the results are an
0.7 nm average of at least 3 measurements. The temperature program, type of
Absorbance (peak area) diluent and dilution ratio as well as modier mass composition were
studied and optimized. These studies were performed by varying the pa-
rameter investigated while all others were kept xed at values shown in
THGA graphite tubes with end caps and Table 1.
integrated platform (Perkin Elmer Part n
B3000655)
Temperature program
Argon (Ar) of 99.999% purity
(White Martins, Brazil)
10/20 In general, sample and modier are successively injected before
Pd (10) + MgNO3 (6) running the temperature program in GF AAS. However, this possibility
Triton X-100 (0.1%) is not practical in matrices like blood, since subsequent drops of sample
1 + 4/1 + 6
and modier dispensed onto the platform do not mix properly, and
 S.V. De Azevedo et al. / Clinical Biochemistry 46 (2013) 123127 125

foaming and ejection are observed during the drying stage, impoverishing
the repeatability of the measurements. Moreover, according to the liter-
ature [15], protein precipitation may occur when blood sample and
modier solutions are mixed previously to the injection in the furnace.
Thus, in the present work, those problems were overcome by injecting
and drying the modier over the platform in the rst stage of the tem-
perature program, before addition of the sample solution. Then, blood
sample solution was added over a dried modier layer onto the plat-
form (stage 2). However, since that problem does not occur with the
urine, sample and modier solutions were dispensed conventionally
by subsequently pipetting of both solutions and running the tempera-
ture program (Table 2).
Pyrolysis and atomization temperature curves were performed in
the presence of blood and urine matrices as well as with aqueous
50 g L 1 Sn solutions (in 0.2% v/v HNO3). For these tests, blood
and urine samples were diluted 1 + 6 and 1 + 4, respectively, with
0.1% (v/v) Triton X-100, and both were also spiked with 50 g L 1
of inorganic Sn. In all cases, 10 L of the modier solution was used,
resulting in 10 g of Pd plus 6 g of Mg(NO3)2 deposited onto the
platform. The analysis of pyrolysis curves demonstrated that 1300 C
was the best choice as pyrolysis temperature in all situations, express-
ing the best compromise between maximum sensitivity and minimum
background. Accordingly, the background was always within the opti-
mum range correction of the equipment. Regarding the atomization
temperature, 2200 C appeared as the best one, resulting in maximum
sensitivity and a peak prole which could be resolved within only 5 s.

Sample dilution
One of the impediments to the direct injection of undiluted bio-
logical uids in GF AAS, especially viscous liquids like blood, is the
poor repeatability of the dispensed volume generating poor repeatability
of the atomic absorption signal. Therefore, sample dilution is necessary,
and consequently the choice of suitable diluents (nature and concentra-
tion) and dilution factors. In this way, 0.1% (v/v) Triton X-100, water and
0.2% (v/v) nitric acid were studied as diluents, while dilution factors
ranging from 3 to 10 for blood and 1.5 to 5 for urine were investigated
(Fig. 1). Those choices were based on previous similar studies found in
the literature [16]. For these tests, all sample solutions were spiked
with 50 g L1 of Sn, after the sample dilution.
According to Fig. 1, 0.1% v/v Triton X-100 showed the best perfor-
mance among the diluent solutions studied, leading to largest sensi-
tivities at the lowest dilution ratio. Triton X-100 helps breaking the
surface tension, facilitating the sample injection and reducing sample
losses to the autosampler capillary walls. Triton X 100 also avoids the
clogging of the capillary, which may occur after many readings, espe-
cially when dealing with samples such as blood. Clogging obstructs
the injection of the correct sample volume onto the platform, leading
to inaccurate results. For blood sample, sensitivity losses due to the
presence of the matrix were observed up to dilution ratios of 1 + 6,
with the signal remaining constant from this ratio. Since no signicant
sensitivity difference due to the presence of the blood matrix was ob-
served with dilution ratios from 1 + 6 on, this dilution ratio was chosen
for the determination of tin in blood, aiming at reaching the lowest
Table 2
Temperature program for the determination of tin in blood and urine.
Stage Temperature (C) Ramp (s) Hold (s) Ar ow (L min1)
Fig. 1. Inuence of the diluent nature and dilution ratio in the Sn analytical response in
the determination of (a) blood and (b) urine by ET AAS: ( ) 0.1% v/v Triton X-100;
( ) Water; ( ) 0.2% v/v HNO3. Other parameters as in Table 1.
possible limit of detection (LOD) in the original sample. In this way,
the background attenuation was well within the correction capacity of
the instrument, and no clogging of the autosampler capillary was ob-
served in the long run.
Regarding the urine sample, no sensitivity loss due to the presence
of the urine matrix was observed in dilution ratios of 1 + 1, 1 + 2 and
1 + 3; however, their use was not advisable since the background
attenuation (in peak height) was equal or higher than 1.0. Thus, the
dilution ratio of 1 + 4 with 0.1% v/v Triton X 100 was chosen in order
to achieve the lowest LOD in the original samples.
Modier
A palladium and magnesium nitrate solution was used as chemical
modier, a mixture originally proposed as universal modier [17].
Performances of three different ratios of Pd+ Mg(NO3)2 were evaluated
using 10 L of these solutions. Masses dispensed onto the platform
were 5 + 3; 10 + 6 and 15+ 10 g of Pd and Mg(NO3)2, respectively.
The most appropriate modier mass composition was assessed by
analyzing sensitivities of aqueous and matrix matched calibration
curves, with concentrations ranging from 2.5 to 50 g L 1 of Sn. A
signicantly lower slope was observed in the presence of lowest modi-
er masses. However, no signicant difference was observed for other
masses investigated. Thus, the intermediary modier mass was chosen.
Calibration studies

1a 110 10 10 250 In general, complex matrices like blood and urine have compo-
2b 130 20 30 250
nents which interfere in the atomization process, thus preventing
3 1300 10 20 250
4c 2200 0 5 0 the use of aqueous solutions for calibration. In such cases analytical
5 2600 1 3 250 curves prepared in the presence of the matrix may help to eliminate
a
Modier (blood) or modier + sample (urine) introduction.
or reduce this interference. Multiplicative matrix effects may be
b
Sample introduction (blood). quantied in terms of the sensitivity ratio (m0ac/m0mc), where m0ac
c
Reading. is the characteristic mass of the aqueous analytical curve and m0mc
126 S.V. De Azevedo et al. / Clinical Biochemistry 46 (2013) 123127

is the characteristic mass of the matrix matched calibration curve, Table 4

taken at similar conditions. Multiplicative matrix effects are not Determination of Sn in blood and urine reference materials by proposed procedures at
optimized conditions.
present if this ratio does not signicantly differ from the unity. In
the present case, sensitivity ratios were signicantly different from Sample Experimental value Reference value
1 (Table 3) for both matrices, conrming that external calibration (g L1) (g L1)

cannot be performed with aqueous calibration solutions. Real sam- TM144-1097 2.9 1.2 3.0 2
ples of blood and urine with low tin content were selected and Seronorm 54.3 1.6 54.6 2.7
spiked for use. The presence of the blood matrix increased the char- TM144-1097: Contox Trace Metal Serum Control (Kaulson Laboratories, USA);
acteristic mass regarding the aqueous medium in a more usual de- Seronorm: Seronorm Urine, lot 0511545 (Sero AS, Norway). Results are averages of
pressive interference. However, the slopes found in the presence of ve replicates and their standard deviation.

the urine matrix were higher than those in the aqueous solution,
showing an unusual situation. Urine is characterized by a high and
variable inorganic salt content such as phosphates, sulfates, chlo-
rides and calcium [4]. According to Borges [18], some of those com-
ponents can change the Sn behavior in the graphite furnace. The
author found that species such as calcium, chloride, and phosphate
interfered in the intensity and shape of the analytical signal when
determining tin in human milk. Phosphate, which is also present in
urine, increased the analytical signal, and it may be the cause of the in-
creased sensitivity in the presence of this matrix. However, analyte addi-
tion curves performed in different blood and urine samples at the
conditions displayed in Table 1 showed slopes (that is, characteristic
masses) with no statistically signicant differences (Table 3). Thus, exter-
nal calibration can be performed using matrix matched calibration solu-
tions, taking samples with the lowest possible Sn content for this
matching.
Analytical gures of merit
For linearity checking, twenty-one calibration solutions, with con-
centrations ranging from 2.5 to 500 g L 1 were used. The aqueous
calibration curve was linear up to 150 g L 1 while linearity was
observed up to 50 g L 1 of Sn in the presence of both matrices. Sen-
sitivity was assessed from the characteristic mass (m0) values, which
were calculated from the linear part of the calibration curves. Limits
of detection (3 , n = 10), calculated [19] for the determination of
tin in blood and urine were obtained from the reading of ten different
sample solutions of the same sample, with no detectable presence of
the metal, after suitable dilution with 0.1% v/v Triton X-100. Values
found were 0.80.1 and 2.70.5 g L1 for urine and blood, respective-
ly, in original samples. Limits of quantication [20] (10 , n=10) were
then 2.50.4 and 9.01.4 g L1 for urine and blood, respectively.
Table 3
Characteristic masses (SD) and sensitivity ratios in the determination of Sn in blood
and urine by GF AAS at optimized conditions.
Matrix Characteristic massa Sensitivity ratio
(g L1) (m0a/m0m)b
The accuracy was assessed by the analysis of standard reference
materials. External calibration was performed with matrix matched
calibration solutions using blood and urine samples with Sn content
below the LOD. Results are shown in Table 4 and there was no statis-
tically signicant difference between experimental and reference
values.
Tin concentrations in blood and urine of environmentally exposed
population
Blood and urine from eleven volunteers not occupationally exposed
to tin have been analyzed using the developed methodology (Table 5).
Levels of tin in blood ranged from 7.4 to 11.2 g L l while urine concen-
trations varied from 0.8 (LOD) to 2.2 g L l. According to the litera-
ture [3], the average concentration of tin in blood of nonexposed
subjects was approximately 140 g L 1, (a much larger range than
that observed in the present study) while the metal in urine ranged
from 0.5 to 5.0 g L 1, closer to the present data. However, other stud-
ies cited by ATSDR reported values ranging from 2 to 9 g L 1 for blood
and 1 to 20 g L 1 for urine [1]. Many of these studies are outdated, and
samples were subjected to processes of pre-treatment such as hot acid
digestion requiring a greater number of reagents and steps and they
may suffer from analytical uncertainties.
Conclusion
A simple, sensitive and direct procedure for the determination of
Sn in whole blood and urine was developed. Just a dilution step
with Triton X100 was necessary as sample pre-treatment as well as
the use of a Pd/Mg solution as modier. Multiplicative matrix effects
were observed, impairing the use of aqueous calibration solutions.
However, the similarity of the analyte addition curves slopes of a series
of samples supported the use of matrix matched calibration solutions,
and, in this way, following the optimized temperature program, an
agreement statistically signicant was observed between experimental
and reference values in the analysis of reference materials Also, limits of
detection obtained showed that the methodology can be used for the

Blood Average 86 5 0.85 0.03 Table 5

Blood 1 90 0.86 Levels of tin in blood and urine (g Ll) of a not occupationally exposed group.
Blood 2 82 0.89
Sample Blood Urine
Blood 3 91 0.82
Blood 4 79 0.85 01 8.0 0.4 0.8
Blood 5 83 0.83 02 7.4 0.4 0.8
Blood 6 90 0.87 03 11.2 0.4 0.8
Urine average 53 3 1.39 0.08 04 10.2 1.8 1.0 0.2
Urine 1 56 1.30 05 7.9 0.4 0.8
Urine 2 53 1.50 06 8.8 0.2 0.9 0.2
Urine 3 49 1.38 07 7.4 0.4 1.0 0.4
Urine 4 56 1.31 08 9.3 0.4 0.8
Urine 5 52 1.43 09 8.7 0.4 1.6 0.5
Urine 6 54 1.44 10 7.4 0.4 2.2 0.7
a 11 8.4 0.9 0.9 0.2
Results are averages of six replicates and their standard deviation.
b Mean 8.4 1.2 1.0 0.5
m0a/m0m: ratio between the characteristic mass obtained from the calibration curves
with aqueous (m0a) and matrix matched standards (m0m) under similar conditions. Results (g Ll) are averages of three replicates and their standard deviation.
 S.V. De Azevedo et al. / Clinical Biochemistry 46 (2013) 123127
determination of Sn in blood and urine of people not occupationally
exposed.
Acknowledgments
Authors dedicate this work to the Dr. Campos for his extreme im-
portant contribution to this paper.
Sayonara Vieira de Azevedo received a scholarship from CAPES.
Authors thank CNPq, the Instituto Nacional de Cincia e Tecnologia
Bioanaltica and FIOCRUZ for the nancial support.
References
[1] Agency for Toxic Substances and Disease Registry (ATSDR), toxicological prole
for tin. Atlanta, GA: U.S.: Department of Health and Human Services, Public Health
Service; 2005.
[2] Rdel H. Case study: bioavailability of tin and tin compounds. Ecotoxicol Environ
Saf 2003;56:180-9.
[3] Ostraknovitch EA, Cherian MG, Tin. In: Nordberg GF, Fowler BA, Nordberg M,
Friberg LT, editors. Handbook on the toxicology of metals. 3rd ed. California,
USA: Academic Press Elsevier; 2007. p. 839-59.
[4] Tsalev DL, Zaprianov ZK. Atomic absorption spectrometry in occupational and en-
vironmental health practice. 2nd ed. Florida, USA: CRC PRESS; 1985.
[5] S.V. Azevedo, Determinao dos nveis de estanho em uidos biolgicos de populao
exposta ambientalmente na Vila Massangana, RO, http://arca.icict.ocruz.br/
bitstream/icict/2426/1/ENSP_Disserta%C3%A7%C3%A3o_Azevedo_Sayonara_Vieira.pdf,
(Electronic Dissertation), July 2009, (acessed april 2011).
[6] A.M.A. Sakuma, Avaliao da exposio humana ao arsnio no Alto Vale do Ribeira,
Brasil, http://libdigi.unicamp.br/document/?code=vtls000341770, (Electronic Thesis),
Feb 2004, (acessed may 2011).
[7] Moreira FR, Mello MG, Campos RC. Different platform and tube geometries and
atomization temperatures in graphite furnace atomic absorption spectrometry:
127
cadmium determination in whole blood as a case study. Spectrochim Acta Part
B At Spectrosc 2007;62:1273-7.
[8] Moreira FR, Neves EB. Use of urine lead level as an exposure indicator and its re-
lationship to blood lead. Cad Saude Publica 2008;24:2151-9.
[9] Neves EB, Mendona-Junior N, Moreira MFR. Exposure assessment to metals in an
armament repair shop of a military organization. Cien Saude Colet 2009;14:
2269-80.
[10] Parsons PJ, Barbosa Jr F. Atomic spectrometry and trends in clinical laboratory
medicine. Spectrochim Acta Part B At Spectrosc 2007;62:9921003.
[11] Taylor A, Branch S, Day MP, Patriarca M, White M. Atomic spectrometry update.
Clinical and biological materials, foods and beverages. J Anal At Spectrom
2011;26:653-92.
[12] Gibson RS, Bailey KB, Romano ABA, Thomson CD. Plasma selenium concentrations
in pregnant women in two countries with contrasting soil selenium levels. J Trace
Elem Med Biol 2011;25:230-5.
[13] Baysal A, Akman S. Determination of lead in hair and its segmental analysis by solid
sampling electrothermal atomic absorption spectrometry. Spectrochim Acta Part B
At Spectrosc 2010;65:340-4.
[14] Gardiner WP. Statistical analysis methods for chemists. A software based approach.
Cambridge, UK: The Royal Society of Chemistry; 1997.
[15] Yin X, Schlemmer G, Welz B. Cadmium determination in biological materials
using graphite furnace atomic absorption spectrometry with palladium nitrate
ammonium nitrate modier. Anal Chem 1987;59:1462-6.
[16] Grinberg P, Campos RC. Iridium as permanent modier in the determination of
lead in whole blood and urine by electrothermal atomic absorption spectrometry.
Spectrochim Acta Part B At Spectrosc 2001;56:1831-43.
[17] Moreira FR, Maia CB, vila A. Titanium as a chemical modier for the determina-
tion of cobalt in marine sediments. Spectrochim Acta Part B At Spectrosc 2002;57:
2141-9.
[18] R.M. Borges, Determinao de Sn em leite humano por espectrometria de absoro
atmica no forno de grate, http://www.maxwell.lambda.ele.puc-rio.br/Busca_etds.
php?strSecao=resultado&nrSeq=14735@1, (Eletronic Dissetation), Dec 2008 (acessed
may 2011).
[19] IUPAC. Compendium of chemical terminology. [http://goldbook.iupac.org/PDF/
goldbook.pdf]Gold Book. Version 2.3.2; 2012 (accessed Set 2012).
[20] V. Thomsen, D. Schatzlein, D. Mercuro, Limits of Detection in Spectroscopy, http://
www.bioforensics.com/conference07/LOD/LODs.pdf, Dec 2010, (accessed Aug 2011).