SciELO - Scientific Electronic Library Online

 
vol.44 número3OBSERVATION OF STABLE 17 e IRON(III) CYCLOPENTADIENYL RADICAL COMPLEXESARCILLAS PILAREADAS II. HIDRODESULFURACION DE TIOFENO SOBRE CATALIZADORES Ni,W SOPORTADOS SOBRE ARCILLAS PILAREADAS índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Boletín de la Sociedad Chilena de Química

versión impresa ISSN 0366-1644

Bol. Soc. Chil. Quím. v.44 n.3 Concepción set. 1999

http://dx.doi.org/10.4067/S0366-16441999000300009 

OPTIMIZATION OF FLAME ATOMIC ABSORPTION SPECTROM-
ETRY WITH PRECONCENTRATION BY FLOW-INJECTION
ON-LINE SORBENT EXTRACTION OF CADMIUM AND
LEAD IN BIOLOGICAL MATERIALS

CARLOS G. BRUHN1*, CAROLINA VILCHES1 AND HERNAN J. CID2

1Depto. de Análisis Instrumental, Facultad de Farmacia, Universidad de Concepción, P.O.
Box 237, Concepción, Chile. Fax: 56-41-231903. E-mail cbruhn@udec.cl
2Centro EULA-CHILE, Universidad de Concepción, P.O. Box 156-C, Concepción, Chile.
(Received: May 4, 1999 - Accepted: June 18, 1999)

ABSTRACT

A flow-injection (FI) system with a minicolumn of bonded silica with octadecyl groups (C-18) to collect diethyldithiocarbamate complexes of Cd and Pb in reference solutions and in acid-digested hair and blood solutions was developed and evaluated by flame atomic absorption spectrometry (FAAS). The system was optimized by multivariate method, based on a factorial experimental design in two levels, selecting eight parameters that mostly affected the expected analytical signal. The detection limits (3xsBL/slope, 60 s preconcentration) were 0.7 and 5 µg/l for Cd and Pb, respectively and the sampling frequency was 40 samples/h. Effects of interfering ions are discussed. The methodology was validated by analysis of certified reference materials of hair and blood for Pb, and by recoveries of Cd and Pb spikes performed in hair and blood samples. Results for Pb agreed well with certified values and recoveries were satisfactory (103% in blood and 100% in hair). Also, the recovery of Cd in hair was fair (107%); however, in blood it was not quantitative (22%).

KEY WORDS: Cadmium, lead, preconcentration by sorption, flow injection flame atomic absorption spectrometry, hair and blood.

RESUMEN

Se desarrolló y evaluó por espectrometría de absorción atómica con llama (EAALL) un sistema de inyección en flujo (IF) con una minicolumna rellena con gel de sílice enlazada con grupos octadecilo (C-18) para retener complejos de dietilditiocarbamatos de Cd y Pb en soluciones de referencia y de digestos ácidos de cabello y sangre. El sistema fue optimizado por método multivariado en base a un diseño experimental factorial en dos niveles, seleccionándose los ocho parámetros que influyen más significativamente en la señal de analito esperada. Los límites de detección 3xsBL/pendiente, con 60 seg de preconcentración) fueron 0.7 y 5 µg/L para Cd y Pb, respectivamente y la frecuencia de muestreo fue 40 muestras/hr. Se discuten los efectos de iones interferentes. La metodología fue validada para Pb por análisis de materiales de referencia certificados de caello y sangre y por estudios de recuperación de Cd y Pb efectuados en muestras de cabello y sangre. Los resultados obtenidos para Pb son concordantes con las concentraciones certificadas y fueron satisfactorios en los estudios de recuperación (103% en sangre y 100% en cabello). También, la recuperación de Cd en cabello fue adecuada (107%); sin embargo, en sangre no fue cuantitativa (22%).

PALABRAS CLAVES: Cadmio, plomo, preconcentración por sorción, espectrometría de absorción atómica con inyección en flujo, cabello y sangre

INTRODUCTION

The determiantion of trace elements such as cadmium and lead in body fluids and tissues provides an important basis for the diagnosis of clinical disorders and intoxication, and for monitoring environmental pollution. The concentrations of Cd and Pb in biological samples are usually below the detection limits of flame atomic absorption spectrometry (FAAS) after their acid digestion and dilution into the final solution volume. Therefore, these detrminations are performed by electrothermal atomic absorption spectrometry (ETAAS) with a graphite furnace1-3).

Separation and preconcentration procedures, such as ion exchange, adsorption, solvent extraction and co-precipitation of both analytes, are often needed before the FAAS determination in these matrices. Conventional off-line procedures for the separation and preconcentration, although effective, are usually time consuming and tedious, require large sample and reagent amounts, and are prone to contamination and analyte loss. Flow injection (FI) on-line preconcentration coupled with spectroscopic techniques has been shown to be very effective in enhancing the sensitivity and selectivity of trace metals in samples with complex matrices4-6). In addition, increased sample frequency and throughput, higher efficiency and better reproducibility than batch or manual procedures are achieved. Several on-line flow injection (FI) ion-exchange or sorbent extraction approaches using microcolumns can achieve 20- to 30-fold signal enhancement in FAAS at sampling frequencies and sample consumption similar to conventional aspiration techniques7-9). In particular, packed column preconcentration by sorbent extraction is attracting much interest because it shows greater selectivity than ion-exchange systems; the latter exhibits small differences between stability constants for various metal ions or even between groups of metal ions10). This often causes problems because, if the matrix contains a large quantity of common cations (e.g., calcium, iron), the ability of a minicolumn to preconcentrate a desired trace element is impaired by its lack of exchange capacity, since the common elements present in the matrix are retained as well10). However, most of these approaches have been applied either to the determination of heavy metals in sea water or in environmental samples. Few reports dealt with applications of FI-FAAS in biological samples11-15) and the latest advances are compiled in a recent monograph16). In some instances peconcentration by co-precipitation-dissolution was used in acid digested blood and urine samples13,14), but this approach is somewhat limited by the hidrophobicity of the precipitate (e.g., during its collection in a knotted reactor), the relatively low phase transfer factor and large dead volume6).

Among packed column preconcentration techniques, sorbent extraction of hydrophobic diethyl ammonium N,N diethyldithiocarbamate (DDC) complexes of heavy metals on reversed-phase bonded silica sorbent with octadecyl functional groups (C-18), followed by elution of the metal chelates in methanol or ethanol shows a great potential for application in acid digested solutions of biological materials such as human hair and blood due to its better selectivity than ion-exchange extraction10,17).

In this work, a minicolumn packed with bonded silica sorbent with C-18, DDC as complexing reagent and methanol as eluent were used in conjunction with a flame atomic absorption spectrometer for the determination of Cd and Pb in acid-digested solutions of hair and blood. A simple time-based FI system with a peristaltic pump was assembled and coupled to the spectrometer, and the FI parameters were optimized by a multivariate method followed by univariate optimization of the more critical parameters.

EXPERIMENTAL
Apparatus

A Perkin-Elmer (PE) 3110 atomic absorption spectrophotometer equipped with deuterium arc background corrector and standard air-acetylene burner system, and an IBM Propinter III printer were used thrughout in FAAS. Hollow-cathode lamps for Cd and Pb, operated at the manufacturer's recommended conditions, were used at their primary resonance line (228.8 and 283.3 nm, respectively). Flame conditions slightly leaner than recommended by the manufacturer were chosen, to compensate for the effect of the organic solvent which was introduced during elution and acted as an additional fuel. Flow spoiler was used in the spray chamber for all measurements. The burner height was adjusted for optimum sensitivity, and the nebulizer uptake rate was regulated to provide optimum absorbance signal for conventional sample aspiration. Absorbance signals were measured in peak height and recorded. The spectrometer was operated with a time constant of 0.2 s.

The flow injection (FI) system (Figure 1) consisted of an Ismatec MV-MS/CA peristaltic pump (P1) with 3 channels to propel sample, chelating agent and wash solutions, respectively; a Masterflex Model 7554-60 peristaltic pump (P2) with one channel to propel the eluent; two Perspex manual injector commutators18) (C1, C2) to handle the solutions flow; pumping tubes, silicone rubber tube (0.8 mm i.d.) for propelling methanol, and Tygon conduit tubes to propel sample (1.85 mm i.d.), complexing agent (diethyldithiocarbamate, DDC) (0.38 mm i.d.) and wash (2% v/v HNO3) (1.85 mm i.d.) solutions, respectively; a Y-shaped Perspex confluence (CF) positioned upstream of C1, to fit polyethylene tubes of 0.8 mm i.d. and join the sample and DDC solution, a linear reactor made of polyethylene tube (R) [0.8 mm i.d., 12 cm (Cd) and 4 cm (Pb)] for mixing the sample solution with DDC. The preconcentration minicolumn (MC) was made from a section of 1 cm length of a white plastic Finnipipette tip (5 mm i.d. at top end and 4 mm i.d. at the opposite end) packed with silica bonded octadecyl, particles size 40-63 µm (Supelco) [60 mg (80 µl) for Cd and 37.5 mg (50 µl) for Pb]. The packing was kept in place by small 2.5 mm thick plugs of white plastic foam and disks of cellulose acetate (0.45 µm, Whatman) at both column ends. The MC was push-fit connected in the manifold of injector commutator C2 and PTFE tape was used to seal the coupling of MC to two conduit tubes (2.5 cm and 5.5 cm in length, respectively and 0.58 cm i.d.). The injector commutator C2 enabled the flow of the sample-DDC soluton mixture through MC in the preconcentration step, as well as the flow of the wash solution in the wash step, and the eluent in the elution step to transfer Cd and Pb sorbed complexes into the flame atomizer through the nebulizer. To minimize dispersion, the sample solution stream mixed with DDC flowed through the column in one direction, and the eluent flow was in reverse direction. The commutator C2 was connected to the nebulizer with a straight 4 cm length Tefzel tubing of 0.35 mm i.d.

The pH measurements were done in an Orion pHmeter, model EA 940 with a combined glass electrode.

A PE Model 1100B atomic absorption spectrometer equipped with deuterium arc background corrector, a PE Model HGA-700 graphite furnace, a PE Model AS-70 autosampler and hollow cathode lamps (PE) for Cd and Pb operated at the manufacturer's recommended conditions was used in ETAAS. The spectrometer was operated with a time constant of 0.02 s. Pyrolytic graphite coated tubes (Part N B3001254) and pyrolytic graphite platforms (Part N B3001256) were used throughout.

Microwave digestions were carried out in a Milestone MLS-1200 MEGA microwave system (Bergamo, Italy), with programmable power control, a MDR-300-S/10 TFM rotor, TFM vessels and a standard cooling system for MDR rotor, following manufacturer's recommended programs. Acid digested samples were subsequently evaporated in a MCR-6-E rotor using a 6 steps evaporation program (36 min) and an acid scrubber unit.

FIG. 1. Flow-injection system and sequence for on-line preconcentration and sorbent extraction of Cd and Pb (see text).

Reagents and materials

All reagents were at least analytica-reagent grade from Merck (Darmstadt, Germany), except for nitric and hydrochloric acid which were further purified in a quartz sub-boiling still (H. Kürner, Rosenheim, Germany) and stored in quartz bottles until use. Hydrogen peroxide (30% m/m), NH4H2PO4, (NH4)2HPO4, MgNO3 x 6H2O and NH3 (25% m/m) were Suprapur. Diethyl ammonium N,N diethyldithiocarbamate (DDC) 0.5 g/l solution (Merck, Darmstadt, Germany) prepared in buffer solution pH = 9.2 (0.01mol/l CH3COOH - 0.02 mol/l NH3) was used as the complexing agent. Ultrapure water (u.p.) (18 Mohm/cm, Millipore Corp., Bedford, Ma, USA) was used throughout this work. Reference solutions were prepared by stepwise dilution of 1000 mg/l Cd and Pb stock solutions immediately prior to use. Certified reference materials (CRM) of human hair GBW07601 (National Research Center for Certified Reference Materials (NRCCRM), China) and lyophilized blood (BCR N 196) were used to validate the analytical procedure for Pb determination in human hair and blood. The lyophilized blood was reconstituted by adding 5.0 ml of u.p. water and homogenized by continuous agitation in a mixing apparatus for 2 h. Extra-pure acetylene and argon (99.998%) were used as the flame and purge gases in FAAS and ETAAS, respectively.

The chemical modifier for the ETAAS measurement of Cd and Pb in human hair and blood was a solution containing 0.050 mol/l NH4H2PO4 + 0.2% v/v HNO3.

Flow system

The FI manifold for the three combinations of injector commutator positions is shown in Figure 1 including the preconcentration, wash and elution l prefill steps. Details for the duration and function of each sequence, as well as the flow conditions are given in Table I.

TABLE I. Sequence of operation for on-line preconcentration of Cd and Pba by FI-FASS

Step
Time
(s)
Flow rate
(ml/min)
Solution
Commutator
position (C1)
Injector
position (C2)
Action
 
 
 

1
60 (60)
4.0 (4.0)
Sample
MC
Fill
C1: On-line mix sample
 
 
 
 
 
 
+ DDC soluton
 
 
0.4 (0.4)
DDC (0.5 g/l
 
 
C2: Load sample mixed
 
 
 
at pH 9.2)
 
 
with DDC on MC
2
20 (20)
4.0 (4.0)
Wash
MC
Fill
C1: Wash reactor
 
 
 
(2% v/v HNO3)
 
 
C2: Wash MC
3
10 (10)
4.0 (4.0)
Sample
Waste
 
C1: Fill tubes sample
 
 
 
 
 
 
+ DDC solution
 
 
0.4 (0.4)
DDC (0.5 g/l
Waste
 
 
 
 
 
at pH 9.2)
 
 
 
 
 
5.0 (4.0)
Eluent
 
Inject
C2: Elute analyte
 
 
 
 
 
 
to FAAS

aValues are shown between parentheses.

Hair and blood study samples

The optimization studies of othermal program conditions were performed in human hair and blood samples obtained respectively, from two volunteers (20- and 24-years old, males and nonsmokers). The hair sample was collected in a polyethylene clean bag after a hair cut done with stainless steel scissors and was cleaned and prepared in powdered form following the procedure described elsewhere19).

Blood (20 ml) was collected by venipuncture with disposable syringe and stainless steel needle, divided in two portions of 10 ml and transferred into precleaned polypropylene tubes (Sarstedt). The blood was mixed slowly for few minutes with an anticoagulant (Na2EDTA, 18.6 mg) and was stored at 4°C until analysis.

Sample pretretment

Hair

Powdered hair (250 mg, reference materials or sample) was microwave digested with 2 ml HNO3 + 0.5 ml H2O2 according to manufacturer's suggested program (25 min) and evaporated in the microwave system ( 36 min), following conditions described elsewhere20). The dry residue was dissolved in 3 ml 1 mol/l HNO3 under heating in a double boiler for 5 min, transferred into a 25-ml volumetric flask and diluted with u.p. water. The solution was filtered through Whatman 41 filter paper and stored at 4°C until analysis.

Blood

Two 1000 µl aliquots of blood (reference material or sample) in separated vessels were microwave digested with 4 ml HNO3 and 1 ml H2O2, following the same program used for hair. After this digestion period, the rotor was removed and cooled; the pressure was carefully relieved, each vessel removed from the rotor and the cover thoroughly separated. Two new 1000 µl aliquots of the same blood were transferred into both vessels, added with 2 ml HNO3 and 1 ml H2O2, and the procedure was repeated. After this second digestion period, the solution digests were evaporated, each dry residue was dissolved likewise in 3 ml 1 mol/l HNO3 diluted into a unique 25-ml volumetric flask with u.p. water, filtered and stored at 4°C as described previously. Hence, this solution contains the acid digest of 4000 µl of blood.

Blank solutions of the acid digestion were prepared likewise throughout.

Procedures

Method development

The FI system was used initially with the conditions described in the literature for a similar approach tested in waters7,8). Preliminary calibration curves were obtained under these conditions, including an independent study of the reactor length and eluent flow rate keeping constant the other parameters. Thereafter, a multivariate approach was applied to optimize the following eight selected parameters for Cd and Pb: sample flow rate, DDC solution flow rate, type of eluent, eluent flow rate, MC packing amount, preconcentration time, reactor length, and transfer tube length. A factorial design was adopted to evaluate the relative weight of the selected parameters based on a multivariate analysis program21) arranged in three blocks with each parameter in study set at two levels (e.g., high and low). In each block beside the 2n experiments, was included a central experiment with the parameters in study set at the mean between the high and low levels, in order to assess the balance of the experimental design and whether the high and low levels selected were adequate or not. In each experment was obtained the calibration curve from the mean peak height absorbance of six measurements performed with each reference solution, and the criterion used to evaluate the relative weight of each experiment was the enrichment factor (E.F.). This figure was calculated as the ratio of the slopes of the calibration graphs obtained with and without preconcentration, and the effect of individual parameters within a block was assessed from the differences between the E.F.s obtained for all experimental pairs in which the specific parameter changed between the high and low level whereas the others were kept constant. These differences are averaged and the predominant level of this parameter is reflected by the sign and its relative weight by the mean difference. The first block comprised a matrix of 5 parametrs [minicolumn packing amount (A), eluent type (B), eluent flow rate (C), reactor length (D) and transference tube length (E)] in two levels corresponding to a 25 factorial design (32 experiments) to cover all possible combinations (Table II). The second one, consisted of a 23 factorial design including the three parameters not considered in the first block [sample flow rate (F), DDC flow rate (G) and preconcentration time (H)] (Table III). The third block was a new matrix with the five parameters (25 factorial design) showing the largest relative weight in two levels for the study analytes; however, the sample flow rate and preconcentration time were coupled and considered as one parameter to keep constant the sample volume. The two parameters showing the largest relative weight in the third block were studied univariately.

TABLE II. Levels of parameters considered in the 1st block of the multivariate analysisa for Cd and Pb by FI-FAAS.

Parameter  
Level

A
MC packing amount (mg)
45
75
B
Eluent flow rate (ml/min)
2
4
C
Eluent type
methanol
ethanol
D
Reactor length (cm)
4
12
E
Transference tube length (cm)
4
8

aExperimental conditions of the 3 additional parameters at constant level:
sample flow rate (3.0 ml/min); DDC flow rate (0.8 ml/min) and preconcentration time (30 s).

TABLE III. Levels of parameters considered in the 2nd block of the multivariate analysisa with variable sample volume for Cd and Pb by FI-FAAS.

Parameter 
Level  
F
Sample flow rate (ml/min
2
4
G
DDC flow rate (ml/min)
0.4
0.8
H
Preconcentration time (s)
30
60

aThe parameters studied in the 1st block were set at their optimum levels.

Analytical performance

The analytical capability of the optimized FI-FAAS was assessed for Cd and Pb by systematic evaluation of the linear working range; the reciprocal sensitivity, the repeatability and reproducibility (variation coefficient, % CV), the detection limit 3xsBL/slope), the column retention efficiency, the E.F. and the concentration efficiency (C.E. = E.F. · samples/min). The selectivity of othe system was studied for both analytes in two concentrations, based on the effects of concomitant cations and anions (Ca2+, Mg2+, Na+, K+, Fe3+, Zn2+, Cu2+, SO42-, Cl- and PO43-) at two concentrations, bracketing the levels present in acid digested solutions of hair and blood. Precision (as relative standard deviation, RSD%) and accuracy (as relative error) were calculated by the determination of Pb in CRMs of hair and blood, by recoveries of Cd and Pb from spiked human hair and blood, and by comparison with the results obtained by ETAAS with a graphite furnace and stabilized temperature platform furnace (STPF) conditions. Quantitation was by the calibration curve method, and in some samples by standard additions too.

RESULTS AND DISCUSSION

Diethyldithiocarbamate behaves as a bidentate univalent anionic ligand, forming extemely stable coordination-saturated chelates with several cations22) which can be quantitatively extracted into water-immiscible organic solvents in a single equilibration step. Instead of the traditional liquid-liquid extraction procedure used for preconcentration and matrix removal before FAAS determination, the organic solvent can be replaced by a suitable hydrophobic sorbent, which when packed in a minicolumn can be used in an appropriate flow injection system to provide on-line separation and extraction of trace metals as was proposed by Ruzicka et al.10). The metal is held as a chelate on the hydrophobic surface of the sorbent and the preconcentrated chelate can be eluted into a small volume of methanol or other water-miscible organic solvent. By eluting the analyte through a change of the polarity of the carrier stream, rather than by breaking the metal-chelate bond by means of strong acids, as was done in the liquid-liquid extraction approach, it is possible to preconcentrate metals from very acidic solutions using reagents with strong functional groups10). This chelating agent is reported to form chelates with metals at high acidities8) as used in this work.

Optimization of the on-line preconcentration system

The FI system was tested initially using the conditions described in the literature and preliminary calibration curves were obtained with linear working ranges of 0-100 µg/l and 0-200 µg/l for Cd and Pb, respectively. Thereafter, the influence of the acidity of the reference solutions (0.2 - 2.0 % v/v HNO3), the linear reactor length (5 - 50 cm), the eluent flow rate of methanol (0.5 - 3.5 ml/min) and the wash step were studied, without change in the other parameters. It was established that an increase in acidity between 0.2 and 2% v/v HNO3 renders a slight decrease in the analytical signal (15% in Cd and 10% in Pb), which is not critical to affect the results obtained in the acid digested solutions of hair and blood because their acidity is below 2% v/v. If a linear reactor is used after the confluence of the sample and DDC solution flows and its length is below 15 cm, the residence time of Cd and Pb complexes within this reactor will not decrease the E.F. The influence of the eluent flow rate affected positively the analytical signals of both analytes between 1.5 and 4.0 ml/min, and the wash step showed unnecessary with reference solutions but was included after the preconcentration step with sample digest solutions.

To optimize the FI parameters a multivariate approach was chosen, based on a factorial experimental design in two levels, divided in three blocks indicated in Table II-IV. The first block includes parameters A,B, C, D, and E at selected high, low and mean levels indicated in Table II, as well as the conditions of the other 3 parameters. The E.F.s obtained in the 34 experiments of the first block (including two central experiments, one with methanol and another with ethanol as eluent) ranged betwen 4.8 and 11.4, and for each parameter in study was obtained the respective table with differences between E.F.s. The mean differences including the sign are summarized for the first block in Table V. For Cd, the parameter with the largest relative weight was the eluent flow rate (C, high level) followed by the eluent type (B, methanol, low level), and with lower and similar weights, the minicolumn packing (A, high level) and the transference tube length (E, low level); the parameter with lowest relative weight was the reactor length (D). No significant synergistic effects were detected between these 5 parameters. The E.F.s obtained for Pb in the first block ranged between 4.7 and 11.6, and according to the mean differences in E.F.s two parameters showed the largest relative weights, the type of eluent (B, methanol, low level) and the eluent flow rate (C, high level), followed by the reactor length (D, low level), the transference tube length (E, low level), and with almost no effect, the minicolumn packing (B, low level). Also, no synergistic effects became evident for Pb between these 5 parameters.

TABLE IV. Levels of parameters considered in the 3rd block of the multivariate analysisa for Cd and Pb by FI-FAAS.

a) Cd 

Parameter
 
Level  

 
F
Sample flow rate (ml/min)
2
4
H
Preconcentration time (s)
60
30
A
MC packing amount (mg)
45
75
B
Eluent flow rate (ml/min)
2.5
5
E
Transference tube length (cm)
4
8
 
 
 
 

aExperimental conditions of the 3 additional parameters at constant level: DDC flow rate (0.4 ml/min); eluent (methanol) and reactor tube length (12 cm).

b) Pb

Parameter
 
Level  

 
F
Sample flow rate (ml/min)
2
4
H
Preconcentration time (s)
60
30
G
DDC flow rate (ml/min)
0.4
0.8
A
MC packing amount (mg)
45
75
B
Eluent flow rate (ml/min)
2.5
5

aExperimental conditions of the 3 additional parameters at constant level: transference tube length (4 cm); eluent (methanol) and reactor tube length (4 cm).

TABLE V. Optimum conditionsa obtained in the first block of the multivariate analysis for Cd and Pb by FI-FAAS.

Element
MC packing
amount
Eluent flow
rate
Eluent
type
Reactor
length
Transference
tube length
 

Cd
0.8
2.8
-1.6
0.1
-0.8
Pb
-0.1
1.9
-2
-1
-0.6

aLevel (+/-) and difference between E.F.s in parenthesis

The second block with three parameters was studied twice with the other five parameters set at the optimum level established in the 1st block. First, a preliminary study was performed in which the sample flow rate (F) was coupled with the preconcentration time (H) in order to keep constant the sample solution volume as one parameter, and the DDC flow rate (G) as second parameter; in this approach, the matrix (22) included five experments (with the central one). No tabe is given for these results because the E.F.s obtained were similar in all experiments for Cd (between 8.3 - 8.4) and with slight differences for Pb (betwen 9.0 - 9.4), showing that under constant sample volume, the coupled sample flow rate - preconcentration time and the DDC flow rate were not significant parameters in the preconcentration system for Cd, whereas for Pb only the DDC flow rate showed somewhat larger relative weight at low level. In the scond and definitive study, the former parameters were considered independently, and the matrix included 23 experiments and the central one. The E.F.s obtained ranged between 3.0 and 15.3 for Cd and between 2.8 and 16.2 for Pb. As expected, according to the differences in E.F.s for Cd and Pb (Table VI) the preconcentration time presented the largest relative weight (H, high level) followed by the sample flow rate (F, high level), and with significantly lower weight the DDC flow rate [G, low (Cd) and high (Pb) level]. In the case of Pb, the DDC flow rate showed contradictory behavior in both studies denoting low level under constant sample volume and high level with variable one. This difference will be elucidated in the third block. In both studies the conditions of the other five parameters were set at the optimum level established in the 1st block.

TABLE VI. Optimum conditionsa obtained in the 2nd block of othe multivariate analysis for Cd and Pb by FI-FAAS.

Element
Sample flow rate
DDC flow rate
Preconcentration time

Cd
4
-0.9
7.2
Pb
4.1
1.3
7.8

aLevel (+/-) and difference between E.F.s in parenthesis.

In the 3rd block the selected parameters were those presenting higher relative weight in the 1st and 2nd blocks. Four parameters were prevalent for both analytes: the eluent flow rate, the minicolumn packing and the sample flow rate coupled with the preconcentration time; the fifth parameter was the transference tube length for Cd and the DDC flow rate for Pb. Since two parameters were coupled, the matrix consisted of 17 experiments including the central one. Methanol was used as eluent and the other 2 parameters were kept at the optimum level established previously for Cd and Pb. The E.F.s obtained ranged between 6.2 and 12.1 for Cd and between 6.5 and 10.5 for Pb. Based on the differences in E.F.s (Table VII), for Cd the minicolumn packing presented the largest relative weight (B, low level) followed by

TABLE VII. Optimum conditionsa obtained in the 3rd block of the multivariate analysis for Cd and Pb by FI-FAAS.

Element
Sample flow rate
with constant
volume
MC packing
amount 
Eluent flow rate
Transference tube
length
 
 
 
 
 

Cd
-0.7
-3.1
+ 1.2
-0.3

aLevel (+/-) and difference between E.F.s in parenthesis

Element
Sample flow rate
with constant
volume
DDC flow rate
MC packing
amount 
Eluent flow rate
 
 
 
 
 
 

Pb
-0.3
-0.04
-2.2
+ 0.9

aLevel (+/-) and difference between E.F.s in parenthesis

the eluent flow rate (C, high level), and with lower weight the sample flow rate (F, low level) and the transference tube length (E, low level). For Pb (Table VII), the minicolumn packing presented the largest relative weight (B, low level) followed by the eluent flow rate (C, high level), and with lower weight the sample flow rate (F, low level); the DDC flow rate showed almost no effect (G). Nevertheless, for both analytes the sample flow rate was kept in the high level to increase the sample throughput, and the preconcentration time was increased to 60 s during sample analysis to lower the detection limits.

The two parameters with largest relative weight (the minicolumn packing and the eluent flow rate) were the same for Cd and Pb and were subjected to an independent study with the other six parameters set at their optimum level. The effect of the minicolumn packing was studied with 10 different packing amounts in the range between 30 mg (40 µl) - 150 mg (200 µl), and no clear trend was attained between E.F. and packing amount. The selected packing amount providing the largest E.F. was 60 mg (80 µl) for Cd and 37.5 mg (50 µl) for Pb. The effect of the eluent flow rate was studied for Cd and Pb with methanol between 3.0 and 6.0 ml/min and the result is shown in Figure 2, including both, the peak height absorbance and the E.F. For Cd, between 2.0 and 5.0 ml/min a direct relationship between absorbance and flow rate becomes evident and beyond this limit a plateau is achieved, probably because the analyte dispersion reaches a stationary state. For Pb, the linear relationship is shorter and above 4.0 ml/min reaches almost constant peak height absorbance. However, in the same flow range the E.F. shows only a slight variation for Cd, with a maximum (10.5) attained at 5.0 ml/min, whereas for Pb it is almost constant with a slight maximum value obtained at 4.0 ml/min 12).

FIG. 2. Effect of eluent flow rate on peak height absorbance (Cd: A; Pb: B) and enrichment factor (Cd: C; Pb: D) for on-line preconcentration of Cd and Pb with the FI manifold in Fig. 1. Preconcentration time 30 (s).

Analytical performance

The Cd2+ and Pb2+ reference solutions were run under optima chemical and flow conditions indicated in Table VIII, using the manifold shown in Figure 1. A study of the capacity of the MC prepared with 60 mg (Cd) and 37.5 mg (Pb) of sorbent was performed with working solutions of 50 µg/l Cd and 500 µg/l Pb. The recovery was quantitave for Cd and Pb up to a sample volume of 5.3 ml (preconcentration time of 80 s), corresponding to 267 and 2665 ng respectively, which is satisfactory for the amount of sorbent and preconcentration time used. With a sample flow rate of 4.0 ml/min and 30 s preconcentration, and with the aforementioned MCs, the retention efficiencies obtained for 75 µg/l Cd and 200 µg/l Pb were 91 and 89%, respectively.

TABLE VIII . Optimum conditions of FI-FAAS with on-line preconcentration for the determination of Cd and Pb.

Parameter
Level 
 
Cd
Pb 

MC packing amount (mg)
60
37.5
Sample flow rate (ml/min)
4
4
DDC flow rate (ml/min)
0.4
0.4
Eluent flow rate (ml/min)
5
4
Reactor length (cm)
12
4
Transference tube length (cm)
4
4
Eluent type
methanol
methanol
Preconcentration timea (s)
60 s
60 s

a30 s for reference solutions

The calibration curve equations and correlation coefficients for 30 (s) preconcentration corresponded to: A = 0.0057 ± 0.00011 x CCd + 0.0024 ± 0.0013; r = 0.9997 (Cd, five study points); A = 0.0005 ± 0.0001 x CPb + 0.0052 ± 0.0004; r = 0.9994 (Pb, six study points), and were considered adequate. In Table IX are indicated the analytical figures of merit obtained for Cd and Pb with the FI system using 60 (s) preconcentration in sample solution analysis. The variation coefficient (CV%) of the slope of the calibration curve is < 10% for both analytes, and the variability of the analytical signal is satisfactory for the concentration levels used with this preconcentration approach. The detection and quantification limits were calculated as 3xsBL and 10xsBL, respectively with 60 s preconcentration (n = 12 runs of 0.2% v/v HNO3 blanks; 4.0 ml sample volume). The method detection limit (n = 15 runs of acid digested blanks) was 0.07 and 0.45 µg/g in human hair and 4.3 and 28 µg/l in blood, and the quantification limit 0.23 and 1.5 µg/g in human hair and 14.4 and 94 µg/l in blood, for Cd and Pb, respectively. These results are satisfactory considering the relatively short preconcentration time (small sample volume) employed. The sample frequency was 40/h for both analytes.

TABLE IX. Analytical figures of merits obtained for Cd and Pb by FI-FAAS with on-line preconcentration by sorption on C-18.

Figure of Merit
Cd
Pb

Reciprocal sensitivity (µg/l)
0.6
8
Linear working range (µg/l)
5 - 75
25 - 500
Mean repeatability (n = 5) C.V. %
3.0
(1.8 - 5.1)a
6.7
(2.2 - 19.3)a
 
Mean reproducibility (3 days) C.V. %
6.8
7.5
Detection limit (µg/l)b
0.7
5
Quantification limit (µg/l)c
2.3
15
Enrichment factor
19
28
Concentration efficiency (EF/min)
12
18
Sample consumption (ml)
4
4
Sample frequency (1/h)
40
40
Retention efficiency %
91
89

aRange of C.V. % in the working curve
b3´SBL/slope for 60 (s) preconcentration
c10´SBL/slope for 60 (s) preconcentration

Interference study

The effects of potential interferents occurring in acid digested solutions of hair and blood samples on the determination of Cd and Pb was investigated using the optimized FI system and FAAS. Cations as nitrates and anions as ammonium solutions in two concentrations, one of them according to the highest levels present in acid digested solutions of the study matrices, were added individually to solutions containing 25 µg/l Cd or 200 µg/l Pb, and the resulting solutions were introduced in the FI system. The results are shown in Figures 3 and 4 for Cd and Pb, respectively. No effects became evident from NH4+ and NO3- on both analytes.

FIG. 3. Interference effects on 25 µg/l Cd analytical signal (expressed with respect to Cd peak height absorbance in 0.2% HNO3 normalized to 100%) by Mg2+, Fe3+, Cu2+, SO42- at 1000 and 10000 µg/l (A); and Na+, K+, Zn2+, Ca2+, Cl-, and PO43- at 10000 and 100000 µg/l (B) in on-line preconcentration by FI-FAAS.

FIG. 4. Interference effects on 200µg/l Pb analytical signal (expressed with respect to Pb peak height absorbance in 0.2% HNO3 and normalized to 100%) by Mg2+, Fe3+, Cu2+, and SO42- at 1000 and 10000 µg/l (A); and Na+, and K+, Zn2+, Ca2+, Cl-, and PO43- at 10000 and 100000 µg/l (B) in on-line preconcentration by FI-FAAS.

The main interference effects on Cd are due to Cu2+ (10000 µg/l), Zn2+ (100000µg/l) and Fe3+ (10000 µg/l), being particularly severe with Cu2+, and in all of them were dependent on the concomitant concentration. This intererence may be associated with formation and stability of DDC complexes of these cations8) because Cu(DDC)2 is more stable than Cd(DDC)2 and can be formed even in stronger acidic medium10). At 400-fold higher concentration the Cu complex could be sorbed on the MC packing reducing significantly the sorption sites for Cd complex, or could affect its kinetic of formation or stability. The interference of Fe3+ is important particularly in acid digested solutions of blood, because its concentration would be ca. 80000 µg/l, considering the acid digestion of 4 ml blood in a final 25 ml solution. Moderate depression on Cd signal was produced by Na+ and K+ at low level (10000 µg/l); however, this effect was meagre at higher concentrations of both cations. Phosphate produces only mild depressive effect in the high level studied (100000 µg/l of P as PO43-) and Mg2+, Ca2+, Cl- and SO42- did not affect Cd absorbance (< 5%),

The main effects on Pb are due to Cu2+ (10000 µg/l) as well, which interferes severely at the high level studied but not at the low level (1000 µg/l), probably by the same reasons given above. This effect is followed in strength by the effects due to Na+, K+ and Fe3+ (10000 µg/l). Sodium interferes by depression at both study levels, whereas K+ interferes likewise except that at low level there is no significant effect. Iron interferes on Pb only at high level and moderately, whereas Zn2+ (100000 µg/l) shows mild depressive effect only at high level. The cations Mg2+ and Ca2+, and anions Cl- and SO42- did not interfere on Pb. Phosphate produces only mild depressive effect, in particular at high level studied (100000 µg/l of P as PO43-).

Accuracy and precision

The analytical procedure was validated by determination of Pb in CRMs of human hair and blood as shown in Table X, and by recovery studies of Cd and Pb spikes added in samples of human hair and blood with intrinsic concentrations below the quantification limits of Cd and Pb. Also, for comparison the CRM of hair was analyzed by Pb by ETAAS with graphite furnace and STPF conditions, applying the standard additions method (Table X). The concentration of Cd in CRMs was below the level necessary to obtain dilute acid digested sample solutions above the limit of quantification for Cd; therefore, it was not possible to determine Cd content in these CRMs. The accuracy expressed as the mean relative per cent error (Erel%) was 2.3% in hair and 0.8% in blood for Pb by the calibration curve method, and the result obtained by ETAAS was slightly higher than the proposed approach. The imprecision expressed as mean relative standard deviation per cent (RSD%) was 2.2% in hair and 6.4% in blood by calibration curve. Both results are considered quite satisfactory because the number of independent subsamples analyzed was relatively low due to limited amount of CRMs available in this laboratory. The recovery study was performed by the calibration curve method on samples of hair and blood spiked with known amounts of Cd and Pb before acid digestion, at concentrations of 10 µg/l and 60 µg/l in the sample solutions. The mean recovery was satisfactory for Pb in both matrices (103% in blood and 100% in hair) and was confirmed by ETAAS (106% in blood and 92% in hair). Also, the mean recovery of Cd in hair was fair (107%) and was confirmed by ETAAS (95%). However, the mean recovery of Cd in blood digests was significantly lower (22%) by FI-FAAS whereas by ETAAS it was quantitative (101%). Possibly, this low recovery could be ascribed to the presence of high concentration of Fe3+ and also low concentration of Cu2+ in acid digested blood solutions as indicated earlier, which interfere significantly either in the complex formation or sorption of Cd(DDC)2 by the minicolumn packing (see interference study). Also, the total concentration of Cd present in acid digested blood wolution with spike (12 µg/l) is close to the limit of quantitation for Cd in the reference solution (ca. 14.4 µg/l considering a dilution factor of 6.25 applied to the sample). Hence, in solutions containing strongly interfering ions (i.e., Fe3+, Cu2+), Cd present at low concentration is not quantifiable by the calibration curve method, and standard additions should be used. However, this approach was not feasible here because it requires large sample solution volume not at hand.

TABLE X. Determination of Pb in Certified Reference Material (CRM) of hair and blood by FI-FAAS with on-line preconcentration and by ETAAS.

CRM
Quantitation
Pb(µg/l)  
 
 
Certified
FI-FAASa
ETAASa

BCR
Calibration curve
772 ± 22
778 ± 50(3)
N.D.b
(Bovine blood)
 
 
 
 
 
 
 
Pb (µg/g, dry weight)
 
GBW 07601
Calibration curve
8.8 ± 0.9
9.0 ± 0.2 (4)
N.D.b
(Human hair)
 
 
 
 
GBW 07601
Standard additions
8.8 ± 0.9
8.8 ± 0.3(3)
9.3 ± 0.7(5)
(Hyman hair)
 
 
 
 

aMean ± standard deviation; number of independent determinations in parenthesis.
bNot determined

Application in blood and hair

The analytical procedure was applied to the determination of Cd and Pb in three human hair samples, and Pb in one blood sample from male volunteers (a student, a laboratory technician (worker 1) and a gas station operator (worker 2)). The samples were analyzed too by ETAAS for comparison (Table XI). Both analytes were present in these samples in concentrations close to the quantitation limits and no

TABLE XI. Determination of Cd and Pb in human hair and Pb in blood by FI-FAAS with on-line preconcentration and by ETAAS.

a) Hair
 
 
 
 
Sample
Cd(µg/g, dry weight)a
 
 
 
FI-FAAS
ETAAS

 
Student
0.08 ± 0.02
0.06 ± 0.01
 
Worker 1
0.13 ± 0.05
0.09 ± 0.02
 
Pb (µg/g, dry weight)
 
Worker 1
2.3 ± 0.2
2.59 ± 0.17
 
Worker 2
1.0 ± 0.3
1.28 ± 0.30
 
 
 
 
b) Blood
 
 
 
 
Sample
Pb(µg/l)a 
 
 
FI-FAAS
ETAAS
 
Worker 2
37 ± 13
33.3 ± 3.0
aMean ± standard deviation; n = 3

statistically significant differences (95% confidence level) were established by Student t-test between the results obtained by both analytical methods. The proposed analytical methodology can be applied reliably to the determination of Cd and Pb in human hair with concentrations above 0.23 and 1.5 µg/g, respectively and for Pb in blood beyond 94 µg/l. However, the methodology is not recommended for the determination of Cd in blood because it requires the digestion of 10 ml sample aliquots and quantitation should be done by the standard additions method to compensate for severe matrix interferences, in particular due to Cu2+, Zn2+ and Fe3+ present in blood.

ACKNOWLEDGEMENTS

Finantial support of Fondo Nacional de Desarrollo Científico y Tecnológico (FONDECYT) (Research grant N 1960664), European Commission (Contract CI1-CT94-014) and Dirección de Investigación of the Universidad de Concepción (DIUC research grant N 95.71.01-4) are gratefully acknowledged.

_________________________

*To whom correspondence should be addressed.

REFERENCES

1. K.S. Subramanian. Prog. Anal. Spectrosc., 9, 237 (1986).         [ Links ]

2. D.T. Miller, D.C. Paschal, E.W. Gunter, P.E. Stroud and J. D'Angelo. Analyst, 112, 1701 (1987).         [ Links ]

3. J.A. Navarro, V.A. Granadillo, O.E. Parra and R. Romero. J. Anal. At. Spectrom., 4, 401 (1989).         [ Links ]

4. J. Ruzicka and E.H. Hansen. Flow Injection Analysis, 2nd Ed., Vol. 62 in Chemical Analysis, John Wiley & Sons, 1988.         [ Links ]

5. J.L. Burguera (Ed.), Flow Injection Atomic Spectroscopy, Marcel Dekker, New York, 1989.         [ Links ]

6. Z. Fang, Flow Injection Separation and Preconcentration, VCH, Weinheim, 1993.         [ Links ]

7. Z. Fang and B. Welz. J. Anal. At. Spectrom., 4, 543 (1989).         [ Links ]

8. Z. Fang, T. Guo and B. Welz. Talanta, 38, 613 (1991).         [ Links ]

9. S. Xu, L. Sun and Z. Fang. Anal. Chim. Acta, 245, 7 (1991).         [ Links ]

10. J. Ruzicka and A. Arndal. Anal. Chim. Acta, 216, 243 (1989).         [ Links ]

11. Z. Fang, S. Xu, L. Dong and W. Li. Talanta, 41, 2165 (1994).         [ Links ]

12. Z. Fang, M. Sperling and B. Welz. J. Anal. At. Spectrom., 6, 301 (1991).         [ Links ]

13. B. Welz, S. Xu and M. Sperling. Appl. Spectrosc., 45, 1433 (1991).         [ Links ]

14. M.F. Enríquez-Domínguez, M.C. Yebra-Biurrun and M.P. Bermejo-Barrera. Analyst, 123, 105 (1998).         [ Links ]

15. Y. Petit de Peña, M. Gallego and M. Valcárcel. J. Anal. At. Spectrom., 12, 453 (1997).         [ Links ]

16. Z. Fang, Flow Injection Atomic Absorption Spectrometry, Wiley, Chichester, 1995.         [ Links ]

17. B. Karlberg. Anal. Chim. Acta, 214, 29 (1988).         [ Links ]

18. F.J. Krug, F° Bergamin and E.A.G. Zagatto. Anal. Chim. Acta, 179, 103 (1986).         [ Links ]

19. C.G. Bruhn, A.A. Rodríguez, C. Barrios, V.H. Jaramillo, J. Becerra, U. González, N.T. Gras, O. Reyes and Seremi Salud. J. Anal. At. Spectrom., 9, 531 (1994).         [ Links ]

20. C.G. Bruhn, J.Y. Neira, G.D. Valenzuela and J. Nobrega. J. Anal. At. Spectrom., 13, 29 (1998).         [ Links ]

21. B. Barros, Y. Spacino, R. Edward "Planejamiento e optimizacao de experimentos", Universidade Estadual de Campinas, SP, Brasil (1996).

22. K.L. Cheng and K. Ueno, Handbook of Organic Analytical Reagents, CRC Press, Boca Raton, FI, 1982.         [ Links ]