Daniela
Borosova
a,
Alena
Manova
b,
Jan
Mocak
bc and
Ernest
Beinrohr
*b
aState Institute of Public Health, Cesta k nemocnici 1, 97 556, Banská Bystrica, Slovak Republic
bSlovak University of Technology, Institute of Analytical Chemistry, Radlinského 9, 812 37, Bratislava, Slovak Republic. E-mail: ernest.beinrohr@stuba.sk
cUniversity of SS. Cyril and Methodius in Trnava, Faculty of Natural Sciences, Námestie Jozefa Herdu 2, 917 01, Trnava, Slovak Republic
First published on 13th October 2010
Nickel in hair samples was determined by GF AAS and stripping chronopotentiometry after a dry ashing mineralisation with a mixture of nitrous oxides, oxygen and ozone. The use of common matrix modifiers brought no significant improvement in the GF AAS measurement. The detection limit, trueness and precision were found to be 0.15 mg kg−1 of the hair sample, 4.3%, and 4.1%, respectively. The linear range was found to be 0.44–3.3 mg kg−1. The stripping chronopotentiometric determination is based on the deposition of nickel ions from an ammonia buffer solution in a porous glassy carbon electrode followed by its stripping into an acidic solution by constant current. Copper interfered and was removed from the acidified sample solution by means of a minicolumn packed with pearl cellulose chemically modified with 8-hydroxychinolinol. Copper was trapped completely whereas nickel passed the column. The detection limit, trueness and precision were found to be 0.032 mg kg−1 in the hair sample, 5.8%, and 6.1%, respectively. The linear range was found to be 0.1–20 mg kg−1. Good agreement between the two methods was found for most hair samples.
Determination of trace elements in hair is important for monitoring the influence of the environment on humans, for epidemiological studies and for evaluation of working place exposures. This assumption is confirmed by the fact that the average trace element content in hair corresponds to their content in blood.1
One of the species of interest in hair samples is nickel which is commonly considered toxic. It enters the environment predominantly as a consequence of its industrial use such as production and discharge of accumulators, applications in metallurgy and electroplating. Smoking is an other common source of nickel contamination. Nickel enters the living organisms from air, digestive tract and epidermis. In hair, the nickel content is in the range of 0.6–6.5 mg kg−1.2 Enhanced Ni content in hair of children indicated a long-term exposition to contaminated environment caused by nearby industrial plant.3
Nickel in hair samples is usually determined by nuclear activation analysis,1 ICP OES,4 X-ray fluorescence5 and PIXE spectroscopy.6 Owing to low detection limits and high selectivity, nickel can advantageously be determined by electrothermal AAS by making use of slurry sampling technique.7
Electrochemical methods can also be used for nickel determination, e.g. adsorptive stripping voltammetry on mercury,8 silver amalgam,9 bismuth film10 and antimony film electrodes.11 The electrochemical methods are based on collection of the Ni(II) dimethylglyoxime complex on the electrode surface followed by cathodic stripping of the deposit. The method is sensitive and selective enough but requires the removal of complex forming substances which interfere.
Nickel can also be determined by anodic stripping voltammetry on solid electrodes such as graphite.12 Here, nickel is deposited as elemental nickel, then is stripped as Ni(II). Owing to the high background due to the electrode material, the detection limits were much worse than in the case of the adsorptive method.
Deposition of metals in porous carbon-based electrodes has been used for their electrochemical preconcentration in AAS13 and MIP OES14 methods. It was found that Ni can be deposited with high efficiency on porous carbon electrode made of crushed glassy carbon particles which have significantly lower background currents than graphite electrodes. Hence, by making use of this electrode and the stripping chronopotentiometric measurement detection limits low enough for hair samples can be expected.
The objective of this paper is to demonstrate the capacities of graphite furnace AAS and stripping chronopotentiometry with the above electrode material in the determination of low nickel concentrations in hair samples.
Step no | Temperature °C | Time/s | Gas flow/L min−1 | Gas type | Read command |
---|---|---|---|---|---|
a HCl lamp: Ni Lumina (Perkin Elmer), wavelength: 232.0 nm, current: 4 mA, slit: 0.2 nm, sample volume: 20 µl, pyrolytically coated tube with platform, signal evaluation: peak area measurement. | |||||
1 | 100 | 8.0 | 3.0 | Normal | No |
2 | 140 | 8.0 | 3.0 | Normal | No |
3 | 700 | 17.0 | 3.0 | Normal | No |
4 | 2300 | 1.0 | 0 | Normal | Yes |
6 | 2600 | 2.0 | 3.0 | Normal | No |
Flow-through chronopotentiometric measurements were carried out by an electrochemical analyser EcaFlow model GLP 150 (Istran, Ltd., Bratislava, Slovakia) equipped with two solenoid inert valves, a peristaltic pump and a microprocessor controlled potentiostat/galvanostat. The block diagram of the system was reported elsewhere.15 The signals were recorded and evaluated by the memory-mapping technique.16The measurement consisted of two main steps: (i) the background signal was measured first by means of a blank sample and (ii) the sample or standard solution was measured: nickel ions were collected from the ammonium sample solution on the electrode as elemental nickel. The flow was switched to the acidic carrier electrolyte to remove the sample matrix from the cell, the flow was stopped and following a quiescence period the deposit was stripped by constant current and the stripping chronopotentiogram was recorded. Finally the cell was rinsed with the carrier electrolyte. The background signal was then subtracted from the signal of the standard or sample yielding a true background corrected net signal.
A compact flow-through electrochemical cell of type 353 with Pt auxiliary, Ag/AgCl reference and E53C microporous glassy carbon working electrodes was used (Istran, Ltd., Bratislava, Slovakia). The effective diameter and length of the electrode were approximately 5 mm and 3 mm, respectively. The electrode surface and void volume were about 25 cm2 and 20 µl, respectively. The operation parameters are listed in Table 2.
Parameter | Value |
---|---|
Deposition current | −4 mA |
Quiescence potential | −1000 mV |
Quiescence time | 10 s |
Terminal potential | 100 mV |
Regeneration potential | 200 mV |
Standby potential | 200 mV |
Stripping current | −50 µA |
Sample volume | 2 ml |
Blank volume | 2 ml |
Rinsing volume | 3 ml |
Flow rate | 3 ml min−1 |
The hair samples were mineralised in a dry digester system APION (TESSEK Praha, Czech Republic) and a microwave assisted wet digester Milestone, MLS 1200 MEGA.
For Cu removal a minicolumn of inner diameter and length of 5 mm and 20 mm, respectively was used. The column was packed with Iontosorb OXIN 100 chelating sorbent with chemically bound 8-hydroxychinolinol groups on pearl cellulose (IONTOSORB, Usti nad Labem, Czech Republic). The column was rinsed with 50 ml of 1 mol dm−3 NaOH, then with water and 1 mol dm−3 HCl until the eluted solution became colourless. Finally, the column was rinsed with water until a neutral pH. After use, the column was regenerated with 1 mol dm−3 HCl and rinsed with water.
Magnesium nitrate solution: 10 g dm−3 Mg(NO3)2 in 15% (v/v) HNO3 (MERCK, No. 5813).
Rhodium matrix modifier, 2.5 mg dm−3 Rh, prepared from Rhodium(III)-chloride-Trihydrate (Degussa).
Carrier electrolyte for chronopotentiometry: 0.1 mol dm−3 NaCl and 0.002 mol dm−3 HCl.
Electrolyte for sample adjustment: 0.1 mol dm−3 NH3 and 0.1 mol dm−3 NH4Cl.
The Ni calibration solutions were prepared from a Ni certified reference material with 1.000 g dm−3 Ni (Slovak Metrological Institute, Bratislava, Slovak Republic).
Two methods were used for sample digestion and preparation:
GF AAS measurements: the formed ash was leached with 10 ml of 1 mol dm−3 HNO3 and the solution was analysed.
Chronopotentiometric measurements: the ash was leached with 5 ml of 0.1 mol dm−3 HCl. This solution was let pass the minicolumn at a flow rate of 1–3 ml min−1. The column was rinsed with 5 ml of 0.1 mol dm−3 HCl giving a total volume of 10 ml of the treated sample solution. On adding 2 ml of 1 mol dm−3 NH4OH the resulting solution was analysed.
The pyrolysis temperatures are influenced in a similar way for all three cases (Fig. 1). The highest sensitivity for the 1 mol dm−3 HNO3 medium is achieved at temperatures of 2300–2400 °C. In the case of the Mg(NO3)2 modifier (10 µl) this temperature was about 200 °C lower and the Rh modifier (10 µl) shifted the atomisation temperature to unacceptably high values. The background signal was significantly influenced by the matrix modifiers (Fig. 2). Without a modifier, the background signal was at a level of 0.02–0.04 absorption units virtually independent on pyrolysis and atomisation temperatures. With the Rh modifier, the background signal decreases with pyrolysis temperatures and is virtually independent on the atomisation temperatures. The Mg(NO3)2 modifier, however, gives rise to a steadily increasing background signal both when increasing the pyrolysis and atomisation temperatures. In fact, this modifier stabilises the sample matrix and is therefore less suitable for this application. These observations confirmed some earlier conclusions19 that there is virtually no need to use a matrix modifier in Ni determination in digester hair samples. Hence, no matrix modifier, pyrolysis and atomisation temperatures of 700 °C and 2300 °C (Fig. 1), respectively, have been used in further measurements.
![]() | ||
Fig. 1 Influence of the pyrolysis and atomisation temperatures on the net signal of Ni. |
![]() | ||
Fig. 2 Influence of the pyrolysis and atomisation temperatures on the background signal. |
The limit of detection obtained by the IUPAC recommended method20 was found to be 0.02 mg dm−3.
Ni2+ = Ni0 − 2e− | (1) |
The deposition can be done either by means of a constant current or constant potential. The former procedure ensured better reproducibility and higher electrochemical yields, so it was used in further experiments.
The measurement is influenced by some species, especially those forming stable complexes with nickel ions and metal ions giving stripping signals near to that of Ni such as copper. Due to the applied digestion method, no complex forming species can be expected in the ashed samples. Cadmium and copper could also give signals near that of nickel, but the cadmium content was negligible in hair samples. However, hair samples inherently contain significant amounts of copper, which adversely influences the accuracy of nickel measurement owing to some overlapping of their stripping peaks. To minimise this interference the carrier electrolyte, Ni is being stripped to, was optimised to achieve the best resolution of the Ni and Cu stripping peaks. The best electrolyte found was that containing 0.1 mol dm−3 NaCl, 0.002 mol dm−3 HCl. Unfortunately, the adverse influence of copper still remained significant (Table 3). Copper concentrations in the analysed sample solutions depress the Ni signal in 10% already at a 1:
1 ratio of Cu
:
Ni, at a ratio of 2
:
1 the Ni signal decreases to about 18% of the original value.
Cu added/µg dm−3 | Ni found/µg dm−3 |
---|---|
a Ni added 50 µg dm−3. | |
10 | 51.7 |
20 | 52.2 |
25 | 48.8 |
50 | 45.5 |
80 | 31.8 |
100 | 17.5 |
200 | 4.3 |
1000 | <0.1 |
Hence, the interfering copper ions had to be removed from the hair sample solutions to gain accurate results. A chelating sorbent Iontosorb Oxin 100 with chemically bound 8-hydroxychinolinol groups on pearl cellulose was tested for this purpose. The sorption of metal ions on such sorbents depends significantly on pH21 which could be used for their separation. It was found that in 0.1 mol dm−3 HCl the sorption of Ni is negligible whereas Cu ions are completely trapped in the sorbent. The procedure was then modified so that the ash obtained from the digestion was eluted with 0.1 mol dm−3 HCl, the resulting solution was let to pass the minicolumn and the column was rinsed with 0.1 mol dm−3 HCl to obtain 10 ml of eluate. Finally, ammonia solution was added to the eluate to obtain an optimum ammonia buffer solution and the resulting solution was analysed (Fig. 3).
![]() | ||
Fig. 3 Chronopotentiometric signals of (a) Ni standard (100 µg dm−3), (b) hair sample, and (c) the same hair sample after removal of Cu by means of chelating sorbent Iontosorb Oxin 100. Ni found 0.438 µg g−1 (Ni found by GF AAS 0.47 µg g−1). The integration boarders set on the stripping peak of the hair sample. |
Parameter | GF AAS | Chronopotentiometry |
---|---|---|
a Obtained by means of a hair sample spiked with Ni. Concentration of Ni in the measured solution was 15 µg dm−3. | ||
Limit of detection/mg kg−1 | 0.15 | 0.032 |
Limit of quantification/mg kg−1 | 0.44 | 0.10 |
Linear range/mg kg−1 | 0.44–3.3 | 0.10–20 |
Truenessa (%) | 4.3 | 5.8 |
Precisiona (%) | 4.1 | 6.1 |
Sample | Ni found/mg kg−1 | |
---|---|---|
Chronopotentiometry | GF AAS | |
1141 | 0.434 ± 0.040 | <0.44 |
7740 | 0.939 ± 0.031 | 1.003 ± 0.020 |
1132 | 0.352 ± 0.004 | <0.44 |
7756 | 1.78 ± 0.13 | 1.500 ± 0.030 |
13![]() |
0.403 ± 0.030 | 0.542 ± 0.011 |
13![]() |
1.56 ± 0.12 | 1.650 ± 0.034 |
13![]() |
0.633 ± 0.047 | 0.614 ± 0.012 |
8539 | <0.10 | <0.44 |
This journal is © The Royal Society of Chemistry 2010 |