Electrothermal atomic absorption spectrometric determination of trace lead, copper and manganese in aluminum and its alloys without preliminary separation

Analytica Chimica Acta, 144 (1982) 189--195 Elsevier Scientific Publishing Company, Amsterdam

-

Printed

in The Netherlands

ELECTROTHERMAL ATOMIC ABSORPTION SPECTROMETRIC DETERMINATION OF TRACE LEAD, COPPER AND MANGANESE IN ALUMINUM AND ITS ALLOYS WITHOUT PRELIMINARY

SEPARATION

KOJI

MATSUSAKI*

Department of Industrial Ube 75.5 (Japan) TAKASHI Department Tof:iwadai, YUROKU

Chemistry,

Yamagrtchi

University,

Tokiwadai,

YOSHINO of industrial Chemistry, Ube 7555 (Japan)

Faculty

of Engineering,

Yamagttchi

University,

YAMAMOTO

Department of Chemistry, Faculty Naka-ku, Hiroshima 730 (Japan) (Received

TechnicalCollege,

25th

of Science,

Hiroshima

University,

Higashisenda-machi,

June 1982)

SUMMARY Graphite-furnace atomic absorption spectrometry is used for the determination of :-- 0.001% of lead, copper and manganese in aluminum and its alloys. The samples are dissolved in hydrochloric acid and analyzed directly after addition of a slight excess of (NH,), EDTA over aluminum. Sample and standard solutions must contain equal amounts of EDT_%

Flame atomic absorption spectrometry (f.a.a.s.) is widely used for the determination of trace metals in aluminum and its alloys, but the procedure usually involves the chemical separation and/or pre-concentration of the analytes from the matrix [l-4]. Graphite-furnace atomization is becoming more popular because higher sensitivities are achieved than by flame atomization. Battistoni et al. [5] reported the determination of trace copper in aluminum alloys by graphite-furnace a.a.s. after liquid-liquid extraction. Such furnace methods would be very useful for determining trace elements in aluminum, provided that chemical separation and pre-concentration of the analytes is not required_ Aluminum and its alloys are readily dissolved in hydrochloric acid. However, interferences arising from chloride matrices are frequently encountered in graphite-furnace a.a.s. [6-g]_ In previous work [lo], the mechanism of halide interferences in the determination of trace elements was investigated; the evidence indicated that these effects arise initially from metal-halide complex formation in solution. On the basis of this mechanism, ammonium-EDT-4 was found to be the most suitable additive 0003-2670/82/0000-OOOO/SO2.75

sij 1982

Elsevier Scientific

Publishing

Company

190

for removal of the interferences, becau se the metal ions are masked by EDTA and the readily volatile ammonium halide is produced. In the present work, ammonium-EDTA is applied to the direct determination of traces of lead, copper and manganese in aluminum and its alloys by graphite-furnace a-as. after sample dissolution in hydrochloric acid. The detection limits were found to be similar to those obtained by flame a.a.s. with liquid-liquid extraction [l--4] _ ESPERIMENTAL

Apparatus and reagents A V&an-Techtron model

63 carbon rod atomizer was used in conjunction with a Varian-Techtron model 1200 atomic absorption spectrometer_ A tube type of graphite cell was used and the absorption was measured in a nitrogen atmosphere. The signal was recorded with a Hitachi model 056 recorder_ Hitachi lead, copper and manganese hollow-cathode lamps were used as the radiation sources. and a Varian-Techtron deuterium lamp was used for background correction_ The applied voltage between the atomizer terminals was measured with a digital voltmeter connected in parallel; the temperature of the center of the graphite tube was measured with a platinum/platinumrhodium thermocouple_ The sample was added from a 5~1 Excalibur Autopet fitted with disposable tips. All solutions were prepared from analytical-reagent grade chemicals and demineralized water, and stored in polyethylene bottles. The lOOO-pg ml-’ stock solutions for lead, copper and manganese were prepared by dissolving each pure metal in the minimum of nitric acid and diluting with 0.1 M nitric acid _

Proced we The sample (ca. 1 g) was weighed accurately and dissolved in the minimum of 6 M hydrochloric acid and several drops of 30% hydrogen peroxide by heating gently on a water bath_ After filtration, the solution was diluted to 100 ml with water. To aliquots of the sample solution containing 1.2-6.0 pg of lead, 0.5-2.5 pg of copper and 0.025-0.12 pg of manganese, 3.7 X lo-’ mol or slightly more (NH4)? EDTA was added per gram of metal sample. The pH was adjusted, if necessary, to a value in the range 2-10 with ammonia, and the volume was made up to 25 ml with water_ Calibration solutions were prepared similarly, using the appropriate stock solutions and an equal amount of (NH&EDTA. A 5~1 aliquot of the sample solution was deposited in the center of the graphite tube, dried, ashed and atomized with nitrogen gas flowing_ The instrumental conditions are summarized in Table 1. The graphite tube was fired at its maximum temperature for 4 s after each sample to clear any residue from the tube s-x-face. The absorption signals during the atomization step were recorded and peak heights were measured. The non-atomic absorp-

191 TABLE

1

Instrumental

conditions

Wavelength N, flow rate

Pb 283.3; Cu 324.6; 6.0 1 min-’

RIn 2i9.5

Dr>-ing

110°C

Ashing Atomization

620’ C (1 .‘i V) for 30 s 5-G V for 4 s

(0.65

V)

for

30

nm

s

tion of the sample was measured by use of the deuterium lamp at the same wavelength as the analyte under the same conditions and a suitable correction applied. The metal contents were determined from calibration graphs. The variation of furnace temperature with applied voltage was measured at the drying and ashing steps, in the absence of any salt. RESULTS

AND DISCUSSION

Inuestigation of analytical conditiom The concentration of the analyte element in the test solution, unless otherwise specified, was set to 0.25, 0.1 and 0.05 r_cgml-’ of lead, copper and manganese, respectively. The effect of ashin g temperature (30 s) on the atomic signal was investigated for the standard solutions in the presence of aluminum chloride. When EDTA present in the graphite tube was ashed completely, no effect of temperature was observed below 850°C (2.0 V). Above S5O”C, a depression of the lead signal was observed, probably because some lead volatilized during the ashing step; the recommended ashing condition is therefore 30 s at 620” C (1.7 V), 0.1 M EDTA being ashed completely_ The effects of drying and atomization conditions and of nitrogen flow rate were examined for each standard solution_ Throughout this study, the same instrumental conditions except for wavelength were applied to all analyte elements in order to simplify the procedure. Under the conditions shown in Table 1, the calibration graphs were linear over the range 0.050-25 pg Pb ml-‘, 0.02-0.1 c(g Cu ml-’ and 0.01-0.05 pg Mn ml-‘_ Prior to the examination of the effect of EDTA, the effects of aluminum chl.oride concentration on the atomic absorption signal of the analytes in the absence of EDTA were investigated_ As shown in Fig. 1, the signal for each analyte is suppressed by 240 pg ml-* aluminum chloride_ Addition of masking reagents such as EDTA, therefore, is necessary for the determination of traces of lead, copper and manganese in aluminum and its alloys after dissolution in hydrochloric acid. The effect of (NH,)2EDTA concentration added was investigated under Although DolinSek and Stupar [ 111 and the recommended conditions_ Ebert and Jungmann [12] reported that EDTA enhanced the lead atomic absorption, this was not confirmed in the present work. Curve [l] in Fig. 2 shows that EDTA above 0.01 M slightly suppresses the lead absorption in

192

10

I

I AI

3b

Fig_ 1. Effects the absorbance (0.05

gg ml-‘)

ICOO

100 j

( w

ml“

10-c

10000

of concentration of aluminum of: (0) lead (0.25 pg ml-*); relative

to that when

10-3 C EDT.4

1

I

I o-2

IO“

( J.&l ml-’

1

(as chloride) in the absence of EDTA on (.a) copper (0.1 rg ml-‘); (3) manganese

the aluminum

is absent.

Fig. 2. Effects of (NH,),EDTA concentraticn on the absorbance of lead (0.25 pg ml-‘) relative to that when EDT-A- and aluminum chloride are absent. AU, concentration (1) nil; (3) 0.01

$1; (3) 0.1 %I.

the absence of aluminum; curves (2) and (3) show that EDTA decreases the depression of lead absorption by aluminum chloride and when equimolar amounts of EDT-4 and aluminum are present, the lead signals are the same as in t.he absence

of

chloride.

Thus

the

interference

of

aluminum

chloride

is

completely removed by addition of an equimolar concentration of EDTA. However, >O.Ol M EDT&I slightly suppresses the lead signal. For copper and manganese, the suppressing effect of large amounts of EDTA was less than for lead, and the interferences by aluminum chloride were also removed completely by addition of an equimolar concentration of EDTA. In order to standardize the amount of EDT-4 added, all samples were regarded as pure aluminum, and the amount of EDTA added was calculated on that basis, i.e., 3 X 10” mol (or slightly more) per gram of sample. For the calibration solutions equal concentrations of EDTA were added to compensate for any interfering effect of EDTA. The maximum concentration of EDTA used in the sample solution was 0.1 M, because at higher concentrations, the reproducibility of the analytical signal was invariably poor. The effect of the pH of the test solution on the atomic absorption signal was investigated for a synthetic sample solution containing aluminum chloride (1 mg Al ml-‘). The pH was adjusted with hydrochloric acid or ammonia. Between pH 2 and 10, no variation in the absorbance of the analytes was observed. Below pH 2, the EDTA precipitated. Therefore, the pH of the solution can be adjusted to any value between 2 and 10 with ammonia; such a wide pH range is very advantageous_ Effect

of other

metals;

detecti-on

limits and recoveries

The conditions described above were used in the investigation of the effects of major metals in aluminum alloys on the recovery of the analytes. The absorbance of the analyte was compared with two sets of solution, ofie

193 TABLE

2

Effects of other metals (present as chloride) (0.1 gg ml-‘) and manganese (0.04 wg ml-‘) pug Al ml-l) Metal tested

Concn.

m

100 100 100 10 10 10 10 10

Zn Cu(II) Fe(II1) Ni

Cr(II1) Sn(I1) hln(II) Pil

(gg ml-’ )

10

Recovery

on the recovery in the presence

of lead (0.1 pg ml-‘), copper of aluminum chloride (1000

(%)

Lead

Copper

99.3 103.2 100.1 98.6 98.9 99.1 98.4 99.9 -

99.2 97.2 96.7 96.3 99.4 100.3 103.8 97.8

Mangan.wz 100.3 101.8 103.6 103.8 10-1.0 101.6 98.3 99.2

having only analyte and EDTA, and the other containing analyte, aluminum chloride (1 mg Al ml-‘), EDTA and another metal chloride. For magnesium, copper and zinc, which usually form alloys with aluminum, the amount added was 10% that of aluminum, i.e., 100 pg ml-‘, and for other metals, 10 pg ml-‘. The amount of added EDTA, therefore, was adjusted to be equivalent to 1100 or 1010 r_cg Al ml-‘. The results are given in Table 2. They show no significant effects of other metals on the determination of lead, copper and manganese by the proposed procedure. The detection limits for each analyte in aluminum were established by use of a synthetic aluminum chloride solution. The detection limits (the concentration of analyte giving an analytical signal twice that of the background signal at the analytical wavelength) were 0.001% for lead and copper and 0.0005% for manganese in aluminum. These values are comparable with those obtained by flame a.a.s. using liquid-liquid estraction [l-4] _ The proposed procedure was applied to the determination of traces (0.001-0.4%) of lead, copper and manganese in several standard aluminum and alloy samples. The results are presented in Table 3. They show good reproducibility, and good agreement with certificate values.

Conc!usions Ammonium-EDTX is shown to be a very suitable additive for the direct determination of traces of lead, copper and manganese in aluminum and its alloys by graphite furnace a.a.s. As the specimen can be dissolved in hydrochloric acid and no chemical separation is required, the analysis is rapid and the risk of contamination of the sample is considerably reduced.

0.24 0.039 0.09 0.001

0,24 f 0.01 (6) 0.037 ?: 0.001(5) 0,088 i 0,003 (7) 0,0010 + 0.0001 (5)

0.0017 + 0.0001(7)

BAM 301

0.36

0,001 0.001

0,OOll + 0.0001 (10) 0.0010 i 0.0001 (10)

0.084

0.15

0.0014 0.21

CcrtiCied vflluc

0.36 ? 0.02 (5)

0.085 ?: 0,002 (6)

O,l(i + 0.01 (G)

0.0014 i 0.0001 (5) 0.23 2 0.01 (G)

FountIn

nMcan and standard deviation with the number of dctcrminations in pnrcnthcscs.

0.0018

0,0014 0.022

nnd its nlloys

Mnn~ntwse (%)

in aluminum

Certified value

0.0014 I 0.0002 (5) 0.021 ?; 0,001 (7)

FOUlld”

Copper (%)

of truces of lend, copper IIII~ mnngnnesc

ALCOA WA.1199.L ALCOA KB.514.C (5% MB-Al nlloy) ALCOA SS.360.AB (10% Si-AI alloy) BCS 26211 (10% MB-AI alloy) BCS 263/l (5% Mg-Al alloy) BCS 196 fi

Sample

Determination

TABLE 3

0,054 !: 0,002 (5)

0.0011 ? 0.0002 (6) 0,062 ?: 0,002 (5)

FOUllCIR

Lead (%)

0,053

0.0010 0.062

VillUC

Certified

w g

195 REFERENCES 1 2 3 4 5 6

R. C. Calkins, Appl. Spectrosc., 20 (1966) 146. A. G6mez Coedo and &I. T. Dorado, Rev. hletal. (IXIacirid). 11 (1955) 61. I. Atsuya, Bunseki Kagaku, 15 (1966) 247. T. Kono, Bunseki Kagaku, 26 (1975) 162. P. Battistoni, P. Bruni, L. Cardeilini, G. Fava and G. Gobbi, Talantn, 2’7 (19SO) 623. J. Smeyers-Verbeke, Y. Michotte, P. Van der Winkel and D. L. Massart, Anal. Chcm., 48 (1976) 125. ‘7 E. J. Czobik and J. P. Matousek, Anal. Chem., 30 (1978) 2. S D. J. Churella and T. R. Copeland, Anal. Chem., 50 (1979) 309. 9 C. W. Fuller, Electrothermal Atomization for Atomic Absorption Spectrometry, The Chemical Society, London, 1977, p_ 62. 10 K. Matsusaki, Anal. Chim. 11 F. DolinSek and J. gtupar, 12 J. Ebert and H. Jungmann,

Acta, 111 (19S2) 233. Analyst, 9s (19’73) S-ll. Fresenius Z. Anal. Chem.,

272

(197-l)

2S7.