Spectrochimica Acta Part B 58 (2003) 1715–1721

Analytical note

Slurry sampling for the determination of silver and gold in soils and sediments using electrothermal atomic absorption spectrometry Ignacio Lopez-Gar

ćıa,

Ńatalia Campillo, Isabel Arnau-Jerez, Manuel Hernandez-Cor

´ doba*

Department of Analytical Chemistry, Faculty of Chemistry, University of Murcia, Campus of Espinardo, Murcia E-30071, Spain Received 23 April 2003; accepted 17 June 2003

Abstract

Procedures for determining silver and gold in soil and sediment samples using electrothermal atomic absorption spectrometry are discussed. Slurries are prepared from the samples and fast-program methodology is used for the heating cycle. To determine silver, suspensions are prepared by weighing 5–200 mg of sample and adding 25 ml of a solution containing 3% vyvconcentrated nitric acid, 50% v yvconcentrated hydrogen peroxide and 25% v

concentrated hydrofluoric acid. For gold determination, slurries are prepared by weighing up to 200 mg of sample and then adding 1 ml of concentrated hydrofluoric acid. For both cases, the slurries are submitted to a 10 min mild heating stage. After cooling to ambient temperature, the suspensions are introduced directly into the atomizer. No modifier other than hydrofluoric acid and hydrogen peroxide (for silver determination) are required and calibration is carried out using aqueous standards. The detection limits are 0.02 and 0.01 mg kgy1 for silver and gold, respectively.

Keywords: Electrothermal atomic absorption spectrometry; Slurry sampling; Silver; Gold; Soils; Sediments 1. Introduction

been reported for silver determination in human

scalp hair w2x and another related to several bio-The use of slurries in electrothermal atomic

logical standard reference materials w3x. A method absorption spectrometry (ETAAS) is a well-estab-for determining gold in high-purity silver by sus-

lished methodology whose advantages have been

pending undissolved particulates of sample in

recently reviewed w1x. However, relatively few nitric acid has also been proposed w4x. In other papers have been published on the determination

cases, the dissolved metal is concentrated using

of silver and gold by means of this approach. To

different materials and then slurried and analyzed

the best of our knowledge, only one procedure has

by ETAAS w5x.

The advantages of slurrying soil and sediments

*Corresponding author. Tel.: q34-968-367406; fax: q34-samples in a medium containing hydrofluoric acid

968-364148.

E-mail address: hcordoba@um.es (M. Hernandez-

have been widely demonstrated w6–8x. This Cor

´ doba).

approach was first used by Bendicho and de Loos-

doi:10.1016/S0584-8547(03)00135-6

1716

I. Lopez-Gar

cıa

ét al. / Spectrochimica Acta Part B 58 (2003) 1715–1721

Vollebregt in their studies on glass analysis w9,10x

land) and diluted as necessary to obtain working

and later proved successful in the analysis of

standards. High quality concentrated (40% myv)

several samples containing silicaysilicate w11x. As hydrofluoric acid, concentrated (40% myv) hydro-far as we know, there are no previous reports

gen peroxide and concentrated (65% myv) nitric

regarding silver or gold determination in soils and

acid were also obtained from Fluka.

sediments using slurry-ETAAS methodology. In

this paper, rapid and reliable procedures for such

2.3. Samples

a purpose are discussed. These procedures are

based on slurrying the samples in a hydrofluoric-

The method was validated by using four refer-

nitric acid mixture followed by measuring the

ence materials, namely, Montana Soil (SRM

metal concentrations by ETAAS, using a fast-

2711), San Joaquin Soil (SRM 2709), domestic

heating program w12x.

sludge (SRM 2881) and estuarine sediment (CRM

277). In addition, two soil samples were obtained

2. Experimental

in the neighborhood of a mining zone and a third

sample was collected from an agricultural soil.

2.1. Instrumentation

Two samples of river sediment obtained in Murcia

A Perkin–Elmer Model 1100B atomic absorp-

(Spain) were also analyzed.

tion spectrometer equipped with deuterium-arc

2.4. Procedures

background correction and an HGA-400 (Perkin–

Elmer) electrothermal atomizer were used. Pyro-

The samples were ground by ball-mill for 15

lytic graphite platforms were obtained from the

min and the resulting powders were kept in tightly

same manufacturer. Measurements were carried out

closed plastic containers until analysis. No sieving

using hollow cathode lamps operated in the exper-

was carried out. For silver determination the sus-

imental conditions given in Table 1. Argon was pensions were prepared by weighing the samples

used as the inert gas, the flow rate being 300 ml

miny1 in all stages except during atomization,

(typical amounts were in the 5–200 mg range) in

plastic beakers and then adding 25 ml of a suspen-

when the flow was stopped. Background-corrected

sion medium containing 3% concentrated nitric

integrated absorbance was used as the analytical

acid, 25% vyvconcentrated hydrofluoric acid solu-

signal.

tion and 50% concentrated hydrogen peroxide. For

A Fritsch Pulverisette (Idar–Oberstein, Germa-

gold and owing to the low content in the samples

ny) ball-mill of 80 ml capacity with 20 agate balls

studied, the suspensions were prepared by slurry-

of 1 cm diameter was used for grinding the

ing up to 200 mg of sample in 1 ml of concentrated

samples. A domestic microwave oven (maximum

hydrofluoric acid

heating power 1450 W) was used to mildly heat

(caution: this is a dangerous

chemical and suitable safety precautions must be

the suspensions. Plastic (polypropylene) vessels

adopted

were found to be suitable for preparing and storing

). No other reagent was necessary for gold

determination. For both analytes, the suspensions

the solutions and the suspensions. Pipette tips were

were submitted to a 30 s mild heating stage using

also of polypropylene. Mineralization of the sam-

a domestic microwave oven at maximum power

ples for comparison purposes was carried out in

or, alternatively, to a 10 min heating stage in a

closed Teflon cups using a MLS-1200 MEGA

steam bath. After cooling, aliquots

microwave oven (Milestone, Bergamo, Italy) and

(20 ml) were

taken while the solution was being continuously

a MDR-1000y6 Rotor (Radiometer, Copenhagen,

stirred with a magnetic stirrer and manually inject-

Denmark).

ed into the electrothermal atomizer. The heating

2.2. Reagents

programs given in Table 1 were used.

To confirm the reliability of the procedures, the

Standard solutions (1000 mg ly1) of silver and

samples, with the exception of SRMs, were pre-

gold were obtained from Fluka (Buchs, Switzer-

viously analyzed. Fractions (50–250 mg) of the

I. Lopez-Gar

cıa

ét al. / Spectrochimica Acta Part B 58 (2003) 1715–1721

1717

Table 1

Instrumental settings and heating programs

Parameter

Gold

Silver

Current lamp (mA)

Wavelength (nm)

242.8

328.1

Bandpass (nm)

0.7

Atomizer type

Platform

Injection volume (ml)

Calibration (ng mly )

0–200

0–20

Characteristic mass (pg)

Detection limit (mg gy1)

0.01

0.02

R.S.D.a(%)

"3.1

"2.3

(100 ng mly1)

(10 ng mly1)

Furnace heating programs

Element

Step

Parameter

Gold

Silver

Dry

T (8C)

300

Ramp (s)

Hold (s)

Atomizationb

T (8C)

2400

1800

Ramp (s)

Hold (s)

Clean

T (8C)

2650

Ramp (s)

Hold (s)

a n s10.

b The flow of argon was stopped.

samples were weighed into Teflon cups and 5 ml

influence of the atomization temperature on the

of concentrated hydrofluoric acid, 0.2 ml of con-

analytical signals was studied in the presence of

centrated nitric acid and 5 ml of water were added.

hydrofluoric acid for both aqueous solutions of the

The program used in the MLS-1200 microwave

metals and suspensions prepared from two differ-

oven consisted of 8 min to reach 1000 W and 7

ent samples. For silver determination, the atomi-

min hold at 1000 W. After this treatment, the

zation temperatures were varied in the 1500–2100

samples were maintained in the closed cups in an

8C range and 1800 8C was selected as optimal.

ice bath for 10 min before being diluted with de-

When the optimal atomization temperature for

ionized water in 25 ml volumetric flasks. Solutions

gold was studied, the aqueous solutions and the

were analyzed by ETAAS.

slurries prepared from the samples behaved differ-

ently, the study being made quite difficult by both

3. Results and discussion

the very low level of analyte and the high back-

3.1. Optimization of the heating programs

ground value. Fig. 1 shows the variation of the analytical signal obtained from a 45 mg ly1 gold

Fast-heating

methodology

w12x

was

used

solution prepared in concentrated hydrofluoric acid

throughout the work. The optimal drying temper-

(curve a) and that found for a 150 mg mly1 slurry

ature and holding time to be used in the modified

prepared from SRM 2711 and spiked with 68 mg

drying step were studied, being the optimal values

ly1 of gold (curve e). It is clear that there was a

300 8C with a 10 s holding time and 20 s ramp

matrix effect and a temperature above that selected

time.

for silver is required. As can be seen in curves b–

To prevent damage to the pyrolytic material

d, the change in the analytical signal in the

from the high silicate content of the samples, the

presence of the solid phase was similar to that

1718

I. Lopez-Gar

cıa

ét al. / Spectrochimica Acta Part B 58 (2003) 1715–1721

when nickel was replaced by others chemical

modifiers. Finally, to facilitate calibration against aqueous standards, no chemical modifier was added and 2400 8C was selected as the optimal

temperature. Under these conditions, no matrix

effect was observed and the background was

lowest.

3.2. Optimization of the chemical agents con-

centrations

As indicated above, the suspensions for gold

determination had to be prepared with a high

proportion of solid matter, which hindered meas-

urement and made the use of a high proportion of

concentrated hydrofluoric acid in the suspending

Fig. 1. Variation of the analytical signal with the atomization medium advisable. This proportion was varied in

temperature. Curves a–d correspond to a 45 mg ly1 gold standard solution in the presence of 0, 200, 400 and 1000 mg ly1

the 0–100% vyvrange and the fractions of analyte

nickel solution, respectively. Curve e was obtained using a 150

mobilized into the liquid phase were estimated by

mg mly1 slurry spiked with 68 mg ly1 gold standard solution.

measuring the gold content of the liquids after

they were filtered through a 0.45 mm membrane

obtained for gold atomization when a classical

filter. Fig. 2 shows gold atomization profiles modifier such as a nickel salt was present, the

obtained in different experimental conditions. As

peak area for gold decreasing as the nickel con-

can be seen in the graphs, when a 150 mg mly1

centration increased. A similar effect was observed

slurry was prepared in a 25% vyvconcentrated

Fig. 2. Atomization profiles for gold obtained from 150 mg mly1 slurries prepared in a medium containing 25 and 100% vyv concentrated hydrofluoric acid (a and b, respectively) and spiked with 200 mg ly1 gold standard. Graphs c and d refer to gold measurement in the supernatants. Graphs e and f refer to gold measurement in the supernatants after the heating treatment.

I. Lopez-Gar

cıa

ét al. / Spectrochimica Acta Part B 58 (2003) 1715–1721

1719

Fig. 3. Effect of the hydrogen peroxide concentration on the peak area of silver. The segments over the bars indicate the S.D.

hydrofluoric acid solution and spiked with 200 mg

A different behavior was observed in the case

ly1 of gold, the analytical signal was very poor

of silver measurements. The extraction of the

and the reliable measurement impossible. When

analyte to the aqueous phase did not increase when

the solid sample was suspended directly in the

the proportion of concentrated hydrofluoric acid in

concentrated hydrofluoric acid solution, a well-

the suspending medium exceeded 25% vyv. How-

shaped atomization profile was obtained. The frac-

ever, an important effect on silver mobilization

tion of gold mobilized was in this case calculated

was observed when hydrogen peroxide was incor-

to be approximately 20%. Graphs 2c and 2d show

porated in the suspension medium. As can be seen

gold atomization profiles obtained from the super-

from Fig. 3, the signal from the silver in the natants. The analytical signal increased (graphs 2e

supernatant considerably increased in the presence

and 2f) when the suspensions were submitted to a

of this chemical. Silver extraction was near 100%

mild heating stage (10 min in a steam bath). No

and so the relative standard deviations (R.S.D.) of

effect was observed in the signal obtained from

the measurements ( n s10) decreased from 8–10%

aqueous standards when the hydrofluoric acid con-

in the absence of hydrogen peroxide, to 2–3% in

centration was increased.

its presence (50% vyvconcentrated solution) in

To make the heating stage easier to carry out,

the suspension medium.

the use of a domestic microwave oven was studied.

It was verified that a short 30 s heating step was

3.3. Influence of the proportion of solid matter in sufficient to achieve maximum signal, approxi-the suspension

mately 45% of gold being extracted to the aqueous

phase when the suspensions were prepared directly

The influence of the suspension concentration

in the concentrated acid solution, as recommended.

on silver determination was studied for three ref-

Lower hydrofluoric acid concentrations proved

erence materials (SRMs 2711, 2709 and 2781),

inadequate. It must be remembered that suitable

covering a wide interval of concentrations (4.63,

safety precautions must be observed when using

0.41 and 98 mg kgy1, respectively). Linearity was

the

concentrated

and

hot

hydrofluoric

acid

obtained for suspensions prepared using up to 4,

medium.

22 and 2 mg mly1 for SRMs 2711, 2709 and

1720

I. Lopez-Gar

cıa

ét al. / Spectrochimica Acta Part B 58 (2003) 1715–1721

Table 2

Results for the determination of gold and silver in soil, sediment and sludge samples Sample

Contenta, mg kgy1

Gold

Silver

Slurry

Mineralization

Slurry

Mineralization

Industrial soil 1

0.12"0.02

0.11"0.02

0.54"0.03

0.59"0.03

Industrial soil 2

0.13" 0.03d

4.94"0.20

4.84"0.15

Agricultural soil

0.06" 0.02e

0.05" 0.02e

0.20"0.05

0.17"0.03

River sediment 1

0.07" 0.03e

3.32"0.02

3.38"0.02

River sediment 2

0.12" 0.02d

0.13" 0.03d

1.20"0.02

1.23"0.03

SRM 2711 (Montana soil)

0.13" 0.03d

0.03c

4.78"0.08

4.63" 0.39b

SRM 2709 (San Joaquin soil)

0.32"0.06

0.3c

0.37"0.03

0.41" 0.03b

SRM 2781 (Domestic sludge)

0.07" 0.03e

94.0"6

98.0" 8b

CRM 277 (Estuarine sediment)

0.12" 0.02d

0.13" 0.03d

3.35"0.06

3.25c

a Mean"S.D. ( n s4).

b Certified value.

c Not certified. Value is given by the supplier for informative purposes.

d Sample spiked with 0.12 mg kgy1 of gold.

e Sample spiked with 0.06 mg kgy1 of gold.

2781, respectively. These limits must be considered

ation). The repeatability was calculated using the

illustrative, since they depend on the concentration

RSDs for ten successive injections. The data

of the metal in the sample. R.S.D.s ( n s10) were shown were obtained by applying the recommend-below"4% when the proportion of solid matter in

ed procedure, which includes heating the suspen-

the suspension was above 0.2 mg mly1. For the

sion. When this stage was omitted, R.S.D.s

samples with a relatively high silver content, dilut-

increased up to "8 and "6%, for gold and silver, ed slurries were prepared and the R.S.D.s were in

respectively.

the 5–10% range.

The results obtained for the different samples

For gold determination, the influence of the

analyzed using the proposed procedures, as well

suspension concentration was studied for two soil

as a reference method based on mineralization

samples of a mining and agricultural origin and a

using a closed system, are summarized in Table 2.

reference soil (SRM 2709). Due to the low gold

The reliability of the method was further corrobo-

concentration in the samples, they were spiked

rated by using four certified reference materials.

with a gold standard solution. Linearity was in this

The results obtained are also shown in Table 2,

case obtained up to 300 mg mly1. The absence of

together with the certified values.

a matrix effect was verified by comparing the

slopes of the aqueous calibration and standard

additions calibration graphs obtained for suspen-

4. Conclusion

sions of all the samples studied at two different

concentration levels.

The slurry-ETAAS approach allows the rapid

determination of silver and gold in soils and

3.4. Calibration graphs, repeatability, results and sediments, confirming the advantages of this meth-accuracy

odology over conventional methods. The presence

of hydrofluoric acid in the suspension medium is

Table 1 shows the concentration range of the important since, in addition to its extracting effect calibration graphs. The detection limits were cal-

(which transfers a significant fraction of the anal-

culated for ten successive injections of the blank

yte into the supernatant), it acts as a true chemical (criterion based on three times the standard devi-modifier by simplifying the matrix during the

I. Lopez-Gar

cıa

ét al. / Spectrochimica Acta Part B 58 (2003) 1715–1721

1721

heating cycle when suspensions are prepared from

and Pseudomonas Putida, J. Anal. At. Spectrom. 8

materials with a high silicaysilicate content.

(1993) 1015–1022.

w6x I. Lopez-Gar

ćıa,

M. Sanchez-Merlos,

M. Hernandez-

Ćor

´ doba, Rapid determination of selenium in soils and

Acknowledgments

sediments using slurry-sampling-electrothermal atomic absorption spectrometry, J. Anal. At. Spectrom. 11

The authors are grateful to the Spanish Ministery

(1996) 1003–1006.

of Science and Technology for financial support

w7x I. Lopez-Gar

ćıa,

M. Sanchez-Merlos,

M. Hernandez-

(Project BQU2000-0218).

Cor

´ doba, Arsenic and antimony determination in soils

and sediments by graphite furnace atomic absorption

References

spectrometry with slurry sampling, Spectrochim. Acta

Part B 52 (1997) 437–443.

w1x M.J. Cal-Prieto, M. Felipe-Sotelo, A. Carlosena, J.M.

8x N. Campillo, I. Lopez-Gar

ćıa,

P. Vinas,

I. Arnau-Jerez,

Andrade, P. Lopez-Mahıa,

Ś. Muniategui, D. Prada,

M. Hernandez-Cor

´ doba, Determination of vanadium,

Slurry sampling for direct analysis of solid materials by molybdenum and chromium in soils, sediments and

electrothermal

atomic

absorption

spectrometry

sludges by electrothermal atomic absorption spectrom-

(ETAAS). A literature review from 1990 to 2000,

etry with slurry sample introduction, J. Anal. At. Spec-Talanta 56 (2002) 1–51.

trom. 17 (2002) 1429–1433.

w2x P. Bermejo-Barrera, A. Moreda-Piner

o, J. Moreda-Pine-

w9x C. Bendicho, M.T.C. de Loos-Vollebregt, The influence ro, A. Bermejo-Barrera, Determination of traces of silver of pyrolysis and matrix modifiers for analysis of glass in human scalp hair slurries by electrothermal atomic materials by GFAAS using slurry sample introduction,

absorption spectrometry, Mikrochim. Acta 129 (1998)

Spectrochim. Acta Part B 45 (1990) 679–693.

71–76.

w10x C. Bendicho, M.T.C. de Loos-Vollebregt, Metal extrac-w3x E.D. Byrd, D.J. Butcher, Determination of trace elemen-tion by hydrofluoric acid from slurries of glass materials ts in biological standard reference materials by graphite in graphite furnace atomic absorption spectrometry,

furnace atomic absorption spectrometry with solid and Spectrochim. Acta Part B 45 (1990) 695–710.

slurry sampling, Spectrosc. Lett. 26 (1993) 1613–1624.

w11x I. Lopez-Gar

ćıa,

É. Navarro, P. Vinas,

M. Hernandez-

w4x M.W. Hinds, Determination of gold, palladium and

Cor

´ doba, Rapid determination of lead, cadmium and

platinum in high-purity silver by different solid-sam-thallium in cements using electrothermal atomic absorp-pling graphite-furnace atomic-absorption spectrometry tion spectrometry with slurry sampling introduction,

methods, Spectrochim. Acta Part B 48 (1993) 435–445.

Fresenius J. Anal. Chem. 357 (1997) 642–646.

w5x L.C. Robles, C. Garcıa-Olalla,

Á.J. Aller, Determination

w12x D.J. Halls, Analytical minimalism applied to the deter-of gold by slurry electrothermal atomic absorption

mination of trace elements by atomic spectrometry, J.

spectrometry after preconcentration by Escherichia Coli Anal. At. Spectrom. 10 (1995) 169–175.

Document Outline

Slurry sampling for the determination of silver and gold in soils and sediments using electrothermal atomic absorption spec ...

srebro i złoto w glebie

Document Outline