Determination of the platinum group elements in geological samples by ICP-MS after NiS fire assay and Te coprecipitation: ultrasound-assisted extraction of PGEs from Te precipitate

Abstract

A method for the determination of the platinum group elements in geological samples by ICP-MS after NiS fire assay and Te coprecipitation has been developed. With the aim of controlling the volatility of OsO₄ and reducing the number of procedural blanks of platinum-group elements (PGEs) as a whole, an ultrasound-assisted extraction technique was attempted for the dissolution of the Te precipitate at low temperature, which prevented the loss of Os. Total blanks were less than 0.12 ng g⁻¹ and method detection limits (3σ, n = 11) were within 0.006–0.039 ng g⁻¹ for all elements. The proposed method was successfully applied to the simultaneous determination of all PGEs in the geological certified reference materials; the results were found to be in reasonable agreement with the certified values. The precisions that were evaluated with the seven replicate results of GBW 07290 and SARM-7 were within the range from 1.1%–7.7% RSD.

---

Introduction

The economic and geological importance of platinum-group elements (PGEs: Ru, Rh, Pd, Os, Ir, Pt) have led to considerable interest being given to developing reliable analytical methods to quantify them in real samples. Various methods have been used for the determination of PGEs in geological samples.^1–2 The widely used procedure is the NiS fire assay coupled with a Te coprecipitation and determination by ICP-MS.^3–7 NiS-FA allows for the use of a large sample size, which might contain discrete or inhomogeneous distribution of these elements in mineral species, and can be analyzed precisely. However, the NiS fire assay separation procedure for ICP-MS analyses also has its own limitations, for example, high numbers of blanks, volatility of some chloride PGEs, poor recovery of some elements and Os loss as OsO₄. In the past few years, new development techniques for determination of PGEs have focused on resolving the above drawbacks.^3–9 However, these conventional methods have a poor recovery of Os³ and quite a number of reports have abandoned the simultaneous determination of Os.^3,9–11 The main difficulty is the loss of Os as OsO₄ in the digestion step of Te precipitation with the oxidant. Generally, temperature controlling digestion of the Te precipitate, which is maintained below 55 °C, is an effective way to avoid the loss of Os,^6,12 but it is time consuming.

Ultrasonic radiation is a powerful aid in the acceleration of various steps of the analytical process. It is an effective way of extracting a number of analytes from different types of samples.^13,14 In this Technical Note, we present ultrasound-assisted extraction of PGEs from Te precipitation at low temperature that prevents the loss of Os. Thus modified, the NiS fire assay–Te coprecipitation method allows the determination of all PGEs (Os included) in geological samples and obviously shortens the analyzing process.

Experimental

Reagents

Sublimated sulfur, sodium tetraborate, sodium carbonate and Silicon dioxide floated powder were used as fusion reagents, all of which were analytical-grade produced by Xi’an Chemical Reagent Co.; carbonyl Ni powder (>99.8% Ni, INCO) was purchased from Shanghai Huazhen Co.; 22.5% m/v tin dichloride was prepared by dissolving 45 g of SnCl₂·2H₂O (Analytical-grade) in 100 ml HCl (1 + 1) and diluting to 200 ml with water; 2 mg ml⁻¹ of Te solution was prepared by dissolving 2 g of Te lumps (high purity) in aqua regia and diluting to 1000 ml with 10% (v/v) HCl. Other reagents include analytical-grade nitric acid (70%), hydrochloric acid (37%) and H₂O₂ (30%).

An Os isotope spike with 97.04% ¹⁹⁰Os and 1.61% ¹⁹²Os was prepared from ¹⁹⁰Os metal powder (Oak Ridge National Laboratory, USA) by the National Research Center of Geoanalysis, and the working spike solution contains 67 ng ml⁻¹ of Os; 10 mg l⁻¹ CLMS-3 (SPEX, Certiprep Inc., USA) and 5 mg l⁻¹ Os (National Standard Central, China) were used as standard stock solutions. Ultrapure water was used throughout.

Silicon dioxide floated powder, tin dichloride and hydrochloric acid were further purified according to the method described by He et al.⁴

Sample preparation

The sample preparation procedures are as follows:

(1) 20 g of the finely powdered sample, 3.5 g of Ni, 1.7 g of S, 10 g of Na₂CO₃, 20 g of Na₂B₄O₇ and 3 g of SiO₂ (7 g for procedural blank) were mixed and transferred into a fire clay crucible. Under reductive conditions³ the sample was fused in a preheated furnace for 75 min at 1050 °C. The NiS button was separated from the slag in an iron mold and crushed into small chips (<1 mm in diameter) in an agate mortar. The NiS small chips were transferred into a 200 ml beaker and dissolved with 100 ml of purified 6 mol l⁻¹ HCl for 2–3 h at 90 °C.

(2) 1 ml of Te solution was added to the complete dissolved solution, the solution was stirred and 2 ml of SnCl₂ solution was added dropwise. A fine, black precipitate formed and was heated to coagulate a Te precipitate at about 80–90 °C for 60 min. A duplicate coprecipitation was required for further collection of the residual PGEs. In this step, the standard series was prepared in parallel with a real sample.

(3) The sample and standard solutions were cooled and then vacuum filtered through a mixed cellulose ester membrane (0.45 μm). The cellulose ester membranes coated with coagulate Te precipitates were rinsed 7–10 times with HCl 10% (v/v). The membranes were transferred into a 25 ml test tube; 0.5 ml of refrigerated HCl, 0.5 ml of refrigerated HNO₃ and 1 ml of refrigerated H₂O₂ were added, and then the open end of the test tube was firmly sealed immediately with a double thickness of Parafilm. During digestion of Te precipitate, the test tube was placed in an ultrasonic cleaner bath, which was added to ice water to keep a low temperature. Ultrasound-assisted extraction of PGEs from Te precipitate was usually carried out after 6–7 min. The test tube was placed in ice water at least 10 min after the complete dissolution, the sealed Parafilm was punctured and iced ultrapure water immediately added into test tube through the hole.

(4) The sample and standard series solutions were diluted to 25 ml with ultrapure water. These final solutions were immediately used for determining Ru, Rh, Pd, Os, Ir and Pt with an ICP-MS.

Instrumentation

A 50 W, 20 kHz ultrasonic cleaner-bath was used for extraction of PGEs from Te precipitate.

Determination was performed with an ICP-MS instrument (X7, Thermo Elemental, USA). The operating conditions are summarized in Table 1. A 10 ng ml⁻¹ Cd and Tl mixed internal standard solution, respectively, close to light (Ru, Rh and Pd) and heavy (Os, Ir and Pt) PGEs, was introduced online by a “Y”-shaped three way linker during ICP-MS measurement. The measured isotopes are given in Table 1. Significant interferences on the determination of PGEs by ICP-MS after nickel sulfide fire assay have been reported by Juvonen et al.⁷ In the present work, the isobaric overlap between ¹⁰⁶Cd and ¹⁰⁶Pd was auto-corrected by the mathematical treatment software of the instrument. An off-line correction⁷ was used to correct the interference of Cu on Rh except that the experimentally determined factor was 0.456 instead of 0.4324. According to the experimental result, no significant interference from Ni on ¹⁰¹Ru was observed; the interferential factor was 0.0 000 003.

Table 1 Instrumental operating conditions

Parameter	Setting
a Correct interference. b Internal standard.
RF power	1350 W
Plasma gas flow rate	14 l min^−1
Auxiliary gas flow rate	1.0 l min^−1
Nebulizer gas flow rate	0.88 l min^−1
Nebulizer	Meinhard, glass, concentric
Spray chamber temperature	3 °C
Sampler cone	Nickel, 1.0 mm orifice diameter
Skimmer cone	Nickel, 0.7 mm orifice diameter
Acquisition mode	Peak jumping
Number of sweeps	50
Channels dwell time per sweep	10 ms
Channels per mass	3
Measured isotopes (isotopic abundance %)
^101Ru (17.0), ^103Rh (100), ^105Pd(22.3), ^106Pd (27.3)^a, ^189Os(16.1), ^193Ir(61.5), ^195Pt(33.7), ^111Cd (12.9)^b, ^205Tl(70.5)^b

---

Results and discussion

Signal response to different oxidation states and memory effect

The sensitivity of the determination of Os by ICP-MS depends strongly on the oxidation states of Os.¹⁵ In this work, we studied the influences of oxidation states on the determination of Os; the results show that the octavalent state of Os gives about 30–40 times higher sensitivity than that of the tetravalent state. The above result suggests that the valence state of Os in sample solutions must be exactly identical to that in the standard, and a mismatch of valence state will cause false results. In this work, the standard series of PGEs were prepared in parallel with real samples beginning with the coprecipitation procedure step. This method ensured that the oxidation states of Os, whether Os was in the real sample solution or standard solutions, were in exactly the same oxidation state. The memory effect is a very serious problem for Os determination by ICP-MS. Ammonia solution is a suitable rinsing agent to effectively clean the Os memory in ICP-MS analysis.¹⁶ In this work, considering other elements, both rinsing solutions—3% ammonia solution and diluted acid (2% HCl + 1% HNO₃)–were used separately to flush the sample introduction system in the time interval between each analytical run.

Optimization of the conditions of ultrasound-assisted extraction

The loss of highly volatile OsO₄ (boiling point 105 °C) is almost unavoidable during the chemical preparation. In order to solve this drawback, we investigated the possibility of eliminating the loss of OsO₄ using ultrasound-assisted extraction of the PGEs from Te precipitate and the parameters of sonication temperature and time were optimized.

In this work, an apparatus (Fig. 1) was designed to monitor the loss of OsO₄ during the dissolution of Te precipitation at different temperature and optimized sonication temperature. The volatile OsO₄ from the test tube is carried by the nebulizer gas and determined by ICP-MS. An electric heater was used to heat the water and a Pt100 thermocouple was used to monitor the temperature of the water-bath, in which the extraction cell (test tube) is immersed. The temperature was autocontrolled by a temperature controller. At first, 1000 ml of ice water (4 °C) was added to the ultrasonic cleaner bath. The loss of OsO₄ was monitored on-line at different temperatures, and the results are shown in Fig. 2. The present experiments clearly demonstrate that the maximum sonication temperature without Os loss was 30 °C; the faint signal of ¹⁸⁹Os was determined when the sonication temperature was in the range of 30–50 °C. At the higher temperature (>50 °C) the intensity of Os was drastically increased. In addition, the relationship between the recovery of Os and sonication temperature was studied and the results are shown in Fig. 3. It can be clearly seen that the Os percent recovery reduced with increasing of the sonication temperature and the Os percent recovery was near completeness (>98%) when the temperature was lower than 30 °C. It is worth pointing out that the temperature of the water-bath continually increases during ultrasound-assisted extraction. In order to control the sonication temperature in the optimal range, it is necessary to add ice to the water-bath.

Fig. 1 Apparatus for monitoring the loss of OsO₄.

Fig. 2 Effect of sonication temperature on the signal intensity of ¹⁸⁹Os. 200 ng Os standard in Te coprecipitation; left standing for 20 s at each monitoring point to measure the ¹⁸⁹Os count.

Fig. 3 Effect of sonication temperature on the recovery of ¹⁸⁹Os. 200 ng Os standard in Te coprecipitation; sonication time: 6 min.

Different sonication times were tested in order to determine the time necessary for total removal of the PGE Te precipitate and the results are shown in Fig. 4. It can be clearly seen that the PGE percent recoveries increased with the length of the sonication time, and maximums were reached at a length of about 6–7 min. With longer extraction time the recoveries of PGEs gradually dropped because the filter became colloid matter, and the colloid matrix affected the sample introduction for the ICP-MS determination. In this work, 6 min was chosen as the extraction time.

Fig. 4 Influence of ultrasound-assisted extraction time on the recoveries of PGEs. Extraction temperature: 4–10 °C; 8 ng ml⁻¹ PGEs standard.

Comparison of two methods for determination of Os

In order to further evaluate the feasibility of the proposed method for the determination of Os, three Chinese geological certified reference materials (GBW 07288, 07291 and 07294) and a South African geological certified reference material (SARM-7) were determined using the present method and compared with the method of isotope dilution-ICP-MS after distillation¹⁵—called method A. The results are listed in Table 2.

Table 2 Comparison of results of Os in real samples with different methods (n = 3, mean ± SD)

	Os/ng g^−1			t-Test value
CRMs	Present method	Method A	Certified values	t 1	t 2
GBW 07288	0.04 ± 0.01	0.08 ± 0.03	(0.05)	1.73	6.93
GBW 07291	10.6 ± 0.7	11.2 ± 1.2	9.6 ± 2.0	2.46	1.45
GBW 07294	0.57 ± 0.08	0.54 ± 0.07	0.64 ± 0.14	1.52	0.65
SARM-7	63.2 ± 3.1	67.5 ± 4.3	63 ± 6.8	0.11	2.40

---

Statistical testing, using the t-test (t₁) on the data of the four CRMs obtained with the present method and certified values, showed that with 95% probability, all the results of the present method were statistically equivalent to the certified values. In addition, the t-test (t₂) showed that the means of the results of the two methods were in agreement for the CRMs, but not for GBW 07288.

Procedural blanks and detection limit

A disadvantage of the nickel sulfide fire assay is the occurrence of relatively large reagent blanks. To overcome this drawback, silicon dioxide, tin dichloride and hydrochloric acid were further purified. The average blank levels of Ru, Rh, Pd, Os, Ir and Pt for the total analytical procedures are 0.013, 0.006, 0.103, 0.016, 0.017 and 0.115 ng g⁻¹ for a 20 g sample, respectively.

The instrumental detection limits (IDL) were calculated as three times the standard deviation of eleven standard blank values; 0.16, 0.09, 0.14, 0.08, 0.13 and 0.23 ng l⁻¹ for Ru, Rh, Pd, Os, Ir and Pt, respectively.

The method detection limits (MDL) were simply calculated as three times the standard deviation of eleven total procedure blank values. 0.009, 0.006, 0.036, 0.009, 0.006 and 0.039 ng g⁻¹ for a 20 g sample size of Ru, Rh, Pd, Os, Ir and Pt, respectively. Of course, real MDL will be affected by spectral interferences originating from elements present in the sample.

Accuracy and reproducibility

In order to investigate the feasibility, accuracy and precision of the proposed method, four certified geological reference materials were analyzed by the present method. All the analytical results are listed in Table 3. It can be seen that the concentrations of PGEs in the certified reference materials obtained by the present method were in good agreement with the certified values of these CRMs including those with very low PGE content (e.g. GBW 07288). The precisions that were evaluated with the seven replicate results of GBW 07290 and SARM-7 were within the range from 1.1%–7.7% RSD (Table 3).

Table 3 Analytical results (ng g⁻¹) of PGEs in the certified reference materials (n = 7)

CRMs		Ru	Rh	Pd	Os	Ir	Pt
a Weighed 5.0 g sample and diluted to 50 ml and diluted 10-fold for the determination of Pd and Pt.
GBW 07288 (arenaceous soil)	Determined value	0.043 ± 0.012	0.014 ± 0.003	0.23 ± 0.02	0.043 ± 0.014	0.051 ± 0.013	0.24 ± 0.03
	Certified value	(0.05)	(0.02)	0.26 ± 0.05	(0.05)	(0.04)	0.26 ± 0.05
GBW 07290 (peridotite)	Determined value	14.2 ± 0.62	1.23 ± 0.10	4.49 ± 0.34	10.1 ± 0.47	4.37 ± 0.31	6.13 ± 0.44
	RSD (%)	4.4	7.7	7.6	4.6	7.1	7.2
	Certified value	14.8 ± 2.7	1.3 ± 0.3	4.6 ± 0.6	9.6 ± 2.0	4.3 ± 0.5	6.4 ± 0.9
GBW 07294 (soil)	Determined value	0.63 ± 0.09	1.03 ± 0.06	14.8 ± 0.7	0.60 ± 0.04	1.14 ± 0.07	14.3 ± 0.9
	Certified value	0.66 ± 0.20	1.1 ± 0.2	15.2 ± 2.3	0.64 ± 0.14	1.2 ± 0.3	14.7 ± 2.5
SARM-7 ^a	Determined value	427 ± 17	246 ± 7	1498 ± 23	66 ± 3.2	78 ± 3.9	3679 ± 41
	RSD (%)	4.0	2.8	1.5	4.8	5.0	1.1
	Certified value	430 ± 57	240 ± 13	1530 ± 32	63 ± 6.8	74 ± 12	3740 ± 45

---

Conclusions

The improved method for determination of PGEs in geological samples by ICP-MS after NiS fire assay and Te coprecipitation is described. With our modifications, the ultrasound-assisted extraction system allows efficient recoveries of PGEs during the Te precipitation dissolution step, without the isotope-dilution technique, the concentrations of Os together with other PGEs can be precisely determined by ICP-MS.

We are grateful to the referees for comments and advice on an early version of this Technical Note.

References

1. R. R. Barefoot and J. C. Van Loon, Talanta, 1999, 49, 1.
2. R. R. Barefoot, J. Anal. At. Spectrom., 1998, 13, 1077.
3. K. Oguri, G. Shimoda and Y. Tatsumi, Chem. Geol., 1999, 157, 189.
4. H. He, C. Lü and Z. Zhou et al., Rock Miner. Analysis (China), 2001, 20, 191.
5. C. Lü, H. He and Z. Zhou et al., Rock Miner. Analysis (China), 2002, 21, 7.
6. M. Gros, J. P. Lorand and A. Luguet, Chem. Geol., 2002, 185, 179.
7. R. Juvonen, T. Lakomaa and L. Soikkeli, Talanta, 2002, 58, 595.
8. M. Sun, J. Jain and M. F. Zhou et al., Can. J. Appl. Spectrosc., 1993, 38, 103.
9. G. S. Reddi, C. R. M. Rao and T. A. S. Rao et al., Fresenius’ J. Anal. Chem., 1994, 348, 350.
10. C. Li, C. Chai and X. Li et al., Geostand. Newsl., 1998, 22, 195.
11. Y. Sun, X. Guan and A. Du, Spectrochim. Acta, Part B, 1998, 53, 1463.
12. S. E. Jackson, B. J. Fryer and W. Gosse et al., Chem. Geol., 1990, 83, 119.
13. I. Lópze, N. Campillo and I. Arnau et al., Anal. Chim. Acta, 2005, 531, 125.
14. M. C. Yebra, S. Cancela and A. Moreno, Anal. Chim. Acta, 2005, 533, 51.
15. Y. Sun, M. Zhou and M. Sun, J. Anal. At. Spectrom., 2001, 16, 345.
16. D. Malinovsky, I. Rodushkin and D. Baxter et al., Anal. Chim. Acta, 2002, 463, 111.

---