Quantification of Ultra-Trace Levels of Pt, Ir and Rh in Polar Snow and Ice Using ICP-SFMS Coupled with a Pre-Concentration and Desolvation Nebulization System

Full text

Notes Bull. Korean Chem. Soc. 2011, Vol. 32, No. 6 2105
DOI 10.5012/bkcs.2011.32.6.2105
Quantification of Ultra-Trace Levels of Pt, Ir and Rh in Polar Snow and Ice Using
ICP-SFMS Coupled with a Pre-Concentration and Desolvation Nebulization System
Tseren-Ochir Soyol-Erdene,†,# Youngsook Huh,†,‡ Sungmin Hong,§ Hee Jin Hwang,# and Soon Do Hur#,*
†School of Earth and Environmental Sciences, Seoul National University, Seoul 151-747, Korea
‡Research Institute of Oceanography, Seoul National University, Seoul 151-747, Korea
§Department of Oceanography, Inha University, Incheon 402-751, Korea
#Korea Polar Research Institute, Songdo Techno Park, Incheon 406-840, Korea. *E-mail: sdhur@kopri.re.kr
Received December 22, 2010, Accepted April 26, 2011
Key Words : Platinum group elements, Antarctic snow, Ice core
A revised analytical method is presented for the quanti-
fication of ultra trace level (fg mL−1) PGEs (Pt, Ir and Rh) in
polar snow and ice samples by focusing on the following
issues: (i) evaluation of the efficiency of a non-boiling pre-
concentration procedure for analyses at the fg mL−1 level of
PGEs in our laboratory; (ii) establishment of the appro-
priate instrumental conditions to obtain low detection limits,
high accuracy and precision; and (iii) verification of the
contributions of possible contamination during the ice core
decontamination process using an artificial ice core as a
sample.
Laboratory Clean Conditions. To avoid possible arti-
ficial contamination of extremely low concentration (fg mL−1)
PGEs in polar samples, all analytical procedures including
sample handling and analyses were carried out in a class-10
clean booth or on a class-10 clean bench inside specially
designed clean laboratories (class 1000) located at the Korea
Polar Research Institute (KOPRI). The cleaning procedures
for the experimental tools, including low- and high-density
polyethylene bottles (LDPE and HDPE), Teflon beakers and
stainless steel tools used for mechanical chiseling during the
decontamination procedure of the ice core sample, were also
performed in the class-10 clean booth according to the
previously established cleaning protocol1 (see supplementary
material).
Ultrapure (UP) water was used for the final cleaning step
and the preparation of reference standard solutions, was
accomplished using a three-stage distillation process:
Millipore RO water purification (Model Elix-3), Milli-Q
purification (Millipore Corp., Model Milli-Q Academic) and
the use of a sub-boiling double distillation still (Berghof
BSB-939-IR, Germany).1 The purest “Optima”-grade nitric
acid (HNO3), which was used in the final cleaning step,
sample acidification and in the preparation of the reference
solutions, was obtained from Fisher Scientific.
Sample Treatment. Pre-concentration of the samples was
performed via a non-boiling evaporation method to increase
the sample concentration by a factor of ~60.1,2 An aluminum
hot plate and pre-cleaned Teflon beakers were used for the
sample evaporation procedure. Before using the beakers,
they were pre-conditioned for 1 h on a hot plate with UP
water. Melted 60 mL samples were pre-concentrated by
sub-boiling to 0.5 mL at ~80 °C. A blank experiment was
performed by using ultrapure water as a sample for each
pre-concentration experiment. One pre-concentration experi-
ment consisted of five samples and one blank. The evapo-
ration of the samples for each pre-concentration set required
approximately 6 h. To adjust the HNO3 concentration to
1% in the final sample volume, 0.5 mL of 2% (m/m) UP
HNO3, which was prepared from UP water and “Optima”-
grade HNO3, was added to pre-concentrated samples (~0.5
mL). The pre-concentrated samples were then transferred to
4 mL HDPE bottles and kept frozen until analyses were
performed.
Instrumental Analyses. Analysis was performed using a
sector-field inductively coupled plasma mass spectrometer
(ICP-SFMS) (Element2, Thermo Finnigan MAT, Germany)
at KOPRI. An Apex high-sensitivity desolvation nebulizer
system (Apex HF, ESA, USA) was incorporated to transport
sample to the ICP-SFMS. The desolvation nebulizer reduced
the oxide production rate (BaO/Ba < 0.02%) and increased
the instrumental sensitivity by approximately 10 times
relative to the standard instrument (~0.9 × 106 cps for 100 pg
mL−1 indium solution). To obtain the lowest blank level,
ultrapure grade (99.999%) argon and nitrogen gases were
used, and the entire analytical system, including the ICP-
SFMS and the sample introduction system, was installed in a
class-10 clean booth located in a class-1000 clean laboratory.
The analysis for Pt and Ir and the analysis for Rh were
performed separately using ICP-SFMS for each pre-concen-
trated sample. Details of the instrumental operating condi-
tions and data acquisition parameters are illustrated in Table
S1.
Calibration and Detection Limits. An external calib-
ration method was applied for the quantification of Pt, Ir and
Rh in the pre-concentrated samples. Reference standard
solutions were prepared by sequential dilution from a 10
ppm PGE multi-element standard solution by weight base
(CMS-2, Inorganic Ventures, Lakewood, New Jersey, USA)
using UP 1% (m/m) HNO3. The concentrations of the
standard solutions used for the calibration curves were 0, 2,
5, 10, 15, 20 and 25 fg mL−1 for Ir quantification and 0, 50,

2106 Bull. Korean Chem. Soc. 2011, Vol. 32, No. 6 Notes
100, 200, 500, 1000 and 2000 fg mL−1 for Rh and Pt quanti-
fication. For these concentration ranges, the correlation
coefficients (r2) of the calibration curves were higher than
0.999 (Figure S1).
195Pt, 193Ir and 103Rh isotopes were measured because 195Pt
and 193Ir have the highest natural abundance and lowest
interference, whereas 103Rh is the only naturally-occurring
isotope of Rh (Table S2). Detection limits were calculated as
three times the standard deviations (1σ) of 10 measurements
of the blank (1% UP HNO3 solutions). The instrumental
detection limits obtained for Pt, Ir and Rh are 26 fg mL−1, 3
fg mL−1 and 9 fg mL−1, respectively. The procedural detec-
tion limits are as low as 0.43 fg mL−1, 0.05 fg mL−1, and 0.15
fg mL−1 for Pt, Ir and Rh, respectively, after correction with
a pre-concentration factor of 60.
Quality Control of Analyses.
Interferences: To avoid contributions by interferents, the
instrumental parameters were optimized daily for the lowest
oxide production rate and the highest sensitivity, and the
effects of potential interferences (Table S2) were checked
daily before the analysis was performed. The interference
check was performed by measuring the analytes (195Pt, 193Ir
and 103Rh) in 0-100 pg mL−1 (pg = 10−12 g) single-element
standard solutions of interferents (Rb, Sr and Cu for 103Rh;
Hf for 195Pt and 193Ir). The desolvation nebulization system
and a good quality design of the instrumental conditions
rendered the interference by interferent concentrations of up
to 100 pg mL−1 negligible. Thus, mathematical corrections
were not required in this concentration range. The Hf
concentrations, which is the potential interferent for Pt and
Ir, were estimated to be lower than 100 pg mL−1 in pre-
concentrated snow samples, whereas the concentrations for
the Rh interferents (Rb, Sr and Cu) were found to exceed
100 pg mL−1 in some pre-concentrated snow samples based
on the results of direct (without pre-concentration) measure-
ments of 80 Antarctic snow samples and a pre-concentration
factor of 60. Therefore, an analytical error range for Rh
determination was estimated in a river water matrix, which
is a matrix with much higher concentrations of potential
interferents, such as Rb and Sr, than the polar samples. This
protocol is described in the next section.
Sensitivity Drift: Internal standards are not preferred for
high-purity polar ice and snow samples because their use
can cause contamination at extremely low concentrations of
analytes. We corrected the sensitivity drift over time by
performing repeated analyses of the standard solutions
between samples. After every five sample analyses (or every
~20 minutes), the instrumental sensitivity drift and blank
levels were checked by measuring the standard (500 fg mL−1)
and blank (1% UP HNO3) solutions. Correction was per-
formed on the sample signals.
Reliability: Because no certified reference materials for
polar ice and snow are available, riverine water certified
reference materials (CRM) (SLRS4 and SLRS5, National
Research Council of Canada, Ottawa, Canada) were used for
accuracy control. First, the concentrations of Pt, Ir and Rh in
the riverine water materials were determined because there
are only one or two comparable studies for Pt and Ir in
SLRS4 and no comparable studies for Rh. The analytical
results for Pt, Ir and Rh from direct measurements (by an
external calibration method) and from a standard addition
method (spiked by PGE standard solutions of 0, 5, 10, 15, 20
and 25 fg mL−1 for Ir analyses and 0, 500, 1000, 1500, 2000
and 2500 fg mL−1 for Rh analyses, Figure S2) for SLRS4
and SLRS5 are presented in Table 1. The Pt concentration in
SLRS4 obtained by direct measurements was in good agree-
ment with those available in the literature.3,4 Thus, it was not
necessary to check the results via the standard addition
method because the concentration was relatively high, and
the results were accurately obtained by direct measurement.
Despite the very low concentration of Ir in SLRS4, the Ir
results obtained by the direct and standard addition measure-
ments were in good agreement with each other within the
range of analytical error. These findings were also consistent
with those reported in a previous study by the authors4 but
were two orders of magnitude lower than the findings
obtained in another study.3 In SLRS5, the Ir concentration
was only determined by the standard addition method, and
the Ir concentration was determined to be 4.8 ± 1.4 fg mL−1.
There is no published data pertaining to Rh concentrations in
either SLRS4 or SLRS5 because of the hindrance of strong
interferences and the ultra-low levels of Rh in riverine water.
The analytical results for Rh in SLRS5 as obtained by direct
measurement and the standard addition method were in
agreement with each other within approximately 20%. The
result for Rh using the standard addition method (1380 ± 90
fg mL−1) is preferred; interference and matrix effects were
eliminated because the samples and standards were both in
the same matrix. In summary, the Pt and Ir results are con-
sistent with the results from the literature and/or the results
of the standard addition method, whereas an analytical error
of approximately 20% for the measurement of Rh must be
considered.
An accuracy check using riverine water CRM, however,
cannot fully verify the reliability of snow and ice analyses
because of considerable differences in the matrix and the
concentration levels of the analytes between river water and
polar snow and ice. For this reason, two reference solutions
were prepared using the riverine water CRMs, which were
more suitable for reliability checks of the PGE analyses of
polar snow and ice samples. These references were (a) the
SLRS4 riverine water CRM diluted 10 times and spiked
Table 1 . Analytical results for river water CRMs (fg mL
−
1
) with
standard deviations (1σ)
PGE CRM This study Literature
Pt SLRS 4 1210 ± 96
a
1250 ± 146
(4)
1300 ± 100
(3)
Ir
SLRS 4
SLRS 4
6.9 ± 1.1
a
8.7 ± 0.8
b
8±4
(4)
8±4
(4)
300 ± 200
(3)
300 ± 200
(3)
SLRS 5 4.8 ± 1.4
b
--
Rh SLRS 5 1690 ± 80
a
--
SLRS 5 1390 ± 90
b
--
a
Direct measurement.
b
Standard addition method.

Notes Bull. Korean Chem. Soc. 2011, Vol. 32, No. 6 2107
with 50 fg mL−1 Ir (as Ir was expected to be too low (~0.8 fg
mL−1) in diluted SLRS4) for Pt and Ir analyses; and (b) the
SLRS5 riverine water CRM diluted 5 times for Rh analysis.
The final PGE concentrations in the reference materials
were calculated to be 115 ± 9 fg mL−1 Pt and 55 ± 5 fg mL−1
Ir (A-solution) and 276 ± 18 fg mL−1 Rh (B-solution).
Despite the quite low concentrations of Pt, Ir and Rh in the
prepared reference materials, all of the analytical results
obtained by repeated measurements between samples agreed
well with the expected concentrations within 10% for Pt
and Ir and within 25% for Rh after blank and signal-drift
corrections.
Recovery and Precision of the Sample Pre-concentra-
tion Procedure. To ensure accuracy, a recovery and pre-
cision test was performed using unacidified PGE multi-
element standard solutions of 0.1, 0.5, 1, 5, 10 and 20 fg mL−1
concentrations (prepared by weight base sequential dilution
from a 10 ppm PGE standard solution) as samples with
similar concentrations to those in polar snow and ice. For
each concentration level, one (for Rh) or three (for Pt and Ir)
standard solutions were pre-concentrated and analyzed. In
order to avoid possible changes in extremely low concen-
trations in unacidified solutions, the solutions were prepared
just before the experiment was conducted, and they were
immediately transferred to acid-soaked PFA beakers (within
10 minutes after final dilution) for pre-concentration. Experi-
mental blanks were evaluated with UP water as a sample and
used as a correction for each pre-concentration set. The
determined PGE concentrations are illustrated in Figure 1 as
a function of the prepared (calculated) concentration. The
slopes of the fitted lines indicate the recovery ratios of the
pre-concentration procedures, which are 93% ± 1%, 101% ±
1% and 111% ± 2% for Pt, Ir and Rh, respectively. The inter-
cepts of the fitted lines represent the average experimental
blank levels from ultrapure water, nitric acid (used for
sample acidification) and from the laboratory materials used
throughout the experimental procedures. The intercepts were
0.11 ± 0.11 fg mL−1, 0.10 ± 0.06 fg mL−1 and 0.26 ± 0.11 fg
mL−1 for Pt, Ir and Rh, respectively. The average relative
standard deviations of single measurements in pre-concen-
trated samples were 14% (0.4-52% for a concentration range
of 1-20 fg mL−1), 14% (0.2-67% for 0.1-20 fg mL−1) and
10% (1.3-34% for 0.1-10 fg mL−1) for Pt, Ir and Rh, respec-
tively. The pre-concentration reproducibility was estimated
based on the relative standard deviations of simultaneous
determinations for the same concentration. On average,
these deviations were as follows: 7% (3.5-10%) for a 1-20 fg
mL−1 concentration range (n = 6 for each concentration (3
samples × 2 measurements)) for Pt, 4% (0.9-15%) for a 0.1-
20 fg mL−1 concentration range (n = 6 (3 × 2)) for Ir and
15% (3.3-27%) for a 0.1-10 fg mL−1 concentration range (n
= 2 (1×2)) for Rh.
The collected results (i.e., recoveries, precisions and blank
levels) for Pt, Ir and Rh in this study are superior to those
obtained from previous studies performed using similar
methods for the determination of PGEs5 and other trace
metals.2
Checking PGE Contributions from Decontamination.
When snow and ice cores are used for reconstructing past
changes in atmospheric composition, decontamination of the
outside of the sample is mandatory because the outer layers
of the core samples are subject to contamination during the
Figure 1. Calibration of the non-boiling evaporation procedure for
(a) Pt, (b) Ir and (c) Rh using 60 mL of 0.1-20 fg mL
−
1
standard
solutions. Slopes of regression lines indicate recovery ratio, and
intercepts indicate experimental blanks.

2108 Bull. Korean Chem. Soc. 2011, Vol. 32, No. 6 Notes
drilling operation. Thus, evaluation of the efficiency and of
the blank contributions of the decontamination procedure is
absolutely necessary. For this purpose, we carried out a
complete analytical procedure, i.e., decontamination, pre-
concentration and instrumental analysis, on artificial ice core
prepared using Milli-Q water. Approximately 2 L of Milli-Q
water was frozen in a previously cleaned 2000 mL Teflon
vessel (Savillex, Minnetonka, USA) for ~48 h. Careful
elimination of the outside layer was then performed by
mechanically chiseling consecutive layers of ice in pro-
gression from the outside toward the center of the artificial
core under ultra-clean conditions. Detailed decontamination
procedures by mechanical chiseling have been reported
previously.6 Through this procedure, the clean inner core of
the artificial core section was obtained after three successive
layers were removed from the outside toward the center. To
confirm that the inner core was free from outside con-
tamination, all of the layers were analyzed for PGEs. A non-
boiling pre-concentration procedure was performed with a
concentration factor of 60 (for the first outer layer) and a
concentration factor of 120 (for the second and third layers
and the inner core). The Pt, Ir and Rh concentrations as a
function of the sample layer from the outside to the inside of
the core are shown in Figure 2. The PGE concentrations in
the outside layer (the first layer) were 2.7, 0.2 and 26 fg mL−1
for Pt, Ir and Rh, respectively. Interestingly, the outside of
the artificial core was found to be slightly contaminated with
the most pronounced contamination for Rh. This result
indicates that contamination of the samples could occur
during handling of the core sections. We then observed well-
established plateau values in the inner layers of the core for
the three PGEs, indicating that no outside PGE contamination
was present in the inner core as a result of the decontami-
nation procedure. The inner core concentrations were close
to or lower than the procedural detection limits of 0.43, 0.05
and 0.15 fg mL−1 for Pt, Ir and Rh, respectively.
Acknowledgments. This work was supported by the
National Research Foundation of Korea (NRF) and by
grants funded by the Korea government (MEST) (No. 2009-
0083899), a KOPRI research grant (PE11090) and an Inha
University research grant (INHA-40890-01). A Brain Korea
21 Graduate Student Fellowship supported T.-O. Soyol-Erdene.
References
1. Hong, S.; Lluberas, A.; Rodriguez, F. Korean J. Polar Res. 2000,
11, 35.
2. Görlach, U.; Boutron, C. F. Anal. Chim. Acta 1990, 236, 391.
3. Rodushkin, I.; Nordlund, P.; Engstro, E.; Baxter, D. C. J. Anal. At.
Spectrom. 2005, 20, 1250.
4. Soyol-Erdene, T.-O.; Han, Y.; Lee, B.; Huh, Y. Atmos. Environ.
2011, 45, 1970.
5. Gabrielli, P.; Varga, A.; Barbante, C.; Boutron, C.; Cozzi, G.;
Gaspari, V.; Planchon, F.; Cairns, W.; Hong, S.; Ferrari, C.;
Capodaglio, G. J. Anal. At. Spectrom. 2004, 19, 831.
6. Candelone, J.-P.; Hong, S.; F. Boutron, C. Anal. Chim. Acta 1994,
299, 9.
Figure 2. Changes in the concentrations of (a) Pt, (b) Ir and (c) Rh
in the sample layers (from outside) and inner core (i.c.) of the
artificial ice core.