Predictive factors for overall quality of life in patients with advanced cancer

Analytical note

Detection of iron species using inductively coupled

plasma mass spectrometry under cold plasma

temperature conditions

Li-Shing Huang

_{, King-Chuen Lin}

a

Department of Chemistry, National Taiwan Uni¨ersity, Taiwan

b

Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 106, Taiwan

Received 8 April 2000; accepted 7 October 2000

Abstract

Ž .

Under the conditions of low radio frequency rf power of 600 W and aerosol gas flow rate of 1.25᎐1.35 lrmin, 56

Ž 54 . q Ž q.

Fe or Fe ions can be detected from the isobaric interference of the ArO or ArN matrix. Using this 56

Ž .

method, the detection limit of Fe can reach 16 ngrl ppt , 60 times smaller than by normal plasma conditions at Ž .

1200 W rf power. The linear dynamic range of the analyte measurement extends to 1000 ngrml ppb . 䊚 2001 Elsevier Science B.V. All rights reserved.

Keywords: Inductively coupled plasma mass spectroscopy; Cold plasma; Iron; Isobaric interference

1. Introduction

In natural waters, the iron cycle involves numerous reactions of electron-transfer catalysis w1᎐3 . Iron detection is important to the researchx into atmospheric chemistry. However, the total iron content in the open ocean is -1 nM. For

U_{Corresponding author. Fax:q886-2-2362-1483.}

Ž .

E-mail address: kclin@mail.ch.ntu.edu.tw K. Lin .

determining such low limits of concentration, development of a sensitive analytical method becomes crucial. The present analytical methods for iron determination, and their limitations, have

w x recently been reviewed by Pehkonen 4 .

Among these methods, e.g. Johnson and co-workers have employed a flow-injection analysis ŽFIA coupled with chemiluminescence detection,. and successfully achieved a subnanomolar limit

Ž . w x

for Fe II detection in seawater 5 . King et al. Ž .

have determined Fe II at the nanomolar level 0584-8547r01r$ - see front matter 䊚 2001 Elsevier Science B.V. All rights reserved.

Ž .

with a spectrophotometric detector, using fer-Ž .

rozine FZ immobilized on a C18 Sep-Pak car-w x

tridge for pre-concentration 6 . In their treat-Ž .2q

ment the Fe FZ3 complex was not separated from the excess FZ, the spectral absorption of FZ at the same wavelength as the complex might cause significant error to a quantitative

measure-Ž .

ment of Fe II species. Brown and co-workers have employed a high-performance liquid chro-Ž .2q matographic method to separate the Fe FZ₃ complex from ferrozine and other contaminants, which were then monitored with a UV-vis

absorp-w x

tion detector at 254 nm 7 . A detection limit of

y10 _{Ž .}

10 M was achieved, when the Fe II species in rainwater or seawater was pre-concentrated 50᎐100 times with a C Sep-Pak cartridge.₁₈

The inductively coupled plasma mass

spec-Ž .

trometer ICP-MS is an alternative sensitive detector widely adopted in elemental trace analy-sis. ICP-MS provides advantages of simultaneous Žor sequential. multi-elemental analysis and extremely sensitive detection for the charged ana-lyte species. The iron has a small ionization potential of 7.86 eV and its ionization efficiency is estimated to be 96% based on the Saha equation w x8,9 . ICP-MS appears to be a powerful tool for iron detection. However, the analyte signal is superimposed on top of the 40_ArOq _background noise, and thereby the analytical sensitivity is considerably reduced. For instance, Uchida and

Ž Ito obtained the iron detection limit of 2 ppb or

. Ž

1 ppb with Ar ICP-MS or air᎐Ar mixed ICP-.

MS . These results were higher than those values

w x q

determined for other elements 10 . Like ArO , the polyatomic matrix ions inherent in ICP-MS have been a severe interference to trace element analysis. In addition to the increased background noise, the resulting space charge effect may reduce the fraction of sample ions passing through

w x

the ion optics toward the detector 11,12 . Thus far, several methods have been demon-strated in an attempt to eliminate these poly-atomic matrix ions. E.g. with the use of mem-brane desolvation or cryogenic desolvation, the water may be removed prior to entering the ICP

w x

torch 13,14 . The ICP may also be coupled with an electrothermal vaporization system, which enables the matrix to be removed prior to the

w x

sample vaporization 15 . The addition of mixed gases into Ar ICP is an alternative method. The added gases such as N , Xe, and H may attenu-_{2 2}

w x

ate the polyatomic ions 16,17 . It has already been

determined that the secondary discharge between the plasma and the grounded sampler cone can enhance the formation of some polyatomic ions such as ArOq, ArHq and other Ar-containing

w x

ions 11 . Therefore, insertion of a grounded metal plate between the outer wall of the torch and the load coil may effectively diminish the secondary discharge, and in turn, substantially reduce the

w x

interference of these matrix ions 18,19 . The Ž .

decrease of radio frequency rf forward power has led to similar effects in reducing the

Ar-con-w x

taining ions 20,21 . Under the cold plasma condi-tions, Houk et al. have reported that the Arqand ArHq ions may be suppressed and thus the obtained measurement of potassium isotope ratio of mrz 39 and 41 is precise with a R.S.D. of

w x 0.3᎐0.9% 20 .

In this work, we have employed an Ar ICP-MS to detect iron in the aqueous solution. The ICP-MS is performed at low rf power and large aero-sol gas flow rate. Under the cold plasma

condi-56 _Ž 54 _.

tions, the intensity ratio of Fe or Fe ion to

q _Ž q_.

ArO or ArN has been characterized. The Ž . dynamic range of a calibration curve for Fe III

Ž .

concentration is obtained. The Fe III can be monitored with a detection limit lower by approx-imately two orders of magnitude than the normal plasma conditions.

2. Experimental

2.1. ICP-MS apparatus

Ž A SCIEX ELAN Model 6000 ICP-MS

Perkin-.

Elmer was used for all data acquisition. It was run in sequential mode, ‘peak hopping’ to masses of interest. A cross-flow nebulizer with a Scott-type double path spray chamber was used. To avoid the conductively coupling between the load coil and the plasma, both ends of the load coil were biased with a high voltage of equal ampli-tude but opposite phase. The plasma potential

may then be minimized. The sampling depth between the sampler tip and the top coil was fixed at 9 mm for all data acquisition. The coolant gas flow rate and the auxiliary gas flow rate were fixed at 15.0 and 1.2 lrmin of Ar throughout the experiment, respectively, while the aerosol gas flow rate was varied. In this work the rf power was varied from 600 to 1200 W, taking advantage of ‘plasmalok’ provided by SCIEX ELAN 6000. Even under the cold plasma operating conditions, the ICP remained very stable.

2.2. Reagents

Ž .

Twice deionized water Millipore was used for

Ž .

the reagent preparation. Methanol HPLC grade was purchased from Mallinckrodt. A 100 ppm Fe3q solution was prepared from the

Ž . Ž .

NH Fe SO4 4 2⭈12H O sample Merck and then2 diluted to various concentrations in 0.1% HNO3 ŽMerck solution..

3. Results and discussion

3.1.56_{FerArO and} 54_{FerArN ratios}

To optimize the ratio of 56_{FerArO, the rf} power, aerosol gas flow rate and potential biased on the ion lens were varied, while the other parameters remained fixed. The coolant gas and auxiliary gas flow rate were fixed at 15 and 1.2 lrmin, respectively. The aerosol gas flow rate was varied from 0.7 to 1.40 lrmin at a step of 0.05 lrmin, while rf power was varied from 600 to 1400 W at a step of 100 W at a fixed potential on the ion lens. The background counts per second Žcps of ArO. qand ArNqin a blank solution were measured, as shown in Fig. 1. Data were averaged

Ž . over five replicates. Similarly, 20 ppb Fe III solu-tion was measured at mrz 56 and 54.

In Fig. 1, it is shown that either decreasing the rf power or increasing the aerosol gas flow rate may effectively reduce the cps of the ArOq and ArNq background ions. However, lowering the plasma temperature similarly reduces the ratio of metal ion to its neutral atom based on the Saha equation. The obtained cps of the Fe ion

de-Ž . qŽ . q

Fig. 1. a rf power dependence of ArO mrz 56 and ArN

Žmrz 54 interference in the blank measurement at an aerosol. Ž .

gas flow rate 1.3 lrmin. b Aerosol gas flow rate dependence

q _{Ž .} q _{Ž .}

of ArO mrz 56 and ArN mrz 54 interference in the

blank measurement at a rf power 600 W.

creases also. As shown in Fig. 2, the net cps of the q _Ž Fe ion is estimated by subtracting the ArO or

q_.

ArN background from the total cps of the

Ž . Ž .

signal at mrz 56 or 54 for the 20 ppb Fe III

56 _Ž

solution. Thus, the ratio of FerArO or 54_{FerArN may be determined as a function of rf.} power and aerosol gas flow rate. The results are plotted in Fig. 3. The ratios tend to increase with a low rf power and a large gas flow rate. In this experiment, the ratio is optimized at a rf power 600 W and aerosol gas flow rate 1.25᎐1.35 lrmin. For the rf power at 600 W and the aerosol gas flow rate at 1.35 lrmin, the ratio of 56Fe ion to ArOq is plotted as a function of the biased po-tential. The optimized voltage is found at 3 V, smaller than that under the normal plasma condi-tions. The factor of biased voltage on the ion lens may be used to qualitatively evaluate the kinetic

Ž . 56 54

Fig. 2. a rf power dependence of Fe and Fe signal in the Ž . 20 ppb solution at an aerosol gas flow rate 1.3 lrmin. b Aerosol gas flow rate dependence of 56_{Fe and}54_{Fe signal in}

the 20 ppb solution at a rf power 600 W.

energy of the ion species. The optimal biased voltage is, therefore, expected to be smaller under the cold plasma condition.

Due to the isobaric interference by ArOq and ArNq, the detection limits of 56_{Fe and}54_{Fe with} ICP-MS are higher than those determined for

w x q q

most elements 10 . The ArO and ArN with weak bond energies are formed predominantly behind the sampler cone. It is found that the secondary discharge produced between the posi-tive plasma potential and the grounded sampler may enhance the ArOqand ArNqformations via

w x

collision-induced chemical reaction 18,19 . Low-ering the plasma temperature is one of the effec-tive methods to remove the secondary discharge w20,21 , and subsequently eliminate these back-x ground interferences. However, one should note that the enhancement of the ratio of sample signal to matrix background is closely related to

Ž .

the ionization potential I.P. of the sample. The Fe ion with a small I.P. decays more slowly than the matrix ions, as the plasma temperature de-creases. Tanner has suggested that the cold plasma technique is most useful for elements having lower ionization potentials, but its sensitiv-ity decreases markedly as the I.P. is above 8 eV w x21 .

3.2. Detection limit and dynamic range

Under the conditions of rf power at 600 W and ion lens biased at 3.0 V, the 56_Fe3q _{ion in 0.1%} HNO solution was detected in the concentration₃ range from 0.05 to 1000 ppb. The calibration curve is plotted in Fig. 4. The detection limit is defined as three times the standard deviation

Ž . 56

Fig. 3. a rf power dependence of the ratios of FerArO and 54_{FerArN in the 20 ppb solution at an aerosol gas flow}

Ž .

rate 1.3 lrmin. b Aerosol gas flow rate dependence of the ratios of56_{FerArO and}54_{FerArN in the 20 ppb solution at a}

Table 1

Comparison of detection limits obtained between cold plasma and normal plasma conditions

Cold plasma Normal plasma

Ž .

RF power W 600 600 1200

Ž .

Ion lens voltage V 600 3.0 7.0

Ž .

Aerosol gas flow rate lrmin 1.30 1.35 1.05

Number of replicates 11 11 11

2 1 6

Ž .

Mean blank signal cps 1.60=10 5.09=10 3.44=10

1 4

Ž .

Standard deviation of blank signals cps 1.10=10 3.16 1.03=10

2 2 4

Ž .

Slope of calibration curve cpsrppb 9.27=10 6.05=10 3.44=10

y2 y2 y1

Ž .

Detection limit ppb 3.6₌₁₀ 1.6₌₁₀ 9.0₌₁₀

Ž3␴ of a blank measurement. Given the slope of. the calibration curve and ␴ value in Table 1, the

Ž . detection limit reaches 16 and 36 ngrl ppt at flow rate of 1.35 and 1.30 lrmin, respectively. The dynamic range of the analyte measurement extends to at least 1000 ppb, covering a very wide concentration range. When the rf power was changed to 1000 W, a normal ICP operating condition, the detection limit of 56_{Fe estimated} similarly gave a value of 0.90 ppb, approximately 60 times larger than that under the cold plasma condition. Comparison of the data is also listed in Table 1.

4. Conclusion

q _Ž q_.

Suffered from the ArO or ArN isobaric

56 _Ž 54 _.

interference, the trace Fe or Fe analysis by

Ž .

Fig. 4. Calibration curve for Fe III standard solution.

the ICP-MS is a difficult task without the use of a w x

high resolution mass spectrometer 22 . Our work provides an alternative method to minimize the

q _Ž q_.

ArO or ArN interference by using a cold plasma technique. In this work, we have

charac-56

Ž 54 . q

terized the ratio of Fe or Fe ion to ArO Žor ArNq. as a function of rf power and aerosol gas flow rate and obtained the linear dynamic

Ž . range of a calibration curve for the Fe III con-centration up to 1000 ppb. The detection limit of iron can reach 16 ppt, approximately 60 times lower than that by the normal plasma operating conditions. The ICP-MS with the use of cold plasma conditions may apparently provide a promising tool to detect the trace iron contents in seawater.

Acknowledgements

This work was supported by the National Sci-ence Council of the Republic of China under the Contract NSC 89-2113-M-002-027.

References

w x1 T.E. Graedel, C.J. Weschler, M.L. Mandich, Influence of transition metal complexes on atmospheric droplet

Ž .

acidity, Nature 317 1985 240᎐242.

w x2 G. Zhuang, Y. Zhen, R.A. Duce, P.R. Brown, Iron photoreduction and oxidation in an acidic mountain

Ž .

stream, Nature 355 1992 537᎐539.

w x3 P. Behra, L. Sigg, Evidence for redox cycling of iron in Ž .

atmospheric water droplets, Nature 344 1990 419᎐421. w x4 S. Pehkonen, Determination of the oxidation states of

Ž .

iron in natural waters, Analyst 120 1995 2655_᎐2663. w x5 V.A. Elrod, K.S. Johnson, K.H. Coale, Determination of

Ž .

subnanomolar levels of iron II and total dissolved iron in seawater by flow injection analysis with

chemilumi-Ž .

nescence detection, Anal. Chem. 63 1991 893᎐898. w x6 D.W. King, J. Lin, D.R. Kester, Spectrophotometric

Ž .

determination of iron II in seawater at nanomolar con-Ž .

centrations, Anal. Chim. Acta 247 1991 125᎐132. w x7 Z. Yi, G. Zhuang, P.R. Brown, R.A. Duce,

High-perfor-mance liquid chromatographic method for the determi-Ž .

nation of ultratrace amounts of iron II in aerosols, Ž . rainwater, and seawater, Anal. Chem. 64 1992 2826_᎐2830.

w x8 R.S. Houk, Mass spectrometry of inductively coupled Ž .

plasmas, Anal. Chem. 58 1986 97A᎐105A.

w x9 J.W. Olesik, Elemental analysis using ICP-OES and ICPrMS, an evaluation and assessment of remaining

Ž .

problems, Anal. Chem. 63 1991 12A᎐21A.

w x10 H. Uchida, T. Ito, Evaluation of an inductively coupled air-argon plasma as an ion source for mass

spectrome-Ž .

try, J. Anal. At. Spectrom. 12 1997 913᎐918.

w x11 H. Niu, R.S. Houk, Fundamental aspects of ion extrac-tion in inductively coupled plasma mass spectrometry,

Ž .

Spectrochim. Acta Part B 51 1996 779᎐815.

w x12 S.D. Tanner, Space charge in ICP-MS: calculation and Ž . implications, Spectrochim. Acta Part B 47 1992 809᎐823.

w x13 I.B. Brenner, A. Zander, M. Plantz, J. Zhu, Characteri-zation of an ultrasonic nebulizer-membrane separation interface with inductively coupled plasma mass spec-trometry for the determination of trace elements by Ž . solvent extraction, J. Anal. At. Spectrom. 12 1997 273_᎐279.

w x14 L.C. Alves, L.A. Allen, R.S. Houk, Measurement of vanadium, nickel, and arsenic in seawater and urine reference materials by inductively coupled plasma mass

spectrometry with cryogenic desolvation, Anal. Chem. Ž .

65 1993 2468᎐2471.

w x15 J.P. Byrne, D.M. Hughes, C.L. Chakrabarti, D.C. Gre-goire, Mechanism of volatilization of tungsten in the graphite furnace investigated by electrothermal vapor-ization inductively coupled plasma mass spectrometry, J.

Ž . Anal. At. Spectrom. 9 1994 913᎐917.

w x16 D.A. Wilson, G.H. Vickers, G.M. Hieftje, Use of the microwave-induced nitrogen discharge at atmospheric pressure as an ion source for elemental mass

spectrome-Ž .

try, Anal. Chem. 59 1987 1664᎐1670.

w x17 J.W. Lam, G Horlick, A comparison of argon and mixed gas plasmas for inductively coupled plasma-mass

spec-Ž .

trometry, Spectrochim. Acta Part B 45 1990 1313᎐1325. w x18 A.L. Gray, Influence of load coil geometry on oxide and doubly charged ion response in inductively coupled plasma source mass spectrometry, J. Anal. At. Spectrom.

Ž .

1 1986 247᎐249.

w x19 N.S. Nonose, N. Matsuda, N. Fudagawa, M. Kubota, Some characteristics of polyatomic ion spectra on induc-tively coupled plasma mass spectrometry, Spectrochim.

Ž . Acta Part B 49 1994 955᎐974.

w x20 S.J. Jiang, R.S. Houk, M.A. Stevens, Alleviation of over-lap interferences for determination of potassium isotope ratios by inductively coupled plasma spectrometry, Anal.

Ž .

Chem. 60 1988 1217᎐1221.

w x21 S.D. Tanner, Characterization of ionization and matrix suppression in inductively coupled ‘cold’ plasma mass

Ž .

spectrometry, J. Anal. At. Spectrom. 10 1995 905᎐921. w x22 J. Wu, E.A. Boyle, Determination of iron in seawater by high-resolution isotope dilution inductively coupled

Ž .

plasma mass spectrometry after Mg OH 2 coprecipita-Ž .