Inductively Coupled Plasma Mass Spectrometry with On-Line

Anal. Chem. 2002, 74, 3924-3928

Inductively Coupled Plasma Mass Spectrometry with On-Line Leaching: A Method To Assess the Mobility and Fractionation of Elements Diane Beauchemin,*,† Kurt Kyser,‡ and Don Chipley‡

Departments of Chemistry and Geological Sciences, Queen’s University, Kingston, Ontario K7L 3N6, Canada

A new technique has been developed to assess the mobility and site of specific elements in complex natural materials such as rocks. Concentration profiles during leaching were obtained by pumping reagents (water, 1% HNO3, 10% HNO3, 30% HNO3), either continuously or with flow injection, through a microcolumn of sample while continuously monitoring analyte signals by inductively coupled plasma mass spectrometry (ICPMS). Compared to batch extraction procedures normally used, the approach involves minimal sample preparation and reduced contamination since the leaching is performed in a closed system. Continuous on-line monitoring also allows a greater resolution of the various phases reacting with given reagent. Compared to continuous leaching, flow injection increased the resolution of the various phases using discrete injections of reagents while reducing reagent consumption and minimizing etching of the MS interface. Furthermore, sensitivity was preserved by injecting into air instead of an aqueous carrier. Whether in the continuous or flow injection modes, the proposed approach provides real-time data on what phases are breaking down and what metals are released. It can therefore be used to design effective leaching strategies and to trace isotopic compositions. However, the resulting spectra are complex and the correct determination of some elements requires high-resolution ICPMS. Information on the chemical speciation of analytes in soils and sediments is required in many instances, such as for risk assessment of contaminants, since it is linked to the mobility and bioavailability of elements. An example is the extent of metal leaching that can occur when dredged masses of sediments are dumped on the land and exposed to air.1 However, the determination of the chemical form of many environmentally important elements is not generally possible.2 For this reason, operational methods are typically used to indicate the fraction of analyte that is potentially labile under a given set of environmental conditions. Although the recommended sequential extraction scheme (BCR * To whom correspondence should be addressed: (e-mail) beauchmn@ chem.queensu.ca; (fax) 613-533-2619. † Department of Chemistry. ‡ Department of Geological Sciences. (1) Peltola, P.; Astrom, M. Sci. Total Environ. 2002, 284, 109-122. (2) Howe, S. E.; Davidson, C. M.; McCarney, M. J. Anal. At. Spectrom. 1999, 14, 163-8.

3924 Analytical Chemistry, Vol. 74, No. 15, August 1, 2002

procedure) has been demonstrated to be reproducible enough for the fractionation of trace metals in environmental impact studies,3 it is both time-consuming and combines information from different phases reacting together. For instance, the first step involves overnight extraction with 0.11 M CH3COOH to obtain the combined exchangeable, water-soluble and acid-soluble phases. Although the time required for the procedure can be significantly reduced by using microwave single extractions,4,5 this batch approach does not distinguish the different phases responsible for the release of elements by a given reagent. This is also true of other modified sequential extraction procedures such as those optimized for a sample type,6 a specific phase,7 or a specific analyte.8 The purpose of this work was to develop a simple and rapid technique to better constrain the mobility and site of specific elements in complex minerals, soils, and sediments. Indeed, the rate of release of elements depends on the sites in which they are held and their solubility in the leach medium. For example, elements adhering to the surface are released much more readily than those in minerals. Therefore, concentration profiles obtained during leaching can distinguish elements released from distinct sources. This is demonstrated here using a sedimentary rock sample. EXPERIMENTAL SECTION Microcolumn Preparation. Two different approaches were explored, without and with flow injection. Both involved the use of a microcolumn of sample that was prepared by loosely packing a small amount (25 mg) in a 3/16-in.-o.d., 1/8-in.-i.d. PTFE tube, between quartz wool plugs. The packing was loose enough so that the original structure of the sample (macropores, etc.) would not be changed. Reagents. The 0.5-1.4-mm size fraction of a sandstone sample from a drill core in the Athabasca Basin of Canada was selected (3) Marin, B.; Valladon, M.; Polve, M.; Monaco, A. Anal. Chim. Acta 1997, 342, 91-112. (4) Pe´rez Cid, B.; Ferna´ndez Albore´s, A.; Ferna´ndez Go´mez, E.; Falque´ Lo´pez, E. Anal. Chim. Acta 2001, 431, 209-218. (5) Pe´rez Cid, B.; Ferna´ndez Albore´s, A.; Ferna´ndez Go´mez, E.; Falque´ Lo´pez, E. Analyst 2001, 126, 1304-1311. (6) Go´mez Ariza, J. L.; Gira´ldez, I.; Sa´nchez-Rodas, D.; Morales, E. Anal. Chim. Acta 2000, 414, 151-164. (7) Krasodebska-Ostrega, B.; Emons, H.; Golimowski, J. Fresenius’ J. Anal. Chem. 2001, 371, 385-390. (8) Wenzel, W. W.; Kirchbaumer, N.; Prohaska, T.; Singeder, G.; Lombi, E.; Adriano, D. C. Anal. Chim. Acta 2001, 436, 309-323. 10.1021/ac025671p CCC: $22.00

© 2002 American Chemical Society Published on Web 06/20/2002

Table 1. Operating Conditions ICPMS instrument UltraMass 700 Ar plasma gas flow rate (L/min) Ar auxiliary gas flow rate (L/min) Ar nebulizer gas flow rate (L/min) sampling depth (mm) rf power (kW) sample uptake rate

15 1.05 0.83-0.90 5.0 1.2 1.3 mL/min

Element 14 0.95 0.90 1.25 100 µL/min

to demonstrate the utility of the procedure. The sample was taken 10 km from the Cigar Lake uranium deposit and consists of quartz, illite, kaolinite, and hematite as major minerals. This sample was chosen because it is near a known metal deposit, is well characterized in terms of mineral chemistry and petrography, has a relatively simple mineralogy, and may have isotopically distinct Pb isotopes that have migrated from the uranium deposit. Reagents (water, 1% HNO3, 10% HNO3, 30% HNO3) were prepared with doubly deionized water (DDW) (18 MΩ cm-1, Millipore, Bedford, MA) and high-purity OmniTrace nitric acid (69.0-70.0%) (EM Science, Gibbstown, NJ). Only nitric acid was considered in this preliminary study because it does not contribute additional spectroscopic interferences to the background spectrum in quadrupole-based ICPMS (Q-ICPMS) and is more selective than HCl as a leaching agent.9 Instrumentation. Q-ICPMS (UltraMass 700, Varian Australia Pty Ltd.) with a standard concentric nebulizer and SturmanMasters spray chamber was used throughout this work with the operating conditions in Table 1. Data were acquired by peak hopping in the time-resolved acquisition mode with 1 point/peak, 1 scan/replicate, and 0.025 amu spacing. The raw time-resolved spectra were then exported and treated with in-house BASIC software for smoothing (with a 7-point Savitzy-Golay polynomial moving window) and peak area determination. Double-focusing high-resolution ICPMS (HR-ICPMS) (Element, ThermoFinnigan) was also used to monitor continuous online leaching. A 0.9 or 1.6 mL/min concentric nebulizer (Meinhard, ThermoFinnigan) was then used to allow multielement analysis despite the slower scanning rate inherent to HR-ICPMS, with the operating conditions listed in Table 1. Data acquisition was performed using a 1.5-ms scan time with 70 runs in two passes and an integration window of 50, at three different resolution settings: 300 (LR), 4000 (MR), and 9600 (HR). The number of samples/peak was 10 for LR and 15 for MR and HR. Continuous Leaching Setup. Continuous leaching involves the insertion of the microcolumn downstream of the peristaltic pump so that reagents (water, 1% HNO3, 10% HNO3, 30% HNO3) may then be continuously pumped sequentially through it. This approach was used with HR-ICPMS. A modified one was used with Q-ICPMS, which allows flushing of the sample introduction system and MS interface with DDW between reagents. It corresponds to that shown in Figure 1 without a reagent injection valve. In either case, leaching profiles (see Figure 2, for example) were readily obtained by continuously monitoring analyte signals by ICPMS. (9) Ochsenku ¨ hn-Petropulu, M.; Lyberopulu, Th.; Ochsenku ¨ hn, K. M.; Parissakis, G. Anal. Chim. Acta 1996, 319, 249-254.

Figure 1. Schematic diagram of the flow injection manifold used for on-line leaching of a soil sample.

Figure 2. Analyte signals observed by Q-ICPMS during the continuous sequential leaches of 25 mg of sandstone sample by DDW, 1% HNO3, 10% HNO3, and 30% HNO3.

Flow Injection Leaching Setup. Flow injection (FI) was also considered because of its numerous advantages;10 particularly discrete injections of reagents may increase the resolution of the various phases being dissolved. Furthermore, FI reduces etching of the MS interface by reagents. Figure 1 shows the setup that was used for this purpose. It involves the addition of two electronically actuated valves (Universal module, Anachem, Luton, U.K.). One, with a 100-µL loop, served to inject the reagent in the air carrier, whereas the other served as a bypass, which can be used to flush the sample introduction system and, by extension, the MS interface between injections of reagents. It also allows the aspiration of the solution used for optimization of the ICPMS instrument without having to remove the soil column. The use of air as a carrier of the discrete reagent plugs corresponds to the discontinuous batch mode (i.e., the sample is in contact with air (10) Beauchemin, D. Discrete Sample Introduction Techniques for Inductively Coupled Plasma Mass Spectrometry; Comprehensive Analytical Chemistry 34; Elsevier: Amsterdam, 2000; Chapter 2, pp 213-346.

Analytical Chemistry, Vol. 74, No. 15, August 1, 2002

3925

Table 2. 207Pb/206Pb Ratio (( Standard Deviation) Measured by Q-ICPMS in Each Reagent after Passage through a Sandstone Sample 207Pb/206Pb

reagent

in continuous mode

in FI mode

water 1% HNO3 10% HNO3 30% HNO3

0.63 ( 0.09 (n ) 113) 0.61 ( 0.14 (n ) 166) 0.46 ( 0.10 (n ) 150) 0.46 ( 0.05 (n ) 126)

0.65 ( 0.09 (n ) 53) 0.61 ( 0.12 (n ) 123) 0.47 ( 0.10 (n ) 73) 0.48 ( 0.10 (n ) 340)

between extractions with different reagents). However, it has the additional advantage of preserving sensitivity. Indeed, the signal obtained by injecting 100 µL of a solution into air is similar to that obtained by continuous nebulization of the same solution.11 RESULTS AND DISCUSSION Continuous Leaching. Comparisons of the profiles of analytes to those of matrix elements simultaneously monitored during continuous leaching and Q-ICPMS indicate which elements are released by the dissolution of which minerals. For example, in Figure 2, Co clearly follows the same profile as Ca, which suggests that it is released by the dissolution of a Ca-containing salt. Furthermore, the fact that so much Co is released with DDW indicates that this element is in a mobile form. Although Figure 2 only shows the profiles for three analytes, numerous others were simultaneously monitored. Still, the Co profile, i.e., a decrease in the amount of leached analyte upon an increase in HNO3 reagent concentration, was observed for several other elements (Ni, Cu, Sr, Zn, Cd, Cr, V, Mn). However, some elements (Pb, As, Th, Sb, Mo, U) behaved differently since an increase in the amount of leached analyte was observed upon changing from 1 to 10 and 30% HNO3. It should be noted that, although the absolute signals varied between 25-mg soil replicates, as expected with small amounts of coarse samples,12 the trends were reproducible. Monitoring several isotopes of an analyte whose isotopic composition varies in nature further helped in discriminating different sources of this analyte. For example, Table 2 shows that 207Pb/206Pb of the Pb released in water and 1% HNO is clearly 3 different than that released in 10 and 30% HNO3, indicating two different sources of Pb. These ratios were obtained by averaging the point-by-point ratios (using Microsoft Excel) over the profiles. Although the fast scanning rate of Q-ICPMS allows the simultaneous determination of a large number of elements, the detection of some elements (such as Fe in Figure 2) requires greater mass resolution. All Fe isotopes suffer from spectroscopic interferences from polyatomic ions; i.e., 40Ar14N interferes with 54Fe whereas ArO, CaO, and CaOH interfere with 56Fe, 57Fe, and 58Fe. ICPMS with either a collision cell or HR capability may be used for verification purposes. In this case, HR-ICPMS was used to confirm the presence of Fe. In fact, it revealed (Figure 3) two different sources of Fe: one readily soluble in water and dilute HNO3 and one requiring more concentrated HNO3. Since the pattern was very similar to that observed for Pb, these data are consistent with Pb residing in two different iron oxides (such as easily soluble and more crystalline forms). In addition, the (11) Craig, J. M.; Beauchemin, D. Analyst 1994, 119, 1677-1682. (12) Gardner, M. J. Anal. Proc. 1995, 32, 115-116.

3926 Analytical Chemistry, Vol. 74, No. 15, August 1, 2002

Figure 3. Analyte signals observed by HR-ICPMS during the continuous sequential leaches of 25 mg of sandstone sample by 1% HNO3, 10% HNO3, and 30% HNO3.

Figure 4. Comparison of the concentrations measured by a single 2-h batch extraction with 2% HNO3 and those obtained after on-line sequential leaches.

determination of additional elements such as S, P (Figure 3), halides, and PGEs are facilitated with HR-ICPMS. Nonetheless, Q-ICPMS is well suited for screening samples since it allows the continuous monitoring of a large number of analytes without significantly increasing the analysis time. Indeed, in HR-ICPMS, a micronebulizer is required if a large number of elements is to be monitored, which further increases the analysis time that is already inherently longer than that with a quadrupole because of the slower scanning rate of HR-ICPMS. For example, the total analysis time for the four sequential leaches was less than 20 min with Q-ICPMS as opposed to 80 min with HR-ICPMS. Compared to batch extraction procedures, the on-line leach procedure involves minimal time for sample preparation, with reduced contamination since the leaching is performed in a closed system. Continuous on-line monitoring also allows a greater resolution of the various phases released by a given reagent. For example, the concentrations measured after a single 2-h batch leaching with 2% HNO3 are qualitatively similar to the sum of the four on-line extractions (Figure 4). Therefore, information on the two different sources of Pb and Fe so evident from the on-line leach was lost by the batch extraction. Because the several phases that are dissolved in a batch process are sequentially dissolved by on-line continuous leach, the latter provides more information on the different sources of analytes in a given sample. Furthermore, since typical sequential extraction procedures use longer

Figure 5. Analyte signals observed during the FI sequential leaches of 25 mg of sandstone sample by DDW, 1% HNO3, 10% HNO3, and 30% HNO3. Several 100-µL injections of each reagent were made.

extraction times, even more fractionation information is lost. Even with HR-ICPMS, the contact time of a reagent with a sample is reduced compared to batch approaches. The continuous monitoring inherent to the proposed approach allows not only the observation of the dissolution of different phases/minerals as they occur but also an indication of the identity of these phases/ minerals since matrix elements are concurrently monitored. This in fact is another major drawback of batch sequential extraction procedures wherein certain phases are assumed to be dissolved by a given extraction step but not generally confirmed by concurrently monitoring matrix elements. Another advantage of the proposed approach is that reagent consumption is reduced since, with on-line monitoring, a switch to a different reagent can be made as soon as analyte signal returns to baseline. The length of a speciation experiment is therefore minimized using this approach. Although only four reagents were used here to demonstrate the concept, several other reagents such as ammonium hydroxide and hydrogen peroxide could also be used without complicating the Q-ICPMS background spectrum. Ultimately, a series of reagents could be selected to achieve the sequential dissolution of the various phases until the whole sample is dissolved. Flow Injection Leaching. Although absolute signals varied, the trends observed in the FI mode with Q-ICPMS were very similar to those seen in the continuous mode (compare Figure 5 to Figure 2). Some irregularities in the FI traces were sometimes observed as a result of tiny air bubbles within the injected reagent plug. In any case, the 207Pb/206Pb average ratios in each reagent are identical to those measured in the continuous mode (Table 2). However, less than 1 mL of each reagent was required in the FI mode compared to over 2 mL in the continuous mode, because a higher resolution of phases was achieved by continuously monitoring individual 100-µL fractions. This is also in contrast to the continuous mode with HR-ICPMS where each point in Figure 3 corresponds to the integration of signal generated by 100 µL of reagent. The manifold also allowed flushing the nebulization system with water between injections to minimize corrosion of the sampling cones by the reagent. These advantages are in

addition to those obtained by on-line leaching in the continuous mode. Some experiments were scaled up to assess the effect of sample size. However, with 250-mg sample and 500-µL injections of reagents, trends were similar to those observed with 25-mg samples, at the expense of a longer analysis time. Furthermore, significant back-pressure problems were experienced with the resulting longer soil column. On the other hand, smaller injection volumes (50 µL) were also tried with the 25-mg samples to determine whether even more information on the various components/phases would be obtained, but the number of injections required with a given reagent simply doubled while the profiles remained similar. Therefore, 25-mg samples with 100-µL reagent injections were used for the remainder of the work. The above findings depend on the sample homogeneity and the surface-tovolume ratio of the sample; different results may be obtained with other sample types. Whether in the continuous or FI mode, calibration was best accomplished using standards prepared in each reagent matrix. A significant suppression was indeed observed upon increasing the acid concentration, which could not be adequately compensated by a single internal standard since it was mass dependent (heavy analytes did not experience as much suppression as lighter ones) as can be seen in Figure 6. In any case, this work demonstrates the many advantages of on-line leach with ICPMS, whether reagents are pumped continuously or discretely injected. Because continuous on-line monitoring allows a greater resolution of the various phases, it provides realtime data on what phases are breaking down, what metals are released, and with whom (i.e., cations and anions). Furthermore, sample preparation is simple and minimal, with reduced contamination. It can therefore be used to design effective leaching strategies and to trace isotopic compositions. It has the potential to become an invaluable asset for environmental studies on soils and sediments, as well as for mining companies in search of ore deposits. The flow injection mode may prove extremely advantageous for the kinetic-based analysis of soils and sediments. For example, in contrast to a method involving the determination of Analytical Chemistry, Vol. 74, No. 15, August 1, 2002

3927

Figure 6. Analyte signals observed for replicate 100-µL injections of 100 µg/L standard solutions prepared in 1% HNO3, 10% HNO3, and 30% HNO3.

aluminum fractionation within 2 h of soil extraction,13 the proposed approach allows the continuous monitoring of species within 20 s of their being directly released from the sample by a given reagent. Furthermore, since air-drying14,15 has been reported to affect the speciation of some elements, the microcolumn approach could be taken to the field. Indeed, microcolumn could be filled with wet soil samples on site, then sealed, and taken back to the laboratory for on-line hookup to ICPMS. However, the resulting spectra are complex. The correct determination of some elements requires HR-ICPMS. Future work will consider the applications to geochemistry and confirm which (13) Powell, K. J. Analyst 1998, 123, 797-802. (14) Wang, Z.; Shan, X.-Q.; Zhang, S. Anal. Chim. Acta 2001, 441, 147-156. (15) Bunzl, K.; Schimmack, W.; Schramel, P.; Suomela, M. Analyst 1999, 124, 1383-1387. (16) Al-Merey, R.; Al-Masri, M. S.; Bozou, R. Anal. Chim. Acta 2002, 452, 143148. (17) Marı´n, A.; Lo´pez-Gonza´lvez; A.; Barbas, C. Anal. Chim. Acta 2001, 442, 305-318. (18) Gu ¨ rleyu ¨ k, H.; Tyson; J. F.; Uden, P. C. Spectrochim. Acta Part B 2000, 55, 933-940.

3928

Analytical Chemistry, Vol. 74, No. 15, August 1, 2002

phase(s) is(are) released by a given reagent. The results for various soil samples from several known sites will be examined to identify pathfinder elements. A long-term goal for the application of this approach is to determine what elements reflect ore deposits and where they reside in anomalous and normal soils and rocks, in both residual and undercover terrains. In addition, efforts will be made to speed up the release of analytes in a given reagent and assist in the release of more refractory phases through the use of ultrasound16,17 or microwave energy.18 ACKNOWLEDGMENT D.B. is grateful to the Natural Sciences and Engineering Research Council of Canada (NSERC) for research funding and to Varian Australia Pty. Ltd. for a contribution to her salary during her academic leave. K.K. is most grateful for funding from NSERC and additional support from Cameco Corp. Received for review April 1, 2002. Accepted May 21, 2002. AC025671P