Direct injection high efficiency nebulizer-inductively coupled plasma mass spectrometry for analysis of petroleum samples

Spectrochimica Acta Part B 58 (2003) 397–413

Direct injection high efficiency nebulizer-inductively coupled plasma mass spectrometry for analysis of petroleum samples Kaveh Kahen, Adelitza Strubinger1, Jose´ R. Chirinos2, Akbar Montaser* Department of Chemistry, George Washington University, Washington, DC 20052, USA Received 24 September 2002; accepted 13 November 2002

Abstract Direct injection high efficiency nebulizer (DIHEN)-inductively coupled plasma mass spectrometry (ICPMS) is investigated for analysis of petroleum samples dissolved in volatile organic solvents. To minimize solvent loading, the solution uptake rate is reduced to 10 mlymin, far less than the level (85 mlymin) commonly used for aqueous sample introduction with the DIHEN, and oxygen is added to the nebulizer gas flow and outer flow of the ICP. Factorial design is applied to investigate the effect of nebulizer tip position within the torch and the nebulizer and intermediate gas flow rates on the precision and the net signal intensity of the elements tested for multielemental analysis. Cluster analysis and principal component analysis are performed to distinguish the behavior of different isotopes, oxide species and doubly charged ions. The best operating conditions at a solution uptake rate of 10 mly min are: RF powers1500 W, nebulizer gas flow rates0.10–0.12 lymin, intermediate gas flow rates1.5 lymin and DIHEN tip positions3–4 mm below the top of the torch intermediate tube. Acceptable recoveries (100"10%) and good precision (less than 3% relative standard deviation) are obtained for trace elemental analysis in organic matrices (a certified gas oil sample and a custom-made certified reference material) using flow injection analysis. Because of high blank levels, detection limits are 1–3 orders of magnitude higher for organic sample introduction than those acquired for aqueous solutions. 䊚 2002 Elsevier Science B.V. All rights reserved. Keywords: DIHEN; ICPMS; Flow injection analysis; Organic solvents; Principal component analysis

1. Introduction

*Corresponding author. Tel.: q1-202-994-6480; fax: q1202-994-2298. E-mail address: [email protected] (A. Montaser). 1 On leave from Department of Analytical Chemistry and Laboratory Optimization, PDVSA INTEVEP, Center of Investigation and Technological Support of Petroleum of Venezuela, Caracas 1041a, Venezuela. 2 ´ ´ On leave from Centro de Quımica Analıtica, Universidad Central de Venezuela, Caracas 1041a, Venezuela.

Inductively coupled plasma mass spectrometry (ICPMS) is a powerful technique for trace multielemental analysis of diverse samples dissolved in aqueous solutions w1x. However, the determination of trace elements in organic matrices (as in solvent extraction and the monitoring of wear metals in lubricating oils and other petroleum products) is often challenging, requiring procedures different

0584-8547/03/$ - see front matter 䊚 2002 Elsevier Science B.V. All rights reserved. PII: S 0 5 8 4 - 8 5 4 7 Ž 0 2 . 0 0 2 6 1 - 6

398

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

from aqueous sample introduction. This is because the aerosol properties and the behavior of the inductively coupled plasma (ICP) with organic solvent introduction differ markedly from those encountered with aqueous solutions w1–9x. For instance, the ICP is usually operated at higher power levels and lower solution uptake rates with organic solvents compared to aqueous solutions because the decomposition of organic compounds in the plasma requires more energy w2,10x. Recently, a simple, relatively low-cost direct injection high efficiency nebulizer (DIHEN) was introduced in the marketplace for the direct nebulization of test solution into the plasma w11–26x. The DIHEN offers 100% analyte transport efficiency for the introduction of 1–100 mlymin of test solutions into the ICP with no spray chamber or solution waste. In comparison to the direct injection nebulizer (DIN) w27–33x, the DIHEN is less expensive, relatively easy to use, and requires no high-pressure pump for solution delivery. Direct injection of sample into the plasma with the DIHEN offers several benefits compared to the conventional nebulizer–spray chamber arrangement: (1) a low internal dead volume and thus rapid response times and reduced memory effects; (2) no volatile analyte lost in the spray chamber; (3) improved precision by eliminating noise sources attributed to the spray chamber; and (4) similar or improved detection limits and sensitivity when operated at microliters per minute compared to conventional nebulizer–spray chamber arrangement requiring milliliters per minute sample. However, the DIHEN introduces nearly four times more solvent (and analyte) w11x, resulting in the deterioration of the plasma excitation and ionization properties and the increase in the spectral interferences, problems common to other direct injection nebulizers w27–33x. This attribute is particularly challenging when volatile organic solvents are introduced because of the plasma instability and the formation of carbon-based products on the nebulizer tip, torch confinement tube and sampling cone of the mass spectrometer interface w34x. To reduce or eliminate carbon formation, oxygen may be added to the nebulizer, intermediate or outer gas flows w31,33,35–42x, but this addition must

be carefully controlled to prevent the degradation of the sampling cone and spectral selectivity. This report is centered on investigation of the DIHEN for the introduction of samples dissolved in xylene using an argon–oxygen ICP to determine key elements important in petroleum industry. To facilitate multielemental analysis with DIHENICPMS, the operating parameters were identified through factorial experiment design. A modified response surface design was used to obtain the best-compromised operating conditions for the plasma. To our knowledge, this report is the first account on the performance of the DIHEN for analysis of organic matrices. 2. Statistical methods used 2.1. Factorial design experiments Several experiments are performed to identify the most influential factors, their ranges of influence, and factor interactions for organic solution introduction w43–46x. All factors are changed simultaneously at a limited number of factor levels. In one factorial design experiment, the screening design, two levels for each factor are used along with central points to check nonlinearities and estimate experimental error. In another factorial design, three levels for each factor are used to map the response surface. The drawback of this design is the large number of experiments that must be performed (3n experiments compared to 2n in the screening design; n is the number of parameters). Modified response surface designs, such as the central composite design and the Box– Behnken design, are more desirable to reduce the number of experiments w43x. In Box–Behnken design, as used in this work, the extreme factor levels are eliminated but the points are computed by approximation. This design is suitable for the investigation of the DIHEN for organic solvent introduction due to the delicacy of the nebulizer and the degradation possibility of the DIHEN tip due to excessive heating. 2.2. Principal component analysis and cluster analysis Principal component analysis (PCA) and cluster analysis (CA) are unsupervised learning methods,

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

that is, they analyze data without the need for additional information about the measurements or numerical data. They are useful in finding structures or similarities (groups, class) in the data. PCA is a multivariate statistical technique that can be applied to a set of variables to reduce dimensionality, that is, to replace a large set of intercorrelated variables with a smaller set of independent variables w46–55x. The new variables, the principal components (PCs), are linear combinations of those in the original data set, but they are uncorrelated with each other. The first PC reflects the original variability as much as possible, while the second one embodies the remaining variability. The last few PCs correspond to random noise. To determine the main components in the PCA, a series of statistical and heuristic approaches can be used. In this study, eigenvalue-one criterion is applied where the average eigenvalue of all PCs is one and only those components with eigenvalues greater than one are considered important w43x. The application of PCA techniques in treating trace elemental analysis with ICPAES w45,52,53x, and ICPMS w54,55x has been well documented. Another multivariate statistical technique used in this study is CA that allows reducing the amount of data through categorization or grouping w43x. Clustering methods can be divided into two basic categories: hierarchical and partitional clustering. Each type utilizes different algorithms to find the clusters. Special rules are used to merge small clusters or split the large ones. The result is illustrated by a cluster tree or a dendrogram, which shows how the clusters are related to each other. A clustering of the data items into different groups may be obtained by cutting the dendrogram at a desired level. Hierarchical clustering is used in this study to reduce the amount of data and pattern recognition. 3. Experimental 3.1. Instrumentation An Elan 6000 ICPMS system (Perkin–Elmery Sciex Corporation, Norwalk, CT) was used under the operating conditions listed in Table 1. The nebulizer and outer gas flows consisted of Ar–O2

399

mixtures. At the solution uptake rate of 10 mly min (see below), the amount of oxygen added to the cited flows (Table 1) was sufficient to provide stable plasma conditions for organic matrices and eliminate the formation of carbon-based products on the nebulizer tip, torch confinement tube and sampling cone of the mass spectrometer interface. Pure argon was used in this work with the DIHEN for the intermediate gas flow. In contrast, in conventional nebulization w35x oxygen is usually added to the intermediate gas flow to avoid carbon deposition on the torch intermediate tube. A maximum RF power of 1500 W was applied throughout this work to sustain stable plasma and avoid its degradation due to the addition of oxygen required to decompose the organic sample introduced directly into the plasma. All analytical data were collected under standard laboratory conditions, i.e. not in a clean-room environment. 3.2. Sample introduction systems The DIHEN (Model DIHEN-170-AA, J.E. Meinhard Associates, Inc., Santa Ana, CA) was constructed from borosilicate glass, based on the nebulizer tip dimensions used earlier for a conventional HEN w11x. To minimize memory effects, the dead volume of the nebulizer was reduced to less than 15 ml by inserting a Teflon䉸 capillary tubing (0.008 inch i.d.=0.016 inch o.d.; SB Fittings Assembly Kit no. 1-Micro, J.E. Meinhard Associates, Inc.) into the nebulizer to the point where the capillary tapers. This setup also allowed easy connection of the nebulizer to a micro flow injection valve. Argon flow rate was controlled by an external mass flow controller (Table 1). The output of the mass flow controller was connected to a gas proportioner (Model MFMR-0800-AA, E300y E200; Matheson Tri-Gas Inc.) where it was mixed with oxygen for delivery to the DIHEN. A computer-actuated, metal-free, six port injection valve (Cetac Technologies, Inc., Omaha, NE) was used to position the sample plug between two air bubbles for delivery to the plasma (Fig. 1) using the timing sequence given in Table 1. Peak profiles were obtained by using Teflon䉸 sample loops (Cetac Technologies, Inc.) for flow injection analysis. To reduce the dead volume, PEEK䉸

400

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

Table 1 Operating conditions for the DIHEN-ICPMS for organic solution introduction ICPMS System RF power (W) Nominal frequency (MHz) RF generator type Induction coil circuitry Sampling depth (above load coil, mm) Sampler (orifice diameter, mm) Skimmer (orifice diameter, mm) Outer gas flow rate (lymin)

Solution flow mode

PE-Sciex Elan 6000 1500 40 Free-running 3-turn coil, PLASMALOK䉸 11 Nickel, 1.1 Nickel, 0.9 15 of Ar with 0.2% vyv O2 (30 mlymin of O2). A Matheson gas proportioner (Model MFMR-0800-AA; E800yE300) was used to mix gases. 1.5 (The gas flow was varied from 1.5 to 2.1 lymin for the factorial design experiments using a Matheson gas Flow meter; Model FM1050;E500) DIHEN (Model 170-AA, J.E. Meinhard Associates䉸, Inc.) 3 below intermediate tube (The position was varied From 3 to 8 mm for the factorial design experiments) 10 0.12 of Ar with 37% vyv O2 (70 mlymin of O2). A Matheson gas proportioner (Model MFMR-0800-AA; E300yE200) was used to mix gases. Argon flow was Varied from 0.10 to 0.15 lymin for the factorial design experiments. Continuous and FIA

Timing parameters for flow injection analysis Initial delay (s) Load time (s) Inject time (s) Number of injections Analysis time (min)

Valve 1 0 60 285 1 5.75

Data acquisition parameters Scan mode Pointsymass Resolution (amu) Sweepsyreading Readingsyreplicate Replicates Dwell timeymass (ms) Integration timeymass (ms)

Peak hopping 1 0.7 5 3 5 50 750

Intermediate gas flow rate (lymin) Sample introduction system DIHEN tip position (mm) Solution uptake rate (mlymin) Nebulizer gas flow rate (lymin)

tubing (0.010 inch i.d.=0.0625 inch o.d.; Upchurch Scientific Inc., Oak Harbor, WA) was used for all the connections on the flow injection valve, except for the DIHEN (see above). The carrier flow was ACS grade xylene (Fisher scientific, Fair Lawn, NJ) which was delivered (10 mly min) to valve 1 using a syringe pump (Model KDS100, KD Scientific, New Hope, PA). With both valves in the load mode (Fig. 1a), this flow passed through the connecting tubing (PEEK䉸, 2inch long) and valve 2 to the DIHEN. In this mode, the sample was pumped into a 20-ml sample

Valve 2 0 155 190 1 5.75

loop (Cetac Technologies) on valve 2 using a fourchannel peristaltic pump (Rabbit, Rainin Instruments Co. Inc., Woburn, MA) and solvent-resistant pump tubing (0.015 inch i.d.; Astoria Pacific International, Clackamas, OR), while the other 20ml loop on valve 1 was filled with air using the same peristaltic pump with an open-end peristaltic pump tubing (Fig. 1a). Once the air loop was full, that is, all the solvent from the previous run was drained out, valve 1 was switched to the inject mode, delivering the air bubble into valve 2 (Fig. 1b). This air bubble approached valve 2 and

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

401

Fig. 1. Schematic diagram of the flow injection system used in this study: (a) Valve 1 and 2 are in the load mode. (b) Valve 1 is in the inject mode and valve 2 is in the load mode. (c) Valve 1 and 2 are in the inject mode.

moved toward the DIHEN as valve 2 was still in the load mode. When approximately half of this air bubble was pumped out of valve 2 (with the other half still in the connecting tubing), valve 2 switched to the inject mode (Fig. 1c). Consequently, the sample was positioned between two air bubbles. With this injection method, clean peaks with well-defined baselines were obtained and the memory effect was minimized. 3.3. Reagents For optimization studies, three oil-based standards containing 50 mgyg of 21 elements, 100 mgyg of As and 100 mgyg of Hg (Conostan䉸

Division, Conoco Inc., Ponca City, OK) were diluted in xylene (Fisher Scientific) to a concentration of 200 ngyg. The elements measured in this study (Table 2) are of interest in petroleum industry, especially for environmental applications. To estimate the relative amount of doubly charged ions and oxide species; Baqq and BaOq ions were also monitored. For elemental analysis, the cited multi-element standard solutions were prepared at concentration levels of 5, 10, 25, 50, 75 and 100 ngyg to obtain external calibration curves and to conduct precision and detection limit studies. A certified gas oil sample (PDVSA INTEVEP, Caracas, Venezuela) was used for spike recovery experiments. To fur-

Table 2 Isotopes under study and potential isobaric interferences Isotope

Natural abundance (%)a

Nominal mass (amu)a

51

99.8 26.2 71.7 100 51.8 12.8 71.7 71.5 29.9 52.4

50.9 59.9 69.0 74.9 106.9 110.9 137.9 153.9 202.0 208.0

V Ni 69 Baqq 75 As 107 Ag 111 Cd 138 Ba 154 BaO 202 Hg 208 Pb 60

a

From Ref. w56x. Potential atomic isobar of naturally occurring isotopes. c Partial list of potential molecular isobars. b

Potential atomic isobarsb

Potential molecular isobarsc HSO, ClO CaO Ga, ClO2, ArP, VO, Ceqq, Laqq ArCl, Smqq, Euqq, Ndqq ZrO, YO MoO

La, Ce Sm, Gd

LaO, CeO WO

402

Factors

Net signal intensity (cps)

DIHEN Nebulizer Intermediate position gas gas (mm) (lymin) (lymin)

51

3 8 3 8 3 8 3 8 5 5 5 5 5a 5a 5a

9675 1655 6215 174 9153 2896 3258 14 6260 3103 97 18 758 890 778

0.10 0.10 0.15 0.15 0.12 0.12 0.12 0.12 0.10 0.15 0.10 0.15 0.12 0.12 0.12

1.70 1.70 1.70 1.70 1.50 1.50 2.10 2.10 1.50 1.50 2.10 2.10 1.70 1.70 1.70

V

60

Ni

99 531 16 821 63 223 2346 95 602 26 256 33 314 323 61 734 31 223 1405 309 9066 10 572 9085

75

As

239 139 36 099 130 719 3750 223 482 58 495 55 079 475 141 355 63 816 2046 311 12 139 15 847 12 451

Ratios

107

Ag

355 069 103 250 311 055 21 195 331 891 169 471 204 772 3434 298 846 178 169 15 240 3625 79 962 91 228 79 581

111

Cd

51 491 16 285 37 808 3915 49 032 26 501 25 681 668 38 959 25 783 2398 694 11 394 13 279 12 245

138

Ba

202

Hg

206

Pb

208

Pb

BaO yBa

640 615 107 004 309 101 585 544 0.2 264 677 34 177 159 315 306 324 5.9 625 292 8565 313 049 590 400 1.3 59 000 7326 46 317 91 520 1.0 635 079 105 686 295 918 567 381 0.1 452 258 51 598 235 443 469 911 4.8 463 508 46 393 236 622 465 070 5.0 9904 1465 10 168 20 225 2.0 644 312 83 311 330 795 650 809 1.3 432 963 50 563 221 450 434 881 3.6 42 945 4338 33 368 62 088 2.4 10 273 1297 9702 18 972 3.2 223 146 19 142 133 876 262 809 8.8 246 122 20 840 148 242 294 245 8.2 236 255 9316 141 802 283 980 8.1

See Table 1 for other operating parameters. a Central points to estimate the experimental errors and to study the nonlinear dependences.

q

Objective function q

Ba

qq

1.5 0.3 0.5 0.2 1.1 0.2 0.4 0.2 0.8 0.3 0.2 0.2 0.2 0.2 0.2

yBa

q

RIsotope

48 816 13 426 34 695 2900 45 555 20 170 22 116 552 37 529 21 305 1929 527 8516 9703 8682

and BaqqyBaq

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

Table 3 Net signal intensities and objective function (R) of several isotopes and ratios of BaOqyBaq and BaqqyBaq in experiments involving response surface design

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

403

Fig. 2. PCA and CA of the data in Table 3. (a) Plot of component weight for net signal intensity of several isotopes, BaOqyBaq and BaqqyBaq. (b) Dendrogram for net signal intensity of several isotopes, BaOqyBaq and BaqqyBaq.

ther verify the accuracy of the method, a custommade certified xylene solution (VHG Labs, Manchester, NH) containing 10 mgyg of As and Hg was also analyzed. 3.4. Experiment involving response surface design The Box–Behnken design was used with three levels for each factor: DIHEN tip position (3, 5 and 8 mm), nebulizer gas flow rate (0.10, 0.12

and 0.15 lymin) and intermediate gas flow rate (1.5, 1.7 and 2.1 lymin) with three central points (DIHEN tip position, 5 mm, nebulizer gas flow rate, 0.12 lymin and intermediate gas flow rate, 1.7 lymin). A nebulizer gas flow rate of 0.10 ly min or higher was selected because below this level droplets accumulated on the nebulizer tip, causing a drift in the aerosol trajectory or nebulizer clogging. Experiments were conducted at a concentration of 200 ngyg for all elements using auto

404

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

Fig. 3. Plot of response surfaces as a function of DIHEN tip position and nebulizer gas flow rate: (a) objective function (R) for the net signal intensity of several isotopes and BaqqyBaq and (b) BaOqyBaq.

lens mode to optimize the lens voltage for all isotopes investigated. A commercial software package (STATGRAPHICS PLUS 5.0, Manugistics, Inc., Rockville, MD) was utilized to perform the factorial design analysis, PCA, CA, and other statistical calculations. 4. Results and discussion 4.1. Optimization of plasma operating conditions for net signal intensity and %R.S.D. The Box–Behnken design was used to investigate dependence of the response function on the

operating parameters using three levels per factor. The influence of the operating parameters on net signal intensity of different isotopes, BaOq yBaq and Baqq yBaq ratios are listed in Table 3. The analysis at the central point (DIHEN tip position, 5 mm, nebulizer gas flow rate, 0.12 lymin and intermediate gas flow rate, 1.7 lymin) was randomly repeated three times to estimate the experimental errors. PCA of these data extracted two components, with the correlations between PC and each original variables presented in Fig. 2a. For all isotopes and Baqq yBaq, a significant correlation

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

405

Fig. 4. Plot of response surfaces as a function of DIHEN tip position and intermediate gas flow rate: (a) objective function (R) for net signal intensity of several isotopes and BaqqyBaq and (b) BaOqyBaq.

was noted to PC1, except for BaOq yBaq, which was correlated to PC2. This figure reveals the similarity between the responses obtained for 51V, 60 Ni, 75As, 107Ag, 111Cd, 202Hg, 208Pb and BaqqyBaq under different operating conditions as compared to BaOq yBaq, which forms a separate group. The corresponding dendrogram (Fig. 2b) obtained by CA also shows two groups formed at a distance value of 6. Considering that CA and PCA reveal two separate categories, an n-dimensional response function (R) was calculated for each group in order to

simplify the results w57x: Rsf(Q1,Q2,«,Qn)

(1)

where R represents the behavior of all the elements in a group, n is the number of factors varied during the optimization, and Qi is a measure of the analytical performance (in this study, net signal intensity and %R.S.D.) of element i at a given combination of parameters w45x. The response function has the form

406

Factors

%R.S.D.

DIHEN position (mm)

Nebulizer gas (lymin)

Intermediate gas (lymin)

51

60

69

3 8 3 8 3 8 3 8 5 5 5 5 5a 5a 5a

0.10 0.10 0.15 0.15 0.12 0.12 0.12 0.12 0.10 0.15 0.10 0.15 0.12 0.12 0.12

1.70 1.70 1.70 1.70 1.50 1.50 2.10 2.10 1.50 1.50 2.10 2.10 1.70 1.70 1.70

4.5 9.3 1.4 10.8 1.2 4.9 7.7 34.3 3.5 2.1 16.3 26.9 5.8 7.4 8.1

1.1 9.2 2.6 10.0 1.7 2.8 4.7 10.1 3.0 2.9 12.7 9.0 5.0 4.6 3.8

7.9 8.5 7.6 6.9 2.1 6.2 6.4 23.4 6.0 5.0 17.2 12.1 10.9 9.2 5.7

V

Objective function Ni

Ba

qq

75

107

111

138

154

202

206

208

R%R.S.D.

1.9 5.3 6.2 6.9 2.0 5.78 5.0 20.0 1.2 4.0 11.2 7.7 6.7 6.1 5.4

1.1 5.5 2.7 8.2 0.8 2.7 3.3 6.6 3.2 2.4 9.5 10.3 5.0 5.7 1.4

1.3 5.3 3.0 6.7 1.8 2.2 3.9 6.8 4.2 0.5 9.9 8.3 6.6 4.7 1.8

6.8 .2 2.4 8.4 4.5 2.6 4.5 13.6 8.2 4.0 12.8 12.3 4.4 5.3 2.5

3.9 2.0 3.7 6.4 5.3 3.6 2.1 11.0 5.9 2.3 10.9 8.4 3.1 2.9 5.1

0.5 4.6 2.7 4.1 0.2 2.2 1.4 6.9 6.4 3.4 8.0 8.4 6.0 5.1 3.3

2.2 5.3 2.9 8.7 1.6 2.8 1.4 8.3 2.5 5.0 6.6 4.0 3.1 2.8 3.6

5.8 3.5 5.3 6.7 3.3 1.7 3.1 7.3 1.0 5.2 8.5 9.5 4.3 3.7 4.2

2.4 5.4 3.3 7.4 1.7 3.1 3.4 11.5 3.4 2.9 10.8 9.6 5.2 4.9 3.6

As

Ag

Cd

See Table 1 for other operating parameters. a Central points to estimate the experimental errors and to study the nonlinear dependences.

Ba

BaO

Hg

Pb

Pb

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

Table 4 Relative standard deviation of analyte intensity and objective function (R) for several isotopes, BaOq, and Baqq in experiments involving response surface design

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

407

Fig. 5. Response surfaces of objective function (R) for %R.S.D. of several isotopes, Baqq and BaOq as a function of: (a) DIHEN tip position and nebulizer gas flow rate; (b) DIHEN position and intermediate gas flow rate. n i Rs2QW i

(2)

is1 i Here R is the geometric mean of the QW i values and Wi is an element-specific weighting parameter. The calculated R-values under different experimental parameters are shown in Table 3. Fig. 3 shows the R function for the net signal intensity of several isotopes and Baqq yBaq (Fig. 3a), and BaOq yBaq (Fig. 3b). The best R-values are obtained for low nebulizer gas flow rates

(0.10–0.12 lymin) and short distances (3–4 mm) between the DIHEN tip and the plasma. Below these values, either the plasma was not stable or the hot region of the mixed-gas plasma could damage the DIHEN tip. For BaOq yBaq ratio, maximum response is obtained at DIHEN tip position of 5–7 mm and nebulizer gas flow rate of 0.12–0.13 lymin. Fig. 4 shows the response surfaces obtained for the DIHEN position vs. intermediate gas flow rate for the net signal intensity of different isotopes

408

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

Fig. 6. Peak profiles: (a) using 5, 10, 20 and 50-ml sample loops for a 10 ngyg Hg in xylene; (b) for 20-ml injections (ns10) of 20 ngyg Hg in xylene; carrier flow rates10 mlymin.

and Baqq yBaq (Fig. 4a), and BaOq yBaq (Fig. 4b) for the data presented in Table 3. The behavior is similar to that noted in Fig. 3, i.e. the best operating condition for the DIHEN position and intermediate gas flow rate are 3–4 mm and 1.5– 1.7 lymin, respectively. In terms of the net signal intensity of analyte, the DIHEN tip position has the most important effect. Increasing the DIHEN tip position from 3 to 8 mm, with respect to the

top of torch intermediate tube would decrease the response function by a factor of 2–3 (Fig. 3a and Fig. 4a). The effect of DIHEN tip position is greater than those observed for nebulizer gas flow rate (from 0.10 to 0.15 lymin) and intermediate gas flow rate (from 1.5 to 2.1 lymin). The maximum in BaOq yBaqq response is acquired at the DIHEN tip position of 5–7 mm and intermediate gas flow rate of 1.7–1.9 lymin. While the oxide

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

ratio is minimized under the optimum operating conditions, the values (;10%) obtained are still higher than those encountered in aqueous solution introduction with the DIHEN w11x and conventional nebulizer–spray chambers (-2%) w1,58x mainly due to the addition of oxygen to the plasma. The influence of the operating parameters on %R.S.D. of analyte signal, BaOq and Baqq are shown in Table 4 along with the R function. PCA and CA of these data showed that all isotopes, BaOq and Baqq form a single group (not shown here) and therefore R%R.S.D. function was calculated for all of them. Fig. 5 shows the response surfaces obtained for the R%R.S.D. function vs. the DIHEN position and nebulizer gas flow rate (Fig. 5a) and the DIHEN position and intermediate gas flow rate (Fig. 5b). The lowest response function for the %R.S.D. values (i.e. the best precision) was obtained at intermediate gas flow rate of 1.5– 1.7 lymin, DIHEN tip position of 3–4 mm, and at nebulizer gas flow rate of 0.12–0.14 lymin. 4.2. Analysis of petroleum samples The optimum operating conditions identified in the previous section were used to develop a mFIA-

409

DIHEN-ICPMS method (Table 1) for the analysis of metals in organic matrices. Fig. 6a shows signal intensities for the nebulization of a 10-ngyg solution of Hg in xylene using 5, 10, 20 and 50-ml sample loops. At carrier flow rate of 10 mlymin the peak height approaches the steady-state signal with the use of 20 or 50-ml sample loops. While peak profiles are very well defined in terms of baseline and height, the appearance of the small tail peak for each main peak is indicative of memory effect for the element under study. A 20ml sample loop was selected for the analysis because the tailing was minimal and signal was reaching the steady-state level. The reproducibility of this method is shown in Fig. 6b for ten injections over a 1-h period for a 20-ngyg solution of Hg in xylene. Similar profiles were obtained for As, Ni and V under the same operating conditions. The precision obtained for several elements is shown in Fig. 7 for different concentration levels and ten injections, indicating a large variation in precision at concentrations less than 10 ngyg (near detection limits). The detection limits obtained for several elements are listed in Table 5. The values for the

Fig. 7. Precision of the FIA-DIHEN-ICPMS method for several elements as function of concentration of several elements in xylene.

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

410

Table 5 Detection limits (ppt) obtained for several isotopes in xylene and aqueous solution using the DIN and the DIHEN Isotope

75

As Cd 202 Hg 60 Ni 208 Pb 51 V 111

Detection limit in xylene

Detection limit in aqueous solution

DINa (1000-ml sample loop)

DIHEN (20-ml sample loop)

DINa (1000-ml sample loop)

DIHEN (85 mlymin)

240 2400 410 2100 330 2800

610 770 1440 810 1460 870

130 17 120 31 24 48

17b 2 37 12b 5b 2b

Based on 3s of blank signal measured at the mass of the isotope. a From Ref. w33x. b From Ref. w11x.

DIHEN were acquired based on the three times S.D. of the blank signal (pure xylene) measured at the mass of the isotope over ten replicates using a six-point external calibration curve (from 5 to 100 ngyg). Because of high blank levels, the detection limits are 1–3 orders of magnitude higher than those obtained for an aqueous solution with the DIHEN but comparable to those obtained by DIN for organic sample introduction using a 1000-ml sample loop w33x. To investigate the accuracy of the method, a certified gas oil sample (PDVSA INTEVEP) was spiked with elements of interest (to 20 ngyg) and a six-point external calibration curve (from 5 to 100 ngyg) was used to calculate the %recoveries

(Table 6A). Clearly, all the elements have been reasonably recovered indicating the good accuracy of the developed technique. To further study the performance of the method, a custom-made CRM was obtained form VHG Labs containing 10 mgy g of As and Hg. This sample was diluted in xylene to 50 ngyg and was analyzed under the optimum operating conditions using external calibration method. The results (Table 6B) show good agreement with the CRM, again documenting the accuracy of the method. 5. Conclusions Optimum parameters were identified for operating a DIHEN with ICPMS for introduction of

Table 6 Analytical applications of DIHEN-ICPMS for analysis of petroleum products (A) Spike %recovery for a certified gas oil sample Isotope

%Recoverya

75

107"7 103"4 100"2 101"4 101"5

As Cd 202 Hg 60 Ni 51 V 111

(B) Analysis of a custom-made certified petroleum reference material Isotope Measured concentration (mgyg) 75

As Hg

202

10.9"0.1 9.7"0.4

All determinations represent an average of four measurements. a Based on the %recovery of a 20-ngyg spike addition to the gas oil sample.

Certified concentration (mgyg) 10.1"0.1 10.1"0.1

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

volatile organic solvents and petroleum products. These conditions are different from those encountered in aqueous solution introduction with the DIHEN. A solution uptake rate of 10 mlymin provided stable plasma conditions for organic matrices and eliminated the formation of carbonbased products on the nebulizer tip, torch confinement tube and sampling cone of the mass spectrometer interface. Compared to the nebulizer and intermediate gas flow rates, the DIHEN position within the ICP torch had a major effect on the precision (%R.S.D.) and net signal intensity of analyte. The oxide levels measured with the DIHEN for organic solution introduction are higher than those measured for aqueous solutions with the DIHEN and conventional nebulization, mainly due to the addition of oxygen to the ICP. At the optimum conditions, the DIHEN-ICPMS approach provided precise and accurate trace elemental analysis information for organic matrices. Acknowledgments This research was sponsored by grants from the US Department of Energy (DE-FG0293ER14320), the National Science Foundation (CHE-9505726 and CHE-9512441), and J.E. Meinhard Associates, Inc. Financial support for AS was provided by Analytical Chemical and Optimization of Laboratory Department, PDVSA INTEVEP S.A. The authors are grateful to Billy W. Acon, Craig S. Westphal, John A. McLean and Su-Ann E. O’Brien for constructive discussions and assistance. We express our appreciation to Bill Rutkowski for excellent machine shop services. References w1x A. Montaser (Ed.), Inductively Coupled Plasma Mass Spectrometry, VCH-Wiley, New York, 1998. w2x A. Montaser, D.W. Golightly (Eds.), Inductively Coupled Plasmas in Analytical Atomic Spectrometry, second ed., VCH-Wiley, New York, 1992. w3x P.W.M. Boumans, M.C. Lux-Steiner, Modification and optimization of a 50 MHz inductively coupled argon plasma with special reference to analyses using organic solvents, Spectrochim. Acta Part B 37 (1982) 97–126. w4x D.G. Wier, M.W. Blades, Characteristics of an inductively coupled argon plasma operating with organic

w5 x

w6 x

w7 x

w8 x

w9 x

w10x

w11x

w12x

w13x

w14x

w15x

w16x

w17x

w18x

411

aerosols. Part 1. Spectral and spatial observations, J. Anal. At. Spectrom. 9 (1994) 1311–1322. D.G. Wier, M.W. Blades, Characteristics of an inductively coupled argon plasma operating with organic aerosols. Part 2. Axial spatial profiles of solvent and analyte species in a chloroform-loaded plasma, J. Anal. At. Spectrom. 9 (1994) 1323–1334. A.C. Lazar, P.B. Farnsworth, Characterization of an inductively coupled plasma with xylene solutions introduced as monodisperse aerosols, Anal. Chem. 69 (1997) 3921–3929. C. Pan, G. Zhu, R.F. Browner, Comparison of desolvation effects with aqueous and organic (carbon tetrachloride) sample introduction for inductively coupled plasma atomic emission spectrometry, J. Anal. At. Spectrom. 5 (1990) 537–542. L. Wang, S.W. May, R.F. Browner, S.H. Pollock, Lowflow interface for liquid chromatography-inductively coupled plasma mass spectrometry speciation using an oscillating capillary nebulizer, J. Anal. At. Spectrom. 11 (1996) 1137–1146. J.W. Olesik, A.W. Moore, Influence of small amounts of organic solvents in aqueous samples on argon inductively coupled plasma spectrometry, Anal. Chem. 62 (1990) 840–845. A.W. Boorn, R.F. Browner, Effects of organic solvents in inductively coupled plasma atomic emission spectrometry, Anal. Chem. 54 (1982) 1402–1410. J.A. McLean, H. Zhang, A. Montaser, A direct high efficiency nebulizer for inductively coupled plasma mass spectrometry, Anal. Chem. 70 (1998) 1012–1020. A. Montaser, J.A. McLean, J.M. Kacsir, A Novel Direct Injection High Efficiency Nebulizer for Atomic Spectroscopy, U.S. Patent No. 6,166,379, 2000. J.S. Becker, H.J. Dietze, J.A. McLean, A. Montaser, Ultratrace and isotope analysis of long-lived radionuclides by inductively coupled plasma quadrupole mass spectrometry using a direct injection high efficiency nebulizer, Anal. Chem. 71 (1999) 3077–3084. B.W. Acon, J.A. McLean, A. Montaser, A large bore direct injection high efficiency nebulizer for inductively coupled plasma spectrometry, Anal. Chem. 72 (2000) 1885–1893. M.G. Minnich, A. Montaser, Direct injection high efficiency nebulization in inductively coupled plasma mass spectrometry under cool and normal plasma conditions, Appl. Spectrosc. 54 (2000) 1261–1269. J.A. McLean, M.G. Minnich, J. Su, W. Lai, A. Montaser, Optical patternation: a technique for three-dimensional aerosol diagnostic, Anal. Chem. 72 (2000) 4796–4804. M.G. Minnich, J.A. McLean, J.A. Montaser, Spatial aerosol characteristics of a direct injection high efficiency nebulizer via optical patternation, Spectrochim. Acta Part B 56 (2001) 1113–1126. B.W. Acon, J.A. McLean, A. Montaser, A direct injection high efficiency nebulizer interface for microbore

412

w19x

w20x

w21x

w22x

w23x

w24x

w25x

w26x

w27x

w28x

w29x

w30x

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413 high-performance liquid chromatography-inductively coupled plasma mass spectrometry detection, J. Anal. At. Spectrom. 16 (2001) 852–857. C.S. Westphal, J.A. McLean, B.W. Acon, L.A. Allen, A. Montaser, Axial inductively coupled plasma time-offlight mass spectrometry using direct liquid sample introduction, J. Anal. At. Spectrom. 17 (2002) 669–675. ´ J.M. Mermet, Evaluation of a direct injecJ.L. Todolı, tion high-efficiency nebulizer (DIHEN) by comparison with a high-efficiency nebulizer (HEN) coupled to a cyclonic spray chamber as a liquid sample introduction system for ICP-AES, J. Anal. At. Spectrom. 16 (2001) 514–520. J.R. Chirinos, K. Kahen, S.E. O’Brien, A. Montaser, Mixed-gas inductively coupled plasma atomic emission spectrometric using direct injection efficiency nebulizer, Anal. Bioanal. Chem. 372 (2002) 128–135. S.E. O’Brien, J.A. McLean, B.W. Acon, B.J. Eshelman, W.F. Bauer, A. Montaser, Determination of memoryprone elements using direct injection high efficiency nebulizer inductively coupled plasma mass spectrometry, Appl. Spectrosc. 56 (2002) 1006–1012. S.F. Boulyga, J.L. Matusevich, V.P. Mironov, V.P. Kudrjashov, L. Halicz, I. Segal, J.A. McLean, A. Montaser, J.S. Becker, Determination of 236Uy238U isotope ratio in contaminated environmental samples using different ICP-MS instruments, J. Anal. At. Spectrom. 17 (2002) 958–964. E. Bjorn, W. Frech, Non-spectral interference effects in inductively coupled plasma mass spectrometry using direct injection high efficiency and microconcentric nebulization, J. Anal. At. Spectrom. 16 (2001) 4–11. V. Majidi, J. Qvarnstrom, Q. Tu, W. Frech, Y.J. Thomassen, Improving sensitivity for CE-ICP-MS using multicapillary parallel separation, J. Anal. At. Spectrom. 14 (1999) 1933–1935. S.F. Boulyga, J.S. Becker, Determination of uranium isotopic composition and 236U content of soil samples and hot particles using inductively coupled plasma mass spectrometry, Fresenius’ J. Anal. Chem. 370 (2001) 612–617. D.R. Wiederin, F.G. Smith, R.S. Houk, Direct injection nebulization for inductively coupled plasma mass spectrometry, Anal. Chem. 63 (1991) 219–225. D.R. Wiederin, R.E. Smyczek, R.S. Houk, On-line standard additions with direct injection nebulization for inductively coupled plasma mass spectrometry, Anal. Chem. 63 (1991) 1626–1631. ´ J.M. Mermet, Minimization of acid effects J.L. Todolı, at low consumption rates in an axially viewed inductively coupled plasma atomic emission spectrometer by using micronebulizer-based sample introduction systems, J. Anal. At. Spectrom. 13 (1998) 727–734. J.S. Crain, J.T. Kiely, Waste reduction in inductively coupled plasma mass spectrometry using flow injection

w31x

w32x w33x

w34x

w35x

w36x

w37x

w38x

w39x

w40x

w41x

w42x

w43x

w44x

w45x

and direct injection nebulizer, J. Anal. At. Spectrom. 11 (1996) 525–528. T.W. Avery, C. Chakrabarty, J.J. Thompson, Characterization and optimization of a direct injection nebulizer for introduction of organic solvents and volatile analyte species into an inductively coupled plasma, Appl. Spectrosc. 44 (1990) 1691–1698. M.R. Plantz, Varian Inc., Wood Dale, IL, Private Communication, December 31, 2001. R.I. Botto, Trace Element Analysis of Petroleum Naphthas and Tars Using Direct Injection ICP-MS, presented at the 27th annual meeting of Federation of Analytical Chemistry and Spectroscopy Societies (FACSS), Nashville, TN, September, 2000. A. Van Heuzen, ICP-MS in the petroleum industry, in: A.R. Date, A.L. Gray (Eds.), Applications of Inductively Coupled Plasma Mass Spectrometry, Blackie and Son Ltd, Glasgow, 1989, pp. 169–188, Chapter 7. C. Pan, G. Zhu, R.F. Browner, Role of auxiliary gas flow in organic sample introduction with inductively coupled plasma atomic emission spectrometry, J. Anal. At. Spectrom. 7 (1992) 1231–1237. A.A. Verbeek, I.B. Brenner, Slurry nebulization of geological materials into argon, argon–nitrogen and argon–oxygen inductively coupled plasmas, J. Anal. At. Spectrom. 4 (1989) 23–26. I.B. Brenner, A. Zander, Sequential multielement analysis of coals by Slurry nebulization using Ar, Ar–O2, and N2 ICPAES, Spectroscopy 13 (1998) 14–16. I.B. Brenner, E. Dorfman, Application of conventional and low-flow torches for the analysis of xylene- and tetralin-based petroliferous materials by argon and oxygen–argon ICP-AES, Microchem. J. 52 (1995) 81–95. J. Wang, E.H. Evans, J.A. Caruso, Addition of molecular gases to argon gas flows for the reduction of polyatomicion interferences in inductively coupled plasma mass spectrometry, J. Anal. At. Spectrom. 7 (1992) 929–936. D. Hausler, Trace element analysis of organic solutions using inductively coupled plasma mass spectrometry, Spectrochim. Acta Part B 42 (1987) 63–73. R.I. Botto, Applications of ultrasonic nebulization in the analysis of petroleum and petrochemicals by inductively coupled plasma atomic emission spectrometry, J. Anal. At. Spectrom. 8 (1993) 51–57. R.I. Botto, J.J. Zhu, Universal calibration for analysis of organic solutions by inductively coupled plasma atomic emission spectrometry, J. Anal. At. Spectrom. 11 (1996) 675–681. M. Otto, Chemometrics Statistics and Computer Application in Analytical Chemistry, VCH-Wiley, New York, 1999. G.E. Box, W.G. Hunter, J.S. Hunter, Statistic for Experimenters. An Introduction to Design, Data Analysis and Model Building, Wiley, Canada, 1978. J. Chirinos, A. Fernandez, J. Franquiz, Multi-element optimization of the operating parameters for inductively

K. Kahen et al. / Spectrochimica Acta Part B 58 (2003) 397–413

w46x

w47x

w48x

w49x

w50x

w51x w52x

coupled plasma atomic emission spectrometry with a charge injection device detector for the analysis of samples dissolved in organic solvents, J. Anal. At. Spectrom. 13 (1998) 995–1000. C. Bayne, I. Rubin, Practical Experimental Designs and Optimization Methods for Chemists, VCH, Weinheim, Germany, 1986. R.J. Howarth, B.J. Coles, M.H. Ransey, The potential of multivariate quality control as a diagnostic tool in geoanalysis, Analyst 125 (2000) 2032–2037. M.T. Fisher, J. Lee, M.K. Mara, Techniques for evaluating control of automated multi-determinant analytical instruments by computer, Analyst 111 (1986) 1225–1229. S.J. Smith, S.P. Caudill, J.L. Pirkle, D.L. Ashley, Composite multivariate quality control using a system of univariate, bivariate and multivariate quality control rules, Anal. Chem. 63 (1991) 1419–1425. S.P. Caudill, S.J. Smith, J.L. Pirkle, D.L. Ashley, Performance characteristics of composite multivariate quality control system, Anal. Chem. 64 (1992) 1390–1395. P.C. Meier, R.E. Zund, Statistical Methods in Analytical Chemistry, vol. 153, Book News Inc, Portland, 2000. A. Lopez-Molinero, A.V. Caballero, J.R. Castillo, Classification of emission spectral lines in inductively coupled plasma atomic emission spectroscopy using principal component analysis, Spectrochim. Acta Part B 49 (1994) 677–682.

413

w53x M. Cave, Improvement of short-term precision in inductively coupled plasma atomic emission spectrometry by principal component analysis modeling, J. Anal. At. Spectrom. 13 (1998) 125–129. w54x I.E. Bechmann, S. Sturup, L.V. Kristensen, High resolution inductively coupled plasma mass spectrometry (HR-ICPMS) determination and multivariate evaluation of 10 trace elements in mussels from 7 sites in Limfjorden, Denmark, Fresenius’ J. Anal. Chem. 368 (2000) 708–714. w55x B. Wyrzykowska, K. Szymczyk, H. Ichichashi, J. Falandysz, B. Skwarzec, S. Yamasaki, Application of ICP sector field MS and principal component analysis for studying interdependences among 23 trace elements in polish beers, J. Agric. Food Chem. 49 (2001) 3425–3431. w56x D.R. Lide (Ed.), CRC handbook of Chemistry and Physics, CRC Press, Boca Raton, FL, 1995. w57x D.A. Sadler, D Littlejohn, C.V. Perkin, Multi-element optimization of the operating parameters for inductively coupled plasma atomic emission spectrometry with a charge-transfer device detection system: a study of the effects of different response functions, J. Anal. At. Spectrom. 11 (1996) 463–468. w58x M.A. Vaughan, G. Horlick, Oxide, hydroxide and doubled-charged analyte species in inductively coupled plasma mass spectrometry, Appl. Spectrosc. 40 (1986) 434–445.