Spatial Emission Profiles for Flagging Matrix Interferences in Axial

Sep 27, 2012 - when viewed axially and a false-color spatial map of the plasma continuum .... spatial emission profile will flag the interference as l...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/ac

Spatial Emission Profiles for Flagging Matrix Interferences in AxialViewing Inductively Coupled Plasma-Atomic Emission Spectrometry: 1. Profile Characteristics and Flagging Efficiency George C.-Y. Chan* and Gary M. Hieftje Department of Chemistry, Indiana University, 800 E. Kirkwood Avenue, Bloomington, Indiana 47405, United States ABSTRACT: Spatially resolved measurements of analyte emission along the cross-sectional axis of an axially viewed inductively coupled plasma (ICP) are utilized to indicate the presence of any of the three major categories of matrix interferences (i.e., plasma-related, sample introduction-related, and spectral interferences). Barium at concentrations of 0.05 or 0.1 M was chosen as a prototype element for plasma-related matrix effects, whereas common mineral acids (nitric, hydrochloric, sulfuric, and phosphoric) at volumetric concentrations from 1% to 20% were used to simulate sample introduction-related matrix effects. Three spectrally interfering line pairs (As and Cd at 228.81 nm, Er and Co at 239.73 nm, and Er and Ce at 302.27 nm) were selected for the study of spectral interferences. Due to dependence on the nature of the interference, the analytical bias at the center of the cross-sectional profile varied between −40% and +50%. In all matrixinterference categories, because plasma characteristics and excitation conditions are heterogeneous along this cross-sectional axis, matrix-induced shifts in analyte emission vary accordingly. As a result, the concentrations determined for an analyte along the cross-sectional plasma axis are not constant but exhibit a position dependence that allows the interference to be flagged. With the exception of spectral interference from emission lines whose total excitation potentials (i.e., the sum of ionization and excitation energies of an ionic emission line) are very close, the spatially resolved concentrations provide an effective indicator for flagging any other matrix interference in axial-viewing ICP-emission spectrometry. The method can be employed under the plasma forward power and carrier-gas flow conditions that are common for robust plasma operation.

I

matrix-effect categories in conventional lateral-viewing (i.e., side-on) ICP-AES. This indicator is based on the fact that plasma behavior and excitation conditions are heterogeneous along the ICP vertical observation axis. As a result, the relative magnitude, and sometimes even the direction, of a matrixinduced change in emission intensity is not constant but is a function of vertical location in the plasma.15,17 The concept and procedure are straightforward. Calibration curves are developed at several vertical locations in the plasma, by means of standard solutions. Samples are then analyzed at each vertical location by reference to the respective calibration curve. If a sample suffers no matrix interference, analytes in the sample will behave in the same way as in the standard solutions, and the determined concentration will be the same at all vertical positions. However, if matrix interferences exist, the determined concentrations will exhibit a spatial dependence, thereby allowing the matrix effect to be detected. System drift can also be exposed in the same way.15 Here we apply the same concept to axially viewed ICP-AES. It is well-established that axial-viewing ICP-AES offers better sensitivities and detection limits than the conventional lateralviewing mode,18,19 but it is also more prone to matrix interferences.20,21 Therefore, from a practical point of view,

nductively coupled plasma-atomic emission spectrometry (ICP-AES) has been widely adopted for multielement analysis of samples of virtually every kind. Although all commercial ICP-AES spectrometers include some automated features (e.g., programmed analysis, automatic calibration, statistical treatment of data and spectral line selection with multiline analysis1), there is one essential component missing in these instruments: the ability to warn the operator when an analytical result is compromised by matrix interferences. The presence of such interferences, without the awareness and subsequent correction by an analyst, will lead to an analytical error that can be 30% or more.2−6 Without dispute, accuracy is the most important quality that an analyst expects.7 One possible reason for the lack of such alerting capability in commercial ICP-AES instruments is the complex nature of matrix interferences. The sample matrix can alter the sample introduction and transport processes8,9 (termed sample introduction-related matrix effects), the physical characteristics of the plasma10,11 (termed plasma-related matrix effects), and can cause spectral overlap with the analyte species of interest12,13 (termed spectral interferences). The characteristics and behavior of these three categories of matrix effects are all different. As a result, different means have been explored in the literature to signal the presence of each category of matrix interference.4,14 Previously, we developed a simple all-in-one indicator15,16 for flagging matrix interferences originating from any of the three © 2012 American Chemical Society

Received: July 13, 2012 Accepted: September 27, 2012 Published: September 27, 2012 50

dx.doi.org/10.1021/ac302095w | Anal. Chem. 2013, 85, 50−57

Analytical Chemistry

Article

Figure 1. Schematic diagram of the optical arrangement for axial-viewing ICP-AES used in this study. The inset shows a photograph of the plasma when viewed axially and a false-color spatial map of the plasma continuum recorded by the CCD.

in this paper were carried out at a plasma forward power of 1100 W and with a plasma torch having an injector diameter of 3 mm. Outer and intermediate gas flows were typical at 14 and 1 L/min, respectively. A sheath flow of 0.10 L/min Ar was added to the central channel at the exit of the spray chamber. The total central channel gas flow, which is the sum of the nebulizer gas and the 0.10 L/min sheath gas, is reported below. The ACTIVA spectrometer combines a Czerny-Turner configuration with a two-dimensional (2D) charge-coupled device (CCD) detector,23 so when the plasma is viewed endon, the entire cross-sectional emission profile of the plasma over a selected spectral range is simultaneously measured (cf. inset in Figure 1). The original entrance optics of the lateralviewing ACTIVA were removed and the plasma was focused onto the entrance slit by a single fused-silica lens (focal length = 100 mm, LUP-50.0-47.0-UV, CVI Laser, Albuquerque, NM) at a 2.7:1 demagnification ratio. The resulting observation length in the plasma is about 19 mm, which is approximately the same as the internal diameter of the torch. Other parameters of the spectrometer (e.g., slit width, resolution) are identical to those listed in our previous report.15 Optical Alignment. A He−Ne laser, used to aid optical alignment, was mounted on a translational stage equipped with an angle adjustment option and was positioned so the beam passed through the entrance slit aligned with the optical axis of the spectrometer. The converging lens was then aligned to the same axis by adjustment of its position so the path of the He− Ne laser beam did not change upon passing through the lens. An adjustable iris was attached to the back of the lens to match the f/no. (f/8.5) of the spectrometer. The torch was then lined up by passing the laser beam through the torch injector. The torch compartment is mounted on a three-dimensional (3D) translational stage (OWIS GmbH, Staufen, Germany). Because of the length of the plasma (approximately 4.5 cm) and the

the development of a matrix-effect indicator for axial-viewing ICP-AES is perhaps even more important than in side-on measurements. Spatially resolved measurements in an axially viewed ICP represent the longitudinally integrated crosssectional profile of the plasma and exhibit heterogeneous plasma behavior and excitation conditions just as do vertical emission profiles in lateral-viewing ICP-AES. Consequently, this cross-sectional emission profile should also be useful for flagging matrix interferences. In this study, the utility of this candidate indicator for axial-viewing ICP-AES is evaluated and compared to the use of a vertical emission profile in lateralviewing ICP-AES. In the companion paper,22 a statistical protocol for automated signaling of interferences is described.

■

EXPERIMENTAL SECTION ICP Spectrometer. A commercial lateral-viewing ICP-AES spectrometer (ACTIVA ICP-AES spectrometer, Horiba JobinYvon, Longjumeau, France) was modified for axial viewing. Figure 1 shows a schematic diagram of the modified optical arrangement. A prototype torch housing for axial viewing, designed and fabricated by Horiba Jobin-Yvon, measures 24 cm × 29 cm × 28 cm and is equipped with ports for optical viewing, radiofrequency power connections, and gas exhausts. The impedance-matching network was demounted from the original lateral-viewing torch box and from the frame of the ICP spectrometer. An elongated torch, with its end extending 15 mm from the top of the load coil, was mounted horizontally in the new axial-viewing housing. The original radiofrequency generator and impedance-matching network were then connected to the axial-viewing torch compartment, which is mounted onto translation stages to simplify positioning. A nitrogen cutoff gas at roughly 25 L/min was applied about 1 cm from the end of the torch to deflect the downstream part of the plasma upward from the optical axis. All experiments reported 51

dx.doi.org/10.1021/ac302095w | Anal. Chem. 2013, 85, 50−57

Analytical Chemistry

Article

emission map of the test element by itself was first measured (referred to below as the reference intensity, Ireference). The analyte emission maps in the presence of selected matrices were then measured in turn and referred to as Imatrix. The effect of a matrix-induced interference was then expressed as relative intensity, defined as the signal from the test element in the presence of the matrix divided by that under the reference conditions (i.e., Imatrix/Ireference). It will be recognized that this ratio (i.e., the relative intensity) will have the same spatial pattern as the determined analyte concentrations in the presence of the interferents.

limited depth of field of the entrance optics, not all of the ICP emission is brought into focus at the entrance slit. Albeit with varying efficiency, emission from these out-of-focus locations was still measured.11 In order to calculate the lens position from the thin-lens formula, the plasma was taken to be a point source located 5 mm downstream from the load coil. Before each set of experiments, the position of the torch box was finely adjusted through the translation stages so emission from the center part of the central channel was focused onto the entrance slit of the spectrometer and registered by the CCD. The inset in Figure 1 displays the end-on view of the plasma and a false-color image of the plasma continuum recorded by the CCD; the x-axis of the CCD detector corresponds to wavelength and the y-axis to cross-sectional location in the plasma. Because the gas flow from the injector cools the plasma,24,25 emission from the center of the plasma is weaker and provides a convenient marker for adjustment of the torch-box position. When the torch box is translated laterally (i.e., left and right) with respect to the entrance slit, different chords of the circular plasma image will fall onto the slit, which alters the false-color display correspondingly. Because the plasma has a donut shape when viewed axially, the vertical length of the cooler, weaker continuum region registered by the CCD (the central bluecolored region on the false-color map in the inset of Figure 1) initially expands, passes through a maximum, and then contracts. The ICP torch-box position that produces the longest region of central-zone emission corresponds to proper alignment; at that position, the diameter of the plasma image can be obtained. Because all studied emission lines were in the UV wavelength range, the spectrometer was set at 250 nm during this alignment process. The torch box was also translated vertically (i.e., up and down with respect to the slit) so that the plasma image falls onto the center of the CCD detector. Furthermore, the distance between the torch box and the focusing lens was fine-tuned for maximum length of weaker plasma-continuum emission of the central channel. Test Elements and Matrices. Sample-preparation procedures for the matrices and test elements were similar to those used in previous experiments with the lateral-viewing ICP.15 For studies of plasma-related and sample introduction-related matrix effects, six elements were selected as analytes. The concentrations of these elements were 0.5 mg/L of Mg, 1 mg/L of Mn, 10 mg/L of Fe, 15 mg/L of Cd, and 25 mg/L of Zn and Pb. Barium, selected because of its moderately serve matrix effect in ICP-AES,6,26,27 was added at concentrations of 0.05 or 0.1 M for experiments involving plasma-related matrix interferences; all matrix and analyte solutions were prepared and diluted with 2% v/v HNO3. For sample introductionrelated matrix-effect experiments, nitric, hydrochloric, sulfuric, and phosphoric acids at various (1%, 2%, 3%, 5%, 10%, and 20%) volumetric concentrations were used. Three analyte− matrix (As−Cd, Er−Co, and Er−Ce) pairs were chosen for the investigation of spectral interferences. Experimental Procedures. Experimental procedures were identical to those employed in the lateral-viewing mode study.15 For each spectral window, the intensity map of the whole 2D CCD (cf. inset, Figure 1) was recorded. All emission lines within a single spectral window of the CCD detector were measured simultaneously, and different spectral windows were collected sequentially. Emission intensities were obtained by integrating the area under each spectral peak with background subtraction taken from nearby pixels. The cross-sectional

■

RESULTS AND DISCUSSION Plasma-Related Matrix Interferences. Figure 2a shows the emission intensity for the Mg II 280.27 nm line at different cross-sectional locations in the plasma with a high injector-gas flow rate (1.5 L/min) and in the absence and presence of a Ba matrix. In the presence of the Ba matrix, a depression in emission clearly occurs in the center of the plasma, whereas an enhancement is registered on both sides of the central channel.

Figure 2. Cross-sectional spatial emission profiles of the Mg II 280.27 nm line in the absence and presence of a 0.1 M Ba matrix at a total central-channel gas flow (including the 0.10 L/min sheath gas flow) of (a) 1.50 L/min and (b) 0.90 L/min. The plasma power was 1100 W, and the injector diameter was 3 mm. 52

dx.doi.org/10.1021/ac302095w | Anal. Chem. 2013, 85, 50−57

Analytical Chemistry

Article

This change in matrix interference along the cross-sectional dimension of the plasma forms the basis of a warning indicator for the interference. For measurements taken at spatial regions where there are matrix-induced signal enhancements (here on the outer edge of the ICP), the determined concentrations of the analyte will be higher than the true value. In contrast, the signal-depression effects in the center of the discharge cause the determined concentration of the analyte to be lower than the true value. Overall, the determined concentration of the analyte changes along the cross-sectional dimension in the plasma. This curvature in the determined concentration along the measurement axis of the plasma is very effective for flagging matrix interferences and can serve as a warning to the analyst. The presence of both matrix-induced enhancement and depression effects results in a spatial location with no apparent matrix interference (at cross-sectional locations ± 0.9 mm in Figure 2a), which is termed the matrix-effect crossover point.17,28,29 However, it is not a prerequisite that the matrix induces both enhancement and depression effects (i.e., the presence of a crossover point is not a necessary condition); the spatial emission profile will flag the interference as long as the magnitude of the interference expresses some spatial dependency, regardless of the direction of the effect. If a matrix-effect crossover point exists and its location is known, an analyst can obtain accurate analytical results by taking analytical data only at the crossover location and neglecting all other spatial locations. Lack of a crossover point will not compromise the ability of the cross-sectional profile to flag interferences, but it will not be able to indicate a location where interference-free analytical results can be acquired. The example in Figure 2a represents an extreme case because the shape of the cross-sectional profile was modified. The profile originally was bell-shaped but transformed to that of a dumbbell in the presence of the Ba matrix; the maximum emission was no longer at the center of the cross-sectional profile but shifted to the two sides. Figure 2b shows the same Mg II cross-sectional emission profiles but obtained with a much lower injector-gas flow of 0.9 L/min. At this lower flow, the general shape of the emission profile remains unchanged, but it is still very clear that depression and enhancement effects arise at the center and on the sides of the central channel, respectively. The observation that emission is attenuated in the center but grows on the sides of the plasma is in agreement with the behavior reported by Olesik and Williamsen.30,31 Of course, in a routine analysis, an operator would not have access to the spatial pattern of analyte emission in the absence of the sample matrix (i.e., the trace with square-shaped symbols in Figures 2a and 2b). However, the operator would be able to obtain working curves at all the lateral positions in the ICP and could determine the apparent analyte concentration in the sample at each of those locations. In the absence of any interference, the determined concentrations would then all be the same, regardless of spatial location, because the sample and standards would behave similarly. In contrast, those calculated concentrations would differ when an interference is present. The resulting pattern of concentrations will be termed here the “relative intensity” because it can be derived merely by dividing the two emission-intensity curves (i.e., a one-point calibration) in Figures 2a or 2b. Figure 3 shows such relative intensity patterns for the Mg II 280.27 nm emission line in the presence of a 0.1 M Ba matrix at injector-gas flow rates from 0.70 to 1.50 L/min. Noticeable curvature, which allows the interference to be recognized, is

Figure 3. Cross-sectional relative-intensity profiles of Mg II 280.27 nm emission in the presence of a 0.1 M Ba matrix as a function of the total central-channel gas flow at 1100 W plasma power. Injector diameter was 3 mm.

observed along the cross-sectional profile of the plasma for all studied gas flows. This behavior is in contrast to that in conventional side-viewing ICP-AES, which requires a slightly higher central-channel gas flow to push the vertical emission profile upward above the load coil.15,32 For the same ICP system as the present one but operated in a side-on mode, it was found15 that central-channel flows of 1.00 and 1.25 L/min (compared to the typical 0.8 L/min) are needed for injector diameters of 2 and 3 mm, respectively. Therefore, low carriergas flows, which typically favor a robust plasma, can be used for flagging matrix interferences in axial-viewing ICP-AES. Figure 4 shows the relative intensities for a pool of emission lines in the presence of a 0.05 M Ba matrix along the crosssectional axis of the plasma. All studied emission lines show curvature in their cross-sectional relative-intensity profiles,

Figure 4. Cross-sectional relative-intensity profiles of selected emission lines in the presence of 0.05 M Ba matrix. The total central-channel gas flow rate was 0.90 L/min; the plasma power was 1100 W, and the injector diameter was 3 mm. 53

dx.doi.org/10.1021/ac302095w | Anal. Chem. 2013, 85, 50−57

Analytical Chemistry

Article

is more than 5%, the cross-sectional profiles showed clear curvature, allowing the interference to be flagged. Spectral Interferences. Similar to the vertical emission profile in lateral-viewing ICP-AES,15 the cross-sectional emission profile is able to flag spectral interferences in the axial-viewing mode. The principle is based on the fact that spectral-line emission is a strong function of plasma temperature as well as excitation and ionization potentials of the analyte element. Because the excitation and ionization potentials of the chosen emission lines of the analyte are almost always different from those of a spectrally interfering element, their normalized spatial emission profiles will show some differences if the span of plasma excitation conditions along the cross-sectional axis is large. As will be shown below, although different spatial emission patterns are observed for many spectrally interfering line pairs, some emission-line pairs do not exhibit large enough differences in their cross-sectional emission profiles to allow the interference to be flagged. This is especially the case when the excitation and ionization potentials of the line pair are close. Figure 6a shows the normalized cross-sectional spatial emission patterns of the As I 228.81 nm and Cd I 228.80 nm lines, which are clearly distinct from each other. Depending on the analyte of interest, either the pure-As or the pure-Cd profile could be used as the calibration standard, with the other considered the interferent. If a sample contains both As and Cd (i.e., with mutual spectral interference), the cross-sectional emission profile of this As−Cd sample will be somewhere between the pure-As or pure-Cd emission patterns, which will be different from the calibration standard. Consequently, the determined concentration behavior of the analyte will express some degree of spatial dependency, which in turn will alert the analyst to the presence of a spectral interference. Figure 6b shows the spatial cross-sectional relative-intensity profile for the As I 228.81 nm line in samples containing Cd at concentrations from 0.01 to 0.3 mg/L. Spectral interference from Cd at a concentration of 0.01 mg/L is negligible, as evidenced by the fact that the relative intensities are practically unity. Visual inspection (cf. Figure 6b) of the cross-sectional profile indicates the lack of interference. At a Cd concentration of 0.03 mg/L, the curvature of the cross-sectional profile already becomes noticeable (cf. Figure 6b). Therefore, the interference can be recognized with confidence at this Cd concentration, even though the analytical error (∼ 3%) is still comparatively minor. In the presence of 0.1 mg/L Cd, the spectral interference causes an averaged error of about 7%. At this Cd concentration, the curvature in the cross-sectional profile is very clear. Of course, the spectral interferences, as well as the curvatures, are even greater at higher Cd concentrations. Figure 6c shows the cross-sectional relative-intensity profile for the Er II 239.73 nm line in a sample with Co added from 0.3 to 10 mg/L. In this case, Co II 239.74 nm is the spectrally interfering line. The analytical bias caused by Co at a concentration of 0.3 mg/L is only about 1% and is considered negligible. At a Co concentration of 1 mg/L, the spectral interference induces an average error of about 5%, and the cross-sectional profile exhibits marginal curvature on the two edges. In the presence of 3 mg/L Co, the spectral interference causes an average error of about 18%. At this Co concentration, the curvature in the Er cross-sectional profile is very clear (cf. Figure 6c). As a result, the interference can be easily recognized. This Er−Co line-pair example demonstrates that the cross-sectional profile is able to flag spectral interference

which allows easy recognition of the interferences. Figure 4 also shows that not all emission lines exhibit crossover points. The relative intensities of some emission lines (e.g., Cd I 228.80 nm, Pb II 220.35 nm, Fe I 371.99 nm, and Mg I 285.21 nm) lie above 1 along the entire cross-sectional profile. This is in strong contrast to the case of conventional lateral viewing, in which all these emission lines exhibit clear crossover points in their vertical emission profiles.15 Sample Introduction-Related Matrix Interferences. The ability of the cross-sectional emission profile to flag interferences caused by the sample introduction system was demonstrated by addition of mineral acids to the sample. In all experiments, the reference standard was prepared in 1% v/v HNO3, whereas the sample contained an additional mineral acid (nitric, hydrochloric, sulfuric, or phosphoric) at a variable concentration. Figure 5 shows the cross-sectional relative-intensity profiles of the Mn II 259.37 nm line in samples containing sulfuric acid

Figure 5. Cross-sectional relative-intensity profiles of the Mn II 259.37 nm line in different additional concentrations of sulfuric acid. The reference sample was prepared in 1% v/v nitric acid. The total central-channel gas flow rate was 0.80 L/min; the plasma power was 1100 W, and the injector diameter was 3 mm.

at different concentrations. Because the density and viscosity of sulfuric acid is much higher than that of water, the sample introduction-related matrix effect from sulfuric acid is strong.33,34 Similar to the case of lateral-viewing ICPAES,15,33 even 1% v/v sulfuric acid caused a noticeable depression in emission intensities (cf. Figure 5). Fortunately, and again just as in lateral-viewing ICP-AES, all the spatial emission profiles showed curvature, which allow for the identification of the acid effects. The result of the drift test is also included in Figure 5. The drift test serves two purposes: it represents the typical repeatability of the cross-sectional profile and ensures that any observed matrix effect is not due to the artifacts of system drift during the course of the experiment. The drift test is performed by measurement of the calibration standard (i.e., no matrix) after the matrix-effect experiment, which typically lasts about 45 min to an hour. Three other common mineral acids (nitric, hydrochloric, and phosphoric acids) were also studied. As long as the matrix effect 54

dx.doi.org/10.1021/ac302095w | Anal. Chem. 2013, 85, 50−57

Analytical Chemistry

Article

Figure 6. (a) Normalized cross-sectional spatial emission profiles of the As I 228.81 nm and Cd I 228.80 nm lines. (b) Cross-sectional relativeintensity profiles of the As I 228.81 nm line under the influence of coexisting Cd from 0.01 to 0.3 mg/L. Concentration of As was 20 mg/L. (c) Cross-sectional relative-intensity profiles of the Er II 239.73 nm line under the influence of coexisting Co from 0.3 to 10 mg/L. The concentration of Er was 200 mg/L. (d) Cross-sectional relative-intensity profiles for the Er II 302.27 nm line under the influence of coexisting Ce from 3 to 100 mg/ L. The concentration of Er was 200 mg/L. Plasma operating parameters were identical to those listed in the caption of Figure 5.

which their total excitation potentials are very close, was selected to illustrate this limitation. The total excitation energies of the Ce−Er line pair at 302.27 nm differ by less than 0.7 eV. Figure 6d shows the relative intensity behavior for the Er II 302.27 nm line in samples containing Ce at concentrations from 3 to 100 mg/L. In all these cases, the cross-sectional profiles appear flat. Accordingly, the cross-sectional profiles are unable to flag the spectral interference from this emission line pair. Interestingly, the vertical-emission profile was able to flag the spectral interference from this line pair in conventional lateral-viewing ICP-AES.15 Because a slightly higher central-channel gas flow is needed for flagging matrix interferences in side-on viewing, similar studies, but with high gas flow rates, were also performed here for axial viewing. However, the cross-sectional profiles in axial viewing were still unable to flag this specific spectral interference, even at central-channel flows of 1.00 or 1.30 L/min.

from emission lines that are both from singly charged ions, as long as the total excitation potentials of the two spectral lines are not too close. The difference in total excitation potentials of this line pair is about 2.1 eV. The normalized cross-sectional profiles of these two ionic emission lines (not shown) indicate that the two profiles are different on the two edges of the central channel, probably due to differences in their excitation energies and the spatially dependent plasma excitation characteristics. However, if the excitation and ionization energies of the spectral-line pairs are too close, their normalized cross-sectional emission profiles become similar. Because the principle for flagging spectral interference depends on different behavior in their spatial emission profiles, the response in the relativeintensity (and determined-concentration) profile will then appear flat even if spectral interference is present. As a result, the interference cannot be recognized. The Ce−Er line pair, in 55

dx.doi.org/10.1021/ac302095w | Anal. Chem. 2013, 85, 50−57

Analytical Chemistry

■

ACKNOWLEDGMENTS The authors are grateful to Horiba Jobin-Yvon (Longjumeau, France) for the loan of the ACTIVA ICP spectrometer used in this work. This research was supported by the U.S. Department of Energy through Grant DE-FG02-98ER14890.

The normalized vertical emission profiles of the Er II and Ce II lines at 302.27 nm were shown to exhibit some dissimilarity in the lateral-viewing mode;15 however, when the plasma is integrated along its apex for end-on observation, the normalized cross-sectional profiles offer no measurable difference. Inspection of the previously published15 vertical emission profiles of the two lines reveals that the normalized Er II emission is stronger than the Ce II line in lower regions of the plasma but weaker at higher observation locations. It appears that axial viewing, in which the whole central channel is longitudinally integrated, averages out the differences in spatial behavior of the two emission lines. Overall, spatial emission profiles for flagging matrix interferences in the axial-viewing mode are slightly less sensitive than those in the conventional lateral-viewing mode.

■

REFERENCES

(1) Mermet, J. M.; Cosnier, A.; Velasquez, S.; Lebouil, S. Am. Lab. (Shelton, CT, U.S.) 2007, 39, 16−17. (2) Iglesias, M.; Vaculovic, T.; Studynkova, J.; Poussel, E.; Mermet, J. M. Spectrochim. Acta, Part B 2004, 59, 1841−1850. (3) Stepan, M.; Musil, P.; Poussel, E.; Mermet, J. M. Spectrochim. Acta, Part B 2001, 56, 443−453. (4) Dennaud, J.; Howes, A.; Poussel, E.; Mermet, J. M. Spectrochim. Acta, Part B 2001, 56, 101−112. (5) Dubuisson, C.; Poussel, E.; Mermet, J. M. J. Anal. At. Spectrom. 1998, 13, 1265−1269. (6) Dettman, J. R.; Olesik, J. W. Spectrochim. Acta, Part B 2012, 76, 96−108. (7) Mermet, J. M. J. Anal. At. Spectrom. 2005, 20, 11−16. (8) Cano, J. M.; Todoli, J. L.; Hernandis, V.; Mora, J. J. Anal. At. Spectrom. 2002, 17, 57−63. (9) Stewart, I. I.; Olesik, J. W. J. Anal. At. Spectrom. 1998, 13, 1249− 1256. (10) Chan, G. C.-Y.; Chan, W. T. Spectrochim. Acta, Part B 2003, 58, 1301−1317. (11) Murtazin, A.; Groh, S.; Niemax, K. Spectrochim. Acta, Part B 2012, 67, 3−16. (12) Boumans, P. W. J. M. Fresenius' J. Anal. Chem. 1986, 324, 397− 425. (13) Mermet, J. M.; Trassy, C. Spectrochim. Acta, Part B 1981, 36, 269−292. (14) Schiavo, D.; Trevizan, L. C.; Pereira, E. R.; Nobrega, J. A. Spectrochim. Acta, Part B 2009, 64, 544−548. (15) Chan, G. C.-Y.; Hieftje, G. M. J. Anal. At. Spectrom. 2008, 23, 193−204. (16) Chan, G. C.-Y.; Hieftje, G. M. J. Anal. At. Spectrom. 2008, 23, 181−192. (17) Blades, M. W.; Horlick, G. Spectrochim. Acta, Part B 1981, 36, 861−880. (18) Brenner, I. B.; Zander, A. T. Spectrochim. Acta, Part B 2000, 55, 1195−1240. (19) Ivaldi, J. C.; Tyson, J. F. Spectrochim. Acta, Part B 1995, 50, 1207−1226. (20) Brenner, I. B.; Le Marchand, A.; Daraed, C.; Chauvet, L. Microchem. J. 1999, 63, 344−355. (21) Sengoku, S.; Wagatsuma, K. Anal. Sci. 2006, 22, 245−248. (22) Chan, G. C.-Y.; Hieftje, G. M. Anal. Chem. 2012, DOI: 10.1021/ ac3022345. (23) Mermet, J. M.; Cosnier, A.; Danthez, Y.; Dubuisson, C.; Fretel, E.; Rogerieux, O.; Velasquez, S. Spectroscopy 2005, 20, 60−66. (24) Demers, D. R. Appl. Spectrosc. 1979, 33, 584−591. (25) de Loos-Vollebregt, M. T. C.; Tiggelman, J. J.; de Galan, L. Spectrochim. Acta, Part B 1988, 43, 773−781. (26) Chan, G. C.-Y.; Hieftje, G. M. Spectrochim. Acta, Part B 2006, 61, 642−659. (27) Chan, G. C.-Y.; Chan, W. T.; Mao, X. L.; Russo, R. E. Spectrochim. Acta, Part B 2001, 56, 77−92. (28) Chan, G. C.-Y.; Hieftje, G. M. J. Anal. At. Spectrom. 2010, 25, 282−294. (29) Chan, G. C.-Y.; Hieftje, G. M. Spectrochim. Acta, Part B 2008, 63, 355−366. (30) Olesik, J. W.; Williamsen, E. J. Appl. Spectrosc. 1989, 43, 933− 940. (31) Olesik, J. W.; Williamsen, E. J. Appl. Spectrosc. 1989, 43, 1223− 1232. (32) Chan, G. C.-Y.; Hieftje, G. M. J. Anal. At. Spectrom. 2009, 24, 439−450.

■

CONCLUSIONS Similar to the vertical emission profile in lateral-viewing ICPAES, in general, the spatially resolved cross-sectional emission profile in axial-viewing ICP-AES is an effective indicator to flag the presence of matrix interferences. A flat, determined concentration profile along the cross-sectional axis of an axialviewing ICP indicates an absence of matrix interference, whereas a dissimilar concentration profile offers a clear warning signal that the analytical results are compromised by interferences caused by any one of the three common matrixeffect categories: plasma-related, sample introduction-related, and spectral interferences. This study utilized an atypical commercial ICP-AES spectrometer, which has the capability of performing simultaneous measurements of the whole plasma spatial emission profile; however, similar measurements can be performed on the more common linear-dispersion or crossdispersion Echelle-grating ICP-AES spectrometers. Instead of a slit, an entrance aperture in the form of a pinhole is usually used in Echelle-grating spectrometers; as a result, only one spatial location in the plasma is measured. However, with simple modification of the entrance optics (e.g., installation of a tilting mirror), different spatial locations of the plasma can be focused onto this entrance aperture, thereby achieving spatially resolved measurements through sequential step-tilting of the mirror. Also, although many commercial ICP-AES spectrometers allow data acquisition at only a single spatial location in the plasma, they are usually equipped with options that permit this observation region to be changed (e.g., through physical movement of the torch via a translation stage or tilting of the optics) either manually or automatically through computer control. The purpose of such observation-volume selection is ordinarily intended for optimization of measurement parameters (e.g., highest signal-to-background ratio, lowest detection limit, best precision), but such controls can clearly be exploited to obtain spatial emission profiles such as those needed to flag interferences in the present study.

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +1 812 855 7905. Fax: +1 812 855 0985. Notes

The authors declare no competing financial interest. 56

dx.doi.org/10.1021/ac302095w | Anal. Chem. 2013, 85, 50−57

Analytical Chemistry

Article

(33) Paredes, E.; Maestre, S. E.; Todoli, J. L. Spectrochim. Acta, Part B 2006, 61, 326−339. (34) Todoli, J. L.; Mermet, J. M. Spectrochim. Acta, Part B 1999, 54, 895−929.

57

dx.doi.org/10.1021/ac302095w | Anal. Chem. 2013, 85, 50−57