Atomization and vaporization of lead from a blood matrix using rhodium-coated tungsten filaments with pseudo-simultaneous electrothermal atomic absorption and inductively coupled plasma mass spectrometric measurements

Abstract

In an effort to understand the pyrolysis and atomization behavior of Pb in a dilute blood matrix when determined using a tungsten filament electrothermal atomic absorption spectrometer (W-filament ETAAS), we directly coupled a W-filament ETAAS instrument (J. Anal. At. Spectrom., 2001, 16, 82) to an inductively coupled plasma mass spectrometer (ICP-MS). The W-filament instrument was operated both as an electrothermal atomizer (ETA) and as an electrothermal vaporizer (ETV). The experimental arrangement was used to monitor simultaneously the analyte (²⁰⁸Pb), the permanent modifier (¹⁰³Rh) coating the W surface, a surrogate for the organic components of the blood matrix (⁴⁰Ar¹²C), and the metal atomizer (¹⁸⁰W, ¹⁸³W and ¹⁸⁴W¹⁶O), while also measuring Pb by AAS pseudo-simultaneously. Electrothermal parameters used in the ETV/ETA experiments were precisely the same as those described previously (Spectrochim. Acta, Part B, 2002, 57, 727) for ETAAS work. This was necessary in order that the ICP-MS data could be used to observe the vaporization of analyte, matrix, modifier, and metallic atomizer throughout the electrothermal cycle. The purge gas used to transport vaporized material from the atomization cell to the ICP was Ar containing 6% H₂. Pseudo-simultaneous ETAAS and ETV-ICP-MS measurements of Pb in blood were achieved using this experimental arrangement, with the MS signals appearing roughly 2 s after those by AAS, consistent with the delay caused by the 15 cm long Teflon transfer line connecting the two instruments. During pyrolysis, a strong signal observed at mass 52 is shown to be due to ⁴⁰Ar¹²C; this signal increases as filament power for pyrolysis increases. When plotted, the data show the effective removal of much of the carbon-based matrix during pyrolysis, but without significant loss of analyte. In contrast, the pre-coated permanent modifier, Rh, is only lost during the final cleaning stage of the program in which maximum power is necessary to remove the carbonaceous matrix. However, when Rh is added to the diluent, it is lost even during pyrolysis, and at a much lower temperature. This explains why addition of Rh directly to the diluent yields no modification benefits, and why periodic re-coating of the filament surface for the analysis is necessary.

---

Introduction

Electrothermal vaporization (ETV) was first coupled to inductively coupled plasma mass spectrometry (ICP-MS) by Gray and Date.¹ As an alternative sample introduction technique, ETV is inherently more sensitive than are traditional nebulizer-spray chamber combinations, because sample transfer into the plasma is much more efficient.² However, the improvements in detection limits when using ETV are not that substantial due to the poorer precision that is associated with transient signal measurement and the poorer reproducibility of sample delivery systems in the ETV. Nevertheless, ETV-ICP-MS is particularly suitable for analyses in which only a limited amount of sample is available, and even solid sample analysis is possible using slurry techniques. Another advantage of ETV-ICP-MS is the freedom for the analyst to optimize a thermal pre-treatment program, thus making it possible to utilize chemistry within the ETV to remove the matrix and to conduct in situ preconcentration, in situ preparation, and even in situ speciation, before the sample enters the plasma. Sturgeon and Lam³ have recently provided a thorough review of this subject.

The graphite furnace is currently the most widely used atomizer for electrothermal atomic absorption spectrometry (ETAAS) and vaporizer for ETV-ICP-MS. Most commercially available ETV systems for ICP-MS instruments are actually modified graphite furnace AAS atomizers. Instead of being used as the atomizer (as it is in ETAAS), the graphite furnace serves as a vaporizer in ETV-ICP-MS, whereby the sample is dried, pyrolyzed, vaporized and subsequently transported into the plasma. Atomization is not required, since the analyte need only be vaporized and introduced by the carrier gas stream into the plasma where atomization and ionization processes occur. Thus, vaporization temperatures required in ETV-ICP-MS are typically lower than the atomization temperatures used in graphite furnace ETAAS.

ETV-ICP-MS arrangements have also been used for some fundamental studies of vaporization and atomization processes that occur within the graphite furnace. Such studies are important not only for determining the optimum experimental conditions required for graphite furnace ETV-ICP-MS, but they may also be useful for probing atomization mechanisms in graphite furnace ETAAS.

Lamoureux et al.^4,5 used a modified commercial graphite furnace atomizer coupled to an ICP-MS to record pseudo-simultaneous AAS and ICP-MS measurements. This instrument arrangement was employed to investigate a MgCl₂ interference reported for Mn determinations by graphite ETAAS.⁴ Although this approach is a potentially powerful tool for elucidating fundamental processes within the graphite furnace, conditions within the ETV device are not quite comparable to those in ETAAS. For example, during atomization in graphite furnace ETAAS, the Ar gas flow is normally stopped to prolong the residence time of ground state atoms in the optical path, thus improving sensitivity. This is considered a fundamental aspect of the stabilized temperature platform furnace (STPF) concept.⁶ However, during the vaporization step in graphite furnace ETV, Ar gas is continually purged to transport the vaporized sample into the plasma. Lamoureux et al.^4,5 used only a very small gas flow rate (50 mL min⁻¹) in their ETV experimental arrangement to transfer vaporized material combined with a make-up gas to achieve a total Ar flow rate of 1 L min⁻¹ into the plasma. In a series of reports, Grégoire and his co-workers used a commercial graphite furnace ETV coupled with an ICP-MS to investigate the vaporization (atomization) of several analytes including W,⁷ B,⁸ Y and rare earth elements,⁹ U,¹⁰ the Pt group elements,¹¹ and Ra and alkaline earth elements.¹² Some studies focused on the vaporization process itself^7,9,12 while others used ETV data to elucidate the atomization mechanisms in graphite furnace AAS.^8,10,11

Metal surfaces, such as W-filaments and the W-tube furnace, have also been used as ETV devices for ICP-MS. Park et al.¹³ used a W or a Re metal ribbon filament as an ETV device with a 5 µL sample capacity. Hauptkorn et al.¹⁴ reported coupling a W-filament ETV device to an ICP-MS to determine trace impurities in high-purity quartz samples using a slurry sampling technique. Their W-filament ETV-ICP-MS arrangement was also utilized to identify combinatorial chemistry compounds from elemental tagging procedures used in drug discovery research.¹⁵ Other groups report using a W boat furnace as an ETV device for ICP-MS.^16,17 The W boat furnace was used for the determination of rare-earth elements in aqueous solution,¹⁸ metal impurities in sulfamic acid,¹⁹ Pb in biological materials,¹⁷ Pb in rock samples,²⁰ and F in aqueous samples.²¹ It has also been reported that strips of W, Ta, Mo or Re have been inserted in a graphite tube and used as vaporization platforms in ETV-ICP-MS.^22,23

We have previously described the feasibility of using W-filaments as electrothermal atomizers for the determination of Pb in blood by AAS.²⁴ A compact prototype AAS instrument equipped with a novel W-filament and a self-reversal background correction system was assembled specifically for blood Pb measurements and was shown to meet current US regulatory requirements for this analysis.²⁵ Further improvements were reported more recently using a Rh-coated W-filament.²⁶ However, the Rh coating is not permanent, and periodic re-coating is required, thus diminishing its utility in non-technical hands. The principal objective of this study was to explore the pyrolysis and atomization behavior of Pb from Rh-coated W filaments by directly coupling a prototype W-filament AAS instrument to an ICP-MS. Using this experimental arrangement made it possible to monitor simultaneously (a) the analyte (²⁰⁸Pb), (b) the permanent modifier (¹⁰³Rh) that coats the W-filament, and (c) a possible surrogate for the organic components of the blood matrix (⁴⁰Ar¹²C) that appears at mass 52, while also measuring Pb pseudo-simultaneously by AAS.

Experimental

Instrumentation

A prototype compact W-filament AAS instrument (Model PBA-10, Exeter Analytical, Inc.,†North Chelmford, MA, USA) designed specifically for blood Pb measurements was used for AAS measurements. For research and development purposes, the prototype instrument was interfaced to a personal computer (DEC 316sx) running a proprietary software program provided by the manufacturer that runs under MS DOS. Complete details of the W-filament AAS instrumentation have been described previously.²⁴ In this study, we used the H-1431 W-filaments obtained unmounted from Osram Inc. (Waldoboro, ME, USA) as atomization devices. These filaments, which we previously called the “long-wire filaments”, were found to be the more suitable filament for blood Pb determinations.²⁴ A Rh coating was applied to these W-filaments to provide long-term protection, and as the permanent modifier for blood Pb measurements by AAS. The Rh coating procedure and the optimized heating program used here are identical to those reported previously.²⁶

For the purposes of this study, the PBA-10 prototype AAS instrument was interfaced to an ELAN 6100 DRC^® inductively coupled plasma mass spectrometer (ICP-MS) (PerkinElmer Sciex, Shelton, CT, USA) running version 2.3.2 of the ELAN software. Although the ELAN 6100 DRC is equipped with dynamic reaction cell (DRC) technology for reducing polyatomic interferences in ICP-MS, this feature was not necessary for this work; therefore, the ELAN was operated in standard, i.e., non-DRC, mode only. The ICP-MS instrumentation in our laboratory is also equipped with an HGA 600MS graphite furnace ETV and an AS 60 autosampler (PerkinElmer Instruments, Inc., Shelton, CT, USA). Initial optimization of the ICP-MS was carried out using a cross-flow nebulizer (PE Part # N8120516), a Scott-type spray chamber (PE Part # N8120124), and aqueous tuning solutions. No further optimization was performed after switching to ETV mode. The carrier gas flow rate was fixed at 1 L min⁻¹ (LPM), consistent with the flow rate used in the W-filament AAS cell.

On the ICP-MS, peak-hopping mode was used for data acquisition, and the optimum dwell time was found to be 10 ms. The length of the actual read time was determined by the number of readings per replicate. Integrated ion intensities, rather than the maximum ion counts, were used for ICP-MS data evaluation. No more than six isotopes were measured in a single ETV event to ensure reliable integration of the transient signals. On the PBA-10 instrument, AAS data were collected pseudo-simultaneously as integrated absorbance (A_i) measurements. The operating parameters of the ICP-MS and the PBA-10 instrumentation are listed in Table 1 unless otherwise stated.

Table 1 Instrumental details for the W-filament ETV-ICP-MS arrangement

PE SCIEX ELAN 6100 DRC^® ICP-MS
ICP rf power:	1100 W
Lens voltage:	7.65 V
ICP supporting gas (Ar):	15 LPM
Auxiliary gas (Ar):	1.1 LPM
Carrier gas (6% H2 in Ar):	1.0 LPM
Dwell time:	10 ms
Scan mode:	Peak hopping
Signal measurement:	Integrated counts
Ions monitored:	^208Pb, ^40Ar^12C, ^13C, ^103Rh, ^183W
EAI PBA 10 W-filament ETV/ETA
Sample volume:	15 µL
Matrix modifier for W-filament ETAAS:	Rh permanent modifier
Heating program:	Dry—160 s at 0.9 W
	Pyrolysis—30 s at 2.6 W
	Cool—10 s at 0 W
	Atomization—2 s at 30 W
	Clean—20 s at 46 W
Purge gas (6% H2 in Ar):	1.0 LPM
Sample transfer from ETV to ICP-MS:	15 cm Teflon tube (6 mm id)

---

W-filament AAS to ICP-MS interface

To interface the two instruments, we directly connected the purge gas outlet from the PBA-10 atomization cell to the quartz injector tube of the plasma torch using a 15 cm length, 6 mm id Teflon tube. As reported previously, we used a mixture of 6% H₂ in Ar as the purge gas in the W-filament atomization cell, and the same gas was used to transport the vaporized sample (matrix and analyte) into the plasma for analysis by ICP-MS. For the purposes of this experimental arrangement, it was necessary to maintain the IEEE 488 electronic interface between the HGA 600MS ETV system and the ELAN 6100 ICP-MS in order to utilize the ETV module of the ELAN software program for data acquisition. Normally, a read cable between the HGA 600MS and the ELAN 6100 carries a contact closure signal that starts data acquisition on the latter at the designated point in the ETV cycle. In the present experimental arrangement, the read cable was disconnected from the HGA 600MS and connected to the PBA-10, thus providing a contact-closure signal to start data acquisition. In version 4.0 of the PBA-10 software provided by the manufacturer, a contact-closure signal started data acquisition at the beginning of the atomization step. Later on, the software was modified by the manufacturer (version 5.0) so that the contact-closure signal triggered data acquisition at the beginning of the dry step, thus enabling the ICP-MS to monitor selected ions throughout the ETV/ETA cycle. In order to utilize the ETV data acquisition module within the ELAN software, it was necessary to use a “phantom” ETV heating program for the HGA 600MS, thus maintaining IEEE communications between the various systems. Since no sample was injected into the HGA 600MS, a low- temperature (0–50 °C) heating program was implemented; data acquisition was triggered by the PBA-10, thus ensuring synchronization of the W-filament AAS and ICP-MS measurements.

Reagents and sample preparation

Ammonium hexachlororhodate(III), (NH₄)₃[RhCl₆] (Alfa Aesar^®, Ward Hill, MA, USA), was used as the source of Rh for coating the filament. Preparation of the Rh solution and details of the Rh coating procedure have been described previously.²⁶ Briefly, a 1 g L⁻¹ solution of Rh is prepared in 10% v/v HCl, and a 50-µL aliquot (50 µg Rh) is deposited on the filament. The filament is heated electrothermally, using a five-step program in which the power is increased from 0.9 to 4.3 W over 4.5 min. This procedure is repeated three more times, such that a total of 200 µg of Rh are deposited on the filament. Following this, the filament is subjected to a seven-step electrothermal heating program in which the power is eventually increased to 46 W for 30 s. In one experiment, we added 0.133 µg µL⁻¹ Rh (equivalent to 2 µg Rh for each 15 µL sample aliquot) in the diluent, to compare its behavior with the pretreated Rh modifier.

For investigations of the W-filament ETAAS coupled to ETV-ICP-MS, we used a New York State Blood Reference Material (NYS-RM Lot 049) with a certified Pb concentration of 42.2 ± 1.8 µg dL⁻¹.‡ These RMs consist of lyophilized whole blood that is obtained from Pb-dosed animals and are certified through interlaboratory studies.²⁷ The lyophilized materials are reconstituted using double-deionized water, and an aliquot is diluted 1 + 4 with a diluent containing 0.5% (v/v) Triton X-100 and 0.2% (v/v) HNO₃. Diluted blood samples (15 µL) are manually deposited on the W-filament using a double-action micropipette (Finnpipette, Helsinki, Finland) or injected into the graphite furnace ETV using the AS-60 autosampler.

Results and discussion

W-filament ETV-ICP-MS optimization studies

Our initial experiments confirmed previous observations by other workers²⁸ that a shorter connecting line between the ETV device and the ICP-MS yields better transfer efficiency, and thus greater sensitivity (measured as ²⁰⁸Pb). Accordingly, we used a 15 cm Teflon tube (6 mm id), i.e., the minimum length that can accommodate the two instruments, to connect the gas outlet from the W-filament atomization cell to the plasma-torch injector tube.

The vertical position of the W-filament atomizer relative to the plasma-torch injector tube was also optimized with respect to ²⁰⁸Pb intensity. Although the integrated counts for ²⁰⁸Pb did not appear to be very sensitive to this parameter, we did observe a small improvement in sensitivity (∼10%) when the atomization cell outlet and the injector tube inlet were aligned, i.e., when the Teflon tube connecting the two instruments was level. However, this improvement may not be statistically significant.

We also varied the id of the Teflon tube connecting the atomization cell to the ICP-MS. Replacement of a 1.5 mm id Teflon tube with a 6 mm id diameter tube markedly improved the sensitivity for ²⁰⁸Pb by ICP-MS. This is consistent with observations reported in the literature for graphite furnace ETV instrumentation.²⁸

Effect of 6% H₂ in the Ar carrier gas on ²⁰⁸Pb sensitivity

After optimization of the interface configuration, the intensity of the ²⁰⁸Pb ICP-MS signals obtained using the W-filament as an ETV device were still an order of magnitude lower than in our previous experience using the HGA 600MS graphite furnace ETV system. A major difference between these two ETV systems, other than the choice of vaporizer materials, is the composition of the purge/carrier gas. With the HGA 600MS system, the graphite furnace is normally purged with pure Ar, and the vaporized sample is carried into the plasma by a stream of pure Ar gas. With the W-filament arrangement, the atomization cell is typically purged with a mixture of 6% H₂ in Ar; this purge gas is also used to transport the vapor generated into the plasma. This difference prompted us to investigate H₂ as the cause of the reduced sensitivity observed for ²⁰⁸Pb with the W-filament ETV-ICP-MS arrangement.

The HGA 600MS furnace ETV system was coupled to the ICP-MS using a 6 mm id Teflon connecting tube of a similar tube length. We compared pure Ar with a mixture of 6% H₂ in Ar as the carrier gas using 15 µL of 1 + 4 diluted NYS RM Lot 049 as the test solution. Sensitivity was monitored at mass 208 and evaluated using integrated ion counts. Addition of 6% H₂ to the Ar carrier gas suppressed the sensitivity for ²⁰⁸Pb by an order of magnitude, from 7.5 × 10⁶ counts·s to 5.1 × 10⁵ counts·s.

We repeated this experiment, comparing pure Ar with 6% H₂ in Ar as the carrier gas, with the W-filament ETV system. Again, the integrated ion intensity for ²⁰⁸Pb, using the same diluted blood sample, was reduced by around an order of magnitude, from 1.5 × 10⁷ counts·s with pure Ar to 1.5 × 10⁶ counts·s with 6% H₂ in Ar as the carrier gas. In a third experiment, we used direct nebulization sample introduction into the ICP-MS with 6% H₂ in Ar as the intermediate carrier gas. The sensitivity for ²⁰⁸Pb in the same 1 + 4 diluted blood sample was reduced by four orders of magnitude, from 2 × 10⁷ counts with pure Ar to around 8 × 10³ counts with 6% H₂ in Ar. Thus, introduction of as little as 6% H₂ in Ar into an Ar plasma ICP-MS greatly reduces the sensitivity for ²⁰⁸Pb, regardless of the sample introduction arrangement.

The effects of molecular gases such as H₂ on Ar plasmas were previously reported by Sesi et al.²⁹ Different impacts were described, depending on whether the molecular gases were added to the plasma outer gas, the auxiliary gas, or the intermediate carrier gas. Sesi et al.²⁹ argue that addition of up to 16.7% H₂ in the Ar carrier gas increases both the electron temperature and the gas-kinetic temperature, thus providing a hotter plasma due both to the much higher thermal conductivity of H₂ than Ar and to the constricted plasma volume after H₂ is added. In contrast, these authors also contend that the concentration of electrons in the plasma is decreased. Sesi et al.²⁹ explain these seemingly contradictory observations as the thermal pinch effect of the plasma. That is, when H₂ is added to the carrier gas, the plasma begins to move away from the torch injector tube, i.e., the distance between the tip of the injector tube and the bottom of the plasma increases. They suggest that the increase in gas temperature induced by addition of H₂ is not sufficient to compensate for the increased residence time of the sample aerosol in the “cooler” torch region, and thus ionization, and consequently sensitivity, is suppressed. Our observations here of reduced sensitivity for ²⁰⁸Pb upon introduction of 6% H₂ to the Ar carrier gas are consistent with those reported by Sesi et al.²⁹

In W-filament AAS, addition of H₂ to the purge gas is considered desirable because it enhances sensitivity for many elements by promoting a reducing environment that improves the signal-to-noise ratio,³⁰ as well as protecting the W-filament from oxidation, thus extending its lifetime. In this study, Pb absorbance signals obtained using a W-filament AAS instrument show a modest improvement (∼10%) in sensitivity, from 264 to 238 pg, upon addition of 6% H₂ to the purge gas. Since a key objective in interfacing a W-filament instrument to the ICP-MS was to investigate the atomization/vaporization processes of Pb in blood, it was necessary to maintain identical conditions to those used in previous ETAAS studies, including the use of 6% H₂ in Ar as the purge gas even though this resulted in less than maximal sensitivity for ²⁰⁸Pb by ICP-MS.

Unlike in graphite furnace ETAAS, in which the Ar purge gas flow is stopped during atomization, the purge gas in W-filament ETAAS flows continuously during the atomization event. Thus, conditions for successful W-filament ETAAS measurements are identical to those typically applied in ETV-ICP-MS, in which vaporized species are swept into the ICP-MS at a purge-gas flow rate of 1 L min⁻¹.

Pseudo-simultaneous AAS and ICP-MS measurements of Pb in blood

By coupling the W-filament ETAAS instrument directly to an ICP-MS, Pb that is atomized from the W-filament surface can be detected almost simultaneously by AAS and by ICP-MS. In addition to monitoring the analyte, in this case ²⁰⁸Pb, use of a W-filament makes it possible to monitor what might otherwise be considered an undesirable polyatomic interference, i.e., the signal at mass 52 which, as we argue below, is primarily due to ⁴⁰Ar¹²C and is therefore a reasonable surrogate for the biological matrix. Fig. 1 shows the pseudo-simultaneous electrothermal vaporization and electrothermal atomization signals for Pb from a 15 µL aliquot of 1 + 4 diluted blood (NYS-RM 049, 42.2 µg dL⁻¹), as measured by ICP-MS and AAS, respectively. In the ICP-MS panel, the most abundant Pb isotope, ²⁰⁸Pb, is monitored from its initial vaporization to well into the cleaning stage of the electrothermal heating program. By contrast, the AAS panel shows a fast transient absorbance signal (<1 s) that is typically associated with W-filament atomization events. There is about a 2-s delay between the start of the atomization/vaporization event and the appearance of ²⁰⁸Pb at the ICP-MS, suggesting that the Pb vapor/atom cloud takes ∼2 s to travel from the atomization cell to the plasma. This is consistent with the use of a 15 cm × 6 mm id transfer line and a gas flow rate of approximately 1 L min⁻¹. Our selection of 1 L min⁻¹ as the gas flow rate was based on previous optimization studies²⁴ for this analysis. In the ICP-MS, the ²⁰⁸Pb signal is detected for more than 10 s, i.e., well into the cleaning step. The broad and tailing Pb MS peak is likely due to the relatively large size of the atomization/vaporization cell (volume ∼244 mL) coupled with diffusion within the transfer line. This is consistent with previous reports that (a) larger cell volumes produce temporal peak broadening due to excessive dilution of the sample vapor by the carrier gas³¹ and (b) peak broadening is due to diffusion of the atomic vapor within the transfer line.⁴ In addition to ²⁰⁸Pb, the signal at mass 52, which is due to ⁴⁰Ar¹²C, is also shown, and indicates that some amount of carbonaceous residue is volatilized during the early part of the cleaning step.

Fig. 1 Pseudo-simultaneous electrothermal vaporization and atomization of Pb from 15 µL of 1 + 4 diluted blood (NYS RM Lot 049, 42.2 µg dL⁻¹) measured by (a) ICP-MS and (b) AAS using a W-filament ETV/ETA instrument coupled to an ICP-MS. Integrated ion intensity for ²⁰⁸Pb by ETV-ICP-MS is 5.5 × 10⁵ counts·s; integrated Pb absorbance by AAS is 0.0222 s, which yields a characteristic mass of (m₀) of 251 pg.

Selected ion monitoring during the ETV heating program

A later modification (version 5.0) of the PBA-10's custom software program synchronized the contact-closure signal for starting data acquisition with the commencement of the ETV heating cycle. This modification enabled selected ion monitoring by ICP-MS across the entire W-filament ETV heating cycle, including drying, pyrolysis, and vaporization/atomization. In Fig. 2, a typical electrothermal heating cycle is shown for a 15 µL aliquot of 1 + 4 diluted blood (NYS-RM 049, 42.2 µg dL⁻¹). Each step in the W-filament electrothermal temperature program, i.e., drying, pyrolysis, cooling, atomization and cleaning, is shown on the time axis in Fig. 2 along with several ions that were monitored. These selected ions include the analyte (²⁰⁸Pb), the permanent modifier (¹⁰³Rh), the atomizer substrate (¹⁸³W) and a surrogate for the blood matrix (⁴⁰Ar¹²C). Nothing of importance (except water) is vaporized during the drying step. However, during the pyrolysis or ashing step, a large transient signal is observed at mass 52 corresponding to the molecular species ⁴⁰Ar¹²C, a surrogate for the carbonaceous blood matrix that is vaporized. Interestingly, Rh monitored as ¹⁰³Rh, and the analyte monitored as ²⁰⁸Pb, are not detected to any great extent during pyrolysis, indicating that Rh is a very effective modifier. However, large amounts of ¹⁰³Rh are detected during the clean step, necessitating periodic re-coating of the W-filament as noted in previous reports,²⁶ while the analyte ²⁰⁸Pb appears mostly as a fast transient signal during the atomization/vaporization period. As noted above, the ⁴⁰Ar¹²C signal observed during atomization is much less intense than that observed during pyrolysis, suggesting that most of the blood matrix is removed during pyrolysis.

Fig. 2 A typical W-filament ETV-ICP-MS intensity–time profile for 15 µL of 1 + 4 diluted blood containing Pb (NYS RM Lot 049: 42.2 µg dL⁻¹ Pb). Note that counts for the ions ¹⁸³W, ⁴⁰Ar¹²C and ¹⁰³Rh are plotted on the right-hand y axis, while those for the analyte ²⁰⁸Pb are plotted on the left-hand y axis. Each ETV step is matched to the corresponding ETA step in the W-filament atomization temperature program.

Spectral interferences in ICP-MS can occur at precisely the mass selected for the analyte. For example at mass 52, not only is the polyatomic species ⁴⁰Ar¹²C expected, but ⁵²Cr also. To ensure that the measurements are not masked by such interferences, we also monitored ¹³C and calculated the isotope ratio ⁴⁰Ar¹²C/¹³C. Fig. 3 shows the change in the ⁴⁰Ar¹²C ion intensity during the pyrolysis event, along with the corresponding ⁴⁰Ar¹²C/¹³C ratio. It is evident that the ⁴⁰Ar¹²C/¹³C ratio remains relatively constant throughout the pyrolysis event independent of ion intensity, suggesting that the ion intensity monitored at mass 52 is primarily contributed by the carbon-containing polyatomic species ⁴⁰Ar¹²C that is a marker of the blood matrix. Equally significant, however, is the fact that Cr is a much more refractory element compared with Pb. Given that Pb is not vaporized at the pyrolysis power setting used, it is inconceivable that Cr would be. Thus, ⁴⁰Ar¹²C appears to be the major species detected at mass 52, and is an appropriate surrogate for the blood matrix.

Fig. 3 Plot showing the ⁴⁰Ar¹²C ion signal and the ratio of ⁴⁰Ar¹²C to¹³C⁺ as a function of time during the entire 30 s pyrolysis event for a 15 µL aliquot of 1 + 4 diluted blood (NYS RM Lot 049: 42.2 µg dL⁻¹ Pb). See the text for an explanation of how these data are plotted.

The amount of W vaporized from the filament and transported into the plasma was monitored at mass 183 (¹⁸³W, natural abundance 14.3%) throughout the entire electrothermal event. As is shown in Fig. 2, a detectable amount of W is clearly vaporized during pyrolysis, during which the power setting, i.e., temperature, is relatively low in relation to the high melting point (3410 °C) and boiling point (5660 °C) of W metal. No significant vaporization is evident during the ETA/ETV step, or even during the clean step, indicating that a more volatile W species must be formed during the pyrolysis step. By contrast, no significant W loss is observed during pyrolysis with blank firings, i.e., in the absence of a diluted blood matrix, suggesting that the matrix is critical to formation of the volatile W species.

In a previous study involving W-filament ETV-ICP-MS, Venable et al.³² detected W at temperatures well below the melting point of the metal and argued that it was due to formation of the more volatile W oxides species. Byrne et al.⁷ investigated the mechanism of volatilization of W from a graphite substrate using ETV-ICP-MS. They found two distinct vaporization processes occur in the temperature range 800–2700 °C, resulting in two distinct peaks for W. One peak that appeared at temperatures as low as 850 °C was attributed to the volatilization of WO₃. Based on our studies, we find it quite conceivable that the blood matrix, and perhaps the nitric acid used in the diluent, could provide a source or sources of oxygen that, during pyrolysis, react(s) with W at the surface of the filament, yielding a mixture of W oxides, such as WO₂ and W₂O₅. These oxide species can undergo sublimation at temperatures between 800 and 900 °C consistent with typical pyrolysis temperatures expected here, thus vaporizing W.³³ According to Venable et al.,³² these W species appear as W and WO in the plasma, despite use of 10% H₂ in Ar as the purge gas to reduce oxidation and to prolong filament lifetimes. Although we used only 6% H₂ in Ar as the purge gas here, we did not detect W in the plasma as W oxide. However, we did detect it as ¹⁸³W, though only during the pyrolysis step.

Rh permanent modifier

Coupling the W-filament AAS instrument to an ICP-MS also provides a unique opportunity to examine the behavior of Rh as the permanent modifier for blood Pb measurements. Fig. 4(i) shows a typical pyrolysis study as carried out in ETAAS; the change in integrated absorbance during the atomization step is plotted as a function of increasing pyrolysis temperature. Superimposed in Fig. 4 as (ii) and (iii) are complementary data obtained pseudo-simultaneously by ICP-MS from the same firing that directly probed the pyrolysis step in real time. However, since the ICP-MS sensitivity for ions such as ²⁰⁸Pb and ⁴⁰Ar¹²C varies as a function of pyrolysis temperature, it is necessary to express the integrated ion intensity during pyrolysis as a fraction of the total integrated ion signal for the entire electrothermal cycle. Thus, in Fig. 4, it is the pyrolysis fraction of the total integrated ion signal for (ii) the analyte (²⁰⁸Pb) and (iii) the matrix surrogate (⁴⁰Ar¹²C) that is plotted as a function of increasing pyrolysis power, i.e., temperature. The reason for the variation in ICP-MS sensitivity is unclear, but it may be a combination of multiple variables such as vapor composition, and transfer efficiency. ICP-MS, which is inherently more sensitive than AAS, reveals minimal losses of Pb during pyrolysis, even below settings of 2.6 W at which the integrated absorbance is unchanged. However, 2.6 W appears to be the optimal pyrolysis setting for blood Pb measurements by AAS. This has now been confirmed with the ICP-MS data, which show that approximately 60% of the vaporized blood matrix (⁴⁰Ar¹²C) reaching the plasma is removed during pyrolysis. At pyrolysis power settings above 2.6 W, analyte loss becomes substantial, but the maximum amount of vaporized blood matrix that can be removed is only ∼70%.

Fig. 4 Pyrolysis studies showing (i) integrated Pb absorbance by AAS and (ii and iii) complementary data obtained pseudo-simultaneously by ICP-MS for a 15 µL aliquot of 1 + 4 diluted blood containing Pb (NYS RM Lot 049: 42.2 µg dL⁻¹ Pb). The ICP-MS data are expressed as the pyrolysis fraction of the total integrated ion signal for (ii) the analyte (²⁰⁸Pb) and (iii) the matrix (⁴⁰Ar¹²C) as a function of increasing pyrolysis power.

As noted above, Rh is only lost during the cleaning step when maximum power (46 W) is applied. The Rh signal, as depicted in Fig. 2, appears to be almost steady-state during the cleaning phase, suggesting that the coating is continuously vaporized throughout the step. Thus, shortening this step will likely improve the durability of the Rh coating and decrease the frequency of recoating. We compared different amounts of blood matrix deposited on the filament. Use of a 15 µL aliquot of 1 + 9 and of 1 + 4 diluted blood yields 1.5 µL and 3 µL of whole blood, respectively. The total integrated ion intensity calculated for ²⁰⁸Pb, based on 1.5 µL whole blood, is 2.8 × 10⁵ counts·s, and is approximately half the integrated intensity found for 3 µL of whole blood (5.5 × 10⁵ counts·s). Similarly, the total integrated ion intensity for ⁴⁰Ar¹²C, based on 1.5 µL of whole blood (1.6 × 10⁶ counts·s) is approximately half that found for 3 µL of whole blood (3.3 × 10⁶ counts·s). This observation further supports our inference that the signal at mass 52 is a valid marker of the blood matrix. However, the intensity of the ¹⁰³Rh signals is roughly the same for both amounts of blood matrix, suggesting that Rh loss is independent of sample loading. Even with blank firings, i.e., in the absence of a diluted blood matrix, significant loss of Rh is observed. These observations are consistent with our previous work that indicated a need for periodic recoating of the W-filament with Rh to sustain the desired performance.²⁶

In our previous study,²⁶ it was observed that the gradual loss of Rh could not be prevented by direct addition of the modifier to the diluent, as is common practice for blood Pb measurements using phosphate modifier. In the present study we were able to probe this characteristic directly. Fig. 5 compares ETV-ICP-MS data collected with (a) Rh pre-coated on the W-filament and (b) Rh added to the blood sample diluent. As noted above, pre-coated Rh is only vaporized during the cleaning step. However, when added to the diluent, Rh is vaporized throughout pyrolysis and atomization, in addition to during the cleaning step. Surprisingly, Rh vaporization appears as numerous, apparently random, spikes throughout the pyrolysis, vaporization/atomization and clean steps. We have no convincing explanation for these observations but they do illustrate why Rh must be thermally pre-coated to, or bound to, the W-surface for successful modification to occur.

Fig. 5 Intensity–time profiles for ²⁰⁸Pb, ⁴⁰Ar¹²C, and ²⁰³Rh observed from a 15-µL aliquot of 1 + 4 diluted blood (NYS RM Lot 049: 42.2 µg dL⁻¹ Pb). Compared are (a) use of a permanent Rh modifier thermally precoated on a W-filament and (b) direct addition of Rh modifier to the diluent. Note that ions other than the analyte (²⁰⁸Pb) are plotted on the right hand side y-axis. Note the color legend assigned in panel (a) is the same for (b).

Conclusions

This study provides additional new information on the characteristics of Rh vaporization from Rh-coated W filaments when they are used as electrothermal atomizers in AAS. Pseudo-simultaneous ICP-MS and AAS measurements of Pb in blood were obtained by interfacing a compact prototype W-filament ETAAS instrument to a commercial PerkinElmer ELAN 6100. Using this arrangement, we were able to probe the atomization/vaporization processes in which Pb is vaporized and atomized from a blood matrix. In addition to the analyte (²⁰⁸Pb), we were able to monitor the matrix (as ⁴⁰Ar¹²C), the modifier (¹⁰³Rh) and the atomizer material (¹⁸³W) across the entire electrothermal heating cycle. Pyrolysis studies included monitoring the analyte and the matrix as a function of increasing pyrolysis power. We confirm that it is during the cleaning stage that most of the Rh coating is vaporized. Thus, careful optimization of the cleaning stage power (i.e., temperature) setting is critical for reducing to a minimum the periodic requirement for Rh re-coating. We have also confirmed that a substantial fraction (60%) of the carbonaceous matrix from blood (measured as ArC at mass 52) is removed during the pyrolysis step without any significant loss of the Pb. However, a small amount of matrix remains on the filament surface even after the pyrolysis and atomization steps, and can only be completely removed by cleaning for at least 30 s at a power setting of 46 W. Interestingly, when Rh is added to the sample diluent, rather than thermally pre-coated in a separate step, most of the modifier is vaporized throughout the pyrolysis, atomization and cleaning steps, appearing as a series of random spikes in the ICP-MS trace. We are unable to provide a convincing explanation for this behavior, but it does explain why this approach was unsuccessful in our previous studies.

Acknowledgements

This work was supported in part by grant R08/CCR208632-04 from the Centers for Disease Control and Prevention. The contents of this paper are solely the responsibility of the authors and do not necessarily represent the official views of the CDC. We thank Mr. Ernest Mills, Mr. Eric Marczak, and the staff of the Wadsworth Center's Lead Poisoning Laboratory for technical assistance with this project.

References

1. A. L. Gray and A. R. Date, Analyst, 1983, 108, 1033.
2. H. E. Taylor, Inductively Coupled Plasma-Mass Spectrometry: Practices and Techniques, Academic Press, San Diego, CA, USA, 2001, pp. 76–79.
3. R. E. Sturgeon and J. W. Lam, J. Anal. At. Spectrom., 1999, 14, 785.
4. M. M. Lamoureux, D. C. Grégoire, C. L. Chakrabarti and D. M. Goltz, Anal. Chem., 1994, 66, 3208.
5. M. M. Lamoureux, D. C. Grégoire, C. L. Chakrabarti and D. M. Goltz, Anal. Chem., 1994, 66, 3217.
6. W. Slavin, D. C. Manning and G. R. Carnrick, At. Spectrosc., 1981, 2(5), 137.
7. J. P. Byrne, D. M. Hughes, C. L. Chakrabarti and D. C. Grégoire, J. Anal. At. Spectrom., 1994, 9, 913.
8. J. P. Byrne, D. C. Grégoire, D. M. Goltz and C. L. Chakrabarti, Spectrochim. Acta, Part B, 1994, 49, 433.
9. D. M. Goltz, D. C. Grégoire and C. L. Chakrabarti, Spectrochim. Acta, Part B, 1995, 50, 1365.
10. D. M. Goltz, D. C. Grégoire, J. P. Byrne and C. L. Chakrabarti, Spectrochim. Acta, Part B, 1995, 50, 803.
11. J. P. Byrne, D. C. Grégoire, M. E. Benyounes and C. L. Chakrabarti, Spectrochim. Acta, Part B, 1997, 52, 1575.
12. R. St. C. McIntyre, D. C. Grégoire and C. L. Chakrabarti, J. Anal. At. Spectrom., 1997, 12, 547.
13. C. J. Park, J. C. Van Loon, P. Arrowsmith and J. B. French, Anal. Chem., 1987, 59, 2191.
14. S. Hauptkorn, V. Krivan, B. Gercken and J. Pavel, J. Anal. At. Spectrom., 1997, 12, 421.
15. B. Gercken, E. Felder, H. Rink and D. Vetter, in Plasma Source Mass Spectrometry, New Developments and Applications, eds. G. Holland and S. D.Tanner, Royal Society of Chemistry, Cambridge, 1999, pp. 222–236.
16. R. Tsukahara and M. Kubota, Spectrochim. Acta, Part B, 1990, 45, 779.
17. Y. Okamoto, Fresenius’ J. Anal. Chem., 2000, 367, 300.
18. N. Shibata, N. Fudagawa and M. Kubota, Anal. Chem., 1991, 63, 636.
19. N. Nonose and M. Kubota, J. Anal. At. Spectrom., 1998, 13, 151.
20. Y. Okamoto, R. Kikkawa, Y. Kobayashi and T. Fujiwara, J. Anal. At. Spectrom., 2001, 16, 96.
21. Y. Okamoto, J. Anal. At. Spectrom., 2001, 16, 539.
22. I. Marawi, L. K. Olson, J. S. Wang and J. A. Caruso, J. Anal. At. Spectrom., 1995, 10, 7.
23. D. M. Goltz, D. C. Grégoire and C. L. Chakrabarti, Can. J. Appl. Spectrosc., 1996, 41, 70.
24. Y. Zhou, P. J. Parsons, K. M. Aldous, P. Brockman and W. Slavin, J. Anal. At. Spectrom., 2001, 16, 82.
25. US Department of Health and Human Services, Fed. Regist., 1992, 57, 7002.
26. Y. Zhou, P. J. Parsons, K. M. Aldous, P. Brockman and W. Slavin, Spectrochim. Acta, Part B, 2002, 57, 727.
27. P. J. Parsons, C. Geraghty and M. F. Verostek, Spectrochim. Acta, Part B, 2001, 56, 1593.
28. A. G. Coedo, T. Dorado, I. Padilla, R. Maibusch and H. M. Kuss, Spectrochim. Acta, Part B, 2000, 55, 185.
29. N. N. Sesi, A. MacKenzie, K. E. Shanks, P. Yang and G. M. Hieftje, Spectrochim. Acta, Part B, 1994, 49, 1259.
30. H. Berndt and G. Schaldach, J. Anal. At. Spectrom., 1988, 3, 709.
31. C. J. Park, J. C. Van Loon, P. Arrowsmith and J. B. French, Can. J. Spectrosc., 1987, 32, 29.
32. J. D. Venable, M. Detwiler and J. A. Holcombe, Spectrochim. Acta, Part B, 2001, 56, 1697.
33. CRC Handbook of Chemistry and Physics,, ed. D. R. Lide, CRC Press, Boca Raton, FL, USA, 75th edn., 1994, pp. 4–109.

---

Footnotes

† Use of trade names is for informational purposes only and does not imply an endorsement by the New York State Department of Health.

‡ To convert µg dL⁻¹ into µmol L⁻¹, multiply by 0.04826; i.e., 10 µg dL⁻¹ equals 0.48 µmol L⁻¹. To convert µg L⁻¹ into µmol L⁻¹, multiply by 0.004826.

---