Preconcentration and Atomization of Arsane in a Dielectric Barrier

Article pubs.acs.org/ac

Preconcentration and Atomization of Arsane in a Dielectric Barrier Discharge with Detection by Atomic Absorption Spectrometry Petr Novák,†,‡ Jiří Dědina,† and Jan Kratzer*,† †

Institute of Analytical Chemistry of the CAS, v. v. i., Veveří 97, 60200 Brno, Czech Republic Charles University in Prague, Faculty of Science, Department of Analytical Chemistry, Hlavova 8, Prague 2 CZ 128 43 Czech Republic

‡

S Supporting Information *

ABSTRACT: Atomization of arsane in a 17 W planar quartz dielectric barrier discharge (DBD) atomizer was optimized, and its performance was compared to that of a multiple microflame quartz tube atomizer (MMQTA) for atomic absorption spectrometry (AAS). Argon, at a flow rate of 60 mL min−1, was the best DBD discharge gas. Free As atoms were also observed in the DBD with nitrogen, hydrogen, and helium discharge gases but not in air. A dryer tube filled with NaOH beads placed downstream from the gas−liquid separator to prevent residual aerosol and moisture transport to the atomizer was found to improve the response by 25%. Analytical figures of merit were comparable, reaching an identical sensitivity of 0.48 s ng −1 As in both atomizers and limits of detection (LOD) of 0.15 ng mL−1 As in MMQTA and 0.16 ng mL−1 As in DBD, respectively. Compared to MMQTA, DBD provided 1 order of magnitude better resistance to interference from other hydride-forming elements (Sb, Se, and Bi). Atomization efficiency in DBD was estimated to be 100% of that reached in the MMQTA. A simple procedure of lossless in situ preconcentration of arsane was developed. Addition of 7 mL min−1 O2 to the Ar plasma discharge resulted in a quantitative retention of arsane in the optical arm of the DBD atomizer. Complete analyte release and atomization was reached as soon as oxygen was switched off. Preconcentration efficiency of 100% was observed, allowing a decrease of the LOD to 0.01 ng mL−1 As employing a 300 s preconcentration period.

A

hydride generation (HG)1 approach is often used as a sample introduction technique prior to arsenic spectrometric detection at ultratrace levels making use of its superb sample introduction efficiency (∼100%). In order to avoid the use of an expensive detector such as inductively coupled mass spectrometry (ICP-MS), a compromise is often chosen which combines cheap and reliable but less-sensitive detectors with a fast and simple preconcentration step to reach satisfactory low detection limits. HG-AAS1 is a favorite and widespread routine method for As determination in analytical laboratories which profits from its user friendliness, reliability, and low investment/running costs. Externally heated quartz tube atomizers (QTA) are commonly used in HG-AAS for As determination.1 A modified version of QTA termed multiple microflame QTA or MMQTA2,3 was designed some 15 years ago. It allows addition of controlled flow rates of air/oxygen into the whole volume of optical arm of the MMQTA to improve sensitivity, linearity, and interference extent of conventional QTA for determination of some hydride-forming elements, especially As and Se. A new kind of low-power microplasma arsane atomizers for AAS4 and AFS5−7 based on dielectric barrier discharges (DBD)8 have been described recently. In addition, a DBDassisted arsane generation has been reported in the literature.9 The fundamental drawback of the DBD papers published up to © XXXX American Chemical Society

the current date is the fact that they can be taken only as a proof-of-concept reports showing possibility of arsane atomization in DBD devices. However, presently, there are no detailed studies performed focusing either on comparison of DBD-based arsane atomizer with other common atomizers or concentrating on estimation of atomization efficiency and investigation of the mechanisms. Various approaches to arsane preconcentration compatible with AAS detection have been reported in the literature including in situ trapping in graphite atomizers10 or a tungsten tube atomizer,11 trapping in a molybdenum foil strip trap coupled to AAS with miniature diffusion flame,12 in-atomizer trapping in the inlet arm of MMQTA,13 or involving the cryotrapping procedure.1,14 The inherent advantage of preconcentration in the cryotrap is the possibility of arsenic speciation analysis.15 Considering the determination of total As content only, the cryogenic preconcentration procedure is too laborious and time-consuming, and procedures involving trapping on metal surfaces (W, Mo)11,12 and quartz traps13 suffer from incomplete arsane preconcentration resulting in Received: April 8, 2016 Accepted: April 29, 2016

A

DOI: 10.1021/acs.analchem.6b01365 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry impaired detection limits and measurement repeatability. In situ trapping of arsane in graphite atomizers remains the only routinely used approach to preconcentration with AAS detection due to its simplicity and compatibility with commercial spectrometers. Although a complete analyte trapping and release is generally assumed for in situ trapping of common hydride-forming elements in graphite atomizers under optimized experimental conditions, severe losses of arsane during preconcentration have been reported employing this approach.16 The aims of this work were to (1) optimize arsane atomization conditions in a planar DBD atomizer, (2) compare DBD atomizer performance and analytical figures of merit including interferences with those of a commonly used MMQTA, (3) estimate the atomization efficiency of arsane in a DBD atomizer and evaluate its applicability in analytical routine, (4) test the possibility of in situ arsane preconcentration in a DBD atomizer.

atomizer) was coupled to a high frequency (25 kHz), highvoltage generator (Lifetech, Czech Republic). The complete apparatus setup is described in detail in ref 18. A schema of the DBD atomizer is shown in Figure 1. The dimensions of the DBD atomizer were the same as those for EHPA atomizer (see below). If explicitly stated, the inner surface of the DBD atomizer was treated with a 5% solution of dimethyldichlorsilane (DMDCS) in toluene in order to passivate its surface. Silanization was performed following the procedure described in ref 18. The DBD atomizer was cleaned if necessary (sensitivity drop or after interferent study) with 10% (v/v) HNO3 for 10 min. Subsequently, it was flushed by deionized water several times and allowed to dry. A T-shaped quartz MMQTA described previously2,3 and depicted in Figure 1 was resistively heated to 900 °C (RMI heating unit, Czech Republic). Its 150 mm (i.d. 6 mm) horizontal arm was aligned within the optical path of the spectrometer. The length and i.d. of the inlet arm was 100 and 2 mm, respectively. A flow of 25 mL min−1 of air as outer gas was employed to enhance its sensitivity. If explicitly stated, a deactivated fused silica capillary (Supelco, Germany, 0.53 mm i.d.), not shown in Figure 1, was centered inside the inlet arm with its tip aligned with the junction of the inlet arm with the optical tube. The capillary was supplied with a flow of 1.5 mL min−1 oxygen, whereas no air as outer gas was employed in this case. Gases from the GLS were in this case introduced into the inlet arm of the multiatomizer through a PTFE tubing placed concentrically around the silica capillary. MMQTA was cleaned if necessary (sensitivity drop or after interferent study) with a mixture of concentrated HNO3 and HF (7:3) for 10 min. Subsequently, it was flushed by deionized water several times and allowed to dry. In a single set of experiments, for comparative purposes, a Tshaped externally heated planar quartz atomizer, termed EHPA further in the text, was employed (see Figure 1). It had the same dimensions and was manufactured in the same way as the quartz body of the DBD chamber. No electrodes were attached to this atomizer. Thus, a rectangular-shaped arm of the atomizer (inner dimensions of 7 mm × 3 mm and length of 75 mm) was aligned with the optical axis of the spectrometer, whereas a quartz tube (20 mm long, 2 mm inner diameter, 4 mm outer diameter) served as its inlet arm. The arm placed in the optical axis of the spectrometer, was resistively heated by a laboratory power supply source (PS 3065-10 B; E-A, ElektroAutomatik GmbH, Viersen, Germany) to 900 °C. Resistive heating was realized by a kanthal wire (4.17 Ω.m−1, 0.65 mm in diameter). The atomizer arm and the resistive wire were wrapped in a Rescor ceramic fiber blanket (Cotronics Corp., U.S.A.) and overlaid with Al foil to serve as thermal insulation. A deactivated fused silica capillary (Supelco, Germany, 0.53 mm i.d.) was centered inside the inlet arm with its tip aligned with the junction of the inlet arm with the rectangular optical arm. The capillary was supplied with a flow of oxygen (1.0 mL min−1) using the same configuration of gases introduction as in the case of MMQTA supplied with oxygen (see Figure 1). Atomic Absorption Measurements. A GBC model SavantAA atomic absorption spectrometer (GBC, Australia) was employed without background correction. A Photron As boosted hollow cathode lamp (superlamp) operated at 193.7 nm analytical line with 2.0 nm spectral bandpass and a lamp current of 20 mA (boost current 7−15 mA). One of the atomizers investigated was located in the optical axis of the

■

EXPERIMENTAL SECTION Chemicals and Standards. All the chemicals and standards used in this work are in detail described in Section S1 of Supporting Information. Hydride Generator. An in-house made, continuous flow hydride generation system based on a peristaltic pump (Ismatec, Switzerland) was employed together with a 3 mL inner volume gas−liquid separator (GLS) with a forced outlet. The dryer based on a polypropylene cartridge (100 mm long, 15 mm i.d.) filled with solid sodium hydroxide beads (further termed as NaOH dryer), identical to that described in ref 17, was inserted downstream from the GLS to prevent aerosol and droplets to enter the atomizers. Gas flows of a carrier/discharge gas (Ar if not explicitly stated otherwise) for both atomizers and auxiliary air for MMQTA were controlled by mass flow controllers (Omega Engineering, U.S.A.). See Table 1 for Table 1. Generation and Atomization Conditions for MMQTA, DBD, and EHPA Atomizers optimum generation conditions HCl concentration (mol L−1) blank/sample channel pump flow rate (mL min−1) NaBH4/KOH concentration reductant channel pump flow rate (mL min−1) optimum atomization conditions MMQTA atomizer temperature (°C) carrier/discharge Ar flow rate (mL min−1) outer air flow rate (mL min−1) or alternatively oxygen flow rate to the capillary (mL min−1) plasma power (W)

1.0 4.0 0.5%/0.4% 1.0 DBD EHPA

900 75 25

45 60 -

900 75 -

1.5 -

17

1.0 -

optimum experimental conditions for arsane generation and Figure 1 for schematic diagram of the hydride generator. The gaseous phase at the GLS outlet always contains, apart from the carrier/discharge gas and arsane, 15 mL min−1 of hydrogen evolved as a side-product from NaBH4 decomposition. For the sake of simplicity, the carrier/discharge gas is termed according to its inert gas major component (i.e., Ar or He, etc.) throughout the remaining text. Atomizers. Either a DBD atomizer or a MMQTA was used in a majority of experiments. A planar configuration of a quartz DBD atomizer (see ref 18 for detailed description of the B


Article

Analytical Chemistry

Figure 1. Scheme of the hydride generator with MMQTA, DBD, and EHPA atomizers.

spectrometer. Measurements were performed under generation and atomization conditions as summarized in Table 1. Procedures. Measurements were performed mostly in the online atomization mode with direct transfer of arsane generated from the GLS into the respective atomizer. Sample introduction time of 30 s was employed followed by introduction of blank to flush the GLS. Switching between standard and blank was performed using a manual three-way valve. If explicitly stated, the in situ collection (preconcentration) mode was employed using the DBD atomizer. The way of gas introduction from hydride generator into the DBD atomizer was modified in the preconcentration mode as shown in Figure 2: a deactivated fused silica capillary (Supelco, Germany, 0.53 mm i.d.) was centered inside the inlet arm of the DBD atomizer with its tip aligned with the junction of the inlet arm with the optical arm. Gases from the GLS were introduced into the inlet arm of the DBD through a PTFE tubing placed concentrically

around the silica capillary. The way of gas introduction was actually the same as with EHPA in online atomization mode (compare Figure 1 and Figure 2). Preconcentration mode consisted of two steps: trapping - analyte was retained in the optical arm of the DBD atomizer; volatilization - trapped analyte is released and atomized in the DBD plasma. Trapping step: Arsane was generated from a sample solution. A sample introduction time between 30 and 300 s was employed, followed by introduction of blank for 30 s. The capillary was supplied with a flow of 7.0 mL min−1 oxygen during the whole period of sample and blank introduction. Volatilization step: Atomic absorption signal reading was initiated at the beginning of the volatilization step followed immediately by switching off the oxygen flow while still running hydride generation from blank solution. Signals were recorded for 90 s, and peak area as well as peak height response were evaluated in both modes of operation. Averages of at least three replicates of signal measurements are presented in figures and in the text with uncertainty expressed as standard deviation (SD). Peak area was invariably employed as the analytical quantity.

■

RESULTS AND DISCUSSION Optimization of Arsane Atomization in a DBD Atomizer. All relevant experimental parameters such as discharge gas type and its flow rate, DBD power supply rate as well as the effect of aerosol removal from gaseous phase by NaOH dryer prior atomization and modification of the inner surface of the atomizer were optimized. Arsane atomization in a DBD atomizer with subsequent AAS detection had already been investigated,4 but the previous study4 was limited only to basic experimental parameters. Moreover, completely different power supply source and slightly altered atomizer design were employed, thus making the results not easily comparable. Unless stated otherwise, the experiments were performed with the NaOH dryer cartridge incorporated downstream from the GLS, in a surface nontreated DBD, with 10 ng mL−1 As standard solution at 17 W DBD power and employing Ar at a flow rate of 60 mL min−1. Effect of Discharge Gas. Argon, helium, nitrogen, hydrogen, and air were investigated as discharge gases (Figure 3A). As detailed in the Experimental Section, the gaseous phase always

Figure 2. Scheme of the DBD “arrangement 1” employed for the preconcentration procedure. C


Article


Figure 3. Optimization of the experimental conditions for arsenic atomization in the DBD in the online atomization mode, 10 ng mL−1 As, if not being subject of optimization, the experimental conditions were fixed at their optimum values: 60 mL min−1 Ar, 17 W DBD power, NaOH dryer tube employed. (A) Effect of discharge gas identity on sensitivity of As determination; (B) Effect of DBD power on peak area of arsenic; (C) Effect of Ar flow rate on peak area of arsenic; (D) Effect of NaOH dryer cartridge on peak shape of arsenic signal.

contained a fraction of hydrogen (15 mL min−1). The highest signal with the best repeatability was observed in Ar, whereas the signal in N2 was approximately 10% lower. Moreover, its repeatability was much worse than in argon. Analyte signal in He and H2, respectively, reached ca. 50% and 40% of the reference signal in Ar. The measurements in hydrogen as a discharge gas were hardly repeatable, resulting in an unacceptable high standard deviation of the results. Argon was thus selected as the best discharge gas for further measurements. Although the discharge performed well in air, no signal was obtained. Similar behavior was observed when replacing air with a mixture of Ar with 20% oxygen. Subsequently, transient signals were observed when air or Ar−O2 mixture were replaced by argon during blank generation. This fact indicates that arsenic species are retained in the DBD in the discharge gas containing oxygen. Moreover, the analyte can be subsequently volatilized and atomized by Ar discharge with hydrogen fraction content from blank generation. Analogous behavior in the air DBD discharge has been recently observed for bismuth.18 Oxygen present in both mixtures (air and Ar−O2) is unambiguously responsible for analyte retention in the DBD since signal of free atoms of As can be detected in N2 and Ar discharges (see Figure 3A). The fact that the analyte can be retained in the DBD atomizer in the Ar−O2 discharge and subsequently released in the Ar−H2 discharge was further studied and employed for the

preconcentration mode measurements (see section Arsane Preconcentration in a DBD Atomizer). Effect of DBD Power Supply Rate. The effect of power between 5 and 19 W employing a flow rate of 60 mL min−1 Ar is depicted in Figure 3B. The discharge is quite inhomogeneous and unstable below 5 W. The arsenic signal reaches a plateau between 10 and 19 W. A power of 17 W was employed for further measurements. Effect of Discharge Gas Flow Rate. Ar serves as a carrier gas to transport analyte hydride from gas liquid separator into the atomizer, but it primarily acts as the discharge gas supporting the plasma. The effect of Ar flow rate was investigated in the range from 40 to 200 mL min−1 (Figure 3C) employing the DBD power of 17 W. The signal of As reaches a plateau between 40 and 75 mL min−1 Ar, followed by a signal decrease at higher Ar flow rates. This signal decrease at higher flow rates is obviously caused by analyte dilution. A flow rate of 60 mL min−1 Ar was chosen as optimum because of high sensitivity and good signal repeatability. Effect of Dryer. Little or no effect of moisture on atomization of arsane and its methylated analogues was observed in a study by Zhu et al.4 when employing a planar DBD atomizer with AAS detection using a cryogenic trap for moisture removal. A dryer tube packed with solid NaOH was employed in this work to study the effect of moisture and aerosol on arsane signal in both DBD and MMQTA atomizers. Its performance has been already demonstrated in previous D


Article


Table 2. Comparison of Detection Limits (LOD) and Sensitivities Reached for DBD, EHPA, and MMQTA Atomizers; As Content in TM-RAIN-04 Certified Reference Material Found by DBD and MMQTA Atomizers measurement mode

online atomization

atomizer

MMQTA

DBD

LOD (ng mL−1) sensitivity (s ng−1) CRM TM-RAIN-04 (ng mL−1 As)

0.15 0.482 ± 0.003 1.09 ± 0.03

0.16 0.475 ± 0.016 1.08 ± 0.01

studies with MMQTA17,19 showing efficient removal of spray droplets and vapor with no losses of arsane. The same observations with MMQTA were reproduced in this work. The fact that same sensitivity was observed in MMQTA regardless of the NaOH dryer use indicates that (i) MMQTA is robust to the spray droplets load owing to its high atomization temperature and (ii) no losses of arsane occur inside the dryer tube. On the contrary, significantly lower signals were observed in the DBD atomizer in absence of NaOH dryer reaching about 75% of the signal observed in the presence of the dryer. An average sensitivity of 0.47 ± 0.01 s ng−1 As (n = 30) was observed using the NaOH dryer tube, whereas it decreased to 0.36 ± 0.02 s ng−1 As (n = 30) in its absence. Moreover, the signal shape was more smooth and regular with the dryer as depicted in Figure 3D. It might be concluded that excessive load of aerosol into the DBD results in impaired atomization of arsane and plasma instability. Therefore, a dryer packed with NaOH was employed for all measurements. Effect of Inner Surface Modification. Modification of the inner DBD surface by silanization has been reported to significantly improve Bi signal in HG-DBD-AAS,18 whereas no effect of DBD silanization on Se signal was observed recently.20 The signal of 10 ng mL−1 As standard solution was measured in a DBD atomizer without any DMDCS-treatment reaching average sensitivity of 0.45 ± 0.02 s ng−1 As (n = 15), whereas sensitivity of 0.48 ± 0.02 s ng−1 As (n = 15) was observed in the same piece of DBD atomizer after silanization. Untreated DBD was therefore used in further measurements. Analytical Performance and Comparison of the Atomizers. Analytical performance of the DBD atomizer was compared to that of MMQTA as a reference since the latter is commonly used in HG-AAS. Moreover, an EHPA atomizer was constructed in order to allow direct comparison of atomization efficiency in two atomizers, that is, DBD and EHPA, being geometrically identical. Although atomization in the DBD atomizer takes place in the plasma environment at a low temperature of the device (below 50 °C), EHPA is heated resistively to 900 °C analogously as MMQTA. All atomizers were operated in the online atomization mode under their optimum atomization conditions and under identical conditions for arsane generation (see Table 1). Sensitivity and LOD. Atomization conditions were initially optimized for all three atomizers. DBD optimization is discussed above. Atomization temperature and oxygen supply either as pure oxygen delivered through the capillary into the junction of inlet and optical arm of EHPA/MMQTA or as outer air delivered to MMQTA (see Figure 1 and Experimental for details) were optimized for MMQTA and EHPA atomizers (data not shown). Addition of air/oxygen to externally heated atomizers has been well-documented to increase sensitivity, improve linearity and resistance to interferences.2,3 Optimum atomization conditions for all atomizers tested are summarized

preconcentration 300 s EHPA

DBD

0.27 0.186 ± 0.002 1.14 ± 0.18 certified value

0.012 0.483 ± 0.023 1.03 ± 0.02

in Table 1. Sensitivity in EHPA is about 60% lower than in MMQTA (Table 2), which is in good agreement with the 1:2 ratio of their optical path lengths. This suggests that EHPA can be taken as a model of an externally heated atomizer. On the contrary, the sensitivity in DBD is not significantly different from that observed in MMQTA, although their optical path lengths are different. Because the noise level in all three atomizers studied was comparable, the LOD values are inversely proportional to the sensitivity as can be seen in Table 2. The accuracy and precision of MMQTA and DBD atomizers is comparable as demonstrated on As determination in certified reference material (Table 2). Interferences. Interferences of Sb, Se, and Bi on As determination were investigated because these hydride-forming elements are effectively cogenerated1,21,22 with arsane (generation efficiency ≥90%). Other important hydride-forming elements (Sn, Pb) were not investigated because generation efficiency of their hydrides under the employed conditions is very low.1,23 The results summarized in Table 3 indicate that Table 3. Effect of Other Hydride Forming Elements on Response from As in MMQTA and DBD Atomizers; Constant Analyte Concentration 5 ng mL−1 As (MMQTA) and 10 ng mL−1 As (DBD) recoveries (%) in the presence of an interferent concentration (ng mL−1) (uncertainties below 2%) interferent III

Sb

SeIV BiIII

atomizer

0

5

50

500

5000

MMQTA DBD MMQTA DBD MMQTA DBD

100 100 100 100 100 100

103 97 97 91 99 96

98 99 93 102 93 98

49 109 47 95 67 94

15 122 12 84 26 32

DBD atomizer provides one order of magnitude better resistance to interference from other hydride-forming elements than MMQTA. A detailed discussion of the results on interferences can be found in Section S2 of Supporting Information. Atomization Efficiency. The EHPA was designed to compare directly the atomization efficiency in the DBD atomizer and externally heated atomizers. Assuming a full atomization of arsane and no free atom decay in DBD atomizer as well as in EHPA, arsenic free atom cross-sectional density should be 3.5 times higher at 45 °C in the former than at 900 °C in the latter (see Table 1). Taking into account a more pronounced collisional broadening of the absorption line at the low temperature in the DBD atomizer,1 the atomic absorption coefficient should be approximately 1.5 times lower compared to that at the temperature of 900 °C in EHPA. In conclusion, the DBD atomizer should yield 2.3 times higher sensitivity than E


Article


repeatability was reached. The peak area of the signal in the volatilization step of the preconcentration mode is related to that in the online atomization mode under the same experimental conditions (standard concentration and sample introduction time) in both modes to quantify the preconcentration efficiency. A preconcentration efficiency of 98 ± 2% was found for 30 s sample introduction time. The peak profiles for the standard solution of 1.0 ng mL−1 As (2 ng As absolute) in the online and preconcentration modes with 30 s sample introduction time, respectively, can be compared in Figure 5.

the EHPA. The fact that 2.6 times higher sensitivity in DBD than in EHPA was found experimentally (Table 2) indicates the same atomization efficiency in both atomizers. Thus, atomization efficiency in DBD is estimated to be 100% of that reached in the MMQTA. Arsane Preconcentration in a DBD Atomizer. Previous experiments in online atomization mode (see section Effect of Discharge Gas) have revealed that arsenic species were retained in the DBD under oxygen-containing atmosphere and could be subsequently volatilized and atomized by Ar−H2 discharge. Usefulness of this approach for in situ preconcentration of As in a DBD atomizer was further studied. See Figure 2 for the apparatus “arrangement 1” employed. Oxygen was introduced into the DBD atomizer via a capillary. This experimental setup is further termed as “arrangement 1” in this section. Another way of gas introduction, termed “arrangement 2” in this section, was used for comparative reasons. It actually corresponds to that used in the online atomization mode (see Figure 1). However, an extra T-connector placed ca. 3 cm upstream the DBD atomizer is included to introduce oxygen to gaseous phase from GLS. It was found experimentally that there was no difference in peak area or peak profile when measurements in online atomization mode (i.e., with no oxygen addition) were performed in “arrangement 1” and “arrangement 2”. The preconcentration was performed (see section Procedures in the Experimental Section) employing “arrangement 1”. Figure 4 shows the effect of oxygen flow rate from 0.5 to 7 mL

Figure 5. Comparison of signal shapes for 1 ng mL−1 As standard solution (2 ng As absolute) in online atomization mode (black line) and preconcentration mode (red line); atomization conditions in both modes: 60 mL min−1 Ar, 17 W DBD power; preconcentration conditions: 30 s preconcentration, 7 mL min−1 O2.

Although the peak shapes are different, their areas are identical, indicating lossless analyte trapping and volatilization during preconcentration procedure. Whereas a broad peak with a full width in half-maximum (fwhm) corresponding to the sample introduction time of 30 s is registered in the online atomization mode, a narrow and high peak (fwhm 2.8 ± 0.2 s) is detected in the preconcentration mode (see Figure 5). Also, the effect of apparatus setup represented by “arrangements” “1 and 2” on arsane preconcentration was investigated. Both arrangements differ from each other only in the way of gases introduction as described above. Interestingly, a significantly lower preconcentration efficiency (89 ± 4%) and different peak profiles (fwhm 3.7 ± 0.2 s) were observed in the preconcentration mode in “arrangement 2” using the same experimental conditions (7 mL min−1 O2 in the trapping step, 30 s sample introduction time, 1.0 ng mL−1 As). The differences in the results between “arrangement 1” and “arrangement 2” might be explained by variances in mixing of the gases and flow patterns. “Arrangement 1” was thus employed in all further experiments with As preconcentration in DBD. The effect of sample introduction time on preconcentration efficiency of As was investigated in “arrangement 1” in a series of experiments employing standard solutions containing 0.1, 0.25, and 1 ng mL−1 As with sample introduction time of 300, 120, and 30 s so that 2 ng of As absolute were generated and preconcentrated in each measurement. No significant change in peak area was observed for sample introduction time between 30 and 300 s, indicating that no analyte losses occur with prolonged sample introduction time. As a consequence, arsane can be generated from larger sample volumes, and detection limits can be improved. Preconcentration efficiency of 102 ±

Figure 4. Effect of oxygen addition to Ar discharge gas on As signal in preconcentration mode; 1 ng mL−1 As, 60 mL min−1 Ar, 17 W DBD power, 30 s preconcentration.

min−1 added to the Ar discharge in the trapping step of the preconcentration procedure on As signal in the volatilization step. The transient signal of analyte released in the volatilization step was evaluated, but in addition, the signal during the trapping step (breakthrough signal) was monitored. The peak area of the volatilization signal steeply increases from 0.5 to 1.0 s when the flow of oxygen supplied to the capillary increases from 0.5 to 4 mL min−1 O2 followed by a plateau between 4 and 7 mL min−1 O2 (see Figure 4). The breakthrough signal (not shown) displayed the opposite trend. Although a breakthrough signal of 1.0 s, which was actually the signal observed in the online atomization mode, was detected in absence of oxygen addition, no breakthrough signals were detected for flow rates above 3.5 mL min−1 O2. An oxygen flow rate of 7 mL min−1 O2 was chosen as optimum for preconcentration because the highest signal with the best F


Article

Analytical Chemistry 4% was found for 300 s. Longer periods were not investigated in order to keep reasonable sample throughput. A calibration curve was constructed for 300 s sample introduction time employing standards with content of 0.03 to 0.2 ng mL−1 As. Analytical figures of merit of the preconcentration mode with 300 s sample introduction time are summarized in Table 2 including method validation by the measurement of certified reference material TM-RAIN-04. It can be deduced from Table 2 that lossless preconcentration can be reached even with 300 s sample introduction time since the sensitivity in this mode is the same as in the online mode. An improvement of detection limit proportional to the sample introduction time can be reached. The possibility of hydride-forming elements preconcentration in a DBD atomizer was first proposed in our study focused on Bi,18 where a preconcentration efficiency of 60% was reached without further optimization. To the best of our knowledge, this is the first report of lossless in situ preconcentration of a hydride-forming element in a planar DBD atomizer with AAS detection. A successful in atomizer preconcentration of arsane in a tubular-shaped DBD atomizer with AFS detection has been described recently by Mao et al.24 based also on oxygen addition to the discharge in order to trap arsane. No effort was made to estimate the preconcentration efficiency in their work.24 Moreover, much more complicated procedure is employed since discharge power supply rate has to be changed between the trapping and volatilization steps as well as gas flows have to be redirected in their setup.24 In situ preconcentration of arsane in a DBD atomizer described in this work is, owing to its 100% efficiency, suited for routine use. Much worse results are reached for in-atomizer preconcentration of arsane in a MMQTA, where only 50% preconcentration efficiency was observed under optimized conditions.13

■

CONCLUSIONS

■

ASSOCIATED CONTENT

■

Chemicals and standards; results of interference study (PDF)

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Czech Science Foundation (project no. P206/14-23532S and Institute of Analytical Chemistry of the CAS, v. v. i. (Institutional Research Plan, project no. RVO: 68081715). The authors are obliged to Mgr. Karel Marschner for theoretical calculations of the atomic absorption coefficients in the EHPA and DBD atomizers.

■

REFERENCES

(1) Dědina, J.; Tsalev, D. L. Hydride Generation Atomic Absorption Spectrometry; John Wiley & Sons, Inc.: Chichester, U.K., 1995. (2) Matoušek, T.; Dědina, J.; Selecká, A. Spectrochim. Acta, Part B 2002, 57, 451−462. (3) Dědina, J.; Matoušek, T. J. Anal. At. Spectrom. 2000, 15, 301−304. (4) Zhu, Z.; Zhang, S.; Lv, Y.; Zhang, X. Anal. Chem. 2006, 78, 865− 872. (5) Zhu, Z.; Liu, J.; Zhang, S.; Na, X.; Zhang, X. Anal. Chim. Acta 2008, 607, 136−141. (6) Xing, Z.; Kuermaiti, B.; Wang, J.; Han, G.; Zhang, S.; Zhang, X. Spectrochim. Acta, Part B 2010, 65, 1056−1060. (7) Yu, Y. L.; Du, Z.; Chen, M. L.; Wang, J. H. J. Anal. At. Spectrom. 2008, 23, 493−499. (8) Meyer, C.; Müller, S.; Gurevich, E. L.; Franzke, J. Analyst 2011, 136, 2427−2440. (9) Yang, M.; Xue, J.; Li, M.; Han, G.; Xing, Z.; Zhang, S.; Zhang, X. Talanta 2014, 126, 1−7. (10) Matusiewicz, H.; Sturgeon, R. E. Spectrochim. Acta, Part B 1996, 51, 377−397. (11) Dočekal, B.; Marek, P. J. Anal. At. Spectrom. 2001, 16, 831−837. (12) Dočekal, B.; Gücer, S.; Selecká, A. Spectrochim. Acta, Part B 2004, 59, 487−495. (13) Kratzer, J.; Dědina, J. Anal. Bioanal. Chem. 2007, 388, 793−800. (14) Braman, R. S.; Foreback, C. C. Science 1973, 182, 1247−1249. (15) Matoušek, T.; Hernandéz-Zavala, A.; Svoboda, M.; Langrová, L.; Adair, B. M.; Drobná, Z.; Thomas, D. J.; Stýblo, M.; Dědina, J. Spectrochim. Acta, Part B 2008, 63, 396−406. (16) Walcerz, M.; Garboś, S.; Bulska, E.; Hulanicki, A. Fresenius' J. Anal. Chem. 1994, 350, 662−666. (17) Svoboda, M.; Kratzer, J.; Vobecký, M.; Dědina, J. Spectrochim. Acta, Part B 2015, 111, 46−51. (18) Kratzer, J.; Boušek, J.; Sturgeon, R. E.; Mester, Z.; Dědina, J. Anal. Chem. 2014, 86, 9620−9625. (19) Taurková, P.; Svoboda, M.; Musil, S.; Matoušek, T. J. Anal. At. Spectrom. 2011, 26, 220−223. (20) Duben, O.; Boušek, J.; Dědina, J.; Kratzer, J. Spectrochim. Acta, Part B 2015, 111, 57−63. (21) Kratzer, J.; Vobecký, M.; Dědina, J. J. Anal. At. Spectrom. 2009, 24, 1222−1228. (22) Hraníček, J.; Č ervený, V.; Kratzer, J.; Vobecký, M.; Rychlovský, P. J. Anal. At. Spectrom. 2012, 27, 1761−1771. (23) Průsǎ , L.; Dědina, J.; Kratzer, J. Anal. Chim. Acta 2013, 804, 50− 58. (24) Mao, X.; Qi, Y.; Huang, J.; Liu, J.; Chen, G.; Na, X.; Wang, M.; Qian, Y. Anal. Chem. 2016, 88, 4147−4152.

Atomization of arsane in a planar DBD atomizer with AAS detection was in detail optimized in this work. Moreover, analytical figures of merit have been compared for DBD and MMQTA as the example of externally heated quartz tube atomizers being commonly used for arsane atomization. A direct comparison between the DBD and the EHPA revealed the same atomization efficiency in the DBD plasma atomizer as in externally heated devices. DBD can compete with MMQTA in terms of sensitivity and LOD. The essential advantage of DBD is that it provides one order of magnitude better resistance to atomization interferences from other hydrideforming elements (Sb, Se, and Bi) than MMQTA, which substantially ameliorates the poor resistance to atomization interferences of conventional externally heated quartz tube atomizers. The other fundamental advantage of the DBD is a simple and fast, 100% efficient approach to in situ preconcentration of As in the atomizer. The trapping and volatilization processes are controlled only by presence and absence of oxygen in the discharge gas, respectively. This opens a way to extremely low detection limits.

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.6b01365. G


Preconcentration and Atomization of Arsane in a Dielectric Barrier

Recommend Documents