Efficient Photochemical Vapor Generation of ... - ACS Publications

Article Cite This: Anal. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/ac

Efficient Photochemical Vapor Generation of Molybdenum for ICPMS Detection Jakub Š oukal,†,‡ Ralph E. Sturgeon,§ and Stanislav Musil*,† †

Institute of Analytical Chemistry of the Czech Academy of Sciences, Veveří 97, 602 00 Brno, Czech Republic Charles University, Faculty of Science, Department of Analytical Chemistry, Albertov 6, 128 43 Prague, Czech Republic § National Research Council of Canada, 1200 Montreal Road, Ottawa, Ontario K1A 0R6, Canada ‡

Downloaded via KAOHSIUNG MEDICAL UNIV on September 23, 2018 at 20:38:37 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Photochemical vapor generation (PVG) of Mo was accomplished using a 19 W high-efficiency flow-through photoreactor operating in a flow injection mode using 30−50% (w/v) formic acid as a reaction medium. The generated volatile product (most probably molybdenum hexacarbonyl) was directed by an argon carrier gas to a plastic gas−liquid separator and introduced into the spray chamber of an inductively coupled plasma mass spectrometer for detection. Particular attention was paid to the determination of overall PVG efficiency relative to that for liquid nebulization. Utilizing a sample flow rate of 1.25 mL min−1, corresponding to an irradiation time of 38 s, PVG efficiencies in the range 46−66% were achieved. The efficiency could be further enhanced by the presence of mg L−1 added Fe3+ ions. A limit of detection of 1.2 ng L−1 and precision of 3% (RSD) at 250 ng L−1 were achieved. Interferences from inorganic anions likely to be encountered during analytical application to real samples (NO3−, Cl−, SO42−, NO2−, and ClO4−) were investigated in detail. The accuracy and applicability of this sensitive methodology was successfully verified by analysis of fresh water Standard Reference Material NIST 1643e, two seawater Certified Reference Materials (NASS-7 and CASS-6), and by analysis of two samples of commercial dietary supplements solubilized in formic acid.

V

conduit of UV transmissible material placed in close proximity to the source. The sample is irradiated while passing through the conduit, and volatile products are efficiently transferred to the gas phase using a gas−liquid separator (GLS).5 An advanced high-efficiency flow-through photoreactor utilizes a modified low pressure mercury discharge lamp wherein the sample is irradiated during transit through an inner quartz channel directly immersed in the discharge.9 Since UV light has only to pass the thin synthetic quartz wall of the inner channel, this photoreactor permits ready access to intense 185 nm radiation, efficiently photolyzing the sample without concurrent ozone generation.5 Apart from elements typically amenable to HG10 and nonmetals (Br, Cl, and I),11−13 PVG has been successfully achieved with several transition metals, including Ni,14−20 Fe,20−22 Co,14,23,24 Cu,25 Cd,26 and Os.27 PVG of Rh, Pt, Pd, Ag, and Au was briefly reported in very early work28 and has only recently been followed up with a more detailed investigation, although absolute efficiencies were not estimated.29 Among the metals listed above, PVG of Ni, Fe, and

apor generation is a sample introduction technique for analytical atomic spectrometry yielding the advantages of efficient separation of an analyte from a matrix and significantly higher introduction efficiency than pneumatic nebulization. Among vapor generation techniques, hydride generation (HG) employing chemical reduction with tetrahydridoborate under acidic conditions remains the most popular and well established for trace elements of IV−VI main groups (i.e., As, Bi, Ge, Pb, Sb, Se, Sn, and Te), mercury, and cadmium.1,2 Several reports have demonstrated use of this reaction for chemical generation of volatile species of about 20 additional transition metals, although the practical applications remain seriously limited due to low generation efficiency, poor reproducibility, unknown mechanism, and identity of generated species as well as their instability.3,4 Photochemical vapor generation (PVG) provides a promising alternative to chemical generation and is applicable to a broader range of elements including metals, metalloids, and even nonmetals.5 With PVG, the analyte is converted to volatile species through the action of UV radiation. The presence of a photochemical agent in the liquid phase is required (e.g., formic or acetic acid).6−8 PVG generators usually consist of a UV source, most often a low-pressure mercury UV lamp emitting primarily 254 nm radiation, and a © XXXX American Chemical Society

Received: July 26, 2018 Accepted: September 7, 2018 Published: September 7, 2018 A

DOI: 10.1021/acs.analchem.8b03354 Anal. Chem. XXXX, XXX, XXX−XXX

Article

Analytical Chemistry

Figure 1. Schematic diagram of the PVG system connected to ICPMS for FI introduction of volatile molybdenum species.

(originally the inlet for makeup argon) to transfer the volatile molybdenum species directly to the ICPMS. Unless otherwise stated, carrier liquid (1% HNO3), mixed with an internal standard (IS) comprising 10 μg L−1 Rh in 2% HNO3, was nebulized into the spray chamber of the ICPMS with a MicroMist nebulizer concurrently with PVG sample introduction in order to provide more robust wet plasma conditions and to monitor plasma stability. The channel for carrier liquid was equipped with a manual injection valve (270 μL) that was utilized exclusively for determination of overall PVG efficiency when the sensitivities obtained with PVG sample introduction and using conventional solution nebulization were compared. Complete plasma settings are summarized in Table S1 of the Supporting Information. Isotopes of 95Mo (dwell time 0.1 s), 98 Mo (0.1 s), and 103Rh (0.05 s) were typically monitored. Measurements were performed in time-resolved analysis mode using He collision cell chemistry (4.1 mL min−1) to exclude the additive effect of any potential polyatomic interferences originating, for example, from BrO+, SeO+, ArCo+, ArFe+, or ArNi+, i.e., other concomitant contaminants in unknown samples that could also be efficiently generated by PVG. The decrease in 95Mo intensity in He mode in comparison to standard mode was not substantial (∼30%). All results are presented using 95Mo signal intensity for evaluation; no statistically significant difference was evident when 98Mo was used for evaluation. A standard tuning procedure was conducted daily to check the sensitivity. This was followed by further tuning while running PVG. The optimum nebulizer Ar flow rate was manually tuned/checked each day. The FI signals were exported to, and integrated in, MS Excel and corrected for any sensitivity drift relative to changes in the 103 Rh IS signal intensity. Reagents and Materials. Deionized water (DIW, < 0.2 μS cm−1, Ultrapur, Watrex, USA) was used for the preparation of all solutions. Formic acid (98%, p.a., Lach-Ner, Czech Republic) was purified in-house in a Teflon BSB-939-IR subboiling distillation apparatus (Berghof). A stock solution of 1000 mg L−1 Mo (as (NH4)2MoO4 in 0.5% NH4OH) was obtained from Absolute Standards, Inc. (USA). Stock solutions of 1000 mg L−1 Cu, Fe, and Ni in 1 M HCl were sourced from

Co from formic acid based media has gained particular attention, including several analytical applications due to its high efficiency and the intrinsic high stability of the generated volatile metal carbonyls that were clearly identified in the gas phase.15,21,30 The aim of this work was the achievement of efficient PVG of molybdenum. Molybdenum has been highlighted as a future target element in the field of PVG (together with W and Cr) in the latest review.5 However, its successful PVG has not yet been reported, although volatile molybdenum hexacarbonyl has been earlier detected in gaseous effluents from landfill and sewage.31,32

■

EXPERIMENTAL SECTION Instrumentation. All measurements were made using an Agilent 7700x single quadrupole inductively coupled plasma mass spectrometer (ICPMS). A schematic diagram of the PVG system connected to ICPMS is depicted in Figure 1. The highefficiency flow-through photoreactor comprised a 19 W lowpressure mercury discharge lamp (Beijing Titan Instrument Co., Ltd., Beijing, China) having a synthetic quartz central channel of ∼720 μL internal volume. Sample solutions were introduced in a flow-injection (FI) mode into a stream of the reaction medium with the aid of an injection valve (490 μL sample loop). Unless otherwise stated, it was delivered at a flow rate of 1.25 mL min−1 to the photoreactor using a peristaltic pump (Reglo ICC, Ismatec), which was also used to evacuate waste from the GLS. The effluent, mixed with a flow of argon (Ar chemifold), was directed to a plastic gas−liquid separator (GLS I, 50 mL internal volume) made from a polypropylene centrifuge vial. This GLS has been described in detail elsewhere.33,34 The only modification was that the tip of the PTFE waste tube was terminated 3 cm from the bottom of the GLS unit in order to maintain a stable liquid level inside the unit (∼10 mL). Occasionally, a modified GLS II was used which was equipped with a bottom frit (medium porosity) to introduce additional argon (Ar frit) directly through the discharged waste solution. Both separators were immersed in an ice-water bath. Unless otherwise stated, the outlet of the GLS was connected to the inlet of a Scott-type spray chamber B


Article

Analytical Chemistry BDH (UK). Other chemicals were obtained from various producers: ammonium hydroxide (25%, p.a.) from SigmaAldrich; acetic (99.8%, semiconductor grade), HNO3, and H2SO4 acids from Lach-Ner (Czech Republic); HCl, HClO4, NaNO3, and Na2SO4 from Merck (Germany); NaCl from Lachema (Czech Republic); and solid MoO2Cl2 from Puralab (Czech Republic). Three certified reference materials (CRMs) were used for method validation: NIST SRM 1643e (Trace Elements in Water), NASS-7 (Seawater Certified Reference Material for Trace Metals and other Constituents), and CASS-6 (Nearshore Seawater Certified Reference Material for Trace Metals and other Constituents). The accuracy of the methodology was also verified using two commercial samples of dietary supplements having declared Mo content (Centrum AZ and Supradyn CoQ10 Energy) purchased from a local pharmacy. Convention. Peak areas of FI signals were employed as a measure of analyte response unless otherwise stated. Averages of at least three replicates are shown in all relevant figures. Uncertainties are presented as plus or minus one standard deviation (SD) or combined SD where results are relative. If sensitivity is mentioned, it refers to peak area related to the analyte mass taken for the measurement. Limits of detection (LOD) are expressed as 3 × SD(blank)/slope of the calibration function. The overall PVG efficiency represents that fraction of the analyte that is generated, phase released, and transported to the ICP (i.e., that actually enters the plasma). Nebulization efficiency represents the traditional fraction of the nebulized analyte that enters the plasma. Safety Precaution. The exact products produced in this system are as yet unidentified [likely Mo(CO)6] and should be considered toxic. Essential safety precautions must be taken during all manipulations and an adequate ventilation/exhaust system used.

Figure 2. Effect of formic acid concentration on peak area from 1 μg L−1 Mo at reaction medium flow rate of 1.25 mL min−1. Uncertainties expressed as SD.

Argon (chemifold) carrier flow was optimized using 30% formic acid and 1 μg L−1 Mo standard to establish conditions for efficiently transferring the volatile species to the ICP from the photoreactor. The gas stream leaving the GLS was mixed with an additional flow of argon (not shown in Figure 1) before it was introduced into the spray chamber; care was taken to keep the total gas flow to the ICP the same, so as to not influence conditions in the plasma or sampling depth. No significant effect of Ar (chemifold) on sensitivity was observed in the range 100−600 mL min−1, suggesting efficient release and stability of the gaseous analyte. A logical 17 s increase in full-width-at-half-maximum (fwhm) of the measured peaks at 100 mL min−1 occurred in comparison to flow rates of 300− 600 mL min−1 due to increased time spent in the GLS (40 mL dead volume). However, 100 mL min−1 was finally chosen for further experiments in order to suppress any potential load of formic acid vapor on the ICP and because the FI peak shape was the most repeatable under these conditions. A typical FI signal at 100 mL min−1 is displayed in Figure 3. The performance of GLS II fitted with a frit at the bottom was examined using Ar flow rates of 0−50 mL min−1 passing through the frit to determine whether some additional fraction of generated species could be further released from the liquid

■

RESULTS AND DISCUSSION Coupling PVG with ICPMS. There are two ways to connect a vapor generator to a standard Agilent 7700x ICPMS instrument: either via a special HMI port placed downstream of the ICPMS spray chamber (see Figure 1) or by use of a port typically used for admitting makeup gas into the spray chamber. In both cases, simultaneous nebulization of liquid IS is possible, creating more robust (wet plasma) conditions in the ICP and permitting the monitoring/correction for any sensitivity drift due to changes in the plasma or interface transmission efficiency.33,35 Efficient generation of volatile molybdenum species was achieved at rather higher concentrations of formic acid in the reaction medium, as shown in Figure 2, in agreement with typical conditions optimized for other transition metals that form volatile metal carbonylsNi, Fe, and Co.15,17,22,23,36 Due to its high concentration, a significant volume flow of gaseous byproducts (CO, CO2, and H2) is generated in the inner quartz tube of the photoreactor by direct photolysis of the formic acid.37 As a result, a segmented flow of the reaction medium arises due to bubbles of these gases, and the supply of analyte to the GLS is thus not uniform. An introduction of volatile species to the ICPMS spray chamber was ultimately adopted because these fluctuations of analyte supply were smoothed by the additional dead volume in the spray chamber. A very similar situation arises with classical hydride generation, as encountered when coupling to an atomic fluorescence spectrometer operating with a diffusion flame atomizer.38

Figure 3. Typical PVG transient signal from 1 μg L−1 Mo in 30% formic acid generated by FI-PVG-ICPMS (black line). Continuous signal from concurrent nebulization of a solution of 10 μg L−1 Rh is included for comparison (red line). C


Article

Analytical Chemistry phase. Argon (chemifold) flow was concurrently changed, depending on the selected flow rate of Ar (frit), such that the total gas flow leaving GLS II remained constant. No influence of such efficient (sparging) bubbling through the liquid phase was found, suggesting that all volatile species have been phase released into the transport gas before entering the GLS, i.e., the generated volatile species appear to be completely partitioned into the gas bubble segments leaving the photoreactor itself. Hence, the simpler GLS I containing no frit was used for all further work. PVG Parameters. The most important parameters influencing PVG were investigated, including concentration and type of reaction medium and sample irradiation time (IT), which is inversely proportional to the flow rate of the sample through the photoreactor. The influence of IT (30% formic acid medium) is depicted in Figure 4. In general, lower flow rates generate larger volumes

which it decreases. This phenomenon has been noted in many other PVG studies5 and is partly the result of a decreased penetration depth of the UV caused by the high absorptivity of the reaction solution. In addition to this effect, an IT of 38 s achieved with 30% formic acid at 1.25 mL min−1 was not maintained at higher concentrations of formic acid due to formation of increasingly larger volumes of bubbles that forced the sample from the photoreactor. At 50% and 70% formic acid, the ITs decreased to 35 and 31 s, respectively; hence the decrease in sensitivity for higher concentrations is also associated with a decrease in IT (as in Figure 4 for higher sample flow rates). For further investigations, 30% formic acid at a flow rate of 1.25 mL min−1 was selected as a compromise between achieved sensitivity on the one hand and limitations arising from analyte contamination in the formic acid (increased blank) compromising limits of detection (cf. Figures of Merit). Effect of pH and Additives. The effect of pH of the reaction medium was tested by the addition of ammonium hydroxide. Various volumes of NH4OH were added, forming ammonium formate (from 0.2 to 1.5 M) while keeping the concentration of formic acid constant at 30% (w/v) in the final solution. The sensitivity gradually decreased from its maximum for 30% formic acid (pH 1.1) to around 50% response for 30% formic acid containing 1.5 M ammonium formate (pH 3.2). This observation is not consistent with that noted for PVG of other transition metal carbonyls, namely, Ni, Fe, and Co.15,22−24 The addition of acetic acid in the range 0.5−5% to the 30% formic acid did not enhance PVG efficiency, and in a mixture containing 10% acetic acid, sensitivity decreased to around 60%. A significant enhancement of PVG efficiency due to the presence of added transition metal ions, namely, Cu2+, Co2+, Ni2+, and Fe3+, has been highlighted by several authors. These ions possibly participate in the formation of more sensitive precursors for UV oxidation of the reaction medium and hence influence the kinetics of reduction.12,19,39,40 Various concentrations (0.01−20 mg L−1) of Fe3+, Ni2+, and Cu2+ added to a 1 μg L−1 Mo standard were examined for potential enhancement of PVG efficiency of Mo. An almost 1.4-fold increase in sensitivity (blank corrected) was evident for Fe3+ at mg L−1 levels, while for Ni2+ and Cu2+ no, or only negligibly positive, effects were found. Detailed results are presented in Figure S1 of the Supporting Information. Unfortunately, the addition of 20 mg L−1 Fe resulted in significant Mo contamination, which was present in the iron AAS standard used, and amounted to a blank of 270 ng L−1(determined by liquid nebulization ICPMS). Hence, a high-purity Fe standard is required for practical utilization of the benefit of its addition for ultratrace analysis and for this reason was not further pursued in this work. Interferences. Previous studies have identified serious interferences from common inorganic anions in PVG systems.5,41 An examination of potential interferences was thus performed considering the possible application of the developed methodology for the analysis of Mo in real water matrices (CRMs), typically stabilized in a dilute nitric acid medium, or in matrices requiring acid digestion of a sample prior to analysis. Figure 5 illustrates the impact of nitrate, nitrite, chloride, perchlorate, and sulfate anions. The influence was studied using mineral acid solutions and/or their available salts added to a sample. Generally, the impact of the respective anions was very similar, regardless of whether salt or acid was

Figure 4. Effect of reaction medium/sample flow rate (30% formic acid reaction medium) on peak area response from 1 μg L−1 Mo. Uncertainties represented as SD are sufficiently small that they cannot be discerned from the data points.

of bubbles due to increased oxidation of the formic acid, resulting in segmented flow of the reaction medium. Hence, the IT is not linearly (inversely) proportional to the flow rate, especially at low values, as the expanding bubbles tend to expel the liquid from the reactor. The highest sensitivity was obtained in the range 1−1.25 mL min−1, corresponding to ITs of 47 to 38 s, respectively, as measured experimentally by following an air bubble introduced into the reaction medium while the PVG process was in operation. These times include not only the time spent by the sample directly in the irradiation zone but also that in two quartz tube segments on either side of the photoreactor (∼0.25 mL) which connect the interior quartz tubes (0.72 mL); these segments do not receive efficient radiation from the discharge. A notable decrease in the signal occurs at higher flow rates, i.e., 1.5, 2, and 2.5 mL min−1 (corresponding to ITs of 33, 25, and 21 s, respectively), logically due to insufficient irradiation of a sample. The notable loss of signal at flow rates below 0.75 mL min−1 (IT = 53 s) may be the result of decomposition of the analyte species in the photoreactor due to the long UV exposure time and/or increase in temperature of the irradiated solution. As noted above, efficient PVG was achieved in formic acid with no response being obtained using acetic acid as the reaction medium (tested up to 90% (w/v)). The sensitivity reaches a maximum in 50% formic acid (Figure 2), above D


Article

Analytical Chemistry

Table 1. Enhancement Factors and Overall PVG Efficiencies under Different Conditions enhancement factora

overall PVG efficiencyb %

formic acid formic acid + 20 mg L−1

8.2 ± 0.6 10.8 ± 0.7

49.5 ± 3.4 64.6 ± 4.3

formic acid formic acid + 20 mg L−1

10.1 ± 0.9 12.2 ± 1.1

60.4 ± 5.4 73.1 ± 6.4

conditions 30% 30% Fe 50% 50% Fe

Enhancement factor = increase in sensitivity (count ng−1) compared to direct solution nebulization. bOverall PVG efficiency = fraction of the analyte that enters plasma; based on comparative 6% nebulization efficiency.

a

supports the high stability of volatile Mo species that ensures its efficient transport into the ICP. Figures of Merit. Calibration functions constructed with 0, 10, 50, 250, 1000, and 4000 ng L−1 Mo standards under chosen PVG (30% formic acid, reaction medium flow rate of 1.25 mL min−1) and ICPMS conditions (Table S1) were linear (R2 = 0.9999) for both 95Mo and 98Mo isotopes. When nondistilled formic acid was used for PVG, high background intensity for both isotopes emerged, reflecting serious Mo contamination, corresponding to approximately 200 ng L−1 Mo in 30% formic acid. Purification by triple sub-boil distillation decreased Mo contamination approximately 25fold, which was reflected in a proportional enhancement in the limit of detection (LOD). Relative and absolute LODs (3σ, n = 10) reached 1.2 ng L−1 and 0.6 pg; however, these were still blank-limited. An estimation of an instrumental LOD, not influenced by Mo contamination from the reaction medium, was carried out to illustrate the ultimate potential of the developed methodology. The 30% formic acid medium was replaced with DIW passing through the photoreactor while the UV lamp was off during measurements of the blank signals. The instrumental LOD obtained in this way was thus influenced only by plasma background and contributions from a nebulized solution and reached 0.2 ng L−1 (3σ, n = 12), i.e., 6-fold lower. Repeatability (RSD) of 14 measurements of a 250 ng L−1 standard over the entire day was approximately 3%. No evident clogging of the cones of the ICPMS or drift in Mo and Rh (IS) sensitivity due to possible carbon deposition (from the introduced formic acid vapors) occurred during long-term use of 30% formic acid. Applications to Real Samples. To validate accuracy of the FI-PVG-ICPMS methodology, three water CRMs were analyzed. Each of these materials is stabilized in nitric acid NIST CRM 1643e in 0.8 M HNO3, NASS-7, and CASS-6 pH stabilized at 1.6 (approximately 0.025 M HNO3). Since nitrates seriously interfere at concentrations above 1 mM (Figure 5), the SRM and CRMs were diluted 1000- and 100fold, respectively, and analyzed using a standard additions technique (comprising two spiked concentrations). Results, summarized in Table 2, are in good agreement with the certified values. The recoveries from the diluted sample media, calculated as a ratio of slopes of the standard additions and external calibration functions, were 55 ± 5% for NIST SRM 1643e, 83 ± 2% for NASS-7, and 85 ± 1% for CASS-6, suggesting an acceptable level of remaining matrix interference. The methodology was also tested on two dietary supplements in tablet form with Mo present at a declared

Figure 5. Relative effects of added inorganic acids and salts on PVG of 1 μg L−1 Mo in 30% formic acid.

added. The only exception may be identified for sulfuric acid and sodium sulfate, although the observed difference is not too significant. Since the PVG system was generally very tolerant to the presence of sulfate ions, this difference could be due to the concurrent change in pH of the reaction medium, which is not negligible at these concentrations of acidic interferences. The most serious suppressive effect was evident for nitrates, starting at concentrations as low as 0.1 mM. The mechanism of interference by nitrate was very recently studied using the Se PVG system with 5% acetic acid as the reaction medium.41 In addition to nitrates, nitrites were found to exhibit a similar suppressive effect, but at even 10−20-fold lower concentrations. Therefore, it was speculated that it was actually nitrite that was directly responsible for the interference during PVG as it is a product of photoreduction of nitrate.41 Nevertheless, it is evident from Figure 5 that the impact of nitrites on PVG of Mo is not as serious as that of nitrites on Se, even at 10-fold higher concentrations. This contrasting behavior must be related to the specific mechanism of PVG and reduction of Mo in the different medium used in this work. Overall PVG Efficiency. The overall PVG efficiency was assessed from comparison of the sensitivity obtained with PVG sample introduction to that using conventional solution nebulization. Molybdenum standards were introduced either with the injection valve available in the PVG system or with a chemifold used for sample delivery for nebulization (see Figure 1). When an injected standard (prepared in 1% HNO3) was nebulized, PVG was concurrently in operation such that the conditions in the plasma were unchanged and vice versa. The chosen optimal conditions yielding the highest sensitivity were 30% and 50% formic acid with and without the addition of 20 mg L−1 Fe3+. The enhancement factors (compared to nebulization) and estimated efficiencies are summarized in Table 1. A nebulization efficiency of 6%, derived from previous work, was used as a reference and lies in the middle of the earlier reported range of 5.1−6.6%.42,43 Since volatile molybdenum species are driven from the GLS through the Scott-type spray chamber chilled to 2 °C, a determination of potential transport loss in the spray chamber was also of interest. An HMI port, close to the torch (see Figure 1), was utilized in this case for vapor introduction instead of the spray chamber, and the sensitivities were compared. The loss of volatile species in the spray chamber was determined to be 6.7% ± 3.0%, which was not critical and E


Article

Analytical Chemistry Table 2. Concentrations of Mo in Water CRMs (μg L−1) Determined by FI-PVG-ICPMS Using Standard Addition Calibration water

found

certified

NIST SRM 1643e (fresh water) NASS-7 (seawater) CASS-6 (nearshore seawater)

127.7 ± 18.9 9.51 ± 0.37 8.98 ± 0.07

121.4 ± 1.3 9.29 ± 0.40 9.15 ± 0.52

atoms by CO derived from photolysis of the formic acid. Papaconstantinou47,48 spectrophotometrically followed the one-, two-, four-, and six-electron stepwise reduction of the isopoly [Mo7O24]6− and [Mo8O26]4− as well as the heteropoly species [P2Mo18O62]6− by organic free radicals derived from γradiolysis of solutions of simple alcohols and formic acid. On this basis, the reduction potential of the (protonated) carboxyl anion radical at pH ≈ 1 (pKa = 1.6) was estimated to be less than −0.89 V (vs SHE) as it was capable of undertaking complete six-electron reduction of all of the Mo species whose redox potentials are all more positive than this. Thermodynamically, the more highly reducing •CO2− radicals arising from deep UV photolysis of formic acid are capable of effecting complete reduction of aquo-molybdate species to Mo(0). In support of this, saturation of Mo standard (1 μg L−1) solutions with CO by bubbling (30 min) before PVG led to a slight enhancement in sensitivity by 26 ± 6% and 7 ± 2% in 10% and 30% formic acid media, respectively. The lower yield of Mo(CO)6 (as well as the carbonyls of Fe, Ni, and Co) from an acetic acid medium may be a consequence of the rapid uptake of CO by •CH3 radicals to yield CH3•CO, subsequently generating CH3COCOCH3, CH3CHO, and CH3COOH.49 The observed enhancement effect of Fe3+ on PVG efficiency may be explained by formation of an Fe3+ formate complex which exhibits significant UV absorption (starting already below 400 nm). Hence, its presence in the reaction medium enhances the rate of photo-oxidation of formic acid5,7 and/or alters the reaction scheme in favor of production of highly reducing radicals such as the aquated electron and •HCO2 as well as •CO2− radicals50 responsible for reduction of the molybdate species. PVG of Mo was also tested using a 15 W germicidal mercury UV lamp (low pressure, Cole-Parmer), a common source used to irradiate sample solution flowing within a 6 m length of 1 mm i.d. and 1.5 mm o.d. PTFE tubing wrapped around the lamp. This is the same experimental arrangement earlier used by Guo et al. and Grinberg et al. for generation of Ni and Co carbonyls.15,30 Although good PVG efficiency for Ni carbonyl using 30% and 50% formic acid and a sample flow rate of 1.5 mL min−1 could be readily achieved with this arrangement, absolutely no signal for Mo was detected using 30% formic acid, and only a negligible response of ∼0.1% (corresponding to an absolute PVG efficiency of only 0.05%) was obtained using 50% formic acid in comparison to a response with the high-efficiency flow-through photorector. As shown by Campanella et al.,51 PTFE of 0.08 mm thickness absorbs 70% incident UV at 185 nm. Taking into account the thickness of PTFE tubing used here for PVG (i.e., ∼0.25 mm), it is obvious that it does not support efficient transmission of VUV radiation. Moreover, the high transmittance of 185 nm radiation through a synthetic quartz tube as compared to other types of quartz material was noted in the original study devoted to the development of the flow-through photoreactor.9 This points to the need for vacuum UV irradiation (185 nm) being required for efficient photoreduction of Mo.

concentration of 50 μg per tablet. Tablets were dissolved in 100 mL of 30% formic acid by sonication in an ultrasonic bath for 2 h at 50 °C. One aliquot was diluted 1000-fold with 30% formic acid and subjected to analysis by FI-PVG-ICPMS using standard additions for calibration. A second aliquot was diluted 100-fold with 1% HNO3 and analyzed using conventional solution nebulization ICPMS with calibration against external standards. Comparison of the results is presented in Table 3, Table 3. Comparison of Concentrations of Mo in Dietary Supplement Samples (μg per Tablet) Centrum AZ Supradyn CoQ10 Energy

FI-PVGa

nebulizationb

declared

56.8 ± 5.1 45.7 ± 2.9

56.7 ± 1.9 46.8 ± 1.6

50 50

a

Determined by FI-PVG-ICPMS using standard additions calibration. Determined by solution nebulization-ICPMS using external calibration.

b

with excellent agreement between both approaches. The results are also close to the declared amount of Mo in the tablets, fulfilling well the regulation of the European Parliament and of the Council concerning tolerances of +45 and −25% on declared content for vitamins and minerals in food supplements.44 Spike recoveries from diluted samples were 102 ± 9% and 100 ± 7%, enabling fit-for-purpose quantification even using external calibration. Possible PVG Mechanism. Although Mo species are present in the environment in several oxidation states, covering the range VI, V, IV, III, II, I, 0, −I, and −II, the VI state is dominant and soluble in water. In aqueous solutions, various polymeric species (polymolybdates) are formed, depending on the Mo concentration and the pH.45 The occurrence of various species may be a concern for PVG based methodology if different pathways and mechanisms of reduction occur, impacting accuracy. In acidic solutions at pH ≈ 1 (equivalent to 30% formic acid) and Mo concentrations lower than 100 mg L−1, only protonated monomeric H2MoO4 or MoO3(H2O)3 forms are likely present.46 A simple experiment was conducted using various AAS standards (Mo either in the form of (NH4)2MoO4 or prepared from pure metal and dilute HCl (1:1) with traces of HNO3) or solid MoO2Cl2 dissolved in DIW, to confirm that the PVG efficiency was the same for all species tested. No significant difference in the enhancement factor/PVG efficiency (in comparison to the value in Table 1) was evident, supporting the assumption that potential speciated forms give rise to identical response. It is assumed that the generated volatile species is Mo(CO)6 (bp of 156 °C) due to its high stability31,32 and similar chemistry to the PVG of carbonyls identified for Ni, Fe, and Co.15,21,30 This remains to be conclusively proven by GC-MS. A likely mechanism for PVG of Mo(CO)6 as a final product would involve the sequential multielectron reduction of Mo(VI) to Mo(0) followed by rapid uptake of the nascent

■

CONCLUSIONS Highly efficient PVG of molybdenum has been demonstrated for the first time using a flow-through UV lamp, permitting a blank-limited LOD of 1.2 ng L−1 to be achieved with ICPMS detection. The generated volatile species is presumably stable Mo(CO)6, but this remains to be verified. The highest sensitivity is achieved using 50% formic acid at a flow rate of F


Article

Analytical Chemistry 1.25 mL min−1. The addition of Fe3+ ions at low mg L−1 concentrations further increases the overall PVG efficiency from 60% to nearly 75%. For analytical applications, compromise conditions employing 30% formic acid were used to minimize the blank, with which 50% efficiency was reached. As with most analytes, application of the PVG methodology to real samples suffers from serious interferences from nitrates while it is quite tolerant to chlorides as well as sulfates. Nevertheless, good accuracy was demonstrated for the analysis of water CRMs using standard addition calibration as well as by the analysis of two dietary supplements of declared Mo content which were readily solubilized with formic acid.

■

(14) de Quadros, D. P. C.; Borges, D. L. G. Microchem. J. 2014, 116, 244−248. (15) Guo, X. M.; Sturgeon, R. E.; Mester, Z.; Gardner, G. Appl. Organomet. Chem. 2004, 18, 205−211. (16) Hu, Y.; Wang, Q.; Zheng, C. B.; Wu, L.; Hou, X. D.; Lv, Y. Anal. Chem. 2014, 86, 842−848. (17) Liu, L. W.; Deng, H.; Wu, L.; Zheng, C. B.; Hou, X. D. Talanta 2010, 80, 1239−1244. (18) Zeng, Y.; Zheng, C. B.; Hou, X. D.; Wang, S. L. Microchem. J. 2014, 117, 83−88. (19) Zheng, C. B.; Sturgeon, R. E.; Hou, X. D. J. Anal. At. Spectrom. 2009, 24, 1452−1458. (20) Zheng, C. B.; Yang, L.; Sturgeon, R. E.; Hou, X. D. Anal. Chem. 2010, 82, 3899−3904. (21) Grinberg, P.; Sturgeon, R. E.; Gardner, G. Microchem. J. 2012, 105, 44−47. (22) Zheng, C. B.; Sturgeon, R. E.; Brophy, C. S.; He, S. P.; Hou, X. D. Anal. Chem. 2010, 82, 2996−3001. (23) Coutinho de Jesus, H.; Grinberg, P.; Sturgeon, R. E. J. Anal. At. Spectrom. 2016, 31, 1590−1604. (24) Deng, H.; Zheng, C. B.; Liu, L. W.; Wu, L.; Hou, X. D.; Lv, Y. Microchem. J. 2010, 96, 277−282. (25) Gao, Y.; Sturgeon, R. E.; Mester, Z.; Pagliano, E.; Galea, R.; Saull, P.; Hou, X. D.; Yang, L. Microchem. J. 2016, 124, 344−349. (26) Nóbrega, J. A.; Sturgeon, R. E.; Grinberg, P.; Gardner, G. J.; Brophy, C. S.; Garcia, E. E. J. Anal. At. Spectrom. 2011, 26, 2519− 2523. (27) Gao, Y.; Li, S. H.; He, H.; Li, T. L.; Yu, T.; Liu, R.; Ni, S. J.; Shi, Z. M. Microchem. J. 2017, 130, 281−286. (28) Guo, X. M.; Sturgeon, R. E.; Mester, Z.; Gardner, G. J. Anal. Chem. 2004, 76, 2401−2405. (29) Macedo de Oliveira, R. M.; Borges, D. L. G. UV photochemical vapor generation of noble metals (Au, Ir, Pd, Pt and Rh): a feasibility study using inductively coupled plasma mass spectrometry and seawater as a test matrix. J. Anal. At. Spectrom. 2018, [Advance Article] DOI: 10.1039/C8JA00179K. (30) Grinberg, P.; Mester, Z.; Sturgeon, R. E.; Ferretti, A. J. Anal. At. Spectrom. 2008, 23, 583−587. (31) Feldmann, J.; Cullen, W. R. Environ. Sci. Technol. 1997, 31, 2125−2129. (32) Feldmann, J. J. Environ. Monit. 1999, 1, 33−37. (33) Matoušek, T.; Currier, J. M.; Trojánková, N.; Saunders, R. J.; Ishida, M. C.; González-Horta, C.; Musil, S.; Mester, Z.; Stýblo, M.; Dědina, J. J. Anal. At. Spectrom. 2013, 28, 1456−1465. (34) Musil, S.; Matoušek, T.; Currier, J. M.; Stýblo, M.; Dědina, J. Anal. Chem. 2014, 86, 10422−10428. (35) Musil, S.; Pétursdóttir, Á . H.; Raab, A.; Gunnlaugsdóttir, H.; Krupp, E.; Feldmann, J. Anal. Chem. 2014, 86, 993−999. (36) Cai, Y.; Li, S. H.; Dou, S.; Yu, Y. L.; Wang, J. H. Anal. Chem. 2015, 87, 1366−1372. (37) Adams, G. E.; Hart, E. J. J. Am. Chem. Soc. 1962, 84, 3994− 3999. (38) Marschner, K.; Musil, S.; Dědina, J. Anal. Chem. 2016, 88, 4041−4047. (39) Gao, Y.; Xu, M.; Sturgeon, R. E.; Mester, Z.; Shi, Z. M.; Galea, R.; Saull, P.; Yang, L. Anal. Chem. 2015, 87, 4495−4502. (40) Wang, Y. L.; Lin, L. L.; Liu, J. X.; Mao, X. F.; Wang, J. H.; Qin, D. U. Analyst 2016, 141, 1530−1536. (41) Lopes, G. S.; Sturgeon, R. E.; Grinberg, P.; Pagliano, E. J. Anal. At. Spectrom. 2017, 32, 2378−2390. (42) Š oukal, J.; Benada, O.; Matoušek, T.; Dědina, J.; Musil, S. Anal. Chim. Acta 2017, 977, 10−19. (43) Vyhnanovský, J.; Kratzer, J.; Benada, O.; Matoušek, T.; Mester, Z.; Sturgeon, R. E.; Dědina, J.; Musil, S. Anal. Chim. Acta 2018, 1005, 16−26. (44) Guidance Document for Competent Authorities for the Control of Compliance with EU Legislation on: Regulation (EU) No 1169/2011 of the European Parliament and of the Council of 25 October 2011 on the provision of food information to consumers,

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.8b03354. ICPMS parameters used for PVG; details on the impact of added transition metal ions, namely, Fe3+, Ni2+, and Cu2+, on net response from a 1 μg L−1 Mo standard (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Stanislav Musil: 0000-0001-8003-0370 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The support of the Czech Academy of Sciences (Institutional support RVO: 68081715) and Charles University (project SVV260440) is gratefully acknowledged. We are also grateful to Prof. Qiuquan Wang (Xiamen University) for assistance with purchasing the photoreactors and Dr. Milan Svoboda for sub-boiling distillation of formic acid.

■

REFERENCES

(1) Dědina, J.; Tsalev, D. L. Hydride Generation Atomic Absorption Spectrometry; John Wiley & Sons: Chichester, UK, 1995. (2) Dědina, J. In Encyclopedia of Analytical Chemistry, Supplementary Vols. S1−S3; Meyers, R. A., Ed.; John Wiley & Sons: Chichester, UK, 2011; pp 897−936. (3) Pohl, P.; Jamroz, P.; Welna, M.; Szymczycha-Madeja, A.; Greda, K. TrAC, Trends Anal. Chem. 2014, 59, 144−155. (4) D’Ulivo, A. Spectrochim. Acta, Part B 2016, 119, 91−107. (5) Sturgeon, R. E. J. Anal. At. Spectrom. 2017, 32, 2319−2340. (6) He, Y. H.; Hou, X. D.; Zheng, C. B.; Sturgeon, R. E. Anal. Bioanal. Chem. 2007, 388, 769−774. (7) Sturgeon, R. E.; Grinberg, P. J. Anal. At. Spectrom. 2012, 27, 222−231. (8) Yin, Y. G.; Liu, J. F.; Jiang, G. B. TrAC, Trends Anal. Chem. 2011, 30, 1672−1684. (9) Qin, D.; Gao, F.; Zhang, Z. H.; Zhao, L. Q.; Liu, J. X.; Ye, J. P.; Li, J. W.; Zheng, F. X. Spectrochim. Acta, Part B 2013, 88, 10−14. (10) Zheng, C. B.; Ma, Q.; Wu, L.; Hou, X. D.; Sturgeon, R. E. Microchem. J. 2010, 95, 32−37. (11) Sturgeon, R. E. Anal. Chem. 2015, 87, 3072−3079. (12) Hu, J.; Sturgeon, R. E.; Nadeau, K.; Hou, X. D.; Zheng, C. B.; Yang, L. Anal. Chem. 2018, 90, 4112−4118. (13) Grinberg, P.; Sturgeon, R. E. J. Anal. At. Spectrom. 2009, 24, 508−514. G


Article

Analytical Chemistry amending Regulations (EC) No 1924/2006 and (EC) No 1925/2006 of the European Parliament and of the Council, and repealing Commission Directive 87/250/EEC, Council Directive 90/496/EEC, Commission Directive 1999/10/EC, Directive 2000/13/EC of the European Parliament and of the Council, Commission Directives 2002/67/EC and 2008/5/EC and Commission Regulation (EC) No 608/2004 and Council Directive 90/496/EEC of 24 September 1990 on nutrition labelling of foodstuffs and Directive 2002/46/EC of the European Parliament and of the Council of 10 June 2002 on the approximation of the laws of the Member States relating to food supplements with regard to the setting of tolerances for nutrient values declared on a label. https://ec.europa.eu/food/sites/food/ files/safety/docs/labelling_nutrition-supplements-guidance_ tolerances_1212_en.pdf (accessed July 24, 2018). (45) Aveston, J.; Anacker, E. W.; Johnson, J. S. Inorg. Chem. 1964, 3, 735−746. (46) Oyerinde, O. F.; Weeks, C. L.; Anbar, A. D.; Spiro, T. G. Inorg. Chim. Acta 2008, 361, 1000−1007. (47) Papaconstantinou, E. Anal. Chem. 1975, 47, 1592−1595. (48) Papaconstantinou, E. Chem. Soc. Rev. 1989, 18, 1−31. (49) Bakač, A. Croat. Chem. Acta 2001, 74, 633−640. (50) Baxendale, J. H.; Bridge, N. K. J. Phys. Chem. 1955, 59, 783− 788. (51) Campanella, B.; Menciassi, A.; Onor, M.; Ferrari, C.; Bramanti, E.; D'Ulivo, A. Spectrochim. Acta, Part B 2016, 126, 11−16.

H


Efficient Photochemical Vapor Generation of ... - ACS Publications

Recommend Documents