Direct Detection of Inorganic Nitrate Salts by Ambient Pressure Helium

Nov 29, 2012 - Yu Haiyan , Chaimbault Patrick , Clarot Igor , Chen Zilin , Leroy Pierre ... Isra Hassan , Julius Pavlov , Ramu Errabelli , Athula B. A...
0 downloads 0 Views 1000KB Size
Download PDF
Article pubs.acs.org/ac

Direct Detection of Inorganic Nitrate Salts by Ambient Pressure Helium-Plasma Ionization Mass Spectrometry Julius Pavlov and Athula B. Attygalle* Center for Mass Spectrometry, Department of Chemistry, Chemical Biology, and Biomedical Engineering, Stevens Institute of Technology, Hoboken, New Jersey 07030, United States S Supporting Information *

ABSTRACT: Inorganic nitrates in solid deposits were detected directly by ambient-pressure helium-plasma ionization-mass spectrometry (HePI-MS), without the need for extensive sample preparation. Nitrates were detected even from complex matrices such as meats and fruit juices. Any electrospray-ionization mass spectrometer can be modified to perform ambient-pressure HePI-MS by simply passing helium through the metal capillary intended for liquid-sample delivery. Nitrates on paper strips, glass slides, or cotton swabs (sometimes wetted with a mineral acid) were inserted directly into the ambient-pressure HePI source. The spectra acquired under negative-ion generating conditions showed a peak at m/z 62 for the nitrate ion, along with a lower-intensity peak at m/z 125 for the nitrate adduct of nitric acid. Apparently, it is nitric acid that is initially transferred to the gas phase, forming an ion−molecule complex with hydroxyl anions present in the plasma. The ion-neutral complex then dissociates by eliminating water to produce gaseous NO3− ions. This hypothesis was supported by the observation that certain solid nitrate salts, which were not readily amenable to HePI (notably the alkali nitrates), were immediately detected as m/z 62 and 125 ions upon acidification by a strong acid. Quantitative evaluations showed that the nitrate-signal response versus the deposited mass is linear for over 3 orders of magnitude. With the use of 15N-labeled nitrate (m/z 63), the limit of detection was determined to be as low as 200 fg.

N

methods require a certain degree of sample preparation and modification by derivatization,2,3 complexation,6 isotope dilution,8 and ion-exchange,12 prior to analysis per se. In addition, most of these methods are prone to significant matrix interferences. Although techniques such as direct analysis in real time (DART)15 have revolutionized sample analysis by ambient-ionization mass spectrometry, because the need for a solvent spray can be eliminated, the detection of inorganic ions has remained particularly difficult because of their low volatilities. In biology and medicine, the analytical determination of NO3− concentration in body fluids is important to monitor the levels of the endogenous signaling molecule NO,4,5,16 which is usually accomplished only by GC/MS or GC-IRMS after lengthy derivatization procedures. Recently described helium-plasma ionization-mass spectrometry (HePI-MS)17,18 has been recognized as one of the most promising novel procedures for detecting volatile and semivolatile organic compounds.19 It is a simple adaptation that enables any conventional electrospray ionization mass spectrometer to be converted to a helium-plasma instrument by passing high-purity helium gas through the sample delivery metal tube.18 When a high potential is applied to the capillary, a zone of plasma is observed at the tip of the tube. We were interested to find out if the HePI source can generate gaseous ions from inorganic salts. In particular,

itrate is one of the primary inorganic nutrients that control the biomass production in aquatic and terrestrial habitats. Thus, potassium, sodium, and ammonium nitrates, generally known as saltpeters, are widely used as nitrogen fertilizers in agriculture. Moreover, many military propellants (e.g., black powder) and commercial blasting agents [e.g., ammonium nitrate−fuel oil (ANFO)] use nitrates because of their powerful oxidative properties. A far more disturbing reality is the misuse of fertilizer-based nitrates as oxidants for the construction of vehicle-borne improvised explosive devices (VBIEDs).1 In actual fact, all major ground terrorist attacks in the Western hemisphere, from the Oklahoma City bombing in 1995 to the events in Oslo in 2011, employed ammonium nitrate-based bombs. To control the damages that terrorists inflict upon civil populations by such bombs, rapid methods to recognize the presence of nitrates are highly sought. It is also critical to detect them specifically amidst background chemical noise to avoid false negatives. The conventional instrumental methods for detection and quantification of nitrates include gas chromatography/mass spectrometry (GC/ MS),2−5 positive-ion electrospray ionization mass spectrometry (ESI-MS) with detection enabled through the addition of a crown ether,6 laser electrospray MS,7 liquid chromatography/mass spectrometry,8,9 chemical-ionization mass spectrometry,10 ion drift-CIMS,11 ion chromatography (IC), and IC/MS.12 The most widely deployed nitrate detection methods are in fact stand-alone IC13 and spectrophotometry14 (e.g., colorimetric kits from Hach Company, Loveland, CO). For aqueous samples, nitrate-specific electrode systems are also available.13 Most of the aforementioned © 2012 American Chemical Society

Received: September 16, 2012 Accepted: November 29, 2012 Published: November 29, 2012 278

dx.doi.org/10.1021/ac3026916 | Anal. Chem. 2013, 85, 278−282

Analytical Chemistry

Article

Co(NO3)2·6H2O, Ni(NO3)2·6H2O, Cu(NO3)2·xH2O, Zn(NO3)2·6H2O, RbNO3, Sr(NO3)2, Y(NO3)3·6H2O, ZrO(NO3)2·xH2O, AgNO3, CsNO3, Ba(NO3)2, La(NO3)3·xH2O, TlNO3, Pb(NO3)2, Bi(NO3)3·5H2O, Hg(NO3)2·2H2O, Ce(NO 3 ) 3 ·6H 2 O, Pr(NO 3 ) 3 ·6H 2 O, Gd(NO 3 ) 3 ·6H 2 O, Dy(NO3)3·6H2O, Th(NO3)4·5H2O, and UO2(NO3)2·6H2O. For the qualitative experiments, aliquots (3 μL of 1% aqueous solutions) were deposited onto a glass slide and dried at room temperature. The slide was placed in the open-air MS source at about 1 cm from the helium plasma region,18 using a ring stand and a clamp, and analyzed directly with or without the addition of 0.5 μL 2.5 M HClO4 or HCl. Liquid samples of biological origin (fruit juices) were taken as 3 μL drops, deposited onto glass slides and analyzed in the same way. Meat samples were taken as small pieces pinched out of the commercial products purchased. The pieces were deposited directly onto glass slides and analyzed as above. For quantification purposes, a series of ammonium nitrate solutions (10−2 to 10−10 M) were prepared and 2 μL aliquots were placed onto glass slides and dried at room temperature to give solid deposits in the range of 20 μg to 200 fg. Each slide was inserted individually into the ion source, and SIR chronograms were recorded by monitoring the m/z 62 signal for the nitrate ion. After one minute, a 0.5 μL drop of 2.5 M aqueous HCl was added to each spot and data acquisition was continued. Hydrochloric acid was preferred because of its higher volatility than that of perchloric acid. With direct insertion methods, the use of integrated chromatographic peak areas (which is the usual practice for quantification by LC-MS or loop injection procedures) was not practical because time-intensity profiles recorded from direct insertion methods are often jagged and serrated (Figure 7), even though a “chromatographic peak” may be obtained using smoothing algorithms. Instead, we summed data from all transients from the moment HCl was added, to the point the signal reached the background level. Using the Waters MassLynx software, we obtained an averaged spectrum by dividing the total intensity value (on an arbitrary scale) by the number of scans. To obtain a peak-intensity value that is representative of the sample amount, we subtracted the average signal recorded for the background from the total sample-plus-background signal. The background-subtracted signal intensities (in arbitrary units) obtained in this way for each sample were used to plot an amount versus signal-response curve (Figure 7). The curve represents

Figure 1. A HePI mass spectrum recorded from a dry deposit drop obtained from an industrial wastewater sample containing the explosive RDX and a large amount of ammonium nitrate.

we wanted to address the acute need for a high-throughput method for detecting nitrates with minimum sample preparation. Here, we report that inorganic nitrates can be efficiently detected and identified by HePI-MS. Although there have been reports on detection of organic nitrates such as pentaerythritol tetranitrate (PETN)20−22 by atmospheric-pressure ionization mass spectrometric methods, to the best of our knowledge this is the first report on determination of inorganic nitrates by a solventless API-MS method.

■

EXPERIMENTAL SECTION The Z-spray ion source of a Waters Micromass Quattro Ultima mass spectrometer was modified into a helium-plasma ionization source by passing a stream (∼30 mL min−1) of high-purity helium (99.999%, Airgas, Radnor, PA), regulated by a needle valve, through the metal capillary held at high voltage (typically, −3 kV). The source temperature was held at 100 °C. The cone voltage was set to 10 V. For CID experiments, the pressure of the argon in the collision cell was held at 2.56 × 10−5 mBar. The heater in the desolvation gas was employed only when it was necessary to heat the sample, at typical temperatures of 150−200 °C. The desolvation heater was not used when samples were acidified, except where the change in relative intensity of perchlorate peaks had to be demonstrated. In addition to nitric acid, the inorganic nitrates tested in our experiments were NaNO3, Mg(NO3)2·6H2O, Al(NO3)3·9H2O, KNO3, Ca(NO3)2·xH2O, Mn(NO3)2·xH2O, Fe(NO3)3·9H2O,

Figure 2. HePI mass spectra from several inorganic nitrates, showing that the relative intensity of the m/z 62 signal strongly depends on the compound analyzed: (a) ambient background from a clean glass slide and (b−d) spectra from solid nitrate deposits on a glass slide without any modification, (b) ZrO(NO3)2·xH2O, (c) Co(NO3)2·6H2O, (d) NaNO3, and (e−f) the samples from (a−d) with 1 μL 10% HClO4 added. 279

dx.doi.org/10.1021/ac3026916 | Anal. Chem. 2013, 85, 278−282

Analytical Chemistry

Article

samples that also contained high concentrations of ammonium nitrate, we noted that the base peak was at m/z 62 for the nitrate anion (Figure 1). The presence of a signal for nitrate was no surprise because ammonium nitrate is known to bear an appreciable vapor pressure at room temperature. However, the unexpectedly high intensity of the nitrate signal generated by the HePI source merited further investigation. We found that even solid deposits of crystalline compounds, such as reagent-grade NH4NO3 and ZrO(NO3)2·xH2O, generated intense signals for the NO3− anion when directly subjected to HePI-MS (Figure 2b). Further experiments carried out with a wide array of nitrate salts revealed that a significant signal at m/z 62 can be obtained from deposits of many transition-metal nitrates, even though the intensities were sometimes low (Figure 2c). In marked contrast, solid deposits derived from the nitrates of Na, K, Rb, Cs, Mg, Ca, Sr, Ba, Tl, Pb, Ag, and the rare-earth nitrates, failed to produce any appreciable m/z 62 signal under similar conditions (Figure 2d). We also noted that whenever a m/z 62 peak was present, it was accompanied by a minor peak at m/z 125. A product-ion spectrum recorded from the m/z 125 ion upon collision-induced dissociation (CID) proved it to be the gaseous nitrate adduct of nitric acid (HNO3·NO3−) (Figure S1 of the Supporting Information). It appeared that nitrates with crystalline water and those with higher acidity produced more intense m/z 62 signals. For example, a saturated aqueous solution of ZrO(NO3)2·xH2O, bears a pH of ∼1.7. In fact, transition metal nitrates are more acidic in solution than those of the alkali and alkali-earth nitrates. Apparently, it is not solely the volatility of inorganic nitrates that determines their detectability by HePI-MS. We envisaged that nitric acid released from the solid nitrate deposits rapidly dissociates in the ion source to produce the gaseous nitrate anions. To test this hypothesis, we first subjected a drop of dilute nitric acid to the helium-plasma ionization and observed both m/z 62 and 125 signals (Figure 3). Subsequently, we added some dilute perchloric acid onto solid nitrate deposits already placed in the source and found that m/z 62 signals could be generated immediately from any nitrate in this way. Perchloric acid releases nitric acid from the salts into the gas phase (Figure 2, panels f−h). In all cases, the nitrate signal was dramatically enhanced to afford the base peak. The results were most conspicuous for the alkali nitrates

the average values of three series of measurements, performed on three different days to reduce the environmental influence.

■

RESULTS While reviewing negative-ion spectra recorded by HePI-MS to detect trace amounts of explosives in industrial wastewater

Figure 3. A HePI mass spectrum from a drop of 1% HNO3.

Figure 4. The effect of desolvation-gas heating on the nitrate and perchlorate signals. (a) A spectrum recorded from a deposit of NaNO3 without desolvation-gas heating [the nitrate signal (m/z 62) dominates the spectrum]. (b) A spectrum recorded with desolvation-gas heating [250 °C; the relative intensities of the perchlorate signals (m/z 99, 101) are enhanced].

Figure 5. HePI mass spectra recorded from meats and vegetable juices: (a) deionized-water blank, (b) fresh Angus beef, (c) a hot dog (cured meat), (d) V8 juice, (e) tomato juice, and (f) celery juice. 280

dx.doi.org/10.1021/ac3026916 | Anal. Chem. 2013, 85, 278−282

Analytical Chemistry

Article

Figure 6. Results from an SIR experiment showing the m/z 62 signal enhancement achieved by adding a strong acid to a solid NaNO3 deposit: (a) the background m/z 62 signal is of very low intensity, and (b) it is dramatically enhanced when 10% HCl is directly applied to the deposit.

occurring nitrate. In the mass spectrum recorded from fresh Angus beef, the nitrate and nitrite peaks were conspicuous by their absence, thereby suggesting that the meat was fresh and had not undergone the curing process. In contrast, the spectrum recorded from a sample of hot dog meat showed an intense peak at m/z 62 for nitrate (Figure 5c). Next, we performed experiments to validate this nitrate-detection method for quantification purposes. First, we performed a simple experiment of repeated insertion and withdrawal of a solid zirconyl nitrate sample into and out of the HePI source. A one-channel selective-ion monitoring (SIR) experiment to detect the m/z 62 for nitrate was performed to obtain a time-intensity profile (“chronogram”) (Figure SI2 of the Supporting Information). Results showed that the m/z 62 signal intensity depends on the presence of nitrate because upon introduction of the sample to the source, the m/z 62 signal immediately intensified, and upon withdrawal of the sample, the signal rapidly decayed. In addition, a dramatic enhancement of the nitrate signal was achieved by the addition of a strong acid to the sample deposit that was already in the source (Figure 6). In another SIR experiment, a sodium nitrate deposit, which does not produce an appreciable m/z 62 signal by itself, generated an intense signal, once a drop of hydrochloric acid was added (Figure 6b). Upon withdrawal of the acidified sample from the source, the signal promptly decayed back to the baseline level. After establishing the specific dependence of the m/z 62 signal on the presence of nitrate, we constructed a nitrate “calibration curve” for different amounts of nitrate (Figure 7). The results showed linearity over 3 orders of magnitude (from 20 μg to 20 ng). A similar series of measurements performed in the same way, using labeled nitrate in the form of Na15NO3, demonstrated that the lowest limit of detection of nitrates is as small as 200 fg per deposit. The results suggest that spiking a sample with a known quantity of Na15NO3 should provide a rapid method for determining nitrates in complex matrices such as food preparations, in a high-throughput manner.

Figure 7. A semilogarithmic plot depicting the relationship between the amount of solid ammonium nitrate deposited on a glass slide versus detector response (N = 3). The SIR signal of nitrate (m/z 62) was monitored, and the average intensity per scan was estimated by summing the mass spectral signal intensities of all scans, from the moment 2.5 M HCl was added to the deposit, to the point when the signal reached the baseline level. The m/z 62 signal responses displayed a linear relationship with the logarithmic value of deposited masses in the range between 20 ng and 20 μg (R2 = 0.997) (see inset).

(Figure 2h). On the other hand, the acidification did not appreciably enhance the signal at m/z 125, thereby rendering support to our hypothesis. It must be emphasized that our very first observation of the m/z 62 signal from ammonium nitrate was made while heating the sample on the glass slide by hot desolvation gas in the MS source. However, adding perchloric acid made heating the sample unnecessary to observe the nitrate signal. In fact, when desolvation gas was employed for heating the sample, after the addition of perchloric acid, the peaks for perchlorate isotopologs (m/z 99, 101) became dominant (Figure 4). Without desolvation-gas heating, the peaks for perchlorate were either not very significant or not observed (Figure 2, panels e−h).We, subsequently, tested the nitrate detection capabilities of HePI-MS on more complex samples of biological origin, such as fruit juices and meats. Some selected results are presented in Figure 5. The peak at m/z 62 was especially prominent for celery juice because of its high level of naturally

■

CONCLUSIONS The results presented above demonstrate that nitrates present in any solid or liquid sample can be rapidly and reliably detected by HPI-MS at ambient conditions with a high degree of sensitivity and quantified over a wide range of concentrations. Minimum sample preparation is involved, and no pretreatment or 281

dx.doi.org/10.1021/ac3026916 | Anal. Chem. 2013, 85, 278−282

Analytical Chemistry

Article

(18) Yang, Z.; Pavlov, J.; Attygalle, A. B. J. Mass Spectrom. 2012, 47, 845−852. (19) Gao, J.; Owen, B. C.; Borton, D. J.; Jin, Z.; Kenttämaa, H. I. J. Am. Soc. Mass Spectrom. 2012, 23, 816−822. (20) Zhang, Y.; Ma, X.; Zhang, S.; Yang, C.; Ouyang, Z.; Zhang, X. Analyst 2009, 134, 176−181. (21) Na, N.; Zhang, C.; Zhao, M.; Zhang, S.; Yang, C.; Fang, X.; Zhang, X. J. Mass Spectrom. 2007, 42, 1079−1085. (22) Nilles, J. M.; Connell, T. R.; Durst, H. D. Anal. Chem. 2009, 81, 6744−6749.

derivatizations are necessary. While the exact limits of detection depend on the specific matrix, generally 20 ng per solid deposit can routinely be detected. For 15N-labeled nitrate, the detection limits are in the femtogram range because the m/z 63 signal has less interference from the background levels. In other words, HePI-MS addresses the need for a high-throughput technique for determination of nitrates as solids (such as white powders at airports). Such methods are urgently sought because many military propellants (e.g., black powder) and commercial blasting agents [e.g., ammonium nitrate−fuel oil (ANFO)] use nitrates. Our overall results point to the general conclusion that HePI-MS has a strong potential to become a facile method to monitor reactions and observe reaction products in statu nascendi in the ion source.

■

ASSOCIATED CONTENT

S Supporting Information *

Additional figures as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by Technogenesis funds from Stevens Institute of Technology. REFERENCES

(1) NATO and Terrorism; Edwards, F. L.; Steinhäuser, F., Eds.; Springer: New York, 2008. (2) Pagliano, E.; Meija, J.; Sturgeon, R. E.; Mester, Z.; D’Ulivo, A. Anal. Chem. 2012, 84, 2592−2596. (3) Pagliano, E.; Onor, M.; Pitzalis, E.; Mester, Z.; Sturgeon, R. E.; D’Ulivo, A. Talanta 2011, 85, 2511−2516. (4) Tsikas, D.; Suchy, M.-T.; Mitschke, A.; Beckmann, B.; Gutzki, F.-M. Methods Mol. Biol. (N.Y., NY, U.S.) 2012, 844, 277−293. (5) Tsikas, D.; Gutzki, F.-M.; Stichtenoth, D. O. Eur. J. Clin. Pharmacol. 2006, 62, 51−59. (6) De Perre, C.; Prado, A.; McCord, B. R. Rapid Commun. Mass Spectrom. 2012, 26, 154−162. (7) Flanigan, P. M.; Brady, J. J.; Judge, E. J.; Levis, R. J. Anal. Chem. 2011, 83, 7115−7122. (8) Li, Y.; Whitaker, J. S.; McCarty, C. L. J. Chromatogr., A 2011, 1218, 476−483. (9) Mathis, J. A.; McCord, B. R. Rapid Commun. Mass Spectrom. 2005, 19, 99−104. (10) Kita, K.; et al. Journal of Atmospheric and Oceanic Technology 2006, 23, 1104−1113. (11) Zheng, J. Atmospheric Chemistry and Physics Discussions; Copernicus GmbH: Göttingen, Germany, 2008; Vol. 8, pp 4877− 4909. (12) Saccani, G.; Tanzi, E.; Cavalli, S.; Rohrer, J. J. AOAC Int. 2006, 89, 712−719. (13) Standard Methods for the Examination of Water and Wastewater, 22nd ed, American Public Health Association, American Water Works Association, Water Environment Federation, 2012. (14) EPA Method 353.1, U.S. Environmental Protection Agency: Washington, D.C., 1978. (15) Cody, R. B.; Laramée, J. A.; Durst, H. D. Anal. Chem. 2005, 77, 2297−2302. (16) Houben, E.; et al. Anal. Chem. 2010, 82, 601−607. (17) Yang, Z.; Attygalle, A. B. J. Am. Soc. Mass Spectrom. 2011, 22, 1395−1402. 282

dx.doi.org/10.1021/ac3026916 | Anal. Chem. 2013, 85, 278−282