Direct analysis of saturated hydrocarbons using glow discharge plasma ionization source for mass spectrometry

Talanta 204 (2019) 310–319

Contents lists available at ScienceDirect

Talanta journal homepage: www.elsevier.com/locate/talanta

Direct analysis of saturated hydrocarbons using glow discharge plasma ionization source for mass spectrometry

T

Yoko Nunomea,∗, Kenji Kodamab, Yasuaki Uekic, Ryo Yoshiied, Ichiro Narusec, Kazuaki Wagatsumae a

Graduate School of Integrated Arts and Sciences, Hiroshima University, 1-7-1 Kagamiyama, Higashi-Hiroshima, 739-8521, Japan X-ray Instrument Division, Rigaku Corporation, 14-8 Akaoji, Takatsuki, Osaka, 569-1146, Japan c Institute of Materials and Systems for Sustainability, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, 464-8603, Japan d Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, 464-8603, Japan e Institute for Materials Research, Tohoku University, Aoba, Katahira 2-1-1, Sendai, 980-8577, Japan b

ARTICLE INFO

ABSTRACT

Keywords: Glow discharge plasma Soft ionization Saturated hydrocarbons Mass spectrometry Ambient air

The ionization source based on glow discharge plasma using ambient air is driven by a pulsed direct-current voltage for soft plasma ionization (SPI). The novelty of this work is that molecular ions [ M+ 13]+ related to the analyte species (M), which may be formed by numerous oxidation, can be dominantly detected as a base peak with little or no fragmentation of them in an air plasma at a pressure of several kPa. The unique ion [ M+ 13]+ was assigned to the oxidation product, [ M+ O 3H]+ , which was confirmed as a deuterated ion [ M+ O 3D]+ ([ M+ 10]+ ) by using a deuterated solvent. The ionization reactions were suggested that the product ion [ M+ O 3H]+ may arise from hydride abstraction reaction of M with O+2 •, dehydrogenation reaction of [ M H]+ • and subsequently oxidation reaction of [ M 3H]+ with O3. n-Alkane mixtures was also measured to evaluate the intermolecular interaction in this system. The limits of detection (LOD) were in the range of 0.126–1.68 ppmv and the relative standard deviation (RSD) for repeatability was approximately 10.0% at the lowest concentration. To our knowledge, this is the first report demonstrating that the spectrum pattern of saturated hydrocarbons could be directly determined without any complicated fragmentation.

1. Introduction Saturated hydrocarbons (alkanes) are difficult to be analyzed by mass spectrometry because of having distinct chemical features of unionizable functional group, low basicity, which is responsible for occurring fragmentation of molecular ions [1,2]. Ionization methods for analysis of alkanes are commonly electron ionization (EI) [3–7], photoionization (PI) [8,9], flame ionization detector (FID) [10–12] and field ionization (FI)/field desorption (FD) [13,14], which are combined with gas chromatography. These methods are time-consuming and tedious for rapid and easy in-situ analysis. Marotta et al. [15] used atmospheric-pressure chemical ionization (APCI) mass spectrometry (MS) to detect several linear, branched and cyclic alkanes produced by corona discharges in VOC-contaminated air. The samples vapors stripped from a small reservoir were introduced into the APCI source, which was kept at near atmospheric pressures. They suggested that the major ion formed from most alkanes (M) was the species [ M H]+, which arose from reactions of M with air plasma

∗

ions, notably O+2 •, NO+, H3 O+, and their hydrates. Usmanov et al. [16] developed a hollow cathode discharge (HCD) ion source for monitoring of alkanes (n-hexane, cyclohexane, n-heptane, n-octane, i-octane) in 28-Torr air plasma. They proposed that the [ M H]+ generated by hydrogen atom abstraction reaction with oxygen molecule, subsequently reacts with ozone molecule, resulting in the formation of [ M+ O H]+ and [ M+ O 3H]+. Sekimoto et al. [17] studied direct analysis in real time (DART) -mass spectrometry associated with corona discharge using He gas (corona-DART). Corona-DART has an ability to detect ions related to alkanes (n-tridecane, n-pentadecane, n-heptadecane), i.e., [ M+ O 3H]+ (M + 13) and/or the analogous monohydrates [ M+ 2 O H]+ (M + 31) which are primarily formed in the plasma jet with a relatively high stability. They suggested that the formation mechanism of the ionic species might occur via hydride abstraction and oxidation. Direct analysis of n-alkanes (C5 – C12) was performed using a laboratory-made microwave-induced plasma ionization (MIPI) source in

Corresponding author. E-mail address: [email protected] (Y. Nunome).

https://doi.org/10.1016/j.talanta.2019.05.115 Received 27 March 2019; Received in revised form 30 May 2019; Accepted 31 May 2019 Available online 04 June 2019 0039-9140/ © 2019 Elsevier B.V. All rights reserved.

Talanta 204 (2019) 310–319

Y. Nunome, et al.

combination with an ion trap mass spectrometer by Li et al. [18]. They observed the peaks for [ M+ O 3H]+ (M + 13) and adducts with a mass shifted higher by 18 u, [ M+ 2 O H]+ (M + 31) in the positive mode with Ar as a plasma working gas. To examine the ionization mechanism of the alkanes, they performed a GC/MS analysis on npentane vapor after reaction with the MIP. A reaction scheme of npentane was proposed such that hydroxyl radical OH• and ground state oxygen O (3P) generated by MIPI would react with n-pentane, leading to generation of hydrogen and cyclopentadiene mainly through H-atom abstraction. In addition, cyclopentadiene reacted with water cluster ions to form protonated cyclopentadiene C5 H+7 (mass-to-charge ratio (m/z) 67, [ M 5]+), adduct ions C5 H6 . H3 O+ (m/z 85, [ M 5 + 18]+) and C5 H6 . (H2 O)2H+ (m/z 103, [ M 5 + 2 × 18]+ ), which were confirmed in MS/MS spectrum. We have successfully developed a soft plasma ionization (SPI) -MS with a glow discharge ionization source driven by a pulsed direct-current (dc) voltage in order to suppress any fragmentation of organic compounds [19]. The SPI source possesses Cu coaxial-hollow electrodes where glow discharge plasma can generate several kinds of reactive ions by using ambient air as a discharge gas. The unique structure of the ion source would contribute to little fragmentation of target molecule. A major ionization reaction of the SPI source is considered that aromatic compounds could be readily ionized by an attachment of NO+ due to high-pressure air plasma [19]. In this study, we attempted to directly detect and determine saturated hydrocarbons with a glow discharge plasma ionization source driven by pulsed direct-current (dc) voltage in order to suppress any fragmentation. To evaluate the basic characteristics of the pulsed dc SPI source, detectable signals for alkanes were examined in an ambient air discharge plasma when varying the discharge parameters, such as discharge current, discharge pressure, sample/flow ratio, frequency and duty ratio. The discharge plasma parameters could be optimized by evaluating the peak intensity of the mass signal for quantitative analysis of alkanes under various experimental conditions. The novelty of this study is that the dominant [ M+ O 3H]+ ion derived from alkanes could be directly detected as a base peak with little or no fragmentation at higher pressures of ambient air.

which are completely isolated by a PTFE insulator. A glow discharge plasma, whose arrangement is a coaxial-hollow, is generated between the electrodes. The flow rate of carrier gas and sample gas was regulated with mass flow controllers (MFC), respectively (See section 2.3). The discharge pressure of 1.0–3.0 kPa in the SPI source was adjusted by monitoring with a capacitance manometer (CCMT-100D, ULVAC, Inc. Components Division, Japan). The dc pulse power supply system includes a highspeed pulse switching circuit (FHVS-2000, Chubu R&D Co., Ltd., Japan), a dc power supply (HVR-2K150P/FG/100, Chubu R&D Co., Ltd., Japan) and a function generator (Model DG535, Stanford research systems, Inc., USA). The q-MS was operated in a full-scan mode or a single-ion recording (SIR) mode. The voltage of a cone, an extractor and a radio frequency (RF) lens (a hexapole ion guide) was set to 3, 0 and 0.1 V, respectively. In the analyzer section, a LM (low mass) and a HM (high mass) resolution were set to be 15.0 and 13.8 (arbitrary units), respectively, and an ion energy of 0.5 V was selected. The photomultiplier tube detector (PMT) was adjusted to be 650 V. The full-scan mass spectrum was acquired over a mass scan range of m/z 2 – 202 at a scan time of 0.2 s (C5 – C10), or m/z 2 – 402 at a scan time of 0.4 s (C11 – C14) and an interscan delay of 0.1 s. Quantitative analysis was performed by monitoring an m/z for target analytes in the SIR mode. The selected ions were m/z 113, 127, 141 and 155 for n-heptane, n-octane, n-nonane and n-decane at a dwell time of 1.5 s and an inter-channel delay of 0.1 s, respectively. The obtained MS spectra were time-integrated over 1.0 min. Each mass spectrum was corrected by subtracting a blank spectrum of ambient air. All data were acquired in a continum mode and processed with MassLynx software (version 3.5, Waters Corp., USA). Typical operating conditions for the SPI-MS system are listed in Table 1. 2.2. Chemicals

2. Experimental

n-Pentane, n-heptane, n-octane, n-nonane, n-decane, n-undecane, ndodecane, n-tridecane, n-tetradecane, 2-methylheptane, 2,4-dimethylhexane and 2,2,4-trimethylpentane were purchased from Wako Pure Chemicals Co., Inc. (Japan). Deuterated chemical, n-heptane-d16 was also obtained from Wako Pure Chemicals Co., Inc. (Japan). nHexane was purchased from Kanto Chemical Co., Inc. (Japan).

2.1. Dc-pulsed soft plasma ionization-quadruple mass spectrometer

2.3. Sample introduction

The experimental apparatus of SPI-MS has been described in detail previously [19]; thus, it will be briefly presented here. Fig. 1 shows a schematic diagram of the SPI source combined with a quadruple mass spectrometer (q-MS) (ZQ-2000, Waters, USA). The SPI source consists of a Cu-hollow anode and a co-axially-arranged inner Cu-mesh cathode,

Ambient air (carrier gas) was introduced into the SPI source at a pre-set pressure, which was monitored with a capacitance manometer while the flow rate was controlled by an MFC (Model 3660-1/4SWN2 -1SLM, Kofloc, Japan). Ambient air was normally used as a discharge gas. A mixture of sample vapor with ambient air was introduced into

Fig. 1. Schematic diagram of the experimental system for measuring saturated hydrocarbons. 311

Talanta 204 (2019) 310–319

Y. Nunome, et al.

Table 1 Soft plasma ionization source-mass spectrometer operating conditions.

Table 2 SIR mode conditions for SPI-MS analysis.

SPI source part Plasma gas Sample

Pulse discharge current (mA) Gas pressure (kPa) Sample/air flow ratio: R (%) Frequency (kHz) Duty ratio (%) q-MS analyzer part Analytical mode Mass range (m/z) Scan time (s) Inter-scan delay (s) Analytical mode Selected mass (m/z) Dwell time (s) Inter-channel delay (s)

Ambient air n-Alkanes (n-pentane, n-hexane, n-heptane, n-heptane-d16 , n-octane, n-nonane, n-decane, n-undecane, n-dodecane, n-tridecane, n-tetradecane) Blanched alkanes (2,4-dimethylhexane, 2-methylheptane, 2,2,4-trimethylpentane) 10 – 70 1.0–3.0 0.4–20 1.0–5.0 5 – 50

Compound

MW

m/z used for quantification

Ion species

n-heptane

100

113

[ M+ O

3H]+

[ M+ O

3H]+

[ M+ O

3H]+

[ M+ O

3H]+

n-octane

n-nonane n-decane

114 128 142

127 141 155

3. Results and discussion 3.1. n-Alkanes (C5 – C7) Owing to larger ionization energy of saturated hydrocarbons, the molecule ion peaks are very low in intensity in the mass spectra. Typical electron ionization (EI) mass spectra of saturated hydrocarbons generally comprise mass peaks of major characteristic fragment ions, including m/z 43, 57, 71 and 85, with each 14 u mass interval which is generated by sequential dissociation of the methylene group of linear saturated alkane. Fig. 2 shows soft plasma ionization (SPI) mass spectra of C5 – C7 nalkanes and deuterated n-heptane: (a) n-pentane, (b) n-hexane, (c) nheptane and (d) n-heptane-d16 . Each mass spectrum was corrected by subtracting a blank of air discharge. The [ M+ NO]+ peak was mainly detected at m/z 102 as a base peak for n-pentane (MW = 72), while not only [ M+ NO]+ peak but also [ M+ 13]+ peak was detected at m/z 116 and 99, respectively, for n-hexane (Fig. 2 (a) and (b)). The NO+ species can be ionized by EI, or by charge transfer ionization and Penning-type ionization with excited N2 molecules [19]. The generated NO+ species would attach with M in high-pressure air discharge, resulting in the formation of [ M+ NO]+ adduct ions by the attachment of NO+ in Eq. (1) [19]. These [ M+ NO]+ ions would be produced and stabilized via three-body association reactions [19].

Scan 2 – 202, 2 – 402 0.2 (2 – 202), 0.4 (2 – 402) 0.1 SIR 113, 127, 141, 155 1.5 0.1

the SPI source before a chamber pressure reached the preset value by controlling an MFC (Model 3660-1/4SW-N2 -100SCCM, Kofloc, Japan). The vapor of an organic solvent sample in a vial was then introduced into the SPI source by switching from a blank gas line to a sample gas line with a 3-way valve. The sample vapor was diluted with ambient air (carrier gas) to a predetermined mixed ratio, which was adjusted by their flow rates (sample/air flow rate, R), and then was introduced directly through the air stream into the SPI source. The pulsed dc power was applied to the SPI source to ignite the plasma while a preset gas pressure was maintained.

NO+ + M

[M + NO]+

(1)

where M is n-alkanes (C5 and C6). As shown in Fig. 2 (c), only the [ M+ 13]+ peak at m/z 113 was detected as a base peak with little or no fragmentation for n-heptane. As the liner chain of alkanes is longer, NO+ attachment to a sample molecule would be less likely to occur [15]. The polarization of molecule would decrease by lengthening the molecular chain; therefore, the NO+ affinity would become lower. It was thus considered that the [ M+ 13]+ ion for n-hexane was detected as another base peak in the mass spectra, differing from the case of npentane having shorter carbon chain. The unique [ M+ 13]+ ion is well known as ion species of alkanes, which is often generated under air plasma discharges. n-Alkanes do not form protonated molecules because of the low proton affinities but can form any oxygenated product ions [20,21]. Therefore, the [ M+ 13]+ ion detected in the mass spectrum of n-hexane was identified to be [ M+ O 3H]+. In our previous paper [19], the mass spectrum of n-hexane provided a [ M+ NO]+ base peak (m/z 116) with a weaker [ M+ O 3H]+ peak (m/z 99) at the sample/air flow ratio (R) of 52.0%. As the R value is higher, the amount of carrier gas (surround air) relatively becoming lower, so that oxidation reaction to M is less likely to proceed. Thus, NO+ attachment [ M+ NO]+ ion rather than oxygenerated [ M+ O 3H]+ ion would produce preferentially when the R is 52.0%. To examine whether the hydrogen atom abstraction reaction occurs in a series of n-alkanes, deuterated n-heptane-d16 (MW = 116) was examined (Fig. 2 (d)) [22–24]. As a result, the peak at m/z 126 was detected as a base peak corresponding to [ M+ 10]+, which was identified as [ M+ O 3D]+. This can be proven that the hydrogen atom abstraction reaction would generally occur on the ionization of n-alkanes in air discharges, resulting in the formation of [ M+ O 3H]+ species.

2.4. Standard preparation Standard gas mixtures (n-heptane, n-octane, n-nonane and n-decane) were used to draw each standard curve. Four gas mixtures were prepared by injecting 1 – 4 μL of each liquid solvent, using a 10-μL syringe, into a 1-L Tedlar bag (CEK-1, GL Science, Japan) which had been filled with 1 L of ambient air with a 200 mL plastics disposal syringe. Then, these gas mixtures were rapidly vaporized and allowed to equilibrate for 5 min. The gas mixtures were subsequently diluted by injecting their aliquots into Tedlar bags through a 2-way stopcock using a plastics syringe. The total gas volume was adjusted to 1 L using surrounding air. The Tedlar bag was then connected to the sample line connected with the SPI ionization source. The standard gas mixtures were measured on three replications: n-heptane concentrations of 0.204, 2.04, 4.08, 8.15, 16.3 ppmv; n-octane concentrations of 0.259, 2.59, 5.18, 10.4, 20.7 ppmv; n-nonane concentrations of 0.404, 4.04, 8.08, 16.2, 32.3 ppmv; n-decane concentrations of 0.616, 6.16, 12.3, 24.7, 49.3 ppmv. Before the use, the Tedlar bags were flushed with high purity nitrogen (G1 grade, 99.99995%, Japan Fine Products Co., Japan) at least three times and the residual nitrogen gas was evacuated by the plastics syringe. To estimate reliability, a blank sample was also measured in the same method as the standard samples. Data acquisition was performed in the SIR mode for the mass signals of m/z 113, 127, 141 and 155 to determine the concentration of analyte compounds (Table 2).

312

Talanta 204 (2019) 310–319

Y. Nunome, et al.

Fig. 2. Soft plasma ionization (SPI) mass spectra of n-alkanes: (a) n-pentane, (b) n-hexane, (c) n-heptane and (d) n-heptane-d16. Discharge parameters were set to the following values: discharge air pressure: 2.5 kPa; discharge current: 30 mA; sample/air flow ratio (R): 4%; modulation frequency: 2.5 kHz; duty ratio: 50%.

3.2. Dependence of current

analyte would mainly pass through the inside of the meth cathode with little coming into contact with the plasma in between the electrodes (Fig. 1) [19]. The decrease in the [ M+ O 3H]+ with high discharge currents may be due to a change in the number densities of reagent ions in the source [25]. The remaining ions, including [M]+ ., NO+ , O+2 and O+3 , monotonically increased with increasing the discharge current. Especially, as O+3 rapidly increased from 35 mA, the ionization reaction of O3 could occur more actively by collision with increased electrons in the SPI source. Therefore, the decreased number density of O3 could be less likely to react with analyte molecules and uncontribute to the formation of the [ M+ O 3H]+.

Fig. 3 shows the current dependance of the SPI source in a pulse discharge plasma on the intensity of [ M+ O 3H]+, [M]+ • (n-heptane), NO+ , O+2 and O+3 at an ion source pressure of 2.5 kPa. Only the mass spectral intensity of [ M+ O 3H]+ is scaled on the right vertical axis. The discharge plasma was not stably maintained below a current of 10 mA. The intensity of the [ M+ O 3H]+ gradually increased and then turned to decrease when the discharge current became larger than 35 mA. In the region of discharge current, the [ M+ O 3H]+ was detected as a base peak with little or no fragmentation. This is because the

Fig. 3. Variations in the intensities of [ M+ O 3H]+ , [M]+ . , NO+, O+2 and O+3 mass peaks as a function of the discharge current. Discharge parameters were set to the following values: discharge air pressure: 2.5 kPa; sample/air flow ratio (R): 4%; modulation frequency: 2.5 kHz; duty ratio: 50%; sample: n-heptane. 313

Talanta 204 (2019) 310–319

Y. Nunome, et al.

species of N2 (N2* ) [19,28,29]. 3.5. Branched alkanes (C8) We examined whether any difference appeared between liner alkanes and branched alkanes on C8 hydrocarbons in air plasma. SPI mass spectra of branched alkanes with eight carbons (MW = 114) were observed, where the following isomers of n-octane: (a) 2-methylheptane, (b) 2,4-dimethylhexane and (c) 2,2,4-trimethylpentane (i-octane), were employed (Fig. 6). Each branched alkane of (a), (b) and (c) contains one-, two- and three-methyl substituent, respectively. Therefore, the sterically bulky compounds are in the ascending order: (a) < (b) < (c). The peaks at m/z 99, 113 and 127 were identified to be [ M CH3]+, [ M H]+ and [ M+ O 3H]+, respectively. The [ M+ O 3H]+ was detected as a base peak in Fig. 6 (a) and (b), but the corresponding peak was not found in Fig. 6 (c). This is likely due to the regioselectivity of oxidation [27], which means that the high steric hindrance of branched alkanes can make the surface area to be smaller compared with liner alkanes; therefore, it is difficult to occur their oxidation process in the sample molecule. Therefore, as for 2,2,4-trimethylpentane with the most bulky compound, the [ M H]+ could be more likely to be produced than the [ M+ O 3H]+ without passing through the oxidations. The main product ions [ M H]+ may be generated via a hydride abstraction reaction of O+2 • (Eq. (3)) [16,20,30], which could be easy to occur regioselectively at a tertiary carbon [27,31]:

Fig. 4. Variations in the intensities of [ M+ O 3H]+ , [M]+ ., NO+ , O+2 and O+3 mass peaks as a function of the discharge pressure. Discharge parameters were set to the following values: discharge current: 30 mA; sample/air flow ratio (R): 4%; modulation frequency: 2.5 kHz; duty ratio: 50%; sample: n-heptane.

3.3. Dependence of pressure Fig. 4 shows air pressure dependence of the [ M+ O 3H]+, [M]+ • (nheptane), NO+ , O+2 and O+3 in a range from 1.0 to 3.0 kPa at a discharge current of 30 mA. Only the mass spectral intensity of [ M+ O 3H]+ is scaled on the right vertical axis. The intensity of the [ M+ O 3H]+ drastically increased with increasing the air discharge pressure. The intensity of NO+ and O+2 tend to decrease with increasing the discharge pressure, whereas the intensity of [M]+ . and O+3 slightly increased beyond a discharge pressure of 2.0 kPa. The SPI source is filled with larger amounts of sample molecule and surround air at higher discharge pressures when the R is kept constant, so that the number density of these air-origin species should be also elevated. The increment of the discharge pressure would reduce the mean free path of constituent particles in the plasma; therefore, the co-existing species could readily collide with sample molecules, thus finally producing the ionized adducts, [ M+ O 3H]+ (described later). The higher pressure condition makes it possible to improve the detection sensitivity (signal to background ratio: S/B) of a sample molecule [19,26].

M + O2+ •

(3)

3.6. Mixture of n-alkanes It was confirmed whether there was any intermolecular interaction between sample molecules in air discharge when n-alkane mixtures were measured. A mixture sample comprising four compounds of nheptane, n-octane, n-nonane and n-decane were prepared with a ratio of approximately 1 : 1: 1 : 1 (v/v). Fig. 7 shows the SPI mass spectrum of the mixed n-alkanes. The peaks at m/z 113, 127, 141 and 155 corresponding to the [ M+ O 3H]+ ions were dominantly detected as the corresponding base peaks with little or no fragmentation, indicating little intermolecular interaction among these n-alkanes in air discharge plasma. That means that each n-alkane can be detected at the same time even if several kinds of alkanes are contained in a sample. It is possible to make simultaneous detection of multiple n-alkanes, which could perform real-time analysis for the practical use. The relative intensities of these peaks seem to be more sensitive in more volatile compounds.

Fig. 5 shows SPI mass spectra of C8 – C14 hydrocarbons: (a) n-octane, (b) n-nonane, (c) n-decane, (d) n-dodecane, (e) n-undecane, (f) ntridecane and (g) n-tetradecane. Fig. 5 (a) – (c) were obtained under the same experimental conditions as Fig. 2, while Fig. 5 (d) – (g) were obtained by changing the R value to 20% due to their low volatility. In four mass spectra of n-octane (MW = 114), n-nonane (MW =128), ndecane (MW = 142) and n-dodecane (MW = 156), each dominant [ M+ O 3H]+ ion was observed as a base peak with little or no fragmentation which appeared at m/z 127, 141, 155 and 169, as similar to the case of n-heptane. Mass spectra shown in Fig. 5 (e) – (g) commonly resulted from not only [ M+ O 3H]+ as a base peak but also [ M+ 2 O 3H]+, [ M+ 2 O H]+ and [ M+ 3O]+ as the co-existing ions. These ions with more oxidized specimens were only observed in the hydrocarbons having more than 12 carbons, implying that the longer liner alkanes influence the oxidative reactivity. These alkanes have a higher molecular size and a higher surface area according to the molecular weight. The obtained results show that the longer linear alkanes could be easily oxidized because the oxidation generally prefers to occur at less steric hindrance and electron-rich sites [27].

[M + 3O]+ •

H ]+ + HO2•

In Fig. 6 (b), 2,4-dimethylhexane may eliminate a methyl group because two tertiary carbons in the molecule are unstably configured, resulting in the formation of [ M CH3]+ via oxidative cracking [27]. These results from the branched alkanes would be roughly consistent with findings observed by X. Xie et al. (2017). Mass spectrometry lacks the ability to perform structural analysis and distinguish isomers, measuring only mass-to-charge ratio of ions. However, this result indicates that the SPI-MS method can makes it possible to distinguish isomers without using MS/MS for the selectivity detection.

3.4. n-Alkanes (C8 – C14)

M+ • + O3

[M

3.7. Performance characteristics The intensities of the base peak for four target alkanes were measured three times under the optimal following conditions for n-heptane: 3.0 kPa, 35 mA, R of 1.6%, 3.0 kHz, duty ratio of 50%. Fig. 8 shows the relationship between the intensity at m/z 113, 127, 141 and 155 in arbitrary units versus concentrations of (a) n-heptane, (b) n-octane, (c) n-nonane and (d) n-decane standard, respectively. The vertical error bars represent the ±1 standard deviation (SD) (where the SD is a standard deviation of a blank sample), which was obtained on the basis of the mean value for the blank sample on three replications The linearity, precision and detection limits for four n-alkanes (C7 – C10)

(2)

where M+ • would be generated by Penning-type ionization with excited 314

Talanta 204 (2019) 310–319

Y. Nunome, et al.

Fig. 5. Soft plasma ionization (SPI) mass spectra of longer-chain n-alkanes: (a) n-octane, (b) n-nonane, (c) n-decane, (d) n-dodecane, (e) n-undecane, (f) n-tridecane and (g) n-tetradecane. Discharge parameters were set to the following values: discharge air pressure: 2.5 kPa; discharge current: 30 mA; sample/air flow ratio (R): (a)–(c) 4%, (d)–(g) 20%; modulation frequency: 2.5 kHz; duty ratio: 50%.

315

Talanta 204 (2019) 310–319

Y. Nunome, et al.

Fig. 6. Soft plasma ionization (SPI) mass spectra of branched octanes: (a) 2-methylheptane, (b) 2,4-dimethylhexane and (c) 2,2,4-trimethylpentane (i-octane). The experimental conditions are the same as in Fig. 2.

ppmv. 3.8. Ionization reactions Ozone, which is a remarkable oxidant, can oxidize organic compounds under ordinary pressure and low temperature conditions without using any catalyst. Any oxidation reaction to analyte molecules by O3 , which would be produced and activated in air plasma [16,33], may proceed in the SPI source. In the nitrogen discharge plasma, the emission bands of NO γsystem and OH radical were observed as shown in Fig. S3 (Appendix), while the related molecular ion, [ M+ O 3H]+, was never detected under the same experimental conditions (MS data not shown). It is thus found that the oxidation reaction by O3 would play an important role in the ionization of the analyte molecules. Furthermore, the NO+ and the hydroxyl radical OH• species, which would be present in the nitrogen glow discharge plasma, were found not to act as a hydride abstraction and a hydrogen atom abstraction reagent, respectively. Actually, the NO+ would be responsible for adduct reactions of the short liner alkanes, such as n-pentane and n-hexane, but it did not cause hydride abstraction reactions simultaneously. Therefore, after the hydride abstraction with O+2 • in Eq. (3), a subsequent reaction of alkanes may be caused mainly by a dehydrogenation process [16,31], as denoted in Eq. (4). The [ M 3H]+ species can further react with O3 , then forming a deprotonated dehydrogenated intermediate [ M+ O H 2H]+ [16]. The [ M+ O 3H]+ species are finally produced by a reaction of Eq. (5).

Fig. 7. Soft plasma ionization (SPI) mass spectra of mixed n-alkanes (C7 – C10). The experimental conditions are the same as in Fig. 2.

determination were shown in Table S1 (Appendix). Standard curves were constructed in a concentration range of 0.204–49.3 ppmv. The coeffcient of determination (R2 ) for each compound was higher than 0.973 (except for n-decane), indicating good linearity, over a range from sub ppmv to hundreds of ppmv with a dynamic range of 10 2 . The relative standard deviation (RSD) was approximately 10.0% at the lowest concentration of within the linear range of three replicate analysis. A limit of detection (LOD) and a limit of qualification (LOQ) were defined as 3.3 SD/a and 10 SD/a (where a is the slope of a standard curve), respectively [32]. The LOD values were obtained in a range of 0.126–1.68 ppmv. The LOQ values were obtained to be lower than 5.10

[M

H ]+

[M

3H ]+

[M + O3

3H ]+ + 2H+ + 2e [M + O

3H ]+

+ O2

(4) (5)

On the basis of our experimental results, successive reactions of Fig. 9 are proposed to explain the formation of [ M+ O 3H]+ for 316

Talanta 204 (2019) 310–319

Y. Nunome, et al.

Fig. 8. Standard curves for (a) n-heptane, (b) n-octane, (c) n-nonane and (d) n-decane. The vertical error bars present ± 1 SD for the mean value at each concentration (n = 3). Discharge parameters were set to the optimal following values: discharge air pressure: 3.0 kPa; discharge current: 35 mA; sample/air flow ratio (R): 1.6%; modulation frequency: 3.0 kHz; duty ratio: 50%.

Fig. 9. Reaction scheme of alkanes using the soft plasma ionization source in air plasma. 317

Talanta 204 (2019) 310–319

Y. Nunome, et al.

alkanes in the SPI source. The [ M H]+ species (1) are formed in a hydride abstraction from M by O+2 • [16,20,30], which could occur regioselectively at a secondary or tertiary carbon [16,27,31]. Furthermore, H2 elimination from the carbocation, [ M H]+ (1), could be probably caused [16,31], so that deprotonated dehydrogenated species ([ M 3H]+) (2), whose allyl cations are resonance-stabilized, would be subsequently generated [16,34]. At the last stage, 1,3-dipolar insertion of O3 would occur regioselectivety at a secondary or tertiary carbon through trioxides (3) as intermediates [16,33,35], resulting in a protonated ketone ([ M+ O 3H]+) (4) [16,33].

[7] [8] [9]

[10]

4. Conclusions

[11]

Direct mass analysis of saturated hydrocarbons including n-alkanes (C5 – C14) and branched alkanes (C8), in which their simple spectrum patterns can be obtained with little or no fragmentation, has successfully conducted by the pulsed dc SPI-MS. Mass spectra of the saturated hydrocarbons, except for n-pentane and 2,2,4-trimethylpentane, comprised their molecular ion peaks of [ M+ O 3H]+ which was clearly identified by using a deuterated solvent. Ionization reactions of the saturated hydrocarbons would depend on the length of the linear carbon chain and sterically bulky on the steric confirmation. From the data presented in this paper, a major ionization reaction of saturated hydrocarbons is suggested as follows: the [ M+ O 3H]+ product ions arise from a hydride abstraction reaction of M with O+2 •, a dehydrogenation reaction of [ M H]+ and subsequently an oxidation reaction of [ M 3H]+ with O3. The mass spectra of n-alkanes mixture provided one peak for each component with little or no fragmentation, which would be capable of performing the simultaneous analysis for the practical use. The LOD values for n-heptane, n-octane, n-nonane and ndecane obtained by this method were in the range between 0.126 and 1.68 ppmv. The RSD for the repeatability triplicate measurements ranged from 9.8 to 10.9% (average; 10.0%) for the lowest concentration.

[12] [13] [14]

[15] [16]

[17]

[18]

[19]

Acknowledgments This work was supported in part by Shimadzu Science Foundation, Japan, in part by the joint research program of the Institute of Materials and Systems for Sustainability, Nagoya University, Japan , and in part by the Inter-University Cooperative Research Program of the Institute for Materials Research, Tohoku University, Japan (Proposal No. 18K0002).

[20] [21]

[22]

Appendix A. Supplementary data [23]

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.talanta.2019.05.115.

[24]

References [25]

[1] C. Wu, K. Qian, M. Nefliu, R.G. Cooks, Ambient analysis of saturated hydrocarbons using discharge-induced oxidation in desorption electrospray ionization, J. Am. Soc. Mass Spectrom. 21 (2) (2010) 261–267, https://doi.org/10.1016/j.jasms.2009. 10.006. [2] T. Zhang, Z.-Y. Li, M.-Q. Zhang, S.-G. He, Gas-phase reactions of atomic gold cations with linear alkanes (C2–C9), J. Phys. Chem. A 120 (25) (2016) 4285–4293, https:// doi.org/10.1021/acs.jpca.6b03836. [3] S.B. Hawthorne, D.J. Miller, J.J. Langenfeld, M.S. Krieger, PM-10 high-volume collection and quantitation of semi- and nonvolatile phenols, methoxylated phenols, alkanes, and polycyclic aromatic hydrocarbons from winter urban air and their relationship to wood smoke emissions, Environ. Sci. Technol. 26 (11) (1992) 2251–2262, https://doi.org/10.1021/es00035a026. [4] S. Dagan, A. Amirav, Electron impact mass spectrometry of alkanes in supersonic molecular beams, J. Am. Soc. Mass Spectrom. 6 (2) (1995) 120–131, https://doi. org/10.1016/S1044-0305(94)00095-H. [5] W.-C. Lai, C. Song, Temperature-programmed retention indices for g.c. and g.c.-m.s. analysis of coal- and petroleum-derived liquid fuels, Fuel 74 (10) (1995) 1436–1451, https://doi.org/10.1016/0016-2361(95)00108-H. [6] F. Troya, M. Lerma-García, J. Herrero-Martínez, E. Simó-Alfonso, Classification of

[26]

[27]

[28]

[29]

318

vegetable oils according to their botanical origin using n-alkane profiles established by GC–MS, Food Chem. 167 (2015) 36–39, https://doi.org/10.1016/j.foodchem. 2014.06.116. M.S. Alam, C. Stark, R.M. Harrison, Using variable ionization energy time-of-flight mass spectrometry with comprehensive GC×GC to identify isomeric species, Anal. Chem. 88 (8) (2016) 4211–4220, https://doi.org/10.1021/acs.analchem.5b03122. L. Hanley, R. Zimmermann, Light and molecular ions: the emergence of vacuum UV single-photon ionization in MS, Anal. Chem. 81 (11) (2009) 4174–4182, https:// doi.org/10.1021/ac8013675. A. Giri, M. Coutriade, A. Racaud, K. Okuda, J. Dane, R.B. Cody, J.-F. Focant, Molecular characterization of volatiles and petrochemical base oils by photo-ionization GC×GC-TOF-MS, Anal. Chem. 89 (10) (2017) 5395–5403, https://doi.org/ 10.1021/acs.analchem.7b00124. U. Schlink, M. Rehwagen, M. Damm, M. Richter, M. Borte, O. Herbarth, Seasonal cycle of indoor-VOCs: comparison of apartments and cities, Atmos. Environ. 38 (8) (2004) 1181–1190, https://doi.org/10.1016/j.atmosenv.2003.11.003. J.K. Volkman, D.G. Holdsworth, G.P. Neill, H. Bavor, Identification of natural, anthropogenic and petroleum hydrocarbons in aquatic sediments, Sci. Total Environ. 112 (2) (1992) 203–219, https://doi.org/10.1016/0048-9697(92) 90188-X. T. Schmitz, D. Hassel, F.-J. Weber, Determination of VOC-components in the exhaust of gasoline and diesel passenger cars, Atmos. Environ. 34 (27) (2000) 4639–4647, https://doi.org/10.1016/S1352-2310(00)00303-4. L. Hejazi, M. Guilhaus, D.B. Hibbert, D. Ebrahimi, Gas chromatography with parallel hard and soft ionization mass spectrometry, Rapid Commun. Mass Spectrom. 29 (1) (2014) 91–99, https://doi.org/10.1002/rcm.7091. W. Genuit, H. Chaabani, Comprehensive two-dimensional gas chromatography-field ionization time-of-flight mass spectrometry (GCxGC-FI-TOFMS) for detailed hydrocarbon middle distillate analysis, Int. J. Mass Spectrom. 413 (2017) 27–32, https://doi.org/10.1016/j.ijms.2016.12.001 sI: Nico Nibbering Issue. E. Marotta, C. Paradisi, A mass spectrometry study of alkanes in air plasma at atmospheric pressure, J. Am. Soc. Mass Spectrom. 20 (4) (2009) 697–707, https:// doi.org/10.1016/j.jasms.2008.12.005. D.T. Usmanov, L.C. Chen, K. Hiraoka, H. Wada, H. Nonami, S. Yamabe, Mass spectrometric monitoring of oxidation of aliphatic C6–C8 hydrocarbons and ethanol in low pressure oxygen and air plasmas, J. Mass Spectrom. 51 (12) (2016) 1187–1195, https://doi.org/10.1002/jms.3890. K. Sekimoto, M. Sakakura, T. Kawamukai, H. Hike, T. Shiota, F. Usui, Y. Bando, M. Takayama, Improvement in ionization efficiency of direct analysis in real timemass spectrometry (DART-MS) by corona discharge, Analyst 141 (2016) 4879–4892, https://doi.org/10.1039/C6AN00779A. D. Li, Y.-H. Tian, Z. Zhao, W. Li, Y. Duan, Ambient ionization and direct identification of volatile organic compounds with microwave-induced plasma mass spectrometry, J. Mass Spectrom. 50 (2) (2015) 388–395, https://doi.org/10.1002/jms. 3540. Y. Nunome, K. Kodama, Y. Ueki, R. Yoshiie, I. Naruse, K. Wagatsuma, Development of soft ionization using direct current pulse glow discharge plasma source in mass spectrometry for volatile organic compounds analysis, Spectrochim. Acta B Atom Spectrosc. 139 (2018) 44–49, https://doi.org/10.1016/j.sab.2017.11.002. R.B. Cody, Observation of molecular ions and analysis of nonpolar compounds with the direct analysis in real time ion source, Anal. Chem. 81 (3) (2009) 1101–1107, https://doi.org/10.1021/ac8022108. A.W. Nørgaard, V. Kofoed-Sørensen, B. Svensmark, P. Wolkoff, P.A. Clausen, Gas chromatography interfaced with atmospheric pressure ionization-quadrupole timeof-flight-mass spectrometry by low-temperature plasma ionization, Anal. Chem. 85 (1) (2013) 28–32, https://doi.org/10.1021/ac301859r. I. Dzidic, D.M. Desiderio, M.S. Wilson, P.F. Crain, J. McCloskey, Mass standards for chemical ionization mass spectrometry, Anal. Chem. 43 (13) (1971) 1877–1879, https://doi.org/10.1021/ac60307a043. E.S.C. Kwok, J. Arey, R. Atkinson, Alkoxy radical isomerization in the OH radicalinitiated reactions of C4–C8 n-alkanes, J. Phys. Chem. 100 (1) (1996) 214–219, https://doi.org/10.1021/jp952036x. J. Arey, S.M. Aschmann, E.S.C. Kwok, R. Atkinson, Alkyl nitrate, hydroxyalkyl nitrate, and hydroxycarbonyl formation from the NOx–air photooxidations of C5–C8 n-alkanes, J. Phys. Chem. A 105 (6) (2001) 1020–1027, https://doi.org/10.1021/ jp003292z. S.A. McLuckey, G.L. Glish, K.G. Asano, B.C. Grant, Atmospheric sampling glow discharge ionization source for the determination of trace organic compounds in ambient air, Anal. Chem. 60 (20) (1988) 2220–2227, https://doi.org/10.1021/ ac00171a012. Y. Wang, L. Hua, Q. Li, J. Jiang, K. Hou, C. Wu, H. Li, Direct detection of small nalkanes at sub-ppbv level by photoelectron-induced O2+ cation chemical ionization mass spectrometry at kPa pressure, Anal. Chem. 90 (8) (2018) 5398–5404, https:// doi.org/10.1021/acs.analchem.8b00595. X. Xie, Z. Wang, Y. Li, L. Zhan, Z. Nie, Investigation and applications of in-source oxidation in liquid sampling-atmospheric pressure afterglow microplasma ionization (LS-APAG) source, J. Am. Soc. Mass Spectrom. 28 (6) (2017) 1036–1047, https://doi.org/10.1007/s13361-016-1550-6. Y. Nunome, H. Park, K. Kodama, Y. Ueki, R. Yoshiie, S.C. Lee, K. Kitagawa, K. Wagatsuma, I. Naruse, Use of soft plasma ionization source at evacuated air atmospheres in time-of-flight mass spectrometry to suppress fragmentation of volatile organic compounds, Spectrosc. Lett. 48 (6) (2015) 436–440, https://doi.org/ 10.1080/00387010.2014.905962. A.A. Abdelaziz, T. Seto, M. Abdel-Salam, Y. Otani, Influence of nitrogen excited species on the destruction of naphthalene in nitrogen and air using surface dielectric barrier discharge, J. Hazard Mater. 246–247 (2013) 26–33, https://doi.org/

Talanta 204 (2019) 310–319

Y. Nunome, et al. 10.1016/j.jhazmat.2012.12.005. [30] E. Marotta, A. Callea, X. Ren, M. Rea, C. Paradisi, DC corona electric discharges for air pollution control, 2–ionic intermediates and mechanisms of hydrocarbon processing, Plasma Process. Polym. 5 (2) (2008) 146–154, https://doi.org/10.1002/ ppap.200700128. [31] S.E. Bell, R.G. Ewing, G.A. Eiceman, Z. Karpas, Atmospheric pressure chemical ionization of alkanes, alkenes, and cycloalkanes, J. Am. Soc. Mass Spectrom. 5 (3) (1994) 177–185, https://doi.org/10.1016/1044-0305(94)85031-3. [32] E.C.L. Cazedey, R. Bonfilio, M.B. Araújo, H.R.N. Salgado, A first-derivative spectrophotometric method for the determination of ciprofloxacin hydrochloride in ophthalmic solution, Phys. Chem. 2 (6) (2012) 116–122, https://doi.org/10.5923/j. pc.20120206.06.

[33] S.T. Ayrton, R. Jones, D.S. Douce, M.R. Morris, R.G. Cooks, Uncatalyzed, regioselective oxidation of saturated hydrocarbons in an ambient corona discharge, Angew. Chem. Int. Ed. 57 (3) (2018) 769–773, https://doi.org/10.1002/anie. 201711190. [34] N. Hourani, N. Kuhnert, Development of a novel direct-infusion atmospheric pressure chemical ionization mass spectrometry method for the analysis of heavy hydrocarbons in light shredder waste, Anal. Methods 4 (2012) 730–735, https:// doi.org/10.1039/C2AY05249K. [35] T.M. Hellman, G.A. Hamilton, Mechanism of alkane oxidation by ozone in the presence and absence of iron(III) chloride, J. Am. Chem. Soc. 96 (5) (1974) 1530–1535, https://doi.org/10.1021/ja00812a042.

319