Atmospheric Pressure Dark-Current Argon Discharge Ionization with Comparable Performance to Direct Analysis in Real Time Mass Spectrometry

Herein, a dark-current discharge state created by combining argon flow with a needle electrode in ambient air is described that has an ionization efficiency and mechanism comparable to those of conventional helium direct analysis in real time (DART), without requiring dopants and DART glow discharge. Using this method, polar compounds such as α-amino acids were ionized in the dark-current argon discharge via (de)protonation, molecular anion formation, fragmentation, (de)protonation with the attachment of oxygen, deprotonation with hydrogen loss and negative ion attachment. In contrast, nonpolar compounds (e.g., n-alkanes) were detected as positive ions via hydride abstraction and oxidation. Major background ions observed were H3O+(H2O)n, O2·+, O2·−(H2O)n and CO3·−. These results indicate that the present dark-current discharge efficiently generates resonance-state argon with an internal energy of ∼14.2 eV, higher than that of the well-known metastable state (∼11.6 eV). It is therefore suggested that ionization reactions occurring there can be attributed to the Penning ionization of atmospheric H2O and O2 by resonance-state argon, analogous to helium DART.


INTRODUCTION
Direct analysis in real time (DART) mass spectrometry was rst reported by Cody et al. in 2005. 1) It is a versatile technique that operates in open air, allowing rapid, noncontact analysis of solid, liquid and gaseous materials without any pre-treatment of samples. Among other things, DART analysis can be performed directly on the surface of clothes, banknotes, 2) fruit, 3) and vegetables. 4) Due to its other appealing features, such as high throughput, lack of memory e ect and simplicity, DART has been applied in the direct analysis of narcotics, 5) addictive drugs, 6) counterfeit drugs, 7) microorganisms, 8) warfare agents, 9) and pesticides, 10) for quality control, 11) in organic synthesis monitoring, 12) and for the surface analysis of living organisms. 13) In common DART, excited helium (mostly the metastable 2 3 S state, He(2 3 S)) is generated inside a ceramic ow chamber by an atmospheric pressure glow discharge using a DC voltage of 5 kV and an electric current of 3 mA. e He(2 3 S) gas is heated by passing it through a heater chamber, and then it is owed into a sampling area through a grid electrode.
e dominant positive-ion formation process is protonation, which results from the Penning ionization of atmospheric water by He(2 3 S). He(2 3 S) has an internal energy of 19.8 eV, which is higher than the ionization energy of water (12.6 eV). Penning ionization results in the generation of oxonium ions, H 3 O + , and its water clusters H 3 O + (H 2 O) n , followed by proton transfer to analytes with proton a nities greater than that of water (691 kJ mol −1 ). In negative-ion mode, analyte ionization can be attributed to proton transfer involving superoxide anion water clusters O 2 · − (H 2 O) n . Although helium DART has been performed with a great amount of success as described above, helium gas is quite di cult to obtain recently, which makes it hard to sustain its use. Argon is a possible alternative gas for DART. Several research groups have investigated how argon works for DART compared to helium. [14][15][16][17] Excited argon stably exists in discharges (including DART glow discharge) in metastable states, such as the 3 P 2 and 3 P 0 states, with internal energies of 11.6 and 11.7 eV, respectively. 15,18) ese energies are lower than the ionization energy of H 2 O, which results in the formation of fewer H 3 O + (H 2 O) n ions and give rise to quite low analyte ionization e ciency in argon DART. 15,16) us, dopant-assisted protonation based on atmospheric pressure photoionization has been used for the e ective operation of argon DART. 15,17) Herein, a novel argon discharge ionization technique under atmospheric pressure is reported in which the analyte ionization e ciency and mechanism are comparable to those of conventional helium DART. e present discharge system was easily established by modifying the conventional DART source: (i) a needle, whose tip surface has a speci c shape, is placed in the sampling area, (ii) heated ground state argon is owed through the sampling area and (iii) low DC voltage (< 2 kV) is applied to this needle using the electrospray voltage source of the mass spectrometer. Notably, the use of dopants and a DART glow discharge are not required. e resulting discharge state in the sampling area is referred to as a 'dark current,' a very low electric current (0.2-1 µA) compared to the DART glow discharge. Excited state argon formed makes it possible to e ciently generate H 3 O + (H 2 O) n and O 2 · − (H 2 O) n reagent ions. ese reagent ions lead to the formation of (de) protonated analytes, the abundances of which are equal to those generated in helium DART.

Mass spectrometry
A schematic illustration of the present experimental setup is shown in Fig. 1a. Hot argon (500°C) or helium (350°C) gas was owed through a DART-SVP source (IonSense, Saugus, MA, U.S.A.). e gas ow rates and purities were 2.0 L min −1 and >99.99% for both the gases. e voltages of the DART glow discharge and the exit grid electrode were −5.0 kV and 350 V, respectively. e DART source exit was directed towards the ceramic ion transfer tube of the Vapur interface or the metallic ori ce of the mass spectrometer, separated by a gap of 10 mm. Vapur interface is an additional pumping system to eliminate the helium gas. e area between the DART source exit and ori ce was surrounded by ambient room temperature air with a relative humidity of 40-60%. e needle used for the dark-current discharge was a 20 mm long stainless-steel pin with a diameter of 200 µm, which has a tip end radius with a curvature of ca. 1 µm and includes a tip end formed into a hyperboloid of revolution (Fig. 1b). is needle was placed perpendicular to the central axis of the mass spectrometer ori ce, with a distance between the needle tip and the ion transfer tube/ ori ce of 4 mm. is needle con guration was optimized for e cient (de) protonation of analytes in low electric eld. e electric eld distribution established on this needle tip surface and resulting ion chemistry have been well characterized elsewhere. 19) DC voltages for the dark-current discharge (< ±2.0 kV) were supplied using the electrospray ionization (ESI) voltage source of the mass spectrometer. e needle counterelectrode was the DART exit grid electrode separated by a gap of 12.5 mm. Analyte desorption and ionization were accomplished by inserting a 1.5 mm ID glass tube containing the solid-or liquid-state analytes (1.0±0.2 mg) into the gas ow at a position of 5.0±0.3 mm from the mass spectrometer ori ce, into which the resulting gas-phase analyte ions were introduced. Mass spectra were acquired with the following three mass spectrometers: LCQ Deca XP ion-trap ( ermo Fisher Scienti c, San Jose, CA, U.S.A.), LCMS-2020 quadrupole (Shimadzu, Kyoto, Japan), or AccuTOF time-of-ight mass spectrometers (JEOL, Tokyo, Japan). All of the instruments comprised an atmospheric pressure ionization (API) interface, two di erential pumping regions and analysers. However, the LCQ ion-trap  35) and LCMS-2020 quadrupole mass spectrometers have a Vapur interface in front of the API interface. e voltages and temperatures of the ori ce, lens and skimmer in the primary di erential pumping region of each instrument were manually adjusted to inhibit excessive fragmentation and clustering processes. ese parameters are summarized in Table 2.
In this work, the following four sets of discharge conditions were used in terms of gas species and turn-on/o DART glow discharge and dark-current discharge (DCD): (i) Ar-DART (turn-on DART operated with Ar), (ii) Ar-DART+DCD (in which argon is owed through turnon DART and DCD), (iii) Ar-DCD (where argon is owed only through turn-on DCD, but argon is heated earlier in a DART source) and (iv) He-DART (conventional DART operated with He). e details of the discharge conditions (i)-(iv) are summarized in Table 3.

RESULTS AND DISCUSSION Background ions formed in the atmospheric pressure dark-current argon discharge
No ions were observed in background mass spectra when performing only Ar-DART (Fig. 2a). However, when the DCD needle was used, a fairly low electric current of 0.2-1 µA was recorded, and background ions were readily observed regardless of whether the DART glow discharge was turned on or not (Figs. 2b and c). e turn-on DART glow discharge was not required for forming background ions by the DCD needle, while argon ow was essentially needed. Notably, the resulting total ion intensities were very similar to those observed in the conventional He-DART technique (Fig. 2d). e turn-on DCD needle voltage of 0.2-1 µA resulted in no visible light on the needle tip or between the DART source exit and the mass spectrometer ori ce.
is non-self-sustaining discharge state can be described as a 'dark current (or Townsend dark).' 20,21) Dark-current discharge is energetically much lower than self-sustaining discharge with visible light and electric current higher than 1 µA, which have been used in atmospheric pressure argon plasma mass spectrometry so far, e.g., corona discharge used in atmospheric pressure Penning ionization (APP e I), 18) glow discharge used in Ar-DART, [14][15][16][17] microwave discharge used in microwave-induced plasma ionization, 22) and inductively coupled plasma (ICP) used in  Table 3. e four sets of discharge conditions used in this work.

Voltage
Gas DART glow discharge ICP mass spectrometry. 23) e present dark-current argon discharge (Ar-DCD) can form background ions similar to conventional He-DART. Figure 3a shows the positive background ion mass spectrum measured by Ar-DCD. e major positive ion observed is the oxonium ion, H 3 O + (m/z 19), and its water cluster, H 3 O + (H 2 O) n (m/z 19+18n). It should be noted that the molecular ion of oxygen O 2 · + (m/z 32) is also present. Taking into account the ionization energy (IE) of oxygen at 12.1 eV, it is most likely that Ar-DCD e ciently generates excited state argon with an internal energy higher than those of metastable states, e.g., the resonance state 5S 3 P 1 at an internal energy of 14.1 eV and 5S 1 P 1 at 14.3 eV. 24   e negative ion mass spectrum also shows a predominantly high ion peak for CO 3 · − (m/z 60). is indicates that ozone, O 3 , is abundant in the discharge area. According to the di erence in the electron a nities of O 2 (0.5 eV) and O 3 (2.1 eV), 26) charge transfer from O 2 · − to O 3 (reaction 9) easily occurs, and the resulting O 3 · − ions formed react with CO 2 in ambient air to generate CO 3 · − (reaction 10). 27)

Analyte ionization in Ar-DCD
Ar-DCD resulted in analyte ionization e ciencies and characteristics comparable to those of conventional He-DART for polar and nonpolar compounds. Figure 4 shows the mass spectra of eight α-amino acids (A) positively and negatively ionized in Ar-DCD. (De) protonated molecules [A±H] ± were dominantly observed for all of the amino acids, and the mass spectral patterns were nearly identical to those observed using the He-DART technique. 28) e other ion species detected were molecular anions A · − , oxygenated (de) protonated molecules [ ), as shown in Fig.  4 and Table 4c. Figure 5 shows the positive-ion Ar-DCD mass spectra of n-pentadecane and n-heptadecane. Both alkanes (Alk) were detected as [Alk+O−3H] + (m/z Alk +13) and their monohydrates [Alk+2O−H] + (m/z Alk +31), and their spectra were found to be analogous to those observed using the He-DART technique. 29,30) An interesting ionization characteristic of Ar-DCD is the high e ciency of protonation, comparable to that of the He-DART technique. To examine how protonation occurs in Ar-DCD, anisole was measured using the four sets of discharge conditions. As anisole has a relatively low ionization energy of 8.2 eV, it can be ionized to a molecular ion by Ar-DART alone via the Penning ionization of metastable argon Ar * meta (reaction 13) 15) : * ⋅+ − * Anisole Ar Anisole Ar e (Ar : 4S P and 4S P states, at 11.6 and 11.7eV, respect ively) Ar-DART also forms protonated molecules with ∼10% intensity of the molecular ion (Table 4), which is consistent with what has been previously shown in the literature. 15) is is most likely due to hydrogen atom transfer from the neutral anisole molecule to the molecular ion (reaction 14) 18) : When the DCD is turned on, the relative intensity of the protonated molecule to the molecular ion increases: 28 and 58% for Ar-DART+DCD and Ar-DCD, respectively (Table  4). e 58% intensity is even higher than that observed in the He-DART technique (47%, as shown in Table 4), and these results suggest that Ar-DCD has a speci c protonation pathway, other than those shown in reactions (13) and (14).
Taking into account the background ions in Ar-DCD described earlier, it is thought that the analyte ionization mechanism is the same as that observed in the He-DART technique. 28) As proton a nities of neutral or deprotonated amino acids (shown in Table 1) are su ciently higher or lower than that of H 2 O (691 kJ mol −1 ) or O 2 · − (1476.9±3.0 kJ mol −1 ), reactions 15 and 17 exothermically proceed at rates close to the e absolute intensities of [A±H] ± observed in Ar-DCD were 1.1-8.1 times higher than those in He-DART (see Table 4c and d). is result may suggest that the abundance of Ar * res is higher than that of He(2 3 S) in the analyte ionization area, and resulting in higher efciency of the background ion formation and analyte (de)protonation in Ar-DCD than He-DART. Its main reason is likely that helium with low mass of 4 can disperse more easily compared to argon with mass of 40.
e oxidation process, in the form of both oxygen at- Fig. 4. Positive-and negative-ion Ar-DCD mass spectra of the eight α-amino acids obtained by the LCQ ion-trap mass spectrometer. AI represents the absolute intensity (arbitrary units) of a given ion. e ion intensities observed here are summarized in Table 4. tachment and hydrogen loss processes, involves mainly HO · hydroxyl radicals. A more detailed explanation of these ionization reactions can be found elsewhere. 28 31) which is consistent with the results obtained here.
An interesting feature of analyte ionization by Ar-DART+DCD is that the formation of oxygenated and/or dehydrogenated (de) protonated molecules, i.e., [A±H+nO] ± , Table 4. Ion species and intensities observed in the (a) Ar-DART, (b) Ar-DART+DCD, (c) Ar-DCD and (d) He-DART mass spectra of the eight α-amino acids, two n-alkanes and anisole.

Excitation of argon in dark-current discharge
In the case of argon, it is known that there are a number of excited states up to the rst ionization potential (15.8 eV), as shown in Fig. 6. Of these, the metastable states at ~11.6 eV (e.g., 4S 3 P 2 and 4S 3 P 0 ) and the resonance states at ~14.2 eV (e.g., 5S 3 P 1 and 5S 1 P 1 ) can occur Penning ionization. 24) As mentioned above, Ar-DCD e ciently generates H 3 O + (H 2 O) n and O 2 · + (Fig. 3a). is indicates that resonance-state argon is dominantly formed in Ar-DCD. e presence of resonance-state argon was also con rmed by direct reaction of benzene with argon excited by the DCD. is reaction resulted in the production of not only molecular ion (C 6 H 6 · + at m/z 78) but also several fragment ions such as C 6 H 5 + (m/z 77) and C 5 H 4 + (m/z 52), as shown in Fig. 7. According to the results of electron ionization (EI), these fragment ions originate from the molecular ion and their appearance energies are 14.1-14.2 eV on average. 32 Energy corresponding to UV radiation is possessed by an argon atom for quite a long time (compared to when this energy behaves as UV radiation), which means that resonance-state argon has a long e ective lifetime in ambient air. 33) e formation of resonance-state argon can be attrib- Electrons with kinetic energies of higher than ∼16 eV can form high-Rydberg state argon with an internal energy of ∼15.6 eV, 34,35) as well as argon ions (Ar + ). High-Rydberg argon easily converts to dimer ions (Ar 2 + ) upon the loss of electrons. 35) However, neither Ar + nor Ar 2 + were detected in the positive background ion mass spectrum (Fig. 3a). is may indicate that in the present discharge conditions, high-Rydberg state atoms and argon-related ions are not formed e ciently or e cient lifetimes of them are too short to detect by the mass spectrometer. e needle tip shape plays a crucial role in enhancing and sustaining the Ar-DCD ionization e ciency. When DCD is created by the present needle, which has a tip end radius with a curvature of ca. 1 µm and includes a tip end formed into a hyperboloid of revolution (Fig. 1b), the ion intensities remain constant even a er a time lapse of 30 min (+1.8 kV in Fig. 9a). However, if a needle, including a tip end formed into a reversed curved surface, is used, the ion intensities signi cantly reduce right a er the application of the DCD voltage (+2.5 kV in Fig. 9b). is is because the tip end radius of the curvature is too small, and hence the shape of the tip end surface changes with the passage of time due to degradation by thermal melting and plasma corrosion, meaning that the electric eld established on the tip end surface cannot be kept constant. In contrast, if a needle with a larger radius of curvature (> 30 µm) is used, ions are not detected in the DCD state (+0.5~+2.9 kV in Fig. 9c). If the tip end radius of curvature is too large, and hence in the DCD state, it is impossible to ensure a region of the tip end position at which su cient amounts of high-energy electrons and resonance-state argon are formed. Only when breakdown discharge involving an emission phenomenon occur beyond the dark-current state, ions are detected with a spike shape (+3.0~+3.2 kV in Fig. 9c). erefore, the Ar-DCD needle has the following conditions: (i) the tip end has a radius of cur-  Finally, we should note the following interesting ndings: Discharge power level of the dark-current state is signicantly lower than breakdown state observed in Ar-DART. However, the dark-current has higher level than breakdown in terms of energies involved in ionization, indicating of more e cient formation of highly excited argons under lower discharge level. To clearly interpret these ndings, further investigation using spectroscopy is needed, e.g., detailed measurements of densities of various excited states under individual discharge states.

CONCLUSION
An atmospheric pressure dark-current discharge state, created by combining argon with a needle electrode in ambient air (Ar-DCD), was found to have an ionization e ciency and mechanism comparable to those of conven- Fig. 9. Total ion chromatogram and average mass spectra of positive background ions obtained using (a) the present DCD needle (Fig. 1b), (b) a DCD needle which has a tip end radius with a curvature of less than 1 µm and includes a tip end formed into a reversed curved surface, and (c) a DCD needle, which includes a tip end formed into a hyperboloid of revolution, but has a tip end radius with a curvature of more than 30 µm. e mass spectrometer used here was the LCQ ion-trap. AI represents the absolute intensity (arbitrary units) of the base peak in a given mass spectrum.
tional helium DART, without requiring the use of a dopant or DART glow discharge. Ar-DCD can ionize polar compounds such as α-amino acids (A) to (de) protonated molecules [A±H] ± , molecular anions A · − , oxygenated (de)protonated molecules [A±H+nO] ± , dehydrogenated deprotonated molecules [A−2H−H] − , fragment ions [A±H−F] ± (F: neutral fragment) and negative ion adducts [A+R] − (R − : negative background ion). e absolute intensities of the (de) protonated molecules were found to be 1.1-8.1 times higher than those observed using the helium DART technique. In contrast, using Ar-DCD, non-polar compounds (e.g., n-alkanes; Alk) were detected as [Alk+O−3H] + and [Alk+2O−H] + ions via hydride abstraction and oxidation processes. Major background ions observed using Ar-DCD were H 3 O + (H 2 O) n , O 2 · + , O 2 · − (H 2 O) n and CO 3 · − , while argonrelated ions were not observed. ese results indicate that Ar-DCD e ciently generates excited state argon with an internal energy higher than those of well-known metastable states (∼11.6 eV), e.g., resonance states such as 5S 3 P 1 with an internal energy of 14.1 eV and 5S 1 P 1 at 14.3 eV. erefore, this suggests that ionization reactions occurring in the Ar-DCD method can be attribute to the Penning ionization of atmospheric H 2 O and O 2 by resonance-state argon, in a similar manner to that in the helium DART method. It was also found that the needle tip shape for DCD plays an important role in enhancing and sustaining the Ar-DCD ionization e ciency. e needle used in this work, i.e., a needle that has a tip end radius with a curvature of ca. 1 µm and includes a tip end formed into a hyperboloid of revolution, is very suitable for this purpose. e present atmospheric pressure dark-current argon discharge ionization technique will contribute to enhancing analytical techniques based on DART.