A systematic comparison between half and full dielectric barrier

Subscriber access provided by UNIV OF NEWCASTLE

Article

A systematic comparison between half and full dielectric barrier discharges based on the LTP and DBDI configurations Felix David Klute, Sebastian Brandt, Pascal Vogel, Beatrix Biskup, Charlotte Reininger, Vlasta Horvatic, Cedomil Vadla, Paul B. Farnsworth, and Joachim Franzke Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b02174 • Publication Date (Web): 20 Jul 2017 Downloaded from http://pubs.acs.org on July 23, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

A systematic comparison between half and full dielectric barrier discharges based on the LTP and DBDI configurations Felix D. Klute†, Sebastian Brandt†, Pascal Vogel†, Beatrix Biskup┼, Charlotte Reininger‡, Vlasta Horvatic⁰, Cedomil Vadla⁰, Paul B. Farnsworth‡, and Joachim Franzke†* †

ISAS - Leibniz Institut für analytische Wissenschaften, Bunsen-Kirchhoff-Str. 11, 44139 Dortmund, Germany Department of Chemistry and Biochemistry Brigham Young University Provo, UT 84602, USA ┼ Experimental Physics II - Reactive Plasmas, Ruhr-Universität Bochum, Universitätstraße 150, 44780 Bochum, Germany ‡

⁰

Institute of Physics, Bijenicka 46, 10000 Zagreb, Croatia

ABSTRACT: Dielectric barrier discharge (DBD)-based analytical applications have experienced rapid development in recent years. DBD designs and parameters and the application they are used for can vary considerably. This leads to a diverse field with many apparently unique systems that are all based on the same physical principle. The most significant changes among DBDs used for chemical analysis are in how the discharge electrodes are separated from the ignited discharge gas. While the official definition of a DBD states that at least one electrode has to be covered by a dielectric to be considered a DBD, configurations with both electrodes covered by dielectric layers can also be realized. The electrode surface plays a major role in several plasma-related technical fields, surface treatment or sputtering processes for example, and has hence been studied in great detail. Analytical DBDs are often operated at low powers and atmospheric pressures, making a direct transfer of insight and know-how gained from the aforementioned well-studied fields complicated. This work focuses on comparing two DBD configurations: the low temperature plasma probe (LTP) and the dielectric barrier discharge for soft ionization (DBDI). The LTP is representative of a DBD with one covered electrode and the DBDI of a design in which both electrodes are covered. These two configurations are well suited for a systematic comparison due to their similar geometric designs based on a dielectric capillary.

The success of atmospheric dielectric barrier discharges for analytical applications in the recent years has brought forth a wide variety of different discharge designs used for an at least as large variety of different applications1-7. While each of these discharges fulfills the criteria of having at least one electrode separated from the rest of the discharge volume, qualifying it as a DBD, the rest of the discharge design can be freely chosen and optimized for each individual application.8 The wide variety of different discharge designs leads to drastically different parameter ranges that can be achieved with DBD discharges. These parameter ranges include pressures from several mbar up to several bar, sizes of several µm to m, voltages of several 100 V up to 10 kV, and excitation frequencies of less than 100 Hz up to 10 MHz. This broadens the field of dielectric barrier discharges, creating an environment in which all the relevant information and applications are hard to track. Some of these DBD parameter ranges, size or pressure, for example, are very striking, and the extremes are obviously unsuited for analytical applications. In other cases the differences between the discharge designs can be much more subtle, making it difficult to exactly explain why a certain design works for one application but not for another. A good example might be the LTP and the DBDI which are quite similar at first glance. Both are operated at atmospheric pressure with excitation voltages of several kV. The voltages are modulated at kHz frequencies and the gas used for the discharge is transported through a glass capillary with flows in the range of 100 – 1000 ml/min. The most

striking difference between the two discharges, aside from the larger size of the LTP, is the fact that both electrodes of the DBDI are mounted outside of the glass capillary, while the grounded electrode of the LTP is inserted into the capillary. Figure 1 shows schematics of the common LTP and DBDI geometries. We have modified them to create a base geometry that is suited for systematic measurements: the so called variable dielectric barrier discharge (VDBD). The focus of these systematic measurements will be to find mechanisms that are present independent of the chosen discharge configuration and can be regarded as a fundamental mechanism of a dielectric barrier discharge in the given parameter range. At the same time the important differences caused by the introduction of an uncovered electrode surface to the discharge will be shown and the possible influence on the discharge efficiency or overall performance will be evaluated. In the end this might help clarify the use of the term DBD in the context of analytical applications and refocus on important similarities or differences of given applications. EXPERIMENTAL SECTION The basic experimental arrangement was already developed and used in previous studies, and only modified for the current work9,10. The key component to the experimental work is the in-house built HV-generator that works with a rectangular voltage waveform ranging from 0 – 3.5 kV and frequencies of 1 – 100 kHz.

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 8

Figure 1: Schematics of the LTP A) and DBDI B) with their common geometries and sizes (to scale). Next to this the VDBD C) is shown. The LTP configuration (version 1) based on the LTP represents the partially covered, the DBDI configuration (version 2) based on the DBDI the fully covered case.

The generator output is optimized and in general operated at 20 kHz. The sharp and reproducible voltage rise of the generator, with typical values of approximately 40 ns per 1 kV with no noticeable overshoot oscillations, is ideal for longterm discharge operations that are needed in this systematic comparison of different discharge configurations.

Figure 2: Mounted VDBD with magnetic manipulator. With the movable grounded wire and cylindrical electrodes one can freely switch between the LTP and DBDI configuration. In addition the influence of the electrode gap distance can be investigated.

Another important part of this work is the development of the variable dielectric barrier discharge (VDBD), which was necessary to find a common base between the standard LTP and DBDI configurations. The size difference between the discharges is significant, as shown in Figure 1, making it necessary to find a common basic size. The final decision was to use capillaries with 1.5 mm outer and 1.0 mm inner diameter, which is between the dimensions of LTP and DBDI. Preliminary tests between a standard DBDI capillary with o.d. 0.9 mm and i.d. 0.46 mm, and the bigger capillaries were done to make sure that they can be used as a substitute. These tests showed that the bigger capillaries behave qualitatively in the same way as a standard DBDI regarding physical parameters, their ignition behavior and their properties as ionization sources for ambient MS. The same attempts were made for comparing a VDBD in LTP configuration with a standard LTP discharge. While the physical parameters showed satisfying agreement, the evaluation of the analytical performance proofed to be difficult. The standard LTP is in general unsuited for the direct ambient ionisation of analytes, due to its bigger size and high gas consumption and is instead used as an ambient desorption source. The smaller VDBD in LTP configuration can be used as source for direct ambient ionisation with a MS but lacks the power and higher gas volumes to be used as an efficient desorption source. The VDBD in LTP configuration can therefore not be compared to it’s bigger pendant the standard LTP analytical wise but purely out of geometric factors that are not a core part of this work. A 3D-printed housing was developed during the preliminary testing, making quick and easy exchange of different discharges possible. These housings are manufactured in a way that the capillaries are mounted with gas-tight seals and


Page 3 of 8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


can be attached to swagelok T-connectors for easier handling. The housings were tested with He gas pressures of up to 4 bar overpressure, and showed no noticeable leakage of gas or burst damage. The standard PLA filament used for the housing was also tested regarding its analytical suitability, and no significant contamination by possible outgassing could be found in preliminary MS measurements. Figure 2 shows a schematic of a mounted VDBD with the addition of a newlydeveloped magnetic manipulator. The manipulator can be used to move a 100 µm diameter tungsten wire inside the mounted VDBD without reopening the gas tight system, making it possible to freely change the gap distance between the powered HV electrode and the grounded wire electrode. As one can see from Figure 1 this electrode gap is small in the case of a LTP and large in the case of a DBDI discharge configuration, making it necessary to systematically and reproducibly change the discharge gap to identify potential influences on the discharge behavior. The VDBD arrangement was investigated with several approaches and different diagnostics. The ignition behavior of the different VDBD configurations was studied by measuring the time-resolved discharge currents, which gives direct information on the formation and duration of different discharge stages, and also gives estimates of the amount of charge produced in the plasma. Previous studies have shown that, in combination with time and space resolved emission measurements, this gives insight to the formation of charged species in the discharge and how relevant species are involved in the processes9,10. A ns gated intensified CCD camera (Andor DH 720 18F-03) was used for the optical measurements, which provided a better time and space resolution compared to photomultiplier measurements carried out in prior studies. The minimum gate time of the camera is 5 ns, which means that an ignition cycle of a plasma with usual times of 1 – 2 µs can be scanned with 200 – 400 single frames. A single frame can be accumulated several times to further increase the SNR. This in turn greatly increases the measuring time and is therefore only possible if the imaged discharge is stable over time. Spectral resolution was realized by using different interference filters matched to the wavelengths of relevant discharge species. MEASURMENTS AND RESULTS A direct optical comparison of the two discharge versions by naked eye already shows clear differences. A strong emission maximum appears that is localized to the tungsten wire, which is introduced directly into the plasma for the VDBD in LTP configuration. Spatio-temporal resolved emission measurements of spectrally integrated emission of all species is shown in Figure 3. They reveal that the development of both discharge configurations inside the capillary is similar. The inner early plasma, which starts from the powered HV electrode around 75 ns, travels towards the grounded inner wire or outer ring electrode. The inner plasma reaches the grounded electrode area around 400 ns after the initial HV rise in both discharge configurations, meaning that the velocity of the excitation front that constitutes the early plasma is nearly the same in both cases. Most striking though is the appearance of an additional emission event occurring in the direct vicinity of the grounded tungsten wire in a time frame of 75 to 250 ns after the voltage rise, and therefore long before the inner early plasma reaches the grounded electrode. The early plasma stage ends after 400 ns have passed and the coincident plasma is

ignited

inside

of

the

capillary.

Figure 3: Spatio-temporal resolved emission measurements of the VDBD in LTP B) or DBDI C) configuration. The time development is shown at different points t1=75 ns, t2=250 ns, t3=325 ns, t4=400 ns and t5=550 ns after the trigger pulse. The discharge current is shown in A). Both discharges were driven with 500 ml/min He 5.0, 3.0 kV @20 kHz square-wave voltage and are shown with the same intensity scaling. The animated movies of these measurements are included in the supplements.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

This plasma stage was called coincident plasma (stage) in previous works9 due to its simultaneous occurrence with the discharge current. It can be identified by the clear reigniting process inside the capillary and the long decay of emission that can last up to several µs. It is noticeable that the emission from this plasma stage is much stronger for the LTP configuration than for the DBDI, and also seems to last longer in time. Another part of the discharge that seems to be stronger in the LTP configuration is the outer early plasma stage referred to as plasma jet in earlier studies. Not only does the plasma jet start 60 ns sooner and its propagation speed is with 37.2 km/s, nearly double as high as the propagation speed of 21.4 km/s that are reached in the DBDI configuration, but also the overall intensity is twice as high. It can therefore be stated that the LTP configuration seems to produce a much stronger and intense plasma when operated under the same experimental conditions as a comparable DBDI discharge. Another indicator that strengthens this statement is the currents shown in Figure 3 A) that were measured as a time reference for the spatio-temporal resolved emission measurements.

Page 4 of 8

electrode size is increased symmetrically, meaning that the size of both the powered as well as the size of the grounded electrode has to be increased. If only one electrode is increased, bottlenecks will appear. Either the charges produced by a very large HV electrode cannot be sufficiently discharged at a very small grounded electrode or a very small powered electrode cannot produce enough charges that could potentially discharge at a very large grounded electrode. This bottleneck effect leads to a reduced effective total electrode size and underlines the importance of surface recombination effects in a full dielectric barrier discharge. Not only will these bottleneck effects not appear in a half dielectric barrier discharge due to the fact that charges can freely recombine and discharge at the surface of the conducting electrode, but also an overall higher amount of charge can be coupled into the discharge. This overall higher amount of charge coupled into the plasma is an explanation for the increase of the overall emission as well as increase of propagation speed of the plasma jet observable in Figure 3B and 3C.

Figure 4: Amount of charge coupled into the discharge depending on the applied voltage. The number of charges was obtained by integrating the time resolved discharge current. The LTP configuration creates a higher amount of charge when directly compared to an equivalent DBDI configuration.

While the currents are qualitatively very similar and show that discharges ignite essentially at the same point in time, they also show that a much higher current is flowing in the LTP configuration. The dependence on applied voltage of the charge coupled into the two discharge configurations is shown in Figure 4. The higher slope for the LTP indicates that, while the charge can be influenced by many factors, including total electrode area and tube diameter, these influences are always weaker in the DBDI. The introduction of a grounded conducting medium into a DBD seems to create a free and therefore more efficient flow of charges that would otherwise be hindered by the dielectric barrier. The presence of the barrier leads to a restriction of total charges that can be coupled into the discharge. The total amount of charges coupled into a complete DBD such as the DBDI can be easily manipulated by the total area of the electrodes as long as the

Figure 5: Absorption measurements of the He* 1083 nm line (2s 3 S1 → 2p 3P2,1,0o transition) and B) VUV emission measurements in the range of 55 – 180 nm. Both measurements show a higher absorption and emission in the LTP configuration compared to the DBDI configuration.

A higher amount of charge means that the discharge will be shifted into a new equilibrium state. In this new state the initial polarization by the high voltage will be stronger leading to higher positive charge that will attract more and faster electrons towards the HV electrode. This leads to the much brighter and stronger formation of the jet that is in principle decoupled from the inner plasma processes, but still influenced by the discharge equilibrium inside. A similar argument can be used to explain the stronger coincident emission of the LTP configuration. The electrons that are reaccelerated from the surface of the former HV electrode can


Page 5 of 8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


freely reach the electrode surface and recombine when they are not restricted by an additional dielectric barrier, as happens in the case of the DBDI configuration. Summarizing these measurements it can be stated that a half dielectric barrier discharge, such as the VDBD LTP configuration, seems to create charges more efficiently than a full dielectric barrier discharge, such as VDBD DBDI configuration. This statement is further supported by absorption measurements conducted in cooperation with the department of chemistry and biochemistry at Brigham Young University and VUV measurements done in cooperation with the group Experimental Physics II - Reactive Plasmas at RuhrUniversity-Bochum. The results of these measurements can be seen in Figure 5. The absorption of the 1083 nm line due to the transition from He* 2s 3S1 metastable to He* 2p 3P2,1,0o state as well as the emission in the range of 60 – 100 nm originating e.g. from the transitions A1 ∑u+ ↔ X1∑g+ and D1∑u+ → X1∑g+ of the He2* 1st and 2nd excimer of the Hopfield continuum11,12, are noticeably higher in the LTP configuration than in the DBDI configuration. This indicates that the He* and He2* species that are involved in these processes have higher populations in the LTP configuration than in the DBDI configuration. Most striking in Figure 5 is that the emission of the 2nd excimer in the Hopfield continuum clearly rises more than the 1st excimer when the discharge configuration is changed from the DBDI to the LTP configuration. A possible explanation might be that the change of the discharge configuration leads to a higher population of higher energy states, such as the D1∑u+, which contributes more to the 2nd excimer originates. This state has an energy of 20.5 eV compared to the 18.1 eV upper level A1 ∑u+ of the 1st excimer. This shift of ratio between emission lines can also be observed in the visible range for He* 587.6 nm and He* 706.5 nm. The intensity of the 587.6 nm line increases more than that of the 706.5 nm line when the discharge configuration is changed from DBDI to LTP. Also in this case the upper level 3d 3D3,2,1 of the 587.6 nm line, with 23.1 eV has more energy than the 3s 3S1 state of 706.5 nm that only has an energy of 22.7 eV. The change of the population ratios of the aforementioned He related energy states that is indicated by the change of emission line ratios in the VUV range and in the visible range (shown in the supplementary data) could be explained by a change in the electron energy distribution function. The distribution of the electron energy could be influenced by a shift in the heating mechanisms of the discharge that is caused by the half dielectric configuration in a LTP like configuration. The extent of this shift in heating mechanisms cannot be determined from the current experiments, and could be a promising topic for additional research. In the context of the current work it suffices to point out the clear differences between half and full dielectric barrier discharges regarding the physical parameters of the discharges. It can be summarized that under all considered points of view, a half dielectric barrier discharge like the VDBD LTP configuration seems to produce a much stronger plasma than a full dielectric barrier discharge like the VDBD DBDI configuration. In the following analytical performance of both discharges VDBD LTP and DBDI configuration will be evaluated via DBD-MS measurements performed with a Thermo LTQ. Similar to the emission experiments described, both discharges were operated at the same conditions, and only the mechanical change from a LTP to a DBDI configuration contributed to the changes observed in the MS

measurements. The VDBD was operated in either LTP or DBDI configuration in front of the LTQ until satisfactory spectra could be measured. While the measurement was still carried out the VDBD was changed into the other version. To make sure that all changes caused by this reconfiguration were properly recorded, all MS parameters that could influence the measurement were deliberately fixed. The scan was defined by a m/z range of 50-400. The AGC was activaed with a maximum injection time of 100 ms. For highly temporal resolved chronograms a micro scan count of 1 in positive ion mode was chosen. A complete list of the parameter setting and tuning parameters of the LTQ is given in the supporting information. The VDBD in both configurations was used with a helium flow of 500 ml/min (purity 99,999 %), a rectangular high voltage with an amplitude of 2.6 kV and an axial configuration facing the MS inlet. The analyte was introduced via headspace capillary (o.d. 360 µm, i.d. 250 µm) with a nitrogen flow of 10 ml/min perpendicular to the ionisation direction as shown in two figures of the supporting information.

Figure 6: Normalized mass spectra of propachlor measured with a Thermo LTQ. A) shows the spectrum measured with the LTP configuration B) of an equivalent DBDI configuration. The base + peak intensity of [M+1] = 212.7 used for normalization was

5.1×104 for A) and 1.3×106 for B) which is around 1.3 times higher. The insets show the representative background signals of the spectra yielded without analyte supply.

Several substances including hexane (C6H14), menthone (C10H18O) and propachlor (C11H14ClNO) were studied. The normalized spectra of propachlor are shown in Figure 6. The spectra qualitatively show the same features with the protonated analyte peak at 212.7 m/z. While the qualitative features of both spectra are quite similar the signal height used for the normalization is noticeable different when comparing



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

both discharge configurations. The intensity of the base peak of propachlor measured at [M+1]+ = 212.7 had an intensity of 1.3×106 for the DBDI configuration and dropped to 5.1×104 for LTP configuration, a factor of 25. In addition the SNR obtained from the spectrum in Figure 6A is only 5.9×103 and the SNR gained from Figure 6B is 7.6×103, a factor of about 1.3 times. The SNR and absolute signal differences illustrated in Figure 6 are representative for a wide range of measurements, only a fraction of which can be presented here. It can be said that while the VDBD in LTP configuration produces the seemingly stronger plasma, the VDBD in DBDI configuration performs better in combination with a MS. During further testing, which was done to make sure that it was really a systematic property of the LTP that it has a seemingly less efficient overall ionisation compared to its DBDI counterpart, the reason for this observed difference was partly discovered. It was found that the grounded electrode has a strong influence on the measured mass spectra, in particular when the discharge is operated in the LTP configuration. The two possible states for the “grounded” electrode of the discharge can either be that it is connected to the ground given by the HV generator, or that it is not connected to anything making it a floating ground. For the first case of having a real grounded electrode, poor absolute signal and SNR for the LTP configuration occurs, as already described. When the electrodes were not connected to anything the mass spectral signals obtained by the LTP and DBDI configuration both increased noticeably. The increase of signal quality shown in Figure 7 is more pronounced for the LTP configuration. While the relative increase of SNR due to the floating ground is only around 7.5 % more for the LTP configuration compared to the DBDI (Figure 7B), the absolute gain of signal intensity is much more pronounced (Figure 7A).

Figure 7: A) Relative signals of the protonated propachlor peak

[M+1]+ = 212.7 in the floating and grounded discharge states. The signals were normalized to the intensity of the grounded LTP. B) Relative increase of the SNR for the floating discharge state in the DBDI (green) and LTP (blue). The SNR was normalized to the grounded discharge state.

Although the relative increase for the LTP is large, its absolute signal intensity is still lower than the signal from the DBDI.

Page 6 of 8

All measurements shown in Figure 3 – 5 were repeated once the influence of the floating or grounded mass electrode was discovered during the MS measurements and the impact of these different discharge states was evaluated. The results are that the discharge consistently and for all the above considered and observed parameters such as charge, emission and excitation propagation speed, decreases by around 10 – 20 % for the floating discharge state. The magnitude of the decrease is the same for the LTP and DBDI configuration and therefore does not explain the much more drastic signal increase for the LTP version. One possible explanation might be found if one once more considers the role of the conducting surface inside the discharge volume. As already discussed, this conducting surface might allow for an increased flow of charges compared to dielectric-covered surface, due to the fact that electrons can leave the discharge volume more efficiently and ions can recombine via surface interactions. A sheath region around the conducting surface, with a corresponding sheath potential, will develop that encumbers the free flow of charges, keeps fast electrons in the bulk volume of the plasma, and transports slow positive ions out. The sheath region regulates the discharge and prevents the plasma from charging up due to an uncontrolled loss of electrons. This sheath potential will greatly decrease when the conducting surface is not connected to a grounded mass due to the fact that it will quickly charge up itself and no continuous loss of charged species will occur. A floating conducting surface can be therefore seen as small capacitor.

Figure 8: Chronograms of Menthon [M+1]+ = 155.3 for a VDBD version 1 with different capacitances connected to the grounded electrode. The electrode and the respective capacitances were connected to GND before the start of the measurement (t < 0) and then than disconnected from it (t = 0). The masstrails show a clear charging behavior which increases in time depending on the capacitor size. The dotted lines in the trails are guides for the eye and the black and red dotted line show the average masstrail level of a grounded or floating discharge without additional capacitors.


Page 7 of 8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


This understanding of the floating electrode’s behavior is supported by the measurements shown in Figure 8. If the capacitance of the grounded surface is artificially increased by adding capacitors to it, the system will show a significant charging behavior when the discharge state is switched from grounded to floating. The MS signal will stagnate on a certain level and only slightly increase, meaning that it is still quasi grounded as long as the capacitor is not fully charged. When the capacitor is charged the signal will rise with a characteristic charging curve.While the increase of the signal for the conducting surface without an additional capacitance is so fast that it is virtually unmeasurable, it takes over 30 min to charge when an additional 1 F capacitor is connected in series to the surface. Figure 8 clearly shows that the floating state is somehow connected to internal charging processes in the plasma. We assume that the internal discharge potentials also change in the way as discussed above. A change of the internal potentials and electric fields might be a reasonable cause for the differences that can be observed for the VDBD in LTP and DBDI configuration with changes in the grounding configuration. Direct measurement of such fields and their possible changes are beyond the scope of this work.

discharges seemingly work better for certain analytical applications and not for others. Based on these findings further studies comparing other discharge parameters with the same systematic approach as was presented might lead to more insights regarding the underlying discharge mechanism that are important for analytical applications. Promising new parameters that will be studied in the future are excitation frequencies, waveforms and voltages and also discharge sizes and different designs, that all might have a huge influence on the produced plasma and its analytical performance.

Conclusion A systematic approach was presented that clearly showed a mechanistic difference between half and full dielectric barrier discharges. A new discharge design, the VDBD, was presented that could be operated in two different configurations: one representing the half dielectric barrier discharges and based on the LTP design and the other representing the full dielectric barrier discharges and based on the DBDI. This discharge design was necessary to exclude all external factors and concentrate on the differences that occur when the plasma is completely or partly separated from conducting surfaces. We also showed that the VDBD in LTP configuration systematically created a more efficient plasma as measured by emission intensity and current. At the same time the VDBD in DBDI configuration performed better in combination with a MS, giving much larger signals and better signal-to-noise ratios. The reason for this difference could partly be found and explained by the fact that the condition of the grounded electrode in the discharge seems to play a major role for the MS signal quality. Contrary to general assumption, it greatly mattered if this electrode was connected to a grounded mass or not. While a disconnected floating electrode leads to generally weaker plasma for the VDBD in both cofigurations, the MS signal intensity increased by more than an order of magnitude for the VDBD in LTP configuration and therefore became comparable to that obtained by the VDBD in DBDI configuration. The reason for this strong influence of the grounded or floating mass electrode cannot be fully explained. A reasonable explanation might be a shifting of internal discharge potentials to the floating electrode that changes the transport and recombination properties of charged species in the discharge. Further studies would be necessary to clear up these findings that would far exceed the scope of this work. The comparison between half and full DBD done in this work shows a noticeable systematic difference between the discharge types and also might help to explain why certain

* E-mail: [email protected] Phone: +49 231 1392-174 Fax: +49 231 1392-120

ASSOCIATED CONTENT Supporting Information animated version B (.m1v) animated version C (.m1v) Supporting Information (SI) (.pdf)

AUTHOR INFORMATION Corresponding Author

ACKNOWLEDGMENT The financial support by the Ministerium für Innovation, Wissenschaft und Forschung des Landes Nordrhein-Westfalen, the Senatsverwaltung für Wirtschaft, Technologie und Forschung des Landes Berlin, the Bundesministerium für Bildung und Forschung, and the Deutsche Forschungsgemeinschaft is gratefully acknowledged. This project is partly supported by the DFG (German Science Foundation) within the SFB-TR 87. This work has been supported in part by the Croatian Science Foundation under the project No. 2753.

REFERENCES (1) Harper, J. D.; Charipar, N. A.; Mulligan, C. C.; Zhang, X. R.; Cooks, R. G.; Ouyang, Z. Anal. Chem. 2008, 80, 9097-9104. (2) Na, N.; Zhang, C.; Zhao, M. X.; Zhang, S. C.; Yang, C. D.; Fang, X.; Zhang, X. R. J. Mass Spectrom. 2007, 42, 1079-1085. (3) Hayen, H.; Michels, A.; Franzke, J. Anal. Chem. 2009, 81, 10239-10245. (4) Meyer, C.; Muller, S.; Gilbert-Lopez, B.; Franzke, J. Anal. Bioanal. Chem. 2013, 405, 4729-4735. (5) Garcia-Reyes, J. F.; Mazzoti, F.; Harper, J. D.; Charipar, N. A.; Oradu, S.; Ouyang, Z.; Sindona, G.; Cooks, R. G. Rapid Commun. Mass Spectrom. 2009, 23, 3057-3062. (6) Almasian, M. R.; Yang, C. D.; Xing, Z.; Zhang, S. C.; Zhang, X. R. Rapid Commun. Mass Spectrom. 2010, 24, 742-748. (7) Albert, A.; Engelhard, C. Anal. Chem. 2012, 84, 10657-10664. (8) Kogelschatz, U.; Eliasson, B.; Egli, W. J. Phys. IV 1997, 7, 4766. (9) Klute, F. D.; Michels, A.; Schütz, A.; Vadla, C.; Horvatic, V.; Franzke, J. Anal. Chem. 2016, 88, 4701-4705. (10) Klute, F. D.; Schütz, A.; Michels, A.; Vadla, C.; Veza, D.; Horvatic, V.; Franzke, J. Analyst 2016, 141, 5842-5848. (11) Hopfield, J. J. The Astrophysical Journal 1930, 72, 133. (12) Smith, A. L.; Meriweth.Jw. J Chem Phys 1965, 42, 2984-&.

Insert Table of Contents artwork here



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment

Page 8 of 8

8

A systematic comparison between half and full dielectric barrier

Recommend Documents