SPECTROMETRY

TECHNICAL FIELD

This disclosure relates to plasma ionization mass spectrometry. In particular it relates to a method and apparatus for dielectric barrier discharge ionization for mass spectrometry.

BACKGROUND ART

Ambient ionization mass spectrometry (MS) is useful for rapid and direct analyses of samples in their native in situ environment, without sample preparation and chromatography. Most ambient ionization methods can be categorized as laser-, spray-, or plasma-based approaches. Plasma sources are advantageous in that both low- and high-polarity small molecules can be detected almost simultaneously by MS. In addition, plasma sources tend to be less susceptible to ion suppression than spray or laser-based methods. Thus, such ion sources are widely used in a range of ambient MS applications. For example, ambient plasma ionization can be used for the rapid detection of drug metabolites, agrochemicals, and in forensics, chemical warfare simulants and emerging environmental contaminants. Plasma ionization MS can also be used to spatially image the distribution of metabolites in live plant and animal tissues. The duration of each pulse can be approximately 500 ns when the gas phase plasma is deprotonated dimethylmethylphosphonate.

An interval between successive pulses can be of the order of micro-seconds (µs). More particularly, the interval between success pulses can be approximately 900µs.

The pulses can form or approximate a square waveform. An amplitude of the pulses can be between approximately 500 to 5000 volts.

The conduit, the first electrode, and the second electrode, can be co-axial with each other. The analyte material can be a liquid analyte, wherein the liquid analyte and a desolvation gas for dissolving said liquid analyte are introduced between the first and second electrodes. The method can comprise delivering said desolvation gas into said conduit, and injecting said liquid analyte into said conduit The desolvation gas can be delivered via a tube, and the liquid analyte can be injected into between the first and second electrodes via a capillary. The tube and the capillary can be arranged at an angle to each other. The angle can be approximately 45 degrees. The capillary can be angled from conduit. The tube can be in axial alignment with the conduit. In a second aspect, the present invention provides an electrical circuit for providing a series of voltage pulses to provide dielectric barrier discharge ionization. The electrical circuit comprises: a high voltage direct current supply; a switching circuit which switches an output of the electrical circuit between the high voltage direct current supply and a low voltage direct current supply; and a command signal source adapted to supply a command signal to the switching circuit, to control switching of the switching circuit. The series of voltage pulses are each of a duration of a nano-second order, there being an interval between successive pulses which is of the order of micro-seconds. The duration of each pulse can be in the range approximately from 100 to 900 ns. The switching circuit can comprise two push-pull switching elements. The switching elements can be solid state components which are synchronously triggered by respective galvanically isolating driver circuits. The high voltage direct current supply can be provided a portable power source which supplies to a high voltage direct current to direct current converter. The lower voltage direct current supply can be provided a portable power source which supplies to a low voltage direct current to direct current converter. An amplitude of the pulses can be between approximately 500 to 5000 volts. A ground of a low voltage direct current supply which supplies the switching circuit can be connected to a ground of the high voltage direct currently supply via a low inductance grounding bar. In the method mentioned above in respect of the first aspect, the series of voltage pulses can be provided by the circuit mentioned in respect of the second aspect. In a third aspect, the present invention provides an apparatus for supplying plasma to a mass spectrometer, including the circuit mentioned above in respect of the second aspect; and a dielectric barrier discharge ionization apparatus comprising an inner electrode, an outer electrode, and a dielectric barrier there between, the dielectric barrier having an inlet for receiving a plasma for analysis, and an outlet for supplying the plasma to the mass spectrometer. An output of the circuit is provided to the inner electrode or the outer electrode of the dielectric barrier discharge ionization apparatus. In a fourth aspect, the present invention provides an apparatus for supplying plasma to a mass spectrometer, wherein the plasma is ionized using the method mentioned above in respect of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS Embodiments will now be described by way of example only, with reference to the accompanying drawings in which: Figure 1 is a schematic showing a DBDI source driven by a circuit which generates nano-second pulses in accordance with one embodiment of the present invention; Figure 2 is a schematic showing an active capillary DBDI source in accordance with one embodiment of the present invention; Figure 3 is a schematic showing a DBDI source driven by a circuit which generates nano-second pulses, showing the circuit components; Figure 4(a) schematically depicts a waveform of the microsecond pulses used in the prior art; Figure 4(b) depicts a waveform of the preferred pulses in accordance with the present invention; Figure 5(a) is a plot of a nanosecond pulsed DBDI mass spectra of methanol solution containing 100 nM of dimethylmethylphosphonate; Figure 5(b) is a plot of a microsecond pulsed DBDI mass spectra of methanol solution containing 100 nM of dimethylmethylphosphonate; Figure 6(a) is a voltage vs time plot for nanosecond pulsed plasma generated using a frequency of 1.1 kHz and a pulse width of 800 ns; Figure 6(b) is a current vs time plot for the nanosecond pulsed plasma mentioned in Figure 6(a); Figure 6(c) is a power vs time plot for the nanosecond pulsed plasma mentioned in Figure 6(a); Figure 6(d) is a voltage vs time plot for microsecond pulsed plasma generated using a frequency of 10 kHz and a pulse width of 50 µs; Figure 6(e) is a current vs time plot for the plasma mentioned in Figure 6(d), where the blue traces correspond to vertical expansions of 20×; Figure 6(f) is a power vs time plot for the plasma mentioned in Figure 6(d), where the blue traces correspond to vertical expansions of 10×; Figure 7 depicts chemical structures of (a) dimethylmethylphosphonate (DMMP), (b) dichlorvos (DDVP), (c) 3-octanone, (d) m-cresol, (e) pentadecafluorooctanoic acid (PFOA), (f) benzylamine (BzA), (g) 4-methoxybenzylamine (4-BzA-OCH ₃ ), (h) 4-methylbenzylamine (4-BzA-CH ₃ ), (i) 4-clorobenzylamine (4-BzA-Cl), (j) 4- fluorobenzylamine (4-BzA-F) and (k) 4-(trifluoromethyl)benzylamine (4-BzA- CF ₃ ); Figure 8(a) is a plot of nanosecond pulsed mass spectrum of vapor from ~1 µL droplets containing 13 pg of dimethyl methylphosphonate; Figure 8(b) is a plot of nanosecond pulsed mass spectrum of vapor from ~1 µL droplets containing 22 ng dichlorovos; Figure 8(c) is a plot of nanosecond pulsed mass spectrum of vapor from ~1 µL droplets containing 13 ng 3-octanone; Figure 8(d) is a plot of nanosecond pulsed mass spectrum of vapor from ~1 µL droplets containing 11 ng benzylamine; Figure 8(e) is a plot of microsecond pulsed mass spectrum of vapor from ~1 µL droplets containing 13 pg of dimethyl methylphosphonate; Figure 8(f) is a plot of microsecond pulsed mass spectrum of vapor from ~1 µL droplets containing 22 ng dichlorovos; Figure 8(g) is a plot of microsecond pulsed mass spectrum of vapor from ~1 µL droplets containing 13 ng 3-octanone; Figure 8(h) is a plot of microsecond pulsed mass spectrum of vapor from ~1 µL droplets containing 11 ng benzylamine; Figure 9(a) is a plot of nanosecond pulsed DBDI mass spectrum of vapor from ~1 µL droplets containing 11 ng of m-cresol, wherein the blue traces correspond to a vertical expansion of 10×; Figure 9(b) is a plot of microsecond pulsed DBDI mass spectrum of vapor from ~1 µL droplets containing 11 ng of m-cresol, wherein the blue traces correspond to a vertical expansion of 10×; Figure 9(c) is a plot of nanosecond pulsed DBDI mass spectrum of vapor from ~1 µL droplets containing 3 ng perfluorooctanoic acid, wherein the blue traces correspond to a vertical expansion of 10×; Figure 9(d) is a plot of microsecond pulsed DBDI mass spectrum of vapor from ~1 µL droplets containing 3 ng perfluorooctanoic acid, wherein the blue traces correspond to a vertical expansion of 10×; Figure 10 show plots of signal-to-noise ratio (S/N) of protonated DMMP in nanopulsed DBDI-MS as a function of (a) frequency and (b) pulsewidth (with a frequency of 1.1 kHz); Figure 11 shows plots of signal-to-background chemical noise (S/N _C ) ratio of protonated DMMP in nanopulsed DBDI-MS as a function of (a) frequency and (b) pulsewidth (with a frequency of 1.1 kHz); Figure 12 shows plots of absolute ion abundance of protonated DMMP in nanopulsed DBDI-MS as a function of (a) frequency and (b) pulse width (with a frequency of 1.1 kHz); Figure 13 shows plots of average background ion signal of protonated DMMP in nanopulsed DBDI-MS as a function of (a) frequency and (b) pulse width (with a frequency of 1.1 kHz); Figure 14 shows nanopulsed (a, b, c) and micropulsed (d, e, f) DBDI mass spectra of 25 pg DMMP in methanol (a, d), water (b, e), and blood plasma (c, f). Inset are the respective control mass spectra of unspiked blood plasma and the asterisks indicate peaks above 2% of ion abundance other than precursor and fragment ions. Figure 15 shows a schematic arrangement of sample introduction and ionization source for the analysis of aldehydes, alkenes, and ketones by nanosecond pulsed plasma ionization mass spectrometry. Figure 16(a) shows the nanosecond pulsed DBDI mass spectrum of acetone, * indicates the formation of protonated dimer ([2M+H] ⁺ ). Figure 16(b) shows the nanosecond pulsed DBDI mass spectrum of 2-butanone. * indicates the formation of protonated dimer ([2M+H] ⁺ ). Figure 16(c) shows the nanosecond pulsed DBDI mass spectrum of 2-pentanone. * indicates the formation of protonated dimer ([2M+H] ⁺ ). Figure 16(d) shows the nanosecond pulsed DBDI mass spectrum of p-xylene. Figure 16(e) shows the nanosecond pulsed DBDI mass spectrum of perillaldehyde. † indicates the formation of [M+NH ₄ ] ⁺ . Figure 17 shows the nanosecond pulsed DBDI mass spectrum of caffeine ([M+H] ⁺ ) from instant coffee sample (solid) by nanosecond pulsed plasma ionization mass spectrometry. DETAILED DESCRIPTION In the following detailed description, reference is made to accompanying drawings which form a part of the detailed description. The illustrative embodiments described in the detailed description, depicted in the drawings and defined in the claims, are not intended to be limiting. Other embodiments may be utilised and other changes may be made without departing from the spirit or scope of the subject matter presented. It will be readily understood that the aspects of the present disclosure, as generally described herein and illustrated in the drawings can be arranged, substituted, combined, separated and designed in a wide variety of different configurations, all of which are contemplated in this disclosure. In the prior art, high voltage alternative current (AC) pulses applied to DBDI are microsecond pulses. Typical AC pulse generating circuits are suited to the generation of microsecond pulses. Also, the pulse length is considered to be sufficient for generating enough ions of an analyte, for detecting the presence and measuring the concentration of the analyte by mass spectrometry. The present invention provides a novel method and apparatus for DBDI, where AC pulses lasting on the order of nanoseconds, orders of magnitude shorter than typical microsecond pulses, are used. Nanosecond pulsed discharges have not been used previously in mass spectrometry, such pulses have become increasingly used in air and water purification, ozone synthesis, surface sterilization, materials processing and biomedical applications. Nanosecond pulsed discharges efficiently couple energy into the form of plasma by significantly reducing the pulse-width of high voltage waveforms compared to microsecond pulsed plasmas and thus, lower power consumption. Moreover, nanosecond pulsed glow discharges are capable of producing plasmas while generating less heat in optical emission spectrometry owing to the narrow pulse-width of the applied waveforms. In this invention, it is recognised that high-voltage nanosecond pulsed discharges have potential for the efficient formation of intact ions in DBDI-MS with a reduction in the formation of background ions in comparison to the use of more conventional microsecond high voltage pulses. However, existing circuitry for generating nanosecond pulses for the other applications (e.g., air and water purification, ozone synthesis, surface sterilization, materials processing, biomedical applications), are not suitable for DBDI. For instance these existing circuits may not be able to generate, or may not be optimal for generating, the required square waveform. They are also designed for different loads than what would be required for the plasma source in DBDI. Disclosed in this specification is the development, extended use, and application of nanosecond pulsed plasma ionization in mass spectrometry. Figure 1 schematically depicts an apparatus 100 to generate DBDI (i.e., a DBDI source) for MS. The apparatus 100 includes an input generator 102 and a DC high voltage source 104. The input generator 102 is adapted to generate nanosecond pulses at amplitudes suitable for plasma discharge ionisation. The input generator 102 is connected to a first electrode 106, in the form of a tube electrode, and a second electrode 108 which is located in the tube electrode 106. The tube electrode 106 is provided around a capillary 110 in which the plasma to be analysed is located. The second electrode 108 is provided inside the capillary 110. The capillary 110 is in communication with the inlet to the mass spectrometer (not shown), where vacuum suction will remove the ionized particles from the capillary toward the mass spectrometer. Figure 1 shows a gasket 112 provided at the interface between the capillary 110 and the mass spectrometer inlet 114 to minimise leakage. A shorter pulse is expected to generate a smaller amount of ionization. However, unexpectedly, when applied in DBDI-MS, this approach was found to significantly improve the signal-to-noise ratio (S/N), sensitivity and detection limits for some analytes compared to using microsecond pulses. In one experiment, six analytes were selected as test analytes of relevance to applications in forensics (Dimethylmethylphosphonate, or “DMMP”), oenology and viticulture (3-octanone and m-cresol), environmental (perfluoro"octanoic acid) and agrochemical (dichlorvos) analysis. In addition, six benzylamines were used to demonstrate that the extent of internal energy deposition in nanosecond plasma ionization is exceptionally low and comparable to microsecond plasma ionization. Thus, this nano-pulsed plasma method is promising for improving the detection of many types of molecules, including those with labile bonds, in dielectric barrier discharge ionization mass spectrometry. Dimethylmethylphosphonate (DMMP), dichlorvos (DDVP), 3-octanone, m- cresol, pentadecafluorooctanoic acid (PFOA), benzylamine (BzA), 4- methoxybenzylamine (4-BzA-OCH3), 4-methylbenzylamine (4-BzA-CH3), 4- clorobenzylamine (4-BzA-Cl), 4-fluorobenzylamine (4-BzA-F), and 4- trifluoromethylbenzylamine (4-BzA-CF3) (Figure S1) were obtained from Sigma Aldrich (Missouri, USA) and used without further purification. These were used as both test analytes and as ‘thermometer’ ions to determine the extent to which molecules with labile bonds can be detected as intact ions by nanosecond pulsed plasma ionization mass spectrometry. Sample solutions were prepared using methanol (HPLC grade; Sentmenat, Spain) by dilution to the desired concentrations (in the experiments, ranging from 13 pg µL-1 to 22 ng µL-1). For thermometer ion experiments, 4-BzA-OCH3, 4–BzA- CH3, 4-BzA-Cl, 4-BzA-F, BzA and 4-BzA-CF3 were diluted in methanol as a mixture at a concentration of 11 ng µL-1 each to evaluate the internal energy deposition by microsecond and nanosecond pulsed plasma ionization. To establish calibration curves, DMMP was diluted into three different matrices: (i) methanol; (ii) deionized water (18 MΩ; Merck Millipore, Victoria, Australia); and (iii) human blood plasma (Australian Red Cross Blood Service) at concentrations of 2.48 to 124 pg µL ^-1 . As shown in Figure 1, the input generator 102 can be coupled with a conventional mass spectrometer. The input generator 102 receives an input voltage from a DC voltage source 104. A DC voltage 105, may also be applied to the input generator 102, to supply the switching circuit. The DC supply and power the circuit at 5V or another voltage. The experiments conducted on the above-mentioned analytes were performed using a linear quadrupole mass spectrometer (LTQ XL, Thermo Fisher Scientific, San Jose, CA, USA) that was equipped with a custom electrodynamic ion funnel (Heartland Mobility, Wichita, KS, USA) and with either an active capillary DBDI (Figure 1) or an electrospray ionization (ESI) source. The active capillary 110 used in the DBDI may be a quartz capillary. In the experiments, quartz capillary with an inside diameter of 1.50 millimetres (mm), an outside diameter of 1.80 mm, and a length of 2 centimetres (cm), from VitroCom, NJ, USA, was used. The capillary 110 is mounted in axial alignment with the capillary entrance to the mass spectrometer. However the capillary 110 may be constructed from another dielectric material to provide the dielectric barrier between the outer electrode 106 and inner electrode 108. The first, or outer, electrode 106 may be a cylindrical electrode. The outer electrode 106 is made from a conductive material such as copper. In the experiments, a copper cylindrical electrode with an inside diameter of 1.7 mm, an outside diameter of 2.5 mm, and a length of 1.5 cm, from GoodFellow Cambridge, UK, was used. The second, or inner, electrode 108 can be provided by a grounded stainless steel wire of 0.30 mm in diameter. However, another conductive material may be used. The grounded stainless wire 108 was inserted 0.75 cm into the quartz capillary 110. Ionization was initiated and maintained by applying a continuous high frequency and high voltage waveform between the inner and outer electrodes 108, 106 across the dielectric barrier using either nanosecond or microsecond pulse widths (see below). The embodiment shown in Figure 1 is suited for the analysis of gas plasma samples. Figure 2 depicts an alternative embodiment, which is modified for the analysis of liquid samples. The modified arrangement includes an active capillary dielectric barrier discharge ion (DBDI) source 130. The active capillary dielectric barrier discharge ion (DBDI) source 130 comprises an inner cylindrical electrode 132, a dielectric barrier 134 provided around the outside perimeter of the inner electrode 132, and an outer electrode 136 provided around the outside perimeter of the dielectric barrier 134. The inner electrode 132, dielectric barrier 134, and the outer electrode 136, are provided in co-axial alignment, that is, sharing the same axis. By way of example only, the following materials and dimensions are provided. The inner electrode 132 can be a cylindrical copper (Cu) inner electrode, with an inside diameter of 1.7 mm, an outside diameter of 2.5 mm, and a length of about 1.5 cm. The dielectric barrier 134 can be a plastic heat-shrink tube, with an inside diameter of 3.0 mm, an outside diameter of 3.4 mm, and a length of 1 cm. The outer electrode 136 can be provided by a stainless steel coiled wire, where the coil has an inside diameter of 3.5 mm, an outside diameter of coil 3.7 mm, and a length of coil 0.7 cm. The diameter of the wire itself may be 0.1 mm. On one end, the inner electrode is positioned to direct the ions toward the mass spectrometer (MS) inlet 114. This is the MS inlet end 131. On the opposite end, the inner electrode 132 is positioned into the dielectric barrier 134, which is in turn positioned partially within the coil (i.e., outer electrode) 136. In the example with the above-mentioned dimensions, the inner electrode 132 is positioned into the dielectric barrier 134 by about 0.7 cm, i.e., they overlap by about 0.7 cm. The inner electrode 132 overlaps with the coil, by about 0.5 cm. Therefore, the coil 136 is positioned farther from the MS inlet end 131 of the inner electrode 132 than the dielectric barrier 134. The MS inlet end 131 of the inner electrode 132 is mounted, in axial alignment, with the capillary entrance 114 to the mass spectrometer using a capillary 138, which may be a quartz capillary having an inside diameter of 1.50 mm, an outside diameter of 1.60 mm, and a length of 9 mm. A gasket fitting 140, such as a Teflon gasket fitting, is provided at the MS inlet end 131 of the inner electrode 132 to provide a seal. High voltage alternating current waveforms are applied between the inner electrode 132 and the outer electrode 136 to generate a continuous plasma between the dielectric barrier 134 and the inner electrode 132. The liquid sample mixture is introduced or infused into the plasma through a capillary emitter 142. The capillary emitter 142 may be made from pulled borosilicate, and in one example has an inner diameter of 76 μm. The capillary emitter 142 is positioned at an angle θ to the axis of the inner electrode 132. This angle θ can be about 45º. The liquid sample is added to the plasma using a pump (not shown), such as a syringe pump. The pump may operate at a flow rate of 2 microlitres per minute (µL min ^-1 ). A continuous flow of a desolvation gas, such as N ₂ (g), is introduced via a tube 144 directed towards the plasma source 130, to facilitate desolvation of the liquid mixture introduced from the emitter 142. The gas flow may be roughly 600 to 900 millilitres per minute (mL min ^-1 ). The tube 144 is provided in axial alignment with the inner electrode 132. In experiments conducted, a gas flow rate of 900 mL min ^-1 resulted in optimal signal levels. The liquid sample may be instead provided in an axial direction toward the plasma source, and the desolvation gas provided at an angle θ, i.e., in opposite arrangement to what is shown in Figure 2. Angular arrangements between these two extremes may be considered. Generally, the desolvation gas and the liquid sample are provided toward the plasma source from directions which are at an angle of θ from each other. However, the directions are not required to be angled from each other. That is, θ can be zero degrees. In another embodiment, θ can be about 90 degrees, to avoid any excess liquid which does not vaporise clogging the plasma source. The above described arrangement allows for nanosecond pulsed plasma ionization by the direct injection of liquid solutions. The signal-to-noise ratio (S/N) of the mass spectrometer can be obtained by taking the ratio of the integrated abundance of the protonated analyte ion (S), and the recorded signal for the precursor ion, measured in m(mass)/z (ion charge), in the blank sample. The signal-to-background chemical noise (S/NC) can be obtained by dividing S by the average chemical background ion signal (NC). The latter value (N _C ) does not include the abundance of the precursor analyte ion or any associated fragment ions. According to this embodiment of the invention, to generate nanosecond pulsed plasmas, a series of high voltage pulses of about 800 ns in duration, with a pulse- to-pulse delay (i.e., interval) of about 900 µs was applied. For comparison, in an experiment, microsecond pulsed plasmas were also generated by applying a high frequency and high voltage waveform with pulse widths of about 50 μs and delays of about 50 μs between pulses. To detect positive ions such as protonated dimethylmethylphosphonate (DMMP) by nanosecond pulsed plasma ionization, the high voltage nanosecond pulses were applied to the inner electrode, and the outer electrode was grounded. In the experimental set-up used, in the microsecond pulsed plasma, the switching of polarity for different ionization modes was not needed since the applied microsecond pulses were bi-directional. To investigate the performance of the nanosecond pulsed plasma ionization by direct injection of liquid solutions, a methanol solution containing a 100 nano- molar (nM) DMMP was prepared, and nanosecond pulsed DBDI mass spectra was obtained. For comparison, microsecond pulsed DBDI mass spectra were obtained for the same solution while keeping the solution injection system and ion source geometry constant. As shown in Figure 5(a), by use of nanosecond pulses, an ion corresponding to protonated DMMP can be readily detected in relatively high abundance (ion counts of about 5×10 ⁴ ) compared to that obtained by using microsecond pulsed plasma ionization (about 3.5×10 ⁴ , see Figure 5(b)). Both spectra indicate the formation of a product ion (m/z 111), which corresponds to a loss of CH ₂ from protonated DMMP due to ion activation in the ion source and/or during ion transmission. The S/N and S/NC calculated for DMMP by nanosecond pulsed plasma ionization was 176 and 1490, respectively which corresponds to an increase in these values by 73% and 63% compared to those obtained from microsecond pulsed plasma ionization (S/N of 47 and S/NC of 550). These results suggest that nanosecond pulsed plasma ionization may significantly improve the analysis of samples in ‘online’ liquid chromatography coupled to nanosecond pulsed dielectric barrier discharge ionization mass spectrometry. Figure 3 depicts an embodiment of the DBDI apparatus 100 in more detail, showing the components of the input pulse generator 102 according to an embodiment of the invention. The input pulse generator 102 includes a function generator or microcontroller 202 and solid state electrical components (including MOSFET switches S1, S2, drivers 206, 208, and a logic controller 210) to output the voltage waveforms needed for DBDI. The function generator or microcontroller 202 provides a short pulse signal, as a command input for a high voltage nano-second pulse generator 204. The nano-second pulse generator 204 includes switches S1, S2 which are controlled to alternatively switch the output of the pulse generator between the high voltage DC supply 104 (“high rail” or “high side”), and a reference voltage or ground (“low rail” or “low side”). The switching element S1 is therefore the high-side switching element, and the other switching element S2 is the low-side switching element. A decoupling capacitor, C, is placed across the high-voltage half-bridge, which is provided by the two push-pull switching elements S1, S2. In the depicted embodiment, the decoupling capacitor C is placed across the high side switch S1 and connected to ground. The switching elements S1, S2 can comprise series- connected power metal oxide semiconductor field-effect transistors (MOSFETs), and triggered synchronously by galvanically isolating driver circuits 206, 208. The output 214 for the pulser circuit 102 is taken between the two series connected switches S1, S2. The command signal generated by the function generator or a microcontroller 202 is provided to a logic controller 210, which produces control signals for the isolated drivers 206, 208. In one embodiment, the logic controller 210 passes the command signal to one driver, and passes the command signal through an inverter 210 before passing it to the other driver. Thus the drivers 206, 208 will provide complementary switching signals for switches S1, S2. The isolated drivers 206, 208 can be provided by a dual channel driver 207. In one embodiment, when the high-side switch S1 is turned on and the low-side switch S2 is turned off, the output of the pulse generator is connected to the positive rail of the high voltage DC supply 104, to supply a high voltage to the plasma source. When the high-side switch S1 is turned off and the low-side switch S2 is turned on, the output of the pulse generator is connected to the ground and thus, no voltage is applied across the plasma electrodes 106, 108. The command waveforms to drive the switching enables a pulsed DC voltage to be applied. A low voltage power supply 105, e.g., a 5V supply, is provided to supply the active components in the nano-second pulse generator 204. The low voltage power supply 105 is connected to the input of an isolated DC/DC converter to provide isolated power supply 216, and is connected to the logic controller 210. The isolated power supply 216 generates two isolated voltages for drivers 206, 208 (or dual channel driver 207), which in turn drive the switches S1, S2 on a half-bridge. As explained above, because of the operation of the logic controller 210, the two switches S1, S2 operate in a complementary mode. That is, when S1 is on, S2 is off, and vice versa. However, in practice, when a semiconductor switch is switched from an “on” state to an “off state”, there is a slight delay before the switch completely turns off. Therefore, the switches S1, S2 cannot be controlled to switch on or off, precisely simultaneously. Otherwise, at the switch over times, both S1 and S2 will stay on momentarily. This can cause a “shoot through”, i.e., short-circuit through S1, S2 and the high-voltage DC supply. To avoid the “shoot through”, a dead time is inserted by the logic controller 210, to the control signals for the switches, between the S1 and S2 switch over. This dead time is essentially an additional delay time, during which both S1 and S2 are off. In one implementation, the low to high (off to on) transition for each control signal will be delayed by the “dead time”, compared with the command input from the function generator. This ensures that one of the switches will be turned off first as the “dead time” periods starts, and the other switch will only be turned on after dead time is expired. The dead time will therefore be chosen on the basis of the time which the switch takes to turn off. The logic controller 210 generates two PWM signals to the drivers 206, 208 (or dual channel half bridge driver 207), which would be complementary but for a dead time inserted at each low to high transition, for both signals. For improved noise immunity, the ground of the low voltage DC supply for the active circuit components in the pulser circuit 204 should be connected with a low inductance grounding bar to the ground of the high-voltage side of the pulser. To detect positive ions using a nanosecond pulsed plasma, a potential gradient is applied between the inner and outer electrodes by applying the pulsed high DC voltage to the outer electrode 106 and grounding the inner electrode 108. The pulsed voltage will vary between the high voltage and ground. For negative ions, the polarity configuration is reversed, and the amplitude of the voltage may also be different depending on the analytes. The DC voltage used is preferably in the range of approximately 1500 to 2500 volts (V) to detect ions in positive and negative modes. However voltages between about 500 V to about 5000 V may be used. The use of DC voltages outside this range resulted in relatively noisy waveforms. In Figure 3, the ionization pulse generator is connected to the arrangement for producing discharge ions from a gaseous plasma. It will be appreciated that it can instead be connected to supply ionization voltage pulses to the arrangement shown in Figure 2. The plasma voltage and current generated by high frequency square-wave generator and high voltage nanosecond pulses may be measured by high voltage differential probe and a current sensing resistance component 212 respectively, using an oscilloscope. The current drawn from the high voltage DC supply by the nanosecond pulsed plasma source is roughly 0.15 microamperes (μA). The resistance component 212 may be provided by a single resistor or two or more resistors connected to provide resistance. The resistance depends on the amount of current being measured. Large resistance values will affect the rising and falling times of the output nano-second pulse voltage waveform. Therefore, any limitation on the acceptable rising and falling times for the waveform will also limit the resistance used. A value between 10 and 200 ohms are typically used. In a particular embodiment, the resistance component 212 has a resistance of 165 ohms. Figure 4(a) depicts a waveform of the microsecond pulses used in the prior art. Figure 4(b) depicts a waveform of the preferred pulses in accordance with the present invention. The prior art waveform includes pulses which are on the order of, for example, 50 microseconds (µs) in duration and which are spaced apart by approximately, for example, 50 µs. The waveform of the preferred pulses according to the invention include much shorter pulses, of between 100 to 900 nanoseconds (ns) in duration, which are spaced apart by about 900 µs. Different pulse durations may be used depending on the analyte to be detected. In practice, a high voltage pulser may be used to provide the switching. An example which may be used is a pulser (FSWP 51-02) from Behlke®. The set up can be modified into a portable version. For instance, the high voltage power supply for the pulser circuit 102 can be provided by a portable power supply such as a battery (or battery and charger), which outputs power to a high voltage DC/DC converter. The portable supply can also output power to a low voltage DC/DC converter, to provide the low voltage power supply for the pulser circuit 102. The high and low voltage DC/DC converters will need to have the specification suitable for the voltages required by the application. The command signal for the pulser 102 may be more preferably provided using a microcontroller instead of a function generator in portable embodiments. The low voltage power supply can also supply other components needed, such as the function generator or the microcontroller which provides the command for the pulser circuit. The current drawn from the high voltage DC supply by the nanosecond pulsed plasma source was measured to be approximately 0.15 μA. For optimal ion signal, the average amplitude of the input voltage pulses applied to the nanosecond pulsed plasma was about 1.5 kilo-volts (kV). The measured plasma source voltage and current are shown in Figure 6(a) and Figure 6(b). For a measured plasma voltage of about 1.5 kV (see Figure 6(a)), the plasma current was measured to peak at values of +110 and -110 mA, at the rising edge and falling edges of the plasma voltage waveform, respectively (see Figure 6(b)). The instantaneous power consumption of the plasma can be obtained by multiplying the plasma voltage and current (see Figure 6(c)). The average power consumption of the plasma source can be obtained by integrating the instantaneous power consumption across an entire waveform period, using the below equation , where T is the period of the plasma voltage, and v(t) and i(t) are the plasma voltage and current as a function of time (t). The instantaneous output power of the nanosecond pulser is mainly composed of two peaks at rising and falling edges of the plasma voltage (see Figure 6(c)). The difference in the instantaneous power at the rising and falling edges of the voltage waveform is power dissipated by the actuator during each complete cycle of the pulse voltage. For the rising edge, the amplitude of the plasma current is higher than 0.1 A, which indicates that the electrical power is drawn from the power supply during this period, which energizes the electrical discharge for the plasma ionization. However, the electrical power consumption is below 0.1 A at the falling edge. Here, the plasma current changes polarity, indicating that power is induced by the electric field due to the space charges remaining at the dielectric surface during positive discharge. The average power consumption to generate the nanosecond pulsed plasma by the nano-second pulser was calculated to be 6.05 ± 0.09 mW under these conditions, based on the integration of the instantaneous power consumption. The quiescent power dissipation from the nano-second pulse generator by operating the system in open circuit was also measured. The current drawn from a 5 V auxiliary DC supply by the circuit was 69 mA, which yielded a power loss of 0.35 W mainly from the control circuit, indicators and gate drivers of the nanosecond pulser module 204. Thus, the power drawn from the function generator and high voltage DC power supply is negligible compared to the quiescent power dissipation of the nano-second pulse generator. As the total power consumption required for the entire setup is 0.35 W, the control circuit can be supplied by a 9 V PP3 (1.2 Ah) battery and the high voltage supply can in principle be replaced by a 9 V to 2.5 kV DC-DC converter. At the maximum, this should allow for more than 12 hours of continuous operation. As mentioned above, the power consumption of the nanosecond pulsed plasma under these conditions is 0.35 watts (W), while the power consumption required by conventional microsecond plasma ionization can be 2.1 W. Thus nanosecond pulsed plasma ionization can be ~83% more power efficient than microsecond plasma ionization. In the experiment, the voltage, current and power consumption using microsecond pulsed plasma ionization (Figures 6(d), 6(e), 6(f)), were also measured. The instantaneous power consumption by microsecond plasma is lower than that for nanosecond plasma using an input voltage that is 500 V lower (see Figure 6). However, the average power consumption that is required for the microsecond plasma source (2.1 W) was a factor of 6 higher than for nanosecond plasma. An input voltage of ~1 kV was optimal for the formation of protonated DMMP under these conditions (data not shown). The average power consumption is lower using nanosecond plasma likely because of the higher switching frequency of microsecond plasma (10 kHz vs. 1.1 kHz) which results in higher switching loss compared to nanosecond pulsed plasma ionization. Thus, nanosecond pulsed plasma ionization can be more power efficient than conventional microsecond plasma ionization by ~380% under these conditions. Nanosecond pulsed plasma ionization can increase the signal-to-noise ratio. To investigate the use of nanosecond pulses in DBDI-MS, mass spectra were obtained for DMMP, dichlorvos (DDVP), 3-octanone, and benzylamine (BzA) (chemical structures shown in Figure 7) that were introduced into the source by vapour pressure from 13 pg to 22 ng of analyte contained in ~1 µL droplets of methanol that were positioned ~3 mm below the plasma. The results are shown in Figure 8 (a to d). For comparison, microsecond pulsed DBDI mass spectra were obtained for the same solutions (see results in Figure 8 e to h). Using nanosecond pulsed plasma for each analyte, ions corresponding to the protonated molecules can be readily detected in relatively high abundances (ion counts of 1.2×10 ⁵ to 4.8×10 ⁶ ) and reasonably high signal-to-noise and signal-to-background chemical noise ratios. The S/N, S/NC data are shown in Table 1 below. In the table, the “mode” refers to the ionization mode, i.e., either protonated (+) or deprotonated (-) ions were detected.

TABLE 1 The signal-to-background chemical noise corresponds to the ratio between the analyte signal and total signal for background chemical noise, not including fragment ions. For protonated benzylammonium (a ‘thermometer ion’), a fragment ion was formed in relatively low abundance relative to the intact precursor ion (i.e. ~20% of precursor ion), which corresponds to the loss of an ammonia molecule partly owing to the relatively low dissociation energy of the C-N bond (163.4 kilo-joules per mole (kJ mol ^-1 ). The absolute signal for protonated DMMP, DDVP, 3-octanone, and BzA ions upon formation by nanosecond pulsed plasma are slightly to moderately higher (by 13 to 56%) than that obtained by using microsecond pulses. The signal-to- noise ratio for these ions formed by nanosecond pulses are 29.7 to 58.1% higher than that obtained using the microsecond plasma (see Figure 8 and Table 1). In addition, the S/N _C values also increased by 20.3 to 58.2% using the nanosecond compared to the microsecond pulsed plasma (see Table 1). This is attributable to the microsecond plasma generating higher background chemical noise than the nanosecond plasma (see Figure 8 and Table 2 below) possibly resulting from the wider pulsewidth of microsecond plasma waveforms (~50 µs) compared to that for the nanosecond plasma (~800 ns). That is, the microsecond high voltage pulses were ‘on’ for a duration which was 125 times longer than the nanosecond pulses, resulting in the formation of background ions in higher abundances, than for the nanosecond pulses. TABLE 2 To investigate the formation of negative ions, mass spectra were obtained of vapour sampled from ~1 µL methanol solutions containing 10.9 nano-grams (ng) of m-cresol and 3.0 ng of pentadecafluorooctanoic acid (PFOA) (see Figure 8) by nanosecond and microsecond pulsed DBDI. The results are shown in Figure 9. For m-cresol and PFOA, ions corresponding to the deprotonated molecules can be readily detected in the same or slightly higher abundances by using nanopulsed than by micropulsed plasmas. Although the absolute signal obtained for deprotonated m-cresol and PFOA using the two different types of plasma are comparable, the S/N and S/NC ratios that were obtained using the nanoplasma were a respective 42.9 to 62.8% and 31.0 to 45.3% higher than that for the microplasma owing to the higher background chemical noise (see Figure 9 and Table 2). For both deprotonated analytes, a primary fragment ion is formed in relatively low abundance (see Figure I), which indicates that ions can be activated during ion formation under these conditions. For m-cresol, an ion corresponding to deprotonated phenol (formed by the loss of CH ₂ from deprotonated m-cresol) is detected at m/z 93 using both nanopulsed and micropulsed plasmas (relative abundances of 13% and 24% to the precursor, respectively). The deprotonated PFOA ion can readily lose CO ₂ to form a decarboxylated ion corresponding to [PFOA–CO2] ^– , owing to ion activation in the ion source and/or during transmission, which is consistent with previous reports for the detection of PFOA in DBDI-MS and electrospray ionization MS. For the microsecond plasma, the decarboxylated product ion is formed in a relative abundance of 37% of the precursor ion compared to 19% for the nanopulsed plasma; i.e. these results are consistent with the extent of internal energy deposition in nanosecond plasma being lower than that in the microsecond plasma under these conditions. Overall, these data indicate that nanosecond pulsed plasma can be used to effectively ionize different analytes in both positive and negative ionization modes. Effects of the frequency and pulse widths of nanosecond pulsed DBDI-MS To investigate the effects of the waveforms used in nanosecond pulsed DBDI on performance, mass spectra of DMMP vapour and the corresponding S/N and S/N _C ratios for protonated DMMP were obtained as a function of the frequency (0.50 to 1.50 kHz) and the widths of the high voltage pulses (100 to 900 ns). Figure 10, Figure 11, and Figure 12 show the effect of the pulse frequency and pulse width, on S/N, S/N _C , and absolute abundance, respectively. For DMMP, both S/N, S/N _C and the absolute ion signal increased when the pulse frequency was in the range from approximately 0.5 to approximately 1.1 kHz. They decreased the frequency is increased further (see Figure 10, Figure 11, and Figure 12). Although the background chemical noise increased slightly when the pulse frequency was increased from 0.5 to 1.25 kHz (Figure 13), the ‘optimal’ S/N and S/NC value is 1.1 kHz. Given that the background chemical noise increases with frequency from 0.5 to 1.25 kHz (see Figure 13), the decrease in absolute ion signal for the analyte by using pulse frequencies higher than 1.1 kHz may partially arise from the fixed capacity in the ion trap, which is filled to a greater extent with background ions at higher frequencies than at lower frequencies. In addition to frequency, the other major parameter of the DBDI-MS ion source that can be optimised is the width of the nanosecond pulses. The S/N, S/N _C and absolute signal for DMMP increased steadily as the pulse width increased from 100 ns to 800 ns, and decreased slightly for 900 ns (see Figure 10, Figure 11, and Figure 12). The background chemical noise was low (~8×10 ² ion counts) and did not depend significantly on the pulse widths under these conditions (see Figure 13). The slight decrease in S/N, S/NC and absolute ion signal of DMMP with increasing pulse widths may arise from the limited capacity of the ion trap and space charge effects on ion transmission. Overall, these results indicate that the frequency and pulse widths can substantially impact the performance of nano- pulsed plasma DBDI-MS (i.e., using pulses with durations on the order of nano- seconds) for the detection of DMMP under these conditions. Rapid detection of DMMP directly from blood plasma by nanosecond pulsed DBDI-MS The performance of ion sources for the detection of analytes can be detrimentally impacted by matrix effects. To investigate matrix effects in DBDI-MS, DMMP was sampled from methanol, water and human blood plasma at a concentration of 25 picograms per microlitre (pg µL ^-1 ) using nanosecond and microsecond pulsed plasmas (Figure 14). For the water sample, relatively intense background chemical noise was generated using the microsecond plasmas compared to nanosecond plasma - see Figure N(b) and Figure 14(e). The water sample also generated peaks corresponding to formation of clusters by water [M+H+H2O] ⁺ and oxygen [M+H+O] ⁺ by both ionization techniques. Nanosecond and microsecond pulsed DBDI-MS of DMMP spiked into blood plasma resulted in the detection of DMMP as the base peak, and the formation of a relatively large number of lower abundance peaks compared to the use of methanol or water as a matrix (< 2% relative abundance; see Figure 14). By use of either nanopulsed or microsecond pulsed plasma, over 7 ions with an abundance more than 2% that of the base peak were detected under these conditions, which were also present in the spectra of blood plasma that was not spiked with DMMP and not in the blank water spectra; i.e. both DMMP and at least 7 unknown blood plasma metabolites can be detected under these conditions (Figure 14). For the detection of DMMP in ~1 µL drops of methanol, water and blood plasma, calibration curves are shown in Figure 5, in which log ₂ (absolute ion abundance) of protonated DMMP is plotted as a function of the log ₂ of the mass of DMMP (2.48 to 124 pg) that was sampled from the matrix solutions. A logarithmic scale was used to ensure that any differences in absolute ion abundance values at low analyte concentrations are clearly shown. For the linear regression ‘best fit’ calibration lines, the R ² values were above 0.998 indicating a high linearity for both ionization methods (see Table 3).

TABLE 3 The plasma matrix strongly affects the sensitivity and limit of detection (LOD) for both nanosecond and microsecond pulsed DBDI-MS. For example, in nanopulsed DBDI-MS, the sensitivity decreased in order of methanol (251.78 ng ^-1 ), water (26.45 ng ^-1 ) and blood plasma (13.99 ng ^-1 ), and a similar general trend was observed for micropulsed plasma under these conditions (see Table 3). The use of the nanopulsed plasma resulted in an increase in the sensitivity by a factor of 43.7, 51.9 and 50.1% in methanol, water and blood plasma in comparison to the use of microsecond pulses (see Table 3). In addition, the LODs also increased in the order of using a matrix of either methanol, water and blood plasma (Table 3). The LODs for DMMP obtained by using nanosecond pulsed plasma were 0.41, 1.09 and 1.77 pg in methanol, water and blood plasma respectively, while those for micro plasma were 0.75, 2.16 and 5.16 pg. That is, the use of nanosecond plasma ionization lowered the LOD for DMMP by a factor of up to 3 compared to the use of microsecond plasma ionization in these matrices. The higher sensitivity and lower LODs for DMMP in blood plasma indicate that nanosecond pulsed plasma ionization can be more tolerant matrix effects from complex mixtures than microsecond pulsed plasma DBDI-MS under these conditions. Extent of internal energy deposition in nanosecond pulsed plasma ionization. An important characteristic of an ion source is the extent of energy that is deposited during ionization. Benzylamonium thermometer ion measurements were used to obtain the extent of internal energy deposition in nanosecond pulsed plasma ionization, conventional microsecond pulsed plasma ionization, and electrospray ionization. Electrospray ionization was used as a comparison because is it considered one of the ‘softest’ known ion sources. These results indicate that the average internal energy deposition in nanosecond pulsed plasma ionization is about the same or slightly lower (by <4.0 kJ mol ^-1 ) than that obtained by use of electrospray and microsecond pulsed plasma ionization. An example of the present invention is the detection of an aldehyde, an alkene and a number of ketones. To form dielectric barrier discharge ionisation, a cylindrical tube made of polyolefin (i.d. 2.7 mm, o.d. 3.1 mm, 7.5 cm long) surrounded a cylindrical copper tube (inner electrode; i.d. 1.7 mm, o.d. 2.5 mm, 15 mm long; GoodFellow Cambridge, UK). The polyolefin tube was wrapped in stainless steel wire (outer electrode; o. d. 0.35 mm, 7.5 mm long). Ionisation was formed by supplying a continuous high frequency waveform and voltage between the inner and outer electrode using nanosecond pulses. To generate nanosecond pulses, a high voltage nanopulse generator (FSWP 51-02, Behlke, Germany) was connected to a high voltage DC supply (TSA4000– 1.2/240SP; Magna-Power Electronics, Flemington, NJ, USA). The current was monitored by a picometer (Keithley 6485 Picoammeter, Oregon, USA). The following parameters were applied to detect ions in positive mode: high voltage of 2-3 kV DC, 1.1 kHz frequency waveform, duty cycle of 0.1 % and 800 ns pulses. Additional details may be found in Anal. Chem. 2020, 92 (6), 4468-4474, the whole of which is Incorporated herein by reference. To ionize the volatile compounds by nanosecond pulsed dielectric barrier discharge ionization, chemicals were introduced through a custom T-shaped headspace sampling unit (Figure 15). Two L-shaped 1/8-inch stainless-steel tubes (for the inlet 301 and outlet 302) are spaced 2 mm apart with a metal bracket 303. The length of the inlet is 40 mm and the outlet is 20 mm. ~ 5 mL of each sample 304 was sealed inside a 50 mL vial 305 via a PTFE silicone septa 306 held in place via a crimp cap 307. The sampling unit was then inserted with a purge N ₂ gas inlet system, where the outlet from the sampling system was introduced to the nanosecond pulsed plasma source. The outlet of the sampling unit was positioned ~ 3 mm from the ionisation source. Chemical vapors of aldehyde, alkene, and ketones were introduced to the plasma source directly from a vial. The arrangement in Figure 2 is similar to that in Figure 15, where the active capillary dielectric barrier discharge ion (DBDI) source comprises an inner cylindrical electrode 132, a dielectric barrier 134 provided around the outside perimeter of the inner electrode 132, and an outer electrode 136 provided around the outside perimeter of the dielectric barrier 134. The inner electrode 132, dielectric barrier 134, and the outer electrode 136, are provided in co-axial alignment, that is, sharing the same axis. 140 is a Teflon seal or gasket and 114 is the MS inlet. Nanosecond pulsed DBDI-MS of each analyte resulted in the formation of protonated molecules in relatively high abundances (Figure 16). The nanosecond pulsed plasma source also resulted in the formation of protonated dimers [2M+H] ⁺ of acetone, 2-butanone, and 2-pentanone under these experimental conditions. We also observed a minor fragment ion of protonated p-xylene (specifically [M- CH ₃ ] ^+● ) and a minor ammoniated adduct ion of perillaldehyde, [M+NH ₄ ] ⁺ . A further example of the present invention is the detection of caffeine from an instant coffee sample by nanosecond pulsed dielectric barrier discharge ionization mass spectrometry To detect caffeine from the instant coffee sample, ~1 g of coffee sample (Woolworths freeze dried classic coffee, medium roast) was placed in a glass vial. The vial was heated by using a heat gun while it was capped. The heated sample vial was de-capped and placed ~10 mm underneath the active capillary nanosecond pulsed DBDI source to ionize the sample vapor. Additional details may be found in Anal. Chem. 2020, 92 (6), 4468-4474, the whole of which is incorporated herein by reference. Here, it can be seen (Figure 17) that the nanosecond pulsed DBDI-MS can readily detect the protonated caffeine (m/z 195) as the most dominant peak in the spectrum. According to the present invention, the use of nanosecond high voltage pulses in dielectric barrier discharge ionization mass spectrometry for small molecule analysis can yield significant performance gains in terms of sensitivity, signal-to- noise ratios and detection limits when compared to more conventional microsecond pulses. For example, the signal-to-noise ratios and sensitivity for some ions formed in DBDI-MS can be increased by up to ~63% and ~52%, respectively, by using nanosecond rather than microsecond pulses. These performance gains can result in a detection limit that is up to 3 times lower for DMMP in a complex mixture (blood plasma) by using nanosecond rather than microsecond pulses. The use of nanosecond pulse widths also decreased power consumption by a factor of 6 under the ‘optimal’ conditions for detecting DMMP compared to microsecond-based pulses. Benzylammonium thermometer ion experiments indicated that the use of nanosecond pulse widths had no detrimental effects on the extent of ion fragmentation during ion formation, transfer and detection compared to the use of microsecond pulses. Moreover, the extent of internal energy deposition in nanosecond pulsed ionization is: (i) comparable or lower (by <4.0 kJ mol-1) than that for both microsecond pulsed DBDI and electrospray ionization; and (ii) substantially lower than that measured previously for direct analysis in real time and atmospheric pressure chemical ionization mass spectrometry. Although the nanosecond pulser module is relatively bulky (105 cm ² ), the size of such solid state pulsing modules continue to decrease relatively rapidly owing to continuing technological improvements. We anticipate that nanosecond pulsed DBDI should prove useful for the chemical analysis of surfaces, ‘online’ liquid chromatography mass spectrometry measurements, and in portable mass spectrometry applications. In Figures 1 to 3, the analyte material is shown as being contained or injected into a conduit. However, it will be appreciated that DBDI according to the present invention is not limited to setups involving dielectric conduits, as long as it involves the application of nano-second high voltage pulses as described herein. It is understood that the requirement is for the dielectric barrier, which may not necessarily be a conduit for an analyte or a gas phase of an analyte, to be located between the two electrodes across which the voltage pulses are to be applied. The analyte material, or the gas phase thereof, will also be located between the two electrodes to enable the DBDI. For instance the material for ionization can be located in an airgap between the electrodes. In some cases, the material for ionization can be directed at a surface by use of a backing pressure (e.g. from compressed gas cylinder) to desorb and ionize analytes from a surface, for detection by mass spectrometry. This would be useful for the chemical analysis of surfaces (e.g. pesticides on fruit, fingerprints, quality control of semiconductors etc). In the claims which follow and in the preceding description of the invention, except where the context requires otherwise due to express language or necessary implication, the word “comprise” or variations such as “comprises” or “comprising” is used in an inclusive sense, i.e. to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention.