A cold atmospheric pressure plasma jet controlled with spatially separated dual-frequency excitations

Z Cao^4Notea011, Q Y Nie^4Notea011,2 and M G Kong^1,3

¹ Department of Electronic and Electrical Engineering, Loughborough University, Loughborough LE11 3TU, UK
² School of Physics and Optoelectronic Technol, Dalian University of Technology, People's Republic of China

³ Author to whom any correspondence should be addressed.

E-mail: m.g.kong@lboro.ac.uk

Received 22 September 2009, in final form 8 October 2009
Published 5 November 2009

Cold atmospheric pressure plasmas (CAPs) configured as a spatially extended jet have commanded much interest, largely due to their increasing use in industrial and medical applications, including polymeric surface modification [1, 2], bacterial and biomolecule inactivation [3, 4], wound healing [5, 6] and nano-structure fabrication [7, 8]. In spite of their strong tendency for the glow-to-arc transition particularly in molecular gases [9], the application success of CAP jets benefits critically from the spatial separation of their plasma generation region from their downstream processing region [10], thus facilitating a near-sample introduction of versatile reaction chemistry with relatively little impact on plasma stability [11]. Spatial extension of CAP jets also makes it possible to treat three-dimensionally shaped objects [12, 13]. However, cold atmospheric plasma jets and jet arrays are relatively new additions to the family of non-equilibrium atmospheric discharges with little understanding of both their technological scope and their underpinning physics. The expanding range of their practical applications heightens two important long-term questions, namely whether future innovations could fundamentally improve their performance as a material processing tool and how their plasma dynamics might be tailored to control downstream reaction chemistry. While these two questions are being pursued sometimes independently, their common goal is to identify effective control strategies with which key plasma parameters can be tuned, separately but not necessarily independently, to influence and control the functionalities of plasma material processing.

The task to achieve controllable plasma processing with physical parameters is particularly challenging for non-equilibrium atmospheric discharges. This is largely due to their very high collisionality that forces a narrow parametric range within which to satisfy multiple and often conflicting requirements such as high electron density and temperature, low gas temperature, robust plasma stability and abundant reactive species. Many of these key plasma parameters are intertwined in strong and sometimes unwanted cross-dependence in most current CAP sources. For example, an attempt to increase plasma density and hence achieve more active reaction chemistry usually leads to a large increase in gas temperature and sometimes plasma instability. The need to reduce gas temperature for processing heat-sensitive materials, on the other hand, is often met by operating CAPs in a low current regime or in an afterglow mode for which reaction chemistry may be significantly compromised. It should be acknowledged that engineering solutions may mitigate some of the unwanted cross-correspondences and achieve an acceptable application efficacy for a specific application. However, it is always desirable to seek a general strategy to weaken these cross-correspondences and enhance opportunities for better process control.

This paper reports an attempt to separate the control of three fundamental plasma quantities in CAP jets, namely plasma density, plasma plume length and gas temperature by means of spatially separated application of dual-frequency excitations. As will be shown below, spatial separation of the dual-frequency excitations is critical to enable separated control of plasma parameters in CAP jets and is consistent with the spatial separation of their plasma generation and plasma processing regions. This work is also motivated by a common dilemma in choosing different CAP jets of contrasting characters. Radio-frequency (RF) CAP jets tend to achieve higher plasma density but with a shorter jet length and a larger gas temperature in comparison with kilohertz excited CAP jets (referred to as ac CAP jets). High plasma density of RF CAP jets is desirable, but their short jet length and large gas temperature tend to restrict the scope of their application to flat and heat-insensitive surfaces. If a second ac excitation is added to its downstream region, the jet plume length could be extended significantly and this physical extension of the plasma plume would also facilitate cooling through heat exchange with the ambient background gas. Dual-frequency excitations are known to be beneficial for low pressure capacitively coupled plasmas [14], but their use for atmospheric plasmas are rare and so far limited to a co-located excitation at two similar radio frequencies in the parallel-plate electrode configuration [15]. In the latter case, the effects of the two excitation frequencies are strongly coupled, encouraged by their co-located application. This is different from our goal of weakening and separating the cross-dependence of plasma density, plasma plume and gas temperature. Spatially separated dual-frequency excitation has not been reported for cold atmospheric plasmas in general and cold atmospheric plasma jets in particular.

To characterize dual-frequency CAP jets, we consider a linear-field CAP jet [16] with three electrodes, namely an upperstream capillary electrode recessed within a quartz tube, a ring electrode wrapped around the quartz tube near the tube nozzle and a downstream plate electrode placed downstream from the tube nozzle as shown in figure 1 [17]. The capillary electrode doubles as the gas inlet. With the ring electrode grounded, the capillary electrode is powered at 5.5 MHz and the downstream plate electrode is powered at 30 kHz. So the influence of the RF excitation is confined mainly within the electrode region between the capillary and the ring electrode, whereas the influence of the ac excitation is confined largely to the open region between the ring and the plate electrodes. This spatially separated application of the dual-frequency excitation is useful to maximize the control of plasma dynamics.

For experiments reported here, we employ a helium flow (99.9999%) at 4 slm (standard litres per minute) to flow through the capillary electrode and then the quartz tube. When the structure is operated as an RF CAP jet at 5.5 MHz with a zero peak–peak ac voltage (i.e. V_ac  =  0 V), the generated plasma remains at the tip of the capillary electrode at a peak–peak RF voltage of V_RF  =  1 kV, as shown in figure 2(a). With increasing V_RF, the plasma starts to extend and at V_RF  =  2 kV its length outside the quartz tube is about 5 mm (see figure 2(b)). A millimetre scale jet length is typical of RF CAP jets [5], and this is due to an electron trapping mechanism for RF atmospheric plasmas [18]. If the CAP jet is operated as an ac device with V_RF  =  0 and the downstream plate electrode powered at 30 kHz, the ignition voltage is found to be above 6 kV as shown in figure 2(c). However, by using a dual-frequency excitation with V_RF  =  1 kV and V_ac  =  3 kV, each alone being unable to extend the plasma outside the quartz tube, a 15 mm long plasma jet is formed outside the tube, as shown in figure 2(d). A further increase in the ac voltage to V_ac  =  5 kV with V_RF  =  1 kV fixed is seen in figure 2(e) to significantly increase the optical emission of the plasma jet. In fact, the emission intensities in figures 2(b) and (e) are comparable, suggesting a similar plasma density of the dual-frequency jet to that of the RF jet in figure 2(b). The dual-frequency excitation is capable of extending the plasma to an outside-tube distance of 30 mm, with the same applied voltages of V_ac  =  5 kV and V_RF  =  1 kV, as shown in figure 2(f). Therefore, the use of dual-frequency excitations can both achieve a similar plasma density to that of an RF excitation and generate spatially extended high-density CAP jets similar to ac CAP jets. In other words, dual-frequency CAP jets capture the advantages of the RF and ac jets in a single device. In fact, this is achieved at a low level of the applied voltages with which single-frequency excitations would be unable to even strike a plasma. Clearly the influences of the two excitations do couple together to influence plasma ignition and subsequent evolution in dual-frequency CAP jets.

Voltage and current measurements indicate that the discharge current is made of a slow-changing ac component superimposed with a fast-oscillating RF component (data now shown). In particular, the upperstream discharge sustained by the RF excitation is dense and provides an additional electron source to the downstream ac-sustained discharge. This increases the plasma density in the downstream. Operating at a much lower excitation frequency, electron wall loss in ac CAP jets is severe with little electron trapping and as a result their electron density tends to be lower than that in an RF CAP jet. However, the reduced electron trapping also allows the ac jet to be spatially extended. With the upperstream discharge providing considerable electrons to the downstream ac plasma jet, the plasma density of the latter is enhanced substantially along its full plume length. One consequence is that the long plume length of dual-frequency CAP jets can be achieved at relatively low applied voltages. To see this more clearly, we place the plate electrode at different downstream locations until the tip of the CAP jet just touches the plate electrode. This is referred to as the longest plume length measured from the tube nozzle to the plate electrode. In figure 3(a), the longest plume length of the RF CAP jet (V_AC  =  0 V) is seen to increase monotonically with V_RF and its maximum value is about 5 mm at V_RF  =  2 kV. For the dual-frequency CAP jet case, its longest plume length is found to be much longer and reach 35 mm at V_AC ≥ 5 kV and V_RF ≥ 1 kV. The dependence of the plume length on V_RF is also different, increasing with V_RF first till V_RF  =  1 kV after which it undergoes much slower changes. This gradual change varies from a small increase at V_AC  =  3 kV to a small decrease at V_AC ≥ 5 kV. These results suggest that the use of the ac excitation substantially changes the plasma plume length in CAP jets.

It should be noted that the downstream plate electrode is powered at ac in figure 2 to maximize the spatial separation of the two excitations for the dual-frequency CAP jet. However, to compare its plume length to that of a conventional ac jet, it is appropriate to power the capillary electrode at ac. In this case, figure 3(b) shows that the maximum value of its longest plume length is 20 mm when V_AC  =  5 kV is applied to the capillary electrode. The plasma jet becomes unstable with a moving root on the downstream electrode if V_AC is raised above 5 kV, and so for the electrode configuration of figure 1 the longest plasma plume length is about 20 mm with ac excitation only. With V_RF switched onto the capillary electrode and V_AC  =  6 kV now applied to the plate electrode, the plume length increases significantly to around 30–37 mm. This is markedly higher than the maximum 20 mm in the ac CAP jet. Therefore, even though the ac excitation influences most significantly the plasma plume length, the RF excitation is also important and in fact extends the plume length. This is likely to benefit from the electron supply of the RF discharge. For the two dual-frequency cases in figure 3(b), their emission intensity is stronger at V_RF  =  2 kV than V_RF  =  1 kV. Yet figure 3(b) suggests that the plume length can be slightly longer at V_RF  =  1 kV particularly for V_AC  =  4 kV. This could be a result of stronger electron trapping in the upperstream region with a larger V_RF [18], and indicates the presence of an optimum operating condition for the upperstream RF discharge.

We now consider the wavelength-integrated optical emission intensity (350–900 nm) and the gas temperature approximated with the rotation temperature, T_rot, measured from the first negative system ( , , v '  =  0, 388–392 nm). Both are obtained for the jet point on the plate electrode. As the plume length of the RF jet is small, the downstream plate electrode is placed 4 mm from the quartz tube nozzle. For the ac and dual-frequency jets, this distance is 15 mm. Figure 4 shows that both the emission intensity and T_rot increase with the applied voltage. The maximum emission intensity is approximately 20 µW cm^–2 and 15 µW cm^–2 for the RF and ac jets, respectively, whereas the corresponding gas temperature is similar at 490 K. In the dual-frequency jet case, a larger V_RF is found to substantially increase the emission intensity. At V_RF  =  2 kV and for T_rot ≈ 490 K, the emission intensity is found to be about 40 µW cm^–2, or 2.0 and 2.7 times those for the RF and ac jets, respectively. This improvement is similar at lower gas temperature. For example, the emission intensity corresponding to T_rot ≈ 400 K is similar at about 7 µW cm^–2 for both the ac and RF jets and about 15 µW cm^–2 for the dual-frequency jet. This represents a factor of 2 in the intensity ratio. Therefore, with little change to the gas temperature, the dual-frequency CAP jet can produce stronger optical emission than its single-frequency counterparts. In other words, dual-frequency CAP jets can be used to achieve separate, not necessarily independent, control of some plasma parameters. The length of the dual-frequency CAP jet is much longer than that of the RF jet and is reached at a much smaller V_AC than that of the ac jet. For applications that require much lower gas temperature than 490 K, a pulsed RF excitation can be used to replace the RF excitation as shown for single-frequency CAPs [19].

If the applied voltage of the single-frequency jet, either RF or ac, is increased above its maximum value shown in figure 4, the plasma jet would become unstable with a moving root. However, for the dual-frequency CAP jet, a further increase in V_AC does not affect plasma stability but may cause the current waveform to deviate from its nominally sinusoidal waveform. As this non-sinusoidal regime is still useful for practical applications, the wavelength-integrated emission intensity in this regime is measured up to V_AC  =  6.5 kV and found to reach 60 µW cm^–2, at least 3 times the maximum of the single-frequency jets (data not shown). This much increased optical emission is indicative of the very high plasma density that can be achieved with the dual-frequency CAP jet. Results based on wavelength-integrated emission intensity in figure 4 would ideally need to be confirmed with direct electron density measurement. For atmospheric plasmas, recent studies suggest that reliable electron density measurement is feasible [20–22] though the methods reported so far would still need further development to fully substantiated, particularly for CAP jets of a millimetre diameter. However, it is possible to strengthen the results in figure 4 with emission intensities of key plasma species. Figure 5 shows the ac voltage dependence of emission intensities of nitrogen dimmer ions at 391 nm, excited helium at 706 nm and two excited atomic oxygen lines at 777 and 845 nm. It is evident that these important species are much more abundant at V_RF  =  2 kV than the ac only case whereas at the same RF voltage the plasma plume cannot even reach the downstream point of 15 mm. Helium line at 706 nm is known to indicate energetic electrons [23], and its larger emission intensity in figure 5(b) supports the conclusion of larger plasma density in dual-frequency CAP jets arrived from the wavelength-integrated emission data in figure 4. Also the higher emission intensities of the dual frequency jet at 777 nm and 845 nm suggest a more active reaction chemistry in the downstream region. Finally it should be noted that the increase in the emission intensity of the dual-frequency jet is different from the ac case, particularly at 706, 777 and 845 nm. This may be indicative of changes in the electron mean energy and indeed the electron energy distribution function. We do not, however, expect much change in ion energy, as it is small in atmospheric plasmas.

The results and their discussions presented here have demonstrated the distinct benefits of dual-frequency CAP jets when the two frequency excitations are applied to spatially separated locations. In comparison with their single-frequency counterparts, dual-frequency CAP jets have been shown to have larger plasma density (as indicated with the emission intensity) and longer plume length at the applied voltages smaller than what is necessary for the single-frequency jet to form outside the quartz tube. These features could be used to achieve longer plume length, as shown in figure 2(f) or higher plasma density as shown in figure 2(e), suggesting the possibility of separate, but not necessarily independent, control of different plasma parameters. The interaction between the influences of the two frequency excitations is both clear and beneficial, for example the upperstream RF discharge feeding electrons to the downstream ac discharge for the latter to acquire higher plasma density and longer plume length without an undue increase in gas temperature. While topics such as this are of interest for further investigation, it is clear that dual-frequency CAP jets offer enhanced plasma properties with potential for greater process control.

[1] 

Klages C P, Hinze A, Lachmann K, Berger C, Borris J, Eichler M, von Hausen M, Zanker A and Thomas M 2007 Plasma Process. Polym. 4 208 
CrossRef

[2] 

Girard-Lauriault P L, Mwale F, Iordanova M, Demers C, Desjardins P and Wertheimer M R 2005 Plasma Process. Polym. 2 263 
CrossRef

[3] 

Laroussi M 2002 IEEE Trans. Plasma Sci. 30 1409 
CrossRef

[4] 

Deng X T, Shi J J, Chen H L and Kong M G 2007 Appl. Phys. Lett. 90 013903 
CrossRef

[5] 

Stoffels E, Flikweert A J, Stoffels W W and Kroesen G M W 2002 Plasma Source Sci. Technol. 11 383 
IOPscience

[6] 

Fridman G, Friedman G, Gutsol A, Shekhter A B, Vasilets V N and Fridman A 2008 Plasma Process. Polym. 5 503 
CrossRef

[7] 

Chen C K, Perry W L, Xu H, Jiang Y and Phillips J 2003 Carbon 41 2555 
CrossRef

[8] 

Richmonds C and Sankaran R M 2008 Appl. Phys. Lett. 93 131501 
CrossRef

[9] 

Meek J M and Craggs J D 1978 Electrical Breakdown of Gases (New York: Wiley)  

[10] 

Walsh J L and Kong M G 2006 Appl. Phys. Lett. 89 161505 
CrossRef

[11] 

Hubicka Z, Cada M, Sicha M, Churpita A, Pokorny P, Soukup L and Jastrabik L 2002 Plasma Source Sci. Technol. 11 195 
IOPscience

[12] 

Ideno T and Ichiki T 2006 Thin Solid Films 506 235 
CrossRef

[13] 

Cao Z, Walsh J L and Kong M G 2009 Appl. Phys. Lett. 94 021501 
CrossRef

[14] 

Boyle P C, Ellingboe A R and Turner M M 2004 J. Phys. D: Appl. Phys. 37 697 
IOPscience

[15] 

Waskoenig J and Gans T 2009 Presented at 19th Int. Symp. on Plasma Chemistry (Bochum, Germany, July 2009)  

[16] 

Walsh J L and Kong M G 2008 Appl. Phys. Lett. 93 111501 
CrossRef

[17] 

Cao Z, Nie Q Y, Walsh J L, Ren C S, Wang D Z and Kong M G 2009 Plasma Sources Sci. Technol. submitted 

[18] 

Liu D W, Shi J J and Kong M G 2007 Appl. Phys. Lett. 90 041502 
CrossRef

[19] 

Ye R, Ishigaki T and Sakuta T 2005 Plasma Sources Sci. Technol. 14 387 
IOPscience

[20] 

Kono A and Iwamoto K 2004 Japan. J. Appl. Phys. Pt 2 43 L1010 
CrossRef

[21] 

Choi J Y, Takano N, Urabe K and Tachibana K 2009 Plasma Sources Sci. Technol. 18 035013 
IOPscience

[22] 

Zhu X M, Pu Y K, Balcon N and Boswell R 2009 J. Phys. D: Appl. Phys. 42 142003 
IOPscience

[23] 

Bibinov N K, Fateev A A and Wiesemann K 2001 J. Phys. D: Appl. Phys. 34 1819 
IOPscience

Notes

Notea01  These authors made equal contributions.