2544 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 39, NO. 11, NOVEMBER 2011 Optical Observation of Nitrogen Propagation in Argon Plasma Daisuke Ogawa, Matthew J. Goeckner, and Lawrence J. Overzet Abstract—While the gas flow patterns in reactors can be visual- ized using models, it has proven that it is difficult to measure them. This paper shows that the gas flow pattern for pulsed injection of nitrogen gas into argon plasma can be readily seen in the optical emission intensity. As N 2 enters the Ar plasma, N 2 molecules cause increased optical emission intensity by the Penning effect. This allows one to measure the N 2 propagation. In this paper, we present the pictures with the measurement details and analysis. Index Terms—Argon, emission, image analysis, nitrogen, optical imaging, optical spectroscopy, plasma measurements, pressure effects, spraying, velocity measurement. L ARGE gas flow rates and fast-pulsed injection of gases at large flow rates are finding increasing importance in de- vice processing, particularly in microelectromechanical-system etching. The large flow rates and the quick switching between gases must be also accompanied by great processing uniformity over large areas. This undoubtedly requires a very uniform gas distribution (or gas flow pattern). While the gas flow pattern can be visualized by modeling, it has thus far proven to be remarkably difficult to measure. In this paper, we show the gas flow pattern for nitrogen injection into argon plasma using time-resolved measurements of the optical emission intensity pattern. Such a method can be used to measure flow patterns in reactors for comparison with models. In order to take the series of time-resolved total emis- sion intensity pictures shown in Figs. 1 and 2, we used an intensified charge-coupled device (ICCD) camera (Princeton Instruments–ITEA/CCD-576GRBEM). The ICCD gate time was set at 0.2 ms, and each picture comes from an average of over 16 N 2 injections. To obtain the change in emission caused by the arrival of the N 2 molecules, the pictures were subtracted, pixel by pixel, from the intensities measured just before N 2 arrived at t =2 ms. White–green indicates a region where the emission increased, dark-blue–black indicates that the emission decreased, and blue indicates that the emission remained constant. All pictures have the same color scale. We visualized the gas propagation in plasma by inject- ing gases through an automotive fuel injector (Denso 23209- 0D040) into a 120-mtorr capacitively coupled Ar plasma. For this initial discharge, the flow rate was 15 standard cu- bic centimeters per minute (SCCM), and the RF power was 10 W at 13.56 MHz. The injector was located 3 ′′ above the plasma region, at the top of a six-way 4(1/2) ′′ cross. The injected gas causes the chamber pressure to rise to Manuscript received November 29, 2010; revised June 3, 2011; accepted June 4, 2011. Date of publication July 18, 2011; date of current version November 9, 2011. This work is supported by the Department of Energy under Grant DE-SC0001355. The authors are with The University of Texas at Dallas, Richardson, TX 75080 USA (e-mail: goeckner@utdallas.edu; overzet@utdallas.edu). Digital Object Identifier 10.1109/TPS.2011.2159625 ∼180 mtorr, 60 ms after the injector opened. The equivalent gas flow rate was 240 SCCM, the reactor was 2.5 L, and the pump speed was 3.6 L/s. The injector opened at ∼3 ms, resulting in a pressure wave arriving at the plasma at ∼3.2 ms. The plasma was driven by the RF electrode, observed at the bottom of the circular window in the pictures. This is most noticeable by the increased optical emission occurring just above it. Although the net plasma power can vary as the gas flows in, it was found to remain constant at ∼10 W during the time period shown in Figs. 1 and 2. Fig. 1 shows the results of the Ar gas injection into the Ar plasma for comparison with the N 2 results. The Ar injection only increased the emission just above the RF electrode surface once the pressure wave arrived [see Fig. 1(b) and (c)]. In addition, the emission decreased in the region just above the brightened line. This is all consistent with the sheath thinning as the pressure increases and the electron mean free path decreases. The emission increases more on the right-hand side of each picture, which is consistent with the orientation of the injection. As shown in Fig. 2, the injection of N 2 into the Ar plasma gives a markedly different emission pattern from that with the Ar injection, allowing one to see how N 2 fills the reactor volume temporally. Here, there is an emission increase that propagates from the top to the bottom on the right-hand side. It takes ∼0.4 ms for N 2 to traverse from top to bottom in these pictures indicating speed ∼160 m/s (one-third the thermal velocity of N 2 ). Furthermore, time-resolved optical emission spectra, taken under experimental conditions similar to those used in Fig. 2, showed that the propagation observed in Fig. 2 was indeed caused by N 2 . Lines from both the first negative system of N 2 (427 nm; C 3 Π u − B 3 Π g , v ′ − v ′′ ; 1–5) [1] and the second positive system of N + 2 (427.8 nm; B 2 σ + u − X 2 σ + g , v ′ − v ′′ ; 0–1) [1] increased as N 2 arrived [2]. Furthermore, an Ar emission line (415.9 nm; 3p 6 − 1s 5 ) [3] decreased at the same time [2]. Therefore, the increased emission was from N 2 , and the implication is that Ar metastables and Ar + cause that increase in emission via excitation reactions [4], [5] of N 2 . Finally, the N 2 addition affects the plasma near the RF electrode nearly immediately. The emission just above the RF electrode well decreases before the N 2 gas arrives. The black region indicates a significant reduction in emission in the main body of the plasma that occurs well before the N 2 gas arrives and continues on the left-hand side even after the N 2 gas reaches the RF electrode. As such, it seems that this decrease cannot be explained by any chemical interaction, such as the increased emission at the right, but must be rather related to a plasma impedance change or other far-reaching and fast effects. In conclusion, we have followed the N 2 propagation in argon plasma by a change in emission. The increased emission ap- pears to have been caused by Penning excitation and ionization 0093-3813/$26.00 © 2011 IEEE