AVS 52nd International Symposium
    Plasma Science and Technology Tuesday Sessions
       Session PS-TuP

Paper PS-TuP18
The Role of Ar Metastables in Distorting Gas Temperature Measurements Derived from Trace N@sub 2@ Optical Emission Rotational Spectroscopy in Ar-Containing Discharges

Tuesday, November 1, 2005, 4:00 pm, Room Exhibit Hall C&D

Session: Plasma Science and Technology Poster Session
Presenter: S.J. Kang, University of Houston
Authors: S.J. Kang, University of Houston
V.M. Donnelly, University of Houston
Correspondent: Click to Email
Gas temperature (T@sub g@) is an important parameter in plasma prosessing. One method for obtaining T@sub g@ is to add small quantities of N@sub 2@ to the discharge and determine the rotational temperature (T@sub rot@) of N@sub 2@ from electron-impact induced emission. The assumption is that collisions with electrons transfer the rotational distribution of the ground state intact to the emitting state. Usually the 0->0 and 1->0 vibronic bands of the C@super 3@@PI@@sub u@ -> B@super 3@@PI@@sub g@ transition are observed. We have found that when the plasma contains large amounts of Ar these emissions do not yield reliable values for T@sub g@. For example, in a 98% Ar/ 2% N@sub 2@ inductively-coupled plasma (ICP), these bands yield an apparent T@sub g@ that decreases with increasing power. This erroneous result is due to a second mechanism of N@sub 2@ C@super 3@@PI@@sub u@ excitation by collisions with Ar metastables (Ar*). Kinetic modeling shows that at low powers, this Penning excitation mechanism dominates over electron-impact excitation, while at high power, the converse is true. Penning excitation releases a large amount of energy into N@sub 2@ (C@super 3@@PI@@sub u@) rotations. When this process dominates, the observed T@sub rot@ is very high and greatly exceeds T@sub g@. Consequently, T@sub rot@ decreases with increasing power. When N@sub 2@ emission is observed from vibrational levels above that of Ar (@super 3@P@sub 0@) (11.72 eV), i.e., for v' > 4, Ar* energy transfer is no longer possible. So T@sub g@ derived from the 3->3 vibronic band is lower and increases slightly with power. This trend is more reasonable than the high values and inverse dependence on power from the 0->0 transition, consistent with emission from the 3->3 vibronic band arising only from collision with electrons. Thus the 3->3 band provides a reliable measure of T@sub g@.