26.03.2013 Views

chemical physics of discharges - Argonne National Laboratory

chemical physics of discharges - Argonne National Laboratory

chemical physics of discharges - Argonne National Laboratory

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

13<br />

field per electron is and per collision ttrereforc<br />

11. 2. iliffusion <strong>of</strong> Ciiargeci Species.<br />

2 e2L 2 ,<br />

in u<br />

e<br />

rnGT *<br />

In the discliarges <strong>of</strong> interest iiere, tile concentration <strong>of</strong> charged species is<br />

greater ellan and a large fractional separation <strong>of</strong> elrctrong alid positive ions<br />

becomes impossible, because it would set up a very large opposiug field. The currents<br />

<strong>of</strong> electrons and ions reaching the wall must tiicn be equal, i = - i) Vn - nu E'<br />

i+ = - D+Vn + nu E' where n = n<br />

+<br />

IJ the diffusion coefficient, LI the mobility, and E' the field due to tile (small) space<br />

charge. The subscripts e and + refer to clectror. and positive ion. Equating i and<br />

i, and eliminating C' one obtains<br />

r<br />

Dew+<br />

I = -<br />

+ U+we<br />

?+ + Ue<br />

= n+ is the electron density, 011 the densitv gradient,<br />

Vn = 1) Vn<br />

which serves as the definition <strong>of</strong> the anbipolar diffusion coefficient, Da. SubstitukTe<br />

ting P - for i)<br />

e c<br />

and the equivalaut expression for 1)+ otic obtains<br />

v+k<br />

which approximately equals - ('1<br />

c e + T ) or D+(1<br />

+<br />

T<br />

+ ")<br />

T+<br />

because the electron mobility<br />

pee' is much larger than the ionic mobility P+. In active glow <strong>discharges</strong> T /T+ is<br />

typically 20 to 100, whereas in the afterglow the electrons thermalize rapidly ana<br />

Te/T+ = 1.<br />

Thus, D 2 20 to 100 D+ in the active discharge, and 2 D+ in the afterglow.<br />

The disappearance <strong>of</strong> charged species bv ambipolar diffusion in the absence <strong>of</strong><br />

6n<br />

a source term is described by the diffusion equation = uavZn where n is a function<br />

<strong>of</strong> r, 8, z, and t. The well-known solution <strong>of</strong> this equation for the case <strong>of</strong> an infinite<br />

m<br />

cylinder is n(r,t) = Z<br />

-kit<br />

AiJo(% I) e where al is the i th root <strong>of</strong> Jo, the Bessel<br />

i=1 0<br />

a.<br />

2 Da<br />

function <strong>of</strong> zero order and k = ($) . The diffusion length, A, therefore<br />

i<br />

Da =<br />

r 0<br />

equals . The first few zeroes <strong>of</strong> Jo are a1 = 2.405, a2 = 5.520. a3 = 8.654,<br />

'i<br />

a,, = 11.792 which shows that, as diffusion proceeds, tile time decay will be increasingly<br />

governed by kl = (-) Da, the first (lowest) diffusion mode, because the next three<br />

r0<br />

higher modes are damped out more rapidly by factors <strong>of</strong> 5.3, 12.9, and 24. After a<br />

short transient, the diffusion-controlled electron decay or the diffusion controlled<br />

loss under steady-state conditions with a spatially well distributed source term can<br />

therefore be closely approximated by a first-order rate constant, k = 5.78 Da/ro2.<br />

This analysis applies when there are only positive ions and electrons present.<br />

h'hen negative ions are also present, their principal effect is to accelerate the<br />

ambipolar diffusion <strong>of</strong> the electrons, (Da)e, which now becomes approximately<br />

T<br />

Te<br />

(Da)e = (1 + A) D+ (1 + -) + X - 1) which for active <strong>discharges</strong> (Te/T+ >> 1)<br />

*+ T+<br />

can be further approximated by (1 + 2 A) Te/T+, where<br />

ratio <strong>of</strong> negative ions and electrons.<br />

= n-/ne. the concentration<br />

It can be seen that for A >> 1 electrons

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!