We have obtained Hubble Space Telescope (HST) Faint Object Camera (FOC) f/4
8 long-slit optical spectroscopy of the inner 2 " of the narrow-line region
(NLR) of the Seyfert 2 galaxy Mrk 3 with a spatial resolution of 0." 06. S
pectra were taken in six locations with the slit approximately perpendicula
r to the radio axis. In the region cospatial with the radio jet, where the
brightest emission line knots are located, the velocity field is highly per
turbed, showing two velocity systems separated by as much as 1700 km s(-1).
In several locations the split lines form almost complete velocity ellipso
ids, implying that we are seeing an expanding shell of gas. The diameter of
this shell (similar to 200 pc) is much larger than the width of the radio
jet (d < 15 pc). We interpret this to be the consequence of the rapid expan
sion of a cocoon of hot gas, shocked and heated by the radio-emitting outfl
ow, which compresses and accelerates the ambient gas. The cocoon mediates t
he energy exchange between jets and the line-emitting gas. The gas motions
within the NLR of Mrk 3 are therefore clearly dominated by the interaction
between the jets and the interstellar medium; the NLR itself Is essentially
a cylindrical shell expanding supersonically. With its current size of 200
pc, the cocoon has expanded to several disk scale heights. Because of the
external gas density stratification, the hot gas located above the plane of
the disk blows out into the halo, puncturing the bubble and fracturing the
velocity ellipsoids. The system is effectively momentum driven. From the s
ize and velocity of the expanding region, we derive an upper limit to the a
ge of the radio source of less than or similar to 1.5 x 10(5) yr, and se lo
wer limit for the jet power of greater than or similar to 2 x 10(42) ergs s
(-1), required to inflate the cocoon; we estimate that the jet minimum adva
nce speed is 3 x 10(-3) pc yr(-1). The total kinetic energy of the high-vel
ocity gas associated with the radio jet can be estimated as similar to 6 x
10(54) ergs, comparable to the total energy carried by the jet over its lif
etime; this quantitatively supports the idea that the NLR gas is accelerate
d by the jet. Radio outflows are associated with at least 50% of Seyfert ga
laxies with typical sizes smaller than a few kpc. If the advance speed of M
rk 3 is representative of the Seyfert population, then these sources must a
lso be short lived and are probably recurrent. Evidence that this is indeed
the case is provided by the fact that the expansion timescale derived for
NGC 1068 is comparable to that seen in Mrk 3. The jet kinetic luminosity of
Mrk 3 is between 2 and 3 orders of magnitude smaller than that derived for
radio-loud AC;Ns with similar emission-line luminosity. On the other hand,
the fraction of jet power dissipated in radio emission is similar. We spec
ulate that the main distinction between radio-quiet and radio-loud active g
alactic nuclei can be ascribed to a difference in jet power rather than to
a different efficiency in synchrotron emission production.