FGQM11

Dr. Faucher-Giguere,

I read with interest your preprint http://arxiv.org/abs/1108.0413 on FeLoBALs.
Very nice work.

I do have some comments and questions about which I'd be interested in
hearing your thoughts.  The length of the email just reflects the strength
of my interest.  I've also posted this email at
http://pathallresearch.blogspot.com/2011/08/fgqm11.html
if you want to discuss it in that forum (I won't post private replies there).


First, I think it's misleading to use the term "quasar blast wave".
It had me thinking you were NOT modelling the case of a steady
quasar wind inflating a bubble, until I got to the appendix.
Wouldn't replacing "blast wave" with "shock wave" or
"quasar-driven shock wave" be more appropriate?


Second, to make the paper more accessible to observers, I think it's
important to introduce the Figure in the Appendix right at the start of
Section 2.  The oversimplified picture observers have (myself included) is
that a quasar wind is in a free expansion phase, driving a shock outwards
at v_sh ~ v_in, with negligible swept up mass.  But as you point out, the
properties of the swept-up gas and not the initial wind govern interactions
with the ambient medium.  So Figure A1 needs to be emphasized in the main
text as _the_ picture to keep in mind.  E.g., without the appendix figure
it's was hard for me to picture how v behind the shock could be < v_sh.

I would also suggest using the term "wind shock" (McKee & Hollenbach 1987,
Fig. 4) instead of "reverse shock".  The latter implies to an observer like
me a shock which is propagating backwards, but that's not the case here
(though not impossible I suppose).

A related comment is that I went looking at the Koo & McKee paper, and
references therein, for a simple outline of why the wind forms a structure
like that of the appendix figure.  Such an outline would be useful to add.
Something as simple as: the wind first freely expands and shocks the ambient
medium; that ambient shock and the leading edge of the wind slow down as
material is swept up; the leading edge of the wind is shocked by the wind
behind it, forming the wind shock interior to the ambient shock.
[Koo & McKee mislead readers at the start of their section 2 by incorrectly
claiming that the velocity of the ambient shock is > than the initial wind
velocity.  It took the above McKee & Hollenbach reference (and Koo & McKee
Figure 4) to clear that up for me.  And it's a point worth making, I think,
that even if the observed FeLoBAL gas is fully accelerated to v=v_hot, that
velocity is less than v_in of the driving wind.  And if I got anything wrong
in my comments above, that just proves my point for adding an outline...]

Also, in the Appendix: clarify that the contact discontinuity separates the
shocked wind and shocked ambient medium, which are assumed to be immiscible.


Third, you conclude the appendix by saying that Mdot_in and v_in can't be
separately inferred; that seems true if all you know is R_s and properties
of the gas near R_s.  But is it impossible in principle, if you knew all
other parameters of the outflow?  Mightn't different (Mdot_in,v_in)
combinations at the same Edot produce different R_sw, for example?
Or could you maybe use momentum arguments to estimate Mdot*v_in and
combine that with Edot_k to get Mdot and v_in?


Fourth, regarding the first sentence of the discussion:
even if FeLoBAL absorption systems are distinct from Hi/LoBAL systems,
FeLoBAL quasars themselves might not be distinct from Hi/LoBAL quasars,
in the sense that FeLoBALs will only be seen along sightlines where quasar
winds (i.e. Hi/LoBALs) exist.  There are of course some workarounds
which could cause FeLoBALs to be seen in objects without Hi/LoBALs:
* if a wind exists but is all so highly ionized it doesn't show CIV,NV,etc.
* if a Hi/LoBAL wind doesn't currently exist, but used to.
* if your idea of Omega=1 for FeLoBALs at kpc scales is correct.

A way to constrain some of the above is to compare the relative numbers of
FeLoBALs with all transitions (Fe II thru C IV) covering the same velocity
ranges, i.e. consistent with absorption only from swept-up clouds, and of
FeLoBALs with HiBAL absorption at a much wider range of velocities.
The former would be non-BALs if you took the FeLoBAL gas away;
the latter would still be BALs.  If your idea of Omega=1 for FeLoBALs is
correct, there should be from 1.5 to 3 times as many of the former as of
the latter (for 40% and 25% BAL fractions, respectively).

Looking at the objects in the paper in terms of the above:
J0838 is a case of the latter; component c is an FeLoBAL system,
but there are other components a and b which are HiBALs.
J0318 is a case of the former.
Q2359 is in between; it has higher-velocity systems only seen in MgII
(it's too low z for C IV coverage from the ground)
but the strongest system is the one with FeII.

Regarding the Omega=1 suggestion, is there a way to estimate the timescale
over which the hot gas will expand laterally, and compare that to the flow
timescale?  Of course a pressure gradient will drive lateral expansion, but
the timescale would have to be comparable to the flow timescale for the
gas bubble to end up with Omega=1 by the time the FeLoBAL region is reached.


Fifth, in section 2.2 it seems to me that one further step would be useful:
giving the pre-existing density above which post-shock gas _would_ cool in
a flow time, and then showing that such densities are almost never plausible.
If I've done the math correctly, post-shock gas will cool in a flow time if
$\bar{n}_H^{pre} > 330 yrs T_{sh,9}^{1/2} (v/10,000 km/s) (R/3 kpc)^{-1}$.


Lastly, a minor point, but I would suggest replacing "QSO" with "quasar"
or "Q" (for variables) throughout, at least if you share my preference for
a made-up word over an acronym...

Cheers
Pat