1: \documentclass[12pt,eqsecnum,preprint]{aastex}
2:
3: \begin{document}
4:
5: \title{The Physical Nature of Polar Broad Absorption Line Quasars}
6: \author{Kajal K. Ghosh\altaffilmark{1}} \and\author{Brian
7: Punsly\altaffilmark{2}} \altaffiltext{1}{Universities Space Research
8: Association, NASA Marshall Space Flight Center, VP62, Huntsville,
9: AL, USA} \altaffiltext{2}{4014 Emerald Street No.116, Torrance CA,
10: USA 90503 and International Center for Relativistic Astrophysics,
11: I.C.R.A.,University of Rome La Sapienza, I-00185 Roma, Italy,
12: brian.m.punsly@L-3com.com or brian.punsly@gte.net}
13:
14: \begin{abstract}
15: It has been shown based on radio variability arguments that some
16: BALQSOs (broad absorption line quasars) are viewed along the polar
17: axis (orthogonal to accretion disk) in the recent article of Zhou et
18: al. These arguments are based on the brightness temperature, $T_{b}$
19: exceeding $10^{12}\,^{\circ}$ K which leads to the well-known
20: inverse Compton catastrophe unless the radio jet is relativistic and
21: is viewed along its axis. In this letter, we expand the Zhou et al
22: sample of polar BALQSOs to the entire SDSS DR5. In the process, we
23: clarify a mistake in their calculation of brightness temperature.
24: The expanded sample of high $T_{b}$ BALQSOS, has an inordinately
25: large fraction of LoBALQSOs (low ionization BALQSOs). We consider
26: this an important clue to understanding the nature of the polar
27: BALQSOs. This is expected in the polar BALQSO analytical/numerical
28: models of Punsly in which LoBALQSOs occur when the line of sight is
29: very close to the polar axis, where the outflow density is the
30: highest.
31: \end{abstract}
32:
33: \keywords{(galaxies:) quasars: absorption lines --- galaxies: jets
34: --- (galaxies:) quasars: general --- accretion, accretion disks --- black hole physics}
35:
36: \section{Introduction}About 15\% - 20\% of quasars show broad UV absorption lines (loosely
37: defined as absorbing gas that is blue shifted at least 5,000 km/s
38: relative to the QSO rest frame and displaying a spread in velocity
39: of at least 2,000 km/s) \citep{wey97,hew03,rei03}. The implication
40: is that large volumes of gas are expelled as a byproduct of
41: accretion onto the the central supermassive black hole.
42: Understanding the details of this ejection mechanism is a crucial
43: step towards understanding the physics of the central engine of
44: quasars. Although evolutionary processes might be related to BAL
45: outflows, it is widely believed that all radio quiet quasars have
46: BAL flows, but the designation of a quasar as a BALQSO depends on
47: whether the line of sight intersects the solid angle subtended by
48: the outflow. The standard model of quasars is one of a hot accretion
49: flow onto a black hole and a surrounding torus of molecular gas
50: \citep{ant93}. The BAL outflow can be an equatorial wind driven from
51: the luminous disk that is viewed at low latitudes, just above the
52: molecular gas, \citet{mur95}, or a bipolar flow launched from the
53: inner regions of the accretion flow \citep{pun99,pun00}. The most
54: general feature of 3-D simulations of accretion flows onto black
55: holes is huge ejections of gas from the central vortex of the
56: accretion flow along the polar axis which rivals the accreted mass
57: flux for rapidly spinning black holes \citep{dev05,haw06}. Thus, it
58: is of profound theoretical interest to look for evidence of these
59: endemic polar ejecta.
60: \par BALQSOs are so distant that direct imaging of the BAL
61: region is beyond the resolution of current optical telescopes. Thus,
62: much of the discussion of BAL geometry is based on deductive
63: reasoning. A novel idea for finding the orientation of BALQSOs was
64: developed in \citet{zho06}. They used radio variability information
65: to bound the size of the the radio emitting gas then deduced that
66: the radio emission must be viewed close to the polar axis and
67: emanate from a relativistic jet, thereby avoiding the well known
68: inverse Compton catastrophe. They found BALQSOs that satisfy these
69: conditions in SDSS DR3. We expand their methods to SDSS DR5 and
70: apply these results to the physics of the BAL wind launching
71: mechanism.
72: \section{The Brightness Temperature} Equation (5) of \citet{zho06} is not well defined since it is unclear
73: what frame of reference is used to evaluate the quantities contained
74: within. Apparently, this gave rise to estimates that are $(1+z)^{3}$
75: larger than what we find in equation (2.4), below. In this section,
76: we derive a formula for $T_{b}$. Physically, it is $T_{b}$ evaluated
77: in the cosmological frame of reference of the quasar, $(T_{b})_{q}$,
78: that is relevant for assessing the "inverse Compton catastrophe." We
79: want to express this in terms of observable quantities at earth
80: designated by the subscript "o." First of all, the brightness
81: temperature is the equivalent blackbody temperature of the radiation
82: assuming one is in the Planck regime, $h \nu \ll k_{b}T$. Consider a
83: source in which the monochromatic intensity has increased by an
84: amount $I(\nu)_{q}$ in a time $\Delta t_{q}$. The brightness
85: temperature associated with the change in luminosity is
86: \begin{eqnarray}
87: && (T_{b})_{q}= \frac{I(\nu)_{q} \mathrm{c}^{2}}{2 k_{b}
88: \nu_{q}^{2}}\;.
89: \end{eqnarray}
90: From the monochromatic version of Liouville's theorem, \citet{gun78}
91: \begin{eqnarray}
92: && \frac{(F_{\nu})_{o}}{\Omega_{o}}\equiv I(\nu)_{o} =
93: \frac{1}{(1+z)^{3}} I(\nu)_{q}\;,
94: \end{eqnarray}
95: $(F_{\nu})_{o}$, is the flux density observed at earth. The solid
96: angle subtended by the source, $\Omega_{o}$, is bounded by the
97: causality requirement that the source could not have expanded more
98: rapidly than the speed of light during a time, $\Delta t_{q}$,
99: \citet{gun78},
100: \begin{eqnarray}
101: && \Omega_{o}=\frac{\mathrm{proper\; area\; perpendicular\; to \;
102: line \; of \; sight}}{(\mathrm{angular\; diameter\; distance})^{2}}
103: \leq \frac{(c \pi \Delta t_{q})^{2}}{4 d_{A}^{2}} \;,
104: \end{eqnarray}
105: where the angular diameter distance is $d_{A}$. In a a cosmology
106: with $H_{0}$=70 km/s/Mpc, $\Omega_{\Lambda}=0.7$ and
107: $\Omega_{m}=0.3$, we can use the expression for $d_{A}$ given by
108: \citet{pen99}, that is accurate to $<1 \% $ relative error, along
109: with (2.1) -(2.3) to find:
110: \begin{mathletters}
111: \begin{eqnarray}
112: && (T_{b})_{q}\approx \frac{8.0\times 10^{12}
113: (1+z)}{(\nu_{o}/\mathrm{1GHz})^{2}(\Delta t_{o}/ 1 \mathrm{yr})^{2}}
114: Z^{2} (\Delta F_{\nu}(\mathrm{mJy}))_{o}\,^{\circ}\,\mathrm{K}\;,\\
115: && Z \equiv 3.31-(3.65) \nonumber \\
116: &&\times\left(\left[(1+z)^{4}-0.203(1+z)^{3}+0.749(1+z)^{2}
117: +0.444(1+z)+0.205\right]^{-0.125}\right)\;,
118: \end{eqnarray}
119: \end{mathletters}
120: where $\Delta F_{\nu}(\mathrm{mJy})$ is the change in flux density
121: in mJy measured at earth at frequency $\nu_{o}$ during the time
122: interval $\Delta t_{o}$.
123: \par When $(T_{b})_{q}> 10^{12}\,^{\circ}\,\mathrm{K}$, the inverse
124: Compton catastrophe occurs. Most of the electron energy is radiated
125: in the inverse Compton regime. The radio synchrotron spectrum from
126: the jet is diminished in intensity to unobservable levels
127: \citep{kel69}. In order to explain the observed radio synchrotron
128: jet in such sources, Doppler boosting is customarily invoked to
129: resolve the paradox. Recall that for an unresolved source the
130: observed flux density, $(F_{\nu})_{o}$, is Doppler enhanced relative
131: to the intrinsic flux density,
132: $(F_{\nu})_{o}=\delta^{3+\alpha}(F_{\nu})_{\mathrm{intrinsic}}$,
133: where the Doppler factor, $\delta$ is given in terms of $\Gamma$,
134: the Lorentz factor of the outflow; $\beta$, the three velocity of
135: the outflow; and $\alpha$, is the spectral index,
136: $\delta=1/[\Gamma(1-\beta\cos{\theta})]$ \citep{lin85}. Thus,
137: Doppler beaming can cause $(T_{b})_{q}$ to be enhanced by a factor
138: of $\delta^{3+\alpha}$ in (2.4). Assuming a flat spectral index for
139: the unresolved relativistic jet, we choose $\alpha=0$. This allows
140: us to find a minimum Doppler factor that avoids the inverse Compton
141: catastrophe, $\delta_{min}[(T_{b})_{q}]=[(T_{b})_{q}/
142: (10^{12}\,^{\circ}\,\mathrm{K})]^{0.333}$. If the jet plasma
143: propagates with $\delta > \delta_{min}[(T_{b})_{q})]$ then the
144: intrinsic $T_{b}$ in the frame of reference of the jet plasma is
145: sufficiently low that there is no inverse Compton catastrophe.
146: \begin{table}
147: \caption{Properties of the Radio Variable BAL
148: Quasars\tablenotemark{a}} {\footnotesize \begin{tabular}{ccccccccc}
149: \tableline \rule{0mm}{3mm}
150: Name & $z$ & $S^{P}_{F}$:date & $S^{P}_{N}$:date & $\sigma$ & BI CIV & BI AlIII & BI MgII & Type \\
151: (SDSS J) & & (mJy) & (mJy) & & (km/s) &(km/s) & (km/s) & \\
152: \tableline \rule{0mm}{3mm}
153: 081102.91+500724.5\tablenotemark{b} &1.84 & 23.07: 05/23/97 & 18.80: 11/15/93 & 5.7 & 1568 & 468 & ... & LoBAL \\
154: 081618.99+482328.4& 3.57 & 72.29: 05/01/97 & 61.10: 11/15/93 & 5.2 & 2576 & ... &....& HiBAL \\
155: 081839.00+313100.1& 2.37 & 8.82: 10/23/95 & 7.2: 12/15/93 & 3.1 & 6445 & ... &....& HiBAL \\
156: 082817.25+371853.7\tablenotemark{c} &1.35 & 21.18: 07/23/94 & 14.50: 12/15/93 & 10.5 & ... & 428 & 1564 & LoBAL \\
157: 093348.37+313335.2& 2.60 &18.35: 10/23/95 & 15.90: 12/15/93 & 3.7 & 4107 & 1405 &....& LoBAL \\
158: 104106.05+144417.4&3.01&27.46: 12/99 & 19.00: 12/06/93 & 11.4 & 2527 & ... & .... & HiBAL \\
159: 113445.83+431858.0 & 2.18 &27.38: 2/20/97 & 24.90: 11/15/93 & 3.4 & 10643 & ... & 1149 & LoBAL \\
160: 134652.72+392411.8& 2.47 &3.60: 08/19/94 & 2.20: 04/16/95 & 3.0 & 1505 & 557 & .... & LoBAL \\
161: 142610.59+441124.0& 2.68 & 6.78: 03/27/97 & 5.0: 03/12/95 & 3.6 & 8652 & 1389 & .... & LoBAL \\
162: 145926.33+493136.8 \tablenotemark{d}&2.37 & 5.22: 04/17/97 & 3.60: 03/12/95 & 3.4 & 9039 & 329 & ....& LoBAL \\
163: 155633.77+351757.3 &1.49 &30.92: 07/03/94 & 26.90: 04/16/95 & 4.3 & e & ... & .... & FeLoBAL \\
164: 165543.24+394519.9 &1.75 &10.15: 08/19/94 & 8.50: 04/16/95 & 3.0 & 5805 & ... & .... & HiBAL \\
165: \end{tabular}}
166: \tablenotetext{a}{The first column in table 1 is the source name,
167: followed by the redshift. Columns (3) and (4) are the FIRST and NVSS
168: peak flux densities at 1.4 GHz, respectively. Column (5) is the
169: statistical significance of the radio variability computed using
170: (3.1). Columns (6) - (8) are the BALnicity indices, \citet{wey91},
171: for CIV, AlII and MgII, respectively. The last column denotes
172: whether the source is a low ionization or a high ionization BALQSO.}
173: \tablenotetext{b}{There is a new spectrum in SDSS with much higher
174: S/N than the spectrum in \citet{zho06} that
175: clearly shows broad Al III absorption}
176: \tablenotetext{c}{from \citet{zho06}, however there is now a much
177: higher signal to noise spectrum in SDSS than the one used in
178: \citet{zho06} and we have recomputed the BALnicity indices}
179: \tablenotetext{d}{from \citet{zho06}} \tablenotetext{e}{identified
180: as FeLoBALQSO in \citet{bec97}}
181: \end{table}
182: \section{Polar BALQSOs} Table 1 is a list of BALQSOs in which the
183: radio variability requires that the jet is propagating well within
184: $35\,^{\circ}$ to the line of sight in order for the jet plasma to
185: satisfy $(T_{b})_{q} < 10^{12}\,^{\circ}\,\mathrm{K}$. Since rather
186: small variations of flux density create these conditions we must
187: first prove that the sources are truly variable. We choose the
188: condition for variability to be
189: \begin{equation}
190: \sigma_{var}=\frac{S_{FP}-S_{NP}}{\sqrt{\sigma^2_{FP}+\sigma^2_{NP}}}>+3\;,
191: \end{equation}
192: where $S_{FP}$ and $S_{NP}$ denote the peak flux density at 1.4 GHz
193: measured by the FIRST and NVSS surveys, respectively, $\sigma_{FP}$
194: and $\sigma_{NP}$ are the FIRST and NVSS peak flux uncertainties,
195: respectively. The beam for the NVSS survey is considerably larger
196: than FIRST ($FWHM\sim 45^{''}$ versus $FWHM\sim 5^{''}$). Thus, a
197: compact source $\ll 5^{''}$ (like a compact radio core which is the
198: putative site of variable activity) should be detected with equal
199: sensitivity if it exceeds the flux limit of the samples. The NVSS
200: beam should pick up more extended flux than FIRST. Variable BALQSO
201: fields in NVSS with confusion from nearby sources were omitted from
202: table 1. Thus, a FIRST peak flux density larger than an NVSS peak
203: flux density by $3\sigma$ (\textbf{hence the plus sign on the RHS of
204: eqn. (3.1)}) is true variability and not an artifact of the
205: different beam sizes. One improvement from \citet{zho06} is that we
206: used the peak NVSS flux density as opposed to the integrated NVSS
207: flux density. This is preferred for the comparison described in
208: (3.1), since the variable radio flux is most likely coming from
209: sub-arcsecond regions of the jet and this choice is less sensitive
210: to low surface brightness noise that can be picked up in the
211: integrated flux \citep{con97}.
212: \par It is crucial to assess the significance of the $3 \sigma$
213: results in table 1. In order to clarify the statistics, we must
214: delineate the super sample from which the sources in table 1 are
215: drawn. The potential variable BALQSOs in DR5 that can be detected by
216: equation (3.1) require two conditions. First of all, the sources
217: must have a BALnicity index $> 0$ as defined in \citet{wey91} for at
218: least one line either, CIV, AlIII, or MgII. Secondly, $ S_{FP} >
219: 3.55$ mJy. The NVSS survey has a peak flux limit of 2.5 mJy in
220: general, but occasionally as low as 2.2 mJy and a minimum
221: measurement uncertainty of $\approx 0.45$ mJy \citet{con98}. Thus, $
222: S_{FP} < 3.55$ mJy could never satisfy eqn. (3.1). There are 116
223: BALQSOs that satisfy these criteria and they form the master sample.
224: Eqn. (3.1) is a one sided probability, so if the errors are Gaussian
225: distributed as in \citet{con97,con98} then there is less than a 20\%
226: chance of expecting a variable BALQSO from this sample of 116.
227: However, we found 20 variable BALQSOs (12 of which made the table
228: because $T_{B}$ was sufficiently high to restrict the line of sight
229: to $< 35^{\circ}$). Thus, based on Gaussian statistics our findings
230: are very significant. However, if the errors in the tail of the
231: distribution are not Gaussian, but arise from unforseen sources of
232: measurement error, then the analysis of \citet{con97,con98} does not
233: apply and we have no means to assess the significance of our
234: results. Note that there are 4 sources in table 1 with a variability
235: that is significant above the $5\sigma$ level.
236:
237: \begin{table} \caption{Line of Sight to Radio Variable BAL
238: Quasars} {\footnotesize
239: \begin{tabular}{ccccc} \tableline \rule{0mm}{3mm}
240: Name & $T_{b}$ & $\delta_{\mathrm{min}}$ & $\theta_{\mathrm{max}}$ & Type \\
241: (SDSS J) & ($10^{12}$$\,^{\circ}$ K) & & (degrees) & \\
242: \tableline \rule{0mm}{3mm}
243: 081102.91+500724.5 & 5.4 & 1.7 & 34.9 & LoBAL \\
244: 081618.99+482328.4& 39.6 & 3.4 & 17.1 & HiBAL \\
245: 081839.00+313100.1& 11.8 & 2.3 & 26.1 & HiBAL \\
246: 082817.25+371853.7& 125.4 & 5.0 & 11.5 & LoBAL \\
247: 093348.37+313335.2& 20.2 & 2.7 & 21.6 & LoBAL \\
248: 104106.05+144417.4& 8.5 & 2.0 & 29.4 & HiBAL \\
249: 113445.83+431858.0 & 5.9 & 1.9 & 33.7 & LoBAL \\
250: 134652.72+392411.8& 73.9 & 4.2 & 13.8 & LoBAL \\
251: 142610.59+441124.0& 11.5 & 2.3 & 26.4 & LoBAL \\
252: 145926.33+493136.8 & 8.8 & 2.1 & 29.0 & LoBAL \\
253: 155633.77+351757.3 & 66.5 & 4.0 & 14.3 & FeLoBAL \\
254: 165543.24+394519.9 & 47.5 & 3.6 & 16.1 & HiBAL \\
255:
256: \end{tabular}}
257:
258: \end{table}
259: \par Our BALQSO identifications and the BALnicity indices in table 1 were derived from spectra in the SDSS DR5
260: public archive. The SDSS spectra of radio-variable BALs were
261: retrieved from the SDSS database and were analyzed using the
262: $IRAF$\footnote{IRAF is the Image Reduction and Analysis Facility,
263: written and supported by the IRAF programming group at the National
264: Optical Astronomy Observatories (NOAO) in Tucson, Arizona.}
265: software. First, the spectrum was de-reddened using the Galactic
266: extinction curve, \citet{sch98}, then the wavelength scale was
267: transformed from the observed to the source frame. The spectra were
268: fitted in XSPEC with a powerlaw plus multiple gaussian model,
269: including fits for both emission lines and absorption lines
270: \citep{arn96}. All the model parameters were kept free. The best fit
271: to the SDSS data was determined using $\chi ^{2}$ minimization. If
272: an emission bump around 2500 $\AA$ is present then the continuum fit
273: in the CIV (1550 $\AA$) region was extrapolated to longer
274: wavelength. The BALnicity indices quoted in table 1 are a
275: consequence of the method of spectral fitting described above and
276: other methods might produce different results. However, the exact
277: BALnicity index is not critical to this discussion, the sources in
278: table 1 clearly show BALs and that is the essential point of
279: relevance here.
280: \par In table 2, we calculate $(T_{b})_{q}$, using equation (2.4)
281: and the peak fluxes from table 1. From this, we calculate
282: $\delta_{min}[(T_{b})_{q}]$ in column (3) as described at the end of
283: section 2. For each value of $\delta_{min}$, one can vary $\beta$ in
284: the definition of $\delta$ to find the maximum value of $\theta$,
285: $\theta_{max}\{\delta_{min}[(T_{b})_{q}]\}$, that is compatible with
286: $\delta_{min}$:
287: \begin{equation}
288: \theta_{max}\{\delta_{min}[(T_{b})_{q}]\}=\mathrm{Max}_{\mid_\beta}\left(\arccos\left\{\left[1-\left(\frac{\sqrt{1-\beta^{2}}}{\delta_{min}[(T_{b})_{q}]}\right)\right]\beta^{-1}\right\}\right)\;.
289: \end{equation}
290: The value of $\theta_{max}$ in column (4) is an extreme upper limit
291: because it is the third in a chain of bounds. \begin{itemize}
292: \item The value of $(T_{b})_{q}$ in table 1 is a lower bound, since
293: NVSS might pick up some extended flux that is missed by the peak
294: FIRST measurement and the radio core was actually weaker at the time
295: of the NVSS measurement than indicated by the peak flux density,
296: i.e., $(\Delta F_{\nu}(\mathrm{mJy}))_{o}$ is underestimated in
297: (2.4). Also, (2.3) is an inequality. From the discussion at the end
298: of section 2, the larger $(T_{b})_{q}$, the larger $\delta_{min}$
299: \item the larger $\delta_{min}$, the lower $\theta_{max}$ by (3.2)
300: and the true value of the jet plasma $\delta$ can be much larger
301: than $\delta_{min}[(T_{b})_{q}]$
302: \item the larger $\delta_{min}$, the lower $\theta_{max}$ by (3.2)
303: and the true value of the jet plasma $\delta$ can be much larger
304: than $\delta_{min}[(T_{b})_{q}]$
305: \item The actual line of sight to earth satisfies
306: $\theta<\theta_{max}$ by definition.
307: \end{itemize} Thus the condition for inclusion
308: into table 2, $\theta_{max}< 35\,^{\circ}$, means that the jet is
309: likely to be propagating very close to the pole.
310:
311: \section{The Theory of Polar BALQSOs}
312: Figure 1 is a small modification to Fig. 9 of \citet{pun00}. In
313: section 4.5 of \citet{pun00}, it is discussed how a relativistic jet
314: (in red) can coexist nested inside of the bipolar BAL wind. This jet
315: can emit radio flux that is beamed to within a small angle of the
316: polar axis. Within the axisymmetric version of the model, LoBALQSOs
317: exist for lines of sight within $15^{\circ}$ of the polar axis. The
318: density of the wind is highest nearest the polar axis as indicated
319: by the gray shading in figure 1 and LoBALQSOs are viewed closer to
320: the polar axis on average than HiBALQSOs. The accretion disk coronal
321: X-rays are screened from the BAL wind by the dense base of the jet
322: which provides hydrogen column densities of $\Sigma_{H}\sim
323: 10^{25}\mathrm{cm^{-2}}$. Similarly, the different lines of sight
324: near the polar axis, through the inhomogeneous wind, naturally
325: provide larger Compton scattering columns and more attenuation of
326: the far UV flux from the inner disk and less attenuation of the near
327: UV and optical flux from the outer disk, making the spectrum appear
328: red. The more extreme and highly polarized LoBALs obviously occur in
329: objects without a perfectly axisymmetric distribution of dust as
330: discussed in detail in \citet{pun00}. Recall that the bipolar BAL
331: wind does not preclude the coexistence of a BAL wind from the outer
332: regions of the accretion disk as envisioned by \citet{mur95}.
333: \par Since the LoBALQSOs are viewed closer to
334: the jet axis than the HiBALQSOs in the axisymmetric model, geometric
335: arguments imply they should typically have jets with larger Doppler
336: factors. Thus, LoBALQSOs should occur in samples of variable BALQSOs
337: at an inordinately high rate if the axisymmetric version of the
338: polar model applies to a significant subpopulation of the BALQSOs.
339: In our master sample of 116 FIRST, DR5 BALQSOs defined in section 3,
340: 69 are HiBALQSOs and 47 are LoBALQSOs. Based on table 1, 4/69 =
341: 5.80\% of HiBALQSOs and 8/47 = 17.02\% of the LoBALQSOs have
342: $(T_{b})_{q}> 5\times 10^{12}\,^{\circ}\,\mathrm{K}$ (this condition
343: is equivalent to $\theta < 35^{\circ}$). The sample is small, but
344: the fact that in table 1, the LoBALQSO likelihood to have a large
345: $T_{b}$ is 2.94 times that of the HiBALQSOs tends to support the
346: notion the polar BALQSO model represents a significant subpopulation
347: of BALQSOs.
348: \section{Conclusion} Using radio variability arguments,
349: we expanded on the sample of known polar BALQSOs begun by
350: \citet{zho06}. In the process, we noted that these radio variable
351: BALQSOs have an inordinately large LoBALQSOs subpopulation. It is
352: interesting that these properties are expected based on an existing
353: detailed theoretical treatment and modeling of bipolar BAL winds
354: \citep{pun99,pun00}. It would be informative to continue to monitor
355: the FIRST BALQSOs at 1.4 GHz, with matched resolution, in order to
356: find more $(T_{b})_{q}> 10^{12}\,^{\circ}\,\mathrm{K}$ BALQSOs and
357: improve the statistical information in table 2.
358:
359: \begin{thebibliography}{}
360: \bibitem[Antonucci(1993)]{ant93} Antonucci, R.J. 1993,
361: Annu. Rev. Astron. Astrophys. \textbf{31} 473
362: \bibitem[Arnaud(1996)]{arn96} Arnaud, K.A., 1996,
363: \textbf{Astronomical Data Analysis Software and Systems V}, eds. Jacoby G. and Barnes J., ASP Conf. Series volume 101, p17
364: \bibitem[Becker et al.(1997)]{bec97} Becker, R.H., Gregg,
365: M.D., Hook, I.M., McMahon, R.G., White, R.L., \& Helfand, D.J. 1997,
366: ApJL textbf{479} 93
367: \bibitem[Condon(1997)]{con97} Condon, J.J. 1997, PASP \textbf{109}
368: 1149
369: \bibitem[Condon et al (1998)]{con98} Condon, J.J. et al 1998, AJ \textbf{115}
370: 1693
371: \bibitem[De Villiers et al (2005)]{dev05} De Villiers, J-P., Hawley, J., Krolik, J.,Hirose, S.
372: 2005, ApJ \textbf{620} 878
373: \bibitem[Gunn(1978)]{gun78}Gunn, J. 1978 in \emph{Observational Cosmology, Eight Advance Course, Swiss Society of Astronomy and Astrophysics}, p. 26 eds
374: A. Maeder, L. Martinet and G. Tammann (Geneva Observatory: Sauverny
375: Switzerland)
376: \bibitem[Hawley and Krolik (2006)]{haw06} Hawley, J., Krolik, K.
377: 2006, ApJ \textbf{641} 103
378: \bibitem[Hewett and Foltz (2003)]{hew03}Hewett, P. and Foltz, C., 2003 AJ
379: \textbf{125} 1784
380: \bibitem[Kellermann \& Pauliny-Toth(1969)]{kel69} Kellermann,
381: K. I., \& Pauliny-Toth, I. I. K. 1969 ApJ, \textbf{155}, L71
382: \bibitem[Lind and Blandford(1985)]{lin85}Lind, K., Blandford, R.
383: 1985, ApJ \textbf{295} 358 345
384: \bibitem[Murray et al(1995)]{mur95} Murray, N. et al 1995, ApJ \textbf{451}
385: 498
386: \bibitem[Pen(1999)]{pen99}Pen, U.-L. 1999, ApJS \textbf{120} 49
387: \bibitem[Punsly(1999a)]{pun99}Punsly, B. 1999, ApJ \textbf{527} 609
388: \bibitem[Punsly(1999b)]{pun00}Punsly, B. 1999, ApJ \textbf{527} 624
389: \bibitem[Punsly and Lipari (2005)]{pun05}Punsly, B. 2005, ApJL \textbf{623}
390: 101
391: \bibitem[Reichard et al.(2003)]{rei03} Reichard, T.A., et
392: al. 2003, AJ \textbf{126} 2594
393: \bibitem[Schlegel et al.(1998)]{sch98} Schlegel, D. J., Finkbeiner, D. P. \& Davis, M. 1998, ApJ \textbf{500}
394: 525
395: \bibitem[Weymann et al.(1991)]{wey91} Weymann, R.J., Morris,
396: S.L., Foltz, C.B., Hewett, P.C. 1991, ApJ \textbf{373}, 23
397: \bibitem[Weymann (1997)]{wey97}Weymann, R. 1997 in ASP Conf. Ser.
398: 128, \textbf{Mass Ejection from Active Nuclei} ed, N.Arav, I.
399: Shlosman and R.J. Weymann (San Francisco: ASP) 3
400: \bibitem[Zhou et al(2006)]{zho06}Zhou, H. et al 2006 ApJ \textbf{639}
401: 716
402: \end{thebibliography}
403: \clearpage
404:
405: \begin{figure}
406: \begin{center}
407: \includegraphics[scale=1.0]{f1.eps}
408: \end{center}
409: \caption{The bipolar wind model of \citet{pun99,pun00}. The gray clouds of the BAL wind has an opening angle of $\sim 25^{\circ}$ - $30^{\circ}$.
410: Nested inside is a relativistic jet (in red). The highest BAL wind densities are near the polar axis. Thus, lines of sight
411: near the polar axis, have the maximum attenuation of X-ray and ionizing UV
412: radiation from the accretion flow. This implies that lines of sight
413: close to the polar axis are more likely to represent LoBALQSOs.}
414: \end{figure}
415:
416: \clearpage
417:
418: \end{document}
419: