astro-ph0103455/ms.tex
1: \documentclass[]{aastex}
2: \usepackage[onecolumn]{emulateapj5}
3: \usepackage{graphicx}
4: \begin{document}
5: \newcommand{\gmet}{{\rm \bf g}}
6: \newcommand{\tmet}{\rm \bf T}
7: \newcommand{\nablamet}{\bf \nabla}
8: \newcommand{\sonicsub}{\bf sonic}
9: \title{On Steady State 
10: Neutrino Heated Ultra-Relativistic Winds from Compact Objects}
11: \def\question#1{{{\marginpar{\tiny \sc }}}}
12: %\def\question#1{{{\marginpar{\tiny \sc #1}}}}
13: 
14: \author{Jason Pruet\altaffilmark{1}, George M. Fuller\altaffilmark{1}, and 
15: Christian Y. Cardall\altaffilmark{2,3,4}}
16: \altaffiltext{1}{Department of Physics, University of California, San Diego, La
17: Jolla, California, 92093-0319}
18: \altaffiltext{2}{Physics Division, Oak Ridge National Laboratory, Oak Ridge,
19:  	TN 37831-6354}
20: \altaffiltext{3}{Department of Physics and Astronomy, University of Tennessee,
21: 	Knoxville, TN 37996-1200}
22: \altaffiltext{4}{Joint Institute for Heavy Ion Research, Oak Ridge National
23: 	Laboratory, Oak Ridge, TN 37831-6374}
24: %\email{jpruet@physics.ucsd.edu, gfuller@ucsd.edu, ccardall@mail.phy.ornl.gov}
25: \begin{abstract}
26: 
27: We study  steady state winds from 
28: compact objects in the regime where the wind velocity at 
29: infinity is ultra-relativistic. This may have relevance to some models of 
30: Gamma-Ray-Bursts (GRB's). Particular attention is paid to the
31: case where neutrinos provide the heating. 
32: Unless the neutrino luminosity is 
33: very large, $L>10^{54} {\rm erg/s}$, the only allowed steady state solutions
34: are those where energy deposition is 
35: dominated by neutrino-antineutrino annihilation at the sonic point.
36: In this case, the matter temperature near the
37: neutron star surface is low, less than 1MeV for typical neutrino 
38: luminosities. This is in contrast to the case for
39: sub-relativistic winds discussed in the context of supernovae 
40: where the matter temperature near the neutron
41: star approximates the temperature characterizing the neutrinos. 
42: We also investigate the setting of the neutron to proton ratio (n/p) in
43: these winds and find that only for large ($>10{\rm  MeV}$) 
44: electron neutrino or electron anti-neutrino temperatures 
45: is $n/p$ entirely determined by neutrino capture. Otherwise,
46: $n/p$ retains an imprint of conditions in the neutron star.
47: 
48: 
49: \end{abstract}
50: 
51: \keywords{Stars: Neutron - Gamma-rays: Bursts - Neutrinos}
52: 
53: \section{Introduction} 
54: \question{added affiliations (CYC)}
55: \question{Commented out email addresses because they were appearing
56: under the authors (CYC)}
57: 
58: We examine properties of ultrarelativistic winds from
59: compact objects and we thereby obtain insights into the special
60: case where these winds arise from neutrino heating. 
61:  Our principal motivation for the present study is to 
62: determine if physically plausible steady state solutions exist in the 
63: ultrarelativistic 
64: regime and to determine the relation between neutrino parameters and the
65: character of the outflow.  A peculiar feature of the steady state ultrarelativistic solutions
66: is that the matter temperature near the neutron star surface is cold. We comment below
67: on whether or not this is physically realizable. We also study how the final neutron to 
68: proton ratio (or alternatively electron fraction, $Y_e=1/(n/p+1)$), is set in these 
69: models. Only for large ($>10{\rm MeV}$) 
70: electron neutrino or electron anti-neutrino temperatures does $Y_e$ in the initial 
71: wind material come into steady state equilibrium with the neutrino fluxes.
72: 
73: 
74: There is a rich literature concerning steady state outflow from neutron 
75: stars. In connection with supernovae the neutrino-driven wind occuring
76: several seconds post core bounce is interesting because it is the favored 
77: candidate for the site of the r-process elements \citep{meyer92,wilsonandmathews,takahashi94}. 
78: The relation between 
79: neutrino parameters and the wind have been extensively discussed \citep{DSW,QW}.
80: These studies were concerned with gentle, modest entropy winds, characterized
81: by Lorentz factors 
82: %far from the star of order unity. 
83: of order unity far from the star.
84: \question{phrase rearranged (CYC)} 
85: Near the 
86: neutron star surface these gentle winds are well approximated as 
87: %rs Files Tools Edit Search Mule TeX Help     
88: \question{random text deleted (CYC)}
89: being in hydrostatic equilibrium. The temperature is set by the competition
90: between neutrino heating and neutrino cooling terms, i.e., the matter
91: temperature is roughly equal to the temperature characterizing the
92: neutrinos. 
93: %Modulo 
94: Apart from
95: \question{word changed (CYC)}
96: effects due to nucleosynthesis, the final electron
97: fraction in these gentle winds is set by the ratio of electron
98: neutrino ($\nu_e$) and electron anti-neutrino ($\bar{\nu}_e$) capture rates
99: on nucleons \citep{Qetal93}. 
100: The general relativistic extension of the wind equations, redshift factors,
101: and neutrino trajectory bending have been discussed by \cite{fullerqian96}, 
102: \cite{Cardall}, and \cite{Salmonson}. 
103: These works also emphasize 
104: %on 
105: subrelativistic outflow velocities.
106: %An important result of
107: %these studies is that the bending of null trajectories in a Schwarzschild 
108: %geometry leads to a substantial enhancement of the neutrino-antineutrino 
109: %heating rate. 
110: \question{sentence deleted (CYC)}
111: Detailed studies of relativistic effects along these lines 
112: for r-process nucleosynthesis have been carried out recently by \cite{japanese1}.
113:  
114: In a pioneering study \cite{Paczynski} examined ultrarelativistic flow from neutron stars
115: relevant for studies of GRB's. In this work no reference was made to a heating
116: term and, except for photon diffusion which is unimportant near the
117: neutron star surface, the flow was assumed adiabatic. This work provided 
118: a detailed picture of the wind evolution far from the neutron star 
119: surface, which is the regime relevant for optical emission in GRB's. There are many 
120: similarities between these steady state ultra-relativistic winds and 
121: the fireballs discussed in connection with GRB's \citep{Piran}.
122:  
123: The present study is an attempt to
124: fill the gap between detailed studies of the relation between neutrino 
125: heating and sub-relativistic outflow and the study of ultra-relativistic
126: outflow without reference to a heating mechanism. This is also
127: a detailed look at the `guts' of a simple model for a GRB central engine.
128: It may be applicable to the model proposed by \cite{SWM}. 
129:  In this model binary neutron stars are compressed and heated as 
130: gravitational radiation saps orbital kinetic energy during the last 
131: seconds of inspiral. The heating drives copious neutrino loss 
132: ($L_{\nu}\sim 10^{52} {\rm erg \, s^{-1}}$) which in turn drives a
133: hot pair plasma from the neutron star surface.   
134: 
135: Lastly, it has recently been argued that the electron fraction in neutrino-heated
136: fireballs may be a clue to neutrino physics at the central engine, and may have
137: observable consequences \citep{FPA}. The present study serves as a ``proof of 
138: concept'' of this idea for a simple case. Interestingly, we find that the electron 
139: fraction in the wind far from the central engine is set by processes very near the
140: central engine. If this is a generic feature of GRB central engine models,
141: information about the electron fraction in GRB's may help pin down the central
142: engine.
143: 
144: 
145: 
146: 
147: 
148: \section{Equations}
149: 
150: 
151: 
152: The general relativistic equations describing the outflow of material from 
153: the neutron star are Euler's equations: 
154: 
155: \begin{eqnarray} 
156: {\rm \bf u}\cdot \left(\nablamet \tmet \right) 
157: &=&- {\rm \bf u}\cdot \left(\nablamet \tmet_{\nu}\right) \cr 
158: \left(\gmet+{\rm \bf u} \otimes {\rm \bf u}\right)\cdot \left(\nablamet  \tmet \right)
159: &=&- \left(\gmet+{\rm \bf u} \otimes {\rm \bf u}\right)\cdot \left(\nablamet \tmet_{\nu}\right),
160: \end{eqnarray}
161: 
162: baryon number conservation,
163: 
164: \begin{equation}
165: \nablamet \cdot \left(\rho_{\rm b} {\rm \bf u}\right)=0,
166: \end{equation}
167: 
168: and the change in electron fraction due to lepton capture on baryons,
169: 
170: \begin{equation} 
171: \label{dye}
172: {\rm \bf u}\cdot \left(\nablamet Y_e\right)=\left(1-Y_e\right)\left(\lambda_{{\nu}_e{\rm n\rightarrow p}}
173:  +\lambda_{e^+{\rm n\rightarrow p}}\right)-
174: Y_e\left(\lambda_{\bar{\nu}_e{\rm p\rightarrow n}}+\lambda_{e^-{\rm p\rightarrow n}}\right).
175: \end{equation}
176: 
177: In the above ${\rm \bf u}$ is the 4-velocity 
178: of the outflowing material, $\tmet$ and $\tmet_{\nu}$ are the stress-energy tensors 
179: for the outflowing matter and the neutrinos respectively, $\gmet$ is the metric 
180: tensor, and $\rho_{\rm b}$ is the rest mass energy density of baryons in the plasma
181: comoving frame. In Eq. (\ref{dye}) $\lambda_{\nu_e{\rm n\rightarrow p}}$ and 
182: $\lambda_{\bar{\nu}_e{\rm p\rightarrow n}}$ are the capture rates for electron anti-neutrinos
183: and electron neutrinos on neutrons and protons respectively, and 
184: $\lambda_{e^+{\rm n\rightarrow p}}$ and $\lambda_{e^-{\rm p\rightarrow n}}$ are 
185: the capture rates for positrons and electrons on neutrons and protons respectively.
186: In what follows we employ the Schwarzschild metric and use 
187: natural units,  $G=c=\hbar=k_b=1$, with $k_b$ the Boltzmann constant.
188: In 
189: %the 
190: \question{word deleted (CYC)}
191: Schwarzschild coordinates the line element is  
192: \begin{equation}
193: ds^2=\gmet _{\mu \nu} dx^{\mu} dx^{\nu}=
194: -\left(1-2M/r\right)dt^2+\left(1-2M/r\right)^{-1}dr^2+r^2d\Omega^2, 
195: \end{equation}
196: Here $M$ is the mass of the neutron star and $d\Omega^2=d\theta^2+\sin^2{\theta}d\phi^2$. 
197: Imposing steady state conditions and writing the radial component of the
198: 4-velocity ${\rm u}^r$ as ${\rm  u}^r=vy$ where $y=\left(1-2M/r\right)^{1/2}/\left(1-v^2\right)^{1/2}$ and 
199: $v$ 
200: is the outflow velocity of the plasma as measured in an inertial frame
201: at rest in the Schwarzschild coordinate ($t,r,\theta,\phi$)  system allows us to rewrite 
202: the equations above as
203: 
204: \begin{mathletters}
205: \begin{eqnarray}
206: \label{arhob}
207: (1-v^2)a \rho_{\rm b}&=&\left(\rho + P\right) \left[{vv' \over 1-v^2} + {M \over r^2}{1 \over \left(1-2M/r\right)}\right]
208: +P'  \\ \cr
209: \label{rhob}
210: q \rho_{\rm b}&=&vy\left[\rho '+\left(\rho + P\right)\left({v'\over v(1-v^2)}+ 
211: {M \over r^2}{1 \over \left(1-2M/r\right)}+{2\over r}\right)\right] \\ \cr
212: \label{mdot}
213: \dot M&=&4\pi r^2\rho_{\rm b}vy  \\ \cr
214:  vy Y_e ' &=& \left(1-Y_e\right)\left(\lambda_{{\nu}_e{\rm n\rightarrow p}}
215:  +\lambda_{e^+{\rm n\rightarrow p}}\right)-
216: Y_e\left(\lambda_{\bar{\nu}_e{\rm p\rightarrow n}}+\lambda_{e^-{\rm p\rightarrow n}}\right).
217: \end{eqnarray}
218: \end{mathletters}
219: In the above, primes denote differentiation with respect to r, and 
220: $\rho$ and $P$
221: are the total energy density (including rest mass) and pressure, respectively, 
222: of the plasma in the
223: comoving plasma frame. We have also introduced 
224: $a=- \left(\left(\gmet+{\rm \bf u} \otimes {\rm \bf u}\right)\cdot \left(\nablamet  \tmet_{\nu}\right)\right)_r/\rho_{\rm b}$ and 
225: $q= -{\rm \bf u} \cdot \left(\nablamet  \tmet_{\nu}\right)/\rho_{\rm b}$ which are respectively the 
226: momentum and energy deposition into the plasma per unit time per unit baryon
227: rest mass. $\dot M$ is the mass outflow rate. The second law of 
228: thermodynamics and some algebra allows us to produce from these expressions two more first integrals of the flow:
229: \begin{eqnarray}
230: \label{firstintegral}
231: \frac{ds}{dr}&=&\frac{{ m_{\rm b}}q}{T vy}\\
232: \label{secondintegral}
233: \frac{d (\hat H y)}{dr}&=&y \left({(1-v^2)a+q/(vy)}\right).
234: \end{eqnarray}
235: 
236: Here we have introduced the nucleon mass ${ m_{\rm b}}$, the 
237: entropy per baryon $s$ (in units of Boltzmann's constant), and the enthalpy
238: $\hat H=(\rho+P)/\rho_{\rm b}$ of the plasma. It is useful to have explicit formulas for these
239: quantities:
240: 
241: \begin{mathletters}
242: \begin{eqnarray}
243: \label{ent}
244: s=5.21 \rho^{-1}_8 \left({T\over {\rm MeV}}\right)^3  \\ \cr
245: \label{enth}
246: \hat H -1 \equiv H={Ts \over m_{\rm b}}
247: \end{eqnarray}
248: \end{mathletters} 
249: 
250: In this equation $\rho_8$ is the density in units of $10^8{\rm g cm^{-3}}$.
251: The number 5.21 above
252: becomes 1.89 as $T$ decreases below the threshold for pair production.
253:  In writing Eqs. \ref{ent} and \ref{enth} we have
254: assumed that the electrons are non-degenerate and that the entropy of 
255: light particles dominates the entropy of nucleons (i.e. $s>20$). This is
256: relevant for our analysis because we will   
257: be concerned principally with conditions near the sonic point, and we
258: will show below that low entropy conditions near the sonic point cannot
259: lead to ultrarelativistic flow at infinity. The quantity H is related 
260: to the quantity $\eta$, the ratio of energy density to rest mass energy 
261: density often discussed in connection with GRB's, through $\eta \sim (3/4)H+1$.
262: 
263: If the flow is to be relativistic at infinity it must either be supersonic
264: at the neutron star surface or pass through a sonic point somewhere above the neutron star. 
265: In the present 
266: work we focus on the case where the flow is not already supersonic at the neutron 
267: star surface. For the case of nondegenerate electrons and negligible baryon pressure ($s>20$) the 
268: equation governing $v'$ can be written as
269: 
270: \begin{equation}
271: {1\over {v\left(1-v^2\right)}}\left(1+H\right)\left(v^2-v_s^2\right)v'={1\over r}\left[{2\over3} H-\left(1+2H/3\right)\left(M/r\right)
272: {1\over {1-2M/r}}+ ar\left(1-v^2\right)-{qr \over 3{\rm  u}^r}\right]
273: \end{equation}
274: From this equation we see that at the sonic point the following conditions
275: must be satisfied, 
276: 
277: \begin{mathletters}
278: \begin{equation}    
279: v^2=v_s^2={H \over 3}{1 \over H+1} 
280: \end{equation}
281:  and
282: 
283: \begin{equation}
284: \label{sonic3}
285: \left(1-g\left(r\right)\right){2 \over 3} H=g(r)+{qr \over 3{\rm  u}^r}-\left(1-v^2\right)ar.
286: \end{equation}
287: \end{mathletters}
288: In this last equation we have introduced 
289: \begin{equation}
290: g\left(r\right)\equiv M/\left(r-2M\right)
291: \end{equation}
292: and ${v_s}^2=(dp/d\rho)$ is the sound speed with the differentiation is taken
293: at constant entropy.
294: 
295: Eq. (\ref{sonic3}) implies that at the sonic point $qr/{\rm  u}^rH<g(r)$. This is interesting
296: because the change in entropy of the outflowing gas past the sonic point is implicitly  
297: given by 
298: 
299: \begin{equation}
300: \label{ineq}
301: \int{ds\over s}=\int{{ q m_{\rm b} \over Ts}{dr\over {\rm  u}^r}}=\int{qdr\over {\rm  u}^rH}<
302: \left({q r \over
303: {\rm  u}^r H}\right)_{\sonicsub} \sim 1,
304: \end{equation} 
305: 
306: 
307: 
308: 
309: Here the subscript $s$ denotes values at the sonic point and the 
310: second to last inequality holds if $q/{\rm  u}^rH$ decreases more rapidly than $r^{-1}$.
311: When the inequality in Eq. (\ref{ineq}) holds, then, the entropy is approximately constant past
312: the sonic point. Because the hydrodynamic equations imply that 
313: $3T'/T=s'/s-2/r-({{\rm  u}^r} '/{\rm  u}^ r)$, the inequalities in Eq. (\ref{ineq}) also imply that the 
314: temperature cannot increase substantially past the sonic point for accelerating 
315: flow. This implies
316: that the enthalpy per baryon is roughly constant or decreasing past the sonic point. We then arrive
317: at the important result that that the final 
318: Lorentz factor of the wind is not 
319: much larger than the value of $H$ at the sonic point when the above inequality 
320: is satisfied.  (The final Lorentz factor of the wind occurs when when the energy in relativistic 
321: particles $\sim H\rho_{\rm b}$ has been converted to baryon kinetic energy.)
322: This will allow us to make important conclusions regarding 
323: the sonic point conditions needed in order that the final flow be relativistic.
324: 
325: 
326: 
327: We can roughly divide the possible sonic point solutions into two families:
328: those for which the terms involving the heating and momentum deposition terms
329: q and a may be neglected (Family I),
330: and those for which these terms are important in determining
331: the conditions at the sonic point (Family II).  As we will show
332: that the neutrino energy deposition terms satisfy the inequalities in 
333: Eq. (\ref{ineq}) when
334: H and ${\rm u}^r$ increase with radius, only family II solutions can
335: support steady state ultrarelativistic flow. An exception to this occurs when
336: the sonic radius is very near 3M.
337: Previous discussions of 
338: steady state winds from neutron stars have focused on family I solutions.
339: For these solutions $H\sim (3/2)M/(r-3M)$ at the sonic point, 
340: and hence the sonic point must be near 3M
341: in order to produce ultrarelativistic flow in the absence of heating terms
342: \citep{Paczynski}. For neutrino driven wind studies relevant for the r-process
343: the heating terms are generally not important near the sonic point simply
344: because those flows are subrelativistic or mildly relativistic,
345:  so that the sonic point occurs far 
346: from the neutron star where neutrino heating has fallen off.
347: 
348: 
349: For Family II solutions (those where heating determines the sonic point conditions
350: and drives the flow relativistic) some general observations can be made. If we 
351: assume that $H(r_{s})\gg H(r_0)$ and $s(r_{\sonicsub})\gg s(r_0)$ (where $r_{\sonicsub}$ is the schwarzschild
352: radial coordinate of the sonic point and $r_0$ is the initial radial coordinate of the flow, usually
353: taken to be the neutron star radius here) then the sonic
354: point condition and Eq. \ref{secondintegral} imply
355: 
356: \begin{equation}
357: \label{yeff}
358: \int^{r_{\sonicsub}}_{r_0}{\frac{yq}{{\rm u}^r}}\equiv y_{eff}\int^{r_{\sonicsub}}_{r_0}{\frac{q}{{\rm u}^r}}
359: \sim \left(\frac{yqr}{2{\rm u}^r(1-g)}\right)_{\sonicsub}.
360: \end{equation}
361: In Eq. \ref{yeff} we have defined $y_{eff}$, we have assumed that $(1-g)H\gg 1$ at the 
362: sonic point, and we have neglected the momentum deposition term $a$. 
363: Eq. \ref{yeff} 
364: serves to specify the position of the sonic radius given a heating term $q$ 
365: and a velocity profile. If $q/{\rm u}^r \propto r^{-\beta}$ (which occurs in steady state
366: winds if the heating rate per unit volume drops as $r^{-\beta-2}$) the sonic point
367: occurs at
368: 
369: \begin{equation}
370: \label{rsonicarbitrary}
371: \frac{r_{\sonicsub}}{r_0} \sim \left(\frac{2(\beta-1)y_{\sonicsub}}
372: {y_{eff}(1-g_{\sonicsub})}+1\right)^{1/(\beta-1)},
373: \end{equation}
374: which implies that the sonic point occurs near $r_0$ for rapidly
375: dropping heating terms. Again the momentum deposition term $a$ has
376: been neglected in this approximation.  Another interesting
377: property of these winds is that the temperature does not vary greatly 
378: within the sonic radius if $q$ is a decreasing function of radius. This is 
379: seen by noting that Eqs. \ref{firstintegral} and \ref{yeff} imply
380: 
381: \begin{equation}
382: \label{teff}
383: T_{s}^{-1}\sim\left(\frac{y_{s}}{y_{eff}}\right) \frac{\int^{r_{\sonicsub}}_{r_0} {q}/{T {\rm u}^r}}  
384:  {\int^{r_{\sonicsub}}_{r_0} {q}/{{\rm u}^r}},
385: \end{equation}
386: which allows us to to define an effective temperature within the sonic point,
387: $T_{eff}=T_{s} y_{s}/y_{eff}$.
388: 
389: The above equations allow us to extract the properties of
390: ultrarelativistic steady state winds for arbitrary heating
391: functions. We now turn to the particular case where neutrinos from the
392: central compact object are responsible for the heating.
393: 
394: \section{Neutrino driven ultra-relativistic winds}
395: 
396: Perhaps the first question raised by a study of steady state
397: neutrino-heated wind solutions is whether or not these
398: winds are physically realizable for conventional compact objects. 
399: There are two parts to this question:
400: i) whether or not steady state can be achieved; and ii) whether or not the 
401: flow can be relativistic. We defer consideration of the second part until
402: later, when we discuss mass ablation from the neutron star surface. 
403: While a hydrodynamic simulation is needed to determine the timescale 
404: to achieve steady state in these winds, a rough estimate is that  
405: the equilibration 
406: timescale is of order the sound crossing time between the sonic radius 
407: and the neutron star surface, $\sim 10^{-4}{\rm s}$. This timescale
408: is short compared to the neutrino diffusion timescale, which is of the 
409: order of seconds. The neutrino diffusion timescale governs the 
410: evolutionary timescale of the neutron star unless it undergoes a phase
411: transition or becomes dynamically unstable. It is known that for most cold
412: neutron star equations of state  general relativistic instability 
413: sets in for $r\sim 3M$.
414:  Below this radius the star becomes dynamically
415: unstable and collapses on a dynamic timescale which is 
416: comparable to the equilibration timescale
417: given above. For these small radius stars, then, it may not be sensible to 
418: put much stock in a 
419: \question{extra ``a'' deleted (CYC)}
420: %a 
421: steady state wind solution.  
422: 
423: 
424: 
425: 
426: 
427: 
428: 
429: 
430: Before describing in some detail properties of the allowed steady 
431: state solutions, we must first examine the neutrino energy deposition terms.
432: To parametrize the neutrinos we make the usual assumption that the neutrinos
433: are emitted from a neutrinosphere with some radius $r_0$ (in this work we
434: do not make a distinction between neutrinosphere radius and neutron star 
435: radius) and 
436: are characterized at the neutrinosphere by a Fermi-Dirac black body 
437: energy distribution
438: with temperature $T$ and zero chemical potential,
439: 
440: \begin{equation}
441: f_{\nu}={{1}\over {e^{-E/T}+1}}.
442: \end{equation}
443: 
444: This is a crude approximation to the actual expected energy spectra but 
445: suffices to mock-up the energy deposition physics (see {\cite{Cardall}}).
446: We further make the free streaming approximation, i.e. we neglect 
447: the fact that the neutrinos suffer a small depletion with increasing 
448: radius owing to interactions with the outflowing plasma.
449: The facts that real neutron stars are characterized by some small but finite
450: decoupling region over which the neutrino distribution function continues to 
451: evolve and that neutrinos from real neutron stars are in general non-thermal
452: or better characterized by degenerate spectra are not important for 
453: our arguments. For a discussion of when it is apropriate to treat the 
454: neutrino as sharply decoupling from a neutrinosphere see \cite{Salmonson}.
455: 
456: Evaluating $\nablamet  T_{\nu}$ for various neutrino plasma 
457: interactions is most easily accomplished by noting that Liouville's
458: theorem implies that in an inertial frame at rest at some Schwarzschild 
459: coordinate r the neutrino distribution function is characterized
460: by the red-shifted temperature 
461: \begin{equation}
462: T_{\nu}\left(r\right)=T_{\nu}(r_0)\left({1-2M/r_0\over 1-2M/r}\right)^{1/2}\equiv h T_{\nu}(r_0),
463: \end{equation}  (see \cite{fullerqian96}) which also
464: serves to define the redshift factor $h$, and a maximum
465: angle of deviation from the radial direction given by 
466: \begin{equation}
467: \cos \theta_{max}\equiv
468:  x=\left(1-\left({r_0\over r}\right)^2{1-2M/r \over 1-2M/r_0}\right)^{1/2}.
469: \end{equation}
470: 
471:  The mass $M$ of the
472: neutron star appears in the expresion for $x$ because of the bending of 
473: null trajectories in the Schwarzschild geometry as has been discussed by 
474: \cite{Cardall} and \cite{Salmonson}. 
475: 
476: 
477: In an inertial frame at rest in Schwarzschild coordinates, the 
478: contribution to $\nablamet  T_{\nu} \equiv 
479: \left(Q^{\hat 0}_{\nu \bar \nu},Q^{\hat r}_{\nu \bar \nu}\right)$
480: from neutrino-antineutrino annihilation is 
481:  
482: 
483: \begin{eqnarray}
484: Q^{\hat 0}_{\nu \bar \nu}&=& {G_{\rm F}^2 \over 9 (2\pi)^5}C T_{\nu}\left(r\right)^4 T_{\bar \nu}\left(r\right)^4F_4\left(0\right)F_3\left(0\right)\bigl(T_{\nu}\left(r\right)+T_{\bar \nu}\left(r\right)\bigr)
485: \Phi \left(x\right) \cr
486: Q^{\hat r}_{\nu \bar \nu}&=& {G_{\rm F}^2 \over 9 (2\pi)^5}C T_{\nu}\left(r\right)^4 T_{\bar \nu}\left(r\right)^4F_4\left(0\right)F_3\left(0\right)\bigl(T_{\nu}\left(r\right)+T_{\bar \nu}\left(r\right)\bigr)
487:  \Psi \left(x\right)/4
488: \end{eqnarray}
489: 
490: In the above $\Phi(x)=(x-1)^4(x^2+4x+5)$, 
491: $\Psi(x)=(x-1)^4(x+1)(3x^2+9x+8)$, $G_{\rm F}$ is the Fermi constant, and hats
492: are used to denote quantities in an inertial frame at rest in the Schwarzschild
493: coordinate sytem. The quantities $F_3(0)=7\pi^4/120$ and $F_4(0)\approx 23.3$ are the  
494: Fermi-integrals of argument zero. The quantity $C$ depends on the neutrino species. For
495: $\nu_e \bar{\nu}_e$ annihilation $C=(1+2 \sin^2 \theta_w)^2+4\sin^4\theta_w$ 
496: where $\theta_w$ is the Weinberg angle 
497: ($\sin^2\theta_w \approx 0.23$), while for $\nu_{\mu} \bar{\nu}_{\mu}$ and  
498: $\nu_{\tau} \bar{\nu}_{\tau}$ 
499: annihilation $C=(1-2 \sin^2 \theta_w)^2+4\sin^4\theta_w$.
500: For simplicity in what follows we will assume that the luminosity of 
501: a single neutrino flavor gives all $\nu \bar{\nu}$ annihilation heating
502: and that for this flavor
503: the neutrino and antineutrino temperatures are equal, 
504: $T_{\nu}=T_{\bar{\nu}}$. We will refer to this single 
505: neutrino flavor as $\mu$ or $\tau$. This is merely a calculational
506: device designed to reproduce the total heating rate from all neutrino flavors
507: and adjusted accordingly for sensitivity studies.
508: 
509: Making a transformation back to Schwarzschild coordinates,
510: we find that 
511: $\rho_{\rm b} q_{\nu \bar{\nu}}=(1-v^2)^{-1/2}Q^{\hat 0}_{\nu \bar \nu}
512: (1-v\Psi/4\Phi)$ and 
513: $\rho_{\rm b} a_{\nu \bar{\nu}}=(1-v^2)^{-1}(1-2M/r)^{-1/2}Q^{\hat 0}_{\nu \bar \nu}
514: (\Psi/4\Phi-v)$. The terms entering the right hand side 
515: of the sonic point condition are then 
516: 
517: 
518: 
519: \begin{eqnarray}
520: {r q_{\nu \bar{\nu}} \over 3{\rm u}^r}&=&{10^{-9} r_6 \over(1-v^2)^{1/2} {\rm u}^r} 
521:  {\left(T_{\nu}(r)/{\rm MeV}\right)^9\over \left(T/{\rm MeV}\right)^4} \Phi H \left(1-{v\Psi 
522: \over 4\Phi}\right)\cr
523: a_{\nu \bar{\nu}} \left(1-v^2\right)r&=& 
524: {3 \cdot 10^{-8} r_6 \over (1-2M/r)^{1/2}}  
525:  {\left({T_{\nu}(r)/{\rm MeV}}\right)^9 \over \left(T/{\rm MeV}\right)^4} \Phi H 
526: \left({\Psi \over 4\Phi}-v\right),
527: \end{eqnarray}
528: where $r_6=r/10^6 {\rm cm}$.
529: 
530: 
531: Evaluation of the contribution to $(\nablamet  T_{\nu})$ 
532: from neutrino-plasma interactions is complicated by the fact that an inertial 
533: observer comoving with the outflowing plasma sees a neutrino distribution
534: function whose temperature varies with direction as $T_{\nu}(r)
535: (1-v \cos \theta)/(1-v^2)^{1/2}$. The effect of this velocity (and direction)
536: dependent redshift is to make the neutrino-plasma heating rates small as
537: the outflow velocity becomes large.
538: We will show that neutrino-antineutrino annihilation
539: is the dominant heating term for the conditions of interest here. It 
540: suffices for us to note that in the limit where $v=0$, the contributions
541: to  $(\nablamet  \tmet_{\nu})$ from neutrino-electron scattering and
542: neurino capture on baryons are, respectively, 
543: 
544: \begin{equation}q_{\nu e} \sim 10 \,q_{\nu \bar{\nu}}  {T^4 \over T_{\nu}^4} {1 \over h^4} 
545: {1-x \over \Phi} 
546: \end{equation}
547: 
548: 
549: if $T<T_{\nu}$, and 
550: 
551: \begin{equation}
552: q_{\nu b} \sim 10^3 q_{\nu \bar{\nu}}  
553: {T^4 \left(T_{\bar{\nu}_e}^6 Y_e +T_{{\nu}_e}^6 \left(1-Y_e\right)\right) \over m_{\rm b} h^3 H T_{\nu}^9}
554: {1-x \over \Phi}.   
555: \end{equation}
556: 
557: Note that all of these heating terms satisfy the inequality in 
558: Eq. (\ref{ineq}) if 
559: $H$ increases with radius. Neutrino-Antineutrino annihilation 
560: satisfies Eq. (\ref{ineq}) because $q_{\nu \bar{\nu}}/H \sim \Phi r^2/(\dot M H)$ which falls 
561: more rapidly than $r^{-6}$. Neutrino electron scattering satisfies 
562: Eq. (\ref{ineq})
563: because $q_{\nu e}/H\sim (1-x) \sim r^{-2}$ Neutrino capture 
564: on baryons also satisfies this condition 
565: because $q_{\nu b}/H \sim (1-x)/H$. These estimates for the scaling
566: of the heating rates with radius neglect the dependence of the heating 
567: on the redshift factor $h$. This implies that the heating rates decrease even
568: more rapidly with increasing radius if the wind is accelerating. 
569: This means that if 
570: the final flow is to be ultra-relativistic, we must have $H>1$ at 
571: the sonic point, 
572: i.e., either $r_{\sonicsub}\sim 3M$, or, 
573: 
574: \begin{mathletters}
575: \begin{equation}
576: \label{nunubardom}
577: T\sim 1\, {\rm MeV}\, \left({r_6 \Phi\left(1-g\right) \left( {T_{\nu}(r)/{\rm 10 MeV}}\right)^9 \over {\rm u}^r}\right)^{1/4}
578: \end{equation}
579: 
580: 
581: for $\nu\bar{\nu}$ annihilation dominating the heating, 
582: or
583: 
584: \begin{equation}
585: \label{nuedom}
586: T_{\nu} > 40 \, {\rm MeV} \, h^{-1} \left({1-g \over r_6(1-x)}\right)^{1/5}
587: \end{equation}
588: 
589: if electron-neutrino scattering dominates the heating rate, or,
590: 
591: \begin{equation}
592: \label{nubdom}
593: T_{\nu} > 30 \, {\rm MeV} \, h^{-1}  \left({1-g \over r_6(1-x)}\right)^{1/6}
594: \end{equation}
595: \end{mathletters}
596: if neutrino capture on baryons dominates. In Eqs. (\ref{nunubardom},\ref{nuedom},
597: and \ref{nubdom}) the vvalues of all quantities 
598: are taken at the sonic point.
599: 
600: Unless the sonic radius occurs near $3M$ ($g\sim1$), the neutrino 
601: temperatures needed to drive steady state ultrarelativistic flow 
602: imply extraordinary neutrino luminosities. (For example, consider the cases inherent in Eqs.
603: \ref{nuedom} and \ref{nubdom}, 
604: $L\approx 10^{54} r_6^2 (T_{\nu}/20{\rm MeV})^4 {\rm erg/s}$.) In what follows
605: we therefore restrict ourselves to the case where $\nu \bar{\nu}$
606: annihilation dominates the heating at the sonic point and where only 
607: smaller, more realistic, neutrino luminosities
608: are required.
609: 
610: \subsection{Steady state flow when $\nu\bar{\nu}$ annihilation dominates at the sonic
611: point}
612: 
613: Having determined that for modest neutrino luminosities 
614: ($L<10^{54} \rm{ergs \,s^{-1}}$) ultrarelativistic steady state flow 
615: only occurs if $\nu\bar{\nu}$ annihilation heating 
616: dominates at the sonic point, we 
617: now turn to a discussion of the behavior of these solutions.
618: 
619: For $\nu\bar{\nu}$ annihilation the heating scales as  
620: $q/{\rm u}^r\propto \Phi r^2 h^9 +O(v)$
621:  and the sonic radius is readily estimated. This is shown in 
622: Fig. 1. From this figure we see that 
623: the sonic radius is generically near $r_0$. Because $q_{\nu \bar{\nu}}/{\rm u}^r$ decreases
624: rapidly with radius we may approximate $y_{eff}\approx y(r_0)$ and $T_{eff}\approx
625: T(r_0)\approx (y_{\sonicsub}/y_{eff})T_{\sonicsub}$, which implies 
626: that the temperature is ``cold'' within the sonic radius so long as
627: thermal 
628: neutrino energy losses are unimportant. Here ``cold'' means smaller than the neutrino 
629: energy distribution temperatures.
630: 
631: The mass ablation rate for $\nu \bar{\nu}$ dominated solutions is 
632: \begin{equation}
633: \dot M \approx 10^{-4} \left({M_{\odot} s^{-1}} \right) 
634: \left(\frac{r_6^3 \Phi (1-g) (T_{\nu}(r)/{\rm 10 MeV})^9}{H}\right)_{\sonicsub}.
635: \end{equation}
636: If $\dot M$ is too large the flow cannot be relativistic. This is the
637: ``baryon loading problem'' discussed in connection with GRB's. The relation
638: between $\dot M$ and the neutrino heating rates depends in detail 
639: on the structure of the compact object atmosphere below, above, and through the neutrinosphere.
640: As argued by \cite{woos1}, obtaining a sufficiently low $\dot M$ may be particularly difficult 
641: for $\nu \bar{\nu}$ energy deposition because this heating rate drops 
642: so rapidly with radius. This requires the density scale height above the
643: neutrinosphere to be small. For type II supernovae, calculations show that 
644: $\dot M \approx \dot E_{\nu}/(GM/r_0)$, where $\dot E_{\nu}$ is the 
645: integrated neutrino energy deposition rate \citep{wilsonandmathews}. 
646: This implies that most of the neutrino energy deposition 
647: goes into extracting baryons from the gravitational potential well 
648: of the neutron star.
649:   For supernovae, then, ultra-relativistic
650: winds are not expected unless the early-time density scale height 
651: is smaller than expected or the late time neutrino luminosities are 
652: larger than expected. \cite{woos1} 
653: have argued that the baryon loading problem
654: cannot be overcome for mass ejection from strange stars resulting
655: from the phase transition of a cooling or accreting neutron star while 
656: \cite{SWM} have argued that neutrino heating during the compression of
657: an inspiralling neutron star might lead to ultrarelativistic flow. 
658: The question must be considered for each proposed GRB site.
659: 
660: 
661: 
662: An estimate of $T_{eff}$ allows a simple determination 
663: of the influence of neutrino
664: capture on the electron fraction in the outflow. Noting that 
665: the number of neutrino captures per baryon $n_c$ is 
666: approximately given by $dn_c=dr \, q_{\nu {\rm b}} /(T_{\nu} {\rm u}^r)$, where
667: $\nu=\nu_e$ or $\bar{\nu}_e$ for neutrino capture on neutrons or protons, gives
668: the total number
669: of neutrino captures per baryon between the neutron star surface and the
670: sonic point 
671: \begin{equation}n_c\approx 
672: {10^{-6}r_6(h^5(1-g_{})(1-x_{}))_{\sonicsub}(T_{\bar{\nu_e}}^5+T_{\nu_e}^5)} 
673: \ln\left({s_{\sonicsub}/s_0}\right),
674: \end{equation}
675: where $s_{\sonicsub}$ and $s_0$ are the entropy at the sonic point and
676: neutron star surface, respectively. With this expression we can solve for 
677: the final electron fraction in the fireball, neglecting $e^+/e^-$ capture
678: because of the low plasma temperature. Representive solutions are shown
679: in Fig.2. This figure serves to illustrate the basic features of how the 
680: final electron fraction 
681: is set: i) For low $T_{\nu_e},T_{\bar{\nu}_e}$, $Y_e$ 
682: simply remains what it was
683: in the neutron star. ii) As  $T_{\nu_e},T_{\bar{\nu}_e}$ increase, the final 
684: $Y_e$ tends
685: toward equilibrium with respect to neutrino capture (i.e. $Y_e\rightarrow
686: T_{\nu_e}^5/(T_{\nu_e}^5+T_{\bar{\nu_e}}^5)$), while still retaining
687: some memory of the value of the electron fraction in the neutron star
688: crust. And iii) For  $T_{\nu_e}$ or $T_{\bar{\nu}_e}$ $>10{\rm MeV}$, 
689: $Y_e$ is set
690: ultimately by the competition between $\nu_e$ and $\bar{\nu_e}$ captures.
691:  If the neutrino temperatures
692: are high and neutrinos set the electron fraction in the outflow, 
693: the final electron fraction is expected to be low because the 
694: neutrinos emitted from a neutron rich star generically satisfy the 
695: temperature hierarchy $T_{\nu_e}\le T_{\bar{\nu_e}}$. Because neutron star crusts
696: typically have $Y_e<0.1$, the electron fraction in the outflow is also 
697: expected to be low even when neutrino capture does not set the final 
698: electron fraction.
699: 
700: 
701: 
702: A numerical example of a steady state neutrino heated wind leading to 
703: ultrarelativistic flow is shown in Fig 3. For this example we have taken
704: $M=1.4M_{\odot}$, $r_0=4M$, $T_{\nu_{\mu}}=T_{\bar{\nu}_{\mu}}=10\,{\rm{MeV}}$, and 
705: $\eta=4/3 H=100$ at the sonic point. As the wind behavior shown in figure 1 is quite
706: independent of $H$ (except of course for the baryon density which scales trivially with $H$)
707: as long as $H>$ a few, we have only shown a single numerical example. 
708: The sonic radius can be chosen to 
709: match a particular entropy at the neutron star surface. Increasing $r_{\sonicsub}$ increases the
710: entropy at the surface. The value of the surface entropy is very sensitive to changes in $r_{\sonicsub}$.
711: Numerically, for $r_{\sonicsub}$ too large, the scale height for velocity changes becomes very 
712: small and the integration fails. Physically, the sonic radius is set by the requirement that 
713: the energy flow $\dot E$ at the sonic radius equals the net neutrino energy deposition rate
714: interior to the sonic radius. Our plot only extends out to 
715: $r \approx 3r_0$. Past this 
716: radius neutrino energy deposition is unimportant and the flow satisfies the 
717: simple scaling laws given in \cite{piran2}. The final Lorentz factor of the wind is approximately 150.
718:  
719: 
720: \section{Conclusions}
721: 
722: We have extended the study of steady state winds from compact 
723: objects to include the case where the wind is driven by a heating
724: term and the wind velocity at infinity is ultra-relativistic. We find that 
725: for heating
726: rates per unit volume which drop more rapidly than $r^{-3}$, the sonic point 
727: condition is dominated by the heating and momentum deposition terms, 
728: the sonic radius is near the compact object radius, and the temperature
729: within the sonic radius is roughly constant. 
730: 
731: Particular attention has been paid to the case where the wind is 
732: driven principally by neutrino energy deposition. For these winds 
733: neutrino anti-neutrino annihilation dominates the heating unless the
734: neutrino luminosity is very large ($\sim 10^{54} {\rm erg s^{-1}}$). Interestingly,
735: then, for a given neutrino luminosity there are an infinite family of 
736: steady state supersonic solutions. On one branch of this family is the usual
737: subrelativistic solution discussed for supernovae, and on the other their
738: exists
739: a continuum of ultra-relativistic solutions dominated by neutrino 
740: heating at the sonic point. Of course, the low mass ablation rate needed
741: for the ultra-relativistic solutions may not be compatible with the
742: structure of the atmosphere of the compact object near the neutrinosphere.
743: This needs to be investigated for each compact object.
744: 
745: The way in which the electron fraction is 
746: set in the steady state winds we have discussed 
747: depends sensitively on the neutrino luminosity, with the number of 
748: neutrino captures per baryon varying as the fifth power of 
749: the neutrino temperature and depending only logarithmically on the 
750: final Lorentz factor of the outflow. For high
751: $\nu_{e}$ or $\bar{\nu}_e$ temperatures the final electron fraction is 
752: set by neutrino capture. Otherwise $Y_e$ remembers its value in the
753: neutron star crust.  At least for a simple case, then,
754: the electron fraction in the outflow is a diagnostic of
755: conditions within the central engine.
756: This is interesting because {\it a priori} one might
757: guess that the electron fraction comes to equilibrium at $Y_e=1/2$, or is 
758: always dominated by neutrino capture, as is the case for 
759: supernovae. It is also interesting because recently \citep{bahcall} it 
760: has been argued that neutrinos arising from pion decay in GRB fireballs
761: may be detectable. This neutrino signal depends, along with the 
762: other parameters characterizing the fireball, on the electron fraction in 
763: the fireball. If more realistic GRB central engine models also 
764: leave their finger prints in the electron fraction in the outflow,
765:  such a signal
766: might be used to distinguish between central engine models.  
767: 
768: \acknowledgements 
769: 
770: This work was partially supported by NSF Grant PHY98-00980 at UCSD, and
771: an  IGPP mini-grant at UCSD. We are indebted to C.Y. Cardall for the 
772: use of an unpublished paper on the relativistic formulation of the 
773: wind equations.
774: 
775: 
776: 
777: 
778: 
779: 
780: 
781: 
782: 
783: 
784: 
785: 
786: 
787: 
788: \begin{thebibliography}{}
789: 
790: 
791: \bibitem[Bahcall \& Meszaros(2000)]{bahcall}
792: Bahcall, J.N. \& Meszaros, P. 2000 \prl, 85, 1362
793: 
794: \bibitem[Cardall \& Fuller(1997)]{Cardall}
795: Cardall, C.Y. \& Fuller, G.~M. 1997 \apjl, 486, L111
796: 
797: \bibitem[Duncan, Shapiro, \& Wassermann(1986)]{DSW}
798: Duncan, R.C., Shapiro, S.L. \& Wasserman, I. 1986
799: \apj, 309, 141
800: 
801: 
802: \bibitem[Fryer \& Woosley(1998)]{woos1}
803: Fryer, C.D., \& Woosley, S.E. 1998, \apj, 501, 780
804: 
805: \bibitem[Fuller, Pruet, \& Abazajian(2000)]{FPA}
806: Fuller, G.~M.,
807: Pruet, J. \& Abazajian, K. 2000, \prl, 85, 2673
808: 
809: \bibitem[Fuller \& Qian(1996)]{fullerqian96}
810: Fuller, G.~M., \& Qian, Y.-Z. 1996, Nuclear Phys. A, 606 167
811: 
812: \bibitem[Meyer {\it et al.} (1992)]{meyer92}
813: Meyer, B.S., Mathews, G.J., Howard, W.M., Woosley, S.E., \& Hoffman, R.D.
814:  1992, \apj 399, 656.
815: 
816: 
817: \bibitem[Otsuki {\it et al.} (2000)]{japanese1}
818: Otsuki, K., Tagoshi, H., Kajino, T., \& Wanajo, S. 2000
819: \apj 533, 424                       
820: 
821: \bibitem[Paczy\'nski (1990)]{Paczynski}
822: Paczy\'nski, B. 1990, \apj, 363, 218
823: 
824: 
825: 
826: \bibitem[Piran, Shemi, \& Narayan (1993)]{piran2}
827: Piran, T., Shemi, A., Narayan, R. 1993, \mnras\ {\bf 63}, 861
828: 
829: \bibitem[Qian {\it et al.} (1993)]{Qetal93}
830: Qian, Y.-Z., Fuller, G.M., Mathews, G.J., Mayle, R.W., \& Woosley, S.E. 1993,
831: \prl 71, 1965.
832: 
833: 
834: 
835: \bibitem[Qian  \& Woosley(1996)]{QW}
836: Qian, Y.-Z. \& Woosley, S.E. 1996, \apj, 471, 331
837: 
838: 
839: \bibitem[Salmonson \& Wilson (1999)]{Salmonson}
840: Salmonson, J.D. \& Wilson, J.R. 1999, \apj, 517, 859
841: 
842: \bibitem[Salmonson, Wilson, \& Mathews (1999)]{SWM}
843: Salmonson, J.D., Wilson, J.R., \& Mathews, G.J. to appear in 
844: \apj, astro-ph/0002312
845: 
846: \bibitem[Shemi \& Piran (1990)]{Piran} Shemi, A. \& Piran, T. 1990,
847: \apj, 365, L55
848: 
849: \bibitem[Takahashi, Witti, \& Janka (1994)]{takahashi94}
850: Takahashi, K., Witti, J., \& Janka, H.T. 1994, Astronomy and Astrophysics
851: 286, 857.
852: 
853: \bibitem[Woosley et. al. (1994)]{wilsonandmathews}
854: Woosley, S.~E., Wilson, J.~R., Mathews, G.~J., Hoffman, R.~D., 
855: \& Meyer, B.~S. 1994, \apj, 433, 229. 
856: 
857: \end{thebibliography}
858: 
859: 
860: 
861: 
862: 
863: \begin{figure*}
864: \epsscale{1.}
865: \plotone{figure1.eps}
866: \caption{\small Approximate position of the the sonic point as a function of $r_0/M$ for 
867: neutrino anti-neutrino annihilation dominating the heating term within 
868: the sonic radius. The momentum depostion term $a$ has been neglected and it 
869: is assumed that at the sonic point $H(1-g)\gg 1$.
870: }
871: \end{figure*}
872: 
873: 
874: \begin{figure*}
875: \epsscale{1.}
876: \plotone{figure2.eps}
877: \caption{\small Final electron fraction in the fireball as a function of 
878: $T_{\bar \nu_e}$ and for different ratios  $T_{\bar \nu_e}/T_{\nu_e}$ (labeled
879: next to the curves). The initial electron fraction is taken to be 0.01, and 
880: the entropy of the plasma is assumed to increase by a factor of $10^5$ as 
881: the plasma travels from the neutron star surface to the sonic point.
882: }
883: \end{figure*}
884: 
885: 
886: 
887: 
888: \begin{figure*}
889: \epsscale{1.}
890: \plotone{figure3.eps}
891: \caption{\small 
892: Evolution of a neutrino heated wind solution as described in the text.
893:  The solid curve is for 
894: temperature (in MeV), the short dashed curve for $q_{\nu \bar{\nu}}$,
895: the long dashed curve for $U^r$, and the dot-dashed curve for
896: $\dot E=\dot M \hat H y$. The asterices correspond to 
897: the position of the sonic point.}
898: \end{figure*}
899: 
900: 
901: \end{document}
902: 
903: 
904: 
905: 
906: 
907: 
908: 
909: 
910: 
911: 
912: